Paradigm Shift Intervention Monitoring
twitter
Commentary
CDC Dual Use of
Inconsistent Flu Antigen Characterization Data
Recombinomics Commentary 13:50
March 27, 2012
CDC has confirmed that the influenza viruses isolated from the cluster of severe respiratory illness in one family in Maryland are seasonal influenza A H3N2 viruses. Genetic sequencing has confirmed that this is a typical human seasonal H3N2 virus that is more than 99% similar to other H3N2 influenza viruses submitted by the state of Maryland this season. While full antigenic testing is pending, based on genetic sequencing of some of the samples, these viruses are close to the H3N2 component of the 2011-2012 seasonal vaccine such that vaccination should offer protection against these viruses.
The above comments represent an Internet rumor (in red) and an official CDC statement (in blue) on the H3N2 sequences from the death cluster in Maryland, suggesting that the sequences represent a Perth/16-like sub-clade and recognized by the current influenza vaccine which uses Perth/16 as the H3N2 target.
However, sequences released by the CDC, including sequences from samples collected in February, strongly suggest that the sequences from Maryland are drift variants with limited reactivity with the current vaccine.
In the week 8 FluView, the CDC reported a large spike in low reactors (isolates producing a 4 fold or greater reduction in titer when compared to the current Perth/16 target). Although the CDC had released 22 H3N2 sequences from 2012, none were classified as low reactors. i.e A/PERTH/16/2009-LIKE (H3N2) LOW GP (although antigen characterization results were listed for 1/3 of the samples). However, shortly thereafter the CDC designated seven 2012 isolates as low reactors with the above designation. Phylogenentic analysis indicated that 5 of the 7 formed a drift variant sub-clade with a number of non-synonymous HA changes, include two that were in or near the receptor binding domain, A198S and V223I. These changes were also present in the newly selected H3N1 vaccine target, A/Victoria/361/2011.
However, the other two low reactors were in another drift variant, which also had a change near the receptor binding domain, S199A, which was in a candidate vaccine target, A/Brisbane/299/2011.
Thus, the seven isolates designated by the CDC as low reactors fell into two distinct sub-clades, and vaccine targets representing each sub-clade had been selected by the CDC, (appropriate constructs had been generated, with sequences released at GISAID). Moreover, the CDC released additional 2012 sequences and virtiually all fell into these two sub-clades, indicating Perth/16-like sequences were no longer circulating in the United States in 2012 (which was alos true for the entire northern hemisphere), and the current H3N2 vaccine component would have limited utility.
The recently released sequences by the CDC increased the number of United States H3 sequences from February, 2012 collections to 11, and the number of January, 2012 H3 sequences to 37. Although none of the 2012 H3 sequences were from Maryland, virtually all fall into the two drift variant sub-clades listed above, with the Victoria/361-like sub-clade dominating, including a February 15 isolate from Delaware, A/Delaware/02/2012. Similarly, the only Maryland H3 sequences from the 2011/2012 season, collected on December 26, A/Maryland/21/2011, also has A198S and V223I and is in the Victoria/361-like sub-clade, although it is characterized as Perth/16 like (A/PERTH/16/2009-LIKE (H3N2) GP), which is also true of a subset of the 2012 sequences which are in the Victoria/361 sub-clade.
This inconsistency is clearly demonstrated within the Victoria/361 sub-clade, as seen in phylogenetic analysis. A/South Carolina/03/2012 (SC/03/12) is designated as a low reactor, yet two isolates which evolved from SC/03/12 (i.e. have all of the synonymous and non-synonymous changes seen in the H3 sequence from SC/03/12), A/South Carolina/01/2012 and A/Delaware/02/2012 are characterized a Perth/16-like. Similarly sequences (A/Minnesota/04/2012), which have evolved from A/South Dakota/02/2012 (designated as a low reactor) are also designated as Perth/16-like. Thus, both February 15 isolates (the most recent public sequences) have evolved directly from sequences designated as low reactors by the CDC, yet are designated as Perth-like, allowing the CDC to use related sequences to justify a new H3 vaccine target, because of a limited cross reactivity with the current target, and also claim that the related sequences are similar to the current target, as indicated in the above CDC statement and internet rumor.
This dual use of antigen characterization data is similar to the data fro H1N1 in the 2007/2008 season, when the vaccine target was Solomon Island/03, which was no longer circulating in the United States. Although Solomon Island/03 had been replaced by Brisbane/59 or Hong Kong/2562, these isolates were called Solomon Island/03-like at the beginning of the 2007/2008 season by the CDC, based on an insensitive antisera used by the CDC. Subsequent antisera developed by Ausatralia demonstrated significant antigenic difference between the three sub-clades, and that assay was used to identify Brisbane/59 isolates, which was the target for the 2008/2009 season. Thus, two different results on Brisbane/59-like sequences were used to claim a vaccine match with Solomon Island/03, as well as justify a change in target from Solomon Island/03 to Brisbane/59.
These internal inconsistencies, which provide evidence for a vaccine match as well as a need to change vaccine targets, are once again used to claim that the Maryland death cluster sequences, which are almost certainly drift variants of Victoria/361 or Brisbane/299 (and likely to be Victoria/361), are not drift variants and likely to react with vaccine against Perth/16, which has been replaced in 2012 in the United States by the two drift variants cited above.
Recombinomics
Presentations
Recombinomics
Publications
Recombinomics
Paper
at Nature Precedings