Cancer Diagnosis Program | Biorepositories and Biospecimen Research Branch

Search BRD Papers

mRNA-seq whole transcriptome profiling of fresh frozen versus archived fixed tissues.

Author(s): Bossel Ben-Moshe N, Gilad S, Perry G, Benjamin S, Balint-Lahat N, Pavlovsky A, Halperin S, Markus B, Yosepovich A, Barshack I, Gal-Yam EN, Domany E, Kaufman B, Dadiani M

Publication: BMC Genomics, 2018, Vol. 19, Page 419

PubMed ID: 29848287 PubMed Review Paper? No

Purpose of Paper

Conclusion of Paper

Studies

1. Study Purpose

   The purpose of this study was to compare mRNA expression profiles generated from matched FFPE and frozen breast tumor specimens using three different next generation sequencing library preparation methods. Sequencing metrics were also investigated in FFPE specimens stored for >10 years. Three matched pairs of FFPE and frozen breast tumors that had been stored for 4-5 years and three FFPE tumor specimens stored for >10 years were obtained (no further details provided). Three 5 μm sections were deparafinized at 90 °C for 5 min and RNA was isolated using a modification of the Qiagen AllPrep Kit. RNA was extracted from frozen specimens using the Trizol kit. RNA was quantified by Qubit fluorometer, and integrity verified by Agilent TapeStation. Sequencing libraries were prepared using the Truseq RNA Sample preparation kit v2 (mRNA-Seq), the Truseq Stranded total RNA with Ribo-Zero Gold (RiboZero) and the Nugen Ovation Human FFPE RNA-Seq Library Kit. Libraries were single end sequenced using an Illumina HiSeq V4.  To investigate the relationship between estrogen receptor mRNA levels and estrogen receptor protein levels, diagnostic FFPE slides were immunohistochemically stained and scored by a pathologist using the American Society of Clinical Oncology (ASCO) and the College of American Pathology (CAP) scoring system.

   Summary of Findings:

   RNA integrity numbers ranged from 7.7-7.4 for the three frozen specimens but were below the limit of detection for two of the FFPE specimens and 1.8 from the third. Compared to frozen specimens, FFPE specimens had a lower percentage of reads mapping to exons (29.2% versus 58% for mRNA-Seq, 8.4% versus 21.4% for RiboZero and 3.9% versus 20.3% for NuGen), a higher percentage of unmapped reads (24% versus 3.8% for mRNA-Seq, 9.6% versus 3% for RiboZero and 58.7% versus 4.8% for NuGen), and for mRNA seq and RiboZero a higher percentage of reads mapping to intragenic regions (28.2% versus 25%  and, 70.2% versus 44.4%, respectively). As use of the NuGen protocol for sequencing of FFPE specimens yielded a low percentage of reads that mapped to exons (3.9%) and a high percentage of reads that were unmapped (58.7%) this method was excluded from further analysis. A low percentage of reads from FFPE specimens mapped to ribosomal RNA (rRNA) when either mRNA-Seq or RiboZero were used (12.7% and 5.3%, respectively), but the percentage of ribosomal reads was much higher for frozen specimens with RiboZero than with mRNA-Seq libraries (24.7% versus 1.3%).  Further mRNA-Seq libraries had a higher exonic mapping rate than RiboZero libraries from both frozen (58% versus 21.4%) and FFPE specimens (29.2% versus 8.4%). The authors state that because of the lower percentage of exonic reads when sequencing with RiboZero a higher number of reads is necessary to detect 11,000 genes (70 million versus 26-42 million for mRNA-Seq). It is important to note that the 3’ bias produced by RiboZero was similar for frozen and FFPE specimens, but for mRNA-Seq a larger 3’ bias was noted in FFPE than frozen specimens. The correlation in gene expression were strong to very strong for matched FFPE and frozen specimens using the mRNA-Seq protocol (correlation coefficient =0.89-0.94), but for RiboZero the correlation coefficients were lower (0.86, 0.83 and 0.33).  Importantly, higher correlations were observed for mRNA-Seq than RiboZero regardless of amount of RNA input (100 ng or 500 ng). The fold change (≥2-fold) between two different tumor specimens was strongly correlated between frozen and FFPE specimens with 80% of genes that were differentially expressed between the two tumors displaying a change in the same direction in frozen and FFPE specimens, regardless of library preparation method. Importantly, levels of 701 mRNAs previously identified to be differentially expressed in breast cancer were strongly correlated between frozen mRNA-Seq libraries and matched FFPE RiboZero (Correlation coefficient =0.92) and mRNA-Seq libraries (correlation coefficient =0.94). Levels of estrogen receptor mRNA and immunohistological scores for estrogen receptor showed a good agreement.

   FFPE specimens stored for >10 years versus 4 years had a lower percentage of reads mapped to exons (20.1% versus 29.2%) and more unmapped reads (45.7% versus 24%), slightly increase in the number of reads needed to detect 11,000 genes.  Nevertheless, the agreement between estrogen receptor mRNA levels and immunohistological scores for estrogen receptor showed a good agreement in the FFPE specimens stored for 10 years.

   Biospecimens

   • Tissue - Breast

   Preservative Types

   • Frozen
   • Formalin

   Diagnoses:

   • Neoplastic - Carcinoma

   Platform:

   Analyte	Technology Platform
   RNA	Automated electrophoresis/Bioanalyzer
   Protein	Immunohistochemistry
   RNA	Fluorometry
   RNA	Next generation sequencing

   Pre-analytical Factors:

   Classification	Pre-analytical Factor	Value(s)
   Biospecimen Preservation	Type of fixation/preservation	Frozen Formalin (buffered)
   Next generation sequencing Specific	Template/input amount	100 ng 500 ng
   Immunohistochemistry Specific	Targeted peptide/protein	Estrogen receptor
   Storage	Storage duration	4-5 years >10 years
   Next generation sequencing Specific	Template modification	Illumina Truseq RNA after poly(A) selection (mNA-seq) Illumina Truseq after ribosomal depletion with Ribo-Zero Nugen Ovation with Ribosomal depletion

   View more