Comparison of the structures and topologies of plasma extracted circulating nuclear and mitochondrial cell-free DNA

Abstract

Introduction: The function, origin and structural features of circulating nuclear DNA (cir-nDNA) and mitochondrial DNA (cir-mtDNA) are poorly known, even though they have been investigated in numerous clinical studies, and are involved in a number of routine clinical applications. Based on our previous report disproving the conventional plasma isolation used for cirDNA analysis, this work enables a direct topological comparison of the circulating structures associated with nuclear DNA and mitochondrial cell-free DNA. Materials and methods: We used a Q-PCR and low-pass whole genome sequencing (LP-WGS) combination approach of cir-nDNA and cir-mtDNA, extracted using a procedure that eliminates platelet activation during the plasma isolation process to prevent mitochondria release in the extracellular milieu. Various physical procedures, such as filtration and differential centrifugation, were employed to infer their circulating structures. Results: DSP-S cir-mtDNA mean size profiles distributed on a slightly shorter range than SSP-S. SSP-S detected 40-fold more low-sized cir-mtDNA fragments (<90 bp/nt) and three-fold less long-sized fragments (>200 bp/nt) than DSP-S. The ratio of the fragment number below 90 bp over the fragment number above 200 bp was very homogenous among both DSP-S and SSP-S profiles, being 134-fold lower with DSP-S than with SSP-S. Cir-mtDNA and cir-nDNA DSP-S and SSP-S mean size profiles of healthy individuals ranged in different intervals with periodic sub-peaks only detectable with cir-nDNA. The very low amount of cir-mtDNA fragments of short size observed suggested that most of the cir-mtDNA is poorly fragmented and appearing longer than ∼1,000 bp, the readout limit of this LP-WGS method. Data suggested that cir-nDNA is, among DNA extracted in plasma, associated with ∼8.6% of large structures (apoptotic bodies, large extracellular vesicles (EVs), cell debris…), ∼27.7% in chromatin and small EVs and ∼63.7% mainly in oligo- and mono-nucleosomes. By contrast, cir-mtDNA appeared to be preponderantly (75.7%) associated with extracellular mitochondria, either in its free form or with large EVs; to a lesser extent, it was also associated with other structures: small EVs (∼18.4%), and exosomes or protein complexes (∼5.9%). Conclusion: This is the first study to directly compare the structural features of cir-nDNA and cir-mtDNA. The significant differences revealed between both are due to the DNA topological structure contained in the nucleus (chromatin) and in the mitochondria (plasmid) that determine their biological stability in blood. Although cir-nDNA and cir-mtDNA are principally associated with mono-nucleosomes and cell-free mitochondria, our study highlights the diversity of the circulating structures associated with cell-free DNA. They consequently have different pharmacokinetics as well as physiological functions. Thus, any accurate evaluation of their biological or diagnostic individual properties must relies on appropriate pre-analytics, and optimally on the isolation or enrichment of one category of their cirDNA associated structures.

Keywords: Q-PCR; circulating DNA; diagnostics; extracellular vesicles; mitochondria; structure; topology; whole genome sequencing.

FIGURE 1

Circulating mitochondrial DNA (cir-mtDNA) size profile in healthy individuals. Relative frequencies of cir-mtDNA fragments from seven subjects (seven colours) as determined by low-pass WGS: (A), DSP sequencing (DSP-S); (B), SSP sequencing (SSP-S). Mean percent of fragments of cir-mtDNA from the seven subjects as determined by DSP-S [green, (C,E,F)] and SSP-S [blue, (D,E,F)]: DSP-S alone (C), SSP-S alone (D), SSP-S in foreground and DSP-S in background (E), DSP-S in foreground and SSP-S in background (F).

FIGURE 2

Comparison of the mean size profiles of circulating mitochondrial DNA (cir-mtDNA) and circulating nuclear DNA (cir-nDNA) from seven healthy individuals with alternate foreground. Low-pass WGS mean size profiles of cir-mtDNA (red) and cir-nDNA (dark blue) obtained with DSP-sequencing (A,B) and SSP-sequencing (C,D): cir-mtDNA in foreground and cir-nDNA in background (A,C); cir-mtDNA in background and cir-nDNA in foreground (B,D).

FIGURE 3

Comparison of the concentrations and variations of circulating nuclear (cir-nDNA, grey) and mitochondrial DNA (cir-mtDNA, blue) from 5 healthy individuals, depending on the physical process applied. The preparation protocol used was without platelet activation (PPw/oPA, (A,B). The physical procedures applied were low-speed centrifugation (LS), filtration (F), and high-speed centrifugation (HS). Samples were either treated with LS only or with LS combined either with F (LS+F) or with HS (LS+HS). The respective variation of the cirDNA concentration in each group was obtained comparing each combination with LS only, used as a reference. Results are presented as boxplots (A) and histograms (B). ns, non-statistically significant; *p < 0.05, **p < 0.01; ***p < 0.001.

FIGURE 4

Effect of incremental centrifugation speed on plasma from five healthy individuals. Comparison of the concentrations [(A), ng/mL], and variations [(B), %)] of circulating nuclear DNA (cir-nDNA, grey) and circulating mitochondrial DNA (cir-mtDNA, blue) in plasma as prepared by the protocol without platelet activation. cirDNA concentration was determined in the supernatant following the subsequent centrifugation steps of the respective supernatant at 16,000 g, 40,000 g and 200,000 g. ns, non-statistically significant; *p < 0.05, **p < 0.01; ****p < 0.0001.