Skip to main content
Premium Trial:

Request an Annual Quote

RNAi and the Off-Target Effect


Sure, RNAi has a lot of potential. After all, it can stop genes from being expressed, and it should soon be able to get to lots of tissues that have been inaccessible to other kinds of treatment. The synthetic small interfering RNA binds to the unwanted mRNA transcripts and they're taken out of circulation. Cleaved and done. Right?

Not quite. "We pick a gene that we think is therapeutically important and we want the small interfering RNA to hit that gene, switch that gene off, and leave other genes alone," says Stuart Pollard, VP of scientific and business strategy at Alnylam Pharmaceuticals.

Over the last few years, weaknesses in the hype of RNAi have surfaced. The small interfering RNAs can silence not only the target gene, but also other genes with a similar enough sequence — a few too many homologous nucleotides and the wrong gene could be quieted. Not only is there that off-target issue, but siRNAs, like double-stranded RNA, are prone to activating the immune system to respond.

The good news is that the problems seem tractable. Each effect can seemingly be avoided by careful screening of target genes, chemical modification of the siRNA, and selection of which lead compounds should be pursued as a potential drug. "We're excited but we've also recognized that this is a mid- to long-term problem," says Alan Sachs, vice president of RNA therapeutics at Merck Research Laboratories.

Off by a bit

In 2006, a team at Rosetta Inpharmatics reported in Nature that small interfering RNAs targeted to silence a specific gene through RNA interference could also affect genes with a sequence similar to the target. "Typically, I call off-target effects those things that are sequence-dependent effects on the expression of genes other than the intended target," says Phillip Zamore, a professor at the University of Massachusetts Medical School. This, he says, is a sequence-dependent down-regulation of genes through a small region of complementarity between the siRNA's six-nucleotide seed sequence and the off-target gene.

"Cross-hybridizing your RNAi would partially match another gene and therefore knock that down some, too. That's made worse with RNAi, potentially, because there's the microRNA pathway which only requires a small match at the 5' end of the anti-sense strand," adds Tod Woolf, president and CEO of RXi Pharmaceuticals.

The Rosetta team worked with Dharmacon on suppressing that off-target activity. Thermo Fisher Scientific's director of biology research and development, Devin Leake, was part of that research effort. At first, he and his collaborators had been interested in modifying siRNAs to increase their potency and specificity. They systematically walked along the siRNA strands, chemically modifying positions along the way. "In our initial chemical modification studies, a couple of very interesting positions appeared," Leake says. "The second position on the guide strand was one of those. It seemed to be a critical feature for guide strand recognition and annealing to a specific target."

In follow-up global expression profiling studies, Leake and his colleagues at Rosetta saw that the second position was important for the siRNA to recognize its target; modifying that position lessened off-target effects. That second spot enables the siRNA guide strand to interact with RISC, as well as to bind the target mRNA. When that spot is chemically modified on siRNA, by adding a 2'-O-methyl to that position, two effects combine to increase specificity: the added chemical occludes off-target interaction while promoting specific target recognition, Leake says.

Though these chemical modifications lessen the off-target effects of siRNAs in number and magnitude, they are not eliminated. "Even with that, the mechanism is inherently not purely specific," points out Merck's Sachs. "This is a problem in the area that will need to be solved."

Zamore isn't so sure that is possible. "I don't see any prospect of making those go away," he says. Furthermore, the chemical modifications have yet to be put to the ultimate test. "Most of those, by the way, remain lab tricks," he adds. "They haven't really been tried in a clinical trial, as far as I know."

In addition to making chemical modifications, Sachs says off-target effects can be assessed empirically. He suggests using gene expression profiling to look at the off-target profiles of lead drug candidates and then following up on the one with the least amount of off-target activity. The process could be successful, Sachs says, with "a combination of chemistry, design, and most importantly, the correct use of gene expression profiling will be needed to choose the best lead [compound] to develop."

Alnylam's Pollard combines an experimental test of a sequence's potential for cross-reaction with a bioinformatics-based approach. "We can compare the sequences that we design with all the sequences it might react with," he says. "We do the experiment to see what level of selectivity does our drug have to the target gene versus these genes that we don't want to hit."

Leake, meanwhile, is still investigating that seed sequence. Not only can it bind to off-target genes, he's found, but it can also attach to 3'-untranslated regions of the human genome, based upon the nucleotide combinations that make up its six nucleotide-long sequence. A seed sequence's ability to bind to an untranslated region correlates to its number of off-target effects. "SiRNAs with low seed frequency correlate with less off-target gene silencing," says Leake. The siRNAs that bind to an untranslated region are then sequestered via the microRNA pathway into the P bodies of the cytoplasm, he adds.

Any difference?

Some researchers say that, in the end, the off-target effect might not be as big a deal as it seems. The effect that they have "are small changes, less than two-fold," says Zamore. "There are comparable kinds of physiological changes caused by small molecule drugs. For the most part, we rely on adaptations in humans — that the body will adapt to that side effect rather than some clever way of engineering them out of small molecule drugs. I suspect that will be generally true for siRNAs also."

Furthermore, a lot of the off-target effects were predictable, says John Rossi at the City of Hope. "If you overexpress the hairpin, you know there's going to be some problems. If you use too much of anything, that's not good," he says. Drug developers have been dealing with side effects for years, and some people think that off-target effects will fall into this category of unwanted-but-still-worth-the-treatment issues. "Off-target effects may play a role in the safety of the drug. They may not," says Sam Reich, Opko Health's executive vice president of ophthalmologics. "As long as the drug is safe, that's OK."

No immunity here

Nucleic acid-based therapies can also activate the immune system. It has nothing to do with their sequences; the very presence of double-stranded RNA can have a class-based effect and trigger cytokine reaction. "That class of non-specific, or undesirable effects, ... clinically, is a much more dangerous problem," says Zamore. "But it's also a much more tractable problem."

Merck's Sachs agrees. "The adverse events that are associated with nucleic acid therapies, including immune stimulation, are avoidable, particularly with significant modification of the RNA sugars," he says.

Initially, the idea was to use small siRNAs. "It was thought that the solution to doing RNAi was to use short ones because they would be too small to induce interferon," says RXi's Woolf.

But a recent Nature paper from the University of Kentucky's Jayakrishna Ambati found that double-stranded RNA just 21 nucleotides long triggered an immune response through toll-like receptor three, which ultimately inhibits angiogenesis. In particular, Ambati studied two siRNAs targeting wet age-related macular degeneration, the ones found in Opko Health's bevasiranib and in Allergan's AGN211745. Both drugs target VEGF to suppress the angiogenesis associated with the disease.

"It's actually known that toll receptor three interacts with long, double-stranded RNAs of viral origin; that's been known for almost a decade now. But it had not been previously shown that even a short-stranded molecule also can interact," says Ambati. "In our system, siRNA has to be 21 nucleotides or longer to active TLR-3. Then there's work from modeling that supports that." The anti-angiogenic properties seen from the drugs, then, could be due to the inflammatory response.

Opko's Reich takes issue with the findings. "We have a tremendous body of data showing that we have the desired effect, the VEGF silencing effect," he says, citing a June paper in Molecular Vision in which his company followed radioactive bevasiranib over time in rabbit eyes and saw it localize to the retina.

Regardless, it seems that just shortening the siRNA to smaller than 18 nucleotides could eliminate TLR-3 recognition. Ambati is cautious, though, since other inflammatory response mechanisms may still be lurking. "It's already known that there are at least half a dozen double-stranded RNA sensors in cells, and I think this is probably just the tip of the iceberg," he says.

Another strategy, then, would be more chemical modifications. "We're working, as many are, to try to understand the impact of TLR-3 activation on the endothelial cells of the eye and more generally in the body. Is the TLR-3 activation something that can be addressed or suppressed by chemical modification of RNA?" says Sachs.

Leake says it can indeed. "For the last several years, chemical modifications have been shown by many different researchers to reduce or eliminate pro-inflammatory effects," he adds.

At RXi, Woolf is already doing that — he's been keeping his eye on these effects since he was a graduate student at Harvard. "The trick is you want to make the RNAi look different so it's not recognized by the immune surveillance, double-stranded RNA detection machinery, yet still act in the RNAi pathway which also recognizes double-stranded RNA. And remarkably, you can do that," he says. "There's a variety of different configurations and chemistries that will do this."

As with off-target effects, there are also experimental ways to discover, and then avoid, triggering an inflammatory response. "We always develop multiple small interfering RNAs to a given target, and we screen to see whether they activate the immune system or not," says Alnylam's Pollard. "We don't develop those products that have that propensity to do it."

Reich also looks at it from the viewpoint of safety and efficacy, those twin benchmarks of approval, rather than unwanted effects. "It's possible that there are class effects. It's possible that there are off-target effects," he says. But the really important questions, he adds, are, "Is the drug efficacious and provide a benefit to the patient that's meaningful? And is it safe enough to give to patients?"


In the Pipeline

Despite these challenges, RNAi-based therapeutics are wending their way through clinical trials to vie for FDA approval. Here's a sampling of what some of the leading therapeutic companies have in their pipelines.

The furthest along the pike is Opko Health's bevasiranib. This VEGF- targeting, anti-angiogenic drug is in phase III trials for treating wet age-associated macular degeneration. "It gets into the cell it needs to, to mediate VEGF silencing. It's safe. We have evidence of anti-angiogenesis," says Sam Reich.

Alnylam Pharmaceuticals has one drug, ALNRSV-1, in phase II trials and others close behind. ALNRSV-1 is being tested to treat people infected with respiratory syncytial virus. The company just finished up a placebo-controlled double-blinded trial of ALNRSV-1 in an experimental infection model. "We showed that ALNRSV-1 directly administered to the upper nasal passage was efficacious compared to a placebo," says Stuart Pollard.

City of Hope's John Rossi is conducting a safety trial of an siRNA targeting HIV. "We've got two patients now that have been transfused with stem cells that are expressing an siRNA targeting HIV, as well as two other antiretroviral RNAs. So far, so good," he says. Through Calando Pharmaceuticals, Rossi is also targeting liver cancer with an siRNA directed against the ribonucleotide reductase gene. He says that will be entering a clinical trial shortly.

RXi Pharmaceuticals is targeting a variety of diseases, including neurology, obesity, and inflammation. Currently, RXi is conducting animal studies of an siRNA against an ALS gene. In metabolism, they are focusing on RIP-150, a gene that when knocked out results in a lean, type II diabetes-resistant trait.

The Scan

NFTs for Genome Sharing

Nature News writes that non-fungible tokens could be a way for people to profit from sharing genomic data.

Wastewater Warning System

Time magazine writes that cities and college campuses are monitoring sewage for SARS-CoV-2, an approach officials hope lasts beyond COVID-19.

Networks to Boost Surveillance

Scientific American writes that new organizations and networks aim to improve the ability of developing countries to conduct SARS-CoV-2 genomic surveillance.

Genome Biology Papers on Gastric Cancer Epimutations, BUTTERFLY, GUNC Tool

In Genome Biology this week: recurrent epigenetic mutations in gastric cancer, correction tool for unique molecular identifier-based assays, and more.