Good things come in small packages

The biological insight held by microRNA promises much – but how do you do about examining the subtleties of these discrete molecules? Sangita Parikh has some answers  

STUDY of microRNA (miRNA) expression profiling is accelerating as scientists investigate the roles of these small, non-coding RNAs in cancer, neurodegenerative disease, developmental biology and other disciplines. miRNAs are approximately 19-30 nucleotides long and serve widespread functions as translation regulators. MicroRNA genes are generally transcribed as long RNA precursors (pri-miRNAs), which are then processed to the shorter hairpin pre-miRNAs (around 70–90 nucleotides), before they are cleaved to form the mature single-stranded miRNAs of around 22 nucleotides (Figure 1).

While these areas of study hold great promise, this class of molecules also pose unique challenges in measurement. The small size of the miRNAs makes them difficult to label with high yields and without sequence bias. Sequences frequently differ by just one nucleotide, requiring highly-specific measurement techniques. New miRNAs are also being discovered continually, so measurement tools must be able to easily acommodate these growing databases. Since less then 1% of total RNAs are miRNAs, most measurement techniques require microgram quantities of total RNA. The samples are usually size-fractionated and amplified for miRNA measurement. Both processes can result in bias in the resulting miRNA expression profile. Further, miRNAs are expressed in large dynamic range, usually over four orders of magnitude. Thus measurement platforms with limited linear dynamic range would compress and distort the measured miRNA expression profile.

A number of approaches can be used to measure miRNA expression, including hybridisation-based techniques such as northern blots, quantitative PCR and microarrays. Due to their high-throughput capabilities, microarrays have become very popular for miRNA expression profile analysis.

This article describes how research at Agilent Laboratories has developed a highly-parallel platform that overcomes the challenges presented by miRNA expression profiling measurement using microarrays. This work produced key technological advances required for the development of a highly-specific, robust microarray-based method for profiling miRNA expression.

Quantitative direct labelling methods that minimise sample manipulations, such as size separation or amplification, are most likely to produce accurate measurement of miRNA profiles. This was an important design criterion for the Agilent miRNA microarray platform. It was achieved using a labelling technology developed by Agilent Laboratories (Wang, Ach and Curry, 2007). For the first time, miRNAs in total RNA could be labelled without amplification or fractionation, producing precise and accurate measurements spanning a linear dynamic range from 0.2amol to 2fmol of input RNA, or five orders of magnitude. The protocol requires just 100ng of total RNA, and the experimental protocol is simple because the sample can be dephosphorylated and labeled in the same tube.

Since the discovery of miRNAs in 1993, the number of discovered miRNAs has rapidly increased. In the Sanger miRNA database, the number of miRNA precursors in animals, plants and viruses has increased from over 200 in 2002 to over 6000 in 2008. To date, more than 800 mature miRNAs have been identified in human. Furthermore, it is thought that as many as one third of all mammalian genes may be miRNA-regulated.

The small size and high sequence homology of miRNAs represents a unique

Figure 1: The mechanism of miRNA transcription, processing, and regulatory activity. miRNA genes are transcribed by RNA polymerase II to form primary miRNA (pri-miRNA) molecules. The ribonuclease, Drosha, then cleaves the pri-miRNA to release the pre-miRNA for cytoplasmic export and processing by Dicer. The mature miRNA product associates with the RNA-induced silencing complex for loading onto the 3' UTR of target mRNAs to mediate translational repression.

Figure 2: Components of the Agilent miRNA microarray probe design. An unmodified microarray probe (black) is a synthesised sequence that hybridises to the target miRNA (red). Probes are anchored to the glass slide surface by a stilt (brown). A. Inclusion of a G residue (black) to the 5' end of the hybridisation sequence complements the 3' end C residue (yellow) introduced in labelling. This additional G-C pair in the probe-target interaction region stabilises targeted miRNAs relative to homologous RNAs. Additionally, all probes contain a 5' hairpin (blue), abutting the probe-target region, to increase target and size miRNA specificity.                        B. Destabilisation of probes that are too stable. For probes requiring it, reduction of probe-target base-pairing is achieved through sequential elimination of base pairing from the 5' end of the miRNA.

challenge for hybridisation-based detection methods. Unlike genomic or messenger RNA targets where sequences with equal melting temperatures (Tm) can be selected from a long sequence segment, miRNAs require a different probe design strategy.  We were able to design efficient miRNA probes using only unmodified DNA oligonucleotides by using several novel design features. First, a G is added to the 5’ end of each probe sequence to complement the 3’ cytosine introduced in labeling (Figure 2A). This added G-C pair stabilises the targeted miRNA relative to homologous RNAs. In situations where the probe-target duplex is too stable (potentially resulting in non-specific interactions) the hybridisation is optimised through reduction from the 5' end of the miRNA (Figure 2B). Since most of the sequence homology occurs in the 5’ end of miRNAs (Griffiths-Jones 2004; Griffiths-Jones et.al. 2006), sequentially eliminating base pairing from the 5’ end has little effect on probe sequence specificity. To help distinguish the targeted miRNA from unintended potential targets, a hairpin structure is incorporated onto the 5’ end of the probe, directly abutting the 3’ end of the hybridised target miRNA. The hairpin discourages the hybridisation to larger nontarget RNAs, and can provide additional stabilisation if the target-probe duplex stacks with the probe hairpin. This design optimisation improves the final specificity of the probes.

The resulting miRNA probes can accurately discriminate between similar sizes and sequences, and this was demonstrated by studies with 19 synthetic human miRNAs with high sequence homology to other miRNAs. We have tested other highly homologous miRNAs as well, but the let-7 family shows the typical and the worst cases of cross-hybridisations. The specificity of the miRNA microarray before empirical Tm balancing is good. However, after Tm-balancing, there is significant improvement in specificity
(Figure 3a). We also found that reducing hybridisation time from 40 hours to 20 did not significantly change the specificity. Agilent currently recommends 20-hour hybridisation in the commercial assays (Figure 3b).

The labelling, probe design and hybridisation technique described here is simple, sensitive and largely unaffected by sample-dependent biases. The assay can measure miRNA directly in complex samples with high yield and specificity. The dephosphorylation and direct labelling of total RNA in the same tube is straightforward and requires very low RNA input. The method also accommodates newly discovered miRNA sequences as they are reported in the continuously evolving miRNA landscape.

This new capability is a unique opportunity to develop a confident and clear picture of the intricate expression networks and systems that impact genomics research. The creation of complete miRNA expression profiles using robust and highly sensitive microarrays allows researchers to be the first to gain broad insight into human, mouse, or rat miRNA expression. miRNA expression signatures are invaluable, and further classification of these different miRNA expression signatures holds great promise in human disease characterisation. This method can serve as the foundation for the future development of a quantitative, fully multiplexed miRNA assay, allowing this diverse yet fundamentally conserved group of small RNAs to be further characterised in their role and involvement in modulating the complex regulatory processes found in cells. Moreover, this may lead to the potential development of robust diagnostic biomarkers for tumour classification, as well as prognostic indicators for chemotherapy and oligonucleotide-based therapeutics.