Induction of Broad-Based Immunity and Protective Efficacy by Self-amplifying mRNA Vaccines Encoding Influenza Virus Hemagglutinin

Abstract

Seasonal influenza is a vaccine-preventable disease that remains a major health problem worldwide, especially in immunocompromised populations. The impact of influenza disease is even greater when strains drift, and influenza pandemics can result when animal-derived influenza virus strains combine with seasonal strains. In this study, we used the SAM technology and characterized the immunogenicity and efficacy of a self-amplifying mRNA expressing influenza virus hemagglutinin (HA) antigen [SAM(HA)] formulated with a novel oil-in-water cationic nanoemulsion. We demonstrated that SAM(HA) was immunogenic in ferrets and facilitated containment of viral replication in the upper respiratory tract of influenza virus-infected animals. In mice, SAM(HA) induced potent functional neutralizing antibody and cellular immune responses, characterized by HA-specific CD4 T helper 1 and CD8 cytotoxic T cells. Furthermore, mice immunized with SAM(HA) derived from the influenza A virus A/California/7/2009 (H1N1) strain (Cal) were protected from a lethal challenge with the heterologous mouse-adapted A/PR/8/1934 (H1N1) virus strain (PR8). Sera derived from SAM(H1-Cal)-immunized animals were not cross-reactive with the PR8 virus, whereas cross-reactivity was observed for HA-specific CD4 and CD8 T cells. Finally, depletion of T cells demonstrated that T-cell responses were essential in mediating heterologous protection. If the SAM vaccine platform proves safe, well tolerated, and effective in humans, the fully synthetic SAM vaccine technology could provide a rapid response platform to control pandemic influenza.

Importance: In this study, we describe protective immune responses in mice and ferrets after vaccination with a novel HA-based influenza vaccine. This novel type of vaccine elicits both humoral and cellular immune responses. Although vaccine-specific antibodies are the key players in mediating protection from homologous influenza virus infections, vaccine-specific T cells contribute to the control of heterologous infections. The rapid production capacity and the synthetic origin of the vaccine antigen make the SAM platform particularly exploitable in case of influenza pandemic.

FIG 1

Immunogenicity and efficacy of SAM(H1-Cal)/CNE in ferrets. Ferrets (n = 6) were immunized i.m. on day 0 and day 56 with 15 or 45 μg SAM(H1-Cal)/CNE, 15 μg of MIIV, 15 μg of MIIV+MF59, or CNE alone (CTRL). Sera were collected on day 28 (post1) and day 80 (post2) and analyzed for HI (A) and VN (B) titers. QL, quantification limit; DL: detection limit. *, P < 0.05; **, P < 0.01 (compared to MIIV). (C) On day 94, ferrets were infected i.n. with 10⁶ PFU of influenza Cal virus per animal. Five days after challenge, nasal washes were collected, and virus titers were determined as focus-forming units (FFU)/ml. *, P < 0.05; **, P < 0.01 (compared to CTRL). (D) Onset of illness as indicated by a loss of 10% of weight relative to the weight on the day of challenge. *, P < 0.05 (compared to CTRL).

FIG 2

Immunogenicity of SAM(H1-Cal) in mice. Mice (n = 12) were immunized i.m. on day 0 and day 56 with SAM(GFP) at 10 μg, SAM(H1-Cal) at 0.1, 1, or 10 μg, MIIV at 1 μg, or MIIV+MF59 at 1 μg. Sera and spleens were collected 2 weeks after the second immunization. Sera were analyzed for H1N1-specific total IgG titers (A), HI titers (B), and IgG2a/IgG1 ratios (C). *, P < 0.05; **, P < 0.01 (compared to MIIV). (D to F) Splenocytes (n = 4) were stimulated in vitro with a H1-Cal peptide pool, and T cells were analyzed for cytokine production by flow cytometry (see Fig. S3A and B in the supplemental material for the gating strategy). The bars represent the cumulative frequency of H1-specific CD4 T cells (D) and CD8 T cells (E) expressing combinations of cytokines, as indicated in the graph. (F) CD107a expression by CD8 T cells. (G to H) In vivo cytotoxicity. CFSE-labeled H1_533–541 or HIV-Gag_107–205 -pulsed target cells were administered i.v. to mice previously immunized with SAM(GFP) or SAM(H1-Cal) or exposed to a sublethal dose of PR8 virus. Splenocytes were harvested 20 h later and analyzed for the presence of CFSE⁺ target cells by flow cytometry, as described in Materials and Methods. (G) Representative histograms showing the percentage of CFSE_high and CFSE_low target cells recovered. (H) Percentage of specific target cell lysis. **, P < 0.01; ***, P < 0.001 [compared to SAM(GFP)]. The data shown are merged data from three independent experiments.

FIG 3

SAM(H1) protects mice from lethal infection. Mice (n = 8) were immunized i.m. on day 0 and day 56. Four weeks after the second immunization, mice were challenged with 20 TCID₅₀ of PR8 virus. Survival (A, B, and C) and body weight (D, E, and F) were monitored for 11 days postinfection. Immunizations were as follows: 10 μg of SAM(GFP) and SAM(HA-PR8) at the indicated doses (A and D), 10 μg of SAM(GFP) and SAM(HA-Cal) at the indicated doses (B and E), and 10 μg of SAM(GFP) or SAM(H1-Cal) or 1 μg of MIIV or MIIV+MF59 (C and F). The percent survival was determined based on humane endpoint criteria. Statistical analysis was performed using Mantel-Cox test. *, P < 0.05; **, P < 0.01; ***, P < 0.001 [compared to SAM(GFP) (A and B) or MIIV (C)]. The data shown are merged data from three independent experiments.

FIG 4

Immune cells are recruited in the lungs after influenza virus challenge. Mice were immunized i.m. on day 0 and day 56 with saline or 10 μg of SAM(H1-Cal) or were exposed to a sublethal dose of PR8 virus. Four weeks after the second immunization, mice were infected with 20 TCID₅₀ of PR8 virus. Lungs (n = 3) were collected at different time points after infection as indicated. (A) Viral loads expressed as the fold change compared to noninfected lungs. *, P < 0.05; **, P < 0.01 (compared to saline). (B) Total cell recruited in the lung. (C) Immune cell recruitment following the gating strategy described in Fig. S5A in the supplemental material. The data shown are merged data from two independent experiments.

FIG 5

T-cell responses in lungs of infected mice. Mice were immunized i.m. on day 0 and day 56 with saline or 10 μg of SAM(H1-Cal) or exposed to a sublethal dose of PR8 virus. Four weeks after the second immunization, mice were infected with 20 TCID₅₀ of PR8 virus. Lungs (n = 3) were collected at different time points after infection and characterized for CD4 (A) and CD8 (D) T-cell recruitment. The frequencies of cytokine-producing (B and E) and CD107a⁺ (C and F) T cells after in vitro stimulation with H1-Cal peptide pool (p) or medium (m) were also determined. (G) H1_533–541-specific CD8 T cells were identified ex vivo with PE-labeled H-2K^d/HA_533–541 pentamer and are characterized as T effectors (Teff, CD44^hi CD62L^low CD127^low), T effector memory (TEM, CD44^hi CD62L^low CD127^high), and T central memory (TCM, CD44^hi CD62L^high CD127^high). The data shown are merged data from two independent experiments.

FIG 6

Role of T cells in mediating protection against lethal heterologous influenza virus challenge. Mice (n = 8) were immunized i.m. on day 0 and day 56 with 10 μg of SAM(H1-PR8) or SAM(H1-Cal). (A) Serum neutralization titers, corresponding to a 50% inhibition of infection (IC₅₀) by PR8 or Cal influenza viruses. The data show the means ± the standard deviations and are representative of four independent experiments. (B) Survival of mice after passive transfer of sera from saline-, SAM(H1-Cal)-, or SAM(H1-PR8)-immunized mice or mice exposed to a sublethal dose of PR8 virus prior to challenge with 20 TCID₅₀ of PR8 virus. **, P < 0.01 (compared to saline). (C and D) Frequency of H1-Cal/H1-PR8 cross-reactive cytokine⁺ CD4 (C) and CD8 (D) T cells in splenocytes from SAM(H1-Cal)-immunized mice. The data show means ± the standard deviations and are cumulative of four independent experiments. *, P < 0.05; **, P < 0.01 (compared to medium). (E and F) Survival of SAM(H1-Cal)-immunized mice after depletion of CD4 and/or CD8 T cells and challenged with PR8 virus at 20 TCID₅₀ (E) or 100 TCID₅₀ (F). The percent survival was determined based on humane endpoint criteria. The data are cumulative for two independent experiments (n = 16). *, P < 0.05; ***, P < 0.001 (compared to treatment with an isotype control).