We know that the presence of carbohydrate sites around the cleavage site of HA are important for pathogenicity. We know that certain mutations in the NS1 protein affect cytokine production in some viruses. And we know that mutations that affect replication and decrease replication—such as in the cold-adapted virus and the so-called PB2 627 mutation found in many of the H5 viruses—lead to a difference in pathogenicity in the mouse.

One important consideration is that the very same residue makes no difference when the studies are done in ferrets. Researchers are now using mice, ferrets, and chickens as animal models. The use of non-human primates for influenza virus studies is somewhat questionable. Although mice are obviously the smallest and cheapest animal, there are major questions about the use of inbred mice, particularly the popular strains that lack the MX1 gene, as those do not produce the induction of a normal antiviral state. The best model animal system is probably the ferret: it requires no adaptation for human viruses. However, we do need to agree internationally on one strain of ferret, because British ferrets and Japanese ferrets and American ferrets are not all the same.

Another problem is that very few immunological reagents exist for the ferret. How would we study a cytokine response? We need a ferret DNA sequence genome project, like those under way for another 11 or so organisms, from zebra fish to mouse, and of course humans before that. We also need to increase the number of ferrets. That doesn’t just mean ordering a few more. We need a very large breeding colony—the approach taken with the breeding of woodchucks for hepatitis-B studies. We should also determine if other small animal models would work, such as hamsters and mini-pigs. An underlying need is for biological containment facilities for both tissue-culture work and animal work.

We need to determine the genes needed for both transmission and pathogenicity, which we would do by making reassortments using reverse genetics—the so-called 6 plus 2, 5 plus 3, etc combinations of 8 genes. We would then study the genetic basis underlying transmission, virulence, and pathogenicity in terms of the function of proteins, the structure of proteins at the atomic level, and how proteins go together to make complexes such as polymerase. We do not completely understand the components of what can be reassorted from one strain versus another. The suggestion is there are incompatibilities among proteins—that not all P proteins can go together.

We were asked whether incremental changes in the genome lead to pandemics. We know that from 1997 to 2005 the H5N1 virus gained the ability to kill ducks, and that it has transferred to cats and killed them. We need studies where a series of viruses have changed in virulence pathogenicity.

We need to obtain the genome sequences of all these viruses and post them on public databases. The new NIAID-NIH–sponsored public database with The Institute of Genome Research (TIGR) sequencing genomes of human viruses is a start, but we need such an approach for other viruses as well. We also need reverse genetics working for all of these viruses. And when we have all those components in place, we can then show that mutations are sufficient and necessary for the property of pathogenicity being examined.

The third question we were asked was whether studies are needed to track the rate of antigenic change in avian and human strains, and to predict the changes that occur. As background, note that only three hemagglutinin subtypes—H1, H2 and H3—cause major human disease



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement