NEW YORK — As scientists struggle to find a vaccine to prevent infection with the AIDS virus, a study in mice suggests hope for a new approach -- one that scientists now want to test in people.

The treated mice in the study appeared to have 100 per cent protection against HIV. That doesn't mean the strategy will work in people. But several experts were impressed.

"This is a very important paper (about) a very creative idea," says the government's AIDS chief, Dr. Anthony Fauci. He didn't take part in the research.

The new study involved injecting mice with a protective gene, an idea that's been tested against HIV infection in animals for a decade.

In the nearly 30 years since HIV was identified, scientists haven't been able to find a vaccine that is broadly effective. One boost came in 2009, when a large study in Thailand showed that an experimental vaccine protected about a third of recipients against infection. That's not good enough for general use, but researchers are now trying to improve it.

Researchers reported the new results in mice online Wednesday in the journal Nature. They hope to test the approach in people in a couple of years. Another research team reported similar success in monkeys in 2009 and hopes to start human tests even sooner.

A traditional vaccine works by masquerading as a germ, training the body's immune system to build specific defences in case the real germ shows up. Those defences are generally antibodies, which are proteins in the blood that have just the right shape to grab onto parts of an invading virus. Once that happens, the virus can't establish a lasting infection and is cleared from the body.

Scientists have identified antibodies that neutralize a wide range of HIV strains, but they've had trouble getting people's immune systems to create those antibodies with a vaccine.

The gene-injection goal is straightforward. Rather than trying to train a person's immune system to devise effective antibodies, why not just give a person genes for those proteins? The genes can slip into cells in muscle or some other tissue and make them pump out lots of the antibodies.

The mouse work is reported by David Baltimore and colleagues at the California Institute of Technology.

Ordinary mice don't get infected with HIV, which attacks the immune system. So the research used mice that carried human immune system cells.

Baltimore's team used a harmless virus to carry an antibody gene and injected it once into a leg muscle. The researchers found that the mice made high levels of the antibody for more than a year. The results suggest lifetime protection for a mouse, Baltimore said, although "we simply don't know what will happen in people."

Even when the mice were injected with very high doses of HIV, they didn't show the loss of certain blood cells that results from HIV infection. Baltimore said THAT researchers couldn't completely rule out the possibility of infection, but that their tests found no evidence of it. He said a few hundred mice appeared to be protected.

The work was funded by the federal government and the Bill and Melinda Gates Foundation. Baltimore said his lab has filed for patents.

"I think it's great," said Dr. Philip R. Johnson of the Children's Hospital of Philadelphia, who reported similar results in monkeys in 2009. "It provides additional evidence this is a concept that's worth moving forward."

Johnson said he has discussed doing a human trial with federal regulators and is preparing an application for permission to go ahead. If all goes well, a preliminary experiment to test the safety of the approach might begin in about a year, he said. Baltimore said his group is also planning human experiments that he hopes will start in the next couple years.

Fauci, who is director of the National Institute of Allergy and Infectious Diseases, cautioned that mouse results don't always pan out in human studies. He also said both the gene approach and standard vaccines should be pursued because it's not clear which will work better.

"We're still in the discovery stage of both of them," he said.

Dr. Harris Goldstein, director of the Einstein-Montefiore Center for AIDS Research in New York, who has done similar research in mice, called Baltimore's result a significant advance if it works in humans because it shows a single injection produces high levels of antibodies for a long time.

It might lead not only to preventing infection, but also a treatment for infected people, he said. If it allowed people with HIV to stop or reduce their medications even for temporary periods, they could avoid the inconvenience and side effects of the drugs, he said.