Trimerized S expressed by modified vaccinia virus Ankara (MVA) confers superior protection against lethal SARS-CoV-2 challenge in mice

ABSTRACT

The reoccurrence of successive waves of SARS-CoV-2 variants suggests the exploration of more vaccine alternatives is imperative. Modified vaccinia virus Ankara (MVA) is a virus vector exhibiting excellent safety as well as efficacy for vaccine development. Here, a series of recombinant MVAs (rMVAs) expressing monomerized or trimerized S proteins from different SARS-CoV-2 variants are engineered. Trimerized S expressed from rMVAs is found predominantly as trimers on the surface of infected cells. Remarkably, immunization of mice with rMVAs demonstrates that S expressed in trimer elicits higher levels of binding IgG and IgA, as well as neutralizing antibodies for matched and mismatched S proteins than S in the monomer. In addition, trimerized S expressed by rMVA induces enhanced cytotoxic T-cell responses than S in the monomer. Importantly, the rMVA vaccines expressing trimerized S exhibit superior protection against a lethal SARS-CoV-2 challenge as the immunized animals all survive without displaying any pathological conditions. This study suggests that opting for trimerized S may represent a more effective approach and highlights that the MVA platform serves as an ideal foundation to continuously advance SARS-CoV-2 vaccine development.

IMPORTANCE

MVA is a promising vaccine vector and has been approved as a vaccine for smallpox and mpox. Our analyses suggested that recombinant MVA expressing S in trimer (rMVA-ST) elicited robust cellular and humoral immunity and was more effective than MVA-S-monomer. Importantly, the rMVA-ST vaccine was able to stimulate decent cross-reactive neutralization against pseudoviruses packaged using S from different sublineages, including Wuhan, Delta, and Omicron. Remarkably, mice immunized with rMVA-ST were completely protected from a lethal challenge of SARS-CoV-2 without displaying any pathological conditions. Our results demonstrated that an MVA vectored vaccine expressing trimerized S is a promising vaccine candidate for SARS-CoV-2 and the strategy might be adapted for future vaccine development for coronaviruses.

INTRODUCTION

Numerous vaccines against SARS-CoV-2 have been developed using various platforms, including inactivated whole virus vaccines, subunit vaccines, nanoparticle vaccines, viral vector vaccines, mRNA vaccines, etc. (1–6). Multiple vaccines, including inactivated virus, mRNA, and adenovirus-vectored vaccines, are acknowledged for their capability to reduce the severity of disease, although they failed to effectively prevent infection and transmission (4, 6, 7). Modified vaccinia virus Ankara (MVA), a replication-deficient virus derived from its parental chorioallantois vaccinia virus Ankara (CVA) by serial passage in chick embryo fibroblast (CEF) cells, has been approved by the FDA as a vaccine for smallpox and mpox (8–10). MVA is a promising vaccine vector as it can efficiently express foreign antigens and displays intrinsic adjuvant activity (11–16). As MVA does not replicate in the vast majority of mammalian cells due to the loss of two crucial host range genes C16 and C12 (17, 18), it is considered an extremely safe vector and can therefore be used for immune-deficient individuals (19). SARS-CoV-2 vaccines with different unique designs using MVA as a carrier have reportedly shown good protection in multiple animal models (5, 20–26). New attempts are constantly being applied to better enhance the immune effect of MVA-vectored vaccines.

The Spike (S) protein of SARS-CoV-2 mainly mediates virus fusion and entry into the host cells and is the predominant antibody-eliciting antigen (27). The S protein exists in its natural conformation as a trimer on the surface of the virus (28). During the fusion process, S depolymerizes, changes its conformation, and is cleaved into S1 and S2 by host protease, such as TMPRSS2 (29). To enhance the immunogenicity of the S protein during vaccine development, a series of modifications have been employed, including the mutation in the furin recognition site and the two proline substitutions (27, 30).

The human C-propeptide of α1(I) collagen (Trimer-Tag) has been reported to spontaneously form trimers (31, 32). When the Trimer-Tag is fused to the S protein from SARS-CoV-2, the latter could polymerize and form trimers, which can stimulate the host immune responses more efficiently in nonhuman primates (33, 34). Here, we fused the Trimer-Tag with S (S-Trimer, ST) from various SARS-CoV-2 variants and inserted the chimeric cassettes into MVA to generate a series of recombinant MVA vaccine candidates (rMVAs) that express S in trimer or monomer. Immunization and challenge experiments showed that the rMVAs induced robust humoral, mucosal, and cellular immune responses in BABL/c mice. Importantly, immunization with the rMVA-ST offered protection against the lethal challenge of SARS-CoV-2 mouse-adapted strain SARS-CoV-2/BMA8. The rMVAs expressing S in trimers confer superior immunity and protection than those expressing S in monomers. Our results demonstrated that rMVA-ST is a highly promising vaccine candidate for SARS-CoV-2 and holds potential for future vaccine development strategies in preventing diseases caused by coronaviruses.

RESULTS

Construction of rMVAs that express modified SARS-CoV-2 S proteins

The full-length S protein from SARS-CoV-2 contains 1,273 amino acids (aa), including the S1 receptor-binding subunit (aa 14–685) and the S2 membrane fusion subunit (aa 686–1,273). The S protein is cleaved into two parts, S1 and S2, during virus internalization. Studies have shown that maintaining the pre-fusion conformation of S facilitates the stimulation of immune responses (35, 36). Therefore, amino acids 682–685 of the S protein were mutated (RRAR-GSAS) to reduce cleavage (27), and the K and V at positions 986 and 987 were mutated to two prolines to maintain the protein at the prefusion conformation (28, 30) (Fig. 1). In addition, as the S protein of SARS-CoV-2 exists in homotrimer on the virus surface, we aimed to maintain this native conformation by fusing Trimer-Tag to the C terminus of the complete sequence of S. To facilitate virus screening, an enhanced green fluorescent protein (eGFP) cassette directed by a poxvirus P11 promoter was added to the 5′ end of the S ORF.

Fig 1.

Construction strategy for rMVA-S and rMVA-ST. The strategy used to engineer rMVA-S and rMVA-ST is illustrated. A cassette comprising a P11-directed eGFP and an mH5-directed S was inserted between MVA067R and MVA070F. Modifications made to S are indicated. Abbreviations: WT, Wild-type; DE, Delta; OM, Omicron; TM, transmembrane domain; CT, C-terminal domain; 3 × Flag, three tandem copies of Flag epitope tag; Trimer-Tag, human C-propeptide of α1(I) collage.

The complete cassette comprising an eGFP and an S sequence was inserted into the MVA genome through homologous recombination between the intergenic site of MVA069R and MVA070F. To investigate cross-protective effects, we also engineered rMVAs expressing S proteins from three representative strains of SARS-CoV-2 (wild type, WT; Delta, DE; Omicron, OM) from different SARS-CoV-2 sublineages. The strategy of rMVA construction is outlined in Fig. 1. All nucleotide sequences inserted into MVA are shown in Table S1.

S fused with a Trimer-Tag (ST) is expressed in a trimeric form on the surface of rMVA-infected mammalian cells

To examine whether the S inserted into MVA could be expressed correctly in human cells, HeLa cells were infected with cushion-purified rMVAs expressing S or ST (names of the rMVAs are indicated in Fig. 2) at three plaque-forming units (PFU) per cell, and total protein was harvested and resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and subsequently subjected to Western blotting analysis. A monoclonal antibody targeting the RBD region was used to detect S, and the expression of S full-length protein with a size of ~180 kDa was observed. No difference in the levels of S between rMVA-S and rMVA-ST was observed, although the S protein with a Trimer-Tag migrated slightly slower than the S with 3 × Flag due to the difference in molecular weight. As expected, no cleavage of S was observed due to the removal of the furin site. VACV-specific anti-serum was also used to detect viral protein synthesis (Fig. 2A). We next assessed whether S with a Trimer-Tag was able to form trimers using a non-reducing native-PAGE. As shown in Fig. 2B, two bands were observed on the gel, including a more evident S-trimer band located at approximately 500 kDa and a much lighter S-monomer band located at~180 kDa. In addition to the VACV-specific anti-serum described above, anti-A3 and I3 antibodies were also used to detect viral late and early protein synthesis, respectively, to ensure comparable input dosages of the viruses. These data illustrated that ST existed predominantly as trimers when fused with a Trimer-Tag. By contrast, S migrated primarily as monomers, and only a small portion displayed as trimers without the use of Trimer-Tag. These data demonstrated that the addition of the trimer tag greatly promoted the trimerization of MVA-expressed S protein (Fig. 2B).

Fig 2.

Viral expression and subcellular localization of modified S proteins. (A) HeLa cells were mock-infected or infected with 3 PFU/cell of purified parental MVA or rMVAs for 24 hours, and lysates were analyzed by SDS-PAGE and Western blotting analysis using S-RBD and β-actin antibodies, and VACV-positive serum. (B) HeLa cells were mock infected or infected with 3 PFU/cell of purified parental MVA or rMVAs for 24 hours, and lysates were analyzed under non-reducing conditions by native-PAGE and western blotting analysis using S-RBD, HSP90, A3, I3, and β-actin antibodies, and VACV-specific anti-serum. S-Trimer and S-Monomer are marked in the figure. (C) HeLa cells were mock infected or infected with 3 PFU/cell of purified parental MVA or rMVAs for 12 hours. Cells were fixed, permeabilized, or left untreated, and then blocked, and stained with primary antibodies to S-RBD followed by fluorescent conjugated secondary antibodies. Hoechst was used to stain DNA. All data are representative of 3 independent experiments.

Next, the subcellular localization of S in rMVA-infected cells was monitored using confocal microscopy. Human HeLa cells were infected with rMVAs indicated in the figure at 3 PFU/cell for 12 hours, and the cells were fixed, permeabilized or non-permeabilized, and incubated with monoclonal anti-RBD antibodies to observe the subcellular localization of S protein using confocal microscopy. It was observed that in permeabilized cells, S was detected predominantly on the plasma membrane of infected cells but could also be observed in the cytoplasm. However, in non-permeabilized cells, the S was found exclusively on the plasma membrane of infected cells (Fig. 2C). In general, the above results demonstrated that S fused with a Trimer-Tag was expressed primarily in trimers on the surface of infected cells.

The insertion of S did not alter the biological characteristics of MVA

To detect whether the insertion of S affected the proliferation of rMVAs on permissive cells, DF-1 and BHK (baby hamster Syrian kidney) cells were infected with rMVAs at 0.01 PFU/cell for 2 hours, and viruses were harvested at indicated time points for quantification and viral growth curves were generated (Fig. 3A and B). It was apparent that the insertion of the S protein did not impact the replication kinetics of rMVAs on permissive cell lines.

Fig 3.

Biological characteristics of rMVAs. DF-1 cells (A) or BHK cells (B) were infected with 0.01 PFU/cell of purified parental MVA or rMVAs and viruses were harvested at 24, 48, and 72 hours post-infection and quantified by plaque assay on DF-1 cells. (C) HeLa cells, A549 cells, or MDCK cells were infected with 0.01 PFU/cell of purified parental MVA, rMVAs, or VACV-WR, and viruses were harvested and quantified by plaque assay on DF-1 cells for MVA and BS-C-1 cells for VACV-WR. (D) HeLa cells were infected with 3 PFU/cell of purified parental MVA or rMVAs for 12 hours. Viral particle morphology was observed by TEM. (E and F) DF-1 cells were infected with 0.1 PFU/cell (E) or 0.01 PFU/cell (F) of purified parent MVA, rMVAs, and passaged rMVAs, and viruses were harvested at 4, 8, and 12 (E) or 24, 48, and 72 (F) hours post-infection and quantified by plaque assay on DF-1 cells. (G) HeLa cells were infected with 0.01 PFU/cell of purified parental MVA, rMVAs, passaged rMVAs, or VACV-WR, and viruses were harvested and quantified by plaque assay on DF-1 cells for MVA and BS-C-1 cells for VACV-WR. (H) The purified parental MVA or rMVAs were diluted to 10⁸ PFU/mL in DMEM and placed at room temperature (25°C). Virus titers were measured every day for 14 days. (I and J) Nude BALB/c mice (female, 6 weeks old) were infected IN with 2 × 10⁷ PFU of unmodified MVA or rMVAs. Body weights (I) and survival (J) of mice were determined daily. (A–C, E–H) Statistical analyses were performed to compare all virus titers to that of MVA using the Student’s t-test. ns, non-significant. All data are representative of 3 independent experiments.

MVA is considered a highly safe vaccine vector as it is replication-deficient in almost all human cells and the vast majority of mammalian cells (14). To ensure the addition of ectopic sequences did not alter the host range of MVA, several mammalian cell lines [HeLa cells, A549 cells, and MDCK (Madin-Darby canine kidney) cells] were infected with rMVAs at 0.01 PFU/cell for 2 hours and total viruses were harvested at 48 hours for quantification by plaque assay (Fig. 3C). VACV-WR (Western Reserve) was included as a positive control. The results demonstrated that VACV-WR was able to replicate in all cells tested, while both MVA and rMVAs expressing various S failed to replicate in these cell lines. The titers observed in rMVAs-infected cells were comparable to those in MVA-infected cells.

The replication blockage of MVA occurred at a late stage of virus replication, most likely during viral morphogenesis (14, 37, 38). Transmission electron microscopy (TEM) was used to observe virion morphology in HeLa cells infected with parental MVA, rMVAs, or VACV-WR as a positive control (Fig. 3D). Infected HeLa cells were fixed and subjected to TEM analysis at 12 hpi. A large number of mature virus particles (MV) were observed in VACV-WR-infected cells, but not in MVA or rMVA-infected cells, where only dense particles and immature viruses were observed. Together, the above results suggested the addition of S proteins did not alter the biological properties in cells and the safety profile of MVA.

To test the genetic stability of rMVAs, 50 consecutive passages of rMVAs were performed on DF-1 cells and the genetic stability of S was determined by Sanger sequencing (Table S2). The rMVAs of the 25th and 50th generations were aligned with the parental virus sequence using SnapGene (Tables S3 to S8). The results suggested that the nucleotide sequences of S for all the 6 rMVAs did not undergo any mutations after 50 consecutive passages. In addition, the growth characteristics of rMVAs and rMVAs after 50 passages were determined, and no significant differences were detected between the passaged viruses and the parental virus or MVA (Fig. 3E and F). No significant changes in the growth characteristics of the virus after passage were found on non-permissive HeLa cells (Fig. 3G). The above results indicated that rMVAs exhibited decent genetic stability. The thermal stability of rMVAs was examined by incubating the purified viruses at room temperature for the durations indicated in the figure prior to virus quantification by plaque assay (Fig. 3H). The results demonstrated that the virus titers decreased by approximately 10 × folds after 7 days and 100 × folds after 14 days, indicating the insertion of S did not alter the thermostability of MVA.

In addition to the in vitro assay described, in vivo evaluation of the safety of the rMVA vaccines was performed using nude BABL/c mice, which are immunodeficient (Fig. 3I). The animals were intranasally inoculated with VACV-WR, MVA, and rMVAs at 2 × times the normal immunization dose (2 × 10⁷ PFU). The body weights were recorded and shown to decrease slightly (~90%), but no mortality was observed throughout the entire experiment (Fig. 3J). All mice infected with VACV-WR as a positive control died within 6 days post-infection.

rMVAs elicit robust humoral and mucosal immune responses in mice

For proper evaluation of the immunogenicity of rMVAs, a lethal animal model in which a SARS-CoV-2 mouse-adapted (SARS-CoV-2/BMA8) virus strain, which was obtained by serial passage of a clinical isolate of SARS-CoV-2 (BetaCov/Wuhan/AMMS01/2020) in aged BALB/c mice for eight times, was employed. The SARS-CoV-2/BMA8 strain was able to infect aged BALB/c mice and resulted in severe respiratory symptoms, rapid weight loss, and eventually death (39, 40). To explore the immunogenicity of rMVA-S and rMVA-ST in this model, 9-month-old BALB/c mice, the same types of animals used for virus challenge, were intranasally administered with 10⁷ PFU of rMVAs for the prime immunization and then boosted 14 days later with the same dosage (Fig. 4A). Mouse serum was collected 14 days after the prime immunization (Day 14) and 14 days after the boost immunization (Day 28). Purified recombinant SARS-CoV-2 S proteins from three different sublineages (WT, DE, or OM) were used for an enzyme-linked immunosorbent assay (ELISA) to measure binding IgG for S proteins (Fig. 4B). Immunization with rMVA-S or rMVA-ST was able to induce high amounts of binding IgG for S proteins from the corresponding virus strains, and the boost immunization enhanced the antibody titers by up to 128 × folds. Notably, the addition of Trimer-Tag further increased the antibody titers by 64 × folds compared to those elicited by monomerized S, suggesting S in a trimerized conformation was superior to those in a monomerized conformation in the induction of IgG responses.

Fig 4.

Induction of binding and neutralizing antibody in response to MVA-vectored SARS-CoV vaccines. (A) Illustration of the design for the animal experiment. Mice were bled on days 0, 14, and 28 to determine S binding and SARS-CoV-2 neutralizing titers. Vaccinated mice 2 weeks after boosting were infected IN with 50 LD₅₀ of SARS-CoV-2 mouse-adapted strain SARS-CoV-2/BMA8. Mice were followed for weight loss and signs of morbidity. Three mice from each group were euthanized on day 3 (*) after a challenge to determine the infectious SARS-CoV-2 virus and subgenomic mRNA. (B–D) Mice were bled before vaccination, at 2 weeks after the prime, and at 2 weeks after the boost. Levels of S-binding IgG (B) or IgA (C) were determined by ELISA. The dotted lines indicate the limit of detection. Pseudovirus neutralizing titers are shown in (D) and plotted as IC₅₀. X1, sera collected after prime immunization; X2, sera collected after boost immunization. (E) Levels of IgG binding to different sublineages of S measured in serum 2 weeks after boost by ELISA. (F) Measurements of neutralizing antibodies of mouse serum against pseudoviruses expressing different sublineages S proteins 2 weeks after the boost immunization. (G–H) IgG2a or IgG1 binding to S was determined by ELISA. (B–H) Statistical analyses were performed using the Student’s t-test from five biological replicates. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001; ns, non-significant. Data in (B–H) are representative of 3 independent experiments.

As an important marker of mucosal immunity, IgA is essential for viral clearance and constraining of the infection in the respiratory system. The levels of S-binding IgA in the serum samples described above were titered by ELISA (Fig. 4C). The overall trend was similar to the result of IgG with the IgA titers being significantly higher in rMVA-ST-immunized mice than rMVA-S-immunized ones. These results demonstrated that S expressed in trimer was able to induce stronger mucosal immune responses than S in the monomer.

Next, the presence of neutralizing antibodies was assessed using a pseudovirus assay reported previously (41). Three different types of SARS-CoV-2 pseudoviruses were manufactured by incorporating S from three different sublineages described above for the measurement of neutralizing antibodies (Fig. 4D). The results exhibited that the neutralizing antibody titers induced by rMVA-ST after the prime immunization were higher than those induced by rMVA-S regardless of the strain, although the differences were not statistically significant. Remarkably, the differences became statistically significant after the boost immunization and the antibody titers from rMVA-ST were all above 10³ IC₅₀. Importantly, the boost immunization significantly enhanced the levels of neutralizing antibodies by up to 16 × fold, suggesting the importance of the booster in conferring successful virus neutralization. The above results indicated that rMVA-ST could elicit higher levels of binding IgG and IgA as well as neutralizing antibodies than rMVA-S after one or two immunizations.

As SARS-CoV-2 evolved rapidly within the human population, we next sought to inquire if rMVA-S and rMVA-ST we engineered were able to provide cross-reactive protection against viruses across different strains by measuring binding IgG and neutralizing antibody levels. The ELISA described in Fig. 2B was performed to measure binding IgGs collected from mice immunized with different rMVA for S proteins from SARS-CoV-2-WT, SARS-CoV-2-DE, or SARS-CoV-2-OM (Fig. 4E). As expected, IgG titers were highest in the animals in which the same S protein sublineage was used for immunization and ELISA quantification. Importantly, binding IgG collected from all animals except for those immunized with MVA alone was able to bind to S proteins from all three sublineages, although with a slightly lower titer in the circumstances of cross-reactive protection (4 to 64 × folds lower). In addition, pseudoviruses expressing S from SARS-CoV-2-WT, SARS-CoV-2-DE, or SARS-CoV-2-OM were used to determine neutralizing antibody titers in the serum collected from boosted mice (Fig. 4F). As expected, mouse serum showed robust neutralizing activity against pseudoviruses with S from different lineages. Similar to the results from binding IgG, the neutralizing antibody titers were higher when the S used for immunization matched with that used for pseudovirus packaging. These results demonstrated that rMVA-S and rMVA-ST might be able to provide cross-reactive protection for different SARS-CoV-2 strains, although to a lesser extent. However, future virus challenge experiments using different virus strains were required to confirm this hypothesis.

The ratio of IgG2a to IgG1 in the serum after immunization reflects the tendency of the immune responses. A ratio greater than 1 indicates the immune response is more inclined to the Th1 type, and a ratio less than 1 indicates a Th2-inclined response. ELISA was used to determine the titers of IgG2a and IgG1 in the serum of immunized animals (Fig. 4G). The results showed that the levels of IgG2a in different immunization groups were on average 3 × to 5 × folds higher than that of IgG1 (Fig. 4H), indicating a Th1-type immune response.

rMVAs stimulate the generation of S-specific CD3+CD8+IFNγ+ T cells

To evaluate cellular immune responses after rMVA-S and rMVA-ST immunization, spleen cells from the animals immunized with the same protocol described in Fig. 4 were collected after boost immunization and were stimulated ex vivo with purified S (final concentration 10 µg/mL) for 6 hours. Next, S-stimulated splenocytes were incubated with antibodies against murine CD3, CD4, CD8, or IFNγ prior to flow cytometry analysis for cytotoxic T cells (Fig. 5A). The results demonstrated both rMVA-S and rMVA-ST were able to induce CD3 +CD8+IFNγ+ T cells for SARS-CoV-2 S proteins, with the spleen cells in the rMVA-ST -immunized animals containing much higher levels of CD3 +CD8+IFNγ+ T cells than those immunized with rMVA-S (Fig. 5B). In addition, CD3 + CD4+ IFNγ+ specific T cells were also tested, although no positive results were detected.

Fig 5.

CD8 +T-cell response to MVA-vectored SARS-CoV2 vaccines. BALB/c mice were vaccinated IN with 10⁷ PFU of unmodified MVA or rMVA for prime immunization and again after 2 weeks for boost immunization. At 2 weeks after the boost, spleen cells pooled from five mice were stimulated with the corresponding S protein. (A and B) Cells were then stained for cell surface markers with mouse anti-CD3-FITC, anti-CD4-APC, and anti-CD8-PerCP. Cells were subsequently stained intracellularly with mouse anti-IFNγ-PE. CD3+ CD8+ IFNγ+ cells were enumerated by flow cytometry. (C and D) Cells were incubated with IFN-γ monoclonal antibody-coated plates for 24 hours. After lysing the cells, biotin-labeled IFN-γ monoclonal antibody and HRP-labeled biotin antibody were added successively for incubation. The number of spots in each well was recorded. (B and D) Statistical analyses were performed using the Student’s t-test from three biological replicates. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. All data are representative of 3 independent experiments.

The amounts of T cells secreting IFN-γ after S protein stimulation were counted using an enzyme-linked immunospot assay (ELISpot, Fig. 5C). The results were consistent with that of the flow cytometry assay. The number of splenocytes secreting IFN-γ after stimulation with S protein in the rMVA-ST group was significantly higher than that in the rMVA-S group (Fig. 5D). As a negative control, the spleens of the MVA-immunized control group did not contain any splenocytes secreting IFN-γ in response to S stimulation. These results exhibited that rMVA-ST induced stronger cellular immune responses than rMVA-S in mice.

Protection of a lethal SARS-CoV-2 challenge with rMVA-S and rMVA-ST

To determine whether the rMVAs engineered were able to protect animals from a lethal SARS-CoV-2 challenge, mice primed and boosted with the dosage described above were challenged with 50 LD₅₀ SARS-CoV-2 mouse-adapted strain SARS-CoV-2/BMA8 (BetaCov/Wuhan/AMMS01/2020). PBS was used as a negative control and MVA was included to rule out non-specific protection caused by innate immune responses to MVA vectors. All animals were observed daily and survival rates as well as body weights were recorded. The lungs and turbinates were harvested from sacrificed animals for virus quantification on day 3. Our analysis showed that animals immunized with MVA exhibited dramatic weight loss starting on the second day after the challenge. Sixty percent of animals died on day 5 and all animals died on day 6 (Fig. 6A and B). In comparison, mice immunized with rMVA-S and rMVA-ST all survived the challenge, although body weight loss was recorded between day 2 and day 5 for animals immunized with rMVA-S, and began to rebound on day 6. Remarkably, animals immunized with rMVA-ST all survived the challenge without displaying any noticeable weight loss.

Fig 6.

Challenge of mice with SARS-CoV-2 following prime and boost vaccinations. (A–D) Vaccinated mice were infected IN with 50 LD₅₀ of SARS-CoV-2 mouse-adapted strain SARS-CoV-2/BMA8. Mice were observed daily for weight loss, signs of morbidity, and survival rate. Three mice from each group were euthanized on day 3 after a challenge to determine viral loads and subgenomic mRNA. Body weights (A) and survival (B) of mice were determined daily. (C) Virus titers in lung homogenates or nasal turbinate obtained on day 3 were determined by endpoint dilution and plotted as TCID₅₀ per gram of tissue. (D) RNA was isolated from lung homogenates or nasal turbinate on day 3. SARS-CoV-2 N subgenomic RNAs were determined by qRT-PCR and plotted as copies per gram of tissue. (E) Lung tissues from mice dissected at 3 dpi were analyzed by hematoxylin and eosin staining. Representative images from each group are displayed. Data in (A–E) are representative of 3 independent experiments. (A, C, and D) Statistical analyses were performed using the student’s t-test from three biological replicates. **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. All data are representative of 3 independent experiments.

Three mice from each group were euthanized on day 3 after the SARS-CoV-2 challenge, viruses were harvested from lung homogenates and turbinates, and the titers were determined and shown in TCID₅₀ (Fig. 6C). The results showed that the virus titers in the lungs and turbinates of mice immunized with rMVAs were significantly lower than those in the MVA-immunized animals. Among them, virus titers in the lungs of mice in the rMVA-ST group were significantly lower than that in the rMVA-S group, which was consistent with the clinical manifestations of the mice. Moreover, the copy number of the SARS-CoV-2 N gene was also analyzed by real-time fluorescence quantitative PCR (RT-qPCR) using specific primers for the N gene (Fig. 6D). Consistent with the results from virus titers, the mRNA levels of N protein in the lungs of mice in the rMVA-ST group were also significantly lower than those in the rMVA-S group. Moreover, hematoxylin and eosin staining were used to detect pathological changes in lung tissues from mice dissected at 3 dpi. The lungs of mice in the rMVA-ST group were almost completely free of lesions, while the lungs of mice in the rMVA-S group showed slight inflammatory infiltration after challenge (Fig. 6E). These data suggested that SARS-CoV-2 exhibited a transient infection in mice immunized with rMVA-ST and was not able to replicate effectively and cause observable clinical symptoms. Overall, our results demonstrated that rMVA-ST completely protected aged BALB/c mice against a lethal challenge of SARS-CoV-2.

DISCUSSION

SARS-CoV-2 has caused a significant loss worldwide since its original outbreak in 2019. Although WHO no longer declares SARS-CoV-2 as a pandemic disease, the virus continues to circulate among the human population despite massive vaccination efforts. Various studies have been published using MVA as a vector to develop the SARS-CoV-2 vaccine although the construction strategy and virus used for challenging were different (5, 20–24). In this study, we generated a candidate SARS-CoV-2 vaccine using MVA as a vaccine vector to express trimerized-S and evaluated its biological property, immunogenicity, and protective effect against a lethal SARS-CoV-2 challenge in a mice model. Our study is different from the others using MVA in two main aspects. First, we used a trimerized S in addition to a monomerized S for construction. Second, the virus challenge model was different and the virus strain was based on the original Wuhan strain and was obtained after continuous passage in mice. To ensure stable and robust expression of S using MVA, a series of modifications were first made on the sequences of S proteins, including mutation of the furin cleavage site (682–685 aa, RRAR to GSAS)(27), mutation of KV 986–987 to PP to maintain pre-fusion conformation (28, 30, 35, 36). To simulate the natural conformation of S protein, we fused S with a Trimer-Tag to express the S protein in a trimeric form. We were able to observe sufficient and robust expression of S from rMVA-S/ST in HeLa cells although the cells are non-permissive to MVA. More importantly, in cells infected with rMVA-ST, S is detected as a high-molecular-weight complex with the protein weight corresponding to S-Trimer, which is consistent with existing reports (33). The S proteins expressed from rMVA-S and rMVA-ST were found predominantly on the plasma membrane of infected cells, which may increase its immunogenicity by facilitating antigen presentation.

To evaluate the safety and stability and ensure the insertion of S and S-Trimer did not alter the biological properties of the vaccine candidates, we first assessed their replicative capability in avian and mammalian cells. Similar to their parental strain MVA, rMVAs could proliferate in DF-1 and CEF cells. However, in all mammalian cells tested, rMVAs could not replicate and failed to form mature virus particles. Therefore, the biological properties of rMVA-S and rMVA-ST did not change in comparison to their parental strain MVA.

SARS-CoV-2 is mainly transmitted through the respiratory tract, and the generation of mucosal immunity is essential for establishing efficacious protection against SARS-CoV-2 infection. We tested the levels of S-binding IgG and IgA in mice immunized with rMVA-S and rMVA-ST and found that both induced robust levels of IgG and IgA after immunization. rMVA-ST elicited significantly higher levels of both antibodies compared to rMVA-S. A recent study suggested that rhesus macaques immunized with adjuvanted S-Trimer were protected from SARS-CoV-2 challenge, indicating that Trimer-Tag promotes the formation of S-trimers and is beneficial to enhance effective immune responses (33). Previous studies and our data demonstrated that the conformation of S contributes to its immunogenicity and should be taken into consideration for future vaccine development (31–34). Importantly, as our recombinant MVAs still contain a functional copy of eGFP for plaque purification, the effect of GFP on immunogenicity and the biological property of the vaccines, a question sometimes omitted for recombinant vaccine studies, needs to be addressed in the future (22, 24, 42). D.S.O. Daian e Silva and colleagues systemically compared the differences in immune activation between MVA and MVA + GFP and concluded that GFP was innocuous to immunized animals and did not alter physiopathological responses to the MVA vector in mice (43).

The constant circulation of SARS-CoV-2 in the human population has led to the emergence of multiple mutated variants of the virus, which may contribute to the attenuation of the virus but also pose a challenge to efficient vaccine design. To evaluate whether rMVA-S/ST vaccines confer sufficient cross-protection, we tested the titers of binding IgG and neutralizing antibodies in animals immunized with different rMVA-S/ST against three different variants of S protein. We observed that S-binding IgG induced by different rMVA-S/STs all exhibited sufficient binding with S proteins from different virus sublineages, and the induced neutralizing antibody could also neutralize pseudoviruses expressing different S proteins. It indicates that there may be cross-protection, although animal challenge experiments with different virus variants are required to confirm it.

Virus-vectored vaccines typically induce stronger cellular immunity than subunit or inactivated virus vaccines. We used flow cytometry and ELISPOT to detect S-specific CD8+ T cells in mouse splenocytes after immunization and found that both rMVA-S and rMVA-ST induced strong cellular immune responses. Importantly, the levels of cellular immune responses elicited by rMVA-ST were significantly higher than that of rMVA-S, which agrees with the trend observed in humoral responses. These results further emphasized that the correct conformation of S contributed to effective protection.

To assess the protective effect of rMVA-S and rMVA-ST, we used a previously established mice model in which a clinical isolate of SARS-CoV-2 (BetaCov/Wuhan/AMMS01/2020) was serially passaged and adapted to BALB/c mice (39, 40). The resulting virus (SARS-CoV-2/BMA8) was able to cause lethality within 1 week of infection while displaying severe respiratory symptoms and dramatic weight loss. It is observed that MVA-immunized mice all died within 6 days after infection and exhibited significant weight loss since day 2. In comparison, mice immunized with rMVA-S or rMVA-ST all survived the challenge. Notably, weight loss was only observed in MVA control as well as rMVA-S-immunized mice but not in rMVA-ST-immunized ones, suggesting a clear advantage of rMVA-ST in reducing symptoms caused by SARS-CoV-2 over rMVA-S. Consistent with this difference, higher virus titers and mRNA copy numbers of SARS-CoV-2 N were also observed in rMVA-S-immunized mice than in rMVA-ST-immunized ones. We monitored all animals daily after the challenge and rMVA-ST-immunized mice did not display any observable clinical symptoms, demonstrating complete protection using this candidate vaccine. In previous studies, other research groups also engineered and evaluated recombinant MVA vaccines expressing monomeric S, although the strategies used were different from ours and no research groups have tried trimeric S before. Liu and Americo et al. observed complete protection using recombinant MVA in an authentic SARS-CoV-2 challenge (22, 24). However, it is important to point out that the virus strain used in the challenge was different from ours and the severity of clinical symptoms is therefore not directly comparable between different studies. It is possible that the virulence of SARS-CoV-2/BMA8 in BALB/c was enhanced during serial passage, but the specific reasons or mechanisms remain elusive (5, 20–26).

MVA is a promising vaccine vector and has been approved as a smallpox vaccine in the United States and Canada. MVA exhibits a decent safety profile and can even be used in immunocompromised patients such as those with hematopoietic stem cell transplants and HIV-positive individuals (19). The SARS-CoV-2 vector of MVA has been engineered and tested in different animal models, demonstrating strong immune effects on SARS-CoV-2 in terms of humoral and cellular immunity, as well as mucosal immunity (5, 20–26). Our study shows the first time that a trimeric S expressed from MVA exhibited enhanced humoral and cellular immune responses, displayed sufficient cross-protective capabilities, and offered superior protection against a SARS-CoV-2 lethal challenge. We believe our vaccine candidate is a nice addition to the vaccine arsenal against SARS-CoV-2 and our strategy may provide valuable insights for vaccine development for emerging and re-emerging diseases caused by coronaviruses in humans and animals.

MATERIALS AND METHODS

Mice

The 9-month-old female BALB/c mice and 6-week-old female nude BALB/c mice were obtained from Beijing Vital River Laboratory Animal Technology Company.

Construction of recombinant viruses

DNA sequences of S from various SARS-CoV2 strains containing a 3 × Flag or a Trimer-Tag at the C-terminus were synthesized (33). Synthetic sequences were cloned and inserted into a modified PRK5 transfer vector, which placed the ORF under the control of a VACV-modified mH5 early/late promoter. The inserted sequences were adjacent to an ORF of eGFP driven by a synthetic poxvirus P11 early/late promoter. To generate rMVAs, 1 µg of linearized plasmids containing the insertions was transfected into CEF cells 1 hour after MVA infection at 3 PFU/cell. Green foci were selected and clonally purified for four rounds and the purified clones were propagated in CEF cells. The purity of the recombinant viruses was further examined and confirmed by PCR amplification (forward, AATTTGGATATTCTATGGCG; reverse, AAACCAGCTGCGATGTATCA) and Sanger sequencing.

Western blotting analysis

Monolayers of HeLa cells were infected with 3 PFU/cell of MVA or rMVAs for 24 hours. Cells were lysed using 1 × cell lysis buffer (Beyotime Biotechnology) containing 1 × phenylmethylsulfonyl fluoride (PMSF). Proteins were resolved on 4%–20% gradient FastPAGE precast gels (Tsingke Biotechnology) or 3%–8% gradient large molecular weight protein precast gels (Beyotime Biotechnology). Proteins were transferred onto nitrocellulose membranes in a Trans-Blot Turbo machine (Bio-Rad). Membranes were blocked in 1 × Tris-buffered saline with 0.1% (vol/vol) Tween 20 (TBST) containing 5% non-fat milk (m/vol) for 1 hour at room temperature and then incubated with primary antibodies [Anti-nCoV-RBD rRmAb (Vazyme, RM3160, 1:1,000), anti-β-actin MmAb (Beyotime Biotechnology, AA128, 1:1,000), anti-HSP90AA1 MmAb (Solarbio, K200046M, 1:1,000), anti-A3 MpAb, anti-I3 MmAb, and VACV-positive serum from rabbit (gifts from Bernard Moss at NIH)] diluted in 5% non-fat milk at 4°C overnight with agitation. The membranes were washed with 1 × TBST three times followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibody (1:10,000) for 1 hour at room temperature. Membranes were washed four times with 1 × TBST and the ECL signals were detected using SuperSignal West Dura substrate (Thermo Fisher Scientific) and visualized by Image J.

Detection of S protein by confocal microscopy

HeLa cells cultured on glass coverslips were infected with rMVAs at 3 PFU/cell for 12 hours. Cells were fixed with 4% paraformaldehyde (PFA) for 15 minutes, and permeabilized with 0.5% Triton X-100 diluted in 1 × phosphate-buffered saline (PBS) for 15 minutes. Fixed cells were then blocked with 3% bovine serum albumin (BSA) diluted in 1 × PBS for 30 minutes at room temperature. Anti-nCoV-RBD rRmAb (Vazyme, RM3160) was diluted in 3% BSA-PBS and used to incubate the cells at 4°C overnight. Samples were washed with 1 × PBS three times and then incubated with secondary antibodies conjugated with Alexa Fluor 555 diluted in 3% BSA-PBS (1:1,000). Nuclei were stained with Hoechst (1:2,000). Cells were washed five times with 1 × PBS prior to visualization. Images were taken on a Leica SP8 confocal microscope and images were processed with the Leica software (Leica Biosystems).

Transmission electron microscopy

HeLa cells were infected with 3 PFU/cell MVA or indicated rMVAs for 12 hours and then fixed with 2.5% glutaraldehyde diluted in 0.1 M cacodylate buffer for 2–3 hours at room temperature or overnight at 4°C. The cells were then embedded in Eponate 12 resin, cut into 80 nm slices, and stained with 5% uronate lactate and 2% lead citrate. After sample preparation, imaging was performed with a JEOL1200EX transmission electron microscope.

Detection of S-binding antibodies by ELISA

Diluted synthetic S protein (1 µg/mL, Vazyme, CG202, CG221, CG234) was added to each well of a 96-well plate for plate coating and incubated for 16–18 hours at 4°C on a shaker. The plates were then blocked with 1 × PBST (Tween 20) containing 5% nonfat milk for 2 hours at room temperature. Serum was collected from immunized animals and a series of eight 4-fold dilutions of each mouse serum sample was prepared in blocking buffer. Diluted serum was added to the appropriate wells and incubated for 1.5 hours at room temperature followed by washing with 1 × PBST three times. Then secondary antibody was added to each well and incubated for 1 hour at room temperature. Plates were washed five times with 1 × PBST. Freshly prepared TMB chromogenic solution (Solarbio) was added to each well. The chemiluminescence reaction was stopped after 10 minutes by the addition of 2 M sulfuric acid (50 µL/well). An automatic enzyme label detector was used to measure the absorbance value (OD₄₅₀) at a wavelength of 450 nm. Calculations were performed to obtain the ratio of the serum OD₄₅₀ value (P) of the experimental group to the serum OD₄₅₀ value (N) of the control group to generate relative titers of the antibodies. The highest serum dilution factor corresponding to P/N > 2 was used as the antibody titer of the ELISA for the sample.

Pseudovirus neutralization assay

Serum samples were heat-inactivated for 30 minutes at 56°C and clarified by high-speed microcentrifugation. Serum samples were then diluted in DMEM containing 5 µg/mL polybrene and 10% heat-inactivated FBS, with a starting dilution of 1:40 and a twofold dilution. Diluted serum samples were then mixed with equal volumes of pseudovirus followed by incubation at 37°C for 1 hour. A total of 200 µL of pseudovirus-serum mix from each sample was used to inoculate huh-7 cells. A positive control was included in which only the pseudovirus was added without serum and a negative control was included using untreated cells. The huh-7 cells were cultured in a 37°C, 5% CO₂ cell culture incubator for 72 hours. The medium was removed from the wells and the cells were lysed with 1 × cell lysis reagent (Promega), followed by shaking at 400 rounds per minute (rpm) for 15 minutes at room temperature. Luciferase reagent (Promega) was added to the wells (100 µL/well and light units were measured on a luminometer (Promega). The relative luciferase unit (RLU) was calculated by setting the positive control as 100%, and the negative control as 0%. The highest dilution ratio of the serum when RLU reached 50% was designated IC₅₀ of serum-neutralizing antibodies.

Stimulation and staining of lymphocytes

Splenocytes from individual mice or those pooled from 5 mice were re-suspended at 1.5 × 10⁷ cells/mL in RPMI-1640 (Solarbio) supplemented with 10% heat-inactivated FBS. Splenocytes (100 µL) were mixed with 100 µL of synthetic S protein (purchased from Vazyme) of corresponding sublineages in 96-well plates to make the final concentration of S protein 10 µg/mL. Cells were incubated at 37°C for 1 hour after which brefeldin A (Sigma Aldrich) was added and incubation was continued for 5 hours. Staining of the cells was performed at 4°C. Fc receptors were blocked with anti-CD16/32 (BD Biosciences, dilutions) for 30 minutes. Surface staining was performed with anti-mouse CD3-FITC (BD Biosciences, dilutions), anti-mouse CD4-APC (BD Biosciences, dilutions), and anti-CD8-PerCP-Cy5.5 (BD Biosciences, dilutions) for 1 hour. Cells were then fixed and permeabilized with Cytofix/Cytoperm solution and stained with IFNγ-PE (BD Biosciences) for 1 hour. Fixed cells were washed with 1 × PBS and then re-suspended in 1 × PBS containing 2% paraformaldehyde. Approximately 10,000 events were acquired on an LSRFortessa cytometer, and analyses were performed with FlowJo (BD Biosciences).

SARS-CoV-2 challenge virus

All virus infection experiments were performed in a bio-safety level 3 laboratory (BSL-3). The mouse-adapted SARS-CoV-2 strain SARS-CoV-2/BMA8 (accession number: OL913103) originated from the Changchun Veterinary Research Institute, Chinese Academy of Agricultural Sciences (Changchun, China) was used in this study (39). The virus was passaged on African green monkey kidney epithelioid cells (Vero-E6) and cultured in DMEM supplemented with 10% heat-inactivated FBS in a humidified atmosphere containing 5% CO₂ at 37°C.

Vaccination and challenge experiments

Before vaccination, the rMVAs were thawed, sonicated twice for 30 s on ice, and diluted to 2 × 10⁸ PFU/mL in 1 × PBS supplemented with 0.05% bovine serum albumin. In the ABSL-2 laboratory, mice were lightly sedated with isoflurane and intranasally immunized with 50 µL of diluted virus to reach a total dose of 10⁷ PFU/animal. All mice scheduled to be infected with SARS-CoV-2 were transferred to an ABSL-3 laboratory 5 days before the virus challenge. The challenge stock of SARS-CoV-2 BAM8 was diluted to 1,000 LD₅₀/mL in 1PBS. Mice were lightly sedated with isoflurane and inoculated IN with 50 µL of SARS-CoV-2 (50 LD₅₀/animal). After infection, morbidity/mortality status and weights were assessed and recorded daily.

Determination of SARS-CoV-2 viral loads in lungs and nasal turbinates

At 3 days post-infection with SARS-CoV-2, lungs and nasal turbinates were removed and placed in 1.5–2 mL of ice-cold PBS and weights of the tissues were recorded. The tissue was homogenized and centrifuged at 4,000 × g for 10 minutes to remove debris from the homogenate, and the supernatants were titrated in triplicate on Vero E6 cells using 10-fold serial dilutions in 96-well microtiter plates. After 72 hours, the plates were stained with crystal violet and scored using the Reed-Muench method to determine TCID₅₀.

Determination of SARS-CoV-2 RNA in lungs and nasal turbinates

After homogenizing the lungs and turbinates, RNA was extracted using a total RNA extraction kit (Fasagen) according to the manufacturer’s instructions. cDNA was synthesized using First-Strand cDNA Synthesis SuperMix (Transgen). SARS-CoV-2 N transcripts were quantified by RT-qPCR with specific primers (forward, GGGGAACTTCTCCTGCTAGAAT; reverse, CAGACATTTTGCTCTCAAGCTG).

ETHICS APPROVAL

All procedures involving live SARS-CoV-2 infection in cells and animals were conducted in a biosafety level 3 (BSL-3)

laboratory and approved by the animal experimental committee of the Laboratory Animal Center, Changchun Veterinary Research Institute (approval number: IACUC of AMMS-11-2023-010). Animals were acclimatized for 3 days prior to infection, given food and water ad libitum, and monitored twice daily. Environmental enrichment was also provided in the cages during the study.

DATA AVAILABILITY

All data generated and analyzed that support the findings of this study are included in this published article (and its supplemental information).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jvi.00521-24.

Supplemental material. jvi.00521-24-s0001.pdf.

Tables S1 to S8.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.