Design of a potential Sema4A-based multi-epitope vaccine to combat triple-negative breast cancer: an immunoinformatic approach

Abstract

Immunotherapy is revamping the therapeutic strategies for TNBC owing to its higher mutational burden and tumour-associated antigens. One of the most intriguing developments in cancer immunotherapy is the focus on peptide-based cancer vaccines. Thus, the current work aims to develop an efficient peptide-based vaccine against TNBC that targets Sema4A, which has recently been identified as a major regulator of TNBC progression. Initially, the antigenic peptides derived from Sema4A were determined and evaluated based on their capability to provoke immunological responses. The assessed epitopes were then linked with a suitable adjuvant (RpfB and RpfE) and appropriate linkers (AAY, GPGPG, KK and EAAAK) to preclude junctional immunogenicity. Eventually, docking and dynamics simulations are performed against TLR-2, TLR-4, TLR-7 and TLR-9 to assess the interaction between the vaccine construct and TLR receptors, as the TLR signalling pathway is critical in the host immune response. The developed vaccine was then exposed to in silico cloning and immune simulation analysis. The findings suggest that the designed vaccine could potentially evoke significant humoral and cellular immune responses in the intended organism. Considering these outcomes, the final multi-epitope vaccine could be employed to serve as an effective choice for TNBC management and may open new avenues for further studies.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12032-023-01970-6.

Introduction

Triple-negative breast cancer (TNBC), which accounts for 15–20% of breast cancer, is an aggressive subtype with a 40% fatality rate. The deficiency of the oestrogen receptor (ER), progesterone receptor (PR) and human epidermal growth factor receptor 2 (HER2) increases the likelihood of recurrence, disease progression and a worse prognosis [1]. Owing to the lack of these receptors, TNBC is insensitive to hormonal or targeted therapies and is often treated with chemotherapy and surgical excision of the original tumour. Apparently, conventional chemotherapy drugs, such as paclitaxel, anthracycline and alkylating agents, can cause systemic toxicity and adverse effects [2]. Besides the typical heterogeneity and aggressiveness of TNBC, its poor prognosis also lies in the scarcity of predictive biomarkers of treatment response [3]. Indeed, this unmet clinical necessity has encouraged the development of immunotherapy in the field of oncology.

Immunotherapy has emerged as a remarkable emphasis in the treatment of cancer, as it provides insights for prolonging the overall survival (OS) of patients and enhancing their quality of life. The other therapies include immune checkpoint inhibitors, adoptive T-cell therapy and dendritic cell vaccines therapeutics have emerged as an important alternative used in the treatment of TNBC. It has been reported that TNBC shows a higher response to immunotherapy because of its immunogenic TME landscape [4]. Evidently, higher proportion of tumour infiltrating lymphocytes (TILs), high tumour mutational load and survival associations with degree of T-cell and B-cell infiltration, is making immunotherapy a promising option against TNBC [5, 6]. Among the anticancer immunotherapies, the multi-epitope peptide cancer vaccines appear as the next generation treatment strategy designed to elicit the immune responses against tumour cells. Either by intensifying the CD4 + helper T lymphocytes (HTLs) and CD8 + cytotoxic T lymphocytes (CTLs) responses or by hampering the immune response suppressors, the anti-tumour immunity activity may be accelerated. Therefore, it is pivotal to construct a novel therapeutic cancer vaccine to foster cellular and humoral immune responses [7].

The major advantage of peptide-based vaccines is that their development is both safe and inexpensive in comparison to conventional techniques of vaccine production [8, 9]. Interestingly, PVX-410, P10s-PADRE and TPIV200 are some of the chimeric peptide-based vaccines for TNBC that are currently in clinical phase trials. The PVX-410 is an HLA-A2-restricted vaccine for patients who overexpress X-box-binding protein 1, with two splice variants, syndecan-1 (CD138) and cell surface glycoprotein SLAM family member 7 (SLAMF7 or CD319) antigens. Similarly, P10s-PADRE, carbohydrate mimetic peptide P10s, is injected prior to chemotherapy and TPIV200 is a penta-epitope vaccine, for patients who overexpress folate receptor alpha (FR α) antigens. These vaccines can only be administered to patients who overexpress these antigens [3, 10]. However, the main limitation of this multi-epitope vaccine is that it can only be administered to patients as a personalised treatment strategy. Thus, vaccine design using novel overexpressed proteins is of immense contribution to the scientific literature towards better treatment. For instance, novel peptide-based vaccines for cancer-testis antigens were developed by Mahdevar et al [11]and Parvizpour et al [12] for the management of TNBC [11, 12]. In another study, Rajendran et al. designed a novel SOX9-based multi-epitope vaccine to battle metastatic triple-negative breast cancer. However, the main limitation of this multi-epitope vaccine is that it can only be administered to patients who have SOX9 overexpression [13].

Cancer immunotherapy can easily target cell surface proteins as they are easily accessible, but finding effective targets is still challenging [14]. Literature studies indicate that Sema4A, acting as a biomarker for dendritic cell activation status and CTL stimulator, promotes the production of IFN-γ and inhibits tumour growth [15, 16]. Exogenous Sema4A, on the other hand, has been shown to protect TNBC cells from hypoxia-induced cytotoxicity, inhibit cell apoptosis and promote cell proliferation [17]. Collectively, these results shed light on the potential use of Sema4A in various fields of cancer research, including vaccine development [18]. Therefore, the current work attempts to design a subunit vaccine targeting Sema4A by applying immunoinformatics approaches in a sequential manner for the treatment and management of TNBC patients.

Methodology

Retrieval of protein and adjuvant sequences

The amino acid sequence of the target protein, Sema4A (Accession No. Q9H3S1) and the adjuvants, RpfB (Accession No. P9WG29) and RpfE (Accession No. O53177), were retrieved from the UniProt database in FASTA format for our analysis [19].

Identification and characterisation of B-cell epitopes and peptides binding to MHC class I and class II alleles

Prediction of cytotoxic and helper T-cell epitopes

The NetMHCpan 4.0 server was used to predict potential CTL-specific epitopes. Using artificial neural networks (ANN), this tool predicts the binding of peptides to any known MHC molecule [20]. In this study, the epitopes were predicted with a peptide length of 9 mer, as the open-binding groove of MHC I protein typically accommodates peptides of 8–11 residues. Of note, twelve major HLA supertypes (A1, A2, A3, A24, A26, B7, B8, B27, B39, B44, B58, and B62) were considered during the epitope prediction in our analysis. Further, NetMHCpan 4.0 achieves its maximum sensitivity when the thresholds for strong and weak binders are set to 0.5 and 2%, respectively. In a similar fashion, the NetMHCIIpan-4.0 server was utilised to scan the HTL-specific epitopes that correspond to HLA-DR, HLA-DQ and HLA-DP MHC class II alleles using ANN algorithm [21]. For this study, the epitopes were predicted using 15-mer peptide length, and the cutoffs for strong and weak binders were set at 1 and 5%, respectively.

Prediction of B-cell epitope

B-cell epitopes include linear and conformational epitopes. The ABCpred server predicts the linear B-cell epitopes with a prediction accuracy of 65.93%. For this analysis, a threshold of 0.51 and window length of 16 were employed, utilising a feed forward and recurrent neural network algorithm [22]. Based on the three-dimensional (3D) structures of the vaccine construct, discontinuous B-cell epitopes were predicted using the ElliPro server on the IEDB interface with default parameters (a minimum residue score of 0.5 and maximum distance of 6) [23]. The residue clustering algorithm and Tornton’s method are applied on this server. Each output epitope receives a score from ElliPro, which is expressed as the protrusion index (PI) value averaged over each epitope residue. The calculation is mainly based on the protein regions protruding from the protein’s globular surface. The regions protruding from the globular surface of the protein are more available for interaction with an antibody. Thus, PI score is directly proportional to solvent accessibility and plays an important role in the structure and function of biological macromolecules.

Screening of predicted epitopes

The predicted epitopes were screened for antigenicity, toxicity, allergenicity and immunogenicity. The antigenicity of the epitopes was predicted using the VaxiJen v2.0 server, which has a prediction accuracy between 70 and 89% [24]. In this analysis, the tumour was selected as the target organism with 0.5 as the threshold. Epitope toxicity was predicted using the ToxinPred server. The toxicity evaluation in this study relied on a support vector machine (SVM)-based approach [25]. The allergenicity of the epitopes was evaluated using the AllerTOP v. 2.0 server. This server relies on the auto-cross-covariance (ACC) transformation and machine learning methods for classification. In particular, Auto-cross-covariance transformation transforms amino acid sequences into a fixed length vectors for analysis. The fixed length vectors are then analysed for its allergenicity using best performing kNN algorithm [26]. Using the IEDB immunogenicity prediction tool with default parameters, the immunogenicity of a candidate epitope was evaluated. Strong immunogenic peptides are more likely to be possible epitopes [27]. Finally, the IFNepitope, an SVM hybrid method-based server was employed to predict the IFN gamma-inducing property of the predicted HTL epitopes. It is trained and tested using 10,433 experimentally validated MHC class II binders from the Immune Epitope Database (IEDB) and classifies them as inducers and non-inducers. In addition, the maximum prediction accuracy of this server is found to be 82.10% [28].

Multi-epitope vaccine construction and evaluation of physicochemical properties

The screened epitopes were coupled with adjuvants and linkers to produce a complete vaccine construct. GPGPG linkers were used for HTL epitopes, AAY linkers for CTL epitopes and KK linkers for B-cell epitopes. The TLR-4 agonists RpfB and RpfE were chosen as an adjuvant to boost the immunogenicity and are linked to each end of the vaccine construct by the EAAAK linker [29]. The antigenicity and allergenicity of the vaccine construct were also assessed. In addition, the Expasy ProtParam server was utilised to predict various physicochemical parameters of the multi-epitope vaccine, including amino acid composition, theoretical pI, molecular weight, instability index, half-life, aliphatic index and the grand average of hydropathicity (GRAVY) [30].

Population coverage of selected T-cell epitopes

Diverse populations have different HLA genotype frequencies. The Immune Epitope Database (IEDB) was used to perform a combined population coverage analysis of HLA classes I and II in the global and Indian populations using default parameters [31]. T-cell epitopes and their respective HLA alleles were provided as input. The programme calculates predicted population coverage, the average number of epitope hits and HLA combinations recognised by populations and the minimum number of epitope combinations recognised by 90% of the population (PC90).

Secondary and tertiary structure prediction and validation

The secondary structure prediction of the designed vaccine was carried out using the PSIPRED 4.0 analysis workbench and the percentages of the helix, strand and coils were determined [32]. Further, the Robetta server was employed for the prediction of the tertiary structure of the vaccine construction [33]. The best predicted model was validated using the SAVES v6.0 server. This server utilises a wide variety of tools such as PROCHECK for visualising the Ramachandran plot, ERRAT for analysing non-bonded atom–atom interactions and Verify 3D for determining the compatibility of an atomic model (3D) [34–36]. The ProSA-web server evaluates the overall quality of the model and identifies any potential errors in the predicted structure [37].

Docking of a designed vaccine with Toll-like receptors

The model structure was uploaded to the ClusPro 2.0 server to investigate the interaction of the vaccine construct with the Toll-like receptors 2, 4, 7 and 9 (PDB IDs: 3A7C, 4G8A, 7CYN and 3WPF). This server adheres to rigid body docking, grouping of least energy structures and structural optimization steps for its docking approach [38]. The docked complex with the lowest energy score and highest binding efficiency was chosen and viewed in the PyMOL visualisation tool. It was further submitted to PDBsum server to investigate the binding residues between the vaccine construct and Toll-like receptors.

Molecular dynamic simulation of the docked complex

A molecular dynamics simulation of the receptor-vaccine construct was performed using GROMACS v2020.1 with the GROMOS96 54a7 force field. Protein topology was generated using the Pdb2gmx module. The complex structures were then put inside a dodecahedron box and solvated using simple point charge (SPC) water molecules using the GROMACS editconf tool. Counter ions were added using the genion tool for neutralising the system, and energy minimization was carried out prior to the MD run with a 10 kJ/mol force constant using the steepest descent algorithm. The LINCS algorithm was applied to retain the bond length, and the particle-mesh Eswald method was used to investigate the electrostatic interactions. The equilibration of the system was done using isothermal-isochoric (NVT) and isothermal-isobaric (NPT) steps. Then the Berendsen algorithm was applied to heat the system to 300 K the optimal physiological temperature of the host with a coupling time of 0.1 ps and a pressure of 1 bar. The complex was finally subjected to a 50 ns MD simulation with a recording interval of 2 fs. The trajectories were analysed using GROMACS utilities for root mean square deviation (RMSD), radius of gyration (Rg) and solvent accessible surface area (SASA) and plotted with xmgrace.

Codon optimization and in silico cloning of the vaccine construct

The peptide vaccine construct was back translated and optimised for expression in host Escherichia coli using the ExpOptimizer and GenScript Codon Optimization tools. This tool generates protein-coding DNA sequence and computes its organism-related properties, such as the codon adaptation index (CAI) and guanine/cytosine (GC) content. This evidence can be used to predict the translation efficiency of the vaccine in E. coli. A GC content of 30–70% and a CAI in the range of 0.8–1.0 are required for the noble expression of the vaccine. To simulate and visualise the in silico cloning, the SnapGene software was used. With the help of BamHI and PflFI restriction enzymes, the sequence of the chimeric protein was put into the pET-28a (+) plasmid. The key benefit of utilising pET expression system is the highly regulated transcription of the cloned protein [39].

Immune simulation profiling

A web-based simulation server, C-ImmSim, was utilised to characterise the effective immune response generated by the vaccine construct. This server uses a position-specific scoring matrix (PSSM) and machine learning to predict both cellular and humoral immune system components in mammals [40]. The simulation volume and simulation steps were set to 50 and 1050, respectively. The remaining parameters, such as the random seed, number of antigens to inject and vaccine proteins (no LPS), were left at their default settings. Three injections at an interval of 4 weeks (1, 84, and 168) were administered.

Results and discussion

Identification and characterisation of immunogenic regions of the selected protein

Cancer vaccines generally work based on the adaptive immunity deduced from CTL, HTL and B-cell responses [41]. Thus, predicting these epitopes to stimulate CD4 + or CD8 + T cells and B-cell immune responses is a crucial step in vaccine development [42]. In the present study, a total of 81 CTL, 303 HTL and 73 B-cell epitopes from the target protein was predicted using the servers. The epitopes were selected, which could bind to at least three MHC class I and II super types. In order to construct a safe and effective vaccine, epitopes must be antigenic, non-allergic, non-toxic and should possess IFN-γ inducing ability (for HTL only). So, the predicted epitopes were screened for its vaccine properties and finally 4 CTL (FNVIRHAVL, WPHFVTVTV, QQSYWPHFV, TAPHIYAVF) (Table 1), 4 HTL (EYTRLAVETAQGLDG, FNVIRHAVLLPADSP, NVIRHAVLLPADSPT, VEYTRLAVETAQGLD) (Table 2) and 4 B-cell (PVLKTDNFLRWLHHDA, KKKSNETQCFNFIRVL, SQDQTLALDPELAGIP, LAAQQSYWPHFVTVTV) promiscuous epitopes were selected for further study (Table 3).

Table 1.

MHC-Class I binding CTL epitopes of Sema4A and its immunogenic properties

CTL Epitopes	HLA allele type	Vaxijen score	Allergenicity	Toxicity
FNVIRHAVL	HLA-B0801	0.7288	Non allergen	Non-toxic
WPHFVTVTV	HLA-B0702, HLA-B3901	0.9582	Non allergen	Non-toxic
QQSYWPHFV	HLA-B3901	1.4537	Non allergen	Non-toxic
TAPHIYAVF	HLA-A2402	0.8982	Non allergen	Non-toxic

Table 2.

MHC-Class II binding HTL epitopes of Sema4A and its immunogenic properties

HTL Epitopes	HLA allele type	Immunogenic properties
		Antigenicity	Allergenicity	IFN-γ	Toxicity
EYTRLAVETAQGLDG	HLA-DRB11001, HLA-DPA10103, HLA-DPB10301	0.6026	Non allergen	Inducer	Non-toxic
FNVIRHAVLLPADSP	HLA-DRB50202, HLA-DPA10103, HLA-DPB11101, HLA-DQA10102, HLA-DQB10501, HLA-DQA10102, HLA-DQB10602, HLA-DQA10103, HLA-DQB10603, HLA-DQA10401, HLA-DQB10301, HLA-DQA10501, HLA-DQB10301, HLA-DQA10505, HLA-DQB10301	0.5224	Non allergen	Inducer	Non-toxic
NVIRHAVLLPADSPT	HLA-DPA10103, HLA-DPB11101, HLA-DQA10102, HLA-DQB10602, HLA-DQA10103, HLA-DQB10603, HLA-DQA10401, HLA-DQB10301, HLA-DQA10501, HLA-DQB10301, HLA-DQA10505, HLA-DQB10301	0.5037	Non allergen	Inducer	Non-toxic
VEYTRLAVETAQGLD	HLA-DRB1_1001, HLA-DRB1_1401, HLA-DPA10103, HLA-DPB10301	0.4105	Non allergen	Inducer	Non-toxic

Table 3.

Linear B-cell epitopes of Sema4A and its immunogenic properties

B-cell epitopes	Epitope score	Vaxijen score	Allergenicity	Toxicity
SQDQTLALDPELAGIP	0.59	0.8717	Non allergen	Non-toxic
PVLKTDNFLRWLHHDA	0.62	1.2293	Non allergen	Non-toxic
KKKSNETQCFNFIRVL	0.68	0.9126	Non allergen	Non-toxic
LAAQQSYWPHFVTVTV	0.75	0.8177	Non allergen	Non-toxic

Assembly of multi-epitope vaccine construct

The nature of the epitopes, adjuvants, linkers and their order and position in the construct are all crucial for multi-epitope vaccine construction [43]. Linkers not only increase the efficacy of epitope presentation but also provide the required separation between the epitopes. The adjuvants are added to boost the immunogenicity of the vaccine construct using EAAAK linkers [44]. CTL, HTL and B-cell epitopes were fused using AAY, GPGPG and KK linkers, respectively (Fig. 1). GPGPG and AAY linkers promote epitope presentation, while also reducing the formation of junctional epitope [45, 46]. Finally, to incorporate the B-cell epitopes KK linkers were utilised, which are also capable of enhancing the immunogenicity [47]. In the present study, multiple constructs were established using distinct TLR-4 agonists as adjuvants to determine the construct with the greatest immunogenicity (Table 4). Owing to the fact that the construct with RpfB (N-terminal) and RpfE (C-terminal) adjuvants had the highest immunogenicity of 11.3848 was chosen for further examination [48].

Fig. 1.

Schematic representation of the final vaccine construct, comprising adjuvant, CTL, HTL and B-cell epitopes linked with EAAAK, AAY, GPGPG and KK linkers, respectively

Table 4.

Physicochemical evaluation of the vaccine constructs with diverse adjuvants

Parameters	CobT + Vaccine construct	Lumazine synthase + Vaccine construct	BCSP31 + Vaccine construct	DnaJ + Vaccine construct	50S ribosomal protein + Vaccine construct	RpfE + Vaccine construct	RpfB + Vaccine construct	RpfB and RpfE + Vaccine construct
Antigenicity	0.5260	0.5544	0.5887	0.6634	0.6543	0.6499	0.5314	0.5554
Immunogenicity	9.1772	4.9275	4.4336	3.7748	3.0339	5.9054	7.9678	11.3848
Molecular weight (kDa)	58.42	39.39	56.31	62.54	35.48	39.49	60.11	78.02
No. of amino acids	564	361	532	531	333	375	565	742
Theoretical pI	6.02	7.88	8.51	7.96	5.89	5.98	6.41	5.68
Instability index	30.64	27.75	24.24	26.49	26.05	37.97	33.08	36.80
Aliphatic index	93.23	88.95	91.79	68.35	92.10	77.68	92.02	86.58
GRAVY	0.125	− 0.018	− 0.026	− 0.426	− 0.009	− 0.194	− 0.013	− 0.075

Evaluation of physicochemical properties

The antigenicity score of the proposed vaccine sequence candidate was found to be 0.5554, indicating that it is likely an antigen capable of evoking an immune response. In addition, the vaccine has proven to be non-toxic and non-allergic, demonstrating its safety. According to the Expasy Protparam server, the final synthetic vaccine consists of 742 amino acids and a molecular mass of 78.02 kDa, which is appropriate as proteins below 110 kDa are easier to purify [49]. The instability index of the constructed vaccine is found to be 36.80, implying greater stability of the protein. Further, the high thermostability and acidic nature are evidenced in the aliphatic index of 86.58 and the theoretical pI of 5.68 [50]. The half-life is the measure of the time taken for half of the protein produced in a cell to degrade. The constructed vaccine has a projected half-life of 30 h in mammalian reticulocytes (in vitro), > 20 h in yeast (in vivo) and > 10 h in E. coli (in vivo), as predicted based on the “N-end rule”, which relates the regulation of the in vivo half-life of a protein to the identity of its N-terminal residue [51, 52]. The GRAVY score is found to be − 0.075, denoting the hydrophilic nature of the vaccine. The values of the physicochemical properties are tabulated in Table 4.

Analysis of population coverage of predicted epitopes

The MHC restricted HLA alleles are highly polymorphic in their peptide binding domains and play a significant role in the adaptive immune system. Thus, it is essential to evaluate the distribution and expression profile of these alleles [53]. In the present study, the IEDB’s population coverage tool was utilised to estimate the population frequency of the predicted epitopes. The estimated global population coverage across 109 countries covering 16 different geographic regions is shown in Fig. S1a, and it is found to be 98.60%. The estimated Indian population coverage is shown in Fig. S1b, and it is found to be 98.69%. Fig. S1c and d represents the population coverage of individual epitopes across the globe and Indian population, respectively. These results suggest that the selected epitopes could be delivered to a wider range of the population.

Prediction and assessment of secondary and 3D structure of the vaccine construct

The secondary protein structure prediction plays an important role in protein function and 3D model of protein structure. The PSIPRED results reveal that the vaccine construct possessed 19.7% alpha-helix, 16.9% extended strands and 63.4% coils. It is to be noted that proteins having an alpha-helix and coils are important as structural antigens for they can be recognised by antibodies [54]. The 3D structure of the final vaccine construct was predicted using the ROBETTA sever, and model 2 was chosen based on the model evaluation using the SAVES server. The overall quality factor of the model determined using ERRAT is found to be 91.64, Verify 3D implies better compatibility of the model and the Ramachandran plot shows 86.2% of the residues in the favoured region. The Z-score of − 9.87 indicates the overall good quality and reliability of the model. Collectively, these results indicate the acceptable quality of the predicted vaccine candidate, and the modelled 3D structure is visualised using PyMOL software.

Discontinuous B-cell epitope prediction

The ElliPro server was employed to predict discontinuous B-cell epitopes from the constructed vaccine 3D structure. The server detected 4 conformational B-cell epitopes in the vaccine protein with scores ranging from 0.667 to 0.779, indicating that these regions will be distinguished by antibodies and plays an important role in humoral-mediated immunity (Table 5).

Table 5.

Predicted discontinuous B-cell epitopes in the final vaccine construct

S. no.	Discontinuous B-cell epitopes	Number of residues	Score
1	MLLVVGALLLVL	12	0.779
2	KVTERLPLPPNARRVEDPEMNMSREVVEDPGVPGTQDVTANVVVTPAHEAVVRVGTKPGTEVPPVNWAINTGNGYYGGVQFDQGTWEANGGLRYAPRAAAATEQAVAEVTRLRQGWGAWP	120	0.765
3	TVTVAAYQQSYWPHFVAAYTAPIYFGPVLKTDNFLRWLHHDAKKKKKSNETQCFNFIRVLKKSQDQTLALDPELAGIPKKLAAQQSYPHFVTVTVEAAAKMKNARTTSPADDAGL	115	0.747
4	DRSRPLQLDGHDAKQVWTTASTVDAALAQLAMTDTAPAAASRASRVPLSGMALPVVSAKTVQLNDGGLVRTVHLPAPNVGLLSAAGVPLLDVATAPIVEGMQIQVTRRKVEVNGVELAPDFLSPPAEEAPPVPVAR	136	0.667

Docking of designed vaccine with Toll-like receptors

The interaction between immune cells and the vaccine is necessary for the development of stable immune response. Toll-like receptors (TLR) are known to express on cancer cells and play a significant role in tumour initiation, progression, and invasion. They also recognise unique molecular patterns on the host and plays a crucial role in innate immunity. Remarkably, the studies indicate that the decreased expression of TLR-2, 4, 7 and 9 may facilitate immune evasion in TNBC and enhance tumour growth [55]. Notably, the TLR-4 and TLR-9 appear to be most promising for clinical application [48, 56]. Thus, we performed ClusPro docking analysis of the modelled vaccine with all these immune molecules (TLR-2, 4, 7 and 9). The docking results showed that the proposed vaccine significantly binds to the TLR-2 (− 1420.5 kcal/mol), TLR-4 (− 1189 kcal/mol), TLR-7 (− 1469.6 kcal/mol) and TLR-9 receptor (− 1372.1 kcal/mol) with the lowest energy, and hence, these models were taken for further analysis. The PyMol software was utilised to visualise the docked complexes. In the present study, PDBsum was used to gain insight into the interaction analysis of the docked complexes. For instance, the range of binding interactions including hydrogen bonds, salt bridges and nonbonding contacts, in the docked complex was explored. The result is shown in Fig. S2. It shows the details of different amino acid residues involved in the stability of TLR-Vaccine complexes. It is evident from the figure that non-bonded interactions dominate in stabilising the complex structure followed by the hydrogen bond interactions. Note that TLR-9-vaccine complex is stabilised by 12 salt bridges, 54 hydrogen bonds and 448 non-bonded interactions. The interactions exhibited by other TLR molecules are reported in Table 6. The higher number of these interacting patterns implicates that the designed vaccine has a strong affinity and structural stability against the TLR-9 receptor compared with other TLR receptors investigated in our analysis. The representative hydrogen bond interactions exhibited by each TLR molecules against vaccine was illustrated using PyMol software (Fig. S3).

Table 6.

The interactions between TLRs and Vaccine of the docked complexes

Complex	No. of salt bridges	No. of Hydrogen bonds	No. of Non-bonded contacts
TLR-2-vaccine	1	16	206
TLR-4 (Chain B)–vaccine	7	34	357
TLR-4 (Chain D)–vaccine	1	3	55
TLR-7-vaccine	1	16	252
TLR-9-vaccine	12	54	448

Molecular dynamic simulation study

Molecular dynamics analysis was employed to examine the binding interactions and flexibility of the binding site. MD simulations were performed for 50 ns using each complex, and the stability of the simulation was evaluated using parameters such as RMSD, Rg and SASA. The RMSD value reveals the structural changes that occurred for all the systems. It is shown in Fig. 2a that the complexes experience an initial increase in RMSD until the 10 ns time point, after which the upward trend for each system ceases and then remained constant throughout the simulation time indicating the stability of the system. It is found that variable RMSDs such as 1.48 nm (TLR-2), 0.67 nm (TLR-4), 1.24 nm (TLR-7) and 0.54 nm (TLR-9) were established by complexes. Comparatively, a lower degree of fluctuation was observed for TLR-9 complex, which is responsible for higher structural integrity than the other complexes. Rg analysis of the vaccine construct in complex with TLR receptors was conducted and is plotted as shown in Fig. 2b. Rg value showed slight fluctuation until 5 ns and then remained equilibrated throughout 50 ns MD simulation. It is found that variable Rg values such as 4.03 nm (TLR-2), 4.81 nm (TLR-4), 5.12 nm (TLR-7) and 3.49 nm (TLR-9) were observed for complexes. It is interesting to observe that the average Rg value of vaccine construct in complex with TLR-9 receptor was lower than the average Rg value of the other TLR vaccine construct complexes which indicates the compactness of the structure after binding. The low Rg profile denotes higher rigidity in the biological system. Solvent accessible surface area (SASA) is used to determine the surface area that can be accessed by solvent molecules. The average SASA values of our vaccine and TLR complexes are found to be 101.575 nm², 208.419 nm², 220.492 nm² and 69.356 nm² for TLR-2, TLR-4, TLR-7 and TLR-9, respectively. Since the SASA trajectories for the complexes remained constant throughout the simulation, this finding provides further evidence that the final vaccine construct maintains a stable conformation in the aqueous environment (Fig. 2c). Collectively, TLR-9 shows stable interactions when compared to other TLR vaccine complexes. The results correlate well with our previous docking analysis and, thus, validate our study.

Fig. 2 .

Molecular dynamics simulation analysis of different TLRs (TLR-2, TLR-4, TLR-7 and TLR-9) and vaccine complexes (a) The RMSD plot of the docked TLR and vaccine complexes at 50 ns. b Radius of gyration plots of TLRs and vaccine. c SASA profile of different TLRs and vaccine

In silico cloning of vaccine construct into pET-28 (+) vector

Codon optimization improves gene translation of the target by considering the codon bias of the host organism. The ExpOptimizer and GenScript optimization tool was utilised for back translation and codon optimization of the vaccine construct in E. coli. The Codon Adaptation Index (CAI) and GC content were estimated to be 0.81 and 56.44%, respectively, signifying optimal expression and better translational efficiency of the vaccine design. The optimised vaccine sequence was then cloned into the pET-28 (+) vector between the BamHI and PflFI restriction sites. Using the SnapGene software, these restriction sites were introduced to the N and C-terminal ends of the vaccine as they are not present in the DNA sequence of the vaccine construct. The overall length of the final vaccine clone was determined to be 4761 bp, and the vaccine clone is shown in Fig. 3.

Fig. 3 .

Representation of the vaccine sequence (red) cloned into the pET-28a (+) vector (black) with the BamHI and PflFI restriction sites at the N and C-terminal respectively

Analysis of vaccine efficacy using immune simulation

C-IMMSim was performed to gain insight of how an immune system would behave when the designed vaccine is introduced into a patient. Immunoglobulins contribute to the proliferation and activation of lymphocytes. The level of IgM concentration illustrated in Fig. 4a is evidence of a primary response. Consecutively, the secondary and tertiary responses also exhibited higher levels of immunoglobulin activity, accompanied by a decline in antigen concentration. Similarly, the level of cytokines in the form of IFN-γ and IL-2 represented in Fig. 4b increased widely showing T-cell-mediated immune response. Thus, the overall immune simulation results indicate the potential of the designed vaccine to target TNBC. The active cytotoxic T-cell (Tc) population per state showed a steady increase in the number of Tc cell population. Since the vaccine contained both CTL and HTL epitopes, it demonstrated stimulation of its corresponding immune cells, which in turn trigger other potential immune cells such NK cells, macrophages, and dendritic cells through complex signalling as illustrated in Fig. 4c. Overall, the immune response grew in parallel with each booster dose, indicating the activation of various immune cells.

Fig. 4.

Immune simulation profiles of the vaccine construct (a) Immunoglobulin expression profile (b) Cytokines and interleukin expression profiles (c) Dendritic cell, Macrophage, NK cell and cytotoxic T-cell population expression profiles

Conclusion

TNBC is the most complex subtype of breast cancer, and the ideal therapeutic choice for TNBC patients remains an unmet challenge. In the current study, diverse immunoinformatic approaches were applied in a sequential manner to design a chimeric subunit vaccine to combat TNBC. Initially, the B- and T-cell epitopes that are antigenic, non-toxic and non-allergic were identified from the Sema4A protein. These epitopes, which could elicit both cell-mediated and humoral immunity, were conjugated with an adjuvant and linkers to create an effective epitope-based peptide vaccine. The three-dimensional model of the vaccine revealed that the construct strongly interacts with the Toll-like receptors (TLR-2, TLR-4, TLR-7 and TLR-9) and induces robust cellular responses. The stability of the vaccine-TLR complex was further confirmed using molecular dynamics simulations. The findings demonstrated that the designed vaccine candidate had a reduced RMSD and Rg value, signifying the integrity and compactness of the molecule’s structure. Additionally, the SASA value of the proposed vaccine was higher, reflecting the molecule’s structural stability in the solvent environment. Finally, the immune simulation studies showed that the designed vaccine could also induce a strong cellular immune response. Given these points, the outcomes of our study indicate that the developed multi-epitope vaccine might be employed to treat triple-negative breast cancer efficiently. However, in vitro and in vivo experimental validations are required to evaluate its efficacy and safety, which will be an exciting avenue to take in the future.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by PP. SV conceived this study and responsible for the overall design, interpretation, manuscript preparation and communication. All authors read and approved the final manuscript.

Funding

The authors declare that no funds, grants or other support were received during the preparation of this manuscript.

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary information file].

Declarations

Conflict of interest

The authors have no relevant financial or non-financial interests to disclose.