Use of Type 5 Single Nucleotide Polymorphisms Allows Noninvasive Prenatal Diagnosis for Consanguineous Families

Noninvasive prenatal tests based on the presence of fetal cell-free DNA (cfDNA) in maternal plasma have only recently been developed. The tests can be divided into 2 groups, the most well-known being the detection of large fetal chromosomal anomalies, such as trisomies and structural abnormalities. This test is often referred to as noninvasive prenatal testing (NIPT), or noninvasive prenatal screening (NIPS). Although the latter name is more adequate as this test is mainly used in a screening setting, either as a first-tier test or contingent after combined testing, it is less commonly used. As the so-called fetal cfDNA is in fact derived from the placenta, and not from the fetus, confined placental mosaicism (CPM) can give false-positive NIPT results, necessitating confirmation of a high-risk result by a diagnostic invasive test. However, prenatal diagnostics based on cfDNA is possible in pregnancies where pathogenic variants are inherited from carrier parents and CPM is not a problem. This leads us to the second group of noninvasive prenatal diagnosis (NIPD), which is mainly used for monogenic disorders. In contrast to NIPT, NIPD is only offered in a few countries by a limited number of specialized laboratories. Performing NIPD is much more complex than NIPT, and the number of tests much smaller. Until now, NIPD could not be offered to consanguineous couples, as their shared genomes hampered reliable analysis.

This is going to change as, in this issue of Clinical Chemistry, authors from the 2 UK-based NIPD laboratories in London and Birmingham, which have been leading the field of NIPD, describe the adaptation and clinical validation of a method already proposed in 2010 to expand access to NIPD for monogenic conditions to consanguineous families (1). A comparable approach was published last year in Clinical Genetics by a group from Geneva, Switzerland (2). Methods used for NIPD for monogenic conditions vary depending on the mode of inheritance. A comprehensive overview of available methods for NIPD and on how and when to use them was recently published by authors from the London-based group (3). The main determinant is whether or not the mother carries the pathogenic variant. For most pregnancies the so-called fetal fraction (FF), or the proportion of fetal/placental cfDNA to maternal cfDNA, is approximately 5% to 10%. As a consequence, the presence of a fetal pathogenic variant needs to be demonstrated against a huge background of maternal cfDNA. In cases where the mother is a heterozygous carrier, slightly less than 50% of cfDNA tested will contain the pathogenic variant from the maternal cfDNA contribution.

If the mother is not a carrier, as is the case with de novo or paternally derived pathogenic variants, simple proof of absence or presence of the variant is sufficient for reliable diagnosis. In those cases, NIPD can be performed using direct methods detecting the pathogenic variant, such as digital PCR or targeted Next Generation Sequencing (NGS). If the mother carries the pathogenic variant a targeted approach can also be used, known as relative mutation dosage (RMD). RMD determines in which way the contribution of the wild-type allele (WT) and the mutated allele is influenced by the fetal genotype. As an example, in case of an autosomal recessive disorder with both parents carrying the same pathogenic variant, an increase in the quantity of the WT is indicative of the fetus being homozygous WT, a balance indicates the fetus is a carrier and a decrease of the WT indicates the fetus will be affected as the pathogenic variant is present on both alleles. Although this approach can be used, it has major drawbacks. Most importantly, even in cases with normal FF, the differences caused by the fetal contribution will be very small, the test less reliable, and the result often inconclusive. In cases with low FF, which are not rare, this will be even worse. The second drawback is that RMD is a variant-specific, and not a generic, method. As in monogenic conditions most families will have private pathogenic variants, this means that a specific RMD test needs to be set up for every pregnancy. For these reasons, most laboratories use relative haplotype dosage analysis (RHDO) for NIPD in cases where the mother is a carrier. RHDO was described as early as 2010 by the group of Dennis Lo in Hong Kong and does not assess the pathogenic variant directly but infers the genotype through haplotyping (4). They also developed a sequential probability ratio test (SPRT), which is now the most commonly used method to perform RHDO. The first step in RHDO is to determine the haplotype of the region surrounding the gene of interest for both parents and a proband (in most cases an affected proband will be used, but depending on the mode of inheritance other possibilities exist). During this step the haplotype of the high- and low-risk allele(s) will be determined and informative single-nucleotide polymorphisms (SNPs) will be called. During pregnancy, haplotyping of the fetal genome as derived from cfDNA will be done based on these informative SNPs. Five types of informative SNPs exist, but until now only 4 of them were used for NIPD. In type 1 SNPs, the mother is homozygous AA and the father homozygous BB. Although these SNPs are not informative for the risk allele, they are very useful during NIPD to determine FF. In type 2, both parents are homozygous for the same AA allele, and use of this type is limited to the determination of the technical quality of the test by looking at background noise. In type 3, the mother is homozygous AA or BB, and the father is heterozygous AB, and these SNPs can be used to determine the paternal haplotype. Type 4 is the other way around: the mother is heterozygous AB and the father is homozygous AA or BB. These are the SNPs that are used to determine the maternal haplotype of the fetus during RHDO. If a dosage imbalance between the A and B alleles is detected, the fetus will be homozygous, a balance between both alleles indicates the fetus will be heterozygous. The final group is called type 5, where both the mother and father are heterozygous AB. Until now this type has not been used, but its possible use for NIPD for monogenic disorders in consanguineous families or for genetic diseases with a strong founder effect was already suggested in 2010 (4).

The main problem with consanguineous parents is that they might share the same haplotype surrounding the gene containing the pathogenic variant. As a result, type 1, 3, and 4 SNPs are rare or absent, whereas there is an increase of type 2 and 5 SNPs. Hanson et al. adapted SPRT in such a way that type 5 SNP information could be incorporated, thereby allowing the determination of the fetal haplotype in these families (1). The idea behind it is comparable to how RMD works, but instead of interrogating one single variant, the signal of all available SNPs is combined into a haplotype block call. In case of a consanguineous family, both parents will be heterozygous Hap1-Hap2, with Hap1 carrying the pathogenic variant. When performing NIPD, an overrepresentation of Hap1 indicates the fetus is affected, an overrepresentation of Hap 2 indicates the fetus will be an unaffected non-carrier, and a balance between Hap 1 and Hap 2 indicates the fetus is an unaffected carrier. The incorporation of type 5 SNPs in SPRT not only allowed the correct identification of the fetal genotype, even at relatively low FF, it also correctly detected the presence of both maternally and paternally inherited meiotic recombination events. This is very important as missing those events may lead to a wrong outcome of the test.

Although this is definitely a major step forward in the clinical use of NIPD, we can expect more improvements in the near future. One of them will be the routine use of long-read sequencing to call the haplotypes without the need of a proband, as exemplified by a recent study on the use of targeted Nanopore sequencing to allow RHDO for β-thalassemia (5). The currently available long-read sequencing technologies have another advantage: they allow determination of the methylation status which might well play a major role in the further development of NIPD (6). Other more indirect methods have been proposed as well, such as targeted locus amplification (TLA)-based phasing of heterozygous variants (7). Although this method works, in my opinion straightforward sequencing (whether or not long read) will be preferred over adapted methods needing more time, technical skills, and equipment. In that view, a logical future step is to use a generic genome-wide approach with bioinformatic filtering of the locus of interest, rather than a locus specific approach. This was already shown to be feasible in 2012 by Stephen Quake's group from Stanford (8). They combined a genome-wide approach with data mining, using the haplotypes of the normal population in the 1000 Genome Project to impute the paternal haplotype. A combination of such a genome-wide approach with long-read sequencing, determination of methylation status, and imputation of missing regions might well be the way we will perform NIPD in the foreseeable future. However, some major hurdles need to be taken first, one of them being sequencing costs, another the need for public databases representing the entire human population, not only the Caucasian population. Until then, every step forward is welcome, one of which you can read about in this issue of Clinical Chemistry.

Nonstandard Abbreviations

cfDNA, cell-free DNA; NIPT, noninvasive prenatal testing; NIPD, noninvasive prenatal diagnosis; FF, fetal fraction; RMD, relative mutation dosage; WT, wild-type allele; RHDO, relative haplotype dosage analysis; SNP, single-nucleotide polymorphism.

Author Contributions

The corresponding author takes full responsibility that all authors on this publication have met the following required criteria of eligibility for authorship: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved. Nobody who qualifies for authorship has been omitted from the list.

Erik Sistermans (Conceptualization-Lead, Writing—original draft-Lead, Writing—review & editing-Lead)

Authors’ Disclosures or Potential Conflicts of Interest

Upon manuscript submission, all authors completed the author disclosure form.

Research Funding

None declared.

Disclosures

E. Sistermans received a Ph.D. grant from MRC Holland.