You are here:
Characterizing Genomic Effects of Formalin Fixation in Archival Samples
Citation:
Wehmas, L., C. Wood, AND S. Hester. Characterizing Genomic Effects of Formalin Fixation in Archival Samples. North Carolina Biotechnology Center, RTP, North Carolina, July 11 - 12, 2018.
Impact/Purpose:
Little is known about potential genomic artifacts caused by formalin fixation, which limits use of archival formalin-fixed paraffin-embedded (FFPE) samples for chemical mode of action studies. This study investigated whether formalin fixation of live tissue samples results in biological or artificial (e.g. nucleic acid damage) gene expression changes and if those changes influenced chemical treatment related gene responses. We found that formalin fixation significantly affected expression of 2946 genes, 98% of which were reduced compared to frozen controls. Freezing samples before fixation resulted in 95% fewer significantly changed genes compared to frozen controls suggesting most of the formalin effect was due to a biological response. This formalin response did not prevent us from detecting highly induced genes known to be affected by the chemical treatment (phenobarbital) compared to vehicle controls within each preservation group. Our results identify gene expression changes from formalin fixation that could impact retrospective mode of action studies with archival FFPE tissue samples.
Description:
Little is known about potential genomic artifacts induced by formalin fixation, which limits use of formalin-fixed paraffin-embedded (FFPE) samples in toxicological and clinical studies. Previously, we identified a transcriptional response in paired frozen and FFPE samples that we hypothesized was due to live tissue being fixed in formalin. To test this idea, livers from B6C3F1 mice treated with 600 ppm of phenobarbital (PB) or vehicle control (Con) for 7 days were divided into three groups:1) fresh-frozen (FR); 2) directly fixed in 10% buffered formalin for 18 hours and processed to FFPE (FIX); and 3) processed as for FIX but initially frozen (FR>FIX) (n=6/group/condition). Ribosomal RNA-depleted, total RNA from these samples was sequenced on an Illumina Hi-seq 2500. Reads were aligned using STAR (2.4) and analyzed in Partek Flow (6.0). FIX vs. FR resulted in 2946 differentially expressed genes (DEGs), 98% of which were down-regulated. FR>FIX vs. FR resulted in 95% fewer DEGs, indicating most of the formalin effect was a transcriptional response from formalin. Ingenuity Pathway Analysis also identified enrichment in mitochondrial dysfunction and transcription elongation pathways associated with formalin exposure. However, the impact of direct fixation on chemical response was less clear. PB treatment induced 180 DEGs within FIX and 159 in FR>FIX of which 120 were shared. The DEGs in each list were consistent with PB exposure, suggesting the formalin signature did not confound chemical response. Our results highlight distinct transcriptional effects of formalin fixation that could impact retrospective mode of action studies with FFPE samples. This abstract does not reflect EPA policy.