Optimized Lipid Nanoparticles for Co-Delivery of mRNA and siRNA Therapeutics in Refractory Liver Cancer.

Hepatocellular carcinoma (HCC) exhibits poor prognosis and rapid resistance to sorafenib, particularly involving p53 loss and Nrf2 hyperactivation. Here, we employ machine learning (ML)-assisted structure-activity relationship (SAR) analysis to guide the engineering of a library of 120 degradable ionizable lipids, enabling the rational design of fluorinated aromatic lipid nanoparticles (LNPs) optimized for combinatorial RNA delivery. ML-based feature-importance analysis prioritizes -CF aromatic tails, and molecular dynamics simulations confirm that these tails enhance RNA binding and nanoparticle stability. The resulting AT-s LNPs, functionalized with lactobionic acid for selective HCC targeting, enable efficient co-delivery of p53 mRNA and Nrf2 siRNA. This strategy restores ferroptosis and induces apoptosis in sorafenib-resistant HCC by suppressing SLC7A11, leading to marked tumor inhibition. Our study demonstrates an ML-assisted LNP optimization strategy, advancing precision RNA therapeutics to overcome resistance in refractory liver cancer.