Measuring gene and protein expression levels throughout the body, researchers can predict who will muster up a rousing immune response to the flu virus just a few days after vaccination, and presumably be better protected against a subsequent infection. The findings, published yesterday (July 10) in Nature Immunology, could explain why vaccines work in some patients but not others, and provide general principles to determine which vaccines will be most effective in a given population.
The approach—using system-wide expression data to predict immune response—is new, and could lead to improved vaccine development, said Sanae Sasaki, an immunologist at Stanford University, who was not involved in the study. “If they can find a key factor that is related [to] immune response, maybe they can find an alternative [vaccine] to induce the immune system” in people who do not have a robust immune response.
Vaccine trials are often slow and expensive, because researchers must vaccinate thousands of people and then wait until some get sick to see if the prevention was successful, said Bali Pulendran, an immunologist at Emory University and co-author of the study. Several years ago, Pulendran and his colleagues began wondering whether system-wide gene and protein expression could predict immune response and thus speed up the clinical trial process.
In 2009, the team used gene and protein expression data to predict immune response to the yellow fever vaccine, identifying a suite of specific markers that could forecast the vaccine’s effectiveness in a given patient. But because that vaccine uses a live virus that replicates inside the host body, it wasn’t clear that the same approach would work for inactivated virus or carbohydrate vaccines.
To test these other vaccine types, Pulendran and colleagues took blood samples from 56 healthy young adults who received either an inactivated influenza or a live attenuated influenza vaccine. At 3 days and a week after vaccination, the researchers measured subjects’ levels of gene expression and inflammatory chemicals known to play a role in immune response. They used mathematical modeling to pinpoint changes in several thousand genes, including several B-cell associated genes, in subjects that went on to have more robust antibody production a month later, regardless of which vaccine they had received.
“We initially began having no preconceived ideas of what genes were important, but as this thing progressed, we could come up with signatures that could predict vaccine efficacy,” Pulendran said. In addition, gene expression analyses revealed noticeable differences that could explain why those vaccinated with the inactivated virus mounted a stronger immune response than those who received attenuated live virus.
The new approach could be used to vet early vaccine candidates in small studies before investing in costly Phase III trials, Pulendran said. Furthermore, researchers can use the technique to retrospectively analyze data to see why some patients responded to a vaccine when others didn’t, he added, which may provide clues for improving the vaccine’s efficacy. Ultimately, the team hopes to deduce general principles about the gene and protein expression that signal an effective immune response for a wide variety of vaccines, and is currently studying whether the approach will work for other existing vaccines, including malaria and shingles.
Follow-up work should also test “individuals who are immune-compromised—the elderly, very young children, or infected individuals—where we need to see why vaccines work less well,” added Rafick-Pierre Sékaly, director of the Vaccine and Gene Therapy Institute of Florida and an author of an accompanying News & Views piece in Nature Immunology.
H. Nakaya, et. al, “Systems biology of vaccination for seasonal influenza in humans,” Nature Immunology, doi:10.1038/ni.2067, 2011.