PRE-EXPOSING CANADA GEESE (BRANTA CANADENSIS) TO A LOW-PATHOGENIC H1N1 AVIAN INFLUENZA VIRUS PROTECTS THEM AGAINST H5N1 HPAI VIRUS CHALLENGE

Wild aquatic birds of the orders Anseriformes (ducks, geese, and swans) and Charadriiformes (gulls and shorebirds) are considered to be the primary reservoir hosts of influenza A viruses in nature (Webster et al. 1992). The known 16 hemagglutinin (HA) and nine neuraminidase (NA) subtypes exist subclinically in these natural hosts and are excreted into the environment, making infection with multiple subtypes of low-pathogenic avian influenza (LPAI) viruses a common occurrence (Alexander 2007). In contrast, outbreaks of highly pathogenic avian influenza (HPAI) in wild birds have been rarely documented. The first reported HPAI outbreak in wild birds involved European common terns (Sterna hirundo) in the Cape of Good Hope area of South Africa in 1961 (Artois et al. 2009). Beginning in 2001, H5N1 HPAI outbreaks, which began in Southeast Asia and subsequently expanded throughout Asia and into Europe and parts of Africa, have also caused mortality events in various wild bird species (Suarez 2010). Some of the long-range spread of H5N1 HPAI virus from Asia to Europe and Africa may have involved migratory waterbirds. As an example, the large outbreak that involved wild birds on Lake Qinghai in May 2005 (Liu et al. 2005; Chen et al. 2006) preceded the rapid spread of the virus to northwestern Asia, Europe, the Middle East, and eventually Africa. Eurasian H5N1 HPAI virus has not been detected in the Americas.

Protective immunity against HPAI virus in birds is provided primarily by subtype-specific antibodies to the viral HA and partially by antibodies to the NA (Swayne and Kapczynski 2008). Although birds also mount antibody responses to other viral proteins such as the matrix and nucleoprotein (NP), these antibodies have not been associated with protection.

The role of heterosubtypic immunity in protecting against H5N1 HPAI virus infection has been studied in various animal models (Heinen et al. 2001; Grebe et al. 2008; Kreijtz et al. 2009; Laurie et al. 2010; Bodewes et al. 2011). Although protection against avian influenza infection is predominantly subtype specific, cross-protective immunity induced by other subtypes has been reported. For example, chickens (Gallus gallus domesticus) that had been previously infected with LPAI A/chicken/HK/G9/97 (H9N2) virus were protected against disease caused by H5N1 HPAI virus. However, protection was not associated with antibody because sera collected from H9N2-infected chickens showed no cross-neutralizing activity against H5N1 (Seo and Webster 2001).

We demonstrated variable levels of susceptibility of Canada Geese (Branta canadensis) to a clade 2.3.2 Eurasian H5N1 HPAI virus (Pasick et al. 2007). This susceptibility was dependent on age (juveniles were more susceptible than adults) and immunologic status. Adult birds that survived an H5N1 challenge had preexisting influenza A virus NP antibodies but were negative for H5 antibodies. Unfortunately, we were unable to determine which HA subtype(s) the birds had prior exposure to. Naïve adult and juvenile Canada Geese that had been pre-exposed to a North American–lineage H5N2 LPAI virus were equally resistant to a lethal H5N1 HPAI challenge. In a follow-up study, we pre-exposed Canada Geese to heterosubtypic LPAI viruses that are commonly isolated from wild waterfowl and challenged them with a clade 2.3.2 H5N1 HPAI virus (Berhane et al. 2010). Canada Geese that were preinfected with a North American lineage H4N6 virus were not protected and died 3–5 days postchallenge. In contrast, some of the birds that were pre-exposed to a North American–lineage H3N8 virus survived H5N1 HPAI virus challenge despite developing clinical signs and shedding virus.

To gain further insight into the potential role that prior exposure of heterosubtypic LPAI viruses might play in the pathogenicity and outcome of H5N1 HPAI virus infection, we pre-exposed Canada Geese to three North American wild-bird–origin LPAI viruses, focusing on those having a higher-order evolutionary relatedness to the H5 HA (Dugan et al. 2008). From an evolutionary standpoint, the H3 and H4 HA subtypes used in our previous study are highly divergent from the H5 subtype. We selected H1, H2, and H6 subtype viruses for this study because they are within the same higher-order clustering on the HA phylogenetic tree and are commonly isolated from waterfowl. We also assessed the role that homologous but antigenically divergent antibodies to the NA subtype 1 glycoprotein (N1) play in protecting against lethal H5N1 challenge.

Low-pathogenic avian influenza viruses A/mallard/Alberta/223/2005 (H1N1), A/mallard/British Columbia/118/2006 (H2N3) and A/American black duck/Quebec/135/2005 (H6N5) were propagated in specific-pathogen–free, 9-day-old embryonating chicken eggs. We determined 50% egg infectious dose (EID₅₀) as described by Pasick et al. (2007). H5N1 HPAI A/chicken/Vietnam/14/2005 (Vietnam/14/05), a clade 2.3.2 virus (GenBank accession EF535027) was propagated and titered in Madin-Darby canine kidney (MDCK) cells and expressed as plaque-forming units (PFU) per milliliter. All viruses used in this study were passage 2.

Twenty 6- to 7-wk-old juvenile Canada Geese were captured at a single location within the city of Winnipeg with the permission of Environment Canada (Canadian Wildlife Service permit 09-MB-SC02). They were housed in biosafety level 3+ animal cubicles at the National Centre for Foreign Animal Disease and cared for in accordance with Canadian Council on Animal Care guidelines and the animal use protocol approved by the Institutional Animal Care Committee. All animals appeared in good health based on a cursory physical examination. The geese were provided ad libitum water and fed duck and goose starter ration (Feedrite, Winnipeg, Manitoba, Canada), crushed alfalfa, and dried corn. The birds were floor housed and had access to a wading pool that held ∼350 L of cold municipal tap water when full. The water in the wading pools was changed daily. Birds were exposed to a 12-hr light∶dark period and were provided with sheltered areas where they could hide and an auxiliary heat source. All birds appeared normal during the 18-day acclimatization period.

After the first week of acclimatization, the birds were leg banded and randomly split into the following groups: H1N1, H2N3, and H6N5 pre-exposure groups; an H5N1 challenge control group; and a contact control group (Table 1). Each group was housed in separate high-efficiency particulate air–filtered animal cubicles in which entry and exit procedures were put in place to prevent cross-contamination. Cloacal swab and blood samples collected from all birds on day 14 of acclimatization were negative for influenza A virus nucleic acid and influenza A virus NP antibodies respectively. After acclimatization, Canada Geese in the pre-exposure groups were inoculated with 10⁶ EID₅₀ of the corresponding LPAI virus using a combination of oral, nasal, ocular, and cloacal routes. Back titrations of all inocula were carried out in 9-day-old embryonated eggs to confirm that they were in the desired range. Control groups were mock inoculated with phosphate-buffered saline (PBS) at that time. Cloacal swabs were collected at 3 and 7 days postexposure (dpe) to assess shedding and blood collected at 7, 14, and 21 dpe to check for seroconversion. Twenty-one days after the first inoculation, the birds were reinoculated with the same dose of virus, and blood was collected 10 days afterwards.

Canada Geese in the three pre-exposure and challenge control groups were inoculated with 1.7×10⁵ PFU/bird of H5N1 HPAI Vietnam/14/05 by instilling 0.1 mL into the nares. At 2 days postinoculation (dpi), two birds from the noninfected contact control group were placed among birds in each of the three LPAI pre-exposure groups. Oropharyngeal and cloacal swabs were collected at 2, 3, 4, and 5 dpi to assess viral shedding. Serum was collected prior to necropsy and at 1 and 2 wk postinoculation to assess seroconversion. To assess environmental contamination, water samples were also collected separately from the top and bottom parts (including fecal and food debris) of the wading pools at 3, 4, 5, 6, and 7 dpi using sterile 50-mL centrifuge tubes.

Serum samples collected at 14 days into the acclimatization and samples collected at 7, 14 and 21 days post–LPAI virus exposure were taken from all groups during the pre-H5N1 challenge portion of the trial and tested for influenza A–specific NP antibodies using a competitive enzyme-linked immunosorbent assay (Yang et al. 2008). These sera were also tested for anti-HA antibodies by hemagglutination-inhibition (HI) test using 4 HA units of homologous virus antigen and 0.5% (v/v) suspensions of chicken erythrocytes.

To test the ability of H1N1, H2N3, and H6N5 antisera to inhibit replication of Vietnam/14/05, plaque reduction neutralization assays were carried out on heat-inactivated sera. In one set of experiments, twofold serial dilutions of serum were incubated with 100 PFU of Vietnam/14/05 for 2 hr at 37 C. The virus-serum mixture was adsorbed to 95–100% confluent overnight cultures of MDCK cells for 1 hr. The inoculum was removed, the cells washed with PBS, and a 1.5% carboxymethylcellulose (CMC; Sigma, Oakville, Ontario, Canada) overlay applied. After 4 days of incubation, the cells were fixed with 10% formalin in PBS and permeabilized with 10% acetone in PBS. Viral plaques were immunostained by sequential incubation with anti-influenza NP monoclonal antibody F28 (Yang et al. 2008), horseradish peroxidase–conjugated goat anti-mouse secondary antibody (Jackson Immunoresearch, Inc., West Grove, Pennsylvania, USA), and 3-amino-9-ethyl-carbazole. Wells that showed ≥90% plaque reduction were considered positive for neutralizing antibodies against H5N1 HPAI virus. In a second set of experiments, the serially diluted sera were included in the CMC overlay in addition to incubating with 100 PFU of Vietnam/14/05 for 2 hr prior to adsorption to MDCK cells.

A semiquantitative real-time reverse transcriptase –PCR assay targeting the M1 gene of influenza A virus segment 7 was carried out as previously described (Spackman et al. 2002). The assay was performed on total RNA extracted from 0.5 mL of 10% (w/v) tissue homogenates or 0.5 mL of clarified swab specimens using the MagMAX™-96 Total RNA Isolation Kit (Applied Biosystems/Ambion, Austin, Texas). Full-length, in vitro–transcribed segment 7 RNA, serially diluted in RNase-free water, was run with each assay to give a semiquantitative estimate of viral load in each tissue.

Moribund geese that did not move, eat, or drink were euthanized by exsanguination under ketamine anesthesia (10 mg/kg intramuscularly). In addition, prescheduled necropsies were performed on birds from each group regardless of clinical presentation. Tissue samples (brain, cervical spinal cord, lung, liver, spleen, cecal tonsil, esophagus, trachea, muscle, kidney, heart, nasal turbinate, pancreas, duodenum, ileum, proventriculus, bursa of Fabricius, and thymus) were collected from each bird and fixed in 10% neutral-buffered formalin for a minimum of 48 hr, embedded in paraffin, sectioned, stained with hematoxylin and eosin, and evaluated for microscopic lesions. Lesions were assessed semiquantitatively as mild, moderate, or severe.

Paraffin-embedded tissue sections were quenched for 10 min in aqueous 3% H₂O₂ and pretreated with proteinase K for 15 min. Mouse monoclonal antibody specific for influenza A NP (F26NP9; Yang et al. 2008) was used at a 1∶10,000 dilution for 1 hr. Influenza antigen in tissue sections was visualized using a horseradish peroxidase–labeled polymer, Envision®+ system (anti-mouse) (Dako, Carpinteria, California, USA), reacted with the chromagen diaminobenzidine. The sections were counter stained with Gill's hematoxylin.

Canada Geese in all three LPAI virus pre-exposure groups exhibited no obvious clinical signs or change in behavior following inoculation with H1N1, H2N3, or H6N5 viruses. Cloacal virus shedding was detected in all four birds in the H6N5 group at 7 dpe; however, viral RNA was not detected in cloacal swab samples in either the H1N1 or H3N2 group at 3 or 7 dpe.

Prior to LPAI virus inoculation, all birds were negative for influenza A virus NP antibodies. Following LPAI virus inoculation, two of four birds in the H6N5 group and one of four in the H1N1 group were positive for influenza A virus NP antibodies at 7 dpe. However, none of the birds from any of the groups had detectable anti-HA antibodies at 14 dpe. By 21 dpe, all birds in the H6N5 group had detectable anti-HA antibodies, but none of the birds in the H1N1 and H2N3 groups were positive. Based on these results, birds in all three groups were reinoculated to generate a better immune response. Ten days following reinoculation, all birds were positive for NP antibodies except for one goose in the H2N3 group. Four of four in the H1N1, two of four in the H2N3, and four of four in the H6N5 group also were positive for antibodies to homologous HA antigen. None of the challenge control or contact control birds had detectable anti-HA antibodies prior to H5N1 challenge and none of the sera collected from birds in the H1N1, H2N3, and H6N5 groups inhibited hemagglutination by Vietnam/14/05 (data not shown). Serologic responses to primary and secondary LPAI virus exposure are summarized in Table 2.

The second day after birds in the LPAI pre-exposure and challenge control groups were inoculated with 1.7×10⁵ PFU of Vietnam/14/05, two contact birds were allowed to mingle with animals in each LPAI pre-exposure group to assess virus transmission. Prescheduled necropsies were performed 3 and 7 dpi as well as on any moribund birds that were euthanized.

At 5 dpi, birds in the H1N1 group were mildly depressed and had green feces. Feed consumption was moderate. Only one bird showed mild neurologic signs consisting of ataxia. These clinical signs lasted for 3 days, after which the birds completely recovered. None of the contact birds showed any signs of disease.

Birds in the H2N3 group began showing clinical signs (depression, ataxia, green feces, and decreased intake of food) at 3 dpi. One goose died at 4 dpi. Contact birds began exhibiting clinical signs, including neurologic disease, on the 4th day of contact and died the next day. Only one bird that was pre-exposed to H2N3 survived the H5N1 challenge. This bird had the highest pre-H5N1 HI titer (128) in the group (Table 2).

Birds in the H6N5 group developed clinical signs as severe as control birds beginning at 3 dpi. One bird was found dead and another euthanized because of severe clinical signs at 4 dpi. Contact control birds began showing clinical signs 2 days after contact. The remaining birds died or were euthanized at 3 days postcontact.

Both challenge control birds began showing clinical signs (depression, ataxia, neurologic signs, green feces, decreased intake of food) at 2 dpi and were euthanized at 5 dpi when moribund.

Only one of the four birds in the H1N1 group shed virus. Virus shedding was also detected in one of the contact control birds at 4 days postcontact. By contrast, H5N1 viral nucleic acid was detected in oropharyngeal and cloacal swab samples of all H2N3 and H6N5 birds and in their respective contacts (Table 3).

As a means of quantifying environmental H5N1 virus shedding, total RNA was extracted from top and bottom parts of the wading pool that was provided for environmental enrichment. Low levels of viral nucleic acid were detected at 3 and 4 dpi in the H1N1 group but not beyond 4 dpi. The inability to detect virus beyond 4 dpi in this group may have been due to the cleaning of the pool every second day. Viral nucleic acid was detected up to 7 dpi from top and bottom parts of the pool in both H2N3 and H6N5 groups, with slightly greater amounts in bottom versus top samples. Because only two birds were in the challenge control group, environmental sampling was not done. Results are summarized in Table 3.

Low levels of viral nucleic acid were detected in the central nervous system, heart, and kidneys of the goose in the H1N1 group that underwent prescheduled necropsy at 3 dpi. By comparison, no viral nucleic acid was detected in any of the tissues of the second bird that underwent prescheduled necropsy at 7 dpi. In H2N3 and H6N5 pre-exposed birds, the virus load was highest in the central nervous system. Virus loads reached maximal levels in tissues by 5 dpi in both contact and pre-exposure animals (Table 4).

Histopathologic lesions characterized by severe meningoencephalitis, encephalomyelitis, pneumonitis, focal pancreatic necrosis, and bursal epithelial cell necrosis were observed in challenge control birds euthanized at 5 dpi. There was positive immunostaining for influenza NP in neurons, glial cells, and ependymal cells in the brain. Positive staining was also present in spinal canal ependymal cells and in neurons of peripheral spinal ganglia. Airway epithelial cells in the lungs, liver hepatocytes, heart myocytes, pancreatic acinar and islet cells, smooth muscle and myenteric plexus ganglion cells in the duodenum and ileum, and epithelial cells of the bursa all showed positive staining.

Histologically, lesions were present in a wide range of organs, with the most severe being in the nervous system, spinal cord, nasal turbinate, and pancreas. Positive immunostaining was found in almost every organ tested, with the highest concentrations in the brain, spinal cord, lung, heart, nasal turbinate, pancreas, and thymus. Lesions and immunostaining in the two contact birds were equivalent to those seen in the animals that were directly challenged.

In this group of birds there was moderate to severe meningoencephalitis, spinal myelitis, and mild focal pancreatic necrosis. Positive immunostaining was present in the brain, spinal cord, lung, cecal tonsil, nasal turbinate, pancreas, duodenum, and thymus, but at a lower level than in the H6N5 group. The same magnitude of lesions was observed in the two contact animals at 5 days postcontact, as was observed in the three animals that were directly challenged.

The birds that underwent scheduled necropsies at 3 and 7 dpi had very few histologic lesions, some of which may not have been H5N1 HPAI specific. Only mild positive immunostaining was observed in the brain and liver of one animal (Table 5; Fig. 1).

To help explain why birds in the H1N1 pre-exposure group survived H5N1 HPAI virus challenge, plaque reduction neutralization assays against Vietnam/14/05 were performed using sera collected from all LPAI exposure groups after the final boost as well as sera collected from Canada Geese that survived H5N1 challenge at 13 dpi. None of the pre– or post–LPAI exposure sera resulted in a reduction of plaque size or number. By comparison, sera collected 13 days post–H5N1 inoculation possessed virus neutralizing activity up to a dilution of 1∶200.

Because all four birds in the H1N1 group were positive for N1 antibodies prior to H5N1 challenge, with titers that ranged from 1∶32 to 1∶128, we asked whether anti-N1 antibodies may have had a role in modulating the response of this group to H5N1 HPAI virus challenge. A second experiment in which serially diluted sera were mixed into the CMC overlay resulted in differences in plaque size but not number. The plaques produced following incubation of Vietnam/14/05 with sera that were collected at the beginning of the experiment (prior to LPAI virus exposure) and sera collected post–H1N1, H2N3, and H6N5 exposure but absent from the CMC overlay were the largest: 523±18 µm mean diameter. The plaques produced following incubation of Vietnam/14/05 with H2N3 and H6N5 postexposure sera that were also present in the CMC overlay were similar in diameter: 305±10 to 333±9 µm respectively. The largest reduction in plaque size (45±10 µm average) was observed when H1N1 postexposure serum was present in the CMC overlay (Fig. 2).

Although work by Harris et al. (2010) indicated that Canada Geese may not significantly contribute to avian influenza virus epidemiology in the wild, this species may nevertheless serve as a useful sentinel for H5N1 HPAI introduction in North America. In a previous study (Berhane et al. 2010) we asked whether heterosubtypic immunity to H3N8 and H4N6 avian influenza viruses, two subtypes that are commonly isolated from wild birds, could protect Canada Geese against H5N1 HPAI. Some protection was observed in H3N8 pre-exposed but not H4N6 pre-exposed birds. Here we expand on that study and ask whether pre-existing immunity to H1, H2, and H6 HA subtype viruses, which possess higher-order evolutionary relatedness to H5 HA, can protect Canada Geese against an HPAI H5N1 challenge. Although prior exposure to H2N3 or H6N5 viruses did not provide protection, a North American lineage LPAI H1N1 virus did.

The principal protective immune response against HPAI viruses has been attributed to subtype-specific HA antibodies, which are thought to act by blocking the attachment of virus to host cell sialic acid receptors and by inhibiting fusion of the viral envelope with host cell membranes. Studies have also shown that anti-NA antibodies can confer cross-protective immunity by inhibiting virus release from infected cells (Chen et al. 2000; Sandbulte et al. 2007; Bragstad et al. 2011; Marcelin et al. 2011).

Although we cannot provide definitive answers regarding the importance of anti-N1 antibodies elicited by A/mallard/Alberta/223/2005 (H1N1) in providing protection against the antigenically distinct Eurasian origin H5N1 HPAI virus, other studies have shown that NA-specific antibodies could afford limited protection through “permissive immunity” by blocking the release of infectious virions from the apical surface of infected cells, thus decreasing virus spread to uninfected cells (Johansson et al. 1989; Johansson and Kilbourne 1993). In this study we hypothesize that H5N1 replication may have been held below the pathogenic threshold by inhibiting the release of progeny virus from infected cells or by the aggregation of virus particles at the apical cell surface thereby limiting both virus spread within the host and shedding into the environment.

The presence of H1N1 postexposure serum in the CMC overlay reduced H5N1 virus plaque size, whereas no significant difference in plaque size and number was observed when H1N1, H2N3, or H6N5 postexposure sera were not included in the CMC overlay. Our in vitro assay results are consistent with other studies showing that anti-NA antibodies affect virus titer and plaque size (Kilbourne et al. 1968) and is consistent with the role that anti-NA antibodies are thought to play in inhibiting release of virus from the cell surface, thus reducing virus spread from infected to noninfected cells.

In our previous study (Pasick et al. 2007) we demonstrated that pre-exposing Canada Geese to a genetically divergent North American origin H5N2 LPAI virus protected them against a H5N1 HPAI virus challenge. The role that an immune response against a genetically divergent homosubtypic N1 plays in affording protection against H5N1 HPAI was not known. Even though Vietnam/14/05 possesses a 20–amino acid deletion in the stalk of its N1 and has only 87% amino acid identity with the N1 of A/mallard/Alberta/223/2005, the H1N1 birds were protected from lethal challenge and did not transmit the infection to naïve contacts. This is in contrast to results for wood ducks (Aix sponsa) that were pre-exposed to H1N1 LPAI A/blue-winged teal/LA/B228/1986 and then challenged with H5N1 HPAI A/whooper swan/Mongolia/244/2005 where a moderate level of mortality was observed (Costa et al. 2011).

Although previous studies conducted in chickens showed that vaccination with a virus-vectored vaccine that expressed a N2 protein having 98% amino acid identity with the challenge virus resulted in 88% protection, vaccination with a genetically distant N2 (85% amino acid identity) provided poor protection (Sylte et al. 2007). Chen et al. (2000) demonstrated that DNA encoding an influenza A virus NA used as a vaccine component can provide effective protection not only against infection to genetically homologous viruses but also against viruses that have undergone significant antigenic drift. Taking the potentially cross-protective properties of anti-NA antibodies into consideration, preexisting immunity to the homologous N1 subtype may modulate the survival and transmission patterns of H5N1 HPAI in infected wild birds.

This project was funded by the Canadian Food Inspection Agency and partially by the Poultry Industry Council. We thank Soren Alexandersen for critical review of the manuscript. We gratefully acknowledge Tamiko Hisanaga, Katherine Handel, Colleen Cottam-Birt, James Neufeld, Estella Moffat, Shelly, Ganske, Brad Collignon, Lynn Burton, Kevin Tierney, Kory Nakamura, Maggie Forbes, and Jaime Bernstein for excellent technical assistance.