嵌合抗原受体
免疫系统
计算生物学
基因传递
体内
细胞疗法
遗传增强
免疫疗法
离体
病毒载体
医学
从长凳到床边
生物
肿瘤微环境
转化研究
翻译(生物学)
生物信息学
T细胞
癌症研究
效应器
多细胞生物
基因组编辑
移植
细胞
重定目标
计算机科学
免疫学
载体(分子生物学)
作者
Yi-Min Yang,Bo Bao,Yu-Hao Cao,Jin Yao,Yu-Fan Ding,Yi-Yang Hu,Fan Fan,Jun-Long Zhao
标识
DOI:10.1186/s43556-026-00447-y
摘要
In vivo chimeric antigen receptor (CAR) cell therapy is undergoing a transformative shift from conventional ex vivo manufacturing toward in situ cellular editing, aiming to generate functional CAR-engineered immune cells directly within patients through targeted vector delivery, thereby significantly enhancing therapeutic accessibility and applicability. While rapid advances have been made in both viral (lentiviral and adeno-associated viral vectors) and non-viral (lipid nanoparticle) delivery platforms, along with the expansion of effector cell lineages including CAR-T, CAR-NK, and CAR-M, critical translational bottlenecks remain. These include insufficient delivery precision, limited cellular persistence, immunosuppressive tumor microenvironment (TME) resistance, and challenges in safety controllability. This review systematically examines the working mechanisms and limitations of current delivery platforms for in vivo gene transfer. It provides a comprehensive comparison of how CAR-T, CAR-NK, and CAR-M platforms employ distinct yet complementary strategies to address tumor heterogeneity, solid tumor physical and immune barriers, and the specificity constraints of in situ editing. Furthermore, we highlight emerging frontiers such as artificial intelligence-guided personalized therapy design, smart delivery systems (logic-gated CARs, circular RNA vectors), and the development of multicellular synergistic "synthetic immune systems." By integrating multidisciplinary perspectives, this review not only offers a comprehensive roadmap bridging fundamental mechanisms to clinical translation but also lays a theoretical and technical foundation for advancing the next generation of safe, precise, and efficacious in vivo CAR therapies.
科研通智能强力驱动
Strongly Powered by AbleSci AI