T Cell Epitope Predictions

Abstract

Throughout the body, T cells monitor MHC-bound ligands expressed on the surface of essentially all cell types. MHC ligands that trigger a T cell immune response are referred to as T cell epitopes. Identifying such epitopes enables tracking, phenotyping, and stimulating T cells involved in immune responses in infectious disease, allergy, autoimmunity, transplantation, and cancer. The specific T cell epitopes recognized in an individual are determined by genetic factors such as the MHC molecules the individual expresses, in parallel to the individual's environmental exposure history. The complexity and importance of T cell epitope mapping have motivated the development of computational approaches that predict what T cell epitopes are likely to be recognized in a given individual or in a broader population. Such predictions guide experimental epitope mapping studies and enable computational analysis of the immunogenic potential of a given protein sequence region.

Keywords: T cells; benchmarking; databases; immune epitopes; machine learning.

Figure 1

Overview of the biological process, experimental assessment, and computational prediction of T cell epitope recognition. (a) Overview of the main cellular mechanisms involved in antigen processing, presentation, and recognition of T cell epitopes that have been included in computational predictions. (Left) MHC class I–restricted T cell epitopes primarily arise from intracellular antigens that are cleaved by the proteasome and transported into the ER by TAP, where they can bind to MHC class I molecules that get transported to the cell surface, where they can be recognized by CD8⁺ T cells. Proteins and peptides are depicted as beads-on-a-string, with red circles indicating amino acids that are C-terminal residues of peptides presented by MHC molecules. In contrast, MHC class II–restricted T cell epitopes (right) are primarily derived from extracellular proteins taken up by professional APCs that are cleaved in lysosomal vesicles, where they can bind to MHC class II molecules, be transported to the cell surface, and be recognized by CD4⁺ T cells. Dark purple circles indicate amino acids at the C-terminal end of the core binding to MHC-II. (b) Three main categories of experimental assays have been utilized to characterize the steps involved in antigen processing and recognition of T cell epitopes. (Left) MHC binding assays that determine the affinity of a synthetic peptide to a specific MHC molecule. (Middle) MHC ligand elution assays that isolate and identify peptides bound to MHC molecules on the cell’s surface as a result of natural antigen processing and presentation. (Right) T cell epitope recognition assays, in which the ability of T cells to interact with and/or respond to a candidate epitope is determined. (c) Approaches to the computational prediction of T cell epitopes, starting with pioneering use of MHC motifs such as SYFPEITHI (left) (47), in which allowed amino acids at anchor positions (blue bolded) and at auxiliary anchor positions (purple) were identified based on a heuristic analysis. This was followed by machine learning approaches that were explicitly trained on quantitative data such as BIMAS (middle) (50), where numeric values would be assigned for each of the 20 conventional amino acids (rows) at each position in a 9-residue peptide (columns), so that they best reproduce measured binding affinities for a set of peptides that were previously tested (the training data). Finally, current neural networks approaches have custom architectures that allow training on combined data from multiple MHC alleles and from both MHC binding and elution data, such as the recent NetMHCpan version 4.0 (right) (127). Abbreviations: APC, antigen-presenting cell; ER, endoplasmic reticulum; TAP, transporter associated with antigen processing; TCR, T cell receptor.