H5N1 Human Avian Sequence Match in Indonesia is Unlikely
Recombinomics Commentary
August 12, 2006
After hoarding for months the genetic blueprints of the viruses isolated from both poultry and people, Indonesian officials have done an about-face and are sharing a large number of both avian and human viral isolates.
A scientist with the UN's Food and Agriculture Organization confirmed Thursday that Indonesia recently sent 91 avian viruses to the Geelong Laboratory in Australia, a reference lab for the FAO.
"We managed to get 91 isolates out to Geelong about two weeks ago. Took us months to get it to happen. But they arrived I think about 10 days ago or something like that - in good condition," Dr. Peter Roeder, an animal health officer with the Rome-based FAO, said in an interview.
"It will make a tremendous difference in the interpretation of the sequence data of the human viruses."
Peiris said the viruses isolated from human cases showed no evidence of having swapped genes - the process is called reassortment - with flu viruses from humans, pigs or other mammals. Reassortment is one of the ways an avian flu virus could acquire the ability to spread easily to and among people, a development that would lead to a flu pandemic.
"There isn't anything (evident in the human isolates) that rings particular alarm bells as such," Peiris said.
But he suggested the viruses that infected people in Indonesia are only one part of the picture of what is going on there.
"It is probably important to have more avian virus sequences from the same geographical areas where the human cases are coming from," he said from Hong Kong.
The above comments provide additional detail about H5N1 bird flu in Indonesia. Although there is no evidence of H5N1 reassortment with human or swine genes, H5N1 has been evolving via incremental changes in the gene sequences. There have been no reports of reassortment between H5N1 and mammalian genes in any country.
Since reassortment swaps full genes, reassort with H5 would not generate a pandemic strain. The WHO definition of a pandemic strain includes a new sero-type. Mammalian flu in human or swine involve H1 and H3, so swapping either gene with H5 would not produce a new H ser-type and most people have immunity to both H1 and H3, so the new virus would not be very virulent. Swapping of internal gene could theoretically increase replication or transmission, but such tests by the CDC involving H5N1 and H3N2 indicated virus with 8 avian or 8 human genes was more efficient at replication or transmission, suggesting that avian human reassortants would not be evolutionarily fit and not selected in a host with a dual infection.
However, dual infections can also lead to recombination, and there are many examples of H5N1 acquisition of mammalian polymorphisms. These acquisitions of human / swine signatures has happened in Vietnam and Thailand as well as Indonesia.
Many of the human specific changes in H5N1 in Indonesia are found in migratory birds and mammals, indicating the human H5N1 is evolving incrementally via recombination. These polymorphisms point away from a poultry source. The 91 samples mentioned above are recent avian samples from Indonesia, but the first six sequences have already been made public, and they do not match the human sequences. Moreover, it is unlikely that the additional samples being sequenced will match the majority of human H5N1 sequences in Indonesia.
These human sequences have a novel cleavage site, which was seen in the first released sequence in March of 2006. The HA sequence was from the first confirmed case in Indonesia. Although the sequence was deposited by Hong Kong on July29, 2005 and the CDC on August 1, 2005, he sequence was not made public until March 25, 2006. Analysis of the sequence quickly identified a novel cleavage site, RESRRKKR, which was not present in any avian sequence from Indonesia, or any sequence at Genbank. However, the HA sequence also had a number of additional changes that were not in Indonesian poultry isolates. These differences were easily seen in phylogenetic trees, which indicated that the isolates with the novel cleavage site form a separate branch that did not include any avian sequences. Last week the sequences were released, and it was clear that all 8 gene segments had polymorphisms that were not present in the Indonesian poultry isolates. The lack of a match was discussed at a recent meeting in Jakarta, and the samples that were sent to Australia were designed to update the avian sequences. Most of the public sequences are from samples collected between 2003 and 2005 and since the first human case was in July, 2005, the number of avian sequences that overlapped in time and location was limited.
However, a careful analysis of the existing data strongly suggests that the more recent avian sequences will also not match the human sequences.
The first human sequence was collected in July, 2005 and have the novel sequences. The second confirmed human case was isolated almost two months later. It had the more common cleavage site, RERRRKKR, and lacked the unique sequences in the index case. However, the second sequence was clearly linked to avian sequences, even though all were collected prior to the second case. The following human cases in 2005, matched the index case and failed to match the avian isolates. In May of 2006 there was a large cluster of human cases in the Karo Regency in north Sumatra. Those sequences also had the wild type HPAI cleavage site and also were missing the unique sequences in the index case, However, even though there were no 2006 avian sequences, linking of the Karo cluster to poultry was much more obvious than the other human cases. As new human isolates were sequenced, the number of differences also increased. The 2006 isolates had additional mammalian specific changes. Some of these specificities may be due to a lack of 2006 avian sequences, but these recent isolates also have the unique changes in the 2005 human isolate and which will do not match the avian sequences.
When the human sequences were release, the CDC also released two avian sequence from October, 2005. These sequences did not match the human sequences with the novel cleavage site. Similarly, the first six sequences from Australian also do not match, although they provide additional evidence linking the Karo cluster to an avian source.
The novel human sequences have been stable. Although new sequences have been added, the mid-2005 changes remain and are present in all human sequences with the novel cleavage site and no avian sequences from Indonesia. This strong segregation indicates the human sequences, which have been detected across a wide geographical range on Java, come from a separate, non-poultry reservoir.
Testing of non-poultry source has been very limited. Human isolates in Indonesia have a number of polymorphisms found in wild birds, yet there are ni wild bird H5N1 sequences from Indonesia. Similarly, there is only one cat isolate, and that sequence matches the human sequences with the novel cleavage site. Although there have bee reports of H5N1 in swine in Indonesia, sequences presented at scientific meetings do not match, but testing has been limited.
The human H5N1 is evolving away from the poultry isolates and more samples from a variety of sources should be sequenced, Recent media reports suggest new human samples may not be sent to Hong Kong or Atlanta for sequencing because WHO now accepts local lab confirmation as evidence of H5N1 infection. However, the local labs in Indonesia have not provided any sequence data, which is critical for monitoring the genetic changes in H5N1 in Indonesia.
Indonesia should continue sending out samples for sequencing ,and those sequences should placed in public databases as soon as possible.
Media Sources
Phylogenetic Trees