We developed an epitope selection method for the design of MHC targeting peptide vaccines. The method utilizes predictions for several clinical checkpoint filters, including binding affinity, immunogenicity, antigenicity, half-life, toxicity, IFNγ release, and instability. The accuracy of the prediction tools for these filter variables was confirmed using experimental data obtained from the Immune Epitope Database (IEDB). We also developed a graphical user interface computational tool called ‘PCOptim’ to assess the success of an epitope filtration method. To validate the filtration methods, we used a large data set of experimentally determined, immunogenic SARS-CoV-2 epitopes, which were obtained from a meta-analysis. The validation process proved that placing filters on individual parameters was the most effective method to select top epitopes. For a proof-of-concept, we designed epitope-based vaccine candidates for squamous cell carcinoma, selected from the top mutated epitopes of the HRAS gene. By comparing the filtered epitopes to PCOptim’s output, we assessed the success of the epitope selection method. The top 15 mutations in squamous cell carcinoma resulted in 16 CD8 epitopes which passed the clinical checkpoints filters. Notably, the identified HRAS epitopes are the same as the clinical immunogenic HRAS epitope-based vaccine candidates identified by the previous studies. This indicates further validation of our filtration method. We expect a similar turn-around for the other designed HRAS epitopes as a vaccine candidate for squamous cell carcinoma. Furthermore, we obtained a world population coverage of 89.45% for the top MHC Class I epitopes and 98.55% population coverage in the absence of the IFNγ release clinical checkpoint filter. We also identified some of the predicted human epitopes to be strong binders to murine MHC molecules, which provides insight into studying their immunogenicity in preclinical models. Further investigation in murine models could warrant the application of these epitopes for treatment or prevention of squamous cell carcinoma.

The current COVID-19 pandemic continues to spread and devastate in the absence of effective treatments, warranting global concern and action. Despite progress in vaccine development, the rise of novel, increasingly infectious SARS-CoV-2 variants makes it clear that our response to the virus must continue to evolve along with it. The use of immunoinformatics provides an opportunity to rapidly and efficiently expand the tools at our disposal to combat the current pandemic and prepare for future outbreaks through epitope-based vaccine design. In this study, we validated and compared the currently available epitope prediction tools, and then used the best tools to predict T cell epitopes from SARS-CoV-2 spike and nucleocapsid proteins for use in an epitope-based vaccine. We combined the mouse MHC affinity predictor and clinical predictors such as HLA affinity, immunogenicity, antigenicity, allergenicity, toxicity and stability to select the highest quality CD8 and CD4 T cell epitopes for the common SARS-CoV-2 variants of concern suitable for further preclinical studies. We also identified variant-specific epitopes to more precisely target the Alpha, Beta, Gamma, Delta, Cluster 5 and US variants. We then modeled the 3D structures of our top 4 N and S epitopes to investigate the molecular interaction between peptide-MHC and peptide-MHC-TCR complexes. Following in vitro and in vivo validation, the epitopes identified by this study may be used in an epitope-based vaccine to protect across all current variants, as well as in variant-specific booster shots to target variants of concern. Immunoinformatics tools allowed us to efficiently predict epitopes in silico most likely to prove effective in vivo, providing a more streamlined process for vaccine development in the context of a rapidly evolving pandemic.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of COVID-19, has challenged public health at an unprecedented scale which has led to a dramatic loss of human life worldwide. To design a protective vaccine against SARS-CoV-2, it is necessary to understand which SARS-CoV-2 specific epitopes can elicit a T cell response and provide protection across a broad population. In this study, PLpro and RdRp, two immunogenic non-structural proteins from an immunodominant gene region ORF1ab, as well as ORF3a and ORF9b are identified as potential vaccine targets against SARS-CoV-2. To select top epitopes for vaccine design, we used various clinical properties, such as antigenicity, allergenicity, toxicity and IFN-y secretion. The analysis of CD8 and CD4 T cell epitopes revealed multiple potential vaccine constructs that cover a high percentage of the world population. We identified 8 immunogenic, antigenic, non-allergenic, non-toxic, stable and IFN-y inducing CD8 proteins for nsp3, 4 for nsp12, 11 for ORF3a and 3 for ORF9b that are common across four lineages of variants of concern: B.1.1.7, P.1, B.1.351 and B.1.617.2, which protect 98.12%, 87.08%, 96.07% and 63.8% of the world population, respectively. We also identified variant specific T cell epitopes that could be useful in targeting each variant strain separately. Including the prediction of mouse MHC affinity towards our top CD8 epitopes, our study revealed a total of 3 immunogenic, antigenic, non-allergenic, non-toxic, stable and IFN-y inducing CD8 epitopes overlapping with 6 antigenic, non-allergenic, non-toxic, stable and IFN-y inducing CD4 epitopes across all four variants of concern which can effectively be utilized in pre-clinical studies. The landscape of SARS-CoV-2 T cell epitopes that we identified can help lead SARS-CoV-2 vaccine development as well as epitope-based peptide vaccine research in the future.

No abstract