CLDN18.2-targeted CAR-NK cells suppress gastric cancer growth in preclinical models.
TL;DR
Preclinical proof-of-concept for a rationally designed CAR4-NK therapy against CLDN18.2-positive gastric cancer, in which CD28H-2B4 co-stimulatory domains enhance the effector function of CAR4-NK cells, and show the potential of irradiated CAR4-NK cells as a safer, controllable "off-the-shelf" product candidate.
Abstract
Targeted immunotherapy for gastric cancer (GC) remains challenging due to safety concerns and limited efficacy against solid tumors. Claudin-18.2 (CLDN18.2) is a highly specific tumor-associated antigen, yet existing CLDN18.2-targeted CAR-T therapies are frequently complicated by cytokine release syndrome (CRS) and on-target/off-tumor toxicity. Natural killer (NK) cells represent a safer "off-the-shelf" alternative, but require optimized CAR architectures for effective solid tumor targeting. Here, we designed five CAR constructs targeting CLDN18.2 through combinatorial optimization of single-chain variable fragments and intracellular co-stimulatory domains (including CD28H, 2B4, 4-1BB, and CD3ζ), and delivered them into NK-92 cells via a baboon envelope glycoprotein-pseudotyped lentiviral vector. The lead candidate CAR4-NK, which harbors a CD28H-2B4-CD3ζ signaling domain, exerted superior in vitro cytotoxicity against CLDN18.2-positive GC cells, accompanied by robust secretion of IFN-γ and granzyme B. In NCG mouse xenograft models, intravenously administered CAR4-NK cells specifically trafficked to tumor sites, significantly suppressed tumor growth, and improved overall survival (100% survival at the study endpoint). The CD28H-2B4 combination was shown to be critical for sustained NK cell activation. Notably, irradiated CAR4-NK cells, despite abrogated proliferative capacity, retained tumoricidal efficacy comparable to unmodified parental NK-92 cells and induced minimal production of the inflammatory cytokines TNF-α and MIP-1α, suggesting a reduced risk of CRS-associated cytokine release within the tested panel. These results provide preclinical proof-of-concept for a rationally designed CAR4-NK therapy against CLDN18.2-positive gastric cancer, in which CD28H-2B4 co-stimulatory domains enhance the effector function of CAR4-NK cells, and show the potential of irradiated CAR4-NK cells as a safer, controllable "off-the-shelf" product candidate.