Systematic optimization of PD-L1 and CLDN18.2 CAR-T designs identifies a bicistronic dual-target, double-CD3ζ architecture with enhanced antitumor activity in gastric cancer

Background Gastric cancer is a leading cause of cancer-related mortality, with limited efficacy of immunotherapies due to intratumoral heterogeneity, an immunosuppressive microenvironment, and antigen-loss escape. The optimization of chimeric antigen receptor (CAR) T-cell therapy targeting PD-L1 and Claudin18.2 (CLDN18.2) is crucial for improving the antitumor response. Methods We performed a stepwise optimization of CAR designs targeting PD-L1 and CLDN18.2, focusing on the costimulatory domains and intracellular signaling architecture. We engineered monospecific PD-L1 and CLDN18.2 CAR-T cells using CD28 or 4-1BB costimulatory domains and assessed their CAR expression, activation phenotype, subset composition, and cytotoxicity. We then engineered three bicistronic dual-target CARs with distinct CD3ζ configurations and evaluated their functionality through in vitro assays and in vivo NUGC4-Luc gastric cancer xenograft models. Results The single-target CLDN18.2-CD28 CAR-T cells exhibited the strongest cytotoxicity in vitro . Among the bicistronic designs, the double-CD3ζ CAR-T cells showed superior expansion, differentiation, and cytotoxicity compared to alternative bicistronic constructs and single-target CAR-T cells. Bulk RNA sequencing revealed enhanced immune activation and reduced exhaustion markers in the double-CD3ζ CAR-T cells. In the NUGC4-Luc xenograft model, the double-CD3ζ CAR-T cells achieved a tumor growth inhibition of 63.23%, outperforming the 48.06% inhibition observed with single CLDN18.2 CAR-T cells. Conclusions The study identifies a bicistronic dual-target, double-CD3ζ CAR-T design that enhances T-cell functional fitness and antitumor efficacy in gastric cancer models. This approach offers a promising strategy for addressing antigen heterogeneity and improving the durability of CAR-T therapies in solid tumors.