Circulating placental RNA in maternal plasma is associated with a preponderance of 5' mRNA fragments: implications for noninvasive prenatal diagnosis and monitoring.
Despite the rapid development of cell-free circulating RNA as a diagnostic tool, much of the biology of these molecular species remains to be elucidated. In view of the well-known lability of RNA and the existence of ribo-nucleases in the circulation (12,13 ), the recent demonstration of stable RNA in plasma and serum (14) came as a surprising and intriguing finding for the field of medical diagnostics. It has been proposed that the stability of such extracellular RNA is conferred by its association with certain types of particulate matter, e.g., apoptotic bodies (15,16). Subsequent studies have also unveiled the molecular aspects of plasma nucleic acids. It was shown that Epstein-Barr virus DNA molecules in the plasma of cancer patients consisted mainly of short DNA fragments <180 by in length (17) and that fetal DNA species in the plasma of pregnant women were relatively shorter than the maternal counterparts (18,19). However, no information is currently available regarding the integrity of plasma RNA.
In this study, we investigated the integrity of plasma RNA by developing real-time quantitative reverse-transcription-PCR (RT-PCR)  assays to target multiple regions along a transcript. Whether a transcript is intact or not would depend on the observed abundances of the different amplicons from the corresponding transcript. In particular, we focused on measurement of the 5', middle, and 3' regions of a transcript. A discrepancy in the measured quantities or detection rates of the 3 regions would suggest the presence of incomplete mRNA fragments. We first analyzed the 5', middle, and 3' regions of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcript in the plasma of healthy persons. This quantitative strategy was validated by 2-step RT-PCR and serial dilution experiments. We then applied the same approach and used the placenta as a model system for studying the integrity of circulating placental RNA. As a panel, GAPDH mRNA and 6 different placental transcripts were systemically quantified for the 5', middle, and 3' regions in placental tissues and for the 5' and 3' regions in the plasma of pregnant women. These transcripts previously were shown to have differential expression in the placenta by microarray analysis (9), including the [beta]-subunit of human chorionic gonadotropin ([beta]hCG), placenta-specific 1 (PLAC1), tissue factor pathway inhibitor 2 (TFPI2), adrenomedullin (ADM), inhibin-[beta] A subunit (INHBA), and pregnancy-associated plasma protein A (PAPPA). Lastly, to demonstrate the clinical relevance of plasma RNA integrity for noninvasive prenatal diagnosis, we compared the rates of detection for the 5' and 3' [beta]hCG assays, using maternal plasma samples from different gestational ages.
Materials and Methods
Peripheral blood samples were collected from 10 healthy nonpregnant persons and 74 healthy pregnant women. The latter group consisted of 8 first-trimester, 25 second-trimester, and 41 third-trimester pregnancies. Term placental samples were obtained from another 10 healthy pregnant women undergoing caesarian sections. All pregnant women were recruited with informed consent at the Department of Obstetrics and Gynecology, Prince of Wales Hospital, Hong Kong. The study was approved by the Clinical Research Ethics Committee of The Chinese University of Hong Kong.
PROCESSING OF BLOOD AND TISSUE SAMPLES
Blood samples were collected into EDTA-containing tubes and centrifuged at 1600g for 10 min at 4[degrees]C (Centrifuge 5810R; Eppendorf). Plasma was then carefully transferred to plain polypropylene tubes to be recentrifuged at 16 OOOg for 10 min at 4[degrees]C (Centrifuge 5415D; Eppendorf) for complete removal of residual cells (20). The supernatants were collected into fresh polypropylene tubes and mixed with TRIzol LS reagent (Invitrogen). Buffy coat was isolated after the first centrifugation step and mixed directly with the TRIzol LS reagent (Invitrogen). Placental samples were stored in an RNAlater (Ambion) stabilizing solution immediately after collection. The solution was removed after storage at 4[degrees]C overnight. All samples were kept at -80[degrees]C until RNA extraction.
RNA EXTRACTION FROM PLASMA AND BUFFY COAT SAMPLES
For plasma RNA extraction, 1.6 mL of plasma was mixed with 2 mL of TRIzol LS reagent (Invitrogen) and 0.4 mL of chloroform (15). For buffy coat RNA extraction, 0.3 mL of buffy coat was mixed with 0.9 mL of TRIzol LS (Invitrogen) and 0.24 mL of chloroform, as suggested by the manufacturer. The mixture was separated into different phases by centrifugation at 12 000g for 15 min at 4[degrees]C (Centrifuge 5417R; Eppendorf). The aqueous layer was then carefully transferred to fresh polypropylene tubes. For RNA precipitation, one volume of 700 mL/L ethanol was added to one volume of the aqueous layer. The mixture was applied to an RNeasy minicolumn (Qiagen) and processed according to the manufacturer's protocols. Total RNA was eluted in 30 [micro]L of RNase-free water and stored at -80[degrees]C until use. DNase treatment was performed with either the RNase-Free DNase Set (Qiagen) or Amplification Grade DNase I (Invitrogen).
RNA EXTRACTION FROM TISSUE SAMPLES
For tissue RNA extraction, the samples were homogenized in TRIzol reagent (Invitrogen) and mixed with chloroform according to the manufacturer's instructions. Briefly, RNA was precipitated from the aqueous layer by mixing with isopropyl alcohol and washed with 750 mL/L ethanol. The samples were then purified with an RNeasy Mini Kit (Qiagen) according to the manufacturer's protocols. Total RNA was eluted in 30 [micro]L of RNase-free water and treated with Amplification Grade DNase I (Invitrogen).
GENERATION OF CALIBRATION CURVES
The use of a single "full-length" cDNA calibrator was important for quantification of different regions within a transcript because potential bias may otherwise result from the use of multiple calibration curves. In this study, DNA calibrators were synthesized by TA cloning. The GAPDH gene was amplified from human pooled organ cDNA (Clontech), whereas amplification of PLAC1, TFPI2, ADM, and INHBA was carried out with human placental cDNA (Clontech). To generate a full-length cDNA, we performed PCR with the forward primer of the 5' assay and the reverse primer of the 3' assay, using the Advantage cDNA Polymerase Mix (Clontech). Similarly, the [beta]hCG gene was reverse-transcribed from chorionic villus RNA by use of the SuperScript One-Step RT-PCR with Platinum Taq System (Invitrogen). PCR-amplified cDNA constructs were cloned into the pGEM T-Easy Vector according to the manufacturer's instructions (Promega). Plasmid DNA was isolated with a Wizard Plus Minipreps DNA Purification Kit (Promega) and confirmed by direct sequencing (BigDye Terminator, Ver. 3.1; Applied Biosystems). Calibration curves were constructed by use of serial dilutions of the extracted plasmid DNA.
Cloning of a full-length cDNA was difficult for the PAPPA gene, which encodes a relatively long mRNA sequence (11 kb). As an alternative method, the targeted regions of the PAPPA transcript were prepared separately by PCR using human placental cDNA (Clontech). The primer sequences for the middle portion were 5'-TTCAAGACTCAGTGTACCCAGGAT-3' (sense) and 5'-ACGCACGGTCACGTTCATT-3' (antisense). The primer sequences for the 5' portion were 5'-TAGAAGCTTGGTCTCCGGCAGTGATCA-3' (sense), which introduced a HindIII restriction site, and 5'-TAGACTAGTCACCATTCACATAGAGCTTCATGA-3' (antisense), which introduced a SpeI restriction site. The primer sequences for the 3' portion were 5'-TAGCTCGAGTTCTCTTGGTCTTATTCCCATCCT-3' (sense), which introduced an XhoI restriction site, and 5'-TAGTCTAGACAGACAGTGTTATCAAATCCCACTC-3' (antisense), which introduced an XbaI restriction site. The middle construct was first inserted into the TOPO TA cloning vector (Invitrogen) at the TA cloning site. The 5'- and 3'-specific PCR constructs were digested at both ends with the appropriate restriction enzymes (BioLabs) and cloned between the HindIII and SpeI sites and between XhoI and XbaI sites of the vector, respectively. The plasmid product was extracted and confirmed by direct sequencing, as described above. The resulting plasmid vector contained all 3 targeted regions of the PAPPA transcript, suitable to be used as quantitative calibrators for each corresponding assay.
REAL-TIME QUANTITATIVE RT-PCR
We used 1-step real-time quantitative RT-PCR to measure the mRNA concentrations of the following genes: GAPDH, [beta]hCG, PLAC1, TFPI2, ADM, INHBA, and PAPPA. Quantitative RT-PCR assays were designed to target the 5', middle, and 3' regions of each transcript. The sequence information of the primers and probes (Proligo; Helios) is given in Table S1 of the Data Supplement that accompanies the online version of this article at http:// www.clinchem.org/content/vol51/issue10/. The primer and probe sequences for the 5' regions of GAPDH and PLAC1 and for the middle region of [beta]hCG have been described previously (8, 9,15). It should be noted that primer design across intron-exon junctions was not attainable for certain assays in this study, taking into consideration that each amplicon must target a particular region of a transcript (e.g., 5' and 3' ends). Nonetheless, precaution was taken to avoid the problem of genomic DNA contamination and to ensure the specificity of our RT-PCR assays. The tested RNA samples were subjected to DNase I treatment, as mentioned above, and controls were included to demonstrate the absence of DNA with the use of AmpliTaq Gold enzyme (Applied Biosystems) and omission of the reverse transcription step.
For valid data comparison in the analysis of plasma RNA, the different systems developed for each transcript were optimized with the corresponding DNA calibrator to acquire similar assay sensitivities. The lower limit of detection was 12 copies/mL for ADM; 24 copies/mL for TFPI2, INHBA, and PAPPA; and 48 copies/mL for GAPDH, [beta]hCG, and PLAC1.
The quantitative RT-PCR was set up according to the manufacturer's instructions (EZ rTth RNA PCR Reagent Kit; Applied Biosystems) with a reaction volume of 50 [micro]L. For the 5' assays, the primers were used at concentrations of 200 nM for GAPDH, [beta]hCG, and PLAC1; 300 nM for ADM and PAPPA; and 600 nM for TFPI2 and INHBA; the probes were used at concentrations of 80 nM for PLAC1; 100 nM for GAPDH, [beta]hCG, ADM, and PAPPA; and 200 nM for TFPI2 and INHBA. For amplification of the middle region, the primers were used at a concentration of 200 nM for all genes; the probes were used at concentrations of 80 nM for [beta]hCG and TFPI2 and 100 nM for the other genes. For the 3' assays, the primers were used at concentrations of 200 nM for GAPDH, [beta]hCG, TFPI2, and INHBA and 300 nM for PLAC1, ADM, and PAPPA. The probes were used at concentrations of 80 nM for TFPI2 and INHBA, and 100 nM for the other genes. For 5' and 3' amplification, extracted plasma RNA was used in volumes of 3 [micro]L for GAPDH; 5 [micro]L for [beta]hCG; 6 [micro]L for ADM, 10 [micro]L for TFPI2, PLAC1 and PAPPA; and 12 [micro]L for INHBA. In addition, 3 [micro]L of extracted plasma RNA was used for measuring the middle region of the GAPDH mRNA sequence. For each transcript, 0.5 ng of extracted placental RNA was used for quantification of the 3 targeted regions. For GAPDH and [beta]hCG, amplification was monitored and analyzed on an ABI Prism 7700 Sequence Detector (Applied Biosystems), whereas all other genes were studied on an ABI Prism 7900 Sequence Detector (Applied Biosystems). Multiple water blanks were included in every analysis. Each reaction mixture was incubated at 50[degrees]C for 2 min to activate uracil N-glycosylase, followed by reverse transcription at 60[degrees]C for 30 min. After an initial denaturation at 95[degrees]C for 5 min, the mixture was cycled 45 times with denaturation at 94[degrees]C for 20 s and 1 min of annealing and extension at 60[degrees]C for GAPDH and [beta]hCG and 56[degrees]C for TFPI2, PLAC1, ADM, INHBA, and PAPPA.
Peripheral blood samples from 10 healthy nonpregnant individuals were used in this experiment. Buffy coat RNA (20 ng) was reverse-transcribed with random and oligo(dT) primers (TagMan Reverse Transcription Reagents; Applied Biosystems). For random priming, the cycling conditions were 25 cycles of 25[degrees]C for 10 min, 48[degrees]C for 30 min, and 95[degrees]C for 5 min. For oligo(dT) priming, the cycling conditions were 25 cycles of 40[degrees]C for 10 min, 48[degrees]C for 30 min, and 95[degrees]C for 5 min. The cDNA products were amplified by real-time quantitative PCR according to the manufacturer's instructions (TagMan PCR Core Reagent Kit; Applied Biosystems) for the 5' and 3' regions of the GAPDH transcript. The primers and probes were used at concentrations of 200 nM and 100 nM, respectively, and the calibrator was generated by TA cloning as described above. For amplification, 3 [micro]L of cDNA was used in a reaction volume of 50 [micro]L. The PCR conditions were 50[degrees]C for 2 min for the activation of uracil Nglycosylase, followed by an initial denaturation at 95[degrees]C for 10 min and then by 45 cycles of denaturation at 95[degrees]C for 15 s and 1 min of annealing and extension at 60[degrees]C.
SERIAL DILUTION EXPERIMENT
Buffy coat RNA samples, from 2 healthy nonpregnant individuals, with a starting concentration of 10 ng/[micro]L were serially diluted 14 times with folds of dilution in the following order: 10-fold, 5-fold, and then 2-fold for 12 more dilutions. These diluted samples were assayed for the 5' and 3' GAPDH mRNA concentrations by 1-step real-time quantitative RT-PCR as described above
Statistical analyses were performed using the SigmaStat 2.03 software (SPSS).
QUANTITATIVE RT-PCR ANALYSIS OF MULTIPLE REGIONS OF GAPDH mRNA IN THE PLASMA OF HEALTHY INDIVIDUALS
To investigate the integrity of plasma RNA, we measured the concentrations of the 5', middle, and 3' regions of the GAPDH transcript in the plasma of 10 healthy individuals, using 1-step real-time quantitative RT-PCR. The results are expressed as fractional mRNA concentrations relative to the absolute concentrations of the 5' amplicon to assess the degree of RNA integrity in plasma. As shown in Fig. 1A, there was a significant gradual decrease in the fractional GAPDH mRNA concentration (Friedman test, P <0.001) as the target site was shifted from the 5' to the 3' end of the transcript, inclusive of the middle region. Pairwise analysis further demonstrated a statistically significant difference for every pair of targeted regions (Student-Newman-Keuls test, P <0.05). The mRNA concentration of the 5' amplicon was 1.9-fold higher than that of the middle amplicon (middle-to-5' ratio = 0.53) and 20-fold higher than that of the 3' amplicon (3'-to-5' ratio = 0.051). These results suggested that there was a preponderance of 5' GAPDH mRNA fragments in the plasma of healthy individuals. Moreover, the 3 sets of quantitative data in terms of absolute mRNA concentrations were significantly correlated with one another [Pearson correlation analysis, r = 0.791 (P <0.01) for 5' end vs middle part; r = 0.689 (P <0.05) for 5' vs 3' ends; and r = 0.770 (P <0.01) for the middle part vs the 3' end; Fig. 1, B-D], as expected for amplicons derived from the same transcript.
VALIDATION OF 1-STEP REAL-TIME RT-PCR ANALYSIS OF RNA INTEGRITY IN CIRCULATION
To exclude the possibility of technical bias in our 5' GAPDH RT-PCR system, we performed 2-step RT-PCR analysis of both the 5' and 3' GAPDH regions, using random and oligo(dT) primers for reverse transcription followed by gene-specific real-time PCR. The rationale for this experiment was based on the fact that the oligo(dT) primer could potentially introduce 3' end bias in the amplified products, as 3'-degraded transcripts would not be successfully reverse-transcribed. Demonstration of 3' mRNA predominance in the oligo(dT) primed system would suggest that our results were not attributable to any analytical bias favoring the 5' system. Buffy coat RNA samples from 10 healthy individuals were assayed for the 5' and 3' GAPDH mRNA concentrations by 2-step RTPCR. As shown in Fig. 2A, the reaction using random primer produced a significantly higher concentration of 5' GAPDH amplicon (Wilcoxon test, P <0.01), consistent with our findings based on 1-step RT-PCR. In contrast, the reaction using oligo(dT) primer produced a significantly higher concentration of 3' GAPDH amplicon (Wilcoxon test, P <0.01), presumably as a result of 3' end bias during the reverse transcription step.
To further validate the observed data, we serially diluted 2 buffy coat RNA samples from healthy individuals (starting concentration, 10 ng/[micro]L) and assessed the detection rates for the 5' and 3' GAPDH amplicons in these samples, using 1-step real-time quantitative RTPCR. The results are expressed as copy number per microliter of total input RNA. As shown in panels B and C of Fig. 2, the 3' GAPDH amplicon was no longer detectable after only 5 consecutive dilution steps. On the other hand, the 5' amplicon was consistently detected in the more dilute samples, with up to 11 and 12 consecutive dilution steps for cases 6 and 7, respectively. The 5' and 3' GAPDH assays had shown equivalent sensitivities with the DNA calibrator, both detecting down to 48 copies/ mL, thereby excluding the possibility of quantitative bias in this experiment. Collectively, these data supported the conclusion that 5' mRNA fragments constitute a majority of the GAPDH mRNA species in the circulation.
[FIGURE 1 OMITTED]
QUANTITATIVE ASSESSMENT OF PLACENTAL RNA INTEGRITY IN PLACENTAL TISSUES
To explore the generality of the predominance of 5' mRNA fragments, we examined a panel of placentaexpressed transcripts. We compared the quantities of the 5', middle, and 3' regions of different mRNAs in placental tissues collected from 10 healthy pregnant women after caesarian section (gestational age, 38-40 weeks). Six placental transcripts, in addition to GAPDH mRNA, were chosen for this experiment, including [beta]hCG, PLAC1, TFPI2, ADM, INHBA, and PAPPA. The 5', middle, and 3' regions of each transcript were systematically quantified by 1-step real-time RT-PCR. Again the results were expressed as fractional mRNA concentrations relative to the 5' region to reflect the degree of RNA integrity of different transcripts in placental tissues (Table 1).
The 3 targeted amplicons for each of the 7 transcripts studied were detectable in all of the studied samples. As shown in Fig. 3, except for PLAC1, the fractional mRNA concentrations differed significantly among the 3 targeted regions of each transcript (Friedman test, P <0.001). For GAPDH and ADM, pairwise multiple comparison showed a significant difference for every pair of targeted regions (Student-Newman-Keuls test, P <0.05). The mRNA concentration of the 5' GAPDH amplicon was 2.1-fold higher than that of the middle amplicon (middle-to-5' ratio = 0.47) and 34-fold higher than that of the 3' amplicon (3'-to-5' ratio = 0.029). Similarly, the mRNA concentration of the 5' ADM amplicon was 3.6-fold higher than that of the middle amplicon (middle-to-5' ratio = 0.28) and 6.7-fold higher than that of the 3' amplicon (3'-to-5' ratio = 0.15). For [beta]hCG, TFPI2, and INHBA, we observed a significant difference between the 5' and 3' amplicons (Student-Newman-Keuls test, P <0.05) and between the middle and 3' amplicons (Student-Newman-Keuls test, P <0.05). The 5' and middle amplicons, however, did not demonstrate any significant difference (Student-Newman-Keuls test, P >0.05). The median 3'-to-5' ratio was 0.62 for [beta]hCG, 0.071 for TFPI2, and 0.40 for INHBA. For the PAPPA gene, pairwise analysis also demonstrated a significant difference for every pair of targeted regions, but with a slightly different trend of expression across the 3 assays (Student-Newman-Keuls test, P <0.05). The mRNA concentration of the 5' PAPPA amplicon was 1.6-fold lower than that of the middle amplicon but 4.0-fold higher than that of the 3' amplicon (3'-to-5' ratio = 0.25). Lastly, for the PLAC1 gene, the fractional mRNA concentration gradually decreased as the target site was shifted from the 5' to the 3' end of the transcript, with the P value close to statistical significance (Friedman test, P = 0.067). These data suggested that, at the expression level, the placental tissue generally contained predominately the 5' fragments of placenta-expressed transcripts.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
QUANTITATIVE ASSESSMENT OF PLACENTAL RNA INTEGRITY IN MATERNAL PLASMA
We next investigated whether the placental RNA circulating in maternal plasma would also have an overrepresentation of the 5'-fragment region. Using 1-step real-time RT-PCR, we measured the abundance of the 5' and 3' ends of different placenta-expressed transcripts in the plasma of pregnant women. To facilitate placental RNA detection in maternal plasma (excluding GAPDH), our samples were chosen based on the expected time during pregnancy when the transcript is most highly expressed, as suggested by previous microarray data (9). Eight first-trimester samples (gestational age, 7-12 weeks) were assayed for [beta]hCG, 8 second-trimester samples (gestational age, 17-19 weeks) for GAPDH, 7 third-trimester samples for PLAC1 (gestational age, 37-40 weeks), TFPI2 (gestational age, 37-38 weeks), and PAPPA (gestational age, 37-41 weeks), and 10 third-trimester samples for ADM (gestational age, 37-40 weeks) and INHBA (gestational age, 37-40 weeks). The detection rates and absolute mRNA concentrations of the respective amplicons for each individual transcript were compared.
We observed a discrepancy in the rates of detection of the 5' and 3' amplicons for all of the studied transcripts (Table 2). Overall, the 5'-specific assays gave higher rates of placenta-derived RNA detection in maternal plasma compared with the 3'-specific assays, given that both types of assays had shown comparable sensitivities with the DNA calibrator during optimization. In terms of the absolute concentrations, we observed a significant decrease in the plasma mRNA concentration for GAPDH (Wilcoxon test, P <0.01), [beta]hCG (Wilcoxon test, P <0.01), TFPI2 (Wilcoxon test, P <0.05), ADM (Wilcoxon test, P <0.01), and INHBA (Wilcoxon test, P <0.01), as the target amplicon was changed from the 5' to the 3' end of the corresponding transcript (Fig. 4). The median 5' mRNA concentrations of GAPDH, [beta]hCG, TFPI2, ADM, and INHBA were 4274 (interquartile range, 2420-4813), 1031 (402-2216), 479 (321-868), 157 (66-263), and 39 (0-53) copies/mL, respectively. On the other hand, the median 3' mRNA concentrations were 114 (0-224), 156 (48-374), 130 (25-232), 0 (0-18), and 0 (0-0) copies/ mL, respectively. Statistical comparisons between the maternal plasma concentrations of the 5' and 3' amplicons of PLAC1 or PAPPA were not performed because of their low detection rates in maternal plasma (Table 2). The 5' amplicons for both PLAC1 and PAPPA were detected in 29% (2 of 7), but the 3' amplicons were not detected in any of the tested samples. The plasma fractional mRNA concentrations (relative to the 5' region) of all transcripts are summarized in Table 1.
[FIGURE 4 OMITTED]
CROSS-GESTATIONAL STUDY OF 5' AND 3' [beta]hCG mRNA IN MATERNAL PLASMA
Apart from biological significance, our findings suggested that targeting the 5' end of a particular transcript could potentially improve the detection rate for RNA in maternal plasma. To illustrate such a clinical application, we compared the 5' and 3' [beta]hCG assays in maternal plasma from different gestation periods. We obtained plasma samples from 8 women in the first trimester of pregnancy (gestational age, 7-12 weeks) and from 17 women in the second trimester of pregnancy (gestational age, 16-20 weeks). As shown in Fig. 5, the circulating [beta]hCG mRNA concentration displayed a decreasing trend with gestational age in both the 5' and 3' systems. Statistical analysis showed that the plasma mRNA concentrations of the 5' amplicon were significantly higher than those of the 3' amplicon for both trimesters (Wilcoxon test, P <0.01). As described above, in the first-trimester samples, the median concentrations of the 5' and 3' [beta]hCG amplicons were 1031 (interquartile range, 402-2216) and 156 (48-374) copies/mL, respectively. In the second-trimester samples, the median concentrations of the 5' and 3' [beta]hCG amplicons were 240 (0-363) and 0 (0-94) copies/mL, respectively. Furthermore, across gestations, the 5' [beta]hCG assay demonstrated a higher detection rate than the 3' [beta]hCG assay in maternal plasma. The detection rates of the 5' [beta]hCG assay were 100% for the first-trimester group and 53% for the second-trimester group, whereas the detection rates of the 3' [beta]hCG assay were only 75% for the first-trimester group and 29% for the second-trimester group. These data confirmed that the use of a 5'-specific system could enhance the sensitivity of mRNA detection in maternal plasma, suggesting potential implications in the future development of new assays for fetal RNA markers for noninvasive prenatal assessment.
[FIGURE 5 OMITTED]
The discovery of placental RNA in the plasma of pregnant women has shown promise for noninvasive prenatal diagnosis and monitoring (8,10,11). However, little is currently known about the biology of circulating RNA. Recently it was demonstrated that cell-free RNA is surprisingly stable in plasma and serum (14). It has been postulated that cell-free circulating RNA is associated with particulate matter and thus protected from degradation by nucleases (15,16). For a better understanding of the molecular characteristics of plasma RNA, we investigated the integrity of these molecular species.
We first studied the integrity of the GAPDH transcript in the plasma of healthy nonpregnant individuals by quantitatively comparing multiple regions (5', middle, and 3' regions) within the transcript in 1-step real-time RT-PCR assays. Our results suggested that circulating GAPDH mRNA population is predominantly 5' fragments rather than intact transcripts. We found that the fractional concentration of GAPDH mRNA (relative to the 5' region) decreased significantly as the amplicon position was moved away from the 5' end of the transcript and that the different target amplicons had significant correlations with one another. To exclude the possibility of analytical bias in the 1-step RT-PCR procedure, we performed 2-step RT-PCR analysis of 5' and 3' GAPDH amplicons in buffy coat RNA samples, using random and oligo(dT) primers in the reverse transcription step. Random priming produced results similar to those obtained by 1-step real-time analysis. In contrast, oligo(dT) priming gave a relatively higher absolute concentration of the 3' amplicon because RNA molecules without a 3' end could not be reverse transcribed in this system. To further validate our data, we used 1-step real-time RT-PCR to test the detection rates for 5' and 3' GAPDH amplicons in 2 sets of buffy coat RNA samples after serial dilutions. This experiment demonstrated the rapid disappearance of 3' fragments, in contrast to the persistent detection of 5' fragments. Given that we optimized both assays so that they were equally sensitive, the serial dilution experiment provided a physical means of demonstrating that there was indeed a higher abundance of transcript molecules with 5' ends than those with intact 3' ends. Collectively, these results confirmed the preponderance of 5' GAPDH mRNA fragments in the circulation.
To study whether the predominance of 5' mRNA fragments could be seen in other transcripts, we used the placenta as a model system for further investigation. Six different placental transcripts, in addition to GAPDH mRNA, were systematically analyzed for the 5' and 3' amplicons in the plasma of pregnant women. Five of the 7 genes (GAPDH, [beta]hCG, TFPI2, ADM, and INHBA) gave significantly higher absolute mRNA concentrations for the 5' amplicon than for the 3' amplicon. Moreover, for every gene under study (including PLAC1 and PAPPA), the 5' assay achieved a higher detection rate than the corresponding 3' assay in the analysis of plasma RNA. These findings provide direct evidence that circulating placental RNA is associated with a preponderance of 5' mRNA fragments in maternal plasma.
We observed a similar phenomenon when we studied the same panel of transcripts in placental tissues. Overall, the 7 genes could be classified into 3 categories with regard to the trend of expression from the analysis of the 5', middle, and 3' regions of each individual transcript. For GAPDH, ADM, and PLAC1, the fractional mRNA concentration (relative to the 5' amplicon) decreased gradually as the target site was changed from the 5' end to the middle part and from the middle part to the 3' end of the sequence. For [beta]hCG, TFPI2, and INHBA, we found no significant difference between the 5' and middle amplicons; however, the concentration of the 3' amplicon was significantly lower than the concentration of the 5' amplicon. Analysis of multiple regions of the PAPPA transcript showed yet another trend of expression in which the middle amplicon showed a significant increase across the 3 systems, with the 3' amplicon being relatively less abundant than the 5' amplicon. These results are consistent with previous reports (21, 22) documenting the detection of mRNA decay fragments in human and animal tissues.
To establish the biological relevance of this study, we addressed several important issues regarding experimental design. First, it is known that [beta]hCG is encoded by 6 highly similar genes ([beta]hCG1 through -3, -5, -7, and -8) clustered on chromosome 19 and is of high homology with the [beta]-subunit of the luteinizing hormone gene (OLH) (23). To avoid introducing quantitative bias in the RTPCR analysis, we ascertained that the 5', middle, and 3' primer sets for [beta]hCG were all complementary with the mRNA sequences arising from [beta]hCG3, -5, and -8, which are the major [beta]hCG mRNA species in the placenta at steady state (24). To minimize nonspecific amplification, we also incorporated in the primer and probe sequences several base mismatches to distinguish between [beta]hCG and OLH. Second, for different RT-PCR systems to be compared on a valid ground, the choice of quantitative calibrators was another key aspect of this study. We constructed our own DNA calibrators by cloning all of the targeted sequences into the same plasmid for each studied transcript. Quantitative analysis of multiple regions within a transcript based on this type of calibrator would accurately reflect the relative changes in the measured mRNA concentrations and eliminate the potential bias from the use of multiple calibration curves. Optimization of assay sensitivity was also critical for accurate data interpretation in the analysis of plasma RNA. We obtained similar detection limits across the compared systems for every transcript under study, excluding the possibility of quantitative bias in our results. Theoretically, variable reverse transcription efficiency between different amplicons attributable to the differences in the primary genetic sequence may affect the quantification and thereby the comparison of concentrations among such amplicons. However, given the universal demonstration of the preponderance of 5' fragments in all of the 7 studied genes, it is unlikely that variation in the reverse transcription efficiency was a major confounding factor that was the sole basis of our observations.
In consideration of the origin of the observed 5' mRNA fragments in maternal plasma, we speculate that this biological phenomenon is likely to be associated with the mechanism of mRNA degradation. Previous studies have reported that the major pathway of mRNA decay in human cells proceeds from the 3' end of a transcript (25-27). One proposed model is that the human exosome, a protein complex consisting of 10 or more 3' exoribonucleases, interacts with an AU-rich element found in the 3'-untranslated region of an RNA transcript (26) through the binding of some AU-rich element-binding proteins (27). It has also been suggested that mRNA decay is correlated with the functional characteristics of human mRNAs (28). In this regard, the functional role, if any, of the observed placental RNA fragments in maternal plasma will require further investigation.
Apart from biological significance, our findings have potential clinical implications for the future study of placental RNA markers in maternal plasma. We have shown that, in addition to the quantitative differences, the detection rates of the 5' RT-PCR assays were higher than those of the corresponding 3' assays for detecting circulating placental RNA. Particularly, the plasma mRNAs of INHBA, PLAC1, and PAPPA could be detected only with the 5' amplicon. In addition, in the cross-gestational study, the 5' [beta]hCG system could detect its target in 100% of first-trimester samples and 53% of second-trimester samples. On the other hand, the 3' [beta]hCG system was positive for only 75% and 29%, respectively, of the corresponding samples. Of note, a recent study reported another [beta]hCG assay that amplified the middle region of the mRNA sequence and detected [beta]hCG mRNA in 100% of first-trimester samples and 42% of second-trimester samples (8). Collectively, these data have suggested that targeting the 5' end of a transcript could potentially increase the performance of a placental RNA marker in maternal plasma for noninvasive prenatal diagnosis. Whether the degraded transcripts detected in maternal plasma could be representative of the pathophysiology of pregnancy-associated diseases is beyond the scope of this study and remains to be investigated.
In conclusion, we have demonstrated that circulating placental RNA is associated with a preponderance of 5' mRNA fragments in maternal plasma. In addition to the biological relevance, the findings of our study have particular implications for the future development of fetal RNA markers for noninvasive prenatal diagnosis. For example, this information could be used to improve the rate of placental RNA detection in maternal plasma, thus facilitating the establishment of more molecular markers for pregnancy-associated diseases, including preeclampsia and fetal chromosomal aneuploidies. Our findings may also have implications for other clinical uses of plasma RNA, particularly in the detection of tumor RNA for cancer diagnosis (29).
We thank Dr. Lisa Y.S. Chan, Katherine C.K. Chow, and Nicole Y.L. Lam for technical assistance. We are also grateful to Dr. Enders K.O. Ng for helpful discussion and advice at the start of the project. This work was supported by a Central Allocation Grant (CUHK 01/03C) from the Research Grants Council of the Hong Kong Special Administrative Region (China). Several of the co-authors hold patents or patent applications on aspects of circulating nucleic acids in plasma.
(1.) Lo YMD, Chiu RWK. The biology and diagnostic applications of plasma RNA. Ann N Y Acad Sci 2004;1022:135-9.
(2.) Wong SC, Lo SF, Cheung MT, Ng EKO, Tse CW, Lai BS, et al. Quantification of plasma [beta]-catenin mRNA in colorectal cancer and adenoma patients. Clin Cancer Res 2004;10:1613-7.
(3.) Silva JM, Dominguez G, Silva J, Garcia JM, Sanchez A, Rodriguez O, et al. Detection of epithelial messenger RNA in the plasma of breast cancer patients is associated with poor prognosis tumor characteristics. Clin Cancer Res 2001;7:2821-5.
(4.) Chen XQ, Bonnefoi H, Pelte MF, Lyautey J, Lederrey C, Movarekhi S, et al. Telomerase RNA as a detection marker in the serum of breast cancer patients. Clin Cancer Res 2000;6:3823-6.
(5.) Kopreski MS, Benko FA, Kwak LW, Gocke CD. Detection of tumor messenger RNA in the serum of patients with malignant melanoma. Clin Cancer Res 1999;5:1961-5.
(6.) Lo KW, Lo YMD, Leung SF, Tsang YS, Chan LYS, Johnson PJ, et al. Analysis of cell-free Epstein-Barr virus associated RNA in the plasma of patients with nasopharyngeal carcinoma. Clin Chem 1999;45:1292-4.
(7.) Poon LLM, Leung TN, Lau TK, Lo YMD. Presence of fetal RNA in maternal plasma. Clin Chem 2000;46:1832-4.
(8.) Ng EKO, Tsui NBY, Lau TK, Leung TN, Chiu RWK, Panesar NS, et al. mRNA of placental origin is readily detectable in maternal plasma. Proc Natl Acad Sci U S A 2003;100:4748-53.
(9.) Tsui NBY, Chim SSC, Chiu RWK, Lau TK, Ng EKO, Leung TN, et al. Systematic micro-array based identification of placental mRNA in maternal plasma: towards non-invasive prenatal gene expression profiling. J Med Genet 2004;41:461-7.
(10.) Ng EKO, Leung TN, Tsui NBY, Lau TK, Panesar NS, Chiu RWK, et al. The concentration of circulating corticotropin-releasing hormone mRNA in maternal plasma is increased in preeclampsia. Clin Chem 2003;49:727-31.
(11.) Ng EKO, El-Sheikhah A, Chiu RWK, Chan KCA, Hogg M, Bindra R, et al. Evaluation of human chorionic gonadotropin [beta]-subunit mRNA concentrations in maternal serum in aneuploid pregnancies: a feasibility study. Clin Chem 2004;50:1055-7.
(12.) Reddi KK, Holland JF. Elevated serum ribonuclease in patients with pancreatic cancer. Proc Natl Acad Sci U S A 1976;73:2308-10.
(13.) Sheid B, Lu T, Pedrinan L, Nelson JH Jr. Plasma ribonuclease: a marker for the detection of ovarian cancer. Cancer 1977;39: 2204-8.
(14.) Tsui NBY, Ng EKO, Lo YMD. Stability of endogenous and added RNA in blood specimens, serum, and plasma. Clin Chem 2002; 48:1647-53.
(15.) Ng EKO, Tsui NBY, Lam NYL, Chiu RWK, Yu SC, Wong SC, et al. Presence of filterable and nonfilterable mRNA in the plasma of cancer patients and healthy individuals. Clin Chem 2002;48: 1212-7.
(16.) Hasselmann DO, Rappl G, Tilgen W, Reinhold U. Extracellular tyrosinase mRNA within apoptotic bodies is protected from degradation in human serum. Clin Chem 2001;47:1488-9.
(17.) Chan KCA, Zhang J, Chan AT, Lei KI, Leung SF, Chan LYS, et al. Molecular characterization of circulating EBV DNA in the plasma of nasopharyngeal carcinoma and lymphoma patients. Cancer Res 2003;63:2028-32.
(18.) Chan KCA, Zhang J, Hui AB, Wong N, Lau TK, Leung TN, et al. Size distributions of maternal and fetal DNA in maternal plasma. Clin Chem 2004;50:88-92.
(19.) Li Y, Zimmermann B, Rusterholz C, Kang A, Holzgreve W, Hahn S. Size separation of circulatory DNA in maternal plasma permits ready detection of fetal DNA polymorphisms. Clin Chem 2004;50: 1002-11.
(20.) Chiu RWK, Poon LLM, Lau TK, Leung TN, Wong EM, Lo YMD. Effects of blood-processing protocols on fetal and total DNA quantification in maternal plasma. Clin Chem 2001;47:1607-13.
(21.) Timofeeva AV, Skrypina NA, Savochkina LP, Beabealashvilli R. Size distribution of the urokinase mRNA decay intermediates in different tissues and cell lines. Biochim Biophys Acta 2000;1517: 33-45.
(22.) Eberding A, Rehaume V, Lee CH. Detection of mRNA degradation intermediates in tissues using the 3'-end poly(A)-tailing polymerase chain reaction method. Anal Biochem 2004;335:58-65.
(23.) Policastro PF, Daniels-McQueen S, Carle G, Boime I. A map of the hCG [beta]-LH [beta] gene cluster. J Biol Chem 1986;261:5907-16.
(24.) Bo M, Boime I. Identification of the transcriptionally active genes of the chorionic gonadotropin [beta] gene cluster in vivo. J Biol Chem 1992;267:3179-84.
(25.) Wang Z, Kiledjian M. Functional link between the mammalian exosome and mRNA decapping. Cell 2001;107:751-62.
(26.) Mukherjee D, Gao M, O'Connor JP, Raijmakers R, Pruijn G, Lutz CS, et al. The mammalian exosome mediates the efficient degradation of mRNAs that contain AU-rich elements. EMBO J 2002; 21:165-74.
(27.) Chen CY, Gherzi R, Ong SE, Chan EL, Raijmakers R, Pruijn GJ, et al. AU binding proteins recruit the exosome to degrade ARE-containing mRNAs. Cell 2001;107:451-64.
(28.) Yang E, van Nimwegen E, Zavolan M, Rajewsky N, Schroeder M, Magnasco M, et al. Decay rates of human mRNAs: correlation with functional characteristics and sequence attributes. Genome Res 2003;13:1863-72.
(29.) Johnson PJ, Lo YMD. Plasma nucleic acids in the diagnosis and management of malignant disease. Clin Chem 2002;48:1186-93.
BLENDA C.K. WONG,  ROSSA W.K. CHIU,  NANCY B.Y. TSUI,  K.C. ALLEN CHAN,  LIN W. CHAN,  TZE K. LAU,  TSE N. LEUNG,  and Y.M. DENNIS LO  *
Departments of  Chemical Pathology and  Obstetrics and Gynecology, The Chinese University of Hong Kong, Prince of Wales Hospital, Shafin, New Territories, Hong Kong SAR, China.
 Nonstandard abbreviations: RT-PCR, reverse transcription-PCR; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; [beta]hCG, [beta]-subunit of human chorionic gonadotropin; PLAC1, placenta-specific 1; TFPI2, tissue factor pathway inhibitor 2; ADM, adrenomedullin; INHBA, inhibin-[beta] A subunit; and PAPPA, pregnancy-associated plasma protein A.
* Address correspondence to this author at: Department of Chemical Pathology, The Chinese University of Hong Kong, Prince of Wales Hospital, Rm 38023, 1/F Clinical Sciences Bldg, 30-32 Ngan Shing Street, Shatin, New Territories, Hong Kong SAR, China. Fax 852-2194-6171; e-mail loym@cuhk. edu.hk.
Received April 11, 2005; accepted July 11, 2005.
Previously published online at DOI: 10.1373/clinchem2005.052340
Table 1. Comparison of the abundances of 5' and 3' mRNA fragments for different placenta-expressed transcripts in placental tissues and maternal plasma. Median 3'-to-5' ratio Gene Placental tissues Maternal plasma (a) GAPDH 0.029 0.034 ADM 0.15 0.20 [beta]hCG 0.62 0.15 TFPI2 0.071 0.21 INHBA 0.40 0 PAPPA 0.25 0 PLAC1 0.86 0 (a) Ratios were calculated only for cases where the 5' amplicon was detectable. Table 2. Detection rates of real-time RT-PCR assays targeting 5 and 3 regions of placenta-expressed transcripts in maternal plasma. No. of 5' amplicon 3' amplicon Gene cases tested positive, % positive, % GAPDH 8 100 63 [beta]hCG 8 100 75 TFPI2 7 100 71 ADM 10 100 40 INHBA 10 60 0 PLAC1 7 29 0 PAPPA 7 29 0
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Molecular Diagnostics and Genetics|
|Author:||Wong, Blenda C.K.; Chiu, Rossa W.K.; Tsui, Nancy B.Y.; Chan, K.C. Allen; Chan, Lin W.; Lau, Tze K.;|
|Date:||Oct 1, 2005|
|Previous Article:||Real-time PCR assays targeting a unique chromosomal sequence of Yersinia pestis.|
|Next Article:||Use of proteomic methods to identify serum biomarkers associated with rat liver toxicity or hypertrophy.|