Published online 24 May 2006 | Nature | doi:10.1038/news060522-10

News

RNA treatment kills mice

Using RNA interference to shut down harmful genes can have fatal flaw.

Some shRNAs are helpful; others are killers.Some shRNAs are helpful; others are killers.

New studies on the safety of a gene-silencing mechanism highlight the need for caution before clinical trials, scientists say.

The research shows that mice can die when injected with high doses of a form of RNA. The RNA molecules, which were folded into structures called short hairpin RNAs (shRNAs), were able to overwhelm a cell's normal RNA-processing mechanisms, with fatal consequences. Mark Kay and postdoctoral fellow Dirk Grimm, both at Stanford University, California, and colleagues report their results this week in Nature1.

Researchers say the work emphasizes the need to proceed carefully towards human trials of RNA interference, a scheme by which RNA is used to shut down harmful genes. Such trials are already underway for a condition known as macular degeneration and a childhood infection called respiratory syncytial virus. And a slew of new trials are planned to treat conditions such as HIV, hepatitis and even bird flu (see 'RNA to the rescue?' ).

Safety concerns about RNA interference have been raised before. Scientists have already reported that some forms of the technique can shut down unintended genes, or trigger a cellular defence mechanism known as the interferon response2. That could lead to harmful side effects.

And Irvin Chen, a molecular virologist at the University of California, Los Angeles, has found that high doses of shRNAs can kill human T cells, which play an important role in the body's natural defences.

"One has to be careful," says John Rossi, a molecular geneticist at the City of Hope's Beckman Research Institute in Duarte, California, who was not involved with the work. "You have to do a lot of preclinical testing; that's what this boils down to."

Too much interference

Grimm and his colleagues delivered 49 different shRNAs into mouse liver cells using stripped-down viruses. They found that 36 of the shRNAs were severely toxic, and 23 killed the mice within two months.

When the scientists looked more closely at the cells, they found out why. The lethal shRNAs blocked the cells' ability to use their own microRNAs, tiny molecules found in most cells that regulate a variety of vital functions. Kay says his study shows that interfering with these microRNAs is a recipe for disaster: "If you override the normal function of these microRNAs, it may not be healthy," Kay says.

Good from the bad

But not every shRNA causes such toxicity, notes Chen, who has studied their effect on the immune system. He says that his lab and others are trying to find out why certain shRNAs are helpful whereas others are harmful. "It's possible to find shRNAs that are not toxic," Chen says. "I'm hoping that in the future we'll be able to make a prediction about which ones would be toxic and which ones wouldn't."

ADVERTISEMENT

In the meantime, scientists are speeding ahead with human trials of RNA interference. Rossi plans to ask the US Food and Drug Administration next month to approve a trial that would combine an shRNA with other gene silencers to fight HIV. And Chen is planning a similar trial. Rossi has found no ill effects from his experimental treatment3.

Companies that are already testing RNA interference in people say they haven't seen any problems either. The company Alnylam, for instance, stresses that it is not using shRNAs, but a different type of molecule, called an siRNA, that has so far proven harmless: "These siRNA molecules have all shown good safety profiles in humans in single and multiple doses," an Alnylam spokesperson said in a statement. Scientists with high hopes for RNA interference hope this good track record stays clean.

Visit our treatmentkills_mice.html">newsblog to read and post comments about this story.  

  • References

    1. Grimm D., et al. Nature, 441. 537 - 541 (2006). | Article |
    2. Sioud M. Trends in Molecular Medicine, 12. 167 - 176 (2006).
    3. Robbins M. A., et al. Nature Biotechnology, 24. 566 - 571 (2006). | Article |