Any cancer immunotherapy modality aims to harness the immune system to eliminate cancer cells. T cells are the primary cell type in charge of this task; however, the contribution of the cells of the innate immune system, such as natural killer cells (NK cells) and natural killer T (NKT) cells, is well documented too (Vivier et al., 2012).

T cells can recognize and kill tumor cells by cognate epitope recognition. Historically, these epitopes are derived from two broad classes of tumor antigens: tumor-associated antigens (TAAs) and tumor-specific antigens (TSAs) (Dave et al., 2022, Mørk et al., 2022). In the context of cancer antigens, the interaction between the CD8+ T cell receptor (TCR) and the peptide-major histocompatibility complex (p-MHC) can induce a tumor-specific immune response. The CD8+ TCR predominantly interacts with the central amino acid residues in an 8- to 11-mer epitope. In contrast, the anchor positions within epitopes contain residues involved mainly in interactions with MHC-I molecules (Oliveira et al., 2021). Subsequently, the MHC-I immunopeptidome repertoire has reduced diversity in anchor position residues and greater diversity in the remaining positions (Capietto et al., 2020). Furthermore, a small portion of the repertoire may be immunogenic, while the rest is poorly immunogenic and, therefore, irrelevant to cancer vaccine development. Accordingly, one of the obstacles in cancer immunotherapy is the identification of the subset of immunogenic tumor antigens that can induce antitumor T cell responses (Ouspenskaia et al., 2022). One of the reasons why T cells can only be activated by a small portion of the immunopeptidome is due to immune intrinsic and extrinsic tolerance mechanisms preventing autoreactive T cell responses that could cause the disease (Alshetaiwi et al., 2020, ElTanbouly and Noelle, 2021).

To deal with immune tolerance, a critical challenge in cancer vaccine development, several strategies, in addition to the application of neoantigen-based vaccines, have been applied, allowing the generation/identification of modified epitopes, altered peptide ligands (APLs), analog peptides, or mimotopes, functional mimics of wild-type epitopes (Buhrman and Slansky, 2013). Thus, combinatorial peptide libraries bearing single defined amino acid positions within a randomized context were employed to identify mimotopes specific for a T-cell clone (Sherev et al., 2003). Positional scanning synthetic combinatorial libraries (PS-SCL) were used as a source to identify T cell peptide ligands (Galloway et al., 2019, Pinilla et al., 2022). And more recently, in vivo positional scanning approach was reported, where peptide mixture(s), each carrying 20 different peptides with amino acid substitutions at defined positions, called “mini” libraries, were used as immunogens for subsequent isolation of individual APLs or mimotopes with enhanced antitumor efficacy (He et al., 2022).

A fundamental requirement for any vaccine to be successful is a similarity of the selected antigen with the pathogen or disease condition; therefore, our Variable Epitope Libraries (VELs) reported previously offer an immunogen construction and vaccine development concept adapted to the antigenic heterogeneity of cancer (Pedroza-Roldan et al., 2009, Manoutcharian et al., 2021). The idea behind the VEL concept is quite simple: simultaneous presentation of complex antigen mixtures to the immune system upon immunization with the VELs will induce a large pool of diverse TCR-bearing antitumor effector T cells targeting multiple both known and unknown antigens, including rapidly evolving collection of neo-antigens.

Although highly unorthodox, the rationale behind VEL immunogens is entirely rational: T cell ontogeny depends on randomly generated TCRs and stochastic methods for CD4+ /CD8+ lineage selection. Furthermore, genomic instability and mutations, inherent to nearly all cancers, can unpredictably lead to antigenic heterogeneity (Fennemann et al., 2019). VEL immunogens are generated by randomly substituting defined non-MHC anchor amino acid positions within T cell-recognized TAA-derived epitopes (Domínguez-Romero et al., 2020). This strategy had improved antitumor effects compared to wild-type TAA treatments (Martínez-Cortés et al., 2021). All VEL immunogens we have tested so far were based on defined TAAs; likewise, to the best of our knowledge, all antigens tested as cancer vaccine candidates by others were also based on TAAs, TSAs, or virus-derived antigens. With this in mind, we decided to explore non-traditional sources of antigens, i.e., an unbiased antigen source to generate undefined epitopes/mimotopes.

The decision on which antigen or how many antigens to include within a potential cancer vaccine is perhaps the most crucial aspect of our endeavor. Therefore, the induction of T cell responses to a broad repertoire of antigen-unbiased epitopes is an effective strategy to eliminate tumor cells and reduce the immune escape effect. In this regard, an antigen-unbiased VEL immunogen is better matched to cancer cells at the antigen level than previous TAA-based VEL immunogens (Servín-Blanco et al., 2018, Jiménez-Chávez et al., 2021).

In this study, we designed an antigen-unbiased 9-mer epitope VEL, G3d, composed of peptides with two defined H2-Dd MHC class I-anchoring residues, substitutions by any of 20 amino acids in three positions, and alanine at the remaining four positions. We hypothesized that the diversity of T cell responses would be primarily due to TCR interaction with the substituted residues rather than with the inert amino acid alanine (Morrison and Weiss, 2001), resulting in a broad, antigen-unbiased immune response that directly mimics tumor antigen heterogeneity, thus limiting immune escape.

We employed the mimotope library G3d as a vaccine immunogen, identified cancer-specific T-cell mimotopes in two different settings of screening assays, and showed differential antitumor efficacy of selected mimotopes.