Background Adoptive cell transfer (ACT) of neoantigen-reactive CD8⁺ T cells has had some success in the clinic; however, mouse models recapitulating neoantigen-reactive CD8⁺ T cell ACT have been limited, especially in poorly immunogenic models such as the murine melanoma model B16F10. Further, direct comparison of neoantigen-reactive CD8⁺ T cell ACT versus ACT utilizing T cells reactive against overexpressed-self or heteroclitic tumor-associated antigen (TAA) peptides has been lacking. To address these gaps, we developed a model system to study neoantigen- and TAA-reactive CD8⁺ T cell ACT in parallel.

Methods Whole exome sequencing and RNA sequencing were employed to predict neoantigens present in B16F10. C57BL/6 mice were then administered charge-modified TLR7/8 conjugate vaccines targeting neoantigenic peptides predicted to elicit T cell responses. Vaccination against neoepitopes and previously characterized TAA epitopes elicited neoantigen- or TAA- reactive CD8⁺ T cells and modest tumor growth control; T cell receptors were isolated from neoantigen- and TAA-reactive CD8⁺ T cell clones. To develop an ACT model, we conducted CRISPR/Cas9-mediated knockdown of endogenous TCR and subsequent transduction (g-retrovirus encoding neoantigen- or TAA-reactive TCRs) in murine CD8⁺ T cells. T cells were expanded in vitro for use in downstream in vitro and in vivo applications.

Results Peptide stimulation in vitro of neoantigen- and TAA-reactive T cells revealed wide ranges of 1) specificity (vs. cross-reactivity to wild type peptide), and 2) avidity for cognate peptide. Neoantigen- and TAA-reactive T cells were able to recognize B16F10 cells in vitro, with the most robust recognition (readout:% T cells IFNg⁺) when target antigen is highly expressed by tumor cells. Ability of neoantigen- or TAA-reactive T cells to kill B16F10 in vitro was strongly dependent upon both tumor antigen expression and T cells’ TCR avidity. Similarly, reduction of tumor growth in vivo required both high tumor antigen expression and transfer of high avidity neoantigen- or TAA-reactive CD8⁺ T cells.

Conclusions To conclude, we have created a novel model of neoantigen- and TAA-reactive ACT in immunotherapy-refractory B16F10 melanoma. Our data suggest that antigen abundance and TCR avidity are parameters that influence ACT efficacy; future research will be conducted to dissect the individual and summative contributions of these parameters and translate this knowledge towards improving ACT design in the clinic.