Amplification of mRNA from Formalin-Fixed Paraffin-Embedded (FFPE) Breast Cancer Tissues without 3' Bias and Multiplex Expression Analysis on Flow-Through Microarrays.

Abstract

Abstract Background: FFPE samples are an abundant sample resource for validation of expression-based assays and are will suited for routine clinical testing. Analysis on microarrays enables the analysis of many transcripts from small samples. However, mRNA in FFPE sections is degraded and cross-linked, with RNA fragment sizes of a few hundred bases or less, and consequently standard methods of microarray sample preparation are not reliable unless probes are very strongly biased to the 3' end. It is not always possible to design high quality probes within this restricted region of mRNAs. Amplification of target sequences any place within transcripts would permit greater flexibility in probe design and optimization of probe sets for reliable multiplex analysis.Material and Methods: Total RNA extracted from FFPE samples was subjected to two rounds of amplification with the ExpressArt TRinucleotide kit (AmpTec), which preferentially primes cDNA synthesis near the 3'-ends of mRNAs (rather than at poly-A sites) and selects for mRNAs against ribosomal RNAs. Biotin-labeled anti-sense RNA (aRNA) was prepared from the second round cDNA with the Illumina® TotalPrep™ RNA kit (Applied Biosystems). The labeled aRNA was hybridized to probes immobilized in singlet on flow-through microarrays on the Ziplex Automated Workstation (Xceed Molecular). Performance was assessed by the analysis of titration mixtures prepared from mRNA from breast cancer and colon cancer FFPE tissue blocks.Results: Median CVs of technical replicates were less than 25%. Linear responses were observed in the analysis of the titration mixtures, and all correlations between predicted and observed expression values of the titration mixtures had R2 values greater than 0.8. The data indicated that the minimum detectable expression difference was less than two-fold. Similar signal intensities were observed for 3'-biased probes and probes several hundred bases from the 3'end, confirming the lack of 3' bias in the amplification.Discussion: The results demonstrate the feasibility of amplifying and quantifying sequences at any position within transcripts in degraded mRNA from FFPE samples. Expression differences of less than two-fold may be analyzed with tens or hundreds of probes for translational research and clinical assay development on the Ziplex Workstation. Many probes may be tested in parallel to optimize probe sets for specific tests. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3052.