Abstract
We recently identified an HLA-A*02:01-restricted CD8 T cell epitope encompassing the H3.3K27M mutation, which is common in diffuse midline glioma, and a corresponding high-affinity T cell receptor (TCR) that recognizes the epitope. While recombinant viral vectors have been widely used for genetic reprogramming of T cells, viral vectors are far from ideal as they typically integrate randomly into the genome and are not governed by the molecular regulatory mechanisms of the cell. We used a non-viral, CRISPR-Cas9-based approach to replace the endogenous TCR with H3.3K27M TCR at the TCR a constant region (TRAC) in human T cells. Co-electroporation of healthy donor-derived T cells with homology-directed repair (HDR) templates encoding the full-length sequence of H3.3K27M TCR along with CRISPR-Cas9 ribonucleoprotein (RNP) resulted in the integration of the new TCR into the TRAC locus by HDR. Antibody staining of TCR α/β and H3.3K27M dextramer showed replacement of endogenous TCR with H.3.3K27M TCR in ~5–10% of TCR+ CD8 T cells. Modifying the HDR template to include a binding site for Cas9, which contains the nuclear localization signal that acts as a “shuttle”, further enhanced the integration efficiency (~15% of TCR+ CD8 T cells). Furthermore, HLA-A2+ H3.3K27M TCR-engineered T cells selectively killed U87 glioma cells expressing the H3.3K27M epitope. In addition, the engineered T cells exhibited a stem memory-like phenotype when expanded in the presence of a cocktail of IL-2, IL-7 and IL-15. Taken together, these data provide evidence for non-viral genome editing as a strategy to engineer T cells with specific TCR for cancer immunotherapy.