This paper compared quantification of miRNA from reference microRNA (miRNA, miR) and pooled plasma using five different platforms. The influence of using different input amounts of miRNA on miRNA expression profiles generated by microarray was investigated using pooled plasma miRNA and reference miRNA.
Defining quantification methods and optimizing protocols for microarray hybridization of circulating microRNAs.
Garcia-Elias A, Alloza L, Puigdecanet E, Nonell L, Tajes M, Curado J, Enjuanes C, Díaz O, Bruguera J, Martí-Almor J, Comín-Colet J, Benito B
Purpose
Conclusion
Based on bioanalyzer electropherograms reference miRNA contained ribosomal RNA not present in plasma specimens. The mean quantified levels of reference miRNA were closest to the expected amount when quantified with the Tecan Infinite 200 PRO Nanoquant platform (Nanoquant), the Thermo Scientific Nanodrop 2000 platform (Nanodrop), or the Agilent Bioanalyzer with the Bio-PicoChip platform (Bio-PicoChip). This was expected as the reference miRNA also contains larger RNA and the aformentioned methods quantify all RNA in the specimen, whereas the Agilent Bioanalyzer with Bio-SmallChip platform (Bio-SmallChip) and the Qubit microRNA Assay kit are specific for small RNA. Nanoquant, Nanodrop, Qubit, Bio-PicoChip showed an acceptable (<10%) coefficient of variance (CV) for replicates when quantifying 10 ng/µL reference miRNA, but the CV was high when quantifying 1 ng/µL reference miRNA with the Nanoquant and Nanodrop methods. The high CV with the Nanoquant and Nanodrop methods was expected as 1 ng/µL is below the detection threshold for these methods. Similar to the results observed for reference RNA, RNA concentration quantified in plasma by Nanoquant and Nanodrop were 8- and 11-fold higher, respectively, than those for case-matched specimens quantified by the Qubit technique; however, but when RNA concentration was determined by Qubit values were higher than with Bio-PicoChip or Bio-SmallChip platforms. Importantly while plasma miRNA concentrations were strongly correlated when measured using Nanoquant and Nanodrop, they were each only weakly correlated with concentrations determined by Qubit. When restricted to the most concentrated specimens (2 ng/µL by Qubit), correlations between Qubit and both Nanoquant and Nanodrop were very strong.
Since only 6 of the 11 plasma aliquots had greater or equal to the 130 ng of miRNA required for the GeneChip Array, the assay was run with 92 ng or 65 ng of reference in addition to the 130 ng where possible . miRNA expression was more strongly correlated when the input amounts of 130 ng and 92 ng miRNA were compared relative to the lower input amount investigated (65 ng), regardless of whether reference or plasma miRNA was used. miRNA expression sensitivity for the assay was also higher and more similar to the 130 ng standard when 92 ng of miRNA was used compared to an input amount of 65 ng.
Studies (1)
Purpose
This study compared quantification of miRNA from reference miRNA and pooled plasma using five different platforms. The influence of using different input amounts of miRNA on miRNA expression profiles generated by microarray was investigated using pooled plasma miRNA and reference miRNA. Plasma was obtained from EDTA blood collected from 12 healthy volunteers by centrifugation at 1500 g for 15 min at 4°C. Plasma was pooled and aliquoted to produce twenty replicate pools which were stored at -80°C. miRNA was extracted using the miRNeasy Serum/Plasma kit. miRNA in plasma pools and the Universal Human miRNA Reference from Agilent Technologies was quantified using the Tecan Infinite 200 PRO Nanoquant platform, the Thermo Scientific Nanodrop 2000 platform, the Qubit microRNA Assay kit, and an Agilent Bioanalyzer with the Bio-PicoChip and Bio-SmallChip platforms. Expression miRNA profiles were generated from 65, 92, or 130 ng (as measured by Qubit) reference miRNA and plasma miRNA using the hybridization GeneChip miRNA 3.0 and 4.0 high-density microarrays. Small RNA libraries were constructed using the NEXTflexTM Small RNA-seq Kit V3 and sequenced on an Illumina Miseq system.
Summary of Findings
Based on bioanalyzer electropherograms, the reference miRNA, but not plasma contained ribosomal RNA. As RIN are based on the presence of larger RNA, the reference miRNA specimen had a much higher RIN than plasma miRNA samples. Quantification of ten replicates of reference RNA by Nanoquant, Nanodrop, Qubit, Bio-PicoChip methods showed acceptable coefficients of variance (CV, 2.03-5.78%), but a higher CV was observed when miRNA was quantified with the Bio-SmallChip platform (12.25%). Quantification of miRNA in replicate plasma pools was highly variable by all methods (Nanoquant CV =29.31%, Nanodrop CV=26.22%, Qubit CV= 26.16%, Bio-PicoChip CV= 67.16% and Bio-SmallChip CV=82.63%). The mean quantified levels were closest to expected level for 10 ng/µL and 1 ng/µL reference miRNA when quantification was with Nanoquant, Nanodrop, or Bio-PicoChip platforms which is consistent with these methods quantifying all RNA present in a specimen. When a 1 ng/uL reference miRNA sample was quantified, the CV for replicated was acceptable when Qubit and Bio-PicoChip methods were used (2.81% and 8.26%, respectively); however, CVs were higher when quantification was by Nanoquant, Nanodrop, or Bio-SmallChip platforms (40.0%, 23.83%, and 44.11%, respectively). The high CVs observed when Nanoquant and Nanodrop methods were used was expected since 1 ng/µL is below the detection threshold for these methods. Similar to results reported for the reference miRNA, miRNA concentrations quantified in plasma samples by Nanoquant and Nanodrop were 8- and 11-fold higher, respectively, than those for the same specimens by the Qubit technique; however, miRNA concentrations determined by Qubit were higher than with Bio-PicoChip or the Bio-SmallChip platforms. The authors postulated that the higher concentrations observed with spectrophotometric methods was attributable to quantification of non-miRNA RNA as well as contaminant interference, which was supported by mean 260/230 ratios and 260/280 ratios of 0.14 and 1.22, respectively. Importantly, while plasma miRNA concentrations were strongly correlated when measured using Nanoquant and Nanodrop (R2 = 0.879) they were each only weakly correlated with Qubit concentrations (R2 = 0.290 and R2 = 0.333, respectively). When restricted to the most concentrated specimens (2 ng/µL by Qubit), correlations between Qubit and both Nanoquant and Nanodrop were strong (R2 = 0.935 and R2 = 0.961, respectively).
Mapping rates for sequencing 20-40 nt RNA from plasma and the reference miRNA were 47% and 61%, respectively. Of the mapped reads, 84.4% and 58.2% mapped to miRNA in the reference miRNA and plasma specimen, respectively. Interestingly 28% of the mapped reads from plasma were mapped to unannotated RNA. miRNA expression was more strongly correlated when the input amounts of 130 ng and 92 ng miRNA were compared relative to the lower input amount investigated (65 ng), regardless of whether reference or plasma miRNA was used. Similarly, the sensitivity and specificity for miRNA expression relative to 130 ng reference miRNA sample was higher when 92 ng was used as input rather than 65 ng (97.8% versus 94.8% and 99.7% versus 99.6%, respectively). While comparatively fewer miRNAs were expressed in plasma compared to the reference miRNA sample (118 versus 615), similar effects of input amount were observed. miRNA expression data generated with 92 ng of plasma miRNA were more strongly correlated to data from 130 ng miRNA than 65 ng of miRNA (ρ=0.961 versus ρ=0.891). Sensitivity but not specificity for expression when 92 ng was used as input was also higher compared to when 65 ng of miRNA from plasma was used as input (92.8% versus 81.4%; 99.5% versus 99.7%, respectively).
Biospecimens
FluidPlasmaPreservation Types
FrozenDiagnoses
NormalPlatforms
RNAAutomated electrophoresis/BioanalyzerRNANext generation sequencingRNADNA microarrayRNAFluorometryRNASpectrophotometryPre-analytical Factors
Classification Pre-Analytical Factor Value(s) Fluorometry Specific Technology platform Agilent Bioanalyzer with the Bio-PicoChipAgilent Bioanalyzer with the Bio-SmallChip platformQubit microRNA Assay kitTecan Infinite 200 PRO Nanoquant platformThermo Scientific Nanodrop 2000 platformDNA microarray Specific Template/input amount 130 ng65 ng92 ng