The first HIV vaccine trial occurred in 1987. Since then, at least 35 different vaccines have been tried in over 65 clinical trials involving thousands of human volunteers. But we're still years off from an effective HIV vaccine. Why?

Vaccines work by activating the body's defenses—the immune system—against a safe form of a disease-causing virus or bacteria. After being injected with either an inactivated whole or a characteristic part of the disease, our immune system is ready and waiting when it experiences the real deal. Soon after HIV was discovered, scientists isolated the outer proteins of the HIV virus—the glycoproteins—and designed vaccines around them. Tests with purified laboratory strains of the virus in animal models showed protection. Yet, when tried in humans against wild virus, they all failed.

Creating a successful HIV vaccine is like making a wanted poster for a constantly changing face. HIV-1 has three major groups: M, O, and N. Group M predominates worldwide, with at least 10 different subtypes circulating in significant numbers. New subtypes are created by mutation and by getting a second infection. Even within an infected individual, the virus changes constantly. The glycoproteins are the face the virus presents to the immune system. About a third of this face is different between subtypes, a quarter different within a given subtype. This is a tremendous difference, large enough to confuse our defenses. Making matters worse, the virus hides the unchanging critical parts of its face behind a hypervariable mask, fooling both our immune systems and most vaccines. Usually, our immune systems can protect us from a second infection from the same virus—this is why we only get chicken pox once. But the immune response to HIV infection is too weak to even stop the original infection, let alone prevent a second one.

After the failure to produce an infection-blocking vaccine, some groups have taken on a more modest goal: Make a vaccine that triggers a more vigorous counterattack, slowing or preventing the progression to AIDS and reducing the risk of transmission. Like being introduced to a blind date by a guy who mugged you last week, these vaccines give the immune system a memorably nasty reintroduction to the HIV proteins. The most promising vaccines stuff multiple HIV proteins, including less variable internal machinery proteins of the virus, in a cold virus engineered to be harmless. In animal tests and preliminary human trials these vaccines work better than initially expected, both slowing infections and in some cases even blocking initial infections. The results from human trials against wild virus are not complete—one major phase II trial's results should be out in 2008.

As the decades have dragged on without success, patience has understandably worn thin. Even understanding the challenges has taken years, requiring advances in virology, molecular biology, genetic engineering, and epidemiology—and the labors of thousands of scientists worldwide. Because of this, when an HIV vaccine comes, it will be among mankind's greatest accomplishments.