The strain of avian influenza virus that has led to the deaths of 140 million birds and 60 people in Asia in the past two years appears to be slowly acquiring genetic changes typical of the "Spanish flu" virus that killed 50 million people nearly a century ago, researchers said yesterday.
How far "bird flu" virus has traveled down the evolutionary path to becoming a pandemic virus is unknown. Nor is it certain that the much-feared strain, designated as influenza A/H5N1, will ever acquire all the genetic features necessary for rapid, worldwide spread.
Nevertheless, the similarities between the Spanish flu virus of 1918 and the H5N1 strain slowly spreading through Asia provide unusually concrete evidence of how dangerous the newer virus is. At least four of its eight genes now contain mutations seen in the deadly strain that circled the globe during and after World War I.
"These H5N1 viruses might be acquiring the ability to adapt to humans, increasing their pandemic risk . . . there is a suggestion there may be some parallel evolution going on," said Jeffery K. Taubenberger, a molecular pathologist at the Armed Forces Institute of Pathology in Rockville.
The comparison of the old and new flu viruses is the first practical use of a science-fiction-like scenario that concluded yesterday with the release of two papers, one by the journal Science and the other by its chief competitor, Nature.
After 10 years of work, Taubenberger and his team succeeded in reconstructing the Spanish flu virus, which was responsible for the deadliest epidemic since the Black Death of the Middle Ages. Reborn in mid-August at a high-security laboratory at the Centers for Disease Control and Prevention in Atlanta, the pathogen appears in animal experiments to be as lethal as it was in humans 87 years ago.
The report came as the United States, many other countries and the World Health Organization are making increasingly urgent preparations for a new flu pandemic.
The Department of Health and Human Services is stockpiling antiviral drugs and is buying enough experimental bird flu vaccine to inoculate 20 million people. President Bush said in a news conference this week that he is considering the use of the military to enforce quarantines, if necessary, and that the government's long-awaited pandemic plan will be released soon.
What makes the accomplishment reported yesterday so remarkable is that no intact samples of the Spanish flu virus exist.
When the pandemic occurred in 1918 and early 1919 -- only American Samoa and parts of Iceland appear to have been spared -- microbiologists did not know for certain what caused it. (The influenza virus was not identified until 1933.) Although biologists were later able to deduce the broad family of influenza viruses the 1918 strain came from, its genetic identity was lost.
Taubenberger and his colleagues were able to piece together the 1918 virus's genes from two unconventional sources. One was fingernail-size pieces of lung tissue, preserved in wax after the autopsies of two soldiers who were among the pandemic's 675,000 American victims. The other source was the frozen body of an Inuit woman who died of influenza in November 1918 and was buried in the permafrost.
The virus's eight "gene segments" -- strands of RNA that are the equivalent of DNA and chromosomes in cells -- were in pieces, like a shelf of ancient vases tipped onto a stone floor. But with gene sequencing and polymerase chain reaction -- the magnifying glass and glue of molecular genetics -- the team reassembled the infamous microbe.
"It is an amazing feat," said Edwin D. Kilbourne, 85, one of the country's leading influenza virologists and a retired professor from Mount Sinai School of Medicine and New York Medical College. "It's a tribute to imagination, perseverance and a great deal of very hard work."
The reconstruction of the 1918 virus is expected to provide immediately useful insights to epidemiologists studying the kind of flu viruses that are being passed among dozens of species of birds and mammals worldwide.
"I think we have been able to unmask the 1918 virus, and it is revealing some of the secrets that will help us prepare for the next pandemic," said Julie L. Gerberding, director of the CDC.
By identifying how the 1918 virus differs genetically from related viruses that do not infect people, researchers hope to pinpoint the mutations an animal flu virus needs to acquire to adapt to human beings. Taubenberger estimates, as a rough guess, that 25 mutations may be essential.
"It could theoretically provide a checklist for surveillance," he said. "You might be able to say: This strain has six of these changes; it's a worrisome virus we need to keep our eye on. Or this one has none."
It will be especially useful to compare the 1918 gene sequence to those of the two flu viruses that have caused pandemics since then -- the "Asian flu" of 1957 and the "Hong Kong flu" of 1968. Hybrid viruses containing genetic features of each can then be constructed and studied in the laboratory.
"We are trying to elucidate the general rules of human adaptation. How does a bird virus become a human virus? Generically put, that is what we're trying to answer," Taubenberger said.
In their paper, Taubenberger, Ann. H. Reid, Thomas G. Fanning and other colleagues at the AFIP describe three genes that make up more than half the 1918 virus's genetic material. Previous papers had described the reconstruction of the virus's five other genes.
Analysis of all eight genes strongly suggests that the Spanish flu virus descended directly from a bird virus and moved into human hosts after slowly accumulating the necessary mutations. It does not appear to have combined with a pig or human flu virus in a process called "reassortment" and to have become a human virus in a sudden step. The 1957 and the 1968 pandemic strains arose through reassortment.
This insight hints that H5N1 might also be capable of adapting to human beings through gradual evolution. In fact, there is evidence the process is already underway.
The polymerase genes described in the new paper produce three proteins that together allow the virus to replicate in cells. Of the 2,232 amino acids in those proteins, 10 are consistently different in the 1918 virus and in other human flu viruses, compared with the proteins in avian flu viruses. The researchers believe the 10 changes may be keys to enabling avian viruses to replicate easily in people. In the past two years, scientists have isolated H5N1 viruses in Asia that have two of the 10 changes.
Similarly, the researchers believe adaptation to humans requires six changes in another gene, called nucleoprotein. H5N1 strains have been found carrying one. In yet another gene -- for a protein called matrix -- four changes appear necessary. Surveillance has turned up H5N1 strains carrying two of the four.
Altogether, the changes suggest that though H5N1 is still a bird virus, it is partway along a path to becoming a human virus.
In a second paper to be published tomorrow in Science, Terrence M. Tumpey of the CDC and Adolfo Garcia-Sastre of Mount Sinai describe experiments with the live, reassembled virus.
They found that mice infected with the microbe died in three days -- unusually fast -- and that unlike most flu viruses, the 1918 strain replicated deep in the lung, not just in the throat and bronchi. That fits the observation that many victims of Spanish flu died when blood or fluid flooded their lungs, often after only a day or two of illness.
The gene sequence of the 1918 virus is now public, which theoretically means it could be reconstructed by others. Publication of the new papers was approved by the federal body that reviews research that could have bioterrorism uses, the National Science Advisory Board for Biosecurity.
Many scientists, however, doubt that Spanish flu could cause a pandemic today if it got out of the laboratory. That is because virtually everyone on Earth has been exposed to many of its descendants in the H1N1 flu virus family -- and consequently are at least partly immune to it.