A live attenuated SARS-CoV-2 vaccine constructed by dual inactivation of NSP16 and ORF3a

Abstract

Background: Live attenuated vaccines against SARS-CoV-2 activate all phases of host immunity resembling a natural infection and they block viral transmission more efficiently than existing vaccines in human use. In our prior work, we characterised an attenuated SARS-CoV-2 variant, designated d16, which harbours a D130A mutation in the NSP16 protein, inactivating its 2'-O-methyltransferase function. The d16 variant has demonstrated an ability to induce both mucosal and sterilising immunity in animal models. However, further investigation is required to identify any additional modifications to d16 that could mitigate concerns regarding potential virulence reversion and the suboptimal regulation of the proinflammatory response.

Methods: Mutations were introduced into molecular clone of SARS-CoV-2 and live attenuated virus was recovered from cultured cells. Virological, biochemical and immunological assays were performed in vitro and in two animal models to access the protective efficacies of the candidate vaccine strain.

Findings: Here we describe evaluation of a derivative of d16. We further modified the d16 variant by inverting the open reading frame of the ORF3a accessory protein, resulting in the d16i3a strain. This modification is anticipated to enhance safety and reduce pathogenicity. d16i3a appeared to be further attenuated in hamsters and transgenic mice compared to d16. Intranasal vaccination with d16i3a stimulated humoural, cell-mediated and mucosal immune responses, conferring sterilising protection against SARS-CoV-2 Delta and Omicron variants in animals. A version of d16i3a expressing the XBB.1.16 spike protein further expanded the vaccine's protection spectrum against circulating variants. Notably, this version has demonstrated efficacy as a booster in hamsters, providing protection against Omicron subvariants and achieving inhibition of viral transmission.

Interpretation: Our work established a platform for generating safe and effective live attenuated vaccines by dual inactivation of NSP16 and ORF3a of SARS-CoV-2.

Funding: This work was supported by National Key Research and Development Program of China (2021YFC0866100, 2023YFC3041600, and 2023YFE0203400), Hong Kong Health and Medical Research Fund (COVID190114, CID-HKU1-9, and 23220712), Hong Kong Research Grants Council (C7142-20GF and T11-709/21-N), Hong Kong Innovation and Technology Commission grant (MHP/128/22), Guangzhou Laboratory (EKPG22-01) and Health@InnoHK (CVVT). Funding sources had no role in the writing of the manuscript or the decision to submit it for publication.

Keywords: COVID-19; Live attenuated vaccine; NSP16; ORF3a; SARS-CoV-2; Sterilising immunity.

Fig. 1

Attenuation of d16i3a in cell culture. (a) Schematic representation of the SARS-CoV-2 genome in WT, d16 and d16i3a. The red asterisk () highlights the D130A mutation in NSP16. (b) Genotyping to verify the inverted ORF3a sequence in d16i3a. P1, P2, and P3 indicate the positions of the three primers on SARS-CoV-2 genome. Black arrows indicate the expected bands. (c) Sequencing results of WT, d16 and d16i3a. The desired point mutation in NSP16 was highlighted by the red box. (d) Immunofluorescent staining. VeroE6-TMPRSS2 cells infected with WT, d16 or d16i3a were fixed at 24 h post-infection (hpi), and SARS-CoV-2 N (red) and ORF3a (green) proteins were detected. Cellular nuclei were stained with DAPI (blue). (e) Western blot analysis. Lysates of VeroE6-TMPRSS2 cells infected with WT, d16 or d16i3a at 0.1 MOI were collected at 24 hpi and probed with RBD- and ORF3a-specific antibodies. (f) Plaque phenotype. VeroE6 cells were infected with the indicated viruses. Cells were fixed and stained with 1% crystal violet. (g) CaCO2, A549-ACE2-TMPRSS2, and Calu-3 cells were infected with WT (black), d16 (blue) or d16i3a (pink) at 0.01 MOI. Viral RNA (vRNA) of the RNA-dependent RNA polymerase (RdRP) region in the supernatant was quantified by RT-qPCR.

Fig. 2

Innate immunostimulatory and inflammasome-activating properties of d16i3a virus. (a–c) GSEA comparing d16i3a versus WT (a), d16 versus WT (b) and d16i3a versus d16 (c). Top 20 enriched molecular pathways were listed. The size of the dot representing each molecular pathway indicated the number of enriched genes. Rich ratio was calculated by dividing the number of enriched genes over the total number of known genes involved in that pathway. (d) Heatmap comparing the differential transcriptome of d16i3a versus WT, d16 versus WT or d16i3a versus d16. P values shown are FDR-adjusted. Differentially expressed genes between d16i3a and WT or between d16i3a and d16 mentioned in the text are highlighted with stars (★) and arrows (⇩), respectively. (e) PMA-differentiated THP-1 macrophages were left untreated or infected with WT or d16i3a at an MOI of 0.1. After 24 h, cells were fixed, stained with DAPI, immunostained for ASC and observed by confocal microscopy. Asterisks represent ASC speck formation. Biological triplicates were performed per sample. (f) The percentages of cells positive for ASC speck were calculated. Student's t test was performed to validate statistical significance. ∗: P < 0.05. (g) Experiment described in (e) was repeated with the addition of DMSO or 10 μM MCC950.

Fig. 3

Attenuation of d16i3a virus in a hamster model. (a) Viral challenge scheme of hamsters. Hamsters were intranasally inoculated with 10,000 PFU (in 50 μl) of WT, d16, or d16i3a virus. On 4 dpi, the lungs, trachea and nasal wash were harvested for viral yield detection. (b, c) Viral loads determined by RT-qPCR (b) and plaque assay (c) on 4 dpi (n = 4/group). RdRP vRNA in the lungs and trachea were normalised to β-actin. All P values were calculated using one-way ANOVA. (d) Cytokine gene expression. Relative expression in lung tissue homogenates normalised to that of β-actin was quantified by RT-qPCR. Results were shown as mean ± SEM. au: arbitrary unit. (e) Representative hamster lung tissue sections harvested on 4 dpi were stained with H&E. Scale bars, 100 μm. (f) Viral challenge scheme of mice. Mice were intranasally inoculated with 1000 PFU (in 20 μl) of WT, d16 or d16i3a virus. (g) Survival analysis of mice (n = 5/group) for 14 days. The survival curves of d16 and d16i3a groups are overlapped. (h) vRNA titre on 4 dpi by RT-qPCR. RdRp vRNA in the lungs was normalised to β-actin mRNA. (i) Lung histopathological analysis. Representative sections of lung tissues from K18-hACE2 mice harvested on 4 dpi were stained with H&E. Scale bars, 100 μm. All P values were calculated using Student's unpaired t-test. ∗: P < 0.05. ∗∗: P < 0.01. ∗∗∗: P < 0.001. ∗∗∗∗: P < 0.0001.

Fig. 4

Protective effect of d16i3a in SARS-CoV-2-infected hamsters. (a) Vaccination and viral challenge scheme. Hamsters were intranasally inoculated with 5000 PFU (in 50 μl) of d16, d16i3a or PBS (placebo) on days 0 and 14, respectively. On day 29, the immunised hamsters were challenged with 10⁵ PFU (in 50 μl) of a clinical isolate of SARS-CoV-2 Delta or Omicron BA.1 variant. Respiratory tissues (trachea and lungs) and the nasal wash were collected on day 31. (b, c) Viral titre determination by RT-qPCR (b) and plaque assay (c) on 2 dpi (n = 5/group). RdRP vRNA in the lungs and trachea was normalised to β-actin transcript. RdRP vRNA in 1 ml of nasal wash was calculated. The dashed line indicates the limit of detection. (d) SARS-CoV-2 neutralisation tests to examine the neutralisation ability of hamster sera collected from different groups on days 14 and 28 post vaccination against Delta, Omicron BA.1 or Omicron BA.5. (e) The cross-neutralisation capability of sera collected from d16i3a-immunised and then Delta- and Omicron BA.1-infected hamsters (1:10 dilution) on day 31 was examined using the SARS-CoV-2 sVNT assay. The dashed line in black represents the assay cutoff at 30% inhibition, whereas the dashed line in red indicates the inhibition by the positive control provided in the assay (left panel). SARS-CoV-2 neutralisation tests were also carried out to examine the neutralisation ability of hamster sera collected from PBS and d16i3a groups on day 31 (right panel). (f) Anti-RBD IgA in bronchoalveolar lavage fluid (BALF) and nasal wash (NW) collected from d16i3a-vaccinated and Omicron-infected hamsters on day 31 post-vaccination. (g) Histopathological changes in the lungs of SARS-CoV-2-challenged hamsters in the d16i3a or PBS group on 2 dpi. Representative sections were stained with H&E. Scale bars, 100 μm. All P values were calculated using Student's unpaired t-test. ∗∗: P < 0.01. ∗∗∗: P < 0.001. ∗∗∗∗: P < 0.0001.

Fig. 5

d16i3a vaccine induces strong intrapulmonary and splenic cellular response in K18-hACE2 transgenic mice infected with Omicron BA.1. (a) Vaccination and viral challenge scheme for T cell assay. K18-hACE2 mice were intranasally vaccinated with d16i3a or PBS (n = 5 mice/group) and subsequently challenged with BA.1 virus on day 29. On day 4 post-infection, flow cytometry was performed on intrapulmonary and splenic T cells. (b–f) Effector phenotype of intrapulmonary CD4⁺ (b, c) and CD8⁺ (d–f) T cells was assessed. Central memory (CM), effector memory (EM), and short-lived effector cells (SLEC) were annotated by CD44⁺ CD62L⁺, CD44⁺ CD62L⁻ and KLRG1⁺ CD127⁻, respectively. Frequencies (Freq) of positive cells were shown as mean ± SEM. Data were analysed by Mann–Whitney U test. (g) Anti-RBD IgG ELISA was performed using d16i3a-vaccinated mouse sera collected on day 33 post-vaccination. Error bars represent mean ± SEM. The cross-neutralisation capability of sera from d16i3a- or d16-vaccinated and Omicron BA.1-challenged mice was examined using the SARS-CoV-2 sVNT assay. The dashed line represents the assay cutoff at 30% inhibition. (h, i) Frequencies (Freq) of splenic germinal centre (GC) B cells (CD95⁺ GL7⁺) (h) and T follicular helper cells (Tfh) (CXCR5⁺ PD1⁺) (i) were assessed. Results were shown as mean ± SEM. Data were analysed by Mann–Whitney U test (ns: not significant; ∗: P < 0.05; ∗∗: P < 0.01).

Fig. 6

Application of d16i3a as a live attenuated vaccine booster in SARS-CoV-2-infected hamsters. (a) Vaccination and viral challenge scheme. Hamsters were intramuscularly injected with inactivated virus of SARS-CoV-2 and subsequently injected with the same dose of SARS-CoV-2 inactivated vaccines or intranasally inoculated with d16i3a SARS-CoV-2 or PBS (placebo) on day 0 or 21. On day 36, the immunised hamsters were challenged with 10⁵ PFU (in 50 μl) of a clinical isolate of SARS-CoV-2 BA.5. Respiratory tissues (trachea and lungs) and the nasal wash were collected on day 38. RdRP vRNA and viral load were detected by RT-qPCR and plaque assay. (b, c) Viral titre determination by RT-qPCR and plaque assay for the nasal wash, trachea and lungs of SARS-CoV-2-challenged hamsters in the d16i3a or PBS group on 2 dpi (n = 4/group). RdRP vRNA in the lungs and trachea was normalised to β-actin mRNA. Data were analysed by one-way ANOVA. (d) Histopathological changes in the lungs of SARS-CoV-2-challenged hamsters in the d16i3a or PBS group on 2 dpi. Representative sections were stained with H&E. Scale bars, 100 μm. All P values were calculated using Student's unpaired t-test. ∗: P < 0.05. ∗∗: P < 0.01. ∗∗∗: P < 0.001. ∗∗∗∗: P < 0.0001.

Fig. 7

Generation and administration of d16i3a-XBB.1.16 as a vaccine booster. (a) A schematic representation of the SARS-CoV-2 genome editing strategy to generate d16i3a-XBB.1.16. (b) Plaque phenotype after 96 h of virus incubation at 37 °C. Plaque sizes were quantified by ImageJ and analysed by one-way ANOVA. ∗∗: P < 0.01; ∗∗∗∗: P < 0.0001. (c) Virus replication kinetics of the indicated recombinant viruses in A549-ACE2-TMPRSS2 cells. Viral subgenomic E RNA (sgRNA) in the supernatant was quantified by RT-qPCR. Infectious virions in the supernatant at 48 hpi were measured by plaque assay. Data were analysed by one-way ANOVA. ∗∗: P < 0.01; ∗∗∗: P < 0.001. (d) Viral challenge scheme. (e) Micro-neutralisation assay against the indicated live viruses using the hamster serum after third dose of vaccination. Data were analysed by Mann–Whitney U test. (f, g) Co-housing experiment was performed on day 87 to determine the blockage of virus transmission (n = 4/5 per group). Viral RNA load (f) and infectious virus titre (g) in lungs and nasal wash of donor hamsters were determined by RT-qPCR and plaque assay, respectively. RdRP vRNA in the lungs was normalised to β-actin mRNA. (h) Viral infectious titre in lungs and nasal wash of recipient hamsters collected 2 days after co-housing was determined by plaque assay. Results were shown as mean ± SEM. Data were analysed by Kruskal–Wallis test (∗∗: P < 0.01).