This article is reproduced with the permission of New Scientist for exclusive use by Nova users.

Turn genes on, turn diseases off
07 April 2007
From New Scientist Print Edition.
Bob Holmes

Enlarge Two ways to turn on a gene

The experiment should have been straightforward. Long-Cheng Li was employing the hottest new technique in biology to turn off genes in cells growing in a dish. The trouble was that the opposite was happening. Instead of blocking production of the protein he was targeting, his RNAs seemed to be revving it up. In fact, levels of the protein shot up fivefold - a massive increase.

"At first I thought there must be something wrong with the samples, that we'd mislabelled them or messed up," recalls Li, a molecular biologist at the University of California, San Francisco. So he had his technician do it all again from scratch - with the same result.

That pulled Li up short. RNA interference, as the technique is called, is supposed to turn genes off, not on. All over the world researchers were using RNAi to turn off almost any gene they wanted, and its potential for treating all kinds of diseases was generating huge excitement. But no one had ever reported RNAi turning genes on.

This would make an immensely powerful technique even more amazing. The possibilities are endless, from treating cancer by switching on our innate anti-tumour genes to restoring the activity of genes that declines as we age. The big question is, have Li and his colleagues made the discovery of a lifetime or simply messed up their experiments?

RNAi has become such a growth industry that it's easy to forget how new it all is. While the effects of gene silencing were spotted in plants in the late 1980s, exactly what triggered these effects wasn't clear. It was only nine years ago that Andrew Fire of Stanford University and Craig Mello of the University of Massachusetts Medical School in Worcester discovered that the key is double-stranded pieces of RNA whose sequence matches part of a gene. Adding just a few of these to a cell will shut down the target gene.

Their finding came as genome sequences were starting to flood in and gave biologists a way to discover what the tens of thousands of newly discovered genes do: if you know a gene's sequence, it's relatively easy to create a piece of RNA that will shut it down so you can see what happens. The technique has become so useful that Fire and Mello shared a Nobel prize in 2006.

Medical applications are not far off either. Trials have already begun of RNAi-based treatments for age-related macular degeneration, the leading cause of blindness in developed countries, and for respiratory syncytial virus, which can cause serious illnesses in children.

Since Fire and Mello's discovery, it has become clear that RNAi exploits an innate mechanism for silencing genes, which may have evolved as a defence against viruses (New Scientist, 14 September 2002, p 28). This mechanism can be triggered by double-stranded RNA molecules just 20 to 30 base pairs long, known as small interfering RNAs or siRNAs. The double strands are unzipped and incorporated into a complex featuring a protein called argonaute, which uses the single-stranded RNA to home in on complementary strands.

"In a sense, the small RNA is the software. It's what makes the machinery addressable. The argonaute protein does the work," says Erik Sontheimer of Northwestern University in Evanston, Illinois.

If the small RNA matches part of the protein-coding portion of a gene, the argonaute complex binds to and destroys the gene's messenger RNAs, the working copies that carry instructions to the part of the cell where proteins are made. No message means no protein and a silent gene. This is the best understood form of RNA interference.

However, it is becoming clear that RNAi comes in more than one flavour. It is also possible to silence a gene with small RNAs that match its promoter sequence - a stretch of DNA before the protein-coding part, to which molecules called transcription factors bind. These determine how active a gene is by either blocking or boosting the production of messenger RNA. When the argonaute complex interacts with a promoter it may block gene expression by triggering the addition of chemical tags to the DNA itself or to the histone proteins around which our DNA is wrapped.

When Li began his work in 2004, this second sort of RNAi had been seen only in plants and yeasts. No one had yet got it to work in mammalian cells. "So I decided to give it a try," Li recalls. He designed an RNA molecule 21 base pairs long that matched the promoter region of a tumour suppressor gene called E-cadherin, then inserted the RNA into prostate cancer cells.

That's when the surprise came. Instead of turning off production of the E-cadherin protein as Li had expected, the researchers found five times as much E-cadherin as before. Baffled, they tried another RNA matched to a different part of the promoter sequence only to get the same result. In fact, with a little tinkering, they got 12-fold increases in gene activity.

Now Li's team had to figure out whether they were looking at a real phenomenon or some niggling little problem with their experiments. Was there something peculiar about their cells? No, the RNAs did the same thing in a different cell line. Was there something weird about E-cadherin? No, they could boost the activity of some, but not all, of the other genes they tried.

Could the activation be a side effect of the RNA targeting some other part of the genome as well? Probably not - Li scoured the human genome database without finding any other match for his RNA sequences. Were there changes in the histone proteins that match what is expected when genes are turned on? Yes - activation corresponded to the loss of a chemical "off" tag, a methyl group on histone protein 3.

As the results piled up, it began to look as though Li and his colleagues were onto something. Their mysterious activation resembled RNAi in some ways - it required an argonaute protein, for example, and some but not all of the 21 base pairs had to be an exact match for the target sequence. But there were intriguing differences, too. For one thing, RNAi usually shuts down gene expression completely within 12 hours, and the effect wears off within a week. Yet it took two to three days for gene activity to peak, and the gene kept humming along for nearly two weeks. Li's team published its results last November (Proceedings of the National Academy of Sciences, vol 103, p 17337).

Meanwhile, at the University of Texas Southwestern Medical Center in Dallas, another team led by chemist David Corey had also been noticing that "gene silencing" experiments occasionally had the opposite effect. The increases in gene activity the team saw were small: 25 to 50 per cent. "Nothing that you would write home about," says Corey, "but they were there, they were reproducible, so it stuck in our minds that this was a possibility." In early 2006, Corey's co-worker Bethany Janowski decided to check it out.

Like most groups working on gene silencing, Corey's team had done most of its work with cell types in which the target genes are quite active - why try to silence an already quiet gene? Janowski realised that this high level of activity might tend to mask any RNA activation. She took one of the mildly activating RNAs they had found earlier, which targets the promoter of the progesterone receptor gene, and tested it in a human cell type in which this gene is not very active. Sure enough, the effect stood out, producing 17-fold increases in gene activity. "They worked beautifully the first time," says Corey. What's more, the method worked in several cell lines and with at least one other gene.

Corey's team found that the line between turning a gene on and turning it off is exquisitely fine: shift the RNA's target just one base pair down the promoter and the effect can switch from silencing to activation. They published their results a few weeks after Li's (Nature Chemical Biology, vol 3, p 166).

As word spreads, other groups are beginning to turn up similar results. David Shames, a colleague of Corey's at Southwestern University, says he has activated at least two of a half-dozen genes he has tested. At City of Hope Medical Center in Duarte, California, Hua Yu says she has seen 5 to 10-fold increases in gene activity after targeting the promoter with small RNAs. The results have yet to be published.

Even so, many leading researchers remain cautious. "There is some scepticism about these papers," says Rob Martienssen, a geneticist at Cold Spring Harbor Laboratory in New York state. In fact, several prominent RNAi researchers declined to speak to New Scientist. This may be because the whole field of promoter-directed RNAi has a whiff of scandal, thanks to a highly publicised case of research fraud involving a biochemist in Japan last year.

While no one thinks Corey or Li have done anything dodgy, there are good reasons for scepticism. For starters, with so many teams working on RNAi, how could such a dramatic effect have gone unnoticed for years? One reason, says Corey, is that most RNAi targets protein-coding sequences, not promoters. The few researchers who are targeting promoters may have dismissed any gene activation as experimental noise. "My bet would be that this has been occurring and people haven't seen it because they haven't looked for it," says Kevin Morris, a molecular biologist at the Scripps Research Institute in La Jolla, California.

A bigger problem is that neither Corey nor Li can explain exactly how small RNAs activate genes. Indeed, they do not even agree entirely on the few details they think they know. For instance, Li found that activation required an argonaute protein to bind to the promoter regions of the genes, while Corey did not.

"These are very important, landmark papers, but it is going to require more work to get the scientific community behind this," says Frank Bennett, who oversees RNAi research at Isis Pharmaceuticals in Carlsbad, California. A step-by-step description of the process would help win over doubters, he says.

In particular, the researchers need to find out if activation is just a backhanded form of normal RNAi. Geneticists are just beginning to realise that a cell's nucleus abounds with RNAs of unknown function, including some that are copies of promoter regions of the DNA (New Scientist, 27 November 2004, p 36). If these RNA copies of promoters somehow silence genes, then RNA activation may merely be silencing the silencers (see Diagram).

While this cannot yet be ruled out, preliminary results suggest otherwise. "Activation is very closely related to RNAi, but there is a distinct difference at a certain stage," says Li's colleague Robert Place, also at the University of California, San Francisco. For one thing, the slower pace of activation compared with silencing suggests something different is happening. The results also suggest one strand of the small RNA molecule is more important than the other - and the crucial strand is the one that pairs with the DNA of the promoter, not with any putative copy, Place says.

Even if RNA activation does turn out to be a backhanded form of silencing, it may not matter. "If you can activate genes, no matter what the mechanism, it could be extremely important," says Martienssen.

As a research tool, RNA activation - dubbed RNAa - could rapidly become as important as RNAi. And it opens up a vast new set of possibilities for treating disease. Li has already tried injecting tumours in mice with small RNAs designed to activate an anti-cancer gene. The treatment made a marked difference in preliminary tests, reducing tumour growth by 47 per cent, he says. Shames, too, has begun to experiment with treating diseases.

It is going to be a long time before any treatments reach patients, because the safety of this approach remains unclear. We know little about the consequences of altering the tags on histones, which adds a whole new dimension of risk. And like conventional RNAi, RNA activation might affect genes other than the target. Small RNAs are not like a conventional drug that can be stopped at the first sign of trouble - the effects will last for weeks. "You could run into some severe toxicities that could be very hard to reverse," says Bennett.

Just a decade ago, RNA was regarded as little more than a messenger for carrying information from gene to protein factory. Now, with the discovery of RNAi and other regulatory roles, the messenger has taken a prominent seat at the controls. "We've got a whole other layer of complexity added to what we never knew was there," says Morris. "RNA is way more funky than we ever thought."

From issue 2598 of New Scientist magazine, 07 April 2007, page 42-45

Why all the fuss about RNAi? Almost all drugs are still discovered rather than designed: millions of compounds are tested to find one or two that might have the desired effect. If a potential wonder drug turns out to have nasty side effects, it's back to square one - no matter how many billions have been spent on testing it. RNA interference, by contrast, relies on small bits of RNA. To turn off - or, it seems, to turn on - any known gene, you just tweak the base-pair sequence of the RNA molecule, which is relatively easy. While a chemist might be lucky to find a few good drug candidates in a lifetime, a biologist could cook up dozens of potentially useful small RNAs in a day. What's more, most conventional drugs work by interacting directly with proteins, which necessitates maintaining high levels of the drug in the body. RNAi lowers or raises protein levels by targeting the source - the gene - and the effects of a single dose last for a week or more. The big catch is delivery. Small RNAs are far larger than conventional drugs, which means that getting them to specific parts of the body and across the cell membrane is a huge challenge. The first treatments to enter human trials target tissues that are relatively easy to access, such as the lungs and eyes. Crucially, though, if researchers can find a way to deliver one small-RNA to a specific tissue, they'll be able to deliver any small RNA to that tissue, because changing the sequence makes no difference to most of the chemical properties of RNA. With dozens of groups around the world working on the problem, and huge fortunes likely to be made by the winners, it may not be very long before the trickle of RNAi treatments entering trials turns into a flood.

For the latest from New Scientiist visit www.newscientist.com