Fighting HIV with antibodies is like battling a forest fire with garden hoses. They can slow the virus, but it vigorously copies itself, making mutant versions that dodge this critical immune response. Now, an experiment suggests that infusing mice with rare, unusually potent antibodies produced in laboratories and then combined into a cocktail might work as a treatment to supplement or even replace antiretroviral (ARV) drugs. It's a long way from mice to humans, and the cost of producing therapeutic antibodies may be prohibitive. But the concept intrigues many researchers.
The most powerful antibodies against HIV are isolated from untreated people who have failed to control their HIV infection for many years. These so-called broadly neutralizing antibodies (bNAbs) do little to help the people who produce them, but their existence indicates that mutations have created an increasingly diverse population of the virus, which in turn has pushed the immune system to evolve a response that is both more potent and works against more variants. Unfortunately, the virus readily mutates around each individual's bNAbs, too. But a team led by immunologist Michel Nussenzweig of the Rockefeller University in New York City has shown that it may be possible to outwit the virus by combining several bNAbs that target different parts of it.
As Nussenzweig and colleagues explain in a paper published online today in Nature, they tested bNAbs in a "humanized" mouse model. HIV cannot copy itself in mouse cells, so they purchased mice that had deliberately crippled immune systems and then rebuilt them with human stem cells. After infecting the mice with HIV, the group tested five different bNAbs recently isolated from humans and then artificially produced as monoclonals in laboratory cultures.
ARV cocktails are extremely effective at treating HIV-infected people, but the drugs have serious long-term toxicities and must be taken every day. Furthermore, several clinical studies have shown that even when ARVs fully suppress HIV for years, viral levels almost always skyrocket within weeks of stopping treatment. If bNAb therapy worked, it would have distinct advantages, Nussenzweig says. "These antibodies are a natural product from humans and shouldn't have many side effects, and they can be very long-lasting," he says. "Potentially, they could be used twice or three times a year."
When the researchers tested the bNAbs as monotherapies, HIV levels typically decreased but rebounded within 2 weeks of stopping the treatment, as the virus mutated to evade the given antibody. A cocktail of three bNAbs that glom onto HIV's surface protein at different locations fared only slightly better. But when investigators combined five different bNAbs (see graphic), the virus remained suppressed in seven of eight mice for 60 days.
The idea of bNAb therapy lost steam several years ago after two small-scale human studies failed. In the June 2005 issue of Nature Medicine, a group led by Alexandra Trkola of the University of Zurich in Switzerland described how it had scant success with a cocktail of three bNAbs given to 14 HIV-infected humans. But Trkola notes that the existing bNAbs then had much less punch than those available today. "The new potent monoclonal antibodies and the new humanized mouse models without a doubt open fantastic possibilities," she says.
Immunologist Dennis Burton, who specializes in bNAbs against HIV at the Scripps Research Institute in San Diego, California, says the "exciting" concept "still has a long way to go" before bNAb therapy proves its worth in HIV-infected humans. For one thing, HIV does not copy itself to particularly high levels in the humanized mouse, which means people infected for many years would have much more diverse swarms of virus that would be more difficult for bNAbs to thwart. These mice also do not have intact immune systems: Humans might develop antibodies to the monoclonals, rendering them ineffective. On a strictly practical front, monoclonals are expensive and difficult to produce in large quantities.
Nussenzweig realizes the challenges. "This doesn't mean this is going to work in people," he says. "I think what it means is it should be tested in people."