Quantitation and Characterization of Mitochondrial DNA in Cell-free Plasma by Deep Sequencing

The presence in plasma of mitochondrial DNA (mtDNA) fragments, a proinflammatory damage-associated molecular pattern (DAMP), is positively associated with outcomes of multiple human disorders. Because of the low mtDNA abundance in plasma, qPCR is typically employed as an analytic strategy. However, this method provides little insight into sequence origins or other characteristics of circulating mtDNA fragments. Here we found that target bait-capture applied to plasma mtDNA derived from severely injured patients enriched recovery by about 1600-fold, thus affording depth sufficient for NextGen sequence analysis. Failure to exclude nuclear mitochondrial (NUMT) pseudogenes from the mtDNA read pool led to overestimation of mtDNA DAMP abundance. After excluding NUMTs using a stringent alignment filtering strategy calling and quantifying NUMTs and polymorphic NUMTs not reported in the human reference, we found 2000-68000X total mtDNA DAMP coverage. Normalization of mtDNA abundance to NUMT coverage reduced batch variability. Two massively-transfused trauma patients and two non-transfused patients displayed time-dependent increases in mtDNA DAMP coverage characterized by leftward shifts and narrowing of fragment length-density curves indicating shorter mean fragment lengths and less dispersion around mean values. Our approach enabled calculation of heteroplasmic diversity, which diminished with time when expressed as a percentage of total mtDNA DAMP abundance. Finally, we found that mtDNA DAMPs originating from certain mt genomic regions harbored oxidatively damaged bases. These findings suggest that target bait-capture and deep sequencing provide unprecedented characterization of cell-free plasma mtDNA and NUMTs.