Identification of efficient multi-epitope combinations against African swine fever virus based on AP205 scaffold-mediated nanodisplay technology

Abstract

Background

African swine fever virus (ASFV), characterized by its large genome and high antigenic diversity, has impeded the development of traditional vaccines. In this study, we employed the AP205 capsid protein as the scaffold to prepare ASFV multi-epitope nanoparticles via sequential steps: fusion of the scaffold gene with ASFV multi-epitope genes, followed by recombinant protein expression, purification, in vitro self-assembly, structural characterization, and immune evaluation in BALB/c mice.

Results

It was found that the dimeric AP205 scaffold enhanced the assembly efficiency of nanoparticles harboring epitopes of ASFV, and the recombinant proteins DB, 22C, and DE, fused with the scaffold and the epitopes, all formed virus-like particles (VLPs). Immunization experiments demonstrated that DE, which carries the linear B-cell epitopes pB438L 51–73, p72 244–262, and p54 132–180, exhibited the optimal immunological efficacy, with outstanding performance in immunoreactivity, immunogenicity, specific antibody titers, immune cell activation, and inhibition of ASFV nucleic acid replication by immune serum. The nanoparticle DF (carrying linear B-cell epitopes p17 5–34, 71–113, and p30 4–15) showed comparable or even superior efficacy in all aspects except for slightly lower antibody levels, while DB (carrying linear B-cell epitopes p17 68–86, pE248R 138–168, 158–185) and 22C (carrying linear B-cell epitopes pE120R 65–117, pE199L 175–189, pO61R 36–56) yielded moderate efficacy.

Conclusions

The dimeric AP205 scaffold tolerates the insertion of exogenous antigen fragments with greater complexity and size, and its simplified fusion display strategy facilitates commercial production. The spatial structure of VLPs confers significant advantages in inducing efficient immune responses. Furthermore, the VLP structures designed and validated in this study provide valuable data for amino acid deconstruction of analogous architectures. Meanwhile, this study identified efficient and precise ASFV epitope combinations, providing informative data to support the application of ASFV multi-epitope nanoparticle vaccines.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12951-026-04608-7.

Introduction

African swine fever virus (ASFV) is an extremely large linear double-stranded DNA virus with an icosahedral structure, approximately 200 nm in diameter. Its genome spans 170–190 kb and encodes more than 150 proteins [1]. ASFV exhibits complex biological characteristics, and research on the mechanisms of immune protection and immune evasion induced by the virus remains limited. However, African swine fever (ASF), the disease caused by ASFV, leads to high mortality rates in pigs and results in substantial economic losses [2]. Currently, extensive research and development efforts are focused on effective vaccines and therapeutics against ASF [3]. Among these, studies on nanoparticle vaccines assembled from recombinant ASFV proteins represent an ideal strategy that prioritizes safe production and application while ensuring robust immunological efficacy [4–6].

The design of multi-epitope vaccines is anchored in the conceptual framework of precision epitope screening, efficient integration, and targeted immune activation. Compared with traditional whole-virus vaccines, they possess prominent advantages in multiple dimensions [7]. First, This type of vaccine elicits a more comprehensive immune protection profile. Specifically, B-cell epitopes can stimulate the host to produce antibodies that directly block pathogen invasion of host cells, whereas T-cell epitopes are able to activate cellular immunity to eliminate infected cells. The coexistence of these two types of epitopes achieves a dual effect of resisting viral adsorption and clearing infected cells, thereby effectively enhancing the immune protection rate [8]. Second, multi-epitope vaccines demonstrate exceptional resistance to pathogen variation [9]. By incorporating multiple conserved epitopes of the pathogen, these vaccines can significantly broaden the scope of immune protection, thereby overcoming a key limitation of whole-virus vaccines: loss of efficacy caused by antigenic drift or shift [10]. Moreover, high safety stands as one of their most prominent advantages. During the design process, virulence genes, immunosuppressive epitopes, and fragments that may induce autoimmune responses are deliberately excluded, with only immunoprotective epitopes retained. This design strategy inherently circumvents potential risks associated with whole-virus vaccines, such as residual virulence, allergic reactions, and autoimmune complications [11].

Nanoparticle vaccine platforms, characterized by unique physicochemical properties and biological functions, have emerged as efficient alternatives to traditional vaccines [12]. Nanoparticles, typically ranging in size from 20 to 200 nm, mimic the natural dimensions of pathogens. This similarity facilitates their uptake by antigen-presenting cells, such as dendritic cells and macrophages, through phagocytosis or endocytosis [13]. Concurrently, they enable the targeted delivery of antigens to antigen-processing compartments in APCs, including lysosomes and endoplasmic reticulum, promoting antigen presentation by both MHC class I and MHC class II molecules [14]. Furthermore, nanoparticles, especially virus-like particles (VLPs), possess ordered and repetitive protein structures that significantly enhance B-cell receptor (BCR) cross-linking, triggering a robust primary signal for B-cell activation [15–18]. Ultimately, these coordinated processes elicit stronger humoral immunity and cellular immunity [19].

In both laboratory research and commercial manufacturing, prokaryotic expression systems remain the gold-standard platform for nanoparticle protein production: Prokaryotes (e.g., Escherichia coli) can be cultured within 1–2 days, with interest protein yields peaking at 50% of total cellular protein [20]. Fermentation processes are readily scalable from shake flasks to thousand-liter bioreactors, with concomitantly low production costs [21]. Critically, prokaryotic systems are particularly well-suited for proteins lacking complex post-translational modifications (PTMs). As core nanocarriers, phage capsids confer distinct advantages: efficient self-assembly, robust structural stability (tolerating pH 3–9 and temperatures ≤ 60 °C) [22], and amenable genetic engineering. Currently, capsids derived from phages such as MS2 and P22 have been leveraged for drug delivery applications: they can encapsulate small interfering RNA (siRNA) and chemotherapeutic agents, thereby facilitating in vivo transport and shielding the cargo from degradation [23, 24]. However, the capsids of most bacteriophages (e.g., Qβ and MS2) lack the capacity to accommodate the insertion of long-chain antigen fragments [25, 26]. In addition, the capsid proteins of porcine circovirus and porcine parvovirus, which can be expressed in prokaryotic systems, are also not suitable for the insertion of ultra-long foreign fragments [27–32]. In contrast, the bacteriophage AP205 capsid uniquely circumvents this limitation: its N- and C-termini not only permit the integration of ultra-long exogenous fragments but also facilitate efficient self-assembly into nanoparticles post-insertion [33]. More importantly, both termini are exposed on the particle surface, thereby enabling the direct and efficient presentation of exogenous fragments [34, 35].

In prokaryotic expression systems, the assembly structure of AP205 contains more subunits compared to other protein scaffolds. Specifically, AP205 nanoparticles are self-assembled from 180 protein subunits into a T = 3 symmetric structure with an approximate diameter of 30 nm [27, 36, 37]. By comparison, bacterial ferritin forms a T = 1 octahedral symmetric structure comprising merely 12 subunits, with a diameter of ~10 nm [37]. Lumazine synthase (AaLS) from the archaeon Aquifex aeolicus adopts a T = 1 icosahedral conformation composed of 60 identical subunits, corresponding to a particle diameter of around 16 nm [38]. Similarly, the computationally designed mi3 scaffold, porcine circovirus capsid protein, and porcine parvovirus capsid protein each consist of 60 subunits [39–42]. Collectively, these structural features imply that the AP205 nanoparticle surface may accommodate a greater number of epitope insertions, potentially enabling more effective B-cell receptor (BCR) cross-linking.

Combining high insertion capacity with efficient surface exposure, the AP205 capsid exhibits dual advantages that render it an optimal vector for displaying ASFV epitopes, particularly linear B-cell epitopes, which aligns seamlessly with the core requirements of epitope display technology.

In prior work from our laboratory, we have successfully screened a series of linear B-cell epitopes of ASFV. These epitopes are derived from a range of key structural proteins including but not limited to p72, pB438L, p17 and p54, with a systematic comparison of the immunoreactivity and immunogenicity of each epitope already completed. The current study will further advance the construction and functional validation of several multi-epitope nanoparticle vaccines targeting ASFV.

In the specific experimental design, the ASFV linear B-cell epitopes previously identified in our laboratory [43] and the ASFV T-cell epitopes reported in the literature [44] were tandemly linked separately. The gene sequences of the linked epitopes were ligated to the N- or C-terminal coding sequence of the bacteriophage AP205 capsid protein to construct expression vectors for the recombinant proteins. The recombinant proteins were expressed using a prokaryotic expression system, followed by in vitro purification, assembly of their nanoparticle structures, structural analysis, and structural validation of epitope display on nanoparticles. After confirming the binding capacity of these nanoparticles to ASFV-positive serum (immunoreactivity), we systematically evaluated the immune efficacy of the multi-epitope nanoparticles in a mouse model. This systematic evaluation included: measuring total serum antibody titers and ASFV epitope-specific antibody levels in immunized mice to determine immunogenicity; evaluating the activation effect of the nanoparticles on bone marrow-derived dendritic cells (BMDCs); assessing the activation status of follicular helper CD4 + T (Tfh) cells; analyzing the activation status of germinal center cells; detecting levels of key cytokines including IFN-γ, IL-4, TNF-α, IL-2; quantifying memory lymphocyte levels; monitoring changes in body weight and assessing the histopathological impact of the nanoparticle vaccine on mouse tissues and organs during immunization. Additionally, the inhibitory activity of immune serum on ASFV nucleic acid replication was verified through an in vitro virus inhibition assay.

Materials and methods

Mice, inactivated virus, serum samples, and the conduct of serum-mediated viral replication inhibition assays

The specific-pathogen-free (SPF) female BALB/c mice, 6–8-week-old, used in the experiments were obtained from the Experimental Animal Center of the Lanzhou Veterinary Research Institute. The inactivation of ASFV (ASFV/II/CN/SC/2019 strain) and the serum-mediated viral replication inhibition assays were both carried out in the Animal Biosafety Level 3 facilities (ABSL-3) at Lanzhou Veterinary Research Institute. Anti-ASFV positive serum was purchased from China Institute of Veterinary Drug Control.

Design of recombinant protein sequences

A total of nine recombinant protein sequences were designed by concatenating multiple linear B-cell epitopes of ASFV previously identified in our laboratory and T-cell epitopes reported in the literature using flexible linkers, which were inserted into either the C-terminus or N-terminus of the AP205 capsid protein (GeneID:956,335, Protein ID: NP_085472.1) scaffold. Specifically, the DB recombinant protein contained concatenated epitopes p17 68–86 and pE248R 138–168, 158–185; the 22 C recombinant protein included pE120R 65–117, pE199L 175–189, and p12 (pO61R) 36–56; 22D and DD shared identical epitope combinations: pB438L 212–230, pH108R 33–46, pE248R 121–136, and p72 352–363, 362–372; 22E and DE also contained the same three epitopes: pB438L 51–73, p72 244–262, and p54 132–180; 22 F and DF shared epitopes p17 5–34, 71–113, and p30 4–15; DTA was composed of T-cell epitopes derived from pp220 (CP2475L, 124–132, 393–401, 788–796, 2125–2133, 1375–1383) and pp62 (CP530R, 227–235, 263–271). The GenBank information and amino acid sequences of these epitopes was provided in Table 1. In addition, conservation analysis was performed on the selected epitopes based on 13 genotype II African swine fever virus (ASFV) isolates from China. The results showed that the vast majority of the screened epitopes were highly conserved among genotype II ASFV strains; only the epitope p30 (4–15) exhibited homology of 58% in one isolate (see Table S10.A and Table S10.B for details). This conservative analysis result references the previously published data of this experiment [43].

Table 1.

GenBank information of the epitopes contained in the recombinant protein

Recombinant proteins	The contained epitopes (sites)	The amino acid sequence of the epitope	The corresponding ASFV protein	Protein ID
DB	63–86	TIDCKSSIPKPPPSYYVQQPEPHH	p17	QBH90595.1
	138–168	GNIVENATQKQSQQIISNCLQGSKQAIDTTT	pE248R	QBH90619.1
	158–185	QGSKQAIDTTTGITNTVNQYSHYTSKNF	pE248R	QBH90619.1
22C	65–117	EKIAEEEDINFLHDTREFTSLVPDEADNKPEDDEESGAKPKKKKHLFPKLSSH	pE120R	YP_009703416.1
	175–189	SSSGGKHNDDKLPPP	pE199L	QBH90617.1
	36–56	MPRQQKKCSKAEECTCNNGSC	p12 (pO61R)	QBH90586.1
DD (22D)	352–363	FGERFITIKLAS	p72 (pB646L)	QBH90570.1
	362–372	ASQKDLVNEFP	p72 (pB646L)	QBH90570.1
	33–46	VTDYIIINTQYTEQ	pH108R	QBH90605.1
	121–136	NCVSSLSGMNVLVVKG	pE248R	QBH90619.1
	212–230	KDVYNILCLAHNLPASVEK	pB438L (p49)	QBH90564.1
DE (22E)	244–262	DLHKPHQSKPILTDENDTQ	p72 (pB646L)	QBH90570.1
	132–180	GPAAAPAAASAPAHPAEPYTTVTTQNTASQTMSAIENLRQRNTYTHKDL	p54 (pE183L)	QBH90613.1
	51–73	NEYTNSPEKAEKGLQLSDLRNFS	pB438L (p49)	QBH90564.1
DF (22F)	5–34	TSPLLSHNLSTREGIKQSTQGLLAHTIAKY	p17	QBH90595.1
	71–113	PKPPPSYYVQQPEPHHHFPVFFRKRKNSTSLQSHIPSDEQLAE	p17	QBH90595.1
	4–15	ILNISMKMEVIF	p30 (CP204L)	QBH90581.1
DTA	124–132	IADAINQEF	pp220 (CP2475L)	QBH90580.1
	393–401	FLNKSTQAY
	788–796	QIYKTLLEY
	2125–2133	SLYPTQFDY
	1375–1383	RSNPGSFYW
	227–235	DFDPLVTFY	pp62(CP530R)	QBH90582.1
	263–271	GTDLYQSAM

Furthermore, in constructs 22 C, 22D, 22E, and 22 F, the ASFV epitopes were fused to the N-terminus of the monomeric AP205 capsid protein scaffold, as illustrated in Fig. 1A. For DB, DD, DE, DF, and DTA, the ASFV epitopes were inserted into the C-terminus of the dimeric AP205 capsid protein scaffold, as shown in Fig. 1B. The positions of the His-tag sequences differed slightly among the recombinant proteins. The linkers used in this study include GGGGSGGGGSGGGGS, GSGTAGGGSGS, GGGGS, GSGESG, GGSGS and GGS. Detailed information on the linkage positions and amino acid sequences of linkers, epitopes, AP205 capsid protein, and His-tags for all recombinant proteins, including the AP205 scaffold protein experimental control, is provided in Table S1. All expression gene sequences of the recombinant proteins were cloned into the pET-28a (+) vector, and plasmid synthesis for recombinant protein expression was carried out by GenScript.

Fig. 1.

Schematic diagram of arrangement of recombinant protein. In the recombinant proteins constructed in this study, the connection modes of the ASFV multi-epitope sequences with the scaffold AP205 amino acid sequence are as follows: A involves the ASFV multi-epitope sequence fused to the N-terminus of the monomeric AP205 capsid protein, while B involves the ASFV multi-epitope sequence fused to the C-terminus of the dimeric AP205 capsid protein sequence. All epitopes and the AP205 scaffold are connected via linkers.

Expression of recombinant proteins

The recombinant plasmids encoding ASFV multi-epitope recombinant proteins were transformed into Escherichia coli Rosetta (DE3) competent cells. The transformed expression strains were inoculated into 2 mL lysogeny broth (LB) liquid medium supplemented with 50 μg/mL kanamycin and incubated at 37 °C with shaking at 220 rpm for 16 h. The resulting culture was then diluted 1:100 into 100 mL of fresh kanamycin-selective LB medium for scale-up cultivation. Subsequently, the culture was further diluted 1:100 and inoculated into 2 L of fresh kanamycin-selective LB medium, followed by incubation at 37 °C until the optical density at 600 nm (OD₆₀₀) reached 0.6–0.8. Protein expression was induced by adding 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) to the culture, which was then incubated with shaking at 16 °C for 16 h. A small aliquot of the culture was collected for analysis of recombinant protein expression via sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE).

Purification of proteins

Escherichia coli cultures were harvested via centrifugation at 6,000 × g for 30 min, and the pellet was collected. The bacterial cells were subsequently resuspended in phosphate-buffered saline (PBS), followed by centrifugation at 8,000 × g for 15 min to collect the bacterial pellet. The pellet was mixed thoroughly with a protein denaturation buffer (8 M urea, 20 mM Tris–HCl, 0.5 M NaCl, 5 mM imidazole, pH 7.9) at a ratio of 1 g bacterial pellet to 10 mL denaturation buffer, and Triton X-114 was added at a ratio of 100:1 (v/v, denaturation buffer: Triton X-114). The mixture was then subjected to ultrasonic disruption on ice (260 W, 3-s pulse, 3-s pause) with the duration determined by the volume of denaturation buffer added (i.e., 1 min per 2 mL of buffer). The sonicated mixture was incubated on a rotator at 4 °C for 30 min, followed by incubation in a 30 °C water bath for 30 min; the supernatant was collected and stored for subsequent use. A purification column was equilibrated with an equal volume of protein denaturation buffer. The prepared denatured sample containing the target protein was mixed with nickel-charged resin at a ratio of 5:1 (v/v, sample: resin) and incubated on a rotator at 4 °C for 2 h. The resin was then washed five times with a wash buffer (6 M urea, 20 mM Tris–HCl, 0.5 M NaCl, 60 mM imidazole, pH 7.9); for each wash, the resin was mixed thoroughly with the wash buffer for 5 min, and the buffer was allowed to flow out completely under gravity. Finally, an equal volume of protein elution buffer (6 M urea, 20 mM Tris–HCl, 0.5 M NaCl, 500 mM imidazole, pH 7.9) relative to the resin was added, mixed by gentle shaking for 30 min, and the resulting eluate was collected. The purified recombinant protein was verified via SDS-PAGE analysis, and its concentration was determined using the Bradford Protein Assay (Beyotime Biotech, Shanghai, China) combined with ImageJ software. The purified protein was stored at −80 °C until further use.

In vitro assembly and purification of nanoparticles

The purified recombinant protein was loaded into a 3 kDa dialysis membrane for in vitro dialysis-induced nanoparticle assembly. First, the protein-containing dialysis membrane was sequentially immersed in three assembly buffers: Assembly Buffer I (4 M urea, 20 mM Tris–HCl, 100 mM NaCl); Assembly Buffer II (2 M urea, 20 mM Tris–HCl, 0.5 M arginine, 5 mM reduced glutathione, 0.5 mM L-oxidized glutathione, 10 mM NaCl, 5% glycerol); Assembly Buffer III (20 mM Tris–HCl, 0.5 M arginine, 5 mM reduced glutathione, 0.5 mM L-oxidized glutathione, 5% glycerol). Two aliquots of Assembly Buffer III were prepared for duplicate dialysis. The pH of all assembly buffers was adjusted according to the isoelectric point (pI) of the recombinant protein (i.e., pH ≠ pI). The pI values of the proteins could be obtained by inputting their amino acid sequences into the website http://isoelectric.org/, and the pI values of the 9 recombinant proteins designed in this study are provided in Table S1. The volume ratio of the recombinant protein to each assembly buffer was 1:200, with dialysis performed at 4–8 °C for 12 h per stage. Subsequently, an endotoxin detection kit (Bioendo, Xiamen, China) was used to determine the endotoxin levels of the dialyzed proteins via the recombinant C-factor fluorescence method. For proteins prone to precipitation, 0.2‰ Triton X-114 was added. After dialysis-induced assembly, the protein sample was centrifuged at 10,000 × g for 10 min to remove precipitates. If the protein concentration post-centrifugation was below 600 μg/mL, the sample was concentrated using 50 kDa cutoff ultrafiltration centrifugal tubes. The assembled protein was stored on ice for immediate use or at −80 °C for long-term storage.

Size exclusion chromatography (SEC) was employed to purify the assembled nanoparticles based on their particle size. First, a CL-4B Sepharose chromatography column (10 mm ×700 mm) installed on an AKTA pure protein purification system (Tissot AC, et al. 2010) was equilibrated with Assembly Buffer III corresponding to the assembled protein [45]. Once the UV280 nm absorbance peak curve stabilized, 500 μL of the assembled protein was loaded into the system’s sample loop for separation, with a flow rate set at 1 mL/min. Fractions of the assembled protein were collected at distinct stages based on alterations in the UV280 nm absorbance profile and stored on ice until subsequent use.

Transmission electron microscopy and validation of nanoparticle stability

Transmission electron microscopy (TEM) was used to characterize the SEC-purified assembled proteins. First, the assembled protein was diluted to a concentration of 30 μg/mL, and 20 μL of the diluted sample was dropped onto the surface of a grid-supported film. After incubating for 10 min at room temperature, the excess liquid on the film surface was blotted dry with filter paper via siphonage. Subsequently, 1% phosphotungstic acid (PTA) was added; after 15 s of incubation, the residual staining solution was blotted away, and the grid was air-dried naturally. TEM was then employed to observe the formation of nanoparticles or VLP structures by the assembled proteins.

Meanwhile, the SEC-purified assembled proteins were subjected to two freeze–thaw cycles. After the cycles, the corresponding samples were prepared and observed via TEM following the same protocol described above.

Immunoelectron microscopy colloid gold labeling experiments

Epitope antibody binding assays were performed on nanoparticles assembled from recombinant proteins carrying linear B-cell epitopes of ASFV, to verify whether the corresponding antigenic epitopes were successfully displayed on the nanoparticle surface.

Preparation of epitope-specific antibodies

In this study, GST-tagged recombinant epitope proteins were constructed and expressed corresponding to each ASFV antigenic epitope contained in the nanoparticles DB, 22 C, DD, DE, and DF. The epitope sequences corresponding to each nanoparticle are detailed in Table 1. The coding genes for the linear B-cell epitopes of ASFV were fused to the GST gene in the pGEX-4 T-1 vector via a flexible linker (GGGGSS) to construct recombinant epitope plasmids with GST as the fusion tag. All plasmids were synthesized and constructed by GenScript Biotech Corporation.

All recombinant plasmids were transformed into Escherichia coli Rosetta (DE3) competent cells, followed by cell culture, protein induction, identification of expressed products, and purification. Finally, GST-fused recombinant proteins harboring each linear B-cell epitope of ASFV were obtained.

Mice were immunized with the above recombinant epitope proteins. Whole blood was collected 7 days after the third immunization, and serum was isolated. To verify the presence of epitope-specific antibodies in serum, an indirect ELISA was performed using inactivated ASFV as the coating antigen. When the difference in Pi/Ni ratios between the experimental group and the negative control group is statistically significant (P < 0.05), the epitope is determined to be immunogenicity, indicating that epitope-specific antibodies are present in the serum. Detailed procedures for the design and expression of recombinant epitope proteins, mouse immunization, and antibody validation were performed according to a previously published study from our laboratory [43]. Subsequently, serum containing epitope-specific antibodies was purified using Protein G agarose resin (APExBIO, USA). The detailed purification protocol is provided in Protocol 1 of Additional file 1.

Immobilization of nanoparticles

First, glow discharge treatment was performed on thin carbon support films (Zhongjing Keyi Technology Co., Ltd., China). The assembled nanoparticle proteins (purified by SEC) were diluted to 7.5 μg/mL. A 25 μL aliquot of the nanoparticle solution was dropped onto a polyethylene blown film, and the thin carbon support film was placed face down onto the droplet surface, followed by incubation for 10 min to allow sufficient adsorption of nanoparticles onto the support film.

Following removal of excess liquid from the support film using filter paper, the film was transferred to a droplet of 4% paraformaldehyde and fixed at room temperature for 40 min, followed by additional fixation with 0.2% glutaraldehyde for 15 min.

After discarding the remaining fixative, the support film was rinsed once with TBS buffer (20 mM Tris, 150 mM NaCl, pH 7.4), and the glutaraldehyde-induced cross-linking was quenched using glycine solution (100 mM glycine in TBS, pH 7.4).

Epitope-specific antibody binding

The support film with immobilized nanoparticle proteins was rinsed three times with TBST buffer (TBS containing 0.05% Tween 20 and 0.05% Triton X-100, pH 7.4), with each rinse incubated for 2 min followed by removal of excess liquid; all subsequent rinsing steps were performed in the same manner.

The support film was then incubated in an epitope-specific antibody solution at a concentration of 1.5 μg/mL at 37 ℃ for 40 min. After incubation with each epitope antibody, the film was rinsed three times with TBST (pH 7.4), and the above incubation procedure was repeated to sequentially incubate the same nanoparticle protein with its corresponding different epitope antibodies. In addition, DB particles incubated with non-specific antibodies were used as the negative control.

Binding of gold-labeled anti-mouse secondary antibody

After incubation with the epitope antibodies, the support film was rinsed once with TBST buffer (TBS containing 0.05% Tween 20 and 0.05% Triton X-100, pH 8.1). The film was then transferred into 6 nm colloidal gold-labeled goat anti-mouse IgG solution (diluted 25-fold; Abcam, Shanghai, China) and incubated at 37 °C for 100 min. Following incubation, the film was rinsed twice with TBST (pH 8.1) and twice with TBS (pH 8.1) in sequence.

Negative staining and TEM observation

The support film was negatively stained with 2% phosphotungstic acid (PTA) for 30 s, rinsed once with pure water, and air-dried prior to observation and imaging using a transmission electron microscope.

Notes

Nanoparticle proteins were diluted with assembly buffer; epitope antibodies were diluted in TBST (pH 7.4); and gold-labeled goat anti-mouse secondary antibody was diluted in TBST (pH 8.1). Environmental humidity was maintained during incubation to prevent drying and cracking of the support film. Immunolabeling experiments for each nanoparticle were performed in replicates of five support films to avoid experimental failure caused by film damage or detachment from copper grids during repeated incubation and washing steps.

Detection of the immunoreactivity of nanoparticles harboring linear B-Cell epitopes of ASFV

The binding capacity of nanoparticle proteins harboring ASFV linear B-cell epitopes to antibodies in ASFV positive serum, i.e., their immunoreactivity, was determined. Nanoparticles (DB, 22 C, DD, DE, DF, and AP205) assembled from recombinant proteins were diluted to 1 μg/mL in carbonate-bicarbonate buffer (CBS). Additionally, equal masses of DB, 22 C, DD, DE, and DF were mixed and diluted to a total concentration of 1 μg/mL to prepare a mixed nanoparticle sample (MIXB). A 100 μL aliquot of each diluted sample was added to each well of an ELISA plate for coating, followed by incubation at 37 °C for 2 h. After discarding the coating solution, the plate was washed three times with phosphate-buffered saline containing Tween 20 (PBST). Then, 100 μL of ASFV-positive serum or negative serum—each diluted 1:100 in serum dilution buffer—was added to the coated wells, and the plate was incubated at 37 °C for 30 min. The plate was subsequently washed four times. A 100 μL aliquot of horseradish peroxidase (HRP)-conjugated goat anti-pig IgG (Abcam, Shanghai, China), diluted 1:10,000, was added to each well, followed by incubation at 37 °C for 30 min. After four additional washes, 3,3’,5,5’-tetramethylbenzidine (TMB, SurModics IVD Inc.) substrate solution was added, and the plate was incubated at 37 °C for 10–15 min. The reaction was terminated by adding 50 μL of stop solution to each well, and the optical density (OD) at 450 nm was measured using a microplate reader (Thermo Fisher Scientific, Inc.).

Immunoreactivity was defined as the ratio of the mean OD value for ASFV-positive serum-incubated wells to that for negative serum-incubated wells, with all wells coated with the identical nanoparticle construct. This was calculated using the formula:

Detection of antibody levels

Mice were immunized with nanoparticles DB, 22 C, DD, DE, DF, AP205, and a mixed nanoparticle sample (MIXBT) composed of equal masses of DB, 22 C, DD, DE, DF, and DTA. Serum antibody levels in the immunized mice were subsequently detected. First, the SEC-purified nanoparticles were adjusted to a concentration of 400 μg/mL. Each nanoparticle sample was mixed and emulsified with Freund’s adjuvant at a 1:1 volume ratio; Freund’s complete adjuvant was used for the first immunization, while Freund’s incomplete adjuvant was used thereafter [46]. A 20 μg dose of the emulsified nanoparticles was administered via subcutaneous injection to each 8-week-old BALB/c mouse. Mice immunized with adjuvant mixed with assembly buffer (Fad) served as negative controls. Booster doses were administered on days 14 and 28 after the primary immunization, and serum samples were collected 7 days after the last immunization.

Detection of IgG1 and IgG2a subtype-specific antibody titers

Enzyme-linked immunosorbent assay (ELISA) was used to determine the titers of IgG1 and IgG2a subtype-specific antibodies against nanoparticle proteins in mouse serum, following the protocol described below:

Sera from mice immunized with the corresponding nanoparticles were subjected to twofold serial dilutions for antibody detection: for IgG1, dilutions were 4 × 10³, 8 × 10³, 1.6 × 10⁴, 3.2 × 10⁴, 6.4 × 10⁴, 1.28 × 10⁵, and 2.56 × 10⁵-fold; for IgG2a, dilutions were 200, 400, 800, 1600, 3200, 6400, and 12,800-fold. Negative control sera were uniformly diluted 50-fold. ELISA plates were coated with 100 μL of 1 μg/mL nanoparticle proteins (DB, 22 C, DD, DE, DF, AP205, and MIXBT) per well. Wells without nanoparticles served as coating negative controls. Plates were incubated at 37 °C for 2 h, then washed three times with PBST. Next, 100 μL of the diluted sera (prepared as above) was added to each well, followed by incubation at 37 °C for 30 min and four washes with PBST. Subsequently, 100 μL of HRP-conjugated goat anti-mouse IgG1 (Abcam, Shanghai, China), diluted 1:50,000, or HRP-conjugated goat anti-mouse IgG2a (Abcam, Shanghai, China), diluted 1:10,000, was added to each well. Plates were incubated at 37 °C for 30 min, then washed four times with PBST. TMB substrate solution was added, and plates were incubated at 37 °C for 10 min. The reaction was terminated by adding 50 μL of stop solution per well, and the optical density (OD) at 450 nm was measured using a microplate reader.

The cutoff value was defined as twice the mean optical density (OD) value of the negative control wells. A test well was deemed positive, indicating the presence of subtype-specific antibodies in the serum, when its OD value exceeded this cutoff.

Detection of ASFV linear B-cell epitope-specific antibodies

ASFV linear B-cell epitope-specific antibodies were detected via ELISA for immunogenicity assessment, which reflects the binding capacity of inactivated ASFV particles to antibodies in serum from nanoparticle-immunized mice, as follows:

The inactivated ASFV particles had undergone one freeze–thaw cycle were diluted 100-fold and coated into ELISA plates at 100 μL per well. Plates were incubated at 37 °C for 2 h, then washed three times. Sera from mice immunized with DB, 22 C, DD, DE, DF, MIXBT, AP205, or Fad (adjuvant mixed with assembly buffer) were diluted 100-fold, while sera from naive mice served as negative controls. 100 μL of each diluted serum was added to individual wells, followed by incubation at 37 °C for 30 min. After washing, 100 μL of HRP-conjugated rabbit anti-mouse IgG antibody (Abcam, Shanghai, China), diluted 1:10,000, was added as the secondary antibody. All subsequent steps (washing, color development with TMB, reaction termination, and optical density (OD) measurement at 450 nm) were identical to those described for the detection of antibodies against nanoparticle proteins. Immunogenicity was evaluated using the Pi/Ni ratio, calculated as the ratio of the OD value of the positive serum to that of the negative control serum.

Stimulatory effects of nanoparticles on bone marrow-derived dendritic cells (BMDCs)

Bone marrow cells were isolated from 6-week-old BALB/c mice, seeded into 6-well plates at a density of 1 × 10⁶ cells/well, and cultured in BMDC induction medium (RPMI 1640 supplemented with 10% fetal bovine serum (FBS, Procell system, Wuhan, China)), 1% penicillin–streptomycin, 50 μmol/L β-mercaptoethanol, 20 ng/mL granulocyte–macrophage colony-stimulating factor (GM-CSF), and 10 ng/mL interleukin-4 (IL-4)) to induce differentiation. The cultures were maintained at 37 °C in a humidified atmosphere containing 5% CO₂, with half of the medium replaced every two days. After 7 days of culture, the BMDCs were harvested for subsequent use [47].

In vitro phagocytic and capture capacities of BMDCs for nanoparticles. Cultured BMDCs were incubated with nanoparticles (DB, 22 C, DD, DE, DF, DTA, MIXBT, and AP205) at a final concentration of 10 µg/mL. Cells treated with assembly buffer solution (ABS) served as the negative control. After a 4 h incubation at 37 °C in a 5% CO₂ atmosphere, the cells were harvested, centrifuged at 500 × g for 5 min, and washed with phosphate buffer solution (PBS). The cells were first blocked with an anti-mouse CD16/32 antibody and incubated at 4 °C for 20 min, then fixed and permeabilized using the BD Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences), and subsequently stained with an Alexa Fluor^® 594-conjugated anti-His tag antibody (BioLegend) for 30 min on ice. Following three washes, the percentage of His⁺ BMDCs was determined via a CytoFLEX flow cytometer (Beckman Coulter, Inc.).

Activating effects of nanoparticles on BMDCs. Cultured BMDCs were treated with nanoparticles (DB, 22 C, DD, DE, DF, DTA, MIXBT, and AP205) at a final concentration of 10 µg/mL. Cells treated with assembly buffer solution (ABS) served as the negative control, while cells treated with 0.5 μg/mL lipopolysaccharide (LPS) were used as the positive control. Following a 24 h incubation at 37 °C in a 5% CO₂ atmosphere, the cells were harvested, blocked with an anti-mouse CD16/32 antibody at 5 μg/mL, and incubated at 4 °C for 20 min. After three washes, the cells were stained with fluorochrome-conjugated anti-mouse CD11c-PE, CD80-APC, CD86-PE/Cy7, and IA/IE (MHC class Ⅱ) -FITC antibodies (all from BioLegend) and incubated at 4 °C for 30 min. Following an additional three washes with PBS, the percentages of CD80⁺CD86⁺/CD11c⁺ and CD86⁺MHCⅡ⁺/CD11c⁺ cells were quantified via a CytoFLEX flow cytometer (Fig. S1).

Follicular helper CD4 + T (Tfh) cell detection

8-week-old BALB/c mice were subcutaneously immunized with nanoparticles (DB, 22 C, DD, DE, DF, DTA, MIXBT, and AP205) or Fad (negative control) on days 0, 14, and 28 at a dose of 20 μg per mouse, as previously described in the antibody detection section.

Seven days after the third immunization, the spleens were harvested from the mice, ground, diluted with lymphocyte separation medium (DAKEWE, Shenzhen, China), and filtered through a 70 μm cell strainer. Subsequently, 1–2 mL of RPMI 1640 medium was carefully layered onto the filtrate, followed by centrifugation at 800 × g for 30 min at room temperature to obtain splenic lymphocytes. The lymphocytes were then treated with red blood cell (RBC) lysis buffer (Solarbio, Beijing, China) and washed with PBS to obtain high-purity splenic lymphocytes.

The fluorescent anti-mouse antibodies, including CD45-APC/Cy7, CD3-FITC, CD4- Alexa Fluor^® 700, and CXCR5-BV421 (all from BioLegend), were added to the lymphocytes resuspended in PBS and incubated at 4 °C for 30 min. After treatment with the BD Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences), the cells were stained with anti-mouse Bcl6-perCP/Cyanine5.5 (BioLegend) fluorescent antibody. Following three washes, the percentage of Bcl6⁺CXCR5⁺/CD4⁺ cells was determined using a CytoFLEX flow cytometer (Fig. S2).

Analyzing the activation status of germinal center cells

Mice were immunized as previously described in the antibody detection section. Seven days after the third immunization, lymph nodes were harvested from the mice and fixed in 4% paraformaldehyde. These tissues were subsequently processed for paraffin embedding and immunofluorescence staining using Elab Fluor^® 488-conjugated anti-mouse B220 antibody (Elabscience, Wuhan, China) and Alexa Fluor^® 647-conjugated anti-mouse GL7 antibody (BioLegend). The preparation of immunofluorescence slides and subsequent staining were performed by Servicebio Technology Co., Ltd.

Detection of lymphocyte cytokine levels in mice immunized with nanoparticles

Mice were immunized as previously described in the antibody detection section. Spleens were harvested from BALB/c mice on 7 days after the third booster immunization. Splenic lymphocytes were isolated following grinding, separation, impurity removal, and washing procedures for subsequent use.

Lymphocytes from each group were seeded into 24-well culture plates at a density of 2 × 10⁶ cells per well and maintained in RPMI 1640 medium (Gibco) supplemented with 20% fetal bovine serum (FBS) and 1% penicillin–streptomycin (P/S, Gibco). Nanoparticles (DB, 22 C, DD, DE, DF, DTA, MIXBT, and AP205) were added to lymphocytes isolated from mice of the corresponding immunized groups at a final concentration of 10 μg/ml. For the negative control group (Fad), lymphocytes were stimulated with assembly buffer solution (ABS). All cells were cultured at 37 °C in a 5% CO₂ atmosphere for 24 h. Culture supernatants were then collected, and the concentrations of cytokines (IFN-γ, TNF-α, and IL-2) secreted by the treated cells were quantified using an ELISA kit (Elabscience, Wuhan, China).

Lymphocytes were seeded into 96-well culture plates at a density of 1 × 10⁶ cells per well in a total volume of 200 μL complete medium (RPMI 1640 medium, 20% FBS, 1% P/S). Then these lymphocytes were stimulated with 5 μL of inactivated ASFV and incubated at 37 °C in a 5% CO₂ atmosphere. After 2 h of incubation, monensin (BD Biosciences) was added to the cells at a 1:1000 dilution to block cytokine translocation. Following an additional 8 h of incubation, the cells were harvested, washed three times with PBS, and blocked with an anti-mouse CD16/32 antibody at a concentration of 5 μg/mL for 20 min at 4 °C. Subsequently, anti-mouse CD3-FITC, CD4-Alexa Fluor^® 700, and CD8-APC fluorescent antibodies (all from BioLegend) were added and incubated at 4 °C for 30 min. After washing with PBS, the cells were fixed and permeabilized using BD Cytofix/Cytoperm Fixation/Permeabilization Kit (BD Biosciences), followed by intracellular cytokine staining with anti-mouse IFN-γ-PE, TNF-α-APC/Cyanine7, and IL-2-PE/Cy7 antibodies (all from BioLegend) for 30 min at room temperature. Cells stimulated with concanavalin A (ConA) (Solarbio, Beijing, China) served as the positive control. The proportions of cytokine-secreting CD4⁺ and CD8⁺ cells were determined using a CytoFLEX flow cytometer (Fig. S3).

For IL-4 detection, anti-mouse CD3-FITC (BioLegend) and anti-mouse CD4- Alexa Fluor^® 700 (BioLegend) antibodies were used for surface staining, and anti-mouse IL-4-PE antibody was added for intracellular staining. All other procedures were identical to those described for the other cytokines, and the proportion of IL-4-secreting CD4⁺ cells was determined using a CytoFLEX flow cytometer (Fig. S3).

Detection of memory cells

8-week-old BALB/c mice were subcutaneously immunized with nanoparticle proteins (DB, 22 C, DD, DE, DF, DTA, MIXBT, and AP205) or Fad (negative control) on days 0, 14, and 28, as previously described in the antibody detection section. Splenic lymphocytes were harvested from the mice seven days after the last immunization, as previously described in the Tfh cell detection section. In addition, peripheral blood was collected at 30 days post the third immunization, and peripheral blood lymphocytes (PBLs) were isolated with a mouse peripheral blood lymphocyte isolation kit (Solarbio, Beijing, China).

The proportions of memory B lymphocytes and memory T lymphocytes were determined. For memory B lymphocyte detection, the cells were stained with anti-mouse CD45-APC/Cy7, B220-Brilliant Violet 605, IgG-Alexa Fluor^® 488, and CD27-BV421 fluorescent antibodies (all from BioLegend). For memory T lymphocyte detection, the cells were stained with anti-mouse CD45-APC/Cy7, CD3-FITC, CD4-AF700, CD8-APC, CD44-PE/Cy7, and CD62L-PE fluorescent antibodies (all from BioLegend). All staining reactions were incubated at 4 °C for 30 min, followed by three washes. The following parameters were quantified using a CytoFLEX flow cytometer: Memory B cells (percentage of CD27⁺IgG⁺ cells within the B220⁺ population, Fig. S4); Effector memory T cells (EM T, percentage of CD44⁺CD62L⁻ cells within CD4⁺ or CD8⁺ populations); Central memory T cells (CM T, percentage of CD44⁺CD62L⁺ cells within CD4⁺ or CD8⁺ populations, Fig. S5).

Verification of the inhibitory activity of immune serum on ASFV nucleic acid replication

Isolation of porcine alveolar macrophages (PAMs)

Intact porcine heart with complete envelope and attached lung-trachea complex were isolated, with care taken to prevent separation of the heart and lungs during dissection. The tracheal opening was tightly ligated with sterile cotton rope, and the entire organ was immersed in room-temperature PBS supplemented with 1% amphotericin B and 1% gentamicin (Solarbio, Beijing, China). Following transfer to the cell culture room, the organ was placed in a tray within a biosafety cabinet. The lung surface and peritracheal area were first rinsed and disinfected with 75% ethanol, then washed with the same antibiotic-containing PBS buffer.

The ligature was loosened and the trachea was elevated. A 50 mL syringe or pipette was used to infuse 500–1000 mL of PBS into the lungs (the first batch of PBS was pre-chilled, and subsequent volumes could gradually warm to room temperature). The trachea was sealed again, and the surface of the lungs was gently massaged for 5 min before the lavage fluid containing alveolar macrophages was drained. This lavage procedure was repeated 4–5 times. The collected lavage fluid was aliquoted into 50 mL centrifuge tubes, balanced, and centrifuged at 500 g for 10 min. The supernatant was discarded, and 5 mL of red blood cell lysis buffer (Solarbio, Beijing, China) was added to each tube to resuspend the cells. After incubation on ice for 3–5 min, lysis was terminated by adding 5 volumes of PBS, and the cells were washed twice with PBS, each time centrifuging at 500 g for 10 min.

After centrifugation, the cells were resuspended in complete medium (RPMI 1640 medium, 20% FBS, 1% P/S) for immediate culture or cell freezing medium (89% FBS, 10% DMSO (Sigma-Aldrich), 1% gentamicin triple-antibiotic (Solarbio, Beijing, China) for cryopreservation in liquid nitrogen. Cell counting was performed to adjust the cell density to 1 × 10⁷ cells/mL. An aliquot of the cell suspension was further diluted to 1 × 10⁶ cells/mL for immediate cell culture and adhesion observation. In this study, only primary porcine alveolar macrophages (PAMs) were used for the propagation of ASFV, and the effective storage period of these cryopreserved cells was 3 months.

Inhibition of viral replication by immune sera

8-week-old BALB/c mice were subcutaneously immunized with nanoparticles (DB, 22 C, DD, DE, DF, DTA, MIXBT, and AP205) or Fad (negative control), on days 0, 14, and 28, as previously described in the antibody detection section. Sera were collected from mice on 7 days after the third immunization. First, the sera were filtered through a 0.22 μm syringe filter for sterilization and inactivated in a 56 °C water bath for 30 min. Subsequently, the sera were serially diluted to 1:4, 1:16, 1:64, and 1:256 in RPMI 1640 medium supplemented with 1% P/S. African swine fever virus (ASFV/II/CN/SC/2019) at a multiplicity of infection (MOI) of 0.01 was mixed with the inactivated mouse sera and incubated overnight at 37 °C in a 5% CO₂ atmosphere. Two additional control groups were set up: sera from non-immunized mice and virus-only solution without sera.

The serum-virus mixtures or controls were added to resuscitated confluent porcine alveolar macrophages (PAMs) in 48-well plates at 200 μL per well, followed by incubation at 37 °C for 1 h with gentle shaking of the plates every 10 min. After incubation, the serum-virus mixtures or controls were discarded, and the cells were washed three times with sterile PBS. Then, 200 μL of complete medium (RPMI 1640 medium, 1.5% FBS, 1% P/S) was added to each well, and the plates were incubated at 37 °C in a 5% CO₂ atmosphere for 48 h. The cell cultures were subjected to three freeze–thaw cycles, and viral nucleic acids were extracted using a DNA virus nucleic acid extraction kit (Solarbio, Beijing, China). The cycle threshold (Ct) values of the extracted nucleic acid samples were determined using an ASFV fluorescent PCR detection kit (LSBIO, Lanzhou, China).

The viral replication inhibition rate was calculated to evaluate the ability of the sera to inhibit ASFV nucleic acid replication, using the formula: Viral replication inhibition rate (%) = [(Ct value of post-immunization serum − Ct value of pre-immunization serum)/Ct value of pre-immunization serum] × 100.

Effects of nanoparticles proteins on mouse physiological functions

8-week-old BALB/c mice were subcutaneously immunized with nanoparticle proteins DB, 22 C, DD, DE, DF, DTA, MIXBT, and AP205 on days 0, 14, and 28, respectively. Two negative control groups were set up: mice immunized with adjuvant containing assembly buffer (Fad) and mice treated with assembly buffer solution (ABS) alone. The body weights of the mice were measured weekly, and the changes in body weight among mice immunized with different nanoparticle proteins were recorded. 7 days after the third immunization, the mice were euthanized, and their visceral organs—including the heart, liver, spleen, lung, and kidney—were harvested. These tissues were fixed in 4% paraformaldehyde, processed into tissue sections, and stained with hematoxylin–eosin (HE) dye (provided by Lilai Biotechnology Co., Ltd.). Subsequent histopathological examination was performed to assess tissue morphological changes.

Statistical analysis

The experimental data were statistically analyzed using GraphPad Prism version 9.5, with one-way ANOVA followed by Sidak’s multiple comparisons test as the post-hoc analysis.

Results

Expression and purification of recombinant proteins

Recombinant proteins DB, 22 C, 22D, 22E, 22 F, DD, DE, DF, DTA, and AP205 were successfully expressed in E. coli Rosetta (DE3) cells. The expressed recombinant proteins were purified via His-tag affinity chromatography and verified by SDS-PAGE. The molecular weights of the expressed recombinant proteins were consistent with the theoretically designed values, as follows: DB, 41.35 kDa; 22 C, 27.41 kDa; 22D, 26.85 kDa; DD, 40.86 kDa; 22E, 27.16 kDa; DE, 42.03 kDa; 22 F, 27.79 kDa; DF, 41.83 kDa; DTA, 40.40 kDa; and AP205, 15.45 kDa (Fig. 2).

Fig. 2.

Expression results of recombinant proteins constructed by fusing ASFV multi-epitope sequences with AP205 capsid protein. Recombinant proteins were verified by SDS-PAGE analysis. The molecular weights of these proteins are as follows: DB (41.35 kDa), 22 C (27.41 kDa), 22D (26.85 kDa), DD (40.86 kDa), 22E (27.16 kDa), DE (42.03 kDa), 22 F (27.79 kDa), DF (41.83 kDa), DTA (40.40 kDa), and AP205 (15.45 kDa)

In vitro assembly and characterization of nanoparticles (proteins)

In this study, we individually verified the in vitro assembly of the recombinant proteins DB, 22 C, 22D, 22E, 22 F, DD, DE, DF, DTA, and the scaffold protein AP205. The pH of the assembly buffer was optimized, with the following values: pH 7.9 for DB, 22 C, 22D, 22E, DD, DE, and DTA; pH 7.5 for 22 F; and pH 7.1 for DF and AP205. After dialysis-mediated assembly, DB, 22 C, and DE formed stable, non-aggregated proteins. In contrast, recombinant proteins 22D, 22E, 22 F, DD, DF, DTA, and AP205 exhibited significant protein aggregation. To mitigate this phenomenon, 0.2‰ Triton X-114 was added to the assembly buffer during protein assembly, which notably reduced the aggregation of these proteins.

Following purification of the assembled proteins by size-exclusion chromatography (SEC), transmission electron microscopy (TEM) was used to visualize the assembled products. The results showed that DB and 22 C formed typical VLP structures; 22D, 22E, and 22 F failed to form nanoparticles. In contrast, the recombinant proteins DD, DE, and DF with the dimeric AP205 scaffold, which share the same ASFV epitopes as 22D, 22E, and 22 F with the monomeric AP205 scaffold, respectively, were successfully assembled into nanoparticles. DE formed VLP structures, DD formed nanostructures with sharper edges, and DF formed nanostructures with relatively dispersed edges. Additionally, DTA and the scaffold control AP205 also assembled into nanoparticles. In stability testing experiments, after the nanoparticles were subjected to two repeated freeze–thaw cycles, DF exhibited structural alterations, forming aggregation-prone protein species, while all other assembled nanoparticles maintained their native structures. The study further demonstrated that the recombinant proteins DB, 22 C, and DE could form VLPs nanostructures without the addition of Triton X-114.

Our findings indicate that the incorporation of distinct amino acid sequences impacts the structural stability of the scaffold itself, with AP205-scaffolded nanoparticles displaying ASFV epitopes exhibiting a predominant size distribution of 30 nm. Moreover, recombinant proteins with the dimeric AP205 scaffold exhibited a significantly enhanced propensity to assemble into uniformly structured nanoparticles compared with those based on the monomeric AP205 scaffold (Fig. 3). These results collectively suggest that the assembly process of AP205-based nanoparticles is influenced by multiple factors, including pH, buffer composition, and inserted epitopes, which provides valuable insights for the design of ASFV vaccines.

Fig. 3.

Transmission electron microscopy (TEM) images of recombinant proteins containing ASFV multi-epitopes following in vitro assembly. The magnification is 40 × 10³. Among the constructs, DB, 22 C, and DE formed annular virus-like particles (VLPs) structures, representing the optimal nanoparticle architecture. Specifically: 22D & DD, 22E & DE, and 22 F & DF (the two protein pairs indicated by red arrows) carry the same ASFV epitopes; 22D, 22E, and 22 F harbor a monomeric AP205 scaffold, while DD, DE, and DF contain a dimeric AP205 scaffold. As shown in the images, 22D, 22E, and 22 F failed to form well-defined and homogeneous nanoparticle structures. In contrast, DD, DE, and DF, their counterparts with identical epitopes, exhibited more distinct and uniform nanoparticle architectures. Both DTA and the AP205 scaffold protein were capable of forming nanoparticle structures. DF (freezing and thawing) denotes the recombinant protein DF after two cycles of repeated freeze-thawing, which formed aggregation-prone nanoparticle structures

Structural validation of ASFV linear B-Cell epitope display on nanoparticles

After immunizing mice with GST-tagged fusion epitopes, the antibodies in the resulting mouse sera exhibited significantly higher binding activity to inactivated African swine fever virus (ASFV) compared with the negative control group immunized with PBS buffer (see Table S2.G), indicating that the mouse sera contained specific antibodies targeting the linear ASFV B-cell epitope.

The aforementioned epitope-specific antibodies were co-incubated with epitope-carrying nanoparticle proteins, and binding was verified using colloidal gold-labeled secondary antibodies. The results demonstrated that the specific antibodies against this linear B-cell epitope could bind to the surface of nanoparticles, with particularly prominent binding on virus-like particle (VLP) structures.

Under transmission electron microscopy, colloidal gold particles were clearly observed to be enriched around VLPs. VLPs assembled by DE exhibited the most prominent epitopes display, followed by those assembled by 22B and 22C. For DD and DF assemblies, their solid white nanoparticle morphology made it easy to confuse them with signal regions from antibody binding. Although clear colloidal gold labeling similar to that on VLPs was not observed, small clusters of gold particles were still visible, confirming to some extent that their epitopes were also displayed on the particle surfaces. As a negative control, DB particles were incubated with non-specific antibodies, and no colloidal gold particle aggregation was observed around them (Fig. 4).

Fig. 4.

Validation of epitope display on nanoparticles via colloidal gold labeling. Red arrows indicate TEM images of nanoparticles bound with epitope antibodies and colloidal gold-labeled secondary antibodies under different viewing angles, with gold particles (the black dots) surrounding the recombinant protein particles. Affected by steric hindrance and imaging angle, gold particles are mostly distributed in partial or semi-encircling patterns. Due to the spatial occupation of epitope-specific antibodies (8–10 nm) and gold-conjugated secondary antibodies, the 6 nm gold particles maintain a certain distance from the particle core, yet are mainly concentrated around the nanoparticles or VLPs. In the negative control, nanoparticles do not bind to non-specific antibodies, and no aggregation of gold particles is observed around them

Furthermore, due to steric hindrance and imaging perspective, most colloidal gold particles were distributed around nanoparticles or VLPs in a partially or semi-encircling manner. Given that the size of epitope-specific antibodies is approximately 8–10 nm, while the nanoparticle protein diameter in this study was around 30 nm, a certain distance was observed between the 6 nm gold particles and the main particle structure in some fields of view [48], yet they were still generally distributed in the peripheral region of nanoparticles or VLPs (Fig. 4).

Based on the above observations, it can be confirmed that the ASFV antigenic epitope was successfully displayed on the surface of AP205-based particles.

Immunoreactivity of nanoparticles harboring linear B-cell epitopes of ASFV

We compared the immunoreactivity of the assembled nanoparticles (DB, 22 C, DD, DE, DF), the mixed nanoparticles (MIXB), and the scaffold control (AP205), defined as the capacity of these nanoparticles to bind antibodies present in ASFV-positive serum, and observed significant differences in their performance. Immunoreactivity was quantified using the Pa/Na ratio, defined as the ratio of the optical density (OD) value of nanoparticle binding to positive serum relative to that to negative serum. The Pa/Na ratios ranked in descending order were DE, MIXB, 22 C, DF, DB, DD, and AP205, with corresponding values of 18.6, 14.75, 11.23, 10.99, 7.597, 4.012, and 1.888. DE exhibited the highest Pa/Na ratio, indicating the strongest reactogenicity. Notably, the Pa/Na ratios of 22 C and DF were significantly higher than those of DB and DD (P < 0.0001; Fig. 5A and Table S2.A). Despite incorporating all ASFV linear B-cell epitopes evaluated in this study, MIXB did not achieve the highest Pa/Na ratio.

Fig. 5.

Immunoreactivity and antibody level detection of nanoparticles-induced immune responses. A: Immunoreactivity of nanoparticles, defined as the binding capacity of assembled nanoparticle proteins to antibodies in ASFV-positive serum. Pa/Na represents the ratio of the OD value of nanoparticle proteins bound to positive serum to that bound to negative serum (n = 6). B: Immunogenicity of nanoparticles, referring to the binding capacity of antibodies induced in mice immunized with nanoparticles to inactivated ASFV particles. Pi/Ni denotes the ratio of the OD value of positive serum in mice bound to inactivated virus to that of negative serum in mice bound to inactivated virus (n = 6). C: IgG1 antibody levels induced in mice immunized with nanoparticles. D: IgG2a antibody levels induced in mice immunized with nanoparticles (n = 8). E: Ratio of IgG1 to IgG2a antibodies (n = 8). F: Relative ratio of IgG1/IgG2a, calculated as the ratio of IgG1/IgG2a values of nanoparticles (DB, 22 C, DD, DE, DF, MIXBT) to that of the AP205 scaffold (n = 8). In the figures: MIXB refers to the mixed nanoparticles of DB, 22 C, DD, DE, and DF; MIXBT denotes the mixed nanoparticles of DB, 22 C, DD, DE, DF, and DTA. Fad represents the negative control group where mice were immunized only with Freund’s adjuvant containing buffer. Mouse sera were collected 7 days after the third immunization with nanoparticles. Data in panels A, B, C, D, E, and F are presented as the mean ± Std. Deviation

Antibody levels generated by nanoparticles

Sera were collected from mice 7 days after the third immunization and analyzed for IgG1 and IgG2a antibody titers against the nanoparticles. The IgG1 (log₁₀) titers of nanoparticles DB, 22 C, DD, DE, DF, MIXBT, and AP205 were 4.975, 4.938, 4.919, 5.371, 4.618, 5.051, and 4.863, respectively (Fig. 5C). The corresponding IgG2a (log₁₀) titers were 3.543, 3.524, 3.693, 3.825, 3.524, 3.599, and 3.618, respectively (Fig. 5D). Among both antibody isotypes, the DE group exhibited significantly higher titers than all other nanoparticles (P < 0.01) (Fig. 5C, 5D, Table S2.C, Table S2.D), while DF showed a moderately lower IgG1 titer (Fig. 5C, Table S2.C). The antibody levels of MIXBT (a mixture containing ASFV T-cell epitope-bearing particle DTA) were not significantly different from those of nanoparticles DB, 22 C, and DD (they carry only ASFV B-cell epitopes) (Fig. 5C, 5D, Table S2.C, Table S2.D).

Furthermore, the IgG1/IgG2a ratios among nanoparticles DB, 22 C, DD, DE, DF, MIXBT, and AP205 were not significantly different (Fig. 5E, Table S2.E). However, when comparing the relative IgG1/IgG2a ratios (normalized to the scaffold AP205), DE showed a significantly higher ratio than AP205 (P < 0.0001), followed by MIXBT (P < 0.05) (Fig. 5F, Table S2.F). These results indicate that recombinant protein-assembled nanoparticles primarily induce humoral immune responses, with DE eliciting the most prominent humoral immunity.

Immune sera were assayed for antibodies against the linear ASFV B-cell epitopes (i.e., assessment of immunogenicity). The results demonstrated that all epitope-bearing nanoparticles elicited epitope-specific antibodies, and no significant differences were observed in their respective Pi/Ni ratios, all of which exceeded 3. In contrast, the binding capacity of immune sera from nanoparticle-immunized mice to inactivated ASFV particles was significantly higher than that of sera from control AP205-immunized mice (P < 0.0001) (Fig. 5B, Table S2.B).

BMDC capture and activation by nanoparticles

In this study, flow cytometry was used to evaluate the capacity of BMDCs to capture nanoparticles and the activating effect of these nanoparticles on BMDCs. Specifically, the capture capacity was determined by binding Alexa Fluor^® 594-conjugated anti-His tag antibodies to the His tags on the nanoparticles, allowing quantification of the percentage of BMDCs that had internalized the nanoparticles. The results showed that the percentages of Alexa Fluor^® 594⁺ BMDCs for nanoparticles DB, 22 C, DD, DE, DF, DTA, and AP205 were 87.94%, 90.17%, 46.26%, 13.62%, 16.98%, 42.89%, and 26.77%, respectively. Thus, BMDCs exhibited the strongest capture capacity for DB and 22 C, followed by DD, DTA, and AP205, while DE and DF exhibited the lowest capture efficiency (Fig. 6A, Table S3.A).

Fig. 6.

Capacity of nanoparticles to induce capture and activation of BMDCs. A: Percentage of Alexa Fluor^® 594-positive (AF 594 +) cells after 4 h of co-incubation of nanoparticles with BMDCs, representing the proportion of BMDCs that captured nanoparticle proteins. B: Percentage of BMDCs with surface co-expression of CD80 and CD86 24 h post nanoparticle treatment. C: Percentage of BMDCs with surface co-expression of MHCII and CD86 24 h post nanoparticle treatment. Panel A was used to verify the nanoparticle capture capacity of BMDCs, while Panels B and C were employed to assess the ability of nanoparticles to activate BMDCs. In the figures: ABS denotes the negative control, where only the buffer solution used for nanoparticle assembly was added to BMDCs; LPS (lipopolysaccharide) served as the positive control for BMDC activation. Each group included four independent replicates (n = 4)

The activating effect of nanoparticles on BMDCs was evaluated by quantifying the percentages of CD80⁺CD86⁺ BMDCs and CD86⁺MHCⅡ⁺ BMDCs. The results showed that the percentages of CD80⁺CD86⁺ BMDCs induced by nanoparticles DB, 22 C, DD, DE, DF, DTA, and AP205 were 25.81%, 27.27%, 27.68%, 25.23%, 28.79%, 22.15%, and 24.91%, respectively. No significant differences were found in percentage across nanoparticles, except DTA, which exhibited a relatively lower level (Fig. 6B, Table S3.B). Meanwhile, the percentages of CD86⁺MHCⅡ⁺ BMDCs induced by the aforementioned nanoparticles were 24.86%, 25.65%, 25.46%, 23.76%, 26.79%, 20.65%, and 22.91%, respectively, and no significant differences were found between each experimental group and the positive control group (Fig. 6C, Table S3.C). These results demonstrate that the prepared nanoparticles possess a significant ability to activate BMDCs.

Tfh cells generation

The proportion of (CXCR5⁺Bcl6⁺) Tfh cells was detected by flow cytometry. The results showed that the proportions of CXCR5⁺Bcl6⁺ cells induced by DF and DE were significantly higher than that of 22C. The percentage of MIXBT was the second highest, only after DF and DE, while DTA, which harbors T-cell epitopes, exhibited a significantly lower percentage compared to the nanoparticle groups containing B-cell epitopes (Fig. 8A, Table S4.A).

Fig. 8.

Nanoparticles regulate Tfh cells generation and lymphocyte cytokine secretion. A: Spleens were collected from mice 7 days after three nanoparticle immunizations to detect the generation of Tfh cells. Flow cytometry was performed, with the percentage of Bcl6⁺CXCR5⁺/CD4⁺ cells used as the indicator for evaluating Tfh cells generation. Each group contained 4 mice (n = 4). B, C, D: Lymphocytes from mice 7 days after three immunizations with the corresponding nanoparticles were stimulated with the same nanoparticles. The secretion levels of cytokines (IFN-γ, TNF-α, IL-2) were determined by ELISA (unit: pg/ml), with 4 mice per group (n = 4). E, F: Lymphocytes from mice 7 days after three nanoparticle immunizations were stimulated with inactivated African swine fever virus (ASFV). The secretion of four cytokines (IFN-γ, TNF-α, IL-2, and IL-4) was analyzed by flow cytometry, and the results were quantified as the percentage of T cells secreting the respective cytokines

The activity of germinal centers induced by nanoparticles

We observed the germinal centers (GCs) formed in the lymph nodes of mice immunized with nanoparticles. The activity of GCs was evaluated based on the intensity and distribution of red fluorescence labeling by GL7. Our results revealed distinct differences in GC activity among the nanoparticle groups: The DE group exhibited the strongest and most widely distributed fluorescence signals in lymph nodes, indicating the highest GC activity; the DF group showed the second-highest activity; significant fluorescence signals were observed in both the DB and DD groups; the 22 C and AP205 groups displayed diffuse fluorescence distribution in GCs; fluorescence signals in the DTA group were negligible; the negative control (Fad) group showed extremely weak, nearly undetectable fluorescence signals in GCs (Fig. 7).

Fig. 7.

Germinal center formation in the lymph nodes of mice immunized with different nanoparticles. Immunized nanoparticles included DB, 22 C, DD, DE, DF, DTA, and AP205. Fad was the negative control (mice immunized with Freund’s adjuvant containing nanoparticle assembly buffer). Lymph nodes were collected 7 days post-third immunization, stained with Elab Fluor^® 488-conjugated B220 and Alexa Fluor^® 647-conjugated GL7, and cell nuclei counterstained with DAPI (blue). Lymph node fluorescence images are shown at 47 × (scale bar = 400 μm) and 120 × (scale bar = 200 μm)

Cytokine levels in mouse lymphocytes

Cell stimulation by nanoparticles

We attempted to stimulate splenic lymphocytes from immunized mice with nanoparticles. The results showed that the levels of cytokines IFN-γ and IL-2 secreted by cells stimulated with nanoparticles DE and DF were generally higher than those stimulated by other nanoparticles, with the secretion levels exceeding 600 pg/ml and 150 pg/ml, respectively. In contrast, DF-stimulated cells secreted significantly more TNF-α (306.2 pg/ml) than DE-stimulated cells (153.9 pg/ml). Additionally, the secretion levels of IFN-γ and IL-2 induced by MIXBT were significantly higher than those by AP205, BD, 22 C, and DD. However, there was no significant difference in TNF-α secretion between MIXBT-stimulated cells and those stimulated by BD, 22 C, DD, or AP205. Among all nanoparticle-stimulated groups, DTA induced the lowest levels of IFN-γ and TNF-α, while its IL-2 level was not significantly different from that of BD, 22 C, or DD. Furthermore, the cytokine levels in all aforementioned nanoparticle-stimulated experimental groups were higher than those in the negative control group (Fad) (Fig. 8B, 8C, 8D, Table S4.B, Table S4.C, Table S4.D). These results indicate that nanoparticles can stimulate mice to mount a robust immune response.

Cell stimulation by inactivated ASFV

In this study, splenic lymphocytes from nanoparticle-immunized mice were stimulated with inactivated ASFV as the antigen. The results showed that the proportions of IFN-γ secreted by CD4⁺ and CD8⁺ T cells were higher than those of TNF-α and IL-2, with CD8⁺ T cells exhibiting higher secretion proportions of all three cytokines (IFN-γ, TNF-α, IL-2) compared to CD4⁺ T cells. Meanwhile, CD4⁺ and CD8⁺ T cells in the negative control group showed no difference in cytokine levels, which remained low. Under stimulation with inactivated ASFV, the proportions of IFN-γ and IL-2 secreted by CD4⁺ T cells and CD8⁺ T cells in the DE, DF, DB, and 22 C groups were generally higher than those in the other groups, while the proportion of TNF-α secretion in all experimental groups remained at a low level (Fig. 8E, 8F, and Table S4.E). Notably, the proportion of IL-4 secreted by CD4⁺ T cells was generally higher than that of IFN-γ, with the DE and 22 C groups exhibiting the most prominent responses, reaching 4.08% and 3.64%, respectively. These results indicate that the immunized mice were successfully induced to generate specific cellular and humoral immune responses against the ASFV epitopes carried by the nanoparticles.

Detection of memory cells

Splenic lymphocytes

Seven days after the third immunization, analysis of memory immune cell proportions in splenic lymphocytes revealed that all nanoparticle-immunized mice generated memory B and T cells.

For memory B cells (CD27⁺IgG⁺ cells), all experimental groups exhibited significantly higher proportions (> 2.2%) than the negative control group (Fad), with the DE (3.72%) and DTA (3.94%) groups showing the highest values, which were significantly higher than the AP205 control group (Fig. 9A, Table S5.A).

Fig. 9.

Proportions of memory immune cells in splenic and peripheral blood lymphocytes of immunized mice. A and D: Proportions of memory B cells (CD27⁺IgG⁺ cells). B and E: Proportions of effector memory T (TEM) cells (CD44⁺CD62L⁻ cells). C and F: Proportions of central memory T (TCM) cells (CD44⁺CD62L⁺ cells). The proportions of TEM and TCM cells were analyzed separately for CD4⁺ and CD8⁺ T cell subsets. Panels A, B, and C show the proportions of relevant lymphocytes in the spleen 7 days after the third immunization, while panels D, E, and F represent those in the peripheral blood 30 days after the third immunization

Regarding effector memory T (EM T) cells (CD44⁺CD62L⁻ cells), the DB and 22 C groups displayed comparable and the highest EM T cell proportions (approximately 40%) in both CD4⁺ and CD8⁺ subsets. The DE and DF groups had slightly higher EM T cell proportions in CD8⁺ cells than in CD4⁺ cells, representing the second-highest values. The DD and DTA groups showed ~25% EM T cells in CD4⁺ cells, which were comparable to the AP205 control group. In CD8⁺ cells, the proportions in the DTA (22.993%) and MIXBT (24.045%) groups were slightly lower than those in the DD (27.668%) and AP205 control (29.873%) groups (Fig. 9B, Table S5.B). For central memory T (CM T) cells (CD44⁺CD62L⁺ cells), the proportions of CM T cells in CD4⁺ cells (6–8%) were consistently lower than those in CD8⁺ cells (11–15%) (Fig. 9C, Table S5.C).

Overall, EM T cell proportions were higher than CM T cells in splenic lymphocytes: the proportion of EM T cells in CD4⁺ cells was more than fourfold that of CM T cells, while in CD8⁺ cells, EM T cell proportions were approximately twice those of CM T cells (Fig. 9B, 9C, Table S5.B, Table S5.C).

Peripheral blood lymphocytes (PBLs)

Thirty days following the third immunization, the proportions of memory immune cells in peripheral blood were analyzed. The results revealed that all experimental groups exhibited a proportion of memory B cells (CD27⁺IgG⁺ cells) exceeding 4%, a value significantly higher than that of the negative control group (Fad). Among the experimental groups, the DD (8.295%), DB (7.653%), and DE (7.033%) groups showed notably higher memory B cell proportions than the AP205 control group (4.918%). These were followed in descending order by the DF (6.335%), 22 C (6.083%), DTA (5.948%), and MIXBT (5.24%) groups (Fig. 9D, Table S5.D).

For effector memory T cells (EM T cells), the proportions of CD4⁺ EM T cells in descending order were the DF, DE, DD, 22 C, DB, AP205, MIXBT, and DTA groups, with the first five groups exceeding 26% and the DTA group showing the lowest proportion (19.505%). Overall, the proportion of CD4⁺ EM T cells in all experimental groups was higher than that of CD8⁺ EM T cells, but none exceeded 30% (Fig. 9E, Table S5.E). For central memory T cells (CM T cells), the overall proportion of CD4⁺ CM T cells was slightly lower than that of CD8⁺ CM T cells, and the proportions of CM T cells in both subsets (ranging from 66 to 80%) were much higher than those of EM T cells. In addition, the proportions of both EM T cells and CM T cells in peripheral blood of the experimental groups were significantly higher than those in the negative control group (Fad) (Fig. 9F, Table S5.F).

Comparative analysis

A comparison of memory cells between splenic lymphocytes collected 7 days after the third immunization and peripheral blood lymphocytes collected 30 days post-immunization revealed that the overall proportion of EM T cells in the former was higher than that in the latter (Fig. 9B and E); however, the proportion of CM T cells in peripheral blood was much higher than that in the spleen, with the maximum proportion in the former exceeding 10 times that of the latter (Fig. 9C and F). Therefore, EM T cells predominated in splenic lymphocytes at 7 days post-immunization, while CM T cells were the dominant subset in peripheral blood lymphocytes at 30 days post-immunization, and the proportion of memory B cells in the latter was higher than that in splenic lymphocytes (Fig. 9A and D).

The ability of nanoparticle-immunized sera to inhibit viral replication

High-concentration sera from mice immunized with ASFV epitope-containing nanoparticles exhibited the ability to inhibit ASFV replication in porcine alveolar macrophages (PAMs). At a serum dilution of 1:4, compared with the negative control AP205 and Fad, sera from mice immunized with ASFV epitope-containing nanoparticles significantly inhibited ASFV replication (Table S6.A and B), with the replication inhibition rate reaching 30%. Notably, the replication inhibition rate of nanoparticle DE (36%) was significantly higher than that of MIXBT (23%). At a serum dilution of 1:16, the viral replication inhibition rates of DB, DE, and DF remained above 20%, which were significantly higher than that of MIXBT. When the serum dilution was increased to 1:64 and 1:256, the replication inhibition rates of ASFV epitope-containing nanoparticles continued to decrease; at 1:256, only DF maintained a certain level of viral replication inhibition (Fig. 10A, Table S6.A and B).

Fig. 10.

Validation of immune efficacy in mouse serum. A: Inhibitory effect of sera from nanoparticle-immunized mice on ASFV nucleic acid replication. Immunized sera were co-incubated with ASFV, followed by inoculation into porcine alveolar macrophages (PAMs). The virus-neutralizing activity of the sera was evaluated by detecting the inhibition efficiency of ASFV nucleic acid replication, with cycle threshold (Ct) values of the corresponding nucleic acid samples determined via quantitative real-time PCR (qPCR). The virus replication inhibition rate (%) was used as the primary endpoint, calculated by the formula: (Ct value of post-immunization serum—Ct value of pre-immunization serum) × 100/Ct value of pre-immunization serum. In the figure, * indicates a statistically significant difference between the compared groups (P < 0.05). B: Comparison of immunogenicity between nanoparticles assembled with GST tag or AP205 capsid protein as carriers (both harboring the same epitope). Red arrows denote that the immunogenicity of nanoparticles with AP205 as the carrier was more than twice that of GST-tagged nanoparticles, while black arrows indicate it was less than twice

Impacts of protein-self-assembled nanoparticles on murine physiology

Tissues including heart, liver, spleen, lung, and kidney were collected from immunized mice, followed by hematoxylin–eosin (HE) staining and observation under a light microscope at 400 × magnification.

The results showed that mild vacuolar degeneration was observed in the livers of all mice immunized with nanoparticles, with vacuoles of varying sizes present in the cytoplasm. Other hepatic structures, such as hepatic cords, sinusoids, and vascular structures in the portal triads of hepatic lobules, remained relatively intact without inflammatory cell infiltration or significant pathological changes. Additionally, we compared the liver tissues of mice immunized with Freund’s adjuvant containing buffer (Fad) and assembly buffer solution (ABS) alone. Vacuolar degeneration was observed in the liver tissues of all mice inoculated with Freund’s adjuvant, whereas almost no vacuolation was detected in the hepatocytes of mice immunized with buffer solution alone (Fig. 11A and Fig. S6).

Fig. 11.

Effects of nanoparticles on the heart, liver, spleen, lung, kidney, and body weight of immunized mice. A: Representative images of liver tissues (400 × magnification) from mice in the DB, 22 C, DD, DE, DF, DTA, Fad, and ABS groups. Mild vacuolar degeneration was observed in hepatocytes of all groups except the ABS group. In addition, A shows representative images of heart, spleen, lung, and kidney tissues (400 × magnification) from the DE and BS groups. Fad served as a negative control (mice immunized with Freund’s adjuvant containing nanoparticle assembly buffer), while BS was another negative control (mice immunized with nanoparticle assembly buffer alone). B: Body weight changes of mice during the immunization period. Each group consisted of 4 mice (n = 4)

For other organ tissues (including heart, spleen, lung, and kidney) of mice immunized with nanoparticles, the tissue structure was intact and distinct, with no obvious pathological changes such as hyperplasia, inflammatory infiltration, or exudation. Figure 11A shows HE-stained images of heart, spleen, lung, and kidney tissues from the DE and ABS groups, while tissue images of other experimental groups and control groups are provided in Fig. S6 and Fig. S7.

Body weight monitoring was performed on mice during the immunization period. The results showed that over the 35-day monitoring period, the body weight of mice mostly increased from 18 to 24 g, and there was no significant difference in body weight changes between the experimental groups and control groups (P > 0.05, Fig. 11B, Table S7).

Discussion

The association between AP205 scaffold-mediated nanoparticle display and immune response

Given the extensive number of ASFV epitopes identified in previous studies (> 70, see Table S8), establishing an efficient and robust system for screening dominant epitopes is highly significant. In practice, random combinations of ASFV multi-epitopes are highly prone to forming inclusion bodies. If a Glutathione S-transferase (GST) tag is used, proteins expressed as inclusion bodies not only prove difficult to purify but also demonstrate extremely low refolding efficiency and poor immunogenicity [49, 50]. In contrast, AP205 offers distinct stability advantages in that it retains structural integrity and biological activity upon refolding, even after exposure to harsh conditions such as extreme pH or denaturants [51]. Therefore, utilizing the dimeric AP205 scaffold for epitopes display ensures the efficient assembly of multi-epitope nanoparticles and preserves their immunological activity.

Moreover, the SpyCatcher/SpyTag conjugation system provides an effective nanoparticle assembly strategy for displaying exogenous antigens on the scaffold surface, with the advantage of preserving the scaffold’s native structure [52]. However, the conjugation efficiency between the antigen and scaffold can be influenced by the amino acid sequence of the antigen itself, potentially leading to mismatched or incomplete binding, thereby affecting antigen presentation [53]. In contrast, while genetically fusing ASFV epitopes to the AP205 scaffold surface does alter the scaffold structure to some extent, this study demonstrated that using dimeric AP205 scaffold fused to different ASFV multi-epitope fragments significantly enhances nanoparticle assembly efficiency (Fig. 3). Furthermore, the SpyCatcher/SpyTag strategy involves separate purification of the scaffold and antigen, followed by in vitro conjugation and additional nanoparticle purification steps. In comparison, recombinant proteins produced via genetic fusion can be directly purified and assembled in vitro, significantly streamlining the process. Therefore, given the abundance of epitopes in ASFV, the genetic fusion approach using a dimeric AP205 scaffold demonstrates advantages in both assembly efficiency and procedural simplicity. This method has previously been successfully applied to display the RBD protein of SARS‑CoV‑2 [26]. In addition, this study demonstrates that the AP205 capsid protein harbors dual functions: it not only serves as a scaffold for epitope display but also provides T-cell epitopes, thereby facilitating the induction of robust humoral immune responses by nanoparticles carrying only ASFV linear B-cell epitopes in mice.

In the present study, the DB, 22 C, and DE constructs self-assembled into VLP structures, whereas the DD and DF constructs only assembled into nanoparticulate structures. The correlation between these aforementioned structural characteristics and immunological efficacy was also validated in our study. Previously, we expressed individual epitope-GST recombinant proteins using GST as the carrier, corresponding to the identical epitopes employed herein, and evaluated their immunoreactivity and immunogenicity [43]. In the current experiments, we calculated the ratios of the immunoreactivity and immunogenicity values of AP205-based multi-epitope nanoparticles to the mean values of the corresponding GST-carried epitopes. The results revealed that nanoparticles capable of VLPs formation consistently exhibited higher ratio values (Fig. 10B and Table S9). Specifically, the immunoreactivity values of the VLPs—forming DB, 22 C, and DE constructs were 2.21-, 2.83-, and 4.05-fold compared to values of the corresponding GST-carried epitopes, respectively, whereas the immunoreactivity ratios of DD and DF were only severally 1.82- and 1.44- fold (Fig. 10B, Table S9.B). With respect to immunogenicity, the corresponding ratio values for these nanoparticles DB, 22 C, DD, DE, DF were 1.22, 1.13, 1.50, 1.14, and 0.91, respectively (Table S9.C). Collectively, these findings indicate that the VLP structures formed using the AP205 scaffold can more effectively enhance the immunoreactivity and immunogenicity responses of the displayed epitopes.

When nanoparticle structures are unstable (e.g., DF), the immunoreactivity of their displayed epitopes exhibits substantial variability: the mean immunoreactivity value of the corresponding GST-epitope exceeded 7, which was significantly higher than that of the other epitope groups; however, the immunoreactivity ratio of AP205-epitope nanoparticles to GST-epitope was lower than those of the other nanoparticle groups. Meanwhile, this study revealed that the levels of IgG1 and IgG2a in the DF group were significantly lower than those in other experimental groups (Fig. 5). These results suggest that the structural instability of DF may not only impair immunoreactivity but also inhibit the ability of effective epitopes to induce normal immune responses. This phenomenon and inference are consistent with the conclusions of previous relevant studies [54, 55], further corroborating the critical role of the structural stability of antigen display systems in regulating immune recognition and response activation.

In this study, the proportions of effector memory T (EM T) cells were found to be the highest in the DB and 22 C groups, followed by the DF and DE groups. To a certain extent, these results indicate that compared with conventional nanoparticles, virus-like particles (VLPs) structures are more conducive to promoting the generation of EM T cells, which is consistent with previous research reports [56–59].

Moving forward, efforts will be made to optimize the fusion strategy of ASFV multi-epitope sequences based on the AP205 capsid protein by employing computer-aided design (CAD) technology. The VLP structures (DB, 22 C, and DE) designed and validated in this work provide a meaningful structural basis for amino acid deconstruction of analogous architectures, thereby facilitating the construction of nanoparticles with enhanced stability and greater capacity to accommodate exogenous fragments. Notably, this technical optimization approach has been validated in previous studies on other viruses [60–62].

Analysis of the immunological effect of nanoparticles

Detection results of IgG1 and IgG2a antibody titers, as well as the IgG1/IgG2a ratio, in mice immunized with nanoparticles revealed that the nanoparticles primarily elicited humoral immune responses (Fig. 5). However, cytokine profiling showed that when comparing the secretion levels of IFN-γ, TNF-α, and IL-2, whether stimulated with nanoparticles or inactivated ASFV, IFN-γ exhibited the highest secretion level or proportion (Fig. 8). This indicates that the immune system tends to support the secretion of IgG2a antibodies during humoral immunity and promote cellular immunity [63, 64]. Furthermore, we detected the proportion of IL-4-secreting cells and found that IL-4 secretion was higher than that of IFN-γ (Fig. 8E). This may directly counteract the inhibitory effect of IFN-γ on IgG1 secretion [65, 66]. Additionally, numerous studies have demonstrated that the presence of IL-4 potently promotes B-cell IgG1 production, with an effect far exceeding the inhibitory role of IFN-γ. Moreover, memory cells can secrete IFN-γ to initiate immune protection mechanisms. In this study, the proportion of memory cells in experimental groups was significantly higher than that in the control group, which may be associated with the high IFN-γ secretion level. Consistent with this, previous studies have shown a correlation between IFN-γ secretion and neutralizing antibody potency [67]. Therefore, the high-level IFN-γ secretion observed in this study is a comprehensive outcome of immune activities such as cellular immunity and memory cell formation, while the presence of IL-4 in the organism maintains robust humoral immune function.

In assessing the ability of nanoparticles to induce the uptake and activation of BMDCs, we found that BMDCs exhibited the strongest uptake capacity for nanoparticles DB and 22 C, whereas their uptake of DE and DF was the weakest. However, this reduced uptake did not impair the ability of DE and DF to activate BMDCs, with DF showing a slight superiority (Fig. 6). Notably, the uptake capacity of BMDCs for nanoparticles was lower than that of macrophages, and their internalization of certain particles was dependent on caveolin-mediated endocytosis [68]. Although DE and DF were taken up at lower levels, the antigens carried by these two nanoparticles may have efficiently bound to MHC molecules on the BMDC surface to form complexes, thereby inducing the upregulation of relevant marker molecules and compensating for their weak uptake capacity [69, 70]. These findings indicate that the epitope molecules present on DE and DF can efficiently activate BMDCs.

In this study, we found that the level of memory B cells in the spleen during the nanoparticle immunization phase was lower than that in the peripheral blood 30 days post-immunization (Fig. 9). It is speculated that within a few days after immunization, the core function of the spleen is to present nanoparticle antigens via antigen-presenting cells (APCs), activating B cells to secrete CXCL13. This in turn recruits CXCR5-expressing Tfh cells and forms germinal centers (GCs) in the spleen. During this stage, B cells mainly differentiate into short-lived antibody-secreting cells, with only a small number of memory B cells generated [71], which is consistent with the activation status of lymph node GCs (GL7-labeled, Fig. 7) observed during the immunization phase. By 30 days post-immunization, the organism enters a non-acute immune stage, where GC reactions subside. Mature memory B cells migrate into the bloodstream under the influence of chemokines, leading to the sustained enrichment and long-term persistence of memory B cells in the peripheral blood.

In addition, the proportion of effector memory T (EM T) cells in splenic lymphocytes was higher than that in the peripheral blood during the immunization phase. In contrast, 30 days post-immunization, the proportion of central memory T (CM T) cells in the peripheral blood was significantly higher than that of EM T cells, reaching over 70% in most cases. These findings indicate that in the early stage of nanoparticle immunization, EM T cells dominate the rapid clearance of invading antigens. As the post-immunization period extends, CM T cells become the core functional cells, and the organism shifts toward long-term antigen memory and proliferative reserve. Through the proliferation and differentiation of CM T cells to replenish effector cells, the sustained intensity of the immune response is maintained [72].

Due to the distinct ASFV linear epitopes carried by each nanoparticle in this study, differences were observed in their performance across various stages of the immune response. These results enhance our understanding of the immunological characteristics of different epitopes and provide a foundation for further investigating epitope-related immune mechanisms. Among the nanoparticles tested, DE, which carries linear B-cell epitopes pB438L 51–73, p72 244–262, and p54 132–180, exhibited the optimal overall immunological efficacy. This was reflected in key immune processes including immunogenicity, immunoreactivity, specific antibody titers, BMDC activation, Tfh cells and germinal center (GC) activation, T-cell cytokine secretion, and inhibition of ASFV viral nucleic acid replication by immune sera. Nanoparticle DF, harboring the epitopes p17 5–34, 71–113, and p30 4–15, demonstrated notably lower immunoreactivity and specific antibody levels. A key advantage of DF lies in its superior immunological performance compared to DE across multiple dimensions—most prominently, the immune sera induced by DF exhibited significantly stronger inhibition of ASFV viral nucleic acid replication, underscoring its distinctive antiviral efficacy. Following DE and DF, nanoparticles DB and 22 C displayed moderate efficacy. DB carries ASFV linear B-cell epitopes p17 68–86 and pE248R 138–168, 158–185, while 22 C carries pE120R 65–117, pE199L 175–189), and p12 (pO61R) 36–56. In contrast, DTA, carrying ASFV T-cell linear epitopes pp220 (CP2475L, 124–132, 393–401, 788–796, 2125–2133, 1375–1383) and pp62 (CP530R, 227–235, 263–271), exhibited the weakest overall immunological efficacy across the tested parameters, though it showed a modest advantage in promoting memory B cell responses.

MIXBT (a mixture of different nanoparticles) did not demonstrate the best immunological performance, despite having the same total dose as individual nanoparticles. This may be attributed to the fact that each nanoparticle component in MIXBT was administered at a dose 1/6 that of individual nanoparticle immunizations, which likely prevented the induction of effective immune responses against each antigen [73, 74]. Therefore, future studies should focus on expanding the epitope-carrying capacity of VLP structures and optimizing immunization dosages.

Furthermore, mild vacuolar degeneration of hepatocytes in mice immunized with nanoparticles combined with Freund’s adjuvant has been reported previously. This phenomenon is reversible, with degeneration typically resolving following the subsidence of inflammation [75, 76].

Limitations and application scope

Currently, the identified high-efficiency ASFV T-cell epitopes remain underrepresented in existing datasets, and additional optimal B-cell epitope combinations are yet to be identified, hindering the development of ASFV multi-epitope nanoparticle vaccines. Meanwhile, some nanoparticles assembled with the AP205 scaffold in this study exhibit a tendency to aggregate. Further screening of suitable dispersants and optimization of the buffer system are therefore required to improve their stability. Moreover, this approach only supports the screening and identification of linear epitopes, but not conformational epitopes. Despite the above limitations, this study confirmed that inclusion bodies of recombinant proteins fused with the dimeric AP205 scaffold, when expressed in a prokaryotic system, can enhance the in vitro assembly efficiency of nanoparticles. Upon immunization in mice, these nanoparticles can effectively induce specific immune responses. This strategy is particularly suitable for the screening and identification of random combinations of multiple ASFV epitopes.

Conclusion

The ASFV multi-epitope nanoparticles generated in this study induced heterogeneous immune responses in mice. Our findings demonstrate that nanoparticles forming VLP structures possess significant advantages during immune processes; however, the intrinsic immunogenicity of the displayed epitopes plays a decisive role in the final efficacy. The highly effective multi-epitope combinations identified herein provide data support for ASFV vaccine development. Furthermore, the structural data of the VLPs designed and validated in this study can support future amino acid deconstruction and analysis of analogous structures. This helps optimize the assembly efficiency of recombinant proteins and further expand the multi-epitope loading capacity of VLPs.

Author contributions

HC and JS performed the conceptualization and funding acquisition. HL conducted the research proposal design, methodology development, animal experiments, data analysis, and manuscript writing. JS, WL, SG, HG, and HD performed the suggestion of the research, tutorship of writing. TZ, JY performed animal experiments. All authors contributed to the article and approved the submitted version.

Funding

This work was supported by the National Natural Science Foundation of China (grant numbers 32273038).

Data availability

All key data related to this study have been made publicly available, including the epitope sequences of ASFV, amino acid sequences of recombinant proteins, conditions for the preparation and assembly of protein nanoparticles, as well as data on the evaluation of immune efficacy in mice immunized with the nanoparticles; part of these data are included in the supplementary materials. Further data will be made available on request.

Declarations

Ethics approval and consent to participate

The animal study was reviewed and approved by Animal Ethics Committee of LVRI, CAAS (experimental animal ethics batch number: LVRIAEC-2025–041 and LVRIAEC-2025–146).

Consent for publication

All authors agree for publication.

Competing interests

The authors declare no competing interests.