Optimization of FFPE preparation and identification of gene attributes associated with RNA degradation.
Author(s): Lin Y, Dong ZH, Ye TY, Yang JM, Xie M, Luo JC, Gao J, Guo AY
Publication: NAR Genom Bioinform, 2024, Vol. 6, Page lqae008
PubMed ID: 38298182 PubMed Review Paper? No
Purpose of Paper
This paper compared the integrity of RNA obtained from matched formalin-fixed, paraffin-embedded (FFPE) specimens that experienced a pre-fixation delay at 4°C or 25°C for up to 12 h (cold ischemia time); were fixed for 12, 24, or 48 h at 4°C or 25°C; and when extraction was from scrolls rather than sections. Additionally, the next generation sequencing transcriptome of frozen and FFPE tumor and peritumor specimens were compared, including specimens with a cold ischemia time of 6 or 48 h at 4°C or 25°C. Differences between FFPE and frozen specimens were further explored using publicly available data from the United States National Institutes of Health (NIH) Biospecimen Preanalytical Variable (BPV) program.
Conclusion of Paper
Specimens fixed for 48 h yielded RNA with a higher percentage of fragments >200 nt (DV200) and >800 nt (DV800) than those fixed for 12 h or 24, but there was no difference in DV200 or DV800 between specimens fixed at 4°C and those fixed at 25°C. A cold ischemia time of up to 12 h did not affect DV200 or DV800 values in FFPE specimens (48 h was not investigated); however, when cold ischemia was >0.5 h, DV200 and DV800 values were higher when the delay occurred at 4°C than to 25°C. Further, gene expression correlation coefficients with the matched frozen specimen were higher when cold ischemia of 6 or 48 h was at 4°C rather than at 25°C. DV200 was lower when RNA was extracted from sections (30-40%) than scrolls (>60%).
While there was no significant 3’ or 5’ bias using either FFPE or frozen specimens, FFPE specimens had a higher percentage of intronic reads and a lower percentage of exonic reads and 3’UTR exonic reads compared to matched frozen specimens. In principle component analysis, frozen and FFPE specimens clustered separately, and the rank order of genes was different between the preservation methods, with 4.4% of genes displaying a rank change > 1000. Analysis of differentially expressed genes in frozen and FFPE specimens from this project and kidney, ovary, and colon specimens from the BPV program identified 1001 genes that had higher transcript levels and 420 genes that had lower transcript levels in FFPE specimens relative to matched frozen specimens in at least two of the four tissue types. Genes with decreased expression (rather than unchanged or increased) in FFPE relative to frozen specimens were found to have a transcript length in the first decile (P<2.22 e-16) and a minimum free energy in the ninth decile (P<2.22 e-16) and to be extracellular. Interestingly, transcripts of mitochondrial genes tended to be higher in FFPE relative to frozen specimens. Of the 2176 housekeeping genes analyzed, 4-14% were differentially expressed between FFPE and frozen specimens, including GAPDH. While only 30-40% of the genes differentially expressed in tumor relative to peritumor specimens were identified using both FFPE and frozen specimens, many of the pathways identified were shared and the authors concluded that FFPE specimens can be useful for comparative transcriptome analysis in cancer specimens.
Studies
-
Study Purpose
This study compared the integrity of RNA obtained from matched FFPE specimens that experienced a pre-fixation delay at 4°C or 25°C for up to 12 h (cold ischemia time); were fixed for 12, 24, or 48 h at 4°C or 25°C; and when extraction was from scrolls rather than sections. Additionally, the next generation sequencing transcriptome of frozen and FFPE tumor and peritumor specimens were compared, including specimens with a cold ischemia time of 6 or 48 h at 4°C or 25°C. Differences between FFPE and frozen specimens were further explored using publicly available data from the United States National Institutes of Health (NIH) Biospecimen Preanalytical Variable (BPV) program. Lung specimens were collected from six patients (diagnosis not specified) and immediately cut into 0.3 cm x 0.3 cm x 0.3 cm specimens. From three patients, matched tumor and peritumor specimens were collected. Matched specimens were then immediately frozen at -80°C or fixed in 4% neutral paraformaldehyde solution at 25°C for 48 h and paraffin-embedded unless otherwise specified. To investigate the effects of cold ischemia on RNA integrity, matched specimens were stored for 0.5, 3, 6, or 12 h at 4°C or 25°C before fixation. Gene expression correlations with frozen specimens were compared among specimens subjected to 0.5, 6, or 48 h cold ischemia at 4°C or 25°C. To investigate the effects of fixation duration and temperature on RNA integrity, matched specimens were fixed for 12, 24, or 48 h at 4°C and 25°C. RNA was extracted from 5 µm scrolls or 5 µm sections of the FFPE specimen using the RNAstorm Kit and by an unspecified method from the frozen specimens. RNA integrity was evaluated by DV200 and DV800 using a bioanalyzer. Sequencing libraries were prepared using the Ribo-off rRNA Depletion Kit and pair-end sequenced using an Illumina Novaseq 6000 instrument. Genes were considered differentially expressed if the log2 of the fold change was >1 and the adjusted P-value was <0.05. Comparison to and further analysis were conducted using publicly available data for matched frozen and FFPE specimens of five renal clear cell carcinomas, seven serous ovarian carcinomas, and five colon adenocarcinomas collected, processed, and analyzed by the National Cancer Institute’s (NCI) Biorepositories and Biospecimen Research Branch (BBRB) BPV program.
Summary of Findings:
FFPE specimens fixed for 48 h yielded RNA with a higher percentage of fragments >200 nt (DV200) and >800 nt (DV800) than those fixed for 12 h (P=0.017 and P=0.0042, respectively) or 24 h (P=0.0075 and P=0.0025, respectively), but there was no difference in DV200 or DV800 between specimens fixed at 4°C and those fixed at 25°C. A cold ischemia time of up to 12 h did not affect DV200 or DV800 values in FFPE specimens (48 h was not investigated); however, when cold ischemia was >0.5 h, DV200 and DV800 values were higher when the delay occurred at 4°C than to 25°C. Most importantly, gene expression correlation coefficients with the matched frozen specimen were higher when a cold ischemia time of 6 or 48 h occurred at 4°C (ρ=0.79 and ρ=0.90, respectively) rather than at 25°C (ρ=0.48 and ρ=0.41, respectively). DV200 was lower when RNA was extracted from FFPE sections (30-40%) than scrolls (>60%).
While there was no significant 3’ or 5’ bias using either FFPE or frozen specimens, FFPE specimens had a higher percentage of intronic reads and a lower percentage of exonic reads and 3’UTR exonic reads compared to matched frozen specimens (P=1.5 e-6, all). Using the BPV data, the authors found DV200 was correlated with gene expression for only 41 genes. In principle component analysis, frozen and FFPE specimens clustered separately, and the rank order of genes was different, with 4.4% of genes displaying a rank change > 1000. Analysis of differentially expressed genes in frozen and FFPE specimens from this project and kidney, ovary, and colon specimens from the BPV program, identified 1001 genes that had higher transcript levels and 420 genes that had lower transcript levels in FFPE relative to matched frozen specimens in least two of the four tissue types examined. Analysis revealed an enrichment of genes involved in chemical stimulus perception (taste receptor genes and olfactory receptor genes) in FFPE specimens. Genes with lower transcript levels (rather than unchanged or increased) in FFPE specimens relative to frozen specimens were found to have a length in the first decile (P<2.22 e-16) and a minimum free energy in ninth decile (P<2.22 e-16) and to be extracellular. Interestingly, transcript levels of mitochondrial genes tended to be higher in FFPE specimens relative to frozen specimens. Of the 2176 housekeeping genes analyzed, 4-14% were differentially expressed between FFPE and frozen specimens, including GAPDH. The authors identified a panel of 40 housekeeping genes that they found to be stable for potential use as normalizers. Among the genes differentially expressed in tumor and peritumor specimens, 217 (38%) of the upregulated and 96 (31%) of the downregulated genes in cancer were identified in both FFPE and frozen specimens. On a pathway level, 7 of the top 10 upregulated pathways and 3 of the top 10 downregulated pathways in cancer were identified using both FFPE and frozen specimens. Analysis of Pan-Cancer annotations found 8 cancer-related pathways that were upregulated in tumor specimens relative to peritumor specimens, regardless of preservation method, and two more in FFPE specimens only. The authors concluded that FFPE specimens can be useful for comparative transcriptome analysis in cancer specimens.
Biospecimens
Preservative Types
- Frozen
- Formalin
Diagnoses:
- Neoplastic - Carcinoma
- Not specified
- Neoplastic - Normal Adjacent
Platform:
Analyte Technology Platform RNA Next generation sequencing RNA Automated electrophoresis/Bioanalyzer Pre-analytical Factors:
Classification Pre-analytical Factor Value(s) Storage Storage temperature 4°C
25°C
Biospecimen Preservation Time in fixative 12 h
24 h
48 h
Biospecimen Acquisition Cold ischemia time <0.5 h
3 h
6 h
12 h
48 h
Biospecimen Preservation Type of fixation/preservation Formalin (buffered)
Frozen
Biospecimen Preservation Temperature of fixation/preservation 4°C
25°C
Analyte Extraction and Purification Specimen format Scrolls
Sections