Small Nucleolar RNA Score: An Assay to Detect Formalin-Overfixed Tissue.
Author(s): Ammerlaan W, Trouet J, Sachs MC, Guan P, Carithers L, Lambert P, Frasquilho S, Antunes L, Kofanova O, Rohrer D, Valley DR, Blanski A, Jewell S, Moore H, Betsou F
Publication: Biopreserv Biobank, 2018, Vol. , Page
PubMed ID: 30234371 PubMed Review Paper? No
Purpose of Paper
The purpose of this paper was to evaluate effects of the duration of formalin fixation on the quantification of more than 950 individual microRNAs (miRNAs) via the qRT-PCR-based SmartChip platform. The authors also sought to identify small RNA targets that may serve as markers of overfixation in formalin-fixed, paraffin-embedded tissue specimens.
Conclusion of Paper
In principal component analysis (PCA) frozen specimens clustered separately from FFPE specimens, and specimens fixed for 72 h clustered separately from those fixed for 6, 12, and 23 h. Differences in expression of individual miRNA between frozen and FFPE specimens were miRNA specific. Differences among FFPE specimens fixed for different durations were minimal in ovary specimens but slightly more prevalent in kidney specimens. Nonetheless, the authors identified four miRNA targets sensitive to overfixation (≥72 h) and developed a scoring system that was then used to identify overfixed specimens with a specificity of ≥85% and a sensitivity of ≥88%. Use of a similar metric on a customized WaferGen miRNA chip allowing for analysis of 40 specimens per run confirmed the original trends observed but resulted in a lower specificity and sensitivity ((≥79% and (≥80%, respectively) than the standard WaferGen miRNA chip.
Studies
-
Study Purpose
The purpose of this study was to evaluate effects of the duration of formalin fixation on the quantification of more than 950 miRNA targets represented on the qRT-PCR-based WaferGen SmartChip assay. Tumor specimens were collected under the National Cancer Institute's Biospecimen Preanalytical Variable (BPV) Program from four different clinical locations and sent to a central location for processing and analysis. The ten kidney and nine ovarian tumor specimens used in this study were collected via surgical resection, aliquoted and snap-frozen in liquid nitrogen or fixed for 6, 12, 23, or 72 h in 10% neutral buffered formalin. Formalin-fixed specimens were then processed for 10-12 h prior to paraffin embedding, and were stored as FFPE tissue bocks at room temperature until sectioning and analysis. For validation of miRNA targets used to discriminate formalin-fixation times, a second set of nineteen tissue specimens (six colon, one kidney, one ovary, two uterus, one rectum, six breast, and two ileum tissue specimens) were collected by the Integrated Biobank of Luxembourg (IBBL), formalin-fixed and paraffin-embedded and stored as blocks at -20°C. RNA was extracted from all specimens with the miRNeasy kit from Qiagen in conjunction with a QIAcube platform. RNA was then quantified by spectrophotometry, and evaluated by RNA integrity number (RIN) using a bioanalyzer. PolyA-tailed RNA was reverse transcribed with miRNA RT primers and analyzed by qPCR using the standard WaferGen SmartChip Human miRNA Panel v3.0 and a custom WaferGen miRNA chip containing 24 miRNA and snoRNA and the WaferGen qPCR v2.5 software. RT-qPCR data underwent global normalization and was subjected to the following quality control criteria: global normalization factor =0.3-3.33; Reference Targets Stability with mean GNorm M < 1.0 and coefficient of variation <0.6. Negative expression was set to be a CNRQ value of 0.050 to decrease overemphasis. Expression patterns of miRNAs were examined among FFPE specimens by principal component analysis (PCA using linear mixed effects models with random intercepts.
Summary of Findings:
Mean RINs were lower in FFPE ovary and kidney specimens compared to corresponding frozen controls (ovary: 2.3 versus 7.8; kidney: 2.8 versus 6.7, respectively). Principal component analysis of data generated with the miRNA WaferGen SmartChip revealed that frozen specimens clustered separately from FFPE specimens, and specimens fixed for 72 h clustered separately from those fixed for 6, 12, and 23 h. The fold difference in FFPE specimens compared to frozen specimens was RNA target-specific but ranged between -4.51 and 3.28. Interestingly, there were more differences in miRNA expression between fixation times in kidney than ovary specimens. Based on CNRQ variability (the percent coefficient of variation), four targets (SCARNA5, SNORA16A, SNORA44, SNORA61) were determined to be the most unstable among overfixed specimens (in formalin for 72 h) based on data generated with the standard WaferGen SmartChip. The authors calculated snoRNA CNRQ score, which is the sum of the logCNRQ values for the four unstable miRNAs identified. The snoRNA CNRQ score was able to discriminate tissue specimens that had been fixed for 72h or longer in both the BPV specimen set (sensitivity=85%, sensitivity= 88%) and the IBBL specimen set (sensitivity=88%, sensitivity= 91%). The authors then investigated the specificity and sensitivity of a customized WaferGen miRNA assay for the identification of overfixed specimens. The customized WaferGen miRNA chip included 8 miRNA and snoRNA targets allowing for more specimens per run. The snoRNA score (sum of CT values for SNORD99, SNORA61, SNORA44, SNORA16A, and SCARNA5 minus the sum of CT values for miR-720, miR-1260, and Let-7A) calculated using the customized array had a lower specificity and sensitivity than the standard WaferGen SmartChip for the BPV specimen set (sensitivity=83%, sensitivity= 82%), the IBBL specimen set (sensitivity=83%, sensitivity= 84%), and the two specimen sets combined (sensitivity=79%, sensitivity= 80%). Although not specifically presented as results, the authors note that overfixation more greatly impacted snoRNAs than miRNAs which were minimally affected by overfixation. The authors report that using TaqMan assays they confirmed the trend toward increased snoRNA CT values with increased fixation duration, but data was not shown.
Biospecimens
Preservative Types
- Frozen
- Formalin
Diagnoses:
- Neoplastic - Carcinoma
Platform:
Analyte Technology Platform RNA Real-time qRT-PCR RNA Automated electrophoresis/Bioanalyzer RNA Low density array Pre-analytical Factors:
Classification Pre-analytical Factor Value(s) Biospecimen Preservation Time in fixative 6 h
12 h
23 h
72 h
Low density array Specific Type of array WaferGen miRNA Chip
Customized miRNA Chip
Biospecimen Preservation Type of fixation/preservation Formalin (buffered)
Snap frozen