by Jeffrey M. Perkel
Cancer diagnostics just ain't what they used to be—and that's a good thing. Classic tests like mammograms, digital-rectal exams, and prostate-specific antigen tests are being joined by a new class of molecular diagnostics based on nucleic acid detection. Promising more individualized and refined medical care, these tests run the gamut, measuring everything from mutations to polymorphisms to gene expression profiles.
Among the most exciting developments in this area is a new type of test, based on a class of molecules whose role in cancer has only recently emerged: microRNAs.
As their name implies, microRNAs (miRNAs) are tiny, just 21 to 23 nucleotides long. Yet these non-coding RNAs pack a big punch, stifling expression of specific messenger RNAs, which they target via base complementarity, by either blocking translation or inducing degradation.
It is an endogenous regulatory process that has captured the attention of such diagnostics firms as Cepheid, Rosetta Genomics, Exiqon, and Combimatrix (among others), all of which have released or are developing human diagnostics based on miRNAs. That's because miRNA properties dovetail perfectly with the needs of diagnostics developers, says Peer Staehler, chief scientific officer at febit.
"MicroRNAs in our hands are quite robust," Staehler says. "Degradation is less of an issue than for mRNA, probably because they are so short. Exactly for that property, they are a pretty interesting molecule for biomarker developers. We can find them in cells, biopsies, and in all major biofluids. And they are different in tissues with and without a major cancer."
Not surprisingly, then, research in this area moves at a blistering pace. PubMed lists more than 4,600 references to the term microRNA, all published since 2000. Since December 2002, the Wellcome Trust Sanger Institute's miRBase Sequence Database (currently at version 13.0, March 2009) has grown from 218 entries to 9,539, representing 103 species.
Nevertheless, says Sangita Parikh, senior product manager for microRNA and gene expression array platforms at Agilent Technologies, the field is "still somewhat in its infancy."
Fortunately, a rich toolset has emerged to aid the field's development.
Several companies offer tools for miRNA profiling, simultaneously assessing the expression of all known miRNAs in a biological sample, using slide-based arrays, bead arrays, or quantitative PCR. Researchers can use this approach to identify potential cancer biomarkers, for instance, by comparing diseased and normal tissue samples or biofluids.
Agilent offers microscope slide-based miRNA profiling arrays. According to Parikh, each slide contains eight arrays (meaning it can profile eight samples at once), each of which contains some 15,000 features (with every miRNA represented by 16 to 20 features). Human, mouse, and rat "catalog" arrays are available, as are custom "eArrays," for which the customer can select any of over 6,000 sequences from the miRBase database.
Affymetrix's GeneChip miRNA array "has one of the most comprehensive coverages" of any miRNA profiling product on the market, says Maher Derbel, miRNA product marketing manager. With 46,228 probes representing 71 organisms, the array includes the complete content of miRBase version 11, plus additional human small nuclear RNAs (e.g., snoRNAs and scaRNAs, which play a role in RNA methylation, among other processes).
Another miRNA profiling product is the nCode array from Invitrogen, a division of Life Technologies. Invitrogen actually offers two forms of the chip, according to Amy Cuneo, product manager for Invitrogen epigenetics: one for multiple species (human, mouse, rat, fruitfly, nematode, and zebrafish), and one restricted to human RNAs. The latter array includes 710 human miRNAs from miRBase version 10, plus an additional 373 novel miRNAs, 29 snoRNAs, and controls.
Those novel miRNAs, Cuneo says, stem from the company's "deep sequencing" of human samples using next-generation DNA sequencing (such as the SOLiD system from Applied Biosystems, another division of Life Technologies). "I believe we are the first company to commercialize products enabled by the results of next-gen sequencing," Cuneo says.
For those without the time or expertise to get microarrays working in their own labs, febit and LC Sciences offer miRNA profiling services based on microfluidic microarrays. febit's array (also available as a product), says Staehler, is basically a traditional spotted array surrounded by microfluidic channels for circulating the hybridization and wash solutions. "It is a very simple device," he says. "The fundamental advantage is that it is very robust and easy to automate."
LC Sciences' array, on the other hand, is actually synthesized on-chip using a photolithographic, microfluidic process, says Chris Hebel, director of business development. The arrays, all of which are custom-built, can have 4,000 or 30,000 elements each. But they are not built from typical nucleic acids, says Hebel.
Unlike the probes on typical gene expression microarrays, miRNAs are too short to allow flexibility in probe design; essentially, the entire miRNA sequence is the probe. That can cause problems down the line, as different microRNAs have very different melting temperatures.
"If you take all the sequences in the Sanger database and you calculate the melting temperature between a miRNA and its capture probe, you get from 45C to 75C, a colossal span," says Keld Sorensen, Executive Director of R&D for the Luminex Bioscience Group.
As a result, Luminex and LC Sciences use modified nucleotides (locked nucleic acids from Exiqon, in Luminex's case) in their capture probes to equilibrate melting temperature values across different miRNA species.
"LNA gives the capture probe a bit of a backbone," says Sorensen, "so the probe is more conducive to hybridization. It increases the melting temperature."
Luminex's FlexmiR assays are based on Luminex's solution-based bead arrays (both custom and catalog arrays are available), in which each miRNA capture probe is associated with a particular bead color. Luminex's LS 200 instrument can distinguish up to 100 different beads in this manner, reading first the bead's identity and then the intensity of hybridization to that bead. Thus, up to 100 different miRNAs can be profiled simultaneously per well of a 96-well plate, though 20 to 30 miRNAs per well is more typical, says Sorensen.
That means FlexmiR occupies the middle ground in terms of miRNA profiling, Sorensen adds.
"Every technology has its sweet spot," he says. "If you need only a couple of samples but 1,000 miRNAs, use a printed array. If you need only two miRNAs, do qRT-PCR. If you need 10 to 20 miRNAs for 1,000 samples, our technology is the right one."
Applied Biosystems addresses the Q-PCR end of the market with its catalog of TaqMan assays, available in both single reaction and multiplexed microfluidic card formats (for measuring up to 384 miRNAs at once). According to Iain Russell, senior product manager of Applied Biosystems miRNA assays, TaqMan assays often are used to validate potentially interesting miRNAs once they have been identified using either arrays or sequencing strategies. (LC Sciences offers miRNA TaqMan assays, as well as Solexa/Illumina next-gen DNA sequencing, as a service.)
To measure miRNAs via microarray, it is necessary first to label them. Most protocols can use total RNA for this purpose, and almost any total RNA isolation system will do (so long as it retains small RNAs); examples include miRvana and TRIzol (Life Technologies), miRNeasy (Qiagen), the Absolutely RNA miRNA isolation kit (Agilent), and the microRNA purification kit (Norgen Biotek). A number of companies have released, or are developing, kits and protocols for purification from archived formalin-fixed, paraffin-embedded (FFPE) tissues, as well—a development that is sure to aid cancer biomarker research as it opens up vast libraries of samples, with associated clinical outcome data, to scientific scrutiny.
But not all labeling systems are the same. Most are based on enzymatic treatment of the RNAs' 3' ends, for instance, by polyadenylation. But, says Heather Kirkpatrick, marketing manager at Mirus Bio, sometimes that 3' end is modified, for instance in plant miRNAs and mammalian piRNAs, which can inhibit labeling and potentially skew findings.
Mirus Bio's LabelIT miRNA labeling kit circumvents this problem, Kirkpatrick says, because it performs chemical, rather than enzymatic, labeling. "Chemical labeling is unbiased relative to enzymatic methods, and that could be an advantage down the road for other classes of RNAs," says her colleague Shannon Bruse, director of scientific operations.
For those who would like to influence gene expression in vivo, either positively or negatively, Life Technologies offers miRNA inhibitors (Anti-miR) and mimics (Pre-miR) through its Ambion product line. And Mirus Bio (among others) offers reagents (for instance, TransIT TKO and TransIT siQUEST) for delivery of those molecules into cultured cells.
Finally, an emerging area of research involves identifying miRNA targets—that is, the genes a particular miRNA silences. Agilent's GeneSpring analysis platform and TargetScan database (current version 5.0, December 2008), use bioinformatics methods (based on base complementarity and folding) to identify putative miRNA targets. Alternatively, LC Sciences offers a genome tiling service to identify miRNA binding sites experimentally.
"It was initially suspected that the miRNAs targeted only the 3’ untranslated region of a gene," says Hebel. "Now we know they target both 3', 5’, promoter regions, and coding regions. So we're learning more all the time about how these miRNAs work."
Given the pace of research and discovery, says Staehler, expect that kind of discovery to continue for some time to come. "We will see a lot of products and success in the next five to 10 years," he says. "I think these heydays are ahead of us."