The new findings should help researchers pick the correct flu strains for each year's vaccine, a process that must be carried out a year ahead of time and that is now analogous to making a long-term weather forecast supported by only limited data.
The group charged with making the decisions about vaccines has been right about 80% of the time, said Dr. Anthony S. Fauci, director of the National Institute of Allergy and Infectious Diseases, and the new findings should lead to an even higher success rate.
"If we have competing candidates, this will help to pick which virus should go in a vaccine," added Dr. Arnold S. Monto, an epidemiologist and flu specialist at the University of Michigan.
Influenza strikes up to 15% of the world's population each year, killing, on average, about 250,000 to 500,000 people, according to the World Health Organization. About 300 million people annually are protected by an influenza vaccine, but producing the vaccine is something of a crapshoot because its ingredients are typically chosen nearly a year before the flu season to allow time for production.
The problem is that the flu virus mutates rapidly, particularly the gene for a surface protein called hemagglutinin that plays a key role in interacting with the human immune system. Those mutations reduce the efficacy of the immune response to the virus, limiting vaccine protection.
For more than a half a century, researchers have debated where these mutations occur. One idea, discredited only about a year ago, suggested that a background infection of the virus remained in each country outside the flu season, allowing mutations to occur.
Other hypotheses were that the virus migrated back and forth between the northern and southern hemispheres following the seasons, that the viruses circulated continuously in the tropics, or that new mutations occurred in China and spread to the rest of the world from there.
The new results, appearing today in the journal Nature and Friday in Science, suggest that there is a little bit of truth in each of those scenarios.
In the Science paper, an international team led by geneticists Colin A. Russell and Derek J. Smith of the University of Cambridge studied more than 13,000 samples of influenza A virus collected on six continents between 2002 and 2007.
They studied physical differences in the hemagglutinin molecules on the surfaces of the viruses by measuring how strongly each one bound to an antibody. They also determined the DNA sequence of the hemagglutinin gene for about 10% of the viruses.
The team found that once the viruses leave Asia, they don't change much and rarely return. The areas outside Asia are "evolutionary graveyards," Russell said.
In the Nature report, biologist Edward Holmes of Pennsylvania State University and his colleagues sequenced the entire genome of 1,302 influenza A viruses collected over a 12-year period.
Holmes' results were very similar to Russell's findings. Holmes' team was able to conclude that the viral mutations originated in only one area in the tropics but could not pinpoint the area.
Several factors contribute to Asia's role in the mutations. While flu season in the Western world typically occurs in winter, flu season in Asia generally is the rainy season. Because of the wide geographic variations in the region, there is generally a rainy season somewhere, allowing the viruses to propagate continuously.
Add to that the high population density, and the conditions are ripe for a high transmission rate and accompanying mutations.
The large rate of air travel from Asia to Europe and North America then carries the mutated viruses to a new environment where there is little resistance to them. Much less direct traffic occurs between Southeast Asia and South America, so the virus must travel via North America, delaying its passage.
Some of the virus is undoubtedly carried back to Asia via air travel, Smith said, but by that time widespread immunity to it has built up in the population and it cannot gain a foothold.
Although the new information will be useful in producing vaccines, Monto said, even more important is the proliferation of new influenza monitoring stations throughout Asia in response to fears of a bird flu pandemic.
Too often, he said, researchers have known which strain should be included in a vaccine but have not been able to obtain a good isolate of the virus. The new monitoring stations should be able to provide those strains, he said.