Articles

Contact | Print version | Recommend article | RSS-Feed

miRNA Extraction and Profiling on FFPE Samples

Poul Erik Høiby1 and Manfred Watzele2*
1Exiqon, Vedbaek, Denmark
2Roche Applied Science, Penzberg, Germany

*Corresponding author

Introduction

RNA silencing is a mechanism of gene regulation whereby small double-stranded RNAs (dsRNAs) of 21–18 nucleotides in length cause mRNA degradation, mRNA translation, or chromatin modification [1]. In the past years, more than 500 novel genes have been identified in animals and plants encoding small transcripts that can form small dsRNA hairpin structures [2–4]. From these hairpin precursors, small 22-nt RNAs – so called microRNAs (miRNA) – are excised by dsRNA-specific endonucleases [5]. The miRNAs are then integrated in ribonucleoprotein particles (RNPs) [6]. Most miRNAs in animals act by binding to partially complementary regions of the 3' untranslated region of their target mRNA, thereby blocking translation [7]. The expression of individual miRNA species was shown to be cell line-specific and directs tissue differentiation in Caenorhabditis elegans [8]. Some proliferative diseases such as B-cell chronic lymphocytic leukemia and colorectal neoplasia were caused by deregulated expression of certain miRNA species [9, 10]. In addition, several physiological processes such as proliferation, axon guidance and morphogenesis seem to be tightly controlled by miRNAs.

To further elucidate miRNA function, an analysis of miRNA expression in specific tissues is necessary. In order to obtain large amounts of high-quality miRNA samples, powerful extraction procedures are needed. Here, we describe the isolation of miRNAs from different formalin-fixed paraffin-embedded (FFPE) tissue and fresh-frozen samples using the HighPure miRNA Isolation Kit.

Materials and Methods

Three commercially available kits fom suppliers A, Q, and Roche Applied Science were used to isolate total RNA from FFPE sections of eight different organs from mice. Half of the organs were fresh-frozen and the other half formalin-fixed and paraffin-embedded. Fresh-frozen samples were extracted with Trizol® reagent for comparison. The same amount of FFPE tissue was used for all kits.

The kits from suppliers A and Q were used as described in the pack insert. For the HighPure miRNA Isolation Kit we used the following procedure: Eight pieces of 10-µm tissue sections were first deparaffinized by vortexing with xylene and washing with ethanol. The tissue was then dissolved in paraffin tissue lysis buffer and ­treated with proteinase K for 3 hours at 55°C. After mixing with binding buffer and binding enhancer, nucleic acids were bound to the spin column and eluted with elution buffer.

Isolated total RNA from each preparation was quantified on a Nanodrop instrument. Then, 1 µg of the total RNA from the FFPE samples was labeled with Hy3 reagent. Total RNA from each fresh-frozen tissue was labeled with Hy5 reagent. miRCURY™ miRNA profiling of each FFPE sample was performed against the fresh-frozen total RNA sample.

Results

Analysis on the Nanodrop instrument showed that a very high amount of total RNA was extracted from almost all the FFPE-treated tissues using the HighPure miRNA Isolation Kit (Figure 1).

By miRCURY™ miRNA profiling of each FFPE sample against fresh-frozen total RNA samples, individual miRNAs­ were identified. Again, using the HighPure miRNA Isolation Kit, a very high number of miRNA ­species could be identified in almost all of the FFPE samples (Figure 2).

miRCURY™ miRNA profiling also yielded high signals for samples treated with the HighPure miRNA Isolation Kit demonstrating the kit’ s exceptional performance (Figure 3).

When comparing the expression levels of the individual miRNAs in FFPE-treated tissue with their levels in fresh-frozen tissues, very high correlation values were obtained with the HighPure miRNA Isolation Kit, whereas supplier A had lower correlation values (Table 1 with 1.0 being 100% correlation).

In the miRNA microarray plot profile, most miRNA levels in FFPE samples are in a range of +/- 2 of the levels in fresh-frozen tissues (data not shown). Outliers are listed in Table 2. As the data show, using the HighPure miRNA Isolation Kit, only very few miRNA levels of the FFPE samples differ from those of the fresh-frozen tissues.

It was further investigated whether the miRNAs with higher abundancy in FFPE samples had specific features that discriminate them from others. Surprisingly, all abnormally­ represented miRNAs detected with the HighPure miRNA Isolation Kit were also detected with the kit from supplier Q (data not shown).

This, however, seemed not to be the case. Comparison of 6 randomly selected probes (Table 3) showed that ­neither the GC content nor the length of the probes was extraordinary. The number of predicted targets was also in the normal range, as was the number of BLAST hits against genome and transcriptome. In addition, a dye-swap experiment was performed and showed that dye-bias was not the cause of abnormal representation.

Summary

The new HighPure miRNA Isolation Kit is a complete kit that contains all reagents necessary for the isolation of miRNA from FFPE tissue (except xylene). Using this kit, the largest amount of total RNA and the highest detection signals were obtained for nearly all tissues tested. In addition, the highest number of miRNA species were detectable with the HighPure miRNA Isolation Kit. Correlations of results for the matching FFPE and fresh-frozen samples were excellent. The repeated up-regulated miRNAs (outliers) cannot be explained by extraordinary probe sequences or by dye-bias.

References

1.   Carrington JC, Ambros V (2003) Science 301:336–338

2.   Lagos-Quintana M et al. (2001) Science 294:853–858

3.   Lagos-Quintana M et al. (2003) RNA 9:175–179

4.   Palatnik JF et al. (2003) Nature 425:257–263

5.   Hutvágner G et al. (2001) Science 293:834–838

6.   Hutvágner G, Zamore PD (2002) Science 297:2056–2060

7.   Reinhart BJ (2000) Nature 403:901–906

8.   Pasquinelli AE, Ruvkun G (2002) Annu Rev Cell Dev Biol 18:495–513

9.   Calin GA et al. (2002) Proc Natl Acad Sci U.S.A. 99:15524–15529

10. Michael MZ et al. (2003) Mol Cancer Res 1:882–891

This article was originally published in Biochemica 3/2008, pages 23-25. ©Springer Medizin Verlag 2008