|Home | About | Journals | Submit | Contact Us | Français|
Influenza viruses attach to cells via a sialic acid moiety (sialic acid receptor) that is α2-3 linked or α2-6 linked to galactose (α2-3SAL or α2-6SAL); sialic acid acts as a receptor for the virus. Using lectin staining, we demonstrated that the α2-6SAL configuration is predominant in the respiratory tract of ferrets, including trachea, bronchus, and lung alveolus tissues. Recombinant wild-type (rWT) influenza A/Solomon Island/3/06 (SI06) (H1N1) viruses were constructed to assess the impact of the hemagglutinin (HA) variations (amino acids 190 or 226) identified in natural variants on virus replication in the upper and lower respiratory tract of ferrets, as well as virus antigenicity and immunogenicity. A single amino acid change at residue 226 (from Gln to Arg) in the HA of SI06 resulted in the complete loss of binding to α2-6SAL and a concomitant loss of the virus's ability to replicate in the lower respiratory tract of ferrets. In contrast, the virus with Gln226 in the HA protein has a receptor binding preference for α2-6SAL and replicates efficiently in the lungs. There was a good correlation between viral replication in the lungs of ferrets and disease symptoms. In addition, we also showed that the 190 and 226 residues affected viral antigenicity and immunogenicity. Our data emphasize the necessity of thoroughly assessing wild-type influenza viruses for their suitability as reference strains and for carefully selecting the HA antigen for vaccine production during annual influenza vaccine evaluation processes.
Influenza viruses cause recurring, highly contagious respiratory disease in humans, resulting in approximately 36,000 deaths (primarily in the elderly) in the United States annually. Influenza viruses continually undergo genetic changes as they replicate in humans, resulting in antigenic drift. Infection with seasonal influenza virus is mainly manifested by respiratory tract infection, with high morbidity and mortality in individuals who have poor pre-existing immunity to influenza viruses, mainly the very young, and in people with inadequate immune responses, such as the elderly. The introduction of new hemagglutinin (HA) subtypes, the 1918-1919 “Spanish” influenza virus (H1N1) and the 1957 (H2N2) and 1968 (H3N2) viruses, caused three pandemics in the last century, as well as the influenza pandemic of 2009 caused by 2009 swine origin H1N1 influenza virus.
Virus pathogenicity is determined by the interplay between virus and host. Multiple factors determine virus virulence, including the ability of the virus to trigger innate immunity (38), the ability to induce degradation of host RNA polymerase II enzyme (31), apoptosis mediated by the PB1-F2 protein (5, 7), contributions of the viral RNA polymerase proteins and the NS1 and M2 proteins (17, 36, 42), and viral receptor binding preference (19, 27).
The binding of influenza viruses to their target cells is mediated by the viral HA protein, which recognizes cell surface glycoconjugate receptors that terminate in sialic acid residues; the sialic acid residues may be α2-3 linked (α2-3SAL) or α2-6 linked (α2-6SAL). Human influenza viruses readily bind to the α2-6SAL receptor on the human respiratory tract epithelium (9, 21). The lack of sustained human-to-human transmission of avian H5N1 viruses is likely due to their α2-3SAL binding preference (34, 40). Ferrets have been proven to be a suitable model to study influenza viruses for virus pathogenicity, transmissibility, attenuation, immunogenicity, and protective efficacy of the vaccine strains and for evaluation of antiviral drugs. The animal is a permissive host for all influenza A and B viruses and develops febrile illness and symptoms that are similar to those of humans. Ferrets also have marked similarities to humans in lung physiology, airway morphology, and cell types present in the respiratory tract, including the distribution of the α2-6SAL receptor for human influenza viruses (26, 41).
During annual influenza vaccine production processes, the HA and neuraminidase (NA) gene segments of wild-type (WT) influenza viruses are reassorted with a vaccine donor strain that donates the six internal protein gene segments (PB1, PB2, PA, NP, M, and NS) required for virus replication. Cold-adapted (ca) A/Ann Arbor/6/60 is used as the master donor virus for live attenuated influenza A virus vaccine strains that have the characteristic cold-adapted, temperature-sensitive, and attenuation phenotypes. The A/Puerto Rico/8/34 (H1N1) virus is normally used as the donor virus for inactivated influenza A virus vaccines. WT influenza viruses are routinely used as reference strains for strain surveillance and for vaccine strain evaluation to ensure that those influenza vaccines are immunogenic and antigenically similar to WT viruses. During the egg expansion process, the HA gene of WT influenza viruses may mutate, allowing the viruses to replicate more efficiently in eggs. The egg-adaptation mutations normally result in amino acid changes in the receptor binding sites to allow virus to bind to α2-3SAL better, but the egg adaptation mutations may also affect virus antigenicity (3, 10, 20).
In this study, we found that WT A/Solomon Island/3/06 (WT SI06) virus that was expanded in eggs contained amino acid variations in HA residues 190 and 226. To evaluate the impact of these amino acid variations on virus receptor binding preference, virus replication in the animal host, immunogenicity, and antigenicity, we constructed recombinant A/Solomon Island/3/06 (rWT SI06) viruses containing single or double amino acid substitutions. We found that a single amino acid mutation at residue 226 of the SI06 HA protein affects not only virus receptor binding preference but also the ability of the virus to replicate in the lower respiratory tract of ferrets. Using lectin staining, we showed that the respiratory tract of ferrets contains predominantly α2-6SAL; the viruses with strong affinity for α2-6SAL replicated efficiently in the lungs, whereas viruses that bound exclusively to α2-3SAL were unable to replicate in the lungs. Our data indicate that virus receptor binding preference plays an important role in viral pathogenicity, immunogenicity, and antigenicity.
The biologically derived WT SI06 (bWT SI06) received from the Centers for Disease Control (CDC) was previously amplified three times in embryonated chicken eggs. The 8 gene segments of bWT SI06 were amplified by reverse transcription (RT)-PCR and cloned into expression vector pAD3000 (15). Each cDNA clone was completely sequenced and compared with the sequence obtained from sequencing of RT-PCR-amplified cDNA using viral RNA as the template. Mutagenesis of plasmids was performed by using a QuikChange site-directed mutagenesis kit (Agilent, La Jolla, CA), and sequence changes were confirmed by sequencing. Recombinant viruses were rescued by transfection of cocultures of 293T cells and Madin-Darby canine kidney (MDCK) cells with 8 plasmids encoding the 8 genomic cDNAs of WT SI06, and the rescued virus was designated rWT SI06. All of the viruses were propagated in the allantoic cavities of 10- to 11-day-old embryonated chicken eggs. The genomic sequence of the recombinant virus was verified by cDNA sequencing.
Chicken red blood cells (cRBCs) (HEMA Resource and Supply, Inc.) were resialylated as previously described (25). One hundred microliters of 10% cRBCs was incubated with 50 mU Vibrio cholerae neuraminidase (Sigma, St. Louis, MO) at 37°C for 1 h to remove sialic acid from cRBCs. Complete removal of sialic acid from cRBCs was confirmed by the lack of hemagglutination. Subsequently, desialylated cRBCs were incubated with 2.5 mU of α2-3(N)-sialyltransferase (Calbiochem, La Jolla, CA) or 2 mU of α2-6(N)-sialyltransferase (Calbiochem, La Jolla, CA) and 1.5 mM cytidine-5′-monophospho-N-acetylneuraminic acid (Sigma, St. Louis, MO) for 1.5 h at 37°C. The resialylated cRBCs were suspended (0.5% [vol/vol]) in phosphate-buffered saline (PBS; pH 7.2 to 7.6). Hemagglutination assays were performed in V-bottom microtiter plates. Fifty microliters of 2-fold serially diluted virus was incubated at room temperature for 1 h with 50 μl of 0.5% cRBCs that had been resialylated (α2-3 or α2-6). The hemagglutination titer was defined as the reciprocal of the highest virus dilution that hemagglutinated cRBCs.
Eight- to 10-week-old male and female ferrets (n = 3/group) from Simonsen Laboratories (Gilroy, CA) or Triple F Farms (Sayre, PA) were used to assess virus replication in the respiratory tract and to evaluate virus immunogenicity. Ferrets were housed individually and inoculated intranasally with 7.0 log10 PFU of virus per 0.2-ml dose. Three days after infection, ferrets were euthanized, and the lungs and nasal turbinates were harvested. Virus titers in the lungs and nasal turbinates were determined by the 50% egg infectious dose (EID50) assay and expressed as 50% egg infectious dose per gram of tissue (log10 EID50/g). Virus-infected ferrets were monitored for body temperature per rectum using a Welch Allyn's Sure Temp thermometer (Welch Allyn, Skaneateles Falls, NY) and observed for clinical symptoms, including nasal symptoms, stool changes, and activity, twice daily. Each ferret's body weight was measured daily. Nasal symptoms were defined by scores 0 to 3 (0, no symptoms were observed; 1, nasal rattling; 2, nasal discharge on external nares; and 3, mouth breathing) as described previously (28). Stool change scores were defined by scores 0 to 3 (0, firmed/normal; 1, soft/watery; 2, mucous/yellow/green; and 3, bloody or absent). Activity scores were also defined by scores 0 to 3 (0, fully playful; 1, only responded to play overtures and did not initiate any play activity; 2, alert but not at all playful; and 3, neither playful nor alert). Ferrets assigned to the immunogenicity studies were bled on days 14 and 28 postinfection, and sera collected on day 14 were assayed for antibody titers by the hemagglutination inhibition (HAI) and microneutralization assays.
Serum antibody levels in postinfection ferret sera against homologous and heterologous viruses were determined by HAI and microneutralization assays. Sera collected on day 14 after intranasal infection were treated with receptor-destroying enzyme (RDE; Denka-Seiken, Tokyo, Japan) at 37°C overnight and heat inactivated at 56°C for 45 min. Treated serum samples (25 μl) were serially diluted 2-fold and incubated with 25 μl of virus containing 8 hemagglutination units in V-bottom microtiter plates for 1 h, followed by incubation with 50 μl of 0.5% turkey RBCs for 45 min. The HAI titer was defined as the reciprocal of the highest serum dilution that inhibited hemagglutination. The microneutralization assay was performed with the RDE-treated serum. Serum serially diluted 2-fold in Eagle's minimal essential medium (EMEM) containing 1 μg/ml of tosylsulfonyl phenylalanyl chloromethyl ketone (TPCK)-trypsin was incubated with 100 μl of virus containing 100 50% tissue culture infective doses (TCID50s) in 96-well U-bottom microplates for 1 h at 33°C. The antiserum-virus mixtures were transferred to MDCK cells in 96-well plates and incubated at 33°C for 4 days. The cytopathic effect (CPE) was observed under a microscope, and the microneutralization titer was defined as the reciprocal of the highest serum dilution that inhibited 50% of the CPE caused by virus infection.
For histology studies, respiratory tract tissues of ferrets, including nasal turbinate (NT), trachea, and lungs (inflated with 10% formalin), were collected and fixed in 10% formalin for a minimum of 24 h. Paraffin-embedded tissue sections were prepared by Cureline, Inc. (Burlingame, CA). The tissue sections were deparaffinized with xylene and dehydrated by using graded alcohols. The sections were then blocked with 3% H2O2 for 15 min, followed by incubation with Carbo-Free blocking solution (Vector Laboratories, Burlingame, CA) at room temperature for 1 h.
Ferret tissue sections were incubated with goat anti-influenza A virus polyclonal antibody (Millipore, Bedford, MA) diluted 1:40 in blocking solution at room temperature for 1 h, followed by incubation with horseradish peroxidase (HRP)-conjugated rabbit anti-goat Ig (Dako, Carpenteria, CA) diluted 1:100 in blocking solution at room temperature for 30 min. The sections were washed 3 times with PBST (0.05% Tween 20 in PBS, pH 7.2) and developed with 3-amino-9-ethylcarbazole (AEC) substrate (Dako) for influenza virus antigen (red) and hematoxylin for nuclear staining (blue-violet).
To block all endogenous avidin/biotin binding sites, ferret tissue sections were treated with an avidin/biotin blocking kit (Vector Laboratories). The sections were then incubated with 10 μg/ml biotinylated agglutinins from Maackia amurensis (MAA1 and MAA2) or biotinylated Sambucus nigra lectin (SNA) (Vector Laboratories) in blocking solution at room temperature for 1 h, followed by incubation with HRP-streptavidin (Vector Laboratories) diluted 1:300 in blocking solution for 30 min. The sections were subsequently developed with AEC substrate for lectin distribution (red) and hematoxylin (Vector Laboratories) for nuclear staining (blue-violet).
Virus labeling was done as previously described (41). Briefly, influenza viruses amplified in embryonated chicken eggs were harvested and clarified by filtration through a 0.22-μm filter (Millipore, Bedford, MA). The viruses were then purified through a 20 to 60% sucrose gradient and suspended in 2 ml PBS. The purified viruses were inactivated by incubation with an equal volume of 10% formalin at room temperature for 1 h, followed by dialysis in PBS overnight. The viruses were mixed with an equal volume of 0.1 mg/ml fluorescein isothiocyanate (FITC) (Fisher Scientific, Pittsburgh, PA) in 0.5 M bicarbonate buffer (pH 9.5) for 1 h with constant stirring, and unbound FITC was subsequently removed by overnight dialysis in PBS.
Formalin-inactivated FITC-labeled viruses were diluted with PBS to a titer of 50 to 100 hemagglutination units per 50 μl. The viruses were then incubated with ferret tissue sections overnight at 4°C, followed by incubation with HRP-conjugated rabbit anti-FITC antibody (Dako) diluted 1:50 in blocking solution at room temperature for 1 h. The tissue sections were subsequently developed with AEC for influenza virus antigen (red) and hematoxylin for nuclear staining (blue-violet).
A/Solomon Island/3/06 (H1N1) received from the CDC had been amplified 3 times in embryonated chicken eggs, and its genomic sequence was determined by sequence analysis. Recombinant WT A/SI/3/06 (rWT SI06) virus was produced from the transfected 8 cDNA plasmids cloned from biologically derived A/Solomon Island/3/06 (bWT SI06) and represented the consensus sequence of bWT SI06. The genomic sequence of rWT SI06 was confirmed to be identical to that of bWT SI06. Both rWT and bWT SI06 virus replicated efficiently in embryonated chicken eggs, with titers of >8.0 log10 PFU/ml.
The replication of bWT and rWT SI06 in the respiratory tract of ferrets was examined. Ferrets were administered 107 PFU of either bWT or rWT SI06 intranasally, and at three days postinfection, virus titers in the nasal turbinates (NT) and lungs were determined by 50% egg infectious dose (EID50) assay. Both bWT and rWT SI06 replicated efficiently in the upper respiratory tract of ferrets at levels of 5.8 and 5.4 log10 EID50/g NT tissues, respectively. However, in contrast to bWT SI06, which replicated at a titer of 5.5 log10 EID50/g tissue in the lungs, rWT SI06 was not detected in the lungs of infected ferrets. To determine if bWT SI06 recovered from lungs of ferrets was identical to the virus inoculum, bWT SI06 isolated from NT and lungs was sequenced. The virus recovered from the lung tissue was found to contain two amino acid changes in the HA protein; at residue 186 (190 in the H3 numbering), there was an amino acid change from aspartic acid to alanine (D190A), and at residue 222 (226 in the H3 numbering), there was an amino acid change from arginine to glutamine (R226Q). The virus recovered from the NT tissue had mixed residues at these two positions (Table (Table11).
To determine whether residues 190 and 226 (in H3 numbering) of the HA protein of bWT SI06 isolated from lung or NT tissues were mutated in vivo or were selected from the existing heterologous viral pool, bWT SI06 was subjected to plaque assay and the HA gene from each individual plaque was sequenced. As shown in Table Table1,1, from 48 plaques sequenced, the majority (90%) of plaques (designated plaque A) had D190 and R226 (DR) residues in the HA protein; 10% of the plaques (a total of 5 plaques, designated plaques B, C, and D) had Q226 in the HA protein, but their residue at 190 was V, N, or A. The genetic stability of these bWT SI06 variants was examined by 3 to 4 rounds of passages in eggs. Except for plaque B that had V190 and Q226 changed to D190 and R226 after its replication in eggs, the rest of the variants maintained their HA protein sequences. Since virus with A190 and Q226 (AQ) was present in the inoculated viral population of bWT SI06 and rWT SI06-DR could not replicate in ferret lungs, we concluded that bWT SI06-AQ isolated from infected ferret lungs was selected from the virus inoculum, not derived from in vivo mutations. This speculation was further supported by the HA sequence analysis of the H1N1 viruses that circulated before, at the same time as, and after A/Solomon Island/3/06 virus. A/Beijing/262/95, A/New Caledonia/20/99 (A/NC/20/99), A/HongKong/2652/06 (A/HK/2652/06), A/St Petersburg/8/06 (A/S.P/8/06), and A/South Dakota/6/07 (A/SD/6/07) are egg isolates, and A/Singapore/23/04 and A/Canada/591/04 are MDCK isolates (Fig. (Fig.1).1). All of these strains have Q226 but with D or N at 190. A/New Caledonia/20/99 vaccine strain (ca) has the N190D change, and A/SI/3/06 ca has the same HA sequence as the HA of the bWT SI06 virus.
To further examine which amino acid was critical for virus to replicate in the lower respiratory tract of ferrets, the replication in ferrets of bWT SI06 and the three bWT SI06 HA variants was evaluated (Table (Table2).2). All viruses replicated efficiently in the upper respiratory tract of ferrets. The titer of bWT SI06-DR (4.4 log10 EID50/g) in the NT was approximately 10-fold lower than those of SI06-AQ (5.5 log10 EID50/g) and bWT SI06-NQ (5.6 log10 EID50/g), and its replication was not detected in the lungs. In contrast, both bWT SI06-NQ and -AQ replicated efficiently in the lungs (5.4 and 5.1 log10 EID50/g).
To confirm that the R226 residue indeed rendered the virus unable to replicate in ferret lungs, rWT SI06 variants that had DR, AR, DQ, or AQ at residues 190 and 226 were produced and evaluated for their ability to replicate in the upper and lower respiratory tract of ferrets (Table (Table2).2). These four variants replicated efficiently in the upper respiratory tract of ferrets, at levels of 5.8, 6.1, 7.2, and 6.9 log10 EID50/g NT tissue, respectively, although the titers of the viruses with DR or AR were more than 10-fold lower than those with DQ and AQ. rWT SI06-DR and -AR were not detected in the lungs, and rWT SI06-DQ and -AQ replicated at titers of 6.3 log10 EID50/g tissue. Thus, these data demonstrate that residue Q226 is critical for the replication of WT SI06 virus in the ferret lungs.
The upper and lower respiratory tracts of ferrets infected with rWT SI06-AQ and rWT SI06-DR were examined by immunohistopathology (Fig. (Fig.2).2). Viral antigens were detected in the nasal cavity and nasopharynx of the ferrets infected with the AQ and DR variants. Viral antigens were barely detected in the trachea, bronchus, and alveoli of ferrets infected with rWT SI06-DR. In contrast, viral antigens were present in the epithelial cells and the glands of the trachea and bronchus and in pneumocytes of the alveoli in the animals infected with rWT SI06-AQ. Thus, consistent with the results from viral replication, the immunostaining results confirmed that rWT SI06-DR virus did not replicate in the lower respiratory tract of ferrets.
rWT SI06 HA variants were assessed for their ability to cause influenza illness in ferrets. Virus-infected ferrets were monitored for body weight gain/loss, body temperature, nasal symptoms, stool changes, and activity (Fig. (Fig.3).3). Ferret body weight gain was the greatest for the PBS control group, with a mean body weight gain of 8.5% on day 3. Ferrets infected with rWT SI06-DR did not lose body weight and had a maximal body weight gain of 7.5% on day 3. Ferrets infected with rWT SI06-AQ and -DQ lost weight on day 1 but gradually gained weight and had maximal body weight gains of 2.2 to 2.7%. These ferrets also had fevers of greater than 39.8°C at 30 h postinfection. Ferrets infected with rWT SI06-DR had temperatures that were higher than those of the PBS control group at 48 to 54 h but did not reach 39.8°C. Except for the PBS group that had a score of zero, some of the virus-infected ferrets had low scores (1 to 2) for nasal symptoms, stool changes, and activity changes; however, these scores did not correlate with temperature and body weight changes. The data obtained for body weight changes and fever demonstrated that viral replication in the lower respiratory tract correlated with disease symptoms.
To examine whether residues 190 and 226 of the SI06 HA protein could affect virus antigenicity and immunogenicity, rWT SI06 variants DR, AQ, AR, and DQ were evaluated for their immunogenicity and cross-reactivity (Table (Table3).3). Serum antibody titers were measured by the HAI and microneutralization assays, and both assays provided comparable results. Overall, rWT SI06-DR-infected ferrets had lower serum antibody titers than the animals infected with other variants, but their sera cross-reacted well with the other three virus variants. The HAI and neutralizing (Nt) antibody titers in sera of ferrets infected with rWT SI06-DQ were about 2-fold higher than the titers in sera of ferrets infected with the DR variant, and their sera reacted well with the other three virus variants. While sera of ferrets infected with rWT SI06-AQ had the highest antibody titers to the homologous virus, they had 7- to 8-fold or 14- to 16-fold lower antibody titers against viruses with DR and DQ, respectively, and yet their titers against the virus with AR were similar to their titers against the homologous virus. rWT SI06-AR serum had antibody titers of 4,096 (HAI) and 6,400 (Nt) and reacted well with the virus with AQ but less well with the viruses with DR or DQ.
The impact of the 190 and 226 residues on the immunogenicity of ca SI06 vaccine virus was also investigated. Ferrets in groups of three were immunized with ca SI06-DR and ca SI06-AQ, and the levels of serum HAI and Nt antibodies collected 14 days postvaccination were determined (Table (Table3).3). Consistent with the data obtained with WT SI06 variants, ca SI06-AQ induced much higher levels of HAI and Nt antibody titers than ca SI06-DR (7- to 8-fold difference). ca SI06-DR antiserum cross-reacted well with the AQ and AR viruses but had a 4-fold reduction against DQ virus relative to the homologous titer. ca SI06-AQ antiserum had very high homologous antibody titers (HAI titer of 6,502 and Nt titer of 14,769) and reacted with the virus with AR, but its reactivity to DR and DQ viruses was reduced by up to 20- and 5-fold, respectively. Thus, the 190 and 226 residues had significant impacts on vaccine virus antigenicity and immunogenicity.
HA protein residue 226 in H3N2 viruses has been previously reported to affect viral preference for receptor binding (33) and influence the severity of illness in ferrets (19). To determine if HA protein residue 226 in the H1N1 virus also affects the receptor binding preference, the H1N1 variants were analyzed by hemagglutination assay using RBCs that contained either α2-3SAL or α2-6SAL. As shown in Table Table4,4, all the variants hemagglutinated untreated cRBCs that contained both α2-6SAL and α2-3SAL. Desialylation removed both types of receptors from the cRBCs, and none of the viruses bound to desialylated RBC. Interestingly, rWT SI06-DR and -AR were only able to bind α2-3SAL-resialylated RBCs. Although rWT SI06-AQ and -DQ both bound α2-6SAL, AQ lost the ability to bind α2-3SAL. rWT SI06-DR and -AQ had distinctive receptor binding preferences, exclusively to α2-3SAL or α2-6SAL, respectively. Thus, amino acid change Q226R abolished virus receptor binding specificity to α2-6SAL.
The distribution of α2-6SAL and α2-3SAL in the respiratory tract of ferrets is similar to their distribution in humans, as reported previously (41). To determine if receptor distribution in the lower respiratory tract of ferrets correlated with viral replication, tracheal, bronchial, and alveolar tissues were stained with MAA1/MAA2 lectins that recognize α2-3SAL and SNA lectin that recognizes α2-6SAL, respectively. To ensure that the three lectins could bind to the respective sialic acid, MDCK cells that contain both α2-6SAL and α2-3SAL were used as the lectin binding assay control (Fig. (Fig.4).4). MAA1 stained the trachea and bronchus poorly, but MAA2 stained the lamina propria or connecting tissue and submucosal areas of the trachea and bronchus well. Both MAA1 and MAA2 stained pneumocytes in alveoli modestly. In contrast, SNA bound strongly to the ciliated cells and submucosal glands of the trachea and bronchus and to pneumocytes and the walls of alveoli. Thus, α2-6SAL is more abundant in the lower respiratory tract of ferrets than α2-3SAL.
rWT SI06-AQ and rWT SI06-DR were further examined for their capacity for binding to the respiratory tissues by using FITC-labeled viruses (Fig. (Fig.5).5). Both viruses were prepared at the same time and under the same conditions. We reproducibly showed that FITC-labeled rWT SI06-AQ bound to the MDCK cells more intensely than rWT SI06-DR, which correlated with the receptor distribution in the MDCK cells. rWT SI06-AQ had strong binding avidity for the submucosal glands of the bronchus, which contain α2-6SAL, as shown earlier, and for pneumocytes of alveoli. rWT SI06-DR bound weakly to the connecting tissue in the bronchus and to the pneumocytes in alveoli. We cannot exclude the possibility that the rWT SI06-DR had poor binding in this assay. Nevertheless, these data confirmed that viral binding correlated well with the distribution of α2-6SAL and α2-3SAL receptors in the respiratory tract. Moreover, the receptor binding preference played a critical role in viral replication in the lungs.
In this study, we showed that the α2-6SAL receptor is much more abundant in the lower respiratory tract of ferrets than α2-3SAL. Thus, WT SI06 virus with a binding preference for α2-6SAL replicated efficiently in the lungs. In contrast, virus with amino acid mutations in the receptor binding regions that confer binding specificity to α2-3SAL did not replicate in the lungs. Replication of virus in ferret lungs correlated with disease symptoms. Hence, the receptor binding preference of an influenza virus is a critical factor that determines viral replication in the lower respiratory tract and viral pathogenicity.
Single amino acid changes of residue 190 or 226 in the HA protein receptor binding region have been previously identified as responsible for the receptor binding preference for either human-like or avian-like receptor (12, 21, 29, 30, 32, 33). By introducing different amino acid variations into the HA protein of rWT SI06 viruses, we demonstrated that only the virus with Q226 in the HA protein could replicate in the lung and that Q226 is a critical residue to enable the virus to replicate in ferret lungs. In contrast to the H1N1 viruses, where Q226 confers binding to α2-6SAL, Q226 confers binding to α2-3SAL for the H3N2 viruses (24, 33). The Q226L change switched the binding specificity of the H3N2 viruses from α2-6SAL to α2-3SAL and influenced the severity of illness of infected ferrets (19). Our recent study of the swine origin 2009 pandemic H1N1 virus also demonstrated that the Q226R change resulted in the loss of α2-6SAL binding (6). Thus, H1N1 viruses with the Q226R change have a different receptor binding preference than the H2, H3, H4, and H9 subtypes of influenza viruses (2, 9, 21, 22, 24). Residue 190 was reported to be important for the receptor binding preference of the 1918 H1N1 pandemic virus and avian H1N1 viruses (12). The loss of α2-6SAL in the 1918 pandemic virus correlated with the loss of ferret-to-ferret transmission (39). We also found that residue 190 influences binding to α2-3SAL. SI06-AQ could not bind to α2-3SAL; in contrast, SI06-DQ could bind to α2-3SAL, indicating that residue 190 also affects the virus receptor binding preference.
The receptor distribution in the respiratory tract of ferrets has not been well studied. MAA and SNA lectins are normally used to determine the cellular localization of ligands by their differential binding of sialic acid-terminated oligosaccharides on the cell surface (24). In humans, conflicting results for receptor distribution in the respiratory tract have been reported; however, these could be a result of using different lectins in the studies (24). MAA1 and MAA2 both bind to α2-3SAL, and MAA1 also binds to non-sialic acid residues. Therefore, both MAA1 and MAA2 were used in our study. In contrast to SNA lectin, which binds avidly to cilial cells in the luminal surface of the airway epithelium, MAA1 and MAA2 bind to the epithelial surface of the trachea less well but bind avidly to the submucosal connective tissue. These results are in agreement with those reported previously for the distribution of receptors in the respiratory tract of ferrets (18, 19). In addition, we also showed that the submucosal glands and alveoli are rich in α2-6SAL, although α2-3SAL was also detected in the alveoli. Kirkeby et al. (18) pointed out that mucinous cells in the airways of ferrets may become metaplastic and change their expression of sialoglycans as a consequence of an influenza infection. However, Kirkeby et al. were unable to detect α2-6SAL and α2-3SAL in the alveoli. Shinya et al. (34) reported that human lungs contained α2-3SAL; they hypothesized that infection with the H5N1 virus, with a binding preference for the avian receptor, is thus restricted to lungs. We discovered that human alveoli also contain abundant levels of α2-6SAL (data not shown), which would allow human influenza virus to replicate efficiently in the lung tissue and cause pneumonia. Our study shows that WT SI06-DR could only bind α2-3SAL and that the virus replicated efficiently in the upper respiratory tract but was unable to replicate in the lungs. This result is different from the findings for the highly pathogenic H5N1 A/Vietnam/1203/2004 (VN04) virus. Despite its preference for the α2-3SAL receptor, VN04 replicates very efficiently in the lungs of infected ferrets (37). The H5N1 VN04 virus has multibasic amino acids in the HA1-HA2 cleavage site, enabling virus to replicate independent of exogenous trypsin and spread systematically. In addition, other viral proteins, such as the PB2, PB1-F2, and NS1 proteins of the VN04 virus, also contribute to virus virulence (4, 8, 17, 36). The H1N1 WT SI06 virus is far less virulent than the H5N1 VN05 virus. We speculate that the inability of the virus with the R226 residue to replicate in the lower respiratory tract is due to the paucity of α2-3SAL in the trachea and lungs. rWT SI06-AQ, with an α2-6SAL specificity, thus has the advantage of replicating in the cells of the trachea and lungs that contain α2-6SAL receptor. Lack of viral replication in the respiratory tract is not necessarily due to lack of viral binding to the cells; other factors may also contribute to virus pathogenicity. rWT SI06 variants, which have identical genomic sequences with the exception of a one or two amino acid difference in the HA protein, replicated drastically differently in ferret lungs, suggesting that the receptor binding preference of a virus is an important factor in determining viral replication in the host.
Human H1N1 influenza virus binds to both α2-6SAL and α2-3SAL, with a preference for α2-6SAL (11, 14). Human H3N2 influenza virus normally binds to α2-6SAL, whereas avian virus preferentially binds to α2-3SAL. Receptor binding preference has been proposed to be an important species barrier; it prevents avian viruses from transmitting to humans (1, 35). Vaccines against influenza virus are mainly produced in embryonated chicken eggs. However, some WT influenza viruses and reassortant vaccine strains do not grow well in eggs. To select variants that grow well in eggs, egg passage is routinely used. The egg adaptation process frequently introduces amino acid changes into the HA protein that affect viral receptor binding and sometimes even viral antigenicity (21, 29, 30, 32, 33). Egg adaptation of human influenza virus results in increased binding of α2-3SAL (avian-like receptor) and impaired ability to bind α2-6SAL (human-like receptor) (16, 21). Recently, Hensley et al. (13) reported that passage of an influenza virus in mice with preexisting immunity frequently selected a single amino acid change at residue 158, 246, or 156 in the HA globular domain which decreased cellular receptor binding avidity and affected viral antigenicity. Thus, influenza viruses evolve to be able to escape from the pressures of existing polyclonal antibodies or other growth pressures through single amino acid changes in the receptor binding region.
Sequence analysis of bWT SI06 virus revealed heterogeneity of the HA protein at residues 190 and 226 that represents sequence changes from viral passages in eggs. Although the virus with residues A190 and Q226 was present as a minor population as determined by sequencing analysis of single plaques, it was the only variant recovered from lung tissue of ferrets infected with bWT SI06. The virus with residues D190 and Q226 in the HA protein was also present in the WT virus pool and replicated well in ferret lungs, but it was not isolated from WT-virus-infected lungs. The original clinical isolate of bWT SI06 most likely contained residue Q226 with affinity to α2-6SAL. Q226 was conserved in H1N1 viruses that circulated in the human population from 1995 to 2007, as shown in Fig. Fig.1;1; however, residue 190 varied among the H1N1 natural isolates (see Fig. Fig.1).1). More recent H1N1 strains have N190 in the HA protein. A/NC/20/90, A/SD/6/07, and A/HK/2652/06 with Q226 replicate well in the lungs of ferrets (data not shown). Residue D, V, N, or A was found in the HA protein of bWT SI06 variants, indicating that residue 190 had a strong pressure to mutate. Some of these residues were the intermediate changes accruing in the egg adaptation process. WT SI06-DR, with a preference for avian-like receptor, must have been selected in the egg amplification process. The Q226R change in the earlier H1N1 influenza viruses that resulted in the loss of binding specificity for α2-6SAL has been reported by Mochalova et al. (23).
The virus with residues DR is less immunogenic than those with residues AQ, DQ, or AR as examined for the rWT SI06 variants, although its antiserum cross-reacts well with other HA protein variants. However, ferret sera generated against viruses with AQ or AR reacted less well with SI06-DR and -DQ. Live attenuated ca SI06-DR vaccine virus replicated slightly less well than ca SI06-AQ in the upper respiratory tract of ferrets and was not as immunogenic as ca SI06-AQ. A/HK/2562/06 (A/Solomon Island/3/06-like strain) that contained N190 and Q226 was also very immunogenic and had a 4-fold reduction in reactivity with rWT SI06-DR relative to the homologous titer against A/HK/2562/06 (data not shown). Because WT SI06 egg isolate with DR residues in the HA protein was used as the reference strain, A/HK/2562/06 was found to be antigenically different from the SI06 strain and was not considered as the vaccine strain. Instead, SI06 with D190 and R226 was chosen as the H1N1 vaccine component for the live attenuated influenza vaccine used in the 2007-2008 influenza season. Our data indicate that A/HK/2562/06 with residues that confer an α2-6SAL binding preference would be a better choice for the H1N1 vaccine than the A/Solomon Island/3/06 strain that has a binding preference to α2-3SAL. In the following season, a later isolate, A/South Dakota/6/07 with residues N190 and Q226, was selected as the vaccine strain. In summary, the results of our studies on the seasonal WT H1N1 viruses further emphasize the need for care in the choice of appropriate WT viruses as reference strains and the need to thoroughly evaluate influenza vaccine variants for their receptor binding preference, antigenicity, and immunogenicity in order to select suitable vaccine strains for immunization.
We are grateful to Joseph Show and Gorazd Drozina for editing the manuscript. We thank Dan Ye and Chin-Fen Yang for their sequencing support, Scott Jacobson, Stephanie Gee, Armando Navarro, Paulyn Cha, and Brett Pickell for ferret studies, Bin Lu and Zhongying Chen for discussions, Jennifer Woo and Chengjun Mo for help with the microscope imaging, and Gary Van Nest for review of the manuscript. We also thank the CDC for providing the WT influenza viruses.
Published ahead of print on 3 March 2010.