Identification of Genetic Variants and Serum Metabolites Associated with Blood‐derived Mitochondrial DNA Copy Number

Abstract

AbstractBackgroundMitochondrial DNA Copy Number (mtDNA‐CN) can be used to estimate mitochondrial function and it varies across tissues and cell types. Blood‐derived mtDNA‐CN has been studied in aging‐related diseases including Alzheimer’s Disease (AD). However, little is known about the genetic contribution to variation of blood‐derived mtDNA‐CN and how this variation affects the level of serum metabolites.MethodWe developed a fast computational pipeline to accurately estimate mtDNA‐CN using whole genome sequence data from blood samples from European ancestry (EA, 2,885 AD cases and 2,777 controls), African American (AA, 1,118 AD cases and 1,725 controls), and Caribbean Hispanic (CH, 1,022 AD cases and 1,949 controls) participants of the Alzheimer’s Disease Sequencing Project (ADSP). In the combined sample and within each population group. We tested association of blood‐derived mtDNA‐CN with 17,569,705 variants (minor allele count≥20) genome‐wide using a linear mixed effect model including covariates for age, sex, sequencing center, PCR, AD status, polygenic risk score of blood cell count, and principal components of ancestry. We also tested the associations of mtDNA‐CN and levels of 175 serum metabolites in 1,521 subjects from Alzheimer’s Disease Neuroimaging Initiative.ResultIn the total sample, we identified genome‐wide significant (GWS) associations with variants from 24 independent loci. The top‐ranked SNP (rs1455130415 in NDUFS8, (p=3.25x10‐42) was also GWS in EAs (p=3.96x10‐28) and AAs (p=2.20x10‐27). SNP rs997412864 (p=5.13x10‐9) in GCAT was found associated with mtDNA‐CN in EAs, but not in other groups. Association of SNP rs1237233197 (p=1.40x10‐8) in NDUFS7 was specific to AAs. Measurements of two long‐chain acylcarnitine, C14:2‐OH (p=4.91x10‐7) and C18:2 (p=1.81x10‐5), were negatively associated with mtDNA‐CN after Bonferroni correction (P<2.9x10‐4). Ornithine, a non‐proteinogenic amino acid, was also negatively associated with mtDNA‐CN (p=2.32x10‐4).ConclusionWe identified GWS association of blood‐derived mtDNA‐CN with a variant in a nuclear‐encoded mitochondrial gene, NDUFS8, which encodes a subunit of Complex I. In addition, mtDNA‐CN was significantly associated with two acylcarnitine metabolites and ornithine whose levels were previously reported to be decreased in blood of AD patients. These results provide more insight about the role of mitochondrial function in AD pathogenesis and suggest that blood‐derived mtDNA‐CN may be a useful biomarker for AD.