Background: Fragments of DNA from cells dying throughout the body are detectable in urine (transrenal DNA, or Tr-DNA). Our goal was the optimization of Tr-DNA isolation and detection techniques, using as a model the analysis of fetal DNA in maternal urine.
Methods: We isolated urinary DNA using a traditional silica-based method and using a new technique based on adsorption of cell-free nucleic acids on Q-Sepharose resin. The presence of Y chromosome–specific SRY (sex-determining region Y) sequences in urine of pregnant women was detected by conventional and real-time PCR using primers/probe sets designed for 25-, 39-, 65-, and 88-bp PCR targets.
Results: Method of DNA isolation and PCR target size affected fetal Tr-DNA detection. Assay diagnostic sensitivity increases as the PCR target is shortened. Shorter DNA fragments (50–150 bp) could be isolated by Q-resin–based technique, which also facilitated fetal Tr-DNA analysis. Using DNA isolated by Q-resin–based method and an “ultrashort” DNA target, we successfully detected SRY sequences in 78 of 82 urine samples from women pregnant with male fetuses (positive predictive value 87.6%). Eleven of 91 urine samples from women pregnant with female fetuses produced SRY false-positive results (negative predictive value 95.2%).
Conclusions: Single-copy fetal DNA sequences can be successfully detected in the urine of pregnant women when adequate methods for DNA isolation and analysis are applied. Strong precautions against sample contamination with male cells and DNA are necessary to avoid false-positive results.
The discovery(1)(2)(3) of cell-free DNA in the bloodstream (so-called circulating cell-free DNA, or ccfDNA1 ) has led to intensive studies of its potential diagnostic application in tumor detection and monitoring(4)(5)(6), prenatal diagnostics(7)(8)(9), and monitoring of trauma(10) and stroke(11). The half-life of ccfDNA is approximately 15 min(12), and it was found that a portion of these circulating DNA fragments cross the kidney barrier and can be found in the urine (transrenal DNA, or Tr-DNA)(13)(14). The existence of Tr-DNA was proven by detection of Y chromosome–specific DNA sequences in urine of women with male fetuses(13), mutant K-ras in urine of patients with colorectal(13)(15)(16) or pancreatic(13) cancers bearing the same mutation, sequences of Mycobacterium tuberculosis in urine of patients with pulmonary tuberculosis(17), and Y chromosome–specific sequences in urine of female recipients of blood transfusion from male donors(13). In all of these studies, DNA was isolated from urine by silica-based methods. The isolated DNA can be resolved into 2 fractions by gel electrophoresis(13)(15). The high-molecular-weight fraction represents DNA from shed cells, whereas low-molecular-weight DNA fragments (150–200 bp) contain most of the Tr-DNA(15).
The amount of specific Tr-DNA sequences isolated from urine may be low, owing to kidney filtering, nuclease activity, loss during purification, etc. In the studies cited above, this problem was overcome by the use of highly sensitive methods, such as nested PCR. Additionally, the assays of fetal DNA in maternal urine used the highly repeated DYZ12 (human Y- chromosome specific repeated DNA family) sequences to facilitate detection. Later efforts by several groups to reproduce the original findings using single-copy SRY (sex-determining region, Y-linked) or TSPY (testis-specific protein, Y-linked; about 10 copies/cell) sequences and conventional or real-time PCR resulted in low test detection rate or failed entirely(18)(19)(20)(21). We applied 2 approaches to improve the diagnostic sensitivity of Tr-DNA analysis, create a reliable test for prenatal sex detection of a fetus, and further develop the basic principles of the Tr-DNA technology in general.
First, we used a new technique for the purification of Tr-DNA based on the binding of cell-free urinary nucleic acid or nucleoproteins to a Q-Sepharose anion-exchange resin, followed by elution with LiCl. This procedure effectively concentrates DNA and partially purifies it, destroying DNA–protein bonds during elution. Electrophoresis of nucleic acids isolated by this method revealed the presence in urine, in addition to the fractions described above, of even shorter DNA fragments ranging from 50 to 150 bp(22). This finding indicated that focusing the detection efforts on very short DNA sequences could increase the diagnostic sensitivity of Tr-DNA assays. Second, we developed several “ultrashort” target PCR techniques and used them for Tr-DNA analysis. Here we describe the results obtained using a combination of these 2 approaches for sex detection of a fetus by PCR amplification of Y chromosome–specific SRY sequences from maternal urinary DNA.
Materials and Methods
urine sample collection
We collected 173 urine samples from pregnant women at Eastern Virginia Medical School under an institutional review board–approved study according to federally regulated guidelines for research involving human subjects. These women were not instructed in the precautions necessary for the proper sample collection to avoid potential contamination with male cells/DNA. However, before sample collection a brief survey was taken to exclude women who had intercourse during the preceding 24 h. The sex of fetuses was determined by ultrasound examination and subsequently verified after delivery; 91 and 82 women were confirmed to have been pregnant with female and male fetuses, respectively. At the time of urine sample collection, 15 women were in the first trimester (8 male, 7 female fetuses), 124 women were in the second trimester (61 male and 63 female fetuses), and 34 women were in the third trimester (13 male and 21 female fetuses). A limited number of urine samples were collected from local pregnant volunteers after special instructions on avoiding male contact. All study participants signed informed consent documents.
The 50–100 mL urine (routinely obtained) was supplemented with EDTA-Na2 up to 10 mmol/L final concentration, divided into aliquots, and stored at −80 °C before being shipped on dry ice for analysis. Received samples were archived at −80 °C for up to 2 years.
Urine samples were thawed to room temperature and mixed by gentle inversion. Two protocols were used for urinary DNA purification.
We placed 10 mL urine in a 50-mL conical centrifuge tube and diluted it with 10 mL nuclease-free water (Ambion). We added 200 μL of Q-Sepharose resin slurry (GE Healthcare) to each sample, and the tubes were rotated at room temperature for 30 min on a rolling drum. The Sepharose resin was pelleted by centrifugation at 1800g for 5 min at ambient temperature, and the supernatant was removed by vacuum aspiration.
We resuspended the pelleted resin in 1 mL of 0.3 mol/L LiCl/10 mmol/L sodium acetate (pH 5.5) and transferred it to a Bio-Rad microspin column. The resuspension/washing buffer was removed by 1-min 800g centrifugation at ambient temperature. The resin was washed with an additional 2.0 mL of 0.3 mol/L LiCl/10 mmol/L sodium acetate (pH 5.5) by 2-min 800g centrifugation at ambient temperature.
Tr-DNA was eluted from the Q-Sepharose resin with 670 μL of 2 mol/L LiCl/10 mmol/L sodium acetate (pH 5.5) using a 3-min 800g centrifugation at ambient temperature. The eluate was added to 2.0 mL 95% ethanol and gently mixed. This mixture was applied onto a Qiagen QIAquick column by 800g centrifugation for 1 min. The column was washed with 1 mL of 2 mol/L LiCl in 70% ethanol by 800g centrifugation for 1 min and further washed with 1 mL of 75 mmol/L potassium acetate (pH 5.0) in 80% ethanol by 800g centrifugation for 1 min. The column was dried by centrifugation in a microfuge (approximately 20 000g) for 3 min. DNA was eluted with 106 μL elution buffer (EB; Qiagen) using 2-min centrifugation in a microfuge (approximately 20 000g) and stored at −20 °C.
We added 10 mL of 6 mol/L guanidine isothiocyanate (GuSCN) (Ambion) to 10 mL urine placed in a 50-mL conical centrifuge tube. We then added 200 μL Wizard resin slurry (Promega) to each sample, and the tubes were rotated at room temperature for 1 h on a rolling drum. We attached the required number of Wizard minicolumns with adapters to a vacuum manifold, each with a 20-mL syringe barrel with the plunger removed.
We pipetted the urine/resin mixture into the 20-mL syringe barrel. The barrel/column was washed with 10 mL of 3 mol/L GuSCN followed by 10 mL of 80% isopropanol wash (Sigma Aldrich). The syringe barrel was removed and discarded. The minicolumn was washed with 200 μL of 80% isopropanol and further washed with 280 μL of 95% ethanol. The minicolumn was removed from the vacuum manifold and dried using a 3-min centrifugation in a microfuge (approximately 20 000g). Tr-DNA was eluted with 106 μL hot (preheated to 60 °C) nuclease-free water (Ambion) using a 2-min centrifugation in a microfuge (approximately 20 000g). Tr-DNA was stored at −20 °C.
We used primers PD-SRY30-F and PD-SRY30-R (Table 1⇓ ) in conventional PCR assays. After a 10-min uracil DNA N-glycosylase (UNG) treatment, the reactions were carried out for 40 cycles in 25-μL volumes in the presence of 300 nmol/L each primer, 3 mmol/L MgCl2, 0.05 U/μL JumpStart Taq DNA Polymerase (Sigma), 10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl, and 200 μmol/L of each dNTP (deoxynucleoside triphosphate). Each cycle consisted of 30-s denaturation phase at 95 °C, 10-s annealing at 60 °C, and 10-s extension at 72 °C. Reaction products were subjected to electrophoresis in 10% polyacrylamide gel and stained with SYBR Gold (Invitrogen) according to manufacturer’s instructions.
In addition to the SRY band, these primers also generate a 100-bp band with both male and female DNA. Sequencing of this band revealed that the 100-bp product originates from human chromosome 2 (GenBank HS2 22329). It was produced consistently for all samples in the absence of male genomic DNA, and so we used this band as an internal control for DNA isolation and the PCR assay. Samples were characterized according to the following rules:
Samples that generate a 50-bp product (30-bp SRY target) were flagged as male fetuses.
Samples lacking a 50-bp product but generating the 100-bp product were flagged as female fetuses only.
Samples lacking evidence of a 50-bp or 100-bp product were flagged as invalid due to PCR inhibitors or failure of DNA isolation or amplification steps.
We designed and optimized real-time PCR assays for 4 targets of differing lengths within the SRY gene (Table 1⇑ ). Of these, the assays for the longer 65- and 88-bp targets were standard TaqMan dual-labeled probe real-time PCR assays. For the detection of the shorter 25- and 39-bp targets, which are too short to be used directly in a standard TaqMan assay, we designed a novel 2-stage, single-tube, real-time PCR method. In the first stage of this assay, the participating forward primer has a special 5′ tail (Fig. 1⇓ ), resulting in an intermediate product that is significantly longer than the original target sequence. The intermediate product is long enough to be used as a template in a TaqMan real-time PCR in the second stage. The special 5′ tail of the first-stage forward primer includes (a) a sequence homologous to the second-stage forward primer, (b) a stem-forming sequence that reduces the primer’s affinity for the intermediate product in favor of the second-stage forward primer during the second stage, and (c) a 5′-methylisocytosine (iso-dC) base, which, in the absence of 2′-deoxy-isoG triphosphate (iso-dGTP), blocks nascent strand elongation, resulting in the intermediate product lacking the stem structure. The stage progression of the reaction was triggered by increasing the temperature of the annealing/extension phase after a preset number of amplification cycles, thereby effectively excluding the first-stage primer from participation in the reaction, thus initiating stage 2.
Each real-time PCR assay was carried out in 25 μL of PCR buffer [10 mmol/L Tris-HCl (pH 8.3), 50 mmol/L KCl] in the presence of 0.05 U/μL JumpStart Taq DNA Polymerase (Sigma), 200 μmol/L deoxyadenosine triphosphate (dATP), deoxycytidine triphosphate (dCTP), deoxyguanosine triphosphate (dGTP) (each), 400 μmol/L deoxyuridine triphosphate (dUTP), following a 10-min treatment with 0.01 U/μL UNG. Other reaction conditions were optimized for each set of oligonucleotides, and therefore differ between the sets. The reactions for 25- and 39-bp targets contained 1.5 mmol/L MgCl2, 200 nmol/L first-stage forward primer, 700 nmol/L second-stage forward primer, 700 nmol/L reverse primer, and 100 nmol/L (for 25-bp target) or 200 nmol/L (for 39-bp target) TaqMan minor groove binder (MGB) probe. The 65-bp PCR target assay contained 3 mmol/L MgCl2, 300 nmol/L forward and reverse primer (each), and 250 nmol/L TaqMan MGB probe. The 88-bp assay contained 2 mmol/L MgCl2, 500 nmol/L forward and reverse primer (each), and 200 nmol/L TaqMan probe. TaqMan MGB probes were obtained from Applied Biosystems, and all other oligonucleotides from Integrated DNA Technologies Inc. The sequences of all oligonucleotides are listed in Table 1⇑ . We used 1, 10, 100, and 1000 genome equivalents of human male DNA (Promega) to construct the calibration curves.
The PCRs for 25- and 39-bp targets were carried out for an initial 10 cycles (stage 1), followed by 40 cycles of stage 2. Stage 1 cycles consisted of 30-s denaturation at 95 °C and 60-s annealing/extension at 47 °C (for 25-bp target) or 53 °C (for 39-bp target). Stage 2 cycles consisted of 30-s denaturation at 95 °C and 60-s annealing/extension at 57 °C (for 25-bp target) or 63 °C (for 39-bp target). Fluorescence was measured during stage 2 at the end of each cycle. The 65- and 88-bp assays were run for 40 cycles, consisting of 30-s denaturation at 95 °C and 60-s annealing/extension at 66 °C (for 65-bp target) or 60 °C (for 88-bp target). Fluorescence was measured at the end of each cycle.
Statistical analyses were performed using the SYSTAT 12 (Systat Software Inc.) and Excel 2003 (Microsoft Corp.) software packages for Windows. To characterize the utility of our method, we calculated its diagnostic sensitivity and specificity, as well as the positive and negative predictive value (PPV and NPV)(23). We estimated the overall correctness of the method by fitting a binomial distribution to the obtained data and comparing the P values to the chosen significance level.
In the first set of experiments, we compared the fetal DNA sequence detection in DNA isolated from maternal urine by the silica-based and new Q-resin–based technique. Primers PD-SRY30-F and PD-SRY30-R, designed to amplify a 30-bp target in the SRY gene, were used to detect this Y-chromosome–specific sequence in the urine of women pregnant with male (10 samples) or female (5 samples) fetuses. Fig. 2⇓ shows that male fetal DNA was successfully detected in all specimens processed by the Q-resin method but only 7 DNA preparations purified by the silica method. All urine samples from pregnancies with female fetuses yielded negative results.
Using 4 sets of primers and probes that amplified 25-, 39-, 65-, and 88-bp sequences within the SRY gene, we performed real-time PCR with DNA isolated from the same urine specimens of women pregnant with male fetuses by 2 techniques, based on silica or Q-resin adsorption. Fig. 3⇓ clearly demonstrates that both the method of DNA isolation and the PCR target size are important. First, the detection rate decreases as the length of PCR target increases. The increase of the target sequence size from 25 to 39 bp decreased the detection rate, and fetal DNA was completely undetectable in all samples with the 88-bp PCR target. Second, the isolation of DNA fragments shorter than 150 bp with the Q-resin–based technique increased the diagnostic sensitivity (the number of correctly detected male fetuses) when 25- and 39-bp sequences were analyzed. In addition, the number of SRY copies (range) detected using primers for the 25-bp target was higher in DNA isolated from same samples with the Q-resin–based method compared to the silica-based one: 196.5 (17.1–451.8) vs 26.3 (0–46.9) genome equivalents/mL urine, respectively. The 2 DNA isolation methods gave similar results with the 65-bp PCR target. The real-time PCR experiments with primers and probes for every target length were performed twice and yielded 100% concordance of the results.
Finally, to evaluate diagnostic sensitivity and specificity of fetal sex detection by analysis of Tr-DNA in maternal urine, we analyzed urine samples obtained from 173 pregnant women. The DNA was isolated using the Q-resin–based method and analyzed by conventional PCR for the presence of the 30-bp SRY target. Table 2⇓ shows that the SRY sequences were successfully detected in Tr-DNA from 78 of 82 women pregnant with male fetuses (PPV 87.6%), and 11 of 91 urine samples from women pregnant with female fetuses gave false-positive results (NPV 95.2%). Most likely this was a result of urine sample contamination with male DNA (see below). To evaluate the reproducibility of these results, we repeated the purification and the conventional PCR analysis for a randomly selected set of 15 of the 173 urine samples. The results obtained were identical to those from the initial experiments.
Thus, we correctly determined sex in 158 of 173 cases, an overall success rate of 91.3%. After further statistical analysis of this result, we estimate at 99% confidence level that the test accuracy is 85%–96%. Surprisingly, for 15 women in the first trimester (7–12 weeks of pregnancy, 8 male and 7 female fetuses) the test was 100% sensitive and 100% specific. Because of a limited number of first-trimester samples, however, the difference between the first and other trimesters is not statistically significant.
The data presented here provide new information on the properties of Tr-DNA, in particular the characteristics of fetal Tr-DNA in maternal urine. First, the higher detection rate of fetal sequence in DNA purified with Q-resin compared to DNA isolated by the silica method demonstrates that DNA fragments 50 to 150 bp long contain fetal Tr-DNA. This difference is seen with 25- and 39-bp PCR targets only. Thus, larger fragments of fetal Tr-DNA, detectable by real-time PCR directed at the 65-bp target, belong to DNA fractions isolated by both methods, most likely to 150- to 200-bp DNA fragments.
Second, although the shortest DNA fragments isolated by the silica method are about 150 bp long, the ability to detect the fetal sequences therein depends on the PCR target length over the range of 25–88 bp. Recently Chan et al.(24) came to the same conclusion comparing median Epstein-Barr virus DNA concentrations, measured by the 59- and 76-bp amplicons, in urine of nasopharyngeal carcinoma patients. The most plausible explanation for these results is the presence of single-strand breaks in the 150- to 200-bp fragments of Tr-DNA, which make amplifiable targets substantially shorter. It is likely that the frequency of such single-strand breaks is dependent on the activity of urinary nucleases and the time between DNA secretion into urine and urine collection, but our data do not provide sufficient information to draw any conclusions. It is also unknown whether ccfDNA in the bloodstream contains such single-strand breaks and very short (50- to <150-bp) DNA fragments, or whether they are formed after its secretion into the urine.
The PPV of fetal DNA detection in urine of women pregnant with male fetuses was 87.6%. One should take into account that testing was performed about 2 years after sample collection. It is possible that a higher assay PPV could be achieved with freshly collected specimens. Analysis of urine samples from 10 local pregnant volunteers gave 100% PPV (data not shown).
Every primer/probe set for the SRY gene used in this study was 100% specific when tested on purified male or female human genomic DNA. However, SRY sequences were detected in about 12.1% of the urine samples obtained from pregnant women carrying female fetuses. Male contact with the subjects of the study was not rigorously controlled for, which may explain the presence of these SRY sequences in the samples.
Our data demonstrate that fetal DNA can be successfully detected in the urine of pregnant women if adequate methods for DNA isolation and analysis are applied. The diagnostic sensitivity of the sex detection test is similar to that demonstrated by plasma DNA-based tests(25)(26)(27). However, the chances of urinary DNA contamination are higher than those for plasma DNA, and therefore effective precautions must be used.
Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 3 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; and (c) final approval of the published article.
Authors’ Disclosures of Potential Conflicts of Interest: Upon manuscript submission, all authors completed the Disclosures of Potential Conflict of Interest form. Potential conflicts of interest:
Employment or Leadership: E.M. Shekhtman, Xenomics Inc.; K. Anne, Xenomics Inc.; H.S. Melkonyan, Xenomics Inc.; D.J. Robbins, Xenomics Inc.; S.R. Umansky, Xenomics Inc.
Consultant or Advisory Role: S.L. Warsof, Xenomics Inc.
Stock Ownership: E.M. Shekhtman, Xenomics Inc.; H.S. Melkonyan, Xenomics Inc.; S.R. Umansky, Xenomics Inc.
Honoraria: None declared.
Research Funding: None declared.
Expert Testimony: None declared.
Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, or preparation or approval of manuscript.
1 Some of the primers have 5′-end extension sequences (tails) not complementary to the targets, resulting in PCR products being longer than the respective targets.
1 Y chromosome–specific SRY sequences were detected by conventional PCR. PPV, 87.6%; NPV, 95.2%; sensitivity, 95.1%; specificity, 87.9%.
↵1 Nonstandard abbreviations: ccfDNA, circulating cell-free DNA; Tr-DNA, transrenal DNA; GuSCN, guanidine isothiocyanate; UNG, uracil DNA N-glycosylase; iso-dC, 5′-methylisocytosine; iso-dGTP, 2′-deoxy-isoG triphosphate; dNTP, deoxynucleoside triphosphate; dATP, deoxyadenosine triphosphate; dCTP, deoxycytidine triphosphate; dGTP, deoxyguanosine triphosphate; dUTP, deoxyuridine triphosphate; MGB, minor groove binder; PPV, positive predictive value; NPV, negative predictive value.
↵2 Human genes: DYZ1: human Y-chromosome specific repeated DNA family; SRY, sex-determining region, Y-linked; TSPY testis-specific protein, Y-linked.
- © 2009 The American Association for Clinical Chemistry