Human blood RNA stabilization in samples collected and transported for a large biobank

Abstract

Background: The Norwegian Mother and Child Cohort Study (MoBa) is a nation-wide population-based pregnancy cohort initiated in 1999, comprising more than 108.000 pregnancies recruited between 1999 and 2008. In this study we evaluated the feasibility of integrating RNA analyses into existing MoBa protocols. We compared two different blood RNA collection tube systems - the PAXgene™ Blood RNA system and the Tempus™ Blood RNA system - and assessed the effects of suboptimal blood volumes in collection tubes and of transportation of blood samples by standard mail. Endpoints to characterize the samples were RNA quality and yield, and the RNA transcript stability of selected genes.

Findings: High-quality RNA could be extracted from blood samples stabilized with both PAXgene and Tempus tubes. The RNA yields obtained from the blood samples collected in Tempus tubes were consistently higher than from PAXgene tubes. Higher RNA yields were obtained from cord blood (3 - 4 times) compared to adult blood with both types of tubes. Transportation of samples by standard mail had moderate effects on RNA quality and RNA transcript stability; the overall RNA quality of the transported samples was high. Some unexplained changes in gene expression were noted, which seemed to correlate with suboptimal blood volumes collected in the tubes. Temperature variations during transportation may also be of some importance.

Conclusions: Our results strongly suggest that special collection tubes are necessary for RNA stabilization and they should be used for establishing new biobanks. We also show that the 50,000 samples collected in the MoBa biobank provide RNA of high quality and in sufficient amounts to allow gene expression analyses for studying the association of disease with altered patterns of gene expression.

Figure 1

Flow chart: experimental design.

Figure 2

Comparison of PAXgene and Tempus tubes. Blood (adult or cord blood) was collected either in the PAXgene tubes or in the Tempus tubes, and stored at room temperature for the indicated number of days, followed by total RNA isolation. Five tubes were used at each time point (n=5). A) The RNA yield from the cord blood and the adult blood samples collected either in the PAXgene tubes or in the Tempus tubes. Each bar represents the average RNA yield per tube ± SE. B) Relative transcript levels of four genes from adult blood samples collected in the PAXgene tubes and stored for up to seven days at room temperature; C) Relative transcript levels of four genes from adult blood samples collected in the Tempus tubes and stored for up to seven days at room temperature. The 0 day samples were used as reference samples (calibrators); hence all other samples were compared against the reference samples. Each bar represents the average log2-transformed fold change values and the error bar indicates ± SE. The dashed lines indicate ± 2 – fold change. The fold change detected for the genes were within ± 2 – fold. D) The normalized expression values of four genes from cord blood samples collected in the PAXgene tubes (n=8) and in the Tempus tubes (n=7). Each bar represents the normalized expression values [2^{-ΔCq (sample)}; where the ΔCq (sample) ± Cq (gene) – Cq (internal control)] and the error bar indicates ± SE.

Figure 3

The effects of transportation on RNA, quantity, quality and transcript stability. Adult blood from one healthy donor was collected in Tempus™ tubes (n = 20). Samples were divided randomly into five groups (n = 4 tubes per group): four groups were sent by standard mail to four different hospitals in Norway, and the fifth group was kept at NIPH (Oslo/lab). Temperature was monitored during transportation. A) The effects of transportation on RNA quality and yield; B) The effects of transportation on RNA transcript levels of six genes. Samples kept at NIPH were used as reference samples; hence all other samples were compared against the reference samples. Each bar represents the average log2-transformed fold change values and the error bar indicates ± SE. The dashed lines indicate ± 2 – fold change.

Figure 4

The effect of suboptimal blood volume collected in the Tempus tubes on RNA quality and transcript stability. A) Cord blood collected in Tempus tubes from maternity units in a period of one year (n = 10 853). B) The effects of suboptimal volume of adult blood collected in the Tempus tubes on relative transcript level for six genes. Adult blood from one healthy donor was collected in Tempus tubes (n = 18). The following blood volumes (3.0, 2.5, 2.0, 1.5, 1.0 and 0.5 ml) were collected in the Tempus tubes (n = 3; three parallels tubes for each blood volume). C) The effects of suboptimal volume of cord blood collected in the Tempus tubes on relative transcript level for six genes. Cord blood from one donor was collected in Tempus tubes (n = 10). The following blood volumes (3.3, 3.0, 2.0, 1.0, and 0.5 ml) were collected in the Tempus tubes (n = 2; two parallels tubes for each blood volume). Samples with 3 ml blood per tube were used as reference samples; hence all other samples were compared against the reference samples. Each bar represents the average log2-transformed fold change values and the error bar indicates = SE. Since some of the samples had low RNA quality, selecting stably expressed reference genes for normalization was a challenge. We therefore used the average Cq value for each gene to normalize the samples: ΔCq (sample) = Cq (gene) – Cq (average); these values where converted into linear scale, 2^-ΔCq. Fold changes were calculated by dividing ΔCq value of the sample by the ΔCq value of the reference sample.