Quantitative prediction of siRNA complexation by ionizable drugs enables their codelivery in nanoparticles

The ionizable lipid in lipid nanoparticles can be replaced with ionizable drugs to encapsulate small interfering RNA (siRNA) and allow intracellular codelivery. We wondered whether we could develop a predictive model to aid in formulation design. A small-scale screening assay was designed to evaluate siRNA complexation by ionizable drugs at low pH and validated experimentally with ionizable drug nanoparticle (IDNP) formulations. We found that siRNA complexation could be predicted by drug hydrophobicity, aromaticity, proximity of nitrogen and oxygen atoms to aromatic rings, and a machine learning model using five molecular descriptors encoding pharmacophore and structural information. For complexing drugs, siRNA encapsulation efficiency in IDNPs was related to hydrophobicity, molar refractivity, chiral centers, hydrogen bond donors, and topological charge. Netarsudil was predicted to encapsulate siRNA at high efficiency and was thus tested experimentally with siRNA targeting connective tissue growth factor (CTGF) in fibrotic human trabecular meshwork cells: Reduced CTGF mRNA expression and actin network density were observed. These predictive tools may unlock combination therapies.