Possible Increased Pathogenicity of Pandemic (H1N1) 2009 Influenza Virus upon Reassortment

Since emergence of the pandemic (H1N1) 2009 virus in April 2009, three influenza A viruses—seasonal (H3N2), seasonal (H1N1), and pandemic (H1N1) 2009—have circulated in humans. Genetic reassortment between these viruses could result in enhanced pathogenicity. We compared 4 reassortant viruses with favorable in vitro replication properties with the wild-type pandemic (H1N1) 2009 virus with respect to replication kinetics in vitro and pathogenicity and transmission in ferrets. Pandemic (H1N1) 2009 viruses containing basic polymerase 2 alone or in combination with acidic polymerase of seasonal (H1N1) virus were attenuated in ferrets. In contrast, pandemic (H1N1) 2009 with neuraminidase of seasonal (H3N2) virus resulted in increased virus replication and more severe pulmonary lesions. The data show that pandemic (H1N1) 2009 virus has the potential to reassort with seasonal influenza viruses, which may result in increased pathogenicity while it maintains the capacity of transmission through aerosols or respiratory droplets.

Since emergence of the pandemic (H1N1) 2009 virus in April 2009, three infl uenza A viruses-seasonal (H3N2), seasonal (H1N1), and pandemic (H1N1) 2009-have circulated in humans. Genetic reassortment between these viruses could result in enhanced pathogenicity. We compared 4 reassortant viruses with favorable in vitro replication properties with the wild-type pandemic (H1N1) 2009 virus with respect to replication kinetics in vitro and pathogenicity and transmission in ferrets. Pandemic (H1N1) 2009 viruses containing basic polymerase 2 alone or in combination with acidic polymerase of seasonal (H1N1) virus were attenuated in ferrets. In contrast, pandemic (H1N1) 2009 with neuraminidase of seasonal (H3N2) virus resulted in increased virus replication and more severe pulmonary lesions. The data show that pandemic (H1N1) 2009 virus has the potential to reassort with seasonal infl uenza viruses, which may result in increased pathogenicity while it maintains the capacity of transmission through aerosols or respiratory droplets.
T he infl uenza virus A (H1N1) that caused the fi rst infl uenza pandemic of the 21st century, pandemic (H1N1) 2009, continues to be detected worldwide (1,2). The pandemic overall has been relatively mild; disease has ranged from subclinical infections to sporadic cases of severe pneumonia and acute respiratory distress syndrome (3)(4)(5)(6)(7)(8). The virus responsible is a unique reassortant virus containing neuraminidase (NA) and matrix genes from the Eurasian swine infl uenza virus lineage, and the other 6 gene segments are derived from the North American triple reassortant swine infl uenza virus lineage (9). From the pandemic's start, there have been concerns the virus may mutate or reassort with contemporary infl uenza viruses and give rise to more pathogenic viruses.
Cocirculation of multiple strains of infl uenza virus A in humans provides an opportunity for viral genetic reassortment (mixing of genes from >2 viruses) (10). Genetic reassortment of pandemic (H1N1) 2009 virus with seasonal infl uenza A (H3N2) or seasonal infl uenza A (H1N1) viruses might thus represent a route to enhanced pathogenicity. No reassortment events between pandemic (H1N1) 2009 and seasonal viruses have been reported in humans. However, a triple-reassortant swine infl uenza virus A (H1N1), distinct from pandemic (H1N1) 2009 virus and containing the hemagglutinin (HA) and NA genes of seasonal infl uenza virus A (H1N1), was described recently (11). Dual infections by seasonal infl uenza A (H1N1) and seasonal infl uenza A (H3N2) viruses have been reported (12), as well as mixed infections of pandemic (H1N1) 2009 and seasonal infl uenza A (H3N2) viruses (13,14), highlighting the potential for reassortment of currently circulating infl uenza viruses. Subtype H1N2 reassortant infl uenza viruses that contain the HA of seasonal infl uenza A (H1N1) and the NA of seasonal infl uenza A (H3N2) viruses have been isolated from humans during previous infl uenza seasons, confi rming that such HA/NA combinations can emerge in humans (15,16).
To investigate the potential for reassortment between seasonal infl uenza A and pandemic (H1N1) 2009 viruses, we used an in vitro selection method using reverse genetics and serial passaging under limited dilution conditions. Pathogenicity and transmission of these viruses were tested Possible Increased Pathogenicity of Pandemic (H1N1) 2009 Infl uenza Virus upon Reassortment by using a ferret model. We report here the identifi cation of 4 reassortants with different gene constellations.

Cells and Viruses
MDCK cells were cultured in Eagle minimum essential medium as described (17) After these viruses were passaged in MDCK cells 2×, all 8 gene segments were amplifi ed by reverse transcription-PCR, cloned in a modifi ed version of the bidirectional reverse genetics plasmid pHW2000 (19,20), and subsequently used to generate recombinant virus by reverse genetics as described elsewhere (19).

Generation of the Reassortant Viruses
Mixtures of reassortant viruses were generated in 293T cells by using reverse genetics, by co-transfecting 8 plasmids that encode the pandemic (H1N1) 2009 virus genome together with 7 plasmids encoding the seasonal infl uenza A (H3N2) or seasonal infl uenza A (H1N1) virus genome. We omitted HA of the seasonal viruses to ensure that only reassortants containing the pandemic (H1N1) 2009 virus HA could arise, against which a large proportion of the human population is still immunologically naïve (21). The 293Tcell supernatants were passaged in quadruplicate under limiting dilution conditions by using 10-fold serial dilutions in MDCK cells 3× to enable selective outgrowth of viruses with high in vitro replication rates. After 3 passages, the genome composition of these viruses was determined by sequencing with conserved primers targeting noncoding regions of each gene segment. Reverse genetics was also used to produce specifi c reassortant viruses (pandemic

In Vitro Characterization of Viruses
Multicycle replication curves were generated by injecting MDCK cells at a multiplicity of infection of 0.01 50% tissue culture infective dose (TCID 50 ) per cell in 2-fold (17). Virus titers from samples of inoculated MDCK cells, as well as nasal and throat swabs or homogenized tissue samples from inoculated ferrets, were determined by endpoint titration in MDCK cells, as described (22).

Ferret Experiments
All animal studies were approved by an independent animal ethics committee. Experiments were performed under animal BioSafety Level 3+ conditions. The ferret model to test pathogenicity and transmission of pandemic (H1N1) 2009 virus was described previously (17,18). To study pathogenicity, 5 groups of 6 infl uenza virus-seronegative female ferrets (Mustella putorius furo) were inoculated intranasally with 10 6 TCID 50 of wild-type pandemic

Immunohistochemistry and Histopathology
Immunohistochemical testing and pathologic examination were performed by using lungs of inoculated ferrets. For each virus, 3 ferrets were euthanized at 3 and 7 days postinoculation (dpi) by exsanguination. Necropsies and tissue sampling were performed according to standard protocol. After fi xation in 10% neutral-buffered formalin and embedding in paraffi n, samples were sectioned at 4 μm and stained with an immunohistochemical method by using a mouse monoclonal antibody against the nucleoprotein of infl uenza virus A (23). Infl uenza virus antigen expression in lung sections was scored for bronchial surface epithelium, bronchial submucosal gland epithelium, bronchiolar epithelium, alveolar type I pneumocytes, and alveolar type II pneumocytes. Scoring was categorized as 0, no positive cells; 1, few positive cells; 2, moderate number of positive cells; and 3, many positive cells. Serial lung sections were stained with hematoxylin and eosin for detection and description of pathologic changes. Samples were scored for infl uenza virus-associated infl ammation in bronchi (bronchitis), bronchial submucosal glands (bronchoadenitis), bronchioles (bronchiolitis), and alveoli (alveolitis). Scoring of severity of infl ammation was 0, no infl ammation; 1, mild infl ammation; 2, moderate infl ammation; and 3, marked infl ammation. Researchers who examined the sections had no knowledge of the identity of the ferrets.

Pathogenicity of the Reassortant Viruses in Ferrets
The mean maximum weight loss was 7% for animals 2009-seasonal infl uenza A (H3N2) NA had a maximum weight loss of 4%, 2%, 2%, and 6%, respectively (data not shown).
Nose and throat swabs were collected daily, and virus titers were determined. Infectious virus shedding continued until 6-7 days dpi from noses ( Figure 2, panels A and C) and throats (Figure 2, panels B and D) of most inoculated animals. Total virus shedding from the nose, as calculated from the area under the curve for ferrets in the experiment for 7 days (n = 3), was signifi cantly lower in animals inoculated with the pandemic (H1N1) 2009-seasonal infl uenza A (H1N1) PB2 reassortant virus (p = 0.003 by t test) and signifi cantly higher in the animals inoculated with the pan-

Pathologic Changes in the Respiratory Tract of Ferrets Inoculated with Pandemic (H1N1) 2009 and Reassortant Viruses
At 7 dpi, virus antigen expression was undetectable in lung tissue of any of the euthanized ferrets, and lesions were absent or resolving. At 3 dpi, neither viral antigen expression nor lesions were detected in lungs of ferrets inoculated with pandemic (H1N1) 2009-seasonal infl uenza A (H1N1) PB2. Only 1 of 3 ferrets inoculated with pandemic (H1N1) 2009 had scant virus antigen expression and mild associated lesions in bronchial submucosal glands and bronchioles at 3 dpi (Table 2, Figure 4). In contrast, all 3 ferrets inoculated with pandemic (H1N1) 2009-seasonal infl uenza A (H3N2) NA had moderate to abundant virus antigen expression in bronchial submucosal glands, bronchioles, or both, associated with moderate to marked infl ammation (Table 2, Figure 4). Virus antigen expression and associated lesions in the lungs of ferrets inoculated with pandemic (H1N1) 2009-seasonal infl uenza A (H3N2) PB1NA were intermediate between those of wild-type pandemic (H1N1) 2009 and pandemic (H1N1)-seasonal infl uenza A (H3N2) NA (Table 2).
Cell types in which virus antigen expression was detected were ciliated epithelial cells of bronchi, epithelial cells of bronchial submucosal glands, ciliated and nonciliated cells of bronchioles, and both squamous and cuboidal epithelial cells (interpreted as type I and type II pneumocytes, respectively) of alveoli ( Figure 4). Virus antigen expression was also seen in desquamated epithelial cells and cell debris in lumina of above tissues.
Lesions associated with virus antigen expression can be categorized as acute, focal or multifocal, necrotizing bronchitis, bronchoadenitis, bronchiolitis, and alveolitis. These lesions were characterized by degeneration and necrosis of epithelial cells, infi ltration of the affected tissues and their lumina by many neutrophils and few eosinophils, and exudation of edema fl uid and fi brin into tissue lumina.

Transmission of Reassortant Viruses in Ferrets
Transmission of pandemic (H1N1) 2009 and reassortant infl uenza viruses through aerosol or respiratory droplets was tested in the ferret model. each of the reassortant viruses. The fi rst day of virus detection in the previously uninfected animals was 2 days post exposure, similar for all viruses tested.

Discussion
We used an in vitro selection method to identify reassortant viruses between pandemic (H1N1) 2009 virus and seasonal infl uenza A (H1N1) and infl uenza A (H3N2) viruses of interest for testing in a ferret model. Studying the effects of reassortment on changes in infl uenza virus phenotype is cumbersome because the number of reassortants that can be generated between 2 viruses is high; 2 8 = 256 different viruses. After 3 passages, a limited number of specifi c virus populations were selected in vitro. Minor virus variants representing <20% of the virus population would remain undetected in our approach of PCR amplifi cation and direct determination of the consensus sequence of the amplicons. However, upon repeating the procedure 4 times for both reassortment combinations, the seasonal infl uenza virus genes that were selected in the pandemic (H1N1) 2009 virus backbone were more or less consistent, with NA of seasonal infl uenza virus A (H3N2) being selected in 4/4 attempts, PB1 of seasonal infl uenza A (H3N2) and PB2 of seasonal infl uenza virus A (H1N1) in 3/4 attempts, and PA of seasonal infl uenza virus A (H1N1) in 2/4 attempts. Replication in MDCK cells may not be the best selection criterion for the identifi cation of reassortants of interest to human health. Nevertheless, we chose this in vitro selection method because previous work has shown that pandemic (H1N1) 2009 outcompetes seasonal infl uenza A (H1N1) and seasonal infl uenza A (H3N2) viruses rapidly, reducing the opportunity for reassortment (24). This growth advantage over seasonal viruses was in agreement with the fact that selected viruses mostly contained pandemic (H1N1) 2009 genes. The use of reverse genetics enables production of all gene segments at approximately similar copy numbers on transfection, whereas after double infection with 2 viruses, in vitro or in ovo viruses may differ in replication capacity, resulting in a bias of reassortants produced.
Notably, the polymerase gene segments of seasonal infl uenza A (H1N1) and seasonal infl uenza A (H3N2) viruses frequently substituted for the polymerase genes of the pandemic (H1N1) 2009 virus in vitro. In minigenome assays, the polymerase complex activity of the wild-type pandemic (H1N1) 2009 virus was relatively low, and replacement of various polymerase genes of the pandemic (H1N1) 2009 virus increased this activity. However, polymerase complexes with the highest activity in minigenome assays were not necessarily the ones detected in the reassortant viruses (data not shown). This apparent discrepancy is probably a result of the different parameters under investigation in the 2 assays, in particular, the production of mRNA vs. all viral RNAs.
Virus   Figure 4). In a previous study, the wild-type pandemic (H1N1) 2009 virus was detected more abundantly in the lower airways of ferrets than in the present study (18). We attribute this difference to the use of a virus isolate rather than a virus generated by reverse genetics and to a different batch of ferrets in the previous study. In the present study, all viruses were produced with reverse genetics and can thus be compared directly. Moreover, the reassortant pandemic (H1N1) 2009 virus with the NA of the seasonal infl uenza virus A (H3N2) was more pathogenic than both sources of pandemic (H1N1) 2009 virus, either the wildtype isolate or the virus derived by reverse genetics. In-creased severity of lesions may be related to higher virus replication in the lung, to stronger host immune responses, or both (25). We conclude that the pandemic (H1N1) 2009 virus has the potential to reassort with seasonal infl uenza virus A (H1N1) and infl uenza virus A (H3N2) and that such reassortment events could result in viruses with increased pathogenicity in ferrets. Although increased pathogenicity in ferrets cannot be extrapolated directly to increased pathogenicity in humans, ferrets are susceptible to natural infection and respiratory disease and lung pathology develop in a manner similar to the that in humans infected with seasonal, avian, or pandemic infl uenza viruses. Thus, the ferret model is generally thought to be a good animal model for infl uenza in humans (26,27  in ferrets and humans (28), and the ferret model has further been used successfully for studies on virus transmission through respiratory droplets or aerosols (18,29). All reassortants were transmitted between ferrets through aerosol or respiratory droplets. These results demonstrate that some reassortants between pandemic (H1N1) 2009 and seasonal infl uenza A (H3N2) were viable, remained transmissible, and were more pathogenic than the wild-type pandemic (H1N1) 2009 virus and emphasize the importance of monitoring reassortant viruses in surveillance programs because reassortment events may affect pathogenicity.
Although viruses with the NA gene (with or without the PB1 gene) of seasonal infl uenza A (H3N2) were identifi ed here as potentially fi t virus reassortants, reassortant viruses with other gene constellations may have selective advantages in humans as well. The 1968 infl uenza virus A (H3N2) pandemic also continued to reassort after the pandemic year, resulting in viruses during 1969-1971 with a different N2 gene than those earlier in the pandemic (30). Reassortants of infl uenza virus A (H1N2) with the HA of seasonal infl uenza A (H1N1) and the NA of seasonal infl uenza A (H3N2) viruses have been isolated from humans during previous infl uenza seasons, thereby confi rming that reassortant infl uenza viruses with such an HA/NA combination can emerge in humans (15,16). Moreover, infl uenza (H1N2) viruses frequently have been detected in pigs around the world (31). Therefore, we recommend that reassortant of pandemic (H1N1) 2009 infl uenza viruses be monitored closely in surveillance programs, particularly when changes in pathogenicity or transmission in humans become apparent.