Advertisement

In silico design and evaluation of a multi-epitope and multi-antigenic African swine fever vaccine

Open AccessPublished:October 12, 2022DOI:https://doi.org/10.1016/j.immuno.2022.100019

      Abstract

      African Swine Fever (ASF) is caused by a highly contagious and fatal hemorrhagic virus, which in 2019 alone in Asia, has killed 8 million pigs with a devastating estimated economic loss amounting to $130 billion. There were attempts to control ASFV transmission; however, developed vaccines failed to produce lasting immunity. Currently, no vaccine has been approved yet. This study designed a novel multi-epitope and multi-antigenic vaccine using open-access bioinformatics tools. B-cell, helper-T and cytotoxic T-cell epitopes were predicted using consensus sequences from ASFV genotypes of antigens p12, p17, p22, p54, p72, and CD2v, and combined with adjuvants and linkers to form the ASF vaccine. Analyses revealed that the ASF vaccine is stable, antigenic, non-allergenic, and not cross-reactive. Docking of SLA-1 to CTL-HTL regions of the developed vaccine revealed that it effectively binds to SLA-1, a vital process in priming an effective immune response. Immune simulations demonstrated that the designed ASF vaccine can elicit primary and secondary immune responses, and stimulate the production of effector immune cells and cytokines. Overall, these results revealed that the designed multi-epitope and multi-antigenic ASF vaccine is potentially effective and warrants further in vitro and in vivo studies to confirm its protective function against ASFV infection.

      Graphical abstract

      Keywords

      Introduction

      African Swine Fever Virus is an emerging DNA arbovirus (arthropod-borne) of global concern in the pig industry, causing a fatal hemorrhagic disease among domestic pigs [
      • Gaudreault NN
      • Madden DW
      • Wilson WC
      • Trujillo JD
      • Richt JA.
      African swine fever virus: an emerging DNA arbovirus.
      ,
      • Revilla Y
      • Pérez-Núñez D
      • Richt J.A.
      African swine fever virus biology and vaccine approaches.
      ]. ASFV is a sole virus member of the Asfarviridae family of double-stranded DNA viruses[
      • Gaudreault NN
      • Madden DW
      • Wilson WC
      • Trujillo JD
      • Richt JA.
      African swine fever virus: an emerging DNA arbovirus.
      ,
      • Revilla Y
      • Pérez-Núñez D
      • Richt J.A.
      African swine fever virus biology and vaccine approaches.
      ]. ASFV mainly infects swines and warthogs, and can be harbored by ticks, highlighting its limited host range and its non-infectivity to humans due to its low mutation rates [
      • Blome S
      • Franzke K
      • Beer M.
      African swine fever – a review of current knowledge.
      ]. The transmission cycle of ASFV is illustrated in Fig. 1. Soft ticks (Ornithodoros) act as vectors and reservoirs of ASFV, where the virus replicates in their midgut [
      • Gaudreault NN
      • Madden DW
      • Wilson WC
      • Trujillo JD
      • Richt JA.
      African swine fever virus: an emerging DNA arbovirus.
      ]. Clinical manifestations of ASFV infection vary depending on infecting strain and host susceptibility [
      • Gaudreault NN
      • Madden DW
      • Wilson WC
      • Trujillo JD
      • Richt JA.
      African swine fever virus: an emerging DNA arbovirus.
      ]. Highly fatal (100% mortality rates in domestic swine) ASFV genotype II was found to produce non-specific manifestations (e.g., diarrhea, fever) in infected swine, while swine with non-fatal ASFV genotype I was often asymptomatic.
      Fig 1
      Fig. 1Transmission Cycle of ASFV. In the tick's life cycle, ASFV can be transmitted in different modes: sexual (male to female), transstadial (offspring to parent) and transovarial (parent to offspring). These infected ticks may transmit the virus to domestic pigs (Sus scrofa domesticus), warthogs (Phacochoerus africanus) and wild boar (Sus scrofa). Other modes of transmission include livestock-wildlife interaction, swill feeding, animal trade, fomites, intra-farm outbreaks, and sylvatic cycle (warthog to ticks and vice versa).
      ASFV has an icosahedral virion (Fig. 2) that spans from 175 to 215 nm in diameter [
      • Ros-Lucas A
      • Correa-Fiz F
      • Bosch-Camós L
      • Rodriguez F
      • Alonso-Padilla J
      Computational analysis of african swine fever virus protein space for the design of an epitope-based vaccine ensemble.
      ]. It is composed of a nucleoprotein core, core shell, inner lipid envelope, icosahedral capsid and host-derived outer lipid envelope [
      • Ros-Lucas A
      • Correa-Fiz F
      • Bosch-Camós L
      • Rodriguez F
      • Alonso-Padilla J
      Computational analysis of african swine fever virus protein space for the design of an epitope-based vaccine ensemble.
      ]. The host-derived outer envelope is composed of several proteins such as p12 and CD2v [
      • Ros-Lucas A
      • Correa-Fiz F
      • Bosch-Camós L
      • Rodriguez F
      • Alonso-Padilla J
      Computational analysis of african swine fever virus protein space for the design of an epitope-based vaccine ensemble.
      ]. The outer capsid is mainly composed of p72 major capsid proteins forming hexameres and other minor capsid proteins such as the penton protein that contributes to its icosahedral shape [
      • Liu Q
      • Ma B
      • Qian N
      • Zhang F
      • Tan X
      • Lei J
      • Xiang Y.
      Structure of the african swine fever virus major capsid protein p72.
      ]. Meanwhile, the inner envelope follows the icosahedral shape of the capsid and is composed of single-pass transmembrane proteins such as p12, p17 and p54 [
      • Alejo A
      • Matamoros T
      • Guerra M
      • Andrés G.
      A proteomic atlas of the african swine fever virus particle.
      ]. The protein core contains the ASFV DNA and viral polyproteins associated with the transcription machinery [
      • Blome S
      • Franzke K
      • Beer M.
      African swine fever – a review of current knowledge.
      ]. The genome size of ASFV ranges from 170 to 190 kilobase pairs which consists of 151 to 167 open reading frames encoding 68 virion-associated proteins [
      • Blome S
      • Franzke K
      • Beer M.
      African swine fever – a review of current knowledge.
      ]. The ASFV genome mainly encodes proteins that function in virion assembly, immune modulation, genome replication, and repair [
      • Blome S
      • Franzke K
      • Beer M.
      African swine fever – a review of current knowledge.
      ]. Diagnosis and genotyping of ASFV strains mainly use p72 gene, whereas recent studies suggested the use of other discriminative genes such as p54 and pB602L genes [
      • Blome S
      • Franzke K
      • Beer M.
      African swine fever – a review of current knowledge.
      ].
      Fig 2
      Fig. 2Schematic illustration of ASFV virion.
      The socio-economic impact of African Swine Fever (ASF) to swine industries worldwide made it a notifiable disease by the World Organisation for Animal Health [
      • Blome S
      • Franzke K
      • Beer M.
      African swine fever – a review of current knowledge.
      ]. Prevalence of ASF in Asia led to the death of 8 million pigs resulting in an estimated economic loss amounting to $130 billion in 2019 alone [

      Kumagai N. African swine fever in Asia and the Pacific. Presented at: African swine fever: An unprecedented global threat - A challenge to livelihoods, food security and biodiversity. Call for action; October 26-30, 2020; Online Webinar. https://www.fao.org/3/cb2145en/cb2145en.pdf.

      ,

      Weaver T, & Habib N (Asian Development Bank). Evaluating losses associated with african swine fever in the People's Republic Of China and neighboring countries. 2021.https://www.adb.org/sites/default/files/publication/680961/eawp-27-losses-african-swine-fever-prc-neighboring-countries.pdf.

      ]. Due to the persisting crisis brought about by ASF epidemics, it is imperative to explore vaccine strategies to protect against ASF and control ASFV transmission. Early attempts to develop vaccines failed to induce protective immune responses against ASFV infection. Recent studies reported conflicting results on ASFV-specific neutralizing antibodies whilst non-neutralizing ASFV antibodies may even lead to swine mortality [
      • Gaudreault NN
      • Madden DW
      • Wilson WC
      • Trujillo JD
      • Richt JA.
      African swine fever virus: an emerging DNA arbovirus.
      ]. Despite the availability of genomic sequences and numerous immunogenic antigens of various ASFV strains, the functions of other ASFV proteins or viral-coding genes remain to be elucidated. These characterized proteins, especially those exposed on the surface of the virions, may be potent candidates for vaccine development. Exploring these immunogenic antigens, this study designed a multi-epitope and multi-antigenic vaccine composed of predicted B-cell, T-helper and Cytotoxic T-cell epitopes from protective ASFV surface antigens. The designed ASF vaccine was evaluated for its immunogenicity, allergenicity, and cross-reactivity. Its physicochemical properties and its secondary and tertiary structures were estimated. Lastly, immune simulations were done to predict its ability to elicit immune responses.

      Methodology

      Retrieval of ASFV sequences

      The methodological framework of the in silico design of the multi-epitope ASF Vaccine is illustrated in Fig. 3. Fifty-six (56) African Swine Fever Virus (ASFV) isolates from different countries were retrieved from the NCBI Virus Database. The protein sequences of p12, p17, p22, p54, p72, and CD2v from these isolates were aligned using Clustal Omega [
      • Madeira F.
      • Pearce M.
      • Tivey A.R.N.
      • Basutkar P.
      • Lee J.
      • Edbali O.
      • Madhusoodanan N.
      • Kolesnikov A.
      • Lopez R.
      Search and sequence analysis tools services from EMBL-EBI in 2022.
      ]. Conserved fragments were identified through protein variability analysis using the Protein Variability Server with a Shannon entropy threshold of 1.0 which were used for screening of B-cell, Helper T-cell and Cytotoxic T-cell epitopes [
      • Garcia-Boronat M.
      • Diez-Rivero C.M.
      • Reinherz E.L.
      • Reche P.A.
      PVS: a web server for protein sequence variability analysis tuned to facilitate conserved epitope discovery.
      ].
      Fig 3
      Fig. 3Methodological framework of the in silico design of the ASF vaccine.

      Linear B lymphocyte epitope prediction

      The identified conserved sequences from six antigens of ASFV were used to screen for linear B lymphocyte (BL) epitopes. Several tools in Immune Epitope Database (IEDB) such as Emini Surface Accessibility (ESA), BepiPred Linear Epitope 2.0 (BLE), Kolaskar & Tongaonkar Antigenicity (KTA), and ABCPred Server were used to generate linear BL epitopes [
      • Emini E.A.
      • Hughes J.V.
      • Perlow D.S.
      • Boger J.
      Induction of hepatitis A virus-neutralizing antibody by a virus-specific synthetic peptide.
      ,
      • Jespersen M.C.
      • Peters B.
      • Nielsen M.
      • Marcatili P.
      BepiPred-2.0: improving sequence-based B-cell epitope prediction using conformational epitopes.
      ,
      • Kolaskar A.S.
      • Tongaonkar P.C.
      A semi-empirical method for prediction of antigenic determinants on protein antigens.
      ,
      • Saha S.
      • Raghava G.P.
      Prediction of continuous B-cell epitopes in an antigen using recurrent neural network.
      ]. Epitopes that overlapped with glycosylation and cleavage sites were excluded. BL epitopes with 7–20 amino acid residues obtained from at least 2 servers were used for further evaluation.

      Linear helper T lymphocyte epitope prediction

      The consensus sequences from the six antigens of ASFV were assessed for predicted MHC II binding using IEDB analysis resource: MHC II Binding Prediction set with IEDB recommended 2.22. Since only human MHC II alleles are available in the prediction tool, common human MHC II alleles that are orthologs to their swine counterpart alleles were used to predict epitope peptides. Helper T-lymphocyte (HTL) epitopes with IC50 < 500 and length of 11 amino acids were considered for further evaluations.

      Linear cytotoxic T lymphocyte epitope prediction

      From the consensus sequences obtained for the six ASFV antigens, proteasomal cleavage/ TAP transport/ MHC class I combined predictor tool in IEDB was employed for cytotoxic T-lymphocytes (CTL) epitopes [
      • Tenzer S
      • Peters B
      • Bulik S
      • Schoor O
      • Lemmel C
      • Schatz MM
      • Kloetzel PM
      • Rammensee HG
      • Schild H
      • Holzhutter HG.
      Modeling the MHC class I pathway by combining predictions of proteasomal cleavage, TAP transport and MHC class I binding.
      ]. In addition to antigen processing (proteasomal cleavage), TAP (transporter associated with antigen processing protein) transport, MHC class I binding, and T cell receptor-MHC interaction are included in the prediction outcomes for potential CTL epitopes. Common SLA class I molecules (SLA -1, -2, and -3) are utilized in the prediction. The output of MHC Class I binding predictions include the possible epitopes with predicted binding affinities (IC50), proteasome, TAP, MHC and total scores. The predicted epitopes were filtered using these parameters: epitope length of 9 to 11 residues, IC50 <500 nM, TAP score >1.0, proteasome-processing score > 1.0, and those covering the MHC I SLA alleles.

      Evaluation of predicted epitopes

      Predicted BL, HTL, and CTL epitopes were evaluated for antigenicity, allergenicity, and cross-reactivity [
      • Doytchinova I.A.
      • Flower D.R.
      Identifying candidate subunit vaccines using an alignment-independent method based on principal amino acid properties.
      ,
      • Doytchinova I.A.
      • Flower D.R.
      VaxiJen: a server for prediction of protective antigens, tumour antigens and subunit vaccines.
      ]. 16–17 Antigenic ASFV epitopes were identified using Vaxijen 2.0 Server with a threshold e” 0.5 and using virus as a target. Epitopes with potential allergenic sites were identified using Allergen FP v.1.0 [
      • Dimitrov I.
      • Naneva L.
      • Doytchinova I.
      • Bangov I.
      AllergenFP: allergenicity prediction by descriptor fingerprints.
      ]. The predicted epitopes were assessed for autoimmune reactions, exploring for possible hits against swine protein sequences (Sus scrofa, Sus scrofa domesticus, and Sus scrofa domestica) in UniProtKB and SwissProt databases using protein-protein BLAST (BLASTp). Default parameters in BLASTp excluding models, non-redundant refseq proteins, and uncultured/environmental sample sequences were used. Epitopes that passed all evaluationswere used in creating the ASF multi-epitope vaccine.

      Multi-epitope ASF vaccine construction

      The ASF vaccine was constructed using the evaluated epitopes. Overlapping epitopes of CTL and HTL from the same protein were shortlisted using BL epitopes as reference. CTL epitopes were also shortlisted using HTL epitopes as template, if the overlapping CTL and HTL epitopes did not overlap with any BL epitopes. Overlapping epitopes for each protein from the same type of lymphocyte were also combined into continuous peptides. The combined epitopes were arranged to form the protein vaccine. Valine was added in the N-terminus and two potential immune-adjuvant peptide sequences were used, specifically the A6 peptide (sequence: TNGDILNYY) and the F3 peptide (sequence: SVDSPTITY). In the constructed vaccine, A6 peptide was linked to F6 using an AAY linker since both peptides are CTL epitopes. The adjuvant peptides are linked to overlapped CTL epitopes using AAY linkers, while HTL and BL overlapping epitopes are linked using GPGPG linkers.

      Evaluation of the constructed multi-epitope ASF vaccine

      Antigenicity and allergenicity of the constructed vaccine was assessed using Vaxijen 2.0 and AllergenFP v1.0 servers, respectively. The vaccine was also checked for cross-reactivity against Sus scrofa (taxid: 9823), Sus scrofa domestica (taxid: 9825), and Sus scrofa domesticus (taxid: 9825) whilst excluding other models (XM/XP), non-redundant RefSeq proteins (WP) and uncultured/environmental sample sequences using BLASTp. The physicochemical properties were obtained using the Protparam tool [
      • Gasteiger E
      • Hoogland C
      • Gattiker A
      • Duvaud S
      • Wilkins MR
      • Appel RD
      • Bairoch A.
      • Walker J.M.
      Protein identification and analysis tools on the ExPASy server.
      ].

      Secondary and tertiary structure analysis and modeling

      The secondary structure composition was determined using both GOR IV and GlobPlot2.3 servers [
      • Linding R.
      GlobPlot: exploring protein sequences for globularity and disorder.
      ]. GOR IV created the secondary structure percentage profile specifically marking regions predicted as random coils, alpha helix, and extended strands. The tertiary structure was modeled using the Galaxy TBM tool and was refined further using Galaxy Refine server [
      • Ko J.
      • Park H.
      • Heo L.
      • Seok C.
      GalaxyWEB server for protein structure prediction and refinement.
      ]. Refined structures were checked for structural validity with ERRAT and Verify 3D [
      • Colovos C.
      • Yeates T.O.
      Verification of protein structures: patterns of nonbonded atomic interactions.
      ,
      • Lüthy R.
      • Bowie J.U.
      • Eisenberg D.
      Assessment of protein models with three-dimensional profiles.
      ]. [
      • Pettersen E.F.
      • Goddard T.D.
      • Huang C.C.
      • Couch G.S.
      • Greenblatt D.M.
      • Meng E.C.
      • Ferrin T.E.
      UCSF chimera-a visualization system for exploratory research and analysis.
      ]The structure was viewed via UCSF Chimera.

      Docking ASF vaccine to SLA-1

      The refined structure was docked to a cleaned crystal structure of swine major histocompatibility complex class 1 SLA-1 0401 from Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB ID: 3QQ3) using ClusPro server [
      • Kozakov D
      • Hall DR
      • Xia B
      • Porter KA
      • Padhorny D
      • Yueh C
      • Beglov D
      • Vajda S.
      The ClusPro web server for protein-protein docking.
      ,
      • Zhang N
      • Qi J
      • Feng S
      • Gao F
      • Liu J
      • Pan X
      • Chen R
      • Li Q
      • Chen Z
      • Li X
      • Xia C
      • Gao GF.
      Crystal structure of swine major histocompatibility complex class I SLA-1*0401 and identification of 2009 pandemic swine-origin influenza A H1N1 virus cytotoxic T lymphocyte epitope peptides.
      ]. The ASF Vaccine-SLA-1 complex with the least binding energy was chosen and modeled in USCF Chimera. The binding energy and affinity for the complex was evaluated at 37 °C by the PRODIGY web server [
      • Xue L.C.
      • Rodrigues J.P.
      • Kastritis P.L.
      • Bonvin A.M.
      • Vangone A.
      PRODIGY: a web server for predicting the binding affinity of protein–protein complexes.
      ].

      ASF vaccine immune simulations

      Immune responses elicited by the ASF vaccine were simulated using the C-ImmSim server [
      • Rapin N
      • Lund O
      • Bernaschi M
      • Castiglione F.
      Computational immunology meets bioinformatics: the use of prediction tools for molecular binding in the simulation of the immune system.
      ]. Default simulation parameters were used in the system for a prophylactic vaccine injected two times in three weeks intervals (time steps set at 1 and 63). The immune response profile of the vaccine was compared with that of the adjuvant alone.

      In silico cloning optimization

      For vaccine production, in silico cloning optimization was employed for the ASF vaccine. Using competent Escherichia coli K12 strains in silico, cloning optimization was performed using Java Codon Adaptation Tool (JCAT) [
      • Grote A.
      • Hiller K.
      • Scheer M.
      • Munch R.
      • Nortemann B.
      • Hempel D.C.
      • Jahn D.
      JCat: a novel tool to adapt codon usage of a target gene to its potential expression host.
      ]. JCAT allows the adaptation of a gene codon to sequenced prokaryotes approximating the codon adaptation index. Additional options which avoided prokaryotic ribosome binding sites, cleavage sites of restriction enzymes and rho-independent transcription terminators were applied. Moreover, the GC-content of the improved sequence of the vaccine was also computed.

      Results

      Conserved sequence analysis, epitope prediction and evaluation

      The conserved amino acid sequences of the ASFV antigens were chosen based on their location in the virion and existing reports on its immunogenicity against ASF. The resulting conserved sequences of p12, p17, p22, p54, p72, and CD2v antigens were used for BL, HTL and CTL epitope prediction. The predicted epitopes were evaluated for antigenicity, allergenicity, and cross-reactivity and those that passed all the evaluations were used in constructing the multi-epitope vaccine. A total of 36 epitopes were obtained: 12 BL epitopes, 20 HTL epitopes and 4 CTL epitopes.

      ASF multi-epitope vaccine construct and properties

      The multi-epitope vaccine (Fig. 4) was constructed using the BL, CTL, and HTL evaluated epitopes connected by linkers. The ASF vaccine consists of 398 residues generated from 36 selected epitopes that were merged into 18 epitope fragments arranged from N-terminus to C-terminus: CTL, HTL, BL epitopes, respectively.
      Fig 4
      Fig. 4ASF Vaccine multi-epitope diagram. The designed vaccine was organized according to the epitopes involved interspersed with respective linkers. The N-terminal contains a valine followed by the adjuvants (pink) with a B cell/HTL linker (beige) in between, while an adjuvant linker (blue) links the adjuvant region to the CTL epitopes (purple) that is interspersed with CTL linkers (green). HTL epitopes (blue green) follows and then the BL epitopes (sky blue) and both are interspersed with B cell/HTL linkers (beige).
      The ASF vaccine was predicted to be antigenic (0.5344) in Vaxijen 2.0 Server using “virus” as a target and non-allergenic with a Tanimoto Similarity index of 0.82. Additionally, the vaccine had no significant similarity with swine proteins. Protparam analysis revealed that the vaccine has a molecular weight of 42, 485. 82 g/mol and a theoretical isoelectric point (pI) of 10.10. The total number of negatively (Asp + Glu) and positively (Arg + Lys) charged residues were 15 and 45, respectively. The computed extinction coefficient was 54,570 M−1 cm−1 at 280 nm measurement in water and absorbance of 1.284 if all pairs of cysteine residues form cystines. On the other hand, if all cysteine residues are reduced the extinction coefficient is 54,320 M−1 cm−1 at 280 nm measurement in water. Its estimated half-life of the vaccine is 100 h in mammalian reticulocytes (in vitro), >20 h in yeast (in vivo), and >10 h in Escherichia coli (in vivo). The resulting instability index is 35.83 (<40 threshold) signifying that the vaccine is classified as a stable protein. The vaccine is hydrophilic as revealed by its grand average of hydropathicity is -0.356.

      ASF vaccine secondary and tertiary structures

      The secondary structure composition of the ASF vaccine is shown in Fig. 5. The most common structures are random coils (purple), followed by extended strands (red) and least being alpha helices (blue). The random coil regions are interspersed within different epitope divisions of the ASF vaccine which correlates with the identified disordered regions (1–7, 93–119, 133–174, 185–243, 255–267, 273–289, 296–308, 318–389). These regions are flexible, less likely to form stable folds, and act as binding regions for B-cell receptors. The best tertiary structure model of the ASF vaccine (Fig. 6) resulted in an elongated peptide with alpha helices interspersed between random coils. It is subdivided into 4 regions: F3-A6 Adjuvant, CTL, HTL and BL epitopes. The conformation with the least interaction energy by ASF Vaccine to SLA-1 docking was shown in Fig. 6. The estimated binding energy (∆G) and binding affinity (Kd) for the complex are -49.3712 kj/mol and 4.5E-09 mol/L, respectively. In addition, residues 42-118 are the interaction points of ASF Vaccine with SLA-1 which coincides with the predicted linear epitopes from the vaccine.
      Fig 5
      Fig. 5Graphical representations of secondary structures of the ASF vaccine. Predicted secondary structures in GOR IV include alpha helices (8.54%) in blue, extended strands (28.64%) in red and random coils (62.81%) in purple.
      Fig 6
      Fig. 6ASF vaccine tertiary structure (top) and ASF Vaccine docked with SLA-1 (bottom)

      Immune response profile and codon optimization of ASF vaccine

      Results from the immune simulations showed that the ASF Vaccine was able to elicit an immune response in contrast to the adjuvant alone (Fig. 7). The primary and secondary immune response was evidently higher in the ASF Vaccine than the adjuvant alone. This trend was also evident with that of the B-cell and cytotoxic T-cell population. Interestingly, populations of helper T-cells and most of the cytokines were relatively higher for the adjuvant alone than the vaccine. Meanwhile for the codon optimization, the resulting CAI-value obtained from the JCAT tool for ASFV multi-epitope vaccine codon usage optimization is 0.9361 which is inside the CAI range of 0.9 to 1. The GC-content of the improved vaccine codon sequence is 54.94% which is within the 30–70% range. Codon optimization of ASFV multi-epitope vaccine with an improved DNA codon sequence length of 1198 base pairs allows the vaccine to be highly expressed in E. coli K12.
      Fig 7
      Fig. 7C-ImmSim simulation profile of the ASF Vaccine and adjuvant alone. A) Immunoglobulin production and specific subclass (colored peaks) in response to injected antigen (black peaks); B) B-cell population per isotype; C) Population of HTL memory and not memory cells; D) CTL populations at the active, resting, and anergic state and: E) Counts induced cytokines and interleukins while the insert plot shows IL-2 level and Simpson Index (D)
      Fig 7
      Fig. 7C-ImmSim simulation profile of the ASF Vaccine and adjuvant alone. A) Immunoglobulin production and specific subclass (colored peaks) in response to injected antigen (black peaks); B) B-cell population per isotype; C) Population of HTL memory and not memory cells; D) CTL populations at the active, resting, and anergic state and: E) Counts induced cytokines and interleukins while the insert plot shows IL-2 level and Simpson Index (D)

      Discussion

      The in silico approach

      Vaccination remains to be a powerful tool in improving animal health, reducing transmission, and mitigating the impacts of diseases on livestock production [
      • Bitsouni V
      • Lycett S
      • Opriessnig T
      • Doeschl-Wilson A.
      Predicting vaccine effectiveness in livestock populations: a theoretical framework applied to prrs virus infections in pigs.
      ,
      • Meeusen EN
      • Walker J
      • Peters A
      • Pastoret PP
      • Jungersen G.
      Current status of veterinary vaccines.
      ]. In ASF, various vaccine platforms that were formerly utilized, failed to confer lasting immunity [
      • Oscherwitz J.
      The promise and challenge of epitope-focused vaccines.
      ]. For example, inactivated vaccines have been proven ineffective independent of the adjuvant used and its method of inactivation [
      • Blome S.
      • Gabriel C.
      • Beer M.
      Modern adjuvants do not enhance the efficacy of an inactivated African swine fever virus vaccine preparation.
      ,
      • Cadenas-Fernández E.
      • Sánchez-Vizcaíno J.M.
      • van den Born E.
      • Kosowska A.
      • van Kilsdonk E.
      • Fernández-Pacheco P.
      • Gallardo C.
      • Arias M.
      • Barasona J.A.
      High doses of inactivated African Swine fever virus are safe, but do not confer protection against a virulent challenge.
      ]. There are also studies that made use of immunogenic ASFV structural proteins, such as p30, p54, and p72, wherein immunization of the purified recombinant proteins, recombinant DNA, or a combination of the two have showed partial to no significant protection and only robust immune responses against the proteins was observed [
      • Argilaguet J.M.
      • Pérez-Martín E.
      • Nofrarías M.
      • Gallardo C.
      • Accensi F.
      • Lacasta A.
      • Mora M.
      • Ballester M.
      • Galindo-Cardiel I.
      • López-Soria S.
      • Escribano J.M.
      • Reche P.A.
      • Rodríguez F.
      DNA vaccination partially protects against African swine fever virus lethal challenge in the absence of antibodies.
      ,
      • Barderas M.G.
      • Rodríguez F.
      • Gómez-Puertas P.
      • Avilés M.
      • Beitia F.
      • Alonso C.
      • Escribano J.M.
      Antigenic and immunogenic properties of a chimera of two immunodominant African swine fever virus proteins.
      ,
      • Gómez-Puertas P.
      • Rodríguez F.
      • Oviedo J.M.
      • Brun A.
      • Alonso C.
      • Escribano J.M.
      The African swine fever virus proteins p54 and p30 are involved in two distinct steps of virus attachment and both contribute to the antibody-mediated protective immune response.
      ,
      • Neilan J.G.
      • Zsak L.
      • Lu Z.
      • Burrage T.G.
      • Kutish G.F.
      • Rock D.L.
      Neutralizing antibodies to African swine fever virus proteins p30, p54, and p72 are not sufficient for antibody-mediated protection.
      ]. This ineffectiveness may be attributed to factors such as high mutability of protein targets and lack of clearly defined targets that could elicit humoral immunity. There also exist research gaps on ASFV infection and immunity, strain variation, and ASFV protein functions, which hampers the development of an effective vaccine. Another aspect to consider is the fact that ASFV contains numerous immunogenic antigens, and some other proteins or viral-coding gene functions remain elusive, which accentuates the need to explore the immune response against these ASFV antigens [
      • Rock DL.
      Thoughts on african swine fever vaccines.
      ]. Moreover, neutralizing antibodies plays an important role in providing a high degree of immune protection and conferring long lasting immunity. Recent studies reported conflicting results on ASFV-specific neutralizing antibodies whilst non-neutralizing ASFV antibodies may even cause swine mortality.
      In the advent of biotechnology, production of subunit vaccines is sought-after since it explores small amounts of protein or glycoprotein virus fragments that are able to induce an effective immune response. Development of subunit vaccines first requires identification of immunogenic antigens; however, the conventional flow of antigen screening is a tedious process. In silico screening prior to conducting in vitro or in vivo assays is advisable to prevent problems on antigen expression [
      • Doytchinova I.A.
      • Flower D.R.
      Identifying candidate subunit vaccines using an alignment-independent method based on principal amino acid properties.
      ]. Open-access immunoinformatics tools were utilized in this study to design, predict and evaluate ASF Vaccine from previously studied immunogenic antigens, prior to actual vaccine production.

      In silico vaccine design/development

      The vaccine was designed using predicted BL, HTL and CTL epitopes from the conserved sequences of six antigens from different ASFV isolates. Conserved sequences from each antigen were identified through the Protein Variability Server, an online based tool that makes use of variability metrics to calculate for the absolute site variability in a given multiple protein-sequence alignments [
      • Garcia-Boronat M.
      • Diez-Rivero C.M.
      • Reinherz E.L.
      • Reche P.A.
      PVS: a web server for protein sequence variability analysis tuned to facilitate conserved epitope discovery.
      ]. Conserved fragments used for screening of B-cell, Helper T-cell and Cytotoxic T-cell epitopes were identified using a Shannon entropy threshold of 1.0, to assure that these sequences are highly conserved. The values of Shannon entropy (H) vary from 0–4.332 [
      • Litwin S.
      • Jores R.
      • Perelson A.S.
      • Weisbuch G.
      Shannon information as a measure of amino acid diversity.
      ]. A Shannon entropy of H ≥ 2.0 is considered variable, while those with H ≤ 2 is considered conserved [
      • Litwin S.
      • Jores R.
      • Perelson A.S.
      • Weisbuch G.
      Shannon information as a measure of amino acid diversity.
      ]. Highly conserved positions are those with H ≤ 1 [
      • Litwin S.
      • Jores R.
      • Perelson A.S.
      • Weisbuch G.
      Shannon information as a measure of amino acid diversity.
      ]. Thus, a Shannon entropy threshold of 1 was in the study to assure that the viral sequences that would be used for downstream analysis are highly conserved. In a rapidly evolving pathogen that makes use of sequence variability for immune evasion, this tool is able to identify distinct patterns of conserved antigenic residues which can be used as protein targets in an epitope vaccine design [
      • Garcia-Boronat M.
      • Diez-Rivero C.M.
      • Reinherz E.L.
      • Reche P.A.
      PVS: a web server for protein sequence variability analysis tuned to facilitate conserved epitope discovery.
      ]. The approach of creating a multiple epitope vaccine is considered advantageous since it includes MHC-restricted epitopes that can be identified by T-cell receptors from a number of T-cell subsets [
      • Zhang L.
      Multi-epitope vaccines: a promising strategy against tumors and viral infections.
      ]. It also consists of overlapping BL, HTL and CTL epitopes which are key elements in inducing immune responses against the target virus [
      • Zhang L.
      Multi-epitope vaccines: a promising strategy against tumors and viral infections.
      ]. This approach also covers several genotypes and adjuvants which could be added in order to enhance vaccine immunogenicity [
      • Zhang L.
      Multi-epitope vaccines: a promising strategy against tumors and viral infections.
      ].
      Immunogenic antigens were chosen for epitope prediction, including p12, p17, p22, p54, p72 and CD2v. These antigens are situated in the capsid, inner envelope, and outer envelope for ease of recognition of immune cells. There have been attempts to identify protective antigens among the ASFV antigens, which include p72 and p54 that were able to induce neutralizing monoclonal and polyclonal antibody production [
      • Rock DL.
      Thoughts on african swine fever vaccines.
      ,
      • Bosch-Camós L
      • López E
      • Rodriguez F.
      African swine fever vaccines: a promising work still in progress.
      ,
      • Wu K
      • Liu J
      • Wang L
      • Fan S
      • Li Z
      • Li Y
      • Yi L
      • Ding H
      • Zhao M
      • Chen J.
      Current state of global african swine fever vaccine development under the prevalence and transmission of ASF in China.
      ]. Capsid protein, p72 was previously described as the “main inducer of interferon-gamma” which is essential in cell-mediated response [
      • Wu K
      • Liu J
      • Wang L
      • Fan S
      • Li Z
      • Li Y
      • Yi L
      • Ding H
      • Zhao M
      • Chen J.
      Current state of global african swine fever vaccine development under the prevalence and transmission of ASF in China.
      ]. Meanwhile, CD2v antigen, involved in hemadsorption of ASFV-infected cells, was shown to induce production of hemagglutination inhibitory antibodies and prevent ASFV infection in vitro [
      • Bosch-Camós L
      • López E
      • Rodriguez F.
      African swine fever vaccines: a promising work still in progress.
      ]. Similarly, surface proteins p12, p17, and p22 were among the newly identified immunogenic antigens upon immunization by DNA primed and recombinant vaccinia virus [
      • Bosch-Camós L
      • López E
      • Rodriguez F.
      African swine fever vaccines: a promising work still in progress.
      ,
      • Jancovich JK
      • Chapman D
      • Hansen DT
      • Robida MD.
      • Loskutov A
      • Craciunescu F
      • Borovkov A
      • Kibler K
      • Goatley L
      • King K
      • Netherton CL
      • Taylor G
      • Jacobs B
      • Sykes K
      • Dixon LK.
      Immunization of oigs by DNA prime and recombinant vaccinia virus boost to identify and rank african swine fever virus immunogenic and protective proteins.
      ].

      Identification of B-cell epitopes

      B lymphocytes are one of the major components of the adaptive immune response that plays a central role in humoral and cellular immunity [
      • Althuwaiqeb SA
      • Bordoni B.
      Histology, B cell lymphocyte.
      ,
      • Hoffman W
      • Lakkis FG
      • Chalasani G.
      B Cells, antibodies, and more.
      ]. Early studies conducted by Zsak et al. reported that antibodies from pigs who recovered from ASFV protection experiment conferred immunity against ASF [
      • Zsak L
      • Onisk DV
      • Afonso CL
      • Rock DL.
      Virulent african swine fever virus isolates are neutralized by swine immune serum and by monoclonal antibodies recognizing a 72-kDa viral protein.
      ]. In vivo and in vitro studies observed the potential protective effects of antibodies through different modes of actions such as complement-mediated cell lysis or antibody dependent cellular cytotoxicity [
      • Takamatsu H
      • Denyer MS
      • Lacasta A
      • Stirling CMA
      • Argilaguet JM
      • Netherton CL
      • Oura CAL
      • Martins C
      • Rodríguez F.
      Cellular immunity in ASFV responses.
      ,
      • Rock DL.
      Challenges for African swine fever vaccine development-"… perhaps the end of the beginning.".
      ]. With these, it is evident that antibodies could provide protection against ASF therefore, BL epitopes were included in the vaccine design. Platforms such as Emini Surface Accessibility (ESA), BepiPred Linear Epitope 2.0 (BLE), Kolaskar & Tongaonkar Antigenicity (KTA) tools found in Immune Epitope Database (IEDB), and ABCPred Server were used to generate linear B-Cell epitopes. Emini Surface Accessibility is a tool based on Emini's surface accessibility scale which predicts the accessibility profile of a hexapeptide sequence showing its probability of being found on the surface [
      • Emini E.A.
      • Hughes J.V.
      • Perlow D.S.
      • Boger J.
      Induction of hepatitis A virus-neutralizing antibody by a virus-specific synthetic peptide.
      ]. BLE is a tool used to predict B-cell epitopes, which is based on a random forest algorithm trained on epitopes from crystal structures [
      • Jespersen M.C.
      • Peters B.
      • Nielsen M.
      • Marcatili P.
      BepiPred-2.0: improving sequence-based B-cell epitope prediction using conformational epitopes.
      ]. KTA is a semi-empirical approach that utilizes the physicochemical properties and frequency of occurrence of amino acid residues in predicting epitopes [
      • Kolaskar A.S.
      • Tongaonkar P.C.
      A semi-empirical method for prediction of antigenic determinants on protein antigens.
      ]. This tool is considered to be 75% accurate in contrast to other known prediction tools [
      • Kolaskar A.S.
      • Tongaonkar P.C.
      A semi-empirical method for prediction of antigenic determinants on protein antigens.
      ]. Additionally, ABCPred is a tool used to predict linear B-Cell epitopes in a given sequence through the use of artificial neural networks. Exploring the combination of these platforms enabled us to predict for B-cell epitopes that can be included in the designed ASF Vaccine [
      • Saha S.
      • Raghava G.P.
      Prediction of continuous B-cell epitopes in an antigen using recurrent neural network.
      ].

      Identification of cytotoxic T-cell epitopes

      Cytotoxic T lymphocytes are important effector cells of the adaptive immune system, which are known to interact with MHC class I molecules present on antigen presenting cells. In swine immunity, MHC-I interacting with CTLs are referred to as SLA Class I. These CTLs cause cytotoxicity to virus-infected cells through the action of perforins, granzymes and cytotoxic granules. As reported by Denyer et al. in 2006, there are two phenotypes of CTLs involved in ASF immune response, both of which cause cytolysis to infected cells [
      • Denyer MS
      • Wileman TE
      • Stirling CMA.
      • Zuber B
      • Takamatsu H
      Perforin expression can define CD8 positive lymphocyte subsets in pigs allowing phenotypic and functional analysis of natural killer, cytotoxic T, natural killer T and MHC un-restricted cytotoxic T-cells.
      ]. ASF outcomes are relative to the infecting viral strain, denoting the existence of more virulent and less virulent ASFV strains. In a study conducted by Oura et al., depletion of CD8+ T lymphocytes resulted in impaired protective immunity leading to increased swine fatalities [
      • Oura CAL
      • Denyer MS
      • Takamatsu H
      • Parkhouse RME.
      In vivo depletion of CD8+ T lymphocytes abrogates protective immunity to african swine fever virus.
      ]. This demonstrates the importance of CTL in conferring protective immunity against ASF. Proteasomal cleavage/TAP transport/MHC class I combined predictor tool in IEDB was employed for CTL epitopes. This tool applies the concept that antigens undergo proteasomal cleavage, transported and finally presented on MHC class I molecules [
      • Fleri W.
      • Paul S.
      • Dhanda S.K.
      • Mahajan S.
      • Xu X.
      • Peters B.
      • Sette A.
      The immune epitope database and analysis resource in epitope discovery and synthetic vaccine design.
      ]. MHC class I molecules are found in most nucleated cells, which present the processed antigens to CTL [
      • Fleri W.
      • Paul S.
      • Dhanda S.K.
      • Mahajan S.
      • Xu X.
      • Peters B.
      • Sette A.
      The immune epitope database and analysis resource in epitope discovery and synthetic vaccine design.
      ]. In addition to antigen processing (proteasomal cleavage), TAP (transporter associated with antigen processing protein) transport, MHC class I binding, then T cell receptor-MHC interaction are included in the prediction outcomes for potential CTL epitopes. MHC molecules are polymorphic and since the aim is to produce a vaccine for swine, common SLA class I molecules (SLA-1, -2, and -3) are utilized in the prediction. All six antigens were used in the SLA-I epitope prediction; however, only p22, CD2v and p72 yielded epitopes that passed the set parameters. After further screening, three p72 epitopes and one p22 epitope remained. To compensate for the reduced number of epitopes, two in silico predicted CTL epitope capable of stimulating interferon-gamma production in vitro were added as vaccine adjuvants [
      • Argilaguet J.M.
      • Pérez-Martín E.
      • Nofrarías M.
      • Gallardo C.
      • Accensi F.
      • Lacasta A.
      • Mora M.
      • Ballester M.
      • Galindo-Cardiel I.
      • López-Soria S.
      • Escribano J.M.
      • Reche P.A.
      • Rodríguez F.
      DNA vaccination partially protects against African swine fever virus lethal challenge in the absence of antibodies.
      ]. Both F3 and A6 peptides were from swine hemagglutinin antigen which is homologous to CD2v [
      • Argilaguet J.M.
      • Pérez-Martín E.
      • Nofrarías M.
      • Gallardo C.
      • Accensi F.
      • Lacasta A.
      • Mora M.
      • Ballester M.
      • Galindo-Cardiel I.
      • López-Soria S.
      • Escribano J.M.
      • Reche P.A.
      • Rodríguez F.
      DNA vaccination partially protects against African swine fever virus lethal challenge in the absence of antibodies.
      ].
      Epitopes with 9–11 residues were considered because MHC I molecules found in vertebrates generally bind with peptides that are 8–11 amino acids long [
      • Pedersen L.E.
      • Harndahl M.
      • Rasmussen M.
      • Lamberth K.
      • Golde W.T.
      • Lund O.
      • Nielsen M.
      • Buus S.
      Porcine major histocompatibility complex (MHC) class I molecules and analysis of their peptide-binding specificities.
      ]. In addition, cytotoxic T-cell epitopes presented by MHC I are commonly linear peptides with 8 to 11 amino acids in length [
      • Nevagi R.J.
      • Toth I.
      • Skwarczynski M.
      • Koutsopoulos S.
      Peptide-based vaccines.
      ]. With such, epitope length of 9–11 amino acids length were considered as part of the cutoffs. IC50 is presented in a percentile rank, which signifies specific binders having lower percentile ranks; hence, IC50 values less than 500 nM are ideal binders, which is the threshold associated with immunogenicity [
      • Fleri W.
      • Paul S.
      • Dhanda S.K.
      • Mahajan S.
      • Xu X.
      • Peters B.
      • Sette A.
      The immune epitope database and analysis resource in epitope discovery and synthetic vaccine design.
      ]. Meanwhile, proteasome-processing scores evaluate the generation of peptides with corresponding C-terminus. TAP transport score, on the other hand, calculates the generation of peptide or N-terminally extended precursors. Both the proteasome-processing and TAP transport scores are predictors that were tested using TOC curves, which utilizes type positive and false positive predictions for every threshold and an area under the ROC curve is measured for the predictor's performance. Hence, a random predictor obtains a score of 0.5 and a perfect predictor would score >1 [
      • Calis J.J.
      • Reinink P.
      • Keller C.
      • Kloetzel P.M.
      • Keşmir C.
      Role of peptide processing predictions in T cell epitope identification: contribution of different prediction programs.
      ].

      Identification of helper T-cell epitopes

      Helper T-lymphocytes orchestrate the activation of macrophages for phagocytosis, BL for antibody secretion, and CTL to eliminate virus-infected cells. This is achieved by interacting with MHC class II molecules on antigen presenting cells (APC). Unlike SLA class I, SLA class II displays strong homology to HLA class II and SLA II genes, DR and DQ are constitutively expressed on professional APC surfaces [
      • Techakriengkrai N
      • Nedumpun T
      • Golde WT
      • Suradhat S.
      Diversity of the swine leukocyte antigen class i and ii in commercial pig populations.
      ]. Physiological similarity between humans and pigs is the rationale the latter is considered as a biomedical model for toxicology, drug screening and heart and gut function [
      • Hammer SE
      • Ho CS
      • Ando A
      • Rogel-Gaillard C
      • Charles M
      • Tector M
      • Tector AJ
      • Lunney JK.
      Importance of the major histocompatibility complex (swine leukocyte antigen) in swine health and biomedical research.
      ]. For this reason, common HLA class II genes were used in the study. The epitopes and their binding affinity with MHC II alleles were determined through IEDB recommended 2.2 found in the IEDB server. This method has an improved performance in peptide binding prediction since it was based on an extended data set of quantitative MHC–peptide binding affinity [
      • Jensen KK
      • Andreatta M
      • Marcatili P
      • et al.
      Improved methods for predicting peptide binding affinity to MHC class II molecules.
      ]. The peptides binding to MHC II are not constrained in size as reported by the works of Yassai et al., and Nelson et al. and as such can vary from 11–30 amino acids long according to Rammensee et al. [
      • Yassai M.
      • Afsari A.
      • Garlie J.
      • Gorski J.
      C-terminal anchoring of a peptide to class II MHC via the P10 residue is compatible with a peptide bulge.
      ,
      • Nelson C.A.
      • Fremont D.H.
      Structural principles of MHC class II antigen presentation.
      ,
      • Rammensee H.G.
      • Friede T.
      • Stevanović S.
      MHC ligands and peptide motifs: first listing.
      ]. For this reason the authors chose to set the parameters to the lowest amino acid parameter to sample the virion peptides in the smallest possible length to detect short binding fragments that are capable of binding with MHC II. The method of sampling the smallest peptide is then compensated by the additional step of peptide conjugation of screened peptides based on IC50 values of less than 500 nm. Effectively, this means that the HTL epitopes that are conjugated are not merely a series of 11 epitopes combined together, but a compilation of Linear HTL epitopes present in each virion part considered.

      Vaccine evaluation

      Aligned with the aim of reducing the tedious process of screening antigens, Vaxijen, an alignment-independent tool applying antigen prediction based on auto-cross covariance transformation approach was utilized. It relies on the chemical properties of the primary structure of antigens [
      • Doytchinova I.A.
      • Flower D.R.
      Identifying candidate subunit vaccines using an alignment-independent method based on principal amino acid properties.
      ]. Each ASFV epitopes obtained and the ASF vaccine peptide were screened for their antigenicity scores. The vaccine yielded a score of 0.5344 which means that is indeed antigenic, signifying the ability of the vaccine to be recognized by antibody binding sites.
      Although considered rare, the possibility of an allergic reaction from vaccinations has been a major health concern [
      • Wood RA
      • Berger M
      • Dreskin SC
      • Setse R
      • Engler RJ
      • Dekker CL
      • Halsey NA.
      An algorithm for treatment of patients with hypersensitivity reactions after vaccines.
      ]. The culprit for hypersensitivity reactions are often associated with the components of the vaccine which can be potential triggers [
      • Chung EH.
      Vaccine allergies.
      ]. Hence, extra care should be done especially with peptide-based vaccines because it could be potentially allergenic. To assess the allergenicity of epitope candidates and the designed vaccine, the study used AllergenFP, an alignment-independent prediction tool which utilizes amino acid descriptors and novel descriptor fingerprints approach based on Tanimoto coefficient similarity search to predict allergenicity [
      • Dimitrov I.
      • Naneva L.
      • Doytchinova I.
      • Bangov I.
      AllergenFP: allergenicity prediction by descriptor fingerprints.
      ]. The designed vaccine scored 0.82 Tanimoto similarity index, signifying that it is non-allergenic and may prevent unwanted hypersensitivity reactions.
      Vaccines have a potential to induce an autoimmunity as a consequence of cross-reactivity [
      • Goodman RE.
      Practical and predictive bioinformatics methods for the identification of potentially cross-reactive protein matches.
      ,
      • Scheurer S
      • Son DY
      • Boehm M
      • Karamloo F
      • Franke S
      • Hoffmann A
      • Haustein D
      • Vieths S.
      Cross-reactivity and epitope analysis of Pru a 1, the major cherry allergen.
      ]. Local alignment algorithms, like FASTA or BLASTp, were used and the results obtained provided the degree of homology between proteins which can be cross-reactive. Screening for the similarity of the predicted epitopes and the ASF Vaccine to swine proteins revealed no significant similarity, thereby ruling out the possibility of cross-reactivity. The GC-content of the improved vaccine codon sequence has been increased, hence with codon optimization, the ASFV multi-epitope vaccine may be highly expressed in E. coli K12 host. The physicochemical properties and in silico cloning optimization are useful later, in vaccine production.
      To probe the possible utility of ASF vaccine in vitro and in vivo, a molecular docking simulation was done to determine if the predicted epitope regions will interact with at least one of its natural receptors such as SLA-1. The 3QQ3 crystal structure used was captured via complexation of SLA-1 0401 with peptides derived from swine-origin influenza A virus which gives a representation of the structural basis of peptide presentation by SLA-1. Cleaning of the PDB file was done on the 3QQ3 crystal structure to ensure that the target protein alone would be used for docking [
      • Kozakov D
      • Hall DR
      • Xia B
      • Porter KA
      • Padhorny D
      • Yueh C
      • Beglov D
      • Vajda S.
      The ClusPro web server for protein-protein docking.
      ]. Default parameters were used in protein-protein docking mode in ClusPro. This tool is a widely used protein-protein docking tool that utilizes Piper, an FFT or Fast Fourier Transform rigid docking program. Essentially, given protein pdb structures it outputs 1000 lowest energy structural confirmation out of 109 possible outcomes that underwent rotations and translations throughout. The clustering of the 1000 ligand positions are determined via the number of neighbors of a ligand in a 9 Angstroms radius from its C-alpha. The exploration of the vaccine epitope-SLA-1 interaction is necessary in the induction of cytolysis of monocytes and macrophages which are preferably infected by ASFV. This demonstrated favorable complex formation which validates the predicted binding region of ASF vaccine and is comparable to epitope-SLA binding results in literature. In future studies, the possible impact of solvents, metabolites, steric effects and allosteric effects on epitope-SLA-1 binding can be explored further, since this study only demonstrated a preliminary study of the possibility of the vaccine epitope-SLA-1 binding [
      • Herrera LR
      • Bisa EP.
      In silico analysis of highly conserved cytotoxic T-cell epitopes in the structural proteins of African swine fever virus.
      ].
      Since swine immune simulations are not freely available, CIMM Server was used. This tool allows the simulation of cellular and humoral responses of the human immune system hence, this tool could provide an overview of the immune response elicited by a mammalian immune system [
      • Rapin N
      • Lund O
      • Bernaschi M
      • Castiglione F.
      Computational immunology meets bioinformatics: the use of prediction tools for molecular binding in the simulation of the immune system.
      ]. Immune simulations showed that the constructed vaccine was able to elicit primary and secondary immune responses and produce memory B cells and antibodies such as IgM and IgG subclasses which are important against infection. It also yielded a slightly higher HTL memory cell population compared to the adjuvant as the antigen (vaccine) concentration increases. This implies that the adjuvant alone can initiate an increased immunity, and when combined with the all epitopes as a vaccine it can induce a lasting immune response by increasing HTL memory cell populations. Lastly, immune simulations revealed that the constructed vaccine and the adjuvant alone were able to stimulate production of cytotoxic T-cells and induce interferon-gamma production and other related cytokines, demonstrating the immunogenicity of ASF Vaccine.
      This study also attempted to do cross protection screening; however, it was deemed not feasible as ASFV was the sole DNA virus infecting swines. Since ASF immune response is strain-specific, consensus sequences from different ASFV genotypes were used to hypothetically provide cross protection among ASFV genotypes. Utilizing open-access bioinformatics tools poses limitations as to epitope prediction, specifically in HTL epitope prediction where we utilized human orthologs of HLA Class II. Nevertheless, utilizing several in silico tools revealed promising results which warrants ASF vaccine for further in vitro and in vivo studies.

      Conclusion

      The designed multi-epitope and multi-antigenic ASF vaccine constructed using predicted B-cell, Helper T-cell, and Cytotoxic T-cell using 6 protective antigens (p12, p17, CD2v, p22, p54, and p72 was shown to be antigenic, immunogenic, non-allergenic and non-cross-reactive which are crucial parameters in designing a vaccine. The ASF vaccine was also recognized by SLA-1 preferentially at the CTL region, which is an essential step in the induction of a cytotoxic immune response against ASFV-infected cells. Immune simulations also showed that ASF vaccine was able to induce pertinent immune response against ASF infection. ASFV sequences were obtained globally and could potentially induce cross protection across ASFV genotypes. Utilizing several in silico tools revealed promising results which can be used as the first step in designing effective and safe ASF vaccines. Hence, the developed multi-epitope and multi-antigenic ASF vaccine warrants further screening in vitro and in vivo to confirm its protective function against ASF infection.

      Funding

      This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

      CRediT authorship contribution statement

      Ara Karizza G. Buan: Visualization, Data curation, Formal analysis, Writing – original draft. Nico Alexander L. Reyes: Visualization, Data curation, Formal analysis, Writing – original draft. Ryan Nikkole B. Pineda: Visualization, Data curation, Formal analysis, Writing – original draft. Paul Mark B. Medina: Visualization, Writing – original draft, Supervision.

      Data Availability

      • The data used for the research were obtained from publicly-available protein sequence databases. The bioinformatics tools for analyses were also accessed free of charge.

      Declaration of Competing Interest

      The authors declare that there has not been any conflict of interest associated with the work and no significant financial support that could influence the results of the study.

      Acknowledgements

      We also would like to show our appreciation to Dr. Leana Rich M. Herrera for sharing her published research work with us on which this research was adapted.

      References

        • Gaudreault NN
        • Madden DW
        • Wilson WC
        • Trujillo JD
        • Richt JA.
        African swine fever virus: an emerging DNA arbovirus.
        Front Vet Sci. 2020; 7: 215https://doi.org/10.3389/fvets.2020.00215
        • Revilla Y
        • Pérez-Núñez D
        • Richt J.A.
        African swine fever virus biology and vaccine approaches.
        Adv Virus Res. 2018; : 41-74https://doi.org/10.1016/bs.aivir.2017.10.002
        • Blome S
        • Franzke K
        • Beer M.
        African swine fever – a review of current knowledge.
        Virus Res. 2020; : 287https://doi.org/10.1016/j.virusres.2020.198099
        • Ros-Lucas A
        • Correa-Fiz F
        • Bosch-Camós L
        • Rodriguez F
        • Alonso-Padilla J
        Computational analysis of african swine fever virus protein space for the design of an epitope-based vaccine ensemble.
        Pathogens. 2020; 9: 1078https://doi.org/10.3390/pathogens9121078
        • Liu Q
        • Ma B
        • Qian N
        • Zhang F
        • Tan X
        • Lei J
        • Xiang Y.
        Structure of the african swine fever virus major capsid protein p72.
        Cell Res. 2019; 29: 953-955https://doi.org/10.1038/s41422-019-0232-x
        • Alejo A
        • Matamoros T
        • Guerra M
        • Andrés G.
        A proteomic atlas of the african swine fever virus particle.
        J Virol. 2018; 92 (-18): e01293https://doi.org/10.1128/jvi.01293-18
      1. Kumagai N. African swine fever in Asia and the Pacific. Presented at: African swine fever: An unprecedented global threat - A challenge to livelihoods, food security and biodiversity. Call for action; October 26-30, 2020; Online Webinar. https://www.fao.org/3/cb2145en/cb2145en.pdf.

      2. Weaver T, & Habib N (Asian Development Bank). Evaluating losses associated with african swine fever in the People's Republic Of China and neighboring countries. 2021.https://www.adb.org/sites/default/files/publication/680961/eawp-27-losses-african-swine-fever-prc-neighboring-countries.pdf.

        • Madeira F.
        • Pearce M.
        • Tivey A.R.N.
        • Basutkar P.
        • Lee J.
        • Edbali O.
        • Madhusoodanan N.
        • Kolesnikov A.
        • Lopez R.
        Search and sequence analysis tools services from EMBL-EBI in 2022.
        Nucleic Acids Res. 2022; 50: W276-W279https://doi.org/10.1093/nar/gkac240
        • Garcia-Boronat M.
        • Diez-Rivero C.M.
        • Reinherz E.L.
        • Reche P.A.
        PVS: a web server for protein sequence variability analysis tuned to facilitate conserved epitope discovery.
        Nucleic Acids Res. 2008; 36: W35-W41https://doi.org/10.1093/nar/gkn211
        • Emini E.A.
        • Hughes J.V.
        • Perlow D.S.
        • Boger J.
        Induction of hepatitis A virus-neutralizing antibody by a virus-specific synthetic peptide.
        J Virol. 1985; 55: 836-839https://doi.org/10.1128/JVI.55.3.836-839.1985
        • Jespersen M.C.
        • Peters B.
        • Nielsen M.
        • Marcatili P.
        BepiPred-2.0: improving sequence-based B-cell epitope prediction using conformational epitopes.
        Nucleic Acids Res. 2017; 45: W24-W29https://doi.org/10.1093/nar/gkx34
        • Kolaskar A.S.
        • Tongaonkar P.C.
        A semi-empirical method for prediction of antigenic determinants on protein antigens.
        FEBS Lett. 1990; 276: 172-174https://doi.org/10.1016/0014-5793(90)80535-q
        • Saha S.
        • Raghava G.P.
        Prediction of continuous B-cell epitopes in an antigen using recurrent neural network.
        Proteins. 2006; 65: 40-48https://doi.org/10.1002/prot.21078
        • Tenzer S
        • Peters B
        • Bulik S
        • Schoor O
        • Lemmel C
        • Schatz MM
        • Kloetzel PM
        • Rammensee HG
        • Schild H
        • Holzhutter HG.
        Modeling the MHC class I pathway by combining predictions of proteasomal cleavage, TAP transport and MHC class I binding.
        Cell Mol Life Sci. 2005; 62: 1025-1037
        • Doytchinova I.A.
        • Flower D.R.
        Identifying candidate subunit vaccines using an alignment-independent method based on principal amino acid properties.
        Vaccine. 2007; 25: 856-866https://doi.org/10.1016/j.vaccine.2006.09.032
        • Doytchinova I.A.
        • Flower D.R.
        VaxiJen: a server for prediction of protective antigens, tumour antigens and subunit vaccines.
        BMC Bioinform. 2007; 8: 4https://doi.org/10.1186/1471-2105-8-4
        • Dimitrov I.
        • Naneva L.
        • Doytchinova I.
        • Bangov I.
        AllergenFP: allergenicity prediction by descriptor fingerprints.
        Bioinformatics. 2014; 30 (Oxford, England): 846-851https://doi.org/10.1093/bioinformatics/btt619
        • Gasteiger E
        • Hoogland C
        • Gattiker A
        • Duvaud S
        • Wilkins MR
        • Appel RD
        • Bairoch A.
        • Walker J.M.
        Protein identification and analysis tools on the ExPASy server.
        The proteomics protocols handbook. Humana Press, 2005: 571-607
        • Linding R.
        GlobPlot: exploring protein sequences for globularity and disorder.
        Nucleic Acids Res. 2003; 31: 3701-3708https://doi.org/10.1093/nar/gkg519
        • Ko J.
        • Park H.
        • Heo L.
        • Seok C.
        GalaxyWEB server for protein structure prediction and refinement.
        Nucleic Acids Res. 2012; 40: W294-W297https://doi.org/10.1093/nar/gks493
        • Colovos C.
        • Yeates T.O.
        Verification of protein structures: patterns of nonbonded atomic interactions.
        Protein Sci. 1993; 2: 1511-1519https://doi.org/10.1002/pro.5560020916
        • Lüthy R.
        • Bowie J.U.
        • Eisenberg D.
        Assessment of protein models with three-dimensional profiles.
        Nature. 1992; 356: 83-85https://doi.org/10.1038/356083a0
        • Pettersen E.F.
        • Goddard T.D.
        • Huang C.C.
        • Couch G.S.
        • Greenblatt D.M.
        • Meng E.C.
        • Ferrin T.E.
        UCSF chimera-a visualization system for exploratory research and analysis.
        J Comput Chem. 2014; 25: 1605-1612https://doi.org/10.1002/jcc.20084
        • Kozakov D
        • Hall DR
        • Xia B
        • Porter KA
        • Padhorny D
        • Yueh C
        • Beglov D
        • Vajda S.
        The ClusPro web server for protein-protein docking.
        Nat Protoc. 2017; 12: 255-278https://doi.org/10.1038/nprot.2016.169
        • Zhang N
        • Qi J
        • Feng S
        • Gao F
        • Liu J
        • Pan X
        • Chen R
        • Li Q
        • Chen Z
        • Li X
        • Xia C
        • Gao GF.
        Crystal structure of swine major histocompatibility complex class I SLA-1*0401 and identification of 2009 pandemic swine-origin influenza A H1N1 virus cytotoxic T lymphocyte epitope peptides.
        J Virol. 2011; 85: 11709-11724https://doi.org/10.1128/jvi.05040-11
        • Xue L.C.
        • Rodrigues J.P.
        • Kastritis P.L.
        • Bonvin A.M.
        • Vangone A.
        PRODIGY: a web server for predicting the binding affinity of protein–protein complexes.
        Bioinformatics. 2016; 32 (2016): 3676-3678https://doi.org/10.1093/bioinformatics/btw514
        • Rapin N
        • Lund O
        • Bernaschi M
        • Castiglione F.
        Computational immunology meets bioinformatics: the use of prediction tools for molecular binding in the simulation of the immune system.
        PloS One. 2010; 5: e9862https://doi.org/10.1371/journal.pone.0009862
        • Grote A.
        • Hiller K.
        • Scheer M.
        • Munch R.
        • Nortemann B.
        • Hempel D.C.
        • Jahn D.
        JCat: a novel tool to adapt codon usage of a target gene to its potential expression host.
        Nucleic Acids Res. 2005; 33 (2005): W526-W531https://doi.org/10.1093/nar/gki376
        • Bitsouni V
        • Lycett S
        • Opriessnig T
        • Doeschl-Wilson A.
        Predicting vaccine effectiveness in livestock populations: a theoretical framework applied to prrs virus infections in pigs.
        PLoS One. 2019; 14e0220738https://doi.org/10.1371/journal.pone.0220738
        • Meeusen EN
        • Walker J
        • Peters A
        • Pastoret PP
        • Jungersen G.
        Current status of veterinary vaccines.
        Clin Microbiol Rev. 2007; 20: 489-510https://doi.org/10.1128/CMR.00005-07
        • Oscherwitz J.
        The promise and challenge of epitope-focused vaccines.
        Hum Vaccin Immunother. 2016; 12: 2113-2116https://doi.org/10.1080/21645515.2016.1160977
        • Blome S.
        • Gabriel C.
        • Beer M.
        Modern adjuvants do not enhance the efficacy of an inactivated African swine fever virus vaccine preparation.
        Vaccine. 2014; 32: 3879-3882https://doi.org/10.1016/j.vaccine.2014.05.051
        • Cadenas-Fernández E.
        • Sánchez-Vizcaíno J.M.
        • van den Born E.
        • Kosowska A.
        • van Kilsdonk E.
        • Fernández-Pacheco P.
        • Gallardo C.
        • Arias M.
        • Barasona J.A.
        High doses of inactivated African Swine fever virus are safe, but do not confer protection against a virulent challenge.
        Vaccines. 2021; 9: 242https://doi.org/10.3390/vaccines9030242
        • Argilaguet J.M.
        • Pérez-Martín E.
        • Nofrarías M.
        • Gallardo C.
        • Accensi F.
        • Lacasta A.
        • Mora M.
        • Ballester M.
        • Galindo-Cardiel I.
        • López-Soria S.
        • Escribano J.M.
        • Reche P.A.
        • Rodríguez F.
        DNA vaccination partially protects against African swine fever virus lethal challenge in the absence of antibodies.
        PloS One. 2012; 7: e40942https://doi.org/10.1371/journal.pone.0040942
        • Barderas M.G.
        • Rodríguez F.
        • Gómez-Puertas P.
        • Avilés M.
        • Beitia F.
        • Alonso C.
        • Escribano J.M.
        Antigenic and immunogenic properties of a chimera of two immunodominant African swine fever virus proteins.
        Arch Viro. 2001; 146: 1681-1691https://doi.org/10.1007/s007050170056
        • Gómez-Puertas P.
        • Rodríguez F.
        • Oviedo J.M.
        • Brun A.
        • Alonso C.
        • Escribano J.M.
        The African swine fever virus proteins p54 and p30 are involved in two distinct steps of virus attachment and both contribute to the antibody-mediated protective immune response.
        Virology. 1998; 243: 461-471https://doi.org/10.1006/viro.1998.9068
        • Neilan J.G.
        • Zsak L.
        • Lu Z.
        • Burrage T.G.
        • Kutish G.F.
        • Rock D.L.
        Neutralizing antibodies to African swine fever virus proteins p30, p54, and p72 are not sufficient for antibody-mediated protection.
        Virology. 2004; 319: 337-342https://doi.org/10.1016/j.virol.2003.11.011
        • Rock DL.
        Thoughts on african swine fever vaccines.
        Viruses. 2021; 13https://doi.org/10.3390/v13050943
        • Litwin S.
        • Jores R.
        • Perelson A.S.
        • Weisbuch G.
        Shannon information as a measure of amino acid diversity.
        Theoretical and experimental insights into immunology. 66. NATO ASI Series, 1992https://doi.org/10.1007/978-3-642-76977-1_17
        • Zhang L.
        Multi-epitope vaccines: a promising strategy against tumors and viral infections.
        Cell Mol Immunol. 2017; 15: 182-184https://doi.org/10.1038/cmi.2017.92
        • Bosch-Camós L
        • López E
        • Rodriguez F.
        African swine fever vaccines: a promising work still in progress.
        Porc Health Manag. 2020; 6https://doi.org/10.1186/s40813-020-00154-2
        • Wu K
        • Liu J
        • Wang L
        • Fan S
        • Li Z
        • Li Y
        • Yi L
        • Ding H
        • Zhao M
        • Chen J.
        Current state of global african swine fever vaccine development under the prevalence and transmission of ASF in China.
        Vaccines. 2020; 8: 531https://doi.org/10.3390/vaccines8030531
        • Jancovich JK
        • Chapman D
        • Hansen DT
        • Robida MD.
        • Loskutov A
        • Craciunescu F
        • Borovkov A
        • Kibler K
        • Goatley L
        • King K
        • Netherton CL
        • Taylor G
        • Jacobs B
        • Sykes K
        • Dixon LK.
        Immunization of oigs by DNA prime and recombinant vaccinia virus boost to identify and rank african swine fever virus immunogenic and protective proteins.
        J Virol. 2018; 92https://doi.org/10.1128/jvi.02219-17
        • Althuwaiqeb SA
        • Bordoni B.
        Histology, B cell lymphocyte.
        StatPearls, Treasure Island FL2021 (StatPearls; 2021. Accessed November 2021)
        • Hoffman W
        • Lakkis FG
        • Chalasani G.
        B Cells, antibodies, and more.
        Clin J Am Soc Nephrol. 2016; 11: 137-154https://doi.org/10.2215/CJN.09430915
        • Zsak L
        • Onisk DV
        • Afonso CL
        • Rock DL.
        Virulent african swine fever virus isolates are neutralized by swine immune serum and by monoclonal antibodies recognizing a 72-kDa viral protein.
        Virology. 1993; 196: 596-602https://doi.org/10.1006/viro.1993.1515
        • Takamatsu H
        • Denyer MS
        • Lacasta A
        • Stirling CMA
        • Argilaguet JM
        • Netherton CL
        • Oura CAL
        • Martins C
        • Rodríguez F.
        Cellular immunity in ASFV responses.
        Virus Res. 2013; 173: 110-121https://doi.org/10.1016/j.virusres.2012.11.009
        • Rock DL.
        Challenges for African swine fever vaccine development-"… perhaps the end of the beginning.".
        Vet Microbiol. 2017; 206: 52-58https://doi.org/10.1016/j.vetmic.2016.10.003
        • Denyer MS
        • Wileman TE
        • Stirling CMA.
        • Zuber B
        • Takamatsu H
        Perforin expression can define CD8 positive lymphocyte subsets in pigs allowing phenotypic and functional analysis of natural killer, cytotoxic T, natural killer T and MHC un-restricted cytotoxic T-cells.
        Vet Immunol Immunopathol. 2006; 110: 279-292https://doi.org/10.1016/j.vetimm.2005.10.005
        • Oura CAL
        • Denyer MS
        • Takamatsu H
        • Parkhouse RME.
        In vivo depletion of CD8+ T lymphocytes abrogates protective immunity to african swine fever virus.
        J Gen Virol. 2005; 86: 2445-2450https://doi.org/10.1099/vir.0.81038-0
        • Fleri W.
        • Paul S.
        • Dhanda S.K.
        • Mahajan S.
        • Xu X.
        • Peters B.
        • Sette A.
        The immune epitope database and analysis resource in epitope discovery and synthetic vaccine design.
        Front Immunol. 2017; 8https://doi.org/10.3389/fimmu.2017.00278
        • Pedersen L.E.
        • Harndahl M.
        • Rasmussen M.
        • Lamberth K.
        • Golde W.T.
        • Lund O.
        • Nielsen M.
        • Buus S.
        Porcine major histocompatibility complex (MHC) class I molecules and analysis of their peptide-binding specificities.
        Immunogenetics. 2011; 63: 821-834https://doi.org/10.1007/s00251-011-0555-3
        • Nevagi R.J.
        • Toth I.
        • Skwarczynski M.
        • Koutsopoulos S.
        Peptide-based vaccines.
        Peptide applications in biomedicine, biotechnology and bioengineering. Woodhead Publishing, 2018: 327-358
        • Calis J.J.
        • Reinink P.
        • Keller C.
        • Kloetzel P.M.
        • Keşmir C.
        Role of peptide processing predictions in T cell epitope identification: contribution of different prediction programs.
        Immunogenetics. 2015; 67: 85-93https://doi.org/10.1007/s00251-014-0815-0
        • Techakriengkrai N
        • Nedumpun T
        • Golde WT
        • Suradhat S.
        Diversity of the swine leukocyte antigen class i and ii in commercial pig populations.
        Front Vet Sci. 2021; 8637682https://doi.org/10.3389/fvets.2021.637682
        • Hammer SE
        • Ho CS
        • Ando A
        • Rogel-Gaillard C
        • Charles M
        • Tector M
        • Tector AJ
        • Lunney JK.
        Importance of the major histocompatibility complex (swine leukocyte antigen) in swine health and biomedical research.
        Annu Rev Anim Biosci. 2020; 8: 171-198https://doi.org/10.1146/annurev-animal-020518-115014
        • Jensen KK
        • Andreatta M
        • Marcatili P
        • et al.
        Improved methods for predicting peptide binding affinity to MHC class II molecules.
        Immunology. 2018; 154: 394-406https://doi.org/10.1111/imm.12889
        • Yassai M.
        • Afsari A.
        • Garlie J.
        • Gorski J.
        C-terminal anchoring of a peptide to class II MHC via the P10 residue is compatible with a peptide bulge.
        J Immunol. 2002; 168: 1281-1285
        • Nelson C.A.
        • Fremont D.H.
        Structural principles of MHC class II antigen presentation.
        Rev Immunogenet. 1999; 1: 47-59https://doi.org/10.4049/jimmunol.168.3.1281
        • Rammensee H.G.
        • Friede T.
        • Stevanović S.
        MHC ligands and peptide motifs: first listing.
        Immunogenetics. 1995; 41: 178-228
        • Wood RA
        • Berger M
        • Dreskin SC
        • Setse R
        • Engler RJ
        • Dekker CL
        • Halsey NA.
        An algorithm for treatment of patients with hypersensitivity reactions after vaccines.
        Pediatrics. 2008; 122: e771-e777https://doi.org/10.1542/peds.2008-1002
        • Chung EH.
        Vaccine allergies.
        Clin Exp Vaccine Res. 2014; 3: 50-57https://doi.org/10.7774/cevr.2014.3.1.50
        • Goodman RE.
        Practical and predictive bioinformatics methods for the identification of potentially cross-reactive protein matches.
        Mol Nutr Food Res. 2006; 50: 655-660https://doi.org/10.1002/mnfr.200500277
        • Scheurer S
        • Son DY
        • Boehm M
        • Karamloo F
        • Franke S
        • Hoffmann A
        • Haustein D
        • Vieths S.
        Cross-reactivity and epitope analysis of Pru a 1, the major cherry allergen.
        Mol Immunol. 1999; 36: 155-167https://doi.org/10.1016/s0161-5890(99)00033-4
        • Herrera LR
        • Bisa EP.
        In silico analysis of highly conserved cytotoxic T-cell epitopes in the structural proteins of African swine fever virus.
        Vet World. 2021; 14: 2625-2633https://doi.org/10.14202/vetworld.2021.2625-2633