Vaccine 28 (2010) Contents lists available at ScienceDirect. Vaccine. journal homepage:

Transcription

1 Vaccine 28 (2010) Contents lists available at ScienceDirect Vaccine journal homepage: The impact of key amino acid substitutions in the hemagglutinin of influenza A (H3N2) viruses on vaccine production and antibody response Zhongying Chen, Helen Zhou, Hong Jin MedImmune, 297 North Bernardo Avenue, Mountain View, CA 94043, United States article info abstract Article history: Received 1 December 2009 Received in revised form 24 March 2010 Accepted 26 March 2010 Keywords: Influenza H3N2 virus Hemagglutinin Egg adaptation Antigenicity Immunogenicity Influenza H3N2 viruses have recently drifted from A/Wisconsin/67/05-like to A/Brisbane/10/07-like viruses. The effect of key amino acid substitutions in the hemagglutinin (HA) protein on the replication, antigenicity and immunogenicity of cold-adapted, live attenuated vaccine strains was examined. A/Brisbane/10/07 egg isolate contained a unique combination of G186 and P194 which were required for efficient virus growth in Madin Darby Canine Kidney (MDCK) cells and embryonated chicken eggs, but the virus induced low level of serum antibody response in ferrets. Substitution of the G186 and P194 in the HA of A/Brisbane/10/07 by V186 and L194 that were present in the HA of A/Wisconsin/67/05-like viruses significantly impaired virus replication but greatly improved the immunogenicity of the vaccine virus to a level comparable to that elicited by the A/Wisconsin/67/05 vaccine. The replication of the variants with impaired growth could be improved by amino acid substitutions at the 195 or 226 residues. The viruses with the G186 and P194 residues were antigenically distinct from the viruses with V186 and L194. Sequence analysis of the HA sequences of the H3N2 viruses from the database and sequencing of the HA gene of a cell-derived A/Brisbane/10/07-like virus before and after egg passage indicated that the P194 residue was most likely derived from egg adaptation. Our results demonstrated the importance of careful evaluation of vaccine strains to ensure that the selected vaccines not only replicate well in eggs, but also retain their antigenicity and are immunogenic in the host. Published by Elsevier Ltd. 1. Introduction Influenza A and B viruses infect 5 15% of the global population annually and cause an estimated half a million deaths worldwide [1]. Vaccination is the most effective way of preventing diseases caused by the influenza virus infection. The current annual seasonal vaccine contains antigens from two type A (H1N1 and H3N2) and one type B strain. The live, attenuated influenza virus (LAIV) vaccine licensed in the US is composed of three 6:2 reassortants containing the hemagglutinin (HA) and neuraminidase (NA) gene segments from the three wild-type influenza strains and the six internal protein gene segments (PB1, PB2, PA, NP, M and NS) from the master donor virus of cold-adapted (ca) A/Ann Arbor/6/60 (MDV-A) or ca B/Ann Arbor/1/66 (MDV-B) that confer the attenuated phenotype [2]. Due to constant antigenic drift of the influenza viruses, the influenza vaccine components require an annual update to antigenically match the circulating strains. The influenza A (H3N2) viruses, which have been associated with the vast majority of severe influenza outbreaks since 1968, evolve more rapidly than other strains. In the past 40 years since the influenza A (H3N2) subtype Corresponding author. Tel.: ; fax: address: chenz@medimmune.com (Z. Chen). entered the human population in 1968, the H3N2 vaccine component has been updated almost 30 times [3,4]. Antigenic variation in influenza viruses is mainly determined by the major surface glycoprotein HA. HA binds to sialic acid (SA)-containing receptors on the cell surface and mediates virus attachment and membrane fusion during virus entry [5,6]. The precursor of the HA protein (HA0) is cleaved into two polypeptides, HA1 and HA2, which is a prerequisite for the virus to be infectious. The three-dimensional structural analysis of the HA trimer showed that the HA receptor-binding site is located on each subunit at the distal end of the molecule [7]. HA is the target for the host immune system to generate neutralizing and protective antibodies; thus the HA protein is the focus of influenza virus surveillance and vaccine selection. Five antigenic sites (A E) have been mapped on the HA1 region [8 10]. The region covers much of the surface of the globular head of HA, including residues around the receptor-binding site. The accumulation of amino acid changes in the antigenic sites of HA causes antigenic drift to escape immune suppression [9,11 14]. It was reported that 18 codons in the HA1 of the human H3N2 viruses associated with antigenic site A or B, or the receptorbinding site, are under positive selection to change the amino acid they encode, leading to antigenic drift [15,16]. We have previously reported that two residues at positions 155 and 156 caused antigenic drift from A/Panama/2007/99 to A/Fujian/411/02, resulting X/$ see front matter. Published by Elsevier Ltd. doi: /j.vaccine

2 4080 Z. Chen et al. / Vaccine 28 (2010) in severe influenza epidemics in 2003 [17]. Cultivation of influenza viruses in embryonated chicken eggs frequently results in the selection of an HA that differs antigenically from naturally occurring viruses [18]. Amino acid substitutions usually cluster at or near the receptor-binding regions which could alter virus antigenicity, immunogenicity, and vaccine efficacy [19 22]. Thus, special care should be taken in the selection of annual reference and vaccine strains. Since 2007, the H3N2 viruses drifted to A/Brisbane/10/07-like viruses with characteristic genetic changes at HA positions 50 and 140. The egg isolate of A/Brisbane/10/07, which was used as a reference strain and was recommended as the vaccine strain, contained two changes (V186G and L194P) at the antigenic site B and the receptor-binding site compared to the previous vaccine strain A/Wisconsin/67/05. Using reverse genetics, we generated 6:2 reassortant vaccine strains which contained different variations at these positions and evaluated the impact of these residues on the replication and immunogenicity of the live, attenuated vaccine strain. In addition, other residues at positions 195 and 226 identified in other egg isolates were also shown to be important for virus replication in host cells. 2. Materials and methods 2.1. Viruses, cells and embryonated chicken eggs Wild-type (wt) influenza A viruses used in this study, A/Wisconsin/67/05, A/Brisbane/10/07, A/Uruguay/716/07, A/Texas/37/07, and A/Malaysia/1199/07 were obtained from the Centers for Disease Control and Prevention (CDC, Atlanta, Georgia). Madin Darby Canine Kidney (MDCK) cells, originally obtained from the European Collection of Cell Cultures (ECACC, Salisbury, UK), were maintained in minimum essential medium (MEM) supplemented with 10% fetal bovine serum (FBS). Embryonated chicken eggs were obtained from Charles River SPAFAS (Franklin, CT) and incubated for days at 37 C prior to virus inoculation Generation of 6:2 recombinant viruses The HA and NA cdnas of wt viruses were amplified by reverse transcription-polymerase chain reaction (RT-PCR) using vrna as the template and cloned between the two BsmBI sites in the pad3000 vector. Recombinant 6:2 vaccine viruses were rescued using the eight plasmid transfection system [23,24]. Briefly, cocultured 293T and MDCK cells were transfected with the six plasmids encoding the internal genes of ca A/Ann Arbor/6/60 (MDV-A) together with the two plasmids encoding the HA and NA genes from wt H3N2 viruses using TransIT-LT1 reagents (Mirus, Madison, WI). Six to fifteen hours later, the DNA-transfection mixture was replaced by Opti-MEM I (Invitrogen, Carlsbad, CA) containing 1 g/ml of TPCK-trypsin (Sigma Aldrich, St. Louis, MO). The transfected cell culture supernatant was collected at 3 6 days post-transfection and used to infect MDCK cells and embryonated chicken eggs as described below. To generate 6:2 recombinant viruses that contained specific amino acid substitution mutations in HA, the HA plasmid was subjected to site-directed mutagenesis by using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA), and viruses were produced by plasmid rescue as described above. The HA sequence of each rescued virus was examined by sequencing analysis Replication of viruses in MDCK cells and eggs MDCK cells in TC-6 plates were inoculated with 0.4 ml of the virus inoculum. After 60 min of adsorption at room temperature, virus inoculum was removed, replaced with 2 ml of Opti-MEM I containing 1 g/ml of TPCK-trypsin. The infected MDCK cells were incubated at 33 C for 7 days or until 80 90% of the cells exhibited cytopathic effects, whichever occurred first. The cell supernatants were collected and subjected to plaque assay in MDCK cells as described before [17]. Ten- to eleven-day embryonated chicken eggs were inoculated with 0.1 ml of transfection supernatants and incubated at 33 C. The allantoic fluids were collected after 3 days. The virus titer was determined by plaque assay [17]. The plaques were immunostained with a chicken anti-ca A/Ann Arbor/6/60 polyclonal antibody Ferret studies Ferrets in groups of three were inoculated intranasally with 7.0 log 10 plaque-forming units (PFU) of egg-amplified viruses. Ferret post-infection sera were collected at 14 days after infection. Hemagglutination inhibition (HAI) assay and microneutralization assay were used to measure serum antibody levels against homologous or heterologous viruses. For HAI assay, 25 l of serial-diluted serum samples treated with receptor-destroying enzyme (RDE, Denka Seiken Co., Tokyo, Japan) were mixed with 4 HA units of the indicated viruses (25 l) in 96-well V-bottom microplates. After incubating at room temperature for 30 min, 50 l of 0.5% turkey erythrocytes (trbc) were added to each well and incubated for an additional 45 min. The HAI titer was defined as the reciprocal of the highest serum dilution that inhibited virus hemagglutination. For the microneutralization assay, 100 l of RDE-treated serum was subjected to a twofold serial dilution in MEM medium containing 1 g/ml of TPCK-trypsin followed by incubation with 100 lofthe indicated viruses at a concentration of 500 TCID 50 /ml in 96-well U- bottom microplates for 1 h at 33 C. The antiserum virus mixtures were transferred to the MDCK cells cultured in 96-well plates and the plates were incubated at 33 C for 6 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% CPE caused by virus infection. 3. Results 3.1. The G186 and P194 residues in the HA of A/Brisbane/10/07 impacted virus replication A/Brisbane/10/07-like viruses were recommended as the vaccine composition for the and influenza season. The G186 V change in HA has been frequently detected in the egg-adapted H3N2 strains with minimal impact on virus antigenicity [25 27]. The A/Wisconsin/67/05 vaccine strain used in the previous influenza seasons also contained the V186 residue (Table 1; GenBank accession No. ACD62287). However, the egg isolate of the A/Brisbane/10/07 reference strain contained a unique combination of G186 and P194 residues. The Proline at residue 194 is a new change from L194 in the corresponding MDCK cell isolates (compare GenBank accession No. ACD62305 with ABW23422) and has not been found in any of the previous H3N2 isolates, indicating this L194P substitution could be introduced by egg adaptation. It was confirmed that egg adaptation of the cell isolate of A/Texas/37/07 caused the L194P change (our unpublished data). The egg isolate of A/Uruguay/716/07 from the CDC also contained the combination of G186 and P194. In contrast, A/Malaysia/1199/07 had the same V186 and L194 as A/Wisconsin/67/05 (Table 1).

3 Z. Chen et al. / Vaccine 28 (2010) Table 1 Comparison of the HA sequences of the recent H3N2 strains. Virus HA residues A/Wisconsin/67/05 G G D S K K V L S I I N A/Brisbane/10/07 E N A I G P V D A/Uruguay/716/07 E N I G P V D A/Texas/37/07 D E N I E G P V D A/Malaysia/1199/07 E N A I E F V N D The HA sequences of the 2007 H3N2 viruses amplified from eggs were compared with the previous vaccine strain A/Wisconsin/67/05. Only the residues different from A/Wisconsin/67/05 are indicated. To investigate the effect of amino acids 186 and 194 in the HA on vaccine virus replication in eggs, 6:2 recombinant vaccine viruses containing the HA and NA gene segments from the H3N2 strains and the 6 internal protein gene segments from the cold-adapted (ca) master donor virus A/Ann Arbor/6/60 (MDV-A), were generated by the eight plasmid transfection method [23,24]. By site-directed mutagenesis of the corresponding cdnas, different HA variants with G or V at position 186 and P or L at position 194 were introduced into both A/Wisconsin/67/05 and A/Brisbane/10/07. Replication of these rescued viruses was evaluated both in MDCK cells and embryonated chicken eggs. As shown in Table 2, ca A/Wisconsin/67/05 (V-L) and A/Brisbane/10/07 (G-P) grew well in both MDCK cells (6 7 log 10 PFU/ml) and eggs (>8.0 log 10 PFU/ml). However, A/Brisbane/10/07 with V-L replacement replicated to a much lower titer in MDCK cells and eggs. In contrast to the large and clear plaques of G-P version, the plaques of A/Brisbane/10/07(V- L) were small and indistinct (compare Fig. 1 No. 1 with No. 2). Similarly, the introduction of G-P into A/Wisconsin/67/05 caused the virus to grow less efficiently and formed tiny plaques (Fig. 1, No. 4), indicating that other amino acid differences between these two strains played an important role in virus replication. In addition to G-P and V-L variants, G-L and V-P were also introduced into both strains. A/Wisconsin/67/05 (G-L) replicated well in MDCK cells, but it replicated to a low titer in eggs, confirming that residue 186 is an egg-adaptation site. The virus that replicated to a titer of 7.8 log 10 FPU/ml in eggs was found to have the Y195H change in the HA. A/Brisbane/10/07 (G-L) replicated poorly both in MDCK cells and eggs. A/Wisconsin/67/05 and A/Brisbane/10/07 containing the introduced V-P residues in their HA could not be rescued, indicating that the L194P change could not cooperate with the 186 V residue. Thus, the 194P residue was likely selected in eggs to allow for efficient replication of A/Brisbane/10/07. The NA proteins of A/Brisbane/10/07 and A/Wisconsin/67/05 differed by nine amino acids. The effect of the NA gene on virus growth was examined by substituting the NA gene from one strain into the other. The 6:2 A/Brisbane/10/07 (V-L) with NA from A/Wisconsin/67/05 still replicated to a very low titer (3.8 and 6.2 log 10 PFU/ml in MDCK cells and eggs, respectively); its plaque size was similar to the virus with the NA segment from A/Brisbane/10/07 (Fig. 1, No. 2 and 3). Similarly, the replacement of A/Wisconsin/67/05 NA with A/Brisbane/10/07 NA had no effect on the growth of 6:2 A/Wisconsin/67/05 (V-L) (Fig. 1, No. 5 and 6). This result indicated that the NA genes from the two H3 strains were closely related and had no significant impact on virus growth The HA residues 186 and 194 affected vaccine immunogenicity and antigenicity in ferrets The HA 186 and 194 residues are located at the antigenic site B. To examine if these two residues affected vaccine immunogenicity and antigenicity, groups of ferrets were inoculated intranasally with ca A/Wisconsin/67/05 and A/Brisbane/10/07 strains containing G-P or V-L in the HA. After 14 days of immunization, serum antibody levels were determined by hemagglutinin inhibition (HAI) and microneutralization assays (Table 3). ca A/Brisbane/10/07 (G-P) induced low antibody response, with a homologous HAI titer of only 29. In contrast, ca A/Wisconsin/67/05 (V-L) induced much higher antibody titers in ferrets, with a homologous HAI titer of A/Brisbane/10/07 with V-L substitution greatly increased the HAI titer to a level similar to that of A/Wisconsin/67/05 (V-L) although its inoculum dose was 10-fold lower due to its low titer in eggs. Similarly, the replacement of V-L with G-P in A/Wisconsin/67/05 significantly reduced the HAI titer from 1290 to 64. The microneutralization assay produced comparable results, demonstrating that the two amino acids at HA positions 186 and 194 were critical to antibody levels induced by the ca viruses. The G-P residues, whether in A/Wisconsin/67/05 or A/Brisbane/10/07 backbone, made the virus less immunogenic; the V-L residues made the virus more immunogenic. Viral replication in the respiratory tract of ferrets was examined in a separate study. Three days after intranasal inoculation, nasal turbinate (NT) suspensions were prepared and titrated by plaque assay. No viruses were detected by plaque assay in ca A/Brisbane/10/07 (G-P) infected NT, yet about 2.9 log 10 PFU/ml of viruses were detected in the NT of ca A/Wisconsin/67/05 (V-L) Table 2 The effect of the amino acids 186 and 194 on virus growth. Virus HA residues Titer (log 10PFU/ml) a MDCK Egg A/Wisconsin/67/05 A/Brisbane/10/07 V L G P G L and 7.8 b V P NR G P V L G L V P NR NR, not rescued. The residues from the original egg isolates are indicated in bold. a Virus titers were expressed as the mean titers from two or more experiments with the standard deviation less than 10%. b The virus with a higher titer (7.8 log 10PFU/ml) contained the Y195H change in the HA.

4 4082 Z. Chen et al. / Vaccine 28 (2010) Figure 1. Plaque morphology of 6:2 reassortants A/Brisbane/10/07 (Bris) and A/Wisconsin/67/05 (WI) with the indicated hemagglutinin (HA) and neuraminidase (NA) segments. The HA variants with G and P or V and L at positions 186 and 194 respectively are indicated as G-P or V-L. Virus titers in eggs are indicated at the bottom. Plaque assay was performed in MDCK cells, incubated at 33 C for 4 days and immunostained with polyclonal antiserum against influenza A virus. infected ferrets. In a more sensitive EID50 assay, the titers of ca A/Brisbane/10/07 (G-P) were reduced by 10-fold compared to that of ca A/Wisconsin/67/05 (V-L), with mean titers of 3.2 (G-P) and 4.3 (V-L), respectively. Thus, viral replication level in the respiratory tract of ferrets correlated with the level of serum antibody response. Furthermore, the antibody cross-reactivity data showed that the two residues also affected virus antigenicity. The variants with the V-L residues were antigenically different from those with the G-P residues. The antiserum against A/Wisconsin/67/05 (V-L) had better cross-reactivity to A/Brisbane/10/07 (V-L) than to A/Wisconsin/67/05 (G-P), with the titer difference >4-fold. Several other H3N2 viruses isolated in 2007 were also examined for their antibody response in ferrets (Table 4). As expected, ferrets infected with ca A/Texas/37/07 and ca A/Uruguay/716/07 containing G-P had lower antibody titers. A/Malaysia/1199/07 containing V-L induced higher homologous HAI titers. The antigenicity of A/Malaysia/1199/07 (V-L) was closer to A/Wisconsin/67/05 (V-L) than to those containing G-P. Thus, the amino acid residues at 186 and 194 not only affected virus growth, but also reduced vaccine virus immunogenicity and antigenicity The HA 195 and 226 residues influenced virus replication Table 2 shows that viruses containing 186G and 194L were unable to grow well in eggs. A change of Y195H in the HA protein was found in one of the egg isolates of A/Wisconsin/67/05 (G-L). To confirm that Y195H could revert the poor growth of A/Wisconsin/67/05 (G-L) in eggs, the Y195H change was introduced into A/Wisconsin/67/05 (G-L) and (G-P). The rescued virus with G- L-H at 186, 194, and 195 replicated well in egg reaching a titer of 8.4 log 10 PFU/ml (Table 5). In contrast, A/Wisconsin/67/05 (G-P) with the Y195H change could not be rescued, indicating that only G-L-H could make the virus to grow well in eggs. Introduction of Y195H into A/Brisbane/10/07 also significantly increased virus fitness in eggs, it increased viral titer from 3.1 to 7.1 log 10 PFU/ml. Examination of the HA sequences of A/Wisconsin/67/05 in the Gen- Bank revealed that one of the A/Wisconsin/67/05 egg isolates also contained G-L-H (GenBank accession No. ABO37599). A/Brisbane/10/07 (V-L) had much lower titers than A/Wisconsin/67/05 (V-L) in eggs (Table 2). The HA proteins of these two viruses differ only in six amino acids at positions Table 3 Serum antibody levels and cross reactivity of Wisconsin and Brisbane ca vaccine virus infected ferrets. Virus antigen HAI GMT of ferret serum immunized with MN GMT of ferret serum immunized with WI Bris WI Bris V-L G-P G-P V-L V-L G-P G-P V-L A/Wiscosin/67/05 V-L G-P A/Brisbane/10/07 G-P V-L Groups of three ferrets were inoculated intranasally with PFU of the indicated egg amplified 6:2 ca viruses except for A/Bribane/10/07 (V-L) that only PFU was used because of its low titer. Serum was collected 14 days after immunization. The level of antibodies against different virus antigens was determined by the HAI assay and microneutralization (MN) assay. Homologous titers are in bold.

5 Z. Chen et al. / Vaccine 28 (2010) Table 4 Serum antibody levels and cross reactivity of ferrets infected with vaccine viruses with different residues at 186 and 194 of the HA protein. Virus antigen HAI GMT of ferret serum immunized with Uru TX a Mal WI G-P G-P V-L V-L A/Uruguay/716/07(Uru) G-P A/Texas/37/07(TX) G-P A/Malaysia/1199/07(Mal) V-L A/Wisconsin/67/05 (WI) V-L a Only one of the three ferrets had detectable HAI titer indicated in the table. The homologous titers are indicated in bold. 50, 122, 138, 140, 223 and 375. Single amino acid substitutions at these residues were introduced into the HA protein of A/Brisbane/10/07 (V-L) to test which amino acid change could improve the growth of A/Brisbane/10/07 (V-L). None of the single amino acid substitutions had much effect on virus growth (data not shown). Since A/Malaysia/1199/07 (V-L) grew well in eggs, the N226 residue present in the HA of A/Malaysia/1199/07 was introduced into A/Brisbane/10/07 (V-L). The I226N change caused A/Brisbane (V-L) to grow much more efficiently in eggs, with a titer of 8.5 log 10 PFU/ml, without affecting virus antigenicity. Thus, A/Brisbane that had V186 and L194 could replicate well in eggs when the 226 residue was altered. 4. Discussion Using reverse genetics, we characterized the effects of several key genetic changes in the HA protein on virus replication, immunogenicity, and antigenicity. By comparing the two recent 6:2 vaccine strains A/Wisconsin/67/05 and A/Brisbane/10/07 which differed only in the HA and NA segments, we identified that variation in the HA amino acid positions 186 and 194 impacted virus replication. Unlike V186 and L194 in Wisconsin/67/05, G186 and P194 are required for efficient growth of the vaccine strain A/Brisbane/10/07 in both MDCK cells and embryonated chicken eggs. None of the other combinations (V-L, V-P or G- L) allowed efficient virus replication, indicating that G186 and P194 retained the HA structure to allow efficient virus entry and replication in the host cells. Amino acids at the 186 and 194 residues also affected virus immunogenicity and antigenicity. A/Bribane/10/07 with G-P was poorly immunogenic in ferrets (HAI titer 29) when ferrets were inoculated with a standard 7.0 log 10 PFU dose. Our previous studies have consistently shown that virus inoculation dose ranging from 5.0 log 10 PFU to 7.0 log 10 PFU produced similar antibody titers. We also showed in this study that the ferrets infected with ca Brisbane/10/07 (V-L) had a high HAI antibody titer even though the inoculation dose was only 6.0 log 10 PFU (Table 3). Thus, the low antibody titer from A/Brisbane/10/07 (G-P) infection was not due to the vaccination dose but rather due to its low level replication in the upper respiratory tract of ferrets. A/Wisconsin/67/05 containing V-L was highly immunogenic (HAI titer 1290). The introduction of V-L into A/Brisbane/10/07 resulted in reduced virus titer in eggs but significantly increased serum HAI antibody titer from 29 to All viruses with G-P tested in this study, including A/Texas/37/07 and A/Uruguay/716/07, had lower antibody titers than the viruses with V-L, such as A/Malaysia/1199/07 and the recent 2008 strains A/Mississippi/4/08 and A/Brisbane/10/08 (data not shown). Furthermore, A/Wisconsin/67/05 (V-L) was antigenetically similar to A/Brisbane/10/07 (V-L) but different from A/Wisconsin (G-P), and vice versa. Similarly, A/Malaysia/1199/07 containing V-L was antigenically closer to A/Wisconsin/67/05 (V- L) than to A/Brisbane/190/07 (G-P) even though it belongs to the A/Brisbane/10/07 lineage genetically. These results further proved that single amino acid substitutions at key immunodominant positions could cause antigenic drift [4,14,17,28,29]. The amino acids at HA positions 186 and 194 are both located at antigenic site B, an area that is constantly under positive selection during evolution of the H3N2 viruses and results in antigenic drift [15,16,30 33] (Fig. 2). Position 186 was known to be an egg adaptation site; previous studies demonstrated that the S186I and G186 V changes were associated with egg adaptation [21,26,34]. L194 is one of the critical residues forming the receptor binding site [7,8]. L194A change severely abolished virus binding to erythrocytes [6,35]. L194I change was also found to be associated with egg adaptation [36]. The mechanism by which G-P and V-L differ in their induced antibody titers is unclear. Viruses with V-L replicated more efficiently in the upper respiratory tract of ferrets than viruses with G-P, suggesting that the level of virus replication correlated with the level of antibody titers in ferrets. Our RBC binding assay showed that viruses with V-L bind to guinea pig RBC better than to chicken RBC, indicating that viruses with the V-L residue may preferentially bind SA( 2,6)Gal receptors better than SA( 2,3)Gal receptors. Our studies with the H1N1 strains have demonstrated that the viruses with SA( 2,6)Gal binding preference replicated more efficiently in the respiratory tract and were more immunogenic in ferrets [37,38]. Thus, the differential levels of serum antibodies in the ferrets vaccinated with different H3N2 vaccine strains may be related to virus receptor binding preference and replication in the respiratory tract. Embryonated chicken eggs are currently the substrate for influenza vaccine production. However, recent clinical isolates of the H3N2 viruses showed poor growth in eggs [39]. Different variants have been isolated from eggs and found to alter virus Table 5 The effect of the HA 195 and 226 residues on virus replication in MDCK cells and eggs. Virus HA residue Titer (log 10PFU/ml) MDCK Egg A/Wisconsin/67/05 A/Brisbane/10/07 G L Y I and 7.8 G L H I G P H I NR G L Y I G L H I V L Y N NR, not rescued.

6 4084 Z. Chen et al. / Vaccine 28 (2010) Figure 2. The location of the HA residues 186, 194, 195 and 226 on the model of the three-dimensional structure of HA complex with sialic acid (SA) (Protein Data Bank 5HMG structure). A: Side view of the HA trimer. The residues were indicated in one of the monomers. B: Detailed view of the amino acids around the receptor-binding site. The figures were generated with Vector NTI 3D Molecule Viewer software. antigenicity, immunogenicity, and vaccine efficacy, posing a challenge for selection of reference and vaccine strains during the annual influenza virus monitoring and vaccine production processes [18 22,40,41]. A/Brisbane/10/07 with 186G and 194L was present in the viruses isolated from MDCK cells, the cell substrate which is considered to better preserve the sequence of clinical specimens than eggs [42]. Although we cannot rule out the possibility of any changes caused by virus replication in MDCK cells, egg adaptation likely caused the L194P change in the HA proteins of A/Brisbane/10/07, A/Uruguay/716/07, and A/Texas/37/07 [36,43]. G-L was shown to result in poor virus replication in eggs for both A/Wisconsin/67/05 and A/Brisbane/10/07. The effect of the L194P change on virus antigenicity was not well evaluated at the time when A/Brisbane/10/07 egg isolate was used as a reference strain. The antigenic difference between A/Brisbane/10/07 (G-P) and A/Wisconsin/67/05(V-L) may not represent the antigenic difference between the clinical isolates of these two viruses. The data obtained in this study suggest that ca A/Brisbane/10/07 (G-P) is not the vaccine strain of choice because of its altered antigenicity and low antibody response despite its efficient replication in eggs. Alternatively, an A/Malaysia/1199/07-like virus or A/Brisbane/10/07 with V186, L194 and N226, both of which grow well in eggs, induce high antibody response and reacted well with the G-L containing viruses antigenically, might be a better choice of a vaccine strain [25,27]. In this study, we also showed the important roles of the 195 and 226 residues in virus replication. Both 195 and 226 are involved in receptor binding (Fig. 2). Substitutions of residues at 195 (Y195F) or 226 (L226P) affected HA receptor binding [6,35]. Residue 226 is critical to virus receptor-binding specificity and host range restriction [6,44,45]. The amino acid substitution of glutamine for leucine at 226 (L226Q) switched HA receptor binding from SA( 2,3)Gal to SA( 2,6)Gal and affected virus host range [46]. Previously, we have shown that the replication of A/Fujian/411/02 in eggs could be improved by changing a minimum of two residues, G186 V and V226I [17]. N226 in the HA of A/Malaysia/1199/07 (also an A/Brisbane/10/07-like virus) allowed efficient virus replication in eggs despite the presence of V186-L194 in the HA. The I226N change also restored efficient growth of A/Brisbane/10/07 (V-L) without affecting virus antigenicity, demonstrating the importance of the 226 residue in receptor binding and virus replication. In summary, our results underscore the importance of the residues at or near the antigenic and receptor binding sites for vaccine virus replication and immunogenicity. The HA sequence should be closely monitored prior to and after the egg adaption process to evaluate the impact of the egg adaptation change on virus growth, antigenicity and immunogenicity and these practices should guide the selection of an appropriate strain for vaccine production. Acknowledgements We thank Scott Jacobson, Stephanie Gee, Armando Navarro and Brett Pickell at MedImmune s animal care facility for the ferret studies; Yang He and Lily Yang in Cell Culture group for providing tissue culture cells; Laura Tan in Process Development, Vanisree Battar in Quality Control, and Haleh Khoshnevisan in Clinical manufacture group for supplying embryonated chicken eggs; Lomi Kim and Weijia Wang for technical assistance; Joseph Shaw and Gary Van Nest for critical review of the paper. References [1] Stohr K. Influenza-WHO cares. Lancet Infect Dis 2002;2(9):517. [2] Murphy BR, Coelingh K. Principles underlying the development and use of live attenuated cold-adapted influenza A and B virus vaccines. Viral Immunol 2002;15(2): [3] Russell CA, Jones TC, Barr IG, Cox NJ, Garten RJ, Gregory V, et al. Influenza vaccine strain selection and recent studies on the global migration of seasonal influenza viruses. Vaccine 2008;26S(4):D31 4. [4] Russell CA, Jones TC, Barr IG, Cox NJ, Garten RJ, Gregory V, et al. The global circulation of seasonal influenza A (H3N2) viruses. Science 2008;320(5874): [5] Lamb RA, Krug RM. Orthomyxoviridae: the viruses and their replication. In: Knipe DM, Howley PM, editors. Fields virology. Philadelphia: Lippincott Williams & Wilkins; p [6] Skehel JJ, Wiley DC. Receptor binding and membrane fusion in virus entry: the influenza hemagglutinin. Annu Rev Biochem 2000;69:

7 Z. Chen et al. / Vaccine 28 (2010) [7] Weis W, Brown JH, Cusack S, Paulson JC, Skehel JJ, Wiley DC. Structure of the influenza virus haemagglutinin complexed with its receptor, sialic acid. Nature 1988;333(6172): [8] Wiley DC, Wilson IA, Skehel JJ. Structural identification of the antibody-binding sites of Hong Kong influenza haemagglutinin and their involvement in antigenic variation. Nature 1981;289(5796): [9] Wilson IA, Cox NJ. Structural basis of immune recognition of influenza virus hemagglutinin. Annu Rev Immunol 1990;8: [10] Wilson IA, Skehel JJ, Wiley DC. Structure of the haemagglutinin membrane glycoprotein of influenza virus at 3 A resolution. Nature 1981;289(5796): [11] Both GW, Sleigh MJ, Cox NJ, Kendal AP. Antigenic drift in influenza virus H3 hemagglutinin from 1968 to 1980: multiple evolutionary pathways and sequential amino acid changes at key antigenic sites. J Virol 1983;48(1): [12] Plotkin JB, Dushoff J, Levin SA. Hemagglutinin sequence clusters and the antigenic evolution of influenza A virus. Proc Natl Acad Sci USA 2002;99(9): [13] Shih AC, Hsiao TC, Ho MS, Li WH. Simultaneous amino acid substitutions at antigenic sites drive influenza A hemagglutinin evolution. Proc Natl Acad Sci USA 2007;104(15): [14] Smith DJ, Lapedes AS, de Jong JC, Bestebroer TM, Rimmelzwaan GF, Osterhaus AD, et al. Mapping the antigenic and genetic evolution of influenza virus. Science 2004;305(5682): [15] Bush RM, Bender CA, Subbarao K, Cox NJ, Fitch WM. Predicting the evolution of human influenza A. Science 1999;286(5446): [16] Bush RM, Fitch WM, Bender CA, Cox NJ. Positive selection on the H3 hemagglutinin gene of human influenza virus A. Mol Biol Evol 1999;16(11): [17] Jin H, Zhou H, Liu H, Chan W, Adhikary L, Mahmood K, et al. Two residues in the hemagglutinin of A/Fujian/411/02-like influenza viruses are responsible for antigenic drift from A/Panama/2007/99. Virology 2005;336(1): [18] Robertson JS. Clinical influenza virus and the embryonated Hen s egg. Rev Med Virol 1993;3(2): [19] Gubareva LV, Wood JM, Meyer WJ, Katz JM, Robertson JS, Major D, et al. Codominant mixtures of viruses in reference strains of influenza virus due to host cell variation. Virology 1994;199(1): [20] Katz JM, Naeve CW, Webster RG. Host cell-mediated variation in H3N2 influenza viruses. Virology 1987;156(2): [21] Kodihalli S, Justewicz DM, Gubareva LV, Webster RG. Selection of a single amino acid substitution in the hemagglutinin molecule by chicken eggs can render influenza A virus (H3) candidate vaccine ineffective. J Virol 1995;69(8): [22] Meyer WJ, Wood JM, Major D, Robertson JS, Webster RG, Katz JM. Influence of host cell-mediated variation on the international surveillance of influenza A (H3N2) viruses. Virology 1993;196(1): [23] Hoffmann E, Neumann G, Kawaoka Y, Hobom G, Webster RG. A DNA transfection system for generation of influenza A virus from eight plasmids. Proc Natl Acad Sci USA 2000;97(11): [24] Jin H, Lu B, Zhou H, Ma C, Zhao J, Yang CF, et al. Multiple amino acid residues confer temperature sensitivity to human influenza virus vaccine strains (FluMist) derived from cold-adapted A/Ann Arbor/6/60. Virology 2003;306(1): [25] Lu B, Zhou H, Chan W, Kemble G, Jin H. Single amino acid substitutions in the hemagglutinin of influenza A/Singapore/21/04 (H3N2) increase virus growth in embryonated chicken eggs. Vaccine 2006;24(44 46): [26] Lu B, Zhou H, Ye D, Kemble G, Jin H. Improvement of influenza A/Fujian/411/02 (H3N2) virus growth in embryonated chicken eggs by balancing the hemagglutinin and neuraminidase activities, using reverse genetics. J Virol 2005;79(11): [27] Widjaja L, Ilyushina N, Webster RG, Webby RJ. Molecular changes associated with adaptation of human influenza A virus in embryonated chicken eggs. Virology 2006;350(1): [28] Hensley SE, Das SR, Bailey AL, Schmidt LM, Hickman HD, Jayaraman A, et al. Hemagglutinin receptor binding avidity drives influenza A virus antigenic drift. Science 2009;326(5953): [29] Lee M, Chen JS, Cho I. Identifying potential immunodominant amino acid positions in hemagglutinin protein of influenza A H3N2 viruses. In: Kawaoka Y, editor. Options for the control of influenza V. San Diego: Elsevier; p [30] Bragstad K, Nielsen LP, Fomsgaard A. The evolution of human influenza A viruses from 1999 to 2006: a complete genome study. Virol J 2008;5:40. [31] Muñoz ET, Deem MW. Epitope analysis for influenza vaccine design. Vaccine 2005;23(9): [32] Sato K, Morishita T, Nobusawa E, Tonegawa K, Sakae K, Nakajima S, et al. Amino-acid change on the antigenic region B1 of H3 haemagglutinin may be a trigger for the emergence of drift strain of influenza A virus. Epidemiol Infect 2004;132(3): [33] Xiu W, Wen Y, Shen X, Xie J, Yang S, Wu B, et al. Molecular evolution of influenza A (H3N2) viruses circulated in Fujian Province, China during the period. Sci China C Life Sci 2008;51(4): [34] Plakokefalos ET, Markoulatos P, Spyrou N, Bei T, Vamvakopoulos NC. Molecular and phylogenetic analysis of haemagglutinin and neuraminidase sequences from recent human influenza type A (H3N2) viral isolates in Southern Greece. Arch Virol 2001;146(10): [35] Martín J, Wharton SA, Lin YP, Takemoto DK, Skehel JJ, Wiley DC, et al. Studies of the binding properties of influenza hemagglutinin receptor-site mutants. Virology 1998;241(1): [36] Mochalova L, Gambaryan A, Romanova J, Tuzikov A, Chinarev A, Katinger D, et al. Receptor-binding properties of modern human influenza viruses primarily isolated in Vero and MDCK cells and chicken embryonated eggs. Virology 2003;313(2): [37] Chen Z, Wang W, Zhou H, Suguitan ALJ, Shambaugh C, Kim L, et al. Generation of live attenuated novel influenza virus A/California/7/09 (H1N1) vaccines with high yield in embryonated chicken eggs. J Virol 2010;84(1): [38] Xu Q, Wang W, Cheng X, Zengel J, Jin H. Influenza H1N1 A/Solomon Island/3/06 virus receptor binding specificity correlates with virus pathogenicity and immunogenicity in ferrets. J Virol; [Epub ahead of print]. [39] Nobusawa E, Ishihara H, Morishita T, Sato K, Nakajima K. Change in receptorbinding specificity of recent human influenza A viruses (H3N2): a single amino acid change in hemagglutinin altered its recognition of sialyloligosaccharides. Virology 2000;278(2): [40] Katz JM, Webster RG. Efficacy of inactivated influenza A virus (H3N2) vaccines grown in mammalian cells or embryonated eggs. J Infect Dis 1989;160(2): [41] Robertson JS, Cook P, Nicolson C, Newman R, Wood JM. Mixed populations in influenza virus vaccine strains. Vaccine 1994;12(14): [42] Katz JM, Wang M, Webster RG. Direct sequencing of the HA gene of influenza (H3N2) virus in original clinical samples reveals sequence identity with mammalian cell-grown virus. J Virol 1990;64(4): [43] Asaoka N, Tanaka Y, Sakai T, Fujii Y, Ohuchi R, Ohuchi M. Low growth ability of recent influenza clinical isolates in MDCK cells is due to their low receptor binding affinities. Microbes Infect 2006;8: [44] Connor RJ, Kawaoka Y, Webster RG, Paulson JC. Receptor specificity in human, avian, and equine H2 and H3 influenza virus isolates. Virology 1994;205(1): [45] Rogers GN, Paulson JC, Daniels RS, Skehel JJ, Wilson IA, Wiley DC. Single amino acid substitutions in influenza haemagglutinin change receptor binding specificity. Nature 1983;304(5921):76 8. [46] Vines A, Wells K, Matrosorich M, Castrucci MR, Ito T, Kawaoka Y. The role of influenza A virus hemagglutinin residues 226 and 228 in receptor specificity and host range restriction. J Virol 1998;72(9):