Peptide antigens derived from tumors have been observed to elicit protective immune responses, categorized as either tumor-associated antigens (TAAs) or tumor-specific antigens (TSAs). Subunit cancer vaccines incorporating these antigens have shown promise in inducing protective immune responses, leading to cancer prevention or eradication. Over recent years, peptide-based cancer vaccines have gained popularity as a treatment modality and are often combined with other forms of cancer therapy. Several clinical trials have explored the safety and efficacy of peptide-based cancer vaccines, with promising outcomes. Advancements in techniques such as whole-exome sequencing, next-generation sequencing, and in silico methods have facilitated the identification of antigens, making it increasingly feasible. Furthermore, the development of novel delivery methods and a deeper understanding of tumor immune evasion mechanisms have heightened the interest in these vaccines among researchers. This article provides an overview of novel insights regarding advancements in the field of peptide-based vaccines as a promising therapeutic avenue for cancer treatment. It summarizes existing computational methods for tumor neoantigen prediction, ongoing clinical trials involving peptide-based cancer vaccines, and recent studies on human vaccination experiments.

The identification and characterization of immunogenic tumor antigens are essential for cancer vaccine development. In light of the impracticality of isolating and evaluating each putative antigen individually, in silico prediction algorithms, particularly those utilizing machine learning (ML) approaches, play a pivotal role. These algorithms significantly reduce the experimental workload necessary for discovering vaccine candidates. In this study, we employed six supervised ML methods on a dataset comprising 212 experimentally validated human tumor peptide antigens and an equal number of non-antigenic human peptides to develop models for immunogenicity prediction. These methods encompassed k-nearest neighbor (kNN), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), support vector machine (SVM), random forest (RF), and extreme gradient boosting (XGBoost). The models underwent validation through internal cross-validation within 10 groups from the training set and were further assessed using an external test set. Remarkably, the kNN model demonstrated superior performance, recognizing 90% of the known immunogens in the test set. The RF model excelled in the identification of non-immunogens, accurately classifying 93% of them in the test set. The three top-performing ML models according to multiple evaluation metrics (SVM, RF, and XGBoost) are to be subsequently integrated into the new version of the VaxiJen server, facilitating tumor antigen prediction through a majority voting mechanism.

Peanut allergy, a prevalent and potentially severe condition affecting millions worldwide, has been linked to specific human leukocyte antigens (HLAs), suggesting increased susceptibility. Employing an immunoinformatic strategy, we developed a “logo model” based on amino acid frequencies in the peptide binding core and used it to predict peptides originating from 28 known peanut allergens binding to HLA-DRB1*03:01, one of the susceptibility alleles. These peptides hold promise for immunotherapy in HLA-DRB1*03:01 carriers, offering reduced allergenicity compared to whole proteins. By targeting essential epitopes, immunotherapy can modulate immune responses with minimal risk of severe reactions. This precise approach could induce immune tolerance with fewer adverse effects, presenting a safer and more effective treatment for peanut allergy and other allergic conditions.

Tumor immunogenicity depends on the ability of peptides to form stable and specific interactions with both HLA molecules and T-cell receptors (TCRs). While HLA binding is essential, not all HLA-binding peptides elicit T-cell responses. This study investigates the molecular features distinguishing immunogenic T-cell epitopes from non-immunogenic HLA binders. Two datasets of nonamer peptides—38 T-cell epitopes and 144 non-epitopes—were compiled and analyzed using sequence logo models and molecular dynamics (MD) simulations of TCR–peptide–HLA complexes. A comparative logo analysis revealed strong amino acid preferences at central positions (p4–p8) in T-cell epitopes and absences in non-epitopes. A representative epitope–non-epitope pair was selected for structural modeling and 100 ns MD simulations. The T-cell epitope formed a more stable complex with the TCR and exhibited greater flexibility, supporting an induced-fit recognition mechanism. It also established a broader and longer-lasting network of hydrogen bonds and π interactions across the residues at positions p4–p8. In contrast, the non-epitope engaged TCR at only two positions. These findings highlight the critical role of the peptide’s central region in TCR engagement and provide structural insights useful for neoantigen prediction, vaccine design, and TCR-based immunotherapies.

Peptide-based cancer vaccines have emerged as a prominent focus in contemporary oncological research, as the quest for innovative cancer treatment modalities continues to gain momentum. A pivotal facet of their development is the precise delineation and characterization of immunogenic tumor antigens. In this context, VaxiJen stands out as one of the most widely used and cited computational servers for predicting immunogenicity, making it an invaluable tool for in silico antigen prediction. However, the database underpinning VaxiJen’s predictions has not undergone a comprehensive update for over fifteen years. To address this, a systematic search of the PubMed database was conducted to identify scholarly articles reporting data on novel immunogenic proteins and peptides undergoing human testing. The corresponding sequences of these proteins and peptides were subsequently curated from UniProtKB. Therefore, in this study, we introduce an updated dataset encompassing a repertoire of tumor immunogens, comprising 546 full-length human proteins and 212 human tumor peptides, as well as tumor non-immunogens, comprising 548 full-length human proteins and 181 human tumor peptides. The recently compiled VaxiGen tumor dataset is openly accessible. Researchers can conveniently download, search, and process it. This dataset, when paired with a suitable negative dataset, can further serve as a valuable training set, thereby facilitating improved predictions of the potential immunogenicity of hitherto uncharacterized protein or peptide sequences.

Identifying and characterizing putative tumor antigens is essential to cancer vaccine development. Given the impracticality of isolating and evaluating each potential antigen individually, in silico prediction algorithms, especially those employing machine learning (ML) techniques, are indispensable. These algorithms substantially decrease the experimental workload required for discovering viable vaccine candidates, thereby accelerating the development process and enhancing the efficiency of identifying promising immunogenic targets. In this study, we employed six supervised ML methods on a dataset containing 546 experimentally validated immunogenic human tumor proteins and 548 non-immunogenic human proteins to develop models for immunogenicity prediction. These models included k-nearest neighbor (kNN), linear discriminant analysis (LDA), quadratic discriminant analysis (QDA), support vector machine (SVM), random forest (RF), and extreme gradient boosting (XGBoost). After validation through internal cross-validation and an external test set, the best-performing models (QDA, RF, and XGBoost) were selected for further evaluation. A comparison between the chosen protein models and our previously developed peptide models for tumor immunogenicity prediction revealed that the peptide models slightly outperformed the protein models. However, since both proteins and peptides can be subject to tumor immunogenicity assessment, evaluating each with the respective models is prudent. The three selected protein models are set to be integrated into the new version of the VaxiJen server.