RNAi adds insight not only to basic biology but also gives researchers a toolkit for RNAi-based gene expression manipulation. The latest findings in the field were presented at Select Biosciences’ “RNAi Europe” conference held recently in Prague.
“One of the big concerns in this field is the specificity of RNAi compounds,” said Kristin Wiederholt, research area manager RNAi/gene regulation at Invitrogen (www.invitrogen.com). “Based on the sequence identity, these compounds can have off-target effects, inhibit unintended targets, or cause nonspecific effects.” One way to circumvent nonspecific targeting is by reducing concentrations of siRNA—but this doesn’t eliminate off-target effects.
Invitrogen’s Stealth™ RNAi, a 25-mer double-stranded RNA, is chemically modified so that only the strand that you are trying to knock down the gene of interest for works. Stealth RNAi designers also use bioinformatics to eliminate duplexes that have the potential for off-target effects by ensuring RNAi antisense reagents have a unique 2–12 region core, confirmed against the Unigene database.
Wiederholt stressed that good experimental design includes testing at least three individual nonoverlapping duplexes per target. “You want to make sure you have the same effect or phenotype or knockdown with multiple duplexes against the same target to increase your confidence level.”
As a test case, the company screened its RNAi kinase collection in a NFkB CellSensor® specific reporter based on GeneBLAzer™ reporter technology. In the primary screen, 57 targets down-regulated NFkB. Six targets were validated with a greater than 95% confidence level, including CHUK1, a known component of the NFkB pathway, and CDC2, which is involved in cell-cycle regulation.
Chemically modified duplexes, multiple duplexes, and good bioinformatic design were some of the upfront components needed for successful RNAi.
siRNA Off-targeting Issues
Researchers originally identified off-target RNAi effects when they coupled siRNA knockdowns with global expression profiling experiments. “In addition to the secondary effects of targeted gene knockdown, there were inconsistent patterns of gene-expression alteration,” said William Marshall, executive vp at Dharmacon (www.dharmacon.com). Marshall’s group found that absolute siRNA sequence complementarity did not predict off-target effects. However, miRNA gave them clues to the root cause.
miRNAs use only the first six to seven nucleotides of the guide or antisense strand, a seed region, which then directs binding into the 3'-UTR at one or multiple landing sites. The Dharmacon group tested the hypothesis that off-target RNAi effects come from complementarity found in six to seven nucleotide seed regions in the siRNA. By combining the microarray expression data and bioinformatic alignment tools, Marshall’s group found a statistically significant correlation between the presence of the seed region in siRNA and the landing site for a microRNA in the 3'-UTR of affected genes that could explain 90% of off-target gene silencing.
“Essentially what we proved was that the siRNA is acting like a microRNA, which leads to unintended gene silencing that is due to the microRNA mechanism, not via the siRNA mechanism,” Marshall said. To overcome this limitation, Dharmacon’s approach, known as On-Target Plus™, is to add a 2' O-methyl group in the seed region of the antisense or guide strand that interferes with binding in the seed region yet maintains binding of the full 19-nucleotide siRNA region. The passenger strand gets further modified with 2' O-methyl groups to completely inactivate the strand.
Apoptosis-related Library Screens
siRNA research can facilitate high-throughput genetic analysis, but requires an appropriate system for the delivery of siRNAs into cells.
Dietmar Lenz, senior scientist, R&D at amaxa (www.amaxa.com), discussed Nucleofection, a technology that is based on electroporation and combines optimized electrical parameters with cell-type specific reagent solutions. The two-component system transports substrates such as DNA or siRNA not only into cytoplasm, but also directly into the cell nucleus. Nucleofector® reagents also protect cells against the typical harms of electroporation.
According to Lenz, due to direct transfer of DNA to the nucleus, primary and slow-dividing cells can now be transfected with up to 90% efficiency. siRNA can be delivered with over 95% efficiency for over 30 mammalian primary cell types, as well as other cell lines. amaxa’s new Nucleofector 96-well Shuttle™ system sizes up current protocols to a 96-well format. “For the first time, you can do siRNA library screens in primary cells,” Lenz said.
Researchers screened an siRNA library from Dharmacon, targeting apoptosis-related genes in difficult-to-transfect Jurkat cells in conjunction with the 96-well format Nucleofector. They phenotyped cells resistant to FAS ligand-induced apoptosis with two readout systems, cell death and Caspase 3 activity, using siRNA targeted against FAS as a positive control. So far, they have identified several positive siRNA pools potentially involved with FAS-mediated apoptosis. They now plan to validate and further analyze ten of these hits.
RNAi technologies will need to adapt as they transfer from in vitro experiments to in vivo applications. Previous work has focused on appropriate RNAi substrates. “The design rules of how to design Dicer substrate siRNAs have now been well-established,” said Mark Behlke, M.D., Ph.D., vp, molecular genetics and biophysics, Integrated DNA Technologies (www.idtdna.com). “We have this new asymmetric design with DNA modifications so that we get directed dicing, and we are able to come up with the potent version on the first try.”
Now IDT is taking this knowledge to animal systems. “One of the real problems with the use of siRNA in vivo is delivery,” said Behlke. Along with the City of Hope National Medical Center, Behlke’s team developed a delivery strategy in a TNF-a RNAi-suppression system that uses intraperitoneal dosing in a cationic lipid-delivery medium.
Results show that because delivery avoided standard intravenous methods there was almost no toxicity. The major target for TNF-a was the macrophages that reside in the peritoneal cavity. Delivery of anti-TNF-a Dicer substrate siRNA to the mice and subsequent challenge with lipopolysaccharide (LPS) resulted in mice that did not go into shock, but rather died more slowly than controls, proof that TNF had been blocked.
Preliminary work in collaboration with University of Iowa researchers shows “some favorable results with pulmonary delivery,” according to Behlke. “Any RNA that you are going to give to an animal is at risk to trigger an immune response, but that is easily evaded by the use of chemical modification. If you are going to experiment in vivo, you really should use modified siRNAs.”
Growing evidence shows that miRNAs can act as key gene-expression regulators. This gives researchers hope that the 21mers that mediate endogenous RNAi could be developed as disease markers. “The hope is that microRNAs, being key regulators of cellular processes and differentiation, will have a better classification power than messenger RNAs,” said Thomas Litman, senior scientist at Exiqon (www.exiqon.com). Litman presented results from Exiqon’s pilot study to classify breast cancer prognosis with the company’s 1,000-target Mircury LNA-based microarrays that include all known 500 miRNAs.
The company’s Mircury microarrays contain more than twice the number of first-generation miRNA microarrays used to study cancers. “What distinguishes our study is the number of available miRNAs and also the technology, because we use LNA technology,” Litman said.
The pilot study included 30 patients, which Litman hopes to increase, especially since Exiqon has access to a few thousand samples from public repositories such as Biobank. Mircury microarrays use LNA, an RNA analogue with a methylene bridge between the oxygen and hydroxyl group that locks the structure and causes an increase of the melting temp around 2–8°C per base. This temperature difference can distinguish single-base mismatches, crucial for miRNA research.
So far, their study confirmed miRNA21, a prognostic marker, that Litman said, “the higher the miRNA21, the worse you are.” Their study, through sequencing, also uncovered around 200 new miRNAs not described in the literature.
Analyzing Drug Modes of Action
High-throughput RNAi-based screens are now allowing researchers to expand their screening strategies as the technique becomes more established.
As direct RNAi-induced loss-of-function screening has become accepted throughout the community in the past few years, researchers are now branching out to explore more sophisticated applications of the technology, such as modifier screens, according to Christophe Echeverri, CEO and CSO of Cenix BioScience (www.cenix-bioscience.com). “They allow you to focus the discovery potential of these screens on a pathway or process of interest that’s being targeted by a compound,” said Echeverri. “You are not only getting insights into the mechanism of action of that compound, but you are also opening up the possibility of identifying either enhancers, which serve as drug sensitizers, or suppressors of the compound’s actions.”
Having encouraged users to consider RNAi-modifier screens for several years now, Cenix has begun conducting such projects for Schering. In the first of these studies, Cenix screened 260 high-priority genes chosen by Schering to identify those targets whose silencing would enhance the cytostatic activity of a Schering drug now under development for oncology indications.
Cenix analyzed the drug- and RNAi-induced phenotypes using a high-content microscopy assay involving multiplexed markers to simultaneously measure cell proliferation, cell-cycle progression, apoptosis, and certain other parameters in fixed human cells. Ultimately, over a dozen targets were identified.
Open Biosystems (www.openbiosystems.com) presented data from its mammalian RNAi libraries designed to deliver transfected lentiviral shRNAmir even in cell lines difficult to transfect. Researchers concluded that transduction and knockdown efficiencies from the lentiviral shRNAmir system ranged from 76–91% in one experiment where researchers transduced Hek293T cells with lentiviral particles that contained shRNAmir to a battery of six targets.