Abstract
Noninvasive prenatal diagnosis (NIPD) for monogenic disorders is relevant to couples that have a high risk of an affected pregnancy due to either a family history of a disorder or because of having been identified as carriers through a screening program. Alternatively, pregnancies where there is no known risk of a disorder that present with scan findings that have a differential diagnosis of a monogenic disorder may in certain circumstances apply for NIPD. As technology improves and costs decrease, testing is now becoming available for some disorders, although access is not yet equitable. This chapter explains the current state of play with NIPD, how the different testing methodologies such as next-generation sequencing and digital PCR are being applied, and the challenges associated with delivering NIPD for monogenic disorders.
Keywords
Monogenic disorder, Mendelian inheritance, Ultrasound scan, Noninvasive prenatal diagnosis, Next-generation sequencing, Digital PCR, Fetal fraction
Introduction
The groundbreaking discovery of the presence of cell-free DNA in maternal plasma was made in 1997 by Dennis Lo . In the years since that, gradual technological development has meant that we can now use a maternal plasma sample as a source of material for prenatal testing for many different applications. One obvious potential application for this is noninvasive prenatal diagnosis (NIPD) of monogenic disorders—so-called Mendelian single gene disorders; however, the clinical application of this testing has lagged behind in comparison to other areas such as cfDNA-based noninvasive prenatal testing, commonly called NIPT, for aneuploidy.
Initial testing for this group of patients was limited to sex-linked and sex-limited disorders, and the use of fetal sexing by detection of Y chromosome markers. This testing was developed and implemented first as it was technologically much simpler to detect a marker that was not present in the mother’s blood. Testing for many monogenic disorders is now becoming possible and available to patients, and this is described in more detail in this chapter.
It is also worth noting at this point that circulating fetal cells have also been explored as a possible source of testing material. They offer certain advantages and disadvantages over analysis of cell-free DNA and are dealt with in more detail in Chapter 19 . At the time of writing this chapter, this is not considered a clinically practical approach to NIPD of monogenic disorders and so is not dealt with further here.
Clinical Applications of NIPD for Monogenic Disorders
Conditions With a Family History
The majority of NIPD assays reported to date have been for conditions where there is a known family history of a disorder, typically with a high recurrence risk in each pregnancy. For example in the case of an autosomal recessive disorder where both parents are carriers there is a one in four risk of an affected child with each subsequent pregnancy. The methods used for variant detection can vary depending upon the mode of inheritance. In the UK, next-generation sequencing (NGS)-based approaches to NIPD are already in routine use, allowing around 30% of all molecular prenatal diagnosis to be performed noninvasively .
Autosomal dominant disorders
NIPD for autosomal dominant conditions that are paternally inherited or have arisen de novo is the least technically challenging scenario; the causative mutation is not present in the maternal DNA, but if present in fetal DNA, would be detectable at low levels in maternal plasma. Several approaches have been taken to develop assays for a variety of autosomal dominant disorders and it should be noted that the same technologies described in the examples later could be modified for a range of conditions. For further technical details of these methods, refer to section “Technical Approaches to NIPD.”
Droplet digital polymerase chain reaction assays (ddPCR) have been described to detect the presence of paternal KCNJ11 and FGFR3 mutations, which cause neonatal diabetes and achondroplasia, respectively. Using this technique, the authors were able to detect paternal alleles at a concentration of < 1%. Orhant et al. demonstrated that it is also possible to detect the same FGFR3 variant through a minisequencing assay, but this technique was only able to detect the disease-causing allele if it is present at > 3% in the maternal plasma.
These genotyping assays can only detect the variant for which they are specifically designed. Although this approach may be appropriate for common mutations or to provide bespoke assays for families with a history of a rare disease, it may not be the most cost-effective way to provide NIPD for common disorders that have a spectrum of associated disease-causing mutations. Other publications have described an alternative approach by designing a single assay that could be used for many families, such as an NGS panel that covers 29 known FGFR3 disease-causing mutations associated with both achondroplasia and thanatophoric dysplasia and a targeted capture-sequencing method that is capable of screening the coding regions of 16 genes known to be associated with skeletal dysplasias .
Monogenic disorders which are caused by expansion of large polymorphic tracts, such as the trinucleotide repeat disorders, pose a particular challenge for NIPD. Van den Oever et al. describe a fragment length analysis technique to determine paternally inherited CAG repeat length in pregnancies at risk of Huntington disease . The authors were able to detect disease causing and intermediate CAG repeats in all patients tested, but could only detect transmission of a normal paternal repeat in 50% of cases as it could not always be distinguished from the maternal allele. In addition, the authors advised caution when testing for very large repeats, as these may not be detectable due to the fragmented nature of cfDNA.
Autosomal recessive disorders
For autosomal recessive conditions, familial mutations may be the same in both parents (homozygous proband), or both parents may carry different mutations in the same gene (compound heterozygous proband). These situations require different approaches to NIPD. An exception to this is when haplotype-based assays are used these methods determine the inheritance of parental mutations in a linkage-type manner and as they do not directly detect the mutations, can be used for any scenario. To date, haplotype-based assays for use in pregnancies at risk of GJB2 -associated hearing loss , congenital adrenal hyperplasia (CAH) , spinal muscular atrophy (SMA) , alpha- and beta-thalassemia , and Gaucher disease have been published. However, these assays are limited to use for couples who have already had an affected child, as DNA from a proband is required to “phase” the analysis. This obstacle has been overcome by Vermeulen et al. , who combined targeted haplotyping of both parents using TLA (targeted locus amplification) with targeted deep sequencing of cfDNA extracted during pregnancy. Using this method, the authors were able to predict the inherited allele with > 98% accuracy in 18 pregnancies at risk of cystic fibrosis (CF), CAH, or beta-thalassemia .
Detection of paternally inherited mutations
In a family where each parent is a carrier of a different mutation in the same gene, the most straightforward approach to NIPD is exclusion of a paternal mutation. If the paternal mutation is excluded, no further testing is indicated as the fetus may be a carrier of the condition but will not be affected. If the paternal mutation is detected by NIPD, further testing by traditional invasive methods is required to look for the presence of the maternal mutation.
As with paternally inherited dominant mutations, this is the least challenging technical scenario for autosomal recessive conditions, as if the mutation is detected in the maternal plasma, it must be fetal in origin. As such, the same techniques that have been used in autosomal dominant conditions can also been used for exclusion of paternal mutations in autosomal recessive conditions.
Debrand et al. describe a ddPCR assay that can detect the common CF causing mutation, DF508, at a level as low as 1.3% of total DNA in pregnancies where the father is a carrier of this mutation . Further demonstrating the range of techniques that can be utilized for NIPD for the same conditions, Galbiati et al. developed a COLD-PCR assay that could detect both the DF508 mutation, as well as an additional three common CF-causing variants . The authors reported that the mutations could not be detected by traditional PCR methods, but that using the COLD-PCR technique, they were able to detect mutations down to a level of < 1%. As mentioned previously, these techniques can be applied to many different conditions, demonstrated by the fact that the same COLD-PCR methodology used for NIPD of CF was also validated for seven common beta-thalassemia-causing mutations . In order to show that different methodologies can be used to confirm NIPD results, the authors also validated a microarray-based methodology for NIPD of CF and beta-thalassemia, which was in complete concordance with the COLD-PCR results . This range of techniques that have been reported in the literature for the same conditions demonstrates the versatility of NIPD.
Detection of maternally inherited mutations
At present, assays to directly detect maternally inherited mutations, either in autosomal recessive or autosomal dominant conditions are not in routine clinical use because of the technical difficulties involved in differentiating between the maternal and fetal DNA in the maternal plasma sample. However, improvements in sensitivity of available techniques mean that advances are being made in this area.
A proof-of-principle study used ddPCR to perform precise allele quantification in maternal plasma and calculate relative dosage (relative mutation dosage, RMD) . This demonstrates the potential of ddPCR for NIPD studies of fetal mutations independently of their parental origin, with 100% accuracy for the detection of paternal alleles and 96% accuracy for the detection of maternal alleles. By using single-nucleotide polymorphisms (SNPs) rather than disease-causing variants, the authors were able to demonstrate the utility of this technique for NIPD analysis of fetal mutations for any inheritance pattern.
An alternative approach was taken by Lv et al., who developed and validated an assay termed circulating single-molecule amplification and resequencing technology (cSMART). In their initial publication, this technique was successfully applied to pregnancies at risk of Wilson disease where inheritance of different maternal and paternal mutant alleles was determined. Further validation work was performed to determine the inheritance of SNP genotypes to demonstrate that this assay can also be used for pregnancies when the parents are both carriers of the same mutation. The authors have since further developed this methodology to determine fetal genotypes in pregnancies where one or both partners were known carriers of an autosomal recessive nonsyndromic hearing loss (ARNSHL) disease causing mutation in either the GJB2 , GJB3 , and SLC26A4 gene .
X-linked and sex-limited disorders
NIPD to determine fetal sex can be offered as a frontline test in pregnancies at risk of serious X-linked conditions, such as Duchenne muscular dystrophy. If this analysis shows the presence of a male fetus, further testing for a definitive diagnosis of the condition in question can be offered. NIPD to directly detect a maternally inherited mutation is technically challenging due to the high background of maternal cfDNA in maternal plasma (see previously), however, haplotype-based approaches have been described for NIPD of Duchenne and Becker muscular dystrophies and hemophilia .
Congenital adrenal hyperplasia (CAH) is a sex-limited disorder where NIPD is used to guide treatment in pregnancy. In affected pregnancies with a female fetus there is a risk of in utero virilization. Antenatal dexamethasone administered very early in pregnancy may prevent or decrease virilization; however, it is important to target its use to at-risk pregnancies. NIPD to determine fetal sex can be used as a first step to target female pregnancies only and this is available from 7 weeks’ gestation. Reports in the literature also describe methods that could be used to determine whether the fetus is affected with CAH, that is, carries the familial mutations or high-risk parental haplotype .
PGD confirmation
It is recommended that prenatal diagnosis is performed to confirm preimplantation genetic diagnosis (PGD), yet many patients are reluctant to undergo testing, due to the risks associated with traditional invasive procedures. In 2010 the ESHRE (European Society for Human Reproduction and Embryology) best practice guidelines included NIPD as an appropriate method for confirming PGD pregnancies. One group has described a customized care trajectory for a family with a history of GJB2-associated hearing impairment that combined PGD with cfDNA-based noninvasive prenatal testing (NIPT) for fetal aneuploidy and a custom-designed NIPD assay, with the results confirmed by amniocentesis . Others have described a combination of haplotype analysis and direct analysis of a paternal mutation to confirm the results of PGD in a pregnancy at risk of Marfan syndrome . Although these management strategies would be prohibitively expensive for standard clinical practice, as different NIPD methodologies become part of routine prenatal diagnosis, these can become incorporated into the PGD care pathway.
For pregnancies at risk of an X-linked disorder where a specific NIPD assay is not available for the condition in question, noninvasive methods can be used to confirm fetal sex following transfer of a female embryo .
Pregnancies With Ultrasound Scan Findings
Another clinical application of NIPD for monogenic disorders is where there is no known family history of a disorder, but when a pregnancy presents with ultrasound scan findings that point toward the possibility of a monogenic disorder. A good example of this is skeletal dysplasia where NIPD provides a useful aid to clinical management, as it allows a definitive diagnosis, differentiation between lethal and nonlethal forms of the disease, and the option of a surgical termination as a postmortem is not required.
Multiple noninvasive methods for the detection of individual common mutations within the FGFR3 gene that cause either achondroplasia or thanatophoric dysplasia have been described; however, skeletal dysplasias are a group of heterogenous diseases and the detection rate is dependent on the accuracy of ultrasound diagnosis. With this in mind, Chitty et al. describe an NGS panel that covers 29 known FGFR3 disease-causing mutations that are associated with both achondroplasia and thanatophoric dysplasia and Dan et al. have developed a targeted capture-sequencing method for the detection of de novo mutations in 16 lethal dysplasia genes .
These same assays allow testing to be performed very early in pregnancy in cases where there is a history of a skeletal dysplasia, either where the mutation is de novo (i.e., not detected in either parent) or paternally inherited.
NIPD where there are abnormalities on ultrasound scan is likely to increase in the future as technology becomes cheaper and it becomes feasible to screen for mutations in panels of genes. One could envisage testing panels of genes associated with particular ultrasound scan features such as “cardiac abnormality,” “brain abnormality,” or even ultimately the whole exome or genome.
Pregnancies at Risk Following Parental Carrier Screening
An increasing number of couples are finding that even in the absence of a family history, they are at risk of having a pregnancy affected with a monogenic disorder. This is due to the development of carrier screening programs and the introduction of commercially available direct-to-consumer carrier testing. With the exception of the haplotype-based approaches, where a proband is required to “phase” analysis, the techniques described previously could also be used to provide NIPD for these pregnancies.
Technical Approaches to NIPD
As already described, many different technical approaches have been employed for the NIPD of monogenic disorders. These are summarized in Table 1 .
| Technique | Disorders | References |
|---|---|---|
| Amplicon-based NGS | ACH | |
| cSMART | Wilson disease | |
| Hearing loss | ||
| Capture-based NGS | Skeletal dysplasias | |
| ddPCR | CF | |
| Neonatal diabetes | ||
| ACH | ||
| Multiple SNPs | ||
| Hemophilia | ||
| COLD-PCR | CF and beta-thalassemia | |
| Microarray | CF and beta-thalassemia | |
| Minisequencing | ACH | |
| Fragment length analysis | HD | |
| CF | ||
| Relative mutation dosage (RMD) | Beta-thalassemia | |
| Haplotype based NGS (RHDO and MG-NIPD) | CF | |
| DMD | ||
| Hearing Loss | ||
| CAH | ||
| Gaucher disease | ||
| SMA | ||
| Alpha- and beta-thalassemia | ||
| Hemophilia |
Detection of Paternal or De Novo Mutations
As described in section “Clinical Applications of NIPD for Monogenic Disorders,” the least challenging NIPD test is determining the presence or absence of a pathogenic variant in maternal plasma that is not present in the maternal genome such as paternally inherited or de novo mutations in cfDNA. As a result, several methods have been developed for this type of assay. The most common of which are described as follows.
Bespoke/targeted NGS
Along with many other areas of genetics, NGS has revolutionized NIPD of monogenic disorders, mainly due to the huge increase in sensitivity NGS provides. Bespoke or targeted NGS involves the selection of a specific chromosomal region or regions of interest for sequencing. This selection of region(s) for sequencing, also known as enrichment, allows for high read depth of a single or series of mutations and can be done using either amplicon- or capture-based methodology.
The term “amplicon-based” NGS is used to describe an NGS protocol where sample preparation and enrichment is solely performed using a polymerase chain reaction (PCR) using PCR primers that are designed to amplify the specific region(s) of interest. Amplicon-based enrichment is relatively straightforward compared to other NGS sample preparation methods (see Fig. 1 ). However, this style of enrichment is limited by the same issues that face conventional PCR and PCR bias can potentially be an issue. These can especially be problematic when designing an assay to analyze multiple variants in a single test.
“Capture-based” NGS is where enrichment is performed via the hybridization of capture probes, designed against the region(s) of interest, to template DNA. This often provides a more uniform enrichment than amplicon-based methods and therefore the capacity for analyzing multiple variants/regions at once is increased. However unlike amplicon-based assays, a sample preparation stage is necessary prior to capture, which often requires multiple steps (see Fig. 1 ). Furthermore, if gDNA sequencing is required alongside cfDNA, then the gDNA must be fragmented prior to the sample preparation stage. This is not necessary when performing amplicon-based NGS as the PCR can be designed so that the PCR product is the correct size for sequencing. The additional sample preparation requirements in capture-based NGS and the use of probes mean it is more time consuming and expensive than amplicon-based NGS.
cSMART
As mentioned previously, an issue with amplicon-based NGS using conventional forward and reverse primers is the possibility of PCR bias. One cause of PCR bias is the preferential PCR amplification of smaller fragments which can skew allelic ratios.
Using a combination of amplicon circularization and the barcoding of individual library molecules, cSMART removes any size bias from individual fragments . The workflow starts with a sample preparation protocol similar to that used in capture-based NGS. Individual library molecules are circularized and uniquely barcoded using a bridging oligonucleotide, and a second PCR is then performed using back to back inverse primers located adjacent to the mutation site(s) of interest. This avoids any size bias from the individual fragments. Duplicate molecules generated by PCR can also be identified and removed due to their identical barcodes and start/stop position.
Droplet digital PCR
As an alternative approach to NGS, several studies have utilized ddPCR for the NIPD of monogenic disorders . ddPCR is a modified form of real-time PCR, which utilizes a water–oil emulsion droplet system. Each water droplet separates template DNA molecules into individual PCR reactions. As a result, thousands of independent amplification events are able to take place within a single sample. These amplification events are then analyzed individually on a droplet reader which counts whether droplets are positive or negative for the mutation of interest ( Fig. 2 ). ddPCR allows the absolute quantification of target DNA copies without the need for standard curves, allowing for significant enhanced sensitivity of standard real-time PCR. One drawback of real time PCR and ddPCR in comparison to NGS is that individual assays are required for each mutation assayed, and multiplexing can be difficult. Therefore for individual mutations ddPCR is a cost-effective technique, but NGS should be considered where panels of mutations are being tested.
