Abstract
Avian influenza has emerged as one of the most ubiquitous viruses within our biosphere. Wild aquatic birds are believed to be the primary reservoir of all influenza viruses; however, the spillover of H5N1 highly pathogenic avian influenza (HPAI) and the recent swine-origin pandemic H1N1 viruses have sparked increased interest in identifying and understanding which and how many species can be infected. Moreover, novel influenza virus sequences were recently isolated from New World bats. Crocodilians have a slow rate of molecular evolution and are the sister group to birds; thus they are a logical reptilian group to explore susceptibility to influenza virus infection and they provide a link between birds and mammals. A primary American alligator (Alligator mississippiensis) cell line, and embryos, were infected with four, low pathogenic avian influenza (LPAI) strains to assess susceptibility to infection. Embryonated alligator eggs supported virus replication, as evidenced by the influenza virus M gene and infectious virus detected in allantoic fluid and by virus antigen staining in embryo tissues. Primary alligator cells were also inoculated with the LPAI viruses and showed susceptibility based upon antigen staining; however, the requirement for trypsin to support replication in cell culture limited replication. To assess influenza virus replication in culture, primary alligator cells were inoculated with H1N1 human influenza or H5N1 HPAI viruses that replicate independent of trypsin. Both viruses replicated efficiently in culture, even at the 30 C temperature preferred by the alligator cells. This research demonstrates the ability of wild-type influenza viruses to infect and replicate within two crocodilian substrates and suggests the need for further research to assess crocodilians as a species potentially susceptible to influenza virus infection.
INTRODUCTION
The emergence of zoonotic influenza A virus, including highly pathogenic avian influenza (HPAI) viruses H5N1 and pandemic H1N1 influenza viruses, has sparked worldwide interest in identifying and understanding which and how many species can be infected and serve as reservoir or vector species for influenza A virus. While influenza A virus has been well documented in mammalian and avian species, new viruses and host species are still being identified, most recently in bats (Tong et al. 2012; Tong et al. 2013). Recent research suggests that insects (Barbazan et al. 2008), reptiles (Huchzermeyer 2002; Davis and Spackman 2008), and amphibians are also susceptible to influenza infection (Huchzermeyer 2002; Mancini et al. 2004; Barbazan et al. 2008; Davis and Spackman 2008). Four species of crocodilian, the Chinese alligator (Alligator sinensis), smooth-fronted caiman (Paleosuchus trigonatus), broad-snouted caiman (Caiman latirostris), and Nile crocodile (Crocodylus niloticus), have evidence of influenza A susceptibility, with portions of the virus genome identified by PCR analysis of blood and serum of captive crocodilians (Davis and Spackman 2008). In addition, influenza C was identified by electron microscopy in the feces of farmed Nile crocodiles (Huchzermeyer 2003). To our knowledge, only samples from captive crocodilians have been analyzed; thus, data are deficient for wild crocodilians. Susceptibility to infection through direct inoculation of crocodilian tissues, cells, or live animals has not been investigated.
Avian species, mainly Anseriformes (e.g., ducks) and Charadriiformes (e.g., gulls), are the natural reservoirs of influenza A viruses (Hubalek 2004; Krauss et al. 2007). Evolutionarily, crocodilians are most closely related to birds (Zardoya and Meyer 2001; Crawford et al. 2012). Crocodilians and birds share several physiologic features such as egg structure and embryonic development (Deeming and Ferguson 1991), similar antibody isotypes (Warr et al. 1995), and homologous pancreatic polypeptides (Lance et al. 1984; Deeming and Ferguson 1991; Warr et al. 1995). Moreover, birds and crocodilians demonstrate similar susceptibility to some pathogens including West Nile virus, caiman pox virus, crocodile pox virus, adenoviruses, Newcastle disease virus, eastern equine encephalitis virus, and coronaviruses (Ritchie 1995; Huchzermeyer 2003; Klenk et al. 2004; Davis and Spackman 2008). Therefore, it is reasonable to hypothesize that crocodilians may be susceptible to other viruses endemic to birds, including influenza viruses, and could potentially serve as a reservoir or mixing vessel for these viruses.
Extant crocodilians are traditionally divided into three families including Alligatoridae, Crocodylidae, and Gavialidae, with the Alligatoridae including alligators and caimans (Janke et al. 2005; St. John et al. 2012). American alligators (Alligator mississippiensis) live in proximity to and opportunistically feed on various avian species (Elsey et al. 2004). Thus, alligators may routinely be exposed to avian influenza virus (AIV) through the ingestion of infected tissues and inhalation of infectious particles during feeding (Reperant et al. 2008). Alligators are also exposed to bird excrement. As the fecal-oral route is considered the primary route of influenza virus transmission in aquatic birds (Wright et al. 1992), this may provide an additional route of infection for alligators.
The shared biologic and ecologic features of alligators and aquatic birds make alligators an important animal to investigate as a potential mixing vessel or reservoir for AIV. In addition, with increasing wild and farm populations, along with habitat encroachment, human-alligator interactions are increasing. We used primary alligator cells and alligator embryos to assess the susceptibly of American alligators to influenza A virus infection.
MATERIALS AND METHODS
Viruses
The low pathogenic avian influenza (LPAI) viruses used in this study were: A/chicken/Texas/167280-4/02 (H5N3), A/blue-winged teal/Louisiana/69B/87 (H4N8), A/mallard/Minnesota/Sg-00692/08 (H2N3), and A/blue-winged teal/Louisiana/Sg-00224/07 (H3N8). Notably, the H3N8 strain was isolated from the same region where the alligator eggs for this study were collected. The LPAI viruses were provided by David Stallknecht (University of Georgia, Athens, Georgia, USA). The human influenza A virus (A/WSN/1933 [H1N1]) is a mouse-adapted virus that replicated in cell culture without trypsin when serum is present (Goto et al. 2001). The HPAI virus, A/Viet Nam/1203/2004 (H5N1), was provided by Richard Webby (Saint Jude Children's Research Hospital, Memphis, Tennessee, USA). The viruses were propagated at 37 C in 9–11-d-old embryonated chicken eggs. Virus titers were assayed in Madin Darby canine kidney (MDCK) cells by 50% tissue culture infectious dose (TCID50) assay (Soboleski et al. 2011; Mooney et al. 2013) and ranged from 104.50 to 106.24 TCID50/mL. All experiments using live HPAI viruses were approved by the institutional biosafety program at the University of Georgia and were conducted in biosafety level 3–enhanced containment following guidelines for use of Select Agents approved by the US Centers for Disease Control and Prevention.
Tissue culture
Primary American alligator embryonic fibroblasts were established from a 41–51-d-old embryo. The embryo was chilled overnight at 4 C, dissected from the egg, washed with alpha-MEM (Gibco, Carlsbad, California, USA) and 1× antibiotics (100 IU/mL penicillin G, 100 µg/mL streptomycin, 0.25 µg/mL amphotericin B), cut into 1×1-mm segments, and washed three times in MEM 1× antibiotic/antimycotic mixture. Segments were digested in a 1∶10 dilution of collagenase B (Roche, Indianapolis, Indiana, USA) in alpha-MEM cell culture media containing antibiotics (penicillin/streptomycin, amphotericin B; alpha-MEM/AB media) for 3 h. Suspensions were filtered through a 70-µm cell strainer (BD Falcon, San Jose, California, USA), washed with alpha-MEM/AB media, and centrifuged at 1,500 × G for 15 min. Cells, which were presumed to be fibroblasts based on morphologic characteristics, were cultured in the alligator cell culture media (ACCM) for all experiments unless otherwise stated (ACCM: 175 mM MEM [Gibco], Primocin (100 µg/mL [Invivogen] San Diego, California, USA), 2 mM L-glutamine (HyClone, South Logan, Utah, USA), 1× penicillin/streptomycin, amphotericin B (CellGro, Manassas, Virginia, USA), 10% fetal bovine serum (FBS), 25 mM HEPES (Gibco), and 1× sodium bicarbonate (7.5 µg/mL; Gibco) at pH 7.5 (Cuchens and Clem 1979). Cells were incubated at 30 C and 6% CO2.
Egg inoculation
Alligator eggs were collected from Rockefeller Wildlife Refuge (Cameron and Vermilion Parishes, Louisiana, USA, 29°73′N, 92°82′W). The University of Georgia Institutional Animal Care and Use Committee approved all protocols involving animals. Eggs were incubated in a mixture of moist vermiculite and peat moss at 28–30 C and 90% humidity.
Viability was determined by candling. Viable eggs were segregated into four groups of 10 infected eggs and one group of 12 control eggs. Of the infected groups, five eggs were incubated at 33 C and another five at 36 C (33 C is the optimum temperature for normal alligator embryonic development; 36 C is the closest optimal temperature for virus replication without inducing embryonic lethality; Smith 1975). The control eggs were divided equally into temperature controls and vehicle controls (three untreated and three given vehicle and incubated at 33 C; three untreated and three given vehicle and incubated at 36 C). The injection site was disinfected with 70% ethanol prior to inoculation. Virus diluted (1∶100) in phosphate-buffered saline (PBS) and antibiotics (100 IU penicillin G, 100 µg/mL streptomycin, 0.25 µg amphotericin B/mL) were injected blindly (the allantoic fluid cavity was targeted but not confirmed) using a sterile 24-mm 18-ga needle. The injection site was sealed with Elmer's glue and eggs were incubated for 5 d at indicated temperatures. Prior to extraction, eggs were chilled overnight at 4 C. Embryos were placed into 50 mL sterile conical tubes and filled with virus transport media (VTM; MEM, 100× antibiotic/antimycotic [10,000 IU/mL penicillin, 10,000 µg/mL streptomycin, 25 µg/mL amphotericin B], 50 µg/mL gentamicin, 50 mg/mL kanamycin, pH 7.4). The allantoic fluid samples and whole embryos were stored at −80 C. The control embryos were cut to reveal viscera and stored at room temperature in 10% formalin.
Virus titration
Virus titration of allantoic fluid and the tissues from embryonated alligator eggs were assayed by TCID50 assay using a cell-based enzyme-linked immunosorbent assay (ELISA). Embryos were thawed at 4 C for 8 hr; tissues were extracted from embryos and homogenized in VTM using Qiagen® Tissue Lyser (Germantown, MD, USA).
For the TCID50 assay, MDCK cells plated in 96-well micro-titer plates were washed and replaced with MDCK infection medium (MEM, 2 mM L-glutamine, 1× antibiotic/antimycotic [100 IU/mL penicillin, 100 µg/mL streptomycin, 25 µg/mL amphotericin B], 50 µg/mL gentamicin, and 1 µg/ml tosyl phenylalanyl chloromethyl ketone [TPCK] trypsin [Worthington, Lakewood, New Jersey, USA]). Tissue homogenates were added in triplicate followed by 10-fold serial dilutions in the remaining rows. Negative (uninfected culture medium) and positive controls constituted the last row, with the positive controls infected with 500–750 TCID50 of virus. Plates were covered and incubated for 36 h or 72 h (37 C, 5% CO2) and fixed with methanol/acetone (80∶20).
Plates were blocked overnight (5% nonfat dry milk, 0.5% bovine serum albumin [BSA], and 1× KPL wash buffer; KPL, Gaithersburg, Maryland, USA) at 4 C, washed three times, and incubated 1 h with mouse anti-influenza nucleoprotein (NP) immunoglobulin G (IgG; H16-L10; provided by Jon Yewdell, NIAID, Bethesda, Maryland, USA). After washing, anti-NP binding was detected using goat anti-mouse IgG horseradish peroxidase (HRP) conjugate (1∶10,000 dilution in KPL wash buffer) and incubated in the dark at room temperature for 1 h, followed by washing and detection using 1-Step Ultra TMB (3,3′,5,5′-tetramentylbenzidine) ELISA substrate solution (Pierce, Rockford, Illinois, USA) following the manufacturer's directions. Sulfuric acid was added to stop the TMB reaction and plates were read at 450 nm using a BioTek Powerwave plate reader (Bio-TEK, Winooski, Vermont, USA). For some samples, the virus titer was also estimated by measuring hemagglutinin (HA) of MDCK culture supernatants with 0.5% turkey red blood cells (Soboleski et al. 2011; Mooney et al. 2013). Briefly, 0.05 mL of supernatant from each well of the TCID50 plate was added to 0.05 mL of 0.5% red blood cells (diluted in PBS) and assayed for agglutination within 1 h. The 50% endpoint was calculated via the Reed and Muench method (Reed and Meunch 1938).
Immunofluorescence staining
Alligator fibroblast cells were trypsinized using 0.05% Trypsin-EDTA and then plated at 7.5×105 cells/mL in ACCM in 12-well plates. Once cells reached 80–90% confluence, the medium was removed and virus diluted (multiplicity of infection [MOI]+0.05) in virus infection media (ACCM without FBS and HEPES plus 1 µg/mL TPCK-trypsin was added to each well). Inoculated plates were incubated for 3 h at 30 C and 6% CO2. Subsequently, virus infection media was removed, plates were washed with sterile PBS, and alligator cell growth medium (without trypsin) was added to each well. Plates were incubated at 33 C under 6% CO2 for 24 h and 72 h or at 37 C under 6% CO2 for 24 h. Following incubation, plates were washed in PBS and fixed for 20 min with methanol/acetone (80∶20). Plates were blocked (5% FBS, 0.1% Tween®20, and 1× KPL wash buffer) overnight at 4 C, washed three times, and incubated for 3 h at room temperature with mouse anti-influenza NP IgG (H16-L10). Plates were washed and goat anti-mouse IgG FITC-conjugated in 1× PBS was added to each well. Plates were incubated at room temperature for 1 h in the dark, washed, and 1 µg/mL of 4′,6-diamidino-2-phenylindole was added to each well to visualize cell nuclei. Plates were incubated for 20 min at room temperature in the dark, washed, and examined using a Zeiss Axiovert 40 CFL fluorescent light microscope (Carl Zeiss Microscopy, Thornwood, New York, USA).
Assay for in vitro replication
Alligator fibroblasts were plated at 104 cells/well in ACCM in a 24-well plate. At 60% confluence, ACCM was removed and cells were washed three times with 1× MEM. Fibroblasts were infected with 100 plaque-forming units (PFU) of A/WSN/33 (H1N1) or A/Viet Nam/1203/04 (H5N1) diluted in ACCM without HEPES for 3 h at 37 C, 5% CO2. Following incubation, cells were washed three times with 1× MEM, and 1 mL of ACCM without HEPES was added to each well and plates were incubated at 30 C, 6% CO2. At 24, 48, and 72 h, supernatants were collected, clarified by centrifugation, BSA added to 0.2%, aliquoted, and stored at −80 C. Infectious virus titers in supernatants was determined by TCID50 assay, as described, except that determination of influenza virus-positive wells was determined by HA.
Immunohistochemistry
Formalin-fixed alligator organs (brain, trachea, lung, heart, liver, intestine, stomach, kidney, spleen, and pancreas) were dissected from embryos, embedded in paraffin, sectioned, and stained with H&E for histologic analysis. Other sections were rehydrated using standard procedures and stained using mouse anti-influenza A NP IgG, biotin conjugated (Bioss Inc., Woburn, Massachusetts, USA), followed by a streptavidin-HRP labeled secondary antibody. Lastly, 3,3′-diaminobenzidine was added to produce a brown precipitate in antigen-positive tissues. Sections were reviewed and scored by a board certified pathologist (E.W.U.).
Real-time reverse-transcriptase (RT)-PCR
Total RNA was extracted from the liver, kidney, and a subset of allantoic fluid samples using a Qiagen RNeasy Mini Kit, according to the manufacturer's protocol, except that a 250-µL aliquot of allantoic fluid was used in the extraction process (Spackman et al. 2002).
The cDNA synthesis and subsequent real-time RT-PCR were performed using a Qiagen OneStep RT-PCR Kit on a Stratagene MX3000p or 3005p thermocycler (Agilent Technologies, Santa Clara, CA, USA) following the protocol of Spackman (2002). This is a well-established TaqMan® (Life Technologies, Carlsbad, CA, USA) protocol used as an influenza A virus diagnostic assay. The primer probe sequences are: M+25, AGA TGA GTC TTC TAA CCG AGG TCG; M–124, TGC AAA AAC ATC TTC AAG TCT CTG; and M+64, FAM-TCA GGC CCC CTC AAA GCC GA-TAMRA (Spackman et al. 2002). Each sample was run in triplicate with influenza A/WSN/33 as a positive control.
Statistics
All statistical analyses were performed using Graph Pad Prism software (GraphPad Software, Inc., La Jolla, California, USA). A Mann-Whitney test or analysis of variance, followed by Bonferroni's multiple comparison test, were used to determine statistical significance. The level of significance for all data was set at <0.05.
RESULTS
In vitro infection of primary embryonic fibroblasts
Alligator primary embryonic fibroblasts were generated as described in the Methods and infected to determine susceptibility to AIV infection. Following inoculation (MOI = 0.05) with A/blue-winged teal/Louisiana/Sg-00224/07 (H3N8), primary embryonic alligator fibroblasts were positive for virus NP (Fig. 1). The NP antigen was present within the cytoplasm but was more strongly co-localized in the nucleus. Cells infected with AIV were positive for NP antigen after incubation at 33 C or 37 C at 24 h postinfection. Infected alligator fibroblasts were also positive for NP at 72 h postinfection at 33 C incubation. Infection with A/blue-winged teal/Louisiana/69B/87 (H4N8), A/chicken/Texas/167280-4/02 (H5N3), or A/mallard/Minnesota/Sg-00692/08 (H2N3) showed similar results (data not shown).
Immunofluorescent staining of nucleoprotein (NP) antigen in primary alligator (Alligator mississippiensis) fibroblasts infected with avian influenza virus H3N8. Positive staining of NP at a multiplicity of infection of 0.05, 24 h postinfection at 33 C, at 40× (A) and 10× (B) magnification. Co-localization of 4′, 6-diamidino-2-pheylindole and NP 24 h postinfection at 33 C at 40× (C) and 10× (D) magnification.
Immunofluorescent staining of nucleoprotein (NP) antigen in primary alligator (Alligator mississippiensis) fibroblasts infected with avian influenza virus H3N8. Positive staining of NP at a multiplicity of infection of 0.05, 24 h postinfection at 33 C, at 40× (A) and 10× (B) magnification. Co-localization of 4′, 6-diamidino-2-pheylindole and NP 24 h postinfection at 33 C at 40× (C) and 10× (D) magnification.
Infection of embryonated alligator eggs with AIV
Embryonated alligator eggs were inoculated with one of the four LPAI (H5N3, H4N8, H3N8, or H2N3) strains to determine susceptibility to virus replication in ovo. Allantoic fluid and embryos were extracted 5 d postinoculation to assess the susceptibility to infection.
Infectious virus titers postinoculation exceeded input virus for all four strains, as indicated by virus titer levels in allantoic fluid of embryonated alligator eggs (Fig. 2). The H5N3, H4N8, and H3N8 strains of influenza virus incubated at 33 C had mean virus titers significantly greater than input virus (P<0.001) with values of 104.79 (±100.77), 103.84 (±100.50), and 103.59 (±100.55) TCID50/mL, respectively (mean [±SEM]), whereas at 36 C these three strains had mean virus titers of 104.84 (±100.65), 104.35 (±100.15), and 104.39 (±100.57) TCID50/mL, respectively. This suggests that temperatures chosen for replication had no significant impact on virus replication when cultured in alligator eggs. In contrast, the H2N3 virus demonstrated improved replication at the higher temperature, with the 33 C group having a mean virus titer of 103.49 (±100.67) TCID50/mL compared to 106.04 (±100.67) TCID50/mL at 36 C (P = 0.032). To validate these results, a subset of allantoic fluid samples was assayed in a standard TCID50 assay using HA as a measure of virus replication (Tompkins et al. 2007; Soboleski et al. 2011). The virus titer, as measured by HA endpoint, corresponded almost directly with the titers measured by cell-based ELISA (data not shown).
Mean virus titer in allantoic fluid from embryonated alligator (Alligator mississippiensis) eggs. Embryonated eggs (n = 5/group) were inoculated with the indicated virus and incubated for 5 d at the indicated temperatures (33 C or 36 C). Allantoic fluid was collected and assayed for infectious virus by 50% tissue culture infectious dose (TCID50) assay (*statistically significant, P<0.05).
Mean virus titer in allantoic fluid from embryonated alligator (Alligator mississippiensis) eggs. Embryonated eggs (n = 5/group) were inoculated with the indicated virus and incubated for 5 d at the indicated temperatures (33 C or 36 C). Allantoic fluid was collected and assayed for infectious virus by 50% tissue culture infectious dose (TCID50) assay (*statistically significant, P<0.05).
Virus genomic RNA was measured in a subset of allantoic fluid samples (n = 5) from the 36 C alligator eggs. The RNA was purified from allantoic fluid and M gene genomic RNA levels were measured using real-time RT-PCR. The mean cycle threshold (Ct) values from allantoic fluid samples (Fig. 3) were below the negative threshold level of 34.79 (±0.17). H5N3-infected samples assayed had a mean Ct value of 13.97 (±0.15), while the H3N8- and H4N8-infected samples had mean Ct values of 15.38 (±0.13) and 15.25 (±0.06), respectively. The H2N3-infected samples had the lowest mean Ct value, 13.23 (±0.27), corresponding to the highest titer measured in the TCID50 assay (Fig. 2). A/WSN/33 (H1N1), used as a positive control, had a mean Ct value of 15.67 (±0.37). Taken together, the infectious virus assay and M gene PCR data indicate influenza virus replication occurred within the alligator eggs.
Real-time reverse-transcription-PCR of M gene levels from alligator (Alligator mississippiensis) allantoic fluid infected with low pathogenic avian influenza virus. Embryonated eggs were inoculated with influenza virus and incubated for 5 d at 36 C. Allantoic fluid was collected and assayed for virus genome. Allantoic fluid was collected from naïve eggs as a negative control. Allantoic fluid from A/WSN/33-infected embryonated chicken eggs was used as a positive control. All infected samples were below the negative control cycle threshold (Ct) of 34.79. Control (n = 9); H5N3, H4N8, and, H3N8 (n = 6); H2N3 (n = 3); and WSN (n = 9).
Real-time reverse-transcription-PCR of M gene levels from alligator (Alligator mississippiensis) allantoic fluid infected with low pathogenic avian influenza virus. Embryonated eggs were inoculated with influenza virus and incubated for 5 d at 36 C. Allantoic fluid was collected and assayed for virus genome. Allantoic fluid was collected from naïve eggs as a negative control. Allantoic fluid from A/WSN/33-infected embryonated chicken eggs was used as a positive control. All infected samples were below the negative control cycle threshold (Ct) of 34.79. Control (n = 9); H5N3, H4N8, and, H3N8 (n = 6); H2N3 (n = 3); and WSN (n = 9).
AIV infection in embryonic tissues
After confirmation of infection and virus replication of all four AIV strains in the allantoic fluid of embryonated alligator eggs, localization of replication was determined. Organs were extracted from embryos for H&E staining, IHC, virus culture, and RNA isolation.
All mock-infected embryos were negative for congestion and necrosis in all tissues and had no indications of antigen staining. However, one vehicle control at 36 C had fluid within the lungs, resulting in nonspecific staining, and so was excluded from further analysis (data not shown). In contrast, the alligator eggs infected with each of the four virus strains had NP staining predominantly in liver and kidneys and, in accordance with previous results, had higher levels of staining at 36 C. Embryos inoculated with each respective strain were positive for NP antigen staining in liver, kidney, or both. Likewise, positive staining was present at 33 C and 36 C. Embryos illustrated necrosis, congestion, and lesions in kidney and liver and in other tissues with all four virus strains (Table 1).
Histopathology from embryonic tissues. Summary of histology and immunohistochemistry data from embryos extracted from inoculated alligator (Alligator mississippiensis) eggs. Numbers in parentheses are number positive/number tested. Dashes (–) indicate no pathology or antigen staining found.

The IHC analysis indicated that the liver and kidney might serve as predominant sites for virus replication. However, attempts to measure tissue virus titers from embryos by TCID50 were unsuccessful. To confirm the results obtained from the IHC, real-time RT-PCR was utilized. Total RNA was extracted from both kidney and liver from the remaining embryos in each group and M gene RNA levels were compared with controls. The mean control Ct values from the uninfected embryonic tissues were used to calculate the negative threshold value (36.40). The control values for the embryos incubated at 33 C were 40.00 (±1.03) for kidney (n = 9) and 42.19 (±1.14) for liver (n = 9; mean [±SD]). Embryos incubated at 36 C had similar values of 40.65 (±1.03) for kidney (n = 12) and 41.80 (±0.75) for liver (n = 12). The mean Ct value of liver and kidney samples from embryos infected with each of the four strains at two respective temperatures were below the negative threshold, indicating the presence of M gene genomic RNA and virus replication in inoculated embryos (Fig. 4). Temperature of incubation did not affect virus replication for H4N8, H3N8, or H5N3 AIV. In contrast, consistent with previous results, H2N3-inoculated embryos incubated at 36 C had increased M gene RNA compared with embryos incubated at 33 C (Fig. 4c).
Reverse-transcription-PCR of M gene in infected embryonic tissues. Real-time RT-PCR results of influenza A genomic RNA specific for M gene from kidney and liver samples of infected and uninfected alligator (Alligator mississippiensis) embryos. Embryonated eggs were inoculated with influenza virus and incubated for 5 d at 33 C or 36 C. Eggs were necropsied and tissues collected and processed for RNA as described in the materials and methods. Tissues were collected from naïve eggs as a negative control. All samples with a cycle threshold (Ct) value >36.40 were classified as negative for M gene. All samples with a Ct <36.40 were classified as positive for influenza A genomic RNA. A + indicates mean Ct value for all samples (n = 9, except for 36 C control kidney and control liver, where n = 12). White box indicates an incubation temperature of 33 C; checkered box indicates incubation temperature of 36 C. (A) H4N8-infected tissues. (B) H3N8-infected tissues. (C) H2N3-infected tissues. (D) H5N3-infected tissues.
Reverse-transcription-PCR of M gene in infected embryonic tissues. Real-time RT-PCR results of influenza A genomic RNA specific for M gene from kidney and liver samples of infected and uninfected alligator (Alligator mississippiensis) embryos. Embryonated eggs were inoculated with influenza virus and incubated for 5 d at 33 C or 36 C. Eggs were necropsied and tissues collected and processed for RNA as described in the materials and methods. Tissues were collected from naïve eggs as a negative control. All samples with a cycle threshold (Ct) value >36.40 were classified as negative for M gene. All samples with a Ct <36.40 were classified as positive for influenza A genomic RNA. A + indicates mean Ct value for all samples (n = 9, except for 36 C control kidney and control liver, where n = 12). White box indicates an incubation temperature of 33 C; checkered box indicates incubation temperature of 36 C. (A) H4N8-infected tissues. (B) H3N8-infected tissues. (C) H2N3-infected tissues. (D) H5N3-infected tissues.
Replication of influenza virus in alligator fibroblast cells
As the LPAI viruses were able to replicate in alligator embryos, we next determined the ability of alligator primary embryonic fibroblasts to support the complete influenza virus lifecycle. Influenza virus generally requires addition of exogenous trypsin or another protease to activate the HA of progeny virions, enabling multiple replication cycles (Shaw and Palese 2013; Wright et al. 2013). These proteases can be damaging to cells in culture, limiting the ability to assess influenza virus replication in many cell substrates. However, A/WSN/33 (H1N1) influenza can replicate efficiently without the addition of trypsin through sequestration of plasminogen in exogenously added serum (Goto et al. 2001). We infected alligator fibroblasts with 100 PFU of A/WSN/33 in media containing serum, cultured the infected cells at 30 C, and collected supernatants over time. The supernatants were assayed for infectious virus by TCID50 assay. At 24 h postinoculation there was limited, albeit measurable, infectious virus detected (Fig. 5a). The lower culture temperature was likely prolonging the eclipse phase of virus replication. However, at 48 h and 72 h postinoculation, there was significantly more infectious virus (P = 0.003), demonstrating replication of A/WSN/33 in alligator cells.
Infectious titers from inoculated primary alligator fibroblasts. Primary alligator fibroblasts were inoculated with 100 plaque-forming units of (left) A/WSN/33 (H1N1) or (right) A/Vietnam/1203/04 (H5N1) influenza virus and incubated for 72 h at 30 C. Culture supernatants were collected at 24, 48, and 72 h postinoculation and assayed for infectious virus by 50% tissue culture infectious dose (TCID50) assay as described in the materials and methods. Error bars indicate SEM of triplicate cultures (*statistically different P<0.05 compared to 24-h time point).
Infectious titers from inoculated primary alligator fibroblasts. Primary alligator fibroblasts were inoculated with 100 plaque-forming units of (left) A/WSN/33 (H1N1) or (right) A/Vietnam/1203/04 (H5N1) influenza virus and incubated for 72 h at 30 C. Culture supernatants were collected at 24, 48, and 72 h postinoculation and assayed for infectious virus by 50% tissue culture infectious dose (TCID50) assay as described in the materials and methods. Error bars indicate SEM of triplicate cultures (*statistically different P<0.05 compared to 24-h time point).
A hallmark of HPAI viruses is the ability to replicate in culture in the absence of trypsin. The HA proteins of HPAI viruses have polybasic cleavage sites (in contrast to the monobasic cleavage sites found in most influenza viruses) that enable cleavage and activation of the HA protein by ubiquitous, furin-like proteases found in the Golgi complex of eukaryotic cells (Shaw and Palese 2013; Wright et al. 2013). To assess the ability of an avian influenza virus to replicate in alligators, we infected fibroblasts with 100 PFU of A/Viet Nam/1203/2004 (H5N1) and collected and assayed supernatants as before. Similar to the H1N1 virus, there was limited virus at 24 h but significant increases in infectious virus at 48 h and 72 h postinoculation (P<0.001), confirming that alligator cells are susceptible to AIV infection and can support the complete infectious lifecycle (Fig. 5b).
DISCUSSION
We demonstrated in ovo and in vitro susceptibility of American alligators to LPAI virus infection, with infection and replication occurring at temperatures much lower than wild-type AIVs would normally replicate (Massin et al. 2001, 2010; Scull et al. 2009). Measurement of virus load in allantoic fluid and culture supernatants indicated robust replication and production of virus progeny after infection with human influenza, LPAI, and HPAI viruses. While we demonstrated in ovo and in vitro susceptibility to AIV in alligators, the ability of AIV to replicate in live crocodilians remains unknown.
Primary alligator fibroblast cells were positive for NP after infection with each strain, demonstrating infection with multiple subtypes and strains. The primary alligator fibroblasts also supported production of infectious progeny viruses at 30 C and 37 C (data not shown). While the trypsin requirement that is common for growth of most influenza viruses in cell culture limited the viruses which could be assayed for replication, both human and avian influenza viruses replicated in the alligator cells, suggesting sialic acid moieties were present for HA binding.
All four AIVs replicated within the embryonated alligator eggs, with all four strains infecting liver and kidneys. All virus-inoculated eggs produced high virus titers in the allantoic fluid. Virus titers were several logs greater than input titers for all four strains and at both temperatures, with the highest production at 36 C. Overall, these temperatures coincide within an optimal temperature range for alligator metabolic activity (Gans 1969). However, reptilian immune function is dependent upon temperature, with extreme temperatures perturbing function; thus, suppressed immune function may account for the increased virus titer at 36 C among certain strains (el Ridi et al. 1988). In contrast to increased titers at 36 C for the H2N3 virus, H5N3 virus replication was less affected by temperature, although the H5N3 virus also had the smallest increase in titer. However, these data are not unexpected as concomitant in vitro and in vivo findings in a murine model illustrate that H5N3 replication may occur at temperatures as low as 33 C (Hatta et al. 2007).
Embryonic organs, notably liver and kidney, of infected eggs were positive for both genomic RNA and NP antigen, suggesting that infection and replication occurred within these organs. In contrast, the most-common sites of infection in avian species are the lower respiratory tract and the gastrointestinal tract (Wright et al. 1992); therefore, we hypothesized that virus infection and replication would occur in similar tissues within embryonated alligators. However, as we restricted our study to in vitro and in ovo infections, the tissue tropism for infections in vivo remain unknown.
The viruses we used were all isolated from birds and geographic regions that overlap with alligator habitats in North America (Stallknecht et al. 1990). Thus, it is possible for these viruses to intersect with alligator populations. The H5N3 LPAI virus we tested was isolated from a poultry farm in Texas (Lee et al. 2004) and replicated in both primary cells and embryos at both low (33 C) and high (36 C) temperatures. An H5N1 HPAI virus replicated in alligator cells in culture, and other HPAI viruses, have been isolated from geographic regions overlapping with crocodilian habitats. While we did not assess the infection and transmission of influenza viruses in live crocodilians, the data suggest that alligators might be susceptible to influenza virus infection and that further study is warranted.
ACKNOWLEDGMENTS
We extend a special thanks to Ruth Elsey of the Louisiana Department of Fish and Game for generously providing alligator eggs, to Frank Michel (University of Georgia, Athens, Georgia, USA) for technical support, Jon Yewdell (National Institute of Allergy and Infectious Diseases, Bethesda, Maryland, USA) for providing the mouse anti-influenza NP IgG hybridoma H16-L10, and to David Stallknecht (University of Georgia, Athens, Georgia, USA), David Suarez (Southeast Poultry Research Laboratory, US Department of Agriculture, Athens, Georgia, USA), and Richard Webby (Saint Jude Children's Research Hospital, Memphis, Tennessee, USA) for providing the influenza viruses used in our study.
LITERATURE CITED
Author notes
These authors contributed equally to this manuscript.