H5N1 Match Failures in Indonesia Raise Pandemic Concerns
Recombinomics Commentary
August 11, 2006
The decision of the government of Indonesia to release the H5N1 sequences at the WHO repository has yielded dramatic changes in the genetic analysis of the H5N1 in Indonesia. The announcement last Thursday was followed by the removal of the password protection on over 500 Indonesian sequences in the database. These new sequences provided additional data indicating the human H5N1 cases in Indonesia are linked to an animal reservoir that has yet to be identified.
The problem began to emerge on March 25, 2005 when the first human sequences were released. Although the first confirmed case was reported in July of 2005, the actual sequence data was sequestered. The first sequence was deposited on July 29, 2005 by the University of Hong Kong, followed by the same sequence from the CDC on August 1, 2005. However, the CDC sequence was released only after the United States had chosen the sequence as a pandemic vaccine target. Analysis of the sequence revealed a novel HA cleavage site as well as a new glycosylation. Although the sequences had many of the sequences that were specific to Indonesian poultry, the sequence did not match any of the know avian sequences and no public sequence had the novel cleavage site.
There were a modest number of avian sequences available from Indonesia and most were from 2003 and 2004. However, new sequences were made public earlier this year and although the new sequences had several isolates from 2005, there was still no sequence that matched the cleavage site. Word then began to leak out of Indonesia indicating that a cat sequence had the cleavage site and matched the human sequence. Phylogenetic trees showed that the growing number of human sequences formed a separate branch which contained no avian sequences.
In July the WHO held a meeting in Jakarta on H5N1 in Indonesia. In that meeting for sequence data was presented showing that virtually all of the human sequence had the novel cleavage site and additional changes in HA that created the separate branch. One sequence was on another branch with avian sequences and the Karo cluster was on another branch with avian sequences, but all of the other sequences formed a separate branch that include the cat sequences. The separate branch was markedly larger and was diverging further from the avian sequences. Other than the one sequence from August of 2005, all human sequences from Java, which included the Jakarta area isolates, were on the separate branch.
The release of the additional sequences provided data for the remaining seven gen segments. All seven indicated that the human sequences were distinct from the avian sequence. Moreover, many of the distinguishing sequences were found in wild birds or mammalian isolates, including human seasonal flu. These data point to an H5N1 reservoir other than domestic poultry and indicate these sequences are being acquired via recombination.
However, since the human sequences were isolate in the 12 months beginning in the middle of 2005, it was possible that there was a dramatic change in the more recent avian sequences that would account for the differences. The lack of recent avian isolates was discussed at the meeting and the shipment of 100 recent avian samples to the WHO affiliated lab in Australia was announced. These new samples were expected to address the lack of overlap in the collection dates.
The sequences released last Friday included two recent isolates from the CDC. The chicken samples were from October 2005 and although there were some matches that were also found in earlier 2005 isolates that were shared by the human isolates, the distinguishing polymorphism, including the HA cleavage site were not found.
Yesterday, more recent r=sequences from patients were released. These sequences showed a continuing evolution of the H5N1 away from the avian sequences. The newly released sequences included isolates from late June, early July of 2006. Yesterday the first six sequences from the samples sent to Australia were also released. These sequences were from 2005, but had several polymorphism in common with the late 2005 chicken sequences released by the CDC. However, the sequence from Australia failed to match the human sequences.
More sequences are expected from Australia and sequences from 2006 could provide the missing matches. However, the consistency of the defining polymorphisms in the human cases continues to point away from domestic poultry. The failure to match raises serious questions about surveillance in Indonesia and elsewhere.
Testing for H5N1 in patients in Indonesia is still tightly linked to exposure to dead or dying poultry. The lack of a match with H5N1 from poultry suggests more patients with bird flu symptoms should be tested for H5N1. Moreover, the large number of H1N1 seasonal flu suggests sequence analysis of those isolates is needed. Similarly, more testing of wild birds and mammals is required. H5N1 positive swine are frequently discussed, but no H5N1 sequences have been made public. Presentations at scientific meetings indicates that sequences from earlier isolates have H5N1 which does not have the novel HA cleavage site, but more testing, especially on more recent isolates is required. Similarly, testing of domestic mammals such as dogs and cats is necessary.
The release of the sequences was a good start. That data indicates that a much more aggressive screening is required to identify the source of the human infections. This aggressive surveillance in Indonesia should begin immediately.
Media Sources
Phylogenetic Trees