Abstract
Numerous studies have reported a potential role for circulating microRNAs as biomarkers in a wide variety of diseases. However, there is a critical reproducibility challenge some of which might be due to differences in preanalytical and/or analytical factors. Thus, in the current study we systematically investigated the impact of selected preanalytical and analytical variables on the measured microRNA levels in plasma. Similar levels of microRNA were found in platelet-poor plasma obtained by dual compared to prolonged single centrifugation. In contrast, poor correlation was observed between measurements in standard plasma compared to platelet-poor plasma. The correlation between quantitative real-time PCR and droplet digital PCR was found to be good, contrary to TaqMan Low Density Array and single TaqMan assays where no correlation could be demonstrated. Dependent on the specific microRNA measured and the normalization strategy used, the intra- and inter-assay variation of quantitative real-time PCR were found to be 4.2–6.8% and 10.5–31.4%, respectively. Using droplet digital PCR the intra-assay variation was 4.4–20.1%, and the inter-assay variation 5.7–26.7%. Plasma preparation and microRNA purification were found to account for 39–73% of the total intra-assay variation, dependent on the microRNA measured and the normalization strategy used. In conclusion, our study highlighted the importance of reporting comprehensive methodological information when publishing, allowing others to perform validation studies where preanalytical and analytical variables as causes for divergent results can be minimized. Furthermore, if microRNAs are to become routinely used diagnostic or prognostic biomarkers, the differences in plasma microRNA levels between health and diseased subjects must exceed the high preanalytical and analytical variability.
Highlights
MicroRNAs are short (~19–24 nucleotides), single-stranded non-coding RNAs acting as posttranscriptional regulators of gene expression
When microRNA levels were normalized to cel-miR-39, no significant differences in levels of miR-92a (p = 0.35), miR-126 (p = 0.26) or miR-16 (p = 0.25) between plasmas obtained by the two centrifugation protocols were found
When normalized to the endogenous miR-16 or to the mean of cel-miR-39 and miR-16, we found no significant differences in levels of miR92a (p!0.52), whereas levels of miR-126 were significantly higher in platelet-poor plasma (PPP) when obtained by dual centrifugation compared to prolonged single centrifugation (P 0.02),Table 1
Summary
MicroRNAs are short (~19–24 nucleotides), single-stranded non-coding RNAs acting as posttranscriptional regulators of gene expression. Preanalytical and analytical conditions are major sources of variation in and between microRNA studies [7,8,9,10,11], and only a minority of the reported results have been reproduced in other studies. There are many reasons for this but, our work, and that of others, has shown that the preanalytical centrifugation conditions have a large impact on the measured plasma/serum levels of many microRNAs [7,10,12]. Another important source of variation is microRNA purification. The choice of normalization strategy may have a significant impact on the reported microRNA levels [14,15]
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call