The team said they had made the surprising discovery that the therapy has a generalised effect of blocking blood-vessel growth, a phenomenon that could harm a wide range of tissues.
The method, rooted in a 1998 breakthrough that earned the Nobel Prize for Medicine in 2006, is one of several to use so-called gene silencing.
Under it, tiny synthetic molecules called small-interfering RNA (siRNA) are designed to act as switches that "turn off" genes, the key protein-controlling stretches of DNA.
By tailoring siRNA to specific genes that are known culprits in disease, a disorder could be blocked or reversed. Cancer, kidney disease, Lou Gehrig's and Parkinson's disease are among the disorders deemed good candidates for this therapy.
Investigators led by Jayakrishna Ambati, an Indian-born professor of ophthalmology at the University of Kentucky, carried out research on mice using siRNA strands designed to tackle age-related macular degeneration (AMD).
AMD is a form of blindness that strikes late in life and affects around 50 million people worldwide. Excess blood-vessel cells build up in the retina, preventing it from functioning correctly.
Two US companies, Sirna Therapeutics and OPKO Health, are conducting advanced trials aimed at fixing AMD by injecting siRNAs into the eye. The idea is that the molecules enter cells and switch off the flawed gene, known as Vegfa, that causes the problem.
These tests are still in their early stages and are under the close scrutiny of US health watchdogs.
Ambati's team, reporting in the British journal Nature, were stunned to find that other siRNAs which were not at all designed to turn off Vegfa also did the switch-off trick.
They conclude that siRNAs do not, as thought, penetrate the cell to switch off the baulky gene, but instead bind to a receptor, or docking point, on the cell's exterior.
There they trigger a response from two guardians in the immune system, which wards off encroaching cells. It is this that stops blood-vessel growth.
And, they suggest, siRNAs work generically, rather than specifically -- in other words, they are more of a blunderbuss than a sniper's rifle.
If so, there are major implications for the theoretical foundation underpinning this therapy, they say.
"The very premise on which this entire class of drugs is based is that they work by this highly specific RNA interference mechanism, whereby only one particular gene is targeted and nothing else is touched," Ambati said in a phone interview with AFP.
"This would make it a remarkable tool, if that were the case. But we've found that that is not the case, not just in mice with this eye disease, but also with mice with skin injury and in human cells, and we have work coming up that this happens in other tissues and organs as well.
"And what we've found is that regardless of the sequences of siRNAs and what the target is, they all work the same way."
Stopping blood-vessel growth was fine in AMD and cancer, but in other cases, it could be harmful, especially if the siRNAs were administered intravenously rather than injected locally, he said.
Ambati said "billions" had been invested in siRNA research, but often on the basis of shaky evidence or in defiance of problems that had been pushed to one side.
One of the biggest hurdles was to find a way of getting siRNAs, which are large and negatively-charged molecules, into the cells.
"It's been known and it's been conveniently ignored, it's sort of as if the market has spoken and you move on," he said.
Referring to the former chairman of the US Federal Reserve, Ambati said: "There has been, to use Alan Greenspan's term, irrational exuberance in pursuing this kind of therapy in, I think, an unnecessarily rapid fashion."
PARIS, March 26, 2008 (AFP)