Rapid evolution of low-pathogenic H9N2 avian influenza viruses following poultry vaccination programmes

HA and NA surface genes.

The HA genes of the 18 H9N2 isolates were sequenced and analysed by the neighbour-joining method (Perrière & Gouy, 1996). Phylograms were constructed together with selected strains available in GenBank, including H5 and H9 viruses currently circulating in Asia. Phylogenetic analysis revealed that all HA genes of the H9N2 viruses clustered under the Ck/Korea/04116/04-like lineage (95.7–99.2 % nucleotide identity), the predominant H9N2 virus found among chicken isolates of LPMs since 2005 (Moon et al., 2010), rather than to the reference vaccine strain (Ck/Korea/01310/01, 91.2–94.2 % nucleotide identities) (Fig. 1a⇓). Noticeably, though, the two 2008 viruses (Dk/Korea/CBU08107/08 and Ck/Korea/SH0802/08) appeared to be more closely related to Ck/Korea/04116/04 (∼98 % nucleotide identity), whilst the most recent isolates were more phylogenetically identical to Ck/Korea/LPM77/06 than to Ck/Korea/04116/04 (98 versus 96.8 % nucleotide identity). The phylogeny of the NA genes showed at least three different lineages. The majority of the N2 genes (n=11) were phylogenetically related to a wild-bird H9N2 virus (Ab/Korea/W03/05-like lineage, 92.3–96.2 % nucleotide identity), whilst five isolates (Ck/Korea/SH0903/09, Dk/Korea/SH0904/09, Ck/Korea/SH0905/09, Ck/Korea/SH0910/09 and Ck/Korea/HC09/09) were grouped together with Korean H3N2 viruses (Dk/Korea/LPM17/05-like lineage, 91.1–94.2 % nucleotide identity). Only Dk/Korea/CBU08107/08 and Ck/Korea/SH0802/08 clustered closely with previous Korean H9N2 viruses under the Ck/Korea/04116/04-like lineage, as was observed for their HA genes (95.7–98.6 % nucleotide identity) (Fig. 1b⇓). These data demonstrated that the donor viruses for the H9 HA genes of the viruses from poultry appeared to have come exclusively from previous Korean Ck/Korea/04116/04-like viruses, whereas the N2 NA genes appeared to have been derived from three different AIV sources.

Internal genes.

Phylogenetic analysis of the internal genes of H9N2 isolates showed that they had the highest identity with AIV isolates circulating recently in South-east Asia, primarily China and Japan. In general, all internal gene trees reflected the Eurasian avian-like lineage with no apparent gene segment contributions from the North American avian lineage.

Alignment based on the NS genes indicated that all H9N2 isolates tested in this study belonged to allelic group A but could be segregated further into two subgroups: seven isolates, including the vaccine strain Ck/Korea/01310/01, appeared to be descendants of Ck/Korea/25232-96006/96 (H9N2)-like viruses and clustered with Ck/Korea/04116/04-like viruses, whereas 11 isolates appeared to be more closely related to Ab/Korea/W03/05-like (H9N2) viruses circulating among wild migratory birds in the same region (Fig. 1c⇑). The M genes could also be divided into two major sublineages. Nine isolates clustered with Ck/Korea/04116/04-like Korean poultry isolates, which included the vaccine strain. The remaining viruses (n=9) appeared to be more closely related to Dk/Hong Kong/Y439/97 (Fig. 1d⇑). Interestingly, only in the NP gene phylogeny were all isolates in the present study nested together as a homogeneous group with previous Korean Ck/Korea/04116/04-like viruses (Fig. 1e⇑).

Three major lines of descent could be observed in the PA genes: an Ab/Korea/W03/05-like wild-bird H9N2 lineage (n=5), a Dk/Hong Kong/Y439/97-like aquatic-bird H9N2 lineage (n=7) and a Ck/Korea/04116/04-like Korean domestic chicken H9N2 lineage (n=6) (Fig. 1f⇑). In contrast, the PB1 and PB2 gene segments could only be divided into two major lineages. Whilst seven of the 18 H9N2 viruses (Dk/Korea/CBU08107/08, Ck/Korea/SH0802/08, Ck/Korea/HC09/09, Ck/Korea/SH0903/09, Dk/Korea/SH0904/09, Ck/Korea/SH0905/09 and Ck/Korea/SH0910/09) clustered consistently with the Ck/Korea/04116/04 virus, the rest (n=11) appear to be more closely related with wild-bird strains, most notably H5 or H9 viruses from Japan and China (Fig. 1g, h⇑). Taken together, the phylogenetic analyses indicated that Korean H9N2 AIVs are continuously evolving in Korean domestic birds through multiple, independent reassortment events with previously reported H9N2 and H3N2 viruses in domestic poultry and migratory wild birds.

Replication and transmission in naïve domestic poultry species.

Antigenic analyses of recent H9N2 viruses indicated that, since 2007, significant antigenic variability had accompanied genetic evolution of the H9N2 viruses studied. Therefore, to investigate the replication and transmission potentials of the newly emerged H9N2 isolates, groups of 5-week-old specific-pathogen-free (SPF) chickens were inoculated intranasally and intratracheally with 400 μl of representative viruses [10⁶ 50 % egg infectious doses (EID₅₀) ml⁻¹] from each genotype and compared with infection by the vaccine strain (Ck/Korea/01310/01). Despite no apparent clinical signs of influenza-like illness, confirming their low-pathogenic nature, all of the inoculated viruses could be detected from tracheal swabs in modest titres (2.3–4.3 log₁₀ EID₅₀ ml⁻¹) at 3 days post-inoculation (days p.i.) (Table 4⇓). However, representative viruses of genotypes B and D could no longer be detected at 5 days p.i., whilst none of the tested viruses could be isolated at 7 days p.i.

In the tissues collected from two infected chickens at 5 days p.i., the tested viruses were detected only in the lungs and caecum of infected chickens at mean virus titres of 1.3–2.7 log₁₀ EID₅₀ (g collected tissue)⁻¹ (Table 4⇑). None of the viruses could be detected in the other tissue samples harvested such as kidney and spleen (data not shown). As expected for chicken-adapted H9N2 viruses, all seronegative contact chickens, co-housed with the infected birds after day 1 p.i., were positive for virus transmission by direct contact at 3 days post-contact (p.c.). These data indicated that the novel genotypes of H9N2 viruses remained well adapted for replication and transmission in chickens.

The phylogeny of the NA genes showed that some viruses possessed the N2 segments of H3N2 viruses (genotypes D and E) circulating predominantly among ducks in Korean LPMs (Song et al., 2008), whilst many internal segments were closely related to wild migratory bird isolates (mainly migratory ducks). These results led us to speculate that reassortment might be occurring in naïve ducks and thus may have played a role in the genesis of the recent novel genotypes of H9N2 viruses. We therefore determined the replication and transmission kinetics of the above selected representative viruses in groups of seven 5-week-old serologically naïve ducks, which were intranasally and intratracheally administered with 400 μl virus inoculum (10⁶ EID₅₀ ml⁻¹). Except for the genotype A isolate and Ck/Korea/01310/01(vaccine strain), all of the viruses could be detected from tracheal and cloacal swabs at modest titres (1.3–3.3 and 1.7–2.7 log₁₀ EID₅₀ ml⁻¹, respectively) at 2 days p.i. (Table 4⇑). Noticeably, H9N2 virus genotypes B, E and F could be detected in both specimens up to 6 days p.i. with comparable virus titres. In contrast, genotype C and D viruses were detected only in cloacal swabs until 4 days p.i. Interestingly, genotype A and Ck/Korea/01310/01 viruses were detected only in tracheal swabs at relatively low titres, and no virus titres were obtained from cloacal swabs (Table 4⇑). In transmission studies, both of the contact birds of the genotype B and F viruses were already positive for virus detection at 2 days p.c., whereas only one contact tested positive for genotypes C, D and E. No evidence of virus transmission by direct contact was observed in the genotype A group. These data clearly demonstrated that some of the novel H9N2 viruses, particularly representative genotype B and F isolates, were considerably better adapted for replication in ducks than the genotype A viruses. Besides phylogenetic analysis, these enhanced replication properties of recent H9N2 viruses in ducks suggested that they might have acted as an intermediate in the generation of novel of H9N2 virus genotypes. Seroconversion in post-infection sera evaluated by haemagglutination inhibition (HI) assays confirmed infection of the tested domestic poultry hosts (data not shown).

Replication of viruses in vaccinated chickens.

Genetic and antigenic analyses of recent H9N2 viruses indicated divergence from H9N2 viruses circulating prior to 2007, and most of the viruses replicated well in chickens and ducks by experimental infection (Tables 2⇑–4⇑). We further examined whether these viruses had the ability to replicate even in chickens that were seropositive as a result of vaccine administration. The representative viruses were intranasally and intratracheally inoculated into groups (n=6) of 8-week-old Ck/Korea/01310/01 (H9N2)-vaccinated chickens (HI titres: 80–160). After day 1 p.i., two seronegative contact birds were added to each group of experimentally inoculated chickens. Surprisingly, tracheal and cloacal swabs collected at 3, 5 and 7 days p.i. for virus titration in 10-day-old embryonated chicken eggs showed that almost all of the viruses could be recovered (highest swab titre: 3.3 log₁₀ EID₅₀ ml⁻¹) at 3 days p.i. with the exception of Ck/Korea/01310/01, the vaccine strain (Table 5⇓). However, Dk/Korea/CBU08107/08 (genotype A) was barely detectable in cloacal swabs in this sampling period. Ck/Korea/SH0902/09 (genotype C), Ck/Korea/HC09/09 (genotype E) and Ck/Korea/SH0912/09 (genotype F) were able to persist for up to 5 days p.i. (at titres of 1.7, 1.3 and 2.3 log₁₀ EID₅₀ ml⁻¹ in tracheal and 1.3, 0.7 and 1.3 log₁₀ EID₅₀ ml⁻¹ in cloacal swabs, respectively), whilst Ck/Korea/SH0906/09 (genotype B) and Ck/Korea/SH0903/09 (genotype D) could only be detected up to 3 days p.i. (at titres of 2.7/2.3 and 1.7/0.7 log₁₀ EID₅₀ ml⁻¹ in tracheal/cloacal swabs, respectively). Furthermore, tracheal swabs from contact birds sampled at 2 days p.c. were already positive for virus detection, except for the genotype A virus. These results demonstrated that these novel genotypes of H9N2 viruses may have the ability to replicate even in H9N2-vaccinated chickens, indicating a lack of efficacy of the current commercial vaccine in protecting immunized hosts from infection with these newly emerged H9N2 viruses, including direct transmission to naïve contact hosts (through respiratory droplets and/or the faecal–oral route).

Virus sampling and isolation.

A total of 615 faecal or cloacal samples from ducks and chickens were collected from various slaughterhouses, where poultry products from most local commercial farms and large-scale backyard farms in Chungcheong province (midwest region of South Korea) are brought for slaughter and processed to be sold as meat birds, from January 2008 to December 2009 (Table 1⇑). In addition, one virus (Dk/Korea/CBU08107/08) was specifically isolated from a commercial poultry farm, whilst two others (Ck/Korea/HC09/09 and Ck/Korea/HC0410/09) were obtained from two different backyard poultry farms. Most samples were collected from non-vaccinated avian domestic species. The collected specimens were suspended in antibiotic solution (200 U penicillin ml⁻¹, 200 μg streptomycin ml⁻¹ and 68 μg amphotericin B ml⁻¹ in PBS) and centrifuged at 6000 g for 15 min at 4 °C. The supernatants were inoculated into 10-day-old embryonated chicken eggs and incubated at 37 °C for 48 h. Allantoic fluid was harvested and clarified by centrifugation at 6000 g for 10 min at 4 °C. The presence of virus was detected by haemagglutination assays performed according to WHO/World Organization for Animal Health recommendations. The allantoic fluid from positive samples was harvested, aliquotted and stored at −80 °C until use. Subtyping was carried out by multiplex RT-PCR assays and confirmed by sequencing as described previously (Chang et al., 2008).

Genetic and phylogenetic analyses.

Viral RNA was extracted from the allantoic fluid of inoculated eggs using an RNeasy Mini kit (Qiagen) according to the manufacturer's instructions. The viral RNA was reverse transcribed at 37 °C for 60 min using Superscript reverse transcriptase (Qiagen) to generate the corresponding cDNA. PCRs were carried out using influenza-specific primers (Hoffmann et al., 2001) and nTaq DNA polymerase (Enzynomics), according to the manufacturer's recommendations. Amplicons were purified using a GeneAll gel extraction kit and sent to Cosmo GeneTech (Seoul, Republic of Korea) for commercial sequencing with an ABI 373 XL DNA sequencer (Perkin-Elmer). DNA sequences were compiled and analysed using Lasergene sequence analysis software package 5.0 (DNASTAR). Multiple sequence alignments were made using clustal v (Aiyar, 2000). Rooted phylograms were prepared using a neighbour-joining algorithm and plotted using NJPlot (Perrière & Gouy, 1996). Branch lengths were proportional to sequence divergence and could be measured relative to the scale bars provided in the figures.

Complete genome sequences of all eight genes of 18 selected viruses isolated from domestic poultry and slaughterhouses between January 2008 and October 2009 were compared with previous Korean wild-bird and domestic H9 and H3 virus isolates. In addition, full-genome sequences of corresponding AIVs isolated from neighbouring China, Hong Kong and Japan were also included in the phylogenetic analysis.

HI assay.

Polyclonal antibodies were obtained from 5-week-old SPF chickens that had been inoculated with Ck/Korea/01310/01 (vaccine strain of H9N2 in Korea from 2007) (Table 3⇑). HI assays were carried out as described elsewhere (Palmer et al., 1975). Briefly, sera were treated with receptor-destroying enzyme (RDE) from Vibrio cholerae (Denka Seiken) to inactivate non-specific inhibitors with a final serum dilution of 1 : 10. RDE-treated sera were serially diluted twofold and equal volumes of virus (8 haemagglutination units in 50 μl) were added to each well. The microplates were incubated at room temperature for 30 min, followed by the addition of 0.5 % chicken red blood cells. The plates were mixed gently and incubated at 37 °C for 30 min. The HI titre was determined by the reciprocal of the last dilution that contained chicken red blood cells with no agglutination. The limit of detection for the HI assays was set to ≤20 HI units.

Virus replication and transmission in domestic poultry species.

To investigate the replication potential of the newly emerged H9N2 genotype influenza viruses in SPF and seropositive chickens (HI titres: 80–160) against Ck/Korea/01310/01 (vaccine strain), representative viruses from each genotype were inoculated intratracheally and intranasally (400 μl) into groups of four 5-week-old SPF or vaccinated chickens (White Leghorn; CAvac) at a concentration of 10⁶ EID₅₀ ml⁻¹. To test for potential virus transmission by direct contact, infected chickens were co-housed with two to three contact chickens after day 1 p.i. and observed for 7 days. On days 3, 5 and 7 p.i., tracheal and cloacal swab samples were collected from all chickens and inoculated into 10-day-old embryonated chicken eggs. After 48 h incubation at 37 °C, the presence of the viruses was detected by a standard haemagglutination assay. Using sterile equipment, individual organ tissue samples, such as lung, caecal tonsil, kidney, caecum and spleen, were collected from infected hosts (two chickens) at 5 days p.i. for virus titration. Mean virus titres were expressed as log₁₀ EID₅₀ ml⁻¹ or log₁₀ EID₅₀ (g tissue sample)⁻¹ computed by the method of Reed & Muench (1938). The limit of virus detection was set at <0.7 log₁₀ EID₅₀ ml⁻¹ or log₁₀ EID₅₀ g⁻¹.

Groups of seven 5-week-old ducks (Anas platyrhynchos domesticus) that were found to be free of detectable influenza virus by serological testing were inoculated intranasally and intratracheally with 400 μl of representative viruses (10⁶ EID₅₀ ml⁻¹). Two uninoculated littermates were co-housed in the same isolator with the experimentally infected ducks to test for transmission. Swabs (cloaca and trachea) were collected at days 2, 4 and 6 p.i. for virus titration in 10-day-old embryonated chicken eggs.