Five species of indigenous North American ducks were used in this study: mallard (Anas platyrhynchos), northern pintail (Anas acuta), blue-winged teal (Anas crecca), redhead (Aythya americana), and wood duck (Aix sponsa). Species were selected to represent the diverse habitat and behavior of ducks in North America and included important AIV reservoirs (mallard), long-distant migrants (northern pintail and blue-winged teal), diving ducks (redhead), and birds that breed in both northern and southern areas of the United States (wood duck). All ducks used in this study were captive-bred and acquired at 10 to 16 weeks of age (Howell's Exotic Waterfowl, Muldrow, OK, USA). This age is consistent with premigration staging in the late summer or early fall when AIV prevalence peaks in wild waterfowl (7). Both male and female ducks were included in each species and were approximately equally represented.

Wild-caught gulls used in this investigation were acquired through the Southeastern Cooperative Wildlife Disease Study, University of Georgia (UGA), under federal permit. Nestling laughing gulls (Larus atricilla) were hand-caught in McIntosh County, Georgia, by personnel from the Georgia Department of Natural Resources and maintained at the College of Veterinary Medicine, UGA. At 12 weeks of age the gulls were transported to biosafety level 3–agriculture (BSL-3-Ag) facilities at the Southeast Poultry Research Laboratory (SEPRL), Agricultural Research Service, United States Department of Agriculture (USDA).

All birds used in this study were cared for in accordance with the guidelines of the Institutional Animal Care and Use Committee, as outlined in the Guide for the Care and Use of Agricultural Animals in Agricultural Research and Teaching (8) and under an animal use protocol approved by the Institutional Animal Care and Use Committee at both SEPRL and UGA. All experiments were performed in the USDA-certified BSL-3-Ag facility at SEPRL (9).

Two viruses were used in this study: A/Whooper Swan/Mongolia/244/05 (H5N1) (Mongolia/05) and A/Duck Meat/Anyang/01 (H5N1) (Anyang/01). The Mongolia/05 isolate was obtained from a dead whooper swan (Cygnus cygnus) and was chosen because of its known lethality in wild waterfowl. The Anyang/01 isolate was chosen on the basis of results from previous experimental infections of Pekin white ducks (Anas platyrhyncos), which did not result in illness or death (10).

Individual stocks of both AIVs used in this study were produced by second passage in 9-day-old embryonated chicken eggs. Allantoic fluid from the inoculated eggs was diluted in brain-heart infusion (BHI) medium to yield a final titer of 10⁶ embryo 50% infectious doses (EID₅₀) per 0.1 mL (single bird inoculum). A sham inoculum was prepared by diluting sterile allantoic fluid 1:30 in BHI.

Preinoculation serum was collected from each bird to confirm they were serologically naïve to influenza A viral antigens by agar gel precipitin test (AGP) and to H5 influenza by specific hemagglutination inhibition (HI) testing by using standard procedures (11). In addition, oropharyngeal and cloacal swabs were collected before inoculation to confirm an AIV-free status. The 5 species of ducks and laughing gulls were each separated into a control group and 2 virus-inoculated groups (Mongolia/05 and Anyang/01), each consisting of 3 birds. Ducks and gulls were inoculated intranasally with a 0.1-mL volume of the designated virus solution or sham-inoculum. All birds were monitored daily for illness or death. Due to the lack of illness exhibited by most ducks, experiments with these species were extended to 20 days postinoculation (DPI) to allow adequate time for seroconversion. Cloacal and oropharyngeal swabs were collected in BHI medium with antimicrobial drugs (100 μg/mL gentamicin, 100 units/mL penicillin, and 5 μg/mL amphotericin B) from all birds at 1, 2, 3, 4, 5, 7, 10, and 14 DPI. Oropharyngeal and cloacal swabs were also collected on 20 DPI from the 5 species of ducks. At 14 DPI (gulls) or 20 DPI (ducks), serum was collected from the surviving birds for serologic testing with HI and AGP, and the birds were humanely killed with intravenous sodium pentobarbital (100 mg/kg bodyweight). Serum was not collected from birds that died during the course of the study (that were not killed at the end of the study). Necropsies were performed on all birds, and routine tissues were collected for histopathologic and immunohistochemical evaluation. In addition, portions of heart, breast muscle, kidney, lung, and brain, and oropharyngeal and cloacal swabs were collected and stored in BHI medium with antimicrobial drugs for virus isolation.

Tissues samples collected at necropsy were preserved in 10% neutral buffered formalin. After fixation, the tissues were routinely processed and embedded in paraffin. Sections were cut at a thickness of 5 μm and stained with hematoxylin and eosin. Duplicate sections were immunohistochemically stained by using a mouse-derived monoclonal antibody (P13C11) specific for type A influenza virus nucleoprotein antigen as the primary antibody (SEPRL, Athens, GA, USA). Procedures used to perform the immunohistochemical testing followed those previously described (12). Fast red was used as the substrate chromagen, and slides were counterstained with hematoxylin. Demonstration of viral antigen was based on chromagen deposition in the nucleus, with or without chromagen deposition in the cytoplasm.

Oropharyngeal and cloacal swabs and tissue samples collected at necropsy were stored at -70°C until virus isolations and titrations were performed. Isolation of virus from swabs and tissues was performed by using embryonated chicken eggs (11). Positive samples were titrated by determining the EID₅₀. The minimal detectable titer was 10^0.97 EID₅₀/mL from swabs and 10^1.97 EID₅₀/g from tissues.

AGP and HI tests were performed on the pre- and postinoculation serum by using standard procedures (11). The HI tests were performed by using a 0.5% suspension of chicken erythrocytes in phosphate-buffered saline.

In addition to the 2 H5N1 viruses used in this study, A/chicken/Hong Kong/220/97 (H5N1) (Hong Kong/97) was included in the phylogenetic analysis because it is the only other H5N1 HPAI virus evaluated in multiple avian species by experimental inoculation (13). Sequence comparisons of these 3 viruses were conducted with the MegAlign program by using the ClustalV alignment algorithm (DNASTAR, Madison, WI, USA), and phylogenetic relationships were estimated by the method of maximum parsimony (PAUP software, version 4.0b10; Sinauer Associates, Inc, Sunderland, MA, USA) by using a bootstrap resampling method with a heuristic search algorithm. Pairwise sequence comparisons were done within the MegAlign program.

Morbidity and mortality data are summarized in Table 1. Wood ducks were the only species of duck to exhibit illness or death after inoculation with either of the HPAI viruses. Severe clinical disease developed in 2 Mongolia/05-inoculated wood ducks, characterized by cloudy eyes, ruffled feathers, rhythmic dilation and constriction of the pupils, severe weakness, incoordination, tremors, and seizures (Figure 1). One of these ducks died at 7 DPI and the other was humanely killed at 8 DPI because of its moribund condition. Two Anyang/01-inoculated wood ducks became ill with clinical signs similar to those described for the Mongolia/05 virus group. One of these ducks died, and the other slowly recovered over 7 days until it exhibited no clinical symptoms. One wood duck in each viral group remained clinically normal for the entire trial. Clinical signs were not observed in the remaining duck species.

All 3 Mongolia/05-inoculated laughing gulls exhibited severe clinical signs consisting of cloudy eyes, ruffled feathers, weakness, and incoordination, torticollis, or both. Two of these gulls died. The remaining gull clinically improved and stabilized over 6 days, but retained a head-tilt for the remainder of the trial. Severe clinical signs developed in all Anyang/01-inoculated gulls, similar to those seen in Mongolia/05-inoculated gulls. The disease progressed to death in 2 of these gulls. The remaining gull exhibited clinical signs for 8 days but gradually recovered until it showed no clinical symptoms.

Viral-induced lesions were found only in the wood ducks and laughing gulls that exhibited clinical signs. Lesions were mild in birds that recovered but were severe and widespread in birds that died or were humanely killed due to severe illness. For each species, the severity and distribution of lesions were the same for both H5N1 viruses, with the following exception described below.

Gross lesions were not present in any of the recovered birds. Wood ducks that died had multiple petechial hemorrhages in the pancreas, whereas laughing gulls had more widely distributed petechial hemorrhages in the ventriculus, apex of the heart, cerebrum, and pancreas.

On histopathologic examination, wood ducks that died had severe, diffuse neuronal necrosis in the cerebrum (Figure 2A) and, less commonly, in the cerebellum. Other common lesions included necrotizing pancreatitis (Figure 2D) and adrenalitis (Figure 2C) and multifocal myocardial necrosis. Myocardial necrosis was only observed in wood ducks inoculated with Mongolia/05 and not with Anyang/01. Necrotizing pancreatitis and cerebral neuronal necrosis were the most common lesions in gulls that died during the study. Necrotizing adrenalitis was also observed in gulls that died but was less common and milder than the changes in the pancreas and cerebrum. Microscopic lesions in wood ducks and laughing gulls that recovered were less severe than in those that died. In both species, the most common lesions in recovered birds were lymphoplasmacytic perivascular encephalitis and heterophilic pancreatitis.

Wood ducks that died during the study had viral antigen in numerous organs, including the brain (Figure 2B), adrenal glands, testicles, kidneys, liver, small intestines, heart, skeletal muscles, pancreas, and air sacs. Viral antigen was most frequently found in cardiac myocytes, parasympathetic ganglia in the submucosal and muscular plexus of the small intestines, and numerous cell types in the brain, including glial cells, ependymal cells, endothelial cells, neurons, and gitter cells. Viral antigen was also detected in the pancreatic acinar cells and cortical and medullary cells of the adrenal gland, although less often than the aforementioned sites. Minimal amounts of viral antigen were detected in the kidney and testis in 1 and 2 wood ducks that died, respectively. The 1 wood duck that recovered had a scant amount of viral antigen in the cerebellar neurons. Laughing gulls that died during the study had viral antigen most frequently detected in the neurons, endothelial cells, glial cells, and ependymal cells in the brain, pancreatic acinar cells, and cortical and medullary cells of the adrenal glands. Laughing gulls that died also had minimal amounts of viral antigen present in other organs including the heart, lungs, air sacs, thymus, kidneys, small intestines, and eyes. Laughing gulls that recovered contained small amounts of viral antigen in the pancreatic acinar cells and cerebral and cerebellar neurons.

The virus isolation results are summarized in Table 1 and Table 2. Viral titers were higher in oropharyngeal swabs than in cloacal swabs in all species and with both H5N1 viruses. Viral titers on cloacal swabs were low, except from birds that died of AIV infection. Oropharyngeal swabs from all species collected at 1 and 2 DPI were positive on virus isolation. Wood ducks and laughing gulls had higher viral titers on oropharyngeal and cloacal swabs and shed virus longer than any of the other species. Virus was isolated from numerous organs in the wood ducks and laughing gulls that died.

Serologic testing results are summarized in Table 3. Both the AGP and HI tests detected postinoculation antibodies in all surviving wood ducks and laughing gulls. However, the effectiveness of these tests in the remaining duck species was variable and dependent on host species and inoculated virus. The HI test detected postinoculation antibodies in multiple avian species that had little serologic response or none as determined by the AGP test (Anyang/01-inoculated mallards, redheads, and northern pintails and Mongolia/05-inoculated mallards). Both serologic tests were least effective in northern pintails and mallards.

In comparing the 3 viruses (Hong Kong/97, Anyang/01, and Mongolia/05) genetically, the hemagglutinin genes are all clearly in the Goose/Guandong/96 lineage. At the amino acid (aa) level they vary by 3.5%–4.8%. They all have the HA cleavage compatible with an HPAI phenotype. The cleavage site is the same for Hong Kong/97 and Anyang/01, but the Mongolia/05 virus has 2 aa changes at the cleavage site. Phylogenetically, Hong Kong/97 and the Anyang/01 are in or close to clade 3, and the Mongolia/05 strain is in clade 2 (14). The Mongolia/05 strain appears to be a representative isolate from the wild bird viruses that have been reported in Asia, Europe, and Africa.

Comparison of the other 7 gene segments demonstrates evidence of reassortment. The viruses from the 1997 outbreak in Hong Kong have a unique subtype-1 neuraminidase gene compared with any of the other H5N1 viruses. The Anyang/01 and Mongolia/05 N1 genes are from the same lineage, and both have an identical 20 aa stalk deletion. For the other 6 internal genes, the Anyang/01 and Mongolia/05 viruses in general were more closely related to each other than to the Hong Kong/97 virus. Except for the H5 gene, the Hong Kong/97 and viruses isolated in Hong Kong in the same year appear to be a unique constellation of genes that has not been seen again. Although the Anyang/01 and Mongolia/05 viruses were more closely related, the internal genes are most likely the result of reassortment with other influenza viruses and not the result of progressive sequence in a single lineage of viral genes.