Mammalian H5N1 Evolution in Indonesia
Recombinomics Commentary
October 7, 2006
The recent announcements of H5N1 bird flu in cats in Indonesia, coupled with results from expanded sequencing of poultry strongly signal the existence of a separate mammalian H5N1 reservoir in Indonesia that is responsible for the vast majority of human cases.
This separate reservoir creates significant problems, because most of the attention has focused on infected poultry, and the mammalian reservoir has been significantly under investigated and under reported.
Therefore, a review of the evolution of this mammalian reservoir is useful. Although Indonesia did not acknowledge H5N1 infections until 2004, the first isolates were from birds in 2003. Sequence analysis of the 2003 and 2004 isolates indicated the H5N1 was Clade 2 and had a number of genetic markers that were specific for Indonesia.
The first human case was reported in July of 2005, and the sequence of the isolate, A/Indonesia/5/2005 had the Indonesia specific markers, but also had a number of unique polymorphisms, including one that created a novel HA cleavage site, RESRRKKR. However, the second human isolate, A/Indonesia/6/2005 had the more common HA cleavage site, RERRRKKR, and was similar to bird isolates. Subsequent human isolates in 2005 had the novel cleavage site, but several had an additional silent change, so although the cleavage site matched the first sequence at the protein level, there were a number of changes at the nucleotide level that divided the sequences with the novel cleavage site into two groups.
However, as the number of sequences increased in early 2006, it became increasingly clear that the human sequences were separating from the poultry sequences, all of which had the common bird cleavage site. The first match of the human sequences was from H5N1 from a throat swab of a cat in Indramayu near a residence were two siblings died from H5N1 infections. The sequence of the cat isolate not only matched the more recent human sequences, but was very close to the sequences of the isolates from the two siblings, A/Indonesia/283H/2006 and A/Indonesia/286H/2006, as well as other human isolates from Indramayu, A/Indonesia/292H/2006 and A/Indonesia/304H/2006. New isolates in 2006 collected from patients in East and West Java were sequenced and all were matches of more recent sequences such as the four human and one cat isolate from Indramayu.
In May of 2006 however, there was a new cluster in the Karo regency in north Sumatra. This outbreak was the largest to date and involve secondary and tertiary transmissions of H5N1. Consequently a meeting was call by WHO and consultants in Jakarta in June. Included in the presentation was a phylogenetic tree that summarized the H5N1 Indonesian isolates as of June 12, 2006.
The tree, which had the human sequences in green and the Karo cluster shaded in pink, clearly showed the match problem. All of the sequences with the novel cleavage site were on the lower portion of the tree and there were no poultry sequences on these two lower branches. Moreover, all of the recent human sequences from Java were on the lowest branch, which was even further from the poultry isolates. Thus, the human isolates were evolving away from the poultry isolates, suggesting the existence of a separate mammalian reservoir.
However, all of the human isolates were from July, 2005 or later, while most of the bird isolates were from earlier dates. Therefore 91 samples were schedule for shipment to Australia for virus isolation and sequencing. The samples were from infections between September, 2005 and March, 2006.
As sequences from these more recent and geographically dispersed isolates began to be published, it was becoming increasingly clear that the vast majority of the human infections on Java were not from domestic poultry. Each human sequence mapped to the lower portion of the tree and which was more distinct from the bird sequences.
The second set of new bird sequences included an isolate with the novel cleavage site. It was from a duck on Indramayu isolated in 2006. However, that isolate match the upper branch of the human sequences, which were composed of six isolates from three patients in 2005. Thus, although every human isolate in 2006 was matching the lower branch, the one duck sequence matched the upper branch.
The third set of poultry sequence had two matched with the lower branch. However, the two matches were from chickens in central Sumatra from 2005. Thus, none of the poultry isolates matched the lower human branch, while all human isolates, as well as the cat isolate, matched the lower human branch.
These data again supported a separate reservoir for the human sequences, and the only matches on Java were from the one cat, and all human isolates. The recent announcement indicates more H5N1 has been detected in cats, but the sequences of those isolates have not been released. Swine H5N1 sequences have been reported, but none match the human sequences.
The match failures pose a major problem because testing of humans is largely limited to patients how have been near dead or dying poultry. However, the poultry association has not been linked to the human infections, so an expanded testing of patients with symptoms is warranted. Similarly, more sequencing of H5N1 from other reservoirs is warranted by the match failures between mammalian and avian sequences.
Media Sources
Phylogenetic Trees