Reliable Detection Of Paternal SNPs Within Deletion Breakpoints For Non-invasive Prenatal Exclusion Of Homozygosity For α~0-thalassemia Using Cell-free Fetal DNA In Maternal Plasma

Backgrounda-Thalassaemia is the most common monogenic inheritance disorder in tropical and subtropical areas in the world with a carrier frequency reaching as high as 10-15% in southeast Asia, the Indian subcontinent, and southern China. This inherited anemia syndrome mainly results from gross deletions of the human a-globin gene cluster mapped to chromosome 16p13.3, and thus is characterized by a reduction or complete suppression of a-globin chain synthesis. Deletion of the double a-globin gene is designated asα0-thalassaemia, such as—SEA,—FIL and—THAI, which is more prevalent in southeastern Asia including southern China. This deletion can result in a fatal condition, hemoglobin (Hb) Bart's hydrops fetalis, when the deletion is homozygous. When both parents areα0-thalassaemia carriers, each pregnancy has a 25% risk for Hb Bart's hydrops. An affected fetus can be effectively detected by prenatal diagnosis for at-risk couples by fetal sampling via amniocentesis and chorionic villus sampling (CVS). However, such traditional techniques are invasive and carry a small but significant risk of miscarriages (0.5%-1%).Fetal nucleated red blood cells or cell-free fetal DNA (cffDNA), that exist in maternal circulation during pregnancy offers a potential source of fetal material for non-invasive prenatal diagnosis (NIPD) of the fetal genetic status. Reliable NIPD tests using cffDNA has recently been used to detect paternally inherited genetic traits, such as Y-linked disorders and rhesus D. A panel of highly sensitive technologies to detect paternally inherited fetal point mutations has been developed and NIPD using cffDNA has been applied to a group of single gene disorders includingβ-thalassemia, cystic fibrosis, and congenital adrenal hyperplasia (CAH), and has been used on the targeted genes in couples who carry different mutations in particular. It has been demonstrated that an alternative approach for detection of different SNPs linked to the mutant allele in parents with identical point mutations is potentially useful in non-invasive diagnosis using cffDNA.It is technically challenging to reliably detect large fetal deletions from cffDNA in maternal plasma when couples share the same deletion due to the lack of specific markers for fetal genotyping. In addition, the fragmented nature of cffDNA is not conducive to efficient detection of large deletions using a PCR-based technique. Recently, Ho et al. reported a method for identification of non-deleted paternally inherited fetal alleles in cfDNA using two microsatellite markers within deletion breakpoints in NIPD for Hb Bart's syndrome. The number of polymorphic STRs in the region of interest is a disadvantage as it is limited, and therefore the method is only applicable for one-third of pregnancies. Using a-thalassaemia as a diagnostic target, we present a novel strategy for the noninvasive prenatal diagnosis of a homozygous deletion using multiplex SNP discrimination of fetal DNA from maternal plasma. In this study we established a protocol for the reliable detection of a panel of paternal SNPs within the (—SEA) breakpoints in cfDNA samples and evaluated 67 pregnancies at risk for a-thalassemia major; moreover, we demonstrated the feasibility and reliability of this NIPD method.Materials and MethodsSample collectionA total of 150 normal unrelated individuals and 150—SEA/aa carriers were recruited for characterization of SNPs within the (—SEA) breakpoints in the a-globin gene cluster (Accession NG₀00006). To establish multiplex PCR-based mini-sequencing for simultaneous scanning of multiple SNPs,40 standard sequenced samples were used in the study. Maternal plasma samples from 67 pregnant women at risk for a-thalassemia major were collected prior to invasive obstetrics procedures (8.5-25 weeks of gestation, mean 19.12 weeks, before chorionic villus sampling in the first trimester and before amniocentesis in the second trimester) between June 2008 and June 2010 in southern China. The patients' age ranged from 20 to 41 years with a mean of 27 years and median of 26 years. Both of the parents of each pregnancy carried an identical (--SEA) deletion except for one case. The study was approved by the Ethics Committee of Nanfang Hospital, an affiliate of Southern Medical University, China. Detailed written informed consent was obtained from all participants.Blood processing and maternal plasma extractionThe buffy coat was collected from blood samples after centrifugation at 1600g for 10 min, whereas maternal plasma was collected after an additional microcentrifugation at 16000g for 10 min. Buffy coat DNA was extracted using a standard phenol-chloroform method. Two milliliters of maternal plasma from each pregnancy were used for DNA extraction using the QIAamp DNA blood Mini Kit according to the manufacturer's instructions with minor modifications. The CVS samples and amniotic fluid samples were handled and analyzed using the multiplex PCR-based mini-sequencing and the a-thalassemia deletion Gap-PCR as previously described by Xiao et al in four clinical centers.Diagnostic strategyIn a-thalassaemia with a cis double gene deletion, the presence of the paternally inherited SNP allele in the maternal plasma would suggest that fetus had inherited the paternal normal a-globin allele and thus would not manifest hemoglobin Bart's hydrops fetalis. A total of 67 at-risk heterozygous couples with the (—SEA) deletion were recruited for this study. Initially, peripheral blood samples were collected from each of the at-risk couples, white blood cells from all couples were then used to isolate genomic DNA, and maternal plasma was used to isolate cfDNA; subsequently, quantitative real-time PCR (Q-PCR) was performed to quantify total cfDNA and cffDNA in the maternal plasma. Secondly, multiplex PCR-based mini-sequencing was used to identify informative markers using genomic DNA from both parents by screening for nine previously known SNPs within the deletion breakpoints. Finally, an allele-specific real-time PCR method was used to detect the normal paternally inherited fetal allele using cfDNA in maternal plasma, according to the detailed information on one or more specific SNPs (generally two sites were selected) that differ in the maternal and paternal genomes. The NIPD results from analyses on all families with informative SNPs were compared with results obtained from conventional invasive prenatal diagnosis (CIPD) in a blinded manner. The CIPD results were uncovered only once the genotypes obtained by NIPD were scored.Screening for informative SNPs within deletion breakpointsGenomic DNA from the 67 couples was used to select informative markers useful for further non-invasive testing. Informative SNP markers were identified by multiplex PCR amplification using a Perkin Elmer Gene Amp PCR system 9600 thermal cycler, followed by mini-sequencing based on a fluorescent-labelled dideoxy single-base extension of an unlabeled oligonucleotide primer or primers that hybridize upstream of the SNP site. The analysis was performed using an ABI3730 Genetic Analyzer and GeneMapper v3.7 software. The specific primer sequences and PCR cycling conditions for the informative SNP markers are listed in Table 1. For the multiplex PCR amplification,50 ng of genomic DNA was used as the template. The final volume was 20μL, containing 10 pmol of primer mixture,2μL 10×PCR buffer, 0.8μL 50 mM MgCl2,0.5μL 10 mM dNTPs, and 1 U Platimum Taq DNA polymerase. After an initial incubation at 95℃for 15 min, the reaction was cycled for 30 s at 94℃,30 s at 56℃, and 1 min at 68℃for 32 cycles, followed by a final extension for 30 min at 68℃. Excess primers and dNTPs were removed by the addition of 5 U shrimp alkaline phosphatase (SAP) and 2 U Exol to 15μL of PCR product and incubation of the mixture at 37℃for 1 hour, followed by 75℃for 15 min. Following the PCR reaction,5μL was run as needed on a 1.5% agarose gel to check efficiency and specificity. Sequence of the single base extension primers including different polymeric T tails,2μL of purified PCR product from parental samples and 2μL from the controls were used as the template. The reaction was carried out in a final volume of 5μL, containing 2.5μL of SNaPshot (?)Multiplex Ready Reaction Mix and 5 pmol of the specific unlabeled primers. The reaction was cycled for 10 s at 96℃,5 s at 50℃, and 30 s at 60℃for 25 cycles. Excess nucleotides were removed by the addition of 1μL (1 U/μL) SAP and 1 U CIP, and incubation of the mixture at 37℃for 1 h followed by 75℃for 15 min. Purified product from control DNA was diluted 1:10 before being charged in the automated sequencer. For the analysis,0.5μL of purified product from the parent's samples and 1μL from the controls were mixed with 9μL of Hi-Di formamide and 0.5μL of GeneScan 120 Liz size standard, followed by incubation of the mixture at 95℃for 5 min and then immediately frozen 5 min.Quantification of circulating DNA in the plasmaReal-time Q-PCR was performed using primers and probe specific to the GAPDH gene in all pregnancies to determine the total amount of circulatory maternal plasma DNA. The male-specific SRY gene was used to monitor the presence of cffDNA and quantify cffDNA. This was achieved according to a method previously described by Birch et al. The amount of circulatory maternal plasma DNA obtained from 67 pregnant women and cffDNA from mothers bearing male fetuses was calculated using real-time PCR threshold cycle values.Specificity and sensitivity of allele-specific real-time PCRTo define the sensitivity and specificity of allele-specific real-time PCR for detection of fetal DNA in the maternal plasma for this study, we conducted a series of experiments performed on artificial mixtures of mutant allele diluted into wild-type allele samples. For each SNP,100 ng genomic DNA was successively diluted at 1:1, 1:10,1:100,1:1000, and 1:10000 in water and used as a starting template for amplification. In addition, DNA samples that affect the ratio of paternal versus maternal alleles were prepared in PCR reaction tubes by mixing --SEA/aa genomic DNA with maternal allele spiked in varying dilutions of 1:1,1:2,1:5,1:10,1:20,1:50, 1:100,1:500,1:1000,1:5000 and 1:10000 in --SEA/aa genomic DNA with paternal allele diluents. The experiment was repeated twice. We optimized the specificity of each of the allele-specific assays for informative SNPs by assessing several factors including buffers composition, temperature, and length of PCR amplification cycles, together with the use of various oligonucletide primers. We then used theΔCT, (Paternal-Maternal) approach to measure the quantitative differences between the paternal allele and maternal allele. We also used melting curve analysis to monitor the presence of the paternal allele.Detection of paternally inherited fetal SNP allele in maternal plasmaAllele specific real-time PCR was performed to detect the paternally inherited normal allele in the presence of an excess of the corresponding maternal sequence. As all informative SNP artificial experiments are functional, the analysis to detect paternally inherited fetal allele by allele specific real-time PCR in the maternal plasma was performed on a Thermocycler sequence detector (Stratagene Mx 3005; La Jolla, CA, USA). A total of 10μL maternal plasma DNA was amplified in duplicate in a final reaction volume of 25μL containing 10μM wild type primer or mutant primer 0.25μL,10μM common primer 0.25μL, and 12.5μL SYBR(?) Green Real-time PCR Master Mix (Applied Biosystems, Foster city, CA). The assay protocol started with an amplification procedure consisting of a denaturation step of 30 sec at 95℃,55 amplification cycles of 10 s at 95℃,10 s at annealing temperature for each SNP marker, and 20 s at 72℃; and this procedure was followed by a melting analysis procedure consisting of a denaturation step of 1 min at 95℃, and a stepwise temperature increase from 60℃to 95℃at 1℃/step with 5 s detection for each step. The fluorescence from SYBR channel was recorded at each step during the annealing and the melting analysis procedure.ResultsAllele frequencies of 16 SNP markers were obtained from 450 chromosomes in Southern Chinese. Nine SNPs with minor allele frequency (MAF)>0.15 were considered to be informative and were used in our protocol. Three SNP markers (rs2514675, rs2974771 and rs2858942) were highly informative (MAF>0.45).The mean concentration of cfDNA obtained from maternal plasma of 65 pregnancies was 611.63 genome-equivalents per milliliter; the mean concentration of fetal DNA in 31 plasma samples from mothers bearing male fetuses was 47.35 genome-equivalents per milliliter. In order for the real-time AS-PCR assay to function reliably, the input DNA concentration were used within a certain range 87.8-1556 genome-equivalents. The amount of circulating DNA in maternal plasma would be sufficient to analyze paternally inherited normal SNP alleles. In addition, the majority of whole blood samples (97%) were processed directly in four clinical centers and the plasma was sent frozen to our center by commercial express courier service. Two samples (case 21 and 38) were sent as whole blood by commercial express courier service. They were placed frozen gel packs directly on tubes of blood, therefore resulted in hemolysis. Based on previously reported results by Birch et al these two cases had to be excluded because the amount of cfDNA exceed 49940 genome equivalents/mL and was substantially higher than would be expected in maternal samples during pregnancy. This method involved multiplex PCR-based mini-sequencing to scan for informative SNPs followed by allele-specific real-time PCR for NIPD results. The mini-sequencing assay demonstrated feasibility and reliability in simultaneously genotyping nine SNPs, and exhibited suitability for high-throughput SNP scanning of genomic DNA samples by multiple PCR-based fragment analysis. We validated the assay using a group of previously sequenced DNA samples (20 unrelated normal individuals and 20 --SEA/ααcarriers). The SNP typing of each case was accurately defined by the mini-sequencing assay at each nucleotide position analyzed, exhibiting 100% concordance with the reference method. SNPs were then assessed in 67 couples, and one or more different SNP(s) within the deletion breakpoints between paternal and maternal alleles was detected in 97% of the couples (65 cases). The SNPs were informative and could be further used as candidate polymorphic sites linked to the paternal-normal allele. Two cases (family 15 and family 35) were non-discriminative as none of the SNPs were informative. Because the a-globin genes are duplicate (a2 and al) with a high degree of homology, SNP ID rs2541669 (g.33004C>T) is located within the Z2 box of a2-globin gene, which is homologous to the Zl box of al-globin gene; while SNP ID rs2238369 (g.35483T>C) is located within the XI box of al-globin gene, which is homologous to the the X2 box of a2-globin gene. The multiplex PCR of these two markers respectively yield two amplicons with identical length. The mini-sequencing amplicons of these two markers show the presence of two peaks of amplification in heterozygous (—SEA) samples whereas all SNPs tested should be hemizygous. However, such expected detection results do not influence our ability to correctly type SNPs for each of these two sites. As illustrated in family 7, g.33004 T from father was detected by both T peak (red) and additional C peak (black), thus indicating a truly positive T allele in this case. The g.35483 T is similar to g.33004 T, with two peaks showing in mother.Reference DNA samples and reliable genotyping samples of recombinant mutant clones were obtained in order to develop the allele specific real-time PCR assay technology. Because a mixture of maternal and paternal alleles was present, a model to test the discrimination capability between paternal and maternal alleles was established by subtracting the respective CT values (ΔCT) of maternal alleles (CT, Matemal) and paternal alleles (CT, patemal) using a serial of artificial model samples for each of the 9 SNPs. These assessments permitted establishment of tentative cut-offΔCT value that could serve to distinguish between samples in which the paternal inherited allele was absent and those samples in which it was present. Following theΔCT, (Patemal-Maternal) analysis, the presence of paternally inherited alleles was at aΔCT value of less than 6. This method permitted a clear discrimination between maternal and paternal allele, even with experimental conditions whereby the maternal allele was diluted in paternal allele of the order expected in maternal plasma samples. The results of the model allele specific real-time PCR experiment demonstrated our optimized protocol could detect one copy of paternal DNA in every 50 copies of maternal DNA. The data from the analysis of the g.31921T>C marker by the allele-specific real-time PCR assay showed unequivocal results. This assessment permitted the establishment of a tentative cut off value that could serve to discriminate between the presence and absence of the paternal-normal allele in the clinical cfDNA samples. Furthermore, based on this cut off value we were able to correctly determine the paternal-normal allele of cffDNA from all 65 cases, revealing the sensitivity and robust accuracy of this technique.A total of 65 cases were examined for the paternal-normal allele and it was detected in 5 out of 5 cases for one discriminative SNP,6 out of 6 cases for two discriminative SNPs,7 out of 7 cases for three discriminative SNPs,5 out of 5 cases for four discriminative SNPs,23 out of 25 cases for five discriminative SNPs,15 out of 15 cases for six discriminative SNPs, and 2 out of 2 cases for seven discriminative SNPs. Two uncertain results were obtained in 2 cases in this study due to sample hemolysis. Homozygosity for the (—SEA) deletion was accurately ruled out in 33 out of 67 (49.3%,95% CI 25.4-78.6%) pregnancies tested. The overall sensitivity for the nine paternal SNP markers detected reached 100% and the specificity was 93.8%. Results for the paternal-normal allele in the maternal plasma from 63 families obtained from analyses in which our protocol was implemented were in concordance with results obtained by conventional prenatal diagnosis.DiscussionIn the present blind study of 67 high genetic risk couples we have shown that PCR-based DNA detection of paternal SNPs within the deletion breakpoints is a rapid and reliable technique for non-invasive prenatal diagnosis of deletionalα-thalassemia. A group of informative SNP markers within the (—SEA) breakpoints that differ in the maternal and paternal chromosomes dependent on the natural variation in the Chinese population studied have been identified. Furthermore, the SNP markers designed in the present strategy are available for exclusion of the paternal deleted chromosome of other two common deletion of (—THAI) and (—FIL) resulting in a-thalassemia in Southeast Asia, thereby potentially extended to include testing for these deletions for NIPD applications. We demonstrated the feasibility of our newly developed method as 97% of the couples (65 cases) had one or more different SNP(s) within the breakpoints of deletions between paternal and maternal alleles.2 of the 67 cases were non-informative, due to the inheritance of identical SNP markers in both parents. Therefore, it is possible to exclude nearly 50% of the at-risk subjects tested by exclusion of the homozygous (—SEA) deletion based on detection of the paternally inherited fetal normal allele if more samples are available to identify informative SNPs. With a potential application in one-half of pregnancies our SNP-based assay thus generally overcomes the limitations of microsatellite markers, although our assay cannot be used to differentiate homozygous deletions from carriers in the remaining 50% of at-risk pregnancies. In some cases, it is possible to detect fetal genetic loci for disease-causing point mutations, as well as SNPs linked to paternally inherited fetal mutant alleles involved in disorders such asβ-thalassemia, achondroplasia, and myotonic dystrophy, although this is highly technically challenging. However, there are very few NIPD applications that can analyze single-gene disorders caused by large deletions through a reliable approach using cffDNA in maternal plasma. This novel approach has been successfully applied to the detection of a-thalassemia in the present study, and could be expanded to other autosomal genetic diseases due to gene deletions, such as 22q11.2 deletion syndrome and deletional (3-thalassemia.In addition to the conventional extraction of cffDNA, the key element of the molecular test involved in our approach is the two-step protocol involving multiplex PCR-based mini-sequencing and allele specific real-time PCR. By the use of the former method, we have shown that it is possible to simultaneously detect nine SNPs within breakpoints of the deletion tested using genomic DNA from high-risk parental individuals. The results obtained from our random samples indicated that two to seven informative SNPs could be detected 89.55% of at-risk couples. Thereby, the test is easy to process since only one SNP is needed in the practical protocol of NIPD for detection of the paternally inherited fetal allele in maternal plasma. The latter is used to detect one informative SNP selected from the above several sites for direct exportation of cfDNA sample NIPD results. Another informative SNP sites can be used to provide double-check information on NIPD for the same plasma sample previously tested if it is necessary. Uncertain results were obtained in 2 cases in this study due to sample hemolysis that could lead to significantly increasing maternal DNA in the tested samples; thereby the amount of maternal DNA in the PCR sample can interfere with the assay. There were several factors observed to affect the sensitivity and reproducibility of the assay system, particularly PCR polymerase, hence AmpliTaq Gold LD polymerase was selected to guarantee consistent results and we were able to consistently detect 2 copies of paternal alleles without interference from the maternal sequences in the model PCR experiment.In our study, plasma samples were collected during 8.5-25 weeks of gestation (mean 19.12 weeks), representing a relatively high level of cffDNA, in which the paternally inherited fetal SNPs, masked by the background of maternal DNA, can be definitely detected through allele-specific real-time PCR, indicating that the sensitivity of the molecular test is able to meet the demands of NIPD for clinical samples. Due to the limited quantities of cffDNA isolated between 8 and 12 weeks of gestation in early pregnancy we believe that it will be important to further enhance detection of SNPs in the application of this technology. Our method could be improved for non-invasive exclusion of fetal large deletions in monogenic disorders by inclusion of an additional step of selective enrichment of cell-free fetal DNA in maternal plasma by size fractionation. Moreover, we would enhance the accuracy of our non-invasive prenatal exclusion of homozygous a-thalassaemia by using SNPs that lie outside the (—SEA) deletion breakpoints as controls in early pregancy.In conclusion, this is the first large scale study reporting a contribution of detecting paternally inherited normal alleles within deletion breakpoints fromα0-thalassemia patient maternal plasma samples. Based on our results, which made use of a combination of multiplex-PCR based mini-sequencing and allele-specific real-time PCR, our strategy could be directly applicable for noninvasive prenatal prediction of homozygous a-thalassaemia by maternal plasma DNA analysis. The advantage of detecting these paternal SNPs within deletion breakpoints in deletional Mendelian disorders is that their presence could be used to exclude pregnancies at risk for these disorders, thereby obviating the need for an invasive prenatal diagnostic procedure. Furthermore, a decrease of these risk-associated procedures could be achieved for diagnosis of these common regional genetic disorders.