In 1918, a devastating flu epidemic swept the world, killing more than 50 million people, including an estimated 675,000 in the U.S. The disease was deadlier than previous flu outbreaks, killing more than 2% of those infected. Some people died within hours, and many families lost at least one member.
Why was the virus so dangerous? Could such a pandemic happen again? Could the strains of avian flu now destroying poultry in Asia and spreading from Russia toward Europe turn into deadly human viruses?
CUTTING-EDGE METHODS. The urgency of those questions inspired scientists to embark on a several-year quest to unearth lung tissue from victims of the 1918 flu and to analyze the viruses' genes. The task has been challenging. Some samples came from tissue saved from long-ago autopsies. Others came from people buried in Alaska and frozen by the permafrost.
In both types of samples, however, the amounts of virus were tiny -- usually just a single copy of DNA fragments. As a result, researchers had to use cutting-edge molecular biology methods to read the genetic codes.
Now they have succeeded. In an Oct. 6 paper in Nature, scientists, led by Jeffery K. Taubenberger of the Armed Forces Institute of Pathology, report that they have read the full genetic sequence of the 1918 virus, along with the codes of scores of other influenza strains, including the avian flu.
DEADLY QUALITIES. Simultaneously, a team at the Centers for Disease Control & Prevention wrote in a paper in the Oct. 7 issue of Science that they used the genetic information to recreate the deadly 1918 strain. They proved how virulent the virus was by giving it to mice. The mice quickly died.
The work of these two teams is more than just a scientific tour-de-force. It also opens the door to understanding why some flu strains are deadlier than others -- and to figuring out how a virus that originally infects only birds is able to evolve to target humans. Soon it should be possible to use this information to learn if today's flu strains are mutating in ways that could bring another pandemic.
BusinessWeek science reporter John Carey asked Jeffery Taubenberger about the research and its implications. Edited excerpts of the conversation follow.
You've been working on this since 1995. What has been the biggest challenge?
The biggest hurdle from the beginning has been the amount and quality of genetic material in the lung tissues of the victims. The quality, either from frozen samples or autopsies, has been just terrible. We had incredibly tiny fragments in tiny quantities, so we had to push the limits of molecular biology to even detect them. Every step was really painful.
We also wanted to be as accurate as possible, so we looked for the sequence of each of the fragments multiple times. We actually found some minor variations [in different examples of the 1918 virus]. One variant was good at binding to human tissue, the other could bind to both human and bird tissue.
That seems to raise a key point, since one huge question is whether epidemics are likely to be caused by a bird flu virus that suddenly mutates so that it can infect humans, or whether a bird virus and a human flu virus get together and mix genes, creating an even more dangerous strain. What is the story for the 1918 virus?
This is the biggest surprise, we think. The data support the conclusion that this was an entirely avian virus that adapted for humans. It was not an assortment or mix of bird and human virus, as in the last two pandemics [in 1957 and 1968].
We think this is big news. It suggests that two mechanisms are possible for the emergence of [epidemic strains of] influenza viruses. And it increases concern about what's happening now with the avian flu situation.
The risk is that [today's] avian flu could go either way. It could acquire the ability to infect humans by reassortment [of its genes with a human strain]. Or it could completely adapt to humans without mixing with a human virus, just by acquiring the necessary mutations.
You have looked at scores of different strains so far. What have you learned about what it takes to jump from animals to humans?
We have been trying to look for changes that distinguish the bird and animal viruses from the human ones. We think there are about 10 [genetic] changes that are really important.
Our concern is that the H5 type [such as the new avian flu strains] might be going down a similar path [of adapting to humans] as the 1918 virus did. What we've found is kind of eerie. Practically any of the bird viruses that we can find with some of these changes tend to be the highly pathogenic avian viruses, like those in Asia.
This suggests to us that these viruses are acquiring mutations that make them more human-adapted. So far it's a lucky thing. These viruses can get into a person, and they can kill. But they can't yet spread from person to person. The race is on now to figure out what changes are crucial to allowing the virus to be transmissible from one person to another.
Everyone is correct to be concerned [about a possible new epidemic]. The 1918 studies only increase the concern. We need to try to catch this thing before it gets out the door.
Why was it also important to recreate such a deadly virus in the lab?
If you just look at the sequence on paper, you can't predict why the virus is so virulent. It clearly has changes that make it behave the way it did, but we're not smart enough to predict which ones are important. By doing studies with a virus that contains the 1918 genes, we can set the bases for modeling the virulence.