Age-related rise in blood mitochondrial DNA variants explained by replication errors and clonal hematopoiesis
Background
The increase with age in mitochondrial DNA (mtDNA) variants detected in blood may be explained by replication-related errors and the clonal expansion of blood cells. This model is supported by a large human-genomics study published in Nature. Its central distinction is between how variants arise and how they increase to detectable levels.
Mitochondrial DNA heteroplasmy means that different mtDNA sequences coexist within a cell or tissue. The researchers examined mtDNA single-nucleotide variants (mtSNVs) that increase with age. Oxidative damage caused by reactive oxygen species has long been considered a source of these variants, but how they increase after arising to levels detectable in blood samples also needed explanation.
Key findings
The team analyzed whole-genome sequencing (WGS) data from 736,038 people in UK Biobank and the US All of Us program. They used the mtSwirl analysis pipeline, which reconstructs each person’s mtDNA reference sequence, to call variants at low heteroplasmy levels. After quality control, 620,385 participants were included in the variant analyses. Age-specific analyses were restricted to ages 18–90 in All of Us and 40–70 in UK Biobank.
Blood mtSNV counts increased particularly markedly after age 60. The variant profile featured C>T and A>G substitutions biased toward the mtDNA heavy strand, while C>A substitutions, a characteristic signature of oxidative damage, were rare. The researchers interpreted this pattern as more consistent with replication-related errors than with oxidative damage. Deamination of DNA exposed as a single strand during replication was proposed as a possible explanation. Errors by DNA polymerase gamma may contribute to some variants, but the findings do not establish errors by a particular enzyme as the cause of all variants.
A clue to the increase in detectable variants came from clonal hematopoiesis (CH), in which a cell population originating from particular blood-forming cells expands. A genome-wide association study of mtSNV counts included up to 547,947 people and identified associations with CH-related variants near TERT, TCL1A, and SMC4. Rare-variant analyses also linked variants in known CH driver genes, including DNMT3A, TET2, and ASXL1, to higher mtSNV counts. The major genetic associations appeared to persist even after 17,504 carriers of known CH driver variants were excluded.
These findings support a two-step explanation. First, replication-related errors give rise to low-level mtDNA variants within cells. Subsequently, independently arising somatic variants in nuclear DNA drive the expansion of blood-cell clones, bringing coexisting mtDNA variants to detectable levels. The age-related mtSNVs analyzed generally had low heteroplasmy levels, and evidence for positive selection was weak. They were therefore interpreted as likely passenger variants that accompany clonal expansion rather than drive it.
Meaning and outlook
The increase in mtDNA variants observed in this study cannot readily be interpreted as mitochondrial dysfunction or a cause of age-related disease. The low-level passenger variants studied here should be distinguished from pathogenic variants that cause mitochondrial disease at high heteroplasmy levels. The researchers propose evaluating mtSNV counts as a sensitive marker of somatic mosaicism associated with CH—that is, the coexistence of genetically distinct cell populations.
However, bulk blood data combining multiple cells could not establish how individual variants were distributed across cells and clones. Filters were applied to reduce the risk of misreading mtDNA-like sequences inserted into the nuclear genome, but a risk of residual contamination remains. Other limitations include the inability to establish whether individual variants were inherited or somatic, and uncertainty about generalizing the findings to other tissues. Single-cell analysis is a next step toward clarifying the distribution of variants across cells and their relationship to clones.
The study methods and interpretation are available in the full original paper.
Nature, Published online: 27 May 2026. DOI: 10.1038/s41586-026-10569-6
Summary: Challenging the historical free-radical dogma where reactive oxygen species-driven oxidative damage was defined as the absolute generator of mitochondrial genomic instability, this landmark population-scale study delineates the structural implementation of high-depth genome-wide analyses over multi-thousand blood cohorts. The deep molecular registry isolates that the accumulation of mitochondrial DNA mutations is programmatically driven by cryptic replication errors rather than environmental oxidation kinetics. These heritable sequence alterations operate initially as sub-clinical passenger mutations, which dynamically expand into high-penetrance arrays due to age-related somatic mosaicism. The data provides a non-oxidative computational baseline establishing DNA polymerase fidelity as the principal driver of transcriptomic aging, optimizing prospective interventions, and liquid biopsy stratification.
Analyzing mtSNV counts in existing blood WGS data could reveal clues to clonal changes that known nuclear DNA driver variants alone do not capture. For example, researchers could select samples with high mtSNV counts but no detected CH driver for follow-up single-cell analysis, then test whether expanded blood-cell clones are actually present.
This is a potential research application of a candidate marker. Clinical diagnostic use would require further evaluation of reproducibility in independent populations, detection criteria, and predictive performance. This study did not establish the performance of an aging score or companion diagnostic, nor did it test the efficacy of antioxidants or anti-aging drugs targeting DNA polymerase gamma.