Branched endosomal disruptor (BEND) lipids mediate mRNA and CRISPR-Cas9 ribonucleoprotein complex gene editing and CAR T cell engineering.

Introduction: Lipid nanoparticles (LNPs) have emerged as the preeminent drug delivery vehicle for nucleic acid therapeutics, as demonstrated by the successful administration of the LNP-based mRNA COVID-19 vaccines. The revolutionary component of LNPs is the ionizable lipid (IL), which serves to both encapsulate the nucleic acids and induce endosome escape, the latter being a major barrier to drug delivery. The canonical IL is composed of straight chain lipid tails; however, recent studies have suggested that terminal branched groups can enhance endosomal escape, creating a more potent LNP. Due to their complexity, only select branched lipids are commercially available, greatly limiting research in this area. Therefore, we have developed a robust two-step synthetic method to produce branched epoxides that can then be reacted with most polyamine core to form branched ILs. These Branched ENdosomal Disruptor (BEND) ILs formulate LNPs that significantly enhance the efficacy of both mRNA and ribonucleoprotein complex mRNA delivery and gene editing in the liver in comparison to both their straight chain versions as well as commercial gold standards. Additionally, we demonstrate that BEND ILs significantly enhance T cell transfection, especially in producing chimeric antigen receptor (CAR) T cells. Lastly, via a combination of artificial endosomes, confocal microscopy, and molecular dynamic (MD) simulations, we discovered that the improved efficacy of BEND lipids is mainly a result of increased endosome destabilization.

Materials and

Methods: Branched epoxides were synthesized via copper-catalyzed Grignard C-C couplings followed by mCPBA-mediated epoxidations (Fig. 1a). The epoxides were then reacted with different lipid cores via SN2 reactions to form the corresponding ILs (Fig. 1b). The resulting ILs were formulated into LNPs encapsulating mRNA using our in-house microfluidic devices with DOPE, C14-PEG2000, and cholesterol (Fig. 1c). LNPs dialyzed in 1X PBS for 2 h and were passed through a 220 nm filter. To analyze LNP biodistribution, C57BL/6J mice were administered the LNPs encapsulating firefly luciferase mRNA at a dose of 0.1 mg/kg. After 6 h, mice were dissected, and luminescence was visualized via an in vivo imaging system (IVIS). Transthyretin (TTR) gene editing was achieved by injecting Cas9 mRNA and TTR sgRNA into C57BL/6J mice at an RNA dose of 1.0 mg/kg. After 7 days, serum TTR levels and liver indels were analyzed by ELISA and next-generation sequencing, respectively. T cell studies were conducted using primary human T cells activated with human CD3/CD28 Dynabeads and dosed at 300 ng of mRNA per 60,000 cells. After 24 h, mCherry and CAR protein expression were evaluated using flow cytometry. To probe endosomal escape, artificial endosomes composed of DOPS, DOPC, and DOPE, and containing a lipid FRET pair, were mixed with the LNPs. The degree of membrane disruption was monitored over time by examining an increase in FRET donor fluorescence. MD simulations were performed with GROMACS using the DOPS/DOPC/DOPE lipid bilayers. ILs were placed 4 nm above the membrane and simulated for 1 µs.

Results, Conclusions, and Discussions: A small library of BEND ILs were successfully synthesized in three steps, and due to the modular nature of the design, the branched epoxides can be utilized for any standard polyamine core. BEND ILs with the “200” ionizable amine core demonstrated up to 5-fold greater liver delivery than C12-200, the gold standard IL that the BEND lipids are modeled after (Fig. 1d). Many of these BEND ILs differ by only five methyl groups, showcasing that small alterations in molecular architecture can drastically improve LNP efficacy. When evaluated for TTR editing, a protein that is overexpressed in TTR-mediated amyloidosis, several BEND LNPs outperformed C12-200 for both TTR serum reduction and liver gene editing, reaching 90% TTR reduction and 60% indels (Fig. 1e-f). When encapsulating Cas9 RNP complex, the top-performing BEND LNPs facilitated liver editing in Ai9 mice, whereas C12-200 mainly edited the lungs (Fig. 1g-j). When BEND ILs were synthesized with the immune-tropic “494” ionizable amine core, the resulting BEND LNPs induced superior transfection in human primary T cells than their linear counterparts (Fig. 1k).

The endosomal escape properties of BEND lipids were investigated by formulating artificial endosomes containing a FRET pair. BEND LNPs elicited superior rupturing compared to the linear controls (Fig. 1l). Additionally, when the same endosomal lipids were modeled as a bilayer through atomistic MD simulations, BEND ILs induced greater penetration into the bilayer than the linear versions (Fig. 1m-n).
These studies reveal essential structure-function relationships that are necessary to accelerate the development of next-generation LNPs. While lipid branching has been utilized in other ionizable lipids designs, our results are the first to explore the possible explanations for this effect in a rigorous and sophisticated manner, pairing both computational and experimental findings to investigate the role of endosomal escape. Moreover, there is a lack of commercially available branched lipids, and thus this project provides a simple and modular synthetic avenue to access lipids with unique terminal branching moieties. We envision that these experiments can be employed as a model for future structure-function studies. Currently, we are investigating the role of lipid branching in enhancing LNP-based vaccines.

Acknowledgements (Optional):