Molecular Signatures in Swine Innate and Adaptive Immune Responses to African Swine Fever Virus Antigens p30/p54/CD2v Expressed Using a Highly Efficient Semliki Forest Virus Replicon System

African swine fever virus (ASFV) causes a devastating viral hemorrhagic disease in domestic pigs and Eurasian wild boars, posing a foremost threat to the swine industry and pig farming. The development of an effective vaccine is urgently needed, but has been hampered by the lack of an in-depth, mechanistic understanding of the host immune response to ASFV infection and the induction of protective immunity. In this study, we report that immunization of pigs with Semliki Forest Virus (SFV) replicon-based vaccine candidates expressing ASFV p30, p54, and CD2v, as well as their ubiquitin-fused derivatives, elicits T cell differentiation and expansion, promoting specific T cell and humoral immunity. Due to significant variations in the individual non-inbred pigs in response to the vaccination, a personalized analysis was conducted. Using integrated analysis of differentially expressed genes (DEGs), Venn, KEGG and WGCNA, Toll-like receptor, C-type lectin receptor, IL17 receptor, NOD-like receptor and nucleic acid sensor-mediated signaling pathways were demonstrated to be positively correlated to the antigen-stimulated antibody production and inversely correlated to the IFN-γ secreting cell counts in peripheral blood mononuclear cells (PBMCs). An up-regulation of CIQA, CIQB, CIQC, C4BPA, SOSC3, S100A8 and S100A9, and down-regulation of CTLA4, CXCL2, CXCL8, FOS, RGS1, EGR1 and SNAI1 are general in the innate immune response post-the second boost. This study reveals that pattern recognition receptors TLR4, DHX58/DDX58 and ZBP1, and chemokines CXCL2, CXCL8 and CXCL10 may play important roles in regulating this vaccination-stimulated adaptive immune response.


Introduction
African swine fever (ASF), caused by ASF virus (ASFV), is a devastating viral hemorrhagic disease in both domestic pigs and Eurasian wild boars. ASFV infects cells of the mononuclear phagocytic system, including differentiated macrophages and specific lineages of reticular system cells [1][2][3]. The highly virulent strains induce extensive damage in affected tissues with symptoms including hemorrhage, edema, ascites, and shock [4,5]. The development of an effective vaccine is urgently required for the prevention and control of this disease [6,7].
Different formats of ASFV vaccine candidates, including inactivated, live-attenuated (LAV), and subunit vaccines, have been attempted, but no effective and safe vaccine has the induction of adaptive immunity by vaccine candidates would be of great help in the rational design and development of an effective ASFV vaccine.
Semliki Forest virus (SFV)-based replicon is derived from SFV, a member in the Alphavirus genus with a positive-stranded RNA genome of approximately 12 kb. The genome contains two main open reading frames (ORF), encoding a viral replicase and structural proteins, respectively. Replacement of the ORF coding for structural proteins by foreign genes creates an SFV-based vector system with a broad host range and high expression efficiency through self-driven replication. As this replication is self-limited due to the lack of viral structural proteins during the replicon package, it is an efficient and safe platform for foreign gene expression and vaccine development. In this study, we construct the SFV replicon-based DNA vaccine candidates expressing ASFV p30, p54, and CD2v proteins with or without fusion to ubiquitin and immunize small groups of non-inbred pigs in the form of DNA or infectious replicon particles (IRP). Comparative studies and personalized analysis on the induction of humoral and cellular immune responses and bioinformatics analysis of transcriptomic data identified genes and signaling pathways relevant to this vaccination-stimulated innate immunity and adaptive immune responses in pigs.

Antigen Expression, Immunization Scheme, and Assessment
The N-terminal Flag-tagged p30, p54, and CD2v (or CD2vED) proteins with or without fusion to ubiquitin were expressed using SFV replicon-based plasmids, pSFV-flagp54, pSFV-flagCD2v, pSFV-flagp30, pSFV-flagUbiCD2vEDp30 and pSFV-flagUbiCD2vEDp54 ( Figure S1). The expression efficiency of these proteins was checked by Western blot, demonstrating their efficient expression with expected molecular masses and reactivities with anti-ASFV sera (Figure 1a) as well as with an anti-Flag antibody. The ubiquitin-fused CD2vEDp30 and CD2vEDp54 were detected as multi-sized bands, indicating their modification by ubiquitination and ubiquitin-mediated protein cleavage, respectively (Figure 1a).
Co-transfection of BHK21 cells with pSFV-flagP54, pSFV-flagCD2v, pSFV-flagP30, pSFV-flagUbiCD2vEDp30, and pSFV-flagUbiCD2vEDp54, respectively, together with pCMV-SFV-helper 1 plasmid that expresses SFV structural proteins for a package of the recombinant SFV genomes, was then carried out to generate IRPs. The expression of target proteins in cells infected with IRPs was confirmed by immunofluorescent microscopy, showing high expression of p30, p54, and CD2v and moderate expression of UbiCD2vEDp30 and UbiCD2vEDp54, respectively ( Figure 1b).
The replicon-based plasmids and IRPs were then used to immunize pigs. As described in the Material and Method section, group1 pigs (#1, #2, #5, and #6) were injected with plasmids or IRPs which express the target proteins without fusion to ubiquitin, while group 2 (#7, #8, #11 and #12) were injected with plasmids or IRPs which express the ubiquitinfused proteins. All experimental pigs were immunized with two boosts after priming at a two-week interval, and whole blood samples were collected (Figure 1c). On day 28 (second boost) and day 29 (24 h post-second boost), an extra whole blood sample was collected from each pig for whole genome mRNA sequencing. In addition, blood samples were collected from group 2 pigs on day 7 to replace samples on day 0, as the original samples collected on day 0 were accidentally mishandled and spoilt during the process.

Specific Cellular Immune Response
To define and compare the T cell expansion stimulated by these antigen combinations, CD4+, CD8+, and CD4+/CD8+DP T cell ratios in PBMC samples, collected on day 0 (day 7 for group 2), 28, and 35, respectively, after priming, were measured by flow cytometry using specific antibodies. Results for individual pigs in Group 1 and Group 2 were presented in Figure 2a,b, respectively. In Group 1, each subtype T cell ratio was obviously increased on day 28, compared to that on day 0, but was largely unchanged after the second boost (day 28 compared with day 35), except in the #2 pig which showed much higher CD8+ and CD4+/CD8+DP T counts and much lower CD4+T count on day 28 and 35. It indicates that immunization of the replicon-based DNA (#1 and #2 pigs) or IRP (#5 and #6 pigs) induced prominent T cell expansion shortly after immunization and remained relatively high CD8+ and CD4+/CD8+DP T cell ratios (Figure 2a). In group 2, the total T cell numbers between day 7 and day 28 were not obviously different and a general reduction of T cell counts on day 35 was also observed, except in the #7 pig which showed a very low total T cell count at day 7 and did not reach the peak before day 35 (Figure 2b). Comparing the two groups, the average numbers of total T cells in group 2 were, in general, higher than those in group 1 on both day 28 (67.9% to 65.7%) and 35 (61.3% to 52.3%), and higher CD4+ counts (40.7% to 34.5%) and higher CD4+/CD8+ DP T cell ratios (10.1% to 7.2%) remained at day 35 (Figure 2a,b and Table 1). However, due to large variations of

Specific Cellular Immune Response
To define and compare the T cell expansion stimulated by these antigen combinations, CD4+, CD8+, and CD4+/CD8+DP T cell ratios in PBMC samples, collected on day 0 (day 7 for group 2), 28, and 35, respectively, after priming, were measured by flow cytometry using specific antibodies. Results for individual pigs in Group 1 and Group 2 were presented in Figure 2a,b, respectively. In Group 1, each subtype T cell ratio was obviously increased on day 28, compared to that on day 0, but was largely unchanged after the second boost (day 28 compared with day 35), except in the #2 pig which showed much higher CD8+ and CD4+/CD8+DP T counts and much lower CD4+T count on day 28 and 35. It indicates that immunization of the replicon-based DNA (#1 and #2 pigs) or IRP (#5 and #6 pigs) induced prominent T cell expansion shortly after immunization and remained relatively high CD8+ and CD4+/CD8+DP T cell ratios (Figure 2a). In group 2, the total T cell numbers between day 7 and day 28 were not obviously different and a general reduction of T cell counts on day 35 was also observed, except in the #7 pig which showed a very low total T cell count at day 7 and did not reach the peak before day 35 (Figure 2b). Comparing the two groups, the average numbers of total T cells in group 2 were, in general, higher than those in group 1 on both day 28 (67.9% to 65.7%) and 35 (61.3% to 52.3%), and higher CD4+ counts (40.7% to 34.5%) and higher CD4+/CD8+ DP T cell ratios (10.1% to 7.2%) remained at day 35 (Figure 2a,b and Table 1). However, due to large variations of the counts and . PBMC cells collected on day 7, 25 and 35 post-priming from group 2 pigs were analyzed and presented as in (a). (c). IFNγ-secretory spots (IFN-γ-SC) (mean±SE) from PBMCs collected on day 28 post-priming and stimulated by p30 (orange), p54 (green) and CD2v (purple), respectively, were determined by ELISpot assay. IFNγ-secretory cell counts were obtained after deducting the basal count (N) from the coordinating sample without stimulation, and control data were obtained by incubation with phytohem agglutinin (yellow). (d). IFNγ-secretory spots (mean ± SE) from PBMCs collected on day 35 post-priming and stimulated by p30 (orange), p54 (green) and CD2v (purple), respectively, were determined and presented as in (C).

Specific Humoral Immune Response
The anti-p30 specific antibody levels in serum samples were measured using an ASFV Sandwich ELISA kit, revealing a high level of anti-p30 response in #1 pig just 14 days post-priming ( Figure 3a). The antibody level in this pig reached the peak in 2 weeks post-second boost and decreased slowly but remained at a relatively high level throughout the duration of the experiment (Figure 3a). A relatively high antibody induction was detected in the #11 pig from day 14 and steadily increased during the whole observation period, although the level was much lower than that in the #1 pig ( Figure 3a). Other antibody-positive pigs included #2, #7, and #8, but the titers were just above the negative control ( Figure 3a). It was uncertain if these levels of reaction would provide significant protection against ASFV. Meanwhile, determination of the anti-p54 response by using a competitive ELISA assay showed a 65-96% inhibition in #1 pig from day14 to 70, but other pigs showed less than 50% inhibition (Figure 3b), suggesting that the anti-p54 response . IFNγ-secretory spots (IFN-γ-SC) (mean ± SE) from PBMCs collected on day 28 post-priming and stimulated by p30 (orange), p54 (green) and CD2v (purple), respectively, were determined by ELISpot assay. IFNγ-secretory cell counts were obtained after deducting the basal count (N) from the coordinating sample without stimulation, and control data were obtained by incubation with phytohem agglutinin (yellow). (d). IFNγ-secretory spots (mean ± SE) from PBMCs collected on day 35 post-priming and stimulated by p30 (orange), p54 (green) and CD2v (purple), respectively, were determined and presented as in (c). ELISpot analysis, an enzyme-linked ImmunoSpot assay, was then used to measure the IFN-γ-secreting cells which is important in regulation of the innate and adaptive immunity by antigen stimulation. IFN-γ responses in vitro revealed a correlation with the degree of cross-protection against heterologous ASFV challenge [30]. PBMCs were stimulated by co-incubation with purified p30, p54, and CD2v proteins, respectively, and IFN-γ secreting cells were scored with an ELISpot reader, showing various levels of IFN-γ secreting cells after two boosts. A moderate induction (2-87 IFN-γ secreting cells per 1 million PBMCs) was elicited in #1, #2, #5, #8, and #11 pigs on day 28 with cognate ligands (Figure 2c), and a more diverse level of the induction (2-200 IFN-γ secreting cells per 1 million PBMCs) was detected in all pigs on day 35 (Figure 2d). The fact that IFN-γ secreting cells were detected in #7 and #12 pigs only on day 35 suggests the requirement for a second boost in some recipients (Figure 2c,d).

Specific Humoral Immune Response
The anti-p30 specific antibody levels in serum samples were measured using an ASFV Sandwich ELISA kit, revealing a high level of anti-p30 response in #1 pig just 14 days postpriming ( Figure 3a). The antibody level in this pig reached the peak in 2 weeks post-second boost and decreased slowly but remained at a relatively high level throughout the duration of the experiment (Figure 3a). A relatively high antibody induction was detected in the #11 pig from day 14 and steadily increased during the whole observation period, although the level was much lower than that in the #1 pig ( Figure 3a). Other antibody-positive pigs included #2, #7, and #8, but the titers were just above the negative control ( Figure 3a). It was uncertain if these levels of reaction would provide significant protection against ASFV. Meanwhile, determination of the anti-p54 response by using a competitive ELISA assay showed a 65-96% inhibition in #1 pig from day14 to 70, but other pigs showed less than 50% inhibition (Figure 3b), suggesting that the anti-p54 response was efficiently induced in pig #1 only. Taken together, these results demonstrate that specific humoral response was elicited in 5 out of the 8 immunized pigs, and the earliest date for positive anti-p30 detection was at day 14 post-priming. was efficiently induced in pig #1 only. Taken together, these results demonstrate that specific humoral response was elicited in 5 out of the 8 immunized pigs, and the earliest date for positive anti-p30 detection was at day 14 post-priming. . Anti-p54 antibody (mean ± SE) levels in sera collected from immunized pigs on day 0, 14, 28, 42, 49, 63 and 70, respectively, were determined with a competitive ELISA kit. An inhibition rate of 50% or above is considered positive according to the manufacturer's instruction.

Transcriptomic Analysis of DEGs Associated with the Immune Response to the Vaccination
To gain insights into the molecular mechanisms underlying the host immune responses to the vaccination, we analyzed the transcriptomic profiles of whole blood cells from each vaccinated pig. As transient innate responses may reach the peak at 24 h postimmunization and the focus of this study was to characterize the innate immune response from memory cells, we collected the samples immediately before (on day 28) and 24 h post-the second boost (on day 29), respectively, for whole transcriptome sequencing. To Anti-p30 antibody levels (mean ± SE) in sera collected from immunized pigs on day 0, 14, 28, 42, 49, 63 and 70, respectively, were detected with ELISA kit. An OD value of 0.15 or above is considered positive according to the manufacturer's instruction. (b). Anti-p54 antibody (mean ± SE) levels in sera collected from immunized pigs on day 0, 14, 28, 42, 49, 63 and 70, respectively, were determined with a competitive ELISA kit. An inhibition rate of 50% or above is considered positive according to the manufacturer's instruction.

Transcriptomic Analysis of DEGs Associated with the Immune Response to the Vaccination
To gain insights into the molecular mechanisms underlying the host immune responses to the vaccination, we analyzed the transcriptomic profiles of whole blood cells from each vaccinated pig. As transient innate responses may reach the peak at 24 h post-immunization and the focus of this study was to characterize the innate immune response from memory cells, we collected the samples immediately before (on day 28) and 24 h post-the second boost (on day 29), respectively, for whole transcriptome sequencing. To confirm the assay reliability, the principal component analysis (PCA) was performed with the data before further biological analysis, showing good technical reproducibility for triplicated samples but poor biological reproducibility due to individual divergence ( Figure S2). Differentially expressed genes (DEGs) from day 28 versus day 29 were then identified with statistical verification, revealing a response of genes with great variations in both quantity and characteristics in individual pigs. A total of 274 to 1216 DEGs with different ratios of up-(red) and down-regulated (green) expression was found in different pigs (Figure 4a). CTLA4, and EGR1, were consistently detected ( Figure S3). These genes also showed relatively high fold-change and their differential expression appeared not to be affected by the vaccine dosage and form. The genes are therefore the most responsive genes to the boost.   To focus on the immune system, genes not relevant to the immune system were filtered out, and analysis of these immune system-related DEGs (Table S1) by Venn diagram assay identified 10 common DEGs in all samples ( Figure 4b). Interestingly, 9 out of the 10 common DEGs were down-regulated, and one (TGM3) varied in different pigs with the range of FC values indicated in Table 2. These common DEGs include lymphocyte regulator CTLA4, CXC chemokine family members, transcription regulators involved in cell growth, differentiation, or death, and undetermined genes. Furthermore, analysis of the top 20 DEGs from each pair of experimental pigs revealed their diverse distribution, but the five common down-regulated genes with known functions, CXCL2, CXCL8, FOS, CTLA4, and EGR1, were consistently detected ( Figure S3). These genes also showed relatively high fold-change and their differential expression appeared not to be affected by the vaccine dosage and form. The genes are therefore the most responsive genes to the boost.

Hub Genes and Molecular Signatures Correlated to the Adaptive Immune Response
Weighted correlation network analysis (WGCNA) has been increasingly used in bioinformatics applications, for finding clusters (modules) of highly correlated genes, for summarizing such clusters using the module eigengene or an intramodular hub gene, for relating modules to one another and to external sample traits (using eigengene network methodology), and for calculating module membership measures [31]. Using this method, we identified the genes associated with the vaccination-stimulated adaptive responses with the transcriptomic data. Taking the anti-p30 antibody OD value measured on day 42 (14 days post-second boost) as a specific trait, we identified a module with an average correlation value of 0.491 and a p-value of 0.0148, representing a hub module with a high possibility of positive correlation to the trait (Figure 5a). Genes in this module were then used to identify gene clusters in individual animals through Venn analysis, and further KEGG analysis revealed the most significant gene clusters (p adjust value < 0.05) were enriched in the IL17 signaling pathway, C-type lectin receptor (CLR) signaling pathway, NOD-like receptor signaling pathway, Toll-like receptor (TLR) signaling pathway, cytosolic DNA-sensing pathway, T cell receptor signaling pathway, Th17 cell differentiation, RIG-I-like receptor signaling pathway, chemokine signaling pathway, and Th1 and Th2 cell differentiation (Tables 3 and S2).
To identify the hub gene and molecular signatures correlated to the cellular immune response, we input the IFN-γ-ELISpot results as the trait and discovered three modules with relatively high correlation coefficients from the total 9 modules. These include a hub module with the highest correlation value [−0.546, p-value = 0.00578] corresponding to CD2vstimulated IFN-γ secretion, a module with a moderately high correlation value (−0.412, p value = 0.0454) corresponding to p30-stimulated IFN-γ secretion, and a module with a correlation value (−0.263, p value = 0.214) corresponding to p54-stimulation ( Figure 5b). All these correlation coefficients in the selected modules were negative, indicating that they are negative regulators correlating to the trait. The first two modules were in the MEturquoise gene-set, the same gene-set correlated to the p30-stimulated antibody production, indicating that genes in these modules may be involved in positively regulating antibody response, but negatively regulating antigen-stimulated IFN-γ secretion. Corresponding to the p54-stimulated IFN-γ secretion, the module in the MEblue gene set showed a slightly higher correlation value (−0.263) than that in MEturquoise (−0.181) (Figure 5b). However, both correlation coefficients were low and the p-value were >0.05, raising uncertainty about their biological significance. Although IFN-γ-ELISpot is often used to measure T cell response, it cannot reflect T cell response accurately as NK cells are also a vital source of IFN-γ in PBMCs. The IFN-γ-ELISpot readout implicated a cellular response, involving both T and NK cells, and the use of CD3+ T cells to measure the T cell response may provide a more specific result.
informatics applications, for finding clusters (modules) of highly correlated genes, for summarizing such clusters using the module eigengene or an intramodular hub gene, for relating modules to one another and to external sample traits (using eigengene network methodology), and for calculating module membership measures [31]. Using this method, we identified the genes associated with the vaccination-stimulated adaptive responses with the transcriptomic data. Taking the anti-p30 antibody OD value measured on day 42 (14 days post-second boost) as a specific trait, we identified a module with an average correlation value of 0.491 and a p-value of 0.0148, representing a hub module with a high possibility of positive correlation to the trait (Figure 5a). Genes in this module were then used to identify gene clusters in individual animals through Venn analysis, and further KEGG analysis revealed the most significant gene clusters (p adjust value < 0.05) were enriched in the IL17 signaling pathway, C-type lectin receptor (CLR) signaling pathway, NOD-like receptor signaling pathway, Toll-like receptor (TLR) signaling pathway, cytosolic DNA-sensing pathway, T cell receptor signaling pathway, Th17 cell differentiation, RIG-I-like receptor signaling pathway, chemokine signaling pathway, and Th1 and Th2 cell differentiation (Tables 3 and S2).

Personalized Analysis to Denote the Genes Involved in Specific Regulation of the Adaptive Immune Response
As the most efficient antibody response was elicited in #1 and #11 pigs, DEGs with p adjust < 0.05 in the trait-correlated genes groups (Table S2) from these two pigs were further analyzed, showing a group of genes, including TLR4, DDX58, DHX58, ZBP1, PTGS2, NFKBIA, TNF, BCL3, IL18, and IL1RAP were up-regulated in #11 pig, but downregulated in #1 pig ( Figure S4, Table 4). Furthermore, no differential transcription of TLR4, DDX58, DHX58, and ZBP1 was observed in #2 and #12 pigs, the two pigs received the same immunization as #1 and #11 pigs, respectively, but did not elicit an efficient antibody response (Table S2). This expression pattern highlighted that TLR4, DDX58, DHX58, and ZBP1 signaling pathways may play an important role in regulating this vaccine-induced antibody production. Up-regulation of these genes in the pig with a lower antibody level and down-regulation in the pig with a higher antibody level indicates a potential feedback regulatory loop formed around these genes. Genes important for regulating the cellular response were identified from the traitcorrelated genes in #5, #7, and #8 pigs as they elicited more prominent IFN-γ secreting, mainly including a group of up-regulated genes (S100A8, C4BPA, PLAUR, PLAU, A2M, IL18, PIK3CB, CCR1) and a group of down-regulated genes (CXCL2, FOS, IL17B, IFN-ALPHA-15, CTLA4) ( Figure S5, Table 4). In addition, as only the #7 pig showed T cell expansion in response to the last boost, unique DEGs in this pig were possibly correlated to this trait. A total of 14 DEGs in the immune system cluster was identified, with about half of them relevant to the CD8+ T cell expansion and/or differentiation by the functional annotation query (marked by "a" superscript in Table 5). All DEGs in this group were up-regulated, except for one uncharacterized gene. It is subunit of NADPH oxidase (NOX), which plays an essential role in the immune system. NCF1 in dendritic cells promotes autoreactive CD8 + T cell activation [33] 03578 FGR FGR is a myeloid Src-family kinase and participates in regulation of acute myeloid leukaemia. [34] 36317 a VASP contributes to naïve CD8 + T cell activation and expansion by promoting T cell-APC interactions in vivo. [35] 02919 TYROBP a key regulator in immune systems by acting as a signaling adaptor in dendritic cells, osteoclasts, macrophages, and microglia. [36] 29521 CD82 a member of transmembrane 4 superfamily and an activation Ag of T-cells. CD82 associates with CD4 or CD8 and delivers costimulatory signals for the TCR/CD3 pathway.
[37] 17920 a CXCL16 CXCL16 is produced by dendritic cells in T cell of lymphoid organs and induced by IFN-7 and TNF-a.This transmembrane chemokine is implicated in activated CD8+ T cell trafficking. [38] 35226 a NA is collected in gene set for CD8 T cell (https://report.majorbio.com/) 37426 a NA is highly enriched genes in myeloid, NK, CD21pB, SWC6gdT, and CD4CD8T-cells. [39] 05638 LCN2 lipocalin 2, works in antibacterial, anti-inflammatory, and protection against cell and tissue stress. [40] 36017 a NA is collected in gene set for CD8 T cell (https://report.majorbio.com/) 44758 NA Undetermined gene * Gene ID: ENSSSCG000000xxxxx; " a ": genes relevant to CD8+ T cell function by the functional annotation query.

Discussion
In this study, we show that immunization of a small number of non-inbred pigs with an SFV replicon-based plasmid DNA or replicon particles expressing ASFV antigens elicits efficient cellular and humoral immune responses in some but not all of the immunized pigs. This inconsistency is due to multiple factors, but the complicated genetic background of the non-inbred experimental pigs and the manual delivery method may be the two main 'culprits'. A much larger-scale animal experiment would be effective to generate statistically significant data and to reduce the variations, but was hampered by many constraints. Alternatively, personalized analysis of immune response to vaccination in individuals would be meaningful to understand the regulatory mechanisms controlling the immune response in animals with limited resources and time. For the same consideration, a 'vector-only control' group was not included in this study. Instead, a strategy to compare and analyze data between pre-and post-immunization (post-boost) in individual animals was adopted.
ASFV structure proteins p30, p54, and CD2v are highly immunogenic using different expression systems, such as baculovirus and adenovirus vectors [14,41,42]. Previous studies with an alphavirus replicon system demonstrated that immunization of pigs with P30 replicon particles induced high anti-p30 antibodies after boost with a naturally attenuated ASFV isolates, OURT88/3 [43], but it was unclear if using IRP alone would elicit an efficient adaptive immune response. In this study, we observed quick T cell differentiation and expansion, efficient antibody production, and effective cellular response in several immunized pigs, demonstrating that the SFV-replicon DNA and/or IRP expression system could indeed induce specific humoral and cellular response efficiently without boost with LAV. It points to the possibility of applying the SFV-replicon system as a useful platform for vaccine development against ASFV. The large amount of production of mRNA, ssRNA, dsRNA, and the targeted protein during the one-cycle replication process may act as PAMPs to drive the innate immune response efficiently. These RNA adjuvant properties and higher levels of antigens produced by the self-replication system would make it a more ideal choice over the conventional CMV promoter-based plasmid DNA system.
Five signaling pathways, including TLR, CLR, NOD-like receptor, cytosolic DNA sensor-mediated, and IL17 receptor-mediated signaling pathways, are shown to be positively correlated to the humoral immune response presented by anti-p30 antibody production and inversely correlate to both p30-and CD2v-stimulated cellular response, highlighting their importance in the navigation of the antigen-stimulated adaptive immunity. Both extracellular and intracellular membrane-associated PRRs, including TLR4, DHX58/DDX58 and ZBP1, were mainly identified in pigs with high levels of specific antibody response (#1 and #11 pigs), but barely detected in other pigs after the second boost. TLR4, DHX58/DDX58 and ZBP1 together with their downstream IRF7 and IRF1, the important type1 IFN regulators, may form a dynamic feedback loop. Antigen-stimulated up-regulation of this group of genes would activate the corresponding signaling pathways and induce the expression of cytokines and chemokines required for the activation and proliferation of T cells essential for an efficient humoral response. An overexpression of these cytokines and chemokines, on the other hand, would induce harmful inflammatory response. As a feedback mechanism, the expression of these PRRs and signaling molecules would be repressed, as observed in the #1 pig, to benefit the development of specific cellular immunity as suggested by WGCNA. In support of this, down-regulation of these genes was not observed in #6 and #11 pigs, the two with much lower IFN-γ secreting cells at day 35. As TLRs were able to activate the innate immune system and subsequently lead to effective adaptive immunity, various TLRs have been added as immune modulators in many subunit vaccines to improve their immunogenicity. TLR4 agonists have emerged as highly potent activators of innate immunity in a number of vaccine adjuvants, such as in the approved vaccines Shingrix, Cervarix, and Fendrix [44]. Our results reported here would support that TLR4 and other PPRs may function as key factors in the induction of innate immunity, which subsequently regulates the adaptive immune response induced by the SFV-replicon-based vaccine candidates. Understanding the reciprocal regulation of the two arms of adaptive immunity and the feedback regulatory mechanisms in these pathways would be particularly important for the rational design of vaccines and vaccination programs against ASFV. In addition, ubiquitination of antigens was reported to reduce the efficiency of antibody production and direct the immune system towards cellular immunity with an unknown mechanism [21,23]. In consistency, our study revealed that pigs (#1, #11) elicited higher antibody responses had much lower IFN-γ production, and pigs (#5, #7, #8) elicited lower antibody responses and had higher IFN-γ production. It seems that, however, ubiquitination of the three antigens in this study was not shown to play a significant role in directing the immune response toward cellular immunity as reported.
The complement system is activated by antibodies while activated complement fragments impact the production of antibodies [45]. The complement system is traditionally considered to be an integral part of the innate immune defense against pathogens, playing a main function in lysis of targets such as bacteria. In the context of viral infections, however, the complement system has been shown to exhibit numerous antiviral mechanisms via direct neutralization of both enveloped and non-enveloped viruses, and/or the promotion of other immune responses [45]. Early studies have shown that complement-mediated lysis performed by neutrophils might act as a significant effector and play a role in the recovery from ASFV infection [46,47]. In this study, we observed that CIQA, CIQB and CIQC were significantly up-regulated in response to the boost in most pigs. CIQA, CIQB and CIQC encode the A-, B-and C-chain polypeptides of serum complement subcomponent C1q, respectively, and the hetero-multimer C1q binds to IgG and IgM to initiate the classical complement activation pathway. The accumulation of C1q together with the specific antibody in the immunized pigs may contribute to fast clearance of ASFV and to prevent viral spread.
IL17 signaling pathway modulates the production of cytokines, chemokines, and anti-microbial molecules, and regulates Th-17 cell differentiation, neutrophil recruitment, and T and B cell sensitization. IL17B is potentially involved in the activation of NF-kB and induction of pro-inflammatory cytokine IL8, the gene product of CXCL8 [48], and induces the production of TNFα and IL1β in a monocytic cell line [49]. This study revealed a group of DEGs including IL17B, TNFAIP3, MAPK14, CXCL2, FOS, FOSB, TNF, and CASP3 were significantly down-regulated whereas S100A8, NF-κBIA, and PTGS2 were up-regulated following the second boost. It was reported that inoculation with attenuated genotype I ASFV strains (NH/P68, OURT88/3) induced enhanced expression of key cytokines (IFNβ, several IFNα subtypes, IL1β, IL12p40, TNFα) and chemokines (CCL4, CXCL8, CXCL10) compared with highly virulent isolates (L60, Benin 97/1, 22653/14, Armenia2007) in in vitro cultured macrophage cells [50][51][52][53][54]. These soluble mediators released by the infected macrophages may promote the development of effective adaptive immune responses [18]. However, this study showed that the down-regulation of both CXCL2 and CXCL8 was detected in all vaccinated pigs, with variable levels (either up, down, or no change) of CXCL10 and CCL4. Further studies would be required to clarify the functional involvement of these genes in the host immune response against ASFV.
CLR signaling pathway, including DC-SIGN (CD209) and CLEC family members CLEC4K, CLEC5A, CLEC7A, CLEC9A, CLEC2, CLEC12A, and CLEC12B, is crucial for tailing immune response to pathogens and induces diverse immune responses. DC-SIGN on myeloid DCs is involved in inducing the differentiation of Th1, Th2, Th17, and regulatory T cells through binding to different ligands [55]. CLEC7A expressed in myeloid DCs, monocytes, macrophages, and B cells induces the differentiation of Th1 and Th17 cells through the induction of IL1β, IL6, IL12, and IL23 as well as TNF and CXCL2 [55]. In this study, LEC7A together with the co-factors and transcription products promoted by CLR signaling pathway, such as MAPK14, CASP1, TNF, RELB, EGR2, BCL3, PTGS2, were differentially regulated in several pigs following the second boost. DC-SING(CD209), a DC cell receptor to initiate the adaptive immune response, was highly up-regulated (Log2FC > 3.9) in #1 and #11 pigs, raising a possibility that it may play a role in the regulation of antibody production in these pigs. BCL3 is a main regulator for Th2 cell differentiation, and RELB, MAPK14, CASP1, TNF, PTGS2, and EGR2 are known to be involved in Th1 and Th17 differentiation. Similarly, variable expression levels of these DEGs were found in different pigs, suggesting a dynamic regulatory mechanism at the gene expression level to dictate a different dominance of the Th1 or Th2 response in individual animals.
Down-regulation of CTLA4 and up-regulation of SOCS3 may play a role in this vaccination-stimulated adaptive immunity. CTLA4 is induced in both CD4+ and CD8+ T cells upon T cell activation and functions as a receptor transmitting an inhibitory signal to T cells to act as an immune attenuator [56]. Down-regulation of CTLA4 identified in all vaccinated pigs in this study would prevent the CTLA4-mediated attenuation of T helper cells. SOCS3 is a member of the SOCS (Suppressors of Cytokine Signaling) family. The two types of SOSC family genes (type I, SOCS4-7; type II, CISH, and SOCS1-3) are key negative regulators of cytokine and growth factor signaling and modulators of the T cell biology [57]. In this study, SOCS3 was generally up-regulated and both EGR1 and EGR2 were down-regulated. Further investigation into the potential interaction between these molecules and their biological significance would shed new light on their involvement in regulating the immune response.
In conclusion, we have demonstrated that SFV replicon-based vaccine candidates expressing a combination of ASFV proteins with or without fusion to ubiquitin are highly immunogenic, inducing rapid T cell expansion/activation, specific antibody response, and antigen-induced IFN-γ secretion from PBMCs. However, the efficiencies of such vaccination were heavily influenced by the genetic background and other unknown intrinsic factors of the immunized hosts, and perhaps an inconsistent antigen delivery. Through personalized analysis, a number of trait-relevant hub genes, molecular signaling pathways, and regulatory mechanisms were uncovered. These findings would be instrumental to guiding future systematic studies using a larger scale of animal experiments, inclusion of more antigens contributed to the protective immunity against ASFV (such as EP153R, MGF505-7R and M448R), a combinatorial vaccine regimen capable of inducing both cellular and humoral immune responses systemically, and viral challenge. In addition, optimization of the delivery route and methods would be essential. Considering the richness of antigenpresenting Langerhans cells (LC) and dermal dendritic cells (DDC) in the skin, intradermal (ID) delivery using microneedles may improve the delivery efficiency and consistency.

Antibodies, Chemicals, Reagents, and Cells
Antibodies and reagents were purchased from the following companies: antibody BHK-21 cells were cultured at 37 • C with 5% CO 2 in DMEM supplemented with 10% FBS and 1% penicillin-streptomycin. PBMCs were prepared by separation from whole blood through density-gradient centrifugation with pig lymphocyte isolates according to the manufacturer's protocol. PBMCs were cultured in RPMI 1640 supplemented with 10% FBS, 50,000 IU penicillin/L, and 50 mg streptomycin/L. Trypan blue staining was used to assess the cell viability.

Preparation of Single Cycle IRPs
Single-cycle IRPs were prepared by co-transfecting BHK21 cells with pSFV-helper-CMV and an antigen-expressing replicon plasmid, pSFV-flagP54, pSFV-flagCD2v, pSFV-flagp30, pSFV-flagUbiCD2vEDp30 or pSFV-flagUbiCD2vEDp54, using TransIntroTM EL Transfection Reagent according to the manufacturer's instruction. At 30 h post-transfection, the culture medium was collected for IRP preparation, and cells were lysed for Western blot or protein purification.
After removal of cell debris from the culture medium by brief centrifugation, the IRP titer was determined with a fluorescent antibody staining technique, and the TCID50 (FFU/mL) was calculated using the Reed-Muench method.

SDS-PAGE and Western Blot Analysis
Approximately 25 µg of total proteins from each sample were mixed with LaemmLi's sample buffer (0.3125 M Tris-HCl, pH 6.8, 10% SDS, 50% glycerol, 25% β-mercaptoethanol, 0.025% bromophenol blue), and separated on an SDS-12% polyacrylamide gel by electrophoresis, followed by transferring onto a nitrocellulose membrane. The membrane was blocked with 5% BSA in TBST (20 mM Tris-HCl pH 7. 4, 150 mM NaCl, 0.1% Tween 20) for 1 h at room temperature, and then incubated with primary antibodies (anti-flag or ASFV serum) at 4 • C overnight. After 3 washes with TBST, the membrane was incubated with 1:15,000 diluted IFCT-conjugated goat anti-mouse IgG or goat anti-swine IgG at room temperature for 1 h. After 3 washes with TBST, the proteins on the membrane were detected with an Azure C600 imager according to the manufacturer's instructions.

Animal Experimental Design and Immunization
Animals were obtained from the Experimental Animal Center of Xinxing Dahua Agricultural, Poultry and Egg Co., Ltd., Yunfu, China. approved number SCXK (Guangdong) 2018-0019. Animal experiments were performed under the approval of the Animal Welfare and Ethical Censor Committee at South China Agricultural University. After the completion of experimental procedures, all animals were euthanized before mechanistic execution.
Eight 30-day-old Sanhua pigs (hybrid of civil pig Dulock and Big Yorkshire) were divided into two groups and fed for 1 week before immunization. Four pigs in Group 1 were immunized with mixed plasmids or IRPs expressing p54, pCD2v, and p30. Two pigs (#1 and #2) in this group were inoculated with 0.6 mg of total plasmid DNA in saline composed of 0.2 mg each of pSFV-flagP54, pSFV-flagCD2v, and pSFV-flagP30; the other two (#5 and #6) were immunized with 1.8 mL of IRP mixture composed of 0.6 mL (10 6 FFU/mL) each of rSFV-flagP54, rSFV-flagCD2v, and rSFV-flagP30. The four pigs in group 2 were immunized with mixed plasmids or IRPs expressing UbiCD2vEDp30 and UbiCD2vEDp54. Two pigs (#7 and #8) were immunized with 0.6mg of total plasmid DNA in saline containing 0.3 mg each of pSFV-flagUbiCD2vEDp30 and pSFV-flagUbiCD2vEDp30; pigs #11 and #12 were immunized with 1.2 mL of IRP mixture containing 0.6 mL (10 6 FFU/mL) each of rSFV-flagUbiCD2vDEp30 and rSFV-flagUbiCD2vEDp54.
The immunization protocol included one priming and two boosts at a 2-week interval, by inoculating one third each of the doses at three different sites (the rectus femoris quadriceps muscle, the trapezius muscle of the neck, and at an ear by a subcutaneous injection). Whole blood samples were collected at 7-to 14-day intervals before inoculation if required.

Measurements of Specific Humoral and Cellular Immune Response
ASFV-specific antibodies in immunized pig sera were measured in duplicate by ELISA assay using the commercially available ASFV sandwich ELISA p30 antibody detection kit and AsurDxTM ASFV p54 Antibodies Test Kit according to the manufacturer's instructions.
IFN-γ ELISpot kit was used to measure the trained cellular response. Briefly, flagP30, flagP54, and flagCD2v proteins were prepared by anti-Flag gel purification from BHK cells transfected with the respective plasmid using the 3X Flag competitive elution method (Beyotime). Non-denaturing proteins without contamination of light and heavy chains from the Flag antibody was obtained with high elution efficiency. After removal of the 3X Flag peptide by filtration with a protein concentrator (10 K MWCO, Thermo) and washing twice with PBS, the purified proteins were resuspended in PBS. 10 4 live porcine PBMCs cells from each sample were stimulated for 20 h with 60µg each of the purified flagP30, flagP54 protein, or flagCD2v protein. RPMI1640 only and 10 g/mL of phytohem agglutinin were used as the negative and positive controls, respectively. Samples were sent to BioStone for score and the frequency of the antigen-specific IFN-γ-secreting cells (IFN-γ-SC) per million PBMCs was calculated after subtracting the spot counts in the negative-control wells from triplicates.

Flow Cytometry
PBMCs were stained with a CD4 monoclonal antibody (74-12-4), FITC and CD8-α monoclonal antibody (76-2-11), and P-phycoerythrin according to the manufacturer's instruction, and analyzed by flow cytometry (CytoFLEX SRT, Backman Coulter Inc., Indianapolis, IN, USA). The measurement was repeated twice and similar results were obtained, one of them was presented.

Transcriptomic Analysis
Whole blood cells were frozen in TRIzol reagent immediately after collection and sent to Shanghai Majorbio Bio-Pharm Technology Co., Ltd., Shanghai, China for RNA extraction and RNA-seq analysis. Technical triplicates were performed for each vaccinated pig in order to reduce systematic error. Total RNA was extracted from whole blood samples using TRIzol ® reagent according to the manufacturer's instructions (Invitrogen) and genomic DNA was removed using DNase I (TaKara). The RNA quality was determined by 2100 Bioanalyser (Agilent) and quantified using the ND-2000 (NanoDrop Technologies, Wilmington, DE, USA). Only high-quality RNA sample (OD260/280 = 1.8~2.2, OD260/230 ≥ 2.0, RIN ≥ 6.5, 28 S:18 S ≥ 1.0, >1 µg) was used to construct sequencing libraries.
RNA-seq transcriptome libraries were prepared following TruSeqTM RNA sample preparation with the Illumina Kit (San Diego, CA, USA) using 1µg of total RNA. Total mRNA was isolated with oligo (dT) beads and fragmented with a fragmentation buffer, and double-stranded cDNA was synthesized using a SuperScript double-stranded cDNA synthesis kit (Invitrogen, CA, USA) with random hexamer primers (Illumina). The synthesized cDNA was then subjected to end-repair, phosphorylation and 'A' base addition. Libraries were size-selected for cDNA target fragments of 300 bp on 2% low-range ultra-agarose separation gel after PCR amplification for 15 cycles using Phusion DNA polymerase (NEB). After quantified by TBS380, the end-paired RNA-seq libraries were sequenced with an Illumina HiSeq xten/NovaSeq 6000 sequencer (2 × 150 bp read length).
To identify differentially expressed genes (DEGs) between two different samples, the expression level of each transcript was calculated according to the transcripts per million reads (TPM) method, and RSEM (http://deweylab.biostat.wisc.edu/rsem/) was used to quantify the gene abundance. Essentially, differential expression analysis was performed using the DESeq2 [4]/DEGseq [5]/EdgeR [6] with Q value ≤ 0.05, DEGs with |log2FC| > 1 and Q value ≤ 0.05 (DESeq2 or EdgeR)/Q value ≤ 0.001 (DEGseq) were considered to be significantly different for the expressed genes. The RNA sequence raw data were submitted to NCBI and the access number is PRJNA936467. In addition, functional-enrichment analyses including GO and KEGG were performed to identify which DEGs were significantly enriched in GO terms and metabolic pathways at Bonferroni-corrected p-value ≤ 0.05, compared with the whole-transcriptome background. GO functional enrichment and KEGG pathway analysis were carried out by Goatools (https://github.com/tanghaibao/Goatools) and KOBAS (http://kobas.cbi.pku.edu.cn/home.do) to summarize the functional annotation of genes/transcripts in NR, Swiss-Prot, Pfam, EggNOG, GO and KEGG databases, and provide two retrieval methods to query the gene information corresponding to specific functions, or retrieve their functions according to the gene information, so as to quickly lock the required information. More technical details in RNA extraction, library preparation, Illumina Hiseq xten/Nova seq 6000 Sequencing, read mapping, differential expression analysis, principal component analysis (PCA), and functional enrichment are available at www.majorbio.com.
To select genes associated with vaccination-induced adaptive immunity, the weighted correlation network analysis (WGCNA) R software package, a comprehensive collection of R functions for performing various aspects of weighted correlation network analysis, was used [32]. In principle, co-expression networks are undirected weighted gene networks. The node of such a network corresponds to gene expression profiles, and edges between genes are determined by the pairwise correlations between gene expressions. The soft threshold was used to ensure a scale-free network, and Pearson's coefficient correlations were calculated to assess the relationships between gene expression data and the trait [32]. A software package available at www.majorbio.com was applied for WGCAN in this study. The transcriptomic datasets collected on days 28 and 29 post-priming were auto-filtered through cluster analysis to remove the outlier, and genes/transcripts were divided into modules according to the expression trend of genes/transcripts with this software package. The correlation of modules with specific phenotypes was obtained by inputting the trait (OD values of anti-p30 or IFN-γ ELISpot readouts) in WGCNA. The modules with the highest correlation and more significant p-value were selected as the characteristic modules for the phenotype. The hub genes were selected from the characteristic modules by Venn with DEGs in pigs eliciting the most obvious trait. The molecular signatures of the hub genes were denoted with KEGG assay and signaling pathways with p-value < 0.05 were considered a significantly enriched term. Data Availability Statement: Data supporting the findings of this study are available in this paper, Supplementary information, or are available from the corresponding authors upon request.