Fetal mitochondrial DNA in maternal plasma in surrogate pregnancies: Detection and topology

Abstract Objectives Due to the maternally‐inherited nature of mitochondrial DNA (mtDNA), there is a lack of information regarding fetal mtDNA in the plasma of pregnant women. We aim to explore the presence and topologic forms of circulating fetal and maternal mtDNA molecules in surrogate pregnancies. Methods Genotypic differences between fetal and surrogate maternal mtDNA were used to identify the fetal and maternal mtDNA molecules in plasma. Plasma samples were obtained from the surrogate pregnant mothers. Using cleavage‐end signatures of BfaI restriction enzyme, linear and circular mtDNA molecules in maternal plasma could be differentiated. Results Fetal‐derived mtDNA molecules were mainly linear (median: 88%; range: 80%–96%), whereas approximately half of the maternal‐derived mtDNA molecules were circular (median: 51%; range: 42%–60%). The fetal DNA fraction of linear mtDNA was lower (median absolute difference: 9.8%; range: 1.1%–27%) than that of nuclear DNA (median: 20%; range: 9.7%–35%). The fetal‐derived linear mtDNA molecules were shorter than the maternal‐derived ones. Conclusion Fetal mtDNA is present in maternal plasma, and consists mainly of linear molecules. Surrogate pregnancies represent a valuable clinical scenario for exploring the biology and potential clinical applications of circulating mtDNA, for example, for pregnancies conceived following mitochondrial replacement therapy.


| INTRODUCTION
Most studies of cell-free DNA in plasma focus on DNA molecules in a linear form derived from the nucleus. These linear plasma DNA molecules display characteristic fragmentation patterns, and are associated with nucleosomal structures. 1 For instance, the modal frequency of fragment sizes is approximately 166 bp, accompanied by multiple 10-bp periodic peaks of small sizes. Such characteristic patterns have inspired recent interest in plasma DNA fragmentomics, encompassing areas such as plasma DNA preferred ends, 2,3 fragment sizes, 1,4,5 end motifs, 6,7 nucleosome footprints, [8][9][10] plasma DNA end orientations 11 as well as biological links between fragmentation processes and nucleases. 6,12,13 In addition to studying the properties of nuclear DNA molecules in plasma, more recently, on the study of subjects following liver and bone marrow transplantation, we have demonstrated that there are two different topologic forms of mitochondrial DNA (mtDNA), namely linear and circular forms, present in plasma DNA. 14 Of note, mtDNA molecules of hematopoietic origin appear to be mainly circular, in sharp contrast to the liver-derived mtDNA molecules that mainly display a linear form. 14 These results suggest that plasma mtDNA topology might vary depending of the tissues of origin of the detected mtDNA.
Topologic study of plasma DNA can also be extended to autosomal DNA molecules. 15 For example, the presence of extrachromosomal circular DNA (eccDNA) was observed in both maternal and fetal DNA molecules in pregnant women, and displaying distinct size patterns. 15 While circular mtDNA and eccDNA in plasma are governed by different biologic mechanisms, our work has highlighted their commonality in their accessibility to being studied using similar methodologies.
In this study, we explore whether tissue-associated topologic forms of mtDNA might be present in the plasma of pregnant women.
Pregnancy offers an attractive model for studying the biology of tissue-specific plasma DNA molecules because there is the coexistence of fetal-and maternal-derived DNA in maternal plasma. 16 However, the study of fetal mtDNA in maternal plasma is complicated by the fact that mitochondria are inherited from the mother. Hence, it would not normally be possible to differentiate fetal and maternal mtDNA molecules in natural pregnancies. In the context of this study, a surrogate mother is a woman engaged by a couple to become pregnant with an embryo fertilized using gametes of the couple. Thus, in this study, surrogate pregnancy was chosen as a model to explore the topologic forms of circulating fetal mtDNA.

| Subjects
The subject recruitment and sample collection were conducted in What is already known about this topic?
• In natural, non-surrogate pregnancies, fetal nuclear DNA in maternal plasma consists of short linear DNA fragments that have a shorter size distribution than the maternal background nuclear DNA.
• Both linear and circular mitochondrial DNA (mtDNA) molecules co-exist in plasma of liver and bone marrow transplant patients.
• Liver tissues contribute a considerable amount of mtDNA predominantly in a linear form to the plasma DNA pool.
• mtDNA molecules of hematopoietic origin are mainly in a circular form in plasma.
What does this study add?
• Fetal and maternal mtDNA molecules could be distinguished in the plasma of surrogate pregnant mothers.
• Linear and circular mtDNA molecules co-existed in the plasma of surrogate pregnant women.
• The fetal mtDNA molecules in surrogate maternal plasma appeared to be predominantly in a linear form, while approximately half of the surrogate maternal mtDNA molecules were in a circular form.
• The fetal mtDNA fraction was less abundant than the fetal nuclear DNA fraction in surrogate pregnant women.

| Sequencing and alignment
Target-captured DNA libraries were sequenced in a 2 × 70 bp pairedend mode on a NextSeq 500 system (Illumina) using the NextSeq 500 High Output Reagent Cartridge v2 Kit (Illumina). Sequencing reads were attributed to multiple samples based on unique dual-index sequences using Picard tools (https://broadinstitute.github.io/picard/).
The demultiplexed reads were aligned to the reference genome database including the human reference genome (NCBI37/hg19) and the mitochondrial genome using BWA. 17 Reads with UMIs information were used to generate consensus reads using fgbio tools. 18 SOAP2 19 was utilized to realign the consensus reads to determine the genomic origins for each paired-end sequenced reads. During the alignment based on SOAP2, 19 we allowed up to two mismatches to make those reads with the single nucleotide variants mappable for the downstream analysis. The fetal and maternal DNA molecules were differentiated from the mapped reads based on informative single nucleotide variants. 1

| Determination of linear and circular mtDNA
The restriction enzyme BfaI was used for differentiating linear and circular mtDNA as previously described. 14 This enzyme cleaves C^TAG sites and is expected to cut DNA once every 256 bp (1/4 4 ). Briefly, after plasma DNA was treated with BfaI, the linear mtDNA molecules in plasma would mainly either remain uncleaved or be cleaved into two fragments both with one BfaI-cleaved end signature. On the contrary, an intact circular mtDNA (~16.5 kb), containing a total of 98 BfaI cutting sites, would be cleaved into linear mtDNA fragments each bearing two cleavage ends. The principle of the topologic analysis of F I G U R E 1 Schematic illustration of the topological assessment of mtDNA in surrogate pregnancies. Fetal mtDNA may be released into surrogate maternal plasma through the placenta during pregnancy. BfaI cleavage signatures were used for differentiating the intact circular mtDNA molecules (~16.5 kb) from pre-existing linear mtDNA molecules (generally <200 bp). BfaI has a 4-bp recognition sequence which by random chance should occur once every 256 bp (1/44). Therefore, fragments originated from intact circular mtDNA molecules would carry two BfaI-cleavage ends (Cyan DNA) after BfaI digestion (Cyan scissor), whereas the majority of the fragments derived from pre-existing linear mtDNA molecules would have no more than one cleavage end (Cyan DNA). Single nucleotide variants in mitochondrial genomes were used to determine the physical forms of fetal and maternal mtDNA in plasma of surrogate pregnant women. Combined with topological assessment of mtDNA, we could identify if the fetal-derived (red-yellow DNA) or maternal-derived (purple-blue DNA) mtDNA molecules were in a circular or linear form according to the patterns of BfaI-cleavage signatures [Colour figure can be viewed at wileyonlinelibrary.com] plasma mtDNA for surrogate pregnant subjects is illustrated in

| Determination of informative variants in mitochondrial genomes and nuclear genomes
The genotypes of fetal and surrogate mother's mitochondrial genomes were determined by sequencing relevant tissue DNA including buffy coat or saliva DNA from oocyte donors or surrogate mother's buffy coat DNA. A cell can contain numerous copies of mtDNA among which some may carry different variants in certain locations of the mtDNA genome, but some may not, which is referred to as mtDNA heteroplasmy. To minimize the influence of the mtDNA heteroplasmy in a tissue, we focused on the informative variants in the mitochondrial genome for which the alleles of a single nucleotide variant site were identical (i.e., homoplasmic) but different between the surrogate mother and oocyte donor. Due to the maternal inheritance of mtDNA, fetal-specific mtDNA variants were represented by oocyte donor specific mtDNA variants. Fetal mtDNA fraction (F m ) was deduced by the allelic ratio between fetal-specific variants (e.g., A) and surrogate maternal specific variants (e.g., B) present in the mitochondrial genome: In our present study, the placental tissues were not available and the fetal genotypes were determined non-invasively. Thus, we could not directly obtain the fetal genotype information. In our analysis, we utilized the nuclear genotypes from tissue genomic DNA from the surrogate mother and the oocyte donor mother, in conjugation with sequencing result from one plasma aliquot, to deduce the informative SNPs allowing the fetal nuclear DNA fraction (F n ) estimation in the second plasma aliquot. For the sake of simplicity, we first focused on the SNPs which were homozygous for the same allele both in the surrogate mother (AA) and the oocyte donor mother (AA   14 However, we did not observe such differences in the size distribution of nuclear DNA molecules between before and after BfaI digestion ( Figure S1B). The data provided further evidence that there may exist circular mtDNA in plasma of surrogate pregnant women.

| Topologic analysis of fetal DNA and maternal mtDNA
We identified single nucleotide variants in the mitochondrial genome whose sequences differed between oocyte donor and paired surro- For the maternal plasma DNA samples without BfaI digestion, the median fetal nuclear DNA fraction was 20% (range: 9.7%-35%), whereas the fetal mtDNA fraction was found to be much lower On the basis of BfaI cleavage end signature analysis, we found that the proportion of linear mtDNA among fetal-derived mtDNA molecules (i.e., placenta-derived; median: 88%; range: 80%-96%) was much higher than that among maternal-derived mtDNA molecules (i.e., mainly of hematopoietic and liver origin; median: 49%; range: 40%-58%) (Figure 4).

| Size distribution of fetal-and maternalderived mtDNA
We next analyzed the size profiles of the linear mtDNA molecules in plasma DNA of surrogate pregnant women without BfaI digestion.
The fetal-derived mtDNA molecules were much shorter than the maternal counterparts ( Figure 5(A) and Figure S2). The fetal-derived nuclear DNA fragments in all five surrogate maternal plasma DNA were found to be shorter than the maternal nuclear DNA ( Figure  However, we could not completely rule out the possibility that fragments carrying two BfaI sites were originated from long linear mtDNA. According to our previous study, there was a median of 18% of pre-existing linear mtDNA carrying at least two BfaI-cleavage sites. 14 However, 50% of mtDNA molecules in plasma mtDNA were found to carry two BfaI sites after digestion, suggesting that the circular mtDNA molecules were present in the surrogate maternal plasma DNA.
In plasma DNA without BfaI digestion, the mtDNA being analyzed are likely to be spontaneously-fragmented linear molecules, as sequencing adaptors could not be ligated to circular DNA molecules.
Among the linear mtDNA molecules, the fetal mtDNA fraction was found to be relatively lower in surrogate maternal plasma compared with the fetal contribution in nuclear DNA fragments, which was opposite to the observation that liver DNA contribution was greatly enriched in linear mtDNA molecules. 14  and~338 bp with sharp 10-bp periodicities. 15 These results suggest that restriction enzyme-based topologic analysis is not only suited for eccDNA identification but can also be used to infer the large circular DNA such as an entire circular mitochondrial genome. It would be interesting to use this approach to study whether circular DNA molecules might be associated with diseases such as cancer in the future study.
We have shown that placental mtDNA molecules are present at readily detectable amounts in maternal plasma and that they are mainly linear in nature. We believe that these realizations may catalyze the development of placental-specific mtDNA assays even in

DATA AVAILABILITY STATEMENT
Research methods are available by contacting the corresponding author. Participants have not consented for sequence data sharing.