Proof-of-principle rapid noninvasive prenatal diagnosis of autosomal recessive founder mutations

Abstract

Background: Noninvasive prenatal testing can be used to accurately detect chromosomal aneuploidies in circulating fetal DNA; however, the necessity of parental haplotype construction is a primary drawback to noninvasive prenatal diagnosis (NIPD) of monogenic disease. Family-specific haplotype assembly is essential for accurate diagnosis of minuscule amounts of circulating cell-free fetal DNA; however, current haplotyping techniques are too time-consuming and laborious to be carried out within the limited time constraints of prenatal testing, hampering practical application of NIPD in the clinic. Here, we have addressed this pitfall and devised a universal strategy for rapid NIPD of a prevalent mutation in the Ashkenazi Jewish (AJ) population.

Methods: Pregnant AJ couples, carrying mutation(s) in GBA, which encodes acid β-glucosidase, were recruited at the SZMC Gaucher Clinic. Targeted next-generation sequencing of GBA-flanking SNPs was performed on peripheral blood samples from each couple, relevant mutation carrier family members, and unrelated individuals who are homozygotes for an AJ founder mutation. Allele-specific haplotypes were constructed based on linkage, and a consensus Gaucher disease-associated founder mutation-flanking haplotype was fine mapped. Together, these haplotypes were used for NIPD. All test results were validated by conventional prenatal or postnatal diagnostic methods.

Results: Ten parental alleles in eight unrelated fetuses were diagnosed successfully based on the noninvasive method developed in this study. The consensus mutation-flanking haplotype aided diagnosis for 6 of 9 founder mutation alleles.

Conclusions: The founder NIPD method developed and described here is rapid, economical, and readily adaptable for prenatal testing of prevalent autosomal recessive disease-causing mutations in an assortment of worldwide populations.

Funding: SZMC, Protalix Biotherapeutics Inc., and Centogene AG.

Figure 5. Noninvasive fetal allele identification based on familial, N370S consensus, and/or N370S near-consensus haplotypes.

Illustrations depict the GBA locus (±2 Mb) and SNPs that were deep sequenced for the construction and typing of fetal alleles (as indicated in the “Haplotype legend”). The numbers shown under DFM denote the distance from mutation (in Mb). The noninvasively identified fetal alleles were (A) WT paternal, (B) N370S maternal, (C) N370S maternal, (D) N370S maternal, (E) WT paternal, (F) WT maternal, (G) N370S paternal, (H) N370S paternal, (I) L444P (non-N370S) maternal, and (J) 84GG (non-N370S) maternal. Note the utility of the N370S consensus haplotype for fetal typing in B, C, and H. The near-consensus N370S haplotype also aided fetal typing in B, C, F–H, and J.

Figure 4. Extended fine mapping of the consensus and near-consensus AJ N370S founder haplotype region as a tool for phasing fetal haplotypes.

Illustration depicting the GBA locus (±2 Mb) and thousands of SNPs that were deep sequenced for the construction and typing of fetal alleles according to the analytical pipeline (as indicated in the key). (A) An extended deep-sequencing panel was used to better fine map the conserved N370S founder haplotype, as in Figure 2. Accordingly, a 301-SNP haplotype (termed “full-consensus N370S haplotype”) was identified in all N370S chromosomes in this study (28 chromosomes altogether). In addition, the consensus haplotype was found to extend 500 kb further downstream of GBA (620 additional SNPs) in 15 of 16 chromosomes from 8 N370S homozygotes. Furthermore, in all N370S homozygotes (but not all N370S carriers), the consensus haplotype was found to extend another 120 kb upstream of GBA (100 additional SNPs). Altogether, these extended haplotypes were termed “near-consensus N370S haplotypes.” (B) The N370S haplotype from each N370S carrier parent in the study was carefully mapped according to homozygous regions and family-based linkage analysis. After comparison to the near-consensus haplotype in A, new parent-specific 5′ and/or 3′ demarcations of the N370S near-consensus haplotype were set (this haplotype was termed the “parent-specific consensus N370S haplotype”). (C) In this example, deep sequencing of the GBA-flanking region in a fetus identified stretches of a linkage-based parental N370S haplotype that resided outside of the consensus N370S region. In addition, some stretches of fetal sequence could not be phased according to family-based linkage. (D) When unphased fetal sequence, such as in C, fell within the parent-specific consensus N370S haplotype (as determined in B), the consensus information was used to phase the fetus (here, with the N370S-linked haplotype), thereby increasing confidence in the diagnostic test result.

Figure 3. Noninvasive fetal allele identification based on familial and/or the N370S founder haplotype.

Illustrations depict the immediate GBA-proximal locus and SNPs that were deep sequenced for the construction and typing of fetal alleles (as indicated in the “Haplotype legend”). (A) In family 1, the paternal WT allele was diagnosed by inference from the family-based N370S-linked haplotype (red squares). The consensus N370S haplotype could not be used to phase the paternal allele in the fetus due to paternal homozygosity in the founder haplotype region. (B) On the other hand, the maternal N370S allele in family 1 was readily identified (in multiple sites) via the consensus haplotype, and this result was corroborated by equivalent matches to the family-based maternal N370S haplotype. (C) For the family 2 maternal allele, the fetal N370S haplotype could only be matched to a single polymorphic site in the family-based haplotype (red square). This site and 2 other fetal SNPs were definitively matched to the N370S mutation by comparison to the founder N370S haplotype. Therefore, in this case, it would not have been possible to reliably diagnose the maternal allele in the fetus without the N370S founder sequence.

Figure 2. Fine mapping of the consensus AJ N370S founder haplotype region.

Hundreds of GBA-flanking SNPs (±250 kb from GBA) were sequenced in order to identify a conserved N370S founder haplotype. (A) NGS-based homozygosity mapping with 7 unrelated homozygote N370S Gaucher patients (denoted as H₁–H₇) (14 N370S chromosomes) was used to identify a preliminary founder haplotype. (B) A representative linkage-based inference of a familial N370S haplotype (hap_N370S). This linkage analysis was performed for 6 different heteroallelic GBA N370S mutation carrier duos (6 N370S chromosomes from 6 sets of 2 first-degree family members carrying the N370S mutation). The resultant alleles were each compared separately to the haplotype from A until a consensus N370S haplotype was demarcated with a 5′ cutoff. (C) Ultimately, the consensus AJ N370S founder haplotype (composed of 153 SNPs) used for NIPD was constructed from 20 different AJ N370S chromosome sequences. Note that this analysis set a 5′ cutoff for the conserved N370S haplotype, but a 3′ cutoff could not be established. WT denotes a WT allele.

Figure 1. Pedigrees of GBA mutation carrier families in this study.

Mutations in GBA are indicated in red. Individuals with unknown genotypes at sample collection are shaded in gray. WT denotes a WT GBA allele; wk, denotes the week of gestation at which maternal plasma was collected.