A Bivalent COVID-19 Vaccine Based on Alpha and Beta Variants Elicits Potent and Broad Immune Responses in Mice against SARS-CoV-2 Variants

With the emergence and rapid spread of new pandemic variants, especially variants of concern (VOCs), the development of next-generation vaccines with broad-spectrum neutralizing activities is of great importance. In this study, SCTV01C, a clinical stage bivalent vaccine based on trimeric spike extracellular domain (S-ECD) of SARS-CoV-2 variants Alpha (B.1.1.7) and Beta (B.1.351) with a squalene-based oil-in-water adjuvant was evaluated in comparison to its two corresponding (Alpha and Beta) monovalent vaccines in mouse immunogenicity studies. The two monovalent vaccines induced potent neutralizing antibody responses against the antigen-matched variants, but drastic reductions in neutralizing antibody titers against antigen-mismatched variants were observed. In comparison, the bivalent vaccine SCTV01C induced relatively higher and broad-spectrum cross-neutralizing activities against various SARS-CoV-2 variants, including the D614G variant, VOCs (B.1.1.7, B.1.351, P.1, B.1.617.2, B.1.1.529), variants of interest (VOIs) (C.37, B.1.621), variants under monitoring (VUMs) (B.1.526, B.1.617.1, B.1.429, C.36.3) and other variants (B.1.618, 20I/484Q). All three vaccines elicited potent Th1-biased T-cell immune responses. These results provide direct evidence that variant-based multivalent vaccines could play important roles in addressing the critical issue of reduced protective efficacy against the existing and emerging SARS-CoV-2 variants.

SARS-CoV-2 virus is composed of nuclear material (RNA genome) surrounded by coat proteins including the spike (S) glycoprotein, which is glycosylated and homotrimeric in nature [25]. The receptor-binding domain (RBD) is located in the S1 subunit of the spike protein, which binds to the ACE2 receptor and mediates virus entry [26]. The genomes of SARS-CoV-2 are prone to mutation, with a speed of two mutations per month [27]. Evolution of the SARS-CoV-2 genome, driven by natural selection or vaccine-resistant mutations, might result in alterations in antigenicity and infectivity [10,28]. Therefore, it is of urgency and significance to develop second-generation vaccines that can stimulate broad-spectrum neutralizing antibodies against existing and future emerging variants of SARS-CoV-2 virus. Monovalent vaccines containing antigens from a single variant strain may not be sufficient to induce cross-protection to other existing and future emerging variants. Multivalent vaccines, a traditional approach providing broad coverage against antigenically variable pathogens, might overcome this obstacle.
To evaluate the potential advantages of multivalent vaccines over monovalent vaccines in broad-spectrum neutralizing capacity, we designed a bivalent vaccine SCTV01C based on the trimeric spike protein extracellular domain (S-ECD) proteins of variants Alpha (B.1.1.7) and Beta (B.1.351), adjuvanted with SCT-VA02B, a squalene-based oil-in-water emulsion similar to adjuvant systems 03 (AS03) and MF59 ® . The immunogenicity of SCTV01C was evaluated in naïve mice in comparison to its two individual antigen components as monovalent vaccines. Comparative studies of SCTV01C or a D614G variant-based monovalent vaccine (mimicking first-generation vaccines targeting the ancestral strain of SARS-CoV-2) as booster shots were also carried out in this study.
Further clinical studies of SCTV01C are ongoing to evaluate its application in unvaccinated populations and as booster doses for previously vaccinated populations. Future development of multivalent vaccines based on the same platform approach can be rapidly deployed to address potential risks posed by evolving SARS-CoV-2 variants.

Vaccine Antigens
SARS-CoV-2 S-ECD was fused to a T4 bacteriophage fibritin motif (i.e., T4 Foldon) to stabilize the trimeric protein conformation [29] produced in CHO cells and purified as vaccine antigens. Briefly, the DNA fragments encoding S-ECD (amino acid 1 to 1208) of SARS-CoV-2 variant Alpha or Beta with the fusion of T4-foldon in the C-terminal were synthesized and constructed into pD2535nt-HDP expression vector. The plasmid was transfected into Chinese hamster ovary (CHO) cells, followed by glutamine synthetase (GS)-based selection. For protein purification, the collected supernatants were captured by cation exchange chromatograph. The eluting peaks were then polished by anion-exchange chromatography and mixed anion-exchange chromatography. Finally, the penetration peak of mixed anion-exchange chromatography was ultra-filtrated to citrate buffer. Preclinical characterization showed the antigens to be 12 × 20 nm nanoparticles with high purities and excellent thermal stability [30]. SCTV01C contains a 1:1 mixture of the trimeric protein antigens of Alpha (B.1.1.7) and Beta (B.1.351) variants.

Vaccine Immunization in Animals
All animals used in this study were raised and cared for in facilities accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care (AAALAC). All experimental procedures were conducted according to Chinese animal use guidelines and were approved by the Institutional Animal Care and Use Committee (IACUC). Female BALB/c and C57BL/6J mice (6-8 weeks old) were purchased from Beijing Vital River Laboratory Animal Technology Co. (Beijing, China). Antigens adjuvanted with SCT-VA02B (2 mg/dose) were used for immunization. A monovalent vaccine based on trimeric S-ECD of the first dominant variant (D614G variant), which shared the same sequence in the receptor-binding domain (RBD) and N-terminal domain (NTD) with the ancestral strain, was used to mimic the ancestral strain-based vaccines in immunogenicity studies of SCTV01C as booster shots.

Antigen-Specific Antibody ELISA Assay
Antigen-specific antibody in mouse serum was detected by ELISA. Briefly, 5 µg/mL of antigen was coated into ELISA plate overnight at 4 • C. After blocking with 2% BSA-TBST buffer, 100 µL of diluted serum samples was transferred to the plates and incubated for 1 h. Anti-mouse IgG F(ab)2-HRP secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA) at a dilution ratio of 1:5000 was incubated at room temperature for 1 h. For Antigen-specific mouse IgG1 and IgG2a analyses, R-mIgG1-R020-HRP (Sino Biological Inc., Beijing, China) or anti-mouse mIgG2c/HRP (Thermo Fisher Scientific, Waltham, MA, USA) were used as secondary antibody.
Plates were washed 5 times, and color development reaction was carried out by adding TMB substrates, and stopped by 2M H 2 SO 4 . Optical absorbance at 450 nm (OD 450 nm) was read in a microplate reader (BioTek). Antibody titer was defined as the highest serum dilution ratio that causes OD450 nm to be 2.1 times higher than the background value of unimmunized mouse serum samples.

SARS-CoV-2 Pseudovirus-Based Neutralization Assay
SARS-CoV-2 pseudovirus was manufactured by Sinocelltech (Beijing, China). Huh-7 cell was obtained from Chinese Culture Tissue Collection Center (CCTCC, China). Luciferase Assay System (E1501) and Passive Lysis 5× Buffer (E1941) were purchased from Promega (Madison, WI, USA). For the quantitative measurement of neutralization antibodies, the SARS-CoV-2 pseudovirus PsV-Luc-Spike carrying the firefly luciferase gene was used. Mutations in S-ECD of each variant pseudoviruses are shown in Table 1. Briefly, vaccine-immunized serum samples (heat-inactivated at 56 • C for 30 min) were serially diluted, incubated with 200 TCID 50 /well pseudovirus (1 h at 37 • C, in a 5% CO 2 incubator), and co-cultured with 2 × 10 4 Huh-7 cells for 20 h. Relative light unit (RLU) was measured to evaluate luciferase activity (CentroXS 3 LB 960 Microplate Luminometer). The calculation formula for the inhibition rate of the pseudovirus entry is: Inhibition (%) = (Postive RLU-Sample RLU)/(Postive RLU-Negative RLU) × 100%. The neutralizing antibody titer (50% inhibitory dilution, NAT 50 ) is defined as the serum dilution at which the RLUs were reduced by 50% when compared with the positive control wells. Positive neutralizing antibody (NAb) was determined as greater than 50% inhibition, and the dose required to achieve this effect in 50% of the animals (ED 50 ) was calculated.

Detection of Th1 and Th2 Cytokines
Antigen-Specific T-cell responses were quantified with IFNγ and IL-4 enzyme-linked immunospot (ELISpot) kit. Mouse IFN-gamma ELISpot PLUS (ALP), strips (3321-4AST-2) and Mouse IL-4 ELISpot PLUS (ALP) (3311-4APW-2) were purchased from Mabtech (Nacka, Sweden). For in vitro stimulation of mouse splenocytes, 15-mer peptides with 11 overlapping amino acids covering the entire SARS-CoV-2 spike protein were used (Sci-Light Biotechnology, Beijing, China), since 90% of the CD4 + T cell epitopes and 97% of the CD8 + T cell epitopes are conserved in the Alpha and Beta variants when compared with the ancestral strain [31]. Briefly, the isolated mouse splenocytes were cultured into ELISpot plate at a density of 2 × 10 5 cells/100 µL per well. Then, Spike S peptides at a concentration of 2 µg/mL, with a volume of 100 µL per well, were added (or not, for negative control wells without stimulation). Cells were incubated for about 20 h, and then IFNγ or IL-4-positive cells were measured. The spots were counted in the Enzyme Linked Spot Analyzer (ImmunoSpot ® S6, CTL, Cleveland, OH, USA). The negative control value was used to subtract the background from the sample value. Values below zero were presented as zero. Data analysis was carried out with GraphPad Prism.

Statistical Analysis
Statistical analysis in this study was performed with GraphPad Prism (version 8.2.1, GraphPad Software, San Diego, CA, USA). Data concerning antibody titers, which followed the skew normal distribution, were presented as geometric mean titer (GMT) ± standard deviation (SD) and were log transformed, resulting in a normal distribution of the data, and then analyzed by unpaired, two-tailed Student's t test. Other data in this paper met the criteria of normal distribution, were presented as Mean ± SD and analyzed using unpaired, two-tailed Student's t test. p < 0.05 was considered statistically significant. The immunogenicity and dose response of the bivalent vaccine SCTV01C were first evaluated with a one-dose immunization scheme. BALB/c mice were injected with a single dose of trimeric B.1.351 S-ECD monovalent vaccine, trimeric B.1.1.7 S-ECD monovalent vaccine, or SCTV01C at five dose levels (0.5, 1, 2, 4 and 8 µg). Four weeks after immunization, serum samples were analyzed for neutralizing activities against the B.1.1.7 and B.1.351 variants using pseudovirus (PsV) neutralization assays ( Figure 1A).

Humoral Immune Responses
As shown in Figure  Dose responses of SCTV01C and the two monovalent vaccines were analyzed. The dose required to achieve greater than 50% inhibition in the PsV neutralization assay in 50% of the mice was defined as the median effective dose (ED 50 ) as a measure of vaccine potency ( Figure 1C). The immunogenicity of SCTV01C and the two monovalent vaccines were further evaluated in C57BL/6J mice with a two-dose prime-boost regimen at a dosage of 1 µg /antigen, based on dose-response study of SCTV01C and its two corresponding monovalent vaccines ( Figure 1). Seven days after the booster immunization, sera were analyzed for antigen-specific antibody titers and neutralizing activities ( Figure 2A     Among the 14 variants evaluated, the B.1.1.7 monovalent vaccine induced higher NAT 50 against five variants, while the B.1.351 monovalent vaccine induced higher NAT 50 against the other nine variants. Apparently, the B.1.351 variant seems to be a better monovalent vaccine candidate against viral variants. Nevertheless, neither monovalent vaccine was able to induce potent and balanced cross-neutralizing antibody titers against all 14 variants, and hence may not be an ideal second generation vaccine candidate. The limited cross-neutralizing activities of monovalent vaccines against diverse variants might be compensated by a multivalent vaccine, as different variants could offer broad coverage on the key mutations in the RBD region. This is supported by published results that broad immunity against wild type (WT) as well as B.1.1.7, P.1, and B.1.351 variants in a murine model was induced by an antigen representing multiple SARS-CoV-2 variants [35].

PsV neutralization
Conversely, the bivalent vaccine SCTV01C induced potent and broad cross-neutralizing antibodies against all 12 non-vaccine variants ( Figure 3 and Table 1). In comparison to the two monovalent vaccines, SCTV01C showed equivalent levels of NAT 50 against the P.1, B.1.1.529 and 20I/484Q variants but significantly higher NAT 50 against the remaining nine variants. The NAT 50 against the B.1.1.529 variant was 437, the lowest NAT 50 among the titers against all 14 variants measured, due to the fact that the B.1.1.529 variant carries 25 distinctive mutations when compared with those mutations harbored in the S protein of other VOCs [36]. The above results clearly demonstrated significant advantages of SCTV01C over the two monovalent vaccines and its potential application as a second generation vaccine candidate with cross-neutralizing activities against existing variants and variants of potential concern emerging in the future.  .1.617.2 B.1.1.529  C.37  B.1.621 B.1.526 B.1.617.1 B.1 Vaccine-induced T cell immunity, especially a Th1-biased immune response, is crucial for the COVID-19 vaccine protective efficacy. A Th2-biased immune response is considered to be the major cause for vaccine-associated enhanced respiratory disease (VAERD), as shown by a preclinical study of an inactivated SARS-CoV vaccine, as well as clinical research of a respiratory syncytial virus (RSV)-inactivated vaccine [37,38].
To evaluate the cellular immune response, splenocytes from vaccinated mice were used for detection of IFNγ-positive (representing Th1 immune response) or IL-4-positive (representing Th2 immune response) splenocytes after stimulating with SARS-CoV-2 spike protein S polypeptides (2 µg/mL) using ELISpot assays. SCTV01C and the two monovalent vaccines induced significant T cell responses as indicated by the number of spot-forming cells (SFC) ( Figure 4B). The ratios of IFNγ + SFCs/IL-4 + SFCs in each group were 2.38, 1.31, 3.96 ( Figure 2C), suggesting the T cell responses induced by the three vaccines were all Th1-biased. Th1-biased responses induced by SCTV01C were further validated by concentrations of Th1 (IFNγ, IL-2) and Th1 (IL-4, IL-5 and IL-13) cytokines in culture supernatants from SCTV01C vaccinated mouse splenocytes in the presence of S peptides ( Figure 4C).

Humoral Immune Responses Induced by Booster Dose of SCTV01C
The necessity of the COVID-19 vaccine booster shots is widely recognized due to waning vaccine immunity over time. A D614G variant-based monovalent vaccine was used to mimic the original strain-based first-generation vaccines. C57BL/6J mice previously immunized with a two-dose schedule of the D614G vaccine received two booster shots of SCTV01C (Mix Regimen) or the D614G vaccine (D614G Regimen), and neutralizing antibody titers against the PsV of B. 1.1.7, B.1.351, B.1.617.2 and B.1.1.529 variants were analyzed 14 days after the third or the fourth immunization ( Figure 5A).  Figure 5C). The Mix Regimen exhibited the highest levels of NAT 50 against all four variants versus the D614G Regimen, suggesting the potential application of SCTV01C as a booster shot for the population previously immunized with the original strain-based first-generation vaccines or previously infected by original strain of SARS-CoV-2.

Discussion
More than 15 first-generation vaccines developed based on the original strain of SARS-CoV-2 have been approved or authorized for emergency use, including two mRNA vaccines, two subunit vaccines, four viral vector vaccines and seven inactivated vaccines [39]. However, the emergence of a large number of SARS-CoV-2 variants with mutations in key neutralizing epitopes not only increases the transmissibility but also reduces the protective effect of the first-generation vaccines and the effectiveness of the therapeutic antibodies based on the original strain of SARS-CoV-2 [33,[40][41][42][43][44][45]. Thus, variant-tailored secondgeneration vaccines, especially those with broadly protective efficacy, might be an effective approach to defend against the pandemic.
Among the VOI and VOC variants designated prior to the emergence of the new variant Omicron (B.1.1.529), the B.1.351 variant showed the most significant impact on the efficacy of first-generation vaccines [40,43,46], and has been the main focus for the development of second-generation vaccines against SARS-CoV-2 variants [32,47]. In this study, the B.1.351 monovalent vaccine induced potent neutralizing activity against the B.1.351 variant, with an equivalent NAT 50 against the P.1 variant, which may be attributed to the fact that both B.1.351 and P.1 contain the K417N/T, E484K, and N501Y mutations [33]. This is consistent with previous studies based on the trimeric S-ECD vaccine [32] and B.1.351-infected sera [34].
However, the neutralizing activity of the B.1.351 monovalent vaccine against the other 13 major variants were markedly decreased, in particular, and no neutralizing activity against the C.36.3 variant was detected. The B.1.1.7 monovalent vaccine also showed severely reduced neutralizing activities against most of the variants tested. These data indicate that vaccines based on a single variant may not be sufficient to induce crossneutralizing activities. Ancestral strain-based vaccines reported to exhibit relatively broad neutralizing activities against SARS-CoV-2 variants also showed relatively diminished cross-reactivity against variants [48][49][50].
A preliminary study has indicated that antigens representing multiple virus strains can generate broad-spectrum immunity against WT as well as VOCs (such as B.1.1.7, P.1, and B.1.351) in mouse models [35], and immunization with multiplexed chimeric spikes can potentially induce broader protection against Sarbecoviruses [51]. Multivalent vaccines, with broad coverage on the key mutations in the spike protein of various SARS-CoV-2 variants, might induce more balanced cross-neutralizing activities. Protein-based or RNAbased bivalent vaccines, composed of antigens originating from the ancestral strain and variant strain of SARS-CoV-2, showed balanced neutralizing activities against both antigenmatched strains in mouse models [32,47,52,53]. Some of the candidates of the bivalent vaccines have entered the clinical stage for further assessment of efficacy against COVID-19 variants (NCT04927065, NCT05004181, NCT04889209, NCT05029856, NCT04904549); however, cross-neutralization activities of these vaccines against antigen-mismatched variants have not been elaborated in preclinical studies thus far.
In this study, we evaluated the neutralizing activity of variants Alpha-and Betabased bivalent vaccine SCTV01C against 14 SARS-CoV-2 variants, including VOCs, VOIs and VUMs. SCTV01C induces potent and relatively balanced broad-spectrum humoral immune responses against all 14 variants tested, with potential advantages over the two corresponding monovalent vaccine ( Figure 3 and Table 1). Furthermore, strong Th1-biased T cell responses were also observed in the mouse immunogenicity studies (Figure 4). Our data also demonstrated potential application of SCTV01C as a booster shot post immunization with the first-generation vaccines ( Figure 5).
The current study has potential limitations. First, due to the requirement for a BSL-3 facility, SCTV01C-vaccinated mouse sera were not used for assessment of neutralization activities against live viruses; however, this will be evaluated in clinical stages with SCTV01C-vaccinated human sera. Second, the protection efficacy of SCTV01C against antigen-matched or antigen-mismatched variants needs to be further evaluated in mouse challenge model.
In summary, our data suggest that the bivalent vaccine SCTV01C could induce broad cross-neutralizing immune responses against SARS-CoV-2 VOCs, VOIs, and VUMs, as well as potent Th1-biased T cell responses. Clinical studies of SCTV01C in previously unvaccinated and vaccinated populations are needed to further evaluate its application in controlling the COVID-19 pandemic.