Characterization of a human H3N8 influenza virus

Summary

Background

In 2022 and 2023, novel reassortant H3N8 influenza viruses infected three people, marking the first human infections with viruses of this subtype.

Methods

Here, we generated one of these viruses (A/Henan/4-10CNIC/2022; hereafter called A/Henan/2022 virus) by using reverse genetics and characterized it.

Findings

In intranasally infected mice, reverse genetics-generated A/Henan/2022 virus caused weight loss in all five animals (one of which had to be euthanized) and replicated efficiently in the respiratory tract. Intranasal infection of ferrets resulted in minor weight loss and moderate fever but no mortality. Reverse genetics-generated A/Henan/2022 virus replicated efficiently in the upper respiratory tract of ferrets but was not detected in the lungs. Virus transmission via respiratory droplets occurred in one of four pairs of ferrets. Deep-sequencing of nasal swab samples from inoculated and exposed ferrets revealed sequence polymorphisms in the haemagglutinin protein that may affect receptor-binding specificity. We also tested 90 human sera for neutralizing antibodies against reverse genetics-generated A/Henan/2022 virus and found that some of them possessed neutralizing antibody titres, especially sera from older donors with likely exposure to earlier human H3N2 viruses.

Interpretation

Our data demonstrate that reverse genetics-generated A/Henan/2022 virus is a low pathogenic influenza virus (of avian influenza virus descent) with some antigenic resemblance to older human H3N2 viruses and limited respiratory droplet transmissibility in ferrets.

Funding

This work was supported by the Japan Program for Infectious Diseases Research and Infrastructure (JP23wm0125002), and the Japan Initiative for World-leading Vaccine Research and Development Centers (JP233fa627001) from the Japan Agency for Medical Research and Development (AMED).

Research in context.

Evidence before this study

Prior to 2022, influenza viruses of the H3N8 subtype were not known to infect humans. However, in 2022 and 2023, three human infections with novel reassortant H3N8 viruses were reported, creating an urgent need to characterize these viruses. Until recently (Sun et al.¹), studies on novel reassortant H3N8 viruses had been limited to only those isolated from avian species; studies of viruses isolated from humans (which may possess human-adapting amino acid substitutions) were lacking.

Added value of this study

Here, we characterized a novel reassortant H3N8 virus isolated from an infected human, and we found that some people have cross-reactive antibodies against novel reassortant H3N8 viruses (perhaps conferred by a past infection with an H3 virus) contrary to what has been reported.

Implications of all the available evidence

Novel reassortant H3N8 viruses can transmit among ferrets via respiratory droplets, albeit with low efficiency. People, especially some older individuals, may have some protection against novel reassortant viruses.

Introduction

Influenza A viruses of the H3N8 subtype circulate in wild waterfowl, the natural reservoir of influenza A viruses, and have formed distinct virus lineages in horses and dogs. Despite the often close contact of people with dogs and horses, the equine and canine H3N8 viruses have not caused outbreaks in humans. Moreover, H3N8 viruses have been sporadically isolated from pigs, seals, and camels (GISAID.org; accessed on 09/09/23). In 2022, two human infections with novel reassortant H3N8 viruses were reported in young boys in China2, 3, 4 (https://www.who.int/emergencies/disease-outbreak-news/item/2022-DON378; accessed on 02/04/2023). Both children were hospitalized and received treatment including antivirals before recovering.^2,3 No H3N8 virus was isolated from close contacts of the two boys,² but two family members of one of the infected children seroconverted.⁴ Moreover, samples collected from a dog and a cat in the same household tested positive for H3N8 virus by RT-PCR.⁴ In March 2023, the first fatal human infection with a novel reassortant H3N8 virus was reported in a 56-year-old woman from Guangdong province, China, with multiple underlying conditions.^5,6

Sequence and phylogenetic analyses revealed that the novel reassortant human H3N8 viruses isolated from the infected children [A/Henan/4-10/2022 (also named A/Henan/4-10CNIC/2022); hereafter called A/Henan/2022 virus; and A/Hunan/CSKFQ-22-5/2022 (also named A/Changsha/1000/2022)] differ from the H3N8 viruses previously circulating in avian species.2, 3, 4^,7 The HA genes of the two human H3N8 viruses share high sequence homology (>98%) and belong to the Eurasian avian H3 lineage, whereas their NA genes are from the North American lineage.2, 3, 4^,7,8 Viruses with a similar NA gene circulated in China for several years.⁷ The remaining viral genes originated from chicken H9N2 viruses of the G57 genotype.2, 3, 4^,7,8 Retrospective analysis revealed that novel reassortant H3N8 viruses (whose HA and NA viral segments bear >98% sequence identify with those isolated from the infected children) had been circulating in live poultry markets and chicken farms in Hong Kong since late 2021.^3,7

The novel reassortant H3N8 viruses detected in poultry in Hong Kong and in the infected children encode a single basic amino acid at the HA cleavage site, indicating that they are low pathogenic avian influenza viruses. A/Hunan/CSKFQ-22-5/2022 encodes HA-226Q and -228G, which confer preferential binding to α2,3-linked sialic acids, the preferred receptor of avian influenza viruses. A/Henan/2022 virus also encodes HA-226Q; however, at position 228, either an unknown amino acid (‘X’) (GISAID EPI_ISL_12277126) or HA-228G/S have been reported.^3,9 Binding assays demonstrated that the novel reassortant H3N8 viruses bind to α2,3- and α2,6-linked sialic acids,^3,10,11 albeit with higher affinity to the former.³ At amino acid 627 of the PB2 protein, A/Henan/2022 virus encodes lysine, a known determinant of mammalian adaptation.^12,13 In contrast, A/Changsha/1000/2022 encodes valine, an amino acid rarely found at this position but known to increase viral growth kinetics compared with glutamic acid, the amino acid commonly found in avian influenza viruses at this position.¹⁴

Novel reassortant H3N8 viruses isolated from chickens or wild birds replicate efficiently in Madin–Darby canine kidney (MDCK) cells and some replicate in human A549 cells.^10,11 Infection of mice with these viruses caused no or moderate weight loss and no mortality.^10,11 Virus titres in the lungs of infected mice ranged from undetectable to 5.5 log₁₀ TCID₅₀/ml on Day 3 post-infection^10,11 and virus replication was efficient in the nasal turbinates of the infected mice.¹⁰ A recent study also found that novel reassortant H3N8 viruses isolated from chickens transmitted in ferrets via direct contact or respiratory droplets, albeit with low efficiency.¹⁵ While this manuscript was in preparation, Sun et al.¹ reported that the authentic human A/Henan/4-10/2022 isolate is highly virulent in mice and ferrets and transmits efficiently among ferrets via respiratory droplets. A reverse genetics-generated variant possessing the same HA amino acid sequence used in our study also transmitted efficiently to exposed ferrets via respiratory droplets.¹ Here, we characterized a reverse genetics-generated A/Henan/2022 virus in vitro and in mice and ferrets; in our experimental settings, the virus transmitted via respiratory droplets among ferrets but less efficiently than reported by Sun et al.¹

Methods

Cells and viruses

To generate A/Henan/2022, we downloaded the virus sequences from GISAID (EPI_ISL_12277126). At amino acid position 228 of HA (mature H3 HA numbering), a degenerate codon (X) is listed. Based on other publicly available sequences of novel reassortant H3N8 virus isolates from humans (GISAID.org; accessed on 03/17/2023), we synthesized an HA gene encoding glycine at HA-228. A/Henan/2022 virus was generated by using reverse genetics,¹⁶ virus stocks were generated in MDCK cells, and the sequence identity of all viral RNA segments was confirmed by Sanger sequencing. Other viruses and sera used in this study are summarized in Supplementary Table S1. MDCK cells were grown in Eagle’s minimal essential medium (MEM) containing 5% newborn calf serum. Human embryonic kidney (HEK) 293T cells were maintained in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum (FCS). hCK cells (i.e., MDCK cells modified to express increased levels of α2,6-linked and reduced levels of α2,3-linked sialic acids¹⁷) were maintained in MEM containing 5% newborn calf serum in the presence of 2 μg/ml puromycin and 10 μg/ml blasticidin. Human lung carcinoma epithelial A549 and chicken embryo fibroblast (CEF) cells were grown in DMEM/Ham’s F12 medium containing 10% FCS. Normal human bronchial epithelial (NHBE) cells (Lonza, Walkersville, MD) were maintained in bronchial epithelial cell growth medium (containing supplements and growth factors, Lonza, Walkersville, MD). All cells were incubated at 37 °C with 5% CO₂.

Growth kinetics of viruses in cell culture

NHBE, MDCK, A549, and CEF cells were infected in triplicate with virus at a multiplicity of infection of 0.001. The inoculum was removed after 1 h of incubation at 37 °C. MDCK, A549, and CEF cells were washed twice, and then incubated at 33 °C and 37 °C (for mammalian cells) or 37 °C and 39 °C (for avian cells) with 1× MEM containing 0.3% bovine serum albumin and 1 μg/ml (0.6 μg/ml for CEF cells) L-(tosylamido-2-phenyl) ethyl chloromethyl ketone (TPCK)-treated trypsin (Worthington Biochemical Corp, NJ). For NHBE cells, the medium was replaced with the bronchial epithelial growth medium containing 0.5 μg/ml TPCK-treated trypsin after removing the inoculum and washing twice. Samples were collected at 12, 24, 36, 48, 60, and 72 h post-infection. Virus titres at the indicated timepoints were determined by performing plaque assays in MDCK cells.

Infection of mice and ferrets

Under isoflurane anesthesia, 15 six-week-old female BALB/c mice were infected with 10⁶ plaque-forming units (pfu) of A/Henan/2022 virus in 50 μl. On Days 3 and 6 post-infection, five animals each were euthanized and various organs (nasal turbinates, tracheas, lungs, brains, spleens, and kidneys) were collected for virus titration. The remaining animals were monitored daily for 14 days for body weight changes.

For replication studies in ferrets, eight 6-month-old female ferrets [confirmed to be serologically negative to the following human and H3N8 influenza viruses: A/Hawaii/70/2019 (H1N1), A/Wisconsin/588/2019 (H1N1), A/Bangladesh/911009/2020 (H3N2), B/Washington/02/2019, B/Phuket/3073/2013, and A/Henan/2022(H3N8); see Supplementary Table S1] were anaesthetized intramuscularly with ketamine and dexmedetomidine (4-5 mg/kg and 10-40 μg/kg of body weight, respectively) and infected with 10⁶ pfu of A/Henan/2022 virus in 500 μl. Four animals each were euthanized on Days 3 or 6 post-infection and various organs (nasal turbinates, tracheas, lungs, brains, spleens, kidneys, hearts, small intestines, and large intestines) were collected for virus titration. Organs and 1 ml of MEM medium containing 0.3% BSA were homogenized at 1850 rpm for six cycles (ON: 6 s; OFF: 4 s) in a multi-beads homogenizer (YASUI KIKAI Corporation). The homogenized samples were then centrifuged at 4500 rpm for 10 min and the supernatant was transferred to a new tube for titration. Body temperature and weight were monitored daily until the animals were euthanized. Potential confounders such as the animal and/or cage location were not controlled for.

Virus transmission studies in ferrets

Ferrets were infected with 10⁶ pfu of A/Henan/2022 virus or A/Isumi/UT-KK001-1/2018 (H1N1pdm09; A/Isumi/2018) in 500 μl (infected ferrets). One day later, four naïve ferrets (exposed ferrets) were each placed in a cage adjacent to an infected ferret. The cages housing the infected or exposed ferrets were separated by about 5 cm. The transmission study was carried out under controlled conditions of 20–25 °C and relative humidity of 38.4% ± 8.8%. The airflow was from the front to the back of the isolator; thus, the airflow direction was perpendicular to the direction of virus transmission between the ferrets (for additional details, see¹⁸). Potential confounders such as the animal and/or cage location were not controlled for.

Nasal swabs were collected from infected ferrets on Day 1 after inoculation and from exposed ferrets on Day 1 after exposure, and then every other day for virus titration. The swabs were pre-soaked in phosphate-buffered saline (PBS), inserted into the ferret’s nasal cavity, and then placed in a tube containing 1.0 ml of DMEM with 50 U/ml penicillin and 50 μg/ml streptomycin and vortexed for 1 min.

Plaque assay

Confluent MDCK cells in 24-well plates were infected with a series of 10-fold dilutions of virus. After a 1-h incubation, the viral inoculum was removed, and the cells were washed once and overlaid with 1% agarose-containing MEM containing 0.3% bovine serum albumin in the presence of TPCK-treated trypsin. The plates were incubated for 2–3 days; then, the agar overlay was removed and the cells were fixed and stained with 20% methanol and crystal violet. Visualized plaques were counted to determine the number of plaque-forming units per ml.

Hemagglutination inhibition (HI) assay

Serum samples were treated with receptor-destroying enzyme (Denka Seiken Co., Ltd., Tokyo, Japan) at 37 °C for 18–20 h, followed by heat inactivation at 56 °C for 1 h and adsorption with turkey red cells for 1 h at room temperature. The treated sera were serially 2-fold diluted with PBS in 96-well V-bottom microtitre plates and mixed with the amount of virus containing four hemagglutination units, followed by incubation at room temperature (25 °C) for 30 min. After the addition of 50 μl of 0.5% turkey erythrocytes to the mixtures, the samples were gently mixed and incubated at room temperature for 1 h. HI titres were calculated as the inverse of the highest antibody dilution that inhibited hemagglutination.

Microneutralization assay

Human sera treated with receptor-destroying enzyme were two-fold serially diluted and mixed with 100 pfu of the respective virus (the virus titre was confirmed by back titration). After incubation at 37 °C for 1 h, the virus/serum mixtures were transferred to MDCK or ‘humanized’ MDCK cells (which express high levels of α2,6-linked sialic acids and low levels of α2,3-linked sialic acids),¹⁷ and the cells were further incubated for 5 days at 33 °C. The neutralization titres were determined as the highest serum dilution that completely prevented cytopathic effects.

Deep sequencing

The complete genomic sequence of the A/Henan/2022 virus stock, and of nasal swab samples obtained from infected and exposed ferrets, were determined by RT-PCR amplification of extracted viral RNA with a mixture of oligonucleotides.

HFadapter (5′- TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGAGCAAAAGCAGG), HF (5′-TGTATAAGAGACAGAGCAAAAGCAGG), HRadapter (5′- GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGAGTAGAAACAAGG), and HR (5′- TGTATAAGAGACAGAGTAGAAACAAGG).¹⁹ The amplified cDNA libraries were fragmented using a Nextera XT kit (Illumina), and barcoded using IDT for Illumina DNA/RNA UD Indexes (Illumina), according to manufacturer’s instructions. DNA libraries were sequenced on a MiSeq System (Illumina) using MiSeq Reagent Kit v3 (600-cycle) cartridges (Illumina). Initial trimming and filtering of reads was performed using Local Run Manager Version 3.0.0 (Illumina).

Demultiplexed deep sequence reads were assembled and variants were counted using IRMA v. 1.0.2.^20,21 Primary assemblies were used for internal genes; secondary assemblies were used for HA and NA. The default FLU configuration was used except that read sorting was conducted with LABEL, no reference elongation was conducted, and the secondary assembly residual assembly factor was set to 400. Preprocessing, running IRMA, and downstream analysis was conducted using version v1.0.0 of a snakemake workflow available at: https://github.com/IRI-UW-Bioinformatics/flu-ngs/releases/tag/v1.0.0.

Ethics statement

This manuscript was reviewed by the University of Wisconsin–Madison Dual Use Research of Concern (DURC) Subcommittee. This review was conducted in accordance with the United States Government September 2014 DURC Policy. The DURC Subcommittee concluded that the studies described herein do not meet the criteria of Dual Use Research of Concern (DURC). All animal experiments and procedures were approved by the Institutional Care and Use Committee of the University of Wisconsin (UW)-Madison School of Veterinary Medicine (protocol #V006426) and followed relevant institutional and American Veterinary Association guidelines.

Role of funders

The funders of the study had no role in the study design, data collection, analysis, interpretation, or writing of this report.

Results

A/Henan/2022 virus replication is attenuated in human cells

Based on published sequences (GISAID EPI_ISL_12277126), DNAs complementary to the A/Henan/2022 viral RNAs were synthesized and cloned into plasmid vectors for reverse genetics.¹⁶ As described in Materials and Method, we synthesized an A/Henan/2022 virus HA cDNA encoding HA-228G. A/Henan/2022 virus was generated by using established reverse genetics protocols.¹⁶ A virus stock was generated and titrated in MDCK cells (virus titre: 4.5 × 10⁷ pfu/ml).

To assess the growth properties of A/Henan/2022 virus in mammalian and avian cells, NHBE (primary human), MDCK (canine), A549 (human), and CEF (chicken) cells were infected at a multiplicity of infection of 0.001 and incubated at 33 °C and 37 °C for NHBE, MDCK, and A549 cells (to mimic the temperature in the human upper and lower respiratory tract, respectively), or at 37 °C and 39 °C for CEF cells (to mimic the temperature of avian species) for the indicated periods of time (Fig. 1). The human H3N2 influenza virus replicated efficiently in mammalian cells but showed delayed replication in avian cells at 37 °C and replicated to low titres in avian cells at 39 °C. The avian H3N8 influenza virus did not replicate efficiently in human A549 cells, in MDCK cells at 33 °C, or in NHBE cells at 33 and 37 °C, but efficient replication was detected in avian CEF cells. A/Henan/2022 virus grew efficiently in MDCK cells, but replicated less efficiently than the human H3N2 virus in human NHBE and A549 cells. In CEF cells, the replication kinetics of A/Henan/2022 virus were in between those of the human H3N2 and the avian H3N8 virus.

Fig. 1.

A/Henan/2022 virus replication in mammalian and avian cells. NHBE, MDCK, A549, and CEF cells were infected with A/Henan/2022 (H3N8), human A/Netherlands/938/1992 (H3N2). Or avian A/blue-winged teal/Iowa/10OS24111/2010 (H3N8) virus at a multiplicity of infection of 0.001. Cells were incubated at 33 °C and 37 °C (for mammalian cells) or 37 °C and 39 °C (for avian cells). At the indicated timepoints, aliquots of supernatant were collected, and virus titres were assessed in MDCK cells. The limit of detection (10 PFU/ml) is indicated by the dashed line. The values presented are the averages ± SD. P values were calculated by using a two-way ANOVA with multiple comparisons (GraphPad Prism version 8.0.0 for Windows; GraphPad Software). Asterisks indicate statistically significant differences between human or avian and H3N8 virus: ∗, P < 0.1; ∗∗, P < 0.01; ∗∗∗, P < 0.001; ∗∗∗∗, P < 0.0001.

A/Henan/2022 virus replication is robust the respiratory tract of mice

To test the pathogenicity of A/Henan/2022 virus in mammalian animal models, we intranasally infected 15 six-week-old female BALB/c mice with 10⁶ pfu of virus. On Days 3 and 6 post-infection, five animals each were euthanized to assess virus titres in various organs; the remaining animals were monitored daily for 14 days for weight loss (Fig. 2) and other signs of disease including ruffled fur, hunched posture, or inappetence. Infection with 10⁶ pfu of A/Henan/2022 virus caused transient weight loss of about 20% in most animals (Fig. 2a); one mouse was euthanized on Day 8 due to weight loss of over 25% of its initial body weight. No other signs of severe disease were detected. On Day 3 post-infection, robust virus replication with titres of 6.3 × 10⁵ to 2.5 × 10⁷ pfu/g was detected in the lungs, nasal turbinates, and tracheas of all infected animals (Fig. 2b). On Day 6 post-infection, virus titres reached 4 × 10² to 3.2 × 10⁵ pfu/g in the lungs of infected animals and 5 × 10⁴ to 2 × 10⁶ pfu/g in the nasal turbinates and tracheas (Fig. 2b). No virus was detected in the brains, spleens, or kidneys of the infected mice.

Fig. 2.

A/Henan/2022 virus replication in BALB/c mice. (a) Body weight changes after viral infection. BALB/c mice (n = 5) were intranasally inoculated with 10⁶ pfu of A/Henan/2022 virus. The body weight of the infected mice was monitored daily for 14 days. On Day 8 post-infection, one mouse was euthanized because its body weight loss was >25% of the initial weight. The values for body weights are means ± SD from live mice (b) A/Henan/2022 virus replication in mice. Ten mice were infected with 10⁶ pfu of A/Henan/2022 virus. On Days 3 and 6 post-infection, five mice each were euthanized and tissue samples (nasal turbinates, tracheas, lungs, brains, spleens, and kidneys) were collected for virus titration in MDCK cells. The limit of detection of for the plaque assays (10 PFU/ml) is indicated by the horizontal dashed line. The values presented are the averages ± SD.

A/Henan/2022 virus replication is efficient in the upper but not lower respiratory tract of ferrets

Similarly, eight 6-month-old female ferrets were intranasally infected with 10⁶ pfu of A/Henan/2022 virus (Fig. 3). The animals experienced a transient increase in body temperature on Day 2 post-infection (Fig. 3a), and mild weight loss on Days 2 and 3 post-infection (Fig. 3b). Three days post-infection, efficient virus replication was detected in the tracheas and nasal turbinates; however, no virus was recovered from the lungs (Fig. 3c). On Day 6 post-infection, virus replication was detected in the nasal turbinates of three animals, and in the trachea of one animal (Fig. 3d). No virus was detected in any other organs on Days 3 or 6 post-infection.

Fig. 3.

A/Henan/2022 virus replication in ferrets. Eight ferrets were intranasally infected with 10⁶ pfu of A/Henan/2022 virus. Body temperature (a) and weight (b) were monitored daily until the animals were euthanized. On Day 3 (c) and Day 6 (d) post-infection, four animals each were euthanized and the tissue samples (nasal turbinates, tracheas, lungs, brains, spleens, kidneys, hearts, small intestines, and large intestines) were collected for virus titration in MDCK cells. The limit of detection for the plaque assays (10 PFU/ml) is indicated by the horizontal dashed line. All values presented are the average ± SD.

Respiratory droplet transmission of A/Henan/2022 virus is inefficient in ferrets

We then asked whether A/Henan/2022 virus could transmit to naïve ferrets via respiratory droplets. Four ferrets were intranasally infected with 10⁶ pfu of A/Henan/2022 virus. Another group of four animals was intranasally infected with 10⁶ pfu of A/Isumi/2018 (H1N1pdm09), which transmits efficiently among ferrets.²² One day later, one naïve ferret each was placed in a cage adjacent to an infected ferret. Nasal swabs were collected on Day 1 after inoculation or exposure, respectively, and then every other day until Day 9. Virus titres in nasal swab samples were determined by performing plaque assays in MDCK cells. For the A/Isumi/2018 control virus, respiratory droplet transmission was detected in all four transmission pairs on Days 3, 5, and 7 post-exposure (Fig. 4a). For A/Henan/2022 virus, titres in the nasal swabs of infected animals reached about 4 log₁₀ pfu/ml on Days 1 and 3 post-infection (Fig. 4b). Lower virus titres were detected on Day 5 post-infection, and the infection was cleared by Day 7 post-infection. Respiratory droplet transmission occurred in one transmission pair (Pair 4) with virus detected on Days 5 and 7 post-exposure (Fig. 4b). Hemagglutination inhibition assays with ferret sera collected on Day 21 post-infection or Day 20 post-exposure confirmed the lack of infection of the exposed animals in Pairs 1–3 (Supplementary Table S2). These data demonstrate that the novel reassortant H3N8 viruses can transmit via respiratory droplets to exposed mammals, albeit with low efficiency.

Fig. 4.

Respiratory droplet transmission in ferrets. Ferrets were infected with 10⁶ pfu of (a) A/Isumi/2018 (H1N1pdm09) or (b) A/Henan/2022 (H3N8) virus (infected ferrets). One day later, four naïve ferrets (exposed ferrets) were each placed in a cage adjacent to an infected ferret. Nasal swabs were collected from infected ferrets on Day 1 after inoculation and from exposed ferrets on Day 1 after exposure, and then every other day for virus titration in MDCK cells. The limit of detection for the plaque assays (10 PFU/ml) is indicated by the horizontal dashed line.

Assessment of virus sub-populations during A/Henan/2022 virus replication and transmission

Virus transmission to a new mammalian host and replication in that host may lead to adaptive mutations. We therefore compared the viral populations of the inoculum used for the ferret infection study with those detected during replication and after transmission. Our analysis focused on non-synonymous amino acid changes detected in $\ge$ 3% of the sequence reads. Compared to the GISAID reference sequence, the inoculum possessed four minor polymorphisms that were detected in 3–9% of the sequence reads (Fig. 5; Table 1); the amino acids encoded by these polymorphisms are not known to affect the biological properties of the virus. No polymorphisms were detected in the Pair 3 infected animal (Fig. 5; Table 1). In the other infected animals, most polymorphisms were detected on Day 5 after infection (Fig. 5; Table 1). The most prominent polymorphisms were detected in HA at amino acid positions 226 and 228 (Fig. 5; Table 1), which determine the preference for α2,3- and α2,6-linked sialic acids. For the Pair 1 infected animal, we detected HA-Q226L (53%) and -G228S (42%) substitutions on Day 5 post-infection; however, only 0.27% of sequence reads encoded both substitutions. For the Pair 2 infected animal, an HA-G228A substitution was detected in 7% of the sequence reads on Day 5 post-infection. For the Pair 4 exposed animal, the HA-G228S substitution was detected in 14% of the sequence reads on Day 7 post-exposure. Viruses isolated on Day 5 post-infection from the Pair 2 infected animal also possessed a polymorphism encoding HA-S186I (18%) (Fig. 5; Table 1). This amino acid position is located at the rim of the receptor-binding site and substitutions here can affect receptor-binding specificity.23, 24, 25 Another viral polymorphism that may affect the biological properties of the virus is PB2-D567N (Fig. 5; Table 1), which was detected in 3% of the sequence reads obtained on Day 5 post-infection from the Pair 1 infected animal. Several computational analyses predicted this to be a host-specific amino acid change with aspartic acid encoded by avian influenza viruses and asparagine encoded by human influenza viruses.26, 27, 28 Another polymorphism that may affect viral properties is NS1-R21Q (Fig. 5; Table 1), which was detected at a frequency of 5% in the Pair 4 infected animal on Day 5 post-infection. The NS1-R21Q mutation abrogates the interaction of NS1 with RIG-I, resulting in upregulated RIG-I signaling.²⁹

Fig. 5.

Sequence polymorphisms detected in the inoculum, and in nasal swab samples from infected and exposed ferrets. The A/Henan/2022 virus stock and nasal swab samples from infected and exposed ferrets were subjected to deep-sequencing analysis. Shown are amino acid variants detected in ≥3% of sequence reads.

Table 1.

A/Henan/2022 polymorphisms detected in the inoculum, infected ferrets, and exposed ferrets.

Pair	Sample	Day	Virus gene	Amino acid position	GISAID reference sequence	Variants detected	GISAID reference sequence (percentage)	Variants detected (percentage)
n/a (inoculum)	Inoculum	n/a	PA	649	L	∗	90%	4%
		n/a	PA	650	Y	F	97%	3%
		n/a	PA	651	A	P	97%	3%
		n/a	NA_N8	2	N	D	91%	9%
1	Infected	5	PB2	52	A	T	96%	4%
		5	PB2	106	T	I	81%	19%
		5	PB2	567	D	N	97%	3%
		5	PB1	648	A	T	96%	4%
		5	PA	627	G	W	84%	16%
		5	HA_H3	102	V	M	95%	4%
		5	HA_H3	226	Q	L	47%	53%
		5	HA_H3	228	G	S	58%	42%
		5	NS1	105	L	I	97%	3%
2	Infected	3	PB1	135	R	G	90%	10%
		3	NA_N8	2	N	D	92%	8%
		5	PB2	570	M	I	95%	5%
		5	PB1	21	T	M	87%	13%
		5	PB1	135	R	G	85%	14%
		5	PB1	352	G	E	96%	4%
		5	PA	378	K	N	94%	6%
		5	PA	690	I	M	92%	8%
		5	HA_H3	186	S	I	82%	18%
		5	HA_H3	228	G	A	93%	7%
4	Infected	3	NP	128	D	Y	96%	4%
		3	NA_N8	385	K	T	94%	6%
		5	PB1	149	V	M	95%	5%
		5	PA	645	V	A	96%	4%
		5	HA_H3	50	K	E	86%	14%
		5	HA_H3	344	E	G	95%	5%
		5	NP	352	M	I	97%	3%
		5	NS1	21	R	Q	95%	5%
	Exposed	7	HA_H3	228	G	S	86%	14%
		7	NP	411	T	A	85%	15%

Serological analyses

To assess the antigenic similarity between A/Henan/2022 virus and other H3N8 viruses, we performed hemagglutination inhibition (HI) assays with several avian H3N8 viruses (Table 2). Overall, we detected substantial cross-reactivity, most likely reflecting the slow evolution of avian influenza viruses. A/Henan/2022 virus was inhibited by sera raised against several avian H3N8 viruses, except for a serum raised against A/yellow-billed pintail/Chile/C2014/2015. Similarly, ferret antiserum raised against A/Henan/2022 virus neutralized most other avian H3N8 viruses tested, except for the A/yellow-billed pintail/Chile/C2014/2015 virus, indicating that A/Henan/2022 virus is antigenically similar to other avian H3N8 viruses.

Table 2.

Antigenic similarity of A/Henan/2022 virus with avian H3N8 viruses.

	A/duck/Ukraine/1/1963 (goat serum)	A/mallard/Alberta/60/1991 (ferret serum)	A/duck/Shantou/1283/2001 (goat serum)	A/northern pintail/Alberta/8/2009 (ferret serum)	A/yellow-billed pintail/Chile/C2014/2015 (ferret serum)	A/Henan/4-10/2022 (ferret serum)
A/duck/Ukraine/1/1963	5120	320	320	160	20	320
A/duck/Chabarovsk/1610/1972	2560	640	640	320	80	640
A/mallard duck/Alberta/257/1987	320	160	320	160	40	160
A/blue-winged teal/Iowa/10OS24111/2010	640	160	160	80	20	320
A/yellow-billed pintail/Chile/C2014/2015	80	40	160	40	40	20
A/Henan/4-10/2022	2560	320	320	160	20	320

Next, we determined whether human sera possess cross-reactive antibodies to the novel reassortant H3N8 viruses. Microneutralization assays were carried out with 90 human sera from donors of different age groups (collected in the spring and summer of 2022; purchased from BioIVT, Westbury, NY, USA) against A/Henan/2022 virus (Fig. 6). As a control, the microneutralization titres against a recent human H3N2 virus (A/Bangladesh/911009/2020) were tested in parallel. Most human sera (78/90) possessed neutralizing antibody titres of ≥40 to the human seasonal A/Bangladesh/911009/2020 virus with no age-related pattern (Fig. 6a). An appreciable number of human sera (62/90) also displayed microneutralization titres of ≥40 against A/Henan/2022 virus (Fig. 6b), especially those from older donors.

Fig. 6.

Microneutralization titres of human sera against A/Bangladesh/911009/2020 (a) and A/Henan/4-10/2022 (b) viruses. Human sera from donors of different ages were tested against the indicated viruses. The limit of detection of the microneutralization titres is indicated by the dashed line.

Discussion

Our characterization of A/Henan/2022 virus, a novel reassortant H3N8 virus, revealed efficient replication in the upper respiratory tract of mice and ferrets and respiratory droplet transmission (although inefficient) among ferrets. The pathogenicity and transmissibility of our reverse genetics-generated A/Henan/2022 virus was lower than that recently reported by Sun et al.,¹ most likely because different variants and/or different experimental conditions were used. Replication in ferrets resulted in the emergence of variants with mutations that may affect the properties of the virus. An appreciable number of human sera displayed low-to-moderate neutralizing antibody titres of 40–320 to A/Henan/2022 virus, especially those from older donors.

Three studies reported very limited reactivity of human sera with novel reassortant H3N8 viruses.^1,7,30 In contrast, we found that about two-thirds of the human sera tested displayed microneutralization titres of ≥40 against A/Henan/2022 virus, especially sera from older donors. Chen et al.¹⁵ reported seropositivity rates of 8%–15% against novel reassortant chicken H3N8 viruses in sera from donors ≥50 years of age, with much lower seropositivity rates among younger donors. Older people may have antibodies that react with the novel reassortant H3N8 viruses, perhaps because the HA (and NA) proteins of earlier human H3N2 viruses (whose HA originated from an avian influenza virus) bear some antigenic similarity with the novel reassortant H3N8 viruses. By contrast, due to the antigenic evolution of H3N2 viruses in humans, sera from younger people show less cross-reactivity with avian H3 viruses.

Previously, two studies^31,32 tested the transmissibility of Eurasian avian H3N8 viruses (with gene compositions different from those of the novel reassortant H3N8 viruses) in guinea pigs and ferrets. Liang et al.³² reported direct contact transmission to one of three guinea pigs for one of the viruses tested, but no transmission was observed for the second virus tested. Zhang et al.³¹ tested eight different Eurasian avian H3N8 viruses, three of which transmitted to contact animals in the same cage. Six of these viruses were also tested for respiratory droplet transmission in guinea pigs. One of them transmitted via respiratory droplets to two of three exposed guinea pigs.³¹ Sequence analysis revealed a PB1–S524G mutation that arose in the infected animal and was also detected in the exposed guinea pigs. The wild-type isolate did not transmit among ferrets via respiratory droplets, but the mutant with the PB1–S524G mutation transmitted to two of three respiratory droplet contact ferrets,³¹ identifying a key role for the PB1–S524G mutation in H3N8 virus transmissibility in ferrets. Recently, Chen et al.¹⁵ tested the transmissibility of novel reassortant H3N8 viruses in ferrets. Three different chicken isolates transmitted to two or three of the three direct contact animals, respectively. Respiratory droplet transmission was detected in none to two of the three exposed animals. While this manuscript was in preparation, Sun et al.¹ published a characterization of the authentic A/Henan/2022 virus isolate. The authentic human isolate (comprising subpopulations encoding HA-228S and HA-228G) efficiently transmitted via respiratory droplets in ferrets.¹ A reverse genetics-generated A/Henan/2022 virus encoding HA-228G (as used in our study) also transmitted to exposed ferrets.¹ Here, we found that our reverse genetics-generated A/Henan/2022 virus (encoding HA-228G) transmitted to one of four exposed ferrets (Fig. 4, Pair 4). The differences in transmission efficiencies likely result from differences in the experimental conditions as described in a comparative ferret transmission study conducted across 11 laboratories, including ours.³³ Factors that may affect transmission efficiency include, but are not limited to, the directionality of the airflow, the number of air exchanges per hour, and the distance between the infected and exposed animals. Collectively, however, these data demonstrate that the novel reassortant H3N8 viruses can transmit among mammals. Avian influenza viruses preferentially bind to α2,3-linked sialic acids, the predominant sialic acid expressed on epithelial cells in the intestinal tract of waterfowl. In contract, human influenza viruses have high binding affinity for α2,6-linked sialic acids, which are predominantly found on epithelial cells in the respiratory tract of mammals. These differences in receptor-binding specificity are conferred by the amino acids at HA positions 226 and 228: HA-226Q/228G (encoded by most avian influenza viruses) confers preferential binding to α2,3-linked sialic acids, whereas HA-226L/228S (typically encoded by human influenza viruses) confers preferential binding to α2,6-linked sialic acids. Several Eurasian avian H3N8 viruses (different from the novel reassortant viruses) bind to both α2,3- and α2,6-linked sialic acids, although they encode HA-226Q/228G (i.e., the avian-like amino acids).^31,34 Similarly, the novel reassortant H3N8 viruses bind to both types of sialyloligosaccharides, even though they encode the avian-type amino acids at amino acid positions 226 and 228 of HA.^3,10,11 However, some of the A/Henan/2022 virus sequences deposited in GISAID list an undefined amino acid (‘X’) at HA position 228, indicating viral subpopulations at this position. By using reverse genetics, we generated an A/Henan/2022 virus encoding HA-228G. Interestingly, variants encoding HA-226L, -228S, or -228A were detected in two of the infected and one of the exposed animals (Fig. 5, Table 1), demonstrating selective pressure on these amino acid positions during replication in ferrets. In particular, the Pair 1 infected ferret encoded high percentages of HA-Q226L (53%) and -G228S (42%) variants on Day 5 post-infection. However, only 0.27% of sequence reads encoded both amino acid substitutions, suggesting that a substitution at one of these positions alone may provide a replicative advantage in ferrets, presumably through more efficient binding to α2,6-linked sialic acids. Our findings match those of Sun et al.,¹ who detected increasing ratios of HA-228S during replication in ferrets and those of Chen et al.,¹⁵ who found that novel reassortant chicken H3N8 viruses acquire HA-G228S or HA-Q226L mutations during replication and/or transmission in ferrets. In addition, we detected an HA-S186I mutation in 18% of sequence reads on Day 5 post-infection for Pair 2. Amino acid changes at the equivalent position of H5^23,24 and H7²⁵ viruses increases α2,6-linked sialic acid binding; however, the effect of an isoleucine residue at this position is currently unknown. Most human H3N2 viruses encode HA-186G, but HA-186I has been detected in approximately 150 human viruses, primarily from the Bangkok ’79 and Sichuan ’87 antigenic cluster of human H3N2 viruses.³⁵

A recent study identified amino acid substitutions at positions 313 (and 52) of the viral nucleoprotein (NP) that overcome the restrictive effect of human BTN3A3 on avian influenza viruses in humans.³⁶ The NP amino acid substitutions detected in the inoculated and exposed ferrets (Fig. 5, Table 1) are not adjacent to amino acid positions 313 and 52 in the NP structure, suggesting that they act through other mechanisms or arose randomly.

In summary, here we characterized A/Henan/2022 virus, a novel reassortant H3N8 virus, in vitro and in mice and ferrets. A/Henan/2022 virus displayed some pathogenicity in mice and inefficient respiratory droplet transmissibility in ferrets. A strength of our study is that we also tested human sera collected from donors of different age groups for antibodies that reacted with this novel reassortant H3N8 virus and found that some people did, in fact, have cross-reactive antibodies; however, we do not know if these cross-reactive antibodies would ameliorate an infection with novel reassortant H3N8 viruses. Therefore, the prevalence of these viruses and their biological properties should be monitored closely.

Contributors

C.G., S.F., G.N., and Y.K. designed the study. C.G., S.F., S.C., L.G., and T.M. performed the experiments. C.G., S.F., G.N., and Y.K. analyzed the data. R.D. deep-sequenced all samples. L.B. and D.P. analyzed the deep-sequencing data. G.N. wrote the manuscript. C.G. and G.N. verified the underlying data. All authors reviewed and approved the manuscript.

Data sharing statement

Data supporting the findings of this study are present in the paper and/or Supplementary Materials. Additional data related to this paper may be requested from the authors.

Declaration of interests

Y.K. has received grant support from Daiichi Sankyo Pharmaceutical, Toyama Chemical, Tauns Laboratories, Inc., Otsuka Pharmaceutical Co., Ltd., Shionogi & Co., Ltd., Otsuka Pharmaceutical, KM Biologics, Kyoritsu Seiyaku, Shinya Corporation, and Fuji Rebio. Y.K. and G.N. are co-founders of FluGen. The other authors have no conflicts of interest.