BCG administration promotes the long-term protection afforded by a single-dose intranasal adenovirus-based SARS-CoV-2 vaccine

Summary

Recent publications have explored intranasal (i.n.) adenovirus-based (Ad) vaccines as an effective strategy for SARS-CoV-2 in pre-clinical models. However, the effects of prior immunizations and infections have yet to be considered. Here, we investigate the immunomodulatory effects of Mycobacterium bovis BCG pre-immunization followed by vaccination with an S-protein-expressing i.n. Ad, termed Ad(Spike). While i.n. Ad(Spike) retains some protective effect after 6 months, a single administration of BCG-Danish prior to Ad(Spike) potentiates its ability to control viral replication of the B.1.351 SARS-CoV-2 variant within the respiratory tract. Though BCG-Danish did not affect Ad(Spike)-generated humoral immunity, it promoted the generation of cytotoxic/Th1 responses over suppressive FoxP3⁺ T_REG cells in the lungs of infected mice. Thus, this vaccination strategy may prove useful in limiting future pandemics by potentiating the long-term efficacy of mucosal vaccines within the context of the widely distributed BCG vaccine.

Graphical abstract

Highlights

• •SARS-CoV-2 intranasal adenovirus vaccine immunity was found to wane over six months
• •BCG-primed mice had enhanced viral clearance/vaccine immunity after COVID-19 vaccination
• •BCG promotes COVID-19 vaccine-specific cytotoxic/T_H1 responses over T_REG responses
• •BCG is safe, widely available, and has the potential to synergize with other vaccines

Biological sciences; Immunology; Immune response; Microbiology

Introduction

The COVID-19 pandemic has resulted in over 766 million cases and over 6.9 million deaths as of May 2023.¹ Infection with SARS-CoV-2 can cause a broad spectrum of disease ranging from mild symptoms to severe lung injury and multi-organ failure, potentially leading to death, especially in the elderly and those with comorbidities.² Additionally, there is evidence that recovered individuals can experience symptoms termed “long COVID,” which can involve multiple organ systems (e.g., lung, heart, kidneys, liver, etc.).³ Consequently, COVID-19 has had severe consequences on global health and the economy and has been the target of novel immunization approaches. Vaccination, in combination with public health measures, has effectively slowed the progression and hospitalization rates associated with SARS-CoV-2⁴; however, these immunization strategies have so far failed to fully prevent viral transmission and infection.⁵ Moreover, vaccine efficacy at preventing infection declines by six months after full vaccination,⁶ and breakthrough infections, especially by novel SARS-CoV-2 variants, have been reported in previously vaccinated individuals.⁷ In addition, there is an urgent need to provide equitable access to affordable and effective vaccines amongst developing nations to prevent the ongoing morbidity and mortality as well as reduce the risk of novel variants emerging.

Among the strategies in preclinical development, a promising approach involves the use of intranasal (i.n.) vaccines, theoretically capable of eliciting local mucosal immune responses within the respiratory tract that can effectively neutralize SARS-CoV-2 entry and prevent viral replication at the site of initial infection.⁸ However, while some vaccines are undergoing phase I/II clinical trials, the success of this approach has thus far been elusive.⁹ For example, i.n. administration of a chimpanzee adenovirus-based vaccine carrying the Spike protein (ChAd-SARS-CoV-2-S) successfully generated neutralizing IgA antibodies and T cell responses in the lungs of hACE-2 mice¹⁰ and provides at least one month of protection in mice and rhesus macaques.¹¹ However, a recent progress report from a phase I clinical trial of the i.n. administration of ChadOx1 has announced a failure to substantially increase mucosal immunity in humans.¹² Nonetheless, the safety results were sufficiently encouraging to pursue approaches that enhance mucosal immunity generated using adenovirus vaccine vectors.

In many countries, the live-attenuated Mycobacterium bovis Bacillus Calmette-Guérin (BCG) vaccine is administered intradermally to newborns soon after birth to protect them from disseminated Mycobacterium tuberculosis infection. In fact, most individuals worldwide are BCG vaccinated; of the approximately 140 million babies born per year, approximately 100 million of them are vaccinated with BCG.¹³ Curiously, vaccination with BCG has also been reported to reduce child mortality¹⁴ from neonatal sepsis and lower respiratory tract infections.¹⁵ There is cumulating evidence that BCG acts on the antiviral immune response by boosting the activity of innate immune cells, a concept known as trained immunity,¹⁶ as well as by promoting heterologous T cell activation.¹⁷ As such, many researchers have speculated on the possibility that BCG could provide immunity against SARS-CoV-2 infection,¹⁸ although there are limited experimental data thus far to support this idea.^19,20 A multi-country randomized trial with over 3,000 healthcare workers compared BCG vaccination to placebo, but it did not show a lower COVID-19 risk in the BCG group.²¹ Still, there remains great promise that BCG can be combined with conventional SARS-CoV-2 vaccination strategies to improve their effectiveness.²² As such, we hypothesized that BCG may provide a novel, cost-effective, and safe priming strategy to enhance the long-term efficacy of i.n. immunization with adenovirus-based COVID-19 vaccines.

Here, we sought to evaluate the effect of prior administration of BCG on the immunogenicity and efficacy of an i.n. human adenovirus (AdV) serotype 5 vaccine expressing the SARS-CoV-2 S (Spike)-protein, referred to as Ad(Spike), in a C57BL/6 mouse model of infection with the B.1.351 variant of SARS-CoV-2. Previous work has shown that a single i.n. immunization with adenovirus-based vaccines against SARS-CoV-2 is sufficient to provide effective protection from infection in naive animals.²³ In this report, we demonstrate that the protection conferred by our intranasally administered Ad(Spike) vaccine alone declines in mice by 6 months post-immunization. Importantly, a single dose of BCG administered systemically prior to Ad(Spike) vaccination was capable of boosting the Spike-specific cytotoxic T cell response in the lungs and, in doing so, significantly decreased the replication and production of infectious viral particles up to 6 months after i.n. immunization with the Ad(Spike) vaccine. This approach is analogous to previous “prime-pull” vaccination strategies that employ a conventional vaccine to elicit systemic T cell responses (“prime”), followed by the recruitment of these activated T cells and associated cytokines to the desired body site via the administration of a localized, secondary vaccine (“pull”).²⁴ However, in this instance, we attempted to exploit the non-specific or off-target immune enhancing properties of BCG as the priming mechanism. Overall, the results herein provide proof-of-concept data that highlight a novel and innovative approach to the use of heterologous vaccine strategies incorporating the widely approved BCG vaccine to protect against current and future viral pandemics.

Results

The in vivo protection provided by a single dose of intranasal recombinant Ad(Spike) attenuates with time

To investigate the long-term protection provided by i.n. vaccine administration, we developed a replication-deficient (ΔE1-, ΔE3-) human adenovirus serotype 5 expressing the full-length S-protein of the ancestral strain of SARS-Co-V-2 (Ad(Spike)) which was codon optimized for expression in human and mouse cell lines (Figure 1A). Exogenous gene expression was verified in vitro by western blot using S-protein specific antibodies (Figure 1B). Female C57BL/6 mice were vaccinated with either PBS or 10⁹ TCID₅₀ Ad(Spike) i.n. The infectious dose of Ad(Spike) was determined using a pilot study and was chosen based on the induction of Spike-specific IgA antibodies in lung homogenates from vaccinated animals (Figure S1). Study animals were challenged 2 or 6 months later with the Beta variant (B.1.351) of SARS-CoV-2. In each case, the animals were followed for 5 days post-infection (dpi) to determine viral shedding from oral swabs (3 dpi) and the viral load in the lungs at 5 dpi (Figure 1C).

Figure 1.

The protection provided by a single dose of intranasal recombinant Ad(Spike) attenuates with time in C57BL/6 mice

(A) Diagram of the SARS-CoV-2 Spike protein (ancestral strain) transgene cassette expressed in our recombinant ΔE1/E3 human adenovirus serotype 5, termed Ad(Spike).

(B) Western blot analyses of antigen expression from SF-BMAd-R cells infected with Ad(Spike). Samples were run against a molecular weight ladder; lanes 1 and 2. The negative control contains SF-BMAd-R cells infected with a ΔE1/E3 adenovirus with an empty gene cassette, termed Ad(e); lane 3. Ad(Spike) was amplified in 3 batches of cells and western blots were run on cell lysates from each batch; lanes 4–6, before being combined and purified. This western blot was run with an RBD-specific antibody, thereby capturing the S1 portion.

(C) Study design schematic. At time 0, animals were vaccinated intranasally (i.n.) with 10⁹ mean tissue culture infectious dose (TCID₅₀) of Ad(Spike) in 30 μL. In the case of the sham control, at time 0 animals were vaccinated i.n. with 30 μL PBS. Mice were then challenged with 10⁶ TCID₅₀ SARS-CoV-2 South African strain (B.1.351) at month 2 or month 6 post-vaccination. In both challenge models, animals were followed for 5 days with nasal swab collection on day 3 and euthanasia on day 5.

(D–F) Infectious viral load in mice challenged with the B.1.351 variant of SARS-CoV-2, 2 and 6 months post-immunization with Ad(Spike). Viral load (TCID₅₀) in (D) oral swabs at 3 days post-infection (dpi), and (E) lungs at 5 dpi, quantified by the Spearman–Kärber method. Viral RNA in mouse (F) lungs at 5 dpi. N = 6. Data points represent individual mice, means ± SD are shown. For (D)–(F), Kruskal-Wallis test with Dunn’s multiple comparisons: ∗p < 0.05; ∗∗p < 0.01; ns = not significant. Schematics made with BioRender.com.

As expected, Ad(Spike)-immunized mice displayed a significant reduction in the production of infectious virus in oral swabs (Figure 1D). Notably, animals challenged 6 months post-vaccination, showed lower viral titers compared to unvaccinated mice, although this did not reach statistical significance (Figure 1D). Similarly, infectious viral titers (Figure 1E) and viral RNA (Figure 1F) assessed directly in the lungs of infected animals at 6 months post-vaccination were not statistically different from unvaccinated mice as opposed to data from mice challenged 2 months after immunization, where significant differences were observed.

BCG administration prolongs the protective effect of Ad(Spike) in immunized mice

We next investigated the effectiveness of a prime-boost vaccination regimen using BCG. Female C57BL/6 mice were pre-immunized with 10⁶ colony forming units (CFU) of the BCG-Danish strain containing an empty plasmid (BCG(e)) intraperitoneally (i.p.), 1 month prior to i.n. Ad(Spike) vaccination (month 0; Figure 2A). Although BCG is routinely administered intradermally in humans, a range of administration routes have been applied to the mouse model including i.n., i.p., aerosol, subcutaneous (s.c.) and intravenous (i.v.).^{19,25,26,27,28,29} For this study, we chose to use the i.p. route for ease of handling and to maximize the systemic distribution of BCG upon a single immunization. In addition, the i.p. route is more immunogenic than the s.c. route and is effective at inducing both systemic and mucosal immunity.^25,30 Naive controls received the vehicle (PBS) in place of both BCG(e) and Ad(Spike). First, we tested the possibility that BCG alone could provide non-specific protection against infection or severe disease following SARS-CoV-2 challenge in our animal model, since reports from human data are controversial.³¹ Animals were pre-immunized with BCG(e) and after one month were vaccinated with an adenoviral vector containing an empty gene cassette (Ad(e)) before challenge with SARS-CoV-2 two months later. Mice pre-immunized with BCG-Danish/Ad(e) displayed no significant reduction in infectious viral titers or viral RNA in oral swabs or lungs (Figure S2) compared to PBS controls, confirming the BCG-Danish vaccine did not provide significant non-specific protection to challenged mice.^19,32

Figure 2.

Exposure to BCG prolongs the protective effect of Ad(Spike) in immunized C57BL/6 mice

(A) Study design schematic. Animals were primed intraperitoneally (i.p.) with 10⁶ cfu BCG containing an empty gene cassette (BCG(e)) or PBS at time −1 month. At time 0, animals were vaccinated intranasally (i.n.) with 10⁹ mean tissue culture infectious dose (TCID₅₀) of Ad(Spike) in 30 μL. In the case of the sham control, at time 0 animals were vaccinated i.n. with 30 μL PBS. Mice were then challenged with 10⁶ TCID₅₀ SARS-CoV-2 South African strain (B.1.351) 6 months post-vaccination. Animals were followed for 5 days post challenge, with nasal swabs collected on days 1, 3, and 5 post-infection (dpi) and euthanasia on day 5.

(B–E) Viral load quantified in oral swabs at 1, 3, and 5 dpi, and (D) in lungs at 5 dpi, quantified by TCID₅₀. Viral RNA in (C) oral swabs at 5 dpi and (E) lungs at 5 dpi. N = 4–6. Data points represent individual mice, means ± SD are shown. For (B), two-way ANOVA with Tukey’s multiple comparisons: ∗p < 0.05; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001. For (C)–(E), Kruskal-Wallis test with Dunn’s multiple comparisons: ∗p < 0.05; ∗∗p < 0.01; ns = not significant. Schematic made with BioRender.com.

We then investigated the effect of prior BCG-Danish administration on the duration of our Ad(Spike) vaccine 6 months after vaccination (Figure 2A) by quantifying the daily variation in viral replication and infectious particles in oral swabs and the pulmonary tissue of mice infected with SARS-CoV-2. Here, we found that while a single dose of i.n. Ad(Spike) could reduce infectious viral particles in oral swabs at 1 and 3 dpi, mean viral titers in animals that were first pre-immunized with BCG were significantly lower than those that were not (Figure 2B). At 5 dpi, we also assessed the reduction of viral RNA in oral swabs (Figure 2C), infectious virus in the lungs measured by TCID₅₀ (Figure 2D), and viral RNA in the lungs (Figure 2E). We found that by 6 months post-vaccination, although a single dose of Ad(Spike) was able to reduce viral burden compared to controls, this reduction was not statistically significant. However, in animals that first received a single administration of BCG-Danish, we did observe a statistically significant protective effect after 6 months, highlighting the fact that BCG is able to potentiate the ability of the i.n. Ad(spike) vaccine to control viral replication within both the upper and lower respiratory tract.

Ad(Spike)-vaccinated animals are protected from SARS-CoV-2-induced histological changes 6 months post-vaccination

SARS-CoV-2 infection in C57BL/6 mice causes pulmonary inflammation characterized by immune cell infiltration, lung atelectasis, and bronchial constriction. Since we observed a greater reduction of viral particle load in BCG-Danish exposed mice, we next assessed the extent of pulmonary damage at 6 months post-Ad(Spike) vaccination, both with and without BCG pre-immunization. Scoring of lung histology showed an overall significant reduction in cellular and tissue damage (CTL), circulatory/vascular damage (CVL), and inflammatory patterns (RIP) in all Ad(Spike) vaccinated animals that was not observably enhanced through pre-immunization with BCG(e) (Figures 3A–3C and Table S1). Thus, we could not distinguish any protective effect of BCG-Danish, as both groups displayed a significant reduction in histological changes within the lung compared to the controls. Collectively, these results confirm that while a single i.n. dose of Ad(Spike) vaccine fails to prevent infection, it does protect against severe SARS-CoV-2-induced lung alterations as long as 6 months post-vaccination.

Figure 3.

SARS-CoV-2 induced pulmonary pathology is well prevented in Ad(Spike) immunized mice, regardless of BCG exposure

(A and B) Heatmap and (B) graphical summary of lung pathology (5 days post infection (dpi)) scored in categories: Fibrosis, CTD, CVL, RIP, RR. CTD: Cell/tissue damage which is comprised of bronchoepithelial necrosis (scored 1–3), inflammatory cells/debris in bronchi (1–3), intraepithelial neutrophils (1–3) alveolar emphysema (Yes = 1/No = 0). CVL: Circulatory/vascular lesions comprised of alveolar hemorrhage (Y/N), significant alveolar edema (Y/N), endothelial/vasculitis (1–3). RIP: Reaction/inflammatory patterns comprised of necrosis/suppurative bronchitis (Y/N), intra-alveolar macrophages (Y/N), mononuclear inflammation around airways (Y/N), neutrophilic/heterophilic inflammation (1–3), mesothelial reaction (1–3). RR: Regeneration/repair, includes alveolar epithelial cell regeneration/proliferation (1–3) and bronchiolar epithelial cells regeneration/proliferation (1–3).

(C) Lungs of vaccinated animals, infected after 6 months, were harvested at 5 dpi and stained with Masson’s trichrome for fibrosis (top row) and H&E staining for pathology scoring (bottom row). Row one is imaged at 100× magnification (scale bar, 100 μm). Row two is imaged at 200× magnification (scale bar, 100 μm) showing airway mononuclear inflammation in control animals. Each image is representative for the group. N = 4–6. Data points represent individual mice, means ± SD are shown. For (B), two-way ANOVA with Tukey’s multiple comparisons: ∗p < 0.05.

Ad(Spike)-induced antibody profiles are not influenced by pre-immunization with BCG

Vaccine protection against SARS-CoV-2 infection is largely attributed to its ability to generate high quantities of neutralizing antibodies, although the potential role of BCG in promoting antibody production remains to be assessed. To determine if BCG influenced the quantity and affinity of the antibody response generated by Ad(Spike), we bled immunized animals at select timepoints pre- and post-vaccination. Antibody titers against the full Spike protein (Figure 4) and the receptor binding domain (RBD) (Figure S3) were assessed via ELISA. As expected, animals in the PBS and BCG(e)+Ad(e) groups did not produce Spike or RBD-specific serum antibodies throughout the study. Spike-specific IgG was detectable at high levels 1 month after vaccination and continued to rise until 2 months after vaccination when levels slowly declined until the end of the study for both Ad(Spike) and BCG(e)+Ad(Spike) groups (Figure 4A). Although RBD-specific IgG was significantly higher in Ad(Spike) animals than those that received BCG(e)+Ad(Spike) at 1 month post-vaccination, this difference was not observed at later time points (Figure S3A). As such, Spike-specific IgG was similar between vaccinated groups.

Figure 4.

Exposure to BCG does not significantly influence the quantity or quality of Ad(Spike) generated antibodies

Spike-specific antibodies in the (A–G) serum and (H–M) bronchoalveolar lavage fluid (BALF).

(A) IgG titers in mouse sera throughout the study schedule determined by ELISA.

(B–D) IgG1 and (C) IgG2c at 6 months post-vaccination. The ratio of Spike-specific IgG1/IgG2c at 6 months post-vaccination is given in (D).

(E) IgG avidity index at −1, 0, 1, 2, and 6 months post-vaccination.

(F) cPass determined antibody neutralization activity in serum at 6 months post-vaccination.

(G) IgA titers in mouse sera calculated at 0, 3, and 6 months post-vaccination. N = 7–10.

(H–K) Spike-specific IgG, (I) IgG1, and (J) IgG2c with the ratio of IgG1/IgG2c given in (K).

(L) cPass determined antibody neutralization activity in BALF at 6 months post-vaccination.

(M) IgA in BALF at 6 months post-vaccination calculated by ELISA. N = 7. Data points represent individual mice, means ± SD are shown. The included legend applies to both serum and BALF data. For (A), (G), two-way ANOVA with Tukey’s multiple comparisons: ∗∗p < 0.01; ∗∗∗∗p < 0.0001. For (B), (F), (K), one-way ANOVA with Tukey’s multiple comparisons: ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns = not significant. For (C), (D), (H)–(J), (L), (M), Kruskal-Wallis test with Dunn’s multiple comparisons: ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001.

Since BCG is known to induce IFNγ,³³ a promoter of IgG2c production,³⁴ we then addressed if BCG influenced the isotype of antibodies produced upon Ad(Spike) vaccination in the animals prior to challenge. Both groups of vaccinated animals similarly expressed Spike-specific IgG1 (Figure 4B) and IgG2c (Figure 4C) and displayed the same ratio of IgG1/IgG2c (Figure 4D). Correspondingly, both groups of vaccinated animals expressed similar levels of RBD-specific IgG1 (Figure S3D) and IgG2c (Figure S3E) in similar ratios (Figure S3F), confirming that BCG-Danish pre-immunization did not influence isotype-switching upon Ad(Spike) vaccination.

We then assessed IgG avidity by ELISA. As expected, Spike- and RBD-specific IgG avidity rose from 0 to 2 months (Figures 4E and S3B, respectively) and Spike-specific avidity was maintained in both vaccinated groups until 6 months, demonstrating that BCG did not influence the avidity of IgG. Interestingly, RBD-specific IgG avidity dropped at 6 months to levels similar to those observed at one-month post-vaccination (Figure S3C), suggesting that neutralization of the ACE2-Spike binding domain declines with time in both groups. In addition, a cPass surrogate virus neutralization assay confirmed that BCG did not influence the abundance of neutralizing antibodies within immunized mice (Figure 4F) compared to mice vaccinated with Ad(Spike) alone. Intranasal vaccination with Ad(Spike) also resulted in detectable levels of antigen-specific serum IgA in both groups (Figure 4G), prompting us to investigate levels of mucosal antibodies. Bronchoalveolar lavage fluid (BALF)-derived Spike-specific IgG antibodies were present and statistically greater than negative controls at 2 and 6 months post-vaccination (Figure 4H). This was observed also for RBD-specific IgG responses in BALF prior to infection (Figure S3H). Specifically, Spike-specific IgG1 (Figure 4I) and IgG2c (Figure 4J), as well as their ratio (Figure 4K), did not differ between the two vaccinated groups. The same pattern was observed for RBD-specific IgG1 (Figure S3I), IgG2c (Figure S3J), and the ratio of the two (Figure S3K). The mean neutralizing activity of BCG-pre-immunized and vaccinated animals was 2-fold greater than those animals that were solely Ad(Spike) vaccinated, though this difference was not significant (Figure 4L). This trend of higher antibody levels in the BALF from pre-immunized and vaccinated animals was again seen with the Spike-specific IgA (Figure 4M) and RBD-specific IgA (Figure S3L) titers; though, again the difference was not statistically significant.

Intranasal vaccination should also induce mucosal immunity in upper branches of the respiratory tract, so we evaluated both IgG and IgA responses directed against Spike and RBD proteins in mouse turbinate (Figure S4). At 6 months post-vaccination, both experimental groups showed Spike and RBD specific immunoglobulins, whereas the PBS and BCG(e)+Ad(e) groups remained at baseline. Interestingly, Ad(Spike) promoted IgA production in mouse nares, whereas IgG was found in animals which were pre-immunized with BCG(e).

Taken together, these data show that pre-immunization with BCG-Danish does not promote Ad(Spike)-induced protection by modulating the generation, quantity, or quality of circulating or mucosal humoral responses.

BCG pre-immunization promotes long-term cellular immunity in the lungs

An important aspect of vaccination is to generate robust and lasting tissue-resident memory T cells (T_RM) in order to confer protection.³⁵ Since we observed that BCG-Danish did not impact the long-term protective antibody response that Ad(Spike) generated against SARS-CoV-2, we next investigated if BCG-Danish pre-immunization potentiated the generation of memory CD4⁺ and CD8⁺ T_RM cells in vaccinated mice. Six months post-vaccination, isolated lung T cells were activated with SARS-CoV-2 Spike protein peptides (Figures 5A and S5). As shown in Figure 5B, at 6 months post-vaccination, the frequencies of lung CD8⁺ T cells from immunized groups were significantly greater than controls regardless of BCG administration. However, when we assessed the production of Granzyme B (GrB) and IFNγ in activated CD69⁺ CD8⁺ T_RM cells, we observed that mice that received BCG(e) prior to immunization with Ad(Spike) produced a higher frequency of GrB- (Figure 5C) and IFNγ-secreting CD8⁺ T cells (Figure 5D) compared to the Ad(Spike)-alone group 6 months post-vaccination, suggesting that their responses are more cytotoxic in nature. In parallel, we observed an increase in activation (CD69⁺) of CD4⁺ T cells upon peptide restimulation in the immunized group that was previously exposed to BCG (Figure 5E). In fact, mice pre-immunized with BCG, regardless of Ad(Spike) vaccination, had increased frequencies of IFNγ⁺ CD4⁺ (Figure 5F) and IL17A⁺ CD4⁺ T cells in the lung (Figure 5G), although these trends did not reach statistical significance. Coincidently, we also observed significantly less FoxP3⁺ regulatory T (T_REG) cells among CD69⁺ CD4⁺ T cells (Figure 5H), suggesting that BCG promotes the generation of cytotoxic over suppressive T cell responses in BCG(e)+Ad(Spike)-immunized mice relative to the other groups. Collectively, these results demonstrate that BCG pre-immunization acts on the long-term potency of the Ad(Spike) vaccine by promoting the generation of cytotoxic and Th1 responses over suppressive FoxP3⁺ T_REG cells in the lungs of infected mice.

Figure 5.

BCG potentiates the generation of long-lasting cellular immunity

Six months post-vaccination, cell mediated responses from lung cells were analyzed following 18 h restimulation with an S-protein peptide pool.

(A) Expression of CD69 and GrB in vaccinated animals compared to controls (Ad-S = Ad(Spike)).

(B–D) Total frequency of CD69⁺CD8⁺ T cells. The frequency of CD69⁺CD8⁺ T cells which are (C) GrB+ and (D) IFNɣ+.

(E) Frequency of CD69⁺ cells among the CD4⁺ population.

(F–H) IFNɣ, (G) IL17a, and (H) FoxP3 expression was also determined from CD69⁺CD4⁺ T cells. N = 7. Data points represent individual mice, means ± SD are shown. For (B)–(H), one-way ANOVA with Tukey’s multiple comparisons: ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; ∗∗∗∗p < 0.0001; ns = not significant.

Ad(Spike) cross-reactive antibodies persist and are not affected by BCG pre-immunization

Next, we assessed if the administration of BCG affected the ability of Spike antibodies generated by an i.n. Ad(Spike) vaccine based on the ancestral (Wuhan) sequence to subsequently recognize epitopes from SARS-CoV-2 alpha (B.1.1.7), beta (B.1.351), gamma (P.1), delta (B.1.617.2), and omicron variants (B.1.1.529, and BA.2). Six months post-vaccination, serum IgG titers against the ancestral S-protein within Ad(Spike)- and BCG(e)+Ad(Spike)-vaccinated animals were not significantly different between the two vaccinated groups (Figure 6A). To ensure BCG pre-immunization does not affect the cross-reactivity of these antibodies, we conducted ELISAs against S-proteins from variant strains of SARS-CoV-2. We found no significant reduction in the antibody binding capacity of serum from vaccinated animals that were vaccinated with Ad(Spike) (Figure 6B) or pre-immunized with BCG and then vaccinated with Ad(Spike) (Figure 6C) against any of the SARS-CoV-2 strains. These results confirm that BCG administration does not affect the generation of cross-reactive anti-Spike antibodies.

Figure 6.

Production of cross-reactive antibodies to variant spike proteins is not altered by BCG

Spike-specific IgG was calculated from serum from vaccinated animals at 6 months post-vaccination by ELISA.

(A–C) Antibodies are shown against Spike Wuhan (Ancestral) strain. Cross-reactive antibodies from (B) Ad(Spike) vaccinated and (C) BCG pre-immunized and Ad(Spike) vaccinated animals were then assessed and compared to the ancestral strain. Antibodies binding strains B.1.1.7 (alpha), B.1.351 (beta), P.1 (gamma), B.1.617.2 (delta), B.1.1.529 (omicron), and BA.2 (omicron) were assessed. N = 7. Data points represent individual mice, means ± SD are shown. For (A), (B), Kruskal-Wallis test with Dunn’s multiple comparisons: ∗p < 0.05; ns = not significant. For (C), one-way ANOVA with Tukey’s multiple comparisons: ∗p < 0.05; ns = not significant.

Discussion

A single intranasal dose of a human adenovirus vectored vaccine has previously been shown to be sufficient to protect mice from SARS-CoV-2 infection and severe disease.³⁶ However, the durability of protection from these vaccines remains to be established as most studies utilize short-term challenge models. Using an AdV vectored vaccine that expresses the Spike protein of the Wuhan isolate, and then challenging with the mouse-permissive Beta variant of SARS-CoV-2, we demonstrated that a single i.n. dose of a human AdV-based vaccine could cross-protect mice from severe disease months after immunization. Specifically, our Ad(Spike) vaccine successfully prevented mice from developing severe pulmonary histological changes when challenged with SARS-CoV-2 at 2 and 6 months post-vaccination. However, the ability of vaccinated mice to limit initial infection was reduced by 6 months post-immunization, confirming what was observed in a recent meta-analysis.⁶ Thus, we aimed to develop a strategy to prolong the immunity provided by i.n. administration of Ad(Spike) by harnessing the non-specific immune-enhancing effects of the widely available BCG vaccine.^37,38 As hypothesized, herein we demonstrate that BCG improves the long-term protection and reduced viral shedding conferred by a single i.n. Ad(Spike) dose by potentiating cellular, rather than humoral, immunity against the Spike antigen in the lungs. Although our initial data to date demonstrates BCG-induced immune-enhancement for up to 6 months, future work will extend the interval of time between vaccination and challenge. In addition, we plan to test additional BCG administration routes (e.g.,: s.c.), as well as varied i.n. Ad(Spike) vaccination doses. The latter is important due to potential safety concerns over the use of AdV-vectored vaccines, and we anticipate that our combination vaccine strategy may enable us to reduce the effective vaccine dose (or dosing frequency) required to maintain a protective effect. Overall, our current research findings may also lay the groundwork for the use of BCG as a more general vaccine enhancement strategy that is applicable to a broad range of viral infections and vaccination modalities, including those based on lipid nanoparticle encapsulated messenger RNA.

As SARS-CoV-2 achieves endemicity and continues to evolve, novel vaccines and immunization routes that increase mucosal immunity^8,39 are needed. Indeed, several groups have demonstrated that when adenovirus vectored SARS-CoV-2 vaccines are administered intranasally, a single immunization is sufficient to confer protection from infection in naive animals.^40,41 One promising option, investigated herein, is based on the ability of the BCG vaccine to act as an indirect promoter of cellular immunity. For many years, BCG has been approved for use in infants across the globe to protect against severe, disseminated forms of tuberculosis. As such, BCG is commonly reported as the world’s most widely used vaccine and has an excellent safety record. Moreover, BCG has also recently been shown to reduce all-cause mortality from unrelated infectious agents, including several viruses, as a result of its non-specific or “off-target”, immune-enhancing effects.^37,42 Consistent with these reports, we demonstrate that pre-immunization with BCG can non-specifically prevent waning immunity from a single-dose, i.n. SARS-CoV-2 vaccine by potentiating local vaccine-specific cell mediated responses without hindering humoral immunity.

The possible contribution of widespread BCG vaccination to SARS-CoV-2 protection in human populations is still unclear; yet, there is growing evidence that BCG vaccination provides ‘trained immunity’ to innate mechanisms,^37,42,43 which can have antagonistic effects on other unrelated pathogens.^19,38,44,45 This is seen in the curious protection from neonatal sepsis and respiratory infections conferred to BCG immunized babies.¹⁵ It has since been demonstrated that due to intrinsic pathogen associated molecular pattern activation of toll like receptors, BCG induces the activation and reprogramming of monocytes,^33,37,42 resulting in increased expression of cell surface markers and the production of pro-inflammatory cytokines and IFNɣ in response to antigenic stimulation.⁴⁶ These data contributed to the hypothesis that BCG may provide heterologous potentiation of antigen-presenting cell (APC) function to improve the immunogenicity and efficacy of vaccines, as made evident in human studies of neonatal vaccinations^47,48 and adult influenza vaccination.⁴⁹ In fact, BCG is a natural adjuvant, at least partly a result of its modified peptidoglycan structure and unusual cell-wall lipid composition,⁵⁰ and has been exploited in several novel vaccine efforts including against SARS-CoV-2.^32,51,52 It is particularly impressive that BCG has the potential to deliver benefit to vaccination strategies even when not directly co-administered, as shown herein. Nevertheless, BCG alone does not seem to provide direct protection from SARS-CoV-2. While an early study in humans suggested protective efficacy against SARS-CoV-2 from BCG vaccination,⁵³ a recent case study demonstrated that this protection was not attributable to BCG over the course of the pandemic due to an underreporting of COVID-19 cases in regions with high BCG-vaccination rates.^31,54 More recently, a randomized trial involving more than 3,000 healthcare workers did not show a reduced risk of severe COVID-19 amongst those who received BCG compared to the placebo group.²¹ Another group found that BCG alone failed to provide protection from SARS-CoV-2 in mice and hamsters,¹⁹ although this may depend upon the timing and route of administration.⁵⁵ Nonetheless, our data support the idea that neither BCG alone (nor when combined with an empty AdV vaccine vector) confers significant protection against SARS-CoV-2 challenge in a mouse model of infection.

As noted above, data show that despite short-term efficacy, protection from SARS-CoV-2 infection wanes over time in humans.^56,57 Consistent with this, we also observed an attenuation of protection by 6 months post-immunization. Although prior exposure to BCG did not change Ad(Spike) protection 2 months post-vaccination (Figure S2), when mice were primed with BCG and then vaccinated with Ad(Spike), significant protection was maintained for at least 6 months post-vaccination. Due to the numerous possible effects of BCG on the innate and adaptive immune responses, we chose to focus on its impact on the type of humoral immunity generated by a single-dose Ad(Spike) vaccine by first assessing the quality and quantity of serum and BALF antibodies. We observed that regardless of their exposure to BCG, mice that were vaccinated with Ad(Spike) displayed robust serum and BALF Spike-specific antibodies, both IgG and IgA which, while slowly declining, were maintained at high titers to at least 6 months post-vaccination. These antibodies maintained high affinity to the receptor binding domain of the Spike protein, confirming that they were capable of neutralizing viral entry. Reassuringly, these antibodies also displayed cross-reactivity against all Spike variants tested, including the beta variant, which was used in our challenge study. However, despite these seemingly high titers of neutralizing IgG1/2c and IgA antibodies, viral RNA and, to a lesser degree, live viral particles remained high in the absence of BCG. Thus, we hypothesized that the BCG bacilli were promoting cellular rather than humoral immunity in the BCG(e)+Ad(Spike)-immunized mice.

Our antigen-specific assay demonstrated that although Ad(Spike) alone showed a trend toward increased expression of cytotoxic activity in antigen-primed T cells, these differences became significant only in mice that were pre-immunized with BCG. Indeed, BCG has been shown to enhance intramuscular vaccine-induced circulating Spike-specific CD4⁺ T cells in humans,²² and our results reveal that it can also potentiate the generation of antigen-specific CD4⁺ T cells in tissues. Many reports suggest BCG acts by training monocytes to have higher expression of MHC-II, CD80, and CD86, thereby enhancing antigen presentation to T cells.^58,59 Concomitantly, adaptive immune responses after BCG vaccination also involves the activation of CD8⁺ T cells.³⁸ Here, we observed a clear potentiation of Spike-specific cytotoxic CD8⁺ T cells up to at least 6 months after vaccination. Furthermore, cytokines secreted by BCG-exposed monocytes, such as IL1β and IL6, are key contributors to CD4⁺ T cell differentiation into Th1 and Th17 subsets^60,61 which is consistent with the increased CD4⁺ T cell expression of IFNɣ or IL17a we observed in BCG-immunized mice even in the absence of Ad(Spike). Thus, it is likely that our assay could not distinguish Spike-specific from bystander Th1 and Th17 cells generated prior to Ad(Spike) administration.

Our study used wild-type C57BL/6 mice, which necessitated using a SARS-CoV-2 strain that binds the mouse ACE-2 (mACE-2) receptor. For this reason, we challenged the C57BL/6 mice with the B.1.351 strain, capable of binding mACE-2 and establishing infection.⁶² Thus, we demonstrated cross-protection delivered by our ancestral strain-based AdV vaccine in a context of SARS-CoV-2 infection due to a heterologous strain, which resembles the situation with human disease.⁶³ Within this study, we offer insight into the effects of prior BCG immunization on reinforcing vaccine efficacy in a mouse model of SARS-CoV-2. Interestingly, due to the long-term persistence of viable BCG in our experimental design (Figure S6), this study also raises the issue of vaccine efficacy in the context of other persistent bacterial co-infections, as well as components of the normal mucosal microbiota, and represents an important future line of inquiry. As preclinical vaccine testing is conducted in naive animals, the role of immunological memory from previous vaccination and persisting infections should be addressed in relation to vaccine efficacy.

Collectively, our results present a vaccination approach that can potentially curb viral pandemics by potentiating the long-term efficacy of a next-generation adenovirus-vectored mucosal vaccine when provided in the context of the safe and widely distributed BCG vaccine. This approach can also have an impact in other mucosal and non-mucosal vaccination strategies as well. Going forward, these strategies may also provide a viable solution to ensure a more rapid and equitable distribution of vaccines among developing nations where BCG is already firmly entrenched within many vaccination programs.

Limitations of the study

Although the C57BL/6 mouse model reflects age- and sex-based differences in human COVID-19 disease,⁶³ it is not the ideal model for SARS-CoV-2 vaccine studies since these mice do not display all the hallmark features of severe pathology and typically recover from infection.⁶⁴ As follows, the lung changes observed in our model were modest, posing a limitation in our ability to discriminate severe disease between vaccinated groups. Furthermore, although lung-cell memory responses were increased in BCG pre-immunized animals, our long-term infection model of 6 months post-vaccination may have been insufficient to clearly assess the synergistic effects of BCG on Ad(Spike). We propose that the roles these memory responses play may become more obvious in longer-term studies when protection from Ad(Spike) alone is further reduced.

In humans, BCG is most typically administered intradermally; however, as noted above, we chose to immunize animals intraperitoneally to maximize systemic responses. An immunization route closer to that used in humans is the s.c. route and this will be explored in future experiments. In addition, our group utilized a human adenovirus serotype 5 (HAdV5), a commonly used vaccine vector due to its capability of inducing strong immune responses, as well as being able to be produced in large quantities. However, some drawbacks include that pre-existing immunity to HAdV5 is common in many developing countries⁶⁵ and that there are potential risks associated with the use of adenovirus vectors (e.g., Guillain-Barre syndrome (GBS)) in some individuals. Although antigen-specific T cell responses have been demonstrated in individuals with pre-existing HAdV5 immunity,⁶⁶ to circumvent this potential challenge, a rarer adenoviral vector could be used, for instance, serotype 26 which showed efficacy in Janssen’s COVID-19 vaccine when deployed in South Africa.⁶⁷ Interestingly, an increase in GBS was not found when using COVID-19 vaccines based on HAdV5, unlike other vaccination platforms.⁶⁸ Despite this, considering the extreme infrequency of vaccine-related serious adverse events, most consider the increased risk of GBS miniscule in comparison to the protection against severe and lethal COVID-19 offered by vaccination.⁶⁹ Using a lower dose of viral-vectored vaccines may help to reduce the risk of adverse events even further and may point to an additional benefit of employing BCG as a means of immune enhancement whereby the combination with BCG may allow for a reduced dosing schedule or a reduction in the viral titer required to achieve effective vaccination.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-SARS-COV2 SARS-COV/SARS-COV-2 Spike RBD Polyclonal Antibody(2019-nCoV) Polyclonal Antibody	MyBioSource, San Diego, CA, USA	Cat: MBS2563840
Anti-mouse IgG HRP	Sigma Aldrich	Cat:A0168; RRID: AB_257867
HRP-conjugated anti-mouse IgA	Sigma Aldrich	Cat:A4789; RRID: AB_258201
Goat anti-mouse IgG1-HRP	SouthernBiotech	RRID:AB_2794426 Cat:1071-05
Goat anti-mouse IgG2c-HRP	SouthernBiotech	RRID:AB_2794462 Cat:1078-05
Fixable viability dye eFluor 780	Thermofisher	Cat:65-0865-14
FOXP3 Monoclonal Antibody (FJK-16s), PE-Cyanine7, eBioscience™	Thermofisher	RRID: AB_891552
BUV395 Hamster Anti-Mouse CD3e	BD Biosciences	RRID: AB_2738278 Cat: 563565
BV510 Rat Anti-Mouse CD8b	BD Biosciences	RRID: AB_2739908 Cat: 740155
BUV737 Rat Anti-Mouse IFN-γ	BD Biosciences	RRID: AB_2870098 Cat: 612769
Alexa Fluor® 700 anti-mouse CD4 Antibody	BioLegend	RRID: AB_493701 Cat: 100536
APC anti-mouse IL-17A Antibody	BioLegend	RRID: AB_536018 Cat: 506916
Brilliant Violet 421™ anti-mouse IL-4 Antibody	BioLegend	RRID: AB_2562594 Cat: 504127
FITC anti-mouse CD69 Antibody	BioLegend	RRID: AB_313108 Cat: 104505
PerCP/Cyanine5.5 anti-mouse TNF-α Antibody	BioLegend	RRID: AB_961434 Cat: 506322
PE anti-human/mouse Granzyme B Recombinant Antibody	BioLegend	RRID: AB_2687032 Cat: 372208
Rockland Immunochemicals Rabbit TrueBlot: Anti-Rabbit IgG HRP - 18-8816-31	Rockland Immunochemicals, Pottstown, PA, USA	Cat: 18-8816-31; RRID: AB_2610847
Bacterial and virus strains
non-replicating Adenovirus (ΔE1, ΔE3; 1st generation) containing no gene cassette	National Research Council of Canada, Montreal, QC, Canada	N/A
BCG Danish	ATCC	Cat: 35733
Chemicals, peptides, and recombinant proteins
SuperSignal West Pico Plus Chemiluminescent Substrate	Thermofisher	Cat:34580
TMB Substrate	Sigma Aldrich	Cat:T0440
GolgiPlug	BD Science, San Jose, CA, USA	RRID:AB_2869014 Cat: 555029
phorbol 12-myristate 13-acetate	Thermofisher	Cat:356150010
Ionomycin	Thermofisher	Cat: J62448.MCR
Fc block	BD Science	RRID:AB_394656
Fixation Buffer	BD Science	RRID: AB_2869005 Cat: 554655
Permeabilization buffer	BD Science	RRID: AB_2869011 Cat: 554723
PepTivator SARS-CoV-2 Prot_S	Miltenyi Biotec, Bergisch Gladbach, North Rhine-Westphalia, Germany	Cat: 130-126-700
PepTivator SARS-CoV-2 Prot_S Complete	Miltenyi Biotec	Cat: 130-127-951
SARS-CoV-2 Neutralization Antibody Standard	GenScript	Cat: A02087
SARS-CoV-2 Spike (Wuhan)	National Research Council of Canada, Montreal, QC, Canada	N/A
SARS-CoV-2 RBD (Wuhan)	National Research Council of Canada, Montreal, QC, Canada	N/A
SARS-CoV-2 Spike (B.1.1.7)	National Research Council of Canada, Montreal, QC, Canada	N/A
SARS-CoV-2 Spike (B.1.351)	National Research Council of Canada, Montreal, QC, Canada	N/A
SARS-CoV-2 Spike (P.1)	National Research Council of Canada, Montreal, QC, Canada	N/A
SARS-CoV-2 Spike (B.1.617.2)	National Research Council of Canada, Montreal, QC, Canada	N/A
SARS-CoV-2 Spike (B.1.1.529)	National Research Council of Canada, Montreal, QC, Canada	N/A
SARS-CoV-2 Spike (BA.2)	BEI Resources, Manassas, VA, USA	Cat: NR-56517
Critical commercial assays
Q-plex Mouse Cytokine – Screen (16-plex)	Quansys Biosciences, West Logan, UT, USA	Cat:110949MS
SARS-CoV-2 Surrogate Virus Neutralization Test (sVNT)	GenScript	Cat:L00847-A
Deposited data
Available upon request	N/A	N/A
Experimental models: Cell lines
AdEasier-1 cells	Addgene, Watertown, MA, USA	GenBank ID:AY370909, Addgene plasmid #16399
One Shot Top10 Chemically Competent E. coli	ThermoFisher Scientific, Waltham, MA, USA	Cat:C404010
HEK293A cells	American Type Culture Collection (ATCC), Manassas, VA, USA	RRID:CVCL_6910
SF-BMAd-R cells	National Research Council of Canada, Montreal, QC, Canada	N/A
Experimental models: Organisms/strains
C57BL/6NCrl mice	Charles River Laboratories, Senneville, QC, Canada	RRID:IMSR_CRL:027
Recombinant DNA
AdSpike construct	Integrated DNA Technologies (Coralville, IA, USA)	N/A
Software and algorithms
FlowJo software (version 10.0.8r1)	Treestar, Ashland, OR, USA	N/A
GraphPad Prism 9 software	(La Jolla, CA, USA)	N/A

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Momar Ndao (momar.ndao@mcgill.ca).

Materials availability

This study did not generate new unique reagents.

Experimental model and study participant details

Mice

Six- to eight-week-old female C57BL/6 mice were ordered from Charles River Laboratories (RRID:IMSR_CRL:027) (Senneville, QC, Canada). Mouse housing, husbandry, and environmental enrichment can be found within the McGill standard operating procedures (SOP) #502, #508, and #509.

Animal ethics

All animal procedures were performed in accordance with the Institutional Animal Care and Use Guidelines approved by the Animal Care and Use Committee at McGill University (Animal Use Protocol 8190). Animals were monitored for adverse events for 3 days post-vaccination and weekly until the end of each experiment. Humane intervention points were monitored according to McGill SOP #410. Challenge experiments were performed in compliance with the Canadian Council on Animal Care guidelines and approved by the Animal Care Ethics Committee at the University of Toronto (APR-00005433-v0002-0). All animals were humanely sacrificed at endpoint by anaesthesia with isoflurane before euthanasia by carbon dioxide asphyxiation, followed by pneumothorax and blood collection by cardiac puncture.

Cell lines and reagents

Cell lines were obtained from commercial sources, passed quality control procedures, and were certified and validated by the manufacturer. SF-BMAd-R cells were validated for identity, as human derived.⁷⁰ All reagents were validated by the manufacturer or has been cited previously in the literature. When available, RRID tags have been listed in the text and in the reagent repository (Table S2).

Mycobacterium bovis BCG Danish culture

Mycobacterium tuberculosis variant bovis BCG (ATCC 35733), provided to us by Dr. Marcel Behr (Research Institute of the McGill University Health Centre), was grown in Middlebrook 7H9 broth (BD, Mississauga, ON, Canada) supplemented with 10% ADC (8.1 g/l NaCl, 50 g/l BSA Fraction V (Millipore Sigma, Billerica, MA, USA), 20 g/l glucose), 0.2% glycerol and 0.05% Tween 80, or on Middlebrook 7H11 agar (BD) supplemented with 10% OADC enrichment (as per ADC plus 0.6 ml/l oleic acid, 3.6mM NaOH). The BCG-Danish strain used in these experiments was transformed with an empty pMV361(hygromycin^R) plasmid and was initially selected in the presence of 50 μg/ml hygromycin (Wisent, Saint-Jean-Baptiste, QC, Canada). This plasmid was integrated into the non-essential L5 phage attachment site (attB) located within the BCG chromosome. We termed this strain BCG(e).

Method details

Author checklists

This study has been submitted with both the ARRIVE and MDAR author checklists.

Preparation of M. bovis BCG-Danish cultures for mouse immunization

100 ml BCG cultures were grown in 7H9/ADC medium to an OD_600nm = 0.6. After spinning at 3000 rpm, cells were washed twice with PBS containing 0.05% Tween-80 (PBS-Tw) and resuspended in 6 ml of this buffer. After passing 10 times through a 22G x1″ and 10 times through a 27G x1/2″ needle, the suspension was mixed with 4 ml of sterile 50% glycerol in PBS. Aliquots were made and frozen at −80°C. Prior to immunization, an aliquot was thawed, and 10-fold serial dilutions were plated on 7H11/OADC+ HYG for quantification, yielding a value of ∼1.5 x10⁸ cfu/ml. On the day of the immunization, these glycerol stocks were diluted (1/20) in PBS-Tw and 200 ul (containing ∼1.5 × 10⁶ cfu) were injected i.p. into the lower right quadrant of the abdomen of the mice using a 28Gx1/2″ needle and an insulin syringe. Inocula were quantified by plating 10-fold serial dilutions on 7H11/OADC+ HYG.

Generation of Ad(Spike) vector

The AdSpike construct was developed following a similar protocol as described.⁷¹ Briefly, the Spike gene cassette combined a Kozak sequence with the full length of the Spike protein (Genbank accession number QHU36824.1), codon optimized to mouse and human expression avoiding restriction sites Bgl2, Pac1, and Pme1, followed by a Kpn1 restriction site and the poly-A signal “TCTAGACTCGACCTCTGGCTAATAAAGGAAATTTATTTTCATTGCAATAGTGTGTTGGAATTTTTTGTGTCTCTCACTCGGAAGGACATATGGGAGGGCAAATCATTTGCGGCCGCGATATC” (GenScript, Piscataway, NJ, USA). The gene cassette was flanked by Bgl2 sites and synthesized by Integrated DNA Technologies (Coralville, IA, USA) then cloned into the vector, pShuttle-CMV-Cuo.⁷² Primers to confirm gene sequence can be found in Table S3. The plasmid containing our recombinant non-replicating human adenovirus serotype 5 (E1 and E3 genes removed (ΔE1-, ΔE3-); 1^st generation) encoding the S-protein gene was made through homologous recombination in AdEasier-1 cells (strain), a gift from Dr. Bert Vogelstein (Addgene plasmid #16399) (Addgene, Watertown, MA, USA).⁷³ It was then linearized with PacI and transfected into HEK293A cells (RRID:CVCL_6910). Our recombinant adenovirus was then amplified using SF-BMAd-R cells⁷⁰ in 3 batches (Ad(Spike) 1–3), combined, and purified by ultracentrifugation on CsCl gradients as described previously,⁷⁴ before titration using a TCID₅₀ assay. A second human adenovirus serotype 5 (ΔE1-, ΔE3-; 1^st generation), lacking a gene cassette, was used as a negative control.

Western blot assays

To determine protein expression by Ad(Spike), cell lysates of Ad(Spike) infected HEK293A cells were assessed. Briefly, cells were infected at a multiplicity of infection of 5 particles per cell and incubated for 48–72 hours, pelleted, and then lysed (0.1M Tris, 10 μL EGTA, 50 μL Triton-100, 0.1M NaCl, 1mM EDTA, 25 μL 10% NaDeoxycholate, 1X protease inhibitor, in ddH₂O). Cell lysates were then resolved on an SDS-PAGE gel under reducing conditions followed by transfer onto a nitrocellulose membrane. The membrane was subsequently blocked in phosphate buffered saline (PBS) with 0.05% Tween 20 (PBS-T) and 5% milk (Smucker Foods of Canada Corp, Markham, ON, Canada) (PBS-TM). The membrane was then incubated with rabbit anti-SARS-CoV-2 Covid-19 Spike RBD coronavirus polyclonal antibody (RRID:AB_258251) diluted 1:5,000 in PBS-TM overnight at 4°C. The membrane was then washed in PBS-T before incubation with horseradish peroxidase (HRP)-conjugated anti-rabbit IgG (IgG-HRP) (Rockland Immunochemicals, Pottstown, PA, USA) diluted 1:10,000 in PBS-T for one hour at room temperature. After incubation the membrane was washed again and developed using SuperSignal West Pico Plus Chemiluminescent Substrate (ThermoFisher Scientific, Waltham, MA, USA) (Figure S7).

Protein expression and purification

SARS-CoV-2 Spike variants Wuhan, B.1.1.7 (alpha), B.1.351 (beta), P.1 (gamma), B.1.617.2 (delta), B.1.1.529 (omicron), and the RBD portion of the Wuhan Spike variant were obtained from the National Research Council of Canada. Recombinant, “tagless” Spike proteins were produced as previously described.⁷⁵ Recombinant RBD protein was produced as previously described.^76,77 Spike variant BA.2 (omicron) was obtained through BEI Resources, NIAID, NIH: Spike Glycoprotein (Stabilized) from SARS-Related Coronavirus 2, BA.2 Lineage (Omicron Variant) with C-Terminal Histidine and Avi Tags, Recombinant from HEK293 Cells, NR-56517.

Immunization and challenge protocol in mice

Each mouse was immunized at weeks 0 and 4 by intraperitoneal (i.p.) injection of 200μL of BCG(e) and i.n. administration of 30μL of adenovirus formulations, respectively. See Table S4 for more precise group descriptions. Group 2 was removed from the long-term challenge experiment since negative control animals displayed similar viremia to the PBS control in the short-term challenge model. Mice were bled from the saphenous vein at weeks 0, 4, 8, 12, and 18. Mice immunized for humoral and cell-mediated immunity assessment (n = 6) were euthanized 6 months after the final immunization and blood, spleens, and lungs were collected. Mice immunized for challenge studies (n = 12) were transferred to the University of Toronto and challenged with 10⁶ TCID₅₀ of SARS-CoV-2 South African strain (B.1.351) 2 (n = 6) or 6 months (n = 6) post-Ad(Spike)-vaccination. TCID₅₀ was determined using the Spearman–Kärber method.⁷⁸ Oral swabs were taken from mice on days 1, 3, and 5 post-challenge in DMEM. Mice were euthanized 5 days after challenge and lungs were collected.

Quantification of viral load

Quantities of infectious virus was determined by determining the median tissue-culture infectivity dose (TCID₅₀) using methods that have been described previously.⁷⁹ Briefly, Vero E6 cells were seeded into plates and incubated overnight at 37°C. On the following day, media was removed, and samples were added and serially diluted using ten-fold dilutions. Plates were incubated at 37°C for 1 h. After incubation, the media was removed and replaced with complete DMEM, and plates were incubated at 37°C for 5 days. Cells were examined for cytopathic effect (CPE) at 5 dpi. TCID₅₀ was defined using the Spearman–Kärber method.⁷⁸

qRT-PCR

Viral RNA loads were calculated as previously described.⁷⁹ Briefly, viral RNA was extracted using the QIAamp viral RNA kit (Qiagen, Hilden, Germany) according to the manufacturer’s guidelines. SARS-CoV-2 viral RNA detection and quantification was performed using the Luna Universal Probe One-Step RT-qPCR kit (New England Biolabs, Whitby, ON, Canada) on the Rotor-gene Q platform (Qiagen). For quantification, standard curves were generated using a synthetic plasmid containing a segment of the E-gene (GenScript) and interpolation was performed as described by Feld et al.⁸⁰ The limit of quantification was determined to be 20 copies/mL.

Spike and RBD-specific IgG, IgG1, IgG2c, IgA quantification and IgG avidity assays

Briefly, high binding 96-well plates (Greiner Bio-One, Frickenhausen, Germany) were coated with recombinant Spike or RBD (0.5 μg/mL) in 100 mM bicarbonate/carbonate buffer (pH 9.6) along with various standard curves (IgG, IgG1, IgG2c, IgA: serially diluted from 2000 ng/mL to 1.953 ng/mL) overnight at 4°C. Then, plates were blocked with 2% bovine serum albumin (BSA; Sigma Aldrich, St. Louis, MO, USA) in PBS-T (blocking buffer) for 1 hour at 37°C before samples diluted in blocking buffer were added in duplicate. Nasal wash samples were run in singlet. When running serum for total Spike/RBD-IgG, an additional set of serum samples was run to determine IgG avidity. Plates were incubated for 1 hour at 37°C then washed with PBS (pH 7.4). For IgG avidity assessment, the additional set of samples received 8M urea, while blocking buffer was added to the first set and the standard curve. Plates were covered and incubated for 15 minutes at room temperature protected from light, washed 4 times, and then blocked again with blocking buffer for 1 hour at 37°C. Next, plates were washed with PBS and anti-mouse IgG-HRP (Sigma Aldrich) was diluted 1:20,000 in blocking buffer and applied for 30 minutes at 37°C. For other immunoglobulins, the same protocol was followed without the additional avidity steps and the appropriate HRP-conjugated antibody was applied. Both IgG1- and IgG2c-HRP were diluted 1:20,000 in blocking buffer and applied for 30 minutes at 37°C. For IgA, HRP-conjugated anti-mouse IgA (Sigma Aldrich) was diluted 1:2,000 in blocking buffer and applied for 1 hour at 37°C. Plates were washed a final time with PBS and 3,3′,5,5′-Tetramethyl benzidine (TMB) substrate (Sigma Aldrich) was added to each well. The reaction was stopped after 15 minutes using H₂SO₄ (0.5M; Fisher Scientific, Waltham, MA, USA) and the optical density (OD) was measured at 450 nm with an EL800 microplate reader (BioTek Instruments Inc., Winooski, VT, USA). Concentrations of Spike/RBD specific antibodies were calculated by extrapolation from respective standard curves and multiplied by the dilution factor. IgG avidity indices were calculated by dividing the IgG titer in the urea conditions by the IgG titer in the non-treated condition.

Surrogate virus neutralization test

Neutralizing antibodies were assessed using the cPass SARS-CoV-2 Neutralization Antibody Detection Kit (GenScript) according to manufacturer’s instructions with the following changes: To collect semi-quantitative results, the kit was run with the SARS-CoV-2 Neutralizing Antibody Calibrator to create a standard curve used to determine the concentration of neutralizing antibodies. BALF samples were run neat or diluted (1:3) and serum samples were diluted (1:150). Samples that gave values above the 30% signal inhibition cut-off value were multiplied by the dilution factor and reported as Units/mL. Data reported according to the manufacturer’s guidelines can be found in Tables S5 (serum) and S6 (BALF).

Turbinate wash, BALF, and lung collection

Six months after the last immunization, unchallenged mice were euthanized, and the lungs were collected. Turbinate washes were collected by combining three washes of 150μL of cold PBS+protease inhibitor (PBS+PI) pushed through the nasal pharynx. Bronchoalveolar lavage fluid (BALF) was collected by combining four lung washes of 0.5mL of cold PBS+PI. Lungs were collected in 1mL cold RPMI. Lungs were digested enzymatically for 30 minutes at 37°C and 5% CO₂ with a cocktail of DNase I (200 μg/mL, Sigma Aldrich), LiberaseTM (100 μg/mL, Roche, Indianapolis, IN, USA), hyaluronidase 1a (1 mg/mL, Life Technologies, Carlsbad, CA, USA), and collagenase XI (250μg/ml; Life Technologies) in RPMI-1640 as described previously.⁸¹ Cells were then washed with RPMI-1640 media containing 1% Penicillin/Streptomycin and 5% FBS. Sterile, filtered ammonium-chloride-potassium (ACK) buffer was used to lyse red blood cells. Filtration through a 0.7μM strainer was performed and the remaining viable cells were recovered.

Quantification of cytokine-secretion in T cells by multi-parametric flow cytometry

Lung lymphocytes were seeded into 96-well flat bottom plates (BD) at 10⁶ cells in 200 uL/well. Duplicate cultures were stimulated with or without a combined preparation of Peptivator Peptide Pools of the complete Spike protein and predicted immunodominant sequences (Miltenyi Biotec, Bergisch Gladbach, North Rhine-Westphalia, Germany) in RPMI (0.3 μg/mL final concentration) for 18 and 96 hours at 37°C + 5% CO₂. For the last 6 hours of incubation, protein transport inhibitor was prepared according to the manufacturer’s guidelines (RRID:AB_2869014, BD Science, San Jose, CA, USA) and added to all samples. Cells stimulated with phorbol 12-myristate 13-acetate (Thermofisher Scientific) and ionomycin (Thermofisher Scientific) were processed as positive controls. All staining and fixation steps took place at 4°C protected from light. Briefly, the cells were washed twice with cold PBS and stained with 50μL/well fixable viability dye eFluor 780 (Thermofisher Scientific) diluted at 1:1000 for 20 minutes. Cells were washed once with PBS. All surface stains were diluted 1:50 in PBS and 50μL/well of extracellular cocktail was applied for 30 minutes. The following antibodies made up the extracellular cocktail: CD3-BUV395 (145-2C11, RRID: AB_27382, BD Biosciences, Franklin Lakes, NJ, USA), CD4-AF700 (RM4-5, RRID: AB_49370, BioLegend, San Diego, CA, USA), CD8b-BV510 (H35-17.2, RRID: AB_2739908, BD Biosciences) and CD69-FITC (H1.2F3, RRID: AB_313108, BioLegend). Cells were then washed as before and fixed with the eBioscience FoxP3 transcription factor staining buffer (Thermofisher Scientific) overnight. The next day, plates were washed with 1X permeabilization buffer (perm buffer) (Thermofisher Scientific) and stained with an intracellular cocktail of antibodies diluted 1:50 in perm buffer applied as 50μL/well for 30 minutes. The intracellular cocktail was made up of: FOXP3-Pe-Cy7 (FJK-16s, RRID: AB_891552, Thermofisher Scientific), IFNγ-BUV737 (XMG1.2, RRID: AB_2870098, BD Biosciences), IL-17A-APC (TC11-18H10.1, RRID: AB_536018, BioLegend), IL-4-BV421 (11B11, RRID: AB_2562594, BioLegend), GrB-PE (QA1602, RRID: AB_2687032, BioLegend), and TNFα-PerCP-Cy5.5 (MP6-XT22, RRID: AB_961434, BioLegend). After staining, cells were washed twice with perm buffer and resuspended in PBS 1X and acquired on a BD LSRFortessa X-20 (BD Science). Flow data were analysed using FlowJo software (version 10.0.8r1) (Treestar, Ashland, OR, USA). Our gating strategy is shown in Figure S5.

Histological analysis

Organs collected from the challenged animals at the time of necropsy were placed in 10% phosphate-buffered formalin. Collected tissues were subsequently processed for histopathology, and slides were stained with haematoxylin and eosin (H & E), to assess tissue architecture and inflammation, and Masson’s trichrome, to determine the progression of fibrosis. Sections of lungs were examined and scored by a pathologist who was blinded to the experimental groups. Lungs were evaluated for fibrosis, the presence or absence of features of cell or tissue damage (CTD: necrosis of bronchiolar epithelial cells (BEC), inflammatory cells and/or cellular debris in bronchi, intraepithelial neutrophils, alveolar emphysema), circulatory changes and vascular lesions (CVL: alveolar hemorrhage, significant alveolar edema, vasculitis/vascular endothelialitis), reactive inflammatory patterns (RIP: necrosuppurative bronchitis, intraalveolar neutrophils, and macrophages, mononuclear infiltrates around airways, presence of polymorphonuclear granulocytes, perivascular mononuclear cuffs, and mesothelial reactivity), as well as regeneration and repair (RR: alveolar epithelial hyperplasia/regeneration, BEC hyperplasia/regeneration) as described elsewhere.^79,82 After the scoring was completed, lung pathology scores were tabulated. Processed and stained lung slides were digitized using Aperio AT2 (Leica Biosystems, Wetzlar, Germany).

Quantification and statistical analysis

Experimental units are defined as individual animals. Sample sizes were empirically estimated based on previous data considering the anticipated variation of the results and statistical power needed, while also minimizing the number of animals used. C57BL/6 mice were randomly attributed to treatment groups. To minimise potential confounders, mice were matched for age and sex. Blinding: For all challenge experiments, staff performing infections and sample harvesting were blinded to the different groups and were only unblinded after data analysis. Inclusion/Exclusion: Aside from a small number of deaths in one of the control groups (Ad(e)+BCG(e)), no other animals were excluded from the analysis. Statistical analysis was performed using GraphPad Prism 9 software (La Jolla, CA, USA) and statistical details of experiments can be found in figure legends. Data were assessed for normality using Shapiro-Wilk tests. Non-parametric data were analysed by Kruskal-Wallis tests with Dunn’s multiple comparisons. When appropriate, one-way and two-way ANOVAs were employed with Tukey’s multiple comparisons. P values <0.05 were considered significant.

Additional resources

Included in the supplementary data are:

• (1)Supplemental figures.
• (2)Reagent repository.
• (3)ARRIVE Checklist.
• (4)MDAR Author Checklist.

Published: August 11, 2023

Data and code availability

• •All data associated with this study are present in the paper or the Supplementary Materials. Requests may be made by contacting the corresponding authors.
• •This paper does not report original code
• •Any additional information required to reanalyze the data reported in this work paper is available from the lead contact upon request.