A novel AI-engineered biparatopic DLL3/CD3 T cell engager demonstrates potent preclinical efficacy and a promising safety profile

Purpose: This study aimed to develop and characterize a novel biparatopic T-cell engager (TCE) targeting DLL3 for the treatment of small cell lung cancer (SCLC) and other neuroendocrine carcinomas (NECs). Methods: Using an AI-engineered multispecific antibody platform, we engineered a biparatopic DLL3/CD3 TCE to bind two distinct epitopes on DLL3 while engaging CD3 on T cells. Binding affinities were systematically optimized to enhance tumor targeting and minimize off-T cell activation. The molecule was evaluated in vitro activity for T cell-dependent cytotoxicity, T cell activation, and cytokine release and in vivo efficacy using human SCLC cell lines-derived xenograft models. Safety was assessed in transgenic mice and cynomolgus monkeys. Results: The biparatopic DLL3/CD3 TCE demonstrated potent and specific cytotoxicity against DLL3-expressing cell lines, while limiting T cell activation and cytokine release. In murine models, it achieved significant tumor growth inhibition at lower doses than a benchmark TCE (tarlatamab analog). The molecule exhibited good tolerability at doses up to 30 mpk in transgenic mice and 10 mpk (administered via step-dosing) in monkeys, without significant adverse effects. Conclusion: The biparatopic DLL3/CD3 TCE exhibits robust preclinical efficacy and a favorable safety profile, supporting its further development as a targeted immunotherapy for DLL3-positive SCLC and NECs.