Laser capture microdissection (LCM) permits isolation of pure cell populations from which RNA can be extracted, amplified, and subjected to microarray analysis, allowing information to be obtained on the gene expression profile of defined cell types. To avoid amplification artifacts and detect genes expressed at different levels, it is important to optimize the choice of both RNA amplification step and microarray platform. We captured by LCM the same colon cancer biopsy and conducted a cross comparison of distinct RNA amplification methods and different chip platforms. We tested two RNA amplification methods with different chemistry: the one-cycle Ovation system (NuGEN) and the two-cycle Ribo OA method (Arcturus). We also compared two different whole genome platforms, based on Affymetrix technology: the U133 plus 2.0 and the X3P array, with probe sets closer to the 3' end of transcripts. After RNA amplification, microarray analysis, and data normalization, we investigated reproducibility and correlation of different methods and arrays. Our results indicate that the Arcturus Ribo OA method is superior for both array choices, especially in combination with X3P arrays, showing the lowest variance and Spearman correlation of 0.986. The quicker NuGEN procedure, when coupled with X3P arrays, also yielded excellent results (correlation of 0.951). These observations will be useful for planning large-scale analyses of LCM-dissected clinical samples.
(c) 2007 Wiley-Liss, Inc.