Tamiflu Resistance Challenges Influenza Genetics Dogma
Recombinomics Commentary 01:49
March 23, 2008
The recently released sequences from H1N1 seasonal flu isolates in the United States by the Centers for Disease Control (CDC) further challenge the influenza genetics dogma, which maintains that such evolution is driven by selection of random mutations. Conceptually, such selection is easily understood when the virus is under attack by antivirals, such as oseltamivir (Tamiflu). Virus which randomly created the mutation which codes for H724Y would be about 1000 fold more efficient at budding from an infected cell, and therefore would have a strong selection advantage in oseltamir treated indivuals.
However, the sudden apperance of the resistance was in countries which usually did not use Tamiflu, and in patients who had not been treated with Tamiflu, which eliminates the role of antiviral selection in these cases. Prior studies had indicated that H274Y would make the virus less fit, so the selective advantage in Tamiflu treated patients would not exist in influenza infected patients who were not under treatment, but the levels H274Y in almost 70% of H1N1 in Norway clearly indicated that H1N1 with H274Y was evolutionarily fit.
This fitness was supported by the recently released sequences by the CDC. All were the Brisbane/59 strain and mapped to the same branch of an H1N1 phylogenetic tree (for both NA and HA), and in many cases the NA sequences were identical, indicating the resistant strain could spread across the country (multiple sequences were released from Arizona in the west and New Jersey in the east).
However, other isolates with H274Y mapped to another branch composed of isolates from western states (Hawaii and California), while other resistant strains were the genetically distinct New Caledonia strain, which was more prevalent in prior years. Thus, not only was the resistant H1N1 evolutionarily fit, but it was formed via multiple independent introductions (at least three in the small number of positives from the US).
Moreover, like the isolates in Europe, the increase in resistance was sudden. Some isolates were identified last season, but the vast majority were from this season, which is also true for countries in Europe, which began to report the detection when multiple H1N1 isolates from Norway were resistant.
The above scenario is difficult to explain by random mutations, even if a “fitness” mutation was widespread and help the resistant strains compete, because the sudden appearance does not correlate with countries that commonly use Tamiflu or recently increased the use of Tamiflu.
However, the change in Tamiflu usage preceding the sudden increase in H274Y in seasonal flu is found in the increased use of Tamiflu to treat H5N1. Although Tamiflu can inhibit the neuraminidase in all nine serotypes, it is less effective in N1. Moreover, it is less effective in H5N1 than H1N1, and in Tamiflu blanket applications, it is used at half the treatment dosage. This set of circumstances lead to the report of Tamiflu resistance Vietnam in 2005 and the genetic change in H5N1 was identical to the change in H1N1.
Moreover, the Tamiflu resistance in Vietnam was also associated with a reduced case fatality rate in northern Vietnam in 2005. In the north, the fatality rate fell from 70% in 2004, to less than 10% in 2005. Moreover, the lower case fatality rate was coincident with more efficient transmission to humans, increasing the likelihood of co-infections of H5N1 and H1N1, including H5N1 with H274Y linked to Tamiflu blanket applications. Although clusters in northern Vietnam were larger, the efficiency of transmission human to human was still markedly below seasonal flu, but recombination in co-infected host would allow for the acquisition of H274Y by H1N1, which could then be more efficiently spread human to human.
In addition, H274Y could be acquired by additional N1 serotypes in birds and swine followed by acquisition of H1N1 at distant locations, such as Norway. In either case, the introduction of H274Y into the human population could then be followed by acquisition of H274Y by other strains via recombination, and would not require new selection of random mutations.
Moreover, the sudden appearance of H274Y in human populations would follow the sudden increase in Tamiflu treatment of H5N1 in multiple locations, including Vietnam, Indonesia, Turkey, and Egypt.
Media Links
Recombinomics Presentations
Recombinomics Publications
Recombinomics Paper at Nature Precedings