Detection of West Nile virus (WNV) has been reported in a variety of wild ducks in the US, but little is known about the pathogenesis and outcome of exposure of the disease in these species. Previous experimental studies of WNV in ducks either have challenged a small number of ducks with WNV or have tested domesticated ducks. To determine susceptibility and immune response, we challenged 7-wk-old Wood Ducks (Aix sponsa) with a 1999 American Crow (Corvus brachyrhynchos) isolate of WNV. Wood Ducks were susceptible to infection with the virus, and, although clinical signs or mortality were not observed, microscopic lesions were noted, particularly in the heart and brain. West Nile virus viremia peaked on day 2 postinfection (pi) at 104.54 plaque-forming units (PFU) of virus/mL serum and WNV was shed orally (between 102 and 102.9 PFU per swab) and cloacally. Specific anti-WNV antibody response was rapid, with anti-WNV IgM detected on day 3 pi followed on day 5 pi by anti-WNV IgG. Neutralizing antibodies were detected by plaque-reduction neutralization assay in one duck on day 4 pi, and in all sampled ducks on day 5. These results indicate that Wood Ducks are susceptible to WNV, but it is unlikely that significant WNV mortality events occur in Wood Ducks or that they play a significant role in transmission. However, WNV viremia was sufficient, in theory, to infect mosquitoes, and oral and cloacal shedding of the virus may increase the risk of infection to other waterbirds.
West Nile virus (WNV) was first detected in North America in the New York City area in 1999 (Centers for Disease Control and Prevention [CDC] 1999; Lanciotti et al. 1999; Komar 2003), and has since spread across the continental US and southern Canada (Petersen et al. 2003; Hofmeister 2011; Kilpatrick 2011; Reisen 2013). West Nile virus has caused significant mortality in North American land birds including corvids (Yaremych et al. 2004; Caffrey et al. 2005; Ladeau et al. 2007), raptors (Saito et al. 2007), and Greater Sage-grouse (Naugle et al. 2004; Walker et al. 2004). Less is known about the effects of WNV on waterbirds, with the exception of large mortality events involving juvenile American White Pelicans (Pelecanus erythrorhynchos) in the northern plains states (Sovada et al. 2008) and, more recently, Eared Grebes (Podiceps nigricollis) at the Great Salt Lake, Utah (Ip et al. 2014). Mortality events have also been reported in captive and farmed ducks and geese (Anatidae) in the US and Canada (Meece et al. 2006; Wojnarowicz et al. 2007; Himsworth et al. 2009). Additionally, in the US, WNV has been detected in several species of wild ducks, including Wood Ducks (Aix sponsa) (CDC 2012). Experimentally, the WNV susceptibility of Mallards (Anas platyrhynchos) (Komar 2003) and Aigamo Ducks (A. platyrhynchos var. domesticus), a cross between a wild duck and a domestic duck, farmed as a meat source, has been reported (Shirafuji et al. 2009). In both studies, ducks were susceptible to WNV and developed viremia titers sufficient to infect feeding mosquitoes and, in one study, potentially capable of infecting other waterbirds through virus shed orally (Shirafuji et al. 2009). A year prior to conducting this study, we had initiated a multiyear WNV serosurvey of freshwater ducks captured at National Wildlife Refuges in North Dakota, South Carolina, and Tennessee. The initial results showed a prevalence of WNV antibody in Mallards and Northern Pintails (Anas acuta) (39/362, 11%) that was more than twice the prevalence in Wood Ducks (7/150, 5%) (E.H. unpubl. data). This observation might result either from a significant difference in risk of exposure to WNV or from a difference in survival following infection. To address the question of survival, we experimentally challenged Wood Ducks with WNV to determine their susceptibility and immune response. Wood ducks inhabit wooded swamps and marshes in North America and nest in natural and artificial cavities, which provide an environment where they may acquire WNV from blood-feeding mosquitoes. The potential for Wood Ducks to contribute to WNV transmission is unknown.
MATERIALS AND METHODS
Wood ducks were obtained as 4-wk-old ducklings from a commercial breeder and were transported to the Biosafety Level 3 Animal Isolation Wing of the National Wildlife Health Center (NWHC) in Madison, Wisconsin, USA. On arrival the birds were given a physical examination, marked with a uniquely numbered leg band, and housed in two groups consisting of five (non–WNV-inoculated controls housed in a separate animal room) and 17 (three contact controls and 14 WNV-challenged) birds. Birds were housed on Tenderfoot flooring (Vittletoe, Inc., Keota, Iowa, USA), with small pools available for exercise and bathing, and were fed a commercial feed (Duck Grower Mash, Purina Mills, Gray Summit, Missouri, USA). The small pools and feed dishes were cleaned and water and food were resupplied daily. Birds were observed twice daily by researchers or animal husbandry staff. All transportation, animal care, WNV challenge, and clinical sampling were approved by the NWHC Institutional Animal Care and Use Committee.
WNV challenge and sampling
One week prior to the WNV challenge, at 6 wk of age, a 500-µL blood sample was obtained from the jugular vein of each duckling for baseline WNV serology. Blood was placed directly into an appropriate amount of BA1 medium for a final serum dilution of 1∶5, assuming a 50∶50 ratio of blood cells to serum in whole blood samples. BA1 medium contains 1× M199 media with Hanks salts (Sigma Chemical Co., St. Louis, Missouri, USA) containing 0.8% Tris HCl, 0.008% sodium bicarbonate, 1% bovine serum albumin, 20% fetal bovine serum, 1,000 U/mL penicillin, and 100 µg/mL streptomycin (Gibco, Grand Isle, New York, USA) and 1 µg/mL amphotericin B (Gibco). Blood samples were allowed to clot on wet ice and centrifuged at 2,000 × G for 10 min at 4 C, and the serum was separated and stored at −80 C. Challenged birds were injected subcutaneously with 100 µL of BA1 medium containing 105 plaque-forming units (PFU) of a low-passage 1999 American Crow (Corvus brachyrhynchos) isolate of WNV (NWHC 16399-2). All control birds were injected with 100 µL of BA1 medium. One WNV-challenged bird died from causes unrelated to the study on day 1 following challenge.
The WNV-challenged and control ducks were divided into two subgroups and sampled on alternate days during days 1–8 postinfection (pi). All birds were also sampled on days 10 and 14 pi and, following the sampling on day 14, were euthanized by inhalation of CO2 gas and necropsied. On each sampling day, a 500-µL blood sample was obtained from the jugular vein and processed as previously described, and an oral and a cloacal sample were obtained using sterile, cotton-tipped swabs that were placed in 1 mL of BA1 medium. The swabs were placed on wet ice until all samples were collected, at which time they were frozen at −80 C. Immediately after euthanasia, the ducks were necropsied and duplicate samples were obtained from the following organs and placed in 1 mL of BA1 and in 10% buffered formalin: skin, trachea, lung, heart, crop, ventriculus, intestine, pancreas, liver, spleen, kidney, bone marrow, sciatic nerve, pectoral muscle, optic tectum, cerebrum, cerebellum, and brain stem. Tissues in BA1 were frozen at −80 C for virus detection by culture and by reverse-transcriptase PCR. Tissues in 10% neutral buffered formalin were trimmed and embedded in paraffin, sectioned at 5 µm, and stained with hematoxylin and eosin.
Virus detection by culture
Serum (1∶5 dilution in BA1) was cultured on confluent Vero cells (ATCC CCL-81, American Type Culture Collection, Manassas, Virginia, USA) plated at a concentration of 2 × 105 cells/mL in complete M199 (Sigma) containing 10% standard fetal bovine serum (Hyclone, Logan, Utah, USA), 1000 U/mL penicillin, 100 µg/mL streptomycin, and 100 µg/mL amphotericin, and plates were incubated at 37 C, 2.5% CO2, for 4 d. Serum and swab samples were diluted in complete M199, added to confluent Vero cell monolayers, and incubated at 37 C for 45 min before complete overlay was added (Beaty et al. 1989). Serum samples were screened at a 1∶10 dilution and positive samples were titered in a twofold series (1∶20–1∶640). Following inoculation, cultures were checked for plaque formation daily between 3 and 5 d, at which time the number of plaques was recorded and virus concentration reported as PFU/mL of serum. Swabs were cultured at a 1∶5 dilution and viral plaques were recorded at 96 h. Cloacal swab samples that produced Vero cell lysis within 24 h were treated by 0.2 µm filtration, by culture at a 1∶40 dilution, or by both methods and were recultured on Vero cells, and viral plaques were recorded at 96 h.
All serum samples were heat inactivated at 56 C for 30 min and tested for specific anti-WNV antibodies by IgM and IgG enzyme immunoassay (EIA). For detection of IgM antibodies, the chicken (Gallus gallus domesticus) WNV IgM-capture EIA (Johnson et al. 2003) was utilized with slight modifications. Briefly, goat anti-duck IgM antibody (Nordic MUbio, Susteren, The Netherlands) was used as the capture antibody and was diluted 1∶500 in carbonate bicarbonate buffer (pH 9.6) and incubated overnight at 4 C. Test sera (50 µL), diluted to 1∶100 in wash buffer, were tested in duplicate on positive and negative WNV antigen (Hennessey Research, Shawnee, Kansas, USA). Captured viral antigen was detected by horseradish peroxidase–conjugated anti-WNV monoclonal antibody 6B6C-1 (MAB8744, EMD Millipore, Billerica, California, USA) reacted with ABTS peroxidase substrate (Kirkegaard and Perry Laboratories, Gaithersburg, Maryland, USA). Absorbance at 405 nm was read using a microtiter absorbance plate reader. Each plate contained a high- and a low-titered anti-WNV duck serum sample, as previously determined by plaque-reduction neutralization test (PRNT), and a negative control duck serum sample. The optimal reagent concentrations were determined by checkerboard titration. To assess specific WNV antibody status, the ratio of the mean optical density (OD) of a sample recorded on wells containing recombinant WNV E protein antigen divided by the OD of the same sample recorded on wells containing negative control antigen was calculated. Secondly, for samples in which the first ratio exceeded 2, the ratio of mean test serum OD to mean negative control serum OD (P/N), recorded on wells containing recombinant WNV E protein antigen, was calculated for each test sample. A P/N ratio of ≥2 was considered positive for anti-WNV IgM in a test sample.
Similarly, an indirect WNV IgG EIA (Johnson et al. 2000) was modified for duck sera as follows. Briefly, flavivirus group–specific monoclonal antibody 4G2 (MAB10216, EMD Millipore) was used as the WNV capture antibody at a 1∶2,000 dilution in carbonate bicarbonate buffer. Test sera (50 µL) diluted to 1∶100 in wash buffer were tested in duplicate on positive and negative WNV antigen (Hennessey Research). Bound duck antibody was detected using 50 µL of a 1∶500 dilution of horseradish peroxidase–conjugated goat anti duck IgG (H+L) antibody (Kirkegaard and Perry Laboratories) reacted with ABTS substrate, and the OD was determined. High-titered and low-titered sera, collected on day 14 pi and determined by PRNT, were used as positive controls and baseline serum was used as negative control on each test plate. To determine IgG-specific anti-WNV antibody, the P/N ratio was calculated.
Neutralizing antibodies to WNV were determined by PRNT (Beaty et al. 1989). Sera were screened at a 1∶20 final dilution on Vero cells grown on six-well plates. Titered stock WNV (NWHC 16399-2) was diluted in BA1 so that the test dose of virus contained between 80 and 100 plaques, and sera resulting in a 90% plaque reduction (PRNT90) were considered positive and were subsequently tested by PRNT using a twofold dilution series until an endpoint titer was reached.
Mean viremia, EIA P/N ratios, and standard errors were calculated using GraphPad Prism version 5.00 for Windows (GraphPad Software, San Diego, California, USA). These were plotted using GraphPad Prism and the R statistical program (R Development Core Team 2014). The geometric mean endpoint titer of final serum samples collected at 14 d pi was calculated.
All ducks challenged with WNV by needle injection became infected with WNV based on viral culture of the virus from serum samples. Mortality or clinical signs of viral infection were not detected in challenged birds. None of the contact controls became infected with WNV and no clinical signs of disease were detected in controls.
Viremia and shedding of WNV
We detected WNV by culture in serum samples from day 1 through day 3 pi (Fig. 1). Viremia titers peaked in ducks on day 2 pi (mean 104.54 PFU/mL, SE 0.168) with a peak viremia titer of 105.33 PFU/mL of serum detected in one bird. However, by day 3 pi viremia titer was above the limit of detection (101.7 PFU/mL) in only two of seven sampled ducks (29%) with viremia titers of 102.82 PFU/mL and 103.6 PFU/mL of serum.
Oral shedding of WNV was detected between day 1 and day 4 pi (Table 1). The initial culture of cloacal swabs at a 1∶5 dilution resulted in cell death within 24 h in many samples and was accompanied by the cell overlay turning yellow. Using filtration, dilution of swab sample, or both of these techniques resulted in cells that remained viable for 3–4 d and revealed discrete plaques. West Nile virus was detected in cloacal swabs on days 3 and 4 pi, and, with the exception of one duck on day 3, the quantity of virus was 101.6–101.9 PFU per swab. A cloacal swab collected on day 3 from one duck contained approximately 103.6 PFU of WNV. Generally, if a bird was shedding WNV orally it was also shedding virus from the cloaca (data not shown).
We detected WNV-specific IgM first on day 3 pi in all challenged birds sampled on that date (Fig. 2). The mean peak IgM P/N ratio occurred at 1 wk pi (mean P/N 37.6) and then decreased significantly to day 14 pi (mean P/N 13.4, paired two-tailed Student's t-test t = 12.1, P<0.001). We detected WNV-specific IgG first by EIA on day 5 pi in all challenged birds (mean P/N 6.1), and the P/N ratio increased significantly through day 14 pi (mean P/N 40.5, posttest for linear trend P<0.001) (Fig. 3). Neutralizing antibodies to WNV were detected in all birds from 5 to 14 d pi (Table 2). Overall, the geometric mean titer increased over the first week pi, but then declined slightly (Table 2).
No gross lesions were detected in any of the ducks on necropsy. Tissues from 13 WNV-challenged ducks, three contact controls, and five non–WNV-challenged control ducks were examined microscopically. Of the 18 tissues examined from the 13 WNV-challenged birds, microscopic lesions were most frequently observed in the heart (n = 10, 77%), cerebellum (n = 9, 69%), brain stem (n = 6, 46%), cerebrum (n = 4, 31%), optic tectum (n = 4, 31%), and skin (n = 4, 31%). Lesions were also found in one sample from the liver and pancreas of one WNV-infected duck, and in samples of intestine from two other birds. Lesions in brain tissue consisted of mild gliosis, perivascular cuffs of mononuclear cells, and, in one case, mild nonsuppurative encephalitis in the cerebellum. In the other affected tissues, lesions consisted of mild to moderate infiltrates of mononuclear cells (primarily small lymphocytes) in the heart and pancreas, mild perivascular cuffs of lymphocytes in the superficial dermis, and mild perivascular cuffs of lymphocytes in the tunica muscularis of the proventriculus and ventriculus. No lesions were observed in the trachea, lung, crop, ventriculus, intestine, pancreas, liver, spleen, kidney, bone marrow, sciatic nerve, or skeletal muscle. Among WNV-challenged birds, lesions were detected in as few as one tissue sample and as many as six tissues. However, there was no correlation between the peak viremia and the number of tissues in which lesions were observed (data not shown).
Wood Ducks, both captive and wild, have been listed among avian species in the US in which WNV has been detected (CDC 2012), but the severity of disease caused by WNV in Wood Ducks has not been assessed. We established that, like domestic ducks previously tested, Wood Ducks are susceptible to WNV infection, but the virus is unlikely to cause significant mortality events in Wood Ducks. We also showed that viremia titers following WNV infection may infect only a small proportion of feeding mosquitoes (Lord et al. 2006), and, although Wood Ducks potentially contaminate their environment through oral and cloacal shedding of the virus, the risk this may pose to other waterbirds is unknown.
Viremia in experimentally infected Wood Ducks followed the same timeline as that reported for Aigamo Ducks (Shirafuji et al. 2009), infected intramuscularly with 103 PFU of a 1999 isolate of WNV, and for mallards (Komar et al. 2003), infected by mosquito feeding. However, the viremia developing in mallards was two logs higher than peak viremia in either Wood Ducks or Aigamo Ducks and may have been due to greater susceptibility of mallards to WNV or to enhancement of WNV infection by mosquito saliva, which has been reported in experimentally infected mice (Schneider et al. 2007; Styer et al. 2011) and chickens (Styer et al. 2006). The proportion of Wood Ducks that shed WNV orally is similar to the range of oral shedding of WNV in Aigamo Ducks; however, Aigamo Ducks shed 1–2 logs more WNV. The quantity of WNV detected in Wood Duck cloacal swabs was lower than that detected in swabs obtained from Aigamo Ducks, and the duration of WNV shedding was slightly shorter (2 d as compared to 4 d) (Shirafuji et al. 2009); however, the sensitivity of detection of WNV in cloacal swabs in our study may have been affected by treatments to avoid contamination. Infected Wood Ducks or Aigamo Ducks may spread WNV by oral and cloacal shedding, which, in their aquatic environment, may result in fecal oral transmission of WNV to other waterbirds.
We detected immunoglobulin class–specific antibody response to using commercial specific anti-duck IgM- and IgG-conjugated antibodies. These class-specific conjugates allowed us to monitor the development of a specific anti-WNV serologic response. A WNV-specific IgM response was detected 2 d before IgG response, and by day 14 the IgM response had decreased whereas the IgG response had increased. The early anti-WNV IgM response in ducks in this study is similar to that reported in needle-inoculated chickens (G. gallus domesticus), in which anti-WNV IgM was detected at 4 d pi (Jankowski et al. 2010) and 5 d pi (Johnson et al. 2003). Similar to previous results in chickens (Jankowski et al. 2010), the IgM response was followed on day 5 by all ducks having detectable anti-WNV IgG. Normally, specific antigen–stimulated B lymphocytes initially produce IgM, but with repeated exposure to antigen and maturation, the constant region of the immunoglobulin molecules are converted to IgG, most properly referred to as IgY, in birds (Hartle et al. 2013).
In WNV-challenged ducks, neutralizing antibody to WNV was detected as early as 4 d in one bird and in all seven birds sampled on day 5. West Nile virus–neutralizing antibodies have been reported as early as 5–7 d pi in experimentally challenged chickens (Senne et al. 2000; Langevin et al. 2001). Endpoint PRNT titers at 14 d in the present study are comparable to those developing in chickens at 14 d (Senne et al. 2000; Langevin et al. 2001).
In agreement with a previous report on a WNV outbreak in Rouen Ducks (A. platyrhynchos), microscopic examination of tissues from experimentally infected Wood Ducks revealed lesions primarily in the heart and brain (Wojnarowicz et al. 2007). Consistent with the absence of clinical disease in experimentally infected Wood Ducks, the histologic lesions were less severe than those described in Rouen Ducks. Lesions in heart, brain, and pancreas of experimentally infected Wood Ducks occurred less frequently than those described in naturally infected Lesser Scaup (Aythya affinis). Lesser Scaup affected by WNV in that outbreak were approximately 3–5 wk younger and likely less immunocompetent than those in our study (Himsworth et al. 2009). Mortalities and clinical neurologic disease were reported in naturally occurring WNV outbreaks in Rouen Ducks (Wojnarowicz et al. 2007) and juvenile Lesser Scaups (Himsworth et al. 2009).
Considering the lack of clinical signs of disease, the rapid specific-antibody response, and the mild microscopic lesions observed, Wood Ducks fall into the susceptibility class suggested by Busquets et al. (2012), in which the species suffers only minor clinical signs with WNV infection and mounts an early and robust immune response. Our study indicates that Wood Ducks are unlikely to develop sufficient viremia to infect a significant number of mosquitoes. Thus, although WNV shed in oral and cloacal secretions of infected Wood Ducks may contaminate the environment, the risk of infection for other birds is unknown, but probably minimal.
We thank Diana Goldberg, Lovekesh Karval, Melissa Lund, and Daniel Shadduck for technical contributions and for the helpful comments on the manuscript of one USGS reviewer. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the US government.