Pathogenicity and Transmission of H5 and H7 Highly Pathogenic Avian Influenza Viruses in Mallards

ABSTRACT

Wild aquatic birds have been associated with the intercontinental spread of H5 subtype highly pathogenic avian influenza (HPAI) viruses of the A/goose/Guangdong/1/96 (Gs/GD) lineage during 2005, 2010, and 2014, but dispersion by wild waterfowl has not been implicated with spread of other HPAI viruses. To better understand why Gs/GD H5 HPAI viruses infect and transmit more efficiently in waterfowl than other HPAI viruses, groups of mallard ducks were challenged with one of 14 different H5 and H7 HPAI viruses, including a Gs/GD lineage H5N1 (clade 2.2) virus from Mongolia, part of the 2005 dispersion, and the H5N8 and H5N2 index HPAI viruses (clade 2.3.4.4) from the United States, part of the 2014 dispersion. All virus-inoculated ducks and contact exposed ducks became infected and shed moderate to high titers of the viruses, with the exception that mallards were resistant to Ck/Pennsylvania/83 and Ck/Queretaro/95 H5N2 HPAI virus infection. Clinical signs were only observed in ducks challenged with the H5N1 2005 virus, which all died, and with the H5N8 and H5N2 2014 viruses, which had decreased weight gain and fever. These three viruses were also shed in higher titers by the ducks, which could facilitate virus transmission and spread. This study highlights the possible role of wild waterfowl in the spread of HPAI viruses.

IMPORTANCE The spread of H5 subtype highly pathogenic avian influenza (HPAI) viruses of the Gs/GD lineage by migratory waterfowl is a serious concern for animal and public health. H5 and H7 HPAI viruses are considered to be adapted to gallinaceous species (chickens, turkeys, quail, etc.) and less likely to infect and transmit in wild ducks. In order to understand why this is different with certain Gs/GD lineage H5 HPAI viruses, we compared the pathogenicity and transmission of several H5 and H7 HPAI viruses from previous poultry outbreaks to Gs/GD lineage H5 viruses, including H5N1 (clade 2.2), H5N8 and H5N2 (clade 2.3.4.4) viruses, in mallards as a representative wild duck species. Surprisingly, most HPAI viruses examined in this study replicated well and transmitted among mallards; however, the three Gs/GD lineage H5 HPAI viruses replicated to higher titers, which could explain the transmission of these viruses in susceptible wild duck populations.

INTRODUCTION

Wild aquatic birds, especially of the orders Anseriformes (ducks, geese, and swans) and Charadriiformes (shorebirds, gulls, terns, and auks) are the natural reservoirs of avian influenza (AI) viruses (1). These AI viruses are highly host adapted, typically replicating in epithelial cells of the gastrointestinal tract and producing subclinical infections. Periodically, these AI viruses transmit from wild aquatic to domestic birds, producing subclinical infections or, occasionally, respiratory disease and drops in egg production, with such transmission and infections being most permissive for domestic waterfowl species (2). This virus phenotype is termed low-pathogenicity or low-pathogenic (LP) AI virus and can be any combination of the 16 hemagglutinin (HA) and 9 neuraminidase (NA) subtypes. However, a few H5 and H7 LPAI viruses after circulating in gallinaceous poultry (chickens, turkeys, quail, etc.) have mutated to produce the highly pathogenic (HP) phenotype of AI viruses (3). These HPAI viruses cause severe systemic disease and high mortality in gallinaceous poultry and are typically easily transmissible among them (4). Historically, HPAI viruses have not caused outbreaks or widespread infections in wild birds except for the die-off that occurred among common terns (Sterna hirundo) in South Africa during 1961 (5). However, since 2002, the A/goose/Guangdong/1/96 (Gs/GD) lineage H5N1 HPAI viruses have caused infections, illness, and death in a variety of captive, zoo, and wild bird species, including waterfowl (6–9).

In both wild and domestic ducks, experimental inoculation with the Gs/GD lineage of H5N1 HPAI viruses can produce a range of clinical outcomes from asymptomatic infections to severe disease with mortality (reviewed in reference 10; see also references 11, to ,18). Both sick and asymptomatic infected ducks shed high virus quantities, increasing the risk of transmission. Furthermore, migratory waterfowl have been infected with the Gs/GD lineage of H5N1 HPAI viruses in the field and, based on field epidemiology, such birds have contributed to long-distance virus spread, including intercontinentally, during three distinct time periods: 2005, 2010, and 2014 (18–28). Experimental studies in some wild duck species, especially mallards (Anas platyrhynchos), have confirmed their potential to be long-distance vectors of these viruses (17, 18).

As the Gs/GD lineage H5N1 viruses continue to circulate and spread, the HA genes have diversified into multiple genetic lineages or clades. In 2005, an H5N1 virus of the Gs/GD lineage was isolated from outbreaks with mortality in several waterfowl species in China (7). This virus lineage, defined as clade 2.2, moved through migratory birds and often was transmitted to poultry, resulting in outbreaks in more than 20 countries (25). A second spillover event from poultry to wild birds of a different H5N1 virus, clade 2.3.2.1, began appearing in wild birds in late 2007, and this lineage of virus, although not spreading as far geographically, resulted in poultry outbreaks in at least seven different countries in Asia. This lineage of virus has been detected in wild birds through 2012 (26, 27). Viruses from this clade 2.3.2.1 probably spread through wild birds to several Asian countries in 2008 and to Europe in 2010 (28). Since 2008, subclade 2.3.4.4 has reassorted with multiple neuraminidase subtypes to form widely circulating H5N2, H5N3, H5N5, H5N6, and H5N8 subtypes of HPAI viruses (22, 29–32). In early 2014, outbreaks of a reassortant H5N8 HPAI virus were reported in South Korea in poultry and wild aquatic birds (33), with migratory aquatic birds strongly suspected in playing a key role in the spread of the virus (23). In late autumn 2014, H5N8 HPAI viruses were detected in Siberia, several countries in Europe, in South Korea, and in Japan (22, 31). Concurrently, this virus was detected in the United States in captive falcons, wild birds, and backyard aquatic and gallinaceous poultry (34). In addition, another novel reassortant HPAI virus of H5 clade 2.3.4.4 (H5N2) was identified as the cause of an outbreak in poultry farms in British Columbia, Canada (35), and was subsequently detected in the United States in wild waterfowl and backyard poultry. From March to June 2015, the H5N2 reassortant virus was found in wild aquatic birds, raptors, and backyard and commercial poultry flocks in the Midwestern region of the United States (36). This reassortant H5N2 virus predominated in the United States, and extensive farm-to-farm transmissions occurred in the Midwestern region. Over 7.5 million turkeys and 42.1 million chickens died or were culled during this outbreak which ended in June 2015 (37).

Although data are limited and reported results are mixed, most H5/H7 HPAI viruses are considered to be adapted to gallinaceous species and therefore less likely to infect, replicate efficiently, and cause disease in domestic or wild ducks (38–40). Experimentally or naturally, mortality in domestic ducks caused by HPAI viruses had been infrequently reported before the Gs/GD H5N1 HPAI outbreaks in Asia (10, 38, 41). Similarly, infections and disease from HPAI viruses in domestic ducks have been rare. Wild aquatic birds are the genetic reservoirs of LPAI viruses but are not genetic or long-term reservoirs of HPAI viruses (42). In most experimental studies, ducks intranasally (i.n.) inoculated with H5 or H7 HPAI viruses showed very mild or no clinical signs (40, 43–48), but virus was isolated in some cases from tracheal and cloacal swabs (44, 47, 48) and recovered from the trachea, gut, liver, brain, and spleen (40). Based on the results of these studies, it is not clear whether domestic or wild ducks can easily become infected with HPAI viruses other than the Gs/GD lineage and transmit the virus to naive ducks.

Recurring outbreaks of H5 and H7 HPAI in poultry and the recent outbreaks of H5N8 and H5N2 HPAI underscore the need to better understand the pathogenesis and transmission of these viruses in wild birds. The goal of the present study was to describe the pathogenicity, viral shedding patterns, and transmissibility of North American clade 2.3.4.4 H5 viruses in mallards and to determine whether they differ from other Gs/GD lineage viruses and historic H5 and H7 HPAI viruses. In this study, we used mallards as a model system since they have been naturally infected with Gs/GD H5 viruses in both Eurasia and North America, they represent the waterfowl species most utilized in experimental infections, and they are closely related to the most commonly reared domestic duck species.

MATERIALS AND METHODS

Viruses.

The following 14 HPAI viruses were used in this study: Gs/GD lineage, clade 2.3.4.4—A/Northern pintail/Washington/40964/2014 (H5N2) (Np/WA/14) and A/Gyrfalcon/Washington/40188-6/2014 (H5N8) (Gf/WA/14), and clade 2.2—A/Whooper swan/Mongolia/244/2005 (H5N1) (Ws/Mongolia/05); A/chicken/Chile/184240-1/2002 (H7N3) (Ck/Chile/02), A/chicken/Canada/314514-2/2005 (H7N3) (Ck/Canada/05), A/chicken/Jalisco/CPA1/2012 (H7N3) (Ck/Jalisco/12), A/chicken/Victoria/85 (H7N4 (Ck/Victoria/85), A/chicken/North Korea/7916/2005 (H7N7) (Ck/North Korea/05); A/chicken/Netherlands/1/2003 (H7N7) (Ck/Netherlands/03), A/turkey/Italy/4580/99 (H7N1) (Tk/Italy/99), 9) A/chicken/Pennsylvania/1370/83 (H5N2) (Ck/PA/83), A/chicken/Queretaro/14588-19/95 (H5N2) (Ck/Queretaro/95), A/turkey/Ireland/1378/83 (H5N8) (Tk/Ireland/83), and A/tern/South Africa/61 (H5N3) (Tern/South Africa/61). Three LPAI viruses were included as controls: A/mallard/Minnesota/410/2000 (H5N2) (Ml/MN/00), A/mallard/Ohio/421/87 (H7N8) (Ml/OH/87), and A/mallard/Sweden/85/2002 (H7N2) (Ml/Sweden/85). Viruses were obtained from the Southeast Poultry Research Laboratory repository. The isolates used were the earliest pass in eggs available for each virus. The viruses were titrated by allantoic sac inoculation of 9- to 10-day-old embryonated chicken eggs (ECEs) according to standard procedures (49). Allantoic fluid was diluted in brain heart infusion (BHI) medium (BD Bioscience, Sparks, MD) in order to obtain an inoculum with 10⁶ 50% egg infectious dose (EID₅₀) per 0.1 ml/bird. All challenge doses were confirmed by retitrating the inocula in ECEs.

Birds.

Mallards were obtained at 1 day of age from a commercial hatchery. Serum samples were collected from 15 ducks per experiment to ascertain that the birds were serologically negative for AI viruses by blocking ELISA (FlockCheck avian influenza MultiS-Screen antibody test; IDEXX Laboratories, Westbrook, ME). At 2 weeks of age, the ducks were housed in self-contained isolation units ventilated under negative pressure with inlet and exhaust HEPA (high-efficiency particulate arrestance)-filtered air and maintained under continuous lighting. Feed and water were provided with ad libitum access. All procedures were performed according to the requirements of protocols approved by the Institutional Animal Care and Use Committee and the Institutional Biosafety Committee.

Experimental design.

Two identical experiments were conducted. In experiment 1, the ducks were inoculated with Np/WA/14 (H5N2), Gf/WA/14 (H5N8), or Ws/Mg/05 (H5N1) HPAI viruses (Table 1). In addition, a H5 North American lineage LPAI virus Ml/MN/00 H5N2 was included for comparison purposes. In experiment 2, seven H7 HPAI viruses and four H5 HPAI viruses were examined, as well as two H7 LPAI viruses for comparison (Table 1). In both experiments, the ducks were separated into a sham-inoculated control group and virus-inoculated groups. Ducks were i.n. inoculated, via choanal cleft, with 10⁶ EID₅₀/0.1 ml of each virus. Sham-inoculated control ducks were i.n. inoculated with 0.1 ml of sterile allantoic fluid diluted 1:300 in BHI. Three naive ducks were added to each group at 1 day postinoculation (dpi) to examine for virus transmission. Body temperatures were taken from all ducks at 2 and 4 dpi, and body weights were determined at 2, 4, 7, and 14 dpi. Two ducks from each group were euthanized at 4 dpi to examine for gross lesions, and portions of the brain, lung, spleen, skeletal muscle, and heart (first experiment) and of the brain, lung, and spleen (second experiment) were collected frozen for subsequent virus detection. In the first study, a full set of tissue samples was also collected for microscopic evaluation as well (nasal cavity, trachea, lungs, air sacs, eye lid, brain, spleen, liver, intestine, pancreas, kidney, adrenal glands, Harderian glands, thymus, bursa, heart, and skeletal muscle). Tissues were fixed in 10% neutral buffered formalin solution, paraffin embedded, sectioned, and stained with hematoxylin and eosin. Duplicate sections were stained by immunohistochemical methods to determine the influenza viral antigen distribution in individual tissues (50). The remaining ducks were observed for clinical signs over a 14-day period during which time clinical signs were recorded. Ducks that showed severe neurological signs, stopped eating or drinking, or remained recumbent were euthanized and were reported as dead on the next day for the calculation of mean death times. Oropharyngeal (OP) and cloacal (CL) swabs were collected at 2, 4, 7, 11, and 14 dpi from directly inoculated birds, and at 1, 3, 6, 10, and 13 days after contact from contact-exposed birds to determine virus shedding. Surviving ducks were bled at 14 dpi for serology and euthanized by the intravenous administration of sodium pentobarbital (100 mg/kg [body weight]).

TABLE 1.

Body temperatures and body weights of mallards inoculated with H5 and H7 LPAI and HPAI viruses

Expt and virus	Avg ± SEMa
	Body temp (oF)		Body wt (kg)
	2 dpi	4 dpi	2 dpi	4 dpi	7 dpi	14 dpi
Expt 1
Controls	107.4 ± 0.3	107.4 ± 0.2	0.34 ± 0.01	0.39 ± 0.01	0.48 ± 0.01	0.64 ± 0.02
Ml/MN/00 (H5N2) LPAIV	107.8 ± 0.2	107.8 ± 0.3	0.34 ± 0.001	0.40 ± 0.01	0.48 ± 0.02	0.65 ± 0.02
Np/WA/14 (H5N2) HPAIV	107.1 ± 0.2	107 ± 0.2	0.29 ± 0.001*	0.32 ± 0.01**	0.40 ± 0.02*	0.57 ± 0.03
Gf/WA/14 (H5N1) HPAIV	108.9 ± 0.2***	107.8 ± 0.5	0.29 ± 0.01*	0.34 ± 0.01*	0.44 ± 0.02	0.64 ± 0.03
Ws/Mongolia/05 (H5N1) HPAIV	109 ± 0.2***	ND	0.27 ± 0.001***	ND	ND	ND
Expt 2
Controls	107.2 ± 0.2	107.2 ± 0.1	0.30 ± 0.01	0.35 ± 0.01	0.44 ± 0.1	0.60 ± 0.02
Ml/OH/87 (H7N8) LPAIV	108.8 ± 0.1	107.7 ± 0.2	0.30 ± 0.01	0.34 ± 0.01	0.43 ± 0.01	0.62 ± 0.01
Ml/Sweden/02 (H7N7) LPAIV	107 ± 0.3	107 ± 0.2	0.28 ± 0.02	0.36 ± 0.02	0.41 ± 0.02	0.60 ± 0.02
Ck/Chile/02 (H7N3) HPAIV	106.9 ± 0.3	107.9 ± 0.2	0.26 ± 0.01	0.29 ± 0.01*	0.42 ± 0.01	0.56 ± 0.01
Ck/Canada/05 (H7N3) HPAIV	107.8 ± 0.1	108 ± 0.3	0.27 ± 0.01	0.33 ± 0.01	0.43 ± 0.02	0.53 ± 0.02
Ck/Jalisco/12 (H7N3) HPAIV	107.6 ± 0.1	107.4 ± 0.2	0.25 ± 0.02	0.30 ± 0.02	0.40 ± 0.02	0.57 ± 0.02
Ck/Victoria/85 (H7N7) HPAIV	106.7 ± 0.3	107.1 ± 0.3	0.26 ± 0.01	0.31 ± 0.01	0.38 ± 0.01	0.49 ± 0.02*
Ck/North Korea/05 (H7N7) HPAIV	107.5 ± 0.1	107 ± 0.3	0.26 ± 0.01	0.32 ± 0.01	0.40 ± 0.01	0.53 ± 0.02
Ck/Netherlands/03 (H7N7) HPAIV	107.3 ± 0.1	107.7 ± 0.2	0.29 ± 0.01	0.33 ± 0.01	0.41 ± 0.02	0.59 ± 0.03
Tk/Italy/99 (H7N1) HPAIV	107 ± 0.2	107.9 ± 0.2	0.29 ± 0.01	0.34 ± 0.001	0.42 ± 0.02	0.59 ± 0.02
Ck/PA/83 (H5N2) HPAIV	106.8 ± 0.1	106.4 ± 0.3	0.27 ± 0.01	0.31 ± 0.02	0.40 ± 0.02	0.55 ± 0.02
Ck/Queretaro/95 (H5N2) HPAIV	107.2 ± 0.2	107.3 ± 0.3	0.25 ± 0.01*	0.30 ± 0.02	0.38 ± 0.02	0.53 ± 0.02
Tk/Ireland/83 (H5N8) HPAIV	106.5 ± 0.2	106.4 ± 0.2	0.27 ± 0.01	0.32 ± 0.001	0.41 ± 0.01	0.58 ± 0.01
Tern/South Africa/61 H5N3 HPAIV	107.4 ± 0.2	106.9 ± 0.2	0.28 ± 0.02	0.33 ± 0.01	0.41 ± 0.02	0.56 ± 0.02

^aSignificant differences in body temperatures or body weights when comparing groups in each experiment are indicated by asterisks (*, P < 0.01; **, P < 0.001; ***, P < 0.0001). ND, not done.

Virus quantifications.

OP and CL swabs were collected in 1 ml of BHI broth with a final concentration of gentamicin (1,000 μg/ml), penicillin G (10,000 U/ml), and amphotericin B (20 IU/ml) and kept frozen at −70°C until processed. Viral RNA was extracted using a MagMAX AI/ND viral RNA isolation kit (Ambion, Austin, TX). Quantitative real-time reverse transcription-PCR (qRT-PCR) for AIV detection was performed as previously described (32). qRT-PCRs targeting the influenza virus M gene (51) were conducted by using a AgPath-ID one-step RT-PCR kit (Ambion) and the ABI 7500 Fast real-time PCR system (Applied Biosystems, Carlsbad, CA). The RT step conditions were 10 min at 45°C and 10 min at 95°C. The cycling conditions were 45 cycles of 15 s at 95°C and 45 s at 60°C. Virus titers in frozen tissue samples were determined by weighing, homogenizing, and diluting tissues in BHI to a 10% (wt/vol) concentration. Viral RNA was extracted using TRIzol LS reagent (Invitrogen, Carlsbad, CA) and a Qiagen RNeasy minikit (Qiagen). Equal amounts of RNA extracted from the tissue samples were used in the qRT-PCR assay (50 ng/μl). For virus quantification, a standard curve was established with RNA extracted from dilutions of the same titrated stock of the challenge virus, and the results are also reported as EID₅₀/ml or EID₅₀/g equivalents. The calculated qRT-PCR lower detection limit for the viruses varied between 10^1.6 and 10^1.9 EID₅₀/ml.

Serology.

Hemagglutination inhibition (HI) assays using homologous antigens were performed to quantify antibody responses to virus infection with serum collected from ducks at 14 dpi as previously described (52). HI titers were reported as reciprocal log₂ titers, with a 3-log₂ (a titer of 1:8) titer or greater considered positive.

Statistical analyses.

One-way analysis of variance, along with Tukey's multiple-comparison tests, was applied to analyze body weights, body temperatures, and titers of virus shedding at 4 dpi using Prism v.5.01 software (GraphPad, La Jolla, CA). A P value of 0.05 was considered significant.

RESULTS

Clinical signs, body temperature, and body weights in mallards infected with HPAI and LPAI viruses.

No clinical signs or mortality was observed in mallards inoculated with any LPAI or HPAI viruses in both experiments, with the exception of those inoculated with H5 Gs/GD HPAI viruses. In experiment 1, ducks inoculated with Ws/Mongolia/05 (H5N1) presented with lethargy, anorexia, and neurological signs, and all of them died by 4 dpi, similar to previous reports with some Gs/GD lineage H5N1 HPAI virus infections in ducks (10). Ducks inoculated with the Gf/WA/14 (H5N8) or Ws/Mongolia/05 (H5N1) HPAI viruses had significantly higher body temperatures at 2 dpi than ducks inoculated with Np/WA/14 (H5N2) HPAI virus or the Ml/MN/00 LPAI virus or the sham-inoculated control ducks (P < 0.0001) (Table 1). Three ducks inoculated with the H5N8 virus still had fever at 4 dpi (Fig. 1). Ducks inoculated with the Np/WA/14 (H5N2), Gf/WA/14 (H5N8), and Ws/Mongolia/05 (H5N1) HPAI viruses had significantly lower body weights than control ducks and ducks inoculated with the LPAI virus at 2 and 4 dpi (Table 1 and Fig. 2) (P < 0.01 to P < 0.0001). Np/WA/14 (H5N2)-infected ducks also had significantly lower weights at 7 dpi (P < 0.01). No differences in body weights were observed between groups at 14 dpi.

FIG 1.

Experiment 1: mean body temperatures of mallards inoculated with LPAI and HPAI viruses at 2 and 4 dpi. Bars represent the standard deviations of the mean. Significant differences in body temperature compared to controls are indicated with asterisks (***, P < 0.0001).

FIG 2.

Experiment 1: mean body weights of mallards inoculated with LPAI and HPAI viruses at 2, 4, 7, and 14 dpi. Bars represent the standard deviations of the mean. Significant differences in body weight compared to controls are indicated with asterisks (*, P < 0.01; **, P < 0.001; ***, P < 0.0001).

In experiment 2, although there were differences in body temperatures at 2 and 4 dpi in ducks inoculated with the HPAI viruses and the LPAI viruses compared to controls, these differences were not significant (Table 1 and Fig. 3). There was also no significant effect of AI virus infection on body weights when examined at 2, 4, 7, and 14 dpi compared to sham-inoculated controls for most of the groups (Table 1 and Fig. 4). Ducks infected with Ck/Queretaro/95 (H5N2), Ck/Chile/02 (H7N3), and Ck/Victoria/85 (H7N7) had lower body weights than control ducks, but at only one time point (2, 4, and 14 dpi, respectively; P < 0.01).

FIG 3.

Experiment 2: mean body temperatures of mallards inoculated with LPAI and HPAI viruses at 2 and 4 dpi. Bars represent the standard deviations of the mean. No significant differences in body temperature among groups were detected.

FIG 4.

Experiment 2: mean body weights of mallards inoculated with LPAI and HPAI viruses at 2, 4, 7, and 14 dpi. Bars represent the standard deviations of the mean. Significant differences in body weight compared to controls are indicated with asterisks (*, P < 0.01).

Lesions induced by HPAIV infections in mallards.

No gross lesions were observed in any of the sham-inoculated control ducks necropsied at 4 dpi in experiments 1 and 2. In experiment 1, no lesions were observed in mallards inoculated with the LPAI virus. Gross lesions were observed in all six ducks inoculated with the H5 HPAI viruses, including mild to moderate dehydration, empty intestines, splenomegaly, and thymus atrophy. Also, nasal discharge, cyanotic bill and toes, dilated and flaccid heart with increased pericardial fluid, renal pallor, and congested brain were observed in ducks inoculated with Ws/Mongolia/05 (H5N1). No gross lesions were observed in any of the ducks necropsied in experiment 2.

Tissues collected from ducks in experiment 1 were examined for microscopic lesions. Ducks inoculated with Ml/MN/00 (H5N2) developed lesions consistent with LPAI virus infection; these were mostly confined to the upper respiratory tract. Mild catarrhal rhinitis and sinusitis, with mucocellular exudates containing sloughed epithelial cells, submucosal edema, and glandular hyperplasia, were observed. The trachea showed mild degenerative changes of the overlying epithelium and mild lymphocytic infiltration in the submucosa and mild edema. Lesions in the gastrointestinal tract consisted of mild proliferation of gut-associated lymphoid tissues. The remaining organs did not show significant histopathologic changes. In contrast, severe and widespread microscopic lesions were observed in tissues from ducks inoculated with Ws/Mongolia/05 (H5N1); these changes were similar to those observed with other Gs/GD lineage H5N1 HPAI viruses (10). The most consistent lesions were moderate to severe rhinitis and sinusitis, moderate tracheitis and bronchitis, mild to moderate interstitial pneumonia, airsacculitis, and moderate multifocal necrosis of cardiac myofibers and, in the brain, randomly scattered foci of malacia with gliosis. Also commonly observed were multifocal pancreatitis, necrosis of the epithelia of the Harderian glands, and moderate multifocal areas of vacuolar degeneration to necrosis of the corticotrophic cells of the adrenal gland. Mild to moderate necrosis of hepatocytes with sinusoidal histiocytosis was found in the liver. The spleen, thymus, bursa, and mucosa-associated lymphoid tissue showed moderate to severe lymphoid depletion. Ducks inoculated with the H5N8 and H5N2 HPAI viruses had moderate rhinitis, sinusitis, tracheitis, bronchitis, mild interstitial pneumonia, and airsacculitis, and mild pancreatitis. Mild necrosis of the epithelia of the Harderian glands was observed in two ducks. The liver of one duck had lymphocytic infiltrations and focal hepatocyte necrosis. Another duck had focal pancreatitis. Mild to moderate lymphoid depletion was observed in all lymphoid organs.

In order to determine sites of virus replication, immunohistochemical staining for AI virus antigen was conducted. In ducks inoculated with the LPAI virus, viral antigen staining was only observed in nasal and trachea epithelial cells (Table 2). Viral antigen staining was present in multiple tissues from ducks infected with the H5 HPAI viruses, indicating systemic infection, but virus staining was more widespread in tissues from ducks inoculated with Ws/Mongolia/05 (H5N1). Viral antigen was present in epithelial cells and macrophages in the nasal turbinates, trachea, lung, air sac, and Harderian and nasal glands, in pancreatic acinar cells, and in resident and infiltrating phagocytes of the thymus and spleen (Fig. 5). In addition, in ducks inoculated with Ws/Mongolia/05 (H5N1), viral antigen was also identified in the autonomic ganglia of the enteric tract, feather epidermal cells, neurons and glial cells of the brain, hepatocytes and Kupffer cells in the liver, fragmented cardiac and skeletal myofibers, and adrenal corticotrophic cells (Fig. 5). In one duck inoculated with Np/WA/14 (H5N2), viral staining was also present in the neuron and ependymal cells in the brain, and in ducks inoculated with Gf/WA/14 (H5N8) we found viral staining in phagocytes in the eyelid submucosa and in hepatocytes.

TABLE 2.

Experiment 1: average distribution of AIV antigen in tissues from ducks infected with H5N8, H5N2, and H5N1 HPAI viruses and with H5N2 LPAI virus

Virus	Detection of AIV antigen in tissuesa
	Nasal epithelium	Trachea	Lung	Air sacs	Eye lid	Brain	Spleen	Liver	Intestine	Pancreas	Kidney	Adrenal gland	Harderian gland	Thymus	Bursa	Heart	Muscle
Ml/MN/00 (H5N2) LPAIV	+/+	+/+	–/–	–/–	–/–	–/–	–/–	–/–	–/–	–/–	–/–	–/–	–/–	–/–	–/–	–/–	–/–
Np/WA/14 (H5N2) HPAIV	++/++	++/+++	++/++	++/++	–/–	++/+	+/+	–/–	–/–	+/+	–/–	–/–	–/+	+/+	–/–	–/–	–/–
Gf/WA/14 (H5N8) HPAIV	++/++	++/++	++/++	+++/++	+/++	–/–	+/+	–/++	–/–	+/+	–/–	–/–	+++/–	+/+	–/–	–/–	–/+
Ws/Mongolia/05 (H5N1) HPAIV	++/+	++/+	+++/++	++/++	–/–	+++/+++	++/++	++/–	+/+	+++/+++	+/–	+++/+++	++/++	++/+	+/+	+++/+++	++/++

^aTissues were examined at 4 dpi (vales are expressed in the form “duck 1/duck 2”). −, no positive cells; +, single positive cells; ++, scattered groups of positive cells; +++, widespread positivity.

FIG 5.

Immunohistochemical staining for AIV antigen in tissues of mallards infected with Gs/GD H5 lineage HPAI viruses. Tissues were collected at 4 dpi. Virus antigen is stained red. Panels A, B, C, D, and E show tissues from ducks infected with Gf/WA/14 (H5N8) HPAIV; panels F, G, and H show tissues from ducks infected with Ws/Mongolia/05 (H5N1) HPAIV. (A) Trachea; (B) air sac; (C) liver; (D) Harderian gland; (E) nasal gland; (F) pancreas; (G) cerebellum; (H) heart. Magnification, ×40. Viral antigen was present in epithelial cells and macrophages in the trachea, air sac, Harderian and nasal glands, in hepatocytes and Kupffer cells in the liver, in pancreatic acinar cells, in neurons and glial cells of the brain, and in cardiac myofibers.

Viral RNA quantification in swabs and tissues.

Oropharyngeal (OP) and cloacal (CL) viral shedding was examined by qRT-PCR, and the results are shown in Table 3 and Fig. 6, 7, 8, and 9. All virus-inoculated ducks became infected, as determined by the detection of viral RNA in swabs, with the exception of three ducks inoculated with Ck/PA/83 (H5N2) HPAI and four ducks inoculated with Ck/Queretaro/95 (H5N2) (Table 3). In experiment 1, mallards inoculated with Ws/Mongolia/05 (H5N1) shed high titers of virus by the OP route at 2 and 4 dpi (10^5.7 to 10^7.8 EID₅₀/ml) (Fig. 6). Mallards inoculated with Np/WA/14 (H5N2) or Gf/WA/14 (H5N8) shed moderate to high titers of virus at 2 and 4 dpi (10^3.8 to 10^6.7 EID₅₀/ml). Mallards inoculated with the Ml/MN/00 (H5N2) LPAI virus shed 10^1.4 to 10^5.1 EID₅₀/ml by the OP route at these same time points, and there was more variability among titers at 2 dpi compared than that observed with the HPAI viruses. In ducks inoculated with the HPAI viruses, CL virus titers were almost 2 log₁₀ lower than what was observed in OP swabs at 2 and 4 dpi. The duration of virus shedding varied among the viruses examined. Most of the ducks inoculated with Gf/WA/14 (H5N8) virus shed low titers, or no virus, by 11 dpi, and most of the mallards inoculated with Np/WA/14 (H5N2) virus or the LPAI virus shed virus up to 14 dpi (Fig. 6). Ducks inoculated with the Ws/Mongolia/05 (H5N1) virus were dead by 4 dpi.

TABLE 3.

Infection, virus shed titers, and seroconversion of ducks inoculated with LPAI and HPAI viruses and contact-exposed ducks^a

Expt and virus	Direct inoculates					Contact-exposed ducks
	No. of infected ducks/total ducks	No. of days virus positive	Oropharyngeal shed titer (mean at 4 dpi)	Cloacal shed titer (mean at 4 dpi)	No. of antibody-positive ducks/total ducks (titer range)	No. of infected ducks/total ducks	Oropharyngeal shed titer (mean at 3 days after contact)	Cloacal shed titer (mean at 3 days after contact)	No. of antibody-positive ducks/total ducks (titer range)
Expt 1
Ml/MN/00 (H5N2) LPAIV	10/10	>14	3.8	3.9	4/8 (8–64)	3/3	4.3	4.0	1/3 (8)
Np/WA/14 (H5N2) HPAIV	10/10	>14	5.8	3.7	8/8 (16–32)	3/3	5.5	4.0	3/3 (16)
Gf/WA/14 (H5N8) HPAIV	10/10	<14	5.7	3.8	7/7 (8–64)	3/3	6.1	3.6	2/3 (16)
Ws/Mongolia/05 (H5N1) HPAIV	10/10	ND	6.6	5.0	ND	3/3	6.3	4.7	ND
Expt 2
Ml/OH/87 (H7N8) LPAIV	10/10	>14	4.8	4.6	8/8 (16–64)	3/3	4.3	4.8	3/3 (16)
Ml/Sweden/02 (H7N7) LPAIV	10/10	>14	4.1	5.3	8/8 (8–128)	3/3	4.0	5.3	3/3 (16–32)
Ck/Chile/02 (H7N3) HPAIV	10/10	>14	4.0	4.1	8/8 (8–64)	3/3	2.8	3.5	3/3 (16–32)
Ck/Canada/05 (H7N3) HPAIV	10/10	>14	4.3	5.1	8/8 (16–256)	3/3	3.8	5.6	3/3 (16–64)
Ck/Jalisco/12 (H7N3) HPAIV	10/10	>14	4.1	5.3	8/8 (16–128)	3/3	4.4	5.0	3/3 (16–64)
Ck/Victoria/85 (H7N7) HPAIV	10/10	<14	4.2	4.0	8/8 (16–64)	3/3	5.3	4.3	3/3 (8–128)
Ck/North Korea/05 (H7N7) HPAIV	10/10	<14	4.5	3.9	8/8 (8–32)	3/3	4.9	3.3	3/3 (16)
Ck/Netherlands/03 (H7N7) HPAIV	10/10	<14	3.7	3.1	8/8 (8–64)	3/3	3.5	3.8	3/3 (32–64)
Tk/Italy/99 (H7N1) HPAIV	10/10	<14	4.4	3.0	8/8 (16–64)	3/3	2.8	2.9	3/3 (16)
Ck/PA/83 (H5N2) HPAIV	7/10	<7	1.8	1.8	6/8 (8)	3/3	1.5	1.7	3/3 (8)
Ck/Queretaro/95 (H5N2) HPAIV	6/10	<7	2.7	1.5	6/8 (8–127)	1/3	–	–	1/3 (8)
Tk/Ireland/83 (H5N8) HPAIV	10/10	<11	4.9	3.1	8/8 (256–624)	3/3	2.9	–	2/3 (8–64)
Tern/South Africa/61 (H5N3) HPAIV	10/10	<14	4.2	3.0	8/8 (16–64)	2/3	–	3.3	1/3 (8)

^aTwo-week-old mallard ducks were i.n. inoculated with 10⁶ EID₅₀ of each virus. Ducks were considered virus positive if RNA was detected at 4 dpi. Virus titers were determined by quantitative real-time RT-PCR and are expressed as the log₁₀ EID₅₀/ml. The mean HI titers (log₂) in ducks are indicated, with the range given in parentheses. The number of birds with positive HI titers is shown (≥ threshold of positivity/total number of sera tested). –, negative; ND, not done.

FIG 6.

Mean viral shedding in mallards inoculated with Ml/MN/00 (H5N2) LPAI, Np/WA/14 (H5N2) HPAI, Gf/WA/14 (H5N8) HPAI, and Ws/Mongolia/05 (H5N1) HPAI viruses. Each data point represents AI virus titers detected in OP and CL swabs at different days after HPAI virus inoculation. Bars represent the standard deviations of the mean. All swabs from which virus was not detected were given a numeric value of 10^1.5 EID₅₀/ml.

In experiment 2, mallards inoculated with the H7 LPAI viruses shed virus by both the OP and CL routes, but shedding occurred for a longer period by the CL route (Fig. 7 and 8). For all H7 viruses, the CL titers at 2 dpi were lower than the OP titers, which is in part explained by the route of virus inoculation, with the viruses initially replicating in the upper respiratory tract and later in the intestine (Fig. 7 and 8). The highest virus titers were detected at 2 and 4 dpi. The pattern of virus shedding was similar between groups of ducks inoculated with Ck/Chile/02 (H7N3) and Ck/Canada/05 (H7N3) HPAI viruses, with most ducks shedding up to 7 dpi by the OP route and until 11 dpi by the CL route (Fig. 7). The Ck/Jalisco/12 (H7N3), Ck/Victoria/85 (H7N7), Ck/North Korea/05 (H7N7), Ck/Netherlands/03 (H7N7), and Tk/Italy/05 (H7N1) HPAI viruses were similarly shed up to 11 dpi by both routes (Fig. 8). Viral RNA was detected in swabs from some individual ducks at 14 dpi in almost all groups.

FIG 7.

Mean viral shedding in mallards inoculated with Ml/OH/87 (H7N8) LPAI, Ml/Sweden/02 (H7N7) HPAI, Ck/Chile/02 (H7N2) HPAI, and Ck/Canada/05 (H7N3) HPAI viruses. Each data point represents AI virus titers detected in OP and CL swabs at days after HPAI virus inoculation. Bars represent the standard deviations of the mean. All swabs from which virus was not detected were given a numeric value between 10^1.5 and 10^1.8 EID₅₀/ml.

FIG 8.

Mean viral shedding in mallards inoculated with Ck/Jalisco/12 (H7N3) HPAI, Ck/Victoris/85 (H7N7) HPAI, Ck/North Korea/05 (H7N7) HPAI, Ck/Netherlands/03 (H7N7) HPAI, and Tk/Italy/99 (H7N1) HPAI viruses. Each data point represents AI virus titers detected in OP and CL swabs at different days after HPAI virus inoculation. Bars represent the standard deviations of the mean. All swabs from which virus was not detected were given a numeric value between 10^1.5 and ^1.8 EID₅₀/ml.

Regarding the H5 HPAI viruses, not all mallards inoculated with Ck/PA/83 (H5N2) or Ck/Queretaro/95 (H5N2) became infected, with only 7 or 6 of 10 mallards shedding the respective viruses (Fig. 9). Although more virus was shed by the OP route than the CL route, the titers were low in both groups, and most ducks only shed until 4 dpi. Mallards inoculated with Tk/Ireland/83 (H5N8) shed mostly by the OP route with some ducks shedding high titers at 4 dpi (Fig. 9). The Tern/South Africa/61 (H5N3) HPAI virus was shed at moderate titers by both routes (Fig. 9).

FIG 9.

Mean viral shedding in mallards inoculated with Ck/PA/83 (H5N2) HPAI, Ck/Queretaro/95 (H5N2) HPAI, Tk/Ireland/83 (H5N8) HPAI, and Tern/South Africa/61 (H5N3) HPAI viruses. Each data point represents AI virus titers detected in OP and CL swabs at different days after HPAI virus inoculation. Bars represent the standard deviations of the mean. All swabs from which virus was not detected were given a numeric value between 10^1.5 and 10^1.8 EID₅₀/ml.

When comparing OP virus titers at 4 dpi in all groups (experiments 1 and 2 combined), significantly higher titers were shed by ducks inoculated with Ws/Mongolia/05 (H5N1), Np/WA/14 (H5N2), and Gf/WA/14 (H5N8) (P < 0.1 to 0.0001) than the rest of the groups, with no significant difference among the three (Table 3). Ducks inoculated with Ck/PA/83 (H5N2) shed significantly smaller amounts of virus than ducks in all other groups, with the exception of Ck/Queretaro/95 (H5N2). Ducks inoculated with Ck/Queretaro/95 (H5N2) also shed less virus than ducks inoculated with Ml/OH/87 (H5N2), Ck/North Korea/05 (H7N7), Ck/Canada/05 (H7N3), Tk/Italy/99 (H7N1), Ck/Victoria/85 (H7N7), and Tern/South Africa/61 (H5N3). With some minor exceptions, there were no significant differences in the titers of virus shed at this time point among the rest of the groups.

The presence of viral RNA was also examined in tissues collected at 4 dpi from virus-inoculated mallards (Table 4). In experiment 1, lung, spleen, brain, heart and muscle were examined; in experiment 2, only the lungs, spleen, and brain were examined. Low viral titers, or titers under the limit of detection, were observed in most tissues from LPAI virus-inoculated groups. In experiment 1, the titers were consistently high in tissues from ducks inoculated with Ws/Mongolia/05 (H5N1). High titers were also observed in some of the tissues from ducks inoculated with the H5 North American Gs/GD lineage HPAI viruses. In general, tissue titers were higher in these two groups than in all other HPAI virus-inoculated groups in experiment 2, with the exception of Ck/Chile/02 (H7N3), Ck/North Korea/05 (H7N7), and Tk/Italy/99 (H7N1), which also had high virus titers in some tissues (Table 4).

TABLE 4.

Comparison of AI virus titers in lungs, spleen, brain, muscle, and heart^a

Expt and virus	Virus titer (log10 EID50/g)
	Lung	Spleen	Brain	Heart	Muscle
Expt 1
Ml/MN/00 (H5N2) LPAIV	–/–	1.3/–	–/1.0	–/–	–/–
Np/WA/14 (H5N2) HPAIV	6.1/6.8	4.7/6.4	4.1/4.2	4.7/4.6	3.7/5.2
Gf/WA/14 (H5N8) HPAIV	5.2/7.2	2.1/4.6	6.1/5.0	3.7/4.3	4.7/5.1
Ws/Mongolia/05 (H5N1) HPAIV	7.6/7.4	6.7/6.8	8.0/8.9	8.5/8.0	ND
Expt 2
Ml/OH/87 (H7N8) LPAIV	3.4/–	3.1/–	–/–	ND	ND
Ml/Sweden/02 (H7N7) LPAIV	–/3.2	–/3.3	2.1/–	ND	ND
Ck/Chile/02 (H7N3) HPAIV	7.3/5.5	5.3/3.4	4.2/3.6	ND	ND
Ck/Canada/05 (H7N3) HPAIV	3.6/4.2	2.7/2.1	–/2.4	ND	ND
Ck/Jalisco/12 (H7N3) HPAIV	3.5/3.7	2.6/4.3	–/5.2	ND	ND
Ck/Victoria/85 (H7N7) HPAIV	2.2/2.1	3.0/1.7	2.1/2.3	ND	ND
Ck/North Korea/05 (H7N7) HPAIV	5.2/5.7	4.6/5.1	2.6/2.8	ND	ND
Ck/Netherlands/03 (H7N7) HPAIV	2.0/2.4	3.0/3.1	2.0/2.0	ND	ND
Tk/Italy/99 (H7N1) HPAIV	6.6/4.9	3.8/4.1	4.0/3.0	ND	ND
Ck/PA/83 (H5N2) HPAIV	1.9/1.7	1.6/1.9	2.1/–	ND	ND
Ck/Queretaro/95 (H5N2) HPAIV	–/–	2.2/1.8	1.8/1.9	ND	ND
Tk/Ireland/83 (H5N8) HPAIV	2.8/–	3.6/1.2	2.9/2.5	ND	ND
Tern/South Africa/61 (H5N3) HPAIV	3.5/2.4	3.8/3.7	2.3/2.5	ND	ND

^aTissues were taken from two ducks per group at 4 dpi (values are expressed in the form “duck 1/duck 2”). –, negative; ND, not done.

Serology.

When examined at 14 dpi, most ducks had detectable antibody titers against the viruses (Table 3). Serology is not available for the group inoculated with the Ws/Mongolia/05 (H5N1) HPAI virus since all of the ducks died.

Transmission to contacts.

To assess virus transmission, naive ducks were added to the experimental groups at 1 dpi. All contact exposure ducks became infected with AIV, as demonstrated by virus shedding and/or seroconversion, with the exception of those in the groups challenged with Ck/Queretaro/95 (H5N2) and Tern/South Africa/61 (H5N3) in which two and one contact ducks, respectively, did not become infected (Table 3).

DISCUSSION

In this study, we compared the pathobiology and transmission of 14 HPAI viruses, including both H5 and H7 viruses in mallards, to assess the potential role of this model waterfowl species in disseminating HPAI viruses. Many studies have examined the pathogenicity of Gs/GD H5N1 HPAI viruses in domestic duck species and captive reared mallards (reviewed in reference 10), but only a limited number of studies have investigated the pathogenicity of other H5 and H7 HPAI viruses in domestic ducks species (38, 40, 43–46, 53, 54) and in nondomestic ducks (17, 18, 47, 48, 55, 56). In most of these studies, virus infection, shedding, and transmission were not thoroughly examined. With the exception of ducks inoculated with A/fowl/Germany/34-Rostock H7N1 (38), clinical signs or mortality were not observed with domestic ducks. However, neurological signs and deaths were reported in Muscovy ducks following infections with a H7N1 HPAI virus during the 1999-2000 outbreaks in Italy (41), and H5N2 HPAI viruses were detected in wild waterfowl in Nigeria (56). All of the HPAI viruses examined in our study infected the mallards and transmitted to direct contacts. Two viruses, Ck/PA/83 (H5N2) and Ck/Queretaro/95 (H5N2), failed to infect all ducks after direct challenge, and two viruses, CK/Queretaro/95 and Tern/South Africa/61 (H5N3), failed to infect all of the contact ducks. Clinical signs were only observed in ducks challenged with Gs/GD lineage H5 HPAI viruses; more specifically, the Ws/Mongolia/05 (H5N1) virus killed all mallards, the Np/WA/14 (H5N2) and Gf/WA/14 (H5N8) viruses affected weight, and the Gf/WA/14 virus induced fever. The high mortality observed in the mallards infected with Ws/Mongolia/05 (H5N1) has also been reported with many other Gs/GD lineage H5N1 HPAI viruses in both domestic and nondomestic ducks, but not all viruses from this lineage cause mortality (10). Although the Ws/Mongolia/05 (H5N1) virus belongs to the clade 2.2 H5N1 viruses that spread from Asia into Europe in 2005, it is not representative of all viruses from this clade. Viruses from clade 2.2 were reported to cause mortality in wild waterfowl (6, 7, 57), but differences in pathogenicity have been described between viruses examined experimentally in domestic ducks (53, 57–61), indicating that ducks could become infected with these viruses and not necessarily show clinical signs but still shed large amounts of virus and transmit virus efficiently.

In ducks, the naturally occurring endemic LPAI viruses are typically enterotropic and are shed primarily through feces (62–64). When wild waterfowl LPAI viruses jump to and become adapted to gallinaceous species, the virus typically changes to become more respiratory-tropic with smaller amounts of detectable virus in feces. When these gallinaceous adapted viruses infect ducks, the virus usually retains the respiratory-tropic replication pattern (65, 66). In our study, the LPAI viruses examined were shed for longer by the CL route than the OP route, although moderate replication in the upper respiratory tract still occurred. As for the HPAI viruses, different patterns of virus shedding were observed. The Ck/PA/83 (H5N2), Ck/Queretaro/95 (H5N2), and TK/Ireland/83 (H5N8) viruses were shed at low titers and for fewer than 7 days, with minimal CL shedding, indicating adaptation to gallinaceous species. Different than the other non-Gs/GD H5 viruses, Tern/South Africa/61 (H5N3) did shed by the cloacal route. This virus caused a die-off in common terns in South Africa in 1961 and was never identified in gallinaceous species, and it therefore remained adapted to wild bird species (5). As for the nine H7 HPAI viruses, surprisingly, most viruses were shed at similar titers by the OP and CL route, with CL shedding lagging 2 days, most likely because the i.n. route of inoculation was used to infect the ducks. Viral RNA was detected in OP and CL swabs for at least 11 dpi and all viruses transmitted to contacts. Based on the pattern of virus shed, most of these H7 HPAI viruses did not seem particularly gallinaceous adapted, which could be explained if these viruses were isolated early in the outbreaks in poultry or if the outbreaks were short-lived. On the other hand, the pattern of virus shedding of the Gs/GD H5 HPAI viruses was what would be expected for viruses well adapted to gallinaceous species with higher titers in OP swabs than CL swabs. These viruses were shed at high titers at 2 and 4 dpi, with the highest titers observed with Ws/Mongolia/05 (H5N1), which explains the mortality seen in mallards infected with this virus. The Gs/GD H5 HPAI lineage is unique because the virus has been endemic for 20 years in gallinaceous and domestic duck species. This complex interplay in multiple species appears to result in the selection of viruses with characteristics of both gallinaceous and duck-adapted viruses. Viral titers in tissues were also the highest for these three viruses compared to the other HPAI viruses used in the study, with the exception of Ck/Chile/02 H7N3, Ck/North Korea/05 H7N7, and Tk/Italy/99 H7N1, which also replicated to high titers in some tissues.

The Gs/GD H5N1 HPAI viruses developed the unique capacity among HPAI viruses to replicate well and cause disease in domestic and wild ducks, producing a range of syndromes from asymptomatic respiratory and digestive tract infections to severe systemic infection and death, as seen in gallinaceous poultry (10). Despite efforts to control the spread of H5N1 HPAI viruses, these viruses continue to evolve, which has led to the emergence of multiple genotypes or sublineages (67). The endemic status of H5N1 HPAI viruses in Asia and Africa has also led to the generation of reassortant H5 strains with novel gene constellations. Recently, new subtypes of H5 HPAI viruses (H5N2, H5N5, H5N6, and H5N8) with the genetic backbone of clade 2.3.4 viruses have been detected in wild birds, ducks, geese, quail, and chickens (23, 29, 32, 67–69). In November and December 2014, HPAI viruses of the H5 subtype originating from China were detected in wild birds and poultry in various countries of Asia and Europe and, for the first time, in North America, where the viruses rapidly spread to wild waterfowl, wild and captive birds of prey, and backyard and commercial poultry (22, 30, 34, 70). The involvement of wild birds in both Europe and North America suggested that these H5 HPAI viruses were better adapted to waterfowl than historic HPAI viruses. This potential adaptation may relate to decreased pathogenicity, increased viral shedding, changes in shedding patterns and duration, or a decrease in infectious dose.

Other experimental studies using clade 2.3.4.4 H5N8 HPAI viruses showed that they replicated systemically and were lethal in chickens but appeared to be attenuated, although efficiently transmitted, in ducks (71). A range of outcomes of infection—from no clinical signs to severe disease—were observed in ducks i.n. inoculated with H5N8 viruses, and the mortality rates varied from 0 to 20% (27, 54, 69, 72, 73). Viral shedding and replication in tissues was high, and virus was shed for more than 5 days (72). Natural infection of domestic ducks with H5N8 has been associated with drops in egg production and a mild increase in mortality (74), and H5N8 has been detected from carcasses of wild birds, including mallards (23, 32, 70). However, with these natural infections, other factors could have affected disease presentation, including coinfection with other pathogens. In our study, although the ducks inoculated with the Gf/WA/14 (H5N8) virus were febrile and the virus replicated well in many tissues, no mortality was observed. Ducks challenged with the reassortant H5N2 virus (Np/WA/14) were not febrile and shed virus longer, which would favor dissemination and transmission of this virus. This virus exhibited the widest geographic spread in the United States. The H5N2 HPAI virus is composed of five gene segments (PB2, PA, HA, M, and NS) related to the Eurasian HPAI H5N8, and the remaining gene segments (PB1, NP, and NA) are related to North American lineage waterfowl viruses (32, 34, 35). Interestingly, clinical signs and mortality have been described in domestic ducks infected with Gs/GD lineage H5 clade 2.3.4.4 HPAI viruses (75). Experimental infection of juvenile Muscovy ducks (Cairina moschata) with another reassortant virus belonging to clade 2.3.4.4, A/chicken/BC/FAV-002/2015 (H5N1), caused neurological signs and death and transmitted to naive contact ducks (75). Muscovy ducks are more susceptible to HPAI virus infection than other domestic ducks, which explains the differences observed when comparing these results to those obtained in mallards infected with the H5 clade 2.3.4.4 viruses in our study (76). The PB1, PA, NA, and NS gene segments of this H5N1 virus were of North American lineage, whereas PB2, HA, NP, and M were derived from the Eurasian lineage H5N8 virus. The role of the genetic makeup of these viruses in the differences observed in pathogenicity remains to be studied.

The ability of these novel reassortant H5 viruses to replicate efficiently without killing the infected ducks allows them to circulate within the duck population and increases the possibility of transmission to poultry. In nature, waterfowl are exposed to many AI viruses, so some degree of homosubtypic and heterosubtypic immunity is expected (77). This immunity could afford the ducks some protection; however, if the AI virus is highly infectious, it might still replicate in the ducks and cause ameliorated clinical signs, and normal bird behavior (e.g., migration) may not be impaired. Based on the viral shedding and transmission results, coupled with the lack of clinical signs, many of the H7 HPAI viruses should be equally fit to transmit efficiently among ducks, yet detection of infection in domestic or wild ducks is rare, restricted to the detection of an H7N1 HPAI virus in domestic ducks during the 1999-2000 outbreaks in Italy. This suggests that the lack of spillover of HPAI viruses from poultry back into wild ducks is not from biological incompatibility but from insufficient exposure to initiate and sustain infection. This underlines the importance of quickly containing HPAI outbreaks in poultry and the separation of domestic and wild birds.

In summary, in this study, 12 of 14 H5 and H7 HPAI viruses infected all inoculated and direct-contact mallards, and virus shed was detected for at least 7 days. The highest titers of virus shed were detected in ducks infected with the Gs/GD lineage H5 HPAI viruses, but in contrast to shedding patterns observed with LPAI waterfowl viruses, higher titers were detected in OP swabs than in CL swabs. The comparison of mallards in a standard challenge model is critical to improve our understanding of virus infection in ducks, but the information should be carefully extrapolated to all duck species. Previous studies have shown differences in clinical disease, shedding, and mortality depending on the duck species tested (17). In addition, for reasons we do not understand, the presence of multiple basic amino acids or insertion of amino acids at the HA cleavage site is predictive of severe disease in gallinaceous species, but it is not predictive of severe disease in ducks. The high virus titers and the minimal clinical signs observed with the clade 2.3.4.4 H5 viruses likely represent factors that facilitate transmission on of these viruses in susceptible wild duck populations. However, transmission and maintenance of avian influenza in wild bird reservoirs is complex, and it remains to be seen whether these viruses are fit enough to persist in these populations.

Funding Statement

This research was supported by USDA/ARS CRIS project 6612-32000-063-00D and by the Center for Research on Influenza Pathogenesis (CRIP) and the NIAID-funded Center of Excellence in Influenza Research and Surveillance (CEIRS; contract HHSN272201400008C).