(M1230-03-17) Strategies for Producing Clinical and Commercial RNA-LNP Drug Products

Purpose: The RNA-lipid nanoparticle (LNP) vaccines for the SARS-CoV-2 pandemic highlight the impact of genomic medicines deployed at scale. RNA-LNP applications also include gene editing, oncology, and rare diseases. Despite this growing momentum, developing RNA-based vaccines and therapeutics still faces significant manufacturing challenges. The mixing process to encapsulate RNA within LNPs is among the most difficult unit operation to scale up to commercial throughput rates and batch sizes, while maintaining critical quality attributes (CQAs) such as size and biological potency. Therefore, we have recently developed a new single-use mixer, enabling pre-dilution RNA-LNP flow rates of up to 800 mL/min. Additionally, we have developed two systems with software that enables 21 CFR Part 11 compliance* (*pending third party audit) and are ATEX/IECEx** (**pending final certification) rated, enabling batch sizes from 0.5 to 3200 L. In this work, we characterize the performance of the new mixer and instrumentation with liposomes and with representative nucleic acid-LNPs. Furthermore, we designed a new SARS-CoV-2 RNA-LNP vaccine candidate, and utilized the complete NanoAssemblr® platform to rapidly scale-up from the research bench to a commercially relevant process.

Methods: POPC (1-palmitoyl-2-oleoyl-glycero-3-phosphocholine):cholesterol liposomes were formulated at flow rates between 2 and 20 mL/min using the NxGen™ mixer; 50 and 200 mL/min using the NxGen 500 mixer; and 200 and 1000 mL/min using the novel NxGen 48 L/h mixer. The size and polydispersity of the liposomes was measured with dynamic light scattering (DLS). Next, plasmid DNA (pDNA) encoding enhanced green fluorescent protein (eGFP) or self-amplifying mRNA (saRNA) encoding the SARS-CoV-2 spike protein, and a custom lipid composition were used as models for an RNA-LNP vaccine. The flow rate for pDNA-LNP formulation was kept constant at 800 mL/min while the pre-dilution formulation volume was stepwise increased from 5 to 50 L using the NxGen 48 L/h mixer. The saRNA-LNP formulation flow rate was stepwise increased from 12 mL/min using the NanoAssemblr® Ignite+™ through to 800 mL/min using the new NanoAssemblr® commercial formulation system and modular commercial formulation skid. N/P ratio, flow rate ratio, and post-formulation diluent were kept constant. The pDNA- and saRNA-LNPs were concentrated with tangential flow filtration (TFF) and diafiltered into cryo-preservation buffer. The formulation and downstream processing steps were accomplished within a 6- hour shift. Size and polydispersity index (PDI) was measured with DLS. For the saRNA-LNP formulations, size and morphology were determined by cryo-electron microscopy. Nucleic acid encapsulation efficiency was determined with either the RiboGreen assay or the PicoGreen assay for saRNA- and DNA-LNPs, respectively. The saRNA integrity, post-encapsulation, was assayed through capillary gel electrophoresis. In vitro expression of the SARS-CoV-2 spike protein was assessed in BHK 570 cells while in vivo expression was assessed in a BALB/c mouse model. Immunization was done by intramuscular injection, with a booster on day 28. Tail bleeds were 1 day before each immunization and mice were culled on day 42. IgG titers were determined by ELISA.

Results: To probe the quality of mixing within the NxGen series of mixers, we used POPC:cholesterol liposomes, which have a size that is sensitive to subtle changes in mixing. We found that 40 nm limit-size liposomes could be formed at flow rates above 6 mL/min on the NxGen mixer, above 80 mL/min on the NxGen 500, and above 400 mL/min on the NxGen 48 L/h. To demonstrate the NxGen 48 L/h mixer with a representative formulation, we then used pDNA-LNPs as a surrogate for an mRNA-LNP vaccine and prepared three batches of increasing volume. The pDNA-LNPs were of high quality from all three batches with particle sizes ranging from 81 to 85 nm, PDI < 0.16, and DNA encapsulation efficiency >98%. Next, we designed a SARS-CoV-2 saRNA-LNP vaccine candidate and stepwise increased the flow rates and batch volumes to mimic the drug development process from discovery through to commercial production. The saRNA-LNPs were of similar quality with sizes ranging from 61 – 83 nm, PDI < 0.18 and RNA encapsulation efficiency > 94%. One way analysis of variance testing showed no significant differences in size, PDI, or encapsulation efficiency when comparing instrument or mixer. In vitro potency assays showed similar dose response curves between samples and consistent EC50 values between 1.3 to 3.6 ng/mL. Finally, in vivo immunization studies showed a robust SARS-CoV-2 specific IgG response in all instrument-mixer combinations with titers varying between 1.94x105 to 2.83x106 ng/mL.

Conclusion: This work demonstrates that production of nucleic acid-LNPs with consistent CQAs can be achieved on a wide range of scales using the NanoAssemblr series of instruments and NxGen series of mixers. Additionally, the NxGen 48 L/h and NanoAssemblr commercial formulation system and modular commercial formulation skid can prepare RNA-LNPs at throughputs needed to meet commercial, regulatory-compliant product goals. Broadly, the introduction of these two new instruments will simplify the production of RNA-LNP drug products for a wide range of applications.