Inherited and acquired mitochondrial DNA damage may directly raise blood cancer risk

Mitochondrial DNA (mtDNA) mutations, called heteroplasmies when mutant and normal copies coexist in a cell, accumulate with age and smoking. They have been linked to blood cancers and higher mortality. A key unanswered question is whether deleterious heteroplasmies drive malignancy or are passengers carried along by clonal expansion of cells with nuclear mutations, as in clonal hematopoiesis of indeterminate potential (CHIP). The answer affects prevention and treatment strategies.

The team used 22,154 first-degree relative pairs from the UK Biobank (mother-offspring, full siblings, identical twins) and 13,831 pairs from All of Us. Heteroplasmies (variant allele fraction 0.05–0.95) were called with MitoHPC. A variant seen in both relatives was treated as inherited; an unshared one was treated as largely acquired. Father-offspring pairs served as a negative control, with a false-positive rate of about 0.34% after removing one common variant. The researchers then used a Mendelian-randomization-style logic. Shared burden, measured by the mMSS score of predicted deleteriousness, is fixed at conception and so cannot simply reflect later clonal expansion.

Shared variants had higher allele fractions (mean 0.39 vs 0.15) but were much less likely to be missense, nonsense, or predicted deleterious, consistent with strong purifying selection. Allele-fraction correlations followed the oogenesis bottleneck: identical twins r=0.96, mother-offspring 0.68, siblings 0.53, with about 3 transmitted mtDNA units. Roughly 90% of the overall deleterious burden came from unshared variants (r²=0.88). Older maternal age at birth raised heteroplasmy count but not mMSS, suggesting the extra variants were largely benign.

Shared and unshared mMSS gave concordant effect sizes for hematological malignancy. Combining UK Biobank with age-restricted All of Us data, shared mMSS was associated with hematological malignancy (RR 2.81, 95% CI 1.29–6.13). Both shared and unshared burden were associated with high-risk CHIP, particularly spliceosome mutations. Only unshared burden was associated with low-risk CHIP. Combined, shared mMSS was associated with high-risk CHIP (OR 5.81, 95% CI 1.44–23.4). The abstract also reports ultra-rare individual mtDNA variants linked to hematological malignancy, a pattern expected of driver mutations. The supplied text cuts off before those results and the discussion.

The results suggest mitochondrial dysfunction may promote clonal expansion of specific high-risk CHIP subtypes and that most of the relevant burden arises during life. That points to possible prevention or monitoring opportunities, especially for myeloid cancers. Caveats: this is a non-peer-reviewed preprint. The confidence intervals are wide and event counts are modest. All of Us effects were smaller and its CHIP detection from whole-genome sequencing was less sensitive. Using shared variants as an instrument assumes no pleiotropy or confounding by maternal lineage and haplogroup, so causality is supported rather than proven.