The primary risk factors for Alzheimers disease (AD) are age >65 years and the presence of the APOE4 genetic variation. Recent genetic work has identified several additional genetic risk factors with effects that pale in comparison to APOE4 effects. However, we reasoned that these other genetic risk factors are essentially identifying bottlenecks in pathways that modulate AD risk. If we can identify how these genetic changes alter these bottlenecks, these risk factors may lead us to drugs that act similarly but with greater effect, altering AD risk dramatically. This premise is the basis of this proposal. In preliminary data, we found that AD-protective variations in genes called CLU and ABCA7 act to increase expression of these genes. CLU produces a protein termed apoJ that is similar to apoE in that both are cholesterol carriers in the brain. Additionally, ABCA7 produces a protein that loads cholesterol onto apoE and apoJ. Hence, we hypothesize that increased cholesterol carrier proteins in the brain, especially those that are well-loaded with cholesterol, may reduce AD risk. Since apoE and apoJ are both known to bind amyloid-ß peptide, a likely causal factor in AD, we further hypothesize that the protective actions of CLU and ABCA7 are mediated by this Aß-binding ability. By screening the literature, we found that the common anticonvulsant drug, valproic acid (VPA), induces CLU robustly, much more than the genetic variation. Additionally, we found that VPA also robustly induces ABCA7 expression. Hence, we propose here to (Aim 1) apply our novel ELISA to quantify apoE/Aß and apoJ/Aß complexes in human CSF and characterize the variables that influence levels of these complexes and (Aim 2) test the extent that VPA increases the levels of these "AD-protective" complexes. Overall, we seek to translate AD genetics into AD therapeutics.