Vaccination inducing broad and improved cross protection against multiple subtypes of influenza A virus

Abstract

Development of an influenza vaccine that provides broadly cross-protective immunity has been a scientific challenge for more than half a century. This study presents an approach to overcome strain-specific protection by supplementing conventional vaccines with virus-like particles (VLPs) containing the conserved M2 protein (M2 VLPs) in the absence of adjuvants. We demonstrate that an inactivated influenza vaccine supplemented with M2 VLPs prevents disease symptoms without showing weight loss and confers complete cross protection against lethal challenge with heterologous influenza A viruses including the 2009 H1N1 pandemic virus as well as heterosubtypic H3N2 and H5N1 influenza viruses. Cross-protective immunity was long-lived, for more than 7 mo. Immune sera from mice immunized with M2 VLP supplemented vaccine transferred cross protection to naive mice. Dendritic and macrophage cells were found to be important for this cross protection mediated by immune sera. The results provide evidence that supplementation of seasonal influenza vaccines with M2 VLPs is a promising approach for overcoming the limitation of strain-specific protection by current vaccines and developing a universal influenza A vaccine.

Influenza A viruses continuously evolve by introducing mutations in the antigenic sites of the surface antigens hemagglutinin (HA) and neuraminidase. Current influenza vaccines that are based on neutralizing antibody responses to the highly variable influenza HA protein provide protection against homologous but not antigenically distinct heterologous viruses. In 2009, a swine-origin H1N1 virus rapidly spread worldwide and became the first pandemic of the 21st century. Previous seasonal influenza vaccination failed to effectively control the emergence and spread of the 2009 H1N1 virus. Since 1997, there have also been recurrent lethal outbreaks of highly pathogenic H5N1 avian origin influenza viruses (1). Development of an influenza vaccine that provides broad cross protection, overcoming the intrinsic limitation of the current inactivated vaccine, has been a scientific challenge since the first vaccination half a century ago.

Human influenza A viruses contain a highly conserved domain of the M2 protein (M2e) exposed on the surface of the virion. M2 is a membrane-anchored tetrameric protein expressed on cell surfaces and incorporated into influenza virions (2). However, despite the presence of the invariant domain of M2, humans lack antibody responses to M2 after vaccination or natural infection, indicating that M2 is poorly immunogenic (3). Previous studies reported immunization with M2e peptide fusion constructs linked to carrier vehicles as vaccine candidates (4–6). Antibodies to M2e have been shown to provide protection against lethal infection in animal models. However, the M2e fusion constructs reported previously provided limited protection against lethal infection as seen by significant weight loss and signs of disease even in the presence of potent adjuvants such as cholera toxin or heat-labile endotoxin derivatives, saponin QS21, Freund adjuvants, or bacterial protein conjugates (4, 5, 7–9). These previous studies suggest that M2 immunity alone might not be sufficient to prevent morbidity, and the adjuvant agents used may not be approved for humans as a result of their potential adverse effects. Moreover, the longevity and breadth of cross protection induced by M2-mediated vaccines largely remain unknown.

An approach presenting M2 on virus-like particles (VLPs) in a membrane-anchored form (M2 VLPs) would be more effective in presenting M2 to the immune system in its native form and at high concentration. We hypothesized that supplementing current vaccines with M2 VLPs as a conserved antigenic target would overcome their limited strain-specific protection. This study demonstrates that an inactivated viral vaccine supplemented with M2 VLPs provides complete cross protection against lethal challenge with heterologous and heterosubtypic influenza viruses by preventing weight loss and apparent disease symptoms, and confers long-lived cross-protective immunity.

Results

M2 VLP Supplementation Enhances M2-Specific Responses Induced by Inactivated Vaccine.

VLPs containing M2 from A/WSN/33 (M2 VLPs) virus were produced in insect cells coinfected with recombinant baculoviruses (rBVs) expressing the WT M1 and M2 proteins (Fig. 1 A and B). M2 VLPs showed a similar size as influenza virions as examined by EM (Fig. S1). The M2 content in VLPs was found to be comparable to that of virus (Fig. 1B). Here, we have investigated these VLPs as a supplement for influenza vaccines to enhance the immune responses to M2 and increase the breadth of protection against influenza A viruses. To determine the effects of M2 VLPs on inducing cross-reactive responses, groups of mice were intranasally immunized with 2 μg of inactivated A/PR/8/34 virus (PR8i) alone or inactivated A/PR8 virus (2 μg) supplemented with 10 μg of M2 VLPs (PR8i+M2VLP). M2-specific antibodies (Fig. 1C) and INF-γ–secreting cell responses (Fig. S2) were induced in the M2 VLP-supplemented group at levels that were significantly higher than those in the PR8i-alone group (Fig. 1C). As controls, vaccination with HIV VLPs as a supplement did not increase M2-specific immune responses above the low levels seen in the PR8i group and M1 VLPs without M2 did not induce M2-specific antibody responses, indicating that the increased immune responses are the result of M2 VLPs (Fig. 1C). Interestingly, vaccination with PR8i+M2VLP induced similar levels of M2e-specific antibodies as M2 VLPs alone (Fig. 1C). These results show that supplementing an inactivated influenza vaccine with M2 VLPs as a conserved antigenic target can significantly improve the immune response to M2.

Fig. 1.

M2 VLP-supplemented inactivated vaccine enhances antibodies to M2 and cross reactivity to heterosubtypic viruses. (A) Diagram of WT M2. EX, external domain (1–24 aa); TM, transmembrane domain (25–43 aa); CT, cytoplasmic domain (44–97 aa). (B) Characterization of M2 VLPs. Influenza M2 proteins in VLPs were detected on Western blot probed with mouse anti-M2e monoclonal antibody (14C2). Lane 1, VLP without M2; Lanes 2 through 5, 0.1, 0.5, 1.0, 2.0 μg of A/WSN virus; lane 6, 0.5 μg of M2 VLPs. (C) M2e-specific IgG antibody. (D) H5N1 A/Vietnam/1203/04-specific IgG1 and IgG2a isotype antibodies. (C and D) Groups of BALB/c mice (n = 9 per group) were intranasally immunized with 2 μg of A/PR8 inactivated vaccine alone (PR8i) or supplemented with 10 μg of VLPs; none, PBS (PR8i); M2, M2 VLPs (PR8i+M2VLP); HIV, HIV VLPs (PR8i+HIVVLP), M2 VLPs without PR8i vaccine (M2VLP), or M1 only VLPs without M2 and PR8i vaccine (M1 or control) at weeks 0 and 4. Serum samples were taken 3 wk after boost immunization. ELISA plates were coated with M2e peptide (2 μg/mL) or purified viruses (2 μg/mL) for determination of M2e or virus-specific antibody levels.

Dosage effects of VLP supplements (1, 5, 10 μg) indicated that a lower dose of M2 VLPs, but not unrelated HIV VLPs, could be used as a supplemental vaccine to enhance immune responses to M2 (Fig. S3A). No significant differences were observed in binding antibody levels to the purified H5 HA protein (Fig. S3B) or a M1 peptide pool (Fig. S3C) between the PR8i+M2VLP and PR8i groups. Importantly, it was noted that immune sera from M2 VLP supplemented vaccination (PR8i+M2VLP) enhanced the breadth of cross reactivities to different subtypes of H3N2 and H5N1 viruses (Fig. S3 D and E). Therefore, immunization with the M2 VLP supplemented vaccine significantly increases the cross-reactive response to heterosubtypic influenza A viruses.

The PR8i+M2 VLP-supplemented group also showed higher levels of IgG2a versus IgG1 specific to the H5N1 A/Vietnam/1203/04 virus (ratio of approximately 10) compared with the PR8i group (ratio of approximately 1; Fig. 1D), but no difference in binding antibody levels to the H5 HA protein (Fig. S3 B and F). The supplementation of inactivated A/PR8 vaccine with M2 VLPs did not induce significant cross-reactive hemagglutination inhibition (HAI) activities against the heterosubtypic H3N2 A/Philippines/82 (A/Phil) or H5N1 A/Vietnam/1203/04 viruses (Fig. S4). Also, immune sera from PR8i+M2VLP or PR8i groups did not show virus neutralizing activity or growth inhibition effects against the H3N2 heterosubtypic virus at significant levels (Fig. S5). Therefore, enhanced binding antibodies to M2e did not correlate with cross-reactive HAI response or heterosubtypic virus neutralizing activity. These results indicate that M2 VLPs can be an effective supplement for influenza vaccines to improve the breadth of cross reactivity to antigenically different viruses by enhancing M2 immunity, and that IgG2a isotype antibody recognizing M2e was preferentially induced by the M2 VLP supplement.

M2 VLP Supplementation Improves Heterologous and Heterosubtypic Protection.

We further investigated the potential effects of M2 VLP supplements on inducing cross protection. At 8 wk after boost, naive or immunized mice were challenged with a lethal dose (6 × LD₅₀) of heterosubtypic H3N2 A/Phil virus. All mice in the control group died from lethal infection (Fig. 2A). Mice immunized with inactivated A/PR8 virus or M2 VLPs alone showed a severe loss of body weight of as much as 16%, although 100% of mice survived infection (Fig. 2A). Also, the HIV VLP-supplemented PR8i vaccine group exhibited significant weight loss, indicating lack of effective protection against the heterosubtypic virus (Fig. 2A). In contrast, mice that were supplemented with M2 VLPs (PR8i+M2VLP) did not show signs of disease or weight loss. To better appreciate the heterosubtypic protective efficacy observed by the addition of M2 VLPs, lung viral titers at day 4 after challenge with A/Phil were determined (Fig. 2B). The PR8i+M2VLP group had significantly lower lung viral titers by more than 200-fold compared with the control group and by 52-fold compared with the PR8i group.

Fig. 2.

M2 VLP-supplemented vaccine improves cross protection against heterosubtypic H3N2 virus. Groups of mice were intranasally challenged with a lethal dose (6 × LD₅₀) of A/Phil (H3N2) virus at 4 wk after boost (n = 9). PR8i, inactivated virus alone; PR8i+M2VLP, M2 VLP supplemented PR8i vaccine; PR8i+HIV VLP, HIV VLP supplemented PR8i vaccine; M2VLP, M2 VLPs only; control, M1 only VLPs. (A) Weight change as a percentage. Bars indicate SDs daily. (B) Lung viral titers were determined by a plaque assay at day 4 after challenge (n = 4 of 9 challenged mice). Asterisk indicates significant difference between PR8i and PR8i+M2VLP groups (**P < 0.01).

To further analyze the breadth of cross protection, we tested protection against a lethal dose of a reassortant H5N1 A/Vietnam/1203/04 virus or the 2009 H1N1 A/California virus (Fig. 3). The PR8i vaccinated group showed a significant loss in weight of as much as 11%, whereas the PR8i+M2VLP group did not show any loss in body weight after challenge with the H5N1 A/Vietnam/1203/04 reassortant virus (Fig. 3A). In the case of challenge with the 2009 pandemic H1N1 A/California/2009 virus (Fig. 3B), the mice immunized with inactivated A/PR8 alone were severely ill, as evidenced by greater than 17% body weight loss, but the M2 VLP supplemented group showed only a slight transient loss in body weight. These results indicate that supplementing an inactivated vaccine with M2 VLPs significantly improves the cross-protective efficacy of the vaccine against heterologous and heterosubtypic influenza A viruses.

Fig. 3.

M2 VLP-supplemented vaccines confer improved and long-lasting cross protection against antigenically different influenza A viruses. A lethal dose (6 × LD₅₀) of reassortant H5N1 A/Vietnam/1203/04 (A) or A/California/04/2009 (H1N1) (B) influenza viruses was used as challenge at 4 wk after boost (n = 6) vaccination. (C) Mice were intranasally challenged with a lethal dose (6 × LD₅₀) of A/Phil (H3N2) virus (n = 9) at 7 mo after vaccination. Body weight changes were recorded daily. PR8i, vaccination with inactivated A/PR8 vaccine alone; PR8i+M2 VLP, inactivated A/PR8 vaccine supplemented with M2 VLPs; control, M1 only VLPs without M2. Bars indicate SDs.

M2 VLP Supplementation Induces Long-Lasting Cross-Protective Immunity.

To determine the duration of cross protection, mice immunized with A/PR8 alone or A/PR8 plus M2 VLPs were challenged with a lethal dose of heterosubtypic A/Philippines virus (6 × LD₅₀) at 7 mo after vaccination. All mice in the control group suffered severe body weight loss and died after challenge (Fig. 3C). The mice immunized with inactivated A/PR8 virus vaccine showed more than 12% loss in body weight and a significant delay in body weight recovery. In contrast, mice in the M2 VLP-supplemented group did not show weight loss and all survived the challenge infection. These results indicate the induction of long-lasting cross-protective immunity by influenza vaccine supplemented with M2 VLPs.

Clodronate Treatment Abrogates Cross Protection by M2 VLP Immune Sera.

M2 VLP immune sera were tested for their protective efficacy against A/Phil virus (H3N2) in naive mice intranasally infected with a lethal dose of the virus mixed with naive or immune sera. The undiluted sera collected from naive mice or mice immunized with inactivated A/PR8 virus did not confer any protection against lethal infection with H3N2 subtype A/Phil virus (Fig. 4A). In contrast, undiluted or twofold diluted immune sera from the PR8i+M2VLP group conferred complete protection against A/Phil virus without weight loss, and 100% survival was observed even with as much as eightfold diluted immune sera (Fig. 4A). To better understand the potential cross-protective mechanisms, naive mice were pretreated with clodronate liposomes to deplete lung airway dendritic cells (DCs) and macrophage cells. The depletion efficiency of alveolar DC/macrophage cells was in the range of 60% to 70% at 1 d after treatment (Fig. S6), which is similar to the depletion efficiency reported in a previous study (10). Cross protection by immune sera from immunization with PR8i+M2 VLP vaccine was not observed in naive mice that were pretreated with clodronate liposomes before infection with a mixture of 2009 H1N1 virus and immune sera (PR8i+M2/CL[+]; Fig. 4 B and C), although this group showed a significant delay in weight loss and mortality compared with a naive group treated with clodronate liposomes. Therefore, DC/macrophage immune cells may be important in immune serum-mediated cross protection. Interestingly, mice treated with heat-inactivated immune sera mixed with a lethal dose of A/Phil virus showed moderate loss in body weight (Fig. S7), indicating that heat-sensitive serum components such as complement might be involved in providing cross protection.

Fig. 4.

Immune sera from PR8i+M2VLP-immunized mice transfer enhanced cross protection in naive mice. (A) Protection against A/Phil (H3N2, 6 × LD₅₀) by immune sera (n = 6, BALB/c mice). Dilutions (in fold) of immune sera are indicated in parentheses. PR8i, immune sera from PR8i group; PR8i+M2VLP, immune sera from PR8i+M2VLP group; control, sera from unimmunized control group. (B and C) Effects of clodronate liposomes on protective efficacy against A/Califonia/04/2009 (H1N1; 6 × LD₅₀) by immune sera (n = 6, BALB/c mice). (B) Body weight changes. P, statistical significance during a period of dates is indicated; p1 between PR8i+M2VLP/CL(+) and naive/CL(+), p2 between PR8i+M2VLP and other groups. (C) Survival rates. PR8i+M2, untreated naive mice with PR8i+M2VLP immune sera and virus; PR8i+M2/CL(+), Clodronate liposome-treated naive mice with PR8i+M2VLP immune sera and virus, naive, untreated naive mice with control immune sera and virus; naive/CL(+), clodronate liposome-treated naive mice with mock immune sera and virus.

Discussion

Current influenza vaccines based on the variable surface antigens do not provide effective protection against pandemic outbreaks such as the 2009 H1N1 virus. Here, we investigated the potential of VLPs containing the highly conserved influenza M2 protein to contribute to cross-protective immunity. We found that vaccination with M2 VLPs alone did not induce strong protective immunity as the immunized mice suffered significant weight loss. We hypothesized that supplementing conventional vaccines with M2 VLPs would confer significantly improved cross protection. The present study demonstrated that addition of M2 VLPs to an inactivated influenza vaccine effectively induced M2-specific humoral and cellular immune responses, indicating that M2 VLPs presenting M2 to the immune system without the immunodominant HA in the same particle are more effective in inducing anti-M2e immune responses. Notably, the M2 VLP-supplemented A/PR8 vaccine conferred cross protection against the distantly related 2009 H1N1 pandemic virus as well as heterosubtypic viruses of the H3N2 and H5N1 subtypes, and completely prevented disease symptoms. Therefore, supplementing conventional vaccines with M2 VLPs can overcome the inefficient cross protection by current inactivated vaccines.

Influenza VLPs are likely to present the M2 protein in a membrane-anchored form mimicking its native conformation. EM images suggest that influenza M1-derived M2 VLPs produced in insect cells have similar sizes as influenza virions, although they do not have spikes on their surfaces. These results on M1-derived VLPs are consistent with previous studies demonstrating a major role of M1 in budding and generating influenza VLPs (11, 12). However, in other studies, influenza M1 did not play a major role in driving particle budding when influenza proteins were expressed in 293T mammalian cells (13). Therefore, production of M1-derived influenza VLPs might be affected by the expression system and/or cell type.

Supplementation of inactivated virus with M2 VLPs resulted in enhanced levels of antibodies that recognize M2e and are cross reactive with influenza A viruses of different subtypes. In particular, we observed that M2 VLP supplemented vaccine induces significantly higher levels of IgG2a antibodies binding to purified virions than those induced by inactivated vaccine alone. IgG2a antibody is known to interact more efficiently with complement and Fc receptors by virtue of its Fc domain properties (14). Therefore, induction of IgG2a as a dominant cross-reactive antibody might contribute to viral clearance, possibly via mechanisms involving the complement system, stimulation of antibody-dependent cellular cytotoxicity and clearance of opsonized virus by macrophages (15). In support of a possible role of complement, we observed that heat treatment of immune sera showed low efficacy in conferring cross protection compared with untreated immune sera that prevented weight loss.

M2 VLP supplemented vaccination did not induce significant levels of heterosubtypic virus-neutralizing activity. These results are consistent with previous studies demonstrating that, in contrast to anti-HA antibodies, M2e binding antibodies did not neutralize the virus (16). An M2e peptide coupled to the immunodominant domain of the hepatitis B core protein was shown to induce antibodies that bound inefficiently to the free virus (16). In contrast, M2 VLP supplemented vaccination induced strong cross-reactive binding antibodies to virions of different subtypes. Nonneutralizing anti-influenza humoral immunity was shown to confer protection (17, 18) and to be dependent on opsonophagocytosis of influenza virions by macrophages (15). It was recently reported that monoclonal antibodies binding to multimeric M2 peptides were more protective than antibodies binding to monomeric M2 peptide (19). In this regard, induction of antibodies cross reactive to heterosubtypic virions by M2 VLP supplemented vaccination seems to be important for cross-protective immunity.

We observed that immune sera from M2 VLP-supplemented vaccines played an important role in providing protection against lethal infection. In contrast, immune sera from mice vaccinated with PR8i alone did not confer cross protection against lethal infection with H3N2 virus. Nonetheless, mice directly immunized with PR8i were partially protected from lethal H3N2 challenge and then recovered. It is possible that the inactivated PR8i vaccine alone is a weak immunogen inducing a threshold of immunity, which is reflected by high lung viral titers. An alternative explanation is that other factors such as cellular immune response to virus internal antigens together with antibodies might contribute to weak cross protection after vaccination.

Multiple mechanisms might be involved in conferring protection by nonneutralizing antibodies (16, 17, 20). DCs and macrophage cells might play a role in mediating cross-protection provided by nonneutralizing antibodies. It was previously reported that macrophages were involved in presenting cross-protective effects observed in vitro by nonneutralizing antibodies to influenza antigens (18). DC/macrophage cells that take up clodronate through liposome vesicles are depleted by apoptosis (21). We used this approach to provide evidence that the cross protection conferred by immune sera from M2 VLP-supplemented vaccination may be mediated by DC/macrophage cells. An alternative possible mechanism is that anti-M2 antibodies rapidly promote the adaptive response by enhancing uptake of antibody-bound virus particles via DC/macrophage cells, as we observed a rapid increase in IFN-γ–secreting cell responses at early times after challenge. Studies are ongoing to further understand the possible roles of CD4 and CD8 T cells, and antibodies possibly recognizing other influenza viral proteins such as nucleoprotein in conferring protection (17). In summary, the present results indicate that M2 antibodies together with alveolar DC/macrophage cells play an important role in conferring cross protection against both heterologous and heterosubtypic influenza A viruses.

Materials and Methods

Viruses and VLPs.

Influenza A viruses A/PR/8/34 (H1N1), A/Philippine/2/82 (H3N2), A/California/04/09 (H1N1), A/WSN/33 (H1N1), and reassortant A/Vietnam/1203/04 (H5N1; HA and neuraminidase were derived from A/Vietnam/1203/04, and the remaining backbone genes from A/PR/8/34 virus) were propagated in chicken eggs and used for challenge studies. M2 VLPs and HIV VLPs were produced in insect cells coinfected with rBVs expressing viral structural proteins. Details are provided in SI Materials and Methods.

Vaccination and Challenge.

Female BALB/c mice (n = 9, 6–8 wk old) were intranasally immunized with inactivated A/PR8 vaccine (2 μg) alone or supplemented with M2 VLPs or HIV VLPs at weeks 0 and 4. To investigate heterosubtypic protective immunity, immunized mice were challenged with a lethal dose (6 × LD₅₀) of different viruses as indicated.

Determination of Antibody Responses, Lung Viral Titers, and INF-Secreting Cells.

Serum antibody responses were determined by ELISA using synthetic M2e peptide (2 μg/mL; SLLTEVETPIRNEWGCR), recombinant H5 HA protein (4 μg/mL), or whole inactivated virus (4 μg/mL) as a coating antigen as previously described (22). Lung viral titers and IFN-γ–secreting cell spots were determined as previously described (23).

Cross-Protective Efficacy of Immune Sera.

To test cross-protective efficacy of immune sera in vivo, 25 μL of a lethal dose of influenza virus (6 × LD₅₀) mixed with 25 μL immune sera with or without heat inactivation (56 °C, 30 min) were used to infect naive mice (n = 4, BALB/c), and body weight changes and survival rates were monitored daily. To deplete DC/macrophage cells, 6 h before infection with a virus and serum mixture, some groups of naive mice (n = 6, BALB/c) were intranasally treated with clodronate liposomes as previously described (24, 25).

Statistical Analysis.

To determine statistical significance, a two-tailed Student t test and one-way ANOVA were used when comparing two or more different groups, respectively. A P value less than 0.05 was considered to be significant.