Tests on lab mice have opened up a new path towards a vaccine against HIV, one of the most frustrating quests in the 30-year history of AIDS, scientists reported on Wednesday.

Genetically modified mice fought back the human immunodeficiency virus

(HIV) after they had been injected with genes to make antibodies, the first line of defence in the immune system, the report said in the journal Nature.

First identified in 1981, AIDS has claimed at least 25 million lives, although the annual toll is falling sharply from the peak of the pandemic in response to drug treatment.

But AIDS campaigners say the pandemic will only be crushed once a vaccine emerges. So far, in clinical trials, only one candidate formula has had even a modest effect, providing a shield of only 31 percent against the risk of HIV infection.

This has prompted researchers to return to the drawing board, to look for "broadly neutralising antibodies" -- Y-shaped proteins that are the immune system's foot soldiers -- among the tiny number of people with an innate ability to resist HIV.

So far, this trawl has turned up around 20 so-called "bNAbs," but there are big unknowns as to how they work and, if so, whether they can be made into a deliverable vaccine.

Delving into this, a team led by David Baltimore at the California Institute of Technology (Caltech) says it has developed a way to deliver bNAb-making genes to lab mice.

The rodents were engineered to carry human cells that allow HIV to penetrate and reproduce.

The approach, called Vectored ImmunoProphylaxis, or VIP, entails using a harmless virus as a "Trojan horse" in which they tucked the genes able to turn out specific bNAbs.

They then injected the virus into the leg muscles of the mice, where it holed up in cells, enabling the bNAb genes to produce antibodies in response to HIV.

The mice were first challenged with just one nanogram of AIDS virus -- enough to infect most non-treated mice that received it -- but the dose was eventually cranked up to 125 nanograms without problems. There were no signs of any side effects.

"VIP has a similar effect to a vaccine but without ever calling on the immune system to do any of the work," said Alejandro Balazs, lead author of the study, in a press release issued by Caltech.

"Normally, you put an antigen or killed bacteria or something into the body, and the immune system figures out how to make an antibody against it.

We've taken that whole part out of the equation."

The team stressed that the jump from mice to humans is large.

"We're not promising that we've actually solved the human problem," said Baltimore. "But the evidence for prevention in these mice is very clear."

He added the team was drawing up plans to cautiously test the method in small-scale human clinical trials.

Baltimore co-won the 1975 Nobel Prize for Medicine at the age of 37 for his work on reverse transcriptase, a key enzyme in the reproduction of retroviruses

-- the family that includes HIV.

In an email exchange with AFP, he said VIP was "like gene therapy, but distinct."

Gene therapy entails slotting a gene into the patient's DNA that corrects a flawed, disease-causing counterpart.

Hopes for this field of research were clouded by several reverses, notably the death of a young volunteer, Jesse Gelsinger, in 1999.

The tragedy raised doubts about where genes should be inserted in the genome and about the safety of the virus that delivered them.

Baltimore explained that VIP used a small, harmless vector, an adeno-associated virus (AAV), which took up residence in the muscle cells but did not slot genes into the mouse's DNA code.

"It's not an 'insertion' but a free plasmid-like element that will exist in muscle cells," he said.

Publication of the study coincided with the eve of World AIDS Day.

The number of people living with HIV currently stands at about 34 million, according to the latest UN figures. (AFP)