ABSTRACT
Influenza A viruses (IAV) sporadically transmit from swine to humans, typically associated with agricultural fairs in the United States. A human seasonal H3 virus from the 2010-2011 IAV season was introduced into the U.S. swine population and termed H3.2010.1 to differentiate it from the previous swine H3 virus. This H3N2 lineage became widespread in the U.S. commercial swine population, subsequently spilling over into exhibition swine, and caused a majority of H3N2 variant (H3N2v) cases in humans in 2016 and 2017. A cluster of human H3N2v cases were reported at an agricultural fair in 2017 in Ohio, where 2010.1 H3N2 IAV was concurrently detected in exhibition swine. Genomic analysis showed that the swine and human isolates were nearly identical. In this study, we evaluated the propensity of a 2010.1 H3N2 IAV (A/swine/Ohio/A01354299/2017 [sw/OH/2017]) isolated from a pig in the agricultural fair outbreak to replicate in ferrets and transmit from swine to ferret. sw/OH/2017 displayed robust replication in the ferret respiratory tract, causing slight fever and moderate weight loss. Further, sw/OH/2017 was capable of efficient respiratory droplet transmission from infected pigs to contact ferrets. These findings establish a model for evaluating the propensity of swine IAV to transmit from pig to ferret as a measure of risk to the human population. The identification of higher-risk swine strains can then be targeted for control measures to limit the dissemination at human-swine interfaces to reduce the risk of zoonotic infections and to inform pandemic planning.
IMPORTANCE A recently emerged lineage of human-like H3N2 (H3.2010.1) influenza A virus (IAV) from swine has been frequently detected in commercial and exhibition swine in recent years and has been associated with H3N2 variant cases in humans from 2016 and 2017. To demonstrate a model for characterizing the potential for zoonotic transmission associated with swine IAV, we performed an in vivo study of transmission between pigs infected with an H3.2010.1 H3N2 IAV and aerosol contact ferrets. The efficient interspecies transmission demonstrated for the H3.2010.1 IAV in swine emphasizes the need for further characterization of viruses circulating at the swine-human interface for transmission potential prior to human spillover and the development and implementation of more robust vaccines and control strategies to mitigate human exposure to higher-risk swine strains.
INTRODUCTION
The population of influenza A viruses (IAV) in swine in the United States is genetically and antigenically diverse. Three subtypes of IAV, H1N1, H1N2, and H3N2, currently circulate in the U.S. swine population that can be further classified into multiple evolutionary lineages and genotypes (1, 2). H3N2 first established in the U.S. swine population circa 1998, with an additional introduction of an N2 neuraminidase (NA) gene segment in 2002 from human seasonal IAV (3–5). The triple-reassortant internal gene cassette (TRIG) was introduced into U.S. swine concurrently with the 1998 human H3N2 glycoproteins (3, 6, 7). Shortly following the emergence of the 2009 H1N1 pandemic (H1N1pdm09), reverse zoonosis of this virus reintroduced the pandemic internal genes into the swine population, in which the pandemic matrix (M) gene segment supplanted the TRIG lineage as the predominant M gene segment (8, 9). Continued, sporadic incursions of H1N1pdm09 resulted in sustained circulation of the additional pandemic internal gene segments, particularly polymerase acidic (PA) and nucleoprotein (NP) (10). In 2012, IAV in swine (IAV-S) were detected with the hemagglutinin (HA) gene segment from a contemporary human seasonal H3N2 virus closely related to strains from the 2010-2011 influenza season (11). Since their introduction, 2010.1 H3N2 viruses evolved into a major lineage of IAV in swine, becoming the predominant H3 clade in 2015 to 2016 (12). The 2010.1 human-like H3N2 viruses were also antigenically distinct from H3N2 clade IV IAV (11, 13) and not contained in licensed swine vaccines; thus, concerns existed over vaccine efficacy and control of this novel IAV lineage in swine.
Swine lineage viruses detected in humans are designated “variant” influenza viruses to differentiate them from human seasonal IAV. Human variant infections with swine lineage H1N1, H1N2, and H3N2, most frequently detected in association with swine exhibitions at agricultural fairs, have been reported since 2005, with H3N2v being the most frequently detected (14, 15). Exhibition swine are a distinct subgroup of pigs, largely separated from the U.S. commercial swine population, that come into contact with a section of the human population different from swine workers with occupational exposure. Exhibition swine are frequently in contact with children, adolescents, and non-occupationally exposed members of the public, while amplifying genetically diverse IAV strains across geographic distances and creating a unique swine-human interface at agricultural fairs in the United States (16–18). To date, 11 genotypes of IAV in exhibition swine have been identified, with the most common being composed of TRIG lineage PB2, PB1, PA, NP, and NS segments, a pandemic M segment, and a 2002 lineage NA segment. Incorporation of the pandemic M segment into the TRIG background was correlated with the detection of H3N2v cases, suggesting that the H3 genotype may have been uniquely adapted to an expanded host range (19, 20). From 2011 to 2017, 426 cases of H3N2v were reported, including 26 hospitalizations and 1 death (14). From July to August 2016, active surveillance at agricultural fairs in Michigan and Ohio identified clade IV and 2010.1 H3N2 viruses in swine that were identical to H3N2v isolated from patients that were in attendance (22). Clade IV and 2010.1 H3N2v viruses have been shown to replicate to high titers in cultured, polarized human airway epithelial cells and transmit readily through respiratory droplets in ferrets, a common model for IAV infection and transmission in humans (23, 24). Further, some portions of the human population have limited immunity against clade IV H3 IAV-S and vaccination with seasonal trivalent inactivated vaccine does not induce neutralizing antibodies against clade IV H3 in humans, nor does it protect ferrets from infection (25–27). Immunity to 2010.1 swine H3N2 virus may differ due to the more recent common human seasonal ancestor.
In 2017, an agricultural fair in Ohio reported pigs with influenza-like illness in the swine barn. Despite efforts by animal health officials to control the outbreak, the county health department subsequently received reports of people with influenza-like illness that had been exhibitors or visitors to the swine barn. Concurrent sampling and subsequent testing of pigs and people exhibiting influenza-like illness found the porcine and human samples to be positive for IAV. There were 11 human cases associated with the swine exhibition outbreak (28). Ten of the cases were in children, and all infected individuals recovered without hospitalization. No subsequent human-to-human transmission was documented during this outbreak. IAV prevalence among the pigs was high: approximately 80% of the pigs showed clinical signs of respiratory illness, and 70% of randomly selected swine (14/20) tested positive using previously described methods (16, 20). Subsequent genetic analysis of the virus isolates identified the viruses as gamma cluster H1N1 and H3N2 IAVs of swine origin. One swine virus, A(H3N2)/swine/Ohio/A01354299/2017, herein referred to as sw/OH/2017, showed high genetic similarity to the human isolates and harbored a unique mutation in the HA1 amino acid sequence, N145, which is known to be a major antigenic determinant of swine H3N2 (29). To establish a zoonotic transmission model and the potential of the swine 2010.1 H3N2 IAV to transmit from swine and cause disease in humans, we assessed the ability of the swine isolate sw/OH/17 to infect ferrets, a model of human IAV infection, via direct inoculation as well as transmission to ferrets following aerosol contact with inoculated pigs, the initial step in zoonotic transmission of IAV-S.
RESULTS
Genetic and phylogenetic analyses of swine H3N2 and H3N2v viruses.To understand the relationship of H3N2v viruses to sw/OH/2017 and other swine viruses collected from agricultural fair sampling, we inferred the best-known maximum likelihood tree for the hemagglutinin (HA) and neuraminidase (NA) gene segments. The HA tree demonstrated a statistically supported monophyletic clade of 2010.1 HA genes collected from agricultural fairs in 2017: within this clade were 9 H3N2v HA genes, including the 7 HA genes collected at the same fair as the sw/OH/2017 isolate used in this study (Fig. 1A). The HA segments of viruses from this clade were most closely related to the 2010.1 H3 lineage, sharing a common ancestor with genes isolated in 2016 from multiple states in the Midwestern United States, including Iowa, Minnesota, Missouri, and Ohio. A second statistically supported monophyletic clade, with HA genes from 2016 agricultural fairs and U.S. Department of Agriculture (USDA) swine IAV surveillance data collected from commercial swine, was also evident. The NA genes collected at agricultural fairs formed a monophyletic clade nested within the swine N2-2002 lineage that was introduced into commercial swine in the United States in approximately 2002, but the 2017 NA data were in a different clade than the 2016 2010.1 swine and H3N2v (Fig. 1B). Closely related IAV NA genes were isolated in 2017 from commercial swine in Illinois and Indiana collected as part of the passive USDA national surveillance swine IAV system. The internal gene constellations of the sw/OH/2017 and all the 2017 H3N2v viruses were TRIG lineage segments (PB2, PB1, PA, NP, and NS) except for the M segments, which were of H1N1 2009 pandemic origin.
Phylogenetic relationships of H3N2 IAV collected from exhibition swine at agricultural fairs and H3N2v from 2016 to 2017. Maximum likelihood phylogenies of HA (A) and NA (B) were inferred and included 19 swine IAV and 29 H3N2v viruses (n = 26 collected between 2016 and 2017) associated with swine exhibitions at agricultural fairs. A/swine/Ohio/A01354299/2017 (sw/OH/2017) is indicated by a red square, and all H3N2v branches are colored red. HA and NA genes collected from 2017 agricultural fairs (cyan) and 2016 agricultural fairs (yellow) are included with a random sample of contemporary commercial swine IAV (2010.1 “human-like” swine H3 HA in blue, clade IV swine HA in black; 2002 N2 NA in blue, 1998 N2 NA in black). Human seasonal H3 or N2 IAV (gray) is included for context. The trees are midpoint rooted for clarity, branch lengths are drawn to scale, the scale bar indicates the number of nucleotide substitutions per site, and statistical support (SH-aLRT) is presented for branches of interest. Trees with tip labels and statistical support are provided in Fig. S1 and S2 in the supplemental material.
To further understand the similarity between sw/OH/2017 and the seven human H3N2v viruses collected in Ohio, we compared the nucleotide and amino acid sequence identities (Table 1). All H3N2v viruses showed nearly 100% identity with swine isolate sw/OH/2017 at the nucleotide and amino acid levels. The exception was the NP from three H3N2v viruses, A/Ohio/13/2017, A/Ohio/14/2017, and A/Ohio/23/2017, that differed at the nucleotide level (99.9%) encoding two point mutations, A178T (A/Ohio/13/2017 and A/Ohio/14/2017) and Y52H (A/Ohio/22/2017). These mutations have not previously been associated with enhanced replication in humans or ferrets. Further molecular characterization showed that the PB2 segment from sw/OH/2017 and H3N2v lacked previously identified markers for enhanced replication in humans, encoding glutamate and aspartate at positions 627 and 701, respectively (30, 31). However, all PB2 segments from swine isolates collected at agricultural fairs in Ohio in 2017 and associated human variant viruses encoded 271A, 590S, 591R, and 661A, which were reported to overcome the phenotype mediated by 627E and 701D (32, 33). Further, the PA genes encoded 669V, a marker shown to enhance the polymerase activity of other zoonotic IAV (34). The receptor binding sites of the sw/OH/2017 and H3N2v HAs encoded 190D, 226I, and 228S, residues known to confer a binding preference for α2,6 sialic acids (35, 36). Together these results suggest the 2010.1 lineage H3N2 found in the exhibition swine population in 2017 contained viruses that have the genotypic markers characterized for potential zoonotic IAV, i.e., for optimized polymerase activity and α2,6 sialic acid receptor binding preference.
Percent identity of A(H3N2)/swine/Ohio/A01354299/2017 compared to 2017 H3N2v isolates from Ohio
Replication and pathogenesis of sw/OH/2017 in ferrets.To better understand the potential of sw/OH/2017 to replicate and cause disease in humans, we intranasally inoculated four ferrets with 106 50% tissue culture infective doses (TCID50) of sw/OH/2017. sw/OH/2017 replicated efficiently in the ferret upper respiratory tract, with high peak virus titers (mean = 6.8 log TCID50/ml) and detection in nasal wash samples until study end at 5 days postinoculation (dpi). Moderate virus titers were detected in the bronchoalveolar lavage fluid (BALF) at 5 dpi (mean = 3.8 log TCID50/ml), indicating that sw/OH/2017 replicated in the upper and lower respiratory tracts of inoculated ferrets (Fig. 2A). Weight loss was observed in all animals beginning at 1 dpi and peaking at 3 dpi (Fig. 2B). All sw/OH/2017-inoculated animals exhibited elevated body temperature compared to baseline, beginning at 1 dpi (increased by 0.3 to 1.8°C), peaking at 2 dpi (increased by 1.5 to 3.9°C), and remaining slightly elevated through the end of the study at 5 dpi (Fig. 2C). Significant increases in temperature or decreases in body weight were not observed in uninfected control ferrets over the duration of the study (data not shown).
Replication of A/swine/Ohio/A01354299/2017 in ferrets. Four ferrets were inoculated intranasally with 106 TCID50 of sw/OH/2017. Nasal wash samples were collected at 1, 3, and 5 dpi and bronchoalveolar lavage (BALF) was collected at 5 dpi and titrated in MDCK cells via TCID50 assay (A). The limit of detection was 101 TCID50/ml and is indicated by a dotted line. Clinical signs were assessed by weight loss (B) and increased body temperature (C) compared to baseline calculated from values collected at −2 to 0 dpi.
Infection of sw/OH/2017 in the lungs of ferrets caused moderate lung lesions typical of IAV infection (Table 2). Lung lesions characterized by areas of consolidation were observed in inoculated ferrets (mean of 3.6% macroscopic pneumonia) but absent in uninfected, control animals. Lung sections taken from affected lobes had a mean composite microscopic lesion score of 11.3 out of 22 in inoculated ferrets. Microscopic lung lesions observed in ferrets challenged with sw/OH/2017 included multifocal bronchitis with severe necrosis of affected bronchi and bronchiolar epithelium and large numbers of neutrophils infiltrating the airway lumen (Fig. 3B). The regions immediately surrounding affected airways were expanded by abundant lymphocytes, macrophages, and neutrophils (Fig. 3B) with occasional peribronchiolar edema, which was absent in the lungs from nonchallenged ferrets (Fig. 3A). Moderate numbers of macrophages and neutrophils extended into adjacent alveolar lumens with mild infiltrates of similar inflammation in the interstitium. There was abundant perivascular edema observed in affected lungs (Fig. 3D) compared to that in nonchallenged control ferrets (Fig. 3C). Submucosal glands were often necrotic and effaced by abundant neutrophils and necrotic inflammatory debris (Fig. 3F), in contrast to normal submucosal glands observed in unchallenged controls (Fig. 3E). The composite pathology score from trachea sections was low (1.3 ± 0.4), but the sections were characterized as having neutrophil infiltration, epithelial cell cilium loss, and mild necrosis (Table 2). Viral antigen was not detected by immunohistochemistry (IHC) in the lung or trachea sections of challenged ferrets.
Lung pathology induced by A/swine/Ohio/A01534299/2017 H3N2 infection in ferrets
Formalin-fixed ferret lung tissue sections were collected at 5 dpi, stained with hematoxylin and eosin, and evaluated for pathology. (A) Nonchallenged ferrets demonstrated normal bronchiolar epithelium and adjacent alveoli; (B) severe bronchiolar epithelial necrosis with abundant neutrophils and marked peribronchiolar lymphocytic cuffing was observed in ferrets challenged with A/swine/Ohio/A01354299/2017; (C) normal blood vessel from a nonchallenged ferret; (D) severe perivascular edema and hemorrhage admixed with large numbers of lymphocytes and neutrophils; (E) normal submucosal glands; (F) affected submucosal glands are necrotic and effaced by large numbers of neutrophils and cellular debris. Magnifications for all panels, ×200.
Swine-to-ferret transmission.To model the swine-human interface present at agricultural fairs and potential for interspecies aerosol transmission, we housed ferrets in individual isolators adjacent to a standard, raised-deck pig enclosure. This configuration barred physical contact and prohibited transfer of potential fomites between species (Fig. 4). To ensure that ferrets were not exposed to virus during inoculation of pigs, the doors on the HEPA-filtered isolators remained in place until 24 h after inoculation of the pigs. The solid doors were then removed to facilitate respiratory droplet exposure at 0 days postcontact (dpc) through the wire cage doors that remained on each isolator. Replication of sw/OH/2017 in the upper respiratory tracts of pigs was similar to that observed with other 2010.1 H3N2 viruses (11), with the highest mean titer at 1 dpi (mean = 4.6 log TCID50/ml), and shedding persisted through 5 dpi (3.1 TCID50/ml). Respiratory transmission of sw/OH/2017 was highly efficient, with virus being detected in the nasal wash fluid from all four contact ferrets beginning at 2 dpc (Fig. 5A). Exposure to infectious porcine respiratory droplets resulted in high levels of virus shedding in the nasal wash fluid from contact ferrets at 2 dpc (mean = 6.1 log TCID50/ml) until 6 dpc (mean = 4.3 log TCID50/ml). Virus was no longer detected in nasal wash samples collected at 8 dpc. Clinical signs of disease were similar between the aerosol contact ferrets and those seen in the primary inoculated ferrets, which included nasal discharge, lethargy, coat roughness, contact avoidance, inappetence, and reduced water intake. Weight loss was observed in contact ferrets beginning at 2 dpc and peaking at 8 dpc, with a full recovery of body weight being observed at 14 dpc (Fig. 5B). An increase in mean body temperature was also observed beginning at 2 dpc (Fig. 5C). All contact ferrets seroconverted by 14 dpc (reciprocal geometric mean titer = 538 [data not shown]).
Experimental design for assessment of swine-to-ferret transmission. Five pigs were housed in a standard, raised pig deck and inoculated intranasally with 2 ml of 106 TCID50 A/swine/Ohio/A01354299/2017. Four naive ferrets, individually housed in isolator cages, were placed adjacent to the pig deck with approximately 7.5 cm of separation between the front of the ferret cages and the pig deck. The impermeable outer doors of the isolators were removed on day 1 postinfection to facilitate exposure of ferrets to infectious porcine respiratory droplets.
Transmission of sw/OH/2017 from inoculated pigs to contact ferrets. Virus replication was assessed via collection of nasal swabs and nasal washes for pigs and ferrets, respectively. Samples were titrated on MDCK cells by TCID50 assay (A). The limit of detection was 101 TCID50/ml and is indicated by a dotted line. Ferrets were monitored for changes in body weight (B) and body temperature (C) compared to baseline calculated from values collected at −2 to 0 dpi, until termination of the study from 14 dpc.
Whole-genome sequencing (WGS) of respiratory samples found no mutations known to confer enhanced pathology or replication in ferrets between viruses from donor pigs and contact ferrets (Table 2; see also Fig. S1 and S2 in the supplemental material). Two amino acid mutations (I61V [A181G] in NP and E433K [G1297A] in NA) were detected in the whole-genome sequence of the 6-dpc nasal wash of one ferret. Three individual mutations in the HA1 were detected in porcine 5-dpi nasal swab samples. Two mutations were detected in samples from two different pigs in known antigenic regions, N145K (C435A) and S193Y (C578A), and one mutation adjacent to the HA receptor binding site (G686T) was also identified in a third pig. Together, these results demonstrate that 2010.1 H3N2 IAV-S currently circulating in U.S. swine were highly efficient in replicating and shedding in the upper respiratory tracts of pigs and ferrets and were also transmitted rapidly to aerosol-contact ferrets through infectious respiratory droplets from pigs.
DISCUSSION
The continued detection of variant influenza cases highlights the need for vigilance at the swine-human interface. Epidemiological studies of swine farmers, slaughterhouse workers, and veterinarians demonstrated serological evidence for interspecies transmission, finding those with frequent, prolonged contact with infected swine to be at greater risk (37–39). Agricultural fairs present a unique swine-human interface bringing together exhibition swine and humans from a large geographical area, often in close proximity to each other. Exhibition swine undergo extensive movement, potentially being exposed to IAV from multiple sources and resulting in a diverse population of IAV comprised of many different genotypes and lineages (17, 40). Further, sampling of exhibition swine and swine barns has found IAV to be consistently detected in clinical and environmental samples (41, 42), underlining the presence of multiple routes for interspecies transmission. Preventing the entry of infected pigs to agricultural fairs has proven challenging, as infected pigs often do not display clinical signs (elevated body temperature or influenza-like illness) (16, 43). Analysis of variant influenza viruses has shown H1N1v and H3N2v to be identical to IAV isolated from pigs at the same fair, suggesting that a subpopulation of swine lineage IAV are capable of replication in humans in situations of high viral titers and close contact (19, 20, 22, 44). Previous risk assessment studies focused on direct inoculation of ferrets with variant IAV isolated from humans (23, 24, 45, 46). In this study, we sought to recapitulate the primary step of interspecies transmission by modeling the swine-human transmission potential using a ferret surrogate. We report the results of a swine-to-ferret transmission study that reproduced the capacity for interspecies transmission of a 2010.1 H3N2 IAV-S, A(H3N2)/swine/Ohio/A01354299/2017, isolated from a pig at an agricultural fair where human variant cases were detected.
Genetic analysis found sw/OH/2017 to be nearly identical (>99% sequence similarity) to a cluster of H3N2v viruses from the same agricultural fair as well as viruses from an additional H3N2v cases that occurred in Maryland in 2017. The hemagglutinin segments were of the human-like 2010.1 H3 lineage that was first detected in commercial swine 2012 and subsequently became the most frequently detected H3 lineage in the United States by the second half of 2015 (12). Being of human origin, sw/OH/2017 and related H3N2v isolates from humans retained receptor binding site residues (190D, 226I, and 228S) shown to confer preferential binding to α2,6 sialic acids despite multiple years of transmission and evolution in swine populations (35, 36). Prior work showed that six amino acids in the swine H3 (145, 155, 156, 158, 159, and 189) had a disproportionate impact on the antigenic properties of the swine H3 clade IV viruses (29); similarly, these six positions and position 193 drive the antigenic phenotype in human seasonal H3 (47). The 2010.1 swine H3 evolved antigenically following introduction to swine with multiple amino acid mutations in HA1, including K145, which was predominant in swine until reverting to N145 in 2017. sw/OH/2017 possessed the antigenic motif N145, T155, H156, N158, F159, K189, and S193, identical to that of the putative human seasonal precursor, A/Victoria/361/2011, but different from current seasonal H3N2 IAV in humans that possess S145, suggesting that individuals exposed to IAV after the antigenic change may lack cross-reactive antibodies capable of neutralizing sw/OH/2017 and like viruses. Additional studies are required to fully assess the impact of major antigenic site changes on escape from preexisting immunity in the human population and segment(s) of the human population potentially at risk. Population immunity due to seasonal vaccination or natural exposure may cross-protect against swine 2010.1 H3N2 infection in humans, but this will likely decrease as the swine 2010.1 H3N2 viruses continue to evolve in swine.
The neuraminidases of sw/OH/2017 and associated 2017 H3N2v were within a statistically supported clade in the 2002 lineage, one of two major NA lineages (N2-1998 and N2-2002) circulating in U.S. swine. The N2-2002 lineage subsequently split into two statistically supported cocirculating subclades. The sw/OH/2017 N2 represents a shift from the 2016 2010.1 H3 NA pairing that incorporated an N2 from a separate statistically supported 2002 clade where 2016 H3N2v NA genes were located (Fig. 1B). Functional balance of HA receptor binding and NA sialidase activity has been shown be important for infection, respiratory droplet transmission, and interspecies transmission (48–51). Additional studies assessing the balance of 2010.1 H3 with NA of each 2002 NA clade are warranted given the fact that these combinations represent the majority of recent variant H3N2 spillover into humans (22). Antibodies inhibiting NA activity are known to reduce the severity and duration of IAV-mediated disease in humans (52, 53) and protect mice from lethal IAV challenge (54, 55). The presence of neuraminidase antibodies in humans following vaccination or natural infection (56, 57) suggests that NA-mediated immunity may provide some protection against zoonotic infection, though a more comprehensive understanding will be required to assess the contributions of NA-inhibiting antibodies in protection, particularly N2 genes that have been evolving in swine for more than 15 years. The N2 from sw/OH/2017 and 2017 H3N2v possessed a histidine residue at position 274 in the NA, a marker associated with sensitivity to neuraminidase inhibitors (58).
The internal gene constellation of sw/OH/2017 was comprised of TRIG segments (PB2, PB1, PA, NP, and NS) and the 2009 pdmH1N1 M segment, the most frequently detected genome pattern according to swine data from USDA surveillance (59). This genome pattern became widespread in the U.S. commercial swine populations and is associated with variant and exhibition IAV cases. The polymerase genes of sw/OH/2017 have markers shown to confer enhanced replication in humans in other virus backbones, 271A, 590S, 591R, and 661A in PB2 and 669V in PA (31–34), indicating that sw/OH/2017 may not need substantial mutation to replicate in humans since human variant viruses have been highly similar to swine isolates from the same swine exhibition site. Further, the pdmH1N1 M segment has been shown to alter virion morphology, increasing filamentous virus particle numbers and NA activity, two phenotypes shown to enhance transmission in the guinea pig model (60–63). The replacement of the TRIG lineage M segment to that of the pdmH1N1 lineage, in tandem with the change of NA clade, potentially confers a transmissible phenotype on swine H3 IAV. However, there was no evidence of human-to-human transmission with the variant cases of 2010.1 H3N2 in 2017 (15). Further studies aimed at elucidating the effects of pdmH1N1 lineage M segment on virion morphology and NA activity in the context of swine-to-swine and interspecies transmission are warranted granted the increasing numbers of variant IAV cases.
The ferret is routinely used as a model for pathogenesis and transmission in risk assessment studies of IAV for humans. Replication of sw/OH/2017 in the upper respiratory tracts of ferrets was observed and was consistent with previously reported findings of human-like H3N2v infection in ferrets, with high virus titers being detected in the nasal wash samples at 1, 3, and 5 dpi (24). We detected moderate (3.8 log TCID50, 5 dpi) virus titers in the lungs of inoculated ferrets, indicating that sw/OH/2017 is capable of replication in the upper and lower ferret respiratory tracts. Inoculated ferrets displayed weight loss and increased body temperature, though not as pronounced as previously reported (23, 24), which may be due to the study endpoint at 5 dpi while maximum weight loss in the prior study was observed between 4 and 10 dpi. The pathology observed in the lungs of inoculated ferrets was moderate, with macroscopic pneumonia and composite histopathology scores of the lungs and trachea typical of infection with seasonal and swine IAV (51, 64). Though viral antigen was not detected in the lungs of infected ferrets by IHC, the presence of lung lesions consistent with IAV infection and viral lung titers detected in the BALF demonstrate that sw/OH/2017 has the capacity to replicate in the lungs of ferrets. Virus titers in the nasal wash and BALF samples were similar to those in previous studies directly inoculating ferrets with human 2010.1 H3N2v (24). Though we did not investigate the presence of virus in extrapulmonary tissue, it is unlikely that virus replication occurred outside the respiratory tract, as this phenotype has not been previously observed in H3N2v challenge studies of ferrets (23, 24, 27).
Previous studies examining the pathogenesis of variant IAV have shown these viruses to be well adapted to replication in the respiratory tissue of ferrets (24, 45, 46). Virus titers and kinetics of virus detection in the nasal swabs from ferrets infected with sw/OH/2017 were similar to previously reported data with H1N1v, H1N2v, and H3N2v. Further, there was no evidence of IAV replication outside the respiratory tract, as was previously observed. This may be attributed to the presence of molecular markers in PB2 (271A, 590S, 591R, and 661A) and PA (669V) that are present in variant IAV previously shown to replicate in ferrets. The major genetic difference between sw/OH/2017 and variant HxN2 IAV in prior ferret infection studies is the lineage of the NA gene segment (23, 24, 26, 45). The NA of sw/OH/2017 is from the N2-2002B lineage that resulted from the spillover of human seasonal IAV into the commercial swine population in 2002 (4, 5). Further, Barman et al. (51) have previously shown that a human lineage HA and NA correlated with efficient respiratory droplet transmission in ferrets, though IAV with a human lineage NA and a swine lineage HA transmitted inefficiently. Virus replication and transmission is known to be influenced by HA-NA balance, and further studies investigating the role of NA-mediated receptor destruction in IAV transmission in the pig-to-ferret and ferret-to-ferret models would further aid in assessing the associated molecular determinants and public health risk posed by spillover of swine IAV to humans.
A facet of risk assessment on variant IAV that has not been documented is assessing the propensity of swine viruses for interspecies transmission. To better understand the capacity of sw/OH/2017 for interspecies transmission, we modeled the swine-human interface by housing ferrets adjacent to a standard pig enclosure. Pigs inoculated with sw/OH/2017 shed virus in nasal swab samples, similar to previous reports of porcine inoculation with 2010.1 swine H3N2 viruses (11). Virus was detected in the nasal wash samples of ferrets exposed to infectious porcine respiratory droplets at the first sampling at 2 dpc and persisted until 6 dpc. These titers were similar to those observed in contact animals in ferret-to-ferret transmission experiments with both clade IV and 2010.1 H3N2v (24). These findings suggest that certain swine IAV have the capacity for interspecies transmission via respiratory droplets without additional mutation. As reported in previous studies, the phenotype of respiratory droplet transmission of 2010.1 H3N2v from ferret to ferret was moderately efficient (24). The efficient interspecies transmission phenotype observed for 2010.1 sw/OH/2017 provides a biological mechanism that corresponds to the increased number of 2010.1 H3N2v cases. Further studies investigating the swine-to-ferret transmission phenotypes of predominant swine and variant IAV will enhance the utility of the swine-to-ferret model to assess relative differences among swine strains.
Exhibition swine harbor a diverse population of IAV of that can sporadically spill over to humans and cause disease. Of particular concern are swine H3N2 strains that have the pandemic H1N1 lineage M segment and molecular markers shown to confer enhanced replication in humans. While these viruses have been shown to replicate and transmit to contact ferrets from directly inoculated ferrets, we demonstrate successful swine-to-ferret transmission of a wild-type 2010.1 H3N2 virus. Future studies further characterizing additional swine IAV in the swine-to-ferret transmission model and elucidating genetic factors that contribute to interspecies transmission will assist public health efforts to control IAV-S spillover to humans and in prioritizing strains for pandemic preparedness through development of candidate vaccine viruses of animal origin.
MATERIALS AND METHODS
Virus.A(H3N2)/swine/Ohio/A01354299/2017 (GenBank accession numbers MF801566 to MF801573) was isolated from a pig at the fair and maintained in the repository held at the National Veterinary Services Laboratories (NVSL) through the U.S. Department of Agriculture (USDA) IAV swine surveillance system in conjunction with the USDA-National Animal Health Laboratory Network (NAHLN). Virus was propagated on Madin-Darby canine kidney (MDCK) cells grown in Opti-MEM (Life Technologies, Waltham, MA) supplemented with 10% fetal calf serum and antibiotics/antimycotics. Virus growth media contained antibiotics/antimycotics and 1 μg/ml of tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-trypsin (Worthington Biochemical Corp., Lakewood, NJ).
Genetic and phylogenetic analyses.H3N2 influenza A virus HA and NA sequences from contemporary swine field strains, World Health Organization human seasonal H3 HA vaccine components, and human H3N2v and swine H3N2 collected from county fairs in 2016 to 2018 were downloaded from GISAID and the Influenza Research Database (65, 66). Where possible, the HA (n = 138) and NA (n = 162) gene sequences were from the same strain; however, unequal sequencing and a requirement to represent additional lineages in the NA-N2 gene tree (human seasonal N2 and swine “1998”) resulted in separate data sets. These data were aligned using MAFFT v7.222 (67, 68), and a maximum likelihood phylogeny for each gene segment was inferred using IQ-TREE v1.6.10 (69) with the best-fit model of molecular evolution automatically determined (for the HA, a transversion model, TVM, with empirical base frequencies and a 3-category FreeRate model of heterogeneity of substitution rates among sites; for the NA, TVM with empirical base frequencies and a 4-category gamma model of heterogeneity of substitution rates among sites [70]). Statistical support for branches within the best-scoring tree were estimated using the Ultrafast bootstrap approximation and the SH-like approximate likelihood ratio test (71). HA1 domain deduced amino acid alignments were used to calculate amino acid mutations between pairs of viruses. Internal gene segment lineage, either TRIG or pandemic, was determined using the octoFLU automated gene classification tool (72).
Virus replication and pathogenesis in ferrets.Twelve 4- to 6-month-old female or neutered male ferrets were procured from Marshall Bioresources (North Rose, NY) and determined to be serologically negative for influenza viruses via IAV nucleoprotein enzyme-linked immunosorbent assay (ELISA; IDEXX Laboratories Inc., Westbrook, ME). Ferrets were housed individually in ferret isolators (Plas-Labs, Lansing, MI) with a HEPA-filtered, negative-pressure environment in a biosafety level 3 containment facility in compliance with an approved USDA National Animal Disease Center (NADC) animal care and use protocol. Ferrets were implanted with a subcutaneous identification and temperature transponder (Biomedic Data Systems Inc., Seaford, DE). Baseline body temperature readings were recorded from −2 to 0 days postinoculation (dpi), and baseline weights were recorded at 0 dpi. Four ferrets were inoculated intranasally with 106 50% tissue culture infective doses (TCID50) of A(H3N2)/swine/Ohio/A01354299/2017 diluted in 1 ml of sterile phosphate-buffered saline (PBS) under isoflurane anesthetic, while another four ferrets were housed under similar conditions in a separate room, serving as uninfected controls. Nasal wash samples were collected at 1, 3, and 5 dpi under ketamine anesthetic. For collection of nasal wash samples, anesthetized ferrets were placed in the prone position and the nasal cavity was irrigated with 1 ml of sterile PBS; samples were collected in petri dishes, transferred to cryovials, and stored on ice until centrifugation at 3,000 rpm at 4°C and storage of the supernatants at –80°C. At 5 dpi, ferrets were humanely euthanized and lungs excised in toto for bronchoalveolar lavage fluid (BALF) collection with 10 ml of minimal essential medium (MEM). Lung and trachea sections were collected and stored in formalin for pathological analysis.
Swine-to-ferret transmission.Ten 3-week-old pigs obtained from a herd free of IAV and porcine reproductive and respiratory syndrome virus were housed in a biosafety level 3 containment facility in compliance with an approved USDA NADC animal care and use protocol. Five pigs were group housed in a raised pig deck (approximately 1.5 m by 2.4 m, raised 0.1 m off the ground) approximately 7.5 cm away from four ferrets housed individually in ferret isolators (Fig. 2). The open pig decks were enclosed by wire fence panels on the two shorter sides and solid panels on the longer sides, all approximately 0.7 m high. The pigs were at the level of the two ferrets in the lowest row of isolators, whereas the two ferrets in the top isolator row were approximately level with the vertical terminus of the pig deck wire panels. Five pigs used as negative controls were housed under similar conditions in a separate room. Pigs were inoculated intranasally with 2 ml of 106 TCID50/ml in PBS. To facilitate respiratory droplet exposure of ferrets to infectious porcine aerosols, the impermeable, outer isolator doors were removed 24 h after infection of pigs while the HEPA filtration motor was left on, allowing ambient air to be pulled into the ferret enclosure through the metal cage door. Additionally, the pig deck was placed near the room air inlet, and the ferret isolators were positioned near the room air outlet. Ferrets were provided routine care and/or handled before pigs, with a change in outer gloves and decontamination of equipment with 70% ethanol between individual ferrets. Nasal swab samples (FLOQSwabs; Copan Diagnostics, Murrieta, CA) were collected from pigs at 0, 1, 3, 5, and 7 dpi, and pigs were humanely euthanized at 15 dpi as previously described (73). To assess virus replication in contact ferrets, nasal wash samples were collected using the methods described above at 2, 4, 6, 8, and 14 days postcontact (dpc), with blood collected prior to euthanasia at 15 dpi/14 dpc from donor pigs and contact ferrets, respectively.
Virus replication and shedding.IAV presence in the nasal swabs (pigs) and nasal washes (ferrets) was assessed by virus isolation and titration on MDCK cells. Isolation of virus from nasal swabs and washes was accomplished by inoculating confluent monolayers of MDCK cells with a 1:2 (swine) or 1:5 (ferrets) dilution of sample material in Opti-MEM (Life Technologies, Carlsbad, CA) supplemented with 1 μg/ml of TPCK-trypsin. The TCID50 were determined by plating 10-fold serial dilutions of nasal swab/wash sample in MEM supplemented with antibiotics/antimycotics, l-glutamine, vitamins, bovine serum albumin, and 1 μg/ml of TPCK-trypsin. Dilutions were plated in quadruplicate onto confluent MDCK monolayers in 96-well plates and allowed to incubate for 72 h at 37°C. The presence of virus was measured by transferring 50 μl of virus supernatant to a V-bottom 96-well plate and incubating for 30 min with 50 μl of 0.5% turkey red blood cell solution. The presence of hemagglutination was considered positive. The TCID50 titer for each sample was calculated using the methods of Reed and Muench (74).
Hemagglutination inhibition.Seroconversion of donor pigs and contact ferrets was assessed using blood collected at 15 dpi and 14 dpc, respectively. Hemagglutination inhibition (HI) assays were performed on receptor-destroying enzyme (RDE)-treated sera in V-bottom 96-well plates with 0.5% turkey red blood cells as previously described (75).
Pathological examination of lungs.Pathological examination of the lungs from inoculated and naive control ferrets was conducted at necropsy at 5 dpi. A weighted lung score was calculated based on the proportion of the weight of each lung lobe to the total lung weight, averaged from four healthy lungs of uninfected controls prior to lavage, with average proportions for the right cranial lobe of 0.19, right middle lobe of 0.09, right caudal lobe of 0.21, right accessory lobe of 0.06, left cranial lobe of 0.24, and left caudal lobe of 0.21. The percentages of the lung surface area displaying purple-red consolidation for each lobe from infected ferrets were visually estimated, weighted based on the proportion of the lobe to the overall lung using the average proportions from the control ferrets described above, and then summed for each individual ferret for a total weighted percentage of macroscopic lung lesions as the lung score. To assess histopathology, tissue sections from the right middle or an affected lung lobe and trachea were excised and then fixed in 10% buffered formalin. Tissue sections were then embedded in paraffin, stained with hematoxylin and eosin, and scored using a previously described scoring system for IAV-infected swine lung and trachea (76). Immunohistochemistry for IAV nucleoprotein presence in the lung and trachea sections was performed at the Iowa State University Veterinary Diagnostic Laboratory using standard protocols (76, 77).
Influenza A virus genome sequencing.Whole-genome sequencing was performed on all samples collected from swine and ferrets from which virus was isolated by growth in MDCK cells. Viral RNA was isolated from the three nasal swab samples collected from each of the five pigs at 1, 3, and 5 dpi (n = 15) and three ferret nasal wash samples collected at 2, 4, and 6 dpc (n = 12) using a MagMax viral RNA isolation kit (Thermo Fisher), and the whole gene segments were amplified as previously described (78). Amplified gene segments were purified using Agencourt AMPure XP beads (Beckman Coulter, Brea, CA). Sequencing libraries were prepared with a Nextera XT DNA library prep kit (Illumina, San Diego, CA), and sequencing was performed on an Illumina MiSeq platform. Full genomes were mapped to the A(H3N2)/swine/Ohio/A01354299/2017 reference genome for analysis in Geneious v9.1.4 (Geneious, Auckland, New Zealand).
Statistical analysis.Changes in ferret weight (percentage of change in time point measurement from prestudy baseline body weight) and temperature (prestudy average baseline temperature subtracted from time point measurement) were assessed using analysis of variance (ANOVA) performed using GraphPad Prism software (GraphPad, La Jolla, CA). A P value less than 0.05 was considered significant.
Data availability.The consensus sequence from each of the samples collected from swine and ferrets is available for download from the USDA Ag Data Commons at https://doi.org/10.15482/USDA.ADC/1518735.
ACKNOWLEDGMENTS
We thank Michelle Harland and Jordyn Zoul for technical assistance and Brett Ashburn, Siu-Yin Virella, and Jean Kaptur for assistance with animal studies.
We were supported by USDA-ARS, USDA-APHIS, and an NIH-National Institute of Allergy and Infectious Diseases (NIAID) interagency agreement associated with CRIP (Center of Research in Influenza Pathogenesis), an NIAID-funded Center of Excellence in Influenza Research and Surveillance (CEIRS; HHSN272201400008C). Influenza A virus surveillance in exhibition swine was supported by the NIH-NIAID CEIRS contract HHSN272201400006C. This research used resources provided by the SCINet project of the USDA Agricultural Research Service, ARS project number 0500-00093-001-00-D. B.S.K., J.B.K., J.C., and T.K.A. were supported by an appointment to the USDA-ARS Research Participation Program administered by the Oak Ridge Institute for Science and Education (ORISE) through an interagency agreement between the U.S. Department of Energy (DOE) and the USDA under contract number DE-AC05-06OR23100.
Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. The USDA is an equal opportunity provider and employer.
FOOTNOTES
- Received 22 May 2020.
- Accepted 28 May 2020.
- Accepted manuscript posted online 10 June 2020.
Supplemental material is available online only.
This is a work of the U.S. Government and is not subject to copyright protection in the United States. Foreign copyrights may apply.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
