Noninvasive prenatal testing to analyze the fetal genome

Prenatal genetic testing has changed markedly since the original introduction of amniocentesis as a means to evaluate the fetal karyotype. Because it has long been recognized that maternal age is highly associated with risk for Down syndrome, maternal age became the first screening test for aneuploidy. Later development of screening with maternal serum analytes and nuchal translucency ultrasound all led to improvements in prenatal genetic testing, with higher detection rates and lower screen-positive rates. The past decade has seen further advances with the discovery of cell-free fetal DNA (cfDNA) in the maternal circulation and development of sequencing and bioinformatics analytic approaches to very accurately assess the risk of Down syndrome. In 2011, this culminated in the clinical introduction of noninvasive prenatal testing (NIPT) for aneuploidy using cfDNA (1). With very-high sensitivity and very-low false-positive rates, this advance represented a tremendous step forward in Down syndrome detection. Subsequently, with higher depth of sequencing and improved bioinformatics analyses, NIPT expanded to include detection of a number of microdeletions (2, 3). Genome-wide screening for copy number variants has also been reported and is now offered clinically (4, 5). Noninvasive identification of fetal single-gene disorders, and ultimately analysis of the fetal genome, has become the “next frontier” in prenatal diagnosis (6) (Fig. 1).

Fig. 1.

An overview of currently available applications of massively parallel sequencing-based noninvasive prenatal testing using cell-free nucleic acids (DNA and RNA) in maternal plasma thought to originate from apoptotic trophoblasts. The fetal genome (from chromosomal to subchromosomal to single-gene levels) may be decoded; the methylome and transcriptome have also been investigated. A, adenine; C, cytosine; G, guanosine; mC, methylated cytosine; MPS, massively parallel sequencing; RNA-seq, RNA sequencing; T, thymine; U, uridine. Reproduced from ref. 17, with permission of Annual Review of Medicine, Volume 67 © 2016 by Annual Reviews.

NIPT can now be used to test for a number of single-gene disorders, including Rh blood type and paternally inherited dominant as well as de novo conditions (7, 8). Fetal sex can be determined using cfDNA analysis by the presence or absence of Y chromosome-specific sequences (e.g., DYS14 or SRY) in the maternal serum (9). Similarly, diagnosis of paternally inherited monogenic traits and disorders can be done by looking for characteristics specific to the fetus and not present in the mother. However, although testing has been reported for some autosomal-recessive and X-linked conditions, this is more difficult given the obligate background of the maternal genomic make-up, and may depend on whether the fetus is at risk for being a homozygote or a compound heterozygote (10).

The potential for use of genomic sequencing to individually optimize medical treatment in diseases such as cancer is just beginning to be realized, and the utility of sequencing healthy individuals is also being explored. Although interpretation of the large amount of data produced is complex, clearly this is the direction of medicine. It has been determined that the entire fetal genome is represented in maternal plasma, and noninvasive measurement of the fetal genome has been reported by a number of investigators (11–13). Still, although the potential benefit of such techniques is promising, in reality this has continued to be limited by the large amount of data obtained, with the large number of candidate de novo fetal mutations leading to a low positive predictive value and limiting the clinical utility of this approach.

In PNAS, Chan et al. (6) sequenced plasma DNA from a pregnant woman to a depth of 270× haploid genome coverage. By comparing the maternal plasma DNA with the parental genomic DNA, and using a series of bioinformatics filters, 85% of fetal de novo mutations were detected at a positive predictive value (PPV) of 74%, a tremendous improvement in PPV over previous strategies. The approach described in this report allowed Chan et al. to determine the maternal inheritance of the fetus for >94% of heterozygous SNPs within the maternal genome. This approach was then validated in a pregnancy diagnosed with cardiofaciocutaneous syndrome as a result of a de novo BRAF mutation in the fetus (6).

In part because of the focus on clinical application of NIPT, there has been less study of the physical and biological characteristics of fetal DNA in maternal plasma. Circulating fetal DNA has been reported to be shorter than maternal DNA (14), but little has been known regarding the molecular basis of this observation. Recent investigation has shown that fragmentation sites of plasma DNA are located in clusters across the genome (15); Chan et al. (6) have now discovered that preferred end sites in plasma DNA exhibit specificity based on their origin (e.g., the placenta versus the mother). This finding opens up new and exciting avenues of investigation, as better understanding of these recurrent end positions may lend insight into the physiology, as well as the pathophysiology, of pregnancy, as well as cancer and other disorders.

This “second generation” approach to produce noninvasive fetal genomes at high resolution using maternal plasma DNA sequencing represents a tremendous advance in prenatal diagnosis. Chan et al. (6) demonstrate the feasibility of detecting fetal de novo mutations across the genome noninvasively from a maternal serum sample at substantially increased resolution and specificity compared with previous techniques. The possibility for clinical implementation is increasingly limited only by the costs involved in very-high-depth sequencing of maternal plasma DNA. It is anticipated that the cost of sequencing will continue to decrease, and the potential availability for whole-exome and whole-genome sequencing of the fetus using cfDNA in the maternal plasma has moved closer to reality.

Prenatal genetic testing has always been associated with ethical debate. With expansion of the range of conditions that can now be detected prenatally, these debates have seen renewed vigor (16). Such discussions usually focus on abortion, and concern that pregnancies will be terminated because of less-severe conditions or because of excess maternal anxiety over variants of uncertain significance. Interpreting the clinical significance of de novo variants continues to present a tremendous challenge, even when the sequencing is performed on fetal tissue, as with amniocentesis. In a child or an adult with a medical condition or serious disorder, such variants may be complex to interpret, but with prenatal diagnosis, decisions regarding pregnancy management may need to be made by expectant parents under conditions of tremendous uncertainty.

Concerns of providers, professional societies, ethicists, and others have often resulted in delayed introduction of genetic testing into the area of prenatal diagnosis. However, patients and providers have clearly jumped on the bandwagon of NIPT, and it is anticipated that expansion of the menu of conditions available for noninvasive testing will be accompanied by increasing demand for such tests. Although discussion of prenatal diagnosis and the severity of conditions that warrant testing generally are based on an assumption that abortion will be the outcome of affected pregnancies, this is not the only rationale for prenatal diagnosis. With increasing availability for postnatal treatments—some of which improve outcomes if started promptly after birth—and the increasing possibility of in utero treatment, the conversation should mature to consideration of the benefits of detecting conditions to impact and improve perinatal care. We are moving ever closer to a time when children are born with their genome known, and their lifelong medical care is more precisely directed. Given the complexity of such tests and their interpretation, strategies and frameworks for their use under different clinical conditions need to be established to assure that the benefits can be realized while minimizing risks of misuse and misunderstanding by patients and their families.