VHH-based CAR-T cells targeting Claudin 18.2 show high efficacy in pancreatic cancer models

Ying Xing¹Gangqiang Shi²Zhengli Li²Xiancheng Liu²Yu Zhang¹Yiqin Song¹Shilin Sun¹Yuping Chen^1*Xuekai Zhu^2*Mike K S Chan³Michelle B F Wong³Thomas Skutella⁴Krista Casazza^5*Jonathan R T Lakey⁶Yunfeng Feng^2*

• ¹Department of Clinical Research for Pediatric Respiratory and Gastrointestinal Diseases, The Children's Hospital of Baoding City, Hebei, China
• ²Celest Therapeutics (Shanghai) Co, Shanghai, China
• ³European Wellness Biomedical Group, Edenkoben, Germany
• ⁴Universitat Heidelberg Institut fur Anatomie und Zellbiologie, Heidelberg, Germany
• ⁵University of California, Irvine, Irvine, United States
• ⁶University of California Irvine, Irvine, United States

Background: Pancreatic ductal adenocarcinoma (PDAC) remains one of the deadliest cancers, with a 5-year survival rate below 10%, largely due to late-stage diagnosis and the limited effectiveness of conventional therapies such as surgery, chemotherapy, and radiation. Claudin 18.2 (CLDN18.2)has emerged as a promising target for PDAC. While single-chain variable fragment (scFv)-based CAR-T cells targeting CLDN18.2 have demonstrated therapeutic potential, CAR-T cells engineered with variable heavy-chain-only domains (VHH) exhibit superior efficacy, highlighting the advantages of VHH-based constructs in targeting this antigen. However, the therapeutic efficacy of anti-CLDN18.2 VHH-CAR T cells remains to be fully elucidated, as previous studies have not comprehensively characterized in vivo performance or mechanistic advantages over scFv-based counterparts. Methods: To characterize the therapeutic potential of anti-CLDN18.2 VHHs, we employed phage display technology to screen a VHH library, resulting in the identification of three positive clones. These candidates were further evaluated and ranked based on binding affinity and multi-round cytotoxicity in Chimeric antigen receptor T (CAR T) cell models. To reduce immunogenicity, the lead VHH was humanized. VHH-CAR T cells incorporating this humanized domain were assessed through in vitro assays measuring cytokine secretion and target cell lysis, followed by in vivo studies to evaluate antitumor efficacy in relevant xenograft models. Results: High-affinity anti-CLDN18.2 VHHs from phage libraries and engineered CAR-T cells using HM2, a humanized VHH, as the antigen-binding domain were successfully identified. Notably, HM2-CAR T cells demonstrated potent and sustained cytokine secretion and cytotoxic activity against CLDN18.2-expressing tumor cells in vitro. More importantly, these VHH-based CAR-T cells achieved significant antitumor efficacy in vivo, underscoring the translational potential of VHH-CAR constructs as a next-generation therapeutic platform with enhanced performance and reduced immunogenicity compared to conventional scFv-based designs. Conclusion: This study establishes an effective framework for developing CLDN18.2-specific VHHs and demonstrates their successful integration into CAR-T cell therapy. The humanized HM2-CAR T cells not only maintain high antigen specificity but also exhibit strong effector functions and pronounced antitumor activity in preclinical models. These findings support the clinical promise of VHH-based CAR-T cells as a next-generation immunotherapy for CLDN18.2-expressing malignancies, particularly PDAC, where effective treatment options remain limited.