A new vaccine can protect macaques against the monkey equivalent of HIV and could provide a fresh approach to an HIV vaccine, a study suggests.
US researchers say the vaccine offered protection to 13 of 24 rhesus macaques treated in the experiment.
In 12 of the monkeys, the vaccine was still effective 12 months later.
They claim the work, published in the journal Nature, could "significantly contribute" to the development of an effective HIV/Aids vaccine.
The researchers gave 24 healthy rhesus macaques a vaccine containing a genetically modified form of the virus, rhesus cytomegalovirus (CMV).
In 13 of the monkeys, the vaccine appeared to offer protection against simian immunodeficiency virus (SIV), the monkey equivalent of HIV. Of these 13, 12 monkeys were still protected one year on.
The researchers say the vaccine works by stimulating the production of a particular type of blood cell, called "effector memory T-cells", which can remain vigilant in the body long after an infection has abated.
Lead author Professor Louis J Picker, of the Vaccine and Gene Therapy Institute in Oregon, compares these cells to armed soldiers at the ready.
"There are soldiers that are back at the base with their rifles in the shed, and then you have the guys out in the field," he told the BBC.
There was also evidence, he said, that the vaccine all but eradicated traces of SIV in the monkeys, something which he said was "unprecedented" in HIV vaccine research.
Researchers in the field welcomed the study, but said safety issues would need to be addressed before similar approaches could be tried in humans.
"I'm excited by the science because it really does demonstrate that it may be possible to eradicate the HIV virus by a strong immune response," said Professor Sir Andrew McMichael of Oxford University.
"But at the same time I'm scratching my head how to take this approach into humans."
Professor McMichael said HIV arose from a type of SIV found in chimpanzees, so the animal model used in the study was a good one. The problem, he said, was the potential safety and regulatory issues with introducing CMV into humans, even though many of us already carry the virus.
"CMV is not totally benign, it does cause a number of diseases. If you're giving people something you're not going to be able to get rid of should it cause problems, then that's quite a difficult risk to manage."
Professor Robin Shattock of Imperial College, London, agreed safety would be key.
"The breakthrough here is in using a viral-delivered vaccine that persists - essentially using an engineered virus to thwart a pathogenic virus. The tricky part will be showing it is safe and effective in humans."
Professor Picker responded by saying such issues would be addressed in forthcoming work, pointing out that early forms of the smallpox vaccine also carried health risks to humans.
"On one level 99% of people in sub-Saharan Africa are CMV-positive and half the people in the developed world are, so we know at lot about it and it's mostly non-pathogenic, except in vulnerable populations like pregnant women," he said.
"We're fully aware to make it available to humans, then the next step is to make a virus which retains or has an enhanced ability to make effector memory cells, but no longer has the capacity to infect vulnerable parts of the population."
Developing an HIV vaccine has so far proved a deeply challenging task, but there have been some promising results.
In 2009, researchers in Thailand published in the Lancet the results of an experimental HIV vaccine, which they said reduced by nearly a third the risk of contracting HIV.
Then last year, a study in the New England Journal of Medicine suggested a drug used to treat HIV-positive patients may offer gay and bisexual men some protection against contracting the virus.
Trials of the combination drug Truvada among nearly 2,500 men suggested it could reduce the chances of male-to-male HIV infection by 44%.
But major breakthroughs remain hard to come by. Indeed, the new Nature study comes as a separate paper in The Lancet Infectious Diseases reports on the failure of an HIV vaccine trial in South Africa.
The MRKAd5 HIV-1 vaccine was trialled in a study involving 801 patients, and no evidence was found that the vaccine was effective.
However, the report authors concede that the study's conclusions may have been compromised by a premature end to the trial.