Programmable trivalent nanocage vaccine confers durable cross-species protection against Bordetella bronchiseptica infection

Abstract

Bordetella bronchiseptica (B. bronchiseptica), a respiratory pathogen endemic in companion animals, poses an escalating zoonotic threat to humans due to intensified human-pet contact, particularly among immunocompromised individuals. Here, we report a programmable trivalent nanocage vaccine, mi3-F1P2Fim2, featuring immunodominant regions of FHA, Prn, and Fim2 covalently fused into a single-chain construct and displayed on a self-assembling mi3 nanocage via the SpyTag003/SpyCatcher003 system, enabling high-density and multivalent antigen presentation. Compared with its soluble monomeric counterpart, mi3-F1P2Fim2 induced significantly enhanced humoral and cellular immune responses, characterized by a 1.98-fold elevation in antigen-specific IgG titers, improved IgG1/IgG2a subclass balance, and heightened dendritic cell activation. The vaccine facilitated efficient uptake by antigen-presenting cells and induced robust germinal center reactions. Notably, mi3-F1P2Fim2 conferred complete protection against B. bronchiseptica challenge in both murine and canine models, achieving > 1000-fold reduction in pulmonary bacterial burden with concomitant alleviation of lung pathology. These data demonstrate that mi3-F1P2Fim2 confers potent, cross-species protective immunity against B. bronchiseptica, establishing it as a promising vaccine candidate for preventing zoonotic respiratory infections. By targeting this clinically significant but understudied pathogen, our findings advance strategies to disrupt B. bronchiseptica transmission and enable targeted immunization aligned with One Health objectives.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04157-z.

Introduction

One Health has emerged as a globally recognized approach that emphasizes the intricate interdependence among human, animal, and environmental health [1–3]. As human lifestyles continue to evolve, companion animals that are increasingly integrated into households are involved in more frequent and intimate interactions with their owners. Consequently, the risk of zoonotic disease transmission has increased due to this close cohabitation [4–6]. Bordetella bronchiseptica (B. bronchiseptica), which possesses multiple virulence factors and immune evasion capabilities, is an emerging zoonotic pathogen that can persistently colonize host organisms. It is widely prevalent in companion animals such as dogs and cats, where it causes upper respiratory tract infections characterized by chronic coughing, dyspnea, and secondary pneumonia [7–9]. In recent years, infections caused by B. bronchiseptica have been increasingly reported in immunocompromised individuals, and this pathogen has been implicated in chronic cough and atypical pneumonia [10, 11]. During the COVID-19 pandemic, prolonged home quarantine and closer human-pet contact, combined with increased respiratory susceptibility to both viral and bacterial pathogens, have led to a rising trend of B. bronchiseptica infections in the population [12–14]. Despite the clinical availability of inactivated vaccines for B. bronchiseptica, these vaccines exhibit substantial limitations regarding safety, cross-protective efficacy and antigenic breadth. Therefore, developing a novel, safer and more controllable vaccine offering multi-antigen protection is crucial for controlling the spread of B. bronchiseptica within the One Health approach.

Conventional subunit vaccines typically rely on a single antigen to elicit immune responses, which often proves insufficient to address the antigenic diversity and complex pathogenic mechanisms of the pathogen [15, 16]. Increasing attention has been directed toward the development of multi-antigen vaccines that can target multiple immune epitopes to enhance protective efficacy and adaptability. For instance, Khan et al. reported that an adjuvanted tetravalent S1 subunit COVID-19 vaccine elicited robust humoral and cellular immune responses and conferred cross-protection against the Omicron variant [17]. In pneumococcal vaccine development, Nakahashi-Ouchida et al. demonstrated that a nasal vaccine containing three pneumococcal surface protein A variants conferred protection against intratracheal pneumococcal infection in nonhuman primates [18]. The diverse virulence factors and immune evasion strategies of B. bronchiseptica pose significant challenges to the efficacy of traditional vaccines [19]. Previous studies have identified filamentous hemagglutinin (FHA) [20], pertactin (Prn) [21], and fimbriae 2 (Fim2) [22] as key virulence components involved in bacterial adhesion, colonization, and the initiation of host immune responses, making them as important antigen candidates in B. bronchiseptica vaccine development. However, the design and development of multi-antigen vaccines remain hindered by numerous challenges, including imbalanced antigen expression, potential immunological interference, and limitations associated with delivery systems, all of which have restricted their widespread application [23–26]. Therefore, there is an urgent need for a technological platform capable of precisely displaying multiple antigens and enabling efficient delivery, in order to fully realize the potential of multi-antigen vaccine strategies.

Recently, nanoparticle-based platforms have emerged as a promising strategy for the development of multi-antigen vaccines, including virus-like particles (VLPs), bacteriophages, polysaccharides, and virosomes [27–33]. Among these platforms, self-assembling protein nanocages have attracted increasing attention for vaccine delivery. A representative example is the mi3 nanocage, one of the most widely utilized nanoplatforms for vaccine delivery, is composed of 60 subunits that spontaneously assemble into a regular icosahedral structure with an approximate diameter of 36 nm. Its highly ordered architecture and nanoscale dimensions enable it to mimic viral particles and elicit virus-like immune recognition. Moreover, the surface of mi3 can be precisely engineered to incorporate antigens at defined positions, enabling orderly and directional antigen presentation [34, 35]. With the growing prominence of structural vaccinology, mi3 has emerged as a structurally programmable protein nanocage and a promising platform for multi-antigen vaccine delivery, capable of enhancing immunogenicity and antibody titers [36–39]. As an example, Hills et al. constructed a quartet nanocage vaccine based on the mi3 nanocage displaying four distinct Sarbecovirus receptor-binding domains (RBDs), which elicited broadly neutralizing antibodies and enhanced cross-protection against emerging variants such as Omicron XBB.1.5 [40]. To enable efficient conjugation between antigens and the mi3 nanocage, the SpyTag003/SpyCatcher003 system was employed, utilizing a covalent protein-pairing strategy based on molecular recognition domains that facilitates the rapid and spontaneous formation of isopeptide bonds under mild conditions. This approach enables structurally controlled and stable antigen display, overcoming the spatial folding limitations commonly encountered in conventional fusion protein designs [41, 42]. The combination of multiple antigens with a nanocage platform, designed to target conserved regions or multiple immune escape sites, can mitigate immune evasion caused by single-antigen mutations, thus offering a novel approach to enhancing vaccine efficacy [43, 44].

In this study, the structurally optimized FHA-Prn-Fim fusion protein was covalently conjugated via SpyTag003 to the SpyCatcher003-modified mi3 nanocage, facilitating efficient antigen display on the nanoparticle surface. The antigen uptake and presentation of the nanoparticle vaccine by monocytes, macrophages and dendritic cells (DCs) were investigated, and its immunogenicity and safety were systematically evaluated in both mouse and dog models. Overall, the multi-antigen nanovaccine strategy not only focuses on mitigating the transmission risk of B. bronchiseptica but also aligns with the One Health approach, providing a broadly applicable platform for vaccine-based control of cross-species pathogens. This approach exemplifies a next-generation vaccine design that is safe, effective, and readily translatable.

Results and discussion

Preparation and characterization of the trivalent nanocage vaccine against B. bronchiseptica

Building upon previous research on the identification and functional domain analysis of protective antigens in Bordetella species, a precise antigen fusion strategy was designated in this study [45–47]. Specifically, the major immunodominant region I of FHA (F1, 1958–2112 aa), region II of Prn (P2, 514–609 aa), and the conserved domain of Fim2 (Fim2, 28–209 aa) were fused in tandem. Compared to traditional multi-antigen mixing or random coupling methods, the tandem design reduces batch-to-batch variability in antigen loading, enhances the consistency of antigen density control, and facilitates directional coupling, which is expected to further improve the immunogenicity [40, 48, 49]. To systematically evaluate the impact of antigen order on nanocage assprotein andmunogenicity, we constructed six fusion proteins with different tandem arrangements: F1-P2-Fim2, F1-Fim2-P2, P2-Fim2-F1, P2-F1-Fim2, Fim2-F1-P2, and Fim2-P2-F1. The SpyTag003 sequence was introduced at the N-terminus of the fusion protein and expressed in Escherichia coli (E. coli) BL21 via the pET-32a vector to form ST003-F1P2Fim2, ST003-F1Fim2P2, ST003-P2Fim2F1, ST003-P2F1Fim2, ST003-Fim2F1P2, and ST003-Fim2P2F1, respectively. The mi3 gene was synthesized according to the codon preference of E. coli, and SpyCatcher003 was introduced at the N-terminus of the mi3 nanocage through a flexible glycine-serine linker (GGS)₄. Expression was carried out in E. coli BL21 strain via the pETStrepII vector to form SC003-mi3 nanocage. ST003-F1P2Fim2, ST003-F1Fim2P2, ST003-P2Fim2F1, ST003-P2F1Fim2, ST003-Fim2F1P2, and ST003-Fim2P2F1 were displayed on the surface of mi3 nanocage through covalent linkage between SpyTag003 and SpyCatcher003. Detailed information on gene amplification is shown in Table S1. The chimeric strategies of ST003-F1P2Fim2, ST003-F1Fim2P2, ST003-P2Fim2F1, ST003-P2F1Fim2, ST003-Fim2F1P2, ST003-Fim2P2F1, and SC003-mi3 are shown in Fig. 1A, while the model for covalent binding of antigens to the mi3 nanocage through SpyTag003 and SpyCatcher003 is shown in Fig. 1B.

Fig. 1.

Design and construction strategy of the trivalent nanocage vaccine. (A) The structural models of B. bronchiseptica antigens, FHA, Prn and Fim2, and mi3 nanocage, with selected immunodominant regions (F1, P2, and Fim2) highlighted. The constructions of fusion proteins with F1, P2, and Fim2 in different sequences were designed. (B) A schematic diagram illustrating the covalent binding of ST003-fused antigens to SpyCatcher003-mi3 to form the trivalent nanoparticle vaccine

The recombinant plasmids encoding ST003-F1P2Fim2, ST003-F1Fim2P2, ST003-P2Fim2F1, ST003-P2F1Fim2, ST003-Fim2F1P2, and ST003-Fim2P2F1 were transformed into E. coli BL21 and purified using a His-tag Ni-NTA column. The recombinant plasmid of SC003-mi3 was transformed into E. coli BL21 strain and purified using StrepII-tag Streptactin-XT column. The purified recombinant proteins were then verified by Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). SDS-PAGE showed that the SC003-mi3 recombinant protein was expressed correctly, with an apparent molecular weight of approximately 47 kDa. The recombinant proteins of ST003-F1P2Fim2, ST003-Fim2F1P2, ST003-Fim2P2F1, ST003-P2Fim2F1, ST003-P2F1Fim2, and ST003-F1Fim2P2 were also expressed correctly, with an apparent molecular weight of approximately 70 kDa (Figure S1). For vaccine generation, SC003-mi3 was incubated with different ST003-antigens in PBS buffer and reacted overnight at 4 ℃. The results showed that SC003-mi3 was covalently linked to each ST003-antigen through SpyTag003 and SpyCatcher003, forming mi3-F1P2Fim2, mi3-F1Fim2P2, mi3-P2Fim2F1, mi3-P2F1Fim2, mi3-Fim2F1P2, and mi3-Fim2P2F1, respectively (Figure S2A, as exemplified by mi3-F1P2Fim2). To evaluate the conjugation efficiency of different ST003-antigens, an excess amount of SC003-mi3 was incubated with each antigen overnight at 4 ℃ in PBS, and the conjugation efficiency was assessed by SDS-PAGE. The results showed that all ST003-antigens were capable of forming covalent conjugates with SC003-mi3. Among them, ST003-F1P2Fim2 exhibited the highest conjugation efficiency, reaching 47.7% (Figure S2B), suggesting that this configuration is more spatially compatible with SC003-mi3. Subsequent mouse immunization experiments revealed no significant differences in total IgG levels induced by the different antigen orders (Figure S2C). However, mi3-F1P2Fim2 demonstrated the most favorable expression yield and coupling efficiency, and it was therefore selected for subsequent vaccine construction and immunoprotective efficacy studies.

To produce mi3-F1P2Fim2, we initially evaluated whether the conjugation efficiency between SC003-mi3 and ST003-F1P2Fim2 is in a ratio-dependent manner. The results revealed that increasing amounts of ST003-F1P2Fim2 led to a progressive decrease in residual SC003-mi3. Notably, when the molar ratio of SC003-mi3 to ST003-F1P2Fim2 reached 2:4, the conjugation efficiency reached 73.6%, no further obvious reduction in SC003-mi3 was observed, indicating saturation of the scaffold and excess unbound ST003-F1P2Fim2 (Fig. 2A). The ST003/SC003 conjugate proteins were subjected to size exclusion chromatography (SEC) to remove unbound antigen and free mi3. SEC showed that the elution peak of SC003-mi3 appeared at 9.5 mL, while that of mi3-F1P2Fim2 appeared at 8.7 mL, suggesting that mi3-F1P2Fim2 was eluted and purified earlier than SC003-mi3 during the SEC process (Fig. 2B). The hydrodynamic diameters of SC003-mi3 and mi3-F1P2Fim2 were measured by using dynamic light scattering (DLS). DLS indicated that the hydrodynamic diameter of SC003-mi3 was 37.84 ± 0.15 nm, while that of mi3-F1P2Fim2 was 50.75 ± 0.54 nm (Fig. 2C). Moreover, negative-staining transmission electron microscopy (TEM) revealed that mi3 nanoparticles exhibited a well-defined spherical morphology, whereas mi3-F1P2Fim2 nanoparticles displayed a more diffuse surface with distinct protruding protein structures representing the F1P2Fim2 antigens, in stark contrast to the mi3 group (Fig. 2D).

Fig. 2.

Preparation and structural characterization of mi3-F1P2Fim2. (A) SDS-PAGE analysis of the assembly products from mixtures of ST003-F1P2Fim2 and SC003-mi3 at different molar ratios. (B) SEC of mi3 and mi3-F1P2Fim2 was performed using a Superose 6 Increase 10/300 GL column. (C) DLS analysis was used to characterize the size distribution of mi3 and mi3-F1P2Fim2. (D) Representative TEM images of the buffer control, mi3, and mi3-F1P2Fim2. Red arrows indicated the protruding surface structures corresponding to the F1P2Fim2 antigens. Scale bar: 200 nm

To further evaluate the physicochemical stability and safety of the mi3-F1P2Fim2 vaccine, we conducted a systematic analysis of its thermal stability, storage stability and in vitro cytotoxicity. Thermal stability experiments showed that mi3-F1P2Fim2 maintained good solubility after incubation at 55 ℃ for 1 h, with no visible precipitation or protein degradation, indicating that the platform had strong thermal resilience and could withstand transportation and environmental fluctuations within a certain range (Figure S3A). In the storage stability test, mi3-F1P2Fim2 was stored at 4 ℃ and 25 ℃ for 30 days, and changes in protein solubility were monitored. The results showed that mi3-F1P2Fim2 underwent significant degradation after 30 days at 25 ℃, with almost complete loss of activity, while it maintained good solubility at 4 ℃, indicating that the vaccine was highly stable under refrigerated conditions (Figure S3B). Additionally, DLS measurements conducted over a 14-day period at 4 °C demonstrated that the hydrodynamic diameter of mi3-F1P2Fim2 remained stable, with no significant changes in particle size or size distribution (Figure S3C). These characteristics were particularly important for ensuring vaccine accessibility and transportability, especially in regions with limited cold chain infrastructure under practical deployment conditions [50, 51]. Cytotoxicity assays demonstrated that mi3-F1P2Fim2 was non-toxic to bone marrow-derived dendritic cells (BMDCs), DC2.4 cells or RAW264.7 macrophages, indicating good biocompatibility of mi3-F1P2Fim2 (Figure S4).

In our study, a SpyTag003/SpyCatcher003-mediated covalent conjugation strategy was employed to construct and characterize a self-assembled trivalent nanocage vaccine, mi3-F1P2Fim2, enabling the presentation of three key immunogens on a single nanocage. The mi3-F1P2Fim2 simplified vaccine composition by reducing the number of components while enabling the display of multiple antigens on each nanocage with favorable physicochemical properties, thermal stability and biocompatibility. Nevertheless, the present study is limited by the absence of direct quantitative measurements of antigen occupancy on individual mi3 nanocages, which is warranted to be further addressed in our future study.

mi3-F1P2Fim2 elicits strong immune response in mice

Humoral immune response is an important indicator for evaluating vaccine immunogenicity [52]. The antigen-specific IgG level is commonly used to reflect the overall antibody response induced by the vaccine, while the changes in the proportion of different IgG subclasses not only reveal the polarization of the immune response but also indirectly reflect the type of helper T cell (Th) response (such as Th1 or Th2) [53–55]. In this study, the humoral immune activation capacity of mi3-F1P2Fim2 vaccine was systematically evaluated by measuring the total IgG and its subclasses in the serum of immunized mice. To determine the optimal immune dose, mice received a prime-boost vaccination with 5 µg, 10 µg, or 20 µg, and sera were collected 14 days after the booster to measure total IgG levels. The results indicated that both 10 µg and 20 µg doses induced significantly higher antibody production, while the 20 µg dose resulted in only a marginal improvement on antibody responses but requiring increased vaccine usage (Figure S5). Therefore, considering both immunization efficacy and application cost, 10 µg was selected as the optimal immunization dose for subsequent animal experiments.

To systematically evaluate the immunogenicity induced by mi3-F1P2Fim2 in the mouse model, mice were immunized with either the F1P2Fim2 recombinant protein vaccine or the mi3-F1P2Fim2 nanoparticle vaccine, with non-immunized mice serving as the control group. Humoral immune activation was analyzed based on blood samples collected at different time points (Fig. 3A). 14 days after the first immunization, a significantly higher level of total IgG was detected in the serum of mi3-F1P2Fim2 immunized mice compared to that in the F1P2Fim2 group. After the booster immunization, total IgG in the sera of mi3-F1P2Fim2-immunized mice increased significantly and was markedly higher than that in the F1P2Fim2 group, indicating that this vaccine platform has strong IgG-inducing capability (Fig. 3B). Additionally, antibody subclass analysis revealed that mi3-F1P2Fim2 elicited IgG2a and IgG1 levels 2.55-fold and 1.2-fold higher, respectively, than those induced by F1P2Fim2 (Fig. 3C). Given that B. bronchiseptica is known to persist within host cells, the induction of a Th1-skewed immune response is considered critical for effective bacterial clearance. In this study, the mi3-F1P2Fim2 nanoparticle vaccine elicited a higher IgG2a/IgG1 ratio compared to the soluble antigen, indicating a Th1-biased humoral response while retaining Th2-associated IgG1 levels. This balanced immune polarization suggests that the dense, ordered antigen display on the mi3 nanocage facilitates more efficient immunological priming, aligning with the immune profile required to control intracellular bacterial or viral infections [56–58].

Fig. 3.

Antibody responses in mice immunized with mi3-F1P2Fim2. (A) A schematic illustration of the prime-boost immunization strategy in a mouse model. Groups of ICR mice (n = 10) were intramuscularly immunized with PBS (control), 10 µg of F1P2Fim2, or 10 µg of mi3-F1P2Fim2, followed by a booster immunization on day 14. Serum was collected on day 0 and day 14 after the primary immunization, and on day 14 after the booster dose to determine antibody levels. (B) IgG antibody responses against B. bronchiseptica in sera of immunized mice (n = 10) were measured by ELISA after primary and booster immunizations with mi3-F1P2Fim2. (C) Levels of IgG1 and IgG2a antibody subclass against B. bronchiseptica in sera of immunized mice (n = 10) were measured by ELISA after booster immunization. Data are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant)

mi3-F1P2Fim2 elicits potent cellular immune responses in mice

Cellular immunity plays a crucial role in protecting against infection by intracellular pathogens through vaccination, and T cells are essential for pathogen clearance and the establishment of immune memory [59, 60]. A single indicator is often insufficient to comprehensively reflect the complex process of T cell activation. Therefore, it should be evaluated from multiple dimensions, including cell function, cytokine secretion, and systemic immune status, to systematically reflect the vaccine’s impact on Th1/Th2 polarization, T cell functional state and overall immune activity [61, 62].

We measured the secretion levels of IFN-γ and IL-4 in splenocytes from each immunization group after antigen stimulation to assess the antigen-specific cellular immune response. Enzyme-Linked Immunospot (ELISPOT) results showed that the number of IFN-γ and IL-4 secreting cells in the spleen of mice immunized with mi3-F1P2Fim2 was significantly higher than that of the F1P2Fim2 group (Fig. 4A). Lymphocyte proliferation, commonly used to reflect the status of immune system, was assessed upon antigen re-stimulation. Results of the stimulation index (SI) indicated that mi3-F1P2Fim2 significantly enhanced splenic lymphocyte proliferation compared to F1P2Fim2 (Fig. 4B). Additionally, cytokine secretion in the splenocyte culture supernatant upon antigen re-stimulation was measured by enzyme-linked immunosorbent assay (ELISA). The mi3-F1P2Fim2 immunization elicited IL-2, IL-4, and IFN-γ levels up to 2.73-fold, 2.08-fold, and 1.68-fold, respectively, of those induced by F1P2Fim2 in splenocyte culture supernatants (Fig. 4C). This data indicates that mi3-F1P2Fim2 promoted both Th1 and Th2 T cell responses. We further measured serum cytokine levels in each group after booster immunization. The levels of IL-2, IL-4, and IFN-γ elicited by mi3-F1P2Fim2 were 1.73-fold, 2.44-fold, and 1.33-fold, respectively, of those induced by F1P2Fim2 (Fig. 4D). Together, these results indicate that the platform not only induces strong local T cell responses but also generates systemic immune activation.

Fig. 4.

Cellular immune responses in C57BL/6 mice induced by mi3-F1P2Fim2 immunization. Two weeks after booster immunization, spleens and serum were collected from immunized mice (n = 5) for assessment of cellular immune responses. (A) The number of IFN-γ or IL-4 secreting cells in splenic lymphocytes was measured by ELISpot assay (n = 5). Representative images from each group are shown. (B) Splenic lymphocyte proliferation in response to antigen stimulation was assessed (n = 5). (C) Levels of IL-2, IL-4, and IFN-γ in the culture supernatant of splenic lymphocytes were measured (n = 5). (D) Serum levels of IL-2, IL-4, and IFN-γ were measured two weeks after booster immunization (n = 5). (E) Percentages of CD4⁺ and CD8⁺ T cells among splenic lymphocytes were determined by flow cytometry (n = 5). Representative images from each group are shown. Data are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant)

To independently assess the differentiation of CD4⁺ and CD8⁺ T cells, splenocytes were collected after booster immunization, and the percentages of CD4⁺ and CD8⁺ T cells were analyzed by flow cytometry. mi3-F1P2Fim2 immunization significantly enhanced the differentiation of CD4⁺ and CD8⁺ T lymphocytes, with levels that were 1.34-fold and 1.89-fold of those induced by F1P2Fim2, respectively (Fig. 4E). These results indicate that mi3-F1P2Fim2, as a multi-antigen nanocage vaccine, can induce robust cellular and humoral immune responses.

mi3-F1P2Fim2 causes efficient antigen uptake and sustained antigen release and promotes the activation and maturation of BMDCs

To elucidate the potential mechanism for the enhanced immune responses induced by mi3-F1P2Fim2, we investigated the main processes of adaptive immunity after mi3-F1P2Fim2 vaccination, including antigen uptake and presentation, and germinal center responses. It is well established that antigen uptake and presentation represent the initial and critical steps in vaccine-induced adaptive immune responses, playing a decisive role in the initiation of T cell responses [63, 64]. Due to their fragile structure and susceptibility to degradation, subunit vaccines often exhibit limited uptake efficiency in antigen-presenting cells (APCs), whereas nanoparticle vaccines, with their optimal size (20–100 nm), high surface antigen density and good stability, significantly enhance antigen uptake efficiency in DCs and macrophages [65–67]. Utilizing stable nanocage platforms, such as mi3, to present antigens can enhance the efficiency of antigen recognition, uptake and sustained release by APCs, thereby boosting subsequent humoral and cellular immune responses [68, 69]. To elucidate whether the efficient antibody response is due to effective antigen uptake and processing, we prepared Cy7-tagged and iFluor™ 594-tagged F1P2Fim2 for antigen tracking. Specifically, iFluor™ 594-tagged mi3-F1P2Fim2 or an equivalent amount of iFluor™ 594-tagged F1P2Fim2 was incubated with BMDCs, DC2.4 cells, or RAW264.7 macrophages, and antigen uptake was observed by fluorescence microscopy. mi3-F1P2Fim2 was efficiently taken up by BMDCs, DC2.4 cells, and RAW264.7 macrophages, exhibiting stronger fluorescence signal intensities than those of F1P2Fim2 (Fig. 5A). Consistent with these observations, confocal microscopy images acquired at 40 × and 60 × magnification further confirmed enhanced cellular uptake of mi3-F1P2Fim2 across a larger field of view (Figure S6). Quantitative analysis revealed that the antigen uptake ratios of mi3-F1P2Fim2 to F1P2Fim2 in BMDCs, DC2.4 cells and RAW264.7 macrophages were 2.01:1, 2.12:1, and 1.79:1, respectively, indicating that the nanoparticle vaccine mi3-F1P2Fim2 exhibited significantly higher uptake efficiency in APCs compared to the sole recombinant protein F1P2Fim2. This may be attributed to the cell’s greater sensitivity to particle size and surface structure, which influences its endocytic properties. The higher uptake levels lay a foundation for subsequent major histocompatibility complex (MHC)-II antigen presentation and CD4⁺ T cell activation, while also contributing to enhanced cross-presentation via the MHC-I pathway, thereby activating CD8⁺ T cell immunity [70]. The multivalent arrangement of mi3-F1P2Fim2 nanoparticle vaccine may also promote multiple receptor crosslinking, enhancing DCs recognition and phagocytosis of the antigen. Therefore, mi3-F1P2Fim2 not only possesses the characteristics of structural stability and high-density antigen presentation but also demonstrates a significant advantage during the antigen uptake phase, thereby significantly enhancing the adaptive immune response.

Fig. 5.

mi3-F1P2Fim2 facilitates antigen uptake, delays antigen release, and promotes activation of BMDCs. (A) F1P2Fim2 (red) distribution in BMDCs, DC2.4 cells, and RAW264.7 macrophages was observed after incubation with equivalent amounts of iFlour™ 594-tagged F1P2Fim2 or mi3-F1P2Fim2 for 6 h. Representative images from each group are shown. Scale bar: 10 μm. Quantification of mean fluorescence intensity (MFI) was performed using ImageJ software (n = 3). (B) Activation markers of BMDCs, including MHC-II and co-stimulatory molecules CD80 and CD86, were analyzed by FCM (n = 5). Normalized MFI values of MHC-II, CD80, and CD86 were quantified (n = 5). Cytokine secretion (TNF-α, IL-1β, and IL-6) in BMDC culture supernatants was measured by ELISA (n = 5). (C) In vivo antigen retention at the injection site visualized using Cy7-tagged F1P2Fim2 or mi3-F1P2Fim2 and monitored with the IVIS Spectrum imaging system. Average radiant efficiency was quantified using Living Image Vision 4.4. p, photons (n = 5). Data are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant)

Furthermore, another key factor influencing vaccine immunogenicity is whether the antigen, once internalized, can effectively induce DC maturation [71]. Mature DCs enhance interactions with T cells by upregulating co-stimulatory molecules, such as CD80, CD86, and MHC-II, thereby facilitating the activation and effector differentiation of naïve T cells [72–74]. Traditional subunit vaccines often struggle to fully induce DC maturation due to low antigen uptake efficiency and transient presentation processes. In contrast, nanoparticle vaccines, through advantages in size, surface multivalent antigen arrangement and enhanced delivery dynamics, can improve the activation of antigen presentation pathways and promote immune cell maturation [75]. Therefore, we further evaluated the effect of mi3-F1P2Fim2 on the expression of MHC-II and co-stimulatory molecules on BMDC surfaces to verify its capability in activating adaptive immune responses. Briefly, BMDCs were incubated with mi3-F1P2Fim2 or an equivalent amount of F1P2Fim2, with LPS used as a positive control. The expression levels of MHC-II, CD80, and CD86 on BMDC surfaces were analyzed by flow cytometry. The results showed that the surface expression levels of co-stimulatory molecules MHC-II, CD80 and CD86 on BMDCs in the mi3-F1P2Fim2 group were 1.40-fold, 1.92-fold, and 2.02-fold of those in the F1P2Fim2 group, respectively (Fig. 5B). Mature DCs secrete various cytokines that not only promote T cell activation and differentiation but also recruit and regulate other immune cells, playing a key role in driving adaptive immune responses [76, 77]. As observed in our study, mi3-F1P2Fim2 induced significantly higher cytokine secretion, with TNF-α, IL-1β, and IL-6 levels increased by 3.92-fold, 1.29-fold, and 0.33-fold, respectively, compared to the F1P2Fim2 group (Fig. 5B). These results indicate that mi3-F1P2Fim2 effectively activates DC maturation and enhances antigen presentation capabilities. The high-density, multi-antigen co-presentation structure of mi3-F1P2Fim2 could greatly improve the probability of antigen recognition and endocytosis by DCs. Additionally, the mi3 platform can release antigens in a sustained manner due to its stable structure, thereby prolonging the antigen exposure time of DCs and enabling more persistent and effective T cell activation. This characteristic is consistent with our observation in the antigen uptake experiment, where mi3-F1P2Fim2 exhibited a significantly higher endocytosis rate in BMDCs than that of F1P2Fim2, further validating the structural advantages of mi3-F1P2Fim2 in antigen processing and presentation.

To validate whether this effect occurs similarly in vivo, the expression of surface molecules on DCs in the inguinal lymph nodes (iLNs) of immunized mice was examined at 3 days post-booster immunization. The expression levels of MHC-II and CD86 molecules on DCs in the lymph nodes of mi3-F1P2Fim2 immunized mice were significantly higher than those in the F1P2Fim2 group, indicating that the nanoparticle enhanced the antigen presentation capability of DCs in vivo (Figure S7). Although CD80 expression was also elevated in the mi3-F1P2Fim2 group, the difference was not statistically significant, which may reflect the more complex regulation of individual co-stimulatory signals in vivo (Figure S7). These results are consistent with the in vivo immune response profile observed in the mouse model: mi3-F1P2Fim2 not only elicited significantly higher titers of total IgG, IgG1, and IgG2a compared to F1P2Fim2, but also potentiated the production of both Th1 (IFN-γ) and Th2 (IL-4) cytokines. Based on the activation of CD4⁺ and CD8⁺ T cells, it can be inferred that the nanoparticle vaccine helps enhance the uptake and presentation of antigens by DCs, thereby more effectively inducing T cell response polarization and expansion. Particularly, under the synergistic action of multiple antigens, the nanoparticle vaccine promoted broader and more durable cellular immune protection.

Moreover, to observe the distribution and retention ability of mi3-F1P2Fim2 in vivo, Cy7-tagged mi3-F1P2Fim2 or an equivalent amount of Cy7-tagged F1P2Fim2 was used to immunize ICR mice. Three mice from each group were randomly selected at 4 h after immunization to harvest inguinal lymph nodes for in vivo imaging using the IVIS^® Spectrum System (PerkinElmer). The remaining mice were monitored for antigen retention at the injection site for 12 days using the imaging system. Compared to F1P2Fim2, mi3-F1P2Fim2 accumulated a strong fluorescent signal in iLNs within a short period after immunization (Figure S8). Additionally, the fluorescence signal of F1P2Fim2 became nearly undetectable by day 9 post immunization. In contrast, mi3-F1P2Fim2 maintained a strong signal at the injection site, with the average radiance in the region of interest (ROI) still exceeding 1.0 × 10⁷ photons/s/cm²/sr at day 12 post immunization, indicating superior antigen retention and lymphatic enrichment properties of mi3-F1P2Fim2 (Fig. 5C). These data are consistent with the in vitro DC activation results, further confirming that the trivalent nanocage vaccine facilitates not only efficient antigen uptake, but also delays antigen degradation and clearance, enhancing the sustained activation ability of APCs in the lymph nodes, thereby providing conditions for more durable T cell memory formation.

mi3-F1P2Fim2 induces germinal center responses

After the presentation of antigens by APCs, the germinal center (GC) forms, which is a critical site for the generation of high-affinity antibodies and the differentiation of memory B cells during the adaptive immune response, playing a decisive role in vaccine-induced humoral immunity [78, 79]. Nanoparticle vaccines, due to their unique antigen arrangement and lymphatic system homing ability, are believed to more effectively activate follicular helper T cells (Tfh) and promote GC formation, thereby enhancing the quality and durability of antibodies. Therefore, we systematically evaluated the ability of mi3-F1P2Fim2 to induce GC reactions in mice. Flow cytometry results showed that the proportion of GC B cells (GL7⁺ CD95⁺) in both lymph nodes and spleen of mi3-F1P2Fim2 immunized mice was 5.19-fold and 2.77-fold higher than those in F1P2Fim2 immunized mice, respectively (Fig. 6A). These results suggest that the nanoparticle platform enhances the antigen presentation efficiency in lymphoid organs, significantly promoting the formation and maintenance of GC structures.

Fig. 6.

GC responses in C57BL/6 mice induced by mi3-F1P2Fim2 immunization. Two weeks after booster immunization, iLNs, spleens, and bone marrow were collected from immunized mice (n = 5) for assessment of GC responses. (A) Representative flow cytometric plots (left) and frequencies (right) of GC B cells (CD95⁺ GL7⁺) (n = 5). (B) Representative flow cytometric plots (left) and frequencies (right) of GC Tfh cells (CXCR5⁺ PD1⁺) (n = 5). (C) Representative flow cytometric plots (left) and frequencies (right) of LLPCs (B220^low CD138⁺) (n = 5). (D) Immunofluorescence staining (left) and quantification (right) of GCs (GL7⁺) in inguinal lymph nodes of immunized mice (n = 3). White arrows indicate GC regions. Representative images from each group are shown. Scale bar: 1000 μm. Scale bar: 1000 μm. Data are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA (*P < 0.05; **P < 0.01; **P < 0.001; ****P < 0.0001; ns, not significant)

In GC B cell response, Tfh cells are the key cell type that regulates B cell affinity maturation, class switching, and long-term antibody production through cytokine secretion and co-stimulatory signals [80]. The proportion of Tfh cells (CXCR5⁺ PD-1⁺) in both lymph nodes and spleens of mi3-F1P2Fim2 immunized mice was 6.65-fold and 2.41-fold of those in F1P2Fim2 immunized mice, respectively (Fig. 6B). This trend is consistent with the upregulation of GC B cells, indicating that mi3-F1P2Fim2, by activating the Tfh response, effectively supports the formation of GC and antibody production. After B cells have completed affinity maturation and class switching, some can differentiate into memory B cells or migrate to the bone marrow, where they further differentiate into long-lived plasma cells (LLPCs). The latter can continuously secrete high-affinity antibodies in the absence of antigen stimulation and represent a key cell population responsible for maintaining long-term humoral immune memory [81–83]. To systematically evaluate the durability of the antibody response induced by the vaccine, we measured the proportion of LLPCs (B220^low CD138⁺) in the bone marrow at day 14 post-booster immunization. The result showed that the proportion of LLPCs in the mi3-F1P2Fim2 group was 1.44-fold higher than that in the F1P2Fim2 group (Fig. 6C). This result indicates that mi3-F1P2Fim2 not only induces a strong GC response early in the immune process but also promotes the differentiation of antibody-secreting cells, thereby establishing long-lasting humoral immunity.

To evaluate the manifestation of GC response at the tissue level perspective, we performed histological analysis of the iLNs at the same time point. Immunofluorescence staining results showed that GC formation was significantly enhanced in the iLNs of mi3-F1P2Fim2 immunized mice, with the number of GCs being 4–8 times higher than that in F1P2Fim2 immunized mice (Fig. 6D). In addition, the expression of the cell proliferation marker Ki-67, as determined by immunohistochemical (IHC) staining, was also significantly higher in mi3-F1P2Fim2 immunized mice than that in F1P2Fim2 immunized mice, suggesting active cell proliferation within the GC regions and stronger germinal center activity following immunization(Figure S9).

Together, the above results indicate that mi3-F1P2Fim2 significantly amplifies the scale and quality of the GC response by enhancing the formation of Tfh cells, GC B cells, and LLPCs, thereby driving high-affinity antibody production and promoting long-lasting humoral immunity. These findings further demonstrate that this multi-antigen nanoparticle vaccine platform has systemic advantages in antigen delivery, presentation, and robust immune response maintenance, making it a powerful tool for developing efficient vaccines.

mi3-F1P2Fim2 confers robust protection against B. bronchiseptica in mice

To assess the efficacy of mi3-F1P2Fim2 in vivo, mice were challenged with 2.6 × LD₅₀ of B. bronchiseptica strain WH1218 four weeks after the booster immunization (Fig. 7A). Briefly, mice were randomly assigned to four groups: PBS (negative control), F1P2Fim2, mi3-F1P2Fim2, and an inactivated vaccine derived from the B. bronchiseptica WH1218 strain (as a positive control of traditional B. bronchiseptica vaccine). The results showed that all mice in the PBS group and F1P2Fim2 group died from 4 to 8 days post challenge, while all mice in the mi3-F1P2Fim2 and inactivated vaccine groups survived (Fig. 7B). Regarding body weight changes, mice immunized with mi3-F1P2Fim2 or inactivated vaccine experienced slight weight loss during the first 4 days, followed by gradual recovery to normal levels, while mice in the PBS or F1P2Fim2 groups exhibited a sharp decline, indicating severe B. bronchiseptica infection (Fig. 7C).

Fig. 7.

The efficacy of mi3-F1P2Fim2 in protecting mice against B. bronchiseptica challenge. (A) Schematic illustration of the prime-boost immunization strategy of mi3-F1P2Fim2 and challenge with B. bronchiseptica in a mouse model. Groups of C57BL/6 mice (n = 15) were i.m. immunized with PBS, 10 µg of F1P2Fim2, 10 µg of mi3-F1P2Fim2, or 10⁸ CFU of inactivated B. bronchiseptica vaccine, followed by a booster immunization on day 14. Mice were challenged with B. bronchiseptica four weeks after the booster immunization. (B) Survival rates were monitored daily for 14 days after challenge (n = 10). (C) Body weight changes post-challenge was monitored and analyzed (n = 10). Mice that lost ≥ 25% of their pre-challenge body weight were euthanized. (D) Bacterial loads in the lungs were quantified at 3 days post-challenge (n = 5). (E) Lung histopathological analysis was performed by H&E staining (n = 3). Representative images from each group are shown, with the lower panel representing a magnified view of the upper image. Scale bars: 100 μm (upper), 50 μm (lower). Data in (B), (C), and (D) are presented as mean ± SEM. Statistical significance in (B) was determined by the log-rank (Mantel-Cox) test; significance in (D) was determined by one-way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant)

Importantly, the lung bacterial load results showed that the average colony-forming units (CFU) in lung tissues of mice in the PBS group was 10^7.5 CFU/g, indicating that in the absence of immunization, B. bronchiseptica can efficiently colonize lung tissues and cause a severe infection. In comparison, the bacterial load in lung tissues of F1P2Fim2 immunized mice showed only a slight decrease, with an average of 10^7.4 CFU/g. By contrast, a dramatic reduction in lung bacterial load (10^4.4 CFU/g) was observed in mice immunized with mi3-F1P2Fim2, which is comparable with that in mice immunized with inactivated vaccine, indicating that the mi3-F1P2Fim2 vaccine can provide protection comparable to traditional vaccines without relying on complete bacterial components (Fig. 7D). Moreover, histopathological analysis of the lungs showed severe pathological changes in mice of PBS group and F1P2Fim2 group, including alveolar septal thickening, bronchiolar necrosis, pulmonary edema and inflammatory cell infiltration. As expected, no significant pathological changes were observed in the lungs of mice immunized with mi3-F1P2Fim2 or inactivated vaccine (Fig. 7E). These results suggest that the mi3-F1P2Fim2 vaccine primarily protects mice from lethal B. bronchiseptica challenge through effective bacterial clearance and reduced lung inflammation. While for broader protection of mi3-F1P2Fim2, challenge with different strains of B. bronchiseptica should be carried out, which is warranted in our further study.

In addition, to assess whether mi3-F1P2Fim2 induces systemic or local toxicity, we performed H&E staining on the heart, liver, spleen, kidneys, lymph nodes, brain and small intestine of immunized mice at 7 days after booster immunization. No obvious pathological signs were observed in the detected tissues, suggesting that the administration of mi3-F1P2Fim2 is potentially safe in vivo (Figure S10).

Immunogenicity and protective efficacy of mi3-F1P2Fim2 in dogs against B. bronchiseptica infection

Dogs are one of the primary natural hosts of B. bronchiseptica, with a high susceptibility to infection and clinical relevance [10, 84, 85]. As companion animals in close contact with humans, dogs provide valuable vaccine evaluation data that offer important insights into strategies for preventing the transmission of infectious diseases between humans and pets. Compared to rodent models, dogs more closely resemble the natural conditions in terms of infection pathways, respiratory clinical manifestations and disease progression, making it an ideal model for evaluating vaccine immunogenicity, safety and efficacy [86, 87]. Therefore, dogs were then employed as a companion animal model to systematically evaluate the immunogenicity and efficacy of mi3-F1P2Fim2. Briefly, dogs were randomly assigned to three groups: PBS (negative control), mi3-F1P2Fim2 and an inactivated vaccine derived from the B. bronchiseptica WH1218 strain (as a positive control of traditional B. bronchiseptica vaccine) (Fig. 8A). The total IgG level in the mi3-F1P2Fim2 immunized dogs was 72.1% of that in the inactivated vaccine immunized dogs at day 14 post-primary immunization, and increased after booster immunization to 91.2% of the level induced by the inactivated vaccine, indicating that mi3-F1P2Fim2 has the ability to activate a strong humoral immune response in dogs (Fig. 8B). Additionally, the levels of IgG2 and IgG1 induced by the mi3-F1P2Fim2 immunized dogs were 1.42-fold and 1.39-fold of those induced by the inactivated vaccine immunized dogs after booster immunization, indicating that it can induce a robust and balanced antibody response in the primary-boost immunization strategy, which helps enhance bacterial clearance efficiency and immune persistence (Fig. 8C).

Fig. 8.

The efficacy of mi3-F1P2Fim2 in protecting dogs against B. bronchiseptica challenge. (A) Schematic illustration of the prime-boost immunization strategy and challenge protocol in a canine model. Groups of dogs (n = 10) were i.m. immunized with PBS, 100 µg of mi3-F1P2Fim2, or 10⁹ CFU of inactivated B. bronchiseptica vaccine, followed by a booster immunization on day 14. Dogs were challenged with B. bronchiseptica four weeks after the booster immunization. (B) IgG antibody responses against B. bronchiseptica in dog serum were measured by ELISA after primary and booster immunizations (n = 5). (C) IgG1 and IgG2 antibody responses against B. bronchiseptica in dog serum were measured by ELISA after primary and booster immunizations (n = 5). (D) Protection rate was monitored for 14 days after challenge (n = 5). (E) Body weight changes post-challenge was monitored and analyzed (n = 5). (F) Clinical symptom scores were recorded after bacterial challenge (n = 5). (G) Bacterial loads in the lungs were quantified at 3 days post challenge (n = 5). (H) Lung histopathological analysis was performed by H&E staining (n = 3). Representative images from each group are shown. Scale bars: 200 μm. Data are presented as mean ± SEM. Statistical significance was determined by one-way ANOVA, statistical significance in (B) was determined by the log-rank (Mantel-Cox) test (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant)

To evaluate the efficacy of mi3-F1P2Fim2 in a dog model, the immunized dogs were challenged with B. bronchiseptica WH1218 strain at 4 weeks after the booster immunization (Fig. 8A). The results showed that the disease incidence of dogs in the PBS group was 100% after challenge, with sustained weight loss. In contrast, both mi3-F1P2Fim2 immunized dogs and inactivated vaccine immunized dogs showed 100% protection, with only slight fluctuations in body weight post-challenge, and most dogs recovered within 4 days (Fig. 8D, E). In terms of clinical symptom scores, dogs in the PBS group generally exhibited noticeable respiratory infection symptoms, including purulent nasal discharge, frequent coughing, sneezing, reduced appetite and occasional vomiting. In contrast, mi3-F1P2Fim2 immunized dogs and inactivated vaccine immunized dogs only showed milder symptoms, primarily consisting of mild nasal discharge, with no significant changes in general condition, and clinical scores were significantly lower in these groups than those in the PBS group, indicating that mi3-F1P2Fim2 conferred complete protection (Fig. 8F). Furthermore, the bacterial load in the lungs of dogs in the PBS group was 10^7.2 CFU/g, indicating widespread bacterial colonization in the lungs in the absence of immunization, while a significant reduction of the bacterial load in the lungs of mi3-F1P2Fim2 immunized dogs (10^4.3 CFU/g) or inactivated vaccine immunized dogs (10^4.1 CFU/g) was observed (Fig. 8G). Consistently, lung tissue from dogs in the PBS group showed severe pathological damage, reflecting large-scale bacterial replication and intense inflammatory response in the host, while the lung tissue structure in dogs immunized with mi3-F1P2Fim2 or inactivated vaccine remained largely intact with no significant pathological abnormalities (Fig. 8H).

Above all, the mi3-F1P2Fim2 nanoparticle vaccine of B. bronchiseptica demonstrated good immunogenicity and complete protection in both mouse and dog models, which is comparable to traditional inactivated vaccine, suggesting that nanoparticle vaccine may serve as a promising alternative to traditional vaccines with greater programmability. Particularly, dogs, as the real host closely associated with humans, further highlight the potential of this vaccine in the prevention of zoonotic diseases and the translation of cross-species vaccine strategies. One limitation of this study is the absence of a systematic scaffold-only (SC003-mi3) control group, which precluded direct quantification of scaffold-specific immune responses, and this will be addressed in future studies.

Conclusions

In conclusion, we successfully constructed a trivalent nanocage vaccine, mi3-F1P2Fim2, targeting B. bronchiseptica, which induced strong humoral and cellular immune responses in mice and dogs. Mechanistic studies revealed that the enhanced immunogenicity of mi3-F1P2Fim2 was primarily driven by the efficient antigen uptake by APCs, sustained antigen retention and robust activation of germinal center reactions. Our study not only highlights the potential of trivalent nanocage vaccine in controlling zoonotic diseases under the One Health approach, but also emphasizes the significant value of a structurally programmable nanoparticle platform, which offers an efficient strategy for developing vaccines against complex pathogens and addressing shared health threats between animals and humans.

Materials and methods

Ethics statement

All animals involved in this study were housed in the Animal Facility at Huazhong Agricultural University. All animal experiments performed were grouped by random procedure and conducted in strict accordance with the recommendations in the Guidelines for the Ethical Treatment of Experimental Animals of the Ministry of Science and Technology of the People’s Republic of China. All experimental procedures were approved by the Animal Experiment Ethics Committee of Huazhong Agricultural University (Approval No: HZAUMO-2024-0355; HZAUDO-2025-0003).

Cell lines

DC2.4 cells were cultured in Roswell Park Memorial Institute 1640 medium (RPMI-1640, BioChannel Biotechnology Co., Ltd.) supplemented with 10% fetal bovine serum (FBS; QmSuero Biotechnology Co., Ltd.) and 1% penicillin-streptomycin (Biosharp). RAW264.7 macrophages were cultured in Dulbecco’s Modified Eagle Medium (DMEM; BioChannel Biotechnology Co., Ltd.) with the same supplements.

Bacterial strains and growth conditions

The B. bronchiseptica strain used in this study was isolated and preserved in our laboratory, designated as WH1218 [88]. B. bronchiseptica was cultured at 37 ℃ on Bordet-Gengou (BG) agar plates supplemented with 6% defibrinated sheep blood or in Tryptic Soy Broth (TSB) medium. As needed, fetal bovine serum (5% v/v) and 0.01% NAD solution (1% v/v) were added to the culture medium. All cultures were initiated from a single colony of the clonal population whenever possible.

Expression constructs

B. bronchiseptica WH1218 strain was used to amplify the main immunodominant type I domain of FHA (designated as F1, 1958–2112 aa), the main immunodominant region II of Prn (designated as P2, 514–609 aa), and the evolutionarily conserved domain of Fim2 (designated as Fim2, 28–209 aa). F1, P2, and Fim2 were codon-optimized for E. coli and synthesized by TsingKe Biological Technology (Wuhan, China). The F1-P2-Fim2 fusion gene was constructed using the Gibson assembly method and cloned into E. coli DH5α. Each domain was separated by a six-residue or eight-residue Gly-Ser linker. Each linker was distinct from others in the construct to reduce potential recombination risks and facilitate sequence analysis. The construction of pET32a-SpyTag003-F1P2Fim2 (ST003-F1P2Fim2) involved sequential insertion of SpyTag003, F1, P2, and Fim2 from the N-terminus to the C-terminus into the pET32a vector with a 6 × His tag at the C-terminus. pET32a-SpyTag003-F1Fim2P2, pET32a-SpyTag003-P2Fim2F1, pET32a-SpyTag003-P2F1Fim2, pET32a-SpyTag003-Fim2F1P2, and pET32a-SpyTag003-Fim2P2F1 (designated respectively as ST003-F1Fim2P2, ST003-P2Fim2F1, ST003-P2F1Fim2, SpyT003-Fim2F1P2, ST003-Fim2P2F1) were constructed based on ST003-F1P2Fim2.

pETStrepII-SpyCatcher003-mi3 (SC003-mi3) was created by replacing the vector in pET28a-SpyCatcher003-mi3 (Addgene 159995, GenBank MT945417) with the pETStrepII vector, which carries the Strep II tag (WSHPQFEK). The pETStrepII vector is a modified version of the pET28a (+) backbone, with the N-terminal His tag replaced by the Strep II tag (Strep-tag II).

Protein expression and purification

The above constructed ST003-F1P2Fim2 plasmid was transformed into BL21 (DE3) E. coli expression receptor cells. After heat-shock transformation, the transformation products were spread onto Luria-Bertani (LB) agar plates containing 50 µg/mL ampicillin and incubated at 37 ℃ overnight. The monoclonal colonies were inoculated into 10 mL of LB medium containing 50 µg/mL ampicillin and incubated at 37 ℃ for 16 h at 200 rpm with shaking. The pre-culture was then diluted 1:100 into 1 L of LB medium (containing 50 µg/mL ampicillin) and incubated at 37 ℃, 200 rpm with shaking. When the incubation reached an optical density (OD) of about 0.6 at 600 nm, isopropyl β-D-1-thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mM, and induction was continued for 12 h at 30 ℃, 200 rpm with shaking. The cells were harvested by centrifugation and resuspended in denaturing lysis buffer (20 mM Tris-HCl, pH 8.0, 300 mM NaCl, 8 M urea, and 20 mM imidazole). The suspension was homogenized and centrifuged at 14,000 g for 20 min at 4 ℃. The target protein was expressed as inclusion bodies and solubilized under denaturing conditions. The resulting supernatant was incubated with a His-tag purification resin column to enrich ST003-F1P2Fim2. The bound protein was eluted using elution buffer (20 mM Tris-HCl, pH 8.0, 300 mM NaCl, 8 M urea, and 300 mM imidazole). The purified protein was dialyzed, and the buffer was replaced with standard PBS buffer. The protein solutions were pooled and concentrated using a 30 kDa molecular weight cutoff centrifugal ultrafiltration device (Millipore). The level of recombinant protein expression was analyzed by 10% SDS-PAGE. The same purification method was applied for ST003-Fim2F1P2, ST003-Fim2P2F1, ST003-P2Fim2F1, ST003-P2F1Fim2, and ST003-F1Fim2P2.

The SC003-mi3 plasmid was transformed into BL21 (DE3) E. coli expression receptor cells. After heat-shock transformation, the transformation products were spread onto LB agar plates containing 50 µg/mL kanamycin and incubated at 37 ℃ overnight. The monoclonal colonies were inoculated into 10 mL of LB medium containing 50 µg/mL kanamycin and incubated at 37 ℃ for 16 h at 200 rpm with shaking. The pre-culture was then diluted 1:100 into 1 L of LB medium (containing 50 µg/mL kanamycin) and incubated at 37 ℃, 200 rpm with shaking. When the incubation reached an optical density (OD) of about 0.6 at 600 nm, IPTG at a final concentration of 0.2 mM was added and induction was continued at 18 ℃ for 16 h at 180 rpm with shaking. Bacteria were collected by centrifugation and resuspended in 50 mL of equilibration buffer (50 mM Tris-HCl, pH7.5, 150 mM NaCl, 1% Triton-100, 1 mM EDTA). The cell suspension was homogenized and centrifuged at 14,000 g for 20 min at 4 ℃, and the supernatant was collected and incubated with Strep-tag II purification resin column to enrich SC003-mi3. Protein was then eluted using elution buffer (100 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 50 mM biotin). The protein solutions were pooled and concentrated using a 100 kDa molecular weight cutoff centrifugal ultrafiltration device (Millipore). The level of recombinant protein expression was analyzed by 10% SDS-PAGE. The concentrations of protein were determined using a BCA protein assay kit (Beyotime). Endotoxin was removed from all of the proteins with the Protein Endotoxin Removal Kit (Beyotime). The amount of endotoxin was < 0.1 EU/mL. All proteins were stored at -80 ℃ until use.

Preparation trivalent nanocage vaccine against B. bronchiseptica

The purified SC003-mi3 and ST003-F1P2Fim2, ST003-F1Fim2P2, ST003-P2Fim2F1, ST003-P2F1Fim2, ST003-Fim2F1P2, ST003-Fim2P2F1 recombinant proteins were mixed at a molar ratio of 1:1 and slowly stirred overnight in PBS buffer at 4 ℃. The covalent couplings between SC003-mi3 and ST003-F1P2Fim2, ST003-F1Fim2P2, ST003-P2Fim2F1, ST003-P2F1Fim2, ST003-Fim2F1P2, ST003-Fim2P2F1 proteins were checked by 10% SDS-PAGE.ST003-F1P2Fim2 was concatenated with mi3 at different molar ratios, where the mi3-to-antigen ratios were 1:2, 2:1, 2:2, 2:3, 2:4, and 2:5, respectively. All conjugates were incubated at 4 ℃ for 6 h and possible aggregates were removed by centrifugation at 6000 g for 20 min at 4 ℃. SDS-PAGE was used to evaluate the binding. To prepare and purify the vaccine, 20 µM SC003-mi3 was incubated with 40 µM ST003-F1P2Fim2 (at 2:4 ratio of mi3 to antigen) overnight at 4 ℃ for an in vitro conjugation reaction. The mixture was then subjected to high-speed centrifugation for 10 min to remove aggregates. Purification was subsequently performed using SEC on a Superose 6 Increase 10/300 GL column pre-equilibrated with PBS buffer. The target protein was collected, concentrated, and stored at -80 ℃.

Stability tests of mi3-F1P2Fim2

Temperature and storage stability analyses were performed here. For thermal stability, aliquots of mi3-F1P2Fim2 were allowed to stand at 4, 16, 37, or 55 ℃ for 60 min and then cooled to 4 ℃ for 10 min. Possible aggregates were removed by centrifugation at 14,000 g for 20 min at 4 ℃. Supernatant samples were loaded onto gels for SDS-PAGE analysis. For storage stability, aliquots of mi3-F1P2Fim2 were stored at 4 or 25 ℃ for 30 days. Possible aggregates were removed by centrifugation at 14,000 g for 20 min at 4 ℃, and the supernatant samples were loaded onto a gel for SDS-PAGE analysis.

Cytotoxicity evaluation of mi3-F1P2Fim2

A 96-well plate was seeded with 100 µL of cell suspension containing either BMDCs, DC2.4 cells, or RAW264.7 macrophages at a density of approximately 5 × 10³ cells/well, and the cells were incubated in a 5% CO₂ incubator at 37 ℃. After the cells had adhered, 0–200 µg/mL of mi3-F1P2Fim2 were added to each well and incubated for 24 h. 10 μL of CCK-8 solution was added to each well, followed by incubation for 4 h. The OD of each well at 450 nm was measured using a spectrophotometer to assess cytotoxicity.

Transmission electron microscopic (TEM) analysis

TEM analysis was performed by placing the purified protein onto glow-discharged carbon-film-coated 300 mesh copper grids at a concentration of approximately 50 µg/ mL in PBS buffer. After adsorption for 1 min, the excess liquid was blotted with filter paper. Then the grids were stained with 2.0% uranyl acetate (UA) for 40 s and washed three times. Imaging was carried out by using a Talos L120C transmission electron microscope (Thermo Fisher Scientific).

Dynamic light scattering (DLS) analysis

DLS analysis was performed to measure the hydrodynamic diameter of SC003-mi3 and mi3-F1P2Fim2 using a Zetasizer Nano analyzer (Malvern Panalytical) at 25 ℃. All samples were centrifuged at 14,000 g and 4 ℃ for 15 min to remove any aggregates and diluted in PBS to a final concentration of 0.5 mg/mL. The Zetasizer Nano analyzer was used to determine the size distribution of the purified protein, with each sample analyzed in triplicate, and the data were analyzed using the Zetasizer software.

Preparation of inactivated B. bronchiseptica vaccine and whole cell protein of B. bronchiseptica

The B. bronchiseptica WH1218 strain was recovered from − 80 ℃ and allowed to reach room temperature before being plated onto BG agar plates for activation. After 48 h of incubation at 37 ℃, single colonies were picked and inoculated into 10 mL of TSB medium, which was then incubated at 37 ℃ for 16 h at 200 rpm with shaking. The pre-culture was diluted 1:100 into 100 mL of TSB medium and incubated at 37 ℃ for 16 h at 200 rpm with shaking. The bacterial concentration was adjusted to 10⁹ CFU/mL, and the bacteria were inactivated with β-propiolactone (1:1000) to prepare the inactivated vaccine.

The B. bronchiseptica were concentrated, washed three times with PBS, and then sonicated. After centrifugation at 14,000 g for 30 min, the supernatant was collected as the B. bronchiseptica whole cell protein. The concentration of the protein was determined with the BCA assay, and the preparation was stored at -80 ℃ until use.

Mouse immunization and challenge test study

All experimental animals were 6–8 weeks old female ICR mice. The mice were randomly divided into 4 groups (n = 15). Immunization with F1P2Fim2, mi3-F1P2Fim2, B. bronchiseptica WH1218 inactivated vaccine (1 × 10⁸ CFU/mouse), and PBS was performed twice at 14-day intervals by hindlimb intramuscular (i.m.) injection with Alum adjuvants (AddaVax). The immunization doses were 10 µg of F1P2Fim2 or the corresponding nanoparticle, ensuring an equimolar amount of antigen protein. At day 42, mice were challenged intranasally with 2.6 × LD₅₀ of B. bronchiseptica. Body weight loss and survival of the attacked mice were recorded daily for 14 days. Body weight loss of more than 25% was used as the humane endpoint. Five mice were randomly selected from each group to be euthanized on day 3, and lung tissues were isolated, observed, and analyzed histologically. Lung tissues (0.2 g) from mice were collected, homogenized in 2 mL of PBS, and appropriately diluted before being plated onto BG agar plates which were incubated at 37 ℃ for 36–48 h. The bacterial load was determined by CFU assay, with results expressed as CFU/g of lung tissue.

Dog immunization and challenge test study

Thirty dogs aged 8–9 weeks were randomly divided into three groups (n = 10). Immunization with mi3-F1P2Fim2, B. bronchiseptica WH1218 inactivated vaccine (1 × 10⁹ CFU/dog), and PBS was performed twice at 14-day i.m. injection with Alum adjuvants (AddaVax). Based on interspecies dose translation principles [89, 90], a 100 µg dose of mi3-F1P2Fim2 was selected for immunization. On day 42, dogs were challenged intranasally with 10⁹ CFU of B. bronchiseptica. Body weight loss, protection rate, and clinical symptoms [87] of the challenged dogs were recorded daily for 14 days. Five dogs were randomly selected from each group to be euthanized on day 3, and lung tissues were isolated, observed, and analyzed histologically. Lung tissues (0.2 g) from dogs were collected, homogenized in 2 mL of PBS, and appropriately diluted before being plated onto BG agar plates and incubated at 37 ℃ for 36–48 h. The bacterial load was determined by CFU assay, with results expressed as CFU/g of lung tissue.

Enzyme-Linked immunospot (ELISpot) assay

The secretion of IFN-γ and IL-4 was detected using a commercial ELISpot kit (DAKEWE). Two weeks after the booster immunization, mouse splenocytes were isolated and resuspended in RPMI-1640. Splenocytes (1 × 10⁵ cells per well) were seeded into pre-treated 96-well plates. The cells were then stimulated with 10 µg/mL of antigen protein and incubated at 37 ℃ with 5% CO₂ for 30 h. Subsequent steps were performed according to the manufacturer’s instructions. Spot counts were detected using an automated spot counter (ImmunoSpot Analyzers), and the spot-forming units (SFU) for each well were further analyzed.

Splenic lymphocyte proliferation assay and cytokine

Two weeks after booster immunization, splenocytes from each group of mice were seeded at a density of 1 × 10⁵ cells/well in 100 µL per well into 96-well cell culture plate. The cells were stimulated with 10 µg/mL of antigen protein and incubated at 37 ℃ with 5% CO₂ for 60 h. Each group included an antigen-stimulated group, a negative control group without antigen stimulation, and a blank control group containing only culture medium. For the splenocyte proliferation assay, after incubation, 20 µL of CCK-8 solution was added to each well, and the cells were incubated for an additional 4 h. The OD450 nm value of each well was measured using a microplate reader. SI = (OD of antigen-stimulated wells − OD of blank control wells) / (OD of negative control wells − OD of blank control wells).

For cytokine ELISA, ELISA kits for mouse IFN-γ, IL-4, and IL-2 (LiankeBio) were used to measure the secretion levels of IFN-γ, IL-4, and IL-2 in the supernatant of antigen-stimulated splenocytes as well as in the serum of mice two weeks after booster immunization.

Enzyme-Linked immunosorbent assay (ELISA)

ELISA was implemented to determine total IgG and antibody subclasses. Whole cell protein of B. bronchiseptica was diluted with coating buffer and added to ELISA plates, which were coated overnight at 4 ℃. The ELISA plate was washed three times with PBS containing Tween (PBST), followed by blocking with 5% skim milk at 37 ℃ for 2 h. After appropriate dilution, serum samples were incubated on the ELISA plate at 37 ℃ for 2. The plate was washed three times with PBST, and then 1 h incubation with horseradish peroxidase (HRP)-conjugated goat anti-mouse antibodies (IgG at 1:5000, IgG1 at 1:5000, IgG2a at 1:5000) was carried out at 37 ℃. Tetramethyl-benzidine substrate was added to the wells at room temperature and incubated for 30 min away from light. The reaction was terminated by adding 2 M H₂SO₄. Absorbance at OD450 nm was immediately measured using a SpectraMax 190 microplate reader (Molecular Devices, CA, USA).

For dog serum, the HRP-conjugated goat anti-mouse antibodies were replaced with goat anti-dog antibodies (IgG at 1:10000, IgM at 1:10000, IgG1 at 1:10000, IgG2 at 1:10000).

Activation and maturation of BMDCs in vitro

Under sterile conditions, bone marrow was isolated from 6 to 8 weeks old C57BL/6 mice and made into a single-cell suspension. These cells were cultured in 10% RPMI-1640 medium containing 20 ng/mL granulocyte-macrophage colony-stimulating factor (GM-CSF) (PEPROTECH) and 10 ng/mL IL-4 (PEPROTECH). On day 2 and day 4, half of the medium was replaced with fresh medium. On day 6, BMDCs were reseeded at a density of 4 × 10⁵ cells/well in a 24-well plate and grown at 37 ℃ in 5% CO₂ for 12 h. The immature BMDCs were incubated with F1P2Fim2, mi3-F1P2Fim2, or lipopolysaccharide (LPS, positive control) at 37 ℃, and supernatants and cells were collected after 24 h. The cells were incubated with anti-mouse CD16/CD32 Fc (eBioscience) block at 4 ℃ for 20 min to prevent nonspecific binding of antibodies, and stained with FITC anti-mouse CD11c, PE-Cy7 anti-mouse I-A/I-E, PE anti-mouse CD86, and APC anti-mouse CD80 (Biolegend). Co-stimulatory factors on BMDCs including MHC-II, CD80, and CD86 were analyzed by Cytek Aurora full-spectrum flow cytometer. The secretion levels of cytokines IL-6, IL-1β, and TNF-α in the culture supernatant were measured using an ELISA kit (LiankeBio) according to the manufacturer’s instructions.

Preparation of fluorescently labeled protein

To fluorescently label F1P2Fim2, iFluor™ 594 Succinimidyl Ester (AAT Bioquest, USA) was dissolved in dimethyl sulfoxide (DMSO) and added to F1P2Fim2 protein in 0.1 M sodium bicarbonate buffer (pH 8.3), with a dye-to-protein molar ratio of 20:1. The reaction was carried out at room temperature with stirring at 600 rpm on a magnetic stirrer for 4 h. Excess unbound dye was removed by Sephadex G-25 gel filtration. The labeled protein was termed 594-tagged F1P2Fim2. The same method was used to prepare Cyanine 7 monosuccinimidyl ester (AAT Bioquest, USA) -tagged F1P2Fim2 (termed Cy7-tagged F1P2Fim2).

Antigen uptake in vitro

BMDCs, DC2.4 cells, and RAW264.7 macrophages were seeded into 15-mm culture dishes (Biosharp) to form a monolayer, followed by incubation with equal amounts of iFluor™ 594-tagged F1P2Fim2 or mi3-F1P2Fim2 for 6 h. Cells were fixed with 4% paraformaldehyde for 30 min, washed three times with PBS, and then permeabilized with 0.1% Triton X-100 at room temperature for 10 min. The cells were then blocked with 10% goat serum at room temperature for 30 min, followed by staining with DAPI (1:500) for 5 min. After three washes with PBS, imaging was immediately performed using a Nikon A1HD25 super-resolution laser scanning confocal microscope equipped with a 100 × NA1.49 oil immersion objective lens. The average fluorescence intensity (MFI) was analyzed using ImageJ software.

In vivo imaging in mice

Twenty-four 6–8 weeks old female ICR mice were randomly divided into three groups, with 8 mice in each group. Each mouse received an intramuscular injection (left hind leg) of 10 µg Cy7-tagged F1P2Fim2, Cy7-tagged mi3-F1P2Fim2, or PBS. In vivo imaging was performed under anesthesia with isoflurane, and the retention time of Cy7-tagged antigen was monitored continuously for 12 days with IVIS Spectrum system (Caliper Life Sciences, USA). Fluorescence intensity at the injection site was semi-quantitatively analyzed using the Living Image 4.4 software. For further analysis, three mice from each group were randomly selected and sacrificed 4 h post-injection to assess the biodistribution of the fluorescent antigen in the draining lymph nodes was studied.

Flow cytometry

To identify DCs, freshly isolated iLNs were collected 3 days after post boost immunization and homogenized into a single-cell suspension using a syringe plunger, then passed through a 40 μm nylon filter (SPL Life Sciences, Korea). After two washes with PBS, the cells were blocked with PBS buffer containing 0.2% bovine serum albumin (BSA). Next, the cells were stained with cocktails of the following fluorescently labeled antibodies: PE anti-mouse CD86, APC anti-mouse CD80, PE/Cyanine7 anti-mouse I-A/I-E, and FITC anti-mouse CD11c. Cells were washed three times and resuspended in PBS for flow cytometry analysis.

To identify CD4⁺ and CD8⁺ T cells, freshly isolated splenocytes were collected 14 days after secondary immunization, homogenized into a single-cell suspension using a syringe plunger, and passed through a 40 μm nylon filter (SPL Life Sciences, Korea). Red blood cells were lysed with ACK lysis buffer (BioSource, California, USA). After two washes with PBS, the cells were incubated with anti-mouse CD16/CD32 Fc block at 4 ℃ for 20 min to prevent nonspecific binding of antibodies. Next, the cells were stained with cocktails of the following fluorescently labeled antibodies: FITC anti-mouse CD4, APC anti-mouse CD3, and PE anti-mouse CD8α. Cells were washed three times and resuspended in PBS for flow cytometry analysis.

To identify total GC B cells, Tfh cells, and LLPCs, freshly isolated iLNs, splenocytes, or bone marrow (BM) were collected 14 days after secondary immunization, processed into single-cell suspensions, and stained with a cocktail of the following surface antibodies: FITC anti-mouse/human CD45R/B220, Alexa Fluor^® 647 anti-mouse/human GL7, PE anti-mouse CD95 (Fas), FITC anti-mouse CD4, PE anti-mouse CD279 (PD-1), APC anti-mouse CD185 (CXCR5), APC anti-mouse CD138. Cells were washed three times and resuspended in PBS for flow cytometry analysis.

Flow cytometry was performed on a Cytek Aurora spectral flow cytometer (Cytek Biosciences, USA), and data were analyzed using FlowJo software (version 10).

Immunofluorescence

The iLNs of mice were fixed in 4% paraformaldehyde for 24 h, dehydrated with 30% (wt/vol) sucrose, and then embedded in OCT and frozen at -20 ℃. The frozen lymph nodes were sectioned into 10 μm-thick sections using a Leica CM1950 cryostat (Leica, Switzerland). The sections were blocked at room temperature for 2 h in PBS containing 10% normal goat serum and incubated with a germinal center staining cocktail containing the following antibodies: Brilliant Violet 421™ anti-mouse/human CD45R/B220 (1:200 dilution), Alexa Fluor^® 594 goat anti-mouse IgG (1:400 dilution), and AlexaFluor^®488 anti-mouse/human GL7 (1:200 dilution). Germinal center numbers were calculated according to the quantity of GL7-positive cell clusters. All images were captured with a high-resolution fast cell imaging fluorescence microscope (Leica, HS0037 Thunder DMi8).

Histopathology and immunohistochemistry

After euthanizing the mice and dogs, major organs were collected and fixed in 4% paraformaldehyde buffer for 24 h. They were then embedded in paraffin, followed by sectioning and staining with hematoxylin and eosin (H&E) for histopathological examination.

For immunohistochemistry, iLNs were collected from mice 14 days after the booster immunization and fixed in 4% paraformaldehyde for 24 h at 4 ℃. The sections were subsequently dehydrated, embedded in paraffin, and sectioned at a thickness of 5 μm. The tissue was deparaffinized and subjected to antigen retrieval by microwave heating in 10 mM citrate buffer (pH 6.0) for 15 min. Endogenous peroxidase activity was blocked by treating the slides with 3% hydrogen peroxide (H₂O₂) for 10 min, followed by three washes with PBS. To prevent nonspecific binding, the slides were blocked with 10% goat serum at room temperature for 30 min. The slides were then incubated overnight at 4 ℃ with anti-Ki67 antibody (Servicebio, 1:500 dilution). After three washes with PBS, the slides were incubated with HRP-conjugated secondary antibody at room temperature for 1 h. The signal was developed with a DAB substrate kit (Servicebio), followed by hematoxylin counterstaining, dehydration, and mounting. The slides were then scanned using a PannoramicSCAN (3DHISTECH, Budapest, Hungary) and images were analyzed and recorded using CaseViewer software (version 2.4). Immunohistochemical staining for Ki67 was evaluated semi-quantitatively. The staining intensity was scored as 0 (negative), 1 (weak), 2 (moderate), or 3 (strong), and the percentage of positively stained cells was scored as 0 (0–5%), 1 (6–35%), 2 (36–70%), or 3 (> 70%). The final IHC score was calculated as the product of these two parameters, and a score of ≥ 4 was considered to indicate high Ki67 expression, while a score < 4 indicated low expression. All evaluations were performed independently by three experienced pathologists in a blinded manner.

Homology modeling

The 3D structure of the B. bronchiseptica FHA protein (GenBank: AF111796), Prn (GenBank: AJ245927), and Fim2 (GenBank: KJ152685.1) were predicted using AlphaFold3 based on their amino acid sequences. To construct the 3D structural models of SpyTag003-SpyCatcher003 (PDB ID: 4MLI) and mi3 (PDB ID: 9FO3), templates from the Protein Data Bank (PDB) were selected. Model visualization and structural alignment were generated using the Python-based molecular viewer PyMOL (The PyMOL Molecular Graphics System, version 1.7.4, Schrödinger, LLC).

Statistical analyses

All data and analyses were performed using GraphPad Prism (version 9.0). Survival differences were analyzed using the log-rank (Mantel-Cox) test. For other data, variance analysis was conducted using one-way or two-way analysis of variance (ANOVA). All statistical tests were considered statistically significant with p-values less than 0.05 (P < 0.05) (*P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant).

Author contributions

W.H. designed the study, analyzed the data, and drafted the manuscript. W.H. performed the animal experiments. L.W., X.Z., Z.R., F.C., and Q.S. assisted in the experiments. J.W., Z.S., Z.F., and M.Z. provided technical advice. L.Z. and M.Z. designed the research, proofread the manuscript, and provided financial support. All authors have approved the final version of the manuscript.

Funding

This work was supported by the National Key Research and Development Program of China (2022YFD1800100) and the Fundamental Research Funds for the Central Universities(2662023PY005).

Data availability

All data associated with this study are present in the article or the Supplementary Materials. Materials can be made available upon reasonable request under a completed Material Transfer Agreement.

Declarations

Competing interests

The authors declare no competing interests.