| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Molecular Diagnostics and Genetics |
1 Department of Pediatrics and 2
Institute for Clinical Research and Health Policy Studies, Tufts-New England Medical Center, Boston, MA.
3 Department of Pathology, Women and Infants Hospital, Providence, RI.
aAddress correspondence to this author at: Division of Genetics, Department of Pediatrics, Tufts-New England Medical Center, 750 Washington Street, Box 394, Boston, MA 02111. Fax 617-636-1469; e-mail kjohnson{at}tufts-nemc.org.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: We obtained 10 mL of fresh AF supernatant from women carrying euploid fetuses (n = 39) and aneuploid fetuses (n = 4). To test the effects of storage and karyotype, samples from euploid fetuses (n = 19) and aneuploid fetuses with trisomies 21 (n = 16), 18 (n = 9), or 13 (n = 3); triploidy (n = 4); or monosomy X (n = 2) were frozen at –80 °C. AF cffDNA was characterized by real-time quantitative PCR amplification of glyceraldehyde-3-phosphate dehydrogenase, gel electrophoresis, and analysis of the DNA fragmentation signature.
Results: We observed a significant correlation of concentration with gestational age for fresh AF cffDNA from euploid fetuses (R2 = 0.77, P <0.0001) but not for frozen cffDNA (P = 0.63). The median amount of cffDNA in frozen euploid samples was significantly lower than in fresh samples (P <0.0001). After adjustment for gestational age, there was a statistically significant decrease in the median amount of cffDNA in frozen aneuploidy samples compared with frozen euploid samples (P = 0.0005). Analysis of the cffDNA size distribution showed different and qualitatively unique patterns for each karyotype.
Conclusions: Gestational age, karyotype, and sample storage time affect concentrations and fragment size of AF cff DNA. These effects may be attributable to fundamental differences in tissue sources, excretion modes, or kinetic pathways. Characteristic signature patterns for each common aneuploidy offer the possibility of using DNA fragmentation analysis as a means of triaging AF samples.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In 2001, we first demonstrated the presence of cffDNA in AF and showed that AF contains larger quantities (100–200-fold) of cffDNA per milliliter than does maternal plasma or serum (10). Subsequently, we showed that cffDNA in AF supernatant could be successfully hybridized to comparative genomic hybridization microarrays to detect aneuploidy (9). Technical improvements in our original extraction protocol resulted in statistically significant higher yields, allowing further study of the biological features and characteristics of AF cffDNA (11).
No study has addressed the biochemical properties of cffDNA in AF, whereas studies of cffDNA in maternal plasma have shown that circulating fetal DNA sequences are smaller than maternal DNA sequences, on the order of <300 bp (12)(13), a property that has been used to increase the yield of fetal DNA extracted from maternal samples to permit noninvasive prenatal diagnosis of ß-thalassemia (14).
We hypothesized that AF cffDNA would have different biophysical properties than cffDNA in maternal plasma. Second trimester AF is composed predominantly of fetal urine. We speculated that passage of cffDNA through the fetal kidneys might affect its qualities. Furthermore we examined additional variables such as karyotype, gestational age, and storage at –80 °C.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
) were frozen at –80 °C. Other than freezing, all samples were processed identically. DNA extraction was performed with the QIAamp DNA Maxi Kit (Qiagen) in combination with 40 mL of viral lysis buffer supplemented with nucleic acid carrier, and 40 mL of 100% ethanol as previously described (11). Final elution was performed with 2 mL of acetate-EDTA buffer. The extracted DNA was stored at –80 °C until further processing. Before storage, the purity of the eluted DNA was assessed with an Eppendorf Biophotometer.
|
To measure the amount of the extracted cffDNA, real-time quantitative PCR analysis was performed in triplicate using the 7700 Sequence Detector (Applied Biosystems), with the mean result of the 3 reactions used for further calculations. Amplification of the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH)2 locus was performed on cffDNA in AF supernatant, as described previously (15). Reactions were set up in a 50 µL volume, using 25 µL of Universal Mastermix and 5 µL of extracted DNA. Primers and probes were used at final concentrations of 300 and 200 nmol/L, respectively. Data were analyzed with Sequence Detection System software, version 1.6.3 (Applied Biosystems). Two samples with no template DNA were included on each reaction plate as negative controls. Cycling conditions for all reactions consisted of a 2-min incubation at 50 °C to allow for UNGerase activity, an initial denaturation step of 95 °C for 10 min, and then 40 cycles at 95 °C for 15 s and 60 °C for 1 min. The results were expressed as genome equivalents per milliliter with a conversion factor of 6.6 pg of DNA per cell (16).
dna electrophoresis and staining
Standard methods were used for the preparation of the 1% agarose gels, using 1x TAE buffer (40 mmol/L Tris acetate, 2 mmol/L Na2EDTA.2H2O, pH 8.5); 20 µL of the eluted, nonamplified cffDNA was added and thoroughly mixed with 5 µL of loading buffer (Blue Juice, Invitrogen), consisting of 650 mg/L sucrose, 10 mmol/L Tris-HCl (pH 7.5), 10 mmol/L EDTA, and 30 mg/L bromphenol blue, which comigrates with 
0.5 kb DNA fragments. To prevent fragments of short length from running off of the gel, we restricted bromphenol blue dye front migration to within several centimeters of the edge of the gel. Double-stranded cffDNA from each sample was separated by electrophoresis in 2 parallel electrophoresis systems (Owi Separation Systems). The gels, 7–8 mm thick, were run from 2.9 V/cm for 60 min, followed by 5.9 V/cm for 60 min, up to 8.75 V/cm for 35 min, using step-wise increases in voltage to improve resolution. For fragment size estimation, a 1-kb extension ladder (Invitrogen) was used. The ladder consisted of 8 bands containing multiples of a 1018-bp DNA fragment, vector bands of 506/517 bp and 1636 bp, and additional bands of 5, 10, 20, and 40 kb. After electrophoresis, the gels were incubated for 20 min, with rocking, in SYBR Gold staining solution (Invitrogen) that had been diluted 1:10 000 fold in 1x TAE buffer.
gel imaging and data analysis
Photographic images were taken during transillumination of the gel at 300 nm (Ultra Lum, model UVB-10) with a Polaroid Model QSP camera with an exposure time of 1 s, aperture of 4.5, and film designed for capturing high-quality electrophoresis images (Polaroid 667 Film ISO 3000/DIN 36). The images were saved in .tif format after scanning with a Hewlett Packard ScanJet 6300c using PrecisionScan Pro software and then transferred to GeneTool software (Syngene). After importing the .tif files into GeneTool, the tracks on the gel were analyzed automatically. For calibration, data from the 1-kb extension ladder were used. Data were created by repeatedly measuring the sum of the pixel values along the band representing each sample (i.e., raw volume). The number of measurements for each sample ranged from 515–536. The gel running distance was expressed as retention factor (Rf) distance, which is equivalent to relative mobility. Relative mobility was defined as the distance migrated by a band divided by the distance migrated by the dye front. The Rf values lie between 0 and 1, with lower Rf values representing larger DNA fragments.
statistical analyses
Descriptive statistics, including medians, 25th, and 75th percentile ranges were generated for all study variables. The nonparametric Kruskal–Wallis test was used to compare unadjusted GAPDH concentrations between trisomy 18, trisomy 21, and euploid pregnancies. Spearman correlation analysis was carried out between GAPDH concentrations and gestational age. Because of the small sample sizes of the other aneuploid samples [trisomy 13 (n = 3), triploidy (n = 4), and monosomy X (n = 2)], we did not perform the separate statistical analysis, although the descriptive characteristics were provided.
The effect of interaction between the karyotype and gestational age on the logarithmically transformed GAPDH concentrations was assessed using multiple linear regression analyses. All statistical analyses were performed using SAS/STAT software (SAS Institute, Inc.). Statistical significance was assigned for P values <0.05.
statistical analysis of fragmentation signature
Fragmentation signatures were analyzed using the trapezoid methods. Area under the curve (AUC) was calculated separately for each sample using all available signal readings. Log-transformed total AUC and AUC for different DNA molecular weights (i.e., distances run by half of the cffDNA fragments through the gel) for frozen euploid and aneuploid samples, as well as fresh and frozen euploid samples, were compared with linear regression analysis after adjustment for the initial amount of PCR product and gestational age. Correlation between AUC and the initial PCR product was assessed with Spearman correlation analysis, simultaneously controlling for gestational age
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
fresh euploid af samples
The data showed that the concentrations of cffDNA from fresh euploid AF samples correlated significantly with gestational age (R2 = 0.77, P <0.0001). Median amounts of cffDNA from fresh AF samples are presented in Table 1
.
fresh versus frozen euploid samples
Data from 19 AF samples from euploid singleton fetuses, which were subsequently frozen, suggested a statistically significant influence of storage time. The median amount of cffDNA in frozen euploid samples, measured by GAPDH, was significantly lower than the median amount in fresh euploid samples (P <0.0001, adjusted for gestational age) (see Table 1
). However, no linear relationship was observed between storage time and concentrations of cffDNA in frozen euploid samples (P = 0.19).
In contrast to fresh euploid samples, for which gestational age is a statistically significant predictor of cffDNA concentrations (P <0.0001), cffDNA concentrations in frozen samples were not statistically associated with gestational age (P = 0.63). However, a significant storage time/gestational age interaction was observed (P = 0.02).
euploid versus aneuploid samples
After adjustment for gestational age, a statistically significant decrease in the median concentration of cffDNA was observed in the subgroups of frozen aneuploid samples compared with frozen euploid samples (P = 0.0005).
The concentrations of cffDNA from aneuploid samples correlated marginally with gestational age in all combined aneuploid samples (R2 = 0.32, P = 0.08). Statistically significant correlations were not found for trisomy 21 (R2 = –0.03, P = 0.93) or trisomy 18 (R2 = 0.04, P = 0.94), although this lack of correlation may be due to small sample sizes.
The small number of fresh aneuploid samples (n = 4) included 1 trisomy 21, 1 triploidy, and 2 monosomy X samples. Although the small number precludes statistical analyses for each aneuploidy type, the median amount of cffDNA in fresh aneuploid samples is 2.3 times higher than that of frozen aneuploid samples [4600 vs 1714 genome equivalents (GE)/mL]. This difference is consistent with that seen in euploid samples [i.e., 2.6 times higher in fresh (1424 GE/mL) vs frozen samples (606 GE/mL)].
fragmentation signature qualitative and quantitative analysis
After gel electrophoresis, scanning, and software analysis, we observed unique qualitative patterns for euploid and each aneuploidy, which we termed "fragmentation signatures" (Figs. 1 through 4
). Figs. 1
and 2
show fragmentation signatures for fresh and frozen euploid samples (stored for 4–6 months), respectively, and Figs. 3
and 4
show fragmentation signatures for trisomies 21 and 13, respectively. For each karyotype group, these patterns were remarkably consistent in different individual samples.
|
|
|
|
To perform quantitative analysis, we developed a measurement in which the discriminative fragmentation signatures of fresh and frozen euploid and aneuploid samples were expressed by the distance (Rf) to which half of the cffDNA fragments have migrated through the gel. Measured values differed significantly between fresh euploid and frozen euploid AF samples (P = 0.002) and among all frozen aneuploidy AF samples (P = 0.0004) (Table 2
).
|
The median AUCs for DNA fragments of different lengths was determined for fresh and frozen euploid AF samples, as well as for aneuploid AF samples. Statistical analysis after adjustment for the initial cffDNA concentration showed highly significant differences in AUC among fresh and frozen euploid and aneuploid samples, as estimated by real-time quantitative PCR analysis using GAPDH (overall P = 0.0003) (Table 2
). The results remained statistically significant after further adjustment for gestational age.
The molecular weight of the cffDNA fragments was significantly higher for fresh euploid samples compared with frozen aneuploid and euploid samples, as determined by analyzing the median percentage of the estimated concentration of cffDNA that migrated in the first 20% of the gel running distance (Rf <0.2) (Table 2
). In addition to a significant overall difference in this measurement among all AF samples (P = 0.0075), a highly significant loss of large fragments was observed in the frozen euploid samples compared with fresh euploid samples (P = 0.0006, unadjusted for gestational age). In addition, a loss of small fragments can be observed by the reduction in peak height at Rf >0.8 between fresh euploid samples (Fig. 1
) and frozen euploid samples (Fig. 2
).
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our results show that there is a unique and consistent qualitative pattern of AF cffDNA fragments in euploid and aneuploid fetuses. The fragmentation signature, which can be demonstrated on standard agarose gels, represents differences in the proportions of different sizes of cffDNA fragments and suggests specific pathognomonic kinetic mechanisms. These results may have clinical application in the rapid triaging of AF. Furthermore, the ability to statistically analyze the data from each sample provides a novel tool for a predictive model of aneuploidy in prenatal diagnosis.
The specific fragmentation signatures may be explained by different apoptotic pathways and/or variable activation of necrotic pathways. DNA degradation is considered to be a defining hallmark of apoptosis. Apoptotic fragmentation is commonly a 2-step process in which DNA is first cleaved into fragments of 50–300 kb, termed high–molecular-weight DNA fragmentation. Subsequently, DNA is cleaved between nucleosomes into smaller fragments of oligonucleosomal size, also described as low–molecular-weight DNA ladders (17).
Fresh euploid AF samples showed a significantly higher percentage of larger DNA fragments than frozen euploid samples, whereas aneuploid samples, such as those from trisomy 21 pregnancies, featured smaller fragments, irrespective of sample storage time. This comparison is confounded by the fact that all of the aneuploidy samples were previously frozen, and direct comparison of euploid and aneuploid samples at identical storage times would be necessary to fully assess differences in fragment length. Indeed, the change in the relative proportion of fragment sizes in the fragmentation signatures between fresh and frozen euploid samples shows that freezing alone contributes to cffDNA fragmentation. To more fully understand the effect of freezing alone on the quantity and quality of cffDNA in AF, the effect of freezing must be addressed in future studies of larger numbers of samples frozen for various periods of time. Nevertheless, we hypothesize that, as in cancer cell lines, in which an asynchronous apoptotic process leads to a decrease in fragment size, the same mechanism can explain the observed differences between the euploid and aneuploid samples (18). Furthermore, the activation of cysteine-dependent aspartate-specific proteases (known as caspases) by upstream pathways, triggered by the underlying karyotype, may initiate apoptosis or enzymatically cleave cellular components.
Up- or down-regulation of genes involved in apoptosis may play an important role in trisomy 21 and may affect detectable cffDNA concentrations. ETS2, a member of the ETS family of transcription factors, which have been proposed to have important functions in immune responses, cancer, and bone development, is located on chromosome 21 (21q22.3) (19). This gene is overexpressed in brains and fibroblasts of individuals with trisomy 21. Overexpression in some of the trisomy 21 samples may lead to an increase of the p53-dependent apoptosis pathway, as seen in prior studies (20). On the other hand, alternative sources of cffDNA release, such as necrosis, may also contribute to the varied and gestational-age independent concentrations of AF cffDNA in aneuploid fetuses. The distinction between apoptosis and necrosis is not always well defined, and in many instances these 2 models may be regarded as a continuum of cell death.
Other pathways, such as necrosis or active secretion, may also contribute to the excretion of cffDNA. Evidence suggests that compared with the release of cffDNA in euploid fetuses, in cases of aneuploidy, nonphysiologic cell death from primary stress signals or secondary to apoptosis (21) contributes, in greater proportion, to the release of cffDNA. Therefore this mode of cffDNA release may contribute substantially to different fragmentation signatures and cffDNA concentrations in abnormal karyotypes.
Specific pathologic processes occurring in fetal organs that are in direct contact with AF may also affect the fragment distribution of cffDNA. Interestingly, 2 distinct fragmentation signatures were observed in the trisomy 18 samples. We speculate that this finding may be explained by differences in the extent of renal dysplasia, a common feature in trisomy 18.
Sample stability at –80 °C is an important variable in basic and clinical research, which often relies on archived samples. Before this study, no data were available about the effect of storage on cffDNA in AF. CffDNA in maternal plasma is reported to be stable at –20 °C for 
4 years (22). Our group previously demonstrated that cffDNA concentrations is maternal plasma undergo storage-related decline of –0.66 GE/mL per month (23). Our present results showed that storage of AF at –80 °C significantly decreases the yield and the integrity of cffDNA. No linear relationship was observed between storage time and concentrations of cffDNA, a finding that suggests that the non-particle-associated form of cffDNA degrades more rapidly than the particle-associated form (24).
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Human genes: GAPDH, glyceraldehyde-3-phosphate dehydrogenase and ETS2, v-ets erythroblastosis virus E26 oncogene homolog 2 (avian).
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
