Commentary
Recombination Drives Global Spread of H1N1 Tamiflu Resistance
Recombinomics Commentary 09:11
August 23, 2008
``What we're seeing is the evolution of the resistance gene and the distribution of it throughout the world,'' said Lance Jennings, a clinical virologist with the Canterbury District Health Board in Christchurch, New Zealand, who is chairman of the Asia-Pacific Advisory Committee on Influenza.
``We have a lot to learn about the molecular epidemiology of influenza viruses.''
The above comments on the dramatic emergence of oseltamivir (Tamiflu) resistance in H1N1 reflect the current state of confusion among those trying to understand the spread, based on the out-dated paradigm of selection of random mutations as a mechanism of antigenic drift. The application of this paradigm to antiviral resistance is conceptually straightforward, since the drug creates a strong selection pressure for the emergence of resistance.
Early reports on one such change, NA H274Y, indicated that this resistance would be limited to patients receiving oseltamivir. Although H274Y was said to generate significant resistance to the drug, influenza with H274Y would not be evolutionarily fit and would not compete favorably with wild type influenza in patients not receiving oseltamivir.
Initial data on H274Y seemed to support this paradigm, as initial reports of the emergence of H274Y was limited to patients in Vietnam who were being treated with Tamiflu, including one patient receiving a sub-optimal prophylactic dose in 2005, as well as a patient in Indonesia who stopped taking a treatment dose in 2006.
However, reports of H274Y in H5N1 in isolates from wild birds in late 2005 demonstrated that H274Y could compete with wild type H5N1 in hosts not taking oseltamivir.
The report of H274Y in H5N1 was then followed by H274Y in H1N1 in seasonal flu cases in the United States at the beginning of the 2006 season. These cases involved the dominant strain in the United States at the time, New Caledonia, which was clade 1. Similarly, in 2006 H274Y was also present in the Hong Kong strain (clade 2C) in China. In both cases, the resistance was found in patients who had not been receiving oseltamivir, demonstrating that H274Y was evolutionarily fit on two distinct H1N1 genetic backgrounds. These isolates shared a region of identity adjacent to the polymorphisms, supporting distribution of the polymorphism via homologous recombination.
At the beginning of the 2007 season in the northern hemisphere, H274Y jumped to another H1N1 genetic background, Brisbane/59 (clade 2B) in Hawaii in the United States. These isolates were closely related to other isolates from Hawaii, but acquired H274Y and also matched the clade 1 and clade 2C sequences on the 3’ side of the acquisition.
This acquisition was followed by a jump to another version of clade 2B, which became the dominant oseltamivir strain in the United States, and became the dominant H1N1 strain in Norway, where the high frequency caught the attention of surveillance groups. The presence o f H274Y on the “northern European” H1N1 background led to widespread reports of high frequencies of H274Y in early 2008.
Thus, the H274Y polymorphism was evolutionarily fit on a number of genetic backgrounds, including H5N1 in wild birds in 2005, followed by H1N1 on clade 1 and clade 2C in 2006, and the dramatic spread onto multiple versions of clade 2B in the 2007/2008 season.
Moreover, at the time of clade 2B expansion, the H1N1 vaccine target switched from clade 1 to clade 2A (Solomon Islands/3). However, in the 2007/2008 season Solomon Island had virtually disappeared, and there were no reports of H274Y on a Solomon Island genetic background. Therefore, the mismatched H1N1 would have reduced effectiveness in blunting the spread of H274Y, and may have accelerated the H1N1 evolution away from the vaccine..
Recent reports of H274Y in the 2008 season in the southern hemisphere include multiple countries where H274Y is being reported on 100% of H1N1 isolates. The first sequences from a country with 100% resistance are from South Africa, where the first 107 H1N1 isolates have H274Y. The dominant sequence from South Africa has a cluster of five polymorphisms near position 190 (H3 numbering) in the receptor binding domain. One of these changes was seen earlier in the “northern European” lineage, and is also present in H1N1 from the 1940’s. The adjacent polymorphism, which is only in the South African isolates, is also in H1N1 isolates from the 1940’s further supporting acquisitions via homologous recombination.
The polymorphisms jumping from one genetic background to another, followed by expansion of the dominant strain, was reported earlier for a polymorphism on NA of H5N1, G743A. The spread of this polymorphism is also dramatic, in the absence of obvious selection, because it is synonymous and therefore does not change the NA sequence.
Like H274Y, G743A was initially reported on multiple genetic backgrounds (all major H5N1 sub-clades). The spread of clade 2.2 out of China in 2005 allowed for further analysis of G743A. In 2006 the polymorphisms was almost exclusively limited to one clade 2.2 sub-clade found in a limited geographical area (southern Germany, northern Switzerland and eastern France).
In early 2007, it appeared in bird isolates in the Nile Delta. The H5N1 in Egypt was well defined by 2006 isolates, which began to diversify in the 2006/2007 season. G743A appeared in multiple isolates in February in the Nile Delta. Plaque purified clones of isolates from one of the birds demonstrated that there were two readily distinguishable sub-clades and both had acquired G743A, which would have been difficult to explain by random mutations, because the parental sequences were present in 2006, and the number of new acquisitions on each background was limited, but included G743A in both instances. Shortly thereafter G743A appeared on additional genetic backgrounds in Egypt, including human isolates in southern Egypt, virtually eliminating the chances of coincidental copy errors on multiple isolates at the same time.
However, the G743A acquisitions were not limited to multiple H5N1 genetic backgrounds in Egypt. At the same time there was a H5N1 outbreak in Moscow and those sequences were closely related to clade 2.2.3 sequences which had been found in Azerbaijan in 2006, without G743A. However, the isoaltes from early 2007 had G743A.
The polymorphism also appeared on another clade 2.2.3 background in Kuwait. This clade 2.2.3 was the Uvs Lake strain which emerged in the summer of 2006 at Uvs lake in Mongolia. This H5N1, which evolved from a massive wild bird outbreak in Mongolia and Russia, migrated to South Korea and Japan in late 2006 and none of the isolates had G743A. However, this genetic background acquired G743A in early 2007 in Kuwait.
The same scenario played out in western Africa. G743A was found in isolates in Ghana on an H5N1 genetic background that had been reported in 2006 in the Ivory Coast. G743 was also on another genetic background related to the H5N1 found in the first human H5N1 case in Nigeria. The presence of G743A is a subset of these related sequences in Nigeria also signaled acquisitions in early 2007.
These outbreaks in early 2007 were followed by outbreaks in Europe, beginning in the summer of 2007. These isolates were the Uvs Lake strain, which became dominant in Europe, and all reported sequences had G743A. Recently the Uvs Lake strain was reported in Nigeria for the first time, and it is likely that G743A will be reported in those sequences also.
Thus, the concurrent acquisition of G743A on multiple genetic backgrounds, and emergence on the dominant strain in Europe, parallels to emergence and spread of H274Y on H1N1 season flu. Neither polymorphism generated clear selection advantages in the avian or human hosts, but became fixed in the dominant clade in circulation, leading to dramatic spread.
These examples of genetic background jumping via homologous recombination are common, and the two examples above illustrate such acquisitions of single nucleotide polymorphisms. This mechanism is the primary driver of influenza evolution and represents a paradigm shift.
These examples will be included in a keynote address to the drug discovery meeting in Beijing in the fall.
Media Links
Recombinomics Presentations
Recombinomics Publications
Recombinomics Paper at Nature Precedings