The laboratory of Dr. Timothy Yeatman at the H. Lee Moffitt Cancer Center and Research Institute (Tampa, FL) investigates the gene expression “signatures” (mRNA expression patterns) that define colon, breast, and other cancers. The laboratory has access to a large collection of formalin-fixed, paraffin-embedded (FFPE) tumor biopsy samples and detailed information on patient treatment and outcomes. Ultimately the Yeatman laboratory plans to use this valuable resource to correlate historical patient outcomes with the gene expression signature observed in the preserved surgical biopsy samples. The eventual goal is to enable physicians to use the specific cancer signatures from fresh biopsies to predict patient prognoses and to design more efficacious treatment regimens.

Preliminary Target Identification

Gene expression profiles in frozen tissue samples from normal tissue and in primary and metastatic tumors were compared using microarray analysis. Statistical analysis of these data is ongoing, but a preliminary set of several differentially expressed genes was identified for a set of experiments to test and validate a workflow for gene expression analysis of FFPE samples. Susan McCarthy, a Senior Biological Scientist in the Yeatman laboratory, confirmed the differential expression of these genes by reverse transcription (RT) and real-time PCR using RNA isolated from the frozen tumor samples, Applied Biosystems High cDNA Capacity Reverse Transcription Kit, and TaqMan Gene Expression Assays. Next, she evaluated whether these differential gene expression results could be duplicated using patient samples that were processed by FFPE rather than by simple freezing.

Development of a Workflow for Gene Expression Analysis of FFPE Samples

FFPE samples are fixed and processed to preserve tissue structure for histological analysis, but the process impedes subsequent enzymatic reactions by trapping nucleic acids and introducing chemical modifications. In order to begin evaluating samples from the center’s FFPE tissue bank, Ms. McCarthy first compared the performance of three commercially available kits for isolation of RNA from FFPE samples. She found that the protocol for the Ambion RecoverAll Total Nucleic Acid Isolation Kit for FFPE could be completed in a single day. This kit also employed the simplest protocol and yielded high quality RNA more reproducibly than the other kits tested.

With its economical pricing and low input RNA requirements, the High Capacity cDNA Reverse Transcription Kit satisfied the laboratory’s need to synthesize cDNA for real-time PCR. Converting prepared FFPE nucleic acid samples to cDNA is also a reliable strategy for long-term sample storage for future experiments.

The researchers needed sufficient starting material for analysis using a panel of several TaqMan Gene Expression Assays—and in future experiments, when a complete set of differentially expressed genes has been identified, they will be analyzing samples with an even larger set of assays using TaqMan Arrays. The Applied Biosystems PreAmp Master Mix expanded their precious FFPE-derived samples to provide enough material for their ambitious research plan.

The researchers used the Applied Biosystems FFPE gene expression workflow (Figure 1) to compare gene expression in normal tissue with that in primary and metastatic colon tumors. They found that their results matched previously published data and confirmed that accurate gene expression data could be obtained from FFPE samples (Figure 2). Data also indicated that gene expression results from frozen tissue samples correlated with TaqMan Gene Expression Assay results from the same tumor samples that were processed using FFPE (Figure 2).

Figure 1.Gene Expression Workflow for Archival Samples

Figure 2.Gene Expression Results Derived from Frozen Tumor Samples or FFPE Samples are Comparable. Osteopontin is a well-documented biomarker of progression in colon cancer. Panel A shows osteopontin expression levels evaluated using microarray analysis of frozen tissue specimens from normal, Duke's A and Duke's D (metastatic) colon cancer samples. Increased osteopontin expression correlated with tumor progression. Panel B shows quantitative osteopontin expression data obtained from 10 Duke's A and 10 Duke's D FFPE colon cancer specimens using TaqMan Gene Expression Assay. (Relative values were normalized to GAPDH, * P < 0.01 compared with Duke's D.) Likewise, Panel C shows TaqMan Gene Expression Assay data comparing relative osteopontin expression in primary tumor to normal colon derived from triplicate FFPE patient samples (same FFPE patient sample processed 3 times). These results correlated well with the microarray data shown in Panel A. (Data courtesy of S. McCarthy, H. Lee Moffitt Cancer Center and Research Institute.)

Next Steps

With the FFPE sample workflow validated, the laboratory has expanded the study to a larger group of tumor samples and genes. TaqMan Arrays, 384-well micro fluidic cards containing TaqMan Gene Expression Assays, are currently being used to validate microarray signatures from colon, breast, and pancreatic cancers, and translate them into qRT-PCR signatures using RNA derived from frozen and FFPE samples.

Scientific Contributor
Susan McCarthy • H. Lee Moffitt Cancer Center and Research Institute, Tampa, FL