Overcoming thermostability challenges in mRNA-lipid nanoparticle systems with piperidine-based ionizable lipids

Messenger RNA (mRNA) therapeutics and vaccines offer rapid development and programmable protein expression, exemplified by SARS-CoV-2 vaccines. Lipid nanoparticles (LNPs) enable in vivo mRNA delivery by protecting cargo and facilitating endosomal escape, with ionizable lipids being critical through electrostatic interactions with mRNA. Despite advances in potency, long-term storage stability of mRNA/LNPs remains a major hurdle, as mRNA undergoes hydrolysis and oxidation and LNP components degrade under thermal stress. Cold-chain solutions at −20 °C or below are costly and logistically burdensome; lyophilization can improve stability but adds complexity and potential reconstitution damage. Recent work indicates that aldehyde impurities formed from ionizable lipid oxidation/hydrolysis can covalently react with mRNA nucleosides and inactivate translation, implicating the ionizable amine headgroup as a source. The authors hypothesized that strategic redesign of the amine structure, particularly heterocyclic amines, could limit aldehyde generation and prevent mRNA-lipid adduct formation, thereby enabling liquid storage at 4 °C.

Prior studies have established LNPs as effective mRNA delivery vehicles, with ionizable lipids enabling endosomal escape and in vivo efficacy (Cullis & Hope 2017; Han et al. 2021). Lyophilization has been explored to enhance thermostability of mRNA/LNPs, though it may affect particle properties and requires additional processing (Muramatsu et al. 2022; Ai et al. 2023; Meulewaeter et al. 2023). A key mechanistic insight by Packer et al. (2021) proposed that oxidative N-oxide formation and subsequent hydrolysis of ionizable lipids yields hydrophobic fatty aldehydes that form covalent adducts with mRNA nucleosides, inhibiting translation. Heterocyclic amine-containing ionizable lipids have been reported to improve immune responses and mRNA delivery (Miao et al. 2019; Ripoll et al. 2022). The authors’ prior work on branching in ionizable lipids (Hashiba et al. 2023) indicated benefits for mRNA delivery. Together, these reports frame the open challenge: reducing aldehyde-driven mRNA inactivation while maintaining or improving delivery potency, ideally enabling refrigerated liquid storage.

Study design: The authors synthesized and evaluated ionizable lipids featuring heterocyclic amine headgroups to mitigate aldehyde impurity formation. They developed a library of 23 N-methyl piperidine-based ionizable lipids (CL15F m–n), varying tail branching and length, and compared them to a prior linear tertiary amine series (CL4F). Lipids were purified by reverse- and normal-phase chromatography and characterized by NMR and MS. LNP formulation: LNPs were prepared by microfluidic mixing (NanoAssemblr Benchtop) of ionizable lipid:cholesterol:DSPC:DMG-PEG2000 at a fixed molar ratio of 50:38.5:10:1.5 with mRNA in citrate buffer (50 mM, pH 3.5), at N/P = 10. Formulations were buffer-exchanged to HEPES/sucrose and stored as liquid at 4 °C or frozen at −80 °C under nitrogen. Apparent pKa was measured by TNS assay; size, PDI, and zeta potential were measured by DLS; encapsulation by RiboGreen. In vitro assays: HEK-293T cells were transfected with LNP-encapsulated firefly luciferase (FLuc) mRNA (100 ng/well) to assess functional delivery vs CL4F LNPs and Lipofectamine MessengerMAX. In vivo studies: For vaccination potential, C57BL/6 mice received intramuscular OVA mRNA LNPs (prime/boost). Anti-OVA IgG titers (ELISA) and OVA-specific IFN-γ ELISpot were measured. For storage stability, Balb/c mice were injected i.v. with hEPO mRNA LNPs (0.25 mg/kg) or FLuc mRNA LNPs (0.1 mg/kg) before and after LNP storage at −80 °C or 4 °C; serum hEPO was quantified by ELISA and liver bioluminescence by IVIS. Analytical chemistry: Lipid component integrity after storage was profiled by HPLC with charged aerosol detection (CAD). Aldehyde impurity levels were estimated by a fluorescence microplate assay using NBD-H, which forms fluorescent hydrazones with carbonyls. Fatty aldehydes were identified by DNPH derivatization followed by UPLC-QTOF-MS (UV at 360 nm, HRMS and isotopic patterns). N-oxidation of lipids was performed with hydrogen peroxide and monitored by 1H-NMR and HPLC/MS; oxidized species localization was probed by high-energy HCD-MS/MS fragment ions diagnostic for tertiary amine N-oxide. Adduct formation assays: Adducts between lipid-derived aldehydes and nucleosides were evaluated by HILIC-MS after incubating nucleosides with CL4F lipids at 60 °C. mRNA–lipid adduct formation was quantified by reversed-phase ion-pair chromatography of extracted mRNA after incubation with lipids or isolated N-oxide species; late-eluting peaks indicated increased hydrophobicity due to adducts. Chemical scavenging of aldehydes was performed using polystyrene sulfonyl hydrazide (PS-TsNHNH2) with acetic acid; secondary amines were removed with PS-isocyanate, and effects on adducts were assessed. Structure–function exploration: Additional cyclic amine headgroup lipids (CL6F, CL16F, CL17F; fixed 14-12 tail) were synthesized to test a mechanistic hypothesis that intramolecular cyclization to 5- or 6-membered imines from N-oxide/hydrolysis would limit release of reactive fatty aldehydes. Their aldehyde levels (NBD-H), adduct formation (ion-pair RP), and in vivo mRNA expression before vs after 60-day 4 °C storage were evaluated.

• Piperidine-based CL15F lipids exhibited strong in vitro mRNA delivery, with most CL15F LNPs outperforming CL4F and some matching Lipofectamine MessengerMAX in HEK-293T cells. Apparent pKa values across CL15F were 6.24–7.15, suitable for endosomal escape.
• Vaccination model: OVA mRNA LNPs using CL15F yielded anti-OVA IgG titers comparable to clinical benchmark lipids ALC-0315 and SM-102, and significantly higher cellular responses; CL15F 9-5 produced a 14-fold increase in IFN-γ ELISpot spots vs SM-102.
• Storage stability: At −80 °C, all LNPs preserved activity up to 5 months. Critically, at 4 °C as liquid, CL15F LNPs maintained in vivo activity over 5 months, whereas CL4F, SM-102, and ALC-0315 LNPs showed a time-dependent decrease with an approximate 2-month half-life. No in vivo toxicity was observed at tested doses.
• Physicochemical properties (size, PDI, encapsulation) and lipid component integrity by HPLC/CAD remained stable during refrigerated storage, indicating functional losses were not due to gross LNP degradation.
• Aldehyde impurity profiling: NBD-H assay showed CL15F lipids generated significantly fewer aldehyde impurities than CL4F and other ionizable lipids. DNPH-UPLC-QTOF-MS identified fatty aldehydes in CL4F samples, confirming lipid-origin aldehydes via N-oxidation and hydrolysis pathways. N-oxidized CL4F produced higher aldehyde levels than parent lipids.
• mRNA adducts: HILIC-MS detected fatty aldehyde–nucleoside adducts for adenosine, uridine, and cytidine (not guanosine) in CL4F incubations. RP ion-pair chromatography showed time-dependent mRNA–lipid adduct formation for linear amine lipids (CL4F and ALC-0315); CL15F showed minimal adducts. Aldehyde scavenging (PS-TsNHNH2) abolished adduct peaks, while amine removal did not, implicating aldehydes as causal.
• Correlation: Across CL4F and CL15F, loss of liver bioluminescence after 60 days at 4 °C correlated with initial aldehyde level by NBD-H (R² = 0.767), linking aldehyde impurities to functional decay.
• Oxidation behavior: 1H-NMR indicated similar N-oxidation kinetics for CL15F and CL4F under peroxide. HCD-MS/MS localized oxidation to the tertiary amine of CL15F. Despite similar N-oxide formation, only CL4F N-oxides increased aldehydes and adduct formation; CL15F N-oxides did not, suggesting hindered hydrolysis or immediate intramolecular cyclization to cyclic imines that sequester aldehydes.
• Generalization with other cyclic amines: CL16F and CL17F headgroups exhibited reduced aldehyde levels and limited adduct formation; CL17F LNPs showed high refrigerated stability similar to CL15F. CL6F generated aldehydes and adducts and showed reduced efficacy after storage, supporting the role of headgroup architecture. Minimizing aldehydes is necessary but not sufficient, as CL16F still lost activity despite low aldehydes, implicating additional factors (e.g., hydrolysis, mRNA structure effects).

The study addresses the key barrier of refrigerated liquid storage of mRNA/LNPs by focusing on a mechanistic source of functional decay: lipid-derived aldehydes that covalently modify mRNA and inhibit translation. By redesigning the ionizable lipid headgroup to a piperidine heterocycle (CL15F), the authors limited aldehyde formation and mRNA–lipid adducts while preserving or enhancing delivery potency. The data support a model wherein N-oxidation occurs for both linear and cyclic amines, but only linear tertiary amines (e.g., CL4F, ALC-0315-like structures) undergo hydrolysis to yield diffusible fatty aldehydes capable of localizing to the LNP core and reacting with mRNA. In contrast, piperidine-based lipids either impede hydrolysis or rapidly form intramolecular cyclic imines, suppressing free aldehyde release and thereby preserving mRNA function. The strong correlation between aldehyde level and loss of activity after 60 days at 4 °C, the elimination of adducts by aldehyde scavenging, and the minimal adduct formation with CL15F N-oxides collectively substantiate the causal chain from lipid structure to aldehyde generation to mRNA inactivation. Importantly, while minimizing aldehydes is pivotal, it does not fully explain all stability outcomes (e.g., CL16F), indicating that additional mechanisms (mRNA hydrolysis kinetics, headgroup basicity, LNP microenvironment, and mRNA higher-order structure in LNPs) also contribute. The improved cellular immune responses with heterocyclic lipids suggest added benefits for vaccine applications where T-cell responses are critical. Overall, the work provides design principles for ionizable lipids that enable practical refrigerated storage of mRNA/LNPs, potentially easing global distribution constraints.

This study identifies and validates piperidine-based ionizable lipids (CL15F), and related cyclic amine architectures (e.g., CL17F), that mitigate aldehyde impurity formation, minimize mRNA–lipid adducts, and preserve in vivo mRNA activity during prolonged 4 °C liquid storage. The findings elucidate how amine headgroup structure governs aldehyde generation pathways following N-oxidation, with cyclic amines impeding hydrolysis or favoring intramolecular imine formation to avoid release of reactive fatty aldehydes. In vitro and in vivo assays demonstrate high delivery potency, vaccine-relevant immunogenicity, and improved thermostability relative to widely used ionizable lipids (ALC-0315, SM-102). These insights provide a rational framework for designing ionizable lipids with enhanced stability, and they are compatible with other stabilization strategies (e.g., aldehyde scavengers, buffer systems, lyophilization). Future work should generalize these principles across broader lipid libraries, dissect contributions from RNA sequence/structure and LNP process variables, and directly quantify mRNA integrity to fully map the determinants of refrigerated stability.

Quantifying the contribution of small aldehydes and their adducts to mRNA inactivation is challenging; DNPH labeling detects small aldehydes but small aldehyde–mRNA adducts are difficult to resolve chromatographically due to modest hydrophobicity changes. Only a limited set of linear (three) and cyclic (four) lipids was examined, so broader generalization requires additional chemotypes. Sequence composition, modifications, and higher-order structure of mRNA in LNPs can modulate adduct formation and hydrolysis, complicating extrapolation. Loss of function can also arise from in-line RNA hydrolysis, potentially catalyzed by lipid headgroups; some stability outcomes (e.g., CL16F) cannot be attributed solely to aldehyde suppression. Comprehensive mRNA integrity analyses are ongoing.