Proc. Natl. Acad. Sci. | AI-driven PROTAC design: 人工智能赋能融合蛋白靶向降解，克服CLIP1-LTK驱动肺癌耐药难题

TL;DR - A PNAS study used AlphaFold3-guided structural modeling and computational scoring to design DCL05, a PROTAC that degrades the lung-cancer driver CLIP1-LTK. It retained activity against kinase-inhibitor-resistant mutants, offering a potential route around acquired drug resistance.

• DCL05 recruits CRBN to degrade CLIP1-LTK through the ubiquitin-proteasome pathway.
• The design workflow modeled LTK-PROTAC-CRBN ternary complexes to optimize linker geometry.
• DCL05 achieved a reported DC50 of about 40 pM and over 99% maximum degradation.
• It suppressed resistant CLIP1-LTK tumor models in vitro and in vivo without obvious toxicity in the reported animal studies.