Featured Article
Wednesday June 02, 2010
by Jeffrey M. Perkel
If you want an inkling of how hot the microRNA field is, just look at miRBase.
In April, the University of Manchester's miRNA database updated to version 15 with the addition of some 4,000 new sequences, including 300 or so new human miRNAs. The database now contains 14,197 records from some 130-plus organisms and viruses, up from 10,883 in September 2009's version 14.
"It was unexpected that there would suddenly be such a sudden jump in known human miRNAs," says Christoph Eicken, head of microarray technical services at LC Sciences, a microRNA service provider. "It was almost stable for one to one-and-a-half years, which is a long time in the microRNA field."
The question for the microRNA research community is: what do those small nucleic acids do? Or, more accurately, what mRNAs do they target, and what is the consequence of that regulation?
To answer those questions, scientists first must figure out which miRNAs are expressed under which conditions, and how those expression patterns change. Fortunately, there exists a broad range of tools to help; the trendiest of them might be "next-generation" sequencing of small or total RNA pools.
Transcriptome sequencing is a popular discovery application of such sequencing platforms as Life Technologies’ SOLiD and Illumina's Genome Analyzer, with which users can catalog expressed genes, characterize splice variants, and identify novel RNAs. But these platforms may also be used to quantify expression levels, explains Bob Setterquist, R&D leader for life technologies research & development at Life Technologies.
"Each read is a signal," Setterquist says. "You tally up the number of reads observed for any microRNA and it will reflect its original concentration."
The technique is time consuming, analytically intense, and relatively expensive. But it also has advantages. For instance, sequencers can discern subtle differences in RNA sequence, a feature other strategies cannot easily match, says Setterquist.
"Sequencing has started to clearly point out to researchers that the [sequence of] nucleotides at the ends of miRNAs varies quite a bit," he says. "It's not clear how they get changed—sometimes these are post-transcriptional changes, sometimes they are in the genome—but the ends seem to be more variable than the early studies suggested."
At service firm LC Sciences, customers can choose from three expression analysis offerings: Illumina-based sequencing, microfluidic microarrays, and quantitative real-time (qRT) PCR.
The buzz over next-gen sequencing notwithstanding, the most popular option, by far, Eicken says, is the company's microarray service.
"People hear about next-gen sequencing and are interested in it, but once you talk to them about it, it turns out that often arrays are still the way to go, because when you talk about differential expression, arrays are a better option," he says.
LC Sciences' microarrays differ from standard printed microarrays in that they are microfluidic and built to order (that is, they are not pre-fabricated). As a result, they can be instantly updated when miRBase refreshes, Eicken says. "Basically the day after the release, we offered miRBase 15 version microarrays."
Initially, those arrays were available with 4,000 features. But in mid-May, the company introduced out a 30,000-feature array format, a density that the company will soon leverage to array both miRNAs and their putative mRNA targets on the same chip, Eicken says.
That's not to say sequencing isn't also popular. Indeed, Eicken says he has observed an increase in customers requesting help with some rather esoteric organisms, which are not included (or at least, not well covered) in miRBase. For these customers, LC Sciences offers a new service called SeqArray. Blending the power of next-gen sequencing with the speed and convenience of microarrays, the SeqArray service first sequences an organism's transcriptome and then designs custom microarrays to profile it.
LC Sciences has applied this approach to some unusual species, Eicken says. "We did this with grizzly bear; we also got a request on sugarcane, and all kinds of plant species."
According to Eicken, the advantage of a service provider comes down to speed, access, and bioinformatics prowess. But, for those researchers who don't mind doing their array work in-house, several companies offer off-the-shelf microarrays for miRNA expression profiling, including Affymetrix, Agilent Technologies, Exiqon, and Invitrogen.
According to Alicia Burt, director of microarray systems at Agilent Technologies, microarrays offer several advantages over next-gen sequencing, including cost, ease of data storage and analysis, and access.
"Access to next-generation sequencers is still a problem for most labs," Burt says.
A recent study by researchers at Exiqon and the University of Copenhagen suggests also that, "for quantification of small RNAs such as microRNAs, microarray expression analysis appears as a both highly specific and very sensitive technology that still surpasses next-generation sequencing with respect to absolute RNA expression quantification." [H. Willenbrock et al., RNA, 2009, doi:10.1261/rna.1699809].
At the same time, however, off-the-shelf arrays may not represent the latest miRBase release, because new content needs to be vetted and (often) probe-design optimized before it can be arrayed. Agilent's off-the-shelf Human miRNA Microarray and Exiqon's miRCURY 5th Generation LNA microRNA Array reflect version 14 of the database, for instance. Affymetrix's GeneChip miRNA Array reflects version 11 content, and Invitrogen's NCode Human miRNA Microarray V3, version 10.
Users can design custom arrays, using Agilent's eArray service, for example. Alternatively, for those uncomfortable with solid-phase microarrays, Qiagen, its subsidiary SABiosciences, Life Technologies, and Exiqon offer off-the-shelf or custom profiling "arrays" based on qRT-PCR.
The advantage of PCR arrays, according to Qiagen R&D Director Eric Lader, (other than the greater familiarity many researchers have with PCR over microarrays) is the greater linear dynamic range of qRT-PCR. Microarray quantitation, based on fluorescence imaging, is accurate over two to three orders of magnitude, Lader says, whereas qRT-PCR is accurate over seven to eight orders of magnitude.
"The benefit of longer linear dynamic range is you can look at very scarcely abundant versus quite abundant transcripts simultaneously with pretty high confidence that both of them will be accurate," Lader says.
In a theme common to such products, SABiosciences' RT2 miRNA PCR array employs a universal miRNA specific RT, Lader says. miRNAs are polyadentylated (tailed) on the 3' end and reverse transcribed. That pool is then distributed into the wells of the microplate-based array, in which one miRNA-specific forward primer per well drives second-strand synthesis of individual miRNAs. Finally, SYBR Green-based qRT-PCR with the universal 3' and specific 5' primers is used to measure expression. The complete RT2 array can profile 704 human miRNAs on eight 96-well plates.
Life Technologies' TaqMan Array miRNA Cards can profile up to 754 individual miRNAs in parallel on two microfluidic cards on the company's 7900HT real-time PCR instrument. Exiqon's locked nucleic acid (LNA)-based miRNA Ready-to-Use PCR Panels can profile up to 730 individual miRNAs.
Microarrays and sequencing both act as profiling tools, useful for measuring the expression of a large number of transcripts on a relatively small number of samples. Once researchers identify a subset of miRNAs they want to test in larger sample sizes, they can switch to some more high-throughput options.
Luminex's FlexmiR v2, based on the company's popular xMap system and released in January, can theoretically profile up to 500 different miRNAs per reaction in parallel, says Keld Sorensen, the company's executive director of R&D. In practice, the method is most often used to profile 20 to 30 transcripts, he says, but can easily do so in thousands of samples.
According to Sorensen, the assay works like this: each of up to 500 differentially coded microbeads is coupled to a unique biotinylated capture probe. Those beads are mixed to create a pool representing all the transcripts the researcher wishes to profile, which is then hybridized with the RNA sample. The beads are washed, unhybridized capture probes and probe-RNA mismatches are removed via RNAse digestion, and streptavidin-phycoerythrin is added as a detection reagent. Finally, the two-color assay is read in the company's proprietary flow-based bead reader.
Researchers interested in only a handful of miRNAs can try PCR-based approaches, such as those available from Life Technologies, Exiqon, and QIAGEN.
According to Iain Russell, senior product manager of Life Technologies miRNA assays, the company now offers some 1,900 TaqMan-based miRNA PCR assays representing six species. Exiqon has 730 validated assays based on its proprietary locked nucleic acid (LNA) technology, whereas QIAGEN offers miScript assays for all human, mouse, and rat miRNAs in miRBase 15.
QIAGEN's product line offers an unusual twist on miRNA expression analysis. Mature miRNAs are processed from longer (inactive) precursor molecules. Most qRT-PCR assays detect the mature form of the molecule, as these are biologically active. But QIAGEN offers users the option to assess expression of the precursors as well, using its miScript Precursor assays.
The key, Lader says, is the universal nature of QIAGEN's reverse transcriptase step. Because it uses universal reverse primers (but sequence-specific forward primers), the assay can uniformly produce cDNA of any RNA present in a sample, whether mRNA, miRNA, or other non-coding species. "They all work with one universal set of conditions," he says. "That's the universal nature of our reverse transcriptase reaction."
Given the evolving nature of RNA biology, it's a good bet such universality will come in handy in the years to come.