Triple-negative breast cancer (TNBC) is a highly aggressive subtype of breast cancer that progresses more quickly and has a higher chance of metastasis and recurrence [1, 2]. The treatment options are limited to chemotherapy because of the lack of expression of standard molecular targets such as estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2). Apparently, traditional chemotherapeutic agents, such as anthracycline, paclitaxel and alkylating agents, may trigger systemic toxicity and disruptive effects [3]. The challenges related to the inherent heterogeneity and aggressive nature of TNBC are further complicated by the limited availability of prognostic biomarkers for treatment response [4]. Immunotherapy has recently emerged as a game-changing technique in the treatment of a wide range of tumors [5]. It should be noted that the immunogenic landscape observed in TNBC has been correlated with an exciting response to immunotherapy. Clearly, the increased prevalence of tumor-infiltrating lymphocytes (TILs), a substantial tumor mutational load and survival correlations with the extent of T-cell and B-cell infiltration rendered immunotherapy an appealing therapeutic choice for TNBC [6, 7].

One of the most cutting-edge immunotherapeutic technique is the use of cancer vaccines to provoke active immune responses in the breast tumor environment. The key benefit of these subunit or peptide vaccines are their low toxicity and targeted adaptive immune response [8]. Interestingly, the ongoing clinical trials exploring chimeric peptide-based vaccines for TNBC includes MUC1 (NCT00986609), CDX-1401 (NCT02661100) PVX- 10 (NCT03362060), P10s-PADRE (NCT02938442) [9]. The incorporation of novel overexpressed proteins into vaccine design has revolutionized the scientific landscape, paving the way for enhanced treatment strategies. The primary downside of this multi-epitope vaccine is its exclusivity to personalized treatment regimen. Certainly, peptide-based vaccines account for 45.5% of clinical studies for breast cancer [10]. For instance, Rajendran et al. developed a novel transcription factor based MZF-1 and SOX9 vaccines for the management of metastatic TNBC [11, 12]. Tingting et al. developed an octamer-binding transcription factor 4 (OCT4) based anti-cancer vaccine, and it’s in vivo investigations have demonstrated significant tumor growth suppression [13]. Recently, Mahdevar et al. utilized immunoinformatics to identify B and T-cell-inducing epitopes for BORIS cancer-testis antigen against breast cancer. Of note, the developed vaccine considerably triggered immune responses both in vitro and in vivo environment [14]. It is worth mentioning that our team devised a successful multi-epitope vaccine targeting Sema4A, a transmembrane protein that has been recognised as a vital regulator of TNBC [15].

Despite the growing interest on immunotherapy, the prospective use of CSC-associated antigens as vaccines are quite limited. Therefore, in order to effectively battle cancer, it is imperative to identify and eradicate CSC in addition to the elimination of tumor cells [16, 17]. NARF associated with hypoxia-inducible HIF-1 target gene, is currently reported to be overexpressed in human breast cancer patients. It recruits histone lysine demethylase 6A (KDM6A) to OCT4 binding sites to activate the NANOG, SOX2, and KLF4 genes, which specify breast CSC. Note that the HIF-1 pathway is highly activated in TNBC relative to other isotypes and is therefore reported to be an excellent candidate for cancer vaccine development [18, 19]. In light of these references, the current study aimed to establish a subunit vaccine combining immunoinformatics techniques and simulation analysis focusing on NARF, a recently identified oncogenic promoter, to overcome the burden associated with TNBC.

The development of a highly effective and therapeutic vaccine targeting TNBC is indispensable due to the scarcity of viable treatment alternatives. Thus, the present study employed a wide range of immunoinformatic methods in a holistic manner to generate a chimeric subunit vaccine against TNBC. Initially, the epitope regions of NARF to stimulate CD4⁺ or CD8⁺ T-cell and B-cell immune responses were retrieved [53, 54]. The extracted epitopes were then evaluated for vaccine-like properties using a variety of descriptors including toxicity and antigenicity. Peptide-based vaccines often exhibit low immunogenicity even with appropriate T and B-cell epitopes [55]. Thus, the selected epitopes were combined with appropriate adjuvants and linkers to produce a robust vaccine construct [56]. Note that the AAY linker boosts immunogenicity, and the GPGPG universal spacer is essential for inducing HTL responses. Additionally, B-cell epitopes were connected using KK linkers to enhance the immunogenicity. While the EAAAK was used to fuse the adjuvant for efficient separation of functional domains.

The vaccine design comprising of 687 amino acids possessed a lower molecular weight of 71.89 kDa (<110 kDa) facilitating its easier purification. The vaccine’s preference for acidic surroundings and its suitability for the body’s polar environment are indicated by its pI of 5.15 and GRAVY score of −0.228, respectively. In addition, the instability index of 36.85 (<40) and aliphatic index of 81.02 implies that the vaccine is a stable protein with high thermostability. Furthermore, the higher solubility index of 0.885 suggests that the vaccine is able to dissolve easily in water and other solvents. The extended half-life of 30 h in mammalian reticulocytes (in vitro), >20 h in yeast (in vivo), and >10 h in E. coli (in vivo) ensures prolonged exposure of our vaccine to the immune system [57–59]. Also, the PI score from ElliPro closely correlates with solvent accessibility and has profound effects on the structure and functionality of biological macromolecules [60]. The presence of alpha-helices and coils in our vaccine construct is remarkable, as they are structural antigens recognized by antibodies [61]. Further, the Z-score of −9.88 for the modelled vaccine suggests that it is structurally analogous to experimentally resolved models of similar sizes, validating its near-native conformation. Importantly, the findings of the population coverage suggest that the selected epitopes could be generalized to the vast majority of the global population [62]. TLRs sense unique host molecular patterns and exhibit a crucial role in innate immunity. Notably, the research suggests that lowered expression of TLR-2, 4, 7, and 9 could promote cancer growth and immune evasion in TNBC [63–65]. It is worth noting that the designed vaccine exhibits greater affinity for TLR-4 and TLR-7 than other targets considered in our analysis. It is obvious from the results that the existence of high number of non-bonded interactions together with hydrogen bonds and salt-bridges facilitate the binding of vaccine with TLR protein. This firm binding will promote MHC presentation, which in turn activate subsequent T cell specific responses [66, 67].

The iMODS findings implied that all the complexes, except vaccine-TLR-2, were directed towards each other using affine-model arrows indicating functional roles and strong binding (Figures 4A, 5A, Supplementary Figure S3, Supplementary Figures S4A and S5A). The deformability peaks in Figures 4B, 5B, Supplementary Figures S4B and S5B indicate the regions of flexibility. Notably, a limited number of peaks in the TLR-4 and TLR-7 complexes imply rigid regions of the residues in the complex. The low B-factor values associated with the TLR-4 and TLR-7 complexes in Figures 4C, 5C, Supplementary Figures S4C and S5C highlight the minimal atomic displacements that contribute to the stability of these complexes. Figures 4D, 5D, Supplementary Figures S4D and S5D shows higher eigenvalues of 1.0942e⁻⁰⁶, 1.2328e⁻⁰⁶ in TLR-4 and TLR-7 complexes, indicating greater stability and antigen presentation. Subsequently, the mode-specific variance, represented in Figures 4E, 5E, Supplementary Figures S4E and S5E, implies their relative contribution to equilibrium motions. It is important to note that the first six normal modes accounted for almost 80% of the variance. Further, the covariance map in Figures 4F, 5F, Supplementary Figures S4F and S5F reveals the coupling between residue pairs in the complex, implying that all complexes are correlated. In addition, Figures 4G, 5G, Supplementary Figures S4G and S5G shows the elastic network analysis of the complexes, with darker grey dots indicating stiffer complexes.

Taken together, TLR-4 and TLR-7 establish stable interactions when compared to other complexes. Evidence from the literature suggests that the firm association of the vaccine with these TLRs signifies a robust immune response, thereby preventing cancer cell evasion [68]. The results are also consistent with our earlier docking analysis. Further, the codon optimized vaccine sequence in E. coli yielded a favourable CAI and GC content values, indicating optimal expression and greater translational efficiency of the engineered vaccine [65, 69]. The overall magnitude of the immune response increased with each booster dose, indicating the strong immune response of our engineered vaccine.

It is worth noting that designing cancer vaccines involves high costs and a time-consuming process. Nevertheless, the benefits of computational (in silico) approaches have significantly reduced the costs and time by using biological modelling techniques. Before delving into laboratory-based experiments for cancer vaccine development, the immunoinformatics approach provides crucial insights into the interaction pattern between the immune system and cancer cells at molecular and cellular levels [70]. For instance, several studies have consistently reported the efficiency of in silico designed vaccine candidates in experimental assessments [13, 17].

Based on evidence from the literature, we hypothesise that our designed vaccine has better antigenic and immunogenic properties, along with better binding affinity for toll like receptors. For instance, the vaccine designed against SOX 9 transcription factor by Rajendran et al. had an antigenicity score of 0.59 and an instability index of 39.84, while our proposed vaccine exhibited an antigenicity score of 0.63 and an instability index of 36.85, demonstrating higher stability and immunogenic properties that evoke the immune response. Similarly, an MZF-1-based vaccine designed for TNBC by Rajendran et al. displayed a lowest immunogenicity score (3.720) and binding affinity score against TLR 2, 4, 7, and 9 (−929.5 kcal/mol, −1060.3 kcal/mol, −975.6 kcal/mol, and −1091.9 kcal/mol), whereas our designed vaccine had an immunogenicity score of 7.92 and binding scores of −1277.5 kcal/mol, −1,120 kcal/mol, −1808 kcal/mol, and −1,353 kcal/mol against TLR 2, 4, 7, and 9. These results imply that the developed vaccine would be efficient in eliciting strong immune responses. Given these considerations, our study on the NARF-based multi-epitope vaccine laid the groundwork for the effective treatment and management of triple-negative breast cancer. In future, the perspectives such as immunosuppressive cell depletion and immunotherapy-based combination therapies could be of important to improve the vaccine’s anti-tumor efficacy [71]. Indeed, in vitro and in vivo experimental validations are crucial to examine the safety and efficacy of the developed vaccine.