Introduction

Traditionally, prenatal diagnosis of fetal genetic status or aneuploidy has been dependent on the use of invasive diagnostic tests (chorionic villus sampling [CVS] and amniocentesis) to obtain a sample of fetal genetic material. The invasive nature of these tests means that there is a small but significant risk of miscarriage of around 1%, and alternative sources of fetal material have long been sought to allow non-invasive prenatal diagnosis (NIPD).

The original focus for NIPD research was fetal cells in the maternal circulation. This line of investigation, however, has not been fruitful, as fetal cells have proven difficult to isolate and may not be pregnancy specific. In 1997, Lo et al. identified the presence of cell-free fetal DNA (cffDNA) in maternal plasma. Cell-free fetal DNA can be detected from 4–5 weeks gestation. Importantly, it is rapidly cleared from the maternal circulation after delivery, which means cffDNA is pregnancy specific. Free-fetal DNA emanates from trophoblast cells in the placenta, and is made up of relatively short fragments of DNA (143 bp in length on average). It is now also evident that the whole fetal genome is represented in the maternal plasma.

Most cell-free DNA (cfDNA) in the circulation is maternal in origin. Digital polymerase chain reaction (PCR) has allowed precise quantification of cfDNA levels, showing that the fetal component contributes around 9% in early pregnancy, increasing to 20% as the pregnancy progresses ; however, this is highly variable between pregnancies. The key limiting factor in the development of specific prenatal tests based on cffDNA has been the difficulty of differentiating the genetic material from the fetus from that of the mother. As a result, the first clinical applications of NIPD have been restricted to the identification of alleles present in the fetus but not in the mother (either inherited from the father or arising de novo ). These include fetal sex determination, paternally inherited single-gene disorders, or those arising de novo , such as achondroplasia. This perspective is rapidly changing with the development of new technologies such as digital PCR and massively parallel sequencing (MPS). These allow researchers to develop NIPD tests for single-gene disorders where parents have the same mutations and for aneuploidies, both of which need to take into account the presence of the mother’s allele or chromosomes.

In addition to testing for genetic and chromosomal disorders, NIPD can also be used to manage other complications of pregnancy; indeed, one of the early success stories for NIPD has been fetal rhesus genotyping in mothers who are rhesus-D (RhD) negative. Analysis of cffDNA in the plasma of RhD-negative pregnant women who have a past history of haemolytic disease of the newborn or have elevated levels of anti-D antibodies has been used clinically to determine the fetal RhD status for almost a decade, and fetal RhD typing of all RhD-negative pregnant women has the potential to become routine clinical practice in the next few years. In addition, complications such as pre-eclampsia, preterm labour and fetal restricted growth, have been shown to be associated with increased concentrations of cell-free fetal DNA and, in the future, at-risk pregnancies may be identified through NIPD.

In this review, we focus on the use of NIPD for genetic and chromosomal disorders. Current research, clinical applications and an overview of research into ethical concerns, social issues and stakeholder viewpoints are described.

Fetal sex determination

Fetal sex determination using cffDNA is a well-established prenatal test that can be carried out reliably from 7 weeks gestation using real-time quantitative PCR (RT-qPCR) to identify the presence or absence of Y chromosome-specific sequences in the maternal plasma. The most common clinical indication for early fetal sex determination is for carriers of X-linked genetic disorders, such as Duchenne muscular dystrophy (DMD) or adrenoleucodystrophy, where male-bearing pregnancies are primarily at risk. For women who are carriers of X-linked conditions, early fetal sex determination using cffDNA can inform the uptake of invasive testing. If a male fetus is identified, the woman can go on to definitive diagnosis of the condition after CVS from 11 weeks. If a female fetus is identified, an invasive test is not needed, and the associated miscarriage risk can be avoided.

Another key clinical indication for early fetal sex determination is for pregnancies at risk of conditions associated with ambiguous development of the external genitalia, such as congenital adrenal hyperplasia (CAH). Female fetuses with CAH require early maternal treatment with dexamethasone to reduce the degree of virilisation. Male fetuses can have CAH, but they are not at risk of genital virilisation. Consequently, if gender can be determined early in pregnancy, women with female fetuses can continue prenatal steroid treatment and go on to definitive genetic diagnosis if requested, whereas those with male fetuses can discontinue the unnecessary steroid treatment and avoid invasive testing.

Fetal sex determination using cffDNA can also be useful in the clarification of some fetal ultrasound findings. These include confirmation of genetic sex if abnormal genitalia are identified and additional information for diagnosing genetic conditions where genital ambiguity or sex reversal is a feature of the condition.

Fetal sex determination can be carried out using ultrasonography, which is highly accurate in the second and third trimesters. It can only be carried out reliably earlier in specialist centres and not before 12 weeks, as the morphological appearance of the male and female external genitalia is similar before 12 weeks. As CVS can be offered at 11 weeks, sex determination at 12 weeks and beyond is too late for those women who would like to have early definitive genetic diagnosis. In addition, where there is a risk of genital ambiguity, ultrasound cannot be used reliably at any gestation. Fetal sex determination using cffDNA is, therefore, a useful alternative with clear clinical benefits.

Fetal sex determination

Fetal sex determination using cffDNA is a well-established prenatal test that can be carried out reliably from 7 weeks gestation using real-time quantitative PCR (RT-qPCR) to identify the presence or absence of Y chromosome-specific sequences in the maternal plasma. The most common clinical indication for early fetal sex determination is for carriers of X-linked genetic disorders, such as Duchenne muscular dystrophy (DMD) or adrenoleucodystrophy, where male-bearing pregnancies are primarily at risk. For women who are carriers of X-linked conditions, early fetal sex determination using cffDNA can inform the uptake of invasive testing. If a male fetus is identified, the woman can go on to definitive diagnosis of the condition after CVS from 11 weeks. If a female fetus is identified, an invasive test is not needed, and the associated miscarriage risk can be avoided.

Another key clinical indication for early fetal sex determination is for pregnancies at risk of conditions associated with ambiguous development of the external genitalia, such as congenital adrenal hyperplasia (CAH). Female fetuses with CAH require early maternal treatment with dexamethasone to reduce the degree of virilisation. Male fetuses can have CAH, but they are not at risk of genital virilisation. Consequently, if gender can be determined early in pregnancy, women with female fetuses can continue prenatal steroid treatment and go on to definitive genetic diagnosis if requested, whereas those with male fetuses can discontinue the unnecessary steroid treatment and avoid invasive testing.

Fetal sex determination using cffDNA can also be useful in the clarification of some fetal ultrasound findings. These include confirmation of genetic sex if abnormal genitalia are identified and additional information for diagnosing genetic conditions where genital ambiguity or sex reversal is a feature of the condition.

Fetal sex determination can be carried out using ultrasonography, which is highly accurate in the second and third trimesters. It can only be carried out reliably earlier in specialist centres and not before 12 weeks, as the morphological appearance of the male and female external genitalia is similar before 12 weeks. As CVS can be offered at 11 weeks, sex determination at 12 weeks and beyond is too late for those women who would like to have early definitive genetic diagnosis. In addition, where there is a risk of genital ambiguity, ultrasound cannot be used reliably at any gestation. Fetal sex determination using cffDNA is, therefore, a useful alternative with clear clinical benefits.

Single-gene disorders

Prenatal diagnosis has seen a rapid increase in the number of disorders that can be tested for as a result of technological advances and research into the genetics underlying many conditions. In the 2009–2010 audit by the UK Clinical Molecular Genetics Society (CMGS), 1744 prenatal diagnoses were carried out in the UK for over 100 different monogenic diseases. Initial reports of NIPD for single-gene disorders were confined to the detection of autosomal dominant conditions that are paternally inherited or arise as a result of a de novo mutation; increasingly, however, new technologies and new approaches are also allowing us to apply NIPD to recessive conditions ( Table 1 ). The first use of cffDNA for detecting a single-gene disorder was for autosomal dominant disorder myotonic dystrophy. In this study, the paternally inherited CTG trinucleotide expansion in the DMPK gene was detected using PCR followed by gel electrophoresis. Non-invasive prenatal diagnosis has also been possible for other autosomal dominant disorders, with the successful diagnosis of cases of paternally inherited early onset primary dystonia and Huntington disease.

Several studies have been published on the use of NIPD for achondroplasia. This condition is inherited in an autosomal dominant fashion, with most cases arising as a de novo mutation in the fibroblast growth factor receptor 3 gene. The mutation creates a restriction site, which can be detected using PCR followed by restriction fragment length polymorphism. Size fractionation followed by matrix-assisted laser desorption and ionisation time-of-flight mass spectrometry (MALDI-TOF MS) and single allele base extension reaction (SABER) was also successfully used for detecting fetuses carrying achondroplasia mutations. Several other conditions that arise as a result of de novo mutations have also proven amenable to non-invasive testing using PCR and restriction fragment length polymorphism, including Apert syndrome, thanatophoric dysplasia and Crouzon syndrome ( Table 1 ).

Non-invasive prenatal diagnosis for autosomal recessive or maternally transmitted autosomal dominant disorders is more complicated, owing to the difficulty of distinguishing between the maternal- and fetal-free DNA in maternal plasma. First attempts involved the exclusion of the paternal mutation in autosomal recessive conditions where the parents were carriers of different mutations ( Table 1 ). If a paternally identified allele can be detected, the fetus has a 50% risk of inheriting the disorder, and therefore invasive testing would be recommended. Alternatively, in the absence of a paternal allele, the fetus could at most be a carrier, and invasive testing would not be indicated. The first exclusion study was carried out for a couple who were carriers of CAH. Using polymorphic markers to show that the paternal allele had not been inherited, Chiu et al. were able to show that the fetus was not affected by CAH. Following this, several studies using a range of different technologies have been successful in detecting or excluding the paternal allele in recessive conditions where parents carry different mutations, including cystic fibrosis, propionic academia, Leber congenital amaurosis and beta-thalassaemia ( Table 1 ).

All of the above exclusion studies were based on determining the absence or presence of a paternal mutation. Notably, Papasavva et al. were able to extend this approach to allow the detection of paternal inheritance of beta-thalassaemia mutations, in which the parents carry either the same or different mutations. An arrayed primer extension microchip assay was used to detect single-nucleotide polymorphisms (SNPs) that could differentiate between the maternal and paternal alleles. Using this assay, the inherited paternal allele was correctly identified in six of seven families. A key limitation of this approach is the number of SNPs that are required for informativity, as only 68% of the 34 families studied had two or more informative SNPs. Also problematic is the necessity to genotype grandparents in order to link the parental alleles.

For recessive conditions where parents carry the same mutation, it is difficult to distinguish the maternally inherited fetal allele from the maternal alleles, and new approaches to NIPD are required. A quantitative approach known as relative mutation dosage (RMD) has been developed on the basis of digital PCR to allow non-invasive diagnosis where parents carry the same mutation. Digital PCR requires the dilution of template DNA to an average concentration of less than one molecule per well, and hundreds to thousands of replicates of a PCR reaction are analysed to allow high copy number counting and quantification of the initial molecules in the template DNA. Because most positive wells contain just one molecule, counting allows quantification of the initial molecules in the template DNA. If a woman is heterozygous and the fetus is also a heterozygote, it is expected that there will be an allelic balance between the wild type and normal alleles; if the fetus is homozygous for either the mutant or the wild type allele there will be an over-representation of one or the other. Sequential probability ratio testing (SPRT) is the statistical method used to determine whether the counts from the digital PCR are in allelic balance or whether there is an imbalance for the over-represented allele. This technique is dependent upon the knowledge of fractional fetal DNA concentration, which can be easily determined in male fetuses using sequences on the Y chromosome, but, in cases where there is a female fetus, are reliant on detection of paternally inherited SNPs.

RMD was first developed for beta-thalassaemia and haemoglobin E mutations. Five out of 10 cases tested were correctly classified (three beta-thalassaemia and two haemoglobin E cases); those that were not classified or were incorrectly classified contained less than 10% cffDNA. More recently, this approach has been used for sickle-cell anaemia. In this study, fractional fetal DNA concentrations for female fetuses were determined using informative insertion and deletion polymorphisms (indels). The sickle-cell genotype was correctly determined in 82% of male fetuses (37 out of 45) and 75% of female fetuses (15 out of 20) where the fractional fetal DNA concentration was greater than 7%. RMD has also been used to determine whether a fetus has inherited mutations for the X-linked disorder haemophilia. The principle of RMD in the case of X-linked disorders differs from that used in the case of the HBB gene mutations in that there are only two possible genotypes: mutant or wild-type. This test was successful for three carriers of haemophilia A and four carriers of haemophilia B. A similar principle to RMD was followed by Sirichotiyakul et al., who successfully used RT-qPCR for detecting homozygous alpha-thalassaemia-1 (Hb Bart’s) in 155 out of 158 pregnancies.

Massively parallel sequencing can also be used to identify single-gene disorders non-invasively. Lo et al. used genome-wide scanning to determine the inheritance of two different beta-thalassaemia alleles in a 12-week old fetus. Over 3.9 billion sequencing reads, providing 65-fold coverage of the genome, were aligned to the non-repeat-masked human genome. The father’s 4bp–CTTT mutation in codons 41/42 of HBB was easily detected. The mother carried the A>G mutation at position −28 of the HBB gene. With the use of SNP homozygous in the father and heterozygous in the mother located near to the mutation, a haplotype could be created, and relative haplotype dosage was used to determine whether there was an increased representation of the haplotype containing the mutation or the haplotype on the unaffected allele. It was found that the fetus was a carrier, having inherited the mutant allele from the father and the normal allele from the mother.