During monthly surveillance of swine IAVs in China during October–December 2020, we collected a total of 376 nasal swab samples from apparently healthy pigs in a slaughterhouse accommodating swine from neighboring regions (Jiangsu, Shandong, and Anhui Provinces in eastern China). We detected H1 subtype swine influenza virus in 9 of those by real-time reverse transcription quantitative PCR; 2 were confirmed as hemagglutinin (HA) positive after inoculating into MDCK cells (12). We further evaluated these 2 swine IAV isolates, A/swine/Jiangsu/HD11/2020 (H1N1) [HD11] and A/swine/Anhui/HD21/2020 (H1N1) [HD21], for their genetic and biologic characteristics.

The genome sequences of HD11 and HD21 deposited in the GenBank database (accession no. OL744678–93) shared 95.4%–99.0% nucleotide identities across the coding regions of 8 genes. We performed searches of those sequences on BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi) and the GISAID database (http://platform.gisaid.org) to present a more comprehensive scene of the homologous reference influenza viruses. As shown by the closest BLAST hits (Table 1), HD11 and HD21 were not only highly related to swine origin IAVs collected during 2012–2018 but also remarkably similar to contemporaneous human H1N1 isolates from 2020 and 2021.

Figure 1

Figure 1. Receptor-binding property of 2 G4 Eurasian avian-like influenza A(H1N1) swine isolates from pigs in China. A) The control virus A/mallard/Huadong/S/2005(H5N1) (HDS05) showed an absolute preference for avian-type SAα-2,3Gal. B) The...

We constructed a phylogenetic gene tree analysis with H1N1 reference strains to confirm the intimate genetic relationship between these 2 swine IAVs and human viruses (Appendix Figure 1). In each tree, HD11 consistently clustered with 3 human H1N1 viruses, A/Tianjin/00030/2020(H1N1), A/Shandong/00204/2021(H1N1), and A/Sichuan/01208/2021(H1N1). As for HD21, the virus aggregated closely with the HD11-involved subbranch in PB2, HA, NP, NA, and M gene trees but gathered more intimately with another 3 human H1N1 viruses containing A/Hubei-Wujiagang/1324/2020(H1N1), A/Gansu-Xifeng/1143/2021(H1N1) and A/Gansu-Xifeng/1194/2021(H1N1) in PB1, PA, and NS gene trees. Taken together, HD11 and HD21 were both closest to contemporaneous human H1N1 strains, and they uniformly possessed the EA H1N1-like HA and NA genes, pandemice influenza–like RNP (PB2, PB1, PA, and NP) and M genes, and TR-like NS gene that made the G4 type gene constellation. We observed that 2 additional swine reference viruses of A/swine/Shandong/LY142/2017(H1N1) and A/swine/China/Qingdao/2018(H1N1) assembled tightly with the HD11/HD21 cluster, further supporting the possibility of IAV interspecies transmission from swine to human.

The 2 G4 genotype EA H1N1 swine isolates both propagated well in specific-pathogen-free chicken embryos with virus titers per 0.1mL >9 log₁₀ 50% egg infectious dose (EID₅₀) (Table 2). However, HD11 replicated much better than HD21 in MDCK cells through the titration of the 50% culture infectious dose (TCID₅₀) value and virus growth at 12-hours intervals across 12–60 hours postinfection (hpi). At >24 hpi, HD11 had generated more than 5 log₁₀ TCID₅₀ and reached a peak of 7 log₁₀ TCID₅₀ at 36 hpi, whereas the titer of HD21 virus remained at the relatively lower level <5 log₁₀ TCID₅₀ until the endpoint.

Subsequently, we conducted a solid-phase direct binding ELISA assay with the synthetic glycopolymer-based receptor mimics Neu5Aca2-3Galb1-4GlcNAcb(3ʹSLN)-PAA-biotin and Neu5Aca2-3Galb1-4GlcNAcb(6ʹSLN)-PAA-biotin (GlycoTech, https://www.glycotech.com) to evaluate the viral receptor-binding preference as previously described (13). We used 1 avian H5N1 virus and 1 human seasonal H3N2 virus as controls; the avian virus displayed a complete 3ʹsialyl-N-acetyllactosamine (SLN) affinity, whereas the human virus possessed a dual binding property to both 3ʹSLN and the more advantageous 6ʹSLN (Figure 1). Unlike HD21, which was endowed with comparable avidity between 3ʹSLN and 6ʹSLN, HD11 resembled the binding feature of the human-origin H3N2 virus that preferentially binds the human-type SAα-2,6Gal (Figure 1).

Figure 2

Figure 2. Pathogenicity of 2 G4 Eurasian avian-like influenza A(H1N1) swine isolates from pigs in China in BALB/c mice. A) Body weight change of infected mice. B) Survival curve of infected mice....

We then investigated the pathogenicity of HD11 and HD21 in mice. We infected 6-week-old BALB/c mice in groups of 5 intranasally with 10^6.0 EID₅₀ virus dose or mock-inoculated them with phosphate-buffered saline (PBS). We monitored body weight changes and clinical symptoms of the mice daily for 14 days. We humanely euthanized an additional 3 challenged mice per group and analzyed them for virus load in tissues at 3 and 5 days postinfection (dpi). Mice in the control group displayed a steady increase in body weight, the HD21 group experienced a slightly transient weight loss on 3 dpi, and all mice survived during the entire experiment (Figure 2). In contrast, HD11 resulted in a steady decrease in body weights starting at 1 dpi, and all died within 8 days. In addition, we observed that both HD11 and HD21 replicated efficiently in the lungs without prior adaptation and readily disseminated into nasal turbinates and the brain (Table 2). Of note, the virus load in respiratory tissues of HD11-infected mice was significantly higher (p<0.01 in lungs and p<0.001 in turbinates) than that of HD21-infected mice on 3 dpi. On 5 dpi, we observed no significant difference in virus titers in the 3 tissues of the mice infected with these 2 isolates. Moreover, HD11 infections increased the mRNA levels of inflammatory cytokines, including interleukin 6 and 10, interferon β and γ, MX1, and C-X-C motif chemokine ligand 10 11 on 3 dpi, 5 dpi, or both, more dramatically than HD21 virus. Both HD21 and HD11 infections increased tumor necrosis factor α expression at relatively low levels (Appendix Figure 2).

Homology alignment and phylogenetic tree construction analysis suggest that HD11 and HD21, two G4 EA H1N1 swine IAVs isolated in 2020 in China, are strongly related to recent human-origin EA H1N1 viruses. In particular, HD11 had higher affinity for human-type 6ʹSLN at the level that is equivalent to the human seasonal H3N2 virus. Moreover, HD11 replicated much faster in vitro in MDCK cells and in vivo in the lung than di HD21 and was highly pathogenic to BALB/c mice, as evidenced by its lethality, higher viral loads in pulmonary tissues, and higher levels of inflammatory cytokines in the lung. We propose that the HD11-like G4 swine isolates whose genomic sequences share great homology with that of contemporaneous human EA H1N1 viruses may lead to interspecies transmission. Therefore, the public health threat from the zoonotic G4 EA H1N1 viruses should not be underestimated.