ABSTRACT
Two subtypes of influenza A virus (IAV), avian-origin canine influenza virus (CIV) H3N2 (CIV-H3N2) and equine-origin CIV H3N8 (CIV-H3N8), are enzootic in the canine population. Dogs have been demonstrated to seroconvert in response to diverse IAVs, and naturally occurring reassortants of CIV-H3N2 and the 2009 H1N1 pandemic virus (pdmH1N1) have been isolated. We conducted a thorough phenotypic evaluation of CIV-H3N2 in order to assess its threat to human health. Using ferret-generated antiserum, we determined that CIV-H3N2 is antigenically distinct from contemporary human H3N2 IAVs, suggesting that there may be minimal herd immunity in humans. We assessed the public health risk of CIV-H3N2 × pandemic H1N1 (pdmH1N1) reassortants by characterizing their in vitro genetic compatibility and in vivo pathogenicity and transmissibility. Using a luciferase minigenome assay, we quantified the polymerase activity of all possible 16 ribonucleoprotein (RNP) complexes (PB2, PB1, PA, NP) between CIV-H3N2 and pdmH1N1, identifying some combinations that were more active than either parental virus complex. Using reverse genetics and fixing the CIV-H3N2 hemagglutinin (HA), we found that 51 of the 127 possible reassortant viruses were viable and able to be rescued. Nineteen of these reassortant viruses had high-growth phenotypes in vitro, and 13 of these replicated in mouse lungs. A single reassortant with the NP and HA gene segments from CIV-H3N2 was selected for characterization in ferrets. The reassortant was efficiently transmitted by contact but not by the airborne route and was pathogenic in ferrets. Our results suggest that CIV-H3N2 reassortants may pose a moderate risk to public health and that the canine host should be monitored for emerging IAVs.
IMPORTANCE IAV pandemics are caused by the introduction of novel viruses that are capable of efficient and sustained transmission into a human population with limited herd immunity. Dogs are a a potential mixing vessel for avian and mammalian IAVs and represent a human health concern due to their susceptibility to infection, large global population, and close physical contact with humans. Our results suggest that humans are likely to have limited preexisting immunity to CIV-H3N2 and that CIV-H3N2 × pdmH1N1 reassortants have moderate genetic compatibility and are transmissible by direct contact in ferrets. Our study contributes to the increasing evidence that surveillance of the canine population for IAVs is an important component of pandemic preparedness.
INTRODUCTION
Influenza A virus (IAV), a member of the family Orthomyxoviridae, is an enveloped virus with eight single-stranded, negative-sense RNA genomic segments (1, 2). Migratory waterfowl are the natural reservoirs of IAVs, although the viruses naturally infect diverse species, including domestic poultry, pigs, dogs, horses, and sea mammals, and cause endemic and pandemic outbreaks in humans (3–7). IAV subtypes are determined by two viral surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), and 18 HA and 11 NA subtypes have been identified (8).
IAV diversity is maintained through the mechanisms of genetic drift and genetic shift. Genetic drift results from the slow accumulation of mutations over multiple replication cycles. Genetic shift is an abrupt change in the genome that occurs when two different IAVs intermix via a coinfected host cell and one or more genomic segments are exchanged. Antigenic shift occurs when the reassortant virus has altered antigenicity. Both mechanisms of diversity can alter the virulence and pathogenicity in the host and/or alter the host range. A zoonotic reassortant virus may generate a pandemic when there is little or no preexisting human immunity and the virus is capable of transmitting efficiently among humans, as occurred in 1957 and 1968 (9) and in 2009 (10).
Dogs warrant special consideration as an influenza virus host due to several factors. Their large global population is difficult to quantify, but a recent estimate is ∼700 million (11), with many of these having close human contact. Dogs are susceptible to avian and mammalian IAVs and may act as mixing vessels for promoting reassortment (12, 13). Two major canine influenza virus (CIV) subtypes, H3N8 and H3N2, are currently circulating in the global canine population. Equine-origin CIV H3N8 (CIV-H3N8) was identified in 2004 at a greyhound racing track in Florida, USA (14), and was subsequently detected in Canada (15), the United Kingdom (16), and Australia (17). CIV-H3N8 is now enzootic in the U.S. canine population, with the virus being reported in dogs in more than 40 states (18). Avian-origin CIV H3N2 (CIV-H3N2) was first identified in South Korea and China in approximately 2006 or 2007 and was later detected in Thailand (19–21) and other areas of Southeast Asia (22–25). CIV-H3N2 was first isolated from dogs outside Asia in the April 2015 outbreak in the midwestern United States, which affected more than 1,000 dogs (http://mediarelations.cornell.edu/2015/04/12/midwest-canine-influenza-outbreak-caused-by-new-strain-of-virus/). The spread of CIV-H3N2 to the United States was confirmed by the Centers for Disease Control and Prevention (https://www.cdc.gov/flu/news/canine-influenza-update.htm).
As early as the 1970s, serological evidence in the United States suggested that dogs could become infected with human strains of H3N2 (26, 27). More recent serological evidence suggests that dogs have been naturally infected with human seasonal H1N1 and H3N2 viruses (28), avian H5N1 (29), H5N2 (30), H9N2 (31), and H10N8 (32) viruses, and the 2009 H1N1 pandemic virus (pdmH1N1) (33). In China, more than 20% of domestic dogs were shown to have serum antibodies to pdmH1N1, and dogs positive for antibodies to both CIV-H3N2 and pdmH1N1 have been identified (23, 34). During a surveillance effort in 2012, researchers identified a virus isolate that contained the hemagglutinin gene segment from CIV-H3N2 and the remaining 7 gene segments from pdmH1N1 (35); more recently, another isolate that contained the matrix gene segment from pdmH1N1 and all other gene segments from CIV-H3N2 was identified (36).
The unique relationship of dogs and humans, the serological evidence that dogs are infected with diverse IAVs (including two enzootic CIV subtypes), and the isolation of naturally occurring reassortant viruses in dogs are all factors that raise concerns that a novel virus with potential risk to public health may emerge in the canine population. Because of these concerns, we conducted a study to (i) characterize the genetic compatibility of pdmH1N1 and CIV-H3N2, (ii) detect the growth phenotype of pdmH1N1 × CIV-H3N2 reassortants in vitro, and (iii) assess the pathogenicity and transmissibility of these reassortant viruses in mice and ferrets.
RESULTS
GD06 is antigenically distinct from historical human H3N2 seasonal IAVs.We performed hemagglutination inhibition (HI) assays to detect the cross-reactivity between CIV strain A/canine/Guangdong/1/2006 (H3N2) (GD06) and ferret antiserum raised against a representative set of human H3N2 seasonal influenza viruses (37). Test serum samples included 25 serum samples harboring H3N2 viruses from 1979 to 2015, including A/Bangkok/1/1979, A/Philippines/2/82, A/Caen/1/1984, A/Mississippi/1/1985, A/Leningrad/360/1986, A/Sichuan/02/1987, A/Sichuan/60/1989, A/Ann Arbor/03/1993, A/Johannesburg/33/1994, A/Nanchang/933/1995, A/Sydney/5/1997, A/Wisconsin/67/2005, A/Brisbane/10/2007, A/Perth/16/2009, A/Victoria/361/2011, A/Texas/50/2012, A/Switzerland/9715293/2013, A/Utah/07/2013, A/Mississippi/17/2013, A/Costa Rica/4700/2013, A/Palau/6759/2014, A/Hong Kong/4801/2014, A/Fiji/2/2015, A/Brisbane/82/2015, and A/Victoria/503/2015. The corresponding homologous titers were 1:2,560, 1:1,280, 1:1,280, 1:2,560, 1:320, 1:320, 1:480, 1:120, 1:480, 1:720, 1:960, 1:1,280, 1:1,280, 1:640, 1:640, 1:1,280, 1:640, 1:1,280, 1:640, 1:320, 1:1,280, 1:1,280, 1:640, 1:1,280, and 1:640, respectively. Results showed weak cross-reactions of 1:40 and 1:20 between GD06 and A/Sydney/5/1997 antiserum and between GD06 and A/Nanchang/933/1995 (H3N2) antiserum, respectively. However, GD06 had a titer of <1:20 with ferret antiserum raised against the remaining human H3N2 seasonal influenza viruses. The homologous titer for GD06 was 1:1,280. In addition, the HI titer between GD06 and ferret antiserum raised against A/canine/Iowa/2005 (H3N8) was 1:40; the homologous titer for A/canine/Iowa/2005 (H3N8) was 1:160.
GD06 and CA09 gene segments showed high compatibility.To characterize genomic compatibility between strain GD06 and pandemic H1N1 strain A/California/4/2009 (H1N1) (CA09), we successfully generated 51 of 127 reassortants (40%). We determined the growth phenotypes of these 51 reassortants and classified them as having high growth (≥106 50% tissue culture infective doses [TCID50]/ml in Madin-Darby canine kidney [MDCK] cells) (n = 19) (Table 1) or moderate growth (<106 TCID50/ml in MDCK cells) (n = 32) (Table 2). A total of 76 of 127 reassortants (60%) could not be generated from a single transfection and propagation attempt (see Table S1 in the supplemental material). Although the genomic constellations of the 51 reassortants showed that all 7 gene segments of each virus were represented, the gene segments were unequally represented. Among the 51 reassortants generated, GD06 gene segments were underrepresented, whereas CA09 gene segments were overrepresented. The NA, PB1, and PB2 gene segments from GD06 were present in only 11 (22%), 14 (27%), and 17 (33%) of 51 reassortants, respectively. The other gene segments were more equally distributed among the 51 reassortants. The NP, M, NS, and PA gene segments from GD06 were present in 24 (47%), 25 (49%), 25 (49%), and 31 (61%) of 51 reassortants, respectively. However, among the high-growth (>106 TCID50/ml) reassortant viruses, all 19 contained the PB1 gene segment from CA09 and 16/19 (∼84%) contained the PB2 gene segment from CA09.
Reassortant viruses expressing a high-growth phenotypea
Reassortant viruses expressing a moderate-growth phenotypea
Reassortant RNP complexes showed enhanced luciferase activity in vitro.The ribonucleoprotein (RNP) complex consists of PB2, PB1, PA, NP, and viral RNA and is responsible for the replication and transcription activity of IAV. Evidence suggests that RNP complex compatibility functions as a restricting factor and critical component of IAV reassortment (38). There is also evidence of a correlation between enhanced polymerase activity and the pathogenicity of IAV (39, 40). We used a minigenome luciferase assay to quantify the polymerase activity of the 16 RNP complex constellations possible from GD06 and CA09. The CA09 RNP complex and GD06 RNP complex did not show a statistically significant difference in luciferase activity (P > 0.05) (Fig. 1). We compared the luciferase activity of the 14 reassortant RNP complexes to those of the 2 wild-type RNP complexes and found that 4 reassortants expressed significantly greater luciferase activity. These four reassortant RNP complexes shared two common traits: they all contained the NP gene segment from GD06 and the PB1 gene segment from CA09. The RNP complex with the greatest luciferase activity was the complex with PB2 of GD06, PB1 of CA09, PA of CA09, and NP of GD06 (GD06PB2CA09PB1CA09PAGD06NP), which had approximately 11 times the luciferase activity of CA09 and GD06 (P < 0.0001). The RNP complex with the next-highest luciferase activity was CA09PB2CA09PB1CA09PAGD06NP, which had approximately six times the activity of CA09 and GD06 (P < 0.0001). Reassortants CA09PB2CA09PB1GD06PAGD06NP and GD06PB2CA09PB1GD06PAGD06NP had approximately three times the activity of CA09 and GD06 (P < 0.001).
Ribonucleoprotein complex luciferase activity and reassortant virus frequency. The luciferase activities of all 16 RNP complexes between GD06 (shaded in brown) and CA09 (shaded in purple) were compared. The ratio of Renilla luciferase to firefly luciferase (Rluc/Fluc) is expressed as the mean + standard deviation from three independent experiments, and statistical analysis was performed using a one-way ANOVA with post hoc Tukey's multiple-comparison test. The reassortant virus RNP data were frequency counts from Tables 1 and 2.
We compared the RNP luciferase activities with the frequency of reassortant viruses generated in vitro to look for possible correlations. A total of 18 out of 51 (∼35%) of the reassortant viruses consisted of one of the four RNP complexes with high levels of luciferase activity (Fig. 1). Three RNP complex constellations were never generated in any of the reassortant viruses, and these also correlated with the lower end of RNP luciferase activity.
Reassortants caused minimal weight loss in mice but replicated to high titers in mouse lungs.To characterize the pathogenicity of the reassortant viruses, a total of 22 viruses were used to intranasally inoculate mice (n = 8), including the 19 high-growth (>106 TCID50/ml) reassortants (Table 1), GD06, CA09, and a GD06 × A/Puerto Rico/8/1934 (H1N1) (PR8) reassortant. Pathogenicity was determined by measuring weight loss and viral lung titers. Overall, weight loss was minimal or did not occur among the infected mice. We used the loss of >5% of the initial body weight as a predefined marker of pathogenicity; however, there was considerable variation among the individual mice, thus limiting the usefulness of the weight loss data. Weight loss of 13.6% ± 5.5% was observed in mice infected with reassortant 111 at 3 days postinoculation (dpi) (Fig. S1). Mice infected with reassortant 109 showed weight loss of 7.6% ± 6.4% at 8 dpi, and mice infected with reassortant 55 showed weight loss of 6.1% ± 4.3% at 7 dpi. Mice infected with GD06 or reassortants 22, 27, 75, 83, 87, 99, 115, and 126 showed no weight loss. Infection with the remaining reassortants and CA09 resulted in minimal weight loss in mice (i.e., <5%).
To detect viral replication in vivo, we humanely euthanized three mice per group at 4 dpi and collected lung tissues for titration in MDCK cells. Mice infected with reassortant viruses 7, 10, 75, 83, 99, and 107 had viral lung titers at or below the level of detection (0.699 log10 TCID50/ml), whereas all other reassortants had titers above the level of detection (Fig. 2). Mice infected with CA09 showed significantly higher lung titers than mice infected with GD06 at 6.2 ± 0.5 log10 TCID50/ml versus 3.0 ± 0.07 log10 TCID50/ml, respectively (P = < 0.0001; one-way analysis of variance [ANOVA] with post hoc Tukey's test). The reassortant that replicated to the maximum titer in mouse lungs was reassortant 109, which was present at 5.8 ± 0.7 log10 TCID50/ml. Reassortant 109 replicated to a significantly higher titer in mouse lungs than 15 of the 21 viruses tested, including reassortants 7, 10, 75, 83, 99, 107, 27, 71, 22, and 31 (P < 0.0001), GD06 (P = 0.0008), CIV-PR8 (P = 0.0077), reassortant 63 (P = 0.0089), reassortant 115 (P = 0.0029), and GD06 × PR8 (P = 0.0077) (one-way ANOVA with post hoc Tukey's test). There was no significant difference in the mouse viral lung titer between reassortant 109 and the remaining six viruses tested, which included reassortants 95, 23, 55, 87, 126, and pdmH1N1 (P > 0.05; one-way ANOVA with post hoc Tukey's test). Reassortant 109 had NP and HA gene segments from GD06 and all other gene segments from CA09.
Viral lung titers of inoculated mice. Mice (n = 3) were intranasally inoculated with 50 μl of 106 TCID50/ml of virus per ml. At 4 dpi they were necropsied and lung tissues were collected for determination of virus titers. Titers are expressed as the mean ± standard deviation.
At 14 dpi, we collected serum from the infected mice to determine whether seroconversion had occurred. Seroconversions were determined by HI assay, as previously described, with titers of ≥20 being defined as seroconversion (41). Of the 110 mice (i.e., 5 mice × 22 viruses), all but 3, including 1 mouse from the reassortant 27 group, another from the reassortant 83 group, and the last one from the reassortant 115 group, seroconverted (Table S2).
Reassortant 109 replicated efficiently in mammalian cells.To characterize the growth phenotype of the viruses, we conducted virus growth kinetic experiments in vitro. We infected canine kidney (MDCK) cells and human lung epithelial (A549) cells with reassortant 109, CA09, or GD06 at a multiplicity of infection (MOI) of 0.001. In MDCK cells, all three viruses replicated efficiently. The maximum titers in MDCK cells were 7.81 ± 0.54, 7.27 ± 0.37, and 6.20 ± 0.17 log10 TCID50/ml for reassortant 109, CA09, and GD06, respectively (Fig. 3A). There was limited viral replication in A549 cells, with the maximum titers being 4.86 ± 0.29, 3.97 ± 0.25, and 2.4 ± 0.06 log10 TCID50/ml for CA09, reassortant 109, and GD06, respectively. The maximum replication titer in A549 cells for all three viruses occurred at 48 h postinoculation.
Growth kinetics of viruses. The growth kinetics of each virus shown were characterized in MDCK cells (A) or A549 cells (B) at an MOI of 0.001. Virus titers are expressed as the mean ± standard deviation from three independent experiments. The limit of virus detection was 100.699 TCID50/ml. R109, reassortant 109.
GD06 showed limited transmission in ferrets.To determine the transmissibility of GD06 through direct contact, we anesthetized and intranasally inoculated three ferrets each with GD06. One day later, we paired three naive ferrets with GD06-inoculated ferrets, which were housed in three individual cages. The three GD06-inoculated ferrets shed virus at titers of 3.93 to 5.032 log10 TCID50/ml (Fig. 4A), and the naive ferrets shed virus at titers of 0.699 to 5.366 log10 TCID50/ml (Fig. 4B). We were unable to detect the viral titer in the naive ferret placed in direct contact with the inoculated ferret in pair 1 (Fig. 4B). At 5 dpi, the virus-inoculated ferrets in pairs 1 and 2 of the direct transmission treatment group were humanely euthanized and tissues were sampled to evaluate the viral titers and pathology. At 21 dpi, the remaining GD06-inoculated ferret showed seroconversion (HI titer, 1:640) (Table 3). The ferrets in direct-contact pairs 2 and 3 seroconverted (HI titer, 1:640 to 1:1,280). The naive ferret in pair 1 did not seroconvert (HI titer, 1:<10).
Viral titers in nasal wash fluids from ferrets. Ferrets were inoculated with 106 TCID50 of the GD06 virus (A, C) or the reassortant 109 virus (E, G). Naive ferrets were exposed either by direct contact with GD06-inoculated ferrets (B) or by aerosol contact with GD06-infected ferrets (D). (F, H) The same scheme described above was followed for reassortant 109, with naive ferrets being exposed by direct contact (F) or by aerosol exposure (H). At 3, 5, 7, and 10 dpi, nasal wash fluids were collected from ferrets and titrated in MDCK cells. The end titers are expressed as the number of log10 TCID50 per milliliter. The limit of virus detection was 100.699 TCID50/ml.
HI titers for serum samples from ferrets inoculated with IAV and from PBS-inoculated control ferrets in aerosol and direct-contact exposure experiments
To determine the aerosol transmission ability of wild-type virus GD06, we intranasally inoculated three ferrets. One day later, we placed three naive ferrets in cages adjacent to the infected ferrets. At 3 and 5 dpi, the infected ferrets shed virus at titers of 4.366 to 5.032 log10 TCID50/ml (Fig. 4C). Virus was not detected in the naive ferrets (Fig. 4D). At 21 dpi, all inoculated ferrets had seroconverted (Table 3). The HI titers ranged from 1:320 to 1:1,280 for the GD06-infected ferrets and from 1:640 to 1:1,280 for reassortant 109-infected ferrets. Serum collected from all naive ferrets was negative for the corresponding viruses (Table 3).
Reassortant 109 showed transmission ability through direct contact but not aerosol transmission.When we evaluated the transmissibility of reassortant 109, the inoculated ferrets shed virus at titers of 4.866 to 6.199 log10 TCID50/ml at 3 and 5 dpi (Fig. 4E), and the three naive exposed ferrets shed virus at titers of 0.699 to 6.032 log10 TCID50/ml (Fig. 4F). At 5 dpi, virus-inoculated ferrets in pairs 1 and 2 of the direct transmission group were humanely euthanized to evaluate the viral titers and tissue pathology. At 21 dpi, serum from the remaining inoculated ferret showed seroconversion (HI titer, 1:2,560) (Table 3), and all three naive direct-contact ferrets seroconverted (HI titers, 1:640 to 1:2,560). These results suggested that reassortant 109 maintained its ability to transmit by contact in ferrets.
When we evaluated the aerosol transmission ability of reassortant 109, we found that even though the infected ferrets shed virus at titers of 4.032 to 5.366 log10 TCID50/ml at 3 and 5 dpi (Fig. 4G), no virus was detected in naive ferrets (Fig. 4H). Furthermore, at 21 dpi, all inoculated ferrets had seroconverted and HI titers ranged from 1:640 to 1:1,280 for reassortant 109-infected ferrets (Table 3), whereas all serum samples collected from all naive ferrets were negative for the corresponding viruses (Table 3).
GD06 and reassortant 109 were pathogenic in ferrets.To detect the pathogenicity of GD06 and reassortant 109 in ferrets, we collected nasal turbinate, tracheal, and lung tissues from the humanely euthanized naive direct-contact ferrets for titration and histopathology. GD06 was detected only in nasal turbinate tissue (virus titer, 3.199 to 3.699 log10 TCID50/g) (Fig. 5). Reassortant 109 was detected in both nasal turbinate and tracheal tissues at titers of 3.032 to 5.032 log10 TCID50/g. No virus was detected from lung tissues. Ferrets inoculated with phosphate-buffered saline (PBS) showed no virus replication.
Viral titers from ferret tissues. Ferrets were inoculated with 106 TCID50 of GD06, reassortant 109, or PBS as a control. At 5 dpi they were humanely euthanized and nasal turbinate, tracheal, and lung tissues were collected. The tissues were homogenized and then titrated in MDCK cells. The limit of virus detection was 100.699 TCID50/ml. No virus was detected in the lungs. Reassortant 109 replicated in nasal turbinate and tracheal tissues, whereas GD06 was detected only in nasal turbinate tissue.
One ferret inoculated with GD06 and both ferrets inoculated with reassortant 109 had evidence of moderate lymphoplasmacellular rhinitis with a loss of cilia and multifocal replacement of the normal respiratory epithelium by stratified squamous epithelium (Fig. 6). The other wild-type GD06-inoculated ferret was minimally affected. No significant lesions were observed in the tracheas of any ferrets. In addition, one of the ferrets inoculated with reassortant 109 had mild focal lymphoplasmacellular peribronchitis and peribronchiolitis that extended slightly into the adjacent lung. Fever (temperature, >40°C) was not present in any of the ferrets, and there was minimal variation in body weight from the initial body weight (data not shown).
Histopathology of ferret nasal turbinate, tracheal, and lung tissues. Ferrets were inoculated with 106 TCID50 of GD06 virus, reassortant 109, or PBS as a control. At 5 dpi they were euthanized and nasal turbinate, tracheal, and lung tissues were collected. One GD06-infected ferret had moderate turbinate pathology (rhinitis with moderate lamina proprial lymphoplasmacellular infiltrates and a loss of cilia or replacement of the respiratory mucosal epithelium by stratified squamous epithelium), whereas the other had minimal pathology (minimal lamina proprial lymphoplasmacellular infiltrates). Both reassortant 109-infected ferrets had moderate to severe turbinate pathology, including lymphoplasmacellular rhinitis and a loss of cilia or replacement of the normal mucosal epithelium by stratified squamous epithelium. No significant lesions were seen in the trachea of any of the ferrets. There was mild lung pathology, including peribronchitis and peribronchiolitis, in one ferret infected with reassortant 109.
DISCUSSION
CIV-H3N2 has become enzootic within the canine population since it was first isolated in the early 2000s. It has also shown interspecies transmission to other companion animals, including cats and ferrets, suggesting that companion animals may serve an important role as intermediate hosts for IAVs (42–44). Serological surveillance indicates that individual dogs have been infected with both pdmH1N1 and CIV-H3N2 (23), and naturally occurring reassortants between the two viruses have been reported (35, 36). Others have demonstrated that a reassortant bearing the HA and NA genes from CIV-H3N2 and all other genes from pdmH1N1 replicated to high levels and produced significant pathological lesions in canine tracheal tissues (12). Additional study of CIV-H3N2 reassortants in the dog may be elucidative; however, there is a current halt on gain-of-function experiments by the U.S. Government.
In our study, antigenic analysis showed that GD06 only weakly cross-reacts with ferret antiserum raised against two of the representative human H3N2 IAVs tested, and cross-reactivity was limited to A/Sydney/5/1997 (H3N2) and A/Nanchang/933/1995 (H3N2). Our results suggest that contemporary seasonal influenza viruses do not cross-react with the novel H3N2 canine influenza viruses. Of course, repeat exposures to antigenically drifted H3N2 viruses may broaden the serological response in humans, and human serological studies are required to confirm herd immunity to GD06-like viruses.
To assess the potential risk of reassortment, we attempted to generate all GD06 × CA09 reassortants with the HA gene of GD06. We were able to successfully generate 51 (40%) of 127 reassortants, suggesting that the gene segments from GD06 and CA09 have moderate genomic compatibility. Consistent with this finding, RNP luciferase reporter assays suggested that specific constellations of GD06 and CA09 genes may enhance the level of viral replication compared to that of the wild-type parental viruses. All 4 RNP complex constellations that showed enhanced luciferase activity contained the NP gene segment from GD06 and the PB1 gene segment from CA09.
All 19 high-growth reassortant viruses contained the PB1 gene segment from CA09, and most (84%) also contained the CA09 PB2 gene segment. However, high in vitro growth did not uniformly translate to efficient replication in mice. Others have described the important role of the PA gene segment in maintaining the stability of CA09 (45). However, our RNP data and our reassortants with high-growth phenotypes showed fairly equal representation of the PA gene segment from GD06 and CA09.
While not all viruses replicated robustly in mice, all but three inoculated mice seroconverted at 14 dpi, suggesting some level of growth. This finding is in contrast to that of a previous study that showed that a CIV-H3N2 isolate was not detected in the respiratory tract of mice after experimental infection and animals remained seronegative (46). The discrepancies in these data may lie in minor variations in the representative viruses used [the previous researchers studied A/Canine/Korea/01/2007 (H3N2)]. Most of the reassortants that replicated efficiently in mouse lungs contained more gene segments from CA09 than from GD06. The maximum replication in mouse lungs by a reassortant was achieved by reassortant 109, which contained only the HA and NP genes from GD06. Overall, there was minimal weight loss in most of the infected mice. The introduction of CA09 gene segments may account for the >5% body weight loss observed in mice inoculated with reassortants 111, 109, and 55. All three of these reassortants consisted of an identical RNA polymerase complex of PB2 of CA09, PB1 of CA09, and PA of CA09. However, other reassortants with the same RNA polymerase complex did not cause a weight loss of >5%, demonstrating that other factors are also important.
Compared to the replication of GD06, reassortant 109 showed enhanced replication in in vitro and in vivo experiments and maintained at least the same level of contact transmissibility between ferrets. In another study, CIV-H3N2 replicated efficiently in the upper and lower respiratory tracts of ferrets, and the animals exhibited clinical signs of disease, such as lethargy, sneezing, unkempt fur, and a loss of appetite (47). In our study, the ferrets showed no clinical signs of illness, possibly related to the lower dose of virus used. In our hands, neither GD06 nor reassortant 109 was detected in ferret lung tissues, although mild lung pathology was seen in one ferret infected with reassortant 109. Both viruses effectively replicated in nasal turbinate tissues, and reassortant 109 also replicated in tracheal tissues.
In our study, ferrets infected with GD06 or reassortant 109 at 106 TCID50 shed virus at titers of 4.032 to 6.199 log10 TCID50/ml at 3 or 5 dpi. These titers are higher than those reported previously in ferrets inoculated with A/canine/Korea/01/2007 (H3N2) at 103.5 50% egg infective doses (EID50) (46). In another study, ferrets with direct exposure to inoculated ferrets shed detectable virus at a mean titer of 2.5 to 2.8 log10 EID50/ml from 10 to 11 dpi, and 2 of 3 exposed ferrets seroconverted (42). In our study, naive direct-exposure ferrets shed virus at titers of 0.699 to 5.366 log10 TCID50/ml at 3, 5, and 7 dpi. Furthermore, all ferrets in direct contact with those inoculated with reassortant 109 seroconverted at 21 dpi. These findings suggest that, in ferrets, CIV-H3N2 and reassortant 109 achieved a relatively high replication efficiency and the ability to be transmitted via direct contact.
In conclusion, reassortants between GD06 and CA09 showed increased viral titers in the lungs of mice and caused more pathological lesions in ferrets than wild-type strain GD06 did. These reassortants warrant the use of enhanced surveillance for CIV in the canine population, especially surveillance for reassortant viruses with zoonotic potential.
MATERIALS AND METHODS
Cells.Madin-Darby canine kidney (MDCK) cells (ATCC CCL-34) and human embryonic kidney epithelial (293T) cells (ATCC CRL-11268) were maintained in Dulbecco's modified Eagle medium (Gibco/BRL) supplemented with 10% fetal bovine serum (Atlanta Biologicals) and penicillin-streptomycin (100 U/ml and 100 μg/ml, respectively; Gibco/BRL) at 37°C with 5% CO2. Human lung adenocarcinoma epithelial (A549) cells (ATCC CCL-185) were maintained in Kaighn's modification of Ham's F-12 (F-12K) medium (ATCC) supplemented with 10% fetal bovine serum and penicillin-streptomycin (100 U/ml and 100 μg/ml, respectively) at 37°C with 5% CO2.
Viruses.We propagated CIV strain A/canine/Guangdong/1/2006 (H3N2), here referred to as GD06, the first H3N2 isolate identified from dogs in southern China (19), in 9- to 11-day-old specific-pathogen-free (SPF) embryonated chicken eggs (Sunrise Farms) and MDCK cells. The pandemic H1N1 strain used for experiments was A/California/4/2009 (H1N1), here referred to as CA09. We also used A/Puerto Rico/8/1934 (H1N1) (PR8).
Mice and ferrets.All animal experiments were performed in Mississippi State University (MSU) AAALAC-accredited facilities and in compliance with MSU Institutional Animal Care and Use Committee (IACUC) and Institutional Biosafety Committee (IBC) protocols. We obtained 4-month-old female SPF ferrets from Triple F Farms and 6- to 8-week-old female SPF BALB/c mice from Harlan Laboratories. All animals were maintained in individually ventilated cages in compatible groups (mice) or in pairs that were individually separated by a partition (ferrets). All ferrets were tested and determined to be free from influenza viral antibodies before initiation of the study. Water and food were available ad libitum.
Ferret antiserum.We generated ferret antiserum against temporally representative human H3N2 IAVs from 1979 to 2015 as previously described (48) for use in evaluating the antigenic similarity between CIV-H3N2 and contemporary human H3N2 IAVs. Ferret antiserum was raised against A/Bangkok/1/1979 (H3N2), A/Philippines/2/82 (H3N2), A/Caen/1/1984 (H3N2), A/Mississippi/1/1985 (H3N2), A/Leningrad/360/1986 (H3N2), A/Sichuan/02/1987 (H3N2), A/Sichuan/60/1989 (H3N2), A/Ann Arbor/03/1993 (H3N2), A/Johannesburg/33/1994 (H3N2), A/Nanchang/933/1995 (H3N2), A/Sydney/5/1997 (H3N2), A/Wisconsin/67/2005 (H3N2), A/Brisbane/10/2007 (H3N2), A/Perth/16/2009 (H3N2), A/Victoria/361/2011 (H3N2), A/Texas/50/2012 (H3N2), A/Switzerland/9715293/2013 (H3N2), A/Utah/07/2013 (H3N2), A/Mississippi/17/2013 (H3N2), A/Costa Rica/4700/2013 (H3N2), A/Palau/6759/2014 (H3N2), A/Hong Kong/4801/2014 (H3N2), A/Fiji/2/2015 (H3N2), A/Brisbane/82/2015 (H3N2), and A/Victoria/503/2015 (H3N2). We also generated ferret antiserum against CIV-H3N8, specifically, A/canine/Iowa/13628/2005 (H3N8).
RNA isolation, PCR, cloning, and plasmid extraction.GD06 viral RNA was isolated using an RNeasy minikit (Qiagen) and amplified using SuperScript one-step reverse transcription-PCR with Platinum Taq DNA polymerase (Invitrogen), and amplicons were cloned into a dual-promoter plasmid vector, pHW2000, as previously described (49, 50). All plasmids were amplified using a GeneJET plasmid midiprep kit (Thermo Fisher Scientific) according to the manufacturer's protocol, and the sequences were confirmed via Sanger sequencing at the Life Sciences Core Laboratories Center (Cornell University, Ithaca, NY, USA).
Generation of reassortant viruses using reverse genetics.We used reverse genetics to generate reassortant viruses with all possible genomic constellations between GD06 and CA09, with the HA gene originating only from wild-type strain GD06 (i.e., 27 − 1, or 127, reassortants, all of which had the HA gene from GD06). We generated the reassortant viruses using transfection and reverse genetics as previously described (51). In brief, 1 μg of each of the eight plasmids expressing the specified genomic constellation was added to Opti-MEM medium (Gibco/BRL) and 16 μl of TransIT-LT1 transfection reagent (Mirus Bio) and mixed gently; the mixture was incubated at room temperature for 45 min. Opti-MEM medium (800 μl) was then added to the mixture, which was then transferred to cocultured MDCK and 293T cells. The transfection medium was removed from the cells 12 h after transfection and replaced with Opti-MEM medium supplemented with tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-treated trypsin from bovine pancreas (TPCK-trypsin; Sigma-Aldrich). At 3 days after transfection, the supernatant was seeded once into MDCK cells. Approximately 72 h later, 50 μl of passage 1 virus was used in a hemagglutination assay with 0.5% turkey erythrocytes. Of the 127 reassortants, 51 showed hemagglutination and were collected and stored at −80°C for additional characterization. All others that we failed to generate were regenerated from the beginning, i.e., transfection. However, no additional reassortants were generated. Titers for the 51 reassortant viruses were determined by using endpoint titration in MDCK cells, and the TCID50 was calculated by using the Reed-Muench method (52). The limit of virus detection was 100.699 TCID50/ml. All viruses with a predetermined cutoff of ≥106 TCID50/ml were selected for further characterization of their pathogenesis in mice. Illumina sequencing confirmed the full gene segments of each virus, and genomic sequencing and assembly are described elsewhere (53).
Growth curve replication kinetics in vitro.To determine virus replication kinetics in vitro, we quantified and infected MDCK and A549 cells at a multiplicity of infection (MOI) of 0.001. After incubation at 37°C for 1 h, cells were washed twice with phosphate-buffered saline (PBS) and incubated in Opti-MEM medium containing TPCK-trypsin at 1 μg/ml (MDCK cells) or 0.5 μg/ml (A549 cells) at 37°C with 5% CO2. Supernatants were collected at 12, 24, 36, 48, and 72 h postinoculation and stored at −80°C. The supernatants were titrated in MDCK cells by TCID50, and TCID50 were calculated by using the Reed-Muench method (52). Titers are shown as the means for the three replicate infections. The limit of virus detection was 100.699 TCID50/ml.
Hemagglutination and HI assays.To detect the virus titer and to determine seroconversion, we performed the hemagglutination assay and hemagglutination inhibition (HI) assay using 0.5% turkey erythrocytes as described elsewhere (41). All HI assays were performed in three replicates.
Luciferase assay to quantify ribonucleoprotein complex activity in vitro.A total of 4 × 104 293T cells were transfected with polymerase PB2, PB1, PA, and NP protein expression plasmids (40 ng) in Corning 96-well plates. Forty nanograms of plasmid phPOLI-RLUC expressing Renilla luciferase and 4 ng of pGL4.13 (luc2/simian virus 40) expressing firefly luciferase (internal control) (Promega, Madison, WI) were also cotransfected in 293T cells. Luciferase activities in lysates from cells harvested at 48 h after transfection were measured using a dual-luciferase reporter assay system (Promega, Madison, WI) per the manufacturer's instruction. We measured the Renilla and firefly luciferase activities for the 16 possible combinations of plasmids expressing the RNP complexes derived from CA09 and GD06, determining the replication efficiency of each combination. The ratio of Renilla/firefly luciferase activities for each RNP combination was normalized to the ratio of the Renilla/firefly luciferase activity for the internal control.
Statistical analysis of luciferase assay.Luciferase data were expressed as means ± standard deviations from three independent experiments. The differences in means were tested using a one-way ANOVA with post hoc Tukey's multiple-comparison test using GraphPad Prism (version 7.02) software (GraphPad Software, Inc., La Jolla, CA). P values of <0.05 were considered statistically significant.
Pathogenesis in mice.We characterized the pathogenesis of the reassortant viruses by inoculating mice with 22 viruses, including 19 reassortants between GD06 and CA09 with the high-growth phenotype (Table 1), one GD06 × PR8 reassortant with the HA and NA gene segments derived from GD06, and the wild-type strains GD06 and CA09. Briefly, we anesthetized 22 groups of mice (n = 8 mice per group) with isoflurane and then intranasally inoculated them with 50 μl of an IAV at a viral load of 106 TCID50/ml or with 50 μl of sterile PBS. All mice were weighed daily, and at 4 days postinoculation (dpi), we humanely euthanized three mice in each group and collected lung tissues under sterile conditions. The lung tissues were stored at −80°C until virus titers could be obtained. We used a pestle (Research Products International) and silicon dioxide (Acros Organics/Thermo Fisher Scientific) to homogenize the lung tissues in 500 μl of sterile PBS and then titrated the supernatants in MDCK cells. The TCID50 was determined by using the Reed-Muench method (52). The limit of virus detection was 100.699 TCID50/ml. The remaining five mice in each group were monitored for weight loss until 14 dpi and then humanely euthanized. Serum was collected from the mice to determine seroconversion. The reassortant virus with the highest lung titer (i.e., reassortant 109) was selected for a transmission and pathogenesis study in ferrets.
Statistical analysis of mouse viral lung titers.Mean viral lung titers were tested using a one-way ANOVA with post hoc Tukey's multiple-comparison test using GraphPad Prism (version 7.02) software (GraphPad Software, Inc., La Jolla, CA). P values of <0.05 were considered statistically significant.
Transmission and pathogenesis in ferrets.To determine the transmissibility and pathogenesis of GD06 and reassortant 109 in ferrets, 4-month-old female ferrets were purchased from Triple F Farms and given 7 days to acclimatize after arrival. Each ferret was weighed, and its temperature was monitored throughout the experiment using implanted transponders (Bio Medic Data Systems, Inc.). Before the experiments were conducted, all 24 ferrets tested seronegative for antibodies against GD06, CA09, A/Perth/16/2009 (H3N2), A/Victoria/361/2011 (H3N2), and A/Minnesota/307875/2012 (H3N2) influenza viruses. Each experimental group included three ferrets that were lightly anesthetized with isoflurane and inoculated intranasally with either GD06 or reassortant 109 (viral load, 106 TCID50 in a 1-ml volume) and three ferrets for evaluation of susceptibility to exposure that were not anesthetized or inoculated (exposure ferrets). The exposure ferrets were exposed to the virus through either ferret-to-ferret direct contact or indirect (i.e., aerosol) contact with the virus-inoculated ferret.
In each of the three direct-contact transmission groups, one virus-inoculated ferret and one exposure ferret were housed as a pair in the same cage without a partition. In each of the three aerosol-transmission groups, one virus-inoculated ferret and one exposure ferret were housed as a pair in the same cage but were separated by a 1-cm-thick, double-layered steel partition with 5-mm perforations (Allentown, Inc.). In all cages, the exposure ferret was placed into the cage 1 day after the virus-inoculated ferret was introduced into the cage. The nonrecirculating airflow in the cage went from the exposure ferret through the partition to the virus-inoculated ferret and exhausted to room air through HEPA filtration.
Ferrets were lightly anesthetized and induced to sneeze in order to collect nasal wash fluids at 3, 5, 7, and 10 dpi to determine viral shedding patterns. Briefly, ferrets were induced to sneeze by inoculating 1 ml of sterile PBS and gently tickling the nasal cavity with a sterile cotton swab. Before performing nasal washes, we measured each ferret's body temperature and weight. We monitored clinical signs daily. The virus titers in the samples were determined by titration in MDCK cells. To evaluate the replication efficiency and the pathology of the viruses in the ferret respiratory tract, we euthanized two of the three virus-inoculated ferrets in each direct-contact-transmission group at 5 dpi. The turbinates, trachea, bronchi, and lungs were collected, and virus titers were determined (limit of detection, 100.699 TCID50/ml). For the remainder of the ferrets, serum was collected at 14 dpi, immediately before they were euthanized.
Ethics statement.Experiments involving mice and ferrets were conducted in compliance with protocols approved by the Institutional Biosafety Committee and the Institutional Animal Care and Use Committee of Mississippi State University (protocols IBC#011-12 and IACUC#13-022) and under the regulations of the Animal Welfare Act (AWA). Viruses were propagated in 9- to 11-day-old SPF embryonated chicken eggs (Sunrise Farms). Laboratory experiments were conducted under biosafety level 2 conditions, with investigators wearing appropriate personal protective equipment. All reassortant viruses were generated before the U.S. Government pause on gain-of-function research.
ACKNOWLEDGMENTS
We are grateful to Elizabeth Bailey and Lucas Ferguson for their assistance with the animal experiments, Lei Li for his assistance with sequence analyses, and Lei Zhong for her assistance with serological assays.
This study was partially supported by the National Institutes of Health (grant number 1R15AI107702).
FOOTNOTES
- Received 18 April 2017.
- Accepted 4 August 2017.
- Accepted manuscript posted online 16 August 2017.
Supplemental material for this article may be found at https://doi.org/10.1128/JVI.00637-17.
- Copyright © 2017 American Society for Microbiology.
