Background: Circulating RNA in plasma/serum is an emerging field for noninvasive molecular diagnosis. Because RNA is widely thought to be labile in the circulation, we investigated the stability and various preanalytical factors that may affect RNA concentrations in blood specimens.
Methods: Blood samples were collected from 65 healthy volunteers. The effects of two preanalytical variables were studied: (a) time delay in processing of EDTA blood and clotted blood after venesection, and (b) freezing and thawing of plasma and serum. The lability of free added RNA in plasma was also investigated. Plasma/serum RNA was measured by a real-time quantitative reverse transcription-PCR assay for glyceraldehyde 3-phosphate dehydrogenase mRNA, whereas DNA was measured by a real-time quantitative PCR assay for the β-globin gene.
Results: No significant difference was found for plasma RNA concentrations obtained from uncentrifuged EDTA blood that had been left at 4 °C for 0, 6, and 24 h (P =0.182). On the other hand, the serum RNA concentrations increased significantly over 24 h when uncentrifuged clotted blood was stored at 4 °C (P <0.05). In comparison, >99% of the free added RNA could no longer be amplified after incubation in plasma for 15 s. Never-frozen plasma, freeze-thawed plasma, and thawed plasma left at room temperature for 1 h showed no significant differences in RNA concentration (P =0.465). No significant difference was observed for freeze-thawed serum (P = 0.430).
Conclusions: Plasma RNA is stable in uncentrifuged EDTA blood stored at 4 °C, but to obtain a stable serum RNA concentration, uncentrifuged clotted blood should be stored at 4 °C and processed within 6 h. A single freeze/thaw cycle produces no significant effect on the RNA concentration of plasma or serum.
Circulating RNA in plasma/serum is an emerging field for noninvasive diagnostic applications. The discoveries of tumor-derived RNA in the plasma/serum of cancer patients (1)(2)(3)(4) and fetal-derived RNA in the plasma of pregnant women (5) have opened up a new field for studying gene expression noninvasively. Recently, real-time quantitative reverse transcription-PCR (RT-PCR) for plasma RNA was developed (4)(6), which makes plasma RNA detection assays more informative and sensitive. The existence of circulating RNA is a remarkable finding because RNA is more labile than DNA and ribonuclease is known to be present in blood (7). At present, the exact mechanisms that protect circulating RNA are still unknown. The RNA may possibly be complexed to lipids, proteins, lipoproteins, or phospholipids (8)(9)(10)(11);bound with DNA in nucleosomes (9)(12); or protected within apoptotic bodies (13) or other vesicular structures. Hasselmann et al. (14) have shown in vitro that RNA within apoptotic bodies is resistant to degradation in serum. Our group has recently shown that most of the circulating RNA in the plasma is particle-associated (6). Therefore, RNA in the circulation may not be as fragile as it was previously assumed to be.
Although circulating RNA is detectable in plasma and serum, little is known about its stability. Theoretically, to maintain RNA integrity, the time and steps between venesection and RNA extraction should be reduced to the minimum. Methods such as snap-freezing of plasma or the addition of a stabilizing reagent have also been used to reduce the chance of RNA degradation. However, detailed studies on the stability of circulating RNA are limited.
In this report, we describe a study on the effects of various preanalytical issues on the stability of plasma RNA. Using real-time quantitative RT-PCR (6), we investigated the effects of two classes of preanalytical factors on plasma RNA concentrations: (a) the time delay in blood processing after venesection, and (b) the effect of freezing and thawing of plasma and serum. Because circulating RNA was recently shown to exist in both particle-associated and non-particle-associated forms (6), both of these forms were studied simultaneously.
Materials and Methods
Blood samples were collected from 65 healthy volunteers. All had given informed consent.
processing of blood samples
Plasma and serum collection.
In this study, all plasma and serum were obtained by the initial centrifugation of the blood samples at 1600g for 10 min at 4 °C (Centrifuge 5810R;Eppendorf). Supernatants were collected and further centrifuged at 16 000g for 10 min at 4 °C (Centrifuge 5415R;Eppendorf).
Delayed blood processing after venesection.
A total of 20 EDTA-blood and 20 clotted blood samples were used in this part of the study. EDTA-blood samples were divided into four portions, each portion with five samples. For the first portion, blood samples were left at room temperature for 0, 6, and 24 h. After that, plasma was subjected to RNA and DNA extraction. Treatment of the second portion was the same as that of the first portion, but the collected plasma was filtered through a 0.22 μm filter (MILLEX-GV; Millipore) before RNA/DNA extraction. For the third portion, blood samples were left at 4 °C for 0, 6, and 24 h. Plasma was then collected for RNA/DNA extraction. The treatment of the fourth portion was the same as that of the third portion except that the plasma samples were filtered through a 0.22 μm filter. The other 20 clotted samples were handled in the same fashion.
Lability of free added RNA in plasma.
The lability of free added RNA in plasma was studied. We added 425 ng of commercially available human RNA (PE Applied Biosystems) to a randomly selected plasma sample and incubated for 0, 5, 10 and 15 s at room temperature. After incubation, Trizol LS reagent (Invitrogen) was immediately added to stop any ribonuclease activity.
Freezing and thawing of plasma and serum.
We studied 16 plasma samples harvested from EDTA blood. Each plasma sample was divided into three parts. One-third was snap-frozen in liquid nitrogen and then thawed at room temperature immediately. One-third was snap-frozen and thawed, after which the thawed plasma was left at room temperature for 1 h. The remaining portion was subjected to no treatment. After the treatments, each of the portions was divided into two halves. One-half was filtered through a 0.22 μm filter, whereas the other half remained unfiltered.
The effect of freezing and thawing of serum was also studied. Six serum samples were collected, which were treated as described above. A freezing temperature of −20 °C was used.
To investigate whether freezing and thawing would have a similar effect on the stability of RNA associated with particles of different sizes, a further experiment was carried out using two plasma samples. For each sample, one-half was frozen at −20 °C and then thawed immediately, whereas the other half remained untreated. Both the freeze-thawed and the untreated samples were then divided into four portions. Three of the four portions were passed through 5, 0.45, and 0.22 μm pore-size filters individually, whereas the fourth portion remained unfiltered.
rna extraction from plasma and serum
We mixed 800 μL of plasma/serum with 1 mL of Trizol LS reagent and 0.2 mL of chloroform (6). The mixture was centrifuged at 12 000g for 15 min at 4 °C, and the aqueous layer was transferred into new tubes. One volume of 700 mL/L ethanol was added to one volume of the aqueous layer. The mixture was then applied to a RNeasy minicolumn (Qiagen) and was processed according to the manufacturer’s recommendations. “On-column DNase digestion” was performed using a RNase-free DNase preparation (Qiagen) as described by the manufacturer. Finally, the RNA was eluted with 15 μL of RNase-free water. RNA was stored at −80 °C until use.
dna extraction from plasma/serum
DNA was extracted from 200 μL of plasma/serum using a QIAamp Blood reagent set (Qiagen) with the “blood and body fluid protocol”as recommended by the manufacturer (15). A final elution volume of 50 μL was used.
real-time quantitative rt-pcr
RNA concentrations were measured using one-step real-time quantitative RT-PCR (6)(16) for glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA. The amplification primers were GAPDHF (5′-GAAGGTGAAGGTCGGAGT-3′) and GAPDHR (5′-GAAGATGGTGATGGGATTTC-3′), and the dual-labeled fluorescent probe was GAPDHP [5′-(FAM)CAAGCTTCCCGTTCTCAGCC(TAMRA)-3′; FAM is 6-carboxyfluorescein, and TAMRA is 6-carboxytetramethylrhodamine]. RT-PCR was set up in a reaction volume of 50 μL using an EZ rTth RNA PCR reagent set (PE Applied Biosystems). Each reaction contained 1× EZ buffer; 200 nM of each primer; 100 nM fluorescent probe; 3 mM Mn(OAc)2;300 μM each of dATP, dCTP, and dGTP; 600 μM dUTP; 5 U of rTth polymerase; and 0.5 U of uracil N-glycosylase. For amplification, we used 3 μL of extracted RNA. Amplification data were collected and analyzed with an ABI Prism 7700 Sequence Detector (PE Applied Biosystems). Each sample was analyzed in duplicate, and multiple negative water blanks were included in every analysis. A calibration curve was prepared using serial dilutions of a commercially available human control RNA (PE Applied Biosystems). The thermal profile was as follows: reaction was initiated at 50 °C for 2 min, followed by 60 °C for 30 min. After denaturation at 95 °C for 5 min, PCR was carried out for 40 cycles at 94 °C for 20 s and 62 °C for 1 min.
real-time quantitative pcr
Purified DNA was quantified by a real-time quantitative PCR for the β-globin gene as described previously (15).
Statistical analyses were performed using SigmaStat 2.03 software (SPSS).
stability of circulating nucleic acid concentrations in blood after venesection
Delayed plasma separation.
To determine whether delayed blood processing would lead to degradation of circulating RNA, we measured the plasma RNA concentrations serially after EDTA blood had been left for 0, 6, and 24 h. At room temperature, RNA concentrations of unfiltered plasma were significantly different among the three time points (Friedman test, P <0.05; Fig. 1A⇓ ).
Pairwise multiple comparison showed a significant increase in RNA concentrations at 6 h compared with 0 h after venesection (Student–Newman–Keuls test, P <0.05). On the other hand, when EDTA blood was left at 4 °C, no significant difference was found for plasma RNA concentrations measured for the three incubation time points (Friedman test, P = 0.182; Fig. 1A⇑ ).
To study the effect of time delay on the concentrations of non-particle-associated RNA, we filtered plasma through a 0.22 μm filter before nucleic acid extraction. Similar results were obtained for EDTA blood left at room temperature and 4 °C. In both cases, no significant differences were observed for filtered plasma RNA concentrations from blood samples incubated for 0, 6, and 24 h (Friedman test, P = 0.124 for room temperature incubation and P = 0.522 for 4 °C incubation; Fig. 1B⇑ ).
Apart from RNA quantification, DNA concentrationswere also measured in the same samples. As shown in Fig. 1⇑ , C and D, no significant differences were observed up to 24 h of incubation in each case (Friedman test, P = 0.367 for blood incubated at room temperature without filtration, P = 0.814 for blood incubated at 4 °C without filtration, P =0.794 for blood incubated at room temperature with filtration, and P = 0.971 for blood incubated at 4 °C with filtration).
Delayed serum separation.
Similar studies were carried out for clotted blood. When the clotted blood was left for 0, 6, and 24 h, the unfiltered serum RNA concentrations demonstrated a statistically significant difference, regardless of whether an incubation temperature of 4 °C or room temperature was used (Friedman test, P <0.05 for both cases; Fig. 2A⇓ ).
Pairwise multiple comparison was performed. At room temperature, the total serum RNA concentrations significantly increased from 0 to 6 h, but then significantly decreased at 24 h when compared with 6 h (Student–Newman–Keuls test, P <0.05 for both cases). At 4 °C, the RNA concentrations during the first 6 h did not significantly change (Student–Newman–Keuls test, P >0.05), but then significantly increased at 24 h when compared with 6 h (Student–Newman–Keuls test, P <0.05).
In contrast, when serum was filtered, the resulting non-particle-associated RNA concentrations showed no significant changes among the three incubation time points (Friedman test, P = 0.093 for room temperature incubation and P = 0.182 for 4 °C incubation; Fig. 2B⇑ ).
The corresponding serum DNA concentrations are shown in Fig. 2⇑ , C and D. Significant differences were found when the uncentrifuged clotted blood was left at room temperature, regardless of whether filtration was applied after serum collection (Friedman test, P <0.05 for both cases; Fig. 2⇑ , C and D). For the unfiltered serum, the DNA concentrations increased significantly during the first 6 h (Student–Newman–Keuls test, P <0.05), but remained unchanged at 24 h when compared with 6 h (Student–Newman–Keuls test, P >0.05). For the filtered serum, the DNA concentrations significantly increased at 24 h when compared with 0 h.
On the other hand, when the uncentrifuged clotted blood was left at 4 °C, a significant difference was found for the DNA concentrations in unfiltered serum (Friedman test, P <0.05) but not the filtered serum DNA concentrations (Friedman test, P =0.124; Fig. 2⇑ , C and D). For the unfiltered serum, the DNA concentrations remained statistically unchanged at the first 6 h (Student–Newman–Keuls test, P >0.05), but significant increased at 24 h when compared with 6 h (Student–Newman–Keuls test, P <0.05).
lability of free added rna in plasma
The change in detectable RNA concentrations after a commercially available extracted RNA sample was added into a plasma sample is shown in Fig. 3⇓ . In the first 5 s, the RNA concentration decreased dramatically, from 67 605 ng/L to 116 ng/L. After a 15-s incubation time, >99% of the RNA became nonamplifiable. The sample with no exogenously added RNA gave an indication of the concentration of endogenous circulating RNA in the tested plasma sample.
effects of freezing and thawing
Experiments were carried out to investigate the RNA concentrations in freeze-thawed plasma. Liquid nitrogen was used in the process to minimize the time of achieving the frozen state. As shown in Fig. 4A⇓ , no significant difference was found for total RNA concentrations among untreated plasma, freeze-thawed plasma, and freeze-thawed plasma that had been left for 1 h at room temperature (Friedman test, P = 0.465). The filtered, non-particle-associated plasma RNA concentrations showed the same result (Friedman test, P = 0.526).
The DNA concentrations in the corresponding freeze-thawed samples were also measured (Fig. 4B⇑ ), and did not demonstrate any significant difference for samples that had been filtered (Friedman test, P = 0.531) or not filtered (Friedman test, P = 0.654).
The effect of freezing and thawing of serum on RNA concentrations is shown in Fig. 4C⇑ . For both filtered and unfiltered samples, no significant differences in serum RNA concentrations were found among the three treatment groups (Friedman test, P = 0.142 for unfiltered serum and P = 0.430 for filtered serum).
The effect of freezing/thawing of plasma on the concentrations of RNA after the thawed plasma had been passed through different-sized filters is shown in Fig. 4D⇑ . For the unfiltered, 5 μm-, 0.45 μm-, and 0.22 μm-filtered plasma, both freeze-thawed and never-frozen paired samples had similar plasma RNA concentrations.
Circulating RNA has been thought to be unstable because of the presence of blood RNases (7). In cancer patients, increased serum RNases has been detected (7). The effect of blood RNases can be demonstrated by the free RNA lability experiment, which showed that >99% of the free added RNA was degraded after a 15-s incubation time (Fig. 3⇑ ). This lability contrasts with the remarkable stability of plasma/serum RNA (Figs. 1⇑ and 2⇑ ), which suggests that plasma/serum RNA is protected by some mechanisms.
In this study, we have demonstrated that circulating cell-free RNA is stable enough to be amplified after the blood is left for 24 h. However, when quantitative study is carried out, precautions should be taken in the storage conditions of unprocessed blood to minimize artifactual fluctuations in nucleic acid concentrations. Such artifactual fluctuations could be attributable to the interplay of two factors: (a) the release of RNA from necrotic and/or apoptotic blood cells (13)(17), and (b) the stability of the original and the newly released RNA mentioned in (a). For example, with regard to delayed processing of EDTA blood, the interplay of these two factors is such that room temperature incubation would lead to an increase in total RNA concentrations within 6 h (Fig. 1A⇑ ). It is also striking to see that for filtered samples, the concentration of RNA is very stable (Fig. 1B⇑ ). This suggests that most of the increase in plasma RNA observed for unfiltered samples at 6 h (when stored at room temperature) is particle-associated. In view of these results, we would recommend storing unfractionated EDTA blood at 4 °C. In contrast to the RNA concentrations, the corresponding DNA concentrations were stable over 24 h (Fig. 1⇑ , C and D). This difference may be attributable to the cytoplasmic localization of RNA, which may facilitate easier assess of the RNA into the plasma than the genomic DNA located in the nucleus.
The RNA concentration in clotted blood is even more variable (Fig. 2⇑ ). To obtain a reliable serum RNA concentration, the unprocessed clotted blood should be stored at 4 °C and processed within 6 h. The variation in RNA concentrations may also be attributable to the two factors mentioned above. For both serum RNA and DNA, the initial increase in concentrations during the first 6 h (room temperature incubation; Fig. 2⇑ , A and C) may be attributable to the release of RNA and DNA from blood cells during the blood clotting process. After that, the RNA concentrations started to decrease, whereas the DNA concentrations remained unchanged. This may be a result of the degradation of the newly released RNA, whereas the DNA remained stable.
In a previous study by Kopreski et al. (2), freezing and thawing of serum was suggested to promote degradation of RNA. However, in the present study, we have shown that freezing and thawing of plasma or serum did not significantly change the RNA concentrations, even after the thawed plasma or serum had been left for 1 h at room temperature. This finding is contradictory to the results reported by Kopreski et al. (2), which showed that a single freeze–thaw cycle led to a 10- to 100-fold reduction of c-abl and tyrosinase mRNA in serum and that these mRNA species were undetectable after the serum was thawed and left for 30 min (2). The difference between the finding of Kopreski et al. and our findings may be attributable to several factors. The first factor is that we used GAPDH mRNA as the mRNA target, whereas the targets used by Kopreski et al. (2) were c-abl and tyrosinase mRNA. It is currently unknown whether different mRNA species may have different stabilities in plasma/serum. In addition, we measured the mRNA concentrations by a real-time quantitative RT-PCR assay, but Kopreski et al. (2) used a semiquantitative method. From one of the experiments in the present study, we showed that the concentrations of RNA associated with particles of different sizes were unaffected by the freezing/thawing of plasma (Fig. 4D⇑ ).
In conclusion, this study has demonstrated that the concentration of circulating RNA is surprisingly stable under different preanalytical situations. This information would probably make the study of both particle-associated and non-particle-associated RNA in plasma and serum simpler and more convenient. However, because this study is based on plasma and serum obtained from healthy individuals, whether the results can be generalized to all cases requires further investigation. In our preliminary observation, the stability of GAPDH mRNA concentrations in the plasma of pregnant women was similar to that of the healthy individuals (not shown). Nonetheless, the stability of tumor-derived (1)(2)(3)(4) and fetus-derived RNA (5) in the plasma/serum of cancer patients and pregnant women remains to be elucidated.
This work is supported by the Innovation and Technology Fund (AF/90/99). We thank Rossa Chiu, Eric Wong, Lisa Chan, Allen Chan, Yanni Lui, Katherine Chow, and Blenda Wong for help and advice during this study.
- © 2002 The American Association for Clinical Chemistry