NEWS the annual vaccine doesn’t match the circulating strain exactly, and in a pandemic, they could reduce deaths until a matched vaccine is ready. A One-Size-Fits-All “Anything that would dampen a pandemic would be useful,” says virologist Robert Couch of Flu Vaccine? Baylor College of Medicine in Houston, Texas. One for all The threat of avian influenza has revived efforts to develop “universal” One of the most hotly pursued strategies for a flu vaccines that protect against all human influenza strains. Although universal vaccine against influenza A is based on a flu protein called M2. This protein forms an that goal remains elusive, vaccines that protect against seasonal flu ion channel crossing the membrane of a virus variants could be closer particle or infected cell, barely jutting out from the surface. It’s an appealing target because the Modern medicine’s main weapon against the To make a universal vaccine for influenza A, 23 amino acids that make up the ectodomain, or influenza virus is woefully unsophisticated. which includes the main seasonal flu strains and protruding part, of M2 (known as M2e) scarcely Each year, companies have to make a new batch bird flu, as well as past pandemic strains, some vary from one human flu strain to the next, even of flu vaccine because unlike, say, polio or scientists are hoping to use “conserved” flu back to the 1918 Spanish flu. chickenpox, flu strains change every year. The proteins that don’t mutate much year to year. Scientists first showed in the late 1980s that vaccine is grown in eggs, a process that takes up (Influenza B, the other type, occurs only in to M2 can slow flu infection in mice. to 9 months, and people have to be vaccinated humans and causes milder symptoms.) Some of In 1999, biochemist Walter Fiers’s team at annually, which many don’t bother to do. More the conserved protein vaccines in the works Ghent University in Belgium reported in Nature troubling, if a pandemic strain of influenza Medicine that it had reduced flu deaths came along, the virus could kill millions of HA (Hemagglutinin) in mice with a vaccine made of M2e people in the time it would take to prepare a fused to another protein, the core of the matching vaccine. hepatitis B virus (HepB). (These proteins What scientists dream of is a vac- clumped into viruslike particles bristling with cine that can protect against any flu NA () M2e that stimulated more antibodies to M2e strain for years or even a lifetime. This than did the protein by itself.) In its M Ion channel so-called universal flu vaccine is still a 2 latest paper in Vaccine in January,

long way off, if it’s even possible. But PB1, PB2, PA Fiers’s lab, now collaborating

many labs are dusting off past projects RNA with the vaccine company Acambis in on broad flu vaccines, spurred by new 1 PB2 Cambridge, Massachusetts, has 2 PB1 funding and fears that H5N1, the 3 PA improved the candidate vaccine by deadly avian influenza that has swept 4 HA attaching three copies of M2e to the 5 NP across half the world, could acquire 6 NA HepB core, delivering it nasally— Lipid bilayer 7 M + M the ability to be transmitted from 1 2 which boosts immune responses 8 NS + NS human to human. Until now, “flu has 1 2 compared to injection—and adding never been before high enough on the an adjuvant, an ingredient that also radar screen” for companies in particu- increases the body’s immune response. lar to follow through with a strong push Although M2e is typically conserved, for a universal vaccine, says Gary Nabel, there’s a small chance that the protein director of the Vaccine Research could still evolve, enabling the virus to evade M matrix protein Center at the U.S. National Institute 1 a vaccine. To assess that risk, Walter Gerhard’s of Allergy and Infectious Diseases group at the Wistar Institute in Philadelphia, (NIAID) in Bethesda, Maryland. Pennsylvania, pushed the virus to mutate by Doing so, however, means coming up with exposing mice with weak immune systems to an an alternative way to stimulate immunity to Weak spots. A universal flu vaccine would H1N1 seasonal flu strain while giving them anti- the virus. The tried-and-true technique for target “conserved” proteins, such as M2 or NP, bodies specific to M2e. As they reported last seasonal flu uses a killed virus vaccine that an inner protein. June in the Journal of Virology, M2e mutants works mainly by triggering antibodies to appeared in some mice after 3 weeks, but there hemagglutinin (HA), the on the stimulate production of antibodies as do conven- were only two types—fewer than might have been virus’s surface that it uses to bind to human tional flu vaccines, whereas others rouse certain expected. “To us, that was reassuring,” Gerhard SCIENCE cells. Hemagglutinin and neuraminidase immune system cells to battle the virus. says, because it should be possible to make an (NA), another surface glycoprotein that helps Other scientists are pursuing a slightly less M2e vaccine to match the few anticipated variants. newly made viruses exit cells, give strains ambitious goal: They are working on vaccines Another major caveat is that although their names (H5N1, for example). The that match a particular HA, such as the H5 in M2 vaccines may prevent deaths from flu, they sequences of HA and NA mutate easily, which H5N1, but that also protect against “drift” strains may not keep people from getting sick, the way is why each season’s flu strain—although it that typically emerge from year to year. conventional vaccines normally do, notes Couch. may be the same in subtype, such as H3N2— It’s not yet clear whether any of these broad That’s because M2 antibodies seem to work by “drifts” slightly from the previous year’s, and vaccines will ever work as well as a traditional, binding to infected cells and promoting their

the annual vaccine must be tailor-made. HA-matched vaccine. But they could help when clearance, instead of blocking the virus (which CREDITS (TOP TO BOTTOM): C. BICKEL/

380 21 APRIL 2006 VOL 312 SCIENCE www.sciencemag.org Published by AAAS SPECIALSECTION

sports few M2 surface proteins) from infecting Researchers at Merck and Vical Inc., a some other mechanism—is possible in new cells, as traditional vaccines are thought to biotech company in San Diego, California, humans. The epidemiological data have been do. Fiers’s mice, for instance, still get sick and reported 13 years ago that a vaccine based on NP scanty. But Epstein recently found suggestive lose some weight, although they do survive. Fiers partially protected mice from seasonal influenza evidence by analyzing old records from a argues that, given the limitations of current A, although some animals still died. Instead of study of 60 Cleveland, Ohio, families who seasonal flu vaccines—a regular flu vaccine immunizing the animals with NP itself, the experienced the 1957 H2N2 flu pandemic. matches the circulating strain only 80% to 90% of researchers used DNA Epstein reports in the January the time and often doesn’t work at all in the encoding the protein as 2006 Journal of Infectious elderly, whose immune systems aren’t good at the vaccine, a strategy Diseases that adults (but not making new antibodies—M2 vaccines are a pos- that often generates a children) who had lab-verified sible replacement. Others, including Gerhard, see more powerful cellular H1N1 flu in the years before M2 vaccines as a backup to regular vaccines, per- immune response. Last that pandemic were one-third haps as an added component in annual flu shots. year, FDA researcher as likely to get sick with the Some experts caution that it’s too early to say Suzanne Epstein and 1957 H2N2 flu. that M2 vaccines will work in people, as opposed others showed that mice to mice. Retired New York Medical College survived seasonal flu Narrowing in virologist Edwin Kilbourne, for instance, and could be partially With the two main approaches questions whether Fiers challenged mice with protected against dying to making a truly universal flu sufficiently high doses of virus. from H5N1 by an NP- vaccine still a question mark, Despite the skepticism, several companies based DNA vaccine some investigators are working hope to commercialize M2 flu vaccines, includ- boosted by ferrying the on a seemingly more approach- ing Switzerland-based Cytos Biotechnology; DNA into cells inside able goal: making HA-specific Acambis expects to submit a clinical trial an adenovirus disabled vaccines that protect against application to the U.S. Food and Drug Admin- so it can’t replicate. drift strains within the same istration (FDA) this year. Acambis’s Ashley Like M2 vaccines, HA family. Birkett agrees that “we need to see how it vaccines based on inter- One way to achieve this performs in the clinic.” Another contender is nal viral proteins won’t Broad thinker. Belgian biochemist broad immunity is with a live Merck, which has done animal tests on an M2 prevent infection alto- Walter Fiers hopes to take advantage of attenuated vaccine, which con- vaccine combined with an influenza B vaccine gether because CTLs the similarities among flu viruses. sists of a virus that can still infect made from a conserved stretch of the virus’s target already infected cells and thus should induce HA, says Merck researcher Antonello Pessi. cells. Still, says Epstein, they could offer some CTL responses. A live attenuated nasal vaccine protection until a pandemic vaccine is produced. made each year for annual flu, FluMist, has been Looking inside Her group is now looking at a DNA vaccine that on the market since 2003 in the United States. Another approach to a universal flu vaccine combines NP and M2, a strategy also being pur- Manufacturer MedImmune says that it protects uses conserved internal proteins such as sued by Vical with NIAID support. Others are against mismatched strains. MedImmune and nucleoprotein (NP) to elicit a different kind of considering the inner proteins not for a stand- NIAID will soon begin clinical testing of FluMist immunity, one based on a type of T cell called alone vaccine but as adjuvants that could broaden versions for potential pandemic strains such as a cytotoxic T lymphocyte (CTL) rather than the immune response to HA-based vaccines. H5N1 and H9N2. The downside of live vaccines, on antibodies. CTLs recognize and kill One overarching question is whether long- however, is that there is a small risk that the virus infected cells expressing viral antigens, lasting immune protection against different could revert to a dangerous form, perhaps even fragments of proteins such as NP. flu subtypes—whether through CTLs, M2, or creating a new pandemic strain. A safer way to achieve protection against drift pandemic strains may be with DNA encoding the HA surface protein delivered by means of a viral vector. Compared to the tradi- tional killed virus vaccine, this should stimu- late broadly protective CTL responses to con- served parts of the HA protein that are shared by related strains. Separate teams at the Uni- versity of Pittsburgh in Pennsylvania and Purdue University in West Lafayette, Indiana, both collaborating with the U.S. Centers for Disease Control and Prevention (CDC) in Atlanta, Georgia, reported in The Lancet and the Journal of Virology in February that an adenovirus-delivered vaccine based on H5 DNA protected against both the 1997 Hong Kong strain of H5N1 and the 2004 Vietnam strain. One advantage compared to conventional egg- grown vaccines: The manufacturing of the vaccine is done with cells, and “you can make

CREDITS: WALTER FIERS (PHOTO) WALTER CREDITS: millions of doses in a month’s time,” says

www.sciencemag.org SCIENCE VOL 312 21 APRIL 2006 381 Published by AAAS Influenza

Suryaprakash Sambhara of CDC, senior deliver the DNA, which have some drawbacks, persuading regulatory agencies—who now author on the Lancet study. instead using gold-coated particles and high- approve flu vaccines based only on HA anti- Both universities are seeking funding for clin- speed injection to get HA-based DNA vaccines body responses—to use CTL responses as a ical tests from NIAID and companies. Sambhara into a person’s skin cells. measure of efficacy instead, notes virologist plans to give the Purdue team an adenovirus- Although the variety of approaches to Albert Osterhaus of Erasmus University in based vaccine containing genes for NP and M, broader flu vaccines can be dizzying, “having Rotterdam, the Netherlands. Still, if univer- which codes for M2 and an inner protein, as well all of those efforts moving forward gives us sal—or at least broader—flu vaccines can as H5HA; and Andrea Gambotto of Pittsburgh more weapons in the arsenal and makes us make it to the market, they could save lives dur- hopes his vaccine, which also worked in chickens, more likely to find the best platform,” NIAID’s ing regular flu season and stave off disaster may be picked up as a bird vaccine. A company Nabel says. Even if one approach rises to the when the next pandemic strikes. called PowderMed avoids using adenoviruses to top, there are many obstacles ahead, such as –JOCELYN KAISER

NEWS tionary, easy, and cheap synthetic pathway that Harvard University chemist and Nobel laureate Oseltamivir Becomes Elias Corey and his team are now reporting. Also unanswered is the question of how well oseltamivir will work against a pandemic virus. Plentiful—But Still Not Cheap The drug is effective against seasonal flu strains, but there aren’t solid data about its efficacy The shortage of oseltamivir may soon be a thing of the past. But whether against H5N1, the avian influenza strain that the drug will become cheap enough for developing countries and how some suspect is a prime candidate to evolve into the next flu pandemic. “We think it’s effective” well it will work against a pandemic remain to be seen against H5N1, says Nikki Shindo, a medical officer at the World Health Organization (WHO) For more than 6 months, German physician Roche announced a series of deals with other in Geneva, Switzerland, “but that’s a feeling. We Tido von Schoen-Angerer has “desperately” companies to dramatically ramp up production of would like to have more evidence.” Some studies tried to order oseltamivir from oseltamivir; in 2007, it says it will on the drawing board aim to get just that. Roche, the Swiss company that pro- be capable of producing 400 mil- duces the anti-influenza drug. “But lion treatment courses (each Change of heart I’ve given up,” he says. Von Schoen- consisting of 10 capsules) yearly. Until less than a year ago, Roche routinely dis- Angerer, research and development That’s up from just 6 million missed suggestions that it sublicense oseltamivir, director of the Campaign for 3 years ago and much more than saying it needed tight control over the complex Access to Essential Medicines at the expected demand. Supply production process, which includes a risky step Médecins sans Frontières in Berlin, will be boosted further because a involving an explosive intermediate compound wanted 100,000 treatment courses to handful of generic drug makers called an azide. It also refused to say how much protect MSF personnel and to treat have started producing their own oseltamivir it produced or how much it was patients, should the organization versions of oseltamivir—some charging governments for it. ever find itself at the cradle of a with a sublicense from Roche, But the company did an about-face last fall. In pandemic. But Roche kept saying it others without. the past 6 months, it has signed deals with more simply didn’t have enough of the Although production may than 15 companies, each of which will help carry drug, Von Schoen-Angerer says. He Drug of choice. More finally meet worldwide demand, out a step in Tamiflu’s production process. eventually picked up 500 treatments than 65 countries are several questions remain. It’s In addition, Roche has sublicensed Shanghai from a Dutch wholesaler last month, stockpiling oseltamivir, unclear whether oseltamivir’s Pharmaceuticals and HEC, both in China, and at a price he says was too high. better known as Tamiflu. price will drop enough for poor Hetero in India, to make oseltamivir from begin- As global demand for oseltamivir, countries to stockpile the drug; ning to end. Those companies will produce better known by Roche’s brand name Tamiflu, many fear that they will be left behind (Science, generic versions for local use; they won’t be reached the stratosphere, many clients found 18 November 2005, p. 1103). That’s why it’s named Tamiflu, and Roche will not control quality, themselves at the end of a long queue. But their essential to find simpler ways to produce the drug, production volume, or pricing policy, says David

frustrating wait should soon be over. Last month, some say. One such method may be a revolu- Reddy, Roche’s pandemic task force leader. YEUNG/REUTERS CREDITS (TOP TO BOTTOM): ROCHE; PAUL

382 21 APRIL 2006 VOL 312 SCIENCE www.sciencemag.org Published by AAAS