A strategic tack in the quarter-century-old effort to devise an AIDS vaccine, which last year became darkly clouded by problems, could be more promising than thought, according to a study published on Sunday.

Research -- among lab primates, not humans -- shows that a vaccine based on priming "killer" T-cells, the heavy artillery of the immune system, can work, its authors said.

This approach was massively hit by the failure last year of what seemed a highly promising candidate, a prototype called V520 developed by the US pharma giant Merck.

V520 used a modified form of a virus for the common cold as a "Trojan horse" to deliver elements of the human immunodeficiency virus (HIV) into the body to prime the immune defences.

The new research, carried out among rhesus monkeys, likewise uses a cold virus to deliver the vaccine.

But unlike the Merck vaccine, it proved emphatically that T-cells could be marshalled into action to tackle simian immunodeficiency virus (SIV), a close cousin to HIV.

When injected with a lethal dose of SIV, monkeys were able to brake replication of the virus and remained healthy for more than 500 days after infection.

T-cell vaccines are not of the preventative kind, like the vaccines for polio or smallpox, which have become famous for shielding us against microbial infection.

Instead, they are "therapeutic" vaccines, meaning that they train killer T-cells to fan out on a search-and-destroy mission after infection, slaughtering enough infected cells to keep the viral invaders in line.

AIDS first emerged in 1981. Swift progress in identifying the virus that caused it unleashed early optimism that a vaccine would quickly emerge.

Out of the 50 candidates that have been evaluated among humans, only two vaccines have made it through all three phases of trials, and both were flops.

About 30 vaccines remain in the pipeline.

They focus almost entirely on an immune-cell response rather than by priming antibodies, an approach that was the first choice for vaccine engineers but has now been largely discarded in the realisation that HIV is a viciously mutating foe.

Merck cut short the V520 trial September 2007 after it was discovered that volunteers who had prior immunity to the Ad5 strain of cold virus used in the formula were more at risk to HIV infection than counterparts who had not been given the vaccine.

"This is a controversial field right now," admitted Daniel Barouch of the Beth Israel Deaconness Medical Center of Boston, Massachusetts, who headed the study published online on Sunday in the British journal Nature.

But, he added: "Despite the disappointing setbacks in HIV-1 vaccine development this past year, our findings suggest that we're not at the end of the road when it comes to T-cell vaccines."

His team said Ad5 is one of the main viruses responsible for the common cold and many people have already encountered it.

As a result, there is no longer a sufficient immune "kick" to it, they believe.

Instead, they used a rarely-encountered cousin, rAD26, which was given a boost in a second injection later, this time using Ad5. Both modified viruses carried part of the so-called Gag surface protein of SIV to seek their immune response.

"This is an extremely important study because it shows that there is still hope for vaccines currently in the pipeline," said Bruce Walker, director of the Harvard University Center for AIDS Research (CFAR).

"It also gives the first clear indication of the level and type of immunity that will likely be needed for an AIDS vaccine to work."