(24c) Self-Amplifying RNA with Modified Nucleotides Improves Potency and Safety

Introduction: Recent advances in messenger RNA (mRNA) technologies have revolutionized vaccines. mRNA expresses viral antigens to trigger immune responses against infection.¹ Lipid nanoparticles (LNPs) provide a delivery platform for mRNA vaccines by stabilizing and protecting mRNA from degradation and allowing mRNA cell internalization and translation. More than 1 billion people worldwide have received at least one dose of COVID mRNA LNP vaccines which have prevented at least tens of millions of deaths.¹ However, current RNA-based technologies are limited by low transfection efficiency in vivo and transient payload expression, and often require frequent or higher dosing to achieve vaccine or therapeutic efficacy. Self-amplifying RNA (saRNA) enables prolonged antigen expression with less frequent dosing.² Inspired by an RNA virus, a saRNA encodes a virus-derived RNA-dependent RNA polymerase (RdRp) on the same RNA transcript as the sequence for the target antigen. The RdRp replicates and amplifies the RNA sequence encoding the antigen, increasing the expression level and duration of the translated target antigen. While wild-type (WT) saRNA shows promise, it triggers a strong innate interferon response which inhibits saRNA replication and protein expression. Incorporating modified nucleotides into saRNA is one approach to address this problem, but they must be compatible with the RdRp machinery. The modified nucleotides used to reduce interferon responses in mRNA vaccines, such as N1-methylpseudouridine, are not saRNA-compatible.³ We hypothesized that substitution with appropriate modified nucleotides will lead to higher and more durable cargo expression than conventional (N1-methylpseudouridine -modified) mRNA and WT saRNA and lower interferon responses compared to WT saRNA.

Methods: We synthesized conventional mRNA and Venezuelan equine encephalitis virus-derived saRNA by in vitro transcription, encoding mCherry or firefly luciferase as the cargo protein. Modified nucleotides such as 5-hydroxymethylcytidine (hm5C) and 5-methylcytidine (m5C) were incorporated during in vitro transcription to produce modified saRNA. Double-stranded RNA (dsRNA) was removed after RNA synthesis by cellulose chromatography and confirmed by a dot-blot analysis.⁴ To screen modified nucleotides, saRNA encoding mCherry was transfected in HEK293T cells using Lipofectamine MessengerMax Transfection Reagent, and transfection efficiency was determined by flow cytometry. Conventional mRNA, WT and modified saRNAs were encapsulated in LNPs composed of 1.5 mol% DMG-PEG2000, 38.5 mol% cholesterol, 10 mol% phospholipids (2 types) and 50 mol% ionizable lipids (14 types). LNP size and polydispersity index were characterized by dynamic light scattering. RNA encapsulation efficiency was quantified by the QuantiFluor RNA System. We screened LNPs in a C2C12 murine myoblast co-culture to determine the optimal formulation for cell transfection. We administered mRNA, WT or modified saRNAs encoding firefly luciferase in the optimized LNPs to female C57BL/6 mice intramuscularly in the hindlimb. We monitored luciferase expressions over time in vivo by IVIS imaging. We drew blood one day after vaccination to determine serum IFN-α level. We weighed mice regularly to monitor weight loss as an indicator of adverse responses to the RNA vaccines.

Results: hm5C-modified and m5C-modified saRNAs led to the highest transfection efficiency among the modified nucleotides screened. LNP screen showed that 1,2-dioleoyl-sn-glycero-3-PE as the phospholipid and SM-102 as the ionizable lipid was the optimal formulation for C2C12 cell transfection. The LNPs had a diameter of ~100 nm, low polydispersity index (~0.1 or less) and RNA encapsulation efficiency >90%. LNPs loading m5C-modified saRNA mediates a 314x luciferase expression compared to mRNA, and a 3.5x expression compared to WT saRNA in C2C12 cells in vitro. At a 2.5 μg dose, m5C-modified saRNA LNPs induced a peak luciferase expression 3.1x higher than mRNA LNPs, and the expression was maintained for greater than a month in vivo, whereas the expression from mRNA LNPs declined ~100-fold in a week. Moreover, m5C-modified saRNA reduced IFN-α response by more than 50% compared to WT saRNA in vivo. We next investigated the effect of dsRNA on luciferase expression in vivo, and observed that while WT saRNA with or without dsRNA removal showed similar expression kinetics, dsRNA removal significantly reduced weight loss. Lastly, we administered 3, 10 or 30 μg of mRNA, WT, hm5C-modified and m5C-modified saRNAs to determine the dose-response relationship and safety profile. Interestingly, all saRNAs induced comparable peak expression levels and areas under the curve across all doses, whereas mRNA reached peak expression and area under the curve with the intermediate 10 μg dose. Furthermore, at the 30 μg dose, both modified saRNAs led to lower weight loss and faster weight recovery in mice than WT saRNA.

Conclusions: Substitution with compatible modified nucleotides in saRNA significantly prolongs payload expression and reduces interferon responses. Modified saRNAs maintain high payload expression and are safe in vivo even at a high dose.

Acknowledgement: The work was supported by the NSF Trailblazer Award and Boston University. We thank the BU IVIS Imaging Core (1S10RR024523-01) for support.

References: 1. Pardi N, Nat Rev Drug Discov. 2024. 2. Bloom K, Gene Ther. 2021. 3. Minnaert AK, Adv Drug Deliv Rev. 2021. 4. Baiersdorfer M, Mol Ther Nucleic Acids. 2019.