Skip to main content

Quantification of microRNA levels in plasma - Impact of preanalytical and analytical conditions.

Binderup HG, Madsen JS, Heegaard NHH, Houlind K, Andersen RF, Brasen CL

Publication: PLoS One (2018) • Vol 13 • Page e0201069PMID: 30024941View on PubMed

Purpose

This paper investigated the effects of draw order, centrifugation steps, and normalization method on the measurement of microRNA (miRNA, miR) levels in plasma from a single patient and compared levels of miRNA in standard and platelet-poor plasma (PPP).

Conclusion

Draw order had no effect on the levels of miRNA in specimens from a single patient, regardless of normalization method. Levels of miR-92a, miR-126, and miR-16 were comparable in plasma obtained by dual centrifugation and prolonged single centrifugation when normalized to cel-miR-39, but levels of miR126 were higher in plasma obtained by dual centrifugation than that obtained by a single prolonged centrifugation step when normalized to miR-16. Centrifugation protocol was found to account for 64-73% of the intra-assay variability when normalized to cel-miR-39, 39-66% when normalized to miR-16, and 55-61% when normalized to both cel-miR-39 and miR-16. Correlations between PPP and standard plasma were miRNA- and normalizer-dependent. Levels of miR-92a were not correlated between individual TaqMan assays and TaqMan low density arrays (TLDA) assays, but modest correlations between methods were observed for miR-16. miRNA levels were very strongly correlated between real-time (q)PCR and droplet digital (dd)PCR when normalized to cel-miR-39, but only modest to very strong when normalized to miR-16 alone or in combination with cel-miR-39.

Studies (1)

Purpose

This study investigated the effects of draw order, centrifugation steps, and normalization method on the measurement of miRNA levels in plasma from a single patient and compared levels of miRNA in standard plasma and PPP. Blood was collected using 21-gauge needles into K2EDTA tubes with intermittent claudication after discarding the first 3 mL. PPP was obtained from 30 blood specimens from a single healthy patient using a standard dual centrifugation protocol (two steps at 3000 x g for 15 min) and by a prolonged single centrifugation (3000 x g for 30 min). Blood from 50 healthy patients was used to obtain standard plasma by centrifugation at 2000 x g for 10 min and PPP by the standard dual centrifugation protocol. Plasma was thawed and recentrifuged at 3000 x g for 15 min and miRNA was purified using the NucleoSpin miRNA Plasma kit. Levels of miR-92a, miR-126, and miR-16 were quantified by TaqMan real-time PCR assays, after preamplification using the TaqMan PreAmp Master kit using TaqMan low density arrays, and by ddPCR.

Summary of Findings

Draw order had no effect on the levels miRNA in specimens from a single patient, regardless of normalization method. Levels of miR-92a, miR-126, and miR-16 were comparable in plasma obtained by dual centrifugation and prolonged single centrifugation when normalized to cel-miR-39, but levels of miR126 were higher in plasma obtained by dual centrifugation than that obtained by a single prolonged centrifugation step when normalized to miR-16 (P<0.02). The coefficients of variation depended on the normalization method and was highest when normalized to cel-miR-39 (10.9-21.1%) and lower when normalized to miR-16 alone (9.5-14.2%) or in combination with cel-miR-39 (9.2-13.3%). Centrifugation protocol was found to account for 64-73% of the intra-assay variability when normalized to cel-miR-39, 39-66% when normalized to miR-16, and 55-61% when normalized to both cel-miR-39 and miR-16.

Cel-miR-39 normalized miRNA levels in PPP were not correlated to those in standard plasma, but miR-92a and miR-126 levels in PPP were modestly correlated to those in standard plasma when normalized to miR-16 (ρ=0.42, P=0.0024 and ρ=0.48, P=0.0004, respectively). Further, levels of miR-126 were modestly correlated in standard and PPP when normalized to cel-miR-39 in combination with miR-16, but there was no correlation observed for miR-92a levels normalized to cel-miR-39 with miR-16.

Levels of miR-92a were not correlated between individual TaqMan assays and TLDA assays, but modest correlations between methods were observed for miR-16 (ρ=0.47, P=0.0005). Inter-assay precision was slightly higher for qPCR than ddPCR, but intra-assay precision was significantly higher with ddPCR.  miRNA levels were very strongly correlated between qPCR and ddPCR when normalized to cel-miR-39 (ρ≥0.92, P<0.0001), but only modest to very strong when normalized to miR-16 alone or in combination with cel-miR-39 (ρ=0.57-0.94, P<0.0001).

  • Biospecimens

    FluidBlood
    FluidPlasma
  • Preservation Types

    Frozen
  • Diagnoses

    Normal
  • Platforms

    RNALow density array
    RNAReal-time qRT-PCR
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    Biospecimen Aliquots and ComponentsAliquot sequential collection
    10th collection
    1st collection
    2nd collection
    3rd collection
    4th collection
    5th collection
    6th collection
    7th collection
    8th collection
    9th collection
    Biospecimen Aliquots and ComponentsBlood and blood products
    Plasma
    Platelet-poor plasma
    Biospecimen Aliquots and ComponentsCentrifugation
    Different number of centrifugation steps compared
    Real-time qRT-PCR SpecificData handling
    Normalized to cel-miR-39
    Normalized to cel-miR-39 and miR-16
    Normalized to miR-16
    Real-time qRT-PCR SpecificTechnology platform
    Single TaqMan assay
    TLDA
    ddPCR