Molecular characterization and pathogenicity of H9N2 avian influenza viruses in poultry in Shandong Province, China, 2021–2023 Authors: Wei Zhang, Libing Sun, Qian Wang, Chenlu Xia, Yuzhong Zhao Published: 2026 ABSTRACT AbstractThe H9N2 avian influenza virus (AIV) is currently widespread worldwide, posing a severe threat to the poultry industry and public health. To investigate the epidemiological characteristics, genomic evolution, and mammalian pathogenicity of poultry-derived H9N2 viruses in Shandong Province, a total of 192 samples from diseased chickens were collected across 10 cities in Shandong Province between February 2021 and December 2023. Real-time quantitative PCR (RT-qPCR) detection identified 10 positive samples, from which 8 virus strains were isolated. Five representative isolates were selected for whole-genome sequencing, phylogenetic analysis, molecular characterization, and mouse pathogenicity evaluation. Infected chickens mainly showed respiratory symptoms and pathological lesions in multiple organs. The five isolates shared 91.3%–99.8% nucleotide identity and were highly homologous (98.26%–99.95%) with recent Chinese H9N2 strains. Phylogenetic results indicated multiple-lineage reassortment: PB2 and M genes belonged to the G1-like lineage, PB1, PA, NP, and NS genes to the SH/F98-like lineage, and HA and NA genes to the Y280-like lineage. All isolates carried the low‑pathogenic HA cleavage site (PSRSSR/GL) and human adapted receptor binding mutations (HA Q226L, I163T, H183N), with HA glycosylation loss at position 218 but retention at 313. In addition, substitutions were observed at antigenic sites, including G90E, S145D, N167G, and T200R. All isolates carried mammalian adaptation related mutations, such as PB2 A588V, PA K356R, PA-X R195K and NS1 F103L/M106I; only the CK/SD/TA1016/22 strain contained the PB2 K526R mutation. All strains carried M2 S31N conferring amantadine resistance, and a new NA glycosylation site at position 306 was found in CK/SD/TA803/22. The pathogenicity to mice showed that the three tested isolates could replicate efficiently in the lungs without prior adaptation, but could not cause mice death. Among them, the CK/SD/TA1016/22 strain showed higher replication levels and more severe lung lesions. In conclusion, H9N2 AIV in Shandong Province continue to evolve and undergo antigenic drift, and they possess evident potential for mammalian adaptation. Continuous epidemiological surveillance and risk assessment are therefore strongly recommended.Keywords: Avian influenza, H9N2, Shandong, Molecular characterization, Pathogenicity Introduction Avian influenza viruses (AIV) belong to the genus Influenza A virus in the family Orthomyxoviridae. Their subtypes are defined based on the antigenic differences of two surface proteins, hemagglutinin (HA) and neuraminidase (NA) (Roy Chowdhury et al., 2019; Sutton et al., 2017). Currently, 19 HA subtypes (H1–H19) and 12 NA subtypes (N1–N12) have been identified. H1–H16 and N1–N9 are predominantly detected in wild waterfowl, the natural reservoirs of AIVs (Webster et al., 1992). Bat-associated subtypes (H17–H18 and N10–N12), as well as H19 in lesser scaups, are also included (Echeverri-De la Hoz et al., 2025; Karakus et al., 2024). H9N2 AIV was first reported in 1966 in turkey flocks in Wisconsin, USA. (Homme et al., 1970). In China, H9N2 AIV was first detected in chicken flocks in Guangdong Province in 1994 and quickly became one of the most widespread and genetically stable AIV subtypes in domestic poultry (Han et al., 2019). Although H9N2 AIV is generally classified as low pathogenic, its long-term circulation and continuous genetic changes make it a potential threat to both poultry farming and public health. The virus mainly spreads through the respiratory route and can infect chickens, ducks, geese, pigeons, quails, and various wild birds (Islam et al., 2025). A single infection usually causes only mild respiratory signs, reduced egg production, and weakened immunity, with little visible organ damage (Hao et al., 2019). However, in commercial poultry farms, H9N2 often co-infects with other respiratory pathogens, such as infectious bronchitis virus or Mycoplasma, leading to more severe clinical signs, higher mortality, and significant economic losses (Cao et al., 2022). H9N2 AIV also poses potential public health risks. Some naturally occurring strains can spread among ferrets through respiratory droplets, suggesting they may cross species boundaries (El Houadfi et al., 2016), and several human infections have been reported (Liu et al., 2020). In addition, H9N2 often acts as a “gene donor” in the formation of new zoonotic influenza viruses. For example, during the 2013 H7N9 outbreak in China, all six internal genes of the virus came from H9N2 AIV (Gao et al., 2013). Later reassortant viruses, such as H5N6 and H10N8, are also closely related to H9N2, highlighting its role in generating viruses with pandemic potential (Yang et al., 2021; Hoffmann et al., 2001). In recent years, the genetic evolution and spread of H9N2 AIV have been influenced by the expansion of poultry farming, frequent movement of poultry across regions, and the ongoing operation of live bird markets. To understand these changes, this study analyzed H9N2 AIVs isolated from 2021 to 2023. We performed whole-genome sequencing and phylogenetic analysis to examine their genetic evolution and molecular characteristics, and used a mouse infection model to evaluate their ability to adapt to mammals and cause disease. This study provides important insights into the recent evolution and potential public health risks of H9N2 AIV, supporting surveillance and control strategies. Materials and methods Sample collection and H9N2 AIV detection To investigate the prevalence of H9N2 AIV in poultry in Shandong Province, samples were collected from a diagnostic laboratory between February 2021 and December 2023. All samples originated from routine diagnostic submissions from poultry farms in ten prefecture-level cities of Shandong Province, including Jinan, Heze, Liaocheng, Linyi, Jining, Weifang, Dezhou, Binzhou, Taian, and Yantai. These samples were submitted by poultry owners as part of routine disease diagnosis, with informed consent obtained prior to submission. Sampling was focused on chicken flocks exhibiting respiratory symptoms. A total of 192 chickens were sampled, with 3–5 chickens collected from each flock. Approximately 0.3 g of tracheal and lung tissues was aseptically collected using sterile scissors and forceps and placed into 1 mL of phosphate-buffered saline (PBS). The samples were centrifuged at 8,000 × g for 3 min, and viral RNA was extracted from the supernatant using the Qiagen RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Reverse transcription was performed using the universal primer Uni12 (5′-AGCAAAAGCAGG-3′) and the ReverTra Ace qPCR RT Kit (Toyobo Biochemicals, Osaka, Japan) according to the manufacturer’s instructions to synthesize complementary DNA (cDNA). The resulting cDNA was stored at −80 °C until use. Detection of H9N2 AIV was performed by real-time quantitative PCR (RT-qPCR) using a commercial amplification kit (Transgene, Beijing, China) in combination with specific primers reported previously (Monne et al., 2008). Detection of Newcastle disease virus (NDV) was performed by reverse transcription polymerase chain reaction (RT-PCR) using primers reported previously (Kant et al., 1997). Virus isolation and identification H9N2-positive tissue samples were homogenized in 1 mL PBS supplemented with penicillin (2,000 U/mL) and streptomycin (2,000 μg/mL), followed by centrifugation at 8,000 × g for 3 min. The clarified supernatants were inoculated into the allantoic cavities of 10-day-old specific pathogen-free (SPF) embryonated chicken eggs obtained from the Shandong Academy of Agricultural Sciences (Jinan, Shandong, China) and incubated at 37 °C for 72 h. Embryos that died within 24 h post-inoculation were considered non-specific deaths and were excluded from further analysis. To obtain purified virus isolates, allantoic fluids from the initial inoculation were subjected to three consecutive blind passages in SPF eggs under the same conditions. After the final passage, the presence of virus in the allantoic fluids was confirmed by hemagglutination assay using 1 % chicken red blood cells, and viral RNA was subsequently extracted to confirm H9N2 AIV by RT-qPCR. Confirmed virus isolates were aliquoted and stored at −80 °C for subsequent experiments. Whole-genome amplification and sequencing For confirmed H9N2 AIV isolates, the full-length genome, including all eight viral gene segments (PB2, PB1, PA, HA, NP, NA, M, and NS), was amplified using a universal primer set as previously described by Hoffmann (Hoffmann et al., 2001). PCR products were examined by electrophoresis on 0.8% agarose gels, excised, and purified using a commercial gel extraction kit (Qiagen, Beijing, China). Purified PCR products were subjected to Sanger sequencing performed by Shanghai Shenggong Biological Engineering Technology Services Co., Ltd. (Shanghai, China). Sequence assembly and phylogenetic analysis Raw sequence data were assembled and edited using SeqMan (DNASTAR software package). Nucleotide sequence homology analysis was performed using MegAlign (DNASTAR software package). Sequence homology analysis was performed using the Basic Local Alignment Search Tool (BLAST) provided by the National Center for Biotechnology Information (NCBI) to identify the most closely related reference sequences. The selection of reference strains was based on previously published studies (Li et al., 2017; Zhao et al., 2023; Sun et al., 2019) and further refined through BLAST analysis to identify and confirm sequences most closely related to the study isolates. In addition, representative reference sequences were retrieved from then Global Initiative on Sharing All Influenza Data (GISAID) and the NCBI Influenza Virus Resource database to ensure comprehensive coverage of genetic diversity across different lineages and geographic origins. Phylogenetic analyses were performed based on the complete open reading frames of each gene segment. Phylogenetic trees were constructed using the neighbor-joining (NJ) method with 1,000 bootstrap replicates implemented in MEGA version 7.0. Molecular feature analysis Key amino acid residues associated with receptor-binding specificity, mammalian adaptation, and viral virulence were analyzed. Amino acid sequence alignments were performed using MegAlign, and molecular markers were identified through comparison with previously reported sequences. Potential glycosylation sites were predicted using the NetGlyc 1.0 Server and were defined as N-X-S/T motifs, where X represents any amino acid except proline (http://www.cbs.dtu.dk/services/NetNGlyc/). Virus titration (EID50) The virus titer was determined by the 50% egg infectious dose (EID50) assay. Briefly, the virus suspension was serially diluted 10-fold in PBS, from 10-1 to 10-9. A volume of 0.1 mL from each dilution was inoculated into the allantoic cavity of 10-day-old SPF embryonated chicken eggs. After incubation at 37 °C for 48 hours, the allantoic fluid was harvested, and virus presence was confirmed by hemagglutination assay. The EID50 was calculated using the Reed and Muench method. Mouse pathogenicity study A total of 32 six-week-old female SPF BALB/c mice were purchased from Jinan Pengyue Biotechnology Co., Ltd. (Jinan, Shandong, China). Mice were randomly divided into four groups, including three virus-infected groups and one mock-infected control group. Animals were housed in standard laboratory conditions, with five mice per cage, a 12 h light/dark cycle, and ad libitum access to food and water. Although the facility was not SPF-certified, all animals were routinely monitored for health status and maintained under strict hygiene and biosafety measures. Mice in the infection groups were intranasally inoculated with 50 μL of virus suspension containing 106 EID50 of H9N2 AIV, while control mice received an equal volume of PBS. Body weight and survival were monitored daily for 14 days post-inoculation (dpi). Mice that lost more than 25% of their initial body weight were humanely euthanized in accordance with institutional animal welfare guidelines. At 3 dpi, mice were euthanized, and lung tissues were collected aseptically. Lung homogenates were prepared for viral titration, and virus titers were expressed as log10 EID50/mL. For histopathological analysis, portions of lung tissues were fixed in 4% paraformaldehyde for 24 h, dehydrated through a graded ethanol series, cleared in xylene, and embedded in paraffin. Tissue sections of 5 μm thickness were cut using a microtome, mounted on glass slides, deparaffinized, and stained with hematoxylin and eosin (H&E) following standard protocols. Stained sections were examined under a light microscope to assess pathological changes. Results Virus detection, isolation, and genome sequencing Between February 2021 and December 2023, RT-qPCR identified 10 samples positive for H9N2 AIV, while all samples were negative for Newcastle disease virus NDV by RT-PCR. The H9N2-positive samples were inoculated into the allantoic cavity of 10-day-old SPF embryonated chicken eggs and subjected to three consecutive blind passages to obtain purified virus isolates. Viral replication in the allantoic fluid was assessed by hemagglutination assay using 1% chicken red blood cells and further confirmed by RT-qPCR. Ultimately, eight H9N2 AIV strains were successfully isolated. Among these, five representative isolates were selected for whole-genome sequencing and further analyses and were designated CK/SD/HZ056/21, CK/SD/DZ078/21, CK/SD/TA803/22, CK/SD/TA1016/22, and CK/SD/LY1437/23. All eight gene segments of the five representative isolates have been deposited in the GenBank database under the following accession numbers: CK/SD/HZ056/21 (PZ177280–PZ177287), CK/SD/DZ078/21 (PZ177294–PZ177301), CK/SD/TA803/22 (PZ177302–PZ177309), CK/SD/TA1016/22 (PZ177310–PZ177317), and CK/SD/LY1437/23 (PZ177318–PZ177325). Gross pathology Chickens infected with H9N2 AIV exhibited pronounced swelling of the head, accompanied by lacrimation and increased ocular discharge (Fig. 1A). Some chickens developed diarrhea, with yellow-green feces (Fig. 1B). The tracheal mucosa was congested and hemorrhagic, and the mucosal surface was coated with abundant sticky secretions (Fig. 1C and D). The kidneys were enlarged to varying degrees, and hemorrhage was observed in some cases (Fig. 1E). The yolk sac also showed congestion and hemorrhage (Fig. 1F). Secondary lesions, including fibrinous pericarditis, airsacculitis, and perihepatitis, were frequently observed in infected chickens. Clinical signs and gross lesions in chickens infected with H9N2 AIV. (A) Swelling of the head with protruding eyes, lacrimation, and ocular discharge; (B) Diarrhea with yellow-green feces; (C-D) Marked congestion and hemorrhage of the tracheal mucosa, with abundant yellowish-white viscous exudate adherent to the tracheal lumen; (E) Enlarged kidneys with hemorrhage; (F) Congestion and hemorrhage of the yolk membrane. Histopathology In the lungs, small arteries and pulmonary capillaries were markedly congested, with capillary lumens filled with red blood cells (Fig. 2A). In the trachea, mucosal epithelial cells were shed, the lamina propria was congested, and infiltration of inflammatory cells was evident; glandular epithelial cells exhibited necrosis (Fig. 2B). The spleen showed focal necrosis, with degenerative and necrotic lymphocytes and macrophages forming necrotic areas containing numerous nuclear fragments (Fig. 2C). Renal tubular epithelial cells exhibited degeneration, necrosis, and sloughing, with interstitial congestion and occasional hemorrhage (Fig. 2D). In the brain, small arteries were congested, and lymphocytes formed multiple layers around the vessels, exhibiting characteristic “perivascular cuffing” (Fig. 2E). In the proventriculus, glandular lumens contained necrotic and sloughed glandular epithelial cells (Fig. 2F). Histopathological changes in various tissues of chickens infected with H9N2 AIV. (A) Pulmonary capillaries and small arteries showing marked congestion. HE × 400; (B) Tracheal mucosal epithelium shedding with inflammatory cell infiltration. HE × 200; (C) Focal necrosis in the spleen with abundant nuclear debris in necrotic areas. HE × 40; (D) Degeneration, necrosis, and desquamation of renal tubular epithelial cells. HE × 200; (E) Perivascular lymphocytic cuffing in the brain. HE × 400; (F) Necrosis and desquamation of gastric glandular epithelium. HE × 400. Homology analysis and phylogenetic analysis In this study, the eight gene segments (PB2, PB1, PA, HA, NP, NA, M, and NS) of five H9N2 AIV isolates obtained in Shandong Province from 2021 to 2023 were sequenced, and nucleotide homology and phylogenetic analyses were performed. Nucleotide homology analysis showed that the PB2, PB1, PA, HA, NP, NA, M, and NS genes of the five isolates exhibited sequence identities ranging from 95.5%–99.6%, 93.1%–99.4%, 95.0%–99.8%, 91.3%–99.4%, 96.3%–99.5%, 95.7%–99.8%, 96.1%–99.7%, and 96.5%–99.0%, respectively. BLAST analysis (Table S1) indicated that all gene segments of the isolates exhibited high nucleotide identity (98.26%–99.95%) with H9N2 AIV strains circulating in China in recent years. The closest matching reference sequences were mainly derived from H9N2 AIVs isolated from poultry or environmental samples in Jiangxi, Shandong, Zhejiang, Guangdong, and Hunan provinces between 2017 and 2021, suggesting that the isolates belong to a domestic lineage with a relatively stable genomic composition. Notably, for the M gene, isolate CK/SD/TA803/22 shared 98.45% nucleotide identity with A/duck/China/9-12/2022 (H6N6), whereas the M genes of the other isolates remained highly homologous to H9N2 AIVs, suggesting a potential inter-subtype reassortment event. Phylogenetic analysis indicated that H9N2 AIVs can be broadly divided into two major lineages worldwide, namely the North American lineage and the Eurasian lineage (Song et al., 2019). The Eurasian lineage has further diverged into multiple sub-lineages, including BJ/94-like (represented by A/chicken/Beijing/1/1994(H9N2)), Y280-like (A/duck/Hong Kong/Y280/1997(H9N2)), G1-like (A/quail/Hong Kong/G1/1997(H9N2)), SH/F/98-like (A/chicken/Shanghai/F/1998(H9N2)), and Y439-like (A/duck/Hong Kong/Y439/1997(H9N2)) (Zhao et al., 2023). The five isolates in this study showed distinct phylogenetic clustering patterns among different gene segments: the PB2 and M genes clustered within the G1-like lineage, the PB1, PA, NP, and NS genes belonged to the SH/F98-like lineage, and the HA and NA genes clustered within the Y280-like lineage (Fig. 3). Overall, all gene segments clustered with representative strains of the predominant H9N2 lineages circulating in domestic poultry in China, and no obvious inter-lineage segregation was observed. Within each gene cluster, topological differences were observed, where TA803/22 and TA1016/22 formed independent branches, while the remaining three isolates clustered within the same sub-branch. This topology was consistent with the nucleotide homology results, indicating genetic variation among isolates while maintaining relatively high overall genetic relatedness. Notably, inconsistent lineage assignments among different gene segments suggest potential segmental evolutionary patterns in these viruses. Overall, H9N2 AIVs circulating in Shandong Province exhibited a genetic background dominated by G1-like, SH/F98-like, and Y280-like lineages across multiple gene segments, with different genes showing distinct predominant evolutionary branches, reflecting ongoing viral evolution and the co-circulation of multiple lineages in this region. Phylogenetic trees were constructed for the nucleotide sequences of PB2, PB1, PA, HA, NP, NA, M, and NS using the neighbor-joining method in MEGA 7.0, with 1,000 bootstrap replications. Solid red circles indicate H9N2 AIV isolated in this study. Molecular characterization To describe the molecular features of the five H9N2 AIV isolates in this study, key amino acid sites in HA, NA, PB2, PB1, PB1-F2, PA, PA-X, NP, M1, M2, NS1, and NEP were analyzed. The HA cleavage site (residues 333–341) of all five isolates was PSRSSR/GL (H9 numbering, used throughout this study), which is characteristic of low-pathogenic AIVs (Gohrbandt et al., 2011). The receptor-binding sites of HA were relatively conserved. Residues 191, 198, 226, and 228 were N, T, L, and G in all isolates. All isolates carried the Q226L mutation, while no G228S substitution was detected, suggesting a potential preference for human-like receptors (α-2,6 sialic acid) (Vines et al., 1998; Li et al., 2014; Sang et al., 2015; Zou et al., 2019). Substitutions at position 163 were observed: four isolates possessed T at this site (I163T), which has been reported to enhance binding to human receptors (Li et al., 2014), whereas CK/SD/TA1016/22 carried N. Although the role of H183N in H9 viruses remains unclear, studies in H3 subtype viruses have shown that this mutation can increase binding to human receptors (Weis et al., 1988). All isolates carried 183N, suggesting a potential role in host adaptation that warrants further investigation. Potential N-linked glycosylation sites in HA were highly conserved among the five H9N2 AIV isolates. Conserved motifs were identified at positions 29 (NST), 82 (NPS), 141 (NVS), 298 (NTT), 305 (NVS), 313 (NCS), and 492 (NGT), which were present in all isolates. Variations were mainly observed at positions 145 and 207. CK/SD/TA803/22 possessed an additional glycosylation motif (NTT) at position 207, whereas CK/SD/TA1016/22 contained a motif (NGT) at position 145. The remaining three isolates lacked glycosylation motifs at these two positions. Previous studies have shown that loss of the glycosylation site at position 218 and acquisition at position 313 can enhance antibody binding and reduce immune escape (Peng et al., 2019). Consistently, all isolates lacked the glycosylation site at position 218 but retained the site at position 313. Previous studies have demonstrated that amino acid substitutions G90E, S145D, D153G, N167G, A168D, T200R, and N201D are closely associated with antigenic variation in H9N2 AIVs (Li et al., 2019). In the present study, substitutions at several antigenic sites were observed among the isolates. Specifically, four isolates carried E90, D145, G167, and R200, respectively. One isolate exhibited a G153 substitution, and three isolates further showed an additional N substitution at this position. At position 168, A168N substitution was detected in four isolates. In addition, an N201D substitution was identified in one isolate. These findings indicate that the antigen-related residues exhibit varying degrees of amino acid variation among the different H9N2 AIV isolates (Table 1). Table 1Key amino acid sites in the HA proteins of the 5 H9N2 AIVs identified in this study.Table 1 dummy alt textVirus nameHACleavage site (H9 numbering)Receptor binding site (H9 numbering)potential glycosylation sites (H9 numbering)Antigenic site(H9 numbering)333-341163191198234236298214114520729830531349290145153167168200201CK/SD/HZ056/21PSRSSR/GLTNTLGNSTNPSNVSNTTNVSNCSNGTEDNGNRNCK/SD/DZ078/21PSRSSR/GLTNTLGNSTNPSNVSNTTNVSNCSNGTEDNGNRNCK/SD/TA803/22PSRSSR/GLTNTLGNSTNPSNVSNTTNTTNVSNCSNGTEDGGNRNCK/SD/TA1016/22PSRSSR/GLNNTLGNSTNPSNVSNGTNTTNVSNCSNGTGNDNATACK/SD/LY1437/23PSRSSR/GLTNTLGNSTNPSNVSNTTNVSNCSNGTEDNGNRN Key amino acid sites in the HA proteins of the 5 H9N2 AIVs identified in this study. NA analysis revealed that neuraminidase inhibitor (NAI) resistance-associated residues at positions 274 and 292 remained conserved as H and R, respectively (N2 numbering, used throughout this study). No known resistance mutations, such as H274Y or R292K, were detected (Pizzorno et al., 2011), suggesting potential sensitivity to neuraminidase inhibitors. N-linked glycosylation sites in NA were also largely conserved at positions 44 (NPS), 69 (NST), 86 (NWS), 146 (NGT), and 234 (NGT). Only CK/SD/TA803/22 exhibited an additional glycosylation motif at position 306 (Table 2). Table 2Key amino acid sites in the NA proteins of the 5 H9N2 AIVs identified in this study.Table 2 dummy alt textNANAI resistance (N2 numbering)potential glycosylation sites (N2 numbering)119292446986146234306ERNPSNSTNWSNGTNGTERNPSNSTNWSNGTNGTERNPSNSTNWSNGTNGTNMSERNPSNSTNWSNGTNGTERNPSNSTNWSNGTNGT Key amino acid sites in the NA proteins of the 5 H9N2 AIVs identified in this study. In PB2, amino acid residues at positions 271, 526, 588, 590, 591, 627, and 701 were T, K, V, G, Q, E, and D, respectively, and were largely conserved among all isolates. Known mammalian-adaptive substitutions PB2 T271A, E627K, and D701N, which enhance viral replication and virulence in mammals (Ge et al., 2017), were not detected in any of the isolates. Previous studies have shown that the PB2 K526R mutation can enhance viral replication and increase pathogenicity in mice (Song et al., 2014). Notably, this mutation was identified exclusively in the CK/SD/TA1016/22 isolate. In contrast, all isolates carried the PB2 A588V mutation, which has been reported in H10N8, H7N9, and H9N2 virus backbones to enhance polymerase activity, thereby promoting viral replication and increasing pathogenicity in mice (Xiao et al., 2016). The PB2 G590S and Q591R mutations, which can compensate for the absence of 627K and enhance mammalian adaptation, were not observed (Tarendeau et al., 2008; Mehle et al., 2009; Yamada et al., 2010). In PB1, no substitutions at positions 368 and 622 were detected, including PB1 I368V and 622G, which have been reported to enhance interspecies transmission and virulence in mammalian models (Ren et al., 2013; Feng et al., 2015). In PB1-F2, all isolates carried the S66 residue, indicating the absence of the N66S mutation associated with enhanced virulence and inhibition of early interferon responses (Conenello et al., 2011). In PA, all isolates possessed the K356R substitution, which has been associated with increased replication and pathogenicity in mice (Xu et al., 2016). In PA-X, all isolates carried K195, corresponding to the R195K mutation, which has been reported to enhance virulence and transmission in mammalian hosts (Sun et al., 2020) (Table 3). Table 3Key amino acid sites in the PB2, PB1, PB1-F2, PA, PA-X, NP, M1, M2, NS1 and NEP proteins of the 5 H9N2 AIVs identified in this study.Table 3 dummy alt textVirus namePB2PB1PB1-F2PAPA-XNPM1M2NS1NEPMammalian adaptationMammalian adaptationVirulenceVirulenceMammalian adaptationVirulenceMammalian adaptationMammalian adaptationDrug resistanceVirulenceMammalian adaptation271526588590591627701368622663561953193021531103106164175CK/SD/HZ056/21TKVGQEDVGNRKNDANLIMYECK/SD/DZ078/21TKVGQEDVGNRKNDANLIMYECK/SD/TA803/22TKVGQEDVGNRKNDANLIMYECK/SD/TA1016/22TRVGQEDIGNRKNDANLIMYECK/SD/LY1437/23TKVGQEDVGNRKNDANLIMYE Key amino acid sites in the PB2, PB1, PB1-F2, PA, PA-X, NP, M1, M2, NS1 and NEP proteins of the 5 H9N2 AIVs identified in this study. In NP, the N319K substitution, which has been reported to enhance viral replication by modulating interaction with host importin-α (Mänz et al., 2013), was not detected, and all isolates carried residue N at position 319. In M1, mutations D30N and A215T in H5N1 viruses have been associated with virion morphology and reduced SUMOylation (Guo et al., 2022). In this study, all isolates possessed residue N at position 30, and no A215T substitution was observed. In M2, all isolates carried the S31N substitution, which is known to confer resistance to amantadine and rimantadine (Belshe et al., 1988). In NS1, substitutions F103L and M106I, which have been shown to enhance viral replication and virulence (Dankar et al., 2011), were present in all isolates. In NEP, mutations M16I, Y41C, and E75G, which have been reported to increase polymerase activity and replication in mammalian cells (Reuther et al., 2014), were not detected in any of the isolates (Table 3). Pathogenicity in mice In the mouse infection study, none of the three viruses (CK/SD/HZ056/21, CK/SD/TA1016/22, and CK/SD/LY1437/23) caused significant body weight loss, indicating low pathogenicity in mice (Fig. 4A). Viral titers were measured in the lungs and nasal turbinates: CK/SD/HZ056/21 had mean titers of 3.92 ± 0.62 log10 EID50 in the lungs and 5.17 ± 0.33 log10 EID50 in the nasal turbinates; CK/SD/TA1016/22 had 4.54 ± 0.34 log10 EID50 in the lungs and 5.79 ± 0.62 log10 EID50 in the nasal turbinates; CK/SD/LY1437/23 had 3.29 ± 0.53 log10 EID50 in the lungs and 5.01 ± 0.33 log10 EID50 in the nasal turbinates (Fig. 4B). Overall, viral replication was higher in the upper respiratory tract (nasal turbinates) than in the lungs, with CK/SD/TA1016/22 showing slightly higher titers in both tissues compared with the other two viruses. Histopathological analysis showed that all three viruses caused pulmonary capillary congestion, thickening of alveolar walls, inflammatory cell infiltration, and interstitial edema. The lung lesions in mice infected with CK/SD/TA1016/22 were more severe than those caused by the other two viruses. Pathogenicity of H9N2 AIV isolated in mice. A. Body weight changes of infected mice, expressed as a percentage relative to the initial body weight. B. Virus titers in lung and nasal turbinate tissues. The mean virus titers in lung and nasal turbinate are shown as log10 EID50 (mean ± SD, n = 3). Overall, virus replication was higher in the upper respiratory tract (nasal turbinate) than in the lung. C. Histopathological changes in the lung (HE staining, × 200). Infection with all three viruses induced capillary congestion, alveolar wall thickening, inflammatory cell infiltration, and interstitial edema, with CK/SD/TA1016/22 causing the most severe lung lesions. Discussion Since the 1990s, H9N2 AIV has been continuously circulating worldwide and has become enzootic in poultry populations. In recent years, it has been frequently detected in growing chickens, laying hens, and indigenous breeds, causing varying degrees of economic losses across multiple countries, particularly in China (Chen et al., 2024). Epidemiological surveillance indicates substantial differences in H9N2 prevalence between poultry farms and live poultry markets. For example, surveillance of live poultry markets in Hubei Province during 2013–2017 detected 641 H9-positive samples out of 4,798 (13.4%) (Hu et al., 2021), whereas a more recent study in Changsha (2021–2023) reported an overall AIV positivity rate of 68.45%, with H9 accounting for 53.51% (Zhang et al., 2025). In contrast, the prevalence in poultry farms is relatively lower, with reported rates of 8.1% (2019–2021, Shandong) (Zhao et al., 2023) and 3.2% (2022–2023, Shandong, Jiangsu, and Guangdong) (Song et al., 2025). In the present study, 10 H9N2 AIV–positive samples were detected from 192 farm samples (5.2%), consistent with previous findings. Collectively, these results indicate that H9N2 AIV prevalence is substantially higher in LPMs than in farms. This disparity is likely attributable to differences in biosecurity practices: farms generally implement stricter biosecurity and have limited contact with external poultry, whereas LPMs, characterized by high poultry turnover and mixed sources, facilitate viral amplification and persistence. In Shandong Province, H9N2 AIV remains endemic in LPMs, commercial farms, and surrounding environments. Given the large-scale poultry industry, frequent interregional movement, diverse indigenous breeds, and long-term vaccination pressure, these factors may collectively drive viral evolution and complicate H9N2 control. Consistent with previous studies across China, the G57 genotype has become the dominant lineage in poultry since approximately 2010, replacing earlier genotypes (Hu et al., 2021; Zhang et al., 2025; Zhao et al., 2023; Li et al., 2017; Wang et al., 2021; Chang et al., 2025). Phylogenetic analysis showed that the HA and NA genes of the isolates belonged to the Y280-like lineage, PB2 and M genes to the G1-like lineage, and the remaining internal genes to the F/98-like lineage, forming a genetic constellation typical of the G57 genotype. Notably, the M gene of strain CK/SD/TA803/22 showed high homology with H6N6 viruses and clustered closely with H9N2 lineages, suggesting a potential inter-subtype reassortment event. Previous studies have demonstrated that H9N2 viruses serve as important gene donors, contributing internal genes—particularly those of G1-like origin—to other subtypes such as H5N1, H7N9, and H10N8, including zoonotic reassortants (Pu et al., 2017; Peacock et al., 2019). Similar reassortment events have also been reported in countries such as India and Nigeria (Tosh et al., 2016; Monne et al., 2015), indicating that H9N2 participates in a complex global gene exchange network. In addition to acting as donors, H9N2 viruses can also acquire external gene segments; for instance, strains circulating in Pakistan and Bangladesh have incorporated genes from H7N3 and H5N1 viruses, and polymerase genes derived from H5N1 have also been identified in Chinese H9N2 isolates (Peacock et al., 2019). These findings collectively highlight the bidirectional gene exchange and adaptive evolution of H9N2 viruses across different ecological settings. Pathological analysis showed that H9N2 AIV infection primarily induced respiratory lesions, including tracheal mucosal congestion, hemorrhage, epithelial shedding, and pulmonary inflammation, consistent with previous reports (Begum et al., 2023; Bóna et al., 2023). In addition, lesions were observed in multiple organs, including the kidney, spleen, proventriculus, and brain. Renal lesions were characterized by degeneration of renal tubular epithelial cells and interstitial congestion. The frequency of renal involvement varies considerably among studies, which is likely dependent on the tissue tropism of the H9N2 strains used for experimental infection (Bóna et al., 2023; Liu et al., 2023). Splenic lymphocyte depletion and necrosis suggested immunosuppression, in agreement with earlier studies (Qiang et al., 2011; Degen et al., 2006). Perivascular lymphocytic cuffing in the brain indicated a potential for central nervous system involvement, as reported in recent studies (Zhang et al., 2021; Kye et al., 2021). Mild epithelial damage in the proventriculus suggested limited gastrointestinal involvement (Kye et al., 2021). These findings, together with recent reports, indicate that contemporary H9N2 viruses are not strictly restricted to the respiratory tract but may exhibit expanded tissue tropism and systemic pathogenicity. Molecular characterization further provided mechanistic insights into viral phenotype and zoonotic potential. The presence of the HA Q226L mutation, along with I163T (in four isolates) and H183N (in all isolates), suggests increased affinity for human-like receptors (Vines et al., 1998; Li et al., 2014; Sang et al., 2015; Zou et al., 2019;Li et al., 2014;Weis et al., 1988). The loss of the glycosylation site at position 218, while retaining that at 313, may alter antigen exposure and immune recognition (Peng et al., 2019). In the polymerase complex, PB2 A588V and PA K356R mutations, identified in all isolates, have been associated with enhanced polymerase activity and virulence in mammals (Xiao et al., 2016;Xu et al., 2016). In addition, NS1 F103L and M106I substitutions may contribute to increased viral replication and pathogenicity (Dankar et al., 2011). Notably, the mammalian-adaptive mutation PB2 K526R was detected only in strain CK/SD/TA1016/22, consistent with its higher replication efficiency in mice, suggesting potential synergistic effects of key mutations (Song et al., 2014). All isolates carried the M2 S31N mutation, indicating widespread resistance to adamantanes (Belshe et al., 1988). Although no known neuraminidase inhibitor resistance mutations were identified (Pizzorno et al., 2011), the emergence of a novel glycosylation site (position 306) in strain CK/SD/TA803/22 warrants further surveillance. Vaccination remains a primary control strategy for H9N2 AIV in many countries, including China, Israel, South Korea, Morocco, Pakistan, Egypt, Iran, and the United Arab Emirates (Peacock et al., 2019). However, antigenic drift and delayed vaccine updates have led to widespread antigenic mismatch between circulating and vaccine strains, facilitating viral persistence in vaccinated flocks. In this study, multiple amino acid substitutions were identified at key antigenic sites (positions 90, 145, 153, 167, 168, 200, and 201), including E90, D145, G167, R200, A168N, and N201A, indicating ongoing antigenic variation. These mutations may alter antigenicity and reduce vaccine-induced immune recognition, thereby promoting immune escape. Consistent with previous reports, long-term vaccination pressure likely contributes to antigenic drift (Kye et al., 2021), highlighting the necessity of continuous antigenic surveillance and timely vaccine strain updates. In the mouse model, all isolates were non-lethal but replicated efficiently in the lungs and nasal turbinates, with higher replication in the upper respiratory tract. Strain CK/SD/TA1016/22 induced more severe pulmonary lesions, likely associated with the PB2 K526R mutation, further supporting the role of PB2 in modulating virulence. From a global poultry health perspective, although H9N2 AIV is generally classified as low pathogenic, its sustained circulation, genetic plasticity, and capacity for cross-species adaptation underscore its potential risk. Continuous monitoring of pathogenicity and evolution is therefore essential for global risk assessment and disease control strategies. This study has several limitations. First, although sampling spanned three years, it was geographically restricted to Shandong Province and may not fully represent the genetic and phenotypic diversity of H9N2 AIV in China or globally. Second, only five representative strains were selected for detailed analysis, which may underestimate population-level variability. Third, the mouse model primarily reflects replication and mammalian adaptation but does not directly assess transmissibility, which requires more suitable models such as ferrets or guinea pigs. Finally, although several molecular markers associated with pathogenicity and host adaptation were identified, their individual and synergistic effects require validation using reverse genetics. Future studies should integrate genomic surveillance with functional characterization, including reverse genetics and antigenic analyses, to better understand H9N2 evolution and its potential public health risks. Funding This research received no external funding. Data availability statement The data that support the findings of this study are included within the article. Ethics approval and consent to participate During the mouse experiments, we strictly followed the regulations of the Regulations on the Administration of Laboratory Animals. In addition, the animal experimental protocols were approved by the Animal Ethical and Experimental Committee of Shandong Agricultural University (Ethics Approval Number: 2015–012). CRediT authorship contribution statement Wei Zhang: Conceptualization, Investigation, Writing – original draft. Libing Sun: Data curation, Investigation. Qian Wang: Data curation, Investigation. Chenlu Xia: Data curation, Investigation. Yuzhong Zhao: Writing – original draft, Software, Funding acquisition, Conceptualization, Writing – review & editing. Disclosures The authors declare that they have no conflicts of interest. Acknowledgements We gratefully acknowledge Sun Honglei (China Agricultural University, Beijing, China) for his valuable guidance and support in the experimental work. Footnotes Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.psj.2026.106936. Supplementary materials References Associated Data Supplementary Materials Data Availability Statement The data that support the findings of this study are included within the article.