CELL INTELLIGENCE · 肿瘤细胞治疗研究

Non-small cell lung cancer (NSCLC) remains difficult to treat with adoptive cell therapies because of antigen heterogeneity, immunosuppressive tumor microenvironments, stromal barriers, and manufacturing variability. Direct evidence for iPSC-derived CAR-NK EV/sEV therapy in NSCLC is currently unavailable. Accordingly, this narrative review evaluates the platform as a translational hypothesis by integrating mechanistic evidence from CAR-engineered EV studies, biological-analog evidence from NK-cell EVs and other engineered EV systems, and clinical-analog evidence from living iPSC-NK or CAR-NK products. The proposed platform could combine a renewable producer-cell source, antigen-directed vesicle binding, and cytotoxic cargo delivery; however, each component, and particularly their integration into a single reproducible product, requires direct experimental validation. Compared with living cell products, CAR-NK sEVs may reduce selected risks related to cellular expansion, persistence, graft-versus-host disease, cytokine release syndrome, and neurotoxicity; however, they do not inherently eliminate antigen-dependent on-target/off-tumor toxicity. Their nanoscale size is hypothesized to improve access to selected stromal barriers in preclinical models, but human solid-tumor penetration remains unproven. Major unresolved issues include rapid systemic clearance and sequestration by hepatic and splenic components of the mononuclear phagocyte system, route-dependent pulmonary deposition, mucus and mucociliary clearance, pulmonary macrophage uptake, subtype- and lesion-specific antigen heterogeneity, EV identity and purity, CAR-positive vesicle quantification, potency-adjusted manufacturing yield, lot comparability, and the absence of potency assays validated across laboratories or linked to clinical outcomes. Thus, iPSC-CAR-NK sEVs should currently be viewed as an investigational translational concept rather than a clinically ready NSCLC therapy.