In research, there is always the risk that someone else will beat you to the finishing line. On top of that, there is the chance that another's findings might undermine your rationale. Dan Barouch of Harvard's Beth Israel Deaconess Medical Center was in the middle of a three-year study testing adenovirus-based vaccine regimens in monkeys when a clinical trial of an HIV vaccine that also relied on an adenovirus failed. But Barouch, whose work centred on the related simian immunodeficiency virus (SIV), persevered, and now reports that his vaccine can protect the animals against infection. His results demonstrate that not all adenovirus vectors need be shelved.

In September 2007, Merck's phase II HIV-vaccine trial, dubbed STEP, was called off because findings showed that it did not prevent HIV infection. The approach involved inserting snippets of HIV genes into a vector — a weakened version of a common-cold-causing adenovirus — which was then injected into human patients. The idea was that proteins produced by the adenovirus would stimulate production of immune-system cells called T cells, and that these would then recognize and kill HIV-infected cells.

The problem with the STEP study was that the kind of adenovirus used — adenovirus serotype 5 (Ad5) — is one of the most common of the 51 known human adenoviruses. As a result, says Barouch, when an Ad5 vector is injected into a person it is recognized by pre-existing antibodies. “In somebody who has pre-existing immunity, Ad5 antibodies may neutralize the vaccine vector before it can do its job,” he explains. He reasoned that a way around the problem would be to use adenoviruses rarely found in humans.

Barouch's group thus sought to identify and develop rare-serotype adenovirus vectors. One candidate, Ad26, is rare in US populations as well as in sub-Saharan Africa, and induced immune responses in animals. In Barouch's study, monkeys received a priming shot of an Ad26-based vaccine and then, several months later, a booster shot of an Ad5-based vaccine. This use of two adenoviruses made the interference of neutralizing antibodies less likely. When exposed to SIV, these monkeys ended up with fewer virus particles in their blood and had a lower risk of dying from infection than those that did not receive the vaccine. By contrast, regimens using Ad5 alone failed to protect the monkeys against SIV infection (see page 87).

That the Ad26/Ad5 regimen worked so well in a setting in which an Ad5-only regimen failed suggests to Barouch that STEP was not the end of the road for T-cell-based vaccines for HIV. His team's work is only a proof-of-concept study, he adds, but it shows that “improved T-cell-based vaccine regimens might be able to do better than the regimen that failed in the STEP study”. And we may soon find out whether this is the case: the Ad26 vector is now in a phase I clinical trial.

The STEP study was so discouraging that some prominent AIDS researchers called for all funding of clinical vaccine research to be redirected to basic research. Barouch advocates a more balanced approach. Although the field of HIV-vaccine research certainly needs much more basic knowledge to solve fundamental scientific problems, he says, both basic research and clinical research should be pursued. “If we had not conducted the STEP study, we would be light years behind where we are today.”