Discovery of novel immunotropic mRNA lipid nanoparticle formulations facilitated by high-throughput in vivo screening

Introduction: Ionizable lipid nanoparticles (LNPs) represent the most clinically advanced non-viral gene delivery vector, forming the basis of clinical gene therapies (such as the Onpattro siRNA therapy for hereditary transthyretin amyloidosis) and vaccines (such as the Moderna and Pfizer/BioNTech COVID-19 vaccines). Beyond vaccines, there are numerous opportunities for RNA LNP therapeutics for immunoengineering applications, including CAR T therapy, solid tumor immunotherapy, autoimmunity, and graft-versus-host disease. While LNPs are an attractive platform for therapeutic gene delivery due to their relative simplicity, tunability, and modularity, relationships between LNP biomaterial structure and functionality remain poorly understood. As such, the LNP subfield is primarily reliant on discovery-based development as work progresses on collecting enough data to enable rational design. Traditional LNP discovery methods rely heavily on low-throughput screening using reporter genes to evaluate functional gene expression. Investigators typically use extensive in vitro or ex vivo screening to identify promising candidates for subsequent in vivo validation; however, in vitro-in vivo correlations in LNP-mediated delivery are notoriously weak. To reduce reliance on in vitro screening approaches and accelerate the LNP discovery process, our group and others have reported the adaptation of next-generation sequencing (NGS) to multiplex the in vivo LNP screening process, greatly increasing throughput compared to traditional screening methods. Generally, LNP discovery has focused on the development of novel ionizable lipids. However, our group and others have shown that the relative composition of lipid excipients within the LNP can also dramatically influence delivery both in vitro/ex vivo and in vivo.

Materials and

Methods: We designed a library of 75 chemically distinct LNPs to simultaneously investigate the influences of both ionizable lipid structure and excipient composition on in vivo mRNA delivery. We used high-throughput, pooled, first-in-vivo screening based on barcoded DNA oligomers to evaluate the ability of these LNPs to deliver mRNA to immune cells (Figure 1a). Each LNP was formulated by combining a unique lipid mixture with an aqueous phase containing both reporter mRNA and a unique DNA oligomer via microfluidic mixing. LNPs were characterized based on size, polydispersity, zeta potential, and entrapment of both nucleic acid cargoes before pooling. We explored two administration routes for immunoengineering applications — i.v. injection, as would be used in e.g., a cancer immunotherapy, and i.m. injection, as in a vaccine. After administering the LNP pool to mice, we isolated blood, spleens, and inguinal lymph nodes and performed fluorescence-activated cell sorting to isolate B cells, T cells, macrophages, and dendritic cells from each sample. We then isolated and amplified barcoded oligomers from these cell populations to produce an NGS library pool. We analyzed the resultant NGS data to identify factors influencing the uptake of LNPs by immune cells in vivo and identify top candidates for LNP-mediated transfection. To confirm the potential of our top candidates for in vivo mRNA delivery, we performed low-throughput validation studies, assessing the intracellular uptake and gross transfection of i.m.-injected LNPs or i.v. injecting Ai9 reporter mice with LNPs encapsulating Cre recombinase-encoding mRNA and measuring tdTomato expression in immune cells using flow cytometry.

Results, Conclusions, and Discussions: Analyzing high-throughput screening results revealed a handful of strong candidates for mRNA delivery to immune cells under i.m. administration, each with dramatic enrichment, suggesting that these new LNP formulations may represent a dramatic improvement in immune transfection (Figure 1b). We observed substantially more promising candidates for mRNA delivery under i.v. injection, though these LNPs generally displayed relatively modest improvement in delivery relative to standard LNP formulations (Figure 1d). Strikingly, we also observed differential influences on LNP uptake by immune cells across injection routes, shining light on fundamental LNP transfection and tropism behavior. Namely, we found ionizable lipid structure to be the primary predictor of immune cell transfection under i.v. administration, with excipient composition seemingly having little effect on LNP uptake. For i.m. injection, however, excipient composition displayed a strong influence on immune cell transfection, with distinct ionizable lipids displaying differential responses to excipient variation, suggesting strong interplay between ionizable lipid structure and excipient composition in determining LNP fate (Figure 1c). Through follow-up validation of lead candidates for i.v. mRNA delivery, we identified potent LNPs capable of transfecting both circulating and splenic leukocytes, observing strong mRNA transfection. Our top identified mRNA LNP candidate demonstrated a nearly 2.5-fold increase in overall peripheral leukocyte transfection rate and over an 8-fold improvement in monocyte transfection rate compared to the clinical control DLin-MC3-DMA (Figure 1e). Altogether, our results provide new insight into design of LNPs for both i.v. and i.m. administration and identify novel LNP formulations with great promise for immunoengineering applications.

Acknowledgements (Optional):