Rapid and potent induction of protective mucosal immunity against SARS-CoV-2 by an adenovirus-vectored nanoparticle vaccine

Abstract

Effective vaccine design ideally elicits potent, durable, and broad-spectrum immunity against rapidly evolving pathogens. However, current platforms, such as recombinant adenovirus vectors and protein nanoparticles, are developed and administered separately, and each has distinct immunological and practical limitations. Here, we introduce a synergistic vaccine framework that intrinsically combines genetic delivery with rational antigen design. We engineer an adenovirus type 5 vector to co-express a prefusion-stabilized SARS-CoV-2 spike and ferritin subunit (Ad5-SpFN), which drives in vivo assembly of spike-decorated nanoparticles. Compared to either platform alone, this unified strategy elicits faster and higher neutralizing antibody titers and robust T-cell responses, and confers complete protection against lethal viral challenge. Mechanistically, Ad5-SpFN enhances early dendritic cell activation and germinal center reactions, leading to a more potent and sustained humoral immune response. Furthermore, it provides exceptional route flexibility, establishing protective mucosal immunity after intranasal administration, which addresses a key limitation of many existing vaccines. This versatile “vector-nanoparticle” hybrid platform can be rapidly reprogrammed, offering a promising strategy for future pandemic preparedness.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04326-0.

Background

Infectious diseases continue to pose a major global public health threat, with pathogens such as coronaviruses and influenza viruses driving recurrent pandemics that cause significant morbidity, mortality, and socioeconomic loss [1, 2]. Vaccination remains one of the most effective countermeasures, where protective efficacy relies on inducing robust and balanced immune responses. Specifically, high titers of neutralizing antibodies (NAbs), primarily driven by B cells, are critical for blocking initial infection, while potent virus-specific T cell responses are essential for controlling viral replication and preventing severe disease [3–5]. An ideal vaccine should therefore synchronize these complementary arms of the adaptive immune system.

To date, multiple vaccine platforms have been successfully deployed, each with distinct immunological profiles and practical trade-offs. Traditional and next-generation platforms range from protein subunits and nanoparticles to viral vectors and nucleic acid-based designs [6–9]. Protein nanoparticle vaccines such as spike-ferritin (SpFN) platform, which utilizes a self-assembling ferritin scaffold to display dense and repetitive antigens, induce stronger NAb titers and broader humoral responses than simple subunit vaccines [10–17]. However, they typically elicit weaker CD8⁺ T cell responses compared to viral-vectored platforms [18, 19]. Although strategies like potent adjuvants or mRNA delivery can enhance cellular immunogenicity, nanoparticle vaccines often face challenges in complex manufacturing, stringent quality control, and cold-chain logistics [20]. In contrast, viral-vectored vaccines, notably adenoviral vectors are well known for eliciting robust CD8⁺ T-cell and Th1-biased CD4⁺ T-cell responses due to their intrinsic ability to infect cells and express antigens endogenously [21, 22]. This platform has been widely validated during the COVID-19 pandemic [23–25]. Nevertheless, adenoviral vectors often induce comparatively lower antibody titers than protein nanoparticles and may be limited by pre-existing vector immunity [26].

Given these complementary strengths and limitations, a rational strategy for next-generation vaccine design involves integrating both modalities. The development of an Ad5 (adenovirus serotype 5)-based nanoparticle vaccine emerges as a promising approach to leverage the potent cellular immunogenicity of viral vectors with the superior antibody-inducing capacity of nanoparticles. Indeed, prior studies using adenoviral vectors to deliver nanoparticle antigens (e.g., influenza hemagglutinin or SARS-CoV-2 RBD displayed on ferritin) have demonstrated high immunogenicity, yet the underlying mechanisms driving these enhanced responses remain incompletely defined [27, 28]. Furthermore, while systemic immunity is crucial, there is growing recognition of the importance of mucosal immunity, especially for respiratory pathogens [29–31]. Most current intramuscular (IM) vaccines, including mRNA and viral vector platforms, do not reliably induce robust mucosal antibody responses in the respiratory tract. Adenoviral vectors, however, possess natural tropism for respiratory tissues and have inherent adjuvant properties, making them particularly suitable for mucosal delivery via intranasal (IN) administration, a route that could potentially establish frontline immune protection at the portal of entry [32, 33].

In this study, we developed Ad5-SpFN, a recombinant Ad5-vectored vaccine encoding the SARS-CoV-2 spike-ferritin nanoparticle antigen designed to promote in vivo assembly. We evaluated its capacity to induce rapid, balanced Th1/Th2 immunity and broad antibody responses via both IM and IN administration in a murine model, and assessed its protective efficacy against viral challenge. Additionally, we investigated the immunological mechanisms underlying the enhanced immunogenicity of this hybrid platform. Our findings aim to inform the rational design of next-generation vaccines that combine broad, potent, and durable systemic immunity with effective mucosal protection.

Methods

Ethics statement

All experiments involving authentic SARS-CoV-2 strains were conducted in an animal biosafety level 3 (ABSL-3) facility of Guangzhou National Laboratory and approved by the biosafety ethics committee of Guangzhou National Laboratory (Approval number: GZLABSL3-A2025-011 A). Animal experiments were approved by the Institutional Animal Care and Use Committee (IACUC: GZLAB-AUCP-2023-06-A03) of Guangzhou National Laboratory, and complied with all relevant ethical regulations regarding animal research. The animals were housed and bred in specific pathogen free (SPF) mouse facilities in compliance with ethical guidelines.

Viral vector production

Recombinant adenoviral vectors were rescued as previously described [34, 35]. Genes encoding control immunogen Spp (Wild-type Spike ectodomain with K986P and V987P mutations), Spp.P2A-Ferritin (Spp.P2A-FN), Spp.EPM-EABR and spike-Ferritin (SpFN) were cloned into pBR322-based adaptor plasmids pAdApt vector. PAdApt together with cosmids pWE.Ad.AflII-rITR.dE3 were linearized and transfected into HEK293A cells together with the linearized cosmids using polyethylenimine (PEI) to produce the virus by homologous recombination. These vectors were then plaque-purified, expanded and purified by double CsCl gradient ultracentrifugation, and dialyzed into phosphate-buffered saline (PBS) containing 5% sucrose. Virus particle (vp) titers were determined by qPCR.

Protein expression and purification

All expression vectors were transiently transfected into 293 F cells using PEI transfection reagent. Cells were grown in polycarbonate baffled shaker flasks at 37 °C and 8% CO₂ at 150 rpm. Cells were harvested 5 days post-transfection via centrifugation at 9,000 rpm for 15 min. Culture supernatants were filtered with a 0.22 μm filter and stored at 4 °C prior to purification. His-tagged proteins were purified using Ni-NTA affinity chromatography from 400 mL of expression supernatant. Ni-NTA resin was equilibrated with 10 column volumes (CV) of 20 mM Tris, 50 mM NaCl (pH 7.4) followed by supernatant loading twice at 4 °C. Unbound protein was removed by washing with 20 CV of PBS buffer. Bound protein was eluted with 250 mM imidazole in 20 mM Tris, 50 mM NaCl buffer (pH 7.4). Ni-NTA resin was used to purify protein from 400 mL of expression supernatant. Addition of 1 mM EDTA and 5% glycerol to the Ni-NTA purified material, prior to SEC or other concentration steps, mitigated the aggregation issue, and increased the nanoparticle formation as judged by size-exclusion chromatography (SEC), and confirmed by Negative-stain electron microscopy (Neg-EM). All proteins were further purified by size-exclusion chromatography using a Superdex 200 10/300 GL purification column. Purity for all the proteins was assessed by SDS-PAGE.

Dynamic light scattering (DLS)

Particle size distributions of adenovirus vectors and protein nanoparticles were obtained from a Zetasizer Ultra instrument. SEC-purified SpFN and Spp.P2A-FN from 293 F cells were diluted to 0.25 mg/mL using 1×PBS buffer, and 40 µL of the prepared NP sample was added to a Disposable Solvent Resistant Micro Cuvette (Malvern, catalog no. ZEN0040). Particle size was measured at 25 °C in back scattering mode. Data processing was performed on the Zetasizer, and the particle size distribution was plotted using GraphPad Prism 9.5.1 software.

Animal study

For the long-term immunogenicity study, groups of C57BL/6J mice aged 6–8 weeks (n = 10, half male and half female) were primed at week 0 by intramuscular injection with 1 × 10⁸ vp of an Ad5 vector vaccine. At week 10, all animals were boosted with 1 × 10⁸ vp of the Ad5 vector vaccine administered via the intranasal route. At week 20, spleens and lungs were collected and processed into single-cell suspensions for flow cytometric analysis of vaccine-induced immune responses.

For the analysis of early immune responses study, groups of female C57BL/6J mice aged 6–8 weeks (n = 5) were immunized at week 0 with 1 × 10⁸ vp Ad5-SpFN or Ad5-Spp, injected intramuscularly at a volume of 100 µL, or with 5 µg of SpFN or Spp protein formulated with Al(OH)₃. At 24 h post immunization, inguinal draining lymph nodes (LN) and injected muscle tissue were harvested to assess dendritic cell (DC) and B cell activation and migration. At 2 weeks post immunization, draining lymph nodes were collected to evaluate germinal center (GC) responses, and spleens were harvested for the assessment of systemic T cell and B cell responses.

To evaluate mucosal immune responses study, groups of female C57BL/6J mice aged 6–8 weeks (n = 5) was immunized at week 0 via the intranasal route with 50 µL of the indicated vaccine formulations. At 24 h post immunization, cervical draining lymph nodes were collected to examine DC and B cell activation and migration. Two weeks after intranasal immunization, lungs were harvested and analyzed for the presence of lung tissue-resident memory T cells (T_RM).

Western blotting

A549 cells were seeded in 6-well plates and, upon reaching 80–90% confluence, were infected with various adenoviral vector vaccines at a dose of 1 × 10⁸ vp per well. After a further 48 h of incubation, cells and culture supernatants were collected and lysed in 500 µL of RIPA buffer supplemented with a protease inhibitor cocktail. The resulting cell lysates and supernatants were separated by SDS-PAGE and transferred onto 0.22 μm polyvinylidene difluoride membranes. Membranes were blocked with nonfat milk in tris-buffered saline supplemented with 0.05% Tween-20 and then incubated with a primary antibody (rabbit polyclonal antibody against SARS-CoV-2 S2, 0.1 µg/mL, Sino Biological, China) for 1.5 h at room temperature, followed by incubation with DyLight 800-conjugated goat anti-rabbit IgG secondary antibody for 45 min. Images were captured with an Odyssey XF Imaging System (Odyssey M, LI-COR).

Isolation of serum, LN, lung, spleen, and BALF samples

Isolation of serum, bronchoalveolar lavage fluid (BALF), and tissues. Prior to tissue collection, mice were euthanized by CO₂ asphyxiation. Blood samples were collected at the indicated time points by submandibular bleeding, and sera was heat-inactivated at 56 °C for 30 min and then stored at −80 °C until analysis. For BALF collection, the lungs were washed three times with 600 µL PBS. BALF samples were clarified by centrifugation and the supernatant were stored at −80 °C. Spleens were gently mashed through a 70 μm cell strainer into cold RPMI 1640 using the plunger end of a 1 mL syringe. The cell suspension was isolated by density gradient centrifugation using Ficoll-Paque PLUS. Cells were then washed, passed through a fresh 40 μm strainer to remove clumps, counted using automated cell counter, and resuspended at the RPMI 1640 medium containing 10% FBS (R10 buffer). Draining lymph nodes (inguinal, cervical) were collected and mechanically dissociated by gently pressing them through a 70 μm cell strainer into RPMI 1640 with 2% FBS. The resulting cell suspension was centrifuged at 500 g for 5 min, the supernatant was discarded, and the cell pellet was resuspended in staining buffer. To reduce blood contamination, the pulmonary vasculature was perfused in situ with 20 mL of cold PBS via the right ventricle until the lungs blanched. Lungs were excised, minced into small pieces using sterile scissors, and transferred into digestion medium consisting of RPMI 1640 containing 2% FBS, collagenase D (1 mg/mL), and DNase I (0.1 mg/mL; Sigma-Aldrich or equivalent). Lung tissue was digested at 37 °C for 45 min with gentle agitation. The cell suspension was isolated by density gradient centrifugation using Ficoll-Paque PLUS. Cells were then washed, passed through a fresh 40 μm strainer to remove clumps, counted using automated cell counter, and resuspended at the R10 buffer for downstream flow cytometric analysis, including the assessment of tissue-resident memory T cells.

Enzyme-linked immunosorbent assay

An enzyme-linked immunosorbent assay (ELISA) was performed to determine antigen-specific binding antibody titers against SARS-CoV-2 RBD, full-length spike, S2 subunit, and the adenoviral vector. Briefly, 96-well high-binding ELISA plates were coated overnight at 4 °C with 100 µL per well of recombinant SARS-CoV-2 RBD protein (in-house produced), spike protein (Sino Biological, Cat# 40589-V08H10), S2 subunit protein, or inactivated Ad5 vector (2.5 × 10⁷ vp per well) diluted in coating buffer. The following day, plates were washed with PBS containing 0.5% Tween-20 (PBST) and blocked with 200 µL per well of 5% nonfat milk in PBST for 2 h at room temperature (RT). After blocking, three-fold serial dilutions of mouse serum or BALF samples (typically starting at 1:100) were added to the plates and incubated for 2 h at RT. Plates were then washed four times with PBST before incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies diluted in PBST containing 5% nonfat milk. Total IgG and IgA were detected using HRP-conjugated anti-mouse IgG or IgA, respectively, whereas IgG subclasses were assessed using HRP-conjugated anti-mouse IgG1 and IgG2a antibodies (all from Sino Biological, China), at dilutions optimized for each antibody (typically 1:4000 for anti-IgG). Following 1.5 h of incubation at RT, plates were washed four times with PBST, and 50 µL of TMB peroxidase substrate solution was added to each well. After a 5–10 min development at RT, reactions were stopped by adding 50 µL 1 M H₂SO₄, and absorbance was read at 450 nm using a Synergy H1 microplate reader (Agilent BioTek, Beijing, China).

Pseudovirus neutralization assay

A pseudovirus neutralization assay was performed to evaluate the neutralizing activity of antibodies in serum and BALF samples [36]. Lentiviral pseudoviruses bearing codon-optimized full-length SARS-CoV-2 spike proteins corresponding to the D614G, Beta, Delta, and Omicron BA.1 variants were generated as described previously. HEK293T-hACE2 cells were seeded into 96-well plates one day before infection. Starting at a 1:50 dilution, serum and BALF samples were subjected to a threefold serial dilution in the assay medium in 96-well plates. An equal volume of pseudovirus was added to each diluted sample, and the virus–antibody mixtures were incubated for 1 h at 37 °C. The mixtures were then transferred onto HEK293T-hACE2 cell. After 48 h of incubation, the culture supernatant was discarded and cells were lysed in 1% Triton X-100, followed by addition of luciferase substrate according to the manufacturer’s instructions. The half-maximal neutralization titer (NT₅₀) was defined as the reciprocal of the highest sample dilution that resulted in a 50% reduction of the mean RLU of the pseudovirus control.

Intracellular cytokine staining (ICS)

PBMC, Spleen and Lungs of immunized mice were harvested and processed. Lung, Spleen and PBMC single-cell suspensions were washed and resuspended in R10. Cells were stimulated with 2 µg/mL spike peptide pool for 1 h at 37 °C before adding Golgi Plug at a final dilution of 1:1000. After a further 6–8 h stimulation, cells were washed in MACS buffer and surface stained with Live/Dead dye Fixable Viability Stain 780, anti-CD16/32 for 15 min, then anti-mouse stained with CD3, CD4, CD8, CD44, CD62L, for 30 min at 4 °C. After washing twice with MACS, cells were fixed and permeabilized for 30 min with BD Cytofix/Cytoperm at 4 °C. After washing twice with perm/wash buffer, cells were stained with IFN-γ (clone XMG1.2; BV711), IL-2 (clone JES6-5H4; PE) and TNFa (clone MP6-XT22; BB 700) (all antibodies diluted 1:100) for 1 h. Cells were then washed twice with 1×Perm Wash buffer. Fixed cells were analyzed by the Agilent Novocyte Advanteon flow cytometer. Data were analyzed using FlowJo v10.8.1.

Enzyme-linked immunospot assay (ELISpot)

IFN-γ ELISpot: Plates were coated with an anti-mouse IFN-γ capture antibody overnight. After blocking, splenocytes (5 × 10⁵ cells/well) were added and stimulated with SARS-CoV-2 spike peptide pools (2 µg/mL) for 24 h. Spots were developed using a biotinylated detection antibody, streptavidin-HRP, and an ACE substrate. Spots were counted using an automated ELISpot reader.

SARS-CoV-2-specific B cell ELISpot: Plates were coated with recombinant spike protein (10 µg/mL). After blocking, splenocytes (1 × 10⁶ cells/well) were added and incubated for 24 h. Antigen-specific IgG or IgA-secreting cells were detected using an anti-mouse IgG antibody, followed by an HRP-conjugated secondary antibody and ACE substrate.

Flow cytometry

For the detection of SARS-CoV-2 spike–specific CD8⁺ T cells, splenocytes were first stained with a fixable viability dye and incubated with Fc receptor blocking reagent. SARS-CoV-2 spike–specific Major Histocompatibility Complex class I (MHC-I) tetramers (H-2Kᵇ: VNFNFNGL) were then added at a final concentration of 5 µg/mL and incubated for 30 min at 4 °C in the dark. After tetramer staining, cells were surface-stained with antibodies against CD3, CD8, CD44, and CD62L for 30 min at 4 °C. For the analysis of T_RM, single-cell suspensions from lung tissue were stained with a fixable viability dye and Fc block, followed by surface staining with antibodies against CD3, CD4, CD8, CD69, and CD103. For DC phenotyping, cells isolated from inguinal and cervical lymph nodes were treated with a fixable viability dye and Fc block, and subsequently stained for surface markers using antibodies against CD11c, CD8α, CD103, CD11b, CCR7, CD86, and MHC-II. For B cell staining, Cells were stained with Fc block and Live/Dead dye Fixable Viability Stain 780, After washing, SARS-CoV-2-spike specific B cells were identified by staining with 1 µg/mL of RBD tetramer at 4 °C for 30 min. All consecutive steps were done either at 4 °C or on ice and washed twice. The cocktail for staining contained CD45, CD45R, Fas, GL-7, IgG1, CD38, IgD, IgM, CD138. The cells were incubated with the staining cocktail at 4 °C for 30 min, then fixed and permeabilized in FoxP3/Transcription Factor Staining Buffer Set according to manufacturer’s instructions before intranuclear staining with Ki67 for 30 min. For follicular helper T (Tfh) and follicular regulatory T (Tfr) cell staining, Cells were first incubated for 30 min at 4 °C with fixable Viability Dye 780 and an Fc receptor block reagent. After washing, surface staining was performed with antibodies against CD3, CD4, CD44, PD-1, CXCR5, and CD40L for 30 min at 4 °C. Cells were then washed with FACS buffer and fixed and permeabilized using the FoxP3/Transcription Factor Staining Buffer Set according to the manufacturer’s instructions, followed by intranuclear staining with Ki67 (and FoxP3 for Tfr staining) for 30 min at 4 °C. All incubations were carried out at 4 °C.

Immunofluorescence staining of lymph node sections

Inguinal LN were harvested into PBS and fixed with 4% PFA plus 10% Sucrose for 30 min at room temperature. Following three washes with PBS, samples were equilibrated in 30% sucrose for 1 h to overnight. Samples were washed three times with PBS and frozen in cryomolds using Optimal Cuting Temperature compound (OCT, Fisher Heathcare 23–730−571) in a dry ice-cooled bath of 2-methylbutane. 10 µm tissue sections were blocked for 30 min at room temperature in staining solution (5% donke’ serum in PBS-Tween) in the presence of Fc receptor block reagent. Sections were stained with the following antibodies: CD21/35, GL7, CD3, IgD. To amplify the signal of the spike-biotin probe, sections were incubated with a Streptavidin-AF647. After staining, slides were mounted with ProLong Diamond Antifade Mountant (Life Technologies). Tiled images were acquired at 10× magnification with a 1 airy unit pinhole using a Zeiss LSM 980 confocal microscope and Zen Black software.

Cell proliferation assay

Spike antigen–specific T-cell responses were assessed using a carboxyfluorescein succinimidyl ester (CFSE) dilution assay. Briefly, freshly isolated splenocytes were resuspended at 1 × 10⁶ cells/mL in pre-warmed 1×PBS and labeled with CFSE using the CellTrace^™ CFSE Cell Proliferation Kit (Invitrogen, Thermo Fisher Scientific) according to the manufacturer’s instructions. The labeling reaction was quenched by adding R10, and cells were washed once. CFSE-labeled cells (200 µL per well) were plated in 96-well round-bottom plates and cultured for 48 h at 37 °C in a humidified incubator with 5% CO₂ in the presence of the spike-peptides stimulation. After culture, cells were harvested, centrifuged to remove the supernatant, and stained first with a fixable viability dye and Fc receptor blocking reagent. Cells were then surface-stained with monoclonal antibodies against CD3, CD4, and CD8. Proliferation of CD4⁺ and CD8⁺ T cells was quantified by CFSE dilution by flow cytometry. Proliferation of B cells in response to spike antigen was evaluated using splenocytes. CFSE-labeled splenocytes were cultured in 96-well round-bottom plates at 1 × 10⁶ cells/mL and stimulated with recombinant spike protein (1 µg/well) in the presence of IL-4 (10 ng/mL) for 72 h at 37 °C and 5% CO₂. Following culture, cells were collected, centrifuged to remove the supernatant, and stained with a fixable viability dye and Fc receptor blocking reagent, followed by surface staining with monoclonal antibodies against CD45R, IgG, and IgD. Proliferation of CFSE-labeled naïve (IgD⁺IgG⁻) and memory (IgD⁻IgG⁺) B-cell subsets was analyzed by flow cytometry based on CFSE dilution, revealing that memory B cells underwent more rounds of cell division than naïve B cells in response to spike antigen stimulation.

SARS-CoV-2 challenge studies

6–8-week-old K18-hACE2 transgenic mice were immunized intramuscularly with one of the following formulations: 1.5 × 10⁸ vp Ad5-SpFN, 1.5 × 10⁸ vp Ad5-Spp, 7.5 µg SpFN adjuvanted with 50 µg aluminum hydroxide, or 7.5 µg Spp adjuvanted with 50 µg aluminum hydroxide. Additional groups received IN immunization with 1.5 × 10⁸ vp Ad5-SpFN, 1.5 × 10⁸ vp Ad5-Spp, or 7.5 µg SpFN formulated with 50 µg aluminum hydroxide. Two weeks after immunization, animals were terminally bled and spike-specific antibody titers were determined. Mice were then challenged in an ABSL-3 facility with 1.5 × 10³ PFU of authentic SARS-CoV-2 virus (Wuhan-Hu-1) administered IN in a final volume of 40 µL PBS. Body weight and clinical signs were monitored daily for 6 days post infection. Animals were humanely euthanized when they lost ≥ 20% of their initial body weight. At 6 days post infection, lungs were collected for histological analysis, immunohistochemistry, and viral load determination.

Quantitative reverse transcription polymerase chain reaction (RT-qPCR)

Lung tissues from individual mice were collected and homogenized in steel bead lysis tubes. Total RNA was extracted from the homogenized tissues using RNAiso Plus (Takara, Cat# 9108) according to the manufacturer’s instructions. Viral RNA copies in lung tissues were then quantified by one-step RT-qPCR using the SARS-CoV-2 RNA detection kit Evo M-MLV One Step RT-qPCR Kit II (Probe) (AG11715), following the manufacturer’s protocol. To generate a standard curve, the SARS-CoV-2 nucleocapsid (N) gene was cloned into a pcDNA3.1 expression plasmid and in vitro transcribed to obtain RNA standards. RT-qPCR reactions were performed on a CFX96 Touch Real-Time PCR Detection System (Bio-Rad) under the following cycling conditions: reverse transcription at 45 °C for 15 min, initial denaturation at 95 °C for 15 min, followed by 43 cycles of 95 °C for 15 s and 60 °C for 30 s. Viral RNA loads in lung tissue homogenates were expressed as copies per mL. The N gene–specific primers and probe were as follows:

N-F: 5’-CAGTAGGGGAACTTCTCCTGCT-3’;

N-R: 5’-CTTTGCTGCTGCTTGACAGA-3’;

N-P: 5’-FAM-CTGGCAATGGCGGTGATGCTGC- BHQ1-3’.

Histopathological analysis

Lungs collected at 6 days post challenge were removed and immersed in 4% neutral buffered formalin for 7 days. After fixation, tissues were processed routinely, embedded in paraffin wax, and sectioned at 5 μm thickness. Transverse sections were taken at the level of the maximal cross-sectional area of the right lung lobe. The sections were then stained with hematoxylin and eosin (H&E) for histopathological evaluation. Finally, sections were dehydrated, cleared, mounted, and examined under a light microscope.

Statistical analyses

Statistical analyses were performed using GraphPad Prism 9.5.1 (GraphPad Software). Data are presented as median with interquartile range (IQR) and represent duplicate measurements. Comparisons between multiple groups were performed using one-way ANOVA with two-sided Mann-Whitney U tests. A p-value less than 0.05 was considered statistically significant.

Results

Strategic design and in vitro characterization of adenoviral-vectored nanoparticle vaccines

To establish proof of concept, we used the full-length SARS-CoV-2 spike as a model antigen, engineering its coding sequence into the ΔE1 region of an Ad5 backbone to create replication-defective adenoviral vectored nanoparticle vaccines. All constructs encoded an N-terminal signal peptide and a His-tag followed by a prefusion-stabilized spike ectodomain (Spp) (Fig. 1A). Based on this backbone, we generated two distinct self-assembling nanoparticle immunogens: SpFN, a ferritin-conjugated nanovaccine we recently developed [10–17], and Spp.EPM-EABR, a novel design displaying Spp on engineered extracellular vesicles [37]. As a control, Spp.P2A-FN was designed with a self-cleaving P2A sequence between Spp and ferritin, permitting ferritin assembly without spike conjugation. Purified recombinant adenoviral vectors showed uniform size (~120 nm in diameter) and morphology by transmission electron microscopy (TEM) (Fig. S1A, S1B), similar particle-to-infectious titer ratios, along with expected viral protein expression profiles (Fig. S1C).

Fig. 1.

Strategic design and in vitro characterization of adenoviral-vectored nanoparticle vaccines. (A) Schematic illustration of four vaccine antigen designs. The SARS-CoV-2 spike (S) protein ectodomain was engineered with an N-terminal signal peptide, a His-tag, a deleted furin cleavage site (dFurin), and proline substitutions (K986P, V987P) to stabilize its pre-fusion conformation. This construct was fused to Helicobacter pylori ferritin gene sequences and synthesized in a codon-optimized DNA plasmid. (B) The antigen expression of the Ad5 vectored vaccine (Ad5-SpFN, Ad5-Spp, Ad5-Spp.P2A-FN and Ad5-Spp.EPM-EABR) in A549 cells and culture supernatant was analyzed 48 h after infection by western blot. (C) Size-exclusion chromatography (SEC) of SpFN and Spp.P2A-FN. The ultraviolet absorptions at 280 nm (OD₂₈₀) were shown. (D) Purified SpFN and Spp.P2A-FN were analyzed under both denaturing (+β-ME) and native (-β-ME) conditions. The ratio of super high molecular weight (SHMW) species to total protein was determined. (E) The particle size of the SpFN and Spp.P2A-FN was measured by a Malvern particle size instrument

Spike expression was quantified in A549 cell lysates and supernatants after transduction at identical multiplicities of infection (MOI). Western blot analysis revealed that SpFN was secreted into the culture supernatant, although a substantial fraction remained intracellular. The other three vectors showed moderately higher secreted spike levels in supernatant, with comparable intracellular expression (Fig. 1B). The secreted SpFN exhibited higher molecular weight than Spp or Spp-P2A-FN, consistent with fusion to ferritin or EPM-EABR domains (Fig. 1B).

To assess nanoparticle assembly, secreted particles were purified by SEC. SpFN eluted as a high-molecular weight peak, distinct from the peak of Spp.P2A-FN (Fig. 1C). Moreover, under native conditions, a super high-molecular-weight fraction account for 41.8% of SpFN was retained at the stacking/running gel interface, accompanied with a band for the monomer account for 58.2% (Fig. 1D, right), confirming the secretion of stable multimers. However, Spp.P2A-FN was almost entirely composed of monomers. TEM revealed correctly assembled nanoparticles approximately 50 nm in diameter (Fig. S1D), a size corroborated by dynamic light scattering (average ~57 nm) (Fig. 1E). Collectively, these data demonstrate that Ad5-vectored delivery enables efficient secretion and in vivo assembly of spike-based nanoparticles.

Ad5-SpFN induces potent and durable antibody responses

To evaluate the immunogenicity, four groups of 6–8-week-old C57BL/6J mice (n = 10) were immunized intramuscularly with 1 × 10⁸ vp of Ad5-SpFN, Ad5-Spp, Ad5-Spp.P2A-FN, and Ad5-Spp.EPM-EABR, respectively. A booster dose was administered at week 10 to assess relative long-term immunogenicity. Serum samples were collected every two weeks following primary or booster immunization (Fig. 2A). Longitudinal ELISA measurements of RBD-specific IgG revealed that the Ad5-SpFN generated the highest antibody titers, which were significantly greater than those of other groups as early as week 2 (Fig. 2B). While the antibody levels in other regimens gradually increased, Ad5-SpFN consistently maintained high antibody titers throughout the study. After the week 10 booster, binding antibodies increased in all groups, with Ad5-SpFN sustaining the highest antibody titers through at least week 20 (Fig. 2B).

Fig. 2.

Ad5-nanoparticle vaccines elicit potent and durable antibody responses. (A) Immunization schema for evaluating the Ad5-SpFN, Ad5-Spp.P2A-FN, Ad5-Spp.EPM-EABR. 6–8-week-old female naive C57BL/6J mice (n = 10/group) were vaccinated with two doses (1 × 10⁸ vp) of the Ad5 vectored vaccines. Mice received an intramuscular prime at week 0 followed by an intranasal boost at week 10. Blood and tissue samples were collected at the indicated time points. (B) SARS-CoV-2 RBD-binding IgG titers in sera of immunized mice at each time point were determined by endpoint dilution ELISA. (C-D) Neutralizing antibody activity against D614G (C) and Delta (D) were assessed by pseudovirus neutralization assays. (E-F) Spike-specific T cell intracellular cytokine staining (ICS) assays. PBMCs at week 10 post-immunization were stimulated with the spike peptide pools at a final concentration of 2 µg/mL. The proportions of IFN-γ⁺, IL-2⁺, and TNF-α⁺ CD4⁺ (E) and CD8⁺ (F) T cells were measured by flow cytometry. (G) Representative images (left) and quantification (right) of IFN-γ spot-forming units (SFU) per 10⁶ PBMCs after restimulation with spike peptide pools. (H-I) IFN-γ ELISpot at week 20 from spleen (H) or IgA ELISA from BALF (I) were determined. Data are presented as median ± IQR. Statistical significance was determined by two-sided Mann-Whitney U tests for comparisons between groups. Dashed lines indicate the lower limit of quantification. (LLOQ: 100 for ELISA, 50 for NAb)

Functional antibody responses were further characterized using a D614G pseudovirus neutralization assay (pNAb), which confirmed superior neutralizing activity in the Ad5-SpFN group (Fig. 2C). Against the Delta variant, Ad5-SpFN induced detectable NAbs by week 2, while the other vaccines elicited titers below the limit of detection (NT₅₀< 20) at this timepoint. Essentially, over the 20-week observation period, the Ad5-SpFN group maintained the highest NAb titers (Fig. 2D), a pattern consistent across serological assays.

At week 10, peripheral blood mononuclear cells (PBMCs) were harvested to evaluate T-cell responses. Intracellular cytokine staining showed that the spike-specific IFN-γ, IL-2 and TNF-α production in CD4⁺ and CD8⁺ T cells did not differ significantly among vaccine groups, though all were substantively higher than those in the Sham control group (Fig. 2E-F). In agreement with the ICS data, IFN-γ ELISpot analysis of total T cell responses showed a similar trend (Fig. 2G), which remained durable through week 20 (Fig. 2H). RBD-specific IgG levels in BALF were highest in Ad5-SpFN group, even though no statistically significant differences were observed among the groups (Fig. 2I).

Collectively, these data indicate that Ad5-SpFN elicits stronger humoral and cellular immune responses than monomeric or eVLP‑based designs, supporting its potential as a promising vaccine candidate.

Ad5-nanoparticle vaccines outperform protein subunit vaccines

Given the higher immunogenicity conferred by SpFN than Spp.EPM-EABR, the former were further chosen for mechanistic characterizations. To assess the immunologic features of vaccine platforms, we next directly compared the immunogenicity of Ad5-SpFN with alum-adjuvanted protein-based vaccine regimens (Fig. 3A). Mice were immunized with the adenoviral vector vaccines, Ad5-SpFN (nanoparticle) or Ad5-Spp (monomer), or with alum-adjuvanted protein subunit vaccines, SpFN (nanoparticle) or Spp (monomer) (Fig. 3A). Blood was collected at week 1, 2, and 4 post-immunization to assess spike-specific antibody and cellular immune responses.

Fig. 3.

Ad5-nanoparticle vaccines generate superior neutralizing antibody titers and robust T cell immunity compared to vectored-subunit and protein vaccines. (A) Immunization schema for evaluating humoral and cellular immunity. C57BL/6J mice were immunized on day 0 with Ad5-SpFN or Ad5-Spp adenoviral vectored vaccines (1 × 10⁸ vp) or with SpFN or Spp protein subunit vaccines adjuvanted with alum. Blood was collected at 1-, 2-, and 4-weeks post-immunization. Spleens and lymph nodes were collected two weeks after immunization for immune cell profiling. (B) SARS-CoV-2 RBD-binding IgG titers in sera from mice at week 1–4 post-immunization (B) and Omicron-specific titers at week 2 post-immunization (C) were measured by ELISA. (D) RBD-specific IgG subclass analysis was performed and IgG1 and IgG2a titers in serum at week 4 were determined by endpoint dilution ELISA. The ratio of IgG2a to IgG1 is compared across groups. (E-F) NAb titers against the D614G (E) and Delta (F) variants were assessed at the indicated time points. (G-I) Proportions of IFN-γ⁺, IL-2⁺, and TNF-α⁺ CD4⁺ (G) and CD8⁺ (H) T cells after stimulation with a spike peptide pool. Polyfunctionality, defined as co-expression of ≥ 2 cytokines (IFN-γ, IL-2, TNF-α) in live CD3⁺CD4⁺CD44⁺ or CD3⁺CD8⁺CD44⁺ T cells, is shown in a pie chart (I). (J-L) Memory T cell analysis. Proportions of T_EM and T_CM CD8⁺ T cells in the spleen (J) and their frequencies among spike-specific (tetramer⁺) cells (K). Proportions of CD4⁺ T_EM and T_CM cells are also shown (L). (M-N) Splenocytes from week 2 were labeled with CFSE and restimulated for 48 h with a spike peptide pool (4 µM). Proliferation was measured by CFSE dilution in CD8⁺ (M) and CD4⁺ (N) T cells via flow cytometry. Data are presented as median ± IQR. Statistical significance was determined by two-sided Mann-Whitney U tests for comparisons between groups

ELISA measurements showed that a single dose of Ad5-SpFN generated the highest serum antibody titers over a 4-week period (Fig. 3B). Notably, just one week after immunization, Ad5-SpFN induced spike-specific antibody titers that were 84.1-, 5.25-, and 354.4-fold higher than those in the Ad5-Spp, SpFN, and Spp groups, respectively, highlighting its early capacity to elicit high-titer binding antibodies (Fig. 3B). Detailed antibody titers for each time point are shown in Fig. S2A. A similar trend was observed for antibodies against the Omicron-Spp protein (Fig. S2B). To explore whether the antibody binding breadth can be derived from S2-directed antibody expansion, we also measured binding antibody titers against the D614G-S2 and Omicron-S2 subunits. While all groups showed comparable responses to WT-S2 (Fig. S2C), the Ad5-SpFN group produced Omicron-S2 specific antibody levels that were 17.2-, 140-, and 4.9-fold higher than those in the Ad5-Spp, SpFN, and Spp groups, respectively, underscoring the role of S2-specific antibodies in broadening coverage (Fig. 3C). As expected, the alum-adjuvanted protein vaccines SpFN and Spp predominantly induced Th2-biased IgG1 antibodies. In contrast, the adenoviral vector vaccines Ad5-SpFN and Ad5-Spp elicited not only high IgG1 titers but also robust Th1-biased IgG2a responses, indicating a more balanced immune profile (Fig. 3D). In alignment with the binding antibody data, Ad5-SpFN also induced stronger NAb against both the D614G and Delta variant compared to other groups (Fig. 3E-F). Detailed pNAb titers for each time point is shown (Fig. S2D-E). The neutralization area under curve (AUC) against the Delta variant was 16.6-, 3.1-, and 16.6-fold greater for Ad5-SpFN than for Ad5-Spp, SpFN, and Spp, respectively (Fig. 3F).

Together, these findings demonstrate that the Ad5-delivered nanoparticle vaccine outperforms protein-based counterparts in triggering early, potent, and broadly neutralizing antibody responses.

Ad5-nanoparticle vaccination stimulates robust T cell immunity

To evaluate T cell immunogenicity, splenocytes from immunized mice were analyzed two weeks post-immunization. Adenoviral vector vaccines (Ad5-SpFN and Ad5-Spp) elicited markedly higher frequencies of spike-specific IL-2⁺, IFN-γ⁺, or TNFα⁺ CD4⁺ and CD8⁺ T cell compared with alum-adjuvanted protein subunit vaccines (Fig. 3G-H). Polyfunctional cytokine-producing T cells are also assessed as an indicator of T cell quality. Ad5-SpFN and Ad5-Spp induced 0.544% and 0.677% polyfunctional CD4⁺ T cells, respectively, compared to only 0.095% in both protein vaccine groups. Within this polyfunctional CD4⁺ T subset, IFN-γ⁺IL-2⁺TNF-α⁺ triple-positive cells accounted for 20.99% and 24.55% in the Ad5-SpFN and Ad5-Spp groups, respectively, but less than 0.1% in the protein vaccine groups (Fig. 3I, left). A similar pattern was observed for CD8⁺ T cells, where Ad5-SpFN and Ad5-Spp induced 0.36% and 0.68% polyfunctional cells, respectively, versus 0.053% and 0.095% in the protein vaccine groups. Within this polyfunctional CD8⁺ T subset, IFN-γ⁺IL-2⁺TNF-α⁺ triple-positive cells accounted for 12.89% and 10.86% in the Ad5-SpFN and Ad5-Spp groups, respectively, but less than 0.1% in the protein vaccine groups (Fig. 3I, right).

Adenoviral vectors also induced higher frequencies of CD4⁺ and CD8⁺ T effector memory (T_EM) and T central memory (T_CM) cells (Fig. 3J-L). Ad5-SpFN and Ad5-Spp induced median frequencies of 34.8% and 30.4% CD8⁺ T_EM and T_CM cells, respectively, compared with 23.0% and 22.4% in protein vaccine groups (Fig. 3J). To further quantify antigen-specific CD8⁺ T cells, we used MHC-I tetramers for the Spike_539−546 epitope. The frequencies of tetramer⁺ CD8⁺ T_EM cells were 11.4%, 13.3% in the Ad5-SpFN, Ad5-Spp groups, respectively, but only 0.45%, and 1.35% in SpFN and Spp groups, respectively. A corresponding advantage was also observed for tetramer⁺ CD8⁺ T_CM cells, reinforcing the superior capacity of adenoviral vectors to induce antigen-specific CD8⁺ T cell responses (Fig. 3K). Similarly, CD4⁺ T_EM frequencies reached 37.0% and 31.7% in Ad5-SpFN and Ad5-SPP groups, versus 25.2% and 27.6% in the SpFN and Spp groups, respectively (Fig. 3L). Of note, within the same delivery platform (adenoviral or protein), nanoparticle versus monomer formats did not significantly alter T_CM or T_EM frequencies (Fig. 3J-L).

To assess immune memory, we performed an antigen-specific recall proliferation assay using CFSE dye dilution to quantify responding T cells. After stimulation with a spike peptide pool, Ad5-SpFN- and Ad5-Sp-primed CD8⁺ T cells showed robust proliferation (76.8% and 81.8% CFSE^low), exceeding responses in protein-vaccinated groups (Fig. 3M). CD4⁺ T-cell proliferation was also strongest in the Ad5-SpFN group (30.0% CFSE^low), consistent with its role in supporting B-cell memory (Fig. 3N).

In summary, these results demonstrate that adenoviral vectors elicit substantially stronger and more polyfunctional T-cell immunity than alum-adjuvanted protein subunit vaccines, irrespective of antigen format.

Ad5-nanoparticle vaccination promotes rapid germinal center B cell reactions

The rapid and robust antibody response elicited by Ad5-SpFN led us to investigate the underlying GC response, a key driver of high-quality humoral immunity. We quantified Tfh and GC B cells in the spleen 7 days post-primary vaccination by flow cytometry (Fig. S3A-B).

Compared to the protein subunit vaccines (SpFN and Spp), adenoviral vector vaccines (Ad5-SpFN and Ad5-Spp) induced higher frequencies of GL7⁺Fas⁺ GC B cells and CD45R⁻CD138⁺ plasma cells, and comparable levels of IgG1⁺CD38⁺ memory B cells in the draining lymph nodes (Fig. 4A-C). Notably, Ad5-SpFN vaccination generated a 5.06-fold higher frequency of spike-specific GC B cells than Ad5-Spp (Fig. 4D). Spike-specific plasma cells and memory B cells were also markedly elevated in the Ad5-SpFN group, showing 9.3- and 5.75-fold increase over Ad5-Spp, respectively (Fig. 4E-F). It has been well established that Tfh provide “fuel” for sustaining GCs [38–41], and that higher Tfh responses are associated with stronger, broader, and more durable neutralizing antibody responses. We therefore analyzed their expansion. The proportion of Tfh cells was 1.38-fold higher in the Ad5-SpFN group than in the Ad5-Spp, and 2.14- to 2.81-fold higher than in the protein vaccine groups, indicating a cellular environment conducive to strong GC reactions and neutralizing antibody production (Fig. 4G). To further characterize the Ad5-SpFN-induced Tfh cells, we analyzed key functional subsets within the Tfh lineage, CD40L⁺ Tfh, Ki67⁺ Tfh (proliferating), and FoxP3⁺ Tfr. Lymph nodes were collected two weeks post-immunization, and the Ad5-SpFN induced a significantly higher proportion of CD40L⁺ Tfh cells after S⁺ peptide pool stimulation compared to other groups (Fig. 4H), indicating a greater potential for CD40L-CD40 help to GC B cells. This enhanced help likely promotes GC B cell survival, class-switch recombination, and affinity maturation. Concurrently, Ad5-SpFN immunization resulted in elevated Ki67⁺ Tfh and reduced FoxP3⁺ Tfr frequency, suggesting an enhanced proliferation and a reduction in FoxP3⁺ Tfr-mediated suppression (Fig. 4I-J).

Fig. 4.

Ad5-nanoparticle vaccination promotes rapid germinal center B cell reactions. (A-C) Frequencies of B cell subsets in draining lymph nodes at week 2 post-immunization. Shown are GC B cells (CD45R⁺GL7⁺Fas⁺) (A), plasma cells (CD45R⁻CD138⁺) (B), and memory B cells (CD45R⁺IgG⁺CD38⁺) (C). (D–F) Frequencies of spike-specific GC B cells, identified with dual-color spike probes (Spike-AF594⁺/Spike-AF647⁺) (D), spike-specific plasma cells (E), and spike-specific memory B cells (F) within their respective gates. (G-J) Tfh and Tfr cell responses in dLNs at day 14. Frequencies of total Tfh cells (CD4⁺CD44^hiPD-1⁺CXCR5⁺) (G), CD40L⁺ Tfh (H), Ki67⁺ Tfh (I), and FoxP3⁺ Tfr cells (J) were shown. (K) Immunofluorescence analysis of dLN architecture. Representative images of frozen sections at week 2 show staining for IgD⁺ (mantle/naïve B cells), CD3⁺ (T cell zones), CD21/35⁺ (follicular dendritic cell network), and GL7⁺ (germinal centers) across groups. Scale bar: 200 μm. (L) Representative images depict spike⁺ B cells (blue) and their spatial distribution relative to GL7⁺ follicles (green) and CD21/35⁺ FDC networks (red). Scale bar: 200 μm. (M) B cell proliferation in response to spike protein. Splenocytes collected at week 2 were labeled with CFSE, stimulated for 48 h with spike protein (1 µg/well), and proliferation was measured as CFSE dilution within IgG⁺ B cells. (N) Experimental schema for evaluating immune responses under CD4⁺ T cell depletion. C57BL/6J mice (n = 5 per group) were administered anti-CD4 monoclonal antibody (GK1.5, 100 µg) every other day for 2 weeks, followed by immunization at week 0 with Ad5-SpFN or Ad5-Spp adenoviral-vectored vaccines (1 × 10⁸ vp) or with SpFN protein subunit vaccines (5 µg formulated with 50 µg Al(OH)₃). Beginning at week 5, mice in the anti-CD4-treated groups received an additional 1-week course of anti-CD4 (GK1.5) treatment and were boosted at week 6. Peripheral blood was collected at the indicated time points for antibody titer determination. (O) RBD-specific antibody titers in serum were measured by ELISA weekly from week 2 to week 7. (P) PNAb titers against the D614G variant were assessed at week 7. Each dot represents an individual animal. Data are presented as median ± IQR. Statistical significance was determined by two-sided Mann-Whitney U tests for comparisons between groups

Immunofluorescence staining of inguinal lymph nodes ten days post-immunization revealed that nanoparticle immunogens, particularly Ad5-SpFN, promoted more prominent GC formation than their monomeric counterparts (Fig. 4K). Staining with biotinylated spike protein further confirmed that Ad5-SpFN induced a higher frequency of spike-specific B cells, with greater colocalization to GL7⁺ GC regions, suggesting active class-switch recombination and affinity maturation (Fig. 4L). We further assessed antigen-specific B cell recall proliferation response. IgG⁺ B-cell proliferation upon spike protein re-stimulation was 1.41- to 1.99-fold higher in the Ad5-SpFN group than in controls (Fig. 4M).

CD4⁺ T cell dependency was further confirmed by depletion studies (Fig. 4N). Neither Ad5-SpFN nor the Ad5-Spp generated detectable antibody responses in CD4⁺ T cell depleted mice post prime (Fig. 4O). Following a booster under continued depletion, the Ad5-SpFN group still elicited binding and neutralizing antibody levels 2.56-fold higher than other CD4-depleted groups, underscoring the resilience of its primed B-cell pool (Fig. 4P).

These data collectively suggest that Ad5-SpFN uniquely fosters a larger and more diverse pool of spike-specific GC B cells, plasma cells, memory B cells, and Tfh cells, correlating with its superior magnitude and quality of antibody responses.

Ad5-nanoparticle immunization triggers early dendritic cell and B cell activation

The rapid emergence of high antibody titers following Ad5-SpFN vaccination led us to investigate early DC activation as DCs are critical for initiating adaptive immune responses, including the Tfh- GC axis. Inguinal lymph nodes and muscles from the injection site were collected 24 h after immunization for analysis by flow cytometry (Fig. S3C).

The frequency of CCR7⁺CD11c⁺ migratory DCs was elevated in all vaccine groups, with Ad5-SpFN showing a proportion of 4.66%, compared to 3.18% for Ad5-Spp, 4.93% for SpFN, and 1.95% for Spp (Fig. 5A). Ad5-SpFN also induced the highest frequencies of CD8⁺CD103⁺ cDC1 and CD8⁻CD11b⁺ cDC2 subsets (Fig. 5B-C), consistent with its enhanced early T cell activation (Fig. 3G-H).

Fig. 5.

Ad5-nanoparticle immunization triggers early dendritic cell and B cell activation. (A–C) Dendritic cell subset frequencies in dLNs at 24 h post-immunization were determined. Shown are the frequencies of CCR7⁺ DCs among total DCs (A), cDC1s (CD8⁺CD103⁺) among total conventional DCs (B), and cDC2s (CD11b⁺CD8⁻) among total DCs (C). (D–E) Frequencies of activated antigen-presenting cells. CD86⁺MHC-II⁺ cells among total DCs (D) and total B cells (E). (F) RT–qPCR analysis of dLN cell pellets at 24 h. Relative mRNA expression of CXCL10, IL6, and ISG15, normalized to housekeeping gene(s) and presented as fold change (2^–ΔΔCt) versus naïve controls. Data are presented as median ± IQR. Statistical significance was determined by two-sided Mann-Whitney U tests for comparisons between groups

Mature DCs, defined as CD86⁺MHC-II⁺, were more abundant in the Ad5-SpFN group, with proportions that were 2.7-, 1.67-, and 5.7-fold higher than in the Ad5-Spp, SpFN, and Spp groups, respectively. The nanoparticle format alone conferred an advantage, as SpFN also induced a 3.4-fold higher frequency than Spp (Fig. 5D). While cDC2s constitute a higher overall proportion of activated DCs, the Ad5-SpFN group induced a significantly higher proportion of activated cDC1s compared to the SpFN and Spp group (Fig. S8), indicating that Ad5-SpFN uniquely harnesses both the adjuvant effect of the Ad5 vector and the immunogenic properties of the SpFN nanoparticle to simultaneously mobilize both cDC1 and cDC2 subsets. This indicates that both the adenoviral platform and the nanoparticle antigen structure promote DC maturation, priming stronger downstream GC and antibody responses.

We next examined whether early B-cell activation was also accelerated. Ad5-SpFN induced CD86⁺MHC-II⁺CD45R⁺ B-cell frequencies that were 2.38-, 8.13-, and 17.2-fold higher than those in the Ad5-Spp, SpFN, and Spp group, respectively (Fig. 5E). To further characterize the early immune environment, we analyzed the expression of key immunomodulatory genes in draining lymph nodes. Il6, a critical signal for B-cell differentiation, was most strongly upregulated by Ad5-SpFN, showing 4.54-, 6.3-, and 30.6-fold higher expression than in the Ad5-Spp, SpFN, and Spp group, respectively (Fig. 5F). While Ad5-SpFN did not further elevate the Th1-associated genes Cxcl10 and Isg15 beyond levels induced by Ad5-Spp, both adenoviral vectors triggered significantly higher expression than the protein vaccines (Fig. 5F).

In addition, analysis of muscle tissue at the injection site revealed comparable frequencies of cDC1, cDC2, and CCR7⁺ DCs across vaccine groups. However, Ad5-SpFN again drove the highest proportion of CD86⁺MHC-II⁺ DCs (2.2- to 100-fold higher than other groups) and activated B cells (Fig. S4A-E), consistent with its enhanced early CD4⁺ T cell activation (Fig. S4F). Together, these data demonstrate that Ad5-SpFN promotes more robust and rapid DC migration, maturation, and early B-cell activation than other vaccine formats, creating an immunogenic milieu conducive to a superior humoral response.

Mucosal immunization of Ad5-SpFN establishes potent lung-resident immunity

To assess the potential of each vaccine platform to establish protective mucosal immunity, a key determinant in blocking respiratory infection and transmission, we compared immune responses following intranasal administration. Mice were immunized intranasally with Ad5-SpFN, Ad5-Sp (1 × 10⁸ vp), or SpFN protein (5 µg plus 50 µg Al(OH)₃). Two weeks later, BALF, lungs, lymph nodes, and spleens were collected for analysis (Fig. 6A).

Fig. 6.

Intranasal immunization induces robust mucosal and systemic humoral and cellular immune responses. (A) Timeline for mucosal immunity analysis. 6–8-week-old C57BL/6J mice received an intranasal immunization on day 0 with Ad5-SpFN or Ad5-Spp adenoviral vectored vaccines (1 × 10⁸ vp) or with SpFN protein subunit vaccines (5 µg adjuvanted with 50 µg Al(OH)₃) in 50 µL PBS. cLNs were collected 24 h post-immunization for DC analysis. Lungs, lymph nodes, and spleens were harvested at day 14 for immune cell profiling. (B-C) RBD-specific IgA and IgG titers at week 2 were measured by ELISA (B). Neutralizing antibody titers against D614G and Delta variants are also shown (C). (D-E) Frequency and absolute number of CD3⁺ T cells (D) and CD45R⁺ B cells (E) in BALF at day 14. (F-H) Lung B cell and antibody-secreting cell responses were further analyzed and frequencies of IgG⁺ B cells and spike-specific (S⁺) IgG⁺ B cells (F), S-specific IgA-secreting cells by ELISpot (G), and IgG1⁺CD38⁺ memory B cells (MBCs) and S⁺ MBCs (H) were shown. (I-K) Frequencies of CD69⁺ CD103⁺ CD4⁺ T_RM (I), CD8⁺ T_RM (J) in the lung and Spike_539−546-tetramer⁺ cells within the CD8⁺ T_RM (K) was shown. Proportions of GL7⁺ Fas⁺ GC B cells and S⁺ GC B cells (L), IgG1⁺CD38⁺ MBCs and S⁺ MBCs (M), and PD-1⁺CXCR5⁺ Tfh cells among CD44⁺ CD4⁺ T cells (N) were determined. Data are presented as median ± IQR. Statistical significance was determined by two-sided Mann-Whitney U tests for comparisons between groups. Dashed lines indicate the lower limit of quantification (LLOQ: 20 for ELISA, 15 for NAb)

Ad5-SpFN induced the highest levels of spike-specific IgG and IgA in both BALF and serum (Fig. 6B and S5A-D). Notably, no spike-specific IgA was detected in the SpFN group. Neutralizing activity in BALF was detectable only in the Ad5-SpFN group, while in serum, Ad5-SpFN elicited D614G neutralizing titers that were 4.04- and 209.2-fold higher than those in the Ad5-Spp and SpFN groups, respectively (Fig. 6C and S5E-F). The SpFN group showed minimal neutralization and failed to neutralize the Delta variant. Of note, intranasal immunization with Ad5-SpFN elicited neutralizing antibodies against BA.1 that were more than 10-fold higher than those in the control groups (Ad5-Spp and SpFN alone). This suggests that the combination of the SpFN nanoparticle platform with the adenoviral vector delivery system confers an immunogenic advantage, enabling the vaccine to maintain cross-neutralization capacity against emerging variants to a greater degree than single-platform vaccines (Fig. S5E).

Analysis of BALF cellularity revealed robust lymphocyte recruitment by adenoviral vectors. CD3⁺ T cells comprised > 50% of BALF lymphocytes in Ad5-SpFN- and Ad5-Spp-immunized mice, compared to only 4.3% in the SpFN group. The absolute number of CD3⁺ T cells was 40.6- and 30.3-fold higher in the Ad5-SpFN group than in the Ad5-Spp and SpFN groups, respectively (Fig. 6D). Similarly, B cells frequencies and numbers were significantly elevated in the Ad5-SpFN group (Fig. 6E), correlating with the observed mucosal antibody titers.

In lung tissues, Ad5-SpFN generated a higher frequency of spike-specific IgG⁺ B cells, 1.99- and 90.9-fold higher than in the Ad5-Spp and SpFN groups, respectively (Fig. 6F). ELISpot confirmed that Ad5-SpFN also induced more spike-specific IgA-secreting cells (Fig. 6G). Phenotypic analysis further showed that Ad5-SpFN promoted higher frequencies of class-switched B cells (IgD⁻IgM⁻CD45R⁺) and memory B cells (IgG⁺CD38⁺) (Fig. 6H).

Adenoviral vectors, particularly Ad5-SpFN, also established stronger lung T_RM cells. Ad5-SpFN induced CD4⁺ T_RM at levels 2.16- and 2.42-fold higher than Ad5-SPP and SpFN, respectively (Fig. 6I). For CD8⁺ T_RM, both adenoviral vaccines induced frequencies 11–16-fold higher than SpFN, with Ad5-SpFN showing a moderate increase over Ad5-Spp (Fig. 6J). Spike-specific CD8⁺ T_RM staining with Spike_539−546 Tetramer⁺ confirmed that adenoviral immunization generated spike-specific CD8⁺ T_RM, which were undetectable in the SpFN group (Fig. 6K).

The cervical lymph nodes (cLNs) are the primary draining sites after intranasal immunization, where antigen and antigen-bearing DCs preferentially accumulate. Analysis of cervical lymph nodes revealed that Ad5-SpFN enhanced germinal center activity. It induced higher frequencies of total and spike-specific GC B cells and memory B cells, along with an expanded population of Tfh cells, compared to other groups (Fig. 6L-N).

Collectively, these data demonstrate that intranasal delivery of Ad5-SpFN elicits superior mucosal and systemic immunity, including robust antibody neutralization, resident lymphocyte populations, and sustained germinal center activity, supporting its potential as an effective mucosal vaccine candidate.

Ad5-SpFN confers robust protection against SARS-CoV-2 challenge

To further evaluate the protective efficacy of the Ad5-SpFN, K18-hACE2 mice were assigned to IM and IN immunization groups (Fig. 7A). In the IM group, mice received either adenoviral vector vaccines (Ad5-SpFN and Ad5-SPP, 1.5 × 10⁸ vp) or alum-adjuvanted protein vaccines (SpFN or Spp, 7.5 µg protein +50 µg Al(OH)₃). The IN was immunized with Ad5-SpFN, Ad5-Spp, SpFN, or sham control at equivalent doses. 14 days after immunization, all mice were challenged intranasally with 1.5 × 10³ PFU of authentic SARS-CoV-2. Either administered via IM injection or IN inoculation, Ad5-SpFN induced stronger NAb against D614G. Interestingly, SpFN only induced high pNAb titer following IM immunization, whereas its antibody titer was markedly lower after IN inoculation (Fig. 7B).

Fig. 7.

Ad5-nanoparticle vaccines protect against SARS-CoV-2 challenge following intramuscular or intranasal immunization. (A) Immunization and challenge schema. Naive 6–8-week-old female K18-hACE2 mice (n = 5 per group) were vaccinated with a single dose of 1.5 × 10⁸ vp Ad5-SpFN, Ad5-Spp or 7.5 µg SpFN, Spp formulated with 50 µg Al(OH)₃ via the intramuscular or intranasal route. Two weeks after the prime dose, mice were challenged with live SARS-CoV-2 (Wuhan-Hu-1) and euthanized 6 days post challenge for lung viral load quantification and histopathological analysis. (B) Serum D614G pseudovirus neutralizing antibody titers were measured two weeks post-immunization. Survival (C, D) and body weight change (E, F) were tracked for 6 days following SARS-CoV-2 challenge in IM- (C, E) and IN-immunized (D, F) groups. (G and H) SARS-CoV-2 nucleocapsid (N) gene copies in lung tissue were quantified by RT-qPCR and are expressed as log₁₀ copies per milliliter of tissue homogenate in IM- (G) and IN-immunized (H) groups. (I and J) Representative hematoxylin and eosin (H&E) (I) and immunohistochemistry (IHC) (J) stained lung sections from each group at day 6 post-challenge. Arrow heads indicate the typical pathological locations. Lung sections were stained with anti-SARS-CoV-2 N antibody. For H&E, scale bar = 200 μm; for IHC, scale bar = 50 μm. Data are presented as median ± IQR. Statistical significance was determined by two-sided Mann-Whitney U tests for comparisons between groups. Dashed lines indicate the higher or lower limit of quantification (HLOQ: 24,300 for NAb, LLOQ: 100 for NAb)

Survival outcomes differed between vaccine platforms and routes. In the IM groups, all mice in the Spp succumbed by day 4 after challenge while the Ad5-Spp group showed 80% survival at day 4 but only 20% by day 6. SpFN conferred 80% survival through day 6, whereas Ad5-SpFN provided 100% protection throughout the observation period (Fig. 7C). In the IN cohort, all sham-treated mice died by day 4, whereas 80% of SpFN-immunized mice were still alive on day 4 but none survived to day 6. In stark contrast, both Ad5-SpFN and Ad5-Spp achieved 100% survival at day 6 (Fig. 7D). Body weight loss closely mirrored survival outcomes. Following IM immunization, mice in the Ad5-Spp and Spp groups exhibited a rapid decline in body weight to ~80% of baseline by day 2, while Ad5-SpFN- and SpFN-immunized mice showed only mild, transient weight loss and recovered to baseline levels (Fig. 7E). In the IN cohort, sham and SpFN immunized mice displayed a rapid lost weight, dropping to 80% of their initial body weight, whereas mice receiving Ad5-SpFN and Ad5-Spp maintained or slightly gained weight, with Ad5-SpFN showing a modest advantage (Fig. 7F).

Viral loads in lung tissue measured on day 6 further highlighted differences in protection. In the IM group, Ad5-SpFN immunized mice had the lowest viral loads, with N gene copy numbers significantly lower than those in the Ad5-Spp (5.3 × 10⁴), SpFN (1 × 10⁴), and Spp (5.6 × 10⁵) groups (Fig. 7G). After IN immunization, lung viral loads in Ad5-SpFN and Ad5-Spp mice were at or near the limit of detection, whereas the SpFN and sham groups showed high viral replication (2.7 × 10⁶ copies) (Fig. 7H).

Histopathological analysis of lung tissue on day 6 confirmed severe alveolar hemorrhage, consolidation, cellular infiltration, and viral-infected cells, particularly around airways and vessels, in non-protected animals (Fig. 7I and J). In contrast, lungs from Ad5-SpFN-vaccinated mice retained normal alveolar architecture and cellular morphology, consistent with the observed virological control and survival. Whether administered via IM injection or IN instillation, the NAb titer right before challenge showed a negative correlation with both the lung viral loads (Fig. S6A) and body weight loss (Fig. S6B) after challenge. Together, these data demonstrate that Ad5-SpFN provides robust, route‑flexible protection against SARS-CoV-2 challenge. Notably, SpFN offers only partial protection when delivered IM and none by IN, while Ad5-Spp confers robust protection via the IN route but is less effective when given by IM.

Discussion

In this study, we successfully engineered an adenovirus-vectored nanoparticle vaccine, Ad5-SpFN, to address an important challenge in vaccinology: combining the potent cellular immunity of adenoviral vectors with the strong, broad humoral responses typical of protein nanoparticles. We found that Ad5-SpFN achieves this integration, inducing rapid, high-titer, and broad-spectrum neutralizing antibody along with Th1-biased T cell responses. Importantly, it also elicits protective mucosal immunity, particularly when administered intranasally, supporting that vectored delivery of nanoparticle immunogens represents a promising strategy for developing vaccines that confer comprehensive systemic and mucosal protection.

Our work aligns with and advances prior studies [27, 42], demonstrating the immunogenic superiority of nanoparticle displays, such as ferritin-based constructs, over monomeric antigens by enhancing lymph node trafficking and B cell activation. Furthermore, it confirms the established capacity of adenoviral vectors to drive potent T-cell responses [5]. Ad5-SpFN uniquely merges these two validated platforms, resulting in a vaccine whose performance exceeds that of conventional Ad5-vectored or protein-subunit vaccines alone, as evidenced by its accelerated antibody kinetics and enhanced neutralization breadth.

The immunological advantage of Ad5-SpFN stems from a powerful synergy between the adenoviral vector and the nanoparticle immunogen, likely creating a multi-stage activation cascade that optimally primes both cellular and humoral immunity. First, the Ad5 vector may act as a potent “danger signal” and efficient gene delivery vehicle, targeting DCs and other antigen-presenting cells to initiate a strong Type I interferon-driven [43, 44], Th1-biased environment. This innate immune activation is further amplified by the nanoparticle antigen itself. Particulate antigens like ferritin-based nanoparticles are inherently more efficient at being taken up by DCs compared to monomeric proteins [45–47]. Our data suggest that the SpFN nanoparticle, upon secretion from transduced cells, provides a multivalent, repetitive antigen array that enhances DC activation, as evidenced by increased CD86 and MHC-II expression. This promotes a mature, migratory DC phenotype (supported by CCR7 upregulation), facilitating the efficient trafficking of antigen to the draining lymph node, a critical step sometimes termed “antigen targeting” in adjuvant science [48, 49].

Within the lymph node, this coordinated delivery likely yields optimal conditions for adaptive immunity. The robust inflammatory milieu and antigen presentation driven by Ad5-primed DCs powerfully activate CD4⁺ T helper cells [50, 51], which are essential for supporting both arms of the adaptive response. Concurrently, the nanoparticle structure delivers the antigen in a form that is ideal for B cell recognition. The dense, repetitive display of spike protein mimics a viral surface, allowing for efficient cross-linking of B cell receptors (BCRs) [52–54]. The optimal size (~50 nm) of the self-assembled SpFN nanoparticles may also facilitates efficient drainage to lymph nodes and uptake by antigen-presenting cells. This high-avidity BCR engagement, combined with ample CD4⁺ T cell help, drives the rapid expansion of spike-specific B cell clones and fosters the formation of large, productive germinal centers [55, 56]. This explains the observed increase in germinal center B cells, T follicular helper cells, and the early, high-titer antibody response. Essentially, the Ad5 vector creates the ideal immunogenic “context” (strong DC and T cell activation), while the nanoparticle provides the perfect “antigen format” for B cell activation and antibody maturation, resulting in the balanced, potent, and broad immunity that defines the Ad5-SpFN vaccine.

In addition, the mucosal application of Ad vectors added another layer of advantage. sIgA, elicited by IN but rarely by IM immunization, is the predominant immunoglobulin at mucosal surfaces and mediates protection by neutralizing virions in the lumen, blocking epithelial attachment and transcytosis, and promoting immune exclusion, thereby reducing the establishment of infection and subsequent transmission.

Beyond the immunological benefits, the Ad5-vectored delivery of a nanoparticle immunogen offers distinct logistical and manufacturing advantages over traditional protein-based nanoparticle vaccines. The production of recombinant protein nanoparticles, while highly effective immunologically, involves complex and costly upstream and downstream processes. This includes large-scale mammalian cell culture for protein expression, followed by intricate and expensive purification steps, chromatography alone can constitute up to 50% of total manufacturing costs for protein biologics [57]. Furthermore, to achieve their full immunogenic potential, purified protein nanoparticles require precise formulation with potent adjuvants, adding another layer of complexity and variability to vaccine production [58, 59]. In contrast, the Ad5-SpFN platform simplifies this pipeline. The nanoparticle is encoded within the viral vector’s genome and produced in vivo by the vaccinee’s own cells, effectively outsourcing the complex steps of antigen expression, folding, and self-assembly to the host’s cellular machinery. This approach circumvents the need for large-scale antigen protein purification, rigorous nanoparticle assembly in vitro, and external adjuvant formulation. By leveraging the established, scalable, and cost-effective manufacturing processes already in place for adenoviral vectored vaccines, this synergized platform is anticipated to significantly lower the barrier for production and global distribution, particularly in resource-limited settings.

Despite these advantages, this study has several limitations. First, while the Ad5-SpFN construct successfully secretes assembled nanoparticles, the efficiency of this process in vivo is likely suboptimal. The rational design and consistent assembly of ferritin-immunogen chimeras remain a known challenge in the field, as the self-assembly process can be stochastic, potentially yielding nanoparticles with variable stoichiometries and affecting immunogen uniformity [60, 61]. Optimizing the antigen design, linkers, and expression cassette to enhance secretion efficiency represents a crucial opportunity for improving immunogenicity and achieving dose-sparing. In addition, although proof-of-concept is firmly established for the SARS-CoV-2 spike antigen, the generalizability of this “Ad5-vectored secretable nanoparticle” platform to antigens from other pathogens requires systematic validation. Different antigens may present unique folding, assembly, or immunodominance challenges when fused to ferritin and delivered via this system. Future work must explore its application to diverse viral families to assess its true versatility as a next-generation vaccine platform.

A significant challenge for the clinical application of Ad5-vectored vaccines is the high seroprevalence of neutralizing antibodies against Ad5 in human populations, which can range from 40% to 80% in some regions. This pre-existing immunity can potentially reduce vaccine efficacy by neutralizing the vector before it can transduce host cells and express the antigen. However, the impact of PEI is not absolute and can be mitigated through several strategies. Clinical experience suggests that extending the interval between natural Ad5 exposure and vaccination allows for the waning of neutralizing antibody titers, thereby reducing immunological interference. Furthermore, the route of administration plays a critical role. Our unpublished data demonstrate that intranasal immunization can effectively overcome the negative impact of pre-existing Ad5 immunity, likely by bypassing systemic neutralizing antibodies and directly engaging the mucosal immune compartment. This finding highlights a key advantage of our intranasally delivered Ad5-SpFN vaccine and supports its potential efficacy even in Ad5-seropositive individuals.

Conclusions

In conclusion, this study demonstrates that the Ad5-SpFN platform successfully combines the strengths of adenoviral vectors and self-assembling nanoparticles to elicit rapid, balanced, and protective systemic and mucosal immunity. The implications are substantial, providing a rational design strategy for next-generation vaccines that require concurrent induction of potent antibodies, robust T cells, and mucosal barriers. This work not only advances vaccine design against SARS-CoV-2 but also offers an adaptable framework for combating a wide range of pathogenic threats.

Abbreviations

NAbs

Neutralizing antibodies

Ad5

Adenovirus serotype 5

IM

Intramuscular

IN

Intranasal

ABSL-3

Animal biosafety level 3

IACUC

Institutional animal care and use committee

SPF

Specific pathogen free

SpFN

Spike-ferritin

Spp

Wild-type spike ectodomain with K986P and V987P mutations

PEI

Polyethylenimine

vp

Virus particle

CV

Column volumes

SEC

Size-exclusion chromatography

TEM

Transmission electron microscopy

DLS

Dynamic light scattering

MHC-I

Major histocompatibility complex class I

Tfh

T follicular helper

Tfr

T follicular regulatory

LN

Lymph node

dLNs

Draining lymph nodes

cLNs

Cervical lymph nodes

DC

Dendritic cell

GC

Germinal center

T_RM

Tissue-resident memory T cells

PVDF

Polyvinylidene difluoride

BALF

Bronchoalveolar lavage fluid

ELISA

Enzyme-linked immunosorbent assay

NT₅₀

Half-maximal neutralization titer

ICS

Intracellular Cytokine Staining

ELISpot

Enzyme-linked immunospot assay

CFSE

Carboxyfluorescein succinimidyl ester

PFU

Plaque-forming units

RT-qPCR

Quantitative reverse transcription polymerase chain reaction

IQR

Interquartile range

MOI

Multiplicities of infection

SHMW

Super high molecular weight

PBMCs

Peripheral blood mononuclear cells

SFU

Spot-forming units

AUC

Area under curve

T_EM

T effector memory

T_CM

T central memory

MBCs

Memory B cells

BCRs

B cell receptors

Author contributions

J.Y. and S.L. designed the study. J.Y. and Z.H. supervised the whole study. S.L. led the immunologic, virologic, and animal work together with Q.Z., J.F., G.J., Y.Z., D.F., Z.N., X.L., and T.C. The paper was written by J.Y., S.L., and X.M. with the involvement of all coauthors.

Funding

This study was funded by the National Key R&D Program of China (No. 2022YFC2304202 to Z.H.), the Major Project of Guangzhou National Laboratory (GZNL2023A01009 and GZNL2023A01005 to J.Y.), the Major Research Plan of the National Natural Science Foundation of China (NSFC) (92469104 to J.Y.) and the NSFC (82371831 to J.Y.), the Major Protect of Guangzhou National Laboratory (GZNL2024A01017 to X.M.), and the NSFC (82572540 to X.M.).

Data availability

All data needed to evaluate the conclusions in the paper are present in the manuscript and/or the Supplementary Materials. Any additional data is available on request from the corresponding author, Dr. Jingyou Yu at yu_jingyou@gzlab.ac.cn.

Declarations

Ethics approval and consent to participate

All of the animal experimental manipulations included in this study were approved by the policies and guidelines of the Animal Ethics Committee of (IACUC: GZLAB-AUCP-2023-06-A03) of Guangzhou National Laboratory. All experiments involving authentic SARS-CoV-2 strains were conducted in an animal biosafety level 3 (ABSL-3) facility of Guangzhou National Laboratory and approved by the biosafety ethics committee of Guangzhou National Laboratory (Approval number: GZLABSL3-A2025-011 A).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.