(20e) Highly Potent mRNA Delivery In Vivo with Intravenously-Administered Ionizable Lipid Nanoparticles

Introduction: Messenger RNA (mRNA) has tremendous therapeutic applications, including protein replacement therapy, cancer immunotherapy, and genomic engineering. However, successful intracellular delivery of mRNA remains challenging, as nucleic acids cannot readily cross cellular membranes alone due to their relatively large size, negative charge, and hydrophilicity. Over the past decade, researchers have developed lipid nanoparticles (LNPs) to encapsulate and deliver small interfering RNAs (siRNAs) to silence protein expression in the liver, but these LNPs have only recently been reported to encapsulate and deliver mRNA to produce protein in vivo. Because of the considerable size, stability, and charge differences between siRNA and mRNA, we hypothesized that LNP formulation parameters of historically-used siRNA-LNPs could be better optimized to more efficiently deliver larger mRNA payloads. To this end, we developed a generalized approach for optimizing LNP formulations for mRNA delivery in vivo using Design of Experiment (DOE) methodologies. After using this strategy to significantly improve mRNA-LNP potency in vivo, we then applied this mRNA-optimized formulation with new classes of ionizable lipid materials and discovered some of the most efficacious intravenously-administered mRNA delivery vehicles reported to date in the literature.

Materials / Methods: The most common siRNA-LNP formulations contain four components: an amine-containing ionizable lipid, phospholipid, cholesterol, and lipid-anchored polyethylene glycol, the relative ratios of which can have profound effects on the formulation potency. We first encapsulated mRNA encoding for Erythropoietin (EPO), a secreted serum protein, into the original siRNA-based LNP formulation (which employed the ionizable lipid C12-200) to establish a baseline from which to improve. LNPs were synthesized via microfluidic mixing of an ethanolic lipid phase with an aqueous RNA phase to form particles approximately 100 nm in diameter. Then, we employed DOE to modulate several formulation parameters simultaneously (e.g. lipid:mRNA weight ratio, phospholipid structure, molar composition of excipients). DOE allowed for the reduction of the total number of formulations required to deduce statistically significant trends in our large, multi-dimensional design space, which is important for the economical screening of LNP formulations due to poor in vitro/in vivo translatability and costs associated with screening large libraries in vivo. EPO mRNA-loaded C12-200 LNPs were injected intravenously at 15 µg total mRNA per mouse, and blood was collected 6 hours post-injection; an ELISA assay was then used to quantify serum EPO levels. Using Definitive Screening and Fractional Factorial DOE Designs, we screened a total of 38 distinct LNPs in vivoand characterized top-performing LNPs via their mRNA encapsulation efficiency, diameter, polydispersity, pKa, zeta potential, and morphology. Then, we applied the formulation parameters from the most efficacious C12-200 LNP formulation to a new class of ionizable lipid materials formed through a ring opening reaction between alkenyl epoxides and a polyamine core; unlike C12-200, which has fully saturated hydrocarbon tails, this new library of alkenyl amino alcohol (AAA) ionizable lipids contained degrees of unsaturation in its tails.

Results: After employing this DOE-based optimization process, we significantly (p < 0.05) increased the efficacy of intravenously-delivered liver-targeting C12-200 LNPs seven-fold for EPO mRNA delivery relative to the formulation previously used for siRNA delivery. At a 15 µg (0.75 mg/kg) mRNA dose, we achieved serum EPO protein concentrations in mice of approximately 7,000 ng/mL, which is 5-6 orders of magnitude greater than physiological levels. We found that increased lipid:RNA weight ratios and the use of the phospholipid 1,2-dioleoyl-sn­-gylcero-3-phosphoethanolamine (DOPE) were the most crucial factors in determining LNP efficacy. Interestingly, when our mRNA-optimized C12-200 LNP formulation was loaded with siRNA, it silenced protein expression in hepatocytes nearly identically to the original siRNA formulation, suggesting that siRNA-loaded LNPs may be more tolerant than mRNA-loaded LNP of formulation design space differences. Lastly, when C12-200 was substituted with a library of novel alkenyl amino alcohol (AAA) lipids and formulated into LNPs with EPO mRNA, lead lipid OF-02 was discovered to afford EPO serum concentrations of approximately 14,000 ng/mL, the highest efficacy yet reported in the literature. Furthermore, mRNA delivered with OF-02 LNPs was expressed predominately in the liver and had a linear dose-response curve ranging from 0.75 to 2.25 mg/kg RNA dose.

Conclusions: In conclusion, we significantly increased the potency of liver-targeting C12-200 LNPs for mRNA delivery using Design of Experiment methodologies, while showing that the same mRNA-optimized LNPs did not confer additional potency when delivering siRNA. The optimized LNP formulation showed applicability to other types of ionizable lipids, such as the novel lipid OF-02, which made highly potent intravenously-delivered mRNA-LNPs. This work provides evidence that some technologies previously used for siRNA delivery may require further optimization for better mRNA delivery. Additionally, the general strategy presented here could be used to screen multi-dimensional design spaces in vivo for optimizing other biomaterials and payloads.