Heterologous chimpanzee adenovirus vector immunizations for SARS-CoV-2 spike and nucleocapsid protect hamsters against COVID-19

Abstract

Available COVID-19 vaccine only provide protection for a limited time due in part to the rapid emergence of viral variants with spike protein mutations, necessitating the generation of new vaccines to combat SARS-CoV-2. Two serologically distinct replication-defective chimpanzee-origin adenovirus (Ad) vectors (AdC) called AdC6 and AdC7 expressing early SARS-CoV-2 isolate spike (S) or nucleocapsid (N) proteins, the latter expressed as a fusion protein within herpes simplex virus glycoprotein D (gD), were tested individually or as a mixture in a hamster COVID-19 SARS-CoV-2 challenge model. The S protein expressing AdC (AdC-S) vectors induced antibodies including those with neutralizing activity that in part cross-reacted with viral variants. Hamsters vaccinated with the AdC-S vectors were protected against serious disease and showed accelerated recovery upon SARS-CoV-2 challenge. Protection was enhanced if AdC-S vectors were given together with the AdC vaccines that expressed the gD N fusion protein (AdC-gDN). In contrast hamsters that just received the AdC-gDN vaccines showed only marginal lessening of symptoms compared to control animals. These results indicate that immune response to the N protein that is less variable than the S protein may potentiate and prolong protection achieved by the currently used S protein based genetic COVID-19 vaccines.

SARS-CoV-2 was first detected in humans towards the end of 2019. Since then, it has infected over 570 million humans and killed nearly 6.4 million. Although vaccines that are highly efficacious against disease caused by the initial isolates have been available for over a year, the virus is constantly evolving and the latest variants, such as omicron BA.4 and BA.5, in part evade vaccine-induced virus neutralizing antibodies (VNAs) [1].

Vaccines that are currently available are either mRNA vaccines [2,3], Ad vector vaccines [[4], [5], [6]], protein vaccines [7], or inactivated viral vaccines [8]. Most vaccines express only the SARS-CoV-2 S protein, which is the target antigen for VNAs [9] that protect against an infection [10]. Two doses of mRNA vaccines are initially over 90% effective in preventing symptomatic disease [2,3]. The Ad vector-based Sputnik V vaccine is equally protective [6] while the single dose Ad vector vaccines from Johnson and Johnson (J&J) [4] and AstraZeneca protect 63–67% of individuals against disease [5], which is similar to the level of protection achieved with the protein vaccine [7] or inactivated viral vaccines [8]. Vaccines efficacy is not sustained necessitating booster immunizations after 6 months [11,12]. Notwithstanding, even fully vaccinated individuals experience breakthrough infections [13,14] especially with omicron [15] and as SARS-CoV-2 will continue to evolve vaccine-induce protection will decline even further.

T cells induced by natural infections with common cold coronaviruses cross-react with SARS-CoV-2 antigens [16] and they have been implicated to provide protection to so-called ‘never COVID’ individuals [17]. The N protein is one of the most abundantly expressed viral antigen in SARS-CoV-2 infected cells, which should make it an easy target for specific T cells [18]. Antibodies to the N protein may also contribute to viral clearance through antibody dependent cellular cytotoxicity (ADCC) or complement-mediated lysis. The benefits of vaccinating against the N protein of SARS-CoV-2 has already been demonstrated by others, who showed that immune responses against the N protein given alone [19], in combination with the membrane protein of SARS-CoV-2 [20] or the S protein [[21], [22], [23]] can provide protection in animal challenge models, which in one study was linked to the ability of N protein-specific antibody-dependent complement deposition, phagocytosis and natural killer cell activation [23].

Here we describe results from a hamster challenge study. Hamsters were immunized sequentially with two serologically distinct replication-defective AdC vectors of serotype SAdV-26, also called AdC6, and SAdV-24, also called AdC7 [24], expressing the S protein or the N protein. Inserts were based on unmodified genes from an early viral isolate. The N protein was expressed as a fusion protein within herpes simplex virus gD, which through inhibition of the early B and T cell attenuator (BTLA) - herpes virus entry mediator (HVEM) checkpoint broadens SARS-CoV-2 N protein-specific CD8⁺ T cell responses [25,26] and augments B cell responses [27]. Vaccines were given in a heterologous prime boost regimen either individually or as mixtures of the same vector backbones expressing the two different inserts. Our data show that vaccines that express the gDN fusion protein induce N-protein-specific antibody and CD8⁺ T cell responses but fail to provide robust protection against infection or disease. As expected, the S protein-expressing vaccines induce VNAs and reduce disease after SARS-CoV-2 challenge but fail to completely prevent infection. Best protection is achieved with the vaccine mixtures indicating that although N protein-specific immune mechanisms by themselves fail to lessen disease upon challenge with a high dose of SARS-CoV-2 they enhance protection mediated by S protein-specific immunity.

1. Material and methods

1.1. Cell lines

HEK 293, BHK-21/WI-2 and A549-hACE-TMPRSS2 (A549 A/T) cells (InvivoGen, cat no. a549-hace2tpsa) were grown in DMEM supplemented with 10% or 5% FBS and antibiotics. Vero TMPRSS2 cells were maintained in DMEM with 10% FBS, 300 μg/ml of hygromycin, 0.5 μg/ml of puromycin and antibiotics.

1.2. Challenge virus

The WT WAS stock was generated from a seed stock (2019-nCoV/USA-WA1/2020). It was obtained from Biodefense and Emerging Infections Research (BEI) resources (Cat # NR-52281, Lot # 70036318) and expanded in Calu-3 cells at 37 °C for 3 days. The stock (Lot # 12152020-1235) was titrated in Vero TMPRSS2 cells.

1.3. Generation and quality control of AdC vectors

Construction and expansion of the AdC-S and AdC-gDN vectors has been described [28]. Briefly the AdC-S vectors carry the gene encoding the full-length S gene from an early viral isolate from Sweden, called SARS-CoV-2/human/SWE/01/2020 (AdC-S_SWE, GenBank number: QIC53204), while the AdC-gDN vectors carry the N gene (GenBank: QIC50514) that upon removal of the start and stop codons had been cloned into herpes simplex virus gD.

1.4. Generation of VSV-S vectors

VSV vectors pseudotyped with S of SARS-CoV-2 were generated in BHK-21/WI-2 cells using a the ΔG-GFP (G∗ΔG-GFP) rVSV kit (Kerafast, Boston, MA, USA) and S sequences cloned into an expression plasmid under the control of the CMV promoter as described previously [29]. For some vectors the S_SWE sequence was modified to incorporate the E484K and N501Y RBD mutations of the B1.351 South African variant (VSV-S_SWE/B1.351), the N501Y and P681H RBD mutations of the B1.1.7 UK variant (VSV-S_SWE/B1.1.7), the L452R and D614G mutations of the Indian B.1.617.2 delta variant (VSV-S _{SWE/B.1.617.2}), or a full-length synthetic sequence of the omicron variant BA.1/B.1.1.529.1 (VSV-S _B.1.1.529.1).

1.5. Hamster immunization and challenge

Hamsters were separated into 4 groups, which were each composed of 4 female and 4 male animals. They were primed by intramuscular (i.m.) injections: group 1 was injected with 1 × 10¹⁰ virus particle (vp) of AdC7-gDN, group 2 with 1 × 10¹⁰ vp of AdC7-S, group 3 with a mixture of 0.5 × 10¹⁰ vp AdC7-gDN and 0.5 × 10¹⁰ AdC7-S and group 4 with 1 × 10¹⁰ vp of the AdC7-HBV2 control. Animals were boosted 2 months after priming with AdC6 vectors expressing the same inserts and used at the same doses.

Prior to challenge, hamsters were anesthetized with 80 mg/kg Ketamine and 5 mg/kg xylazine given i.m. The animals were challenged with 6 × 10³ PFU of SARS-CoV-2 in 100 μL per animal (50 μl/nostril). ∼20 min post challenge antisedan (0.04 ml, 1 mg/kg) was given i.m. to all hamsters.

1.6. ELISA for anti-S1 and anti-S2 antibodies

Sera from individual hamsters were tested for S-specific antibodies by ELISA on plates coated with 100 μl of a mixture of S1 and S2 (Native Antigen Company, Kidlington, UK) diluted to 1 μg/ml in bicarbonate buffer. The next day plates were washed and blocked for 24 hours at 4 °C with 150 μl of a 3% BSA-PBS solution. Sera were diluted in 3% BSA-PBS and added to the plates, which were incubated at room temperature for 1 h and washed 4 × with 150 μl of PBS. A goat anti-hamster IgG (H + L)-alkaline phosphatase antibody (Sigma–Aldrich, St Louis, MO) diluted 1:1000 in 3% BSA-PBS was added for 1 h at room temperature. Plates were washed and phosphatase substrate (Sigma–Aldrich, St Louis, MO) in diethanolamine buffer was added. Plates were read in an ELISA reader at 405 nm. Background data were subtracted from the experimental data. Data are expressed as area under the curve (AUC) for the different dilutions.

1.7. ELISA for receptor binding domain (RBD)-binding antibodies

Sera were tested for inhibition of ACE binding to the RBD of the S protein by the Anti-SARS-CoV-2 Neutralizing Antibody Titer Serologic Assay Kit from Acro Biosystems (Newark, DE) following the manufacturer's instructions. A positive standard with a known concentration provided by the kit was used to extrapolate anti-S antibody concentrations into μg per ml of serum.

1.8. Neutralization assay with S protein pseudotyped VSV vectors

A549 A/T cells were plated into Terasaki plate. The following day, sera were serially diluted in DMEM with 10% FBS. VSV-S vectors were diluted in medium containing 1 μg/ml of a mouse monoclonal antibody against VSV glycoprotein (sc-365019, Santa Cruz Biotechnology, Dallas, TX) and then incubated with the sera for 90 min at room temperature. 5 μl aliquots of the mixtures were transferred onto Terasaki plate wells with A549 A/T cells. Each serum dilution was tested in duplicates, an additional 4–6 wells were treated with VSV-S that had been incubated with medium rather than serum. Plates were incubated for 48 hours and then numbers of fluorescent cells were counted. Titers were set as the last serum dilution that reduced numbers of green-fluorescent cells by at least 50%.

1.9. ELISA for N protein-specific antibodies

Sera were tested for antibodies to the N protein by an ELISA on plates coated with 1 μg of a purified N protein (Abbexa, abx163973) using the same procedures as for antibodies to the S protein. A monoclonal antibody to N protein (Abbexa, MCA6373, lot# 156962) served as a positive control (data not shown).

1.10. T cell assays

PBMCs were purified by Ficoll® Plaque Plus (GE Healthcare, Chicago, IL) gradient centrifugation for 30min at 2800 rpm. Cells were washed and seeded into 96-roundbottom well plates (0.2–1 × 10⁶ cells per well). Lymphocytes were stimulated with three pools of peptides representing the N sequence. Peptides were 15 amino acids in length and overlapped by 10 amino acids with the adjacent peptides. For stimulation ∼10⁶ lymphocytes plated in medium containing 2% fetal calf serum and 1.5 μl/ml Golgiplug (BD Bioscience; San Jose, CA) were cultured with peptides each present at a final concentration of 2 μg/ml for 5 h at 37 °C in a 5% CO₂ incubator. Control cells were cultured without peptides. Following stimulation cells were incubated with anti-CD8b-PE (clone eBio341, eBioscience 53-6.7 and violet live/dead dye (Thermo Fisher Scientific) at 4 °C for 30 min in the dark. Cells were washed once with PBS and then fixed and permeabilized with Cytofix/Cytoperm (BD Biosciences, San Jose, CA) for 20min. Cells were then incubated with an anti–IFN–γ-FITC antibody (clone XMG1.2 BioLegend, San Diego, CA) at 4 °C for 30 min in the dark. Cells were washed and fixed in 1:3 dilution of BD Cytofix fixation buffer (BD Pharmingen, San Diego CA). They were analyzed by a BD FACS Celesta (BD Biosciences, San Jose, CA) and DiVa software. Post-acquisition analyses were performed with FlowJo (TreeStar, Ashland, OR). Data shown in graph represent % of IFN-γ production by CD8⁺ cells upon peptide stimulation. Background values obtained for the same cells cultured without peptide(s) were subtracted.

1.11. Genomic mRNA PCR assay and sub-genomic (sg)mRNA assay

Viral RNA from oral swabs was isolated with the Qiagen MinElute virus spin kit (cat. no. 57704). Tissue RNA was extracted with RNA-STAT 60 (Tel-test”B″)/chloroform, precipitated and resuspended in AVE Buffer (Qiagen 1020953). The qRT-PCR assay utilizes reagents that bind to a conserved region of the N gene of SARS-CoV-2. The amount of RNA was determined by O.D. reading at 260, numbers of copies were calculated and 10⁸ copies per reaction were amplified. The mRNA PCR assay samples were amplified with TaqMan RT-PCR kit (Bioline cat# BIO-78005). Serial dilutions of a standard were included. Samples were amplified in an Applied Biosystems 7500 Sequence detector at 48 °C for 30 min, 95 °C for 10 min followed by 40 cycles of 95 °C for 15 s, and 1 min at 55 °C.

Primers/probe sequences:

2019-nCoV_N1-F:5′-GAC CCC AAA ATC AGC GAA AT-3’.

2019-nCoV_N1-R: 5′-TCT GGT TAC TGC CAG TTG AAT CTG-3’.

2019-nCoV_N1-P: 5′-FAM-ACC CCG CAT TAC GTT TGG TGG ACC-BHQ1-3’.

The sgmRNA assays used the TaqMan RT-PCR kit (Bioline #BIO-78005 and was performed with an Applied Biosystems 7500 Sequence detector at 48 °C for 30 min, 95 °C for 10 min followed by 40 cycles of 95 °C for 15 s, and 1 min at 55 °C.

Primers/probe sequences:

sg-N-F: 5′-CGATCTCTTGTAGATCTGTTCTC-3’.

sg-N-R: 5′-GGTGAACCAAGACGCAGTAT-3’.

Sg-N-P: 5′-6-FAM/TAACCAGAA/ZEN/TGGAGAACGCAGTGGG/3IABkFQ/

1.12. Tissue viral burden by TCID₅₀

Infectious titers of SARS-CoV-2 were measured on Vero TMPRSS2 cells using serial dilutions of strained supernatants from cleared homogenized tissues. After incubation for 4 days, wells were visually inspected for cytopathic lesions. TCID₅₀ of all samples were calculated using the Read-Muench formula. All viral assays were conducted using the right lung lobes of each animal.

1.13. Lung histology

The entire left lobe of each lung was placed in 10% neutral buffered formalin for histopathologic analysis. Lung was processed to hematoxylin and eosin (H&E) stained slides that crossed the entire lobe. One slide for each animal was examined by a board-certified pathologist at Experimental Pathology Laboratories, Inc. (EPL®) in Sterling, Virginia. Findings were graded from one to five, depending upon severity as follows: The severity of tissue lesions are graded as follows: Grade 1 (Minimal) corresponds to a histopathologic change ranging from inconspicuous to barely noticeable. This grade was used for processes where less than approximately 10% of the tissue in an average high-power field was involved. Grade 2 (Mild) corresponds to a histopathologic change that is a noticeable but not a prominent feature of the tissue. This grade was used for processes where between approximately 10% and 25% of the tissue in an average high-power field was involved. Grade 3 (Moderate) corresponds to a histopathologic change that is a prominent but not a dominant feature of the tissue. This grade was used for processes where between approximately 25% and 50% of the tissue in an average high-power field was involved. Grade 4 (Marked) corresponds to a histopathologic change that is a dominant but not an overwhelming feature of the tissue. This grade was used for processes where between approximately 50% and 95% of the tissue in an average high-power field was involved. Grade 5 (Severe) corresponds to a histopathologic change that is an overwhelming feature of the tissue. This grade was used for processes where greater than approximately 95% of the tissue in an average high-power field was involved.

1.14. Statistical analyses

Data were analyzed by 2-or 1-way ANOVA or Kruskal–Wallis test with Dunn correction for multiple comparisons. Correlations were carried out by one-tailed Spearmanʼs rank correlation tests. P values equal or less than 0.05 were considered significant. The analyses were carried out by GraphPad Prism.

2. Results

2.1. Experimental design

Sixteen female and sixteen male Syrian golden hamsters were separated into 4 groups of 8 animals each (4 females, 4 males/group). Hamsters were primed by i.m. injections: group 1 was injected with 1 × 10¹⁰ vp of AdC7-gDN, group 2 with 1 × 10¹⁰ vp of AdC7-S, group 3 with a mixture of 0.5 × 10¹⁰ vp AdC7-gDN and 0.5 × 10¹⁰ AdC7-S and group 4 with 1 × 10¹⁰ vp of the AdC7-HBV2 control. We only used half of the vaccine dose for each component of the mixture to keep the total amount of vector that was given to each animal constant, as in a clinical setting higher vaccine doses increase adverse events. Animals were boosted 2 months after priming with AdC6 vectors expressing the same inserts and used at the same doses. Hamsters were bled at baseline, 14 days after the prime and the boost to determine SARS-CoV-2 S protein-specific antibody responses in groups 2–4. N-specific antibody and CD8⁺ T cell responses in groups 1, 3, and 4 were measured from blood 14 days after the boost (Fig. 1 A). Animals were challenged intranasally with SARS-CoV-2 4 weeks after the boost. Animals were then checked daily for symptoms. Oral swabs were collected on days 2, 4, 7 and 14 after viral challenge to determine viral loads. Four animals in each group were euthanized 4 days and the others 14 days after challenge. After euthanasia, lung viral titers were determined, and lung sections were screened for pathology.

Fig. 1.

Immune responses to the vaccine vectors. [A] Experimental design. [B-H] Results for female animals are shown as circles, those for males are shown as squares. [B,C] Sera harvested at baseline, 2 weeks after the prime or the boost were tested for antibody responses. [B] Reactivity against SARS-Co-V2 S1/S2 proteins tested for by an ELISA. Data show AUC for dilution curves generated with sera of individual animals. Control animals scored negative with adsorbance values below background and these data are shown as an AUC of 0.1. Lines show geometric means (GM). Significant differences were calculated using two-Way repeated measures ANOVA with Tukey correction for multiple comparisons between time points by group. In this and all subsequent graphs lines with stars above indicated significant differences: (∗) p-value between 0.01 and 0.05, (∗∗) p-value between 0.001 and 0.01., (∗∗∗) p-value between 0.0001 and 0.001, (∗∗∗∗) p-value <0.0001. For the statistical analysis data from males and females were combined. [C] Sera were tested by a neutralization assay using VSV-S vectors pseudotyped with the same S protein as expressed by the vaccines. Data are shown for individual animals. Negatives are shown as a titer of 10. Lines indicate GMs. Lines with stars above show significant differences by two-Way repeated measures ANOVA with Tukey correction for multiple comparisons between time points by group, and between groups by time point. [D] Sera harvested after the boost were tested for inhibition of ACE binding to the RBD of S1 by an ELISA. Data are shown for individual animals as μg of antibody/ml calculated based on an internal standard. Negatives are shown as a titers of 1 μg/ml. Lines indicate GMs. Significant differences were calculated by one-way ANOVA with Tukey's correction. [E,F] Cross-reactivity against SARS-Co-V2 variants was assessed with sera collected after the prime [E] or the boost [F]. They were tested for neutralization of VSV-S vectors pseudotyped with the indicated S protein variants. Group 2 [light grey circles] and group 3 [dark grey squares] are indicated by different symbols. For the statistical analysis by one-way ANOVA with Dunnett's correction data for the two groups were combined. [G] Sera harvested 2 weeks after the boost were tested for N protein-specific antibodies by an ELISA. Results are shown as AUC for individual sera from female and male hamsters. Lines show GMs. Significant differences were calculated by Kruskal Wallis test with Dunn's multiple comparisons test. [H] Frequencies of CD8⁺ T cells in blood producing IFN-γ in response to N protein-derived peptides. Graph shows sum of responses to the 3 peptide pools. Lines show median responses. Lines with stars above show significant differences between animals that received the COVID vaccine and control animals by Kruskal Wallis test with Dunn's multiple comparisons test. [I] Proportions of N-specific CD8⁺ T cell responses to the 3 peptide pools.

2.2. Immune responses to the vaccine vectors

Sera from groups 2, 3, and 4 hamsters harvested at baseline and at 2 weeks after the prime or boost were tested by ELISA on plates coated with S1 and S2 proteins. They were also tested in a neutralization assays using vesicular stomatitis virus (VSV) vectors pseudo-typed with SARS-CoV-2 S protein. Sera harvested 2 weeks after the boost were in addition tested by an ELISA, which measures antibodies specific to the RBD of S1 (Fig. 1A). All pre-immunization samples and samples from the control groups point scored negative in either one of these assays. All but one hamster of group 3 scored positive by the S1/S2-specific ELISA by 2 weeks after the prime. All hamsters seroconverted after the boost and overall responses showed significant increases (Fig. 1B): geometric mean (GM) titers of group 2 females increased 1.6-fold, while those of males increased 6.8-fold; GM titers of group 3 females increased 1.5-fold while those of males increased 1.9-fold. There were no significant differences in responses of females compared to males. All hamsters developed antibodies that neutralized the pseudotyped VSV-S virus with GM titers of 109 in group 2 and 167 in group 3 after the prime, which changed to 166 and 365 after the boost, respectively. The superior booster response in group 3 was driven by female hamsters, which showed a 4-fold increase compared to the 1.2-fold increase in males (Fig. 1C). After the boost all animals scored positive in the RBD-specific ELISA (Fig. 1D) and there were no significant differences between females and males or groups 2 and 3.

Sera were tested for reactivity against alpha, beta, delta, and omicron variants. Sera collected after the prime (Fig. 1E) or after the boost (Fig. 1F) showed significantly reduced titers comparing neutralization of VSV-S vectors expressing wild-type S protein to those with S proteins of variants.

Sera from hamsters of groups 1, 3, and 4 harvested after the boost were tested for antibodies to the N protein. All AdC-gDN vaccinated animals scored positive with no differences between groups 1 and 3 or female and male hamsters (Fig. 1G).

Peripheral blood lymphocytes were tested two weeks after the boost for production of IFN-γ by CD8⁺ T cells in response to three peptide pools representing the N protein sequence. Most animals of groups 1 and 3 developed N protein specific CD8⁺ T cell responses (Fig. 1H). Responses showed no preference to any of the 3 peptide pools representing different segments of the N protein (Fig. 1I).

In summary, animals mounted robust immune responses to the expressed SARS-CoV-2 antigens following vaccination and as far as tested most types of responses increased after the boost.

2.3. Clinical symptoms and weight loss after SARS-CoV-2 challenge

After challenge hamsters were scored twice daily for clinical COVID-19 symptoms (Fig. 2 A). Most animals developed benign symptoms such as mildly ruffled fur, hunched over posture and/or closed or squinted eyes. None of the vaccinated hamsters exceeded a COVID-19 disease score of 1. One female in the control group 4 was given a disease score of 2 on days 5–7 after challenge. One male animal in the control group died 7 days after challenge. His lung pathology suggested that his death was most likely caused by SARS-CoV-2. All female animals developed symptoms by day 4 after challenge. All males of groups 2 and 4 and 3 animals of group 4 had COVID-19 symptoms by day 2. In males of group 3 symptoms were not detected till day 4. Thereafter the disease course was markedly different for vaccinated female and male animals as has been reported for humans [30]. With some fluctuations, half of the females of group 1 and 2 remained symptomatic throughout the observation period while group 3 females became and remained symptom-free by days 5 or 6 (Fig. 2B). Control hamsters of either gender as well as males of the vaccine groups exhibited symptoms until the day of their euthanasia. Although these results are based on small numbers of animals, they suggest that vaccinated females recover more rapidly than vaccinated males upon breakthrough infections and that the combination vaccine compared to the individual vaccines accelerates recovery in females.

Fig. 2.

Disease score and weight loss after challenge. [A] Experimental design. [B] Disease score over time after challenge in female (top) and male (bottom) animals of groups 1 to 4 (left to right). Data are shown for individual animals with bars indicating means. [C] Weight loss after challenge is shown as % weigh reduction over the weight on the day of challenge. Data are shown as means ± SEM. Significant differences were compared by One-Way ANOVA with Tukey correction for day 4 and 14 data separately, which are indicated by connecting lines next to the legend. Day 4 comparisons are shown first followed by/and then day 14 comparisons. [D] Median days with maximum weight loss. Significant differences were calculated using one-Way ANOVA with Tukey correction.

These data were mirrored by levels of weight loss. Control and group 1 animals rapidly started losing weight after challenge (Fig. 2C) with maximal weight loss around days 6–7 (Fig. 2D). Weight loss in group 1 animals tended to be lower than that of the controls. Hamsters of groups 2 and 3 lost only a minimal amount of weight and started regaining weight by day 2 or 3 after challenge. Animals in groups 2 and 3 showed significantly less weight loss compared to group 4 animals when tested on days 4 or 14. Group 3 animals showed significant differences to group 1 animals on days 4 and 14 while group 2 animals only differed from group 1 animals on day 4 after challenge. Females recovered their weight 1–2 days earlier than males (Fig. S1A). In summary, the AdC-S vaccines given alone or in combination with the AdC-gDN vaccines offered significant protection against symptoms and weight loss. In female hamsters addition of the AdC-gDN to the AdC-S vaccines accelerated recovery from symptomatic disease.

2.4. Effect of vaccination on viral loads and lung histology

Viral titers were determined from oral swabs collected on days 2, 4 (8 animals for each group), 7, and 14 (4 animals for each group) and from lung tissues collected at euthanasia on days 4 or 14 after challenge (Fig. 3 A). Samples were screened for viral RNA and sgRNA, the latter has been suggested to determine titers of infectious virus more accurately [31]. Immunization with the vaccine mixture reduced oral viral RNA titers (Fig. 3B) while both the AdC-S vaccines and the vaccine mixture caused significant reductions in sgRNA titers (Fig. 3C) titers on day 2 after infection. No significant differences were observed for other time points. Vaccination with either the AdC-S or the AdC-S + AdC-gDN regimens reduced lung viral RNA titers (Fig. 3D) on day 4 after challenge. AdC-S vaccination also reduced sgRNA titers at that time point (Fig. 3E). By day 14 viral loads in lungs were markedly reduced in all groups and although group 3 continued to show lower RNA and sgRNA titers compared to the others this failed to reach significance. To ensure that titers based on viral genomes reflected presence of infectious virus, lungs and nares of the early euthanasia groups were tested 4 days after challenge for the TCID₅₀ of SARS-CoV-2 (Fig. 3F and G). TCID₅₀ in nasal tissues showed a ∼2 log₁₀ reduction in groups 2 and 3 as compared to the control group 4 (Fig. 3F) while lungs of groups 2 and 3 showed a 3 or 2 log₁₀ reduction, respectively (Fig. 3G). One group 2 animal (male) and two group 3 animals (1 female, 1 male) no longer harbored infectious virus in their nares by day 4 after challenge. Complete clearance from lungs was observed in the two group 2 males, one of which also had undetectable titers in nares. Although these data do not allow us to conclude that the vaccine-induced immune responses completely prevented infection in some animals, they confirm that vaccination not only protected against disease but also blunted viral spread and accelerated clearance.

Fig. 3.

Viral loads and lung pathology. [A] Experimental design. [B,C] Viral RNA [B] and sgRNA [C] loads in oral swabs collected after challenge on days 2 and 4 for all animals and on days 7 and 14 for the day 14 euthanasia group animals. Significant difference between groups for each time point were calculated by two-Way repeated measures ANOVA with Tukey correction; they are indicated with lines and stars above as in legend to Fig. 1. [D,E] Lung viral RNA [D] and sgRNA [E] titers are shown for individual hamsters of the day 4 and day 14 euthanasia groups. Differences were calculated by two-Way ANOVA with Tukey correction, with lines and stars showing significant differences as in legend to Fig. 1. [F,G] SARS-CoV-2 TCID₅₀ titers in nares [F] and lungs [G] were determine on day 4 after challenge. Data were analyzed by an uncorrected Kruskal–Wallis test. [H] Sum of lung lesions for the day 4 and day 14 euthanasia group. Differences were calculated by two-Way ANOVA with Tukey correction, with lines and stars above as in legend to Fig. 1. [I] Severity of the different types of lesions according to gender of the animals.

Lungs were analyzed for pathological changes (Fig. 3H). Macroscopically lungs showed red or dark red patches, dark spotted, mottled/red lobes in groups 1, 2, and 4 but not 3 (Fig. S1B). Tissue lesions were graded microscopically and lungs from most animals showed grade 1–3 histopathology. Especially on day 4, lungs of control animals showed perivascular edema, alveolar hemorrhage, bronchiolo-alveolar hyperplasia, mixed or mononuclear cell inflammation (bronchoalveolar, alveolar, and/or interstitial), mononuclear vascular/perivascular inflammation, mesothelial hypertrophy, and/or pleural fibrosis (Fig. 3H and I). In groups 2 and 3 lesions were milder and more multifocal than in groups 1 and 4 on day 4. The sum of the score for the individual symptoms showed on day 4 significant differences between groups 2 and 3 compared to the animals of groups 1 and 4. On day 14 lungs of the control animals continued to show marked pathology while group 3 animal lungs no longer showed any lesions; group 2 animals showed residual mild disease. Group 1 male animals, which had only received the AdC-gDN vaccine, also showed reduced pathology compared to group 4 animals (Fig. 3H). Lung pathology showed gender specific differences (Fig. 3I). By day 4 after infection the two group 3 females only showed grade 1 hyperplasia while the 2 males of this group had additional lesions. By day 14 both females of the control group showed a lessening of symptoms, which was not observed in the surviving male.

In summary, the AdC-S vectors given alone or in combination of with the AdC-gDN vectors significantly decreased viral loads and reduced lung pathology and the latter was more pronounced upon immunization of animals with the combination vaccine.

2.5. Correlations

We analyzed the data for correlations by Spearman separately for hamsters that were tested for N protein specific T and B cell responses (group 1, 3, and 4) or S protein specific antibody responses (groups 2, 3, and 4). Data for the early day 4 euthanasia groups were analyzed separately from those of the late day 14 euthanasia groups (Fig. 4 ). As expected, in all groups viral titers measured by the different methods showed positive correlations with each other and with disease parameters, i.e., weight loss and lung pathology. Immune responses showed inverse correlations between viral loads and disease parameters. For groups 1, 3, and 4 correlations were more pronounced for the late euthanasia groups where both N protein specific T and B cell responses showed significant inverse correlations with weight loss and levels of lung pathology. For groups 2, 3, and 4 S protein specific antibody responses after the boost showed significant inverse correlations not only for the two tested disease markers but also for viral loads. The data indicate that both types of immune responses, i.e., antibodies to the S or N proteins and N protein specific CD8⁺ T cells reduce viral loads and severity of disease.

Fig. 4.

Correlations between antibodies to N proteins and disease parameters and T cell responses. The graphs show as heatmaps the R-values of correlations by Spearman between the indicated parameters. Red squares show positive R values, blue squares show negative R values as indicated in the legend. Yellow numbers of R within colored squares are only shown for R values with p-values ≤0.05. [A] Analysis for the early euthanasia animals of groups 1, 3, and 4. [B] Analysis for the early euthanasia animals of groups 2, 3, and 4. [C] Analysis for the late euthanasia animals of groups 1, 3, and 4. [D] Analysis for the late euthanasia animals of groups 2, 3, and 4.

3. Discussion

In a remarkable achievement, SARS-CoV-2 vaccines were developed and approved for use in humans within less than a year after onset of the COVID-19 pandemic [[2], [3], [4], [5], [6], [7], [8]]. Nevertheless, even in countries with easy access to the vaccines the pandemic has been a roller coaster ride with cycles of case reductions followed by rapid increases due to the emergence and spread of more transmissible variants. Available COVID-19 vaccines protect well against disease but fail to prevent infections, which allows for spread of the virus even in populations with high vaccine coverage [13,32]. In addition, protection wanes after a few months. Resistance to COVID-19 disease can be restored at least temporarily by a booster immunization but after for example a third dose of an RNA vaccine [33], protection declines after four months [33,34], and it is currently unknown if frequent boosts with such genetic vaccines within short intervals will continue to restore antibody titers and protect against new variants that progressively more escape vaccine-induced VNAs.

SARS-CoV-2 has thus far mainly accumulated mutations within the S protein. As other viral proteins are less variable, we decided to explore vaccines that express the N protein in addition to the S protein. The N protein is a target for non-neutralizing antibodies that might play a role in combating a SARS-CoV-2 infection [35]. The N protein also induces CD8⁺ T cells that by rapidly killing cells early after they become infected have been implicated to contribute to protection against severe COVID-19 or to improve resistance in individuals with suboptimal VNA titers [10,36,37].

For the hamster challenge study, we used two AdC vectors that belong to distinct serotypes to prevent that Ad capsid-specific antibody responses induced by the prime neutralize the boosting Ad vector. Humans rarely have VNAs to AdC viruses and those who are seropositive have low titers [38]. VNAs to Ad viruses can prevent infection of cells with Ad vectors and thereby expression of the transgene product [38]. This in turn reduces the antigenic load and blunts immune responses to the vaccine antigen, which can be problematic for human serotype Ad vectors or for the repeated use of the same Ad vector [38,39]. The AdC vectors for the S protein carry the full-length unmodified gene from an early SARS-CoV-2 isolate and as we showed earlier induce after a boost sustained antibody responses in mice [40]. These data were recapitulated in hamsters, which developed binding as well as neutralizing antibodies after vaccination with the AdC-S vectors. A booster effect was seen for binding antibodies for groups 2 and 3 while VNA responses only increased in group 3 animals that received the AdC-S/AdC-gDN vector mixtures. The discrepancy in the booster responses using the two different types of assays could potentially reflect that affinity maturation after the boost may have increased responses measured by ELISA but not by the biological assay or alternatively that increases were mainly seen for antibodies to non-neutralizing epitopes. The better response of group 3 to the boost could reflect increased T help due to responses to the N protein or a difference in vector dosing with group 2 hamsters receiving double the dose of the AdC-S vector. As also reported by others, VNAs cross-reacted albeit with significantly reduced activity against viral variants of concern [41,42].

The N protein is expressed by the AdC vectors as a fusion protein within HSV-gD, which as an inhibitor of the early BTLA-HVEM T cell checkpoint enhances B cell responses and broadens CD8⁺ T cell responses as we showed with several antigens including the SARS-CoV-2 N protein in mice [[25], [26], [27],40]. Expressing the N protein by an AdC vector within a heterologous viral protein that is expressed on the surface of an infected cells may promote B cell responses to the N protein's linear epitopes [43] but could potentially distort the proteins’ structure and destroy conformation-dependent epitopes. Nevertheless, we obtained strong N protein-specific antibody responses after the boost in hamsters that received the AdC-gDN vaccines.

The N protein as expressed by the AdC vectors induces potent and broad CD8⁺ T cell responses in mice [40]. Responses are readily detectable but modest in hamsters, which may in part relate to the paucity of hamster-specific antibodies to T cell determinants and key cytokines, which only allowed us to test for circulating CD8⁺ lymphocytes producing IFN-γ in response to the N peptide pools. Hamsters were not tested for T cell responses to the S protein.

Hamsters were challenged after vaccination to assess if the vaccines protected against disease or infection. Mirroring humans, male hamsters were more susceptible to severe disease than female hamsters – one male hamster in the control group died and all male hamsters regardless of vaccination exhibited clinical symptoms throughout the 14-day observation period. S protein-immune female hamsters showed accelerated recovery from symptoms especially if they had also received the AdC-gDN vaccine. Both the combination vaccine and the AdC-S vaccine provided solid protection against weight loss. Despite reducing symptoms none of the vaccine regimens appeared to prevent infection; in all animals viral RNA and sgRNA were recovered at high titers from their oral cavities with only minor reductions in groups 2 and 3 on day 2 after challenge. Results obtained by testing for genetic material appeared to at least in part reflect detection of non-infectious viral material as testing of lungs and nasal tissue harvested on day 4 after challenge by an infectivity assay showed >3 or 2 log₁₀ reductions in titers in group 2 and 3 animals, respectively. Lung pathology mirrored lung viral titers. Females tended to have less lung disease compared to males. By 14 days after challenge, group 3 animals no longer showed any lung lesions while half of the group 2 animals exhibited residual hyperplasia. Overall, these data show that although AdC-gDN vectors by themselves induce marginal protection their addition to AdC-S vectors increases the vaccines’ effectiveness. This contrasts data from another group; they reported T cell-mediated protection against COVID-19 disease and a reduction in viral loads upon intravenous immunization of hamsters with a human serotype 5 Ad vector expressing the N protein [19]. Differences in results may reflect the distinct immunization protocols or differences in the challenge.

In our study, the contribution of the AdC-gDN vaccine to protection induced by the AdC-S vaccines was not solely mediated by T cells, which have the advantage that they persist for very long periods of time upon antigen-driven activation [44], which would likely extend duration of protection well beyond the 4–6 months that current vaccines offer. Antibodies to the N protein had strong inverse correlations with disease parameters such as weight loss, viral loads and lung pathology in the late euthanasia group indicating that the N protein-specific antibodies contribute to the vaccines’ efficacy. Additional studies will be needed to assess the efficacy of the multivalent AdC vaccines against SARS-CoV-2 variants.

Although available vaccines still avert serious COVID-19 disease and death, considering that we are now in the 3rd year of the pandemic with the near certainty that additional variants with even more mutations, which will further escape VNAs induced by current vaccines, will evolve, new vaccines expressing additional SARS-CoV-2 antigens, such as the N protein to augment VNA mediated protection through additional immune mechanisms should be explored.

Author contributions

Conceptualization: KS, HE; Methodology: MH, MN, RA, AC, DN, ZX, XZ, KS, HE; Investigation: MH, MN, RA, AC, DN, JD, WG, ZX, XZ, KS, HE; Supervision: KS, HE; Writing – original draft: HE; Writing – review & editing: MH, MN, RA, AC, DN, JD, WG, ZX, XZ, KS, HE; Statistics: QL, JD.

Funding

This work was funded by grants from The G. Harold and Leila Y. Mathers Charitable Foundation, the Commonwealth of Pennsylvania, and the Wistar Science Discovery Fund. MH was the recipient of a Fellowship from Janssen Scientific Affairs. Support for Shared Resources utilized in this study was provided by Cancer Center Support Grant (CCSG) P30CA010815 to The Wistar Institute.

Data and materials availability

All data are available in the main text or the supplementary materials.

Declaration of competing interest

HCJE holds equity in Virion Therapeutics. She serves as a Consultant to several Gene Therapy companies.

Appendix A. Supplementary data

The following is the Supplementary data to this article: