A blood cholesterol polygenic score in two independent AD cohorts: Genetics/genetic factors of Alzheimer's disease. (7th December 2020)

• Record Type: Journal Article
• Authors: Nilsson, Nathalie I.V.; Picard, Cynthia; Labonte, Anne; Poirier, Judes
• Is Part Of: Alzheimer's & dementia. Volume 16(2020)Supplement 3
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-12-07
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.044985
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: While blood total cholesterol (TC) levels is a recognized factor contributing to Alzheimer's disease (Alz. Ass., 2017), its role is debated (Wood et al., 2014). E.g. several studies have shown that midlife hypercholesterolemia is associated with increased risk of developing AD (Reiman et al., 2010; Solomon et al., 2009; Toro et al., 2014), but others have shown no association (Li et al., 2005; Tan et al., 2003). High TC has also been associated with increased amyloid load in the hippocampus (Pappolla et al., 2003) and hypometabolism in brain regions affected by AD (Reiman et al., 2010). Studies with statins have shown similar discrepancies with retrospective studies indicating a protective effect but have had no consistent positive effects in randomized controlled trials (Shobab et al., 2005). Thus, in an attempt to address these discrepancies, we constructed a polygenic score (TC‐pgs) capturing some of the variance in peripheral TC levels and evaluated it for associations with AD risk and biomarkers. Method: A weighted TC‐pgs was constructed using summary data from Willer et al., (2013; http://lipidgenetics.org/ ) in two AD cohorts; PREVENT‐AD (https://preventad.loris.ca/ ) and ROSMAP (https://www.radc.rush.edu/home.htm ). The TC‐pgs was optimized for correlation with TC levels in the PREVENT‐AD cohort, and then evaluated for correlations with AD risk and CSF biomarkers using both cohorts. Results: We found that by stratifying for statin use and sex we more than doubled the amount of variance explained by the score (from ∼7.5% to ∼17.5%) specifically in statin free females. Furthermore, the TC‐pgs improved the prediction of hypercholesterolemia (AUC 0.805 vs 0.646, p = 0.0016) in the same group. The TC‐pgs were evaluated for associations with CSF Aβ42, p‐tau and tau in PREVENT‐AD and for clinically and pathologically defined AD in ROSMAP. We did not find any significant associations, but for a trend towards a positive association between TC‐pgs and CSF p‐tau levels (p = 0.09). Conclusion: Our TC‐pgs did improve prediction of hypercholesterolemia but failed to correlate with AD or AD biomarkers. Thus, the findings do not support that an increased cumulative genetic risk of hypercholesterolemia influence the risk of AD.