Influenza A and B viruses expressing altered NS1 proteins: A vaccine approach

NS1–99 virus was generated by ribonucleoprotein transfection (5) by using the 25A-1 helper influenza virus as described for delNS1 virus (4). NS1–99 virus contains a 154-nt insertion between nucleotide positions 324–478 of the NS1–99 viral NS gene. This insertion creates a frame shift and results in the generation of NS1 protein containing the first 99 amino acids of wild-type NS1, with three additional C-terminal amino acids (HisAspSer). This insertion arose as a PCR artifact during repair of a pUC19-T3/NS construct by site-directed mutagenesis and contains seven 22-nt repeats (5′-CATGACTCTTGAGGAAATGTCA-3′). Influenza B/Yamagata/1/73 clone 201 (B/201) and influenza B/Yamagata AWBY-234 (B/234) have been described previously (6–9).

Influenza A viruses were grown in embryonated chicken eggs between the ages of 6 and 14 days old (SPAFAS, Preston, CT) at 37°C for 48 h. Influenza B viruses were grown in embryonated chicken eggs at 35°C for 72 h. One hundred plaque-forming units (pfu) of virus were injected into the allantoic cavity of each egg. Allantoic fluid from influenza A or B virus-infected eggs was serially diluted in PBS and assayed for hemagglutination (HA) of chicken red blood cells (0.5%; Truslow Farms, Chestertown, MD) in 96-well plates. Plaque assay of influenza A or B virus stocks was performed on Madin–Darby canine kidney cells (MDCK) cells in the presence of 2 μg/ml trypsin (Difco) at 37°C (influenza A viruses) or 35°C (influenza B viruses). It should be noted that the MDCK plaque size of delNS1 virus is markedly reduced as compared with wild-type PR8 virus. This is likely to be a reflection of major difference in the permissiveness of MDCK cells for the replication of these two viruses (4).

Dishes (3.5 cm) of 70% confluent MDCK cells were infected [multiplicity of infection (MOI) = 10] with wild-type PR8 virus, the recombinant NS1–99 virus, or mock infected with PBS for 1 h at room temperature. Cells were then incubated in 1 ml labeling media (MEM–cys/met + 250 μCi ³⁵S-cys/met) for 8 h at 37°C and lysed on ice in 200 μl RIPA buffer. Lysates were clarified by centrifugation for 10 min at 12,000 × g. Cell supernatant was precipitated for 2 h at 4°C by using a polyclonal antibody (2 μl) against the influenza A viral NS1 protein (8). Fifty microliters of Protein A Sepharose beads was added to mixture and antibody complexes were pelleted after incubation at 4°C for 1 h. Pellets were washed twice with RIPA buffer (0.15 mM NaCl/0.05 mM Tris⋅HCl, pH 7.2/1% Triton X-100/1% sodium deoxycholate/0.1% SDS), and half of the samples were loaded onto 17.5% SDS/PAGE gel. Separated bands were visualized by autoradiography.

Thirty-five-millimeter dishes of confluent Vero cells were mock infected or infected at an MOI of 2 with either NS1–99, delNS1, or PR8 virus for the indicated time point. Cell extracts were made in RIPA buffer, and 10% were loaded on an SDS–polyacrylamide gel. Separated proteins were probed for Western analysis with a rabbit polyclonal against the viral nucleoprotein or a rabbit polyclonal against the nuclear export protein (NEP) (10).

BALB/c mice are an established model for influenza viral infection (11–14). Four-week-old female BALB/c mice (Taconic Farms) were sedated with ether and infected intranasally with indicated amounts of influenza A or B virus (50 μl volume). Recovery from ether occurred within 1–2 min after infection. For viral lung titers, mice were killed at either day 3 or day 6. Lungs were homogenized and resuspended in sterile PBS (100 mg lung tissue per 1 ml PBS) and titered on MDCK cells in the presence of 2 μg/ml trypsin.

Four weeks after immunization, mice were challenged by intranasal infection with either wild-type PR8 virus (10⁵ or 5 × 10⁶ pfu) or wild-type B/Yamagata virus (3 × 10⁵ pfu), as indicated in Tables 1 and 2.

All procedures were in accordance with National Institutes of Health guidelines for the care and use of laboratory animals.

To quantitate the amounts of virus-specific antibodies present in immunized mice, ELISA analysis of the reactivity of diluted (1:1,000) serum against purified viral antigen was performed. Blood was drawn from mice 4 wk after immunization and before challenge with wild-type virus. Blood was incubated at 4°C for 2 h and spun at 200 × g for 10 min. Supernatant (serum) was removed and stored at −20°C. Fifty microliters of sucrose gradient purified influenza A/PR8 or B/Yamagata virus (25 mg/ml) was used to coat 96-well ELISA dishes (Immulon4, Dynex, Chantilly, VA). After PBS wash, coated wells were blocked with PBS 1% BSA and incubated with diluted (1:1,000) serum. After 1 h incubation with serum at room temperature, wells were rinsed with PBS and incubated with a secondary anti-mouse IgG peroxidase (Boehringer-Mannheim). Rinsed wells were incubated with colorimetric substrate (2,2′-azino-di-[3-ethylbenzthiazoline sulfonate], Boehringer-Mannheim) for 20′ and read with ELISA reader (OD595, Bio-Tek Instruments, Burlington, VT).

The ELISPOT assay used to detect antigen-specific cytotoxic T lymphocytes was performed as described previously (15, 16). This assay measures the number of IFN-γ-producing spleen cells in cells exposed to MHC class I-presented NP peptide. Ninety-six-well nitrocellulose plates (Milititer HA, Millipore) were coated with 7 μg/ml of a monoclonal antibody against mouse IFN-γ (no. R4–6A2, PharMingen) in 75 μl PBS and incubated at room temperature overnight. Wells were washed six times with culture medium and incubated with DMEM supplemented with 10% FCS for 1 h at 37 °C. Pooled splenocytes from immunized mice were added to antibody-coated wells in serial dilutions. P815 cells (a mastocytoma cell line that expresses only MHC class I molecules) were used as antigen-presenting cells. P815 cells (1 × 10⁵ cells/ml) were pulsed with 1 × 10⁻⁶ M of the synthetic NP peptide TYQRTRALV for 1 h at 37°C. After repeated washings with culture medium, cells were treated with 50 μg/ml mitomycin C (Sigma) for 1 h. NP-peptide treated P815 cells were added to each well. As a control, untreated P815 cells were used. Plates were incubated for 24 h at 37°C with 5% CO₂ and then washed extensively with PBS + 0.05% Tween-20 (PBS/T). Wells were incubated with a solution of 2 μg/ml biotinylated anti-mouse IFN-γ monoclonal antibody (XGM1.2, PharMingen) in PBS/T for 1 h at room temperature. Plates were washed with PBS/T and incubated with peroxidase-labeled streptavidin (1:1,000; Kirkegaard & Perry Laboratories) for 1 h at room temperature. Wells were washed with PBS/T and PBS, and 1 μg/ml substrate (3,3′-diaminobenzidine tetrahydrochoride dihydrate in 50 mM Tris⋅HCl, pH 7.5, containing 0.015% hydrogen peroxide) was added. Spots in each well were counted with the aid of a microscope.

DelNS1 and NS1–99 viruses were generated by ribonucleoprotein transfection of altered viral influenza NS genes by using the temperature-sensitive helper virus 25A-1 as described previously (Fig. 1A) (4). DelNS1 virus lacks the NS1 gene, and the NS1–99 virus encodes a mutant NS1 protein containing only the N-terminal 99 amino acids of wild-type NS1, creating an NS1 protein of approximately 11.2 kDa (Figure 1B). By contrast, the wild-type virus encodes an NS1 protein of 230 amino acids. We infected Vero cells, a cell line that is known to lack the type I IFN loci (17, 18), with NS1–99, delNS1, or PR8 virus at an MOI of 2. Despite the alterations in the NS gene, all three of these influenza A viruses express comparable levels of the viral NEP (NS2), the second protein encoded by the viral NS gene (Fig. 1 A and C).

Others have demonstrated that the IFN inducibility of both cultured chicken embryo cells and the intact embryonated egg increases with age (19–21). Specifically, Morahan and Grossberg established a strong correlation between age-related IFN induction and resistance of older intact chicken embryos to influenza (NWS) virus infection (21). We wished to determine the growth properties of influenza A and B viruses encoding either wild-type or altered NS1 proteins in eggs of varying age to examine the relationship between alteration in the NS1 protein and its functionality as an IFN antagonist. Our prediction was that alteration of this virally encoded IFN antagonist would compromise the ability of a virus to grow in older eggs but would still allow growth in younger eggs.

Equivalent doses of influenza A/PR8, NS1–99, or delNS1 viruses were injected into the allantoic cavity of embryonated chicken eggs ranging in age from 6 to 14 days old. Infected eggs were incubated at 37°C for 48 h. Allantoic fluid was harvested and assayed for HA activity as described in Materials and Methods. As shown in Fig. 2A, wild-type influenza A/PR8 virus was able to grow to high HA titers in embryonated eggs regardless of age, whereas growth of an isogenic virus lacking the viral NS1 gene (delNS1 virus) is compromised in its ability to grow to high HA titers in eggs older than 6 days. The NS1–99 virus showed intermediate growth characteristics, growing to high HA titers in 6- and 8-day-old eggs, and to moderate HA titers in 10- and 12-day-old eggs. The NS1–99 virus was unable to grow in 14-day-old eggs and consistently grew to titers that were slightly lower than the wild-type parent virus (Figure 2A). Virus titers from allantoic fluid of eggs infected with delNS1 or PR8 virus were also measured by plaque assay (Fig. 2B). Yield of delNS1 virus decreased with the age of the infected egg, corresponding to the reduction observed in HA titers, with 6- and 7-day-old eggs yielding a maximum of 1 × 10⁷ pfu/ml allantoic fluid. PR8 virus consistently grew to titers greater than 1 × 10⁸ pfu/ml, regardless of the egg age.

Influenza A/delNS1 virus has been shown previously to be nonpathogenic in wild-type mice. When mice were inoculated intranasally with 5 × 10⁴ pfu of delNS1 virus, all survived without showing any weight loss (4). We extended our pathogenicity studies by using the NS1–99 virus and examined whether delNS1 and/or NS1–99 viruses induced protective immune responses. As summarized in Table 1, 4-wk-old female BALB/c mice were inoculated with various dilutions of either delNS1 (groups A-B) or NS1–99 (groups C-D) virus and monitored for 4 wk. Mice infected with 2 × 10³ pfu (2 LD₅₀) of wild-type influenza PR8 virus or mock infected with PBS were included as controls. All mice inoculated with delNS1 virus survived, even at the highest inoculum (1 × 10⁶ pfu). Eight of ten mice inoculated with the highest dose of NS1–99 virus (group C) survived, and mice immunized with a lower dose (3.3 × 10⁴ pfu, group D) demonstrated a 90% survival rate. In contrast, only one of five mice survived after administration of 2 × 10³ pfu of wild-type PR8 (group E). To determine virus replication levels in lungs of infected mice, three groups of six mice each were also infected with 1 × 10⁴ pfu of either PR8, NS1–99, or delNS1 virus and killed at day 3 or day 6. Neither NS1–99 nor delNS1 virus-infected mice showed detectable lung titers at either day 3 or day 6 (<10 pfu/ml), whereas wild-type PR8 virus-infected mice had average lung titers of 2 × 10⁶ pfu/ml and 3 × 10⁴ pfu/ml at day 3 and day 6, respectively. One of six mice inoculated with 1 × 10⁴ pfu PR8 virus died before lung titration at day 6.

Four weeks after immunization, surviving mice were challenged with 1 × 10⁵ pfu (100 LD₅₀) or 5 × 10⁶ pfu (5,000 LD₅₀) of wild-type PR8 virus. As shown in Table 1, mice immunized with 1 × 10⁶ pfu of either delNS1 (group A) or NS1–99 (group C) were completely protected against lethal challenge with 100 or even 5,000 LD₅₀ of influenza A/PR8 virus. Reduction in the immunizing dose of NS1–99 virus to 3.3 × 10⁴ pfu (group D) resulted in 78% protection, with 7 of 9 mice surviving 2 wk after challenge. Reduction in the immunizing dose of delNS1 virus to 3.3 × 10⁴ pfu (group B) resulted in a loss of protection against lethal (100 or 5,000 LD₅₀) PR8 virus challenge. None of the mice mock immunized was protected against wild-type challenge, and the one mouse surviving PR8 virus infection was protected when reexposed to this virus.

Before challenge with wild-type influenza A/PR8 virus, serum samples were taken from delNS1 and NS1–99 immunized mice (five mice each from groups A–D in Table 1) and from mock-immunized mice (group F) and analyzed for reactivity against purified influenza A/PR8 virus by ELISA. As shown in Fig. 3, immunization with 1 × 10⁶ pfu of either delNS1 virus (group A; mice A1-A5) or NS1–99 virus (group C; mice C1-C5) gave measurable antibody titers against wild-type virus. Reduction in the immunizing dose of either delNS1 virus (groups B; mice B1-B5) or NS1–99 virus (groups D; mice D1-D5) resulted in a concomitant reduction in PR8-reactive immune sera. As expected, serum from mice inoculated with PBS (group F) did not react with purified influenza A/PR8 virus.

We also examined the production of NP peptide-specific T cells in delNS1 or NS1–99 virus immunized mice. Groups of three BALB/c mice were immunized intranasally with 1 × 10³ pfu of either delNS1, NS1–99, or PR8 virus. At day 10, spleen cells were harvested and analyzed by enzyme-linked immunospot for reactivity to an influenza virus NP peptide as previously described (15, 16). As shown in Fig. 4, mice immunized with either delNS1 or NS1–99 virus stimulated the production of NP-peptide-specific spleen cells to levels comparable to those seen in mice exposed to wild-type influenza A/PR8 virus. As expected, intranasal inoculation of mice with PBS did not induce the production of spleen cells reactive to influenza A NP peptide. These results are indicative of at least abortive viral replication, as immunization with inactivated virus has been shown not to induce a detectable cytotoxic T lymphocyte response (22).

Influenza B/201 and influenza B/AWBY/234 (B/234) viruses are laboratory variants of influenza B/Yamagata/1/73 that contain alterations in the NS1 ORF as compared with the NS1 ORF of wild-type influenza B/Yamagata (6, 7). As shown in Fig. 5, nucleotide deletions in the NS gene of either B/201 or B/234 create frame shifts that result in the generation of shortened NS1 proteins. The influenza B/201 and B/234 viral NS1 proteins contain, respectively, 110 and 89 amino acids corresponding to the N terminus of the B/Yamagata viral NS1 protein.

We hypothesized that the NS1 of influenza B virus would show IFN antagonist activity in vivo. We therefore examined whether influenza B/Yamagata virus derivatives that code for shortened NS1 proteins would show growth attenuation in older immunologically mature embryonated eggs as compared with wild-type influenza B/Yamagata virus. Both B/201 and B/234 viruses grow in 6-day-old eggs albeit to lower titers than B/Yamagata virus (Fig. 6). However, whereas B/Yamagata grew equally well in 8-day-old eggs, growth of the other two viruses was markedly reduced in embryos of this age.

We also compared the virulence of influenza B viruses in BALB/c mice. Because influenza B/Yamagata virus is not mouse adapted and does not cause disease or death in mice, we examined viral lung titers as an index of the ability of different viruses to replicate within the mouse host. We found that at 3 days after infection, B/201 and B/234 viruses showed more than two logs reduction in viral titers as compared with the wild-type B/Yamagata virus (Table 2). By 6 days after infection, mice in all infected groups had cleared the virus, and lung titers were below the limit of detection by plaque assay. In a separate experiment, we tested whether previous immunization of mice was able to protect against infection by wild-type B/Yamagata virus. Mice immunized with either B/201 or B/234 viruses were indeed protected, as shown by lack of replication at day 3 by the wild-type virus in the lungs of these mice.

Four weeks after immunization and before challenge with wild-type virus, serum was taken from mice infected with B/Yamagata, B/201, or B/234 virus and analyzed for antibodies directed against purified B/Yamagata virus. Infection with either B/Yamagata, B/201, or B/234 virus resulted in a significant induction of influenza B/Yamagata virus-specific serum IgG (Fig. 7). Furthermore, although viral lung titers vary significantly among groups of mice infected with these influenza B viruses, B/Yamagata-specific antibody titers appear to be generated to roughly similar levels.

Our results indicate that alterations in the NS1 ORFs of influenza A viruses affect the growth properties of these viruses in embryonated eggs. Influenza A/delNS1 is a recombinant virus that completely lacks the anti-IFN NS1 protein. This virus grows poorly in embryonated eggs older than 7 days (Fig. 2 A and B). We speculate that this reduction in growth in older eggs is because of maturation of the host innate immune system and the inability of the virus to counteract the increasing IFN response in older embryonated eggs (21). Influenza A/NS1–99 virus encodes a truncated NS1 protein containing 99 N-terminal amino acids of the wild-type protein. NS1–99 demonstrates intermediate growth characteristics in eggs as compared with delNS1 and wild-type PR8 viruses (Fig. 2A), suggesting that the shortened NS1 protein encoded by NS1–99 virus retains partial function. We have not yet determined the precise mechanism by which the NS1 protein counteracts the IFN response of the cell. The NS1 protein of influenza A virus has several reported activities (reviewed in ref. 23), including the inhibition of the IFN induced protein kinase, PKR (24, 25). The influenza A NS1 protein has also been shown to bind to several RNA species including double-stranded RNA (25–27), a potent activator of the IFN response (28). Reduction or loss of one or more of these NS1 activities may contribute to virus attenuation. Despite the reduction in growth of NS1-attenuated influenza A viruses in older eggs, younger eggs have proven to be suitable media for preparative growth of these viruses to titers sufficient for vaccination study.

Another important characteristic for a potential live vaccine is attenuation in the host. When administered to mice, delNS1 and NS1–99 viruses are much less virulent than wild-type influenza A/PR8 virus in BALB/c mice (Table 1). Intranasal inoculation with 1 × 10⁶ pfu of delNS1 does not cause disease (data not shown) or death in infected mice. Furthermore, the LD₅₀ of NS1–99 virus is significantly greater than the LD₅₀ for PR8 virus (i.e., 1 × 10³ pfu), because 80% of mice infected with 1 × 10⁶ pfu of NS1–99 virus survived.

Like the immunologically mature embryonated egg, the wild-type mouse represents an IFN competent environment. Without the ability to fully counteract this IFN response, delNS1 and NS1–99 viruses are severely weakened within the mouse host. The argument that alteration of virally encoded IFN antagonists correlates with attenuation is strengthened by recent work on paramyxoviruses (29–31).

We also wanted to assess the ability of influenza A viruses encoding altered NS1 proteins to induce a protective immune response. We found that although delNS1 and NS1–99 viruses are attenuated in the mouse host, they are still able to induce an immune response capable of providing a significant degree of protection against challenge with a highly lethal dose (5,000 LD₅₀) of wild-type influenza A/PR8 virus. Mice immunized with 1 × 10⁶ pfu of delNS1 or NS1–99 virus gave measurable antibody titers and were 100% protected against lethal challenge. As expected, reduction in the immunizing dose of either delNS1 or NS1–99 virus resulted in lower levels of PR8-specific serum IgG and a decrease in the number of mice protected against wild-type challenge (groups A–D, Fig. 3, and Table 1). The antibody levels in groups B and D do not correlate with protection. However, vaccination with this dose of NS1–99 virus appears to induce better protection than an equivalent dose of delNS1 virus. We plan to investigate whether other NS1 mutants may be more immunogenic and also whether the protection induced by delNS1 and NS1–99 viruses may be improved by a boosting protocol.

Both delNS1 and NS1–99 viruses induced influenza NP peptide-specific CD8 + spleen cells to numbers comparable to wild-type influenza PR8 virus (Fig. 4), indicating that immunization with these NS1-altered viruses results in the stimulation of this cell-mediated component of the immune system. Although cell-mediated immunity may not be sufficient for complete protection, it has been shown to be critical in the clearance of influenza virus from the infected host (32).

Influenza A and B viral NS1 proteins share little sequence homology. However, like the influenza A NS1 protein, influenza B NS1 is able to inhibit the activation of the IFN-induced protein kinase (PKR) and binds to double-stranded RNA in vitro (33).

We hypothesized that the influenza B NS1 protein acts as an IFN antagonist. We demonstrate here that growth in eggs of influenza B viruses containing altered NS1 proteins (B/234 and B/201 viruses) is inversely proportional to previously reported IFN inducibility (Fig. 6) (20, 21). B/234 and B/201 viruses were also attenuated in BALB/c mice (Table 2), showing dramatic reduction in pfu titers in the mouse lung at 3 days after infection as compared with wild-type B/Yamagata virus. Although attenuated, influenza B/234 and B/201 viruses induced an immunological response in BALB/c mice capable of reducing the infectivity of wild-type B/Yamagata virus (Fig. 7 and Table 2).

We speculate that growth attenuation of B/234 and B/201 viruses is a result of a decrease in the ability of altered NS1 proteins encoded by these viruses to act as IFN antagonists. However, the ability of influenza B viral NS1 to act as an IFN antagonist awaits further functional analysis.

To counteract the rapid and efficient induction of an antiviral state by type I IFNs, many viruses encode IFN antagonists (reviewed in ref. 34). The NS1 protein of influenza A virus has been shown to play an important role in this aspect of the replicative cycle of this virus (4). We show that influenza A and B viruses containing alterations in the NS1 protein are attenuated and provide protective immunity against challenge with wild-type virus. We propose that deletion of virally encoded IFN antagonists or mutagenesis of these proteins to reduce activity can be used as a general strategy to construct live viral vaccines that are optimally attenuated and immunogenic.