Mass Spectrometry–Based Detection of Hemoglobin E Mutation by Allele-Specific Base Extension Reaction

The discovery of circulating fetal nucleic acids in maternal plasma has opened up exciting possibilities for noninvasive prenatal diagnosis(1)(2). The recent development of the mass spectrometry-based single-allele base extension reaction (SABER) protocol has enabled sensitive differentiation of fetal-specific alleles down to a single-nucleotide level(3)(4)(5). In this report, we intended to develop a mass spectrometry–based method for the noninvasive prenatal diagnosis of the hemoglobin E (HbE) mutation. Unexpectedly, the lack of specificity of SABER for the HbE mutation was discovered during assay development, and systematic optimization on an artificial model has been carried out. This development has resulted in an alternative protocol, the allele-specific base extension reaction (ASBER).

HbE disease is an autosomal recessive hemoglobinopathy caused by a (GAG→AAG) missense mutation in codon 26 of the β-globin gene(6). It is the most common thalassemic hemoglobinopathy in Southeast Asia(7). Although homozygotes of HbE are mildly affected by the mutation, compound heterozygotes of the HbE and the β-thalassemia mutation will result in severe anemia (HbE/β-thalassemia). Therefore, detection of the HbE mutant allele is critically important in the prenatal diagnosis of thalassemia major in Southeast Asia. In noninvasive prenatal diagnosis of HbE, the absence of the paternally transmitted fetal HbE mutant allele in the plasma of a pregnant female carrier of a β-thalassemic mutation negates the fetal inheritance of HbE/β-thalassemia. Theoretically, this suggests that 50% of invasive procedures are unnecessary and can thus be avoided.

All samples in this study were collected with informed consent, and approval was granted by the institutional ethics committee. Venous blood (6 mL) was collected into EDTA tubes from each couple referred for prenatal diagnosis. Plasma and buffy coat were harvested from blood samples after a 1st centrifugation at 1600g for 10 min and a further microcentrifugation of plasma aliquot at 16 000g for 10 min as previously described(8). DNA was extracted from buffy coat and 800 μL of plasma by a Nucleon Blood DNA Extraction Kit (GE Healthcare) and a QIAamp Blood Mini Kit (Qiagen) with elution volume of 50 μL H₂O, respectively, according to the manufacturers’ recommendations.

The principle of standard homogenous MassEXTEND (Sequenom) and SABER protocols have been described previously(3). HotStar Taq polymerase (Qiagen) was used in the PCR at a final volume of 25 μL, containing 10 μL of plasma DNA and 0.2 μmol/L PCR primers (Integrated DNA Technologies). The thermal profile was 95 °C for 15 min for hot start, 45 cycles of denaturing at 95 °C for 20 s, annealing at 58 °C for 30 s, and extension at 72 °C for 1 min, followed by a final incubation at 72 °C for 3 min. PCR products were then treated with shrimp alkaline phosphatase (Sequenom) for 40 min at 37 °C to remove unreacted dNTPs. Base extension reaction was carried out with thermosequenase (Sequenom) on 10 μL of shrimp alkaline phosphatase–treated PCR product in a final reaction volume of 14 μL, with 1.54 μmol/L of the extension primer (Integrated DNA Technologies) and a terminator mix of dideoxy/deoxynucleotides, each at 64 μmol/L. The thermal profile consisted of 94 °C for 2 min, followed by a rapid thermocycling for 75 cycles at 94 °C, 52 °C, and 74 °C, all for 5 s. Products were then analyzed by the MassARRAY™ Analyzer Compact Mass Spectrometer (Brucker), a MALDI-TOF system. Details of the PCR primers, termination mix, and extension primers are listed in the Data Supplement that accompanies the online version of this Technical Brief at http://www.clinchem.org/content/vol53/issue12.

Thirteen HbE-negative pregnant participants (gestational age: 16–22 weeks) with male partners being HbE carriers were recruited. Genotypes of the couples and the fetuses were confirmed by analysis of buffy coat and cord blood samples, respectively, by the standard homogenous MassEXTEND protocol. Five of the fetuses were shown to be HbE positive.

Genotyping results of the SABER protocol on maternal plasma showed nonspecificity (see Fig. 1 in the Data Supplement). All 8 informative-negative pregnancies were misclassified as positive. To evaluate the important analytical parameters for the nonspecificity of SABER, an artificial model was established. Genomic DNA of a male HbE carrier was mixed with wild-type maternal genomic DNA to mimic the plasma DNA of an affected pregnancy. Direct dilution of wild-type maternal genomic DNA was used to mimic plasma DNA of normal pregnancy. Systematic evaluation revealed that the terminator concentration and the annealing temperature in the base extension reaction were critical for the reduction of false-positive detection in SABER. A 50-fold dilution of the standard terminator concentration and annealing temperature at 66 °C were shown to be optimal for both specific and sensitive detection of the HbE mutation in the artificial model (Table 1⇓ ).

We explored the development of an alternative protocol, ASBER. The 3′ end of the extension primer for ASBER was engineered to be complementary to the fetal mutant allele. Hence, primer extension of the maternal wild-type allele would be inhibited by 3′-primer-template mismatch (Fig. 1A⇓ ). Initially, the use of a single relevant terminator (ddA only) at standard concentration also revealed nonspecificity (data not shown). Thus, terminator dilution and the introduction of competing terminators (ddA, ddC, ddG, and ddT, each at 64 μmol/L) in the ASBER protocol were evaluated. Both the use of competing terminators and 20-fold dilution of the relevant terminator were found to provide specific genotyping results. However, the use of competing terminators appears to offer better detection sensitivity on the artificial model with different percentage mix of mutant DNA (Fig. 1B⇓ ). The design of ASBER with competing terminators was adopted to reanalyze the 13 maternal plasma samples, and the fetal genotypes were all correctly determined except 1 case of false-negative result.

In this study, we have shown that for certain mutational context, SABER might generate nonspecific results due to incorporation of the single terminator in the extension reaction mixture despite it being noncomplement to the template PCR product. Although the mechanism behind this nonspecificity or misincorporation is still unclear, we hypothesize that the intrinsic specificity of SABER is a subtle balance between the fidelity of the polymerase and the degree of excess of the single relevant terminator in the primer extension reaction mixture(9). The lower the polymerase fidelity and the higher the terminator concentration promote misincorporation. Based on this reasoning, optimization can be achieved by adjusting the terminator concentration as described above (Table 1A⇓ ). Others have shown that the polymerase fidelity could be improved by using a proofreading polymerase(9). However, this method did not offer any improvement in the specific detection of the trace amount of fetal DNA in maternal plasma in our preliminary findings (data not shown). On the other hand, we showed that nonspecificity of SABER was reduced by increasing the annealing temperature for base extension (Table 1B⇓ ). This result might be because the environment becomes less favorable for misincorporation of the noncomplementary terminator.

To overcome the intrinsic tendency of terminator misincorporation in SABER, we explored and developed ASBER based on the principle of allele-specific primer extension, which has been reported to offer successful mass spectrometry–based mutation genotyping(10)(11). However, such design has not been applied to specific fetal allele detection in noninvasive prenatal diagnosis. Compared with previous studies(10)(11), specific priming of the fetal-specific allele in noninvasive prenatal diagnosis is less favorable due to the overwhelming background of maternal allele and the lack of priming competition from alternative allele-specific primer. Yet we expect better specificity in ASBER than the conventional allele-specific PCR(12), because the ASBER extension products would not serve as templates for further amplification. Thus, a misprimed ASBER product, if it occurs, would not be exponentially propagated. Nevertheless, false-positive detection is also detectable in ASBER with the single relevant terminator and therefore required optimization. False-positive detection was not observed in ASBER with competing terminators. This finding further suggested that an overdominance of any single terminator in the extension mixture promotes misincorporation. Further comparison of detection sensitivity of the 2 ASBER protocols shows that improving specificity by reduction of the fractional concentration of a terminator (i.e., inclusion of 3 other types of dideoxynucleotide as competing terminators) is better than dilution of the relevant terminator (Fig. 1B⇓ ).

Reanalysis of the 13 pregnancies at risk for HbE/β-thalassemia with the ASBER protocol showed substantial improvement in specificity compared with the SABER protocol, with 1 false-negative result. The diagnostic performance of the HbE ASBER assay needs to be confirmed with larger-scale studies.

In summary, we have successfully developed ASBER, which confers advantages in terms of specificity and sensitivity over SABER for the detection of the HbE mutation. Using this approach, noninvasive prenatal diagnosis of the HbE mutation has been achieved with a specificity of 100% (8 of 8) and a sensitivity of 80% (4 of 5). To further improve the sensitivity of the assay, potential fetal allele enrichment steps—for example, PCR clamping by PNA probe and size fractionation—can be included(13). We speculate that the application of the ASBER protocol can be extended to other fields of circulating nucleic acids, for instance, detection of circulating tumor-specific DNA in cancer patients such as KRAS point mutations(14) and detection of circulating donor-specific DNA in transplant recipients(15).

• Download figure
• Open in new tab

Figure 1.

Comparison between the SABER and ASBER protocols.

(A), Schematic diagram comparing the principles of the SABER and ASBER protocols. The site of the point mutation is indicated in capital letters. The boxed letter indicates the type of dideoxynucleotide terminator in the base extension reaction. (B), grouped bar chart comparing detection sensitivity of different protocol designs on artificial plasma model with different concentrations. Eight separate PCRs were performed for a particular concentration of artificial mutant and wild-type allele mix followed by 8 separate base extension reactions. The base extension products were then analyzed in duplicate by the MassARRAY™ system. A reaction was scored positive if 1 of the duplicates was called by the TyperAnalyzer software (Sequenom). The sensitivity was then calculated as the percentage of positive detection out of the 8 separate PCRs.

• © 2007 The American Association for Clinical Chemistry