Mitochondrial Genetic Determinants of Peripheral Blood Transcriptomics, DNA Methylation and Blood Pressure

Human mitochondrial DNA (mtDNA) is a ∼16.5 kb circular molecule encoding 37 genes, including 13 protein subunits of the oxidative phosphorylation complexes responsible for producing the majority of adenosine triphosphate that cells use for chemical energy. The vast majority of the mitochondria-destined proteins, however, are encoded by nuclear DNA, and in order to adapt to the ever-changing cellular milieu, these two genetic compartments maintain a bi-directional regulation, leading to alterations in DNA methylation (DNAm) and gene expression. Genetic variation in the mtDNA is also capable of affecting DNAm and gene expression, but the effects in natural populations are not completely known. As mitochondria lie at an interface between the bioenergetic processes of the human body, they are potent mediators of common metabolic disorders. The maternal inheritance of mtDNA has also created a hypothesis that mtDNA variation may have sex-specific effects on specific traits.

The aims of this study were to examine the associations of mtDNA single-nucleotide polymorphisms (mtSNPs) with genome-wide peripheral blood transcriptome and DNA methylation profiles and to investigate whether these associations show sexual dimorphism or are affected by the onset of prediabetes, a condition preceding type 2 diabetes mellitus. The mitochondrial genetic determinants of blood pressure were also examined.

In study I, genome-wide peripheral blood transcript data and mtSNPs obtained by next-generation sequencing from a population-based Young Finns Study (YFS) cohort (n = 955) were used. In study II, a discovery association study on nuclear DNAm was performed in the YFS population (n = 926), and replication was sought in the Ludwigshafen Risk and Cardiovascular Health (LURIC) study (n = 2, 317). In study III, mitochondrial genetic associations with blood pressure were studied in the YFS (n = 1, 150) and in the Finnish Cardiovascular Study (FINCAVAS; n = 3, 112) by a meta-analysis. In the LURIC and FINCAVAS populations, mtSNPs were obtained with microarrays. The sexual dimorphism was examined by applying the analyses separately to males and females and testing for significant differences in the effect size (I–III). The prediabetes-specific effects were investigated similarly by performing the analyses separately to individuals with prediabetes and normoglycaemia (I–II).

Study I identified 53 associations between mtSNPs and gene transcripts, corresponding to 7 genes and 31 variants. Some associations were also replications of previously reported results. Eight associations remained significant after conditional analysis. In addition, five genes showed differential expression between haplogroups. One association demonstrated prediabetes-specific effects, while no evidence for sex- specific effects on gene expression was observed. In study II, numerous methylation quantitative trait mtDNA loci were observed in the discovery phase. Of these, 19 variant-based and four haplogroup-based associations were replicated and reached epigenome-wide significance when the results were combined in a meta-analysis. Two differentially methylated sites associated with gene transcripts in the YFS. Discovery analysis also showed that several associations had sex- or prediabetes-specific effects, but none of them were replicated in the LURIC population. Study III did not identify any variants associated with blood pressure levels nor any evidence of sexual dimorphism.

These results both reveal new and replicate previously reported genetic associations with peripheral blood transcriptomics. For the first time on a cohort level, mitochondrial genetic determinants of DNAm are presented. Some of the transcripts and CpG sites may be linked to the biological processes taking place in mitochondria, which suggests that the associations represent the mitochondrial–nuclear communication in order to maintain cellular homeostasis. The variant associations with transcriptomics and DNAm did not correspond to each other, which indicates that, if there is a causal relationship between mtDNA variation and peripheral blood transcriptomics, the regulatory mechanisms are not mediated by changes in DNAm. As suggested by several previous studies, mtDNA variation does not seem to have a significant role in the regulation of blood pressure. Finally, no convincing evidence was found of sex- and prediabetes-specific effects of mtDNA variation.