Abstract B134: AI-designed photo-immunolipid therapy (PILT) remodels the pancreatic tumor microenvironment to potentiate immune checkpoint blockade

Abstract

Pancreatic ductal adenocarcinoma (PDAC) remains refractory to immune checkpoint blockade because of extensive desmoplasia and a markedly immunosuppressive tumor microenvironment. Multi-modal treatment approaches are needed to remodel the pancreatic tumor microenvironment to a more immunopermissive state. Our team is developing photo-immunolipid therapy (PILT), a bifunctional lipid nanoparticle platform capable of remodeling the pancreatic tumor microenvironment, while simultaneously inducing immunogenic cell death and PD-L1 blockade. Here, we implement supervised and unsupervised machine learning to optimize near-infrared light-activatable lipid nanoparticle composition for maximal immunogenic cell death. In immune-cold syngeneic pancreatic tumor models derived from genetically engineered mouse models, AI-designed PILT disrupted tumor vasculature and stromal barriers to improve intratumoral delivery and distribution of the bifunctional nanoparticle. Bulk RNA sequencing revealed extensive immune transcriptional remodeling of tumors toward an immunopermissive state, which was accompanied by >50% tumor growth inhibition, and overall survival was prolonged by >60%. These findings establish AI-designed PILT as a modular nanomedicine platform for remodeling the pancreatic tumor microenvironment to potentiate immune checkpoint blockade in PDAC.