Low‐pathogenicity influenza viruses replicate differently in laughing gulls and mallards

Abstract Wild aquatic birds are natural reservoirs of low‐pathogenicity avian influenza viruses (LPAIVs). Laughing gulls inoculated with four gull‐origin LPAIVs (H7N3, H6N4, H3N8, and H2N3) had a predominate respiratory infection. By contrast, mallards inoculated with two mallard‐origin LPAIVs (H5N6 and H4N8) became infected and had similar virus titers in oropharyngeal (OP) and cloacal (CL) swabs. The trend toward predominate OP shedding in gulls suggest a greater role of direct bird transmission in maintenance, whereas mallards shedding suggests importance of fecal‐oral transmission through water contamination. Additional infectivity and pathogenesis studies are needed to confirm this replication difference for LPAI viruses in gulls.


| ME THODS
North American LPAIVs used in experiments are listed in Table 1; they were propagated in specific pathogen-free (SPF) 9-to 11-day-old embryonating chicken eggs (ECE) following standard procedures. 6 Low-passage virus stocks were used for challenge.
Laughing gulls (7-10-days-of-age) were obtained under federal permit and reared for 12 weeks in captivity until challenged. Ten-to 16-week-old mallards were purchased from a commercial hatchery (Chenoa Waterfowl). For challenge (Table 1), birds were grouped and housed in negative pressure high-efficiency particle air (HEPA) ventilated cabinets with ad libitum access to feed and water.
Laughing gulls and mallards were divided in groups and inoculated with respective gull-and mallard-origin LPAIV (Tables 1 and 2) via the choanal cleft, which provides exposure to upper respiratory tract and drainage into oral cavity for swallowing and exposure to gastrointestinal tract, with approximately 10 6 mean embryo infectious doses (EID 50 ) in 0.1 ml per bird. Back titers were reported in Table 1. Oropharyngeal (OP) and cloacal (CL) swabs were collected on 1, 2, 3, 4, 7, and 10 days of post-inoculation (dpi) and placed in Becton-Dickinson BBL brain heart infusion (BHI) medium with 2× concentration of antibiotics (10,000 U/ml Penicillin G, 10,000 μg/ml Streptomycin, 25 μg/ml Amphotericin B) (HyCone Laboratories, Inc). Samples were stored at −8°C until tested. Blood was collected pre-(0-day) and post-inoculation (10 days) to assess serum antibody responses. Birds were observed daily for clinical signs and euthanized at 10 dpi following approved protocols. These studies were reviewed and approved by the USNPRC Institutional Animal Care and Use Committee (IACUC) and conducted with appropriate biocontainment and biosafety measures.
OP and CL swabs were processed to determine viral shedding titers by quantitative real-time PCR (RRT-PCR). Briefly, the RNA was extracted using MagMAXTM-96 AI/ND Viral RNA Isolation Kit ® (ThermoFisher Scientific) following the manufacturer's instruction. Further, RRT-PCR that targets the matrix gene of avian influenza was performed with the AgPath-ID OneStep RT-PCR kit (ThermoFisher Scientific) using 7500 FAST Real-time PCR System (Applied Biosystems), as previously described. 7 Virus quantity was established with a standard curve from RNA extracted from 10-fold dilutions of the challenge virus in duplicate.
Serum was tested for anti-AIV antibodies using hemagglutination inhibition (HI) assay and blocking enzyme-linked immunosorbent assay (bELISA). The homologous antigens were prepared as previously described 6 and the HI assay performed following standard procedures. 6 Titers were calculated as the reciprocal of the last HI positive serum dilution and were converted to log 2 . Titers were expressed as geometric mean titers (GMT-log 2 ). Samples were con-

| RE SULTS
The sham controls were not infected based on lack of pre-and postchallenge AIV antibodies and negative virus detected in OP and CL swabs after challenge. Clinical signs or mortality were not observed in any inoculated laughing gulls and mallards.  Only LG/H7N3-inoculated gulls shed to end of the 10 dpi while

TA B L E 1 Avian influenza viruses and back titers used in the experimental challenge of laughing gulls and mallards
LG/H6N4-inoculated gulls shed to 3 dpi. In contrast, laughing gulls infected with LG/H3N8 and HG/H2N3 ( Figure 1C-D) had low-OP virus titers (highest mean virus titer detection of 1.5 and 1.7 log 10 EID 50 /ml, respectively), and virus was not detected in CL swabs.
Overall, in laughing gulls viral shedding was predominantly associated with OP swabs, with highest titers observed in the first two days of post-inoculation. Virus was detected less frequently in CL swabs and titers were low ( Figure 1). Mallards inoculated with M/ H5N6 ( Figure 1E), and M/H4N8 ( Figure 1F) had viral shedding detected over the 10-days study. In the first three dpi, the mean virus shedding was higher in the OP than in the CL samples for M/H5N6inoculated mallards ( Figure 1E). This shedding pattern changed after 4 dpi with M/H5N6 virus excretion in the CL reaching peaks as 4.7 log 10 EID 50 /ml ( Figure 1E). Mallards inoculated with M/ H4N8 had a constant mean shedding titer in the OP samples during the 10 dpi, and CL shedding oscillated between days and birds with titers similar between OP and CL samples or CL slightly lower than OP swabs ( Figure 1F).
None of the laughing gulls or mallards had pre-existing anti-AIV antibodies prior to LPAIV inoculation ( None of the LPAIV-infected gulls or mallards in our experiments experienced morbidity or mortality ( Figure 1 and Table 2) which was expected with LPAIV in these host species. 1  It is possible that differences in shedding patterns between ducks and gulls may be related to the expression of α2,3-linked sialic acid (SA) receptors in tissues. 15,16 Ducks show similar expression of SA receptors in the respiratory and intestinal tract, which may explain the equal respiratory and fecal shedding pattern of our study. 15,16 In vitro studies demonstrated that SA receptors' stronger expression in respiratory tract of ring-billed gulls and laughing gulls which was consistent with predominant respiratory shedding in our study. 15,16 However, other factors present in the host and the virus strain may be also involved in the differences in AIV prevalence, viral shedding, and the disease's outcome, 5,15,16

CO N FLI C T O F I NTE R E S T
Authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are provided in the figures and tables of the article. Additional information is available from the corresponding author upon reasonable request.

O RCI D
David E. Swayne https://orcid.org/0000-0001-7472-1992 F I G U R E 2 Scatter plot of HI titers in laughing gulls and mallards pre-and post-challenge with different LPAI virus strains. The HI titers were analyzed using the challenge virus as antigen for the experimental infection in (A) laughing gulls and (B) mallards. Titers were expressed as geometric mean titers (GMT-log 2 ), and samples with titers below 3 log 2 GMT were considered negative. Dotted horizontal lines indicate limit of detection