ABSTRACT
The complement pathway is involved in eliminating antigen immune complexes. However, the role of the C3 complement system remains largely unknown in influenza virus M2 extracellular (M2e) domain or hemagglutinin (HA) vaccine-mediated protection after vaccination. Using a C3 knockout (C3 KO) mouse model, we found that complement protein C3 was required for effective induction of immune responses to vaccination with M2e-based or HA-based vaccines, which include isotype class-switched antibodies and effector CD4 and CD8 T cell responses. C3 KO mice after active immunization with cross-protective nonneutralizing M2e-based vaccine were not protected against influenza virus, although low levels of M2e-specific antibodies were protective after passive coadministration with virus in wild-type mice. In contrast, C3 KO mice that were immunized with strain-specific neutralizing HA-based vaccine were protected against homologous virus challenge despite lower levels of HA antibody responses. C3 KO mice showed impaired maintenance of innate immune cells and a defect in innate immune responses upon exposure to antigens. The findings in this study suggest that C3 is required for effective induction of humoral and cellular adaptive immune responses as well as protective immunity after nonneutralizing influenza M2e vaccination.
IMPORTANCE Complement is the well-known innate immune defense system involved in the opsonization and lysis of pathogens but is less studied in establishing adaptive immunity after vaccination. Influenza virus HA-based vaccination confers protection via strain-specific neutralizing antibodies, whereas M2e-based vaccination induces a broad spectrum of protection by immunity against the conserved M2e epitopes. This study revealed the critical roles of C3 complement in inducing humoral and cellular immune responses after immunization with M2e or HA vaccines. C3 was found to be required for protection by M2e-based but not by HA-based active vaccination as well as for maintaining innate antigen-presenting cells. Findings in this study have insight into better understanding the roles of C3 complement in inducing effective innate and adaptive immunity as well as in conferring protection by cross-protective conserved M2e vaccination.
INTRODUCTION
Influenza viruses are members of the family Orthomyxoviridae and contain eight segmented negative-sense RNA genomes (1, 2). Influenza A viruses are classified into different subtypes based on their major surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA) (3). Currently, 18 HA (H1 to H18) and 11 NA (N1 to N11) subtypes have been known to exist and continue to mutate in various hosts, including humans, birds, and pigs (1, 2). Current influenza vaccines are targeting strain-specific HA antigens and confer protection against homologous virus so that annual updates of influenza vaccines are required to match the antigenicity of the virus strains which are predicted to circulate (4). This influenza vaccine strategy is not effective in preventing drift or shift mutant seasonal viruses and pandemic outbreaks, raising the need for developing broadly cross-protective influenza vaccines (5).
Influenza virus M2 is an ion channel protein incorporated into the surface of the virion, playing a role in viral entry (6–8). The extracellular domain of M2 (M2e) is a highly conserved antigen across human influenza A subtypes (9, 10). Therefore, targeting M2e has been considered a promising strategy for the development of broadly cross-protective influenza vaccines (5). Previously, we demonstrated that virus-like particles (VLP) containing heterologous tandem repeat M2e (M2e5x VLP) confers cross-protection against multiple subtypes of influenza viruses (11). The mechanism of cross-protective immunity by M2e vaccines has not been fully understood yet. M2e-specific antibodies are considered a main immune correlate for conferring protection against multiple strains of influenza virus infection, even though M2e antibodies lack the virus-neutralizing activity (12, 13). In addition to the M2e-specific antibodies, M2e-specific T cell responses are also important for optimal cross-protection against influenza virus infection (14, 15). Here, we studied the complement-dependent mechanism of M2e-mediated immunity compared with HA-based immunity.
Complement is a primitive surveillance system and contributes to lowering the burden of infected pathogens during an early phase of infection (16, 17). The complement system directly mediates viral clearance, including neutralization, opsonization, lysis, and phagocytosis, via complement receptors (18). Moreover, the complement system regulates both humoral and T cell immunity (19). The complement system is involved in the B cell responses via complement receptors CD21 and CD35 (20) and by localizing antigens to follicular dendritic cells (DCs), which are specialized cells secreting chemoattractant chemokines for B lymphocytes (21). Complement C3 protein was reported to play a role in inducing CD4 and CD8 T cell responses and in lung viral clearance after influenza virus infection, whereas mice deficient for complement receptors CR1 and CR2 (Cr2−/− mice) cleared the infection normally (22). C3 was required for effective control and protection against influenza virus infection, as reported in C3 knockout (C3 KO) mouse studies (23, 24). The mechanism by which C3 controls innate and adaptive immunity remains not fully understood.
The roles of C3 in inducing adaptive immunity and conferring protection after vaccination remain largely unknown. In this study, using a C3 KO mouse model, we investigated the possible roles of C3 in inducing immune responses and protection after immunization with cross-protective nonneutralizing M2e5x VLP or strain-specific neutralizing H5 HA VLP vaccines and compared them to those after virus infection. C3 was found to play an important role in inducing innate and adaptive immune responses to influenza virus infection or VLP vaccination. C3 was also required for nonneutralizing immune-mediated protection by M2e5x VLP but not for neutralizing immune-mediated protection by H5 HA VLP. Possible underlying mechanisms have been explored and discussed.
RESULTS
Early induction of IgG antibodies to viral infection is severely impaired in C3-deficient mice.To determine the IgG antibody levels and protection in C3 KO mice under influenza virus infection, groups of mice were infected with two different doses of A/Philippines/2/1982 H3N2 virus (Fig. 1A to D). C3 KO mice were much more sensitive to influenza virus infection than wild-type (WT) (C57BL/6) mice (Fig. 1A to D). The sublethal dose (0.8× 50% lethal dose [LD50]) of A/Phil H3N2 virus for WT mice leads to 100% lethality in C3 KO mice, and a 10-fold lower dose (0.8× LD50) induced a comparable level of weight loss (∼22%) in C3 KO mice, whereas no significant weight loss (2 to ∼3%) was seen in WT mice (Fig. 1A to D). C3 KO mice rarely induced IgG and IgG isotype antibodies at day 7 postinfection with a low dose of A/Philippines/2/1982 H3N2 (A/Phil H3N2) (Fig. 1E to H), suggesting a delay in inducing IgG antibodies specific for virus. At a later time point, day 14 after viral infection, C3 KO mice could induce moderate levels of H3N2 virus-specific IgG, IgG1, IgG2a, and IgG2b antibodies, which were significantly lower than those in WT mice (Fig. 1I to L).
C3 KO mice show a defect in inducing virus-specific IgG antibodies and high susceptibility to influenza virus infection. Each group of mice (n = 5) was infected with 0.8× LD50 (high dose in panels A and B) or 0.08× LD50 (low dose in panels C and D) for WT and C3 KO of influenza virus (A/Phil and H3N2). Body weight changes (A and C) and survival rates (B and D) were monitored for 14 days after infection. IgG, IgG1, IgG2c, and IgG2b were detected by serum ELISA at 7 days after infection (E to H) and at 14 days after the same low-dose infection (I to L). Sera were serially diluted and IgG antibodies determined by ELISA coated with inactivated virus (A/Phil) antigen. Error bars indicate means ± SEM. (M) Total serum IgM, IgG, and isotype-switched IgG antibodies were measured in naive WT and C3 KO mice. Total natural antibodies were captured using goat anti-mouse IgM, IgG, and isotype-switched IgG antibodies.
To determine the antibody phenotypes as a result of C3 deficiency in naive mice, we compared natural antibody levels between WT (C57BL/6) and C3 KO mice (Fig. 1M). There were no differences in IgM and IgG isotypes (IgG1, IgG2c, and IgG2b) in sera from WT and C3 KO mice in the absence of antigen exposure, indicating no defect in maintaining non-antigen-specific natural antibodies in C3 KO mice. These results suggest that C3 plays a role in inducing virus-specific IgG antibodies and immunity in response to viral infection.
C3-deficient mice are less effective in inducing M2e-specific IgG responses after M2e5x VLP immunization.To evaluate the role of C3 in mediating cross-protective nonneutralizing M2e immunity, we compared IgG antibody levels induced in WT and C3 KO mice after M2e5x VLP vaccination (Fig. 2). Groups of mice in WT and C3 KO mice were intramuscularly immunized with 10 μg of M2e5x VLP by a prime-boost strategy. At 2 weeks after prime (Fig. 2A to D) and boost (Fig. 2E to H) immunization, IgG isotype antibodies specific for M2e were measured by enzyme-linked immunosorbent assay (ELISA). Overall, M2e5x VLP-immunized WT mice showed high levels of M2e-specific IgG1 and IgG2c as well as IgG2b antibodies after both prime (Fig. 2A to D) and boost (Fig. 2F to H) immunization. However, the levels of M2e-specific IgG isotypes were significantly lower in C3 KO mice, by 6-fold after prime and by 36-fold after boost immunization, than those in WT mice (Fig. 2A to H). In particular, the IgG1 isotype, T helper type 2 (Th2), was not induced above the level of detection in C3 KO mice (Fig. 2F). These results indicate that C3 protein is important for the production of antigen-specific IgG antibodies after M2e5x VLP vaccination.
C3 KO mice have a defect in inducing M2e-specific IgG antibodies after vaccination. Groups of mice (n = 5) were boost immunized with 10 μg of M2e5x VLP at a 3-week interval. IgG (A), IgG1 (B), IgG2c (C), and IgG2b (D) from prime immune sera and IgG (E), IgG1 (F), IgG2c (G), and IgG2b (H) from boost immune sera were detected by ELISA coated with M2e peptide antigen. Error bars indicate means ± SEM.
M2e5x VLP-immunized C3-deficient mice are not protected after challenge.To determine whether the C3 complement system has a role in M2e-mediated cross-protection, the groups of WT and C3 KO mice were prime and boost immunized with M2e5x VLP at an interval of 3 weeks, followed by challenge infection with a sublethal dose (0.8× LD50) of A/Phil H3N2 virus 3 weeks later (Fig. 3). Since C3 KO mice showed approximately 10-fold higher susceptibility to influenza virus infection, we used a 10-fold lower dose (0.08× LD50) to infect C3 KO groups than that of WT groups in order to evaluate the protective efficacy of vaccination in C3 KO mice. As expected, WT mice immunized with M2e5x VLP showed a moderate weight loss (10 to 12%) (Fig. 3A and B), whereas influenza virus infection resulted in significant weight loss in the C3 KO mice that were immunized with M2e5x VLP, displaying similar severe weight loss and survival rates in both the immunized and naive C3 mouse groups (Fig. 3C and D). Protective efficacy was further confirmed by lung viral titers at day 7 postinfection. M2e5x VLP vaccination of WT mice conferred 100-fold lower lung viral titers than the naive control, but this protection of viral clearance was not observed in vaccinated C3 KO mice (Fig. 3E). These results suggest that C3 is an essential immune component in active M2e vaccination-mediated protection.
C3 KO mice immunized with M2e5x VLP are not protected against influenza virus. Groups of mice (n = 5) were boost immunized with 10 μg of M2e5x VLP. At 4 weeks after boost immunization, WT and C3 KO mice were challenged with influenza virus (A/Phil and H3N2) at 0.8× LD50 and 0.08× LD50, respectively. Body weight changes (A and C) and survival rates (B and D) were monitored for 14 days. Lung viral titers were determined by an egg inoculation assay at 7 days after challenge. Statistical significance was calculated by 2-way ANOVA and a Bonferroni's multiple-comparison test. *, P < 0.05; **, P < 0.01, ***, P < 0.001; ****, P < 0.0001. Significance among the groups is indicated.
C3 KO mice display a defect in generating antibody-secreting cells and germinal center phenotypic cells.Developing long-lived antibody-secreting cell responses is important for providing long-term protection. We determined whether C3 would play a role in generating long-lived plasma cells and germinal center (GC) phenotypic B cells. In vitro antibody production was measured from cultured cells of bone marrow (BM) and spleens collected from immune mice at 7 days after influenza virus challenge (Fig. 4). As expected, both bone marrow cells and splenocytes from the WT group immunized with M2e5x VLP secreted high levels of M2e-specific IgG antibodies, indicating effective development of M2e-specific antibody-secreting cell responses (Fig. 4A and B). However, C3 KO mice showed significantly lower production of M2e-specific IgG antibodies than WT mice in both bone marrow cells and splenocytes (Fig. 4A and B). In addition, C3 deficiency resulted in approximately 4-fold lower levels of germinal center phenotypic (CD19+ IgD− B220+ GL7+) B cells in spleens than in those of the WT control (Fig. 4C and D). These results suggest that C3 plays a key role in the development of antigen-specific antibody-secreting cells and germinal center phenotypic B cells after vaccination.
M2e5x VLP vaccination induces lower levels of IgG antibody-secreting and germinal center B cell responses in C3 KO mice than those in WT mice. (A) IgG production from in vitro cultures of bone marrow cells from vaccinated WT and C3 KO mice. (B) IgG production from in vitro cultures of spleen cells of immunized WT and C3 KO mice. Bone marrow and spleen cells were harvested on day 7 postinfection. Cells were cultured for 1 day in the presence of M2e peptide antigen. IgG levels were detected by ELISA. (C and D) Germinal center B cells in mediastinal lymph nodes (MLN). MLN cells were harvested, stained with B220, GL7, IgD, and CD19, and analyzed by flow cytometry at 7 days after infection. Statistical significance was calculated by 2-way ANOVA and a Bonferroni's multiple-comparison test. ***, P < 0.001; ****, P < 0.0001. Significance among the groups is indicated.
C3 is required to develop effective M2e-specific cellular immune responses.It has been known that T cell responses contribute to cross-protective immunity in M2-mediated immune responses (14, 25). Thus, we determined whether C3 would be required for developing M2e-specific T cell immunity. Lung cells and splenocytes were collected at 7 days after challenge and cultured in vitro. After stimulation with an M2e peptide, cytokine-producing cell spots were counted as an indicator of T cell responses (Fig. 5). The M2e5x VLP-immunized WT group showed more than 7-fold higher levels of gamma interferon (IFN-γ)-secreting and 5-fold higher levels of interleukin-4 (IL-4)-secreting lung cells than the naive WT control group (Fig. 5A and B). However, C3 KO mice failed to develop cytokine-producing T cells, showing no difference between the M2e5x VLP-immunized and naive C3 KO groups (Fig. 5A and B). A similar pattern was observed in splenocyte cultures. High levels of IFN-γ- and IL-4-secreting splenocytes were induced in WT mice immunized with M2e5x VLP but not in C3 KO mice (Fig. 5C and D).
C3 KO mice induce lower cellular immune responses to M2e5x VLP vaccination. (A) IFN-γ-secreting cells in lungs. (B) IL-4-secreting cells in lungs. (C) IFN-γ-secreting cells in splenocytes. (D) IL-4-secreting cells in splenocytes. Lung cells and splenocytes were isolated from WT (C57BL/6) and C3 KO mice previously immunized with M2e5x VLP at 7 days after challenge with A/Phil H3N2. Cytokine-producing cell spots were counted by an ELISPOT assay reader. Statistical significance was calculated by 2-way ANOVA and a Bonferroni's multiple-comparison test. **, P < 0.01; ****, P < 0.0001. Significance among the groups is indicated.
To investigate further details of M2e-specific T cell phenotypes, intracellular cytokine staining and flow cytometry assays were applied by a strategy of cytokine-positive T cell gating (see Fig. S1 in the supplemental material). Lung cells and splenocytes were harvested from WT and C3 KO mice previously immunized with M2e5x VLP at 7 days after virus infection. M2e5x VLP-immunized WT mice showed significantly high numbers of IFN-γ-secreting CD4+ and CD8+ T cells in lungs (Fig. 6A and B) and spleens (Fig. 6C and D). These results suggest that C3 is required to develop effective M2e-specific T cell responses.
C3 KO mice show a defect in inducing M2e-specific IFN-γ-secreting CD4+ and CD8+ T cells after vaccination. (A) IFN-γ-secreting CD4+ T cells in lungs. (B) IFN-γ-secreting CD8+ T cells in lungs. (C) IFN-γ-secreting CD4+ T cells in spleens. (D) IFN-γ-secreting CD8+ T cells in spleens. Lung and spleen cells were isolated from WT (C57BL/6) and C3 KO mice previously immunized with M2e5x VLP at 7 days after challenge with A/Phil H3N2. Isolated cells were stained with CD45, CD4, and CD8α and then intracellularly with IFN-γ antibodies and analyzed by flow cytometry. Statistical significance was calculated by 2-way ANOVA and a Bonferroni's multiple-comparison test. **, P < 0.01; ****, P < 0.0001. Significance among the groups is indicated.
M2e-specific antibodies induced in C3 KO mice are protective.Since M2e5x VLP-immunized C3 KO mice were not protected (Fig. 3), we determined whether M2e-specific antibodies induced in C3 KO mice would have protective capability. Naive WT mice (BALB/c) were infected with a mixture of M2e5x VLP immune sera and a lethal dose of H3N2 influenza A/Phil virus, and then body weight changes were monitored for 14 days (Fig. 7). Naive mice that received M2e5x VLP immune sera from WT (Fig. 7A) and C3 KO (Fig. 7B) mice together with H3N2 virus showed 19% and 24% weight loss, respectively, and then recovered normal weight with 100% survival rates. A lower level of M2e-specific antibodies in C3 KO mouse immune sera appears to contribute to delaying weight loss and 100% survival protection (Fig. 7B). The mice that received WT naive sera with H3N2 virus showed a delay of 2 days in weight loss and survival rates compared to those of the mice with C3 KO naive sera and H3N2 virus. Naive mice that received naive sera and virus did not survive, suggesting that M2e-specific antibodies induced in C3 KO mice confer survival benefits of protection.
Sera from M2e5x VLP-vaccinated C3 KO mice are protective in wild-type naive mice. Immune sera collected from M2e5x VLP-immunized WT (C57BL/6) (A) and C3 KO (B) mice were incubated with a lethal dose of influenza virus (A/Phil H3N2). Naive BALB/c mice were intranasally infected with influenza virus mixed with immune sera or naive sera. Final infection doses were 0.8× LD50 for WT and 0.08× LD50 for C3 KO. Body weights were monitored for 14 days.
HA VLP-vaccinated C3 KO mice induce hemagglutination inhibiting antibodies and protection.Since C3 KO mice that were immunized with M2e5x VLP could not confer protection, we determined whether HA VLP vaccines would induce protection in C3 KO mice. Groups of WT and C3 KO mice were intramuscularly immunized with 10 μg of H5 HA VLP containing HA derived from A/Indonesia/05/2005 virus. H5 HA VLP-immunized WT mice showed 6-fold higher levels of virus-specific IgG and IgG2b antibodies after prime (Fig. 8A) and 36-fold higher levels of IgG2c and IgG2b isotype IgG antibodies after boost immunization (Fig. 8E to H) compared to those in C3 KO mice. Although C3 KO mice induced smaller amounts of virus-specific IgG isotypes than those in WT mice after H5 HA VLP vaccination, the levels of IgG, IgG2b, and IgG2c isotype antibodies induced in C3 KO mice were substantially higher (Fig. 8A to H). Consistent with the levels of binding IgG antibodies, immune sera from C3 KO mice showed significant levels of hemagglutination inhibition (HI) titers in a range of 128 to 256, although these levels of HI titers in C3 KO mice were significantly lower than those in WT immune sera (Fig. 8I). C3 KO mice immunized with H5 VLP were protected against homologous rgH5N1 influenza virus (1.5 LD50) without displaying weight loss (Fig. 8J and K). These results suggest that production of IgG and HI antibodies is affected in the absence of C3; however, as the overall HI titers remained above the threshold necessary for neutralization, no impact on survival was observed in C3 KO mice.
C3 KO mice have a defect in HA-specific IgG production but not in HA-mediated homologous protection. Groups of mice (n = 5) were boost immunized with 10 μg of H5 HA VLP. IgG (A), IgG1 (B), IgG2c (C), and IgG2b (D) from prime immune sera and IgG (E), IgG1 (F), IgG2c (G), and IgG2b (H) from boost immune sera were detected using ELISA plates coated with vaccine-specific antigen, inactivated rgH5N1 virus (A/Indonesia/05). Error bars indicate means ± SEM. Vaccinated mice (n = 5/group) with 10 μg of H5 VLP and control mice challenged with 0.8× LD50 of rgH5N1 influenza virus (rgA/Indonesia H5N1) for WT and 0.08× LD50 for C3 KO mice. (I) HI titers against rgH5N1 were determined from immune sera of H5 VLP. Body weight changes (J) and survival rates (K) were monitored for 14 days. The dotted line indicates the detection limit. Statistical significance was calculated by 2-way ANOVA and a Bonferroni's multiple-comparison test. **, P < 0.01; ****, P < 0.0001. Significance among the groups is indicated.
C3 deficiency leads to an abnormal cellularity of innate immune cells.To better understand the roles of C3 in inducing adaptive immune responses, we examined the innate immune cell phenotypes in the peritoneal cavity by flow cytometry gating of cell populations based on phenotypic cell surface markers (Fig. S2). Naive C3 KO mice showed significantly lower levels in the cellularity of innate immune cells than WT mice. Total macrophages, activated macrophages, plasmacytoid dendritic cells (pDC), CD11bhigh DC, and CD11blow DC (Fig. 9A) subsets were found to be present at significantly reduced levels in naive C3 KO mice than in WT mice.
Cellular phenotypes in WT and C3 KO mice. (A) Cellularity of different phenotypic cells in peritoneal exudates from WT and C3 KO mice (n = 5). Cells in peritoneal exudates were harvested and their phenotypes and cellularity were determined. (B to G) Cellularity of lung cells at day 5 postinfection. The phenotypic markers used to identify immune cells are the following: macrophages, CD11b+ F4/80+; MHC-IIhigh macrophages, CD11b+ F4/80+ MHC-IIhigh; monocytes, CD11b+ Ly6chigh F4/80+; neutrophils, CD11b+ Ly6clow F4/80−; eosinophils, CD11b+ SiglecF+ CD11c− F4/80−; pDCs, CD11c+ B220+ MHC-IIhigh; CD11bhigh DCs, CD11c+ CD11bhigh MHC-IIhigh; CD11blow DCs, CD11c+ CD11blow MHC-IIhigh. Statistical significance was calculated by 2-way ANOVA and a Bonferroni's multiple-comparison test. *, P < 0.05; **, P < 0.01; ***, P < 0.001. Significance among the groups is indicated.
In addition, the cellularity of lung cells was determined 5 days after H3N2 A/Phil influenza virus infection of WT and C3 KO mice (Fig. 9B to G). C3 KO mice displayed lower levels of innate immune cells, including macrophages (Fig. 9B), monocytes (Fig. 9C), neutrophils (Fig. 9D), eosinophils (Fig. 9E), pDC (Fig. 9F), and CD11blow DC (Fig. 9G) subsets than those of WT mice after infection. Before infection, naive WT and C3 KO mice show similar levels of these innate immune cells present in the lungs (Fig. 9B to G). Similarly, C3 KO mice showed defects in recruiting innate immune cells (monocytes, neutrophils, and DC subsets) in the mediastinal draining lymph nodes after influenza vaccination and challenge (data not shown). These results suggest that low cellularity and less effectiveness in recruiting innate immune cells are correlated with the defects in developing adaptive immunity in C3 KO mice after influenza virus infection or vaccination.
DISCUSSION
It has not been well understood how complement C3 plays a role in inducing immune responses and protection after active neutralizing or nonneutralizing antigen vaccination. The present study provides evidence that the complement system has a critical role in inducing humoral and cellular immune responses to vaccination and in conferring protection by immunization with cross-protective but conserved nonneutralizing M2e vaccine or strain-specific neutralizing HA antigen vaccine. M2e5x VLP immunization of C3 KO mice induced moderate levels of IgG antibodies, which were significantly lower than those in WT mice. The WT C57BL/6 mice with M2e5x VLP active immunization showed significantly better recovery from weight loss than naive WT mice after challenge. In contrast, the C3 KO mice with M2e5x VLP active immunization displayed severe weight loss and low survival rates, similar to those in naive C3 KO mice even after challenge at a 10-fold lower dose, indicating no evidence of protection. In contrast, C3 deficiency did not affect the neutralizing antibody-mediated homologous protection. However, the IgG and HI antibody levels were significantly lower in actively vaccinated C3 KO mice than in WT C57BL/6 mice. The results in this study provide evidence that C3 is required for effectively generating adaptive immune responses after vaccination as well as conferring protection with influenza M2e vaccination inducing nonneutralizing immunity.
Previously, passive transfer approaches of M2e-specific antibodies have demonstrated controversial results in describing the roles of the complement system. When C3 KO mice were passively administered immune sera of M2e conjugate (M2e-HBc) vaccination, a survival rate similar to that of WT mice was observed, indicating that complement C3 is not required for M2e antibody-mediated passive protection (26). In contrast, Wang et al. reported that complement C3 is important for lung viral clearance of infected mice by passive transfer of anti-M2e monoclonal antibodies (27). We found that there was no significant difference in conferring protection by M2e-specific serum antibody itself generated from C3 KO mice compared to that in WT mice. A possible explanation is that M2e-specific antibodies induced in C3 KO mice are sufficient to confer the in vivo survival protection against influenza virus infection when a mixture of virus and sera was intranasally inoculated into naive WT BALB/c mice with the intact C3 complement system.
We observed that priming of T cells and generation of GC B cells in response to vaccination were severely compromised in C3 KO mice. Complement receptor CR2 also might be involved in retaining antigens on follicular dendritic cells playing a critical role in GC formation and class switching of activated B cells (28). Particularly after boost vaccination of C3 KO mice, a more significant difference was observed in inducing M2e-specific IgG1 and IgG2b class-switched isotype antibodies than in WT mice. A prominent defect in IgG1 (Th2 type) antibody induction is partially consistent with a report that C3 was involved in Th2-dependent CD4 T cell induction for eosinophil infiltration and IL-4 production in a murine model of pulmonary allergy (29). In contrast to M2e5x VLP antigen, induction of IgG antibodies was not detected after immunization of C3 KO mice with T-dependent bacteriophage antigen (30). This difference between VLP and bacteriophage antigens might be due to the fact that VLP was able to induce antigen-specific IgG and isotype-switched antibodies in CD4-deficient mice after vaccination (31), suggesting that T-independent IgG-inducing vaccine antigens partially overcome a defect in C3 complement. Reduced but moderate levels of antigen-specific IgG production under the C3-deficient condition were also observed in a different type of antigen, such as live influenza virus and H5 HA VLP vaccine. Immunized C3 KO mice with H5 HA VLP vaccine did not show weight loss after homologous influenza virus infection. Reduced HI titers of 128 to 256 might be sufficient for homologous protection even in the absence of C3. We previously reported a similar finding under the Fc receptor-deficient condition, demonstrating that an inactivated virus vaccine (A/PR/8/34 H1N1) showed an HI titer around 128 and conferred homologous protection independent of Fc receptors in an FcγR KO mouse model (32).
A previous study has demonstrated that C3 deficiency caused inefficient delivery of viral antigens to professional antigen-presenting cells, leading to impaired priming of CD4 helper T and CD8 T cells after acute influenza virus infection (22). Our data demonstrated that C3 deficiency leads to maintaining significantly lower levels of antigen-presenting innate immune cells, including macrophages, an activated form of macrophages, and major subsets of dendritic cells in the peritoneal cavity in naive C3 KO mice than in WT mice. This impairment in maintaining innate immune cells due to C3 genetic deficiency might have caused an inefficient innate microenvironment for effective antigen presentation or delivery of vaccine antigens to antigen-presenting cells, resulting in a defect in effective priming of CD4+ T cells. We also found that the frequency and number of M2e-specific IFN-γ-producing primed CD4+ and CD8+ T cells were reduced in C3 KO mice after vaccination and infection. Moreover, it was reported that impaired T cell priming by C3 deficiency is linked to a defect in pulmonary DC migration from the lungs to the draining lymph nodes (33). A model of complement receptor Cr2 KO mice with influenza virus infection displayed normal T cell priming (22). In another study, treatment of mice with antibodies blocking complement receptor 1 and 2 signaling did not inhibit priming T cells, yet humoral responses were impaired (34). A possible explanation is that blocking the interaction of CD21 and CD19 B cell coreceptors reduces B cell activation but not T cell priming. This study also provides evidence that C3 deficiency resulted in a significant delay in inducing virus-specific IgG antibodies, as evidenced by no detectable levels of IgG antibodies at day 7 postinfection, suggesting a slower kinetics of IgG induction and resulting in severe weight loss after infection compared to weights of WT mice. This suggests that C3 complement is involved in early and rapid induction of IgG antibodies, which appears to be critical for initial control of viral replication and protection. Our observation is consistent with previous studies that C3 KO mice are highly susceptible to viral infection (22, 24). It is possible that low responses to vaccination and high susceptibility of C3 KO mice to viral infection are due to an overall defect in the innate immune system, resulting in ineffective induction of humoral and cellular adaptive immunity.
In summary, this study demonstrates that complement protein C3 is required for effective development of adaptive immune responses to influenza virus M2e and HA VLP vaccination as well as virus infection. In this study, we found multiple roles of C3 in both innate and adaptive immune systems using a C3 KO mouse model. (i) C3 KO mice showed a significant defect in inducing IgG antibodies to immunization with M2e and HA VLP vaccines as well as virus infection. (ii) C3 KO mice displayed impaired induction of effector CD4+ and CD8+ T cells. (iii) C3 is required for conferring protection by active immunization with nonneutralizing M2e-based vaccines but not with neutralizing HA-based vaccines. (iv) C3 KO mice have a defect in retaining innate immune antigen-presenting cells and in recruiting innate immune cells to the site of antigen entry.
MATERIALS AND METHODS
Viruses and vaccines.Reassortant H5N1 (rgH5N1) containing H5 HA with polybasic residues removed from H5N1 A/Indonesia/05/2005 and NA and six internal genes from A/PR/8/1934 and A/Philippines/2/1982 (A/Phil and H3N2) were propagated in embryonated hen's eggs as previously described (13, 35). M2e5x VLP containing tandem repeats of heterologous M2e derived from human (2xM2e), swine (1×), and avian (2xM2e) influenza virus was prepared as detailed in a previous study (11). H5 VLP presenting the H5 subtype of HA protein from A/Indonesia/05/2005 was previously described (36). Briefly, Sf9 insect cells were coinfected with recombinant baculoviruses expressing influenza M1 matrix core protein and M2e5x or H5 HA. M2e5x VLP and H5 VLP vaccines were purified from the cell culture supernatants containing released VLP by sucrose gradient ultracentrifugation and characterized as reported (11, 36).
Immunization and challenge.Wild-type (WT) C57BL/6 and mutant C3 KO mice (B6.129S4-C3tm1Crr/J), used at 6 to 10 weeks old in this study, were obtained from the Jackson Laboratory (Sacramento, CA). Groups of each strain of mouse (n = 5, males and females) were intramuscularly (i.m.) immunized with 10 μg (total proteins) of M2e5x VLP or H5 VLP by a prime-boost regimen at a 3-week interval. At 4 weeks after boost immunization, immunized WT mice were challenged intranasally with a sublethal dose of A/Phil H3N2 (0.8× LD50) or rgH5N1 (0.8× LD50) virus, whereas a 10-fold lower dose of virus was used to infect C3 KO mice in certain experimental sets. Survival rates and body weight changes were monitored daily for 14 days upon infection. All animal experimental procedures in this study were approved by the Georgia State University Institutional Animal Care and Use Committee review boards.
Determination of antibody responses.Influenza virus-specific or M2e-specific antibody levels were determined by enzyme-linked immunosorbent assay (ELISA). Immune sera were serially diluted and then applied to a 96-well plate (Corning Incorporated, Tewksbury, MA) coated with M2e peptide, inactivated A/Indonesia rgH5N1, or inactivated A/Philippines/2/1982 H3N2 virus as previously described (32, 37). IgG and IgG isotype levels were determined by horseradish peroxidase (HRP)-conjugated anti-mouse IgG, IgG1, IgG2a, IgG2b, or IgG2c (SouthernBiotech, Birmingham, AL), with tetramethylbenzidine (eBioscience, San Diego, CA) as a substrate (38). For analysis of long-lived antibody-secreting cell responses, splenocytes and bone marrow (BM) cells were collected from mice at day 7 after virus infection and incubated in culture plates coated with M2e peptide for 1 day. Antibody levels secreted in vitro were measured by ELISA.
Lung virus titers.Lung samples were harvested from the groups of mice at 7 days after challenge. Viral titers were determined as described previously (11). Briefly, lung extracts were serially diluted 10-fold and injected into 10-day-old embryonated chicken eggs. The 50% egg infectious dose (EID50) was calculated by the Reed-Muench method (11).
HI assay.Hemagglutination inhibition (HI) assay was performed as previously described (39). Immune sera were mixed with a receptor-destroying enzyme (Sigma-Aldrich, St. Louis, MO) and then incubated at 37°C. At 16 h after incubation, samples were heat inactivated at 56°C for 30 min. Serially 2-fold-diluted sera were incubated with 8 HA units of A/Indonesia rgH5N1 for 30 min, followed by adding 0.5% chicken red blood cells (Lampire Biological Laboratories, Pipersville, PA) to determine HI titers.
Cytokine ELISPOT assay.To detect gamma interferon (IFN-γ) and interleukin-4 (IL-4) spot-forming cells (SFCs), splenocytes (5 × 105 cells/well) and lung cells (2 × 105 cells/well) were cultured on 96-well enzyme-linked immunospot (ELISPOT) plates coated with anti-mouse IFN-γ (clone R4-6A2) or IL-4 (clone BVD4-1D11) monoclonal antibody (BD Biosciences, San Diego, CA) in the presence of M2e peptide (4 μg/ml). The cytokine spots were developed with biotinylated mouse IFN-γ, IL-4 antibodies, and alkaline phosphatase-labeled streptavidin (BD Pharmingen, San Diego, CA). The spots were visualized with a 3,3′-diaminobenzidine substrate and counted by an ELISPOT reader (BioSys, Miami, FL).
In vivo protection assay of immune sera.To test whether M2e5x VLP immune sera contribute to cross-protection, an in vivo protection assay was performed as described previously (11). Briefly, sera that were heat inactivated at 56°C for 30 min were diluted and mixed with a lethal dose (2× LD50) of A/Philippines/2/1982 (H3N2). Naive BALB/c mice were intranasally infected with a mixture (50 μl) of virus and sera, and the survival rates and body weight changes were monitored daily for 14 days.
Intracellular cytokine staining and flow cytometry assay.For intracellular cytokine analysis, harvested cells were stimulated with M2e peptides and then stained with fluorescence-labeled anti-mouse CD4 and anti-mouse CD8 antibodies. Subsequently, the cells were made permeable by using the Cytofix/Cytoperm kit (BD Biosciences, San Diego, CA), and intracellular cytokines were stained with anti-mouse IFN-γ-allophycocyanin (APC)/Cy7 (clone XMG1.2) (BD Biosciences, San Diego, CA). Stained cells were analyzed using an LSRFortessa (BD Biosciences, San Diego, CA) and FlowJo software (TreeStar).
Germinal center B cell staining and flow cytometry assay.To determine the germinal center B cell phenotype (CD19+ IgD− GL7+ B220+), cells were harvested from mediastinal lymph nodes (MLN) and then stained with anti-mouse B220-APC/Cy7 (clone RA3-6B2), GL7-biotin (clone GL-7), streptavidin-fluorescein isothiocyanate (FITC), IgD-phycoerythrin (PE) (clone 11-26c), and CD19− PE/Cy5.5 (clone eBio1D3) antibodies. All antibodies were purchased from eBioscience. Stained cells were analyzed using a LSRFortessa (BD Biosciences, San Diego, CA) and FlowJo software (TreeStar).
Flow cytometry assay.Peritoneal exudates from naive WT (C57BL/6) and C3 KO mice (n = 5 each) were harvested at 24 h after injection. Isolated peritoneal cells were treated with Fc receptor blocker (anti-CD16/32) and then stained with fluorescence-labeled antibodies, namely, CD11b-APC (clone M1/70), CD11c-PE/Cy7 (clone N418), F4/80-FITC (clone BM8), major histocompatibility complex class II-PE/Cy5 (clone M5/114.15.2), Ly6c-A700 (clone HK1.4), and B220-APC/Cy7 (clone RA3-6B2). All antibodies were purchased from eBioscience. Stained cells were analyzed using an LSRFortessa (BD Biosciences, San Diego, CA) and FlowJo software (TreeStar).
Statistical analysis.All results are expressed as the means ± standard errors of the means (SEM). Significant differences among treatments were evaluated by 2-way analysis of variance (ANOVA). P values of less than or equal to 0.05 were considered statistically significant.
ACKNOWLEDGMENTS
This work was supported by NIH/NIAID grants AI105170 (S.-M.K.), AI119366 (S.-M.K.), and AI093772 (S.-M.K.).
FOOTNOTES
- Received 4 June 2018.
- Accepted 20 July 2018.
- Accepted manuscript posted online 1 August 2018.
Supplemental material for this article may be found at https://doi.org/10.1128/JVI.00969-18.
REFERENCES
- Copyright © 2018 American Society for Microbiology.
