Noninvasive Prenatal Diagnosis for Single-Gene Disorders

Introduction

The identification of cell-free fetal DNA (cffDNA) circulating in maternal plasma, which is present from early in pregnancy, has paved the way for the development of noninvasive prenatal diagnosis (NIPD), avoiding the small miscarriage risk associated with invasive prenatal diagnosis. The cffDNA is rapidly cleared from the maternal circulation after delivery, and because the whole fetal genome is represented, it offers scope for comprehensive prenatal diagnosis. NIPD for monogenic disorders is diagnostic, and results do not require confirmation on chorionic villi or amniocytes obtained by invasive testing as diagnosis is undertaken in pregnancies known to be at increased risk because of family history or sonographic findings and is targeted at a specific gene or panel of genes. This contrasts with noninvasive prenatal testing or screening (NIPT or NIPS) for aneuploidy, which is considered a very sensitive screening test requiring an invasive test to confirm a positive result ( Table 22.1 ). This is because cffDNA emanates from the placenta, and testing maternal plasma can detect confined placental mosaicism, and because the majority of cell-free DNA (cfDNA) in maternal plasma comes from the mother herself, NIPT can reveal unexpected maternal chromosomal rearrangements. Although maternal mosaicism can complicate NIPD for monogenic disorders, this can be taken into account by testing maternal genomic DNA (gDNA) in parallel with the cfDNA, but as far as we are aware, placental mosaicism for cell lines containing mutations found in single-gene disorders has not been reported.

Differences between NIPD and NIPT also apply to the timing of testing, largely because any cffDNA-based test is affected by the fetal fraction. The fetal fraction increases as pregnancy progresses, usually reaching a level that can be used for testing in straightforward applications such as the detection or exclusion of an allele not present in the maternal genome, for example, fetal sex determination, by 7 weeks’ gestation. However, for more complex applications that require accurate quantification (e.g., aneuploidy screening), accurate testing usually requires higher levels of cffDNA, which are attained around 10 to 12 weeks in most pregnancies. It is therefore critical that the gestational age is determined using ultrasonography before embarking on cfDNA testing. Furthermore, ultrasonography should be used to look for evidence of multiple pregnancies, which may complicate results. If an empty sac or ‘vanishing twin’ is identified, this may lead to false-positive results because the placenta continues to shed cffDNA in the absence of a fetal pole. NIPD is possible in multiple pregnancies, but in the absence of any abnormal ultrasound findings, it is not possible to tell which fetuses are affected, and invasive testing will be needed for definitive diagnosis. However, in dizygotic twins, it may be possible to analyse single nucleotide polymorphisms (SNPs) to measure variations in the fetal fraction between genomic regions to determine zygosity and analyse the fetal fragments for each fetus.

It is not possible to separate maternal and fetal cfDNA and, as described earlier, any test based on analysing cfDNA in maternal plasma will analyse both the maternal and fetal fractions, the maternal fraction being the most abundant. Consequently, the early clinical applications of NIPD focussed on the detection of paternal alleles or those arising de novo that were therefore not present in the mother (e.g., fetal sex determination, fetal RHD typing in RhD-negative mothers, paternally inherited single-gene disorders , or single-gene disorders arising de novo , such as achondroplasia ). NIPD can also be used in pregnancies at risk for autosomal recessive conditions if the parents are heterozygous for different mutations. In these cases, termed ‘paternal exclusion testing’, if the paternal allele is identified in maternal plasma, an invasive test is still required for definitive prenatal diagnosis to determine inheritance of the maternal allele and see whether the fetus is affected or not.

Many different laboratory approaches have been reported for NIPD, largely in research settings ( Table 22.2 ). New technologies, such as massively parallel sequencing (MPS), that allow accurate quantification of specific sequences has meant that tests are now being developed that take into account the presence of the maternal cfDNA, thus allowing diagnosis of autosomal recessive X-linked conditions.

Confirming the presence of cffDNA is critical to delivering a reliable NIPD result. In situations in which testing requires determination of the presence or absence of an allele not present in the mother, detection allows definitive diagnosis, but if absent, although it may reflect a true negative. it may also result from absence or very low levels of cffDNA with consequent failure to amplify the target sequence. Inclusion of a fetal specific marker in an NIPD assay is required to confirm the presence of cffDNA and allows a negative result to be definitive. In male-bearing pregnancies, Y-chromosome sequences can be used, but there is no straightforward approach for female-bearing pregnancies. Paternally inherited SNPs, short tandem repeats or small insertion or deletions (indel markers) that are either absent or different in the mother could be used but add to the cost of the test as maternal and paternal DNA require analysis, and even a large panel of SNPs may not always be informative. Alternative approaches to confirm the presence of cffDNA take advantage of epigenetic differences between the fetus and the mother which arise as a result of the differential expression of maternally and paternally inherited alleles. The Ras-association domain family member 1, transcript variant A (RASSF1A) which is hypomethylated in the mother and hypermethylated in the fetus, is the allele most frequently used, but the assay is complex and not ideally suited for use in a busy service laboratory. Developing robust assays to measure fetal fraction may be particularly important in women with high body mass indexes because cffDNA levels tend to be lower in obese women, probably because of increase shedding of maternal cfDNA from adipose tissue. In addition, strategies that optimise sample collection and processing (e.g., separation of the plasma within a few hours of blood draw or the use of cell stabilising tubes) are important to optimise cffDNA levels and minimise inconclusive or failed cases.

Although the commercial sector has driven the development of NIPT for aneuploidy, which is now available worldwide, development of NIPD for monogenic disorders has been led by the academic sector, presumably because of the rarity of most disorders together with the cost and time required to develop individual tests do not make NIPD an attractive commercial prospect. In this chapter, we discuss the development of a clinical service focused on NIPD for monogenic disorders, which includes NIPD for fetal sex determination to inform the need for definitive diagnosis in pregnancies at risk for sex-linked disorders. Implementation of any new test into clinical practice requires robust validation, both in the laboratory but also in terms of clinical validity and economic viability, and must meet service users’ needs. Development of an NIPD clinical service in the United Kingdom has been based on the UK Genetic Testing Network (UKGTN) standards, as required by the UK National Health Service (NHS), to commission and thus fully fund any molecular genetic test for a monogenic disorder ( Table 22.3 ). Here we discuss current applications of NIPD (in the United Kingdom and elsewhere) and provide an overview of clinical utility, economic and social issues and stakeholder viewpoints, all of which are required for UKGTN approval.

Fetal Sex Determination

Fetal sex determination was one of the first applications of cfDNA to be used in clinical practice with a variety of methods being applied across centres in Europe from the late 1990s. This test, now well-established in several countries, can be performed reliably from 7 weeks’ gestation to guide the management of pregnancies at risk for X-linked conditions, congenital adrenal hyperplasia (CAH) and in cases in which there is genital ambiguity. In pregnancies at risk for X-linked conditions, such as Duchenne muscular dystrophy (DMD) or adrenoleukodystrophy, identification of a male fetus using NIPD indicates the need for an invasive test to determine inheritance of the maternal mutant allele and subsequent definitive diagnosis, but this is not required if the fetus is found to be female. In pregnancies at risk for CAH, early maternal treatment with dexamethasone, although controversial, can reduce the degree of virilisation of external genitalia in affected female fetuses. Because this is a recessive condition, males can have CAH, but they are not at risk for genital virilisation. Early determination of fetal sex can be used to guide maternal dexamethasone therapy and direct the need for invasive testing for definitive diagnosis, which, as will be discussed later, is now possible in many families using NIPD for CAH. NIPD for fetal sex determination is also helpful for the clarification of fetal ultrasound findings such as genital ambiguity or to provide additional information for diagnosing genetic conditions such as campomelic dysplasia or Smith-Lemni-Opitz syndrome in which genital ambiguity or sex reversal is a feature of the condition.

Noninvasive prenatal diagnosis for fetal sex determination is relatively straightforward as it is performed by identifying the presence or absence of Y-chromosome sequences in maternal plasma. Detection of the Y-chromosome sequence indicates that the fetus is male. If the Y-chromosome sequence is not detected, it is assumed that the fetus is female. A systematic review and meta-analysis that included 57 studies with 3524 male- and 3017 female-bearing pregnancies tested over a wide range of gestations and with a variety of laboratory methodologies, largely performed on a research basis, indicated that NIPD for fetal sex determination is reliable after 7 weeks in pregnancy. A large national audit of NIPD for sex determination performed in UK public-sector laboratories using the most common technique, real-time quantitative polymerase chain reaction (RT-qPCR) of Y-chromosome targets (the single copy SRY gene or the multicopy DYS14 sequence located within the TSPY gene) showed a sensitivity of 99.5% (95% confidence interval, 98.2–99.9%) . A failure rate of individual tests of approximately 4% to 5% has been seen in many studies, a rate which some suggest may be reduced by using digital PCR or by using assays that target both the SRY and DYS14 genes. However, these latter studies only report small numbers, and further evaluation of these methods is required. Test failures and false-negative results caused by technical issues are likely to occur with any method and, as discussed earlier, the concurrent use of a universal cffDNA marker to confirm the presence of cffDNA will minimise false-negative results by identifying pregnancies in which very low fetal fraction results in failure to amplify the Y-chromosome sequences rather than absence of Y-markers because the fetus is female.

Because NIPD for fetal sex determination has been in clinical practice for several years, it has been possible to demonstrate clinical utility. In the UK audit, the rates of invasive testing in women at risk for very severe X-linked conditions (excluding haemophilia) were only 43%, and 38% in those at risk for CAH. However, the same audit failed to demonstrate clinical utility in pregnancies at risk for haemophilia, in which invasive testing was infrequent (16.9% compared with 43%) for other severe X-linked disorders. This is largely because in the majority of these pregnancies, knowledge of fetal sex is required to direct management of labour rather than inform parental decisions regarding invasive testing for definitive diagnosis and possible termination of pregnancy. Sex determination in these pregnancies can readily be performed using routine midtrimester ultrasonography; thus clinical utility could not be demonstrated as an alternative, more cost-effective approach is available. Clinical utility has been further demonstrated in a French multicentre study of fetal sex determination in 258 pregnancies at risk for CAH (134 male and 124 female fetuses), which showed that prenatal maternal steroid treatment had been avoided in 68% of male-bearing pregnancies.

Health economic aspects must be considered when implementing new tests, particularly in publicly funded health care systems. A detailed health economic analysis in the United Kingdom demonstrated that NIPD for sex determination was cost effective for severe X-linked conditions such as DMD because the costs of NIPD are offset by the smaller proportion of women who require invasive testing, reducing costs by an average of £87 (-£303 to £131) per pregnancy. Cost savings were slightly greater, a reduction of £193 (-£301 to -£84), in pregnancies at risk for CAH because of the additional savings related to maternal dexamethasone treatment. It must be noted that these costs relate to the United Kingdom, where the cost of labour is high and laboratory consumables may be lower than in other countries. Thus the labour costs related to invasive testing are relatively high. In health economies with a different balance of labour: consumable costs, benefits and costs may be different. UK service users welcome the availability of the NIPD and report practical benefits from safe early testing as well as psychological benefits, such as a feeling of having control over the pregnancy and peace of mind. Health professionals also welcome the advent of safer and earlier fetal sex determination with NIPD.

The UKGTN gave approval for NIPD for fetal sex determination in the UK NHS, but their recommendation, and thus subsequent commissioning and public-sector funding, was limited to NIPD for serious X-linked disorders (excluding haemophilia) and CAH. Since gaining UKGTN approval, there has been a significant increase in the use of NIPD for fetal sex determination, with it now being one of the most frequently performed prenatal molecular tests in pregnancies at risk for monogenic disorders. However, there remains significant discordance in practice in pregnancies at risk for haemophilia with some services routinely offering NIPD to all carriers of haemophilia or to all carriers of severe haemophilia but others primarily offering NIPD as a first step to invasive testing. Development of definitive NIPD for haemophilia may resolve this inequality of access as this would enable safe optimisation of pregnancy management for all pregnancies.