Phylogenetic and epidemiological characteristics of H9N2 avian influenza viruses in Shandong Province,China from 2019 to 2021

ZHAO Yi-ran ,ZHAO Yu-zhong ,LlU Si-dang ,XlAO Yi-hong ,Ll Ning ,LlU Kui-hao ,MENG Fan-liang,ZHAO Jun,LlU Meng-da,Ll Bao-quan#

1 College of Animal Science and Veterinary Medicine,Shandong Agricultural University/Shandong Provincial Key Laboratory of Animal Biotechnology and Disease Control and Prevention,Shandong Agricultural University,Tai’an 271018,P.R.China

2 Division of Zoonoses Surveillance,China Animal Health and Epidemiology Center,Qingdao 266033,P.R.China

Abstract H9N2 avian influenza virus (AIV) has widely circulated in poultry worldwide and sporadic infections in humans and mammals. During our surveillance of chicken from 2019 to 2021 in Shandong Province,China,we isolated 11 H9N2 AIVs. Phylogenetic analyses showed that the eight gene segments of the 11 isolates were closely related to several sublineages of Eurasian lineage: BJ/94-like clades (HA and NA genes),G1-like clades (PB2 and M genes),and SH/F/98-like clades (PB1,PA,NP and NS genes). The isolates showed mutation sites that preferentially bind to humanlike receptors (HA) and mammalian fitness sites (PB2,PB1 and PA),as well as mutations in antigen and drug resistance sites. Moreover,studies with mice revealed four isolates with varying levels of pathogenicity. The average antibody titer of the H9N2 AIVs was 8.60 log2. Based on our results,the epidemiological surveillance of H9N2 AIVs should be strengthened.

Keywords: influenza virus,phylogenetic analyses,mutation sites,pathogenicity,antibody titer,epidemiological surveillance

1.lntroduction

Avian influenza (AI) is a highly contagious viral respiratory disease caused by influenza A viruses of the family Orthomyxoviridae (Nodaet al.2006;Bahariet al.2015). Avian influenza viruses (AIVs) encoding at least 10 proteins: the two surface proteins (HA and NA),nucleoprotein (NP),three polymerase proteins (PB2,PB1 and PA),matrix protein (M1),ion channel protein(M2),nonstructural protein (NS1),and nuclear export protein (NS2) (Calderet al.2010;Xueet al.2014;Guet al.2017;Zhaoet al.2022). AIVs are further divided into subtypes based on the antigenic properties of HA and NA. At present,expect the H17N10 and H18N11 subtypes found only in bats,16 HA subtypes (H1-H16)and 9 NA subtypes (N1-N9) have been distinguished in avian species (Hinshawet al.1982;Kawaokaet al.1990;R?hmet al.1996;Wanget al.2022). In turn,such variety supports the wide spread of AIVs in different combinations among avian species.

H9N2 AIVs were first detected in 1966 (Hommeet al.1970),they have spread widely in poultry around the world. The poultry infected show no clinical illness or slight respiratory signs,which will be significantly worsening if infected with other pathogens (Kandeilet al.2014;Kandeilet al.2017). As phylogenetic analysis has indicated,H9N2 AIVs are divided into two primary lineages: North American and Eurasian (Websteret al.1992;Guoet al.2000). The Eurasian lineage is subdivided into various sublineages: BJ/94-like (the prototype virus is A/chicken/Beijing/1/1994(H9N2)) or Y280-like (the prototype virus is A/duck/Hong Kong/Y280/1997(H9N2),G1-like (the prototype virus is A/quail/Hong Kong/G1/1997(H9N2)),SH/F/98-like (the prototype virus is A/chicken/Shanghai/F/1998(H9N2))and Y439-like (the prototype virus is A/duck/Hong Kong/Y439/1997(H9N2)) (Luet al.2005;Fusaroet al.2011;Guet al.2017;Xuet al.2007;Yanet al.2017). In Chinese mainland the H9N2 AIV was first isolated from chicken in Guangdong in 1994 and has since been responsible for a local epidemic in most parts of China (Zhanget al.2019).

To prevent the circulation of H9N2 AIV infection in chickens,China has implemented an extensive vaccination program for A/chicken/Guangdong/SS/1994(CK/GD/SS/94),A/chicken/Shandong/6/1996(CK/SD/6/96),and A/chicken/Shanghai/F/1998(CK/SH/F/98)(Liet al.2005,2019;Zhanget al.2008;Wuet al.2010).However,H9N2 AIVs continue to persist in chickens,even in vaccinated ones (Liet al.2005;Liuet al.2016),which indicates antigenic differences between epidemic strains and vaccine strains. In our study,we performed a systematic investigation involving virus isolation,genetic and pathological studies,and serological assays to gather useful reference data for the future prevention and control of H9N2 AIVs.

2.Materials and methods

2.1.Sample collection

From January 2019 to June 2021,specimens including trachea,lung,glandular stomach,spleen,and brain were collected from 136 chickens of suspected AI in chicken farms in Shandong Province,China. All tissue samples were collected and separated into two parts. One part was immersed in cooled viral glycerol-phosphate-buffered saline (PBS) transport medium (10% glycerol,90%PBS,penicillin 2 000 U mL-1,and streptomycin 2 000 U mL-1) and frozen at -70°C for virus isolation,whereas the other part was fixed in 10% formalin solution for histopathological examination.

2.2.Virus isolation

Thawed tissue samples were ground and centrifuged at 5 000 r min-1for 3 min at 4°C. Next,0.1 mL of the supernatant was inoculated into allantoic cavities of 10-day-old specific-pathogen free (SPF) chicken eggs(SPF Chicken Research of Poultry Institute,Shandong Academy of Agricultural Sciences,Jinan,China). After 72 h of incubation at 37°C,the allantoic fluid was harvested and tested for HA activity. For samples with HA activity,the isolates were determined by hemagglutination inhibition (HI) assay using antisera against H5,H7 and H9 subtype AIV,and Newcastle disease viruses (NDV).The allantoic fluid containing the H9 subtype viruses were stored at -70°C until use.

2.3.Nucleotide sequencing

Viral RNA was extracted by using a Viral RNA Mini Kit(QIAGEN,Germany) according to the manufacturer’s instructions,and was reverse transcribed by using the Uni12 primer (5′-AGCAAAAGCAGG-3′) with the AMV reverse transcriptase (TaKaRa,Dalian,China). Complete genome amplification was performed using specific primers. Last,the PCR products were sequenced by Sangon Biotech Co.,Ltd.(Shanghai,China).

2.4.Analysis of viral sequencing

Sequencing data were compiled and edited using the program SeqMan (DNAStar,Madison,WI,USA). The reference virus sequences were downloaded from the databases of the Influenza Virus Resources at the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov),and nucleotide homology and protein sequences were analyzed using the program MegAlign(DNAStar,Madison,WI,USA). Next,phylogenetic trees were generated using the software MEGA 7.0 and by applying the neighbor-joining method with the Tamura-Nei model and 1 000 bootstrap replicates,in which a distinct phylogenetic lineage with bootstrap support of 70%was considered to indicate a common origin. Using the NetNGlyc 1.0 Server (https://services.healthtech.dtu.dk/) to predict the potential glycosylation sites (N-X-S/T motifs)in HA and NA proteins,X could be any amino acid except proline.

2.5.Virus titer

The virus titer was measured in 50% egg infectious doses(EID50) mL-1. Briefly,serial 10-fold dilutions of allantoic fluid were made in PBS,and 0.1 mL of each dilution was inoculated into allantoic cavities within 10-day-old SPF chicken eggs. Four chicken eggs were inoculated with each virus dilution. After 48 h of incubation at 37°C,the EID50values were calculated by the Reed-Muench.

2.6.Pathogenicity test in mice

To evaluate the potential pathogenicity of four H9N2 AIV isolates,we used 6-8-week-old female BALB/c mice,which we divided into five groups of eight mice each.Mice in Groups 1-4 were lightly anesthetized with dry-ice and inoculated intranasally with 106EID50of each virus at a volume of 50 μL. Meanwhile,mice in Group 5 were inoculated intranasally with 50 μL of PBS as a control.Three mice in each group were euthanized at 3 days post infection (dpi),and their brain,turbinate,lung,spleen and kidney tissues were collected and maintained at -70°C to determine the virus titers. Parts of the tissue samples were fixed in 4% neutral formaldehyde and stained with hematoxylin-eosin for histopathological observation. The rest of the mice were monitored daily for 14 days for body weight,inappetence,emaciation,inactivity,ruffled fur and mortality. Mice were considered to be dead when they lost more than 30% of their body weight. The mice were humanely euthanized with CO2at the end of the 14-day observation period.

2.7.Serological assays

We collected 7 240 chicken serum samples from Shandong Province. HI assays were performed for antibodies against H9N2 AIVs using specific antigens,and the antibody titer of each serum sample was recorded.The antigen was purchased from Harbin Guosheng Biotechnology Co.,Ltd.,China.

3.Results

3.1.Clinical signs and gross lesions

The clinical signs of sick chickens included depression,loss of appetite or weight loss,swelling of the head,edema around the eyelids,prominent eyes,lacrimation and discharge (Fig.1-A),along with a swollen,purple comb and meat beard (Fig.1-B). Other symptoms such as nasal discharge (Fig.1-C),dyspnea,water-like white stool,decreased egg production rate,irregular egg shape,and thin egg shell were also observed. As the disease progressed,symptoms of nervous disorder appeared.Autopsies revealed nasal cavities full of mucus,congested and bleeding tracheal mucosa,abundant sticky secretions on the mucosal surface (Fig.1-D),and solid,dark-red lungs (Fig.1-E). The glandular gastric papillae were swollen or hemorrhagic (Fig.1-F),and the duodena were hemorrhagic as well (Fig.1-G). The kidneys were swollen,congested and bleeding,with uric acid deposits in the form of splotchy kidney lesions (Fig.1-H). The egg follicles of the hens were liquefied and hemorrhagic,purplish-red and purple-black,and the yolk fluid flowed into the abdominal cavities following yolk rupture,which formed yolk peritonitis (Fig.1-I).

Fig.1 Clinical signs and gross lesions observed in the chickens.A,swelling of the head,edema around the eyelids. B,comb and meat beard are purple and swollen. C,nasal discharge.D,congestion of tracheal mucosa with mucus. E,lung red consolidation. F,glandular gastric papillae were swollen or hemorrhagic. G,congestion of the duodenum and pancreas.H,kidney swelling and urate deposition. I,yolk peritonitis.

3.2.Histopathological changes

Among other findings,the epithelial cells of the tracheal mucosa were detached (Fig.2-A),the lamina propria were congested (Fig.2-B),the inflammatory cells were infiltrated,and the glandular epithelial cells were necrotic(Fig.2-C). The tertiary bronchi and lung chambers were filled with serous-fibrous exudate,the interstitium of the lung lobules and the walls of the respiratory capillary were widened and congested and showed inflammatory cell infiltration (Fig.2-D-F). The submucosal layer of the glandular stomachs were vasodilated and congested,with fatty degeneration of the glandular epithelial cells and the necrosis and detachment of the mucosal epithelial cells(Fig.2-G-I). The splenic parenchymal cells (lymphocytes,macrophages) had undergone focal degeneration and necrosis,which had formed necrotic foci rich in nuclear fragmentation (Fig.2-J-L). Congestion and hemorrhage were observed in the meninges. The small arteries of the brain tissue were congested with multiple layers of lymphocytic infiltration around the vessels that had formed a cuff phenomenon. Neuronal degeneration and necrosis were observed,as were glial cells surrounding neurons to form satellite phenomenon or neuronophagia (Fig.2-M-O).

Fig.2 Histopathological changes observed in avian influenza cases. A,the epithelial cells of the tracheal mucosa were detached,inflammatory cells were infiltrated (HE×400). B,the lamina propria was congested,inflammatory cells were infiltrated (HE×400). C,the glandular epithelial cells were necrotic (HE×400). D,serous-fibrinous inflammation (HE×100). E and F,interstitial pneumonia(HE×400). G,necrosis and fall off of the mucosal epithelial cells (HE×400). H,fatty degeneration of the glandular epithelial cells(HE×400). I,the submucosal layer of the glandular stomach was vasodilated and congested (HE×100). J-L,focal necrosis of the spleen (J,HE×100;K and L,HE×400). M,congestion and hemorrhage in the meninges (HE×400). N,cuff phenomenon (HE×400).O,satellite phenomenon or neuronophagia (HE×400).

3.3.Virus isolation and identification

From January 2019 to June 2021,we got 11 H9N2 AIV isolates from chickens suspected of having AI on farms in Shandong Province. The isolated viruses did not react with H5,H7 and Newcastle disease virus antisera. The full genomes of all isolates were sequenced,and the nucleotide sequences were deposited in GenBank (the accession numbers appear in Table 1). The epidemiologic information for the 11 H9N2 AIV isolates also appears in Table 1.

3.4.Homology analysis

The gene sequences of the H9N2 AIVs were compared for homology by using the program MegAlign. The PB2,PB1,PA,HA,NP,NA,M,and NS gene isolates showed 94.6-99.6,93.5-100,94.8-100,90.6-99.6,94.5-99.7,95.7-99.4,96.1-99.8 and 96.3-99.2% homology,respectively. Homology was extremely low between the isolated and classic reference strains (DK/HK/Y280/97,QL/HK/G1/97,DK/HK/Y439/97,and CK/SH/F/98). By contrast,they were more similar to the G57 genotype reference strain (CK/ZJ/HJ/07). Comparing the isolated strains with the corresponding sequences of the highest nucleotide homology strains in GenBank revealed that the reference sequences with the highest gene homology to each isolate were primarily from viruses since 2016,with the gene similarity exceeding 97.4%. Those reference strains were isolated from a wide geographic area (Tables 2 and 3).

Table 1 Details of samples collected for 11 H9N2 avian influenza virus (AIV) isolates in Shandong Province,China from 2019 to 2021

Table 2 Nucleotide homogenous analyses of 11 H9N2 avian influenza virus (AIV) isolates with reference strains1)

Table 3 The influenza viruses in NCBI with highest nucleotide homology with 11 H9N2 avian influenza virus (AIV) isolates

3.5.Phylogenetic analysis

To understand the molecular evolution of the 11 isolates,we constructed phylogenetic trees using the entire genomes of the isolates and the reference AIVs published in GenBank. The results showed that the eight gene segments of all viruses were clustered into two major lineages of Eurasian lineage and North-American lineage.The Eurasian lineage further blooms into various virus clusters,which include BJ/94-like or Y280-like,G1-like,SH/F/98-like,and Y439-like lineage. The HA and NA genes belong to the BJ/94-like lineage,the PB2 and M genes to G1-like lineage,and the PB1,PA,NP and NS genes to SH/F/98-like lineage from the isolates. The composition of the eight gene segments of the isolates was the same as the G57 genotype (Zhuet al.2018;Sunet al.2019;Wanget al.2021),which indicates that all belong to the G57 genotype (Fig.3).

Fig.3 Phylogenetic trees of the PB2,PB1,PA,HA,NP,NA,M,and NS genes of H9N2 avain influenza viruses (AIVs). Phylogenetic trees were generated with the MEGA 7.0 Software by using neighbor-joining analysis and reliability of the tree was assessed by bootstrap analysis with 1 000 replications. Phylogenetic trees were based on the comparison of nucleotide sequences of the isolates in this study to the reference AIV sequences published in GenBank. The scale bar represents the distance unit between sequence pairs. The isolates in this study were marked in solid red circles.

Fig.3 (Continued from preceding page)

3.6.Molecular features analysis

The multi-base insertion of the HA cleavage site is a primary determinant of the pathogenicity of influenza A viruses,which is helpful for HA protein cleavage and viral replication in a wide range of host cells (Gultyaevet al.2019;Chanet al.2020).The isolates in our study possessed a single basic amino acid (PSRSSR/GLF,PSRSNR/GLF,or PSKSSR/GLF) in the HA cleavage site (H9 numbering,which is used throughout this work)(Table 4). Their cleavage sites belong to the molecular features of lowly pathogenic avian influenza (LPAI) viruses.

Different influenza virus strains show significantly different preferences for the sialic acid (SA) receptor,human and swine influenza viruses prefer to bind to SA-α-2,6-terminal saccharides,whereas AIVs prefer to bind to SA-α-2,3-terminal saccharides (Rogerset al.1989;Suzukiet al.2000;Trebbienet al.2011). Such receptor binding specificity is primarily determined by a few key amino acid sites of the HA protein. Compared with the reference strains,the receptor binding sites of the HA protein from the isolates were conserved at amino acid positions Y109,W161,L202,Y203 and G236 and mutation at amino acid positions N163 and K198. Past studies have shown that the substitution of I163T,H191N,A198V,Q234L and G236S in the HA protein will lead to enhanced binding to human receptors (Srinivasanet al.2013;Liet al.2014;Zouet al.2019;Tenget al.2016;Obadanet al.2019). Meanwhile,T163,N191,V198 and L234 were observed in two isolates;T163,N191 and L234 were observed in seven;and N191,V198 and L234 were observed in two (Table 4).

Other past studies have revealed that the substitution of G90E,S145D,D153G,N167G,A168D,T200R,and N201D related to the antigenic variation of H9N2 AIVs(Liet al.2019). E90,D145,G153,G167 and R200 were observed in three isolates;D145,G153,G167 and R200 were observed in two;and E90 was observed in one. In addition,the antigenic variation N168 appeared in five isolates,whereas the antigenic variation S/T201 appeared in three,which indicates further variation in the antigenicity of H9N2 AIVs (Table 4).

Changes in certain glycosylation sites of the HA and NA proteins may directly affect the biological characteristics of influenza virus strains. Compared with the reference strains,the glycosylation sites of the HA protein from the isolates were conserved at positions NST29 and NGT492.CK/SD/099/20 was missing a potential glycosylation site at position 82,while CK/SD/127/21 add a potential glycosylation site at position 145. Studies have shown that deleting the glycosylation site at position 218 and increasing the glycosylation site at position 313 increased antibody binding and moderately prevented the virus from escaping neutralization with homologous antisera (Penget al.2019).Potential glycosylation sites were deleted at position 218 of the HA protein of the isolates,and a potential glycosylation site was added at positions 313. Compared with the reference strains,the glycosylation sites of the NA protein from the isolates were conserved at positions 69 and 234(N2 numbering,which is used throughout this work). CK/SD/062/20 and CK/SD/079/20 were missing a potential glycosylation site at position 44,while CK/SD/097/20 and CK/SD/099/20 were missing one at position 86 (Tables 5 and 6).

The hemadsorption sites (HBS’s,366-373,399-403 and 431-433),active center (140-157),and antigenic determinants (153,197-199,328-336,339-347,367-370,400-403,and 431-434) in the NA protein were also analyzed (N2 numbering,which is used throughout this work) (Zhuet al.2018). Compared with the reference strains,the HBS’s of the NA protein from the isolates were changed at positions 366,399,400,403 and 431,and no mutations occurred at other positions. The active center of the NA protein had changed at positions 140,147,150 and 152,while its antigenic site had changed at positions 328,343,400,403 and 431,with no mutations at other positions.The HA protein’s receptor binding region is foveolar,with 146-150 amino acids forming the right margin and 232-237 amino acids forming the left margin (Liet al.2019).Compared with the reference strains,the left edge of the receptor-binding pockets of the HA protein from the isolates had changed at positions 149 and 150,whereas the right edges had changed at positions 233,234 and 235 (Table 7).

Mammalian adaptation mutations in the PB2 protein(V588,R591,K627 and N701) were identified (Barberiset al.2020;Xiaoet al.2016),and all the 11 AIV isolates had such a mutation at positions V588. It is known that 368V in the PB1 protein exists in the H5N1 virus adapted to ferrets and it has been shown that this mutation is dominant in the H9N2 viruses (Sunet al.2020). Compared with the reference strains,except CK/SD/062/20 and CK/SD/079/20,all nine AIV isolates had the mutation at positions 368. K577E in the PB1 protein of H9N2 viruses is a determinant of pathogenicity in mice and could be a signature for mammalian adaptation of AIV(Kamikiet al.2018). All isolates did not have mutations at position 577. K356R in the PA protein of the H9N2 AIVs can increase mammalian replication and pathogenicity(Xuet al.2016). All 11 H9N2 AIV isolates had mutations at position 356. No H274Y or R292K substitution was observed in all isolates,which indicates that isolated viruses would be sensitive to NA inhibitors (Gubarevaet al.2001;Montoet al.2006;Tanget al.2019). Last,S31N was observed in the M2 protein in all isolates (Table 8),which suggests that the isolated viruses would be resistant to amantadine (Zhuet al.2018).

3.7.Replication and virulence of the H9N2 viruses in mice

No mice infected with four H9N2 AIV isolates showed obvious weight lost during the observation period. All isolates replicated effectively in the turbinate,with titers of 2.5 to 6.75 log10EID50mL-1. Only two viruses,CK/SD/005/19 and CK/SD/079/20,replicated effectively in the lungs of mice with titers of 3.25 to 7.5 log10EID50mL-1.CK/SD/102/19 was detected in the lung of a mice with titers of 2.25 log10EID50mL-1,whereas CK/SD/099/19 was not. No virus was detected in the spleen,kidneys,or brain of any mice (Fig.4).

Fig.4 Weight variation (A) and replication (B) of four H9N2 avian influenza virus (AIV) isolates in mice. Mice in each group were intranasally inoculated with 106 EID50 of each virus in a 50-μL volume. The body weight of mice was measured over 14 days.Visceral tissues were harvested on 3 dpi. Viral titer was determined by endpoint titration in 10-day-old SPF chicken eggs.

3.8.Histopathological damage to the H9N2 viruses in mice

Histopathological changes in the lungs showed similar changes of alveolar wall thickening and inflammatory cell infiltration in all infected groups (Fig.5-B-E).

Fig.5 Histopathological damage of avian influenza virus (AIV) isolates in mice. Mice in each group were intranasally inoculated with 106 EID50 of each virus in a 50-μL volume. Lung tissues were harvested on 3 days post-infection (dpi). Lung tissues fixed in 10% formalin,embedded in paraffin,sectioned,stained with hematoxylin and eosin,and observed under a microscope (HE×400).A,a indicates the control group. B-E,lung sections of mice infected with CK/SD/005/19,CK/SD/079/20,CK/SD/099/20,and CK/SD/102/21,respectively.

3.9.Analysis of H9N2 AlV antibody test

Among the 7 240 serum samples,the average titer of the HI antibody was 8.60 log2,with a dispersion of 23.56% and antibody pass rate was 98.52% (≥4 log2),which indicatethat the overall prevention of H9N2 in large-scale chicken farms in Shandong Province could protect the chickens from the same type of influenza virus. However,107 serum samples had antibody titers less than 4 log2. The antibody titers of 1 989 serum samples ranged from 4 to 7 log2. 3 689 serum samples had antibody titers ranging from 8 to 10 log2,and 1 455 serum samples had antibody titers greater than 11 log2. The serum antibody levels were diverse,which may be due to a variety of factors,including differences in individual flocks,feeding environments and vaccination routes (Table 9).

Table 4 The key amino acid sites of HA protein in 11 H9N2 avian influenza virus (AIV) isolates and reference strains

Table 5 The potential glycosylation sites of HA protein in 11 H9N2 avian influenza virus (AIV) isolates and reference strains

Table 6 The potential glycosylation sites of NA protein in 11 H9N2 AIV isolates and reference strains

Table 7 The key amino acid sites of NA protein in 11 H9N2 avian influenza virus (AIV) isolates and reference strains

Table 8 The key amino acid sites of PB2,PB1,PA,NA and M protein in 11 H9N2 avian influenza virus (AIV) isolates and reference strains

Table 9 Antibody titers of serum samples in different collecting times

4.Discussion

The H9N2 AIV was first reported in Wisconsin in North American turkey flocks in 1966 (Hommeet al.1970). InChina,the H9N2 AIV was first isolated from chickens in Guangdong Province in 1994 (Sunet al.2015;Guet al.2017). However,the long-term latent existence of the virus in chickens that have not shown any characteristic clinical symptoms could increase the likelihood of gene reassortment with other AIVs during co-infection. In addition,because H9N2 AIVs can infect humans and other mammals (Peiriset al.1999;Puet al.2017;Sunet al.2013;Yuet al.2011;Zhanget al.2015). This highlights the ability of the H9N2 AIV to cross-species transmission. In our study,11 H9N2 AIVs had mutation sites that preferentially bind to human-like receptors and increase mammalian fitness sites,such as T163,N191,V198,and L234 in the HA protein,V588 in the PB2 protein,V368 in the PB1 protein,and R356 in the PA protein. At present,H9N2 AIVs represent the primary subtype of AIVs prevalent in chickens in China,which has caused continuous harm to the country’s poultryindustry. Among them,the genotype G57 virus has been the primary epidemic strain in China since 2013 (Zhuet al.2018). All isolates in our study also belonged to the confirmed G57 genotype,which indicates that the G57 genotype remains the primary epidemic strain in chicken in China.

Vaccination is the primary means of controlling the H9N2 AIV infection in chickens. Although China once used CK/GD/SS/94,CK/SD/6/96 and CK/SH/F/98 as vaccine strain for H9N2 AIV and thus somewhat reduced the positive rate of H9N2 AIV,those viruses remain in circulation and continue to evolve in China. Moreover,currently available vaccines may provide only limited protection. Our data indicate that there are antigenic variants related sites (E90,D145,G153,G167,and R200)exist in the HA protein of H9N2 AIV isolate. Although our data also show high levels of antibodies to H9N2 AIVs in flocks from Shandong Province,H9N2 AIV infection in high antibody immunized chicken flocks occurred nevertheless,possibly due to the antigenic mutation of the H9N2 AIVs,though may be related to many other predisposing factors. Those mutation sites may cause H9N2 AIVs to evade vaccine immunity.

The M2 protein inhibitors amantadine and rimantadine and the neuraminidase inhibitors (NAIs) oseltamivir and zanamivir been used to treat influenza virus infections in many countries;however,the drug resistance of M2 inhibitors has been widely reported (Yuet al.2011;Wanget al.2018;Liet al.2019;Kodeet al.2019;). In our study,all isolated H9N2 AIVs contained 31N amantadine resistance mutations but no mutations of NAIs resistance.If NAIs resistance emerges in the future amid a lack of new anti-influenza drugs,then it will significantly impact the treatment of influenza viruses. Therefore,updating vaccines and anti-influenza drugs seed strains based on data from continuous monitoring is particularly important for the future control of H9N2 AIVs.

To evaluate whether the H9N2 AIVs isolates could cross interspecies barriers and infect mammalian hosts,we examined viral replication and virulence in BALB/c mice. Some H9N2 AIVs cannot replicate in mice and cause symptoms,whereas others can replicate in mice and cause signs of disease,even death (Liet al.2005;Biet al.2010). In our study,no mice infected showed obvious weight lost or death. The difference in the pathogenicity of H9N2 AIVs in mice may be mainly due to differences in the characteristics of the HA and polymerase (PB2,PB1 and PA) proteins in determining the host range and adapting to mammalian species.At the same time,other proteins of H9N2 AIVs might also play important role,and it is necessary to further investigate the specific factors affecting the pathogenicity of H9N2 AIVs in mammals. Among other results,virus titer indicated that the three H9N2 AIV isolates replicated better in turbinate than in lungs. These results showed that most isolates could replicate better in the upper respiratory tract than in the lower respiratory tract. The reasons for the different tissue tropisms of the isolated viruses to the turbinates and lungs of mice also require further investigation.

Due to the characteristics of the influenza virus,genetic mutations and recombination will continue to occur,which will likely spawn new influenza pandemics such as the 2009 swine-origin H1N1 influenza pandemic. Therefore,it is necessary to understand the molecular biological characteristics of H9N2 AIVs to facilitate the prediction of mutants with epidemic potential and in turn,assess the potential risks of H9N2 AIV to public health.

5.Conclusion

In summary,our findings suggest that the isolates exhibit mutant sites that preferentially bind to humanlike receptors and mammalian fitness sites,as well as mutations in antigenic and drug resistance sites.Moreover,studies with mice identified four isolates with different levels of pathogenicity. Antibody levels in serum samples in this study showed a high degree of diversity.Continued surveillance of the H9N2 virus is needed to monitor further increases in virus evolution and its potential threat to public health.

Acknowledgements

We are grateful for Prof.Sun Honglei (China Agricultural University) who kindly provided us experimental guidance.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Ethical approval

All animal experiments in our study were approved by the Animal Care and Use Committee of Shandong Agricultural University,China (Project Identification Code 2015-012).