introduction: Chimeric antigen receptor T (CAR-T) cell therapy presents a promising treatment option for various cancers, including solid tumors. Carcinoembryonic antigen (CEA) is an attractive target due to its high expression in many tumors, particularly gastrointestinal cancers, while limited expression in normal adult tissues. In our previous clinical study, we reported a 70% disease control rate with no severe side effects using a humanized CEA-targeting CAR-T cell. However, the selection of the appropriate single-chain variable fragment (scFv) significantly affects the therapeutic efficacy of CAR-T cells by defining their specific behavior towards the target antigen. Therefore, this study aimed to identif... More
introduction: Chimeric antigen receptor T (CAR-T) cell therapy presents a promising treatment option for various cancers, including solid tumors. Carcinoembryonic antigen (CEA) is an attractive target due to its high expression in many tumors, particularly gastrointestinal cancers, while limited expression in normal adult tissues. In our previous clinical study, we reported a 70% disease control rate with no severe side effects using a humanized CEA-targeting CAR-T cell. However, the selection of the appropriate single-chain variable fragment (scFv) significantly affects the therapeutic efficacy of CAR-T cells by defining their specific behavior towards the target antigen. Therefore, this study aimed to identify the optimal scFv and investigate its biological functions to further optimize the therapeutic potential of CAR-T cells targeting CEA-positive carcinoma.
methods: We screened four reported humanized or fully human anti-CEA antibodies (M5A, hMN-14, BW431/26, and C2-45), and inserted them into a 3rd-generation CAR structure. We purified the scFvs and measured the affinity. We monitored CAR-T cell phenotype and scFv binding stability to CEA antigen through flow cytometry. We performed repeated CEA antigen stimulation assays to compare the proliferation potential and response of the four CAR-T cells, then further evaluated the anti-tumor efficacy of CAR-T cells ex vivo and in vivo.
results: M5A and hMN-14 CARs displayed higher affinity and more stable CEA binding ability than BW431/26 and C2-45 CARs. During CAR-T cell production culture, hMN-14 CAR-T cells exhibit a larger proportion of memory-like T cells, while M5A CAR-T cells showed a more differentiated phenotype, suggesting a greater tonic signal of M5A scFv. M5A, hMN-14, and BW431/26 CAR-T cells exhibited effective tumor cell lysis and IFN-γ release when cocultured with CEA-positive tumor cells , correlating with the abundance of CEA expression in target cells. While C2-45 resulted in almost no tumor lysis or IFN-γ release. In a repeat CEA antigen stimulation assay, M5A showed the best cell proliferation and cytokine secretion levels. In a mouse xenograft model, M5A CAR-T cells displayed better antitumor efficacy without preconditioning.
discussion: Our findings suggest that scFvs derived from different antibodies have distinctive characteristics, and stable expression and appropriate affinity are critical for robust antitumor efficacy. This study highlights the importance of selecting an optimal scFv in CAR-T cell design for effective CEA-targeted therapy. The identified optimal scFv, M5A, could be potentially applied in future clinical trials of CAR-T cell therapy targeting CEA-positive carcinoma.
