A single dose of an adenovirus-vectored vaccine provides complete protection of the upper and lower respiratory tracts against SARS-CoV-2 challenge


 The unprecedented coronavirus disease 2019 (COVID-19) epidemic has created a worldwide public health emergency, and there is an urgent need to develop an effective vaccine to control this severe infectious disease. Here, we found that a single vaccination with a replication-defective human type 5 adenovirus encoding the SARS-CoV-2 spike protein (Ad5-nCoV) protected mice completely against SARS-CoV-2 infection in the upper and lower respiratory tracts. Additionally, a single vaccination with Ad5-nCoV protected ferrets from SARS-CoV-2 infection in the upper respiratory tract. This study suggested that a combination of intramuscular and mucosal vaccination maybe provide a desirable protective efficacy and different Ad5-nCoV delivery modes are worth further investigation in human clinical trials.

vaccine (Ad5 vector) by the intramuscular (IM) or intranasal (IN) route at week 0. S-speci c IgG antibody, anti-SARS-CoV-2 speci c neutralizing antibody (NAb), IgA and cellular immune responses were detected in each group. ELISA IgG titres peaked at day 28 in the IM groups and then slightly decreased afterwards, while those in the IN groups remained at a steady peak from week 4 to week 8 (Fig. 1c, 1f). There were higher IgG titres in the high-dose IN groups than in the IM groups (P<0.0001 at week 6, and P=0.0001 at week 8) and no difference between the middle-or low-dose IM and IN groups (P>0.05) at weeks 6 and 8 post immunization (Extended Data Fig. 1a). NAbs were detected by a virus-speci c microneutralization assay, and NAb titres reached peaks at week 6 or week 8 in the IN or IM group, respectively (Fig. 1d, 1g).
NAb titres were signi cantly higher in the high-dose IN groups than in the IM groups (P<0.0001 at week 4 and week 6, and P=0.0021 at week 8) from week 4 to week 8, while there was no difference in the middledose group at weeks 6 and 8 and no difference in the low-dose group at any time point (Extended Data   Fig. 1b). The trend of SARS-CoV-2 pseudovirus NAb (PNAb) titres was similar to that of NAb titres (Fig. 1e,  1h; Extended Data Fig. 1c), and a relatively good correlation was shown among IgG titres, NAb titres and PNAb titres at week 6 and week 8 post immunization (Extended Data Fig. 2). S-speci c IgG in the trachealung wash was detected in both the IM and IN groups at week 2 (Extended Data Fig. 3b) and week 10 (Extended Data Fig. 4a), but S-speci c IgA was found only in the IN groups (Extended Data Fig. 3c, 4b).
NAbs and PNAbs were also detected in the trachea-lung washes in both high-dose groups but not in either low-dose group (Extended Data Fig. 4c, 4d). Both the middle-dose IM and IN groups exhibited signi cant induction of IFNγ, TNFα and IL-2 responses in splenic CD8 + T cells and CD4 + T cells at week 2, with a higher level in the IM group than in the IN group (Fig. 1i, 1j). Dose-dependent cellular immune responses were found in the IM groups but not in the IN groups at week 10 (Extended Data Fig. 4e, 4f).
Seven of ten vaccinated mice in every group were inoculated intranasally with a mouse-adapted SARS-CoV-2 virus (HRB26M) at a dose of 10 3.6 plaque-forming units (PFU)) per mouse at week 10 post immunization. Four and three out of seven mice in every dose group were sacri ced for viral load quanti cation in the lungs and turbinates at 3 and 5 days post inoculation (dpi), respectively. No virus was detected in the lungs or turbinates in all the IN vaccinated groups at 3 and 5 dpi by quantitative polymerase chain reaction (qPCR) and PFU assays, while all the IN control group animals were infected with a mean viral load of 1.2×10 4 PFU/g in the turbinates and 5.6×10 5 PFU/g in the lungs at 3 dpi (Fig. 2). No virus was detected in the lungs in all the IM-vaccinated groups at 3 and 5 dpi, while all the IM control group animals' lungs were infected with a mean viral load of 3.3 ×10 6 PFU/g at 3 dpi. Virus was detected in the turbinates of some mice of the IM group by PFU assays and qPCR, with a signi cant reduction in the three vaccinated groups compared with the load of the IM control animals (Extended Data Fig. 5). These data demonstrated that a single low dose of Ad5-nCoV can completely protect both the upper respiratory tract and lungs of mice from infection and it seemed that the infection prevention in the upper respiratory tract is more di cult than in the lower respiratory tract from the different protection via different vaccination mode in mice.
The ferret is a mammalian model in which SARS-CoV-2 can replicate in its upper respiratory tract, but not in the lungs 9 .We next evaluated the protective e cacy of the mucosal vaccination and IM vaccination of Ad5-nCoV in the upper respiratory tracts of ferrets, Eighteen ferrets were equally grouped into the IM vaccination group (5×10 10 VP), the mucosal vaccination group (simultaneous oral delivery with 5×10 10 VP and IN delivery with 5×10 10 VP for one ferret) and the control group and challenged on week 4 after a single vaccination of Ad5-nCoV. All vaccinated ferrets produced S-speci c serum IgG antibodies and NAbs at week 4 post vaccination, which were not detected in control animals, with no difference between the two vaccination groups (Fig. 3a, 3b). We also observed cellular immune responses in 5 out of 6 ferrets in the IM group and 3 out of 6 ferrets in the mucosal vaccination group by IFNγ ELISpot assays at week 4 ( Fig. 3c). These animals were challenged intranasally with 10 5 PFU of SARS-CoV-2 at week 4 post vaccination, and nasal washes were collected every two days for viral load analysis of SARS-CoV-2 by qPCR and PFU assays. No virus was detected in the nose washes of mucosal vaccination group animals by the qPCR and PFU assays from 2 to 8 dpi, in contrast to all the infected control animals (Fig. 3d, 3e).
Virus was detected in the nose washes of 3/6, 2/6, 0/6 and 0/6 of the IM-vaccinated animals by PFU assay at 2, 4, 6, and 8 dpi, respectively, and a signi cant reduction in viral load was found at each day between the IM vaccination group and the control group animals (Fig. 3d, 3e).
We have demonstrated for the rst time, to our knowledge, that complete protection for the upper and lower respiratory tracts against SARS-CoV-2 infection can be achieved using a single mucosal inoculation of Ad5-nCoV in mice and ferrets. A single intramuscular inoculation of Ad5-nCoV can protect the lungs of mice from SARS-CoV-2 infection and signi cantly reduce viral replication in the upper respiratory tract of mice and ferrets. The challenges administered to these animals with a high dose of virus via the intranasal route do not re ect realistic human exposure. Whether a lower challenge dose would result in more e cient protection of the upper respiratory tract by IM vaccination remains to be determined. Although additional work is needed to fully unravel the protective mechanisms associated with the different vaccination delivery modes, we speculate that S-speci c mucosal antibodies or mucosal tissueresident T cells play important roles in protection 10 . Mucosal vaccination, which induces pathogenspeci c mucosal immunity in addition to systemic immunity, can provide a rst line of protection at the SARS-CoV-2 port of entry. Combining the rapid systemic immune response of intramuscular injection with the local mucosal immunity of the mucosal vaccination may be more effective than a single inoculation model. This work gives us a hint that the different routes of vaccination should be considered in human clinical trials for the development of the SARS-CoV-2 vaccine.
Virus titers were determined by standard plaque assay on Vero E6 cells, and virus stocks were stored in aliquots at -80 °C until use.

Construction of Ad5-nCoV expressing SARS-CoV-2 S
The full spike protein gene of SARS-CoV-2 based on the Wuhan-Hu-1 strain (NC_045512.2) was codon optimized by UpGene software, and the signal peptide was substituted with tPA for increased expression in mammalian cells. The gene was synthesized with EcoRI and HindIII upstream and downstream of the open reading frame, respectively, and cloned into the shuttle plasmid of the AdMax adenovirus system (Microbix Biosystem, Canada) by enzyme digestion and ligation. After sequencing identi cation, the shuttle plasmid with the target gene was cotransfected into HEK293 cells with the backbone plasmid (pBHGloxΔE1, 3Cre) by TurboFect transfection reagent (Thermo Scienti c, USA) according to the manufacturer's instructions. The transfected cells were passaged when they were overgrown and collected until Ad-related cytopathic effects were observed. The cells were lysed by three freeze-thaw cycles to release the recombinant viruses. The recombinant adenoviruses were con rmed by target gene sequencing, monocloned by agarose plaque selection, ampli ed by serial passage on HEK293 cells, and puri ed by ion exchange chromatography and size exclusion. The number of total VP was measured by ultraviolet spectrophotometer analysis, with one OD 260 equal to approximately 1. Vaccination and challenge of mice BALB/c mice (n = 10 per group) were immunized intramuscularly or intranasally with 5×10 9 VP (high dose), 5×10 8 VP (middle dose) or 5×10 7 VP (low dose) of Ad5-nCoV at day 0 or 5×10 9 VP of Ad5 vector as a control. Sera were collected for S-speci c ELISA, SARS-CoV-2 NAb titration (MN 50 ) and SARS-CoV-2 PNAb titration at different time points. Another three groups of mice immunized with a middle dose of Ad5-nCoV by the IM or IN route or with the control vaccine were euthanized at day 14 post immunization for splenic cellular immune response detection and trachea-lung wash antibody detection. At 10 weeks after vaccination, three of ten vaccinated mice in every group were euthanized for splenic T cell response detection and trachea-lung wash antibody and NAb detection. The remaining mice (seven per group) were challenged intranasally with SARS-CoV-2 HRB26M strain at a dosage of 10 3.6 PFU in a volume of 50 μL.
Four and three out of seven mice in every dose group were sacri ced for viral load quanti cation in the lungs and turbinates at 3 and 5 dpi, respectively.

Vaccination and challenge of ferrets
Ferrets (6/group) were randomized by body weight, sex, and age and grouped into the IM vaccination group (5×10 10 VP), the mucosal vaccination group (simultaneous oral delivery with 5×10 10 VP and IN delivery with 5×10 10 VP for one ferret) and the control group. Blood was collected at week 4 for analysis of antibody responses by ELISA, plaque reduction neutralization assays, and IFNγ ELISpot. The immunized animals were challenged intranasally with a dose of 10 5 PFU of SARS-CoV-2 HRB25 strain at day 28, and the viral load of the nasal wash was detected by qPCR and PFU assay every two days post infection.
Quantitative RT-PCR Viral loads were determined by quantitative real-time PCR. Viral RNA was extracted by using a QIAamp vRNA Minikit (Qiagen, Germany). Reverse transcription was performed by using HiScript II Q RT SuperMix for qPCR (Vazyme, China). qPCR was conducted by using an Applied Biosystems QuantStudio 5 Real-Time PCR System (Thermo Scienti c, USA) with Premix Ex Taq for probe qPCR (TaRaKa, China). N genespeci c primers (forward, 5'-GGG GAA CTT CTC CTG CTA GAA T-3'; reverse, 5'-CAG ACA TTT TGC TCT CAA GCT G-3') and probe (5'-FAM-TTG CTG CTG CTT GAC AGA TT-TAMRA-3') were utilized according to the information provided by the National Institute for Viral Disease Control and Prevention, China (http://nmdc.cn/nCoV). The amount of vRNA for the target SARS-CoV-2 N gene was normalized to the standard curve from a plasmid (pBluescript II SK-N, 4,221 bp) containing the full-length cDNA of the SARS-CoV-2 N gene. The assay sensitivity was 1000 copies/mL. A double antigen sandwich ELISA kit (ProtTech, China) was used for SARS-CoV-2 S-speci c IgG assays for the ferrets. Brie y, 100 μL of serum was added to an antigen-coated microtitre plate, and the plate was incubated at RT for 30 min and washed with PBST. Then, the plate was incubated with HRP-conjugated antigen at 37 °C for 30 min and washed with PBST. The optical density (OD) was measured at 450 nm after the addition of the substrate solution and the subsequent stop solution. Seropositivity was de ned as an OD value ≥0.2.

SARS-CoV-2 neutralization assay
The neutralizing activity of sera from the mice was assessed using a microneutralization (MN) assay. Serial dilutions of heat-inactivated sera were incubated with 100 TCID 50 of SARS-CoV-2 IME-BJ01 strain at 37 °C for 2 h. Antibody-virus complexes were added to pre-plated Vero E6 cell monolayers in 96-well plates and incubated for 48~72 h. The cells were stained with 0.05% crystal violet for 30 min. The OD was measured at 570 nm/630 nm after the addition of the decolorization solution. Neutralization results were analysed by Reed-Muench method to estimate the dilution of sera required for half-maximal neutralization of infection (EC 50 titre). The initial dilution of sera (1:16) was set as the limit of con dence of the assay. Seropositivity was de ned as a titre ≥16.
The neutralizing activity of heat-inactivated sera from the ferrets was assessed using a plaque reduction neutralization test (PRNT) assay. Serial dilutions of sera were incubated with 50 PFU of the SARS-CoV-2 HRB25 strain at 37 °C for 21 h. Antibody-virus complexes were added to pre-plated Vero E6 cell monolayers in 24-well plates and incubated for 48 h with agarose overlay. Neutralizing antibody titres were calculated as the maximum serum dilution yielding a 50% reduction in the number of plaques relative to that for control serum prepared from uninfected animals. Seropositivity was de ned as a titre ≥ 8. ELISpot Kit (Mabtech, Sweden) following the manufacturer's instructions. In brief, 1x10 5 of peripheral blood mononuclear cells (PBMCs) of vaccinated ferrets were produced by density gradient sedimentation and stimulated with inactivated SARS-CoV-2 in a pre-coated ELISpot plate for 16 h in a 37 °C humidi ed incubator with 5% CO 2 . The next day, the plate was washed 5 times with PBST, incubated for 1 h at RT with the biotin conjugated detection antibody, washed, incubated for 1 h at RT with streptavidin-HRP, washed, and developed with AEC substrate (BD Pharmingen, San Diego, CA). The plate was washed extensively in deionized water to stop color development and dried in the dark, and the spots were counted in an AID ELISPOT reader (AID GmbH, Strassberg, Germany). The result was expressed as the number of SARS-CoV-2-speci c spots per 1 million PBMCs.

Intracellular cytokine staining
Splenocytes of BALB/c mice were prepared by pushing the spleen through a 70-μm cell strainer, followed by red blood cell lysis and several washes. The cells were stimulated for 6 h at 37 °C with or without 1 μg/mL of overlapping 15-amino-acid peptides covering the S protein and with BD GolgiStop TM and BD GolgiPlug TM to block cytokine secretion. Following peptide pool stimulation, the splenocytes were washed and stained with a mixture of antibodies against lineage markers, including anti-CD3 PerCP-Cy5.5 (clone 17A2), anti-CD4 Alexa Fluor 700 (clone RM4-5), and anti-CD8 FITC (clone 5H10-1), and the viability dye Near-IR to exclude dead cells from data analysis. After one wash with PBS, the cells were xed and permeabilized with Cyto x/Cytoperm (BD Biosciences, USA), washed with Perm/Wash buffer (BD Biosciences, USA), and stained with anti-IFNγ PE (clone XMG1.2), anti-TNF PE-Cy7 (clone MP6-XT22) and anti-IL-2 Brilliant Violet 421 (clone JES6-5H4). The cells were washed successively with Perm/Wash buffer and PBS and resuspended in PBS, and data were acquired on a FACS Canto TM (BD Biosciences, USA). At least 200,000 events were collected for each sample, and the data were analysed by FACS Diva software. CD8 + and CD4 + T cells were gated from single cells (FSC-A vs FSC-H), lymphocytes (FSC-A vs SSC-A) and live CD3 + T cells (CD3 + vs Near-IR -), successively, and the detection results were de ned as the percentage of cytokine-positive cells among CD8 + or CD4 + T cells.

Statistical analysis
The analyses were performed with GraphPad Prism v.8.0.2. Two-tailed nonparametric Mann-Whitney's rank tests were conducted to compare differences between groups. Correlations were assessed by Spearman rank-correlation tests. Figure 1 Page 13/15

Figure 2
Protective e cacy of Ad5-nCoV in mice. BALB/c mice (n = 10 per group) received a single immunization with different doses of Ad5-nCoV or Ad5 vector by the IM or IN route, and seven of ten mice were challenged at week 10 with 103.6 PFU of SARS-CoV-2 HRB26M strain in a volume of 50 μL by the IN route. Tissue live virus numbers (a) and tissue viral loads (b) were detected at 3 dpi (n=4) and 5 dpi (n=3) in the lungs and turbinates, respectively. Black bars re ect geometric means; 3 dpi = red circles; 5 dpi = blue squares; dotted line = the limit of detection. Immunogenicity and protective e cacy of Ad5-nCoV in ferrets. Ferrets (n = 6 per group) received a single immunization by the IM route or two simultaneous doses by the IN and oral routes with 5×1010 VP per dose of Ad5-nCoV or received a sham vaccine. a, b, Humoral immune responses were assessed at week 4 following vaccination by S-speci c ELISA (a) and NAb titration (PRNT90) (b). c, Cellular immune response was assessed by IFNγ ELISpot assays. d, e, Live virus numbers (d) and viral load (e) of SARS-CoV-2 in nasal washes obtained from ferrets after challenge. Bar at the geometric mean, and connecting lines re ect geometric means; *=p-value < 0.05, **=p-value < 0.01, IN and oral inoculation animals = red circles; IM injection animals = blue squares; control animals = brown triangles; dotted line = the limit of detection.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.