The therapeutic success of mRNA critically depends on delivery systems capable of robust and organ-selective in vivo targeting. However, achieving such precision remains difficult because the relationships between nanoparticle composition and in vivo delivery outcomes are highly nonlinear and cannot be fully resolved through conventional empirical screening. Here, we present a quantitative design framework that integrates an I-optimal design-of-experiments (DOE) strategy with predictive regression modeling to define the formulation-biodistribution relationships of polymer-lipid integrated nanoparticles (PLINs). This PLIN architecture, composed of amphiphilic polyesters and cationic/ionizable lipids and designed without cholesterol or helper lipids, expands the accessible chemical design space beyond that of conventional lipid nanoparticles. The I-optimal DoE identifies a minimal yet information-rich set of 15 formulations, enabling the construction of predictive regression models (R2 > 0.96) that quantitatively link formulation parameters to organ-specific mRNA expression in vivo. Model-guided optimization yields PLINs with up to 91% lung or 96% spleen selectivity, and mechanistic studies reveal the key physicochemical determinants underlying these organ-tropic behaviors. The resulting design rules are generalized across diverse therapeutic contexts, including somatic genome editing, cytokine delivery for acute lung injury, and antigen-specific cancer vaccination. Together, this model-driven and material-divergent framework establishes a quantitative and generalizable principle for engineering organ-targeted mRNA delivery systems.
Wang et al. (Fri,) studied this question.