Subclinical Infection and Transmission of Clade 2.3.4.4 H5N6 Highly Pathogenic Avian Influenza Virus in Mandarin Duck (Aix galericulata) and Domestic Pigeon (Columbia livia domestica)

Since 2014, H5Nx clade 2.3.4.4 highly pathogenic avian influenza viruses (HPAIV) have caused outbreaks in wild birds and poultry in multiple continents, including Asia, Europe, Africa, and North America. Wild birds were suspected to be the sources of the local and global spreads of HPAIV. This study evaluated the infectivity, pathogenicity, and transmissibility of clade 2.3.4.4 H5N6 HPAIV in mandarin ducks (Aix galericulata) and domestic pigeons (Columbia livia domestica). None of the birds used in this study, 20 mandarin ducks or 8 pigeons, showed clinical signs or mortality due to H5N6 HPAI infection. Two genotypes of H5N6 HPAIV showed replication and transmission by direct and indirect contact between mandarin ducks. H5N6 HPAIV replicated and transmitted by direct contact between pigeons, although the viral shedding titer and duration were relatively lower and shorter than those in mandarin ducks. Influenza virus antigen was detected in various internal organs of infected mandarin ducks and pigeons, indicating systemic infection. Therefore, our results indicate mandarin ducks and pigeons can be subclinically infected with clade 2.3.4.4 H5N6 HPAIV and transfer the virus to adjacent birds. The role of mandarin ducks and pigeons in the spread and prevalence of clade 2.3.4.4 H5N6 viruses should be carefully monitored.


Introduction
After the discovery of A/goose/Guangdong/1/1996(H5N1) (Gs/GD), a highly pathogenic avian influenza (HPAI) virus in China, global spread of Gs/GD-lineage H5 HPAI viruses has been observed, and the viruses have evolved into phylogenetically distinct clades (0-9) and subclades by genetic drift and shift [1]. Before 2008, H5N1 was the major subtype of the Gs/GD lineage; however, novel H5Nx subtypes, such as H5N2, H5N5, H5N6, and H5N8, have emerged as a result of reassortment with prevailing low-pathogenic avian influenza viruses (LPAI) [2][3][4]. Since 2014, H5N6 and H5N8 subtypes of viruses bearing the hemagglutinin (HA) gene of clade 2.3.4.4 have been detected in wild birds and poultry in multiple continents, including Asia, Europe, Africa, and North America. They have become the dominant subtypes of the Gs/Gd lineage [5][6][7][8][9][10]. All birds were kept in self-contained isolation units ventilated under negative pressure with HEPA-filtered air. Commercial chicken feed and freshwater were provided daily. All animal experiments were conducted in a biosafety level 3 (BSL-3) facility at Konkuk University.

Experimental Design
Both genotypes C1 and C4 were used to infect mandarin ducks. Twenty-four mandarin ducks were divided into two groups, and each group was divided into three subgroups: seven for inoculation, three for contact exposure, and two for indirect exposure. Genotype C1 was used to infect pigeons experimentally. Ten pigeons were divided into two groups: seven for inoculation and three for contact exposure.
Fourteen mandarin ducks and seven pigeons in the inoculation groups were inoculated with 0.2 mL of 10 7.0 EID 50 /mL via choanal cleft. To evaluate transmissibility through contact, three naïve birds from each group were co-housed with the inoculated birds 24 h later. We used a cage with two rooms divided by a barrier with a window of 3-layered wire mesh (30 × 35 cm) to evaluate the transmissibility through indirect transmission, allowing common and indirect airflow. The two outer layers had 1.2 cm and an inner layer had 2 mm of grid size which can prevent direct contact between the two rooms. Two inoculated and naïve mandarin ducks were housed in each room of a mesh-divided cage. All birds were monitored daily for clinical signs and mortality.

Viral Quantification
To quantify viral shedding, oropharyngeal and cloacal swabs were collected from the birds at 2, 4, 6, 8, and 12 days post-inoculation (dpi) and suspended in 1 mL of phosphate-buffered saline (PBS) containing antibiotics (400 mg/mL of gentamicin). Viral RNA was extracted from 200 µL of the suspension using MagNA Pure 96 DNA and Viral NA Small Volume Kit on the MagNA Pure 96 instrument (Roche Applied Sciences, Penzberg, Germany) according to the manufacturer's instructions. Quantitative analysis of viral RNA was conducted using the cycle threshold (Ct) value using rRT-PCR targeting the M gene [43]. It has been demonstrated that Ct values correlate strongly with infectious viral titer [46]. To convert the Ct value into the EID 50 equivalent unit, we generated a standard curve between them as previously described [36]. Briefly, the HPAI H5N6 challenge virus with a known EID 50 titer was serially diluted 10-fold and quantified using rRT-PCR as described above. Ct values of viral 10-fold dilutions were plotted against viral EID 50 titers. The resulting standard curve showed a high correlation (r 2 > 0.99). The detection limit was 10 2 EID 50 equivalent/mL. Virus titers in samples were calculated by extrapolation of the standard curve equation.

Statistical Analysis
The presence of statistically significant differences in the viral shedding titers was identified using one-way ANOVA with Tukey's post hoc test in GraphPad Prism 5 software (GraphPad Software Inc., San Diego, CA, USA). The samples from which viral shedding was not detected by rRT-PCR were treated as 10 1.9 EID 50 equivalent/mL for statistical purposes. The p-values of < 0.05 were regarded as statistically significant.

Serology
To verify seroconversion, serum samples were collected from the birds at 1 day before and 14 days after the infection. For homologous anti-H5 antibody detection, the serum samples were treated with receptor-destroying enzyme (Denka Seiken Co, Ltd., Gosen, Japan) according to the manufacturer's instructions and screened for HI antibody. Formalin-inactivated homologous antigens and 1% chicken red blood cells were used in the HI assay. A commercially available competitive NP-ELISA kit (Bionote) was used to detect anti-influenza A NP-specific antibodies according to the manufacturer's instructions.

Immunohistochemistry
At 5 dpi, two birds from each inoculation group were euthanized. Organs (brain, trachea, heart, lung, liver, spleen, pancreas, kidney, cecal tonsil, and intestine) were sampled, fixed in 4% paraformaldehyde, and embedded in paraffin. The resulting blocks were cut into 5 µm-thick sections and subjected to immunohistochemical staining to detect influenza virus-specific signals, as described previously [47]. The tissue sections were incubated overnight at 4°C with goat anti-influenza A virus antibody (25 µg/mL, Merck, Darmstadt, Germany), for 2 h at room temperature with biotinylated anti-goat IgG (Vector Laboratories Inc., Burlingame, CA, USA), and for 1 h at room temperature with horseradish peroxidaseconjugated streptavidin (Vector Laboratories Inc.). Positive signals were visualized using diaminobenzidine and counterstained with methyl green. Negative control staining was performed by omitting the primary antibody incubation step.

Clinical Signs of Infection, Viral Shedding, and Transmission
No mortality or clinical signs were observed in both mandarin ducks and pigeons (Table 1), except for a pigeon in the inoculation group that died at 13 dpi. The pigeon did not exhibit any clinical signs of HPAI, gross lesion at autopsy, or viral shedding. Thus, it was difficult to conclude whether HPAI-related fatality was present.  3 One pigeon died at 13 dpi before bleeding for serological analysis.
Both genotypes C1 and C4 were able to efficiently replicate in all mandarin ducks used in the experiment. Successional peak viral titers were in the order of inoculation, contact exposure, and indirect exposure, indicating that both genotype C1 and C4 clade 2.3.4.4 H5N6 HPAI viruses can successfully replicate in mandarin ducks and be transmitted through direct contact as well as indirect exposure (Table 1, Figure 1). However, excretions of the genotype C1 virus were detected from 2 to 12 dpi in inoculated birds, and the genotype C4 virus was detected from 2 to 8 dpi in inoculated birds. The peak amount of viral shedding was 10 7.19 EID 50 /mL and 10 8.23 EID 50 /mL for genotypes C1 and C4, respectively. Both contact-and indirectly exposed birds excreted detectable amounts of viruses from 4 dpi, except one bird in the contact exposure group for the C4 genotype, which excreted the virus via the oropharynx from 2 dpi. These results show that both genotypes C1 and C4 can replicate in and transmit between mandarin ducks via direct or indirect exposure. The viral shedding titers were not significantly different between the viral genotypes or sample types possibly due to the limited number of animals used in the study.
Both genotypes C1 and C4 were able to efficiently replicate in all mandarin ducks used in the experiment. Successional peak viral titers were in the order of inoculation, contact exposure, and indirect exposure, indicating that both genotype C1 and C4 clade 2.3.4.4 H5N6 HPAI viruses can successfully replicate in mandarin ducks and be transmitted through direct contact as well as indirect exposure (Table 1, Figure 1). However, excretions of the genotype C1 virus were detected from 2 to 12 dpi in inoculated birds, and the genotype C4 virus was detected from 2 to 8 dpi in inoculated birds. The peak amount of viral shedding was 10 7.19 EID50/mL and 10 8.23 EID50/mL for genotypes C1 and C4, respectively. Both contact-and indirectly exposed birds excreted detectable amounts of viruses from 4 dpi, except one bird in the contact exposure group for the C4 genotype, which excreted the virus via the oropharynx from 2 dpi. These results show that both genotypes C1 and C4 can replicate in and transmit between mandarin ducks via direct or indirect exposure. The viral shedding titers were not significantly different between the viral genotypes or sample types possibly due to the limited number of animals used in the study. Domestic pigeons exhibited lower levels of or no viral shedding compared to mandarin ducks. Three out of five birds (60%) in the inoculation group and two out of three (66.67%) in the contract exposure group of domestic pigeons showed viral shedding (Table 1). Viral shedding was detected from 2 to 6 dpi in the inoculation group ( Figure 2). The peak amount of viral shedding was 10 5.26 EID50/mL and 10 6.35 EID50/mL for the inoculation and contact exposure groups, respectively. One contact-exposed pigeon shed the virus from 4 to 12 dpi, with a peak viral titer of 10 6.4 EID50/mL at 8 dpi. The other pigeon in the contact group shed the virus only at 4 dpi via the oropharynx. These results show Domestic pigeons exhibited lower levels of or no viral shedding compared to mandarin ducks. Three out of five birds (60%) in the inoculation group and two out of three (66.67%) in the contract exposure group of domestic pigeons showed viral shedding (Table 1). Viral shedding was detected from 2 to 6 dpi in the inoculation group ( Figure 2). The peak amount of viral shedding was 10 5.26 EID 50 /mL and 10 6.35 EID 50 /mL for the inoculation and contact exposure groups, respectively. One contact-exposed pigeon shed the virus from 4 to 12 dpi, with a peak viral titer of 10 6

Serology
Seroconversion was confirmed in all mandarin ducks by NP-ELISA and HI assay. Seroconversion was not detected by NP-ELISA in one mandarin duck in the inoculation group of genotype C4; however, the results of viral detection by rRT-PCR and the HI assay indicate viral replication in the bird (Table 1).
In the HI assay, only one pigeon in the inoculation group showed seroconversion. However, seroconversion in all pigeons in the inoculation group and one out of three in the contact exposure group was confirmed by NP-ELISA, indicating viral replication and transmission via direct contact in pigeons (Table 1).

Immunohistochemistry
In the immunohistochemical (IHC) analysis, positive signals for influenza virus from various organs indicated systemic tissue tropism of the clade 2.3.4.4 H5N6 HPAIV in mandarin ducks and domestic pigeons. Influenza virus was localized in all tested organs except for the lungs of genotype C1-infected mandarin ducks, and the heart of genotype C4infected mandarin ducks and genotype C1-infected domestic pigeons ( Table 2, Figures 3  and 4). Influenza antigen was not detected in any of the negative control tissue sections. Due to the limited number of birds used for this study, it is not possible to confirm that the viruses have no tropism in the organs negative for an influenza signal in the IHC analysis.

Serology
Seroconversion was confirmed in all mandarin ducks by NP-ELISA and HI assay. Seroconversion was not detected by NP-ELISA in one mandarin duck in the inoculation group of genotype C4; however, the results of viral detection by rRT-PCR and the HI assay indicate viral replication in the bird (Table 1).
In the HI assay, only one pigeon in the inoculation group showed seroconversion. However, seroconversion in all pigeons in the inoculation group and one out of three in the contact exposure group was confirmed by NP-ELISA, indicating viral replication and transmission via direct contact in pigeons (Table 1).

Immunohistochemistry
In the immunohistochemical (IHC) analysis, positive signals for influenza virus from various organs indicated systemic tissue tropism of the clade 2.3.4.4 H5N6 HPAIV in mandarin ducks and domestic pigeons. Influenza virus was localized in all tested organs except for the lungs of genotype C1-infected mandarin ducks, and the heart of genotype C4-infected mandarin ducks and genotype C1-infected domestic pigeons (Table 2, Figures 3 and 4). Influenza antigen was not detected in any of the negative control tissue sections. Due to the limited number of birds used for this study, it is not possible to confirm that the viruses have no tropism in the organs negative for an influenza signal in the IHC analysis.

Discussion
Based on the viral genome sequence data in the GISAID database (https://platform.gisaid.org/, accessed on 19 April 2021), clade 2.3.4.4 H5N6 HPAI viruses have been detected in multiple species of wild birds, including mandarin duck (Aix galericulata), northern pintail (Anas acuta), Eurasian wigeon (Anas penelope), common teal (Anas crecca), heron (Ardeidae sp.), and domestic pigeon (Columbia livia domestica) in many other countries in Asia, including Japan, China, and Vietnam. Clade 2.3.4.4 H5Nx HPAI viruses have been detected in multiple continents since 2014. Wild birds have been suggested to play a major role in the prevalence, evolution, and dissemination of the viruses. Seasonal movement between and aggregation in the breeding and wintering sites of migratory wild birds have contributed to the long-distance dissemination and reassortment of HPAI viruses [48][49][50]. Subclinical infection of wild birds with low levels of clinical signs and prolonged

Discussion
Based on the viral genome sequence data in the GISAID database (https://platform.gisaid.org/, accessed on 19 April 2021), clade 2.3.4.4 H5N6 HPAI viruses have been detected in multiple species of wild birds, including mandarin duck (Aix galericulata), northern pintail (Anas acuta), Eurasian wigeon (Anas penelope), common teal (Anas crecca), heron (Ardeidae sp.), and domestic pigeon (Columbia livia domestica) in many other countries in Asia, including Japan, China, and Vietnam. Clade 2.3.4.4 H5Nx HPAI viruses have been detected in multiple continents since 2014. Wild birds have been suggested to play a major role in the prevalence, evolution, and dissemination of the viruses. Seasonal movement between and aggregation in the breeding and wintering sites of migratory wild birds have contributed to the long-distance dissemination and reassortment of HPAI viruses [48][49][50]. Subclinical infection of wild birds with low levels of clinical signs and prolonged

Discussion
Based on the viral genome sequence data in the GISAID database (https://platform. gisaid.org/, accessed on 19 April 2021), clade 2.3.4.4 H5N6 HPAI viruses have been detected in multiple species of wild birds, including mandarin duck (Aix galericulata), northern pintail (Anas acuta), Eurasian wigeon (Anas penelope), common teal (Anas crecca), heron (Ardeidae sp.), and domestic pigeon (Columbia livia domestica) in many other countries in Asia, including Japan, China, and Vietnam. Clade 2.3.4.4 H5Nx HPAI viruses have been detected in multiple continents since 2014. Wild birds have been suggested to play a major role in the prevalence, evolution, and dissemination of the viruses. Seasonal movement between and aggregation in the breeding and wintering sites of migratory wild birds have contributed to the long-distance dissemination and reassortment of HPAI viruses [48][49][50]. Subclinical infection of wild birds with low levels of clinical signs and prolonged viral excretion may facilitate efficient conveyance of the virus to unaffected areas or other susceptible hosts. Terrestrial birds can freely interact with wild birds and poultry through water and food Seroconversion was confirmed in all inoculated and one contact-exposed pigeon by NP-ELISA; however, H5-specific antibody was detected in one pigeon by HI assay. Previous studies reported similar results that the HI test had lower sensitivity than the commercially available competitive NP-ELISA for non-gallinaceous species [33,53], suggesting an underestimation of the immune response in AIV-exposed pigeons.
Influenza virus antigens were detected in many internal organs of mandarin ducks and pigeons infected with clade 2. 3 [35,37,39,54]. Despite the systemic distribution of the virus in mandarin ducks and pigeons, the birds did not show any detectable clinical signs in this study.
Pigeons are potential intermediate hosts of H5 HPAIV with less possibility of effective viral propagation and transmission than species of Anseriformes or Galliformes and are suggested not to play a significant role in the maintenance of the virus [32][33][34][35]. Cladedependent differences in pathogenicity and mortality in the course of H5 HPAI infection have been reported in previous experimental infection studies [36][37][38][39][40][41]55], and natural infection cases of H5 HPAIV in pigeons have been identified [56][57][58]. However, few studies have reported intra-species transmission of the H5 HPAIV in pigeons after infection [42]. In our previous experiment using clade 2 infection depending on host species has been reported [37,[59][60][61]. In contrast to Gs/GD H5N1 HPAIV, which can efficiently replicate and induce disease with various symptoms in wild birds and domestic poultry [62], clade 2.3.4.4 H5Nx HPAIV exhibited mild or no clinical signs of infection in wild waterfowl [63,64]. This suggests better adaptation of clade 2.3.4.4 H5Nx HPAIV to waterfowl and a consequent influence on long-distance dispersal by wild migratory birds. The potential adaptation to a specific host species may lead to decreased pathogenicity, increased viral shedding, changes in the viral shedding pattern or duration, or a decreased infectious dose [63]. Clade 2.3.4.4 H5N6 HPAIV with different internal gene origins may exhibit altered pathogenicity in avian and mammalian animal models [65]. Therefore, the emergence of novel genotypes due to reassortment should be carefully monitored during HPAIV outbreaks, and possible alterations in pathogenicity and transmissibility in susceptible host species need to be evaluated. Subclinical infection and replication of clade 2.3.4.4 H5N6 HPAIV in mandarin ducks and domestic pigeons, and reports of human infection with highly similar H5N6 viruses, including fatal cases from China, highlight the potential for the maintenance and dissemination of zoonotic H5N6 viruses to distant geographical areas. Previous HPAI surveillance strategies in Korea focused on migratory waterfowl. However, HPAI infections in wild terrestrial birds are of concern as they might facilitate the transmission of viruses between wild and domestic birds. Considering the possibility of potential dispersal and maintenance of HPAI viruses through these wild bird species, enhanced active surveillance in both migratory and terrestrial wild birds should be implemented.