Evaluation of a DoE based approach for comprehensive modelling of the effect of lipid nanoparticle composition on nucleic acid delivery

Abstract

Therapeutic nucleic acids (TNAs) comprise an alternative to conventional drugs for cancer therapy. Recently, stable nucleic acid lipid particles (SNALPs) have been explored to deliver TNA efficiently and safely both in vitro and in vivo. Small interfering RNA (siRNA) and messenger RNA (mRNA) based drugs have been suggested for a wide range of pathologies, and their respective lipid nanoparticle (LNP) formulations have been optimised using a Design of Experiments (DoE) approach. However, it is uncertain as to whether data obtained from DoE using simple experimental outputs can be used to generate a general heuristic for delivery of diverse TNA both in vitro and in vivo. Using plasmid DNA (pDNA), for which limited DoE optimisation has been performed, and siRNA to represent the two extremities of the TNA spectrum in terms of size and biological requirements, we performed a comparative DoE for both molecules and assessed the predictive qualities of the model both in vitro and in vivo. By producing a minimum run of 24 SNALP formulations with different lipid compositions incorporating either pDNA or siRNA, DoE models were successfully established for predicting the effect of individual lipid composition on particle size, TNA encapsulation and transfection both in vitro and in vivo. The results showed that the particle size, and in vitro and in vivo transfection efficiency of both pDNA and siRNA SNALP formulations were affected by lipid compositions. The encapsulation efficiency of pDNA SNALPs but not siRNA SNALPs was affected by the lipid composition. Notably, the optimal lipid compositions of SNALPs for pDNA/siRNA delivery were not identical. Furthermore, in vitro transfection efficiency could not be used to predict promising LNP candidates in vivo. The DoE approach described in this study may provide a method for comprehensive optimisation of LNPs for various applications. The model and optimal formulation described in this study can serve as a foundation from which to develop other novel NA containing LNPs for multiple applications such as NA based vaccines, cancer immunotherapies and other TNA therapies.