Recent headlines have promised that a ‘universal flu vaccine’ may be within reach, pointing to antibodies that offer broad protection in animal studies. But the scientists behind this effort had to first overcome great skepticism from their peers—as well as an imperfect laboratory test. Hannah Hoag reports on one virologist’s 20-year effort to challenge the tenets of the field.
Influenza is the Lady Gaga of viruses: it reinvents itself each year, often in unexpected ways. But the flu virus is far more dangerous than an infectious tune. Although the flu usually manifests as a mild illness, the virus kills as many as 500,000 people worldwide each year, and it continues to provide a challenge from a vaccination standpoint. Whereas most vaccines for illnesses such as measles or polio offer years or decades of protection, influenza vaccines tend to work for only one season. The relentless refashioning means new influenza vaccines must be routinely reformulated, all at a cost to consumers and global health systems of more than $4 billion each year.
A new type of vaccine could be on the way. In the past few years, a flurry of papers has provided firm evidence of antibodies capable of neutralizing multiple subtypes of the influenza virus. Immunologists say that isolating such antibodies is the first step toward the creation of a universal influenza vaccine that protects against seasonal flu year after year—and possibly prevents hundreds of millions of deaths when the next influenza pandemic sweeps across the globe. Several such universal flu vaccines are already in early human clinical testing. But convincing the biology community of the existence and potential of such antibodies was an uphill battle, and one complicated by a ‘gold standard’ test that masked the key findings.
Yoshinobu Okuno, who has chased the dream of a universal antibody against flu since 1989, knows these challenges well. Okuno, a virologist at Osaka University in Japan, is now viewed by many experts in the field as an important and early champion of the idea. Yet his discovery two decades ago of a broad-acting antibody called C179 didn’t make waves at the time. “People didn’t pay attention to it,” says Ian Wilson, a structural biologist at the Scripps Research Institute in La Jolla, California. “In those days, most people weren’t thinking about broadly neutralizing antibodies that you could develop for flu.”
The very test that prompted Okuno to look for these special antibodies—a tool known as the hemagglutination inhibition assay—tripped up the efforts of others in the field. In hindsight, the fault in the assay provides a cautionary tale of how the shortcomings of a test can mean that biomedical researchers miss what they are not looking for.
Put it to the test
The ultimate aim of any vaccine is to prompt the body to produce antibodies; and the ultimate role of an antibody is to bind harmful invaders in the body and trigger their destruction. In 1941, George Hirst, a prominent flu researcher working at the Rockefeller Foundation in New York, developed a test that could detect antibodies against the influenza virus in serum samples. Hirst noticed that mixing the influenza virus with red blood cells caused the cells to clump together. Hemagglutinins—lollypop-shaped glycoproteins protruding from the virus—attached to receptors on the red blood cells, joining them together. Hirst found that he could inhibit this ‘hemagglutination’ by adding serum that contained influenza antibodies. Antibodies bound to the hemagglutinins prevented the virus from attaching to the red blood cells and impeded the cells from clumping.
Today, the hemagglutination inhibition assay is considered to be the gold-standard serologic test to type influenza antibodies in humans and animals. It’s also used by vaccine manufacturers to determine vaccine efficacy and support the approval of seasonal influenza vaccines by the US Food and Drug Administration (FDA).
In 1989, at the age of 42, Okuno decided to drop the decades of work he’d done on flaviviruses and focus his research on the influenza virus. He wanted to find an antibody he could use to develop a better flu vaccine. Okuno thought that he could isolate promising influenza antibodies that might lead him to a target for a universal flu vaccine by comparing results from the hemagglutination inhibition assay and the rapid microneutralization test. “I speculated that if there were conserved neutralizing epitopes on the hemagglutinin, they must be located in the stem region, not in the head,” Okuno recalls. If the immune system produced antibodies that bound the hemagglutinin stem, he reasoned, the hemagglutination inhibition assay wouldn’t ‘see’ them.
Okuno worked doggedly, but often alone. “In Japan, in the first half of the 1990s, it was a dark period for influenza research. Many influenza researchers had changed their [research focus] to HIV or to other viruses,” he says. “People thought influenza was not an important disease.”
Okuno began his work in mice, infecting them with an H2N2 flu virus called A/Okuda/57, a strain that had been isolated from a patient named Okuda by Okuno’s father in 1957. Okuno pulled out hundreds of antibodies and used the hemagglutination inhibition assay and a microneutralization test modified for influenza viruses to sort though them. Most influenza antibodies—whether those produced naturally or induced by vaccines—bind the spherical head of the hemagglutinin protein to stop the influenza virus from attaching to the cell. But this part of the virus evolves steadily, making it an imperfect target. Okuno was looking for an antibody that attached to the hemagglutinin stem, which changes little from one flu season to the next because mutations within the stem decrease the evolutionary fitness of the virus. It was a “tricky method,” he says.
Mice given the antibody were protected from A/Okuda/57 but also another type of flu, H1N1, Okuno found. In fact, the antibody neutralized several other subtypes of the virus, including H1, H2, H5, H6 and H9 viruses. It was the first time anyone had found such a conserved site on the hemagglutinin stem. When he discovered C179, Okuno recalled a Japanese proverb: You can realize what you wish for. “Okuno did some very nice work, with lots of details and three follow-up studies in good journals,” says Antonio Lanzavecchia, director of the Institute for Research in Biomedicine in Bellinzona, Switzerland, and cofounder of Humabs Biomed. But for the next 15 years, the work of Okuno and his collaborators went largely unnoticed by the influenza community.
In hindsight, part of the reason the work stayed under the radar may have been because the hemagglutination inhibition assay was still used widely to detect antibodies against the influenza virus, and antibodies like C179 failed the test. “The surrogate for neutralization is the hemagglutination inhibition test. It’s something the flu folks have been using for years,” says Kanta Subbarao, chief of the division of emerging infections at the US National Institute of Allergy and Infectious Diseases (NIAID) in Bethesda, Maryland. Okuno’s discovery seemed improbable to many. “It just sat there,” Subbarao says.
Only recently have scientists accepted the idea that one antibody could cross-react with several types of influenza. “The dogma was that antibodies could only neutralize a given subtype of influenza and only some isolates within a given subtype,” says Lanzavecchia. But starting four years ago, half a dozen independent groups began to publish their discoveries of human antibodies capable of neutralizing many different subtypes of the influenza virus. Some assert that they might have been unearthed sooner if the hemagglutination inhibition assay hadn’t been used as a stand-in for neutralization.
In recent years, scientists have taken advantage of new molecular and immunological tools to isolate promising antibodies, and they have become more comfortable with microneutralization tests instead of relying on the tests used in the past. These renewed efforts have begun to bear fruit. Near the end of 2008, scientists at Crucell, a Dutch vaccine company now owned by Johnson & Johnson, the New Jersey–based pharma giant, reported on the discovery of a human antibody much like Okuno’s mouse antibody. CR6261 protected mice before and after they had been exposed to lethal doses of H5N1 and H1N1, and it could neutralize many influenza subtypes. But it didn’t inhibit hemagglutination. Initially, the Crucell scientists puzzled over the failed hemagglutination inhibition assay. Although they were sure of their results—the antibody worked in the animal models—they didn’t have direct evidence that CR6261 bound the stem region. “It was impossible to publish,” says Jaap Goudsmit, now the director at the Crucell Vaccine Institute. “No one believed us.” After sending the paper to several journals, they finally published in PLoS ONE.
Meanwhile, in February 2009, immunologist Wayne Marasco and his colleagues at the Dana-Farber Cancer Institute in Boston described a set of antibodies that neutralized three clades of H5N1 ‘bird flu’ as well as the H1 influenza controls. “We thought we had an artifact. It took us a while to recognize the breadth of our discovery,” says Marasco. But the antibodies also protected mice from lethal doses of two strains of H5N1 influenza.
The shake-up finally took hold a week later, when Wilson’s group at the Scripps, working with researchers at Crucell, published in Science. The team showed the cocrystal structures of CR6261 bound to viruses responsible for the 1918 H1N1 influenza pandemic and the more recent H5N1 avian flu. CR6261 was almost identical to Marasco’s antibody. “We were dumbstruck,” says Damian Ekiert, a postdoc at the University of California–San Francisco who was a graduate student in Wilson’s laboratory at the time. “People talk about antibodies like they’re snowflakes, and here we had each pulled the same antibody out of different people.”
By 2011, Lanzavecchia’s group had also found an antibody in human plasma cells that recognized all 16 hemagglutinin subtypes of influenza A. Wilson and Ekiert, along with scientists at Crucell, simultaneously published data on a broad-acting human monoclonal antibody, CR8020, that broadly neutralized many ‘group 2’ strains of influenza A.
These antibodies have spurred companies, including Crucell, to pursue the development of universal flu vaccines and drugs to protect frontline workers in an influenza pandemic or treat serious illness caused by seasonal influenza. “I think there is a niche for the right antibody at the right efficacy that could do a lot of good for people at high risk,” says Gary Nabel, chief scientific officer of the French drugmaker Sanofi and the former director of the NIAID’s Vaccine Research Center.
When it comes to the 2009 pandemic flu, the testing problems pile up. John Schrader, an immunologist at the University of British Columbia in Vancouver, found one during his study of H1N1 ‘swine flu’ infection. He discovered that people exposed to H1N1 or vaccinated against it produced some antibodies that protected and treated mice exposed to the H5N1 avian flu. None of the antibodies were picked up by the hemagglutination inhibition assay. But then two experienced public health laboratories failed to show that the antibodies could neutralize the H1N1 pandemic virus using a standard World Health Organization influenza neutralization assay. Schrader was puzzled for months until he realized that unless he modified the neutralization assay, it couldn’t identify antibodies targeting the stem either. It’s hard to say for certain that the test misled other scientists whose experiments had tried to capture similar broadly neutralizing antibodies, “but I suspect it did,” says Subbarao. “Obviously, we now know that you’d never pick it up.” Meanwhile, the reach of broad-acting antibodies continues: in 2012, a group that included Ekiert, Wilson and Crucell scientists reported data showing that they found antibodies that can protect against both influenza A and influenza B.
Okuno, now 65, says he’s surprised by how many different scientists have managed to capture broadly neutralizing human monoclonal antibodies that are so similar to C179. “It seemed very difficult to obtain these antibodies,” he says. He continues to work in the field, and although he’s pleased to see references to C179 in “the great journals” and to meet the scientists behind the work, Okuno says he wishes he’d found the human form first.
The trials and tribulations of the hunt for a universal flu antibody should serve as a cautionary tale for the entire field of immunology, asserts Lanzavecchia. “It’s important to build up a consensus around some ideas, and at the same time, there comes a time to change that consensus,” he says. “Right now, we’re changing it. We now know there are rare antibodies that are neutralizing viruses that are very different and [that] were not thought to be cross-neutralizable. After influenza, there are other viruses as well: dengue, rabies, other respiratory viruses.”
A version of this story appeared in the January 2013 print issue of Nature Medicine.