The relative mutation dosage approach

By using dPCR for maternal plasma DNA analysis, the relative mutation dosage (RMD) approach could be applied to assess maternally inherited fetal mutations for NIPT of monogenic diseases ( Figure 1 A). This strategy is particularly relevant to couples who are carriers of the same autosomal recessive mutation, as frequently observed in consanguineous marriages and diseases with a large founder effect, or if the mother carries the mutation inherited in an autosomal or X-linked fashion. The rationale lies in the identification of slight quantitative alterations of the relative proportions of the mutant and wild-type alleles caused by the additional contribution of maternally inherited fetal alleles in maternal plasma. In fact, one could detect a balanced ratio of the mutant and the wild-type alleles (1:1 allelic ratio) in the plasma of non-pregnant woman heterozygous at the locus (i.e., possessing one mutant and one wild-type allele). During pregnancy, however, the fetal DNA released into maternal circulation would alter the relative allele dosage, with a magnitude of the difference depending on the fetal DNA fraction. For example, in autosomal recessive diseases, the fetus homozygous for a mutant allele would skew the relative allele dosage toward the mutant allele, while the fetus homozygous for the wild-type allele would cause skewing of the relative dosage toward the wild-type allele. A fetus that has inherited both the wild-type and the mutant alleles would not alter the relative dosage. In RMD for an autosomal dominant disease, the fetus either inherits the mutant allele from the mother (the allelic ratio would be equal) or the maternal mutant allele would not be passed onto the fetus, resulting in an over-representation of the wild-type allele. Only two combinations of male fetal genotypes are possible for diseases with an X-linked inheritance. Ultimately, the type of inheritance patterns influences the degree of allelic skewing in relation to the fetal DNA fraction. For example, the expected degree of skewing is different for the autosomal recessive model of inheritance as compared to the X-linked disease model for male fetuses . The RMD approach has been applied to the detection of maternally inherited fetal mutations causing β-thalassemia , sickle-cell anemia , hemophilia , and methylmalonic acidemia .

Non-invasive prenatal diagnosis of monogenic diseases using the relative mutation dosage (RMD) approach by droplet digital PCR (ddPCR) and the relative haplotype dosage (RHDO) approach by massively parallel sequencing (MPS). A) RMD analysis examines alterations of the relative proportions of the mutant (M) and wild-type (N) alleles. (1) Maternal and paternal mutation status are assessed by parental DNA genotyping, which guides the design of the digital PCR assay. For illustration, parents who are carriers of the same autosomal recessive mutation (M/N) are shown. (2) Maternal plasma DNA is analyzed using ddPCR where most of the reaction droplets contain zero or one template DNA molecule. The mutant allele proportion (M) is determined by the mutation assay and the fetal DNA fraction (F) is determined by the ZFXY assay. (3) The concentrations of the mutant and wild-type DNA, as well as ZFX and ZFY DNA are measured. The M and F are calculated and used in the sequential probability ratio test (SPRT) analysis to determine whether the mutant or wild-type alleles in maternal plasma are in allelic balance or imbalance. (4) Samples with the proportion of positive droplets containing the M allele by RMD analysis or with the proportion of total sequence reads contributed by Hap I alleles by RHDO analysis above the upper boundary and below the lower boundary are classified as affected and normal, respectively. B) RHDO analysis interrogates alterations of the relative proportions of maternal haplotypes through the use of heterozygous single-nucleotide polymorphisms (SNPs). (1) Paternal genotype data and maternal haplotype data are determined. SNPs being heterozygous in the mother (A/B) and homozygous in the father (A/A) are selected. (2) Maternal plasma DNA is sequenced, sequence tags are aligned to the reference human genome, and the SNPs are located. (3) The number of sequenced reads for each allele of the respective informative loci is counted and the fetal DNA fraction is determined. The SPRT analysis statistically evaluates whether a haplotype dosage balance or imbalance for Hap I (haplotype associated with the mutant allele) or Hap II (haplotype associated with the wild-type allele) is observed in the maternal plasma as a result of plasma DNA contribution originating from the fetus. The counts of each informative SNP on the same haplotype are added until reaching a statistical confidence for SPRT classification.

The relative haplotype dosage approach

The current sequencing technologies enable obtaining the whole mutational profile of the fetus from maternal plasma, which opens new exciting opportunities to diagnose virtually any monogenic disease by using only one genetic test . The single-nucleotide polymorphism (SNP) information detected in the paternal and maternal genomes facilitates the interpretation of DNA sequencing data for NIPT of monogenic diseases. First, the paternal inheritance is deduced through detection of paternally inherited SNPs and/or mutations present in maternal plasma but absent from the mother’s genome, which are heterozygous in the father and homozygous in the mother. On the other hand, the maternal inheritance is established by interrogating SNP markers as a series of inheritance blocks, so-called haplotypes, selected from genomic regions physically linked to the mutation, for which the father is homozygous and the mother is heterozygous. This strategy has been termed as a relative haplotype dosage (RHDO) analysis and is affected by the fetal DNA fraction ( Figure 1 B). However, unlike RMD that measures the allelic alterations related only to one particular mutation or SNP, the RHDO approach assesses the balance or imbalance of maternal haplotypes. The robustness of such a test increases by analyzing multiple SNP sites assigned to one haplotype. Recent advances use targeted sequencing protocols, which allow one to analyze only the genomic region of interest, removing the need for studying the whole genome . In addition to its application in β-thalassemia, which will be discussed in the next section, a targeted sequencing approach coupled with RHDO analysis has been successfully tested for the early assessment of CAH fetal status in 14 at-risk pregnancies . The protocol captured a 6-Mb region flanking the CYP21A2 gene, and the parental haplotypes in each family were established by using the genomic information of the parents and the proband. Fetal inheritance of the parental haplotypes was then deduced by maternal plasma DNA analysis using sequence reads from the covered genomic region, which allowed correct prediction of the fetal CAH status in all families as early as 5 weeks and 6 days of pregnancy. The NIPT approach for the assessment of mutations in CYP21A2 has great clinical impact because early diagnosis allows for evaluating whether dexamethasone treatment is required in the pregnancy. The robustness of the haplotype-based analysis using targeted MPS has been further proven in NIPT of other families with CAH or Duchenne and Becker muscular dystrophies , and in the detection of autosomal recessive founder mutations for type I Gaucher disease .

Exclusion of paternally inherited β-thalassemia mutations in pregnancies with parents carrying different mutations

Thalassemias have been widely studied as a model for NIPT of monogenic diseases. β-thalassemia major is a severe form of the disease requiring regular blood transfusions, whereas the carriers of β-thalassemia are clinically asymptomatic . Chiu et al investigated the use of cell free DNA in maternal plasma in pregnancies at risk of β-thalassemia major . The authors successfully employed an allele-specific real-time PCR for the detection or exclusion of the paternally inherited mutation, a 4-bp deletion at codon 41 and 42, CD 41/42 (−CTTT). The clinical significance of such an approach was substantial, as the identification of fetuses that have not inherited the paternal mutant allele would eliminate the need for an invasive diagnosis in couples carrying different β-thalassemia mutations.

Exclusion of paternally inherited β-thalassemia mutations shared by both parents

For populations where the parents are found to share the same mutation, such an approach might not be applicable due to the inability to distinguish between the maternal and paternal alleles. In these clinical scenarios, the analysis of paternal SNP markers offers the solution, provided that a set of informative SNPs for which the mother is homozygous and the father heterozygous can be obtained. As compared to the mutation-specific assays, the SNP-based approach can target not only the mutant but also the wild-type allele, if the linkage to the paternal SNP has been established from a pedigree analysis. However, careful choice of technologies is needed so that the one-nucleotide changes between the maternal and fetal DNA sequences in maternal plasma could be robustly investigated. Ding et al developed a system for single-nucleotide discrimination of maternal plasma DNA by using single-allele base extension reaction (SABER) and mass spectrometry . The authors explored the exclusion-based approach proposed by Chiu et al for analysis of the four most common southeast Asian mutations causing β-thalassemia. Besides the direct mutation detection from maternal plasma, this study has for the first time demonstrated NIPT of a monogenic disease in couples that shared the same mutation by assessing the fetal inheritance of paternal SNP alleles linked to the disease locus. A microarray system, called an allele-specific PCR and arrayed primer extension (AS-APEX), has later provided parallel detection of multiple paternally inherited SNPs linked to the HBB locus for NIPT of β-thalassemia in the Cypriot and southern Chinese population .

The researchers proposed various molecular strategies to improve the sensitivity of conventional methods, such as real-time PCR, in detecting fetal mutations in maternal plasma. One approach took advantage of the size differences between the fetal and maternal DNA for a selective enrichment of fetal DNA fragments , the latter being generally shorter in size as compared to maternally derived DNA . A size-based separation of plasma cell free DNA in agarose gel coupled with a suppression of maternal alleles using a peptide-nucleic-acid (PNA) clamp PCR has improved the sensitivity to detect paternally inherited fetal point mutations in couples carrying different β-thalassemia mutations. Later, the same group tested the size-fractionated cell free DNA by mass spectrometry to analyze a paternally inherited CD 39 mutation in the plasma of a pregnant carrier of the IVSI-110 mutation . Although the size fractionation step significantly improved the sensitivity, the susceptibility to contamination using a gel-based approach posed a limitation to its routine use. The PNA-clamping method has also been tested in conjunction with a microelectronic microchip technology in order to detect seven HBB mutations frequent in Mediterranean regions in parents carrying different mutations . The PNA-clamping technology, however, requires a mutation-specific optimization of PNA concentration, which might prohibit its use in routine and large-scale testing. An alternative approach for fetal DNA enrichment focused on the applicability of full co-amplification at lower denaturation temperature-PCR (COLD-PCR) for the detection of two paternally inherited β-globin mutations common in the Mediterranean populations . In COLD-PCR, the mutation-containing DNA sequences become preferentially amplified during PCR due to their reduced melting temperature. The authors indeed observed an increased sensitivity for the detection of paternally inherited mutant alleles with a full COLD-PCR. Liu et al have recently developed a new enrichment strategy for NIPT of paternally inherited fetal mutations causing β-thalassemia . The authors designed a primer-introduced restriction analysis PCR (PIRA-PCR), which uses restriction digestion to selectively eliminate wild-type alleles in the detection of fetal β-thalassemia mutations in maternal plasma. This approach, however, depends on a limited number of restriction enzymes not suitable for every mutation.

Detection of maternally inherited β-thalassemia mutations

To broaden the spectrum of pregnancies amenable to prenatal testing for β-thalassemia using maternal plasma DNA analysis, one had to develop strategies for the detection of maternally inherited mutations. This goal has been achieved by using the RMD approach, which enabled the identification of maternal mutations in the HBB gene, CD 41/42 (−CTTT), and HbE (G>A) in the plasma of pregnant carriers with healthy homozygous male partners . The protocol correctly predicted the fetal status in 50% of the cases, although it was unable to classify 4 cases with fetal DNA fractions lower than 10% and incorrectly classified one case. The authors therefore developed another size-based approach, called a digital nucleic acid size selection (NASS) to enrich for fetal DNA fragments, which enhanced the RMD classification.

The MPS technology was adopted for the NIPT of β-thalassemia in a couple, where the father was a carrier of the CD 41/42 (−CTTT) mutation, and the pregnant woman carried the A>G mutation at nucleotide −28 of the HBB gene . The detection of DNA sequence reads containing the SNPs linked to the paternal mutation confirmed that the fetus had inherited the mutation from the father. The maternal haplotype was then constructed by using genotyping data of the mother and the fetus, which assigned the −28 mutation to maternal Hap II and the wild-type allele to maternal Hap I. The over-representation of the Hap I in maternal plasma using the RHDO analysis indicated that the fetus had inherited the maternal wild-type allele, and therefore was a β-thalassemia carrier.

A targeted sequencing approach, which selectively analyzes genomic regions containing the disease-causing gene, provides a solution to reduce the costs of the MPS-based tests by increasing the proportion of informative data. Lam et al explored the use of solution-phase hybridization, which captured a 288-kb region of the HBB gene cluster in couples who were carriers of different β-thalassemia mutations . Without the need for a pedigree analysis, the digital PCR strategy enabled a physical deduction of parental haplotypes, which were then used in RHDO analysis to reveal the fetal β-thalassemic status. This study further demonstrated the applicability of RHDO analysis to establish maternal inheritance for a genomic region in which the parents had similar haplotype structures.

The MPS technology has later been examined for NIPT of β-thalassemia in the Cypriot population using amplicon sequencing, which targeted SNPs across the β-globin cluster linked to the paternal mutant allele . By selecting highly heterozygous SNPs, this approach was applicable in 80% of pregnancies, but the presence or absence of the paternal mutant allele was correctly determined only in 27 out of 34 samples, with four false negative and three false positive results.

More recently, Xiong et al developed a targeted MPS approach to detect the HbE mutation as well as the four most common β-thalassemia mutations found in southeast Asia . Eighty-three families with couples being carriers of the mutations were analyzed by using three overlapping PCR amplicons for each of the five tested mutations, to assess the presence of paternally inherited mutant alleles in maternal plasma. The overall sensitivity and specificity for the detection of paternal mutations was 100% and 92.1%, respectively. The protocol, however, was not applicable to one-third of the pregnancies when the parents carried the same mutation.