A Protease-Mediated Regulatory Chimeric Antigen Receptor (CAR): "Double Arm" CAR System Improves Tumor-Cell-Specificity of CAR-T Cell Therapy

【Introduction】 CAR-T cell therapy has shown excellent therapeutic effects against some malignancies including refractory B cell lymphoma or leukemia. This adoptive T cell transfer therapy is an attractive methodology; however, there are several disadvantages to be overcome. Current CAR-T cell therapy targets single cell-surface molecule, which leads to damage of normal cells expressing the target protein. Such "On-target / off-tumor" effect is one of the major adverse effects. Improvement of target-cell-specificity is needed to avoid this serious adverse effect and to expand target diseases of this therapy. Here, we show the protease-mediated "Double-Arm" CAR-T cell system, which improved the specificity of CAR-T cell therapy by recognizing two distinct cell-surface proteins. We designed two types of CARs: "Effector CAR" and "Scissors CAR". The "Effector CAR" is constituted of a single chain Fv fragment (scFv) targeting a cell-surface protein (protein X) on tumor cells, Human Immunodeficiency Virus protease (HIVPR) recognition polypeptide sequence, and a functional domain of CD3-zeta. The "scissors CAR" is constituted of a recognition portion targeting another protein (protein Y) and HIVPR. The HIVPR induces cleavage of the recognition polypeptide sequence in the effector CAR leading to inactivation of the effector CAR when the CAR-T cells contact with cells expressing both proteins X and Y.

【Material and Methods】 For proof of principle, we first constructed "anti-CD19 mCherry CAR" harboring mCherry fluorescence protein in the cytoplasmic region under the HIVPR recognition polypeptide sequence. Also, we constructed "anti-CD19 scissors CAR", low affinity "anti-HER2 (4D5-3) scissors CAR" and high affinity "anti-HER2 (4D5-8) scissors CAR". To analyze the target-cell-dependent cleavage of mCherry CAR, 293T cells expressing these CARs were co-cultured with target cells, including K562 (CD19-, HER2-), Raji (CD19+, HER2+), or SK-BR-3 (CD19-, HER2+). To obtain target cells expressing both CD19 and HER2, Raji and SK-BR-3 cells were molecularly manipulated. (1) To evaluate efficiency of this system, after co-cultivation of CAR-transduced 293T and the target cells (K562, Raji, SK-BR-3), the localization of mCherry was examined under the microscopy and Western blotting. (2) To assess the T cell activation, we constructed "anti-CD19 effector CAR" and established Jurkat cells expressing both the "effector CAR" and the "anti-HER2 scissors CAR". These cells were co-cultured with wild type or engineered Raji or SK-BR-3 cells. T cell activation was analyzed with flowcytometric analysis. (3) To assess the CAR activity of this system, we transduced these CARs to primary T cells from healthy donors. After co-cultivation with the target cells, we measured target-specific cytotoxic activity.

【Results】 (1) Transduced "anti-CD19 mCherry CAR" was detected as a membrane-bound protein in 293T cells. Co-cultivation of 293T cells expressing both the "anti-CD19 mCherry CAR" and "anti-HER2 scissors CAR" with engineered Raji cells expressing both CD19 and HER2 induced cleavage of the recognition site and translocation of the mCherry from the membrane to the cytoplasm. These results suggested that this system would regulate activities of CAR-T cell through HIVPR-mediated cleavage of the "effector CAR" in vitro. (2) Jurkat cells expressing "anti-CD19 effector CAR" were activated through a target-cell-dependent manner. Furthermore, "anti-HER2 scissors CAR" attenuated T cell activation driven by "anti-CD19 effector CAR" when Jurkat cells expressing both the "anti-CD19 effector CAR" and the "anti-HER2 scissors CAR" contacted with target-cells expressing both CD19 and HER2. In addition, high affinity 4D5-8 scissors CAR showed more potent attenuation than the low affinity 4D5-3 scissors CAR. (3) Primary human T cells expressing "anti-CD19 effector CAR" showed CD19-expressing target-cell-specific cytotoxic activity. We also demonstrated that "Double Arm" primary human CAR-T cells showed tumor-cell-specificity as seen in cell-line models described above.

【Discussion】 Our "Double-Arm" CAR-T cell system has improved tumor-cell-specificity even in primary human T-cells as we expected. This system would attenuate the adverse effects of clinical CAR-T cell therapies.