Modulation of G3BP1 influences the functional efficacy of CAR T cells against glioblastoma

Background: Recent clinical trials of disialoganglioside GD2 CAR T cells have demonstrated encouraging efficacy in high-risk neuroblastoma and diffuse intrinsic pontine glioma. However, durable responses in solid tumors remain rare, largely due to T cell exhaustion driven by chronic antigen exposure and hostile tumor microenvironments. Antigenic, metabolic, and oxidative stresses activate translational control programs that converge on stress granule assembly in T cells. G3BP1 is a central regulator of stress granule formation, sequestering translationally repressed mRNAs and RNA-binding proteins to reshape protein synthesis during stress. We hypothesize that G3BP1 functions as a stress-buffering checkpoint that tunes CAR T cell fate in response to antigenic and metabolic stress. To test this hypothesis, we engineered GD2-specific CAR T cells with modulated G3BP1 expression and tested their function in vitro and in a preclinical glioblastoma tumor model. Methods: We generated GD2-specific CAR constructs incorporating a GFP-tagged G3BP1 marker (GFP::G3BP1), enabling real-time visualization of G3BP1 localization and stress granule assembly in CAR T cells. Inducible G3BP1 knockdown was achieved using an NFAT promoter-driven shRNA system. Primary human T cells were genetically engineered to express GD2 CAR together with either inducible G3BP1 knockdown or GFP::G3BP1 overexpression. Cytotoxic activity was assessed in coculture assays with GD2-expressing tumor cell lines. Antitumor efficacy was evaluated in a subcutaneous xenograft model of glioblastoma. Results: Exposure of GD2-specific CAR-expressing Jurkat cells to target cells with varying GD2 expression induced robust G3BP1-positive stress granule formation, which was specific to GD2-positive targets. Modulation of G3BP1 expression did not impair CAR T cell expansion during manufacturing or proliferation in vitro. In effector-to-target cytotoxicity assays, all GD2-CAR T cell variants mediated antigen-specific tumor lysis, and altering G3BP1 levels did not affect acute cytolytic capacity, indicating that G3BP1 does not regulate immediate killing function. In a subcutaneous GBM tumor with high GD2 expression, G3BP1 overexpression improved GD2-CAR T cell-mediated tumor control, while NFAT-driven G3BP1 knockdown led to impaired CAR T responses and performed worse than the parental GD2-CAR T product. Conclusions: These studies establish proof-of-concept that rational modulation of G3BP1 and stress granule dynamics can enhance CAR T cell efficacy in solid tumors, including GBM, without compromising acute cytotoxic function.