From November 2023 to February 2024, during the winter migration season of whooper swans, we collected a total of 945 fresh fecal samples from whooper swans in Pinglu Wetland, Shanxi Province (34° 49′ 51″ N, 111° 8′ 53″ E). Each fresh fecal drop was collected and placed into 2 mL of PBS supplemented with penicillin and streptomycin. The samples were vortexed, centrifuged, and inoculated into 10-day-old chicken embryos for virus isolation in an ABSL-2 laboratory at Shanxi Agricultural University. The allantoic fluid was harvested after 72 h and tested via a hemagglutination (HA) assay. The HA subtype was determined via the HI test, whereas the NA subtype was identified via PCR and sequencing as described previously (Liang et al., 2024). The virus A/swan/Shanxi/2143/2023 (H12N2) used in this study was isolated from fresh fecal samples of whooper swans collected in Shanxi Province.

RNA from the H12N2 virus was extracted using TRIzol reagent, followed by the synthesis of complementary DNA (cDNA) via reverse transcription with the Uni12 primer (AGCAAAAGCAGG). The eight segments of the H12N2 virus were amplified via PCR using specific primers. Sequencing was conducted by Sangon Biotech Co., Ltd. (Shanghai, China). The nucleotide sequence data were subsequently edited using the SeqMan program (DNAstar, v7.1.0). Molecular markers for each segment were identified using the MegAlign program (DNAstar, v7.1.0).

Markov chain Monte Carlo (MCMC) trees with molecular clocks of the HA and NA genes of H12N2 viruses and their internal genes (PB2, PB1, PA, NP, M, and NS) were constructed using BEAST (v1.10.4) (Suchard et al., 2018). Briefly, the sequences were aligned via MAFFT (v7.453) (Katoh and Standley, 2013), and duplicate or highly similar sequences were removed using BioAider (v1.423) (Zhou et al., 2020). The model of nucleotide substitution was selected in ModelFinder of IQ-tree, and a relaxed log-normal clock was set in the BEAUti program. The best-fitting nucleotide models GTR + F + I + R3 for the HA data and TVM + F + G4 for the NA data were selected. The MCMC chain was executed for 20 million steps and sampled at 10,000-step intervals to generate a BEAST file. Phylogenetic trees were constructed for the six internal genes (PB2, PB1, PA, NP, M and NS) using the same methodology, with the best model used for each gene. Tracer (v1.7.1) was used to ascertain whether the parameters converged (effective sample size values ≥200). A target tree was obtained by selecting the tree with the largest posterior probability with a 10% burn-in using the Tree Annotator program.

The growth kinetics of the H12N2 virus were determined in Madin–Darby canine kidney (MDCK) cells and chicken embryo fibroblast (DF1) cells. Briefly, confluent monolayers of cells were infected with H12N2 virus (10⁴ EID₅₀/well) and incubated at 37°C. The cell supernatants were collected at 12, 24, 36, and 48 h post infection (h.p.i.), the viral titers were determined in chicken embryos, expressed as 50% egg infectious dose (EID₅₀). The experiments were performed in triplicate.

Three five-week-old ducks were inoculated intranasally with 10⁶ EID₅₀ of the H12N2 virus in a volume of 200 μL. On day 3 post infection (p.i.), the ducks were euthanized, and tissue samples, including trachea, lung, pancreas, liver, spleen, kidney, intestine, brain, and bursa of Fabricius, were collected for viral titration in chicken eggs. To evaluate the transmission of the virus in ducks, three ducks were inoculated intranasally with 10⁶ EID₅₀ of the virus in a volume of 200 μL. An additional three naive ducks were placed in the same isolator at 24 h.p.i. to serve as the contact group. Oropharyngeal and cloacal swabs were collected from both the inoculated and contact ducks on days 1, 3, 5, 7, 9, and 11 p.i. Viral titers from the swabs were determined via titration in chicken embryos. The sera of the inoculated and contact ducks were collected on days 10, 15, and 21 p.i., and the antibody titers were assessed via the hemagglutination inhibition (HI) assay.

Specific pathogen-free (SPF) chickens (5 weeks old) were purchased from Beijing Boehringer Ingelheim Vital Biotech Co., Ltd. (Beijing, China). Three chickens were inoculated intranasally with 10⁶ EID₅₀ of the H12N2 virus in a volume of 200 μL. On day 3 p.i., the chickens were euthanized, and tissue samples, including trachea, lung, pancreas, liver, spleen, kidney, intestine, brain, and bursa of Fabricius, were collected for viral titration in chicken embryos. To evaluate the transmission of the virus in chickens, three chickens were inoculated intranasally with 10⁶ EID₅₀ of the virus in a volume of 200 μL. An additional three naive chickens were placed in the same isolator at 24 h.p.i. to serve as the contact group. Oropharyngeal and cloacal swabs were collected from both the inoculated and contact chickens on days 1, 3, 5, 7, 9, and 11 p.i. Viral titers from the swabs were determined through titration in chicken embryos. The sera of the inoculated and contact chickens were collected on days 10, 15, and 21 p.i., and the antibody titers were assessed via the HI assay.

Five-week-old BALB/c mice were obtained from Vital River Laboratory (Beijing. China). To evaluate the replicability of the virus in mice, groups of eleven BALB/c mice were mildly anesthetized with CO₂ and inoculated intranasally with 10⁶ EID₅₀ H12N2 virus in a volume of 50 μL. Three mice were euthanized on days 3 and 5 p.i., and the brain, nasal turbinate, spleen, kidney, and lung were collected and titrated in chicken embryos. A portion of the lung tissue was collected for histopathological observation. The remaining five mice were continuously monitored for body weight changes and survival for 14 days. Five mice inoculated with PBS were used as the control group to observe body weight changes.

H12 influenza viruses have been detected in wild birds since 1975; however, the epidemiology and ecology of H12 viruses remain poorly understood. To elucidate the epidemiology of H12 AIVs, we categorized the available sequence data of H12Nx (N1–N9) viruses from the GISAID and NCBI databases. After excluding duplicated and highly similar entries, we identified a total of 516 H12Nx isolates up to November 22, 2024. Statistical classification of H12Nx virus subtypes isolated between 1975 and 2024 revealed nine distinct combinations of HA and NA subtypes. H12N5 (n = 316) was identified as the predominant subtype, whereas H12N2 (n = 30) was relatively less common. The other identified subtypes included H12N1 (n = 28), H12N3 (n = 21), H12N4 (n = 50), H12N6 (n = 16), H12N7 (n = 13), H12N8 (n = 16), and H12N9 (n = 8) (Figure 1A). Notably, H12 viruses were highly prevalent among Anas (n = 186) and Arenaria (n = 208) (Figure 1B). Moreover, H12N2 viruses have been identified in various wild bird hosts (Figure 1C). Geographically, H12 isolates were predominantly found in North America (n = 354), followed by Asia (n = 67), Europe (n = 58), Australia (n = 21), South America (n = 8), Central America (n = 6) and South Africa (n = 2) (Figure 1D). These data not only enhance our understanding of H12 virus epidemiology but also facilitate improved monitoring of wild bird populations.

The Yellow River Wetland of Pinglu (Shanxi Province) is located along the East Asian–Australasian (EA) flyway of migratory birds. In recent years, approximately 8,000 whooper swans (Cygnus cygnus) have been observed wintering in this area each year. We conducted surveillance for AIVs from November 2023 to February 2024 in the Yellow River Wetland of Pinglu. A total of 945 fecal samples were collected and inoculated into embryonated chicken eggs. From these samples, one H12N2 virus was successfully isolated and designated A/swan/Shanxi/2143/2023 (H12N2), abbreviated as SX/2143. The entire genome of the virus was sequenced, and the sequence data were submitted to the GISAID database (accession number: EPI3408601-EPI3408608).

On the basis of the complete genomic sequence of the SX/2143 virus, we analyzed the molecular characteristics of key amino acid sites in its viral proteins. These results indicate that the amino acid motif PQVQNR↓GLF is present at the cleavage site of the HA protein, reflecting typical features associated with low pathogenicity in poultry. Several amino acid substitutions were observed in the identified strain SX/2143, including PB1-D3V, H436Y and D622G, PA-A515T, M1-N30D, T215A, NS1-I106M, and C138F. These substitutions suggested that the H12 virus may have acquired a certain degree of adaptability to mammalian hosts. Additionally, the mammalian host adaptive mutations PB2-E627K and D701N were not detected in SX/2143 (Table 1).

To investigate the genetic evolution of the H12N2 virus isolated from the whooper swan in this study, we performed an evolutionary analysis of the HA and NA genes. A phylogenetic tree for the SX/2143 virus was constructed using reference sequences retrieved from the public influenza virus database. As shown in Figure 2, the H12 subtype HA gene can be divided into two lineages: the North American lineage and the Eurasian lineage. Most of the lineage types of the isolated strains are associated with their geographical regions; however, there were exceptions, such as certain Asian isolates whose HA genes belong to the North American lineage (Figure 2). The HA gene of SX/2143 isolated in this study belongs to the Eurasian lineage and is closely related to the H12N5 strain isolated from a duck in the Taiwan region (Figure 2). The nucleotide homology with the isolate exhibiting the highest homology reached 98.23% (Supplementary Table S1). Phylogenetic analysis of the NA genes of HxN2 subtypes revealed that the N2 genes can also be divided into North American and Eurasian lineages, with the N2 genes of the H12N2 virus forming several distinct subbranches. Notably, the N2 gene of the SX/2143 strain isolated in this study was most closely related to the N2 gene of a duck-derived H9N2 strain, suggesting that the N2 gene may have originated from the H9N2 virus (Figure 3). This relationship is supported by a nucleotide identity of 98.44%, with the isolate showing the highest homology (Supplementary Table S1).

We further constructed a phylogenetic tree of the internal genes (PB2, PB1, PA, NP, M and NS) of SX/2143 to analyze the evolutionary process and potential origins of the H12N2 strain. The PB2 and PB1 genes were most closely related to the PB2 and PB1 genes of H3N8 strains isolated from the environment or from wild birds, respectively (Supplementary Figures S1A,B). The PA gene of the H12N2 virus was likely derived from the PA gene of waterfowl H2N3 strains (Supplementary Figure S1C). The NP gene is closely related to the H8 subtype strain derived from ducks, whereas the M gene shows the highest similarity to the M gene of H4N6 viruses isolated from mallards in China (Supplementary Figures S1D,E). The NS gene had the closest relationship with H1N1 strains from wild birds (Supplementary Figure S1F). The results indicated that the internal genes of the H12N2 virus have a complex origin. The H12N2 virus may have emerged through intricate reassortment events between H12 subtype viruses and other influenza virus subtypes carried by wild birds, either during their circulation among wild bird populations or through the migration of these birds.

Wild birds and waterfowl are the primary natural hosts of the H12N2 virus. However, the replication capacity and horizontal transmission potential of the H12N2 virus in ducks and chickens remain to be thoroughly investigated. We initially assessed the replication capacity of the whooper swan-derived H12N2 virus in ducks and evaluated its transmission dynamics. The SX/2143 virus did not cause mortality or noticeable clinical symptoms in ducks throughout the observation period. The virus was detected in the trachea, pancreas, and bursa of Fabricius in two inoculated ducks, with viral titers ranging from 0.75 to 2.25 log₁₀ EID₅₀/ml (Figure 4A). Additionally, the virus was detected in the lung, liver, spleen, and kidney of one of the three inoculated ducks, with titers ranging from 0.75 to 1.5 log₁₀ EID₅₀/ml. No virus was detected in the intestine or brain of the inoculated ducks (Figure 4A). In transmission experiments, oropharyngeal and cloacal swabs were collected into the same tube and titrated in chicken embryos, revealing that the virus was detected in two of the three inoculated ducks on days 3 and 7 p.i., with viral titers ranging from 0.75 to 1.25 log₁₀ EID₅₀/ml (Figure 4B). The virus was also detected in one duck from the contact group, with a viral titer of 0.75 log₁₀ EID₅₀/ml on day 7 post-contact (p.c.) (Figure 4B). Serum samples were collected from both the inoculated and contact groups of ducks on days 10, 15, and 21 p.i. HI tests were performed to determine the serum antibody titers. The results indicated that one duck from the inoculated group and one duck from the contact group exhibited seroconversion, with antibody titers reaching 1 log2 (Figure 4C). The infection and transmission experiments in ducks demonstrated that while the H12N2 virus can infect domestic ducks, it replicates poorly and shows limited potential for contact transmission.

We further assessed the ability of this whooper swan-derived H12N2 virus to infect and transmit in chickens. As shown in Figure 4D, the virus was detected in the trachea of all three inoculated chickens, with viral titers ranging from 1.25 to 1.75 log₁₀ EID₅₀/ml. Low viral titers were also detected in the lung, liver, spleen, bursa of Fabricius, and intestine, whereas no virus was detected in the pancreas, kidney, or brain. The results from the transmission study indicated that the virus was detected in swab samples from two of three chickens in both the inoculated and contact groups, with viral titers ranging from 0.75 to 1.5 log₁₀ EID₅₀/ml (Figure 4E). Notably, the HI assay revealed that the sera from all inoculated and contact chickens tested positive for the H12N2 virus, with titers ranging from 1 log2 to 4 log2 on days 10, 15, and 21 p.i. (Figure 4F). In summary, the H12N2 virus derived from the whooper swan can replicate at low levels in both ducks and chickens, primarily within the respiratory system, including the lungs and trachea. In ducks, the H12N2 virus exhibited moderate transmission efficiency, with two out of three ducks showing evidence of transmission. However, in chickens, the virus has a high capacity for transmission, suggesting that the H12N2 virus may be better adapted to chickens than to ducks.

Currently, there is a lack of relevant reports regarding the replication ability and pathogenicity of the H12N2 virus in mammals. To evaluate the potential infection risk of this wild bird-derived H12N2 virus in mammals, we conducted a study using a mouse model. BALB/c mice were inoculated with the SX/2143 virus at a dose of 10⁶ EID₅₀, and organs were collected for viral titration on days 3 and 5 p.i. As shown in Figure 5, on days 3 and 5 p.i., the virus was detected in the nasal turbinate and lungs of the mice, with viral titers ranging from 0.75 to 2.25 log₁₀EID₅₀/ml (Figures 5A,B). No virus was detected in the spleen, kidneys, or brain. During the 14-day observation period, the mice inoculated with the H12N2 virus did not exhibit significant body weight loss, and no apparent clinical signs were observed (Figure 5C). Mild to moderate pathological lesions were observed in the lung tissue through hematoxylin and eosin (HE) staining (Figures 5D,E). These results suggest that the wild bird-derived H12N2 virus can replicate in the lungs and nasal turbinates of mice without prior adaptation.

To evaluate the replication capacity of the H12N2 virus at the cellular level in vitro, we conducted experiments using MDCK cells and DF1 cells, which are commonly used in influenza virus research. Following infection of MDCK or DF1 cells with the SX/2143 (H12N2) virus, supernatants were collected at 12, 24, 36, and 48 h.p.i. for viral titration. The results indicated that the virus exhibited poorer replication capacity in DF1 cells, and the viral titer increased from 2.17 log₁₀ EID₅₀/ml to 2.5 log₁₀ EID₅₀/ml from 12 to 48 h.p.i. (Figure 6). Notably, the H12N2 virus replicates efficiently in MDCK cells, with viral titers at various time points significantly higher than those in DF1 cells. In MDCK cells, the viral titer was 5.17 log₁₀ EID₅₀/ml at 12 h.p.i. and increased to 8.58 log₁₀ EID₅₀/ml at 48 h.p.i. (Figure 6). These results indicated that the H12N2 virus replicates efficiently in MDCK cells, suggesting that the virus has the potential for cross-species transmission to infect mammals.