Influenza A (H5N1) Viruses from Pigs, Indonesia

TOC summary: Pigs may serve as intermediate hosts in which this avian virus can adapt to mammals.

Pigs have long been considered potential intermediate hosts in which avian infl uenza viruses can adapt to humans. To determine whether this potential exists for pigs in Indonesia, we conducted surveillance during 2005-2009. We found that 52 pigs in 4 provinces were infected during 2005-2007 but not [2008][2009]. Phylogenetic analysis showed that the viruses had been introduced into the pig population in Indonesia on at least 3 occasions. One isolate had acquired the ability to recognize a human-type receptor. No infected pig had infl uenza-like symptoms, indicating that infl uenza A (H5N1) viruses can replicate undetected for prolonged periods, facilitating avian virus adaptation to mammalian hosts. Our data suggest that pigs are at risk for infection during outbreaks of infl uenza virus A (H5N1) and can serve as intermediate hosts in which this avian virus can adapt to mammals.
A highly pathogenic avian infl uenza virus A (H5N1) was fi rst recognized among geese in Guangdong Province, southern People's Republic of China, in 1996 (1). Within a year, this goose virus underwent reassortment with viruses circulating in other avian species. By 1997, the virus had become widespread among poultry in Hong Kong, and direct avian-to-human transmission of infl uenza A (H5N1) viruses was reported (2,3). Since late 2003, infl uenza A (H5N1) viruses have spread to domestic poultry in other Southeast Asian countries (4). Since mid 2005, they have been detected across Asia, Europe, and Africa, causing severe damage to the poultry industry and infecting >490 humans, resulting in a mortality rate of 60% (5)(6)(7)(8). Indonesia has been particularly affected by these viruses; >160 cases of human infection (i.e., about one third of the total confi rmed human infl uenza A (H5N1) infections worldwide) and a mortality rate >80% have been reported (8). Hence, understanding prevalence and adaptation of infl uenza A (H5N1) infl uenza viruses in Indonesia is crucial.
Infl uenza viruses attach to host cells by binding their hemagglutinin (HA) to cell-surface oligosaccharides containing a terminal sialic acid. The HA of avian infl uenza viruses preferentially binds to sialic acid linked to galactose by α-2,3 linkages (SAα2,3Gal); that of human viruses binds to SAα2,6Gal (9). Correspondingly, epithelial cells in the upper respiratory tracts of humans mainly bear SAα2,6Gal receptors (10,11), and those in duck intestines (the major replication site for duck viruses) mainly possess SAα2,3Gal (12). Virus receptor specifi cities and expression patterns of receptors on host cells are thought to be major determinants of the host range restriction of infl uenza viruses (13). Thus, the recognition of human-type receptors by avian viruses appears to be necessary for these viruses to replicate in the upper respiratory tract and be transmitted effi ciently from human to human. Given that infl uenza A (H5N1) viruses isolated from humans are not transmitted effi ciently despite their ability to recognize human-type receptors (14), mutations in the polymerase and other viral genes may also be needed for replication of infl uenza A (H5N1) viruses in the upper respiratory tract (15).
Traditionally, pigs have been considered as "mixing vessels" (16)(17)(18)(19) because they support replication of avian Infl uenza A (H5N1) Viruses from Pigs, Indonesia and human infl uenza viruses (17). Their tracheal epithelial cells reportedly bear SAα2,3Gal and SAα2,6Gal receptors (18). However, recent studies have shown that despite SAα2,3Gal and SAα2,6Gal receptors in pig respiratory tracts, SAα2,3Gal is found only in the smaller airways (bronchioli and alveoli) and not in the trachea (20,21). Kuchipudi et al. (22) found SAα2,3Gal and SAα2,6Gal receptors in the bronchi, bronchioli, and alveoli of chickens and ducks; however, SAα2,6Gal was dominant in chicken tracheal epithelium, and SAα2,3Gal, in duck trachea. Given that infl uenza A (H5N1) viruses have been transmitted directly from birds to humans, the central dogma of pigs as a mixing vessel may no longer stand. Moreover, under experimental conditions, pig susceptibility to infection with avian infl uenza A (H5N1) viruses is low (23). Nevertheless, the pandemic (H1N1) 2009 virus is a reassortant that originated from 4 genetically distinct viruses and appeared to be generated in pigs (24), suggesting their role in the generation of pandemic infl uenza viruses. Infection of pigs with infl uenza A (H5N1) viruses has been reported in Vietnam (25) and China (26); however, the infection status of pigs in Indonesia remains unknown. We, therefore, explored whether pigs in Indonesia had been infected with infl uenza A (H5N1) viruses and, if so, whether the viruses were transmitted multiple times and had acquired the ability to recognize human-type receptors.

Specimen Collection
Virologic and serologic surveillance was conducted during 3 rainy seasons during 2005-2009: January-February 2005, October-February 2007, and, November 2008-April 2009. Nasal, fecal, and serum samples were collected from apparently healthy pigs in various districts of Indonesia ( Figure 1). The nasal and fecal samples were injected into 10-day-old embryonated eggs, and the allantoic fl uid was tested for hemagglutination. Hemagglutination-positive allantoic fl uid was subjected to reverse transcription-PCR by using H5 HA-specifi c and N1 neuraminidasespecifi c primers; only positive samples were tested further.
Serum was analyzed to estimate the prevalence of infl uenza virus A (H5N1) infection.

Serologic Analysis
Swine serum samples were tested for neutralizing antibodies against infl uenza A/swine/East Java/UT6040/2007 (H5N1) and A/duck/Czechoslovakia/56 (H4N6) viruses. Subtype H4N6 was used as a negative control. Serum was mixed with 3 volumes of receptor-destroying enzyme (Denka Seiken Co., Ltd, Tokyo, Japan) overnight at 37°C and inactivated at 56°C for 30 min. A 2-fold serial dilution series of serum (1:4-1:512) was mixed with an equal volume of infl uenza virus at 100 TCID 50 (50% tissue culture infectious doses) and incubated at 37°C for 30 min. Viruses were inoculated to monolayers of MDCK cells for 1 h, washed 2×, and incubated with MEM containing 0.3% BSA for 2 d at 37°C in a 5% CO 2 incubator. Cytopathic effects were observed to determine the neutralizing activity of the test serum. The detection limit for the neutralizing antibody was <4 dilutions of serum.

Sequence Analysis
To characterize the swine infl uenza A (H5N1) viruses isolated in Indonesia, we sequenced the HA genes of 39 viruses isolated from pigs in Banten, East Java, North Sumatra, and South Kalimantan provinces and grouped them according to their genetic similarities. Viral RNA was extracted with ISOGEN (Nippon Gene, Tokyo, Japan) ac- cording to the manufacturer's instructions. Extracted RNA was reverse transcribed with SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA, USA) and an oligonucleotide complementary to the 12-nt sequence at the 3′ end of the viral RNA and amplifi ed by PCR with Pfu-ultra (Stratagene, La Jolla, CA, USA) or Phusion (Finnzymes, Espoo, Finland) high-fi delity DNA polymerase and primers specifi c for each segment of the infl uenza virus A (H5N1). Primer sequences are available upon request. The PCR products were separated by agarose gel electrophoresis, purifi ed by using a MinElute Gel Extraction Kit (QIAGEN, Hilden, Germany), and then sequenced. The nucleotide sequences obtained in this study are available from GenBank, accession nos. HM440051-HM440154.

Phylogenetic Analysis
We phylogenetically analyzed 13 representative swine infl uenza A (H5N1) viruses for all 8 viral genes and compared these sequences with publicly available sequences. All sequences were assembled and edited with BioEdit 7 software (28). Neighbor-joining tree analysis was conducted by using ClustalW (www.clustal.org). Estimates of the phylogenies were calculated by performing 100 neighborjoining bootstrap replicates.

Receptor Specifi city Assays
During replication in pigs, avian infl uenza viruses may adapt to recognize human-type receptors because such receptors are present in the epithelial cells of pig trachea (18). We therefore analyzed the receptor specifi city of representative viruses from each of the 3 swine groups: A/swine/ Banten/UT3081/2005 for the 2005 swine group, A/swine/ East Java/UT6012/2007 for the 2006-07 swine (A) group, and A/swine/Banten/UT6001/2006 for the 2006-07 swine (B) group. We also analyzed A/swine/Banten/UT3062/2005 clone 6 and A/swine/Banten/3063/05 clone 1, each of which possesses a single amino acid change in HA that distinguishes it from other clones. The receptor specifi city of these infl uenza A (H5N1) viruses was determined by use of an assay that measures direct binding to sialylglycopolymers possessing either SAα2,3Gal or SAα2,6Gal. We used this solid-phase binding assay with the sodium salts of sialylglycopolymers (poly α-L-glutamic acid backbones containing N-acetylneuraminic acid linked to galactose through either an α-2,3 or -2,6 bond (Neu5Acα2,3Galβ1,4GlcNAcβ-pAP and Neu5Aαc2,6Galβ1,4GlcNAcβ-pAP) as described (29,30). Briefl y, microtiter plates (Polystyrene Universal-BIND Microplate, Corning, NY, USA, USA) were incubated with glycopolymer in phosphate buffered saline (PBS) at 4°C for 3 h and then irradiated under UV light at 254 nm for 2 min. After removal of the glycopolymer solution, the plates were blocked with 0.1 mL PBS containing 2% BSA (Invitrogen) at room temperature for 1 h. After being washed 5× with PBS, the plates were incubated in a solution containing infl uenza virus (128 hemagglutination units in PBS) at 4°C for 12 h. After 3 more washes with PBS, antibody to the virus was added to the plates, which were then incubated for 2 h at 4°C, washed 3× with ice cold PBS, and then incubated with horseradish peroxidase-conjugated protein A (Organon Teknika N.V; Cappel Products, Turnhout, Belgium; 2000-fold dilution in PBS) at 4°C. After being washed 4× with ice-cold PBS, the plates were then incubated with o-phenylenediamine (Sigma-Aldrich) in PBS containing 0.01% H 2 O 2 for 10 min at room temperature, and the reaction was stopped by adding 0.05 mL of 1N HCl. Absorbance was determined at 490 nm.

Virus Prevalence
Of 702 nasal swabs, 52 (7.4%) collected in 2005-2007 yielded infl uenza A (H5N1) viruses (Table 1; online Appendix Table); no virus was isolated from fecal samples of the same pigs. All 35 viruses isolated in 2005 were from 5 pig farms in the Tangerang District of Banten Province, near an area in which an infl uenza A (H5N1) outbreak among poultry had been confi rmed in 2004 (31) and where the virus has since remained enzootic. Samples collected from a slaughterhouse in the Surabaya district of East Java Province were negative for infl uenza virus A (H5N1). In the subsequent surveillance period, October 2006-February 2007, we detected viruses in pigs on 4 farms in the Tangerang, Kediri, and Medan districts of Banten, East Java, and North Sumatra, respectively, and in slaughterhouses in the Surabaya and Banjarmasin districts of East Java and South Kalimantan; all sites were near previous outbreaks of infl uenza virus A (H5N1) infection among poultry. Pigs from which these viruses were isolated did not show any signs of infl uenza-like illness at the time of sample collection. During the November 2008-April 2009 surveillance period, virus was not isolated from any nasal swabs from 300 pigs tested. However, the 300 serum samples tested indicated that 3 (1%) pigs had neutralizing antibodies against a subtype H5N1 virus but not subtype H4N6, suggesting limited exposure to infl uenza A (H5N1) viruses. These positive samples were obtained from a farm in the Malang District of East Java Province; neutralizing titers were 4-16 (Table 1).

Virus Sequences
Among the 39 viruses sequenced, the fi rst group comprised 24 isolates collected in Banten Province during January 2005-February 2005; the HA genes of these viruses were either identical or differed by no more than 2 nt. The second group comprised 9 isolates collected during October 2006-February 2007 and also differed by only 2 nt, al-though the viruses were collected in different provinces: Banten, East Java, North Sumatra, and South Kalimantan. The HA genes in these 2 groups differed from each other by 49-53 nt. The third group included 6 isolates collected in Banten, East Java, and North Sumatra during October 2006-February 2007; these HAs were identical except for 1 nt and differed from those of the fi rst and second groups by 42-45 and 58-61 nt, respectively. Thus, the swine infl uenza A (H5N1) viruses collected in our surveillance study could be classifi ed into 3 distinct groups on the basis of their HA gene sequences, irrespective of the province from which they were isolated, suggesting extensive movement of pigs among provinces.  swine (B) group. Analyses of the other 7 genes demonstrated that the phylogenetic relationships established for the HA gene were maintained; that is, the swine viruses in each group possessed nearly identical genes, and each group of swine viruses was most closely related to a chicken virus isolated near the site where the swine viruses were collected (Figure 2,

Receptor Specifi city
Sequence analysis of the PCR products of the HA genes of A/swine/Banten/UT3062/2005 and A/swine/Banten/UT3063/2005 indicated that nucleotides were heterogeneous at certain positions, prompting us to plaque purify the viruses in MDCK cells to obtain viral clones with distinct HA sequences (Table 2). We found that most of the swine infl uenza subtype H5N1 isolates bound to only SAα2,3Gal, whereas the plaque-purifi ed clone 6 of A/swine/Banten/ UT3062/05 bound to SAα2,3Gal and SAα2,6Gal ( Figure  3), indicating that during their replication in pigs, avian infl uenza A (H5N1) viruses can acquire the ability to recognize human virus receptors.

Discussion
In contrast to the few reported cases of infection of pigs with highly pathogenic avian infl uenza A (H5N1) viruses (17,25,26), our surveillance study of 7 provinces in Indonesia during 3 periods shows that 7.4% of pigs surveyed during 2005-2007, but not 2008-2009, were infected with infl uenza A (H5N1) viruses. Phylogenetic analysis indicated that the viruses were transmitted to pigs on several different occasions, probably from poultry on Because the phylogenetic relationships established for the HA gene extended to all viral genes, we conclude that the 3 groups of viruses identifi ed in this survey were likely established independently, suggesting at least 3 separate avian-to-pig episodes of transmission of infl uenza A (H5N1) viruses during 2005-2009 in Indonesia. Our fi ndings confi rm sporadic reports of the susceptibility of pigs to infl uenza A virus (H5N1) infection in natural (25,26) and experimental settings (23,34) and suggest that when an outbreak of infl uenza A virus (H5N1) infection occurs on poultry farms, pigs on nearby farms should be evaluated for infection.
We also found evidence of pig-to-pig transmission of infl uenza A virus (H5N1), particularly among animals sampled during the fi rst surveillance period. Many viruses possessing almost identical genes were isolated from pigs on the same farms (Tables 1, 2). Pig-to-pig transmission would likely prolong the duration of infl uenza A (H5N1) virus infection within a pig population, thereby increasing the likelihood of adaptation and the subsequent generation of infl uenza A (H5N1) viruses that replicate effi ciently in humans.
The lack of infl uenza-like signs in pigs infected with infl uenza A (H5N1) viruses has several public health implications. In Indonesia, pigs are transported to different locations according to market needs. This movement is refl ected in our fi nding that clusters of swine viruses collected after 2006 were not consistent with those common to the sampling region. Indeed, viruses collected in North Sumatra, South Kalimantan, East Java, and Banten provinces showed identical or nearly identical genes, indicative of extensive transport of infected pigs throughout Indonesia. Thus, pathogenic infl uenza A (H5N1) viruses could easily evade detection as they spread through Indonesia in asymptomatic pigs being transported from province to province.
Our analysis of viral receptor specifi cities showed that 1 plaque-purifi ed clone of A/swine/Banten/UT3062/2005 bound to avian-type and human-type receptors. Serine at position 134 was responsible for the human-type receptor recognition. This position is located within the 130-loop structural component of the receptor-binding pocket (35). Hence, the amino acid change at this position may affect receptor binding. Because serine at position 134 is never seen in avian infl uenza A (H5N1) viruses (alanine is highly conserved at this position in avian infl uenza A [H5N1] viruses), the Ala134Ser mutation probably occurred during adaptation of the virus to pigs. According to a previous report (36), human isolates possessing valine at this position could also bind to the human-type receptor, although a mutation at position 129 (L129V) was also required for the human-type receptor recognition in this strain. Therefore, mutations at position 134 probably correlated with human-type receptor recognition and may serve as molecular markers for assessing the pandemic potential of infl uenza virus A (H5N1) isolates.
Although infl uenza virus A (H5N1) infection was not reported among swine workers in Indonesia while we were collecting our pig specimens, a previous cohort study showed that such workers, as well as their unexposed spouses, had increased levels of antibody to swine infl uenza A (H1N1) viruses (37), suggesting that humans are indeed susceptible to swine-adapted viruses (38). The recent swine-origin pandemic (H1N1) 2009 further demonstrates that pigs can be a potential source of virus capable of causing a human infl uenza pandemic (24). These fi ndings suggest that as infl uenza A (H5N1) viruses spread among pigs and adapt to recognize human-type receptors, farmers, swine workers, and their families will be at greatest risk for infection by the newly adapted viruses.
In summary, we found that infl uenza A (H5N1) viruses have been transmitted multiple times to pig populations in Indonesia and that 1 virus has acquired the ability to recognize human-type receptors. Of particular concern is that pigs infected with infl uenza A (H5N1) viruses showed no signifi cant infl uenza-like signs and were likely transported to and from different provinces in Indonesia. On the basis of our fi ndings, we encourage the Indonesian government to control the transport of pigs within Indonesia. Otherwise, opportunities for this avian virus to adapt to mammals will increase, as will the risk for emergence of a new pandemic infl uenza virus. Dr Nidom is a director of the Faculty of Veterinary Medicine and the Collaborating Research Center-Emerging and Reemerging Infectious Diseases, Tropical Disease Centre, Airlangga University, Surabaya, Indonesia. His research interests include the pathobiology and epidemiology of poultry pathogens and zoonoses transmitted by poultry.