Tips from the Bench: Does Fractionation of Blood Samples Affect mRNA Expression Levels?

When studying gene expression there is always the concern that sample manipulation prior to RNA isolation may lead to changes in mRNA expression patterns. An example of a common sample manipulation is fractionation of blood. Fractionation of white blood cells (WBCs), aka leukocytes, offers several advantages over use of whole blood, including concentration of the sample, removal of the RNases found in plasma, and the ability to use a more homogeneous population of cells for analysis. Here we discuss an experiment that addresses whether blood cell fractionation affects the levels of several key mRNAs expressed in leukocytes.

Blood fractionation to obtain the “buffy coat” fraction containing the WBCs involves centrifuging anticoagulated blood to separate the cell components according to density, allowing recovery of the total leukocyte population. After centrifugation, the blood separates into an upper plasma fraction, comprising ~60% of the sample volume, a lower red blood cell fraction comprising ~40%, and a thin interface layer, called the “buffy coat”. The buffy coat is typically recovered by aspiration after removing the plasma; contamination of the buffy coat fraction with red blood cells is unavoidable. The extent to which centrifugation affects mRNA profiles in WBCs recovered in the buffy coat fraction is a concern.

Real-time RT-PCR was used to assess relative levels of several sentinel mRNAs in unmanipulated whole blood and in blood that was first fractionated using centrifugation and then remixed. This strategy was used so that any differences in mRNA levels could not be attributed to using isolated fractionated leukocytes versus whole blood, but would instead relate to differences due to the experimental treatment.

Several tubes of blood were collected from a healthy volunteer into evacuated tubes containing EDTA anticoagulant (Terumo). Triplicate aliquots (0.5 ml) of whole blood were removed from each tube, mixed with 1.3 ml of RNAlater® Tissue Collection:RNA Stabilization Solution (see below, What is RNAlater?) to stabilize the mRNA profiles, and stored at −20°C. The remaining samples were centrifuged for 10 min at 3,000 rpm in a clinical centrifuge. The plasma was aspirated into a 5 ml pipet and then was returned to the sample (to mimic the procedure used to collect buffy coat). These samples were mixed by gentle inversion, and then triplicate 0.5 ml aliquots were mixed with RNAlater and stored at -20°C.

Total RNA was extracted from all 6 samples (triplicate untreated and triplicate centrifuged samples) using the RiboPure™-Blood Kit, and 100 ng of each was used in qRT-PCR assays using Applied Biosystems Taqman® probes and primers. The panel of mRNAs amplified was chosen to include genes whose expression in previous studies was shown to be especially volatile in blood (e.g., p53 tumor suppressor gene, cytokine IL-8, and c-jun oncogene). mRNA levels in each sample were expressed as fold-difference relative to one of the controls arbitrarily chosen to be the calibrator sample, as calculated using the ΔΔCt method.

Figure 1 demonstrates the minimal variation observed between replicates and between control samples and centrifuged samples. Variations were less than 1.5-fold in almost every case. The greatest variation was seen for TNF (tumor necrosis factor), especially for one of the centrifuged samples, sample “f”, in which the mRNA level was lower (~0.4-fold of control) in one of the three replicates.

Mock Buffy Coat = blood was centrifuged and remixed; HPRT = Hypoxanthine-Guanine; Phosphoribosyl Transferase, TNF = Tumor Necrosis Factor; TGFb = Transforming Growth Factor β Casp = Caspase 3; PTGS2 = prostaglandin-endoperoxide synthase 2; SD = standard deviation.

Figure 1. Fractionation of Blood Cells by Centrifugation Minimally Affects mRNA Expression Levels. RNA was isolated using the RiboPure™-Blood Kit from triplicate aliquots of Whole Blood or Mock Buffy Coat (blood centrifuged, buffy coat aspirated and then returned to tube) samples stabilized in RNAlater®. qRT-PCR (100 ng input RNA) was performed on each sample using Applied Biosystems Taqman® probes and primers. The first replicate was the calibrator and 18S rRNA was used for normalization. Ct values were averaged among triplicate samples; values were very similar for every gene targeted. One of the triplicate Mock Buffy Coat reactions showed a significant TNF expression change (note higher standard deviation and coefficient of variance); this data point may be considered an outlier.

Overall the data generated in this experiment suggest that centrifugation of whole blood using conditions required to obtain the WBCs within the buffy coat fraction does not significantly alter mRNA expression levels.

Scientific Contributors
Marianna Goldrick, Nathan Harris • Ambion, Inc.