Human Clade 2.3.4.4 A/H5N6 Influenza Virus Lacks Mammalian Adaptation Markers and Does Not Transmit via the Airborne Route between Ferrets

Avian influenza A viruses are a threat to human health, as they cross the species barrier and infect humans occasionally, often with severe outcome. The antigenic and genetic diversity of A/H5 viruses from the A/goose/Guangdong/1/96 lineage is increasing, due to continued circulation and reassortment in poultry, posing a constant risk for public health and requiring regular risk assessments. Here we performed an in-depth characterization of the properties of the newly emerged zoonotic A/H5N6 virus in vitro and in ferrets. The lack of airborne transmission in the ferret model indicates that A/H5N6 virus does not pose a direct public health threat, despite the fact that it can replicate to high titers throughout the respiratory tracts of ferrets and cause more severe disease than other clade 2.3.4.4 viruses.

A/H5N6 GZ/14 HA is unstable. Upon virus attachment to SA receptors on the cell surface and internalization into endosomes, a low-pH-triggered conformational change of HA mediates fusion of the viral and endosomal membranes to release the viral genome in the cytoplasm (18). In order to assess the acid stability of the HA of A/H5N6 GZ/14, a syncytium formation assay was performed to measure the pH threshold required for HA-mediated cell-to-cell fusion. Vero cells were transfected with plasmids expressing different HAs and exposed to trypsin to cleave and activate the HA, followed by acidification of the cell culture by a pH gradient ranging from pH 4.8 to 5.9. Visual inspection of the cell cultures for the presence of syncytia (multinucleated cells) was used to determine the pH threshold triggering the conformational change and subsequent membrane fusion (Fig. 1A). Control HAs, the A/H5N1 IN/05 wild-type (WT) HA (unstable), the A/H5N1 IN/05 HA carrying airborne-transmission substitutions (H103Y, T156A, Q222L, and G224S [H5 numbering used throughout]) (stable), and the human A/H3N2 NL/03 HA (stable), were included. The pH values at which cell-to-cell fusion was triggered by A/H5N6 GZ/14 HA and A/H5N8 ck/NL/14 HA were 5.6 and 5.8, respectively, which were similar to or higher than those of the unstable A/H5N1 IN/05 WT HA (pH 5.6) and higher than the two stable HAs, A/H5N1 IN/05 airborne (pH 5.3) and A/H3N2 NL/03 (pH 5.3).
The conformational change of HA from a metastable nonfusogenic state to a stable fusogenic state can also be triggered at neutral pH when the HA is exposed to increasing temperatures (19). Therefore, the thermostability was assessed using a temperature sensitivity assay to further assess HA stability (Fig. 1B). Viruses were exposed to a temperature gradient ranging from 50 to 60°C for 30 min, and the HA titer was subsequently recorded. In agreement with the results of the fusion assay, A/H5N6 GZ/14 and A/H5N8 ck/NL/14 viruses were less stable than the human A/H3N2 and the airborne transmissible A/H5N1 virus.
A/H5N6 GZ/14 possesses a high polymerase activity, mediated by the E627K substitution in PB2. To measure the polymerase activity of A/H5N6 GZ/14, minireplicon assays were performed by transfecting 293T cells with plasmids expressing the polymerase genes (PB2, PB1, and PA) and the nucleoprotein (NP) of different GsGd A/H5N6 GZ/14 does not transmit between ferrets via the airborne route. Human influenza viruses are known to be transmitted from person to person via the airborne route by aerosols or respiratory droplets. Therefore, it is of importance to investigate the airborne transmissibility of emerging influenza A viruses. Here, airborne transmission experiments were performed in ferrets. Briefly, one day after donor ferrets were inoculated with A/H5N6 GZ/14, naive recipient ferrets were placed in an adjacent cage, separated from the donor cage by a steel grid, allowing viruses to be transmitted only via the airborne route. Virus shedding from the inoculated ferrets was high, with titers up to 10 5.25 50% tissue culture infective doses (TCID 50 )/ml, and long-term, until 7 days postinoculation (dpi) (Fig. 3). Upon exposure, no replicating virus was detected in any of the recipient ferrets. Lack of transmission was confirmed by the absence of seroconversion 14 days postexposure in all recipient ferrets (data not shown).
A/H5N6 GZ/14 is more pathogenic than a clade 2.3.4.4 A/H5N8 virus. In order to assess the pathogenicity of A/H5N6 GZ/14, six ferrets were inoculated intranasally with 10 6 TCID 50 of virus. Daily, the animals were monitored for body weight and clinical signs. Nose and throat swabs were also collected daily. At 3 and 6 dpi, three animals were sacrificed, and necropsies were performed. Data from this experiment were compared to historical data on intranasal challenge of ferrets with A/H5N8 ck/NL/14 and A/H5N1 IN/05 (14).
A/H5N6 GZ/14-inoculated animals lost on average 13% of their starting body weight (ranging from 9 to 18% at 6 dpi) (Fig. 4A). This body weight loss was intermediate between that of A/H5N8-inoculated animals (3%) and that of A/H5N1-inoculated animals (21%). At 3 dpi, all animals were less active and exhibited ruffled fur and watery eyes. One ferret was particularly weak, especially in the hind legs and did not move when stimulated. At 4 dpi, all ferrets presented with ruffled fur and with weak hind legs, showed inappetence, and kept their eyes slightly closed. At 6 dpi, all animals regained 293T cells were transfected with plasmids containing PB2, PB1, PA, and NP genes, a luciferase (firefly) reporter plasmid, and a control luciferase (Renilla) reporter plasmid. After transfection, the 293T cells were cultured at 33°C (light gray) and 37°C (dark gray) for 24 h, and luciferase activity was evaluated from the cell extracts. Results are presented as the means Ϯ standard deviations (error bars) of polymerase activity of virus compared to the polymerase activity of the human A/H1N1 virus, HK/98, which was set at 100%, and are averages from three independent experiments.
Characterization of a Novel Clade 2.3.4.4 A/H5N6 Virus activities, and their body weight stabilized. To summarize, A/H5N6 GZ/14-inoculated animals showed more severe clinical signs than A/H5N8-inoculated animals, such as inactivity or ruffled fur, as reflected by the global clinical score (Fig. 4A). However, they did not progress to visible signs of pneumonia as observed in A/H5N1-inoculated animals (14).
The shedding of A/H5N6 GZ/14 from the respiratory tracts of ferrets, as measured by virus titration of throat and nose swab samples, was high and prolonged, similar to that of A/H5N1 IN/05, and longer and higher in titers than those of A/H5N8 ck/NL/14 (Fig. 4B).
A/H5N6 GZ/14 virus was detected by virus titration at high titers in the nasal turbinates, both at 3 dpi and 6 dpi (mean titers of 10 8.3 and 10 6.8 TCID 50 /g of tissue, respectively) ( Fig. 4C and D). Moreover, A/H5N6 GZ/14 virus was detected at 3 dpi in more tissues of the respiratory tract, such as the lung and tonsils, and at higher titers than of A/H5N8 ck/NL/14. At 3 dpi, A/H5N6 GZ/14 was indeed detected at a high titer (10 7.1 TCID 50 /g of tissue) in the lung of one animal. A/H5N6 GZ/14 was also detected at 3 dpi in extrarespiratory tissues such as the olfactory bulb (mean titers of 10 4.2 TCID 50 /g of tissue). At 6 dpi, A/H5N6 GZ/14 was detected only in the lung of one animal, which was the animal that had lost the most weight (18%) (Fig. 4D). The low virus titers in the brain, pancreas, and intestine at 3 dpi and 6 dpi were not confirmed by immunohistochemical analysis (IHC) detecting with the viral NP ( Fig. 4C and D; Table 2). At 3 dpi, high titers in the nasal turbinates correlated with the presence of virus antigen in the caudal respiratory epithelium and more abundantly in the olfactory  Table 2). Virus antigen expression was associated with moderate epithelial necrosis and infiltration of some neutrophils and macrophages in the caudal respiratory epithelium and with moderate to severe necrosis and inflammation in the olfactory epithelium. Gross pathology revealed multifocal moderate gray-red consolidation and multifocal mild emphysema in the lung of one ferret, in which virus titers were also detected (see above), with 50% of the lungs affected and a relative lung weight (RLW) of 0.9%, expressed as a percentage of the total body weight ( Table 3). The RLW is a measure of the level of pulmonary inflammatory infiltrate and edema. In comparison, the median RLW of noninfected ferrets is 0.6% of total body weight, with a range from 0.5% to 0.8% (23). In the lungs of this ferret, virus antigen expression was observed in few pneumocytes, more of the type II than of the type I, and in a few alveolar macrophages (Table 2). Virus antigen expression was associated with a mild to moderate infiltration of neutrophils, alveolar macrophages, edema, and epithelial necrosis. Moreover, there was a multifocal peribronchiolar cuffing with lymphocytes, macrophages, and plasma cells. The other two ferrets presented with only 0% and 5% of the lung tissue affected (RLWs of 0.4% and 0.6%, respectively), and no virus antigen expression nor microscopic lesions were observed . Average activity status is represented by the solid line and the closed black circles (right y axis). Activity status was scored daily as follows: 0, alert and playful; 1, alert and playful only when stimulated; 2, alert but not playful when stimulated; 3, neither alert nor playful when stimulated. (B) Average virus titers in the throat and nasal swabs are shown by the gray circles and white circles, respectively. (C) Virus titers in tissues collected at 3 dpi are shown for individual ferrets in black, light gray, and dark gray. (D) Virus titers in tissues collected at 6 dpi are shown for individual ferrets in black, light gray, and dark gray. All virus titers were determined by endpoint titration in MDCK cells. In panel B, the lower limit of detection is indicated by the dotted line. In panels C and D, the lower limit of detection is indicated by the short horizontal lines spanning the three bars for each tissue. All error bars represent standard deviations. Nasal turb., nasal turbinates; Tracheobr. ln, tracheobronchial lymph node; Olf. bulb, olfactory bulb.  (Tables 2 and 3). In the olfactory bulbs of two ferrets, virus antigen expression was observed in some cells that were associated with mild gliosis. In other tissues, antigen expression was not demonstrated, and there were no significant lesions ( Table 2).
At 6 dpi, virus antigen distribution in the nasal epithelium (respiratory and olfactory) was similar to that of animals euthanized at 3 dpi albeit with fewer positive cells ( Table 2). Virus antigen expression was also associated with inflammation but with less necrosis compared to 3 dpi. Moreover, one ferret presented lesions in 10% of his lungs characterized by mild consolidation (RLW of 0.7%), while the other two ferrets showed no lesions (RLWs of 0.7% and 0.8%). Microscopically, no virus antigen or lesions were observed in the lungs of animals euthanized at 6 dpi. In the olfactory bulbs of two  ferrets, few cells in the glomerular zone were virus antigen positive associated with mild gliosis. A/H5N6 GZ/14 causes severe pneumonia upon intratracheal inoculation of ferrets. Most humans infected with an HPAI A/H5 GsGd virus developed severe pneumonia resulting in acute respiratory distress syndrome. Differences in virus tropism and pathogenicity after different routes of inoculation of ferrets have already been reported (24), and it has been shown that intranasal inoculation does not necessarily result in the development of pneumonia, as seen in most human cases of infection. Therefore, in order to study the ability of A/H5N6 GZ/14 to cause pneumonia, three ferrets were inoculated intratracheally with 10 6 TCID 50 of virus. At 1 dpi, all animals were alert, but they were playful only after stimulation (activity status [AS] of 1). At 2 dpi, in the morning, one ferret was found dead, while the two other ferrets were less active than the day before (AS of 2). Later in the afternoon, these two ferrets were neither alert nor playful when stimulated (AS of 3). One animal presented with severe dyspnea and succumbed to the infection shortly after. The other ferret was very weak, demonstrated swollen conjunctiva and mild dyspnea and was euthanized because of its poor condition. The animal that died at 1 dpi had lost 6% of its starting body weight, and the other two ferrets had lost 12% and 10% of their starting body weights.
At the time of death or euthanasia, between 1 dpi and 2 dpi depending on the animal, virus titers were very high throughout the respiratory tract and with the highest titers in the lungs (mean titers of 10 7.2 TCID 50 /g of tissue) (Fig. 6A).
Gross pathological analysis revealed a severe diffuse to coalescing pulmonary consolidation affecting 90 to 100% of the lungs of all animals, characterized by dark red and depressed lung parenchyma. The RLWs of these ferrets were between 1.3% and 1.6% (Table 3). The ferrets had enlarged tracheobronchial lymph nodes, approximately two to three times the normal size. Microscopically, there was multifocal coalescing to diffuse expression of virus antigen in many type II pneumocytes, in a moderate number of type I pneumocytes, and few macrophages. Virus antigen expression was associated with multifocal coalescing to severe diffuse alveolar damage (DAD) (Fig. 6D to G). DAD was characterized by severe epithelial necrosis, large amounts of intraluminal proteinaceous exudate (edema), fibrin, extravasated erythrocytes (hemorrhage), cellular debris, and infiltration of moderate numbers of viable and degenerated neutrophils and increased numbers of alveolar macrophages. Around the bronchioles in the perivascular space, there was a mild to moderate infiltration of macrophages, lymphocytes, and a few plasma cells and edema. In two ferrets, there was antigen present in a few mononuclear cells of the tracheobronchial lymph node. In the other collected tissues, virus antigen expression was not demonstrated, and significant lesions were absent.
A/H5N6 GZ/14 is not recognized by ferret antisera raised against available candidate vaccine viruses. To study the antigenic relationship between A/H5N6 GZ/14 and other clade 2.3.4.4 viruses and the neutralization of A/H5N6 GZ/14 by postinfection antisera raised against available candidate vaccine viruses (CVVs), we performed a hemagglutination inhibition (HI) assay (

DISCUSSION
Here, we report on a detailed in vitro and in vivo characterization of the clade 2.3.4.4 A/H5N6 GZ/14 virus in order to evaluate the risk for humans. Sequence analysis of A/H5N6 GZ/14 virus revealed a few amino acid substitutions that have previously been described as mammalian adaptation markers for other A/H5 viruses of the GsGd a Ferrets are indicated as follows: F1, ferret 1; F2, ferret 2, etc. b The relative lung weight was expressed as a percentage of the total body weight at the time of the necropsy. The median relative lung of noninoculated female ferrets is 0.6 (range, 0.5 to 0.8).
lineage: 94N (25), 133A (25), and 235P (26) in HA, which have been associated with increased binding of A/H5N1 viruses to human-type receptors, and E627K in PB2, which has been associated with increased replication of influenza viruses in vitro and in vivo at temperatures equivalent to those of the mammalian upper respiratory tract (URT) (21,22,27). Despite the fact that our knowledge of the genetic markers of mammalian adaptation is growing, performing risk assessment using genetic data alone might be misleading because of context dependency. Therefore, we sought to perform a phenotypic characterization of A/H5N6 GZ/14 by focusing on the phenotypes that have been associated with airborne transmission of A/H5N1 GsGd viruses (28,29): a change in receptor binding preference of HA from avian-type to human-type receptors, ␣2,3-SA and ␣2,6-SA, respectively, increased HA thermostability and acid stability and increased polymerase activity. Using a resialylated TRBC assay, A/H5N6 GZ/14, as well as A/H5N8 ck/NL/14, was shown to display a typical pattern of avian influenza viruses by binding exclusively to ␣2,3-SA. These data are in accordance with previous studies, in which A/H5N6 GZ/14 and another human A/H5N6 isolate, A/Sichuan/26221/2014, were found to bind only to ␣2,3-SA (30, 31). Interestingly, A/H5N6 viruses isolated from healthy ducks in China bound in a similar fashion to both ␣2,3-SA and ␣2,6-SA as well as to both upper (trachea) and lower (alveoli) respiratory epithelium (32). A/H5N6 viruses isolated from wild bird carcasses in Hong Kong also displayed a dual binding phenotype (31). The substitutions that mediated the binding of these avian A/H5N6 viruses to human-type receptors were not identified. Although most clade 2.3.4.4 viruses have retained ␣2,3-SA specificity, it was shown that the emergence of clade 2.3.4.4 viruses was accompanied with the capacity of these viruses to bind to fucosylated sialilosides, mediated by K218Q and S223R, substitutions that are conserved in clade 2.3.4.4 HAs (33). Of note, such modifications cannot be detected using the resialylated TRBC, in which only the nature of the bondage between the terminal sialic acid and the antepenultimate sugar is evaluated.
HA acid stability has been described as an important host range factor, associated with adaptation of avian and swine influenza viruses to the human host, airborne transmission of avian viruses, and pandemic potential (28,29,(34)(35)(36)(37). The HAs of A/H5N6 GZ/14 and of A/H5N8 ck/NL/14 were unstable compared to those of a human A/H3N2 virus and an avian-origin A/H5N1 transmissible virus assessed in syncytium formation and thermostability assays. A/H5N6 GZ/14 and A/H5N8 ck/NL/14 do not possess known substitutions that have been reported to increase the acid stability and/or thermostability of A/H5 viruses from other clades of the GsGd lineage, such as H103Y, T315I, or K58I (HA2 numbering) (28,29,38,39).
Using a minireplicon assay, we demonstrated that A/H5N6 GZ/14 possessed a high polymerase activity, which was mediated by E627K in PB2. Out of the 16 A/H5N6 human isolates, seven isolates carry the PB2-E627K, and one isolate, A/Sichuan/26221/ 2014, carries the PB2-D701N substitution, which has also been associated with increased replication and transmission in mammalian hosts (20,40,41). Acquisition of PB2-E627K might be the result of adaptation to the human host, although this substitution has also been detected in A/H5 avian virus isolates from other clades, as well as in other avian-origin influenza viruses such as A/H7N9 (42).
Of the three virus traits that conferred airborne transmissibility to A/H5N1 viruses between ferrets, A/H5N6 GZ/14 possesses only one of these traits, which is high polymerase activity. On the basis of the results of these in vitro assays, we predicted that A/H5N6 GZ/14 would not be transmissible via the airborne route, which was confirmed using the ferret model. Ferrets are used as animal models to study airborne transmission of influenza viruses, as it was shown that human influenza viruses transmit via the airborne route between ferrets, while avian influenza viruses do not,  (14)(15)(16). However, in contrast to these viruses that had acquired internal genes from LPAI viruses from wild birds, A/H5N6 GZ/14 possesses internal genes from the A/H5N1 GsGd lineage and also the adaptive substitution E627K in PB2 that might increase its pathogenicity in mammalian hosts. Ferrets are used as a model organism to study the pathogenesis of influenza viruses, as their susceptibility to infection with human and avian influenza viruses is similar to that of humans, and they develop respiratory disease similar to that observed in humans. Using the ferret model, we showed that the pathogenicity of A/H5N6 GZ/14 largely depends on the route of inoculation, as described previously for A/H5N1 viruses (24). Upon intranasal inoculation, the pathogenicity of A/H5N6 GZ/14 was intermediate between the pathogenicities of A/H5N1 IN/05 and A/H5N8 ck/NL/14 (14). The virus was detected in the olfactory bulb as early as 3 dpi, and it is possible that A/H5N6 GZ/14 would also be able to spread to the brain as seen with A/H5N1 IN/05, 7 days after intranasal inoculation (24,43). Ferrets lost on average 13% of their body weight over the time of the experiment, displayed more clinical signs than A/H5N8 ck/NL/14-inoculated ferrets, but did not progress to pneumonia as seen in A/H5N1 IN/05-inoculated animals. Differences between A/H5N6 and A/H5N1 might just be due to the route of inoculationintranasal-which results in virus deposition in the lower respiratory tracts of some ferrets but not necessarily all (23,24). Therefore, to assess the ability of A/H5N6 GZ/14 to cause pneumonia, which is observed in most humans infected with A/H5 GsGd viruses, ferrets were inoculated intratracheally with 10 6 TCID 50 of virus. Upon intratracheal challenge, all ferrets developed a severe and fatal pneumonia characterized by severe DAD. High levels of replication of A/H5N6 GZ/14 in the lungs of infected ferrets upon deposition of the virus in the lungs correlates with the study by Hui et al., in which A/H5N6 GZ/14 was shown to replicate to very high titers in lung explants (31). Sun et al. (32) also showed that A/H5N6 viruses isolated from healthy ducks were not as pathogenic in ferrets as a 2.3.4 A/H5N1 virus, but these A/H5N6 viruses carried internal genes from different origin.
To conclude, the results from this study demonstrate that an A/H5N6 virus of the G1 genotype is not fully adapted to mammalian hosts and therefore presents a low risk to humans, although it caused more severe disease in ferrets than other clade 2.3.4.4 viruses. However, considering the diversity of genotypes among A/H5N6 viruses, more extensive characterization and risk assessment of A/H5N6 viruses of different genotypes is required in order to completely assess the risks. Moreover, owing to its pathogenicity in ferrets and its antigenic divergence from other 2.3.4.4 A/H5 Gs/Gd viruses, A/H5N6 GZ/14 virus could be used as a challenge candidate for vaccination/ challenge studies to assess the ability of different vaccination strategies to elicit protective responses to emerging Gs/Gd HPAI viruses.
Viruses. All eight gene segments of a human virus isolate, A/H5N6 A/Guangzhou/39715/2014 (GZ/14), were amplified from copy DNA by PCR using specific primers and cloned in a modified version of the bidirectional reverse genetics plasmid pHW2000 (44). Recombinant viruses were rescued by reverse genetics upon transfection of 293T cells as previously described (44). Titrations. MDCK cells were inoculated with 10-fold serial dilutions of virus stocks, nose swabs, throat swabs, or homogenized tissue samples. The cells were washed with phosphate-buffered saline (PBS) 1 h after inoculation and cultured in infection medium, consisting of EMEM supplemented with 100 U/ml PEN, 100 U/ml STR, 2 mM L-Glu, 1.5 mg/ml NaHCO 3 , 10 mM HEPES, 1ϫ NEAA, and 20 g/ml trypsin (N-tosyl-L-phenylalanine chloromethyl ketone [TPCK]-treated trypsin; Sigma). Three days after inoculation, supernatants of cell cultures were tested for agglutinating activity using TRBCs as an indicator of virus replication. Infectious virus titers were calculated from 4 replicates each of the homogenized tissue samples, nose swabs, and throat swabs and from 10 replicates of the virus stocks by the method of Reed and Muench (45).
Ferret experiments. Ferrets were housed and experiments were conducted in strict compliance with European guidelines (European Union [EU] directive on animal testing 86/609/EEC) and Dutch legislation (Experiments on Animals Act, 1997). All animal experiments were approved by the independent animal experimentation ethical review committee "stichting dier experimenten commissie (DEC) consult" (Erasmus MC permit number EMC4028). Research projects involving laboratory animals can be executed only if they are approved by the animal experiments committee (DEC). The DEC considers the application and pays careful attention to the effects of the intervention on the animal and its discomfort and weighs this against the social and scientific benefit to humans or animals. The researcher is required to keep the effects of the intervention to a minimum, based on the three R's (refinement, replacement, reduction).
(i) Ferret transmission experiments. Four ferrets were inoculated intranasally with a total dose of 10 6 TCID 50 of virus distributed into the two nostrils (250 l per nostril). Each donor ferret was then placed in a transmission cage. One day after inoculation, one naive recipient ferret was placed opposite of each donor ferret. Each transmission pair was housed in a separate transmission cage designed to prevent direct contact but to allow airflow from the donor ferret to the recipient ferret. Nose and throat swabs were collected at 1, 3, 5, and 7 days postinoculation (dpi) from donor ferrets and at 1, 3, 5, 7, and 9 days postexposure (dpe) from the recipient ferrets and subjected to endpoint titration in MDCK cells. Donor ferrets were euthanized at 7 dpi and recipient ferrets were euthanized at 14 dpe to allow assessment of seroconversion.
(ii) Ferret pathogenesis experiments. Six ferrets were inoculated intranasally with a total dose of 10 6 TCID 50 of A/H5N6 GZ/14 by applying 250 l of virus suspension dropwise to each nostril. Clinical scores in all groups were assessed every day. Activity status was scored as follows: 0, alert and playful; 1, alert and playful only when stimulated; 2, alert but not playful when stimulated; 3, neither alert nor playful when stimulated. Dyspnea was characterized by open-mouth breathing with exaggerated abdominal movement (46). Body weight was recorded daily. Throat and nose swabs were collected every day and were stored at Ϫ80°C in transport medium (Hanks' balanced salt solution containing 0.5% lactalbumin [Sigma], 10% glycerol [Sigma], 200 U/ml PEN, 200 mg/ml STR, 100 U/ml polymyxin B sulfate [Sigma], and 250 mg/ml gentamicin [Gibco]) for endpoint titration in MDCK cells. At 3 and 6 dpi, three animals from each group were euthanized by exsanguination under anesthesia and were necropsied according to a standard protocol (46). Tissues harvested for virological examination (right nasal turbinates, trachea, right bronchus, right lungs, tracheobronchial lymph nodes, right tonsils, right olfactory bulb, right cerebrum, right cerebellum, heart, liver, spleen, kidney, adrenal gland, pancreas, and jejunum) were homogenized in transport medium using the FastPrep system (MP Biomedicals) with two onequarter-inch ceramic sphere balls, centrifuged 1,500 ϫ g for 10 min, aliquoted, and stored at Ϫ80°C for endpoint titration in MDCK cells. Tissues harvested for histological examination (nasal turbinates, larynx, trachea, left bronchus, left lung, tracheobronchial lymph nodes, heart, liver, spleen, kidney, adrenal gland, esophagus, duodenum, jejunum, colon, pancreas, mesenteric lymph node, left olfactory bulb, left cerebrum, left cerebellum, and eye lid) were fixed in 10% neutral buffered formalin, embedded in paraffin, sectioned at 4 m, and stained with hematoxylin and eosin (HE) for examination by light microscopy.
In a second experiment, three ferrets were inoculated intratracheally with 3 ml of virus dilution containing 10 6 TCID 50 . The experimental design and procedures were similar to those performed for the intranasal inoculation.
Immunohistochemistry (IHC). For detection of influenza A virus antigen, sequential slides of all tissues were stained with a primary antibody against the influenza A virus nucleoprotein (NP) as described previously (46). In each staining procedure, an isotype control was included as a negative control, and a lung section from a cat infected experimentally with A/H5N1 was used as a positive control.
Resialylated turkey red blood cell assay. All sialic acids (SA) were removed from the surfaces of TRBCs by incubating 1% TRBCs in PBS with 50 mU of Vibrio cholerae NA (VCNA) (Roche) in 8 mM calcium chloride (CaCl 2 ) at 37°C for 1 h. After two washing steps with PBS, VCNA-treated TRBCs were resuspended in PBS containing 1% bovine serum albumin (BSA) to a final concentration of 0.5% TRBCs. Removal of SA was confirmed by observation of a complete loss of hemagglutination of the TRBCs by control influenza A viruses. Subsequently, resialylation was performed using 0.5 mU of ␣2,3-N-sialyltransferase (Sigma) or 25 mU of ␣2,6-N-sialyltransferase (Sigma) and 1.5 mM CMP-sialic acid (Merck) at 37°C for 2 h in a final volume of 75 l to produce ␣2,3-TRBCs and ␣2,6-TRBCs, respectively. After two washing steps, the TRBCs were resuspended in PBS containing 1% BSA to a final concentration of 0.5% TRBCs. Resialylation was confirmed by hemagglutination of viruses with known receptor specificity: A/H5N1 IN/05 and A/H3N2 NL/13. The receptor specificity of A/H5N8 ck/NL/14 and A/H5N6 GZ/14 was tested by performing a standard hemagglutination assay with the modified TRBCs.
Syncytium formation assay. Influenza virus HA-induced cell fusion was tested in Vero-118 cells transfected with 5 g of HA cloned in pCAGGs expression plasmid using Xtremegene transfection reagent according to the manufacturer's instructions (Roche). One day after transfection, the cells were harvested using trypsin-EDTA and plated in 24-well plates. The next morning, the cells were washed, and medium was replaced with IMDM (Lonza) containing 10 g/ml of trypsin. After 1 h, the cells were washed with PBS and exposed to PBS at pH 4.8, 4.9, 5.0, 5.1, 5.2, 5.3, 5.4, 5.5, 5.6, 5.7, 5.8, or 5.9 for 10 min at 37°C. Subsequently, the PBS was replaced by IMDM supplemented with 10% FBS, 100 U/ml PEN, 100 mg/ml STR, and 2 mM L-Glu. The next day, the cells were fixed using 80% ice-cold acetone, washed, and stained using a 20% Giemsa solution (Merck Millipore, Darmstadt, Germany). Visual inspection of the cell cultures for the presence of syncytia (multinucleated cells) was used to determine the pH threshold required for fusion.
Thermostability assay. In short, viruses were diluted to 128 HA units (HAU)/25 l using PBS. The samples were incubated in a thermal cycler for 30 min at temperatures of 50°C, 51.7°C, 54.3°C, 56°C, 58.5°C, and 60°C. Subsequently, the HA titer was determined by performing a hemagglutination assay using TRBCs.
Polymerase assay. 293T cell monolayers were transfected with 125 ng of luciferase reporter plasmid (firefly open reading frame flanked by the noncoding regions of the M segment under control of a PolI promoter) and 12.5 ng of an internal control plasmid (Renilla gene under the cytomegalovirus [CMV] promoter) together with the mix of PB2, PB1, PA, and NP plasmids, cloned in a bidirectional reverse genetics plasmid, in quantities of 125, 125, 125, and 250 ng, respectively. The transfected cells were incubated at 33°C or 37°C. After 24-h incubation, the supernatants were discarded, and the cell extracts were prepared in 100 l of lysis buffer. The luciferase levels were assayed with luciferase assay system (Promega) and detected by a luminometer. Polymerase complexes of control viruses were included in the assay: a human H1N1 virus A/Hong Kong/54/1998 (HK/98), two human A/H5N1 viruses A/Vietnam/ 1203/2004 (VN/1203/04) and A/Indonesia/5/05 (IN/05), and an avian virus A/H5N2 A/mallard/Netherlands/3/1999 (ml/NL/99).
Hemagglutination inhibition assay. Ferret antisera were prepared by intranasal inoculation with 500 l recombinant virus containing six gene segments of PR/34 (47), the corresponding HA without the MBCS, and the corresponding NA. Fourteen days postinoculation, ferrets were boosted by injecting subcutaneously 500 l of a 1:1 mix of concentrated virus (Ͼ2,048 HAU) and Titermax Gold adjuvant (Sigma-Aldrich, St. Louis, MO, USA). Antisera were collected 2 weeks later. Ferret antisera were then pretreated overnight with receptor-destroying enzyme (VCNA) at 37°C and incubated at 56°C for 1 h the next day. Antisera were subsequently treated with 10% TRBCs at 4°C for an hour. Twofold serial dilutions of the antisera, starting at a 1:20 dilution, were mixed with 25 l of a virus stock containing 4 hemagglutinating units and were incubated at 37°C for 30 min. Subsequently, 25 l of 1% TRBCs was added, and the mixture was incubated at 4°C for 1 h. Hemagglutination inhibition (HI) was read and was expressed as the reciprocal value of the highest dilution of the serum that completely inhibited agglutination of virus and erythrocytes. The detection limit corresponds to half of the first antisera dilution (Ͻ10).
Biosafety. All experiments were conducted within the enhanced animal biosafety level 3 (ABSL3ϩ) facility of Erasmus MC. The ABSL3ϩ facility consists of a negative-pressure laboratory (Ϫ30 Pa) in which all in vivo and in vitro experimental work is carried out in class 3 isolators or class 3 biosafety cabinets, which also have negative pressure (less than Ϫ200 Pa). Although all experiments are conducted in closed class 3 cabinets and isolators, special personal protective equipment, including laboratory suits, gloves, and FFP3 facemasks, is used. Air released from the class 3 units is filtered by high-efficiency particulate air (HEPA) filters and then leaves via the facility ventilation system, again via HEPA filters. Only authorized personnel that have received the appropriate training can access the ABSL3ϩ facility. For animal handling in the facilities, personnel always work in pairs. The facility is secured by procedures recognized as appropriate by the institutional biosafety officers and facility management at Erasmus MC and Dutch and U.S. government inspectors.