Paradigm Shift Intervention Monitoring
twitter
Live feed of underlying pandemic map data here
Commentary
WHO Obsolete Reports
on H1N1 D225G Raise Concerns
Recombinomics
Commentary 22:08
February 03, 2010
The mutations appear to occur sporadically and spontaneously. To date, no links between the small number of patients infected with the mutated virus have been found and the mutation does not appear to spread.
The D222G substitution has been detected in virus isolates from around 20 countries, areas and territories in the Americas, Asia, Europe and Oceania. These changes have been found since April 2009 but have not been associated with temporal or geographical clustering, strongly suggesting the mutation in these viruses has occurred sporadically as opposed to the emergence and sustained transmission of a variant virus.
The three statements above are from various WHO announcements on the receptor binding domain change D225G (aka D222G) which has a high association with fatal H1N1 cases. The comments above were obsolete or incorrect at the time of public release and raise serious questions about WHO’s ability to understand the evolution of pandemic H1N1. They are still tethered to an outdated and incorrect notion that genetic drift in influenza is largely driven by random copy errors which emerge through selection. This notion has been thoroughly discounted by influenza sequences, which support homologous recombination as the key driver.
The above comments on D225G were largely dictated by events in Ukraine. In late October media reports described the death of about 100 young adults who were appearing at emergency rooms, coughing up blood, and dying within days with severe lung damage. Although similar descriptions had been reported previously for fatally infected H1N1 patients, the high number of cases in a small region in western Ukraine led to a WHO investigative team traveling to Ukraine to gather information.
The above circumstances led to the prediction that the cases would be due to H1N1 with receptor binding domain changes, and these changes would be D225G and D225N. This prediction was based on sequences that had just been released from Brazil, Australia, and China. The HA sequences had D225G, D225N or both and were largely associated with severe or fatal cases and the H1N1 in each are represented a distinct genetic background.
The movement of the same polymorphism to multiple genetic backgrounds was best explained by recombination, which had been previously detail in clade 2.2 H5N1 as well as multiple clades of seasonal H1N1. For H1N1 the same synonymous polymorphism, NA G743A, suddenly appeared on multipe clade 2.2 sub-clades in Egypt in early 2007. This change appeared concurrently on different genetic backgrounds in Russia, Kuwait, Ghana, and Nigeria, which was most easily explained by recombination with a host sequence carrying G743A.
Similar results were seen in seasonal flu. Tamiflu resistance H274Y also appeared on multiple H1N1 clade 2B backgrounds in 2007/2008, but more detailed analysis indicated it had began to appear in 2006 in patients who had not received Tamiflu. Like the H5N1 polymorphism, it jumped from sub-clade to sub-clade. In 2007/2008 addition polymorphisms appeared on clade 2B which had been on a co-circulating sub-clade (2C), and the new acquisitions included adjacent polymorphisms. On NA there were three consecutive polymorphisms acquired, which included a synonymous polymorphisms, which would have had less selection pressure. Thus, the background switching, which included sequential polymorphisms, strongly supported recombination.
Moreover, the fixing of H274Y was linked to the acquisition of receptor binding domain changes. For H274Y, A193T was the key acquisition, which was then accompanied by flanking changes at positions 187, 189, and 196, which were associated with fixing H274Y worldwide.
Similar associations were seen in the fixing of adamantine resistance in H3N2. That anti-viral marker (M2 S31N) as also associated with an HA change at position 193 (S193F) as well as a second RBD change, D225N.
There was concern that similar events in pandemic H1N1 would lead to H274Y jumping from seasonal H1N1 to pandemic H1N1, which would then pair up with one or more receptor binding domain changes. The first example of H274Yacquisition which was not linked to Tamiflu usage was in an isolate that had H274Y and D225E. This H274Y linkage with a position 225 change increased concerns for fixing of H274Y in pandemic H1N1.
This concern was increased further in the fall of 2009 when sequences were released from fatal cases in Sao Paulo Brazil. Two sequences from lung or throat necropsy tissues had D225N, while two other sequences from lung had D225G. Sequences had also been released from a severe case in Zhejiang, China and it also had D225G, but was on a different genetic background. Similarly Australian sequences with D225G and D225N were on a third genetic background. The appearance of D225G and D225N on different genetic backgrounds led to the prediction that the same changes would be found in Ukraine, because both markers were “in play” and were associated with severe and fatal cases.
Consequently the involvement of D225G and D225N was predicted for the severe and fatal cases in Ukraine.
The team WHO sent collected representative samples, which were sent to the Mill Hill regional lab in London and subsequently to another WHO regional lab at the CDC in Atlanta.
On October 17, WHO released a Ukraine update, which state that “tests reveal no significant changes in the Pandemic (H1N1) virus” as noted above. However, the next day Mill Hill released 10 HA sequences from Ukraine. One was from an earlier collection in Kiev, while nine were from western Ukraine. Five were nasal washes from milder cases, while four were from throat or lung tissues from deceased cases. All four sample from deceased cases had the predicted D225G. Moreover, the D225G was linked to a "low reactor" clasification.
Others saw this receptor binding domain change in four of four fatal cases as significant because D225G was rare and only found in about 1% of public H1N1 HA sequences. Norway had generated sequences matching the western Ukraine sub-clade and found D225G in three. Two had died and one was severe, so Norway issued an alert.
WHO responded to the alert with a briefing note on October 20 by noting that D225G had been seen in previous mild and severe cases and characterized D225G as "spontaneous and sporadic and does not appear to spread", as quoted above.
The CDC then released five more HA sequences. Three were from the five mild cases sequenced by Mill Hill, but two were unique and appeared to be from fatal cases. Both had D225N, which had also been predicted. Now the fatal cases with D225G/N was up to six and although were from the same clade, there were difference that divided the sequences into sub-clades, and D225G/N was in multiple sub-clades, again pointing toward spread via recombination.
Mill Hill then released to more sequences from lung samples and both sequences had D225G and D225N, demonstrating that the polymorphisms were spreading and were not spontaneous.
However, after the first six fatal sequences had been made public, WHO issued a preliminary evaluation of D225G and although 26 of the 52 samples were with D225G were from fatal cases, WHO maintained that the clinical significance was unclear and maintained that the cluster in Ukraine was not a cluster, and stated that the isolates were sporadic, as noted above.
Subsequently, multiple labs in Russia released sequences from lung and throat samples from fatal cases and they too had D225G and D225N. Several of the sequences were from the same general sub-clade, but small difference required multiple introductions, which further discounted the WHO working hypothesis that the changes were random, spontaneous, sporadic, and didn’t transmit.
WHO then e-mailed the same report via its WER, even though the Russian sequences invalidated virtually all of the major points in the report.
The WER was e-mail on a Friday, and Mill Hill released 28 samples from autopsy lung on the following Monday. 21 of the samples had D225G or D225N, and 10 of the samples had both. Thus, the linkage with fatal cases was higher than the frequencies cited by WHO, and the samples were clustered in time, space, and phylogeny. More independent introductions were required to explain the data which included identical genetic backbones with a wild type RBD or one with D225G, D225N, or both added. Similarly both polymorphisms were on branches with wild type sequences, signaling even more independent introductions, which was not consistent with the WHO working hypothesis and fully supported movement of these polymorphism by recombination.
Thus, in spite of the overwhelming data discounting WHO’s working hypothesis, they have not made any new comment since their statement last year.
This silence and linkage to the unsupported paradigm of random mutations continues to be hazardous to the world’s health.
Media Links
Recombinomics
Presentations
Recombinomics
Publications
Recombinomics
Paper
at Nature Precedings