Pathogenicity of Highly Pathogenic Avian Influenza A(H5N1) Viruses Isolated from Cats in Mice and Ferrets, South Korea, 2023 (2024)

Materials and Methods

Cells

We grew and maintained MDCK cells (American Type Culture Collection, https://www.atcc.org) in Eagle’s Minimum Essential Medium (WELGENE, https://www.welgene.com) containing 5% fetal bovine serum, 1 mmol/L L-glutamine, and penicillin/streptomycin (Thermo Fisher Scientific, https://www.thermofisher.com). We incubated the cells at 37°C in 5% CO₂ until use.

Virus Distribution

The Animal and Plant Quarantine Agency (APQA), South Korea, provided 3 HPAI H5N1 virus isolates and deposited their whole-genome sequences in the GISAID EpiFlu database (http://www.gisaid.org). We propagated the isolates in specific pathogen-free embryonated chicken eggs (second passage) and confirmed that their sequences were identical to those provided by the APQA (Table 1). A/duck/Korea/H493/2022(H5N1) (GISAID accession no. EPI_ISL_15647834) originated from a duck farm in the Yecheon area in October 2022. A/feline/Korea/M302–6/2023(H5N1) (accession no. EPI_ISL_18819809) was from the Yongsan District, and A/feline/Korea/M305–7/2023(H5N1) (accession no. EPI_ISL_18819807) was from the Gwanak District; both of those viruses from cats were obtained from animal shelters during July 2023.

Genetic and Phylogenetic Analysis

We extracted virus RNA by using the RNeasy Mini Kit (QIAGEN, https://www.qiagen.com) and performed gene amplification and library preparation by using the Illumina Microbial Amplicon Prep-Influenza A/B kit (Illumina, https://www.illumina.com). Subsequently, we sequenced whole genomes of the viruses on a MiSeq instrument by using MiSeq Reagent Kit v2 (Illumina) to obtain 2 × 150-bp read lengths. For phylogenetic analysis, we searched for sequences, other than those analyzed in this study, in the GISAID database. We inferred phylogenetic relationships of sequences obtained in this study by using the maximum-likelihood method, 1,000 bootstrap values, and MEGA 7 software (17).

Virus Titrations

We determined virus titers of oropharyngeal and cloacal swab samples, nasal washes, and hom*ogenized tissue samples by performing endpoint titrations in MDCK cell monolayers. We inoculated MDCK cells with 10-fold serial dilutions of each sample prepared in fetal bovine serum–free medium containing L-1-tosylamido-2-phenylethyl chloromethyl ketone–treated trypsin and penicillin/streptomycin. After a 72-hour incubation at 37°C, we detected viruses in a standard hemagglutination assay by using 0.5% turkey erythrocytes. We expressed mean virus titers as log₁₀ 50% tissue culture infectious dose (TCID₅₀). The detection limit was 0.5 log₁₀ TCID₅₀/mL. We estimated virus titers by using t-tests and 2-way analysis of variance in GraphPad Prism 9 (GraphPad Software Inc., https://www.graphpad.com).

Neuraminidase Inhibitor Resistance

We used the neuraminidase (NA) inhibitors oseltamivir and peramivir (Cayman Chemical, https://www.caymanchem.com) and zanamivir (Sigma-Aldrich, https://www.sigmaaldrich.com) to assess drug susceptibility of the 3 virus isolates. We used a fluorescence assay containing the 2′-(4-methylumbelliferyl)-α-D-N-acetylneuraminic acid substrate (Sigma-Aldrich) (18,19). We normalized influenza viruses to equivalent NA activities and incubated virus samples with 10-fold serial dilutions (0–30,000 nmol/L) of oseltamivir, zanamivir, or peramivir. We measured the fluorescence signal by using a Mithras LB 940 reader (Berthold Technologies, https://www.berthold.com) at excitation/emission wavelengths of 355/460 nm. We estimated the 50% inhibitory concentration (IC₅₀) for each sample from dose-response curves by using the sigmoidal, 4-parameter, logistic nonlinear regression equation in GraphPad Prism 9. To assess neuraminidase inhibitor (NAI) resistance, we divided the IC₅₀ value of the virus being analyzed by the IC₅₀ value of the NAI-sensitive influenza A(H1N1)pdm09 virus strain, which has the amino acid H274 in neuraminidase, making it NAI-susceptible.

Experimental Infections of Mice and Ferrets

We anesthetized groups of 6-week-old BALB/c mice (SAMTAKO, http://www.samtako.com) (n = 5/group) with ketamine and intranasally inoculated them with 50 µL of 10⁰–10⁶ TCID₅₀/mL of virus. After virus inoculation, we weighed the mice and monitored them for clinical signs and death for 14 days. For virus replication studies, we intranasally inoculated 15 mice per group with 50 µL of 10³ 50% median lethal dose (MLD₅₀)/mL. We euthanized 5 mice per group on days 3, 5, and 7 postinoculation and assessed virus titers in brain, trachea, nasal turbinate, lung, heart, liver, kidney, spleen, and intestinal samples.

We anesthetized 20–22-week-old ferrets (IDBio, http://www.idbio.co.kr) (n = 12/group) with ketamine and intranasally inoculated them with 1 mL of 10³ MLD₅₀/mL of virus. After virus inoculation, we weighed and monitored 3 ferrets per virus group for clinical signs and death for 14 days. We used the remaining 9 ferrets per group for virus replication studies. We euthanized 3 ferrets per virus group on days 3, 5, and 7 postinoculation and assessed virus titers in brain, trachea, nasal turbinate, lung, heart, liver, kidney, spleen, and intestinal samples. To assess virus transmission via contact infection, we intranasally inoculated 1 ferret (per virus) with 1 mL of 10³ MLD₅₀/mL virus and then housed serologically-naive ferrets (n = 2) in the same cage the next day (1 cage/virus). We collected nasal wash samples from each ferret on days 3, 5, and 7 postinoculation and measured virus titers. We euthanized mice and ferrets showing >20% body weight loss, which we considered a humane endpoint.

Serologic Tests

We collected blood samples from ferrets in the infection groups 14 days postinfection and in the transmission groups 14 days after contact with a virus-infected ferret. We determined seroconversion by using a microneutralization assay.

Results

Genetic Characterization of H5N1 Viruses Isolated from Cats

The Korea Disease Control and Prevention Agency (KDCA) received 2 different HPAI H5N1 viruses from cats in animal shelters that were collected by APQA, which we used to characterize infections in other mammals. We used the 2 H5N1 viruses, A/feline/Korea/M302–6/2023(H5N1) from Yongsan (abbreviated as YS/2023) and A/feline/Korea/M305–7/2023(H5N1) from Gwanak (abbreviated as GA/2023), to determine how the pathogenicity and transmission of those H5N1 viruses differed from previously prevalent H5N1 viruses. We also analyzed the virus isolated from a duck, A/duck/Korea/H493/2022(H5N1) (abbreviated as YC/2022), representing the first poultry outbreak in autumn 2022.

We conducted whole-genome sequence analysis to confirm genetic characteristics of the YC/2022, YS/2023, and GA/2023 viruses and observed polybasic residues (REKRRKR/GLF) within the cleavage sites of hemagglutinin (HA), classifying all 3 viruses as HPAI. Phylogenetic analysis confirmed the viruses belonged to clade 2.3.4.4b (Table 1; Appendix Figure). YS/2023 and GA/2023 viruses shared 99.9%–100% genetic similarity (Appendix Table). They also showed close genetic relatedness to H5N1 viruses that have been circulating in wild birds and poultry in Asia since 2022, including in South Korea, China, and Japan (15).

We identified amino acid substitutions related to mammal adaptation in both YS/2023 and GA/2023 viruses. In both viruses, we found mutations S123P, S133A, and T156A (H5 numbering), which enhance binding affinity of the HA protein to α2,6-sialic acid on the host cell surface and contribute to increased mammal receptor tropism (20; J. Yang et al., unpub. data, https://doi.org/10.1101/2024.07.09.602706). In addition, in the polymerase basic 2 (PB2) gene segment, we identified mutations encoding D701N in YS/2023 and E627K in GA/2023; both substitutions are mammal-adapting mutations known to increase polymerase activity and virulence in mammals (Table 2) (21–23). However, mutations associated with antiviral drug resistance, such as H274Y in NA and S31N in the matrix protein, were not detected (21,24).

Virus Pathogenesis in a Mouse Model

We intranasally inoculated 10-fold serial dilutions of infectious dose for each virus into 6-week-old BALB/c mice. After a 2-week observation period, the MLD₅₀ values were 10^1.5 TCID₅₀/mL for YS/2023 and 10^0.5 TCID₅₀/mL for GA/2023, whereas the MLD₅₀ was 10^4.8 TCID₅₀/mL for YC/2022. The 2 viruses isolated from cats had ≈10-fold difference in MLD₅₀ values between them, but their MLD₅₀ values were >1,000-fold lower than that of the duck isolate (Table 1, Figure 1).

To assess detailed clinical symptoms and virus replication in internal organs, we intranasally inoculated 50 µL of 10³ MLD₅₀/mL (YS/2023, 10^4.5 TCID₅₀/mL; GA/2023, 10^3.5 TCID₅₀/mL; YC/2022, 10^7.8 TCID₅₀/mL) of each virus into 6-week-old BALB/c mice (n = 15 in each group). All infected mice exhibited clinical symptoms, such as weight loss, ruffled fur, lethargy, and ataxia, within 5 days postinfection. Virus infection was confirmed in the respiratory tract of all mice on day 3 postinfection, and only viruses isolated from cats were detected in all organs (including the brain) by day 5 postinfection (Figure 2). Furthermore, all virus-infected mice had virus titers in lung, trachea, and nasal turbinate samples beginning on day 3 during the early stage of infection (Figure 2). By day 5, only mice infected with both cat isolates (YS/2023 and GA/2023) had virus titers in 9 organs; we observed high virus titers in brain, nasal turbinate, trachea, lung, and heart samples. In particular, mice infected with GA/2023 exhibited high titers in lung (10^5.4 TCID₅₀/mL) and trachea (10^5.1 TCID₅₀/mL) tissue on day 3 and in lung (10^4.9 TCID₅₀/mL) and brain (10^4.7 TCID₅₀/mL) tissue on day 5, but all mice died before day 7. Mice infected with YS/2023 had high titers (10^3.9–10^4.9 TCID₅₀/mL) in lung, trachea, and brain tissues on day 5 and in brain (10^4.8 TCID₅₀/mL) and nasal turbinate (10^4.25 TCID₅₀/mL) tissue on day 7. Both cat-derived virus isolates used to infect mice showed a systemic infection pattern and high lethality; high virus titers occurred in most organs, including the brain. The duck isolate, YC/2022, was detected only in the brain and respiratory organs (nasal turbinate, trachea, and lungs) by day 7; virus titers were lower than those for the cat isolates (Figure 2).

Virus Pathogenesis and Transmission in a Ferret Model

Ferrets intranasally infected with the 2 cat-derived H5N1 virus isolates showed severe clinical symptoms, including sneezing, nasal discharge, diarrhea, and neurologic complications. They also exhibited a mean peak reduction in bodyweight of 5.5%–25.7% and fever of 0.7°C–2.6°C above baseline temperature. All ferrets died by day 8 in the GA/2023 infection group and by day 9 in the YS/2023 group (100% mortality rate) (Figure 3). The ferrets infected with the GA/2023 and YS/2023 viruses showed a systemic infection pattern. In ferrets infected with GA/2023, the virus was detected in all organs except the kidneys by day 5, and virus titers of 10^0.8–10^2.9 TCID₅₀/mL were detected in the brain and respiratory organs. Those titers were lower than titers observed in ferrets infected with YS/2023 (10^2.3–10^3.4 TCID₅₀/mL) that survived longer (Figure 4). The ferrets infected with YC/2022 showed the highest virus titers in the respiratory tract (10^2.3–10^5.4 TCID₅₀/mL) until day 5, and infection was observed in all organs. However, 1 ferret died in the YC/2022 group on day 11, resulting in a survival rate of 66.6% (Figure 3).

To assess virus transmission via contact with YS/2023, GA/2023, and YC/2022, each virus was intranasally inoculated into 1 ferret and 2 serologically naive ferrets were moved into the same cage as the infected animal. The ferrets inoculated with either cat-derived virus died on day 8 postinoculation. In contrast, only 1 of 2 ferrets in contact with the YS/2023-infected ferret died on day 13; no seroconversion was observed in the surviving ferret. The 2 ferrets in contact with the GA/2023-infected ferret died by day 9 (100% mortality rate) (Figure 5). Virus concentrations increased (10^2.3–10^5.5 TCID₅₀/mL) in nasal wash samples collected from ferrets exposed to the cat-derived viruses on days 3, 5, and 7, confirming transmission and infection in naive ferrets through contact with infected animals (Figure 6). In contrast, the ferret infected intranasally with YC/2022 did not die, although the virus was detected in nasal washes. Both ferrets in that contact group survived; no virus was detected in nasal washes and no seroconversion was observed, inferring that contact transmission of YC/2022 did not occur.

Antiviral Drug Susceptibility of Influenza A(H5N1) Viruses Isolated from Cats

We experimentally analyzed NAI susceptibility to evaluate the effectiveness of existing influenza antiviral drugs against YS/2023, GA/2023, and YC/2022 viruses. We compared the IC₅₀ values of the 3 viruses with that of an antiviral drug–susceptible human influenza A(H1N1)pdm09 reference virus. The high sensitivity of YS/2023, GA/2023, and YC/2022 viruses to NAIs confirmed the effectiveness of specific antiviral drugs (Table 3).

Discussion