Peter Benn and Howard Cuckle
In 1997, Lo and colleagues reported that plasma from pregnant women carrying male fetuses contained cell-free (cf) DNA derived from the Y-chromosome (1). The clinical implications of this important observation were self-evident; the presence of conceptus-derived DNA in maternal plasma could be the basis of a novel approach to noninvasive prenatal testing. Follow-up studies demonstrated the utility of cfDNA for determining fetal sex and Rhesus blood group type (2). Using cfDNA to test for chromosome imbalances required additional techniques that would allow quantification of specific chromosome regions, and this was not achieved until 2008 when advances in DNA sequencing were applied to cfDNA (3,4).
Today, testing based on cfDNA provides the most effective method to screen for Down, Edwards, and Patau syndromes (trisomies 18 and 13). It can also be applied to sex chromosome abnormalities and select microdeletion syndromes. For some monogenetic disorders, cfDNA testing can be fully diagnostic. In the future, it is anticipated that whole-genome and whole-exome sequencing will be performed on cfDNA to identify a broad range of genetic conditions.
Biology of cfDNA
In healthy pregnant women, there are two classes of cfDNA present in the maternal plasma. The first type is derived from the mother. It mostly consists of DNA sequences approximately 166 bp in size, corresponding to the length that winds around one nucleosome plus an approximately 20 bp linker (5). This appears to be derived from apoptotic hematopoietic cells (6) and adipose tissue (7). In some women, there may be an additional maternal component from necrotic tissue or a malignancy (8). The second type of cfDNA is derived from trophoblasts or cytotrophoblasts, is approximately 143 bp in length, one nucleosome wrap, but lacking the linker sequence (5). It is commonly, although imprecisely, referred to as “fetal” cfDNA.
The relative proportion of cfDNA that is fetal is referred to as the fetal fraction (FF). Fetal cfDNA can be detected as early as 4 weeks in pregnancy, and the FF usually increases rapidly early in pregnancy, more slowly in the late first trimester through the second trimester, and again rapidly thereafter (9,10). By the time when most genetic screening will begin, 10–12 weeks’ gestation, the average FF is approximately 10%. Following delivery, fetal cfDNA rapidly disappears from the maternal circulation (11). There is, therefore, no concern that results of screening based on cfDNA could reflect a prior pregnancy.
FF can vary considerably from patient to patient, and there are a number of factors, in addition to gestational age, that are known to affect the levels. FF is lower in obese patients due to increased maternal adipose cell degradation and/or dilution of fetal cfDNA into a larger plasma volume (7). FF is also lower in pregnancies with fetal trisomy 13, 18, or digynic triploidy, which is most likely due to the presence of a smaller placenta and reduced production of fetal cfDNA (12,13). Maternal conditions such as intrahepatic cholestasis (14), autoimmune disease (15,16), use of medications such as heparin (17), and pregnancy complications (18,19) have been associated with altered FF.
Various approaches have been proposed for the detection of aneuploidy based on the analysis of cfDNA, and three methods are in widespread clinical use.
Whole-genome massively parallel sequencing (MPS) is based on sequencing and counting large numbers of unique cfDNA fragments in the plasma and assigning the fragments to the chromosome from which they originated (3,4). Both maternal and fetal DNA is sequenced, and a relative excess (trisomy) or deficiency (monosomy) for a chromosome of interest indicates aneuploidy. The sequencing may be carried out at a single end of each fragment or may involve both ends (20). The latter can be advantageous because the analysis will identify the fragment size, which differs depending on whether the DNA fragment is maternal or fetal in origin, and therefore can provide a relative measurement of the FF. Alternatively, a separate assay is used for FF quantification. There are various enhancements to the process including adjustments based on the GC content of the DNA, emphasis of the most informative sequences (principle component analysis), and averaging data from sliding widows along the chromosome or other noise reduction methods. However, counting a high number of sequences is a basic requirement when the FF is low or for detection of a small imbalance such as a microdeletion or microduplication (21). Since the sequencing involves DNA fragments from the entire genome, in principle, any large imbalance could potentially be detected (22).
A second approach targets specific chromosome regions of interest for enrichment and amplification. Initially sequencing, but more recently a microarray, is then used to determine if there is an excess or deficiency for one particular chromosome or region relative to others (23). As with genome-wide cfDNA screening, the test does not distinguish between maternal and fetal imbalances. FF is established by evaluating regions where there is a high probability of single nucleotide polymorphisms (SNPs) that differ between maternal and fetal genomes. Potential advantages of this targeted methodology are a lower cost, because not all regions are evaluated for copy number, and evaluation of copy number based on higher numbers of DNA fragments from specific chromosome regions of interest. Expanding the scope of the test to look for additional abnormalities is possible but would require a more fundamental redesign with incorporation of new probes.
The third method that has been developed for cfDNA screening for fetal aneuploidy relies on analyzing SNPs and determining the relative quantitative contributions of maternal and fetal DNA in the plasma. One version of the test involves the multiplex polymerase chain reaction (PCR) and analysis of over 13,000 polymorphic regions on chromosomes 13, 18, 21, X, and Y plus additional sites for the detection of microdeletion syndromes (24). All informative loci are incorporated into an algorithm that provides both a maximum likelihood estimate of the copy number at each region and a measure of FF. The approach also identifies diandric triploidy and molar pregnancies (25,26). The method is expandable to include additional imbalances, but this does require sufficient informative SNPs within the region of interest.
There are other methods that are in clinical use in some countries, but these generally are based on more limited validation studies or scant published data on clinical experience. One is based on quantitative PCR of specific target sequences (27). Another relies on incorporation of target regions into circular DNA that can undergo rolling circle replication, fluorophore labeling of the products, and detection following nanoparticle filtration (28). It is reasonable to anticipate that other approaches to cfDNA screening will emerge. While all methods appear to show high detection rates and low false-positive rates for Down syndrome, it should not be assumed that they are all equivalent.
False-positive and false-negative results
One of the most common reasons for discordance between cf-DNA results and fetal karyotype is the presence of a cell line that is substantially confined to the trophoblast cell lineage and is not present, or not detectable, in the fetus (“confined placental mosaicism,” CPM) (Table 34.1) (29). Based on cytogenetic studies of chorionic villus samples, the chance for CPM will depend on the specific chromosome under consideration; for example, CPM involving chromosome 13 is far more common than chromosome 21.
Table 34.1 Biological explanations for false-positive and false-negative results for cytogenetic abnormalities or fetal sex mis-assignment
Confined placental mosaicism
Copy number variant
Maternal, somatic (distal to a fragile site)
Sex chromosome abnormality
Residual cytotrophoblasts from nonviable co-twin
Low fetal fraction