| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Molecular Diagnostics and Genetics |
1 Department of Chemical Pathology, 2
Centre for Emerging Infectious Diseases, 3
Li Ka Shing Institute of Health Sciences, and 4
Department of Obstetrics and Gynaecology, Chinese University of Hong Kong, Shatin, New Territories, Hong Kong SAR, China.
5 Harris Birthright Research Centre for Fetal Medicine, King’s College Hospital, London, United Kingdom.
aAddress correspondence to this author at: Department of Chemical Pathology, Rm. 38023, 1/F, Clinical Sciences Bldg., Prince of Wales Hospital, 30-32 Ngan Shing Street, Shatin, Hong Kong SAR, China. Fax 852-2194-6171; e-mail loym{at}cuhk.edu.hk.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: Placental-derived fetal-specific unmethylated maspin (SERPINB5) promoter sequences on human chromosome 18 were detectable in placental–maternal DNA mixtures and in maternal plasma by bisulfite modification followed by methylation-specific PCR (MSP) and primer extension. The ratios between the extension products of the 2 alleles were calculated for heterozygous placentas, placental–maternal blood cell DNA mixtures, and maternal plasma samples. The allelic ratios were compared between pregnancies carrying trisomy 18 and euploid fetuses.
Results: The epigenetic allelic ratios of all tested trisomy 18 samples deviated from the reference range obtained from euploid samples (placental DNA, 1.135 to 2.052; placental–maternal DNA mixtures, 1.170 to 1.985; maternal plasma, 0.330 to 3.044; without skew correction on the raw mass spectrometric data). A theoretical model was established and validated that predicted that a minimum of 200 copies of genomic DNA after bisulfite conversion were required for distinguishing euploid and aneuploid fetuses with confidence.
Conclusion: Epigenetic allelic ratio analysis of maternal plasma DNA represents a promising approach for noninvasive prenatal diagnosis of fetal chromosomal aneuploidies.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The use of circulating fetal DNA for prenatal diagnosis has been well described for paternally inherited variations, in scenarios such as fetal Rhesus D blood group (7) and ß-thalassemia major (8)(9) testing. In addition, quantitative aberrations of circulating fetal DNA have been reported in preeclampsia (10)(11), preterm labor(12), and certain fetal chromosomal aneuploidies (13). Despite increased circulating fetal DNA concentrations in trisomy 21 pregnancies, there is a large overlap in fetal DNA concentrations between healthy and aneuploid pregnancies (13)(14). Moreover, these applications are based on the detection of paternally inherited variations or Y chromosomal sequences and thus are useful for only a fraction of all pregnancies unless a large panel of polymorphic markers is used. Such limitations could be overcome by the discovery of sex- and variation-independent fetal DNA markers in maternal plasma. Epigenetic differences could be used to differentiate fetal and maternal DNA in maternal plasma. For example, differential DNA methylation could be used as the basis for methylation-specific approaches to detect only the fetal DNA and not its maternal counterpart in maternal plasma.
Recently, Chim et al. reported that the maspin (SERPINB5) gene, 1 with its epigenetic differences between the placenta and maternal blood cells, could serve as a universal fetal DNA marker (15). Researchers had previously found that maspin was expressed in the placenta (16), and its tissue-specific expression was regulated by promoter CpG methylation (17). Our previous results demonstrated that unmethylated maspin promoter (U-maspin) 2 sequences could be found in the placenta but not in maternal blood cells. The latter has previously been proposed to constitute the major source of maternal DNA in maternal plasma (18). Furthermore, the placental-derived U-maspin sequences were readily detectable in maternal plasma and were confirmed to be fetus specific. The discovery of the location of the maspin gene, on chromosome 18q21.33, provides a valuable opportunity to test the potential application of plasma epigenetic markers for the noninvasive prenatal detection of fetal chromosomal aneuploidies, using trisomy 18 (T18) as a model system.
T18 (Edward syndrome) is the second most common form of chromosomal aneuploidy, with an average incidence of 1 in 6000 pregnancies (19). We have previously demonstrated that fetal U-maspin DNA could be amplified with high specificity by use of methylation-specific PCR (MSP) even when such fetal DNA molecules were present among an excess of background plasma DNA of maternal origin (15). We hypothesized that chromosome 18 copy number could be inferred by assessing the allelic ratio of a single-base variation present within these MSP-amplified fetal DNA molecules, the so-called epigenetic allelic ratio (EAR) approach. In this paper, we present a theoretical and practical analysis of this approach.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
processing of blood and tissue samples
Maternal peripheral blood samples were centrifuged at 1600g for 10 min at 4 °C, and the plasma portion was recentrifuged at 16 000 g for 10 min at 4 °C (20). The blood cell portion was recentrifuged at 2500g, and any residual plasma was removed. DNA from the peripheral blood cells was extracted in accordance with the NucleonTM DNA Extraction Kit (GE Healthcare). DNA from maternal plasma was extracted with the blood and body fluid protocol of the QIAamp DNA Blood Mini Kit (Qiagen). DNA was extracted from 1.6 mL of plasma for the 3rd-trimester cases and at least 4.2 mL of plasma for all other cases. DNA samples from the CVS and placentas were extracted with the QIAamp DNA Mini Kit (Qiagen) according to the manufacturer’s tissue protocol.
genotyping of the maspin –156 single-base variation
Genotyping of the maspin –156 single-base variation was performed by use of a primer extension protocol (15) for genomic DNA from both the fetal and the maternal materials. Details of the procedures are described in Data 1A in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol52/issue12. Sequences of the PCR forward and reverse primers and the Homogenous MassEXTENDTM (hME) primer are listed in Table 1A in the online Data Supplement. The extension reactions were designed to generate products of distinct masses for each allele that were readily resolvable by matrix-assisted laser desorption and ionization time-of-flight mass spectrometry analysis (21). Data acquisition from the SpectroCHIP was done in the MassARRAY Analyzer Compact Mass Spectrometer (Sequenom). Mass data were imported into the MassARRAY Typer (Sequenom) software for analysis.
u-maspin ear determination
To determine the U-maspin EARs, we performed bisulfite conversion of the samples and then subjected them to MSP (22) followed by allele-specific primer extension. Bisulfite conversion, a procedure in which unmethylated cytosine was converted to uracil while methylated cytosine remained unchanged (22)(23), was performed with the MethylampTM DNA Modification Kit (Epigentek) according to the manufacturer’s instructions. MSP was performed as previously described (15), with a slight modification of the primer sequences (see Table 1B in the online Data Supplement), to specifically amplify the fetal-specific U-maspin (15). The extension primer was U-maspin specific (see Table 1B in the online Data Supplement). Details of the procedures are described in Data 1B in the online Data Supplement. The experimental procedures and the allelic ratio determination from the mass spectrometric data are outlined in Fig. 1
. All allelic frequencies were directly calculated and exported from the Typer (Sequenom) software. No skew correction of the peak frequency was used.
|
specificity of the u-maspin hme assay and establishment of euploid reference interval
We used 2 approaches to confirm the specificity of the U-maspin hME assay. We analyzed maternal blood cell DNA, which is hypermethylated at the maspin promoter and therefore should not be detected by the assay. We also analyzed DNA mixtures composed of 5% placental DNA and 95% maternal blood cell DNA from fetal–maternal pairs with different maspin genotypes to determine whether only the fetal alleles would be detected. After confirming the specificity of the U-maspin hME assay, we determined the EAR among euploid and T18 samples. First, DNA from 3 heterozygous 1st-trimester CVS and 28 heterozygous term placental tissues were used to establish a reference interval for euploid pregnancies. We next determined the allelic ratios of DNA from CVS or placental tissues of 7 T18 pregnancies.
theoretical modeling for ear determination
Because the ultimate aim of the present study was to transfer the established EAR protocol for noninvasive detection of fetal T18 from maternal plasma, we used a theoretical model to assess whether robust allelic ratio discrimination between euploid and T18 fetuses could be achieved at typical plasma DNA concentrations. If the total copy number of a particular DNA sequence is high enough, the DNA allelic ratio can be accurately determined by use of a single-base variation marker. At low DNA copy numbers, however, the allelic ratio fluctuates. The allelic ratio fluctuation can be modeled by binomial distribution because most variations are biallelic. The maspin –156 single-base variation is an A/C variation. Therefore, if N is the total number of fetal U-maspin molecules in the PCR before amplification and P is the theoretical A allele frequency, for euploid fetuses P should be 0.5, and for T18 fetuses, P should be either 
or 
, depending whether the extra chromosome carries a C or A allele, respectively. For each allelic ratio determination experiment, the probability of getting exactly K copies of A allele (or with an A allele frequency of K/N) is given by the probability mass function:
 |
The distributions of A allele frequency for euploid and T18 cases for which the total numbers of fetal U-maspin template molecules available for MSP amplification are 20, 50, 100, and 200 copies are shown in Fig. 2
. According to the theoretical model, a minimum of 200 copies of the U-maspin molecules would be required to achieve 97% diagnostic sensitivity and 97% diagnostic specificity to distinguish euploid from aneuploid cases. This calculation takes into account potential inaccuracies in allele frequency determination by the MassARRAY system (Sequenom), which has previously been shown to be 
3% (24). To test this model, we performed a serial dilution experiment on the placental–maternal blood cell mixtures to investigate the distribution of the EARs at low U-maspin copy numbers. DNA degradation is a well-known consequence of bisulfite conversion. We prepared DNA mixtures that initially contained 1000 to 20 000 copies of placental DNA and assumed that 95% of the DNA molecules would be degraded after bisulfite modification (25); hence, after bisulfite conversion, the "effective" number of placental DNA molecules in these mixtures would be 50, 100, 200, 500, and 1000 copies, respectively (see Data 2 in the online Data Supplement). The effective copy number is the residual number of DNA molecules after bisulfite conversion. The bisulfite-converted mixtures were then subjected to MSP and primer extension analysis. Two euploid and 2 T18 (1 with the AAC genotype and the other with the ACC genotype) cases were used in the serial dilution experiment. For each mixture, analyses were performed in quadruplicate at each template fetal DNA concentration to evaluate the degree of spread of the ratios.
|
ear determination in maternal plasma
As a final evaluation, we assessed the feasibility of applying the EAR approach to fetal DNA analysis in maternal plasma. Before the direct assessment of maternal plasma, placental and maternal blood cell DNA mixtures in a 5:95 proportion were prepared to simulate the fetal DNA concentrations and proportions in maternal plasma. DNA mixtures involving placental DNA from euploid (n = 16) or T18 (n = 6) heterozygous fetuses were prepared and analyzed. We next proceeded to the testing of maternal plasma. Eight cases of predelivery maternal plasma (1.6 mL) with the corresponding postdelivery samples were subjected to the U-maspin EAR determination. Furthermore, because TOP for T18 pregnancies was usually performed during the first or second trimester, the U-maspin hME assay was performed on gestationally matched (15 to 18 weeks) maternal plasma from 2 euploid and 2 T18 pregnancies. At least 4.2 mL of 2nd-trimester maternal plasma was used for EAR determination because fetal DNA concentration in maternal plasma was expected to be lower in early gestation (5).
statistical analysis
Statistical analyses were performed with SigmaStat 3.0 software (SPSS).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
specificity of the u-maspin hme assay
When maternal blood cell DNA was analyzed by the U-maspin hME assay, no primer extension products were detected (Fig. 3
, A and B). These results, indicating insensitivity to the detection of methylated molecules in maternal blood cells, confirmed the specificity of the assay (15). The genotypes revealed by the U-maspin hME assay for the mixtures of placental and maternal blood cell DNA from the 15 fetal–maternal pairs with different genotypes were concordant with those of the CVS and placental DNA (see Table 2 in the online Data Supplement).
|
u-maspin ear determination
We used 31 heterozygous placental tissues from euploid pregnancies to establish a reference interval for the U-maspin EAR, as shown in Fig. 4A
. The reference interval, defined as 1.96 SD away from both sides of the mean U-maspin EAR in euploid placentas, was 1.135 to 2.052. Three of the euploid placentas fell outside this interval, giving a false-positive rate (misclassification as T18) of 9.7%. When we subjected the 7 heterozygous T18 placental samples to the same analysis we observed either a higher or a lower EAR than the reference interval for all cases (Fig. 4A
). A higher or lower ratio was expected when the T18 fetus had an AAC or ACC genotype, respectively. Thus, T18 cases were distinguishable from euploid cases by use of the EAR approach on placental DNA. Representative mass spectra comparing euploid and T18 samples are shown in Fig. 5A
.
|
|
theoretical modeling for ear determination
Results for the serial dilution on placental–maternal DNA mixtures are shown in Fig. 6
. The distribution for the 4 replicates from each sample remained tight from a placental DNA input equivalent to an original amount of 20 000 down to 4000 copies of the maspin gene before bisulfite conversion. From the plot, the scatter became wider at 2000 copies of U-maspin, and the EARs for 1 T18 mixture showed a 2-fold difference among the 4 replicates at 1000 copies. In addition, at 1000 copies, the EAR of 1 replicate from a euploid DNA mixture was out of the reference interval established as described above. The EARs became unstable as the input DNA decreased, and poorer differentiation between the euploid and T18 cases would be expected. Assuming a 95% template destruction rate after bisulfite conversion, these results agreed with the mathematical prediction (Fig. 2
) of the need for at least 200 effective copies of template fetal DNA molecules for the differentiation of euploid from aneuploid pregnancies.
|
ear determination in maternal plasma
Mixtures of 5% placental DNA from either euploid or T18 heterozygous fetuses with 95% maternal blood cell DNA were prepared to simulate the concentrations and proportions of fetal DNA in maternal plasma (5). Mixed DNA samples were prepared with placental DNA from 16 euploid heterozygous fetuses. Eleven of these placental DNA samples were mixed with blood cell DNA from mothers homozygous for the maspin variation. The remaining mixtures involved heterozygous maternal blood cell DNA. Results from the former subgroup showed that the paternally inherited fetal-specific allele was detectable in all cases (see Table 2 in the online Data Supplement). We established a reference interval of 1.170 to 1.985 based on the EARs obtained from these 16 euploid DNA mixtures. EARs of the mixed DNA from 6 heterozygous T18 CVS or placentas deviated from the euploid reference interval (Fig. 4B
). No overlapping between the EARs for the euploid and the trisomy groups was observed in this mixing experiment. Representative mass spectra are shown in Fig. 5B
.
We next proceeded to direct testing of maternal plasma samples. On the basis of EARs obtained from 8 predelivery plasma samples and the pooling of 2 2nd-trimester samples, we established a plasma reference interval of 0.330 to 3.044. A plot for the maternal plasma data is shown in Fig. 4C
. No primer extension products were detectable from the same volume of postdelivery plasma from the same woman (Fig. 3C
). Because circulating fetal DNA concentrations are relatively low in early pregnancy (<100 genome-equivalents/mL of plasma) (5), a total of 8.7 mL of plasma from 2 2nd-trimester euploid pregnancies were pooled for the EAR determination. Maternal plasma from an AAC T18 pregnancy (4.8 mL of plasma) gave an EAR of 3.412, and that from an ACC T18 pregnancy (4.2 mL of plasma) gave a ratio of 0.255 (Fig. 4C
). Representative mass spectra are shown in Fig. 5C
.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We first evaluated and confirmed the placental (fetal) specificity of the U-maspin hME assay (15). The assays were designed in such a way that there was a dual level of specificity toward the detection of the U-maspin sequences, first during the PCR step and then during primer extension. We then assessed the feasibility of the EAR approach to distinguish euploid from trisomy DNA samples in placental DNA samples, placental–maternal blood cell DNA mixtures, and, finally, maternal plasma samples.
We calculated an EAR reference interval from 31 euploid placentas, 3 of which fell outside this interval. The degree of placental hypomethylation of the maspin promoter varies from sample to sample (15). If the number of U-maspin molecules in those 3 euploid placental tissues was particularly low, there might not be enough unmethylated molecules in the reaction mixture for a reliable EAR to be determined. The relationship between U-maspin concentration and the reliability of the resulting EAR is further evident from the theoretical model and serial dilution experiment.
Our serial dilution experiments illustrated that when the input DNA decreased, a wider spread of allelic ratios was observed. The wider data dispersion as a result of the increased imprecision of the assay at low input DNA amounts would reduce the magnitude of the ratio difference between euploid and aneuploid cases. The likelihood for false classification would therefore increase.
Results from the theoretical model were taken into consideration when we designed the protocols for maternal plasma EAR analysis. Because the maternal plasma fetal DNA concentrations in early pregnancies were lower than those at later gestation, we performed EAR determination with 1.6 mL predelivery plasma but at least 4.2 mL of 2nd-trimester plasma. Although we were able to determine the U-maspin EAR in all the analyzed plasma samples, the euploid reference interval obtained from the maternal plasma analysis had a much wider spread than that determined from the placental DNA samples or DNA mixtures. The wider dispersion of the plasma range was most likely attributable to the imprecision because of the low input DNA amount. Median fetal DNA concentration in late pregnancy was previously reported to be 250 copies/mL of maternal plasma (5). Thus, the use of 1.6 mL of predelivery plasma would mean that only 20 (250 copies/mL x 1.6 mL x 5% DNA remaining after bisulfite conversion) effective fetal DNA template molecules would be available for EAR determination. Considering our theoretical model, it is not surprising that the dispersion of the plasma ratios was much greater.
Our study and the theoretical modeling also highlight the difficulty of achieving robust euploid and aneuploid EAR discrimination in face of the problem of DNA degradation as a result of bisulfite modification. Bisulfite modification was reported to cause 84%–96% DNA degradation (25). The prenatal diagnosis for fetal chromosomal aneuploidy should ideally be performed as early in pregnancy as possible; however, it would be technically difficult to ensure the availability of 200 copies of fetal-specific U-maspin molecules for EAR determination during early pregnancy, in view of the low fetal DNA concentration in early gestation maternal plasma and the need for bisulfite modification.
In spite of the technical difficulties, our study has nonetheless demonstrated that a fetal epigenetic marker together with an informative polymorphism could be a potential strategy for the noninvasive detection of fetal chromosomal aneuploidy. The current study was essentially a demonstration of principle. The main reason for our small sample size was that the maspin –156 single-base variation was rare. Although we have seen this variation in the Chinese and African populations (unpublished data), we did not find the C allele among 129 Caucasian placentas genotyped, a difficulty that was further compounded by the incidence of T18 being 1 in 6000 pregnancies. From a clinical point of view, the discovery within other loci of more highly polymorphic single-base variations exhibiting differential methylation between the placenta and maternal blood cells would be a logical next step. From a technical point of view, one challenge is to develop methylation analysis methods that do not rely on bisulfite conversion and thereby avoid the associated DNA degradation effect. If and when all of these challenges are overcome, we envision that the EAR approach might become an attractive method for the noninvasive prenatal detection of several fetal chromosomal aneuploidies.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Nonstandard abbreviations: U-maspin, unmethylated maspin promoter; T18, trisomy 18; MSP, methylation-specific PCR; EAR, epigenetic allelic ratio; CVS, chorionic villus sample; TOP, termination of pregnancy; hME, Homogenous MassEXTENDTM.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  S. S.C. Chim, S. Jin, T. Y.H. Lee, F. M.F. Lun, W. S. Lee, L. Y.S. Chan, Y. Jin, N. Yang, Y. K. Tong, T. Y. Leung, et al. Systematic Search for Placental DNA-Methylation Markers on Chromosome 21: Toward a Maternal Plasma-Based Epigenetic Test for Fetal Trisomy 21 Clin. Chem., March 1, 2008; 54(3): 500 - 511. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. M. Dennis Lo and R. W. K. Chiu Noninvasive Prenatal Diagnosis of Fetal Chromosomal Aneuploidies by Maternal Plasma Nucleic Acid Analysis Clin. Chem., March 1, 2008; 54(3): 461 - 466. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. K. Tong, R. W.K. Chiu, T. Y. Leung, C. Ding, T. K. Lau, T. N. Leung, and Y.M. D. Lo Detection of Restriction Enzyme Digested Target DNA by PCR Amplification Using a Stem-Loop Primer: Application to the Detection of Hypomethylated Fetal DNA in Maternal Plasma Clin. Chem., November 1, 2007; 53(11): 1906 - 1914. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. W. Y. Chan, C. A. Graham, T. H. Rainer, N. Y. L. Lam, R. W. K. Chiu, K.-W. Chik, V. Lee, and Y. M. D. Lo Use of a Bone Marrow Transplantation Model System to Demonstrate the Hematopoietic Origin of Plasma S100B mRNA Clin. Chem., October 1, 2007; 53(10): 1874 - 1876. [Full Text] [PDF]
|
|
|
|
 |
|
 Y. M. D. Lo, F. M. F. Lun, K. C. A. Chan, N. B. Y. Tsui, K. C. Chong, T. K. Lau, T. Y. Leung, B. C. Y. Zee, C. R. Cantor, and R. W. K. Chiu From the Cover: Digital PCR for the molecular detection of fetal chromosomal aneuploidy PNAS, August 7, 2007; 104(32): 13116 - 13121. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. W. Bianchi Will Epigenetic Allelic Ratio Analysis Turn Prenatal Diagnosis of Trisomy 18 on Its EAR? Clin. Chem., December 1, 2006; 52(12): 2182 - 2183. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, predelivery maternal plasma; +, second trimester maternal plasma. The reference intervals for various sample types are marked with dotted lines across the plots.
