Skip to main content

MicroRNA Stability in FFPE Tissue Samples: Dependence on GC Content.

Kakimoto Y, Tanaka M, Kamiguchi H, Ochiai E, Osawa M

Publication: PLoS One (2016) • Vol 11 • Page e0163125PMID: 27649415View on PubMed

Purpose

This paper compared next-generation sequencing (NGS) results in three paired formalin-fixed paraffin-embedded (FFPE) and frozen cardiac specimens and examined the effect of GC content on miRNA stability by real-time PCR using ten paired FFPE and frozen left ventricular wall specimens.

Conclusion

FFPE specimens had shorter read lengths, lower percentage of reads mapped to the human genome and to miRNA, and more degradation at the ends of the reads than frozen specimens. However, normalized read counts of the 240 miRNAs with more than 10 reads per million mapped in at least one specimen were strongly correlated between matched frozen and FFPE specimens (r=0.88-0.92). While lists of the top ten most abundant miRNA were similar in FFPE and frozen specimens, levels of 30 miRNAs in FFPE specimens were less than 50% of the levels in the matched frozen specimen and the levels of 69 miRNAs were 2-fold higher in FFPE specimens than the matched frozen specimens. miRNAs with higher GC content were more likely to have higher expression in FFPE than frozen specimens as determined by real-time PCR and NGS.

Studies (1)

Purpose

This study compared NGS results in three paired FFPE and frozen cardiac specimens and examined the effect of GC content on miRNA stability by real-time PCR using ten paired FFPE and frozen left ventricular wall specimens. Ten matched specimens of left ventricular wall obtained at autopsy (post mortem interval of 20-120 h) were immersed in liquid nitrogen and stored at -80˚C for less than 2 months or fixed in 10% formalin for 5 days to 1.5 months, embedded in paraffin, and stored at room temperature for up to 13 months. All specimens appeared microscopically normal. RNA was isolated from 100 mg of crushed frozen specimens using mirVana miRNA Isolation Kit and from one or two 20 µm xylene-deparaffinized FFPE sections using the RecoverAll kit. RNA concentration was determined by spectrophotometry and RNA integrity and purity using a bioanalyzer. RNA was stored at -80˚C. RNA libraries were prepared from 24 ng of RNA from three matched ventricular wall specimens using the Ion Total RNA-seq kit V2 and sequenced on an Ion Torrent PGM system. Sequencing results of miR-99b-5p, miR133a-3p, miR-133b, miR-22-3p, miR21-5p, miR-1-3p, and let-7e-5p were verified by TaqMan real-time PCR in all specimens.

Summary of Findings

The most common read length was 22 nt for frozen specimens but was only 11 nt for the matched FFPE specimens. For frozen specimens, 92.1% of the reads were mapped to the human genome and 32% of these mapped to miRNA, but for FFPE specimens only 87% mapped to the human genome, of which only 9.4% mapped to miRNA. Importantly, the specimen with the longest fixation duration of the three also had the lowest mapping rate to the genome and miRbase. Examination of the reads showed degradation at both ends but more degradation was observed at the 3’ end than the 5’ end and more degradation was noted in FFPE specimens than frozen specimens. The normalized read counts of the 240 miRNAs with more than 10 reads per million mapped in at least one specimen were strongly correlated between matched frozen and FFPE specimens (r=0.88-0.92). However, levels of 30 miRNAs in FFPE specimens were less than 50% of the levels in the matched frozen specimen and levels of 69 miRNAs were 2-fold higher in FFPE specimens than the matched frozen specimens. There was a modest correlation of the fold change between FFPE and frozen specimens and the GC content of the miRNA (r=0.42) such that miRNAs with higher GC content were more likely to have higher expression in FFPE than frozen specimens. Further analysis showed the average GC content was significantly lower in specimens with a <0.5-fold change than a 0.5 to 2-fold change (0.36 versus 0.47, P<0.0001) and lower in specimens with a 0.5 to 2-fold change than those with >2-fold change (0.47 versus 0.54, P<0.0001). Real-time PCR analysis confirmed that the ratio of FFPE to frozen miRNA levels was higher for miRNAs with GC content of 41-64% than those with a GC content of 27-36% (P<0.0001). While lists of the ten most abundant miRNAs in each specimen were similar between FFPE and frozen specimens, differences included miR-1-3p accounting for 19-31% of miRNA in frozen specimens but only 4-9% in FFPE specimens and miR-133a-3p accounting for 11-14% of miRNA in FFPE specimens but only 3-5% of the miRNA in frozen specimens.

  • Biospecimens

    TissueHeart
  • Preservation Types

    FormalinFrozen
  • Diagnoses

    Autopsy
  • Platforms

    RNANext generation sequencing
    RNAAutomated electrophoresis/Bioanalyzer
    RNAReal-time qRT-PCR
  • Pre-analytical Factors

    ClassificationPre-Analytical FactorValue(s)
    Biospecimen PreservationType of fixation/preservation
    Formalin (buffered)
    Snap frozen
    Real-time qRT-PCR SpecificTargeted nucleic acid
    let-7e-5p
    miR-1-3p
    miR-22-3p
    miR-99b-5p
    miR133a-3p
    miR133b
    miR21-5p
    Next generation sequencing SpecificTechnology platform
    Real-time PCR