Erin Newburn, MS, PhD
Senior Manager, Field Applications Scientist

Can gene and variant expression from RNASeq be done well using FFPE specimens?

As immunotherapy translational research studies using RNA data provide new clues to patient response, identifying the upfront technical factors affecting success is vital.

For most oncological studies, the sample format available is formalin-fixed paraffin-embedded (FFPE) samples.  While formalin can greatly preserve precious tissue (potentially for decades!), it has historically been challenging for DNA, but especially RNA studies (Hedegaard et al 2014).  These perils are well known as the chemical modification and crosslinking process can damage the integrity of nucleic acids.  Additionally, methods such as poly-A selection and ribodepletion for mRNA enrichment, suffer limitations including the inability to work with degraded material or inconsistent removal of ribosomal RNA.

A robust, reliable method for analyzing gene expression in FFPE

Developed by the scientists at Personalis, the ACE Cancer Transcriptome™ assay uses targeted capture methodology in order to overcome the limitations mentioned above.  In this approach, multiple probes target each transcript so that the assay captures transcripts even when the poly-A tail is lost due to RNA degradation, making it amenable to working with challenging starting materials, like FFPE.

Examining sequencing QC metrics such as 1) the percentage of ribosomal RNA free reads and 2) the percentage of reads that map to the exonic region can show initial clues when assessing RNA data quality from FFPE. To further understand the quality of gene expression profiles from FFPE tissue, the team acquired paired FFPE and matched adjacent Fresh Frozen (FF) tumor biopsy samples.  These tissues were from colon, lung and rectal cancer patients.  The availability of matched adjacent FF tissue provided a comparator for “good” gene expression results.  We performed our ACE Cancer Transcriptome assay on the tumor samples and compared gene expression values (Log2 Transcripts Per Million (TPM)) among matched pairs.  High correlation (R2 values 0.88–0.95) was found between FF and FFPE Log2 normalized gene expression values (TPM) across all tissue pairs.  These results illustrate that an optimized transcriptome approach can provide results from FFPE concordant with and “on par” of those obtained from FF specimen.

Do’s and Don’ts for FFPE Fixation

 Storing and collecting cancer samples as FFPE specimens can significantly increase the breadth of materials for retrospective molecular characterization.  As mentioned above, assessing the quality of both gene and variant level expression data from FFPE specimens is necessary for determining the accuracy of the downstream analytics.   While in the planning stages for fixing specimens, next generation sequencing laboratories can provide best practice guidelines in preserving specimens that are designated for future use in genomic and transcriptomic studies.

Fixation behaviors that can alter DNA and RNA quality include:

  • Duration of fixation. Fixing samples for >24 hours can be detrimental
  • Fixation Temperature. Fixation should occur at room temperature, as elevated temperatures are not recommended
  • Fresh reagents. Beware of expiration dates and prepare fresh buffer at the start of a study.
  • Proper storage of tissue. Samples should be fixed immediately, or snap frozen until fixed.

In general, at Personalis we recommend freshly prepared 10% formalin in PBS solution and fixing overnight at room temperature.  This procedure can optimize successful isolation of both high quality DNA and RNA from these precious patient samples.

Having accurate transcriptomic data will accelerate molecular characterization efforts in critical areas of cancer research.  For immuno-oncology, these data will specifically fuel a greater understanding of tumor progression, the immune response, and treatment resistance in therapeutic development of checkpoint blockade.  The recent advances in sample extraction and NGS library preparation for RNA are already helping to facilitate the discovery of transcriptomic signatures to guide efficacious immunotherapy treatment, limit patient exposure to adverse events, and understand the fundamentals of resistance.  Thus, comprehensive tumor profiling through RNASeq data will allow for far-reaching applications for immuno-oncology.