Methods to analyze MicroRNAs

Tissue, RNA preparation and storage

The storage and preparation of miRNAs from samples are crucial for microarray gene expression analysis. Recently, it has been shown that miRNAs within tissues can be kept in a satisfactory state for prolonged periods using formalin. Xi et al. (2007) found that the expression of miRNAs from formalin-fixed paraffin-embedded (FFPE) specimens was in good correlation with fresh frozen samples, raising the possibility of archiving fixed clinical specimens for further miRNA analysis at a later date. The first step in obtaining miRNA samples is the isolation of RNA from different cells and tissues. Many different in-house protocols and commercial kits can be used. At GPCF we have gathered experience with acidic phenol methods (home-made and commercial (TRIzol, Invitrogen) as well as using silica-based membrane-spin column (miRNeasy, Qiagen). These methods can provide high quality RNA for further investigations.

Although the relative abundance of small RNAs in a total RNA preparation is typically in the range of 0.01-0.05%, it is important that all RNA molecule types are isolated in a single step – no small RNA enrichment should be done! The sensitivity of miRNA detection is ensured by applying an amplification step before hybridization.

MicroRNA labeling and data analysis

The analysis of mature miRNA expression is key to understanding its physiological functions and pathological implications. In 2003, Krichevsky et al. were the first to design an oligonucleotide array that could detect miRNAs in mammalian brain tissues by labeling low molecular weight RNA with radioactive isotopes. Other labeling and hybridization technologies followed quickly, see table below.

Method	Pros	Cons	References

miRNA arrays probe design	Low cost; high throughput	Radioactive isotopes	Krichevsky et al., 2003
Nucleotide modification	Tm balance	Needs special nucleotide	Castoldi et al., 2006
RNA length trim	Tm balance, sensitivity (0.2 amol–2 fmol)	Needs labeled donors and RNA ligase	Wang et al., 2007
miChip labeling	Tm balance	Needs LNA	Beuvink et al., 2007
Indirect	Potential for amplification	Random primers, precursor contamination	Liu et al., 2004
RAKE approach	High specificity	Lower sensitivity	Nelson et al., 2004
Cisplatin-based	Chemical label	Hybridization interference	Babak et al., 2004
PCR and Cy3-label	Sample amplification, sensitivity	Multiple steps, high background	Miska et al., 2004
Electrocatalytic label	Sensitivity (0.2 pM–400 pM)	Low-throughput	Gao et al., 2007
Quantum dots	Sensitivity (156 pM–20 nM)		Liang et al., 2005
T4 RNA ligase	Simple	ATP interference	Thompson et al., 2005
Poly(A) polymerase	<3 fmol	Various limitations in labeling efficiency	Shingara et al., 2005



Other technologies

Northern blot	The most reliable technology – the ‘Gold standard’	Radioactive isotopes, low throughput, low sensitivity	Valoczi et al., 2004
Cloning	Discovery of new miRNAs	High cost for sequencing	Lau et al., 2001
Real-time PCR	Rapid detection of miRNAs and precursors	High cost	Schmittgen et al., 2004, Chen et al., 2005
In situ detection	Ability to detect miRNA levels in tissue	Low throughput	Kloosterman et al., 2006
Bead-based	Multiple sample test, high speed and low cost	Specific conditions	Lu et al., 2005
Single molecule detection	Sensitivity (500 fmol)	High cost, special instrument	Neely et al., 2006