Dual-Targeting CAR-NK Cells for Recurrent/Progressive Glioblastoma and High-Grade Glioma

Rationale: GBM/HGG is molecularly heterogeneous and may evade single-antigen immunotherapy via antigen loss or heterogeneous antigen expression. This study evaluates a dual-target CAR-NK strategy that can recognize two tumor-associated antigens selected from IL13Rα2, EGFR/EGFRvIII, and B7-H3 (CD276). Target assessment and cohort assignment: At screening, archived and/or fresh tumor tissue (from clinically indicated resection or biopsy) will be evaluated by immunohistochemistry (IHC) for IL13Rα2, EGFR, and B7-H3, and by molecular testing (e.g., RT-PCR or NGS) for EGFRvIII when applicable. Participants will be assigned to one of three biomarker-defined cohorts based on the two highest-priority antigens detected above protocol-defined thresholds. An internal review committee may recommend prioritizing one construct for later expansion based on emerging safety/feasibility/activity signals. Investigational product: The investigational products are off-the-shelf allogeneic CAR-NK cells manufactured from a standardized NK-cell source (e.g., iPSC-derived NK cells or a qualified NK master cell bank) and genetically modified to express a tandem (dual-target) CAR. The constructs include a built-in safety switch (e.g., inducible caspase-9) and may include an NK-support cytokine module (e.g., IL-15) to enhance persistence. The exact dual-target construct used for each participant depends on the screening antigen profile.

Route of administration: CAR-NK cells will be delivered locoregionally to the surgical cavity wall and/or into the ventricular system via a neurosurgically placed catheter/reservoir, to maximize exposure at the tumor site while limiting systemic exposure. Dose escalation and expansion: Each biomarker-defined cohort follows a modified 3+3 dose-escalation schema with up to three dose levels (e.g., 1×10^7, 3×10^7, and 1×10^8 CAR-NK cells per infusion). Dose-limiting toxicities (DLTs) will be assessed during the first 28 days after the first infusion. After an RP2D is identified within a cohort, an expansion stage will further characterize safety and preliminary activity and may allow repeat infusions at the RP2D. Follow-up: Participants will be monitored closely for adverse events including cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). Tumor response will be assessed with MRI using RANO criteria at regular intervals. Long-term follow-up for gene-modified cell therapy safety will be conducted per applicable guidance.