There is a chance that seasonal flu shots would confer some protection against the avian flu virus H5N1, which has raised fears of a deadly global outbreak. Researchers vaccinated mice against a common form of the flu and found that although the animals got sick they were less likely to die when infected with the related H5N1.

The vaccine caused the mice to create antibodies against neuraminidase, a flu protein that lets newly born virus particles escape from infected cells. The researchers also found signs that some humans carry similar antibodies. "It's hard to prove but my gut feeling would be, if people had high enough levels of this antibody, there certainly would be a reduction in severity" from H5N1 infection, says virologist Richard Webby of St. Jude Children's Research Hospital in Memphis, Tenn., whose group performed the research. "But that's the million dollar question: How much of this antibody do you have to have?"

Researchers name flu viruses based on the type of hemagglutinin (HA) and neuraminidase (NA) proteins they containhence the numbers after "H" and "N" in H5N1. Flu vaccines are designed to prevent infection by eliciting antibodies against HA, which the virus uses to break into cells lining the airways.

But experts have speculated that antibodies against one type of neuraminidase could provide protection against multiple flu viruses sharing the same NA type. Some studies suggest, for example, that the 1968 H3N2 flu pandemic, which killed 1 million people worldwide, was less severe than it might have been because of neuraminidase antibodies left over from the 1957 H2N2 pandemic, which killed twice as many people.

Neuraminidase antibodies would not prevent a person from getting sick with the flu, because they do not stop the virus from infecting cells. To see if they would suffice to make H5N1 infection less severe, Webby and his co-workers injected mice with DNA for the neuraminidase gene from human H1N1, one of three flu subtypes covered by this winter's flu shot. Next they infected the mice with avian H5N1. After two weeks, five out of 10 of these mice survived, but none of the control mice lived.

The researchers also looked at blood serum samples from human volunteers. Of 38 samples, 31 contained antibodies against H1N1 neuraminidase, presumably from past infections or vaccinations. In test tubes, seven of the serum samples inhibited the activity of neuraminidase from H5N1.

The results are "very intriguing" but "it is premature to conclude that immunity induced by the [H1N1] virus will provide significant protection from illness associated with avian influenza H5N1," caution Laura Gillim-Ross and Kanta Subbarao of the National Institute of Allergy and Infectious Disease in an editorial accompanying the report, published online February 12 by PLoS Medicine. They note that with less than 300 confirmed human cases of H5N1 infection, researchers would be hard pressed to determine the amount of antibodies needed to confer protection.

"There's no doubt we've got to focus on hemagglutinin" for developing pandemic flu vaccines, Webby says. The amount of neuraminidase in seasonal flu shots, he says, is unknown and likely varies from batch to batch.