Reliable Gene Expression Profiling from Small and Hematoxylin and Eosin-Stained Clinical Formalin-Fixed, Paraffin-Embedded Specimens Using the HTG EdgeSeq Platform.
Author(s): Qi Z, Wang L, Desai K, Cogswell J, Stern M, Lawson B, Kerkar SP, Vitazka P
Publication: J Mol Diagn, 2019, Vol. 21, Page 796-807
PubMed ID: 31255795 PubMed Review Paper? No
Purpose of Paper
This paper investigated the effects of prior H&E and immunohistochemical (IHC) staining on gene expression analysis in formalin-fixed paraffin-embedded (FFPE) non-small cell lung carcinoma (NSCLC) and small cell lung carcinoma (SCLC) specimens and investigated if expression patterns were correlated with histological results. Gene expression in matched frozen and FFPE colorectal specimens was also compared. The effects of bead-to-DNA ratio on dimer removal and effects of DNA input amount on gene expression were also investigated.
Conclusion of Paper
Gene expression in unstained sections was strongly to very strongly correlated with that in matched H&E sections but only weakly to modestly correlated with that in IHC-stained slides. A closer look at individual genes showed similar expression of programmed death-ligand 1 (PD-L1) or programmed death-ligand 2 (PD-L2) and the immunotherapy target programmed-death 1 (PD-1) in unstained and H&E stained specimens but much larger differences in expression patterns in unstained and IHC-stained sections. Interferon gamma (IFN-γ) signature scores were very strongly correlated between unstained and H&E-stained section from 42 NSCLC specimens and were also modestly correlated with the immune scores. The correlation in gene expression between RNA from frozen specimens and that from FFPE colorectal specimens was strong to very strong (r=0.79-0.90) similar to the correlation between unstained and H&E stained specimen.
Adapter dimer removal was best when the bead-to-DNA ratio was ≤1.5 but a ratio of ≤1.3 led to a lower library yield. Library yield increased with increasing DNA input and alignment rates were >69% when the equivalent of 0.31-5 mm2 DNA was used for input. Gene expression in libraries generated with input amounts between 0.31-20 mm2 were strongly to very strongly correlated (r>0.85) but correlations decreased slightly and alignment rates dropped when the equivalent of <0.15 mm2 was used.
Studies
-
Study Purpose
This study investigated the effects of prior H&E and IHC staining on gene expression analysis in FFPE NSCLC and SLC specimens and investigated if expression patterns were correlated with histological results. Gene expression in RNA from colorectal frozen specimens and matched FFPE specimens was also compared. Serial sections (5 µm) from FFPE blocks from 45 surgical NSCLC specimens and three SCLC were used for H&E and PD-L1 staining but each experiment used only a few specimens. IHC staining included deparaffinizaton, antigen retrieval, peroxidase blocking, anti-PD-L1 labelling and visualization steps, but details were not specified. Prior to sequencing, H&E and IHC-stained slides were submerged in xylene for 2 days, tissue was scraped into microfuge tubes, incubated at 95°C in lysis buffer (demodification), incubated with proteinase K at 50°C for 2 h, and placed in a 96 well plate. Gene expression profiles were generated using the EdgeSeq system which combines a nuclease protection assay and next -generation sequencing. Nuclease protection probes for 2560 genes were added to the lysates and then S1 nuclease was added to degrade non-targeted RNA and non-hybridized probes. Hybridized RNA and probes were purified using the Agencourt AMPure XP and quantified using the KAPA Library Quantification Kit. Libraries were sequenced on an Illumina MiSeq. RNA from nine frozen colorectal adenocarcinoma specimens and RNA extracted from matched FFPE blocks using the AllPrep DNA/RNA FFPE Kit was sequenced.
Summary of Findings:
Gene expression for three SCLC and three NSCLC in unstained slides was strongly to very strongly correlated with that in matched H&E sections (r=0.86-0.97) but only weakly to modestly correlated with that in IHC-stained slides (r=0.21-0.56). When the correlation in gene expression between unstained slides and H&E-stained slides for all 48 specimens was examined, correlations of > 0.85 were observed for 90% of pairs (43 of 48). The correlation in gene expression between RNA from frozen specimens and that from FFPE colorectal specimens was strong to very strong (r=0.79-0.90), similar to the correlation between unstained and H&E-stained specimen. Further, analysis of the expression of PD-L1, PD-L2, and PD-1 in the three SCLC and three NSCLC showed that the maximum difference between unstained and H&E-stained was 1.4 log2 counts per million (CPM) but 9.6 log2 CPM for unstained and IHC-stained specimens. Expression of the 25 IFN-γ was weakly to very strongly correlated in the 42 pairs of stained and unstained specimens (r=0.29-0.92) with weak correlations observed for genes with low expression. IFN-γ signature scores were very strongly correlated between unstained and H&E-stained sections from 42 NSCLC specimens and were also modestly significantly correlated with the immune scores (r=0.45, P=0.0011 for unstained and r=0.48, P=0.0025 for H&E-stained).
Biospecimens
Preservative Types
- Frozen
- Formalin
Diagnoses:
- Neoplastic - Carcinoma
Platform:
Analyte Technology Platform RNA Next generation sequencing Pre-analytical Factors:
Classification Pre-analytical Factor Value(s) Biospecimen Preservation Type of fixation/preservation Formalin (buffered)
Frozen
Next generation sequencing Specific Type of tissue stain Unstained section
H&E stained section
IHC stained section
-
Study Purpose
The purpose of this study was to optimize the bead-to-DNA ratio and the DNA input amount for the EdgeSeq analysis. Serial sections (5 µm) from FFPE SCLC blocks (number unspecified but <3) were incubated at 95°C in lysis buffer (demodification) and incubated with proteinase K at 50°C for 2 h. Lysate was diluted such that the wells had an equivalent to 20, 10, 5, 2.5, 2.0 1.25, 0.625, 0.31, 0.15, and 0.08 mm2. Nuclease protection probes for 2560 genes were added to the lysates and then S1 nuclease was added to degrade non-targeted RNA and non-hybridized probes. Hybridized RNA and probes were purified using the different amounts of beads in the Agencourt AMPure XP (1.0, 1.5, 2.0, and 2.5 bead-to-DNA ratio). Libraries were quantified using the KAPA Library Quantification Kit. Libraries were sequenced on an Illumina MiSeq. RNA from nine frozen colorectal adenocarcinoma specimens and RNA from matched FFPE blocks extracted using the AllPrep DNA/RNA FFPE Kit was sequenced.
Summary of Findings:
Adapter dimer removal from the library was best when the bead-to-DNA ratio was ≤1.5 but a ratio of 1.3 or 1.0 led to a higher cycle threshold (CT) value (lower yield) for the library. Library yield increased with increasing DNA input. Gene expression in libraries generated with input amounts between 0.31-20 mm2 were strongly to very strongly correlated (r>0.85) but correlations decreased slightly when the equivalent of <0.31mm2 was used (r=0.77-0.96). While a good alignment rate (>69%) was observed when the equivalent of 0.31-5 mm2 DNA was used for input, the alignment rates were <40% when the equivalent of <0.15 mm2 was used.
Biospecimens
Preservative Types
- Formalin
Diagnoses:
- Neoplastic - Carcinoma
Platform:
Analyte Technology Platform RNA Next generation sequencing Pre-analytical Factors:
Classification Pre-analytical Factor Value(s) Analyte Extraction and Purification Analyte purification Bead to DNA ratio of 1.0
Bead to DNA ratio of 1.3
Bead to DNA ratio of 1.5
Bead to DNA ratio of 1.8
Bead to DNA ratio of 2.0
Bead to DNA ratio of 2.5
Next generation sequencing Specific Template/input amount 20 mm2
10 mm2
5 mm2
2.5 mm2
1.25 mm2
0.625 mm2
0.31 mm2
0.15 mm2
0.08 mm2