Pinning a protein snippet into shape with a protein scaffold triggered the immune system to make an antibody to match. Credit: Proceedings of the National Academy of Sciences

Preventing shape-shifting in a key segment of protein from HIV can prime an immune system to develop antibodies against the virus. Such antibodies, elicited against specific protein segments, could one day serve as the basis for a vaccine to fight many different strains of HIV or other swiftly mutating viruses.

Although a small percentage of individuals infected with HIV develop 'broadly neutralizing antibodies' that disarm different strains of the virus, researchers have so far been unable to develop a vaccine that coaxes the immune system into making such antibodies.

To get around this problem, researchers have tried to extract a key bit of protein — or epitope — that a neutralizing antibody recognizes, in the hope that the immune system will react more strongly to the epitope in isolation. However, without the rest of the protein to hold it in place, the segment generally loses its recognizable stable structure.

Now, scientists in the United States have devised a computer model for identifying a protein that could serve as a type of scaffold, locking an epitope into the structure to which a neutralizing antibody can bind. "We've figured out how to pull out those snippets, retain their structure, and teach the immune system to recognize them," says Peter Kwong, a structural biologist at the National Institute of Allergy and Infectious Diseases in Bethesda, Maryland, and an author of the study, published in Proceedings of the National Academy of Sciences today1.

The researchers tested the approach on an epitope of gp41, a protein on the surface of HIV that helps the virus to invade and infect host cells. The epitope is recognized by an antibody called 2F5, which can neutralize several strains of the virus.

After extracting the epitope from the virus, the researchers transplanted the epitope onto a protein with a complementary structure that acted as the scaffold.

The immune systems of guinea pigs injected with this scaffold–epitope complex developed antibodies very similar in structure to 2F5 antibodies.

Ten times stronger

"They're getting some good targeting," says Jamie Scott, a molecular immunologist at Simon Fraser University in Burnaby, Canada, noting that all of the scaffolds appear to work well, and one in particular is giving a response that's 10 times stronger than the others. That scaffold combined with the epitope produced the best mimicry of the 2F5 antibody. "It's a beautiful show of what at this point in time can be done with modelling and using scaffolds," she adds.

In a parallel study published earlier this month in Structure2, William Schief, a computational structural biologist at the University of Washington in Seattle, and his colleagues designed a protein scaffold to boost the immune system's response to an epitope recognized by a different neutralizing antibody. Schief and David Baker, a structural biologist also at the University of Washington, are coauthors on both papers.

"The sort of antibodies they made using their scaffolds were much more like the original antibody than you'd get just by putting peptides in vaccine candidates, for instance," says Dennis Burton, an HIV researcher at the Scripps Research Institute in La Jolla, California. "That's a huge advance, because it gives us much more control in the design of a vaccine than we have ever had before."

The strategy isn't yet ready to be applied to HIV vaccine design. So far, it falls short in a crucial way: antibodies elicited by these epitopes, although similar to the neutralizing antibodies the researchers were aiming to duplicate, don't actually neutralize HIV.

Kwong says that this is because the 2F5 antibody is thought to neutralize gp41 by also binding at a second spot closer to the viral membrane.

Burton agrees. "For HIV, for this type of antibody, they are going to have to modify their scaffold somehow" to take account of this extra mechanism, he says.

But getting the antibody to work may not be as straightforward as adding that second step, says Scott.

For one thing, she notes, the structure of gp41 at the membrane is still unknown. What's more, she says, neutralizing antibodies are generally isolated from people who have had a chronic infection for about two years, but antibodies in people with more acute infections have a very different structure. Whether antibodies elicited by a vaccine — which are likely similar to an acute response — can have the properties of HIV broadly neutralizing antibodies is just one of the biological mysteries of neutralizing antibodies that structural work may not address, Scott explains. "Is this taking HIV vaccinology ahead? I don't know, because of these other issues," she says.

Nonetheless, says Burton, "this is a big step for rational vaccine design". More immediately, the epitope–scaffold strategy could work for viruses in which the binding between an epitope and a neutralizing antibody is less complicated, adds Scott.

"The most obvious candidate is a universal flu vaccine," says Burton, noting that several antibodies described in the past two years have shown a broad effect against different strains of the influenza virus.