New technologies for FFPE samples: Improved RNA isolation and novel cDNA priming for RT-qPCR and for microarray analysis

Abstract
Archival FFPE samples harbour a wealth of information. Although FFPE RNA is severely degraded plus challenges due to inter- and intramolecular cross-linking and base modifications, mining of gene expression data is still possible and extracted information about differential gene expression is comparable to data from frozen samples, even at quantitative level. The combination of a novel procedure for RNA liberation and demodification results in highly reproducible data in RT-qPCR studies (Oberli et al 2008) and can be used to determine RNA profiles of cancer samples (Schobesberger et al 2008). Our novel TR priming combines the advantages of oligo-dT (priming near the 3’end and selection against rRNAs) with random priming (works with mRNA fragments without poly-A: no requirement for a universal target sequence). Oligo-dT primers provide cDNA from regions next to the 3‘-poly-A tail. Internal transcript regions are (partially) lost and effects on different mRNAs and PCR amplicons are variable. Accordingly, gene expression results (RT-qPCR) vary widely with different RNA qualities. For example, upon RNA degradation, the Ct for actin mRNA increases 9.9 cycles, with much lower effects on other mRNAs, e.g. an increase of only 3.9 for YMHAZ. Although there is good selectivity against rRNA sequences in intact RNA: Ct for 18S rRNA is only -4.1 vs actin mRNA (17-fold more rRNA). Since rRNA is more stable, the difference increases to 9.0 in degraded RNA (500-fold more rRNA). With random primers (or a mix of random and oligo-dT), the recovery of internal transcript regions is improved. Accordingly, there is less variability with different RNA qualities: Ct for beta-actin increases by 5.4 cycles, comparable to 4.7 for YMHAZ. Selectivity against rRNA sequences is poor. Intact RNA: Ct for 18S rRNA vs actin is -9.2 (588-fold more rRNA), with further increase to -10.7 (1663-fold more rRNA). With our novel TR primers, preferential priming occurs at the 3‘-end of RNA fragments (independent from poly-A) and different to random priming, further “cutting“ of RNA fragments into multiple, short cDNAs is prevented; combined with improved recovery of internal transcript regions. Ct for beta-actin increases by 4.1, very comparable to 4.4 for YMHAZ. Furthermore, selectivity against rRNA sequences is good. Intact RNA: Ct for 18S rRNA vs actin is -7.5 (181-fold more rRNA), a small increase to -8.6 (388-fold more rRNA, lower than for oligo-dT with this RNA sample). TR priming can be used for FFPE RNA templates, followed by qPCR analyses or combined with mRNA amplification and labelling for genome-wide microarray gene expression profiling. Selectivity against rRNA sequences makes TR primers an ideal tool for the novel “Exon Microarrays“. Presented data will demonstrate advanced recovery of FFPE RNAs: RNA profiles (Agilent Bioanalyzer), comparison of expression profiles obtained with RNAs from fresh-frozen and from FFPE samples, as determined by RT-qPCR and by Affymetrix microarrays.