Premium
Genome‐wide study of the human lipidome and links to Alzheimer’s disease risk
Author(s) -
Meikle Peter J.,
Giles Corey,
Cadby Gemma,
Huynh Kevin,
Mellett Natalie A.,
Olshansky Gavriel,
Smith Alexander,
Nguyen Anh,
Chatterjee Pratishtha,
Martins Ian,
Laws Simon M.,
Bush Ashley I.,
Rowe Christopher C.,
Villemagne Victor L.L.,
Ames David,
Masters Colin L.,
Arnold Matthias,
Kastenmüller Gabi,
Nho Kwangsik,
Saykin Andrew J.,
Baillie Rebecca,
Han Xianlin,
Inouye Michael,
Martins Ralph N.,
KaddurahDaouk Rima F.,
Moses Eric
Publication year - 2020
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1002/alz.045600
Subject(s) - lipidome , genome wide association study , lipidomics , lipid metabolism , biology , disease , genetic association , genetics , bioinformatics , single nucleotide polymorphism , medicine , gene , genotype , endocrinology
Background Dysregulation of lipid metabolism is as an important – and modifiable – risk factor for the initiation and progression of Alzheimer’s disease (AD). Although lipid metabolism is established as crucial to numerous biological processes, the genetic factors that influence inter‐individual variation are still not well understood. To address this issue, we apply an integrative approach to link genetic variants with altered lipid metabolism and AD. Method Lipidomic profiling was performed using liquid chromatography coupled electrospray‐ionization tandem‐mass spectrometry on 4,492 genotyped individuals from the Busselton Family Health Study. We performed genome‐wide association analysis of 596 lipid species and 33 lipid classes using linear‐mixed models, correcting for age, sex, their interactions, 10 genomic principal components and an empirical genetic relatedness matrix. To account for lipoprotein mediated associations, the analysis was repeated with HDL‐C, triglycerides and total cholesterol as covariates. Additionally, collider bias was avoided by conditioning using multi‐trait‐based conditional and joint analysis. Validation of genome‐wide significant associations is supported by replication in the Australian Imaging, Biomarker & Lifestyle Study of Ageing (n = 1,112) and Alzheimer’s Disease Neuroimaging Initiative (n = 757) cohorts. Result Over 70,000 genome‐wide significant (p<5e‐08) associations were identified, with 451 lipid species having at least one significant association. Approximately 70% of the observed associations are independent of lipoprotein measures. We observed associations (p<1.0E‐6) between lipid species and seven loci containing genes previously identified as risk factors for AD, including ApoE, ABCA7 and SLC24A4. We further identify 1252 variants (92 independent gene regions) that associated (p<5.0E‐8) with the 71 lipid species we identified as significantly associated with incident AD in a combined meta‐analysis of the AIBL and ADNI cohorts. Conclusion By linking lipid species to AD, through genetic associations, we highlight potential therapeutic targets for monitoring, prevention and treatment.