Recombinant Bacillus Calmette-Guérin Expressing SARS-CoV-2 Chimeric Protein Protects K18-hACE2 Mice against Viral Challenge

Abstract

COVID-19 has accounted for more than 6 million deaths worldwide. Bacillus Calmette-Guérin (BCG), the existing tuberculosis vaccine, is known to induce heterologous effects over other infections due to trained immunity and has been proposed to be a potential strategy against SARS-CoV-2 infection. In this report, we constructed a recombinant BCG (rBCG) expressing domains of the SARS-CoV-2 nucleocapsid and spike proteins (termed rBCG-ChD6), recognized as major candidates for vaccine development. We investigated whether rBCG-ChD6 immunization followed by a boost with the recombinant nucleocapsid and spike chimera (rChimera), together with alum, provided protection against SARS-CoV-2 infection in K18-hACE2 mice. A single dose of rBCG-ChD6 boosted with rChimera associated with alum elicited the highest anti-Chimera total IgG and IgG2c Ab titers with neutralizing activity against SARS-CoV-2 Wuhan strain when compared with control groups. Importantly, following SARS-CoV-2 challenge, this vaccination regimen induced IFN-γ and IL-6 production in spleen cells and reduced viral load in the lungs. In addition, no viable virus was detected in mice immunized with rBCG-ChD6 boosted with rChimera, which was associated with decreased lung pathology when compared with BCG WT-rChimera/alum or rChimera/alum control groups. Overall, our study demonstrates the potential of a prime-boost immunization system based on an rBCG expressing a chimeric protein derived from SARS-CoV-2 to protect mice against viral challenge.

Figure 1.. Immunodominant B and T cell epitopes identified in the Nucleocapsid and Spike proteins for vaccine formulation.

Schematic representation of the Nucleocapsid (N) and Spike (S) structural proteins from SARS-CoV-2. Protein domains are represented by N- and C-terminal domains (NTD and CTD, respectively) for the Nucleocapsid, including the serine-glycine-arginine rich domain (SGRD) and the serine-arginine rich domain (SRD); and by S1, S2 and the Receptor Binding Domain (RBD) for Spike. Immunogenic peptides previously identified by our group are represented as numbered dashes. Identified N- and S-derived amino acid sequence and position are also shown. The complete immunoinformatic analysis is described elsewhere (12, 14). Figures were prepared using DOG illustrator of protein domains structures (88). N and S accession numbers are YP_009724397.2 and YP_009724390.1, respectively.

Figure 2.. N- and S-derived peptides are stable and reactive to SARS-CoV-2 infected sera.

NMR 3D structures for peptides (A) Npep and (B) Spep. The ten structures of lowest energy are represented with the superposition of the backbone traces. The residues’ side chains are colored by the Kyte-DoLittle hydrophobicity scale, labeled at right, being dodger blue the most hydrophobic residues and orange red the most hydrophilic ones. The highlighted side chains are in the peptide’s concavities. Furthermore, Npep (C) and Spep (D) were tested against human sera by ELISA. Sera from individuals infected with SARS-CoV-2 prior to mass vaccination and positive by molecular diagnosis (qPCR) were used as Positive individuals (n = 35). Samples from healthy people collected prior to 2015 were used as Negative individuals (n = 35). Sera were diluted 1:100 for the assay and total IgG detected values are plotted. Graphs are representative of two independent experiments. Statistical analysis were performed using Student’s t-test and * represents p value < 0.05. Dotted lines represent the cut-off between the samples and were obtained by ROC curve analyses.

Figure 3.. Recombinant Chimera is reactive to human sera from SARS-CoV-2 infected patients and also from CoronaVac immunized individuals.

A) Schematic representation of rChimera. The N-domain is composed of 45–206 amino acids from N protein (YP_009724397.2) while the S-domain incorporates 223–514 amino acids from the S protein (YP_009724390.1). The positions of peptides #1, 2, 3 and 4 are shown by dashed lines. B) Reactivity of rChimera against human sera was assessed by ELISA. Sera from SARS-CoV-2 infected individuals (from year 2020, prior to mass vaccination and positive by molecular diagnosis) and sera from healthy individuals (from year 2015, prior to the pandemic) were used as positive and negative controls of the assay (n = 35). Sera were diluted 1:100 for the assay. C) Sera from non-SARS-CoV-2-infected individuals (negative by molecular and serological diagnosis) were assessed to rChimera’s reactivity before and after CoronaVac immunization. Samples were collected at day zero (before immunization) and three weeks after CoronaVac booster dose (n = 13). Both results are representative of total IgG antibodies against rChimera detected in the samples. Graphs are representative of two independent experiments. Statistical analysis were performed using Student’s t-test and * represents p value < 0.05. Dotted lines represent the cut-off between the samples and were obtained by ROC curve analyses.

Figure 4.. Immunization of K18-hACE2 mice with rBCG-ChD6 and boosted with rChimera protein induces IgG and IgG2c/IgG1 isotype responses, neutralizing antibodies, IFN-γ and IL-6 production.

A) Schematic representation of K18-hACE2 mice vaccination. Mice (n = 5–7 mice) were immunized at day 0 with PBS, BCG WT or rBCG-ChD6 (s.c.). At day 30, mice were boosted with formulations containing rChimera+Alum or PBS+Alum. At day 50, blood was collected and spleens were harvested. Immunization groups are indicated next to the vaccination schedule. Sera collected at day 50 were used to determine antibodies against rChimera by ELISA. B) Sera was diluted serially and IgG endpoint-titers were determined by ELISA. The dilution of each sample that reached the average of the control sera (± 2 standard deviation units) was represented in bars by mean ± S.D and the dotted line indicates the limit of detection. C) Sera diluted 1:100 were assayed for IgG1 and IgG2c antibody isotype determination against rChimera. IgG2c/IgG1 ratio is represented by the heatmap and the color scale ranges from 0.0 in black to 1.5 in white, indicating the IgG2c/IgG1 ratio of detected absorbances. D-E) Spleens were harvested 20 days after booster dose. Culture cell supernatant was assayed for (D) IFN-γ and (E) IL-6 cytokine productions in response to stimulation with rChimera, LPS or ConcanavalinA. Bars represent mean cytokine concentration ± SD for each group. F) Neutralizing rates were assessed by incubating the mice sera for one hour with an HIV-1-based pseudovirus at different dilutions. Pseudovirus-sera solutions were used to infect 293T/ACE2 cells in 96-well plate and NanoLuc (RLU) activity was detected. The neutralization rate against wildtype and variant viruses is obtained by the formula described in the methods section and final results are plotted as the observed percentage of neutralization. Statistical analyses were performed using 1-way ANOVA (for B and F) and 2-way ANOVA (for D and E) followed by the Bonferroni post-hoc test. * represents p value < 0.05. Results are representative of two independent experiments. Limits of detections are described in the methods section.

Figure 5.. Prime-boost immunization regimen with rBCG-ChD6 and rChimera reduced weight loss in SARS-CoV-2 infected K18-hACE2 murine model.

A) Schematic representation of K18-hACE2 mice (n = 5–7) prime-boost vaccination. Mice were immunized subcutaneously at day 0 with PBS, BCG WT or rBCG-ChD6 (s.c.). At day 30, mice were boosted with formulations containing PBS or rChimera with Alum. SARS-CoV-2 challenge was performed intranasally at day 50 and lungs were harvested at day 56. Results are representative of two independent experiments. B) Mice were observed for the following six days and relative body weight is represented. Statistical analyses were performed using 2-way ANOVA followed by the Bonferroni post-hoc test. An asterisk represents statistical difference when rBCG-ChD6/rChimera+Alum group (indicated by black circles connected by straight lines) was compared to PBS/PBS+Alum (indicated by white circles connected by straight lines), # when it is compared to PBS/rChimera+Alum (indicated by black triangles connected by dotted lines) and & when it is compared to BCG WT/rChimera+Alum (indicated by black squares connected by dashed lines) (p value < 0.05).

Figure 6.. rBCG-ChD6/rChimera+Alum immunization provides protection against SARS-CoV-2 infection in the lungs of K18-hACE2 mice.

Harvested lungs from SARS-CoV-2 infected mice were processed after 6 days. A) RNA extraction followed by RT-qPCR using 100 ng of total RNA determined the viral load of each sample and is presented as Log₁₀ of nanograms of RNA. B) Supernatant from lung homogenates containing viable viruses were serially diluted and used to infect VERO-E6 cells for 1 hour. After applying CMC and letting plaque formation, cells were fixed with formol and stained with crystal violet. PFUs were counted and viable viral titers are represented as PFU/g. Statistical analysis were performed using 1-way ANOVA followed by the Bonferroni post-hoc test and * represents p value < 0.05. Results are representative of two independent experiments. LOD represents the limit of detection for each experiment.

Figure 7.. rBCG-ChD6 priming followed by rChimera+Alum boosting ameliorates lung pathology in K18-hACE2 mice infected with SARS-CoV-2.

The lungs from previously immunized K18-hACE2 mice were harvested 6 days post-infection with 2×10⁴ PFU SARS-CoV-2 for histopathological analyses. Panels A-D indicate representative images from lungs of each immunized group analysed at 10- and 20-times magnification with a micrometer scale depicted at each picture for comparison. Black arrows indicate inflammatory infiltrates. Panels A-C represent diffuse pneumonia while panel D represents overall preservation of pulmonary alveoli spaces. Panel E demonstrates alveoli wall thickening analyses using ImageJ software plotted as free alveolar space percentage. Twenty images were analysed for each lung and 5–7 mice were used for each group for the whole histopathological analysis. Statistical analyses were performed using 1-way ANOVA followed by the Bonferroni post-hoc test and * represents p value < 0.05. Results are representative of two independent experiments.