Box 1. Deadly combinations

Dog kidney cells can help experts to assess whether a deadly avian flu virus will mix its genes with a common human strain to create a virus that could kill millions. In these cells, virologists at the US Centers for Disease Control and Prevention (CDC) in Atlanta, Georgia, are now running tests using various genetic combinations of the H5N1 avian flu virus and a common human flu strain.

In January, under tight security, CDC virologists began to mimic natural genetic 'reassortment' using reverse genetics (see main text). Using the H5N1 strain as a backbone, they have been substituting various combinations of eight genes from the H3N2 human flu virus.

There are 254 possible combinations, so early experiments are simply screening them to see whether they can survive in mammals — which is where the canine cells come in. "We're trying to approach this in a systematic way," explains Nancy Cox, who heads the CDC's influenza branch.

In the autumn, Cox's team plans to test the most viable and dangerous blends in live animals, to see which ones are readily transmissible. The idea is to get a preview of what a pandemic strain might look like, so workers in the field can be primed to look for novel viruses that may pose a particular threat.

Virologist Albert Osterhaus of Erasmus University in Rotterdam, the Netherlands, expects to get approval from national authorities in the next two months to begin similar research. The animal studies will prove the most challenging, he says: "Transmission experiments are notoriously difficult. You have to mimic the normal human situation."

Osterhaus also aims to conduct transmissibility studies of other strains of bird flu, including H7N7 — which jumped to people in the Netherlands in 2003, causing symptoms including conjunctivitis, and killing one person. "The attention the H5 virus has received is logical, but we should not forget that other avian viruses could do the same thing," he says.

Such research is intensely controversial, because of the potential for giving bioterrorists a recipe to create mayhem. Cox and Osterhaus point out that the genetic engineering involved is beyond the reach of most terrorist groups. "But if we found something that was truly horrific, we might not release that information," Cox adds.

The bigger danger, in fact, is that nature will do the job first. Cox and Osterhaus know they're in a race against time that they can't be sure of winning. "We just don't know when a pandemic strain will strike," says Osterhaus.

Roxanne Khamsi

 CDC

Nancy Cox is seeking the likely form of a pandemic strain of flu.