Peptide-based vaccines for cancer: are we choosing the right peptides?

The T-cell receptor (TCR) interaction with MHC molecules is a central event in T-cell-mediated responses. During thymic development, MHC molecules educate T cells, guiding their positive and negative selection. Outside the thymus, protein antigens from viral or bacterial sources are presented in the context of MHC molecules to T cells after intracellular processing to peptides. TCRs on lymphocytes recognize foreign antigens bound to the MHC and proliferate in the periphery in response to these stimuli. The MHC loci encode class I and II MHC molecules. Class I MHC molecules are found on the surface of nucleated cells, where they form noncovalently linked heterodimers between a MHC-encoded, membrane-spanning variable heavy-chain and a non-MHC-encoded, invariant light-chain β₂-microglobulin. Class I molecules bind short peptides (usually ∼8–10 amino acids), mostly derived from endogenously synthesized proteins, but also from exogenous antigens, in the endoplasmic reticulum and present them to cytotoxic T lymphocytes (CTLs) that express the CD8 coreceptor.

Since an individual expresses only a limited number of MHC molecules, these antigen-presenting molecules must be able to interact with a large diversity of peptide sequences. Nevertheless, the binding of peptides cannot be overly ‘promiscuous’ since individual MHC molecules have their own particular peptide-binding motifs. The peptide-binding groove can be subdivided into various pockets (A–F). Early biochemical studies indicated the presence of conserved (consensus) sequences in high-affinity binding peptides. In this process, low-affinity peptides were lost and do not feature in the considerations of Rammensee et al., who set the ‘rules’ for peptide binding for class I MHCs based on the sequence findings [1]. It is most important to note that the crystallographic studies indicated that the conserved regions were hydrophobic ‘anchors’ that were bound in ‘hydrophobic’ pockets. The nonconserved (i.e., variable) region of peptides were likely to be presented to T cells. Although these anchor residues are required for stabilization and high-affinity binding of the peptide to MHC, it has been demonstrated that some peptides that do not contain the respective canonical anchor residues can still bind and be presented by class I MHC molecules, allowing cell lysis to occur [2–8].

Peptides that bind to class I MHC molecules with high affinity usually induce high-avidity CTLs. Identification of high-affinity peptides of foreign antigens (e.g., virus or bacteria) and induction of high-avidity CTLs is necessary for their elimination. However, if the antigen is self, such as tumor-associated antigens, their CTL repertoire for high-affinity binding peptides would most probably be deleted, as has been demonstrated for p53 and MUC1, which leads to tolerance. Thus, for tumor immunotherapy, the most appropriate peptides for immunization would be the low-to-medium-affinity binding peptides. However, two significant problems exist in the use of lower-affinity epitopes. First, peptide epitopes of lower affinity are unlikely to conform with the predicted epitope motifs and, thus, are difficult to identify. Since such low-affinity peptides cannot be detected by elution studies and prediction algorithms, the only effective methods for their identification are systematic binding studies and recognition of the peptide–MHC by the TCR only. Second, peptide affinity for the MHC and stability of the peptide–MHC complex has been demonstrated to be a significant factor in overall immunogenicity. In order to overcome this problem, attempts have been made to improve the affinity of peptides for the MHC through the replacement of ‘anchor’ residues with the previously determined canonical amino acids. Although this can result in the enhancement of peptide–MHC interactions and reduced likelihood of tolerance, in many cases, mutations to the MHC anchor residues have resulted in CTLs that do not recognize their natural counterpart. These results highlight the importance of balance between MHC affinity and TCR crossreactivity for effective epitope enhancement.

One such example of where mutations were made to the anchor amino acid positions to include the canonical high-affinity amino acids to a self tumor-associated antigen in order to increase affinity, was to the MUC1–8 peptide. The MUC1–8 peptide binds with low affinity to H-2K^b and is devoid of the canonical anchor amino acids [4]. CTLs were induced in C57BL/6 mice where MUC1 is foreign; however, in MUC1 transgenic mice where MUC1 is self, CTLs were not induced. Mutations were made at positions 5 and 8, which yielded a high-affinity binding peptide that induced high-avidity CTLs in MUC1 transgenic mice [9].

Low-affinity peptides, which induce robust immune responses, are autoimmune peptides particularly presented by class II MHC molecules, although there is some evidence that low-affinity autoimmune peptides can also be presented by class I MHC. Thus, since self low-affinity peptides are responsible for strong autoimmune responses, why are we not choosing the low-affinity peptides as candidates for cancer vaccines? Are we choosing the wrong peptides in peptide-based cancer vaccines?

In summary, the choice of candidate peptides for use in immunotherapeutic cancer vaccines is vital. A plethora of knowledge regarding interactions between the peptide and MHC, and between the peptide–MHC complex and TCR is required. Although high-affinity peptides represent an obvious choice, they have the disadvantage of being poorly immunogenic – the cognate T cells having been deleted during thymic maturation. Hence, low- or medium-affinity peptides present a more viable alternative but are less likely to induce an immune response. Thus, anchor modifications are required in tumor-associated peptides to improve receptor binding. Immunizing humans with low-affinity mutated peptides that could induce CTLs may have a therapeutic benefit. These peptides would be of interest in human clinical trials and in order to determine if such peptides are more potent than attempting to induce immunity to self high affinity peptides. The ability to find low-affinity peptides from a self antigen (where the T-cell repertoire would not have been deleted) and to make appropriate mutations in order to yield highly immuogenic higher affinity binding peptides that can be used to immunize patients, is an interesting prospect, and may lead the way for alternative approaches in peptide-based cancer immunotherapy studies.

Financial & competing interests disclosure

The author has no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.