Secondary changes allow spread of oseltamivir resistant influenza virus

The influenza virus neuraminidase (NA) protein is required for virus release from the cell, a property exploited by the antiviral drugs oseltamivir (Tamiflu) and zanamavir (Relenza). During clinical testing of oseltamivir in 2001, some individuals shed drug-resistant viruses with an amino acid change from histidine to tyrosine (H274Y) in NA. Such viruses are not inhibited by oseltamivir because the amino acid change leads to  decreased binding of the drug. But these viruses replicated less well in cell culture, and had reduced infectivity in ferrets. It was concluded that oseltamivir resistant influenza virus mutants would not spread in the population. Why was this conclusion wrong?

During the 2008-09 flu season oseltamivir resistant influenza H1N1 viruses with the H274Y change began to spread, and within a year they were found in most seasonal isolates. It was hypothesized that these viruses contained other amino acid changes that masked the deleterious effect of H274Y. The H274Y mutation does not affect the catalytic activity of the NA: the ability to cleave sialic acid from glycoproteins. However it does lead to a decease in the amount of NA protein that is transported to the surface of infected cells.

Computational methods were used to identify amino acids in NA that could potentially compensate for the effect of H274Y. A single amino acid change at position 194 of NA, when present with H274Y, restored NA on the cell surface to normal levels.

Did a similar amino acid change in seasonal H1N1 strains allow the spread of oseltamivir resistant viruses with H274Y? Introduction of this amino acid change into the seasonal H1N1 strains A/Texas/91 and A/New Caledonia/99 causes a decrease in surface NA. However the same change has a lesser effect on surface NA in cells infected with A/Solomon Islands/2006. Two amino acid changes were identified in the NA protein of recent oseltamivir-resistant seasonal H1N1 viruses that restore surface levels of NA in the presence of H274Y: V234M and R222Q.

It seems likely that the amino acid changes V234M and R222Q emerged first in the NA of seasonal H1N1 viruses. Why these changes appeared is unknown, but they could be a consequence of random drift, antigenic selection, or a need to balance HA and NA activities. Once these changes were in place, oseltamivir resistant viruses with the H274Y could be selected, and because they had no defect in fitness, they spread globally.

The conclusion is that H274Y in NA attenuates the fitness of influenza virus by reducing the amount of NA on the cell surface. Spread of such viruses in the population is impossible without secondary amino acid changes that restore adequate levels of surface NA. H274Y probably causes a defect in NA folding or transport that is balanced by the secondary mutations.

These findings are another example of how drug resistance frequently comes with a cost to protein stability or folding, and prevents evolution unless compensated by secondary mutations.

There have been scattered isolations of oseltamivir-resistant, pandemic 2009 H1N1 influenza virus with the H274Y change. Will these viruses spread globally, or are they less fit, evolutionary dead ends? Introduction of the H274Y change into the NA of 2009 pandemic H1N1 virus leads to a large decrease in surface NA. Unless the 2009 swine-origin viruses already produce excess NA, viruses with the H274Y change are not likely to spread without secondary mutations that rescue NA surface expression.

Bloom JD, Gong LI, & Baltimore D (2010). Permissive secondary mutations enable the evolution of influenza oseltamivir resistance. Science (New York, N.Y.), 328 (5983), 1272-5 PMID: 20522774