P06: Gene expression profiling in toxicological studies using FFPE tissues

Large numbers of formalin-fixed paraffin-embedded (FFPE) tissues from toxicological studies have been accumulated over the years in the archives. Although RNA extracted from FFPE tissue is significantly degraded, new technologies are allowing their use for gene expression analysis. In this study, the cDNA-mediated annealing, selection, extension and ligation (DASL ™) assay from Illumina Inc. and the WT-Ovation ™ FFPE-System from NuGEN Inc. were used. Both methods, compared to standard technologies, do not need intact mRNA poly-A tails for cDNA synthesis. The impact of formalin fixation time and RNA extraction method on RNA quality and gene expression in both technologies was evaluated. Furthermore, the comparability of gene expression data from FFPE to fresh frozen (FF) tissue was determined. Technically, livers from 3 untreated rats were fixed for different times in formalin. Using RNA extracted with either High Pure RNA Paraffin (Roche) or RNeasy ® FFPE Kits (Qiagen), the DASL ™ assay was performed, and samples hybridized onto a custom DASL-BeadChip (512 liver toxicity genes). A subset of samples was processed using the NuGEN system and hybridized onto Affymetrix Rat 230 2.0 GeneChips ® . Secondly, FFPE livers from rats treated with a former drug candidate, who was positive for hepatotoxicity, and control animals were analyzed using the DASL ™ assay and compared to FF samples (Illumina whole genome rat chip). Gene expression from FFPE tissue with RNA extracted with the High Pure RNA Paraffin kit showed the best correlation to FF samples. Overall, correlation coefficients dropped with increased formalin fixation times in both expression platforms. Longer fixation times resulted in increased alterations of signal intensities, correlating to increased degradation of mRNA. For the second study with treated animals, a dose - and time -dependent clustering was observed from the gene expression data. Focusing on the high dose treated animals it was clearly shown that two animals behaved differently than others, confirming the fresh frozen data as well as the histopathological findings, where these two animals showed the most severe hepatocellular hypertrophy, bile duct hyperplasia and bile duct/hepatocyte necrosis. In conclusion, results of gene expression analysis are heavily dependent on RNA extraction and fixation times. These studies showed that ideally tissues should not be fixed for longer than 1 week to obtain good gene expression data. Nevertheless, both techniques are appropriate methods for gene expression analysis when working with degraded RNA.