All mRNAs used consisted of CAP-carrying linear mRNA with a 5′untranslated region (UTR) from the human hydroxysteroid 17-beta dehydrogenase 4 gene (HSD17B4), a 3’ UTR from the human proteasome 20S subunit beta 3 gene (PSMB3) and a histone stem-loop followed by an A100 poly(A) tail (Gebre et al., 2022), either encoding for Rabies virus glycoprotein G (RABV-G) (N1-methylpseudouridine (m1Ψ)-modified, RABV-G mRNA) or encoding for enhanced green fluorescent protein (unmodified, eGFP mRNA). Furthermore, all open reading frames were GC-optimized (Thess et al., 2015). All mRNAs were produced in-house (CureVac SE) (Tübingen, Germany).

Lipid Nanoparticles (LNPs) were prepared using the microfluidics mixing technique, where an ethanolic lipid solution is mixed with an aqueous mRNA solution. The lipid solution contained ionizable lipid H (Broadpharma, USA), cholesterol (Avanti Polar Lipids, USA), Dioctadecanoyl-sn-glycero-3-phosphocholin (DSPC) (Avanti Polar Lipids, USA), and 1,2-Dimyristoyl-rac-glycero-3-methylpolyoxyethylene glycol-2000 (DMG-PEG2000) (NOF Corporation, Japan) (mol ratio 48.5:38.9:11.1:1.5) or N-methyl-2-(N4', N4′-di(tetradecyl)succinamide)-poly[(N-acetyl)ethylamine] (PMOz-DM-amide) (4778.20 g/mol, purity >95% (LC-UV); CureVac) replacing PEG at various mol% (1.0%, 1.5% and 2.5%), all dissolved in ethanol, while the mRNA was diluted in 50 mM citrate (pH 4.0) (see more details in Figure 1). The device used for the mixing was the Nanoassemblr^® Ignite™ (Precision Nanosystems, Vancouver, Canada) and the solutions were mixed at a volume ratio of 3:1 (mRNA solution:lipid solution) and at a flow rate of 20 mL/min. After mixing, the LNPs were transferred to dialysis cassettes (Slyde-a-Lyzer G3) (ThermoFisher, USA) and dialyzed overnight, with 2 buffer exchanges in between, against a physiological buffer. After dialysis, the LNPs were concentrated using Vivaspin^® ultrafiltration units (Satorius, Germany) and physicochemical features were characterized. Hydrodynamic diameter size and polydispersity index (PDI) were measured by dynamic light scattering using the DynaPro Plate Reader III (Wyatt, USA). The zeta potential was measured with a Zetasizer Nano ZS90 (Malvern, United Kingdom). mRNA concentration and encapsulation efficiency were determined via Quant-iT Ribogreen^® RNA assay (ThermoFisher, USA) following the manufacturer’s recommendations. Briefly, the mRNA is diluted with TE buffer in the presence or absence of Triton X-100 and the Ribogreen^® dye was added to the samples, followed by their measurement at 485 nm excitation and 535 nm emission. Samples treated with Triton X-100 represent the total mRNA (LNPs are broken down in the presence of the detergent) and the samples in absence of Triton X-100 comprise “external” mRNA. The encapsulation efficiency is defined by E E   % = RNA total − RNA external RNA total ∙ 100 % . The composition and physicochemical properties of all LNPs are summarized in Figure 1.

Cell lines used in this study include HeLa (DSMZ ACC57), HepG2 (DSMZ ACC 180), 293T (ATCC CRL-3126) and THP-1 Null2 (InvivoGen) cells. Blood was purchased from the Zentrum für Klinische Transfusionsmedizin (Tübingen, Germany) and human peripheral blood mononuclear cells (PBMCs) were isolated with Ficoll-Paque PLUS (Cytiva) densitiy gradient medium and cryopreserved in 90% FCS and 10% DMSO until required. HeLa, HepG2 and PBMCs were maintained in RPMI (brand/specification; supplemented with 10% FCS, 2 mM L-Glutamine, 100 μg/mL Pen-Strep), while 293T cells were cultured in DMEM (brand/specification; supplemented with 10% FCS, 2 mM L-Glutamine, 100 μg/mL Pen-Strep). THP-1 Null2 cells were maintained in RPMI (brand/specification; supplemented with 10% FCS, 25 mM HEPES, 100 μg/mL Pen-Strep, 100 μg/mL Normocin, 100 μg/mL Zeocin). All cells were cultured at 37°C, 5% CO₂.

HeLa, 293T and HepG2 cells were respectively seeded overnight (37°C, 5% CO₂) in 6-well plates at 2.5 × 10⁵ cells/well, 3.0 × 10⁵ cells/well and 5.0 × 10⁵ cells/well, and THP-1 Null2 cells in 24-well plates at 5.0 × 10⁵ cells/well. LNP-formulated eGFP mRNA was added to the culture medium to a final concentration of 10 ng/mL (for transfection of HeLa, 293T and HepG2 cells) and 500 ng/mL (for transfection of THP-1 Null2 cells), respectively. Cells were incubated for another 18 h at 37°C, 5% CO₂, stained with Fixable Viability Stain 780 (BD Biosciences) and the eGFP translation product was detected via flow cytometry. Data were respectively acquired and analyzed with the NovoCyte Penteon flow cytometer (Agilent Technologies) and FlowJo software (v10.8.1).

PBMCs were seeded at 1.5 × 10⁶ cells in 24-well plates and treated with LNPs at 500 ng/mL or control (medium) for 18 h at 37°C, 5% CO₂. Supernatants were collected, cleared by centrifugation and chemokines and cytokines were quantified (see section 2.6). The cells were stained with Fixable Viability Stain 780 (BD Biosciences, #565388), anti-HLA-DR (R718, clone L243, BD Biosciences, #568578), anti-CD14 (BUV395, clone MφP9, BD Biosciences, #563561), anti-CD3 (BUV496, clone SK7), anti-CD19 (BUV563, clone SJ25C1, BD Biosciences, # 612916), anti-CD45 (BUV805, clone HI30, BD Biosciences, #612891), anti-CD11c (BV480, clone B-ly6, BD Biosciences, #566135), anti-CD16 (BV786, clone 3G8, BD Biosciences,# 563690), anti-CD56 (PE-Cy7, clone B159, BD Biosciences, #557747) and anti-CD123 (PE, clone 7G3, BD Biosciences, # 568155) to identify different immune cell populations. Staining was additionally performed with anti-CD80 (BV605, L307.4, BD Biosciences, #563315), anti-CD83 (PE-Cy5, clone HB15e, BD Biosciences, #551058), anti-CD86 (BV421, clone 2331 (FUN-1), BD Biosciences, # 562432) to identify activated antigen presenting cells. The fixable viability dye and antibodies were obtained from BD Biosciences. Data were respectively acquired and analyzed with the NovoCyte Penteon flow cytometer (Agilent Technologies) and FlowJo software (v10.8.1).

LEGENDplex multi-analyte Flow assays (Biolegend) were used to quantify cytokines and chemokines in the present study. Cell culture supernatants were collected after 18 h (at 37°C, 5% CO₂) of treatment with 500 ng/mL of LNPs. The resulting supernatants were analyzed with the Human Anti-Virus Response Panel (13-plex) and enabled quantification of IL-1β, IL-6, CXCL8 (IL-8), IL-10, IL-12p70, IFN-α2, IFN-β, IFN-λ1 (IL-29), IFN-λ2/3 (IL-28a/b), IFN-γ, TNF-α, CXCL10 (IP-10) and GM-CSF. For in vivo studies, sera were collected 14 h after primary immunization and analyzed with the Mouse Anti-Virus Response Panel (13-plex) and enabled quantification of IFN-γ, CXCL1 (KC), TNF-α, CCL2 (MCP-1), IL-12p70, CCL5 (RANTES), IL-1β, CXCL10 (IP-10), GM-CSF, IL-10, IFN-β, IFN-α and IL-6.

Spleens were harvested on day 28 and single cell suspensions prepared. The resulting cells were stained with Zombie UV Fixable Viability dye (Biolegend, #423108), α-CD90.2 (AF700, 53-2.1, Biolegend, #140324), α-CD8a (BUV737, clone 53-6.7, BD Biosciences, #612759), α-CD4 (BV786, clone RM4-5, BD Biosciences, #563727) and a class I (H-2Ld) pentamer (APC, ProImmune Ltd.) loaded with 453-461aa of RABV-G. Data were respectively acquired and analyzed with the NovoCyte Penteon flow cytometer (Agilent Technologies) and FlowJo software (v10.8.1).

Animal experiments were performed at Synovo GmbH (Tübingen, Germany). Female BALB/cJRj mice 8–9 weeks old were randomly divided into five groups (n = 7) and injected intramuscularly (Musculus tibialis) with 25 µL of mRNA LNP formulation (1 µg of mRNA per dose) or buffer control at days 0 (primary immunization) and 21 (secondary immunization). Blood was collected 14 h after the primary immunization, allowed to clot and serum prepared by centrifugation and the resulting material was used for chemokine and cytokine determination (see section 2.6). Serum was prepared in a similar manner on day 28 for quantification of virus neutralizing titers (VNTs) by EUROVIR Hygiene-Labor GmbH (Luckenwalde, Germany). Splenocytes from day 28 were analyzed for the presence of antigen-specific CD8⁺ T-cell responses (see section 2.7).

Mice were injected i.m. with 25 µL of LNP 1.5% PEG and LNP 1.5% PMOZ (1 µg of mRNA per dose) or buffer and blood was collected after 14 h. The blood was allowed to clot and sera were prepared via centrifugation. The resulting sera were diluted 1:10 in 0.9% NaCl solution and AST/ALT levels were quantified with the response^®910VET Veterinary Chemistry Analyzer (DiaSys Diagnostic Systems, USA) according to the manufacturers’ instructions.

Statistical analysis was performed using GraphPad Prism software (version 10.1.0). Data were assessed by a one-way ANOVA with Tukey’s multiple comparison test or by Brown-Forsythe and Welch ANOVA with Dunnett’s T3 multiple comparison tests. Only statistically significant data are illustrated on figures and significant differences are defined as *p < 0.1, **p < 0.01, ***p < 0.001, ****p < 0.0001.

To study PMOZ as a PEG alternative we selected SM-102 LNP (Jackson et al., 2020) as a model system. Towards this end, SM-102 LNP (hereafter simply referred to as LNP) was formulated with ionizable lipid H (Hassett et al., 2021), DSPC, cholesterol and polymer-lipid (Hald et al., 2022). The polymer-lipids used in this case were either PEG2000-DMG or PMOZ-DM-amide. Both contain a C14 lipid anchor, while PEG2000-DMG or PMOZ-DM-amide respectively comprise polymer repeating units of 44 or 50 (Figure 1A). PEG2000-DMG was formulated at 1.5 mol% and the resulting LNP served as reference material for the present study. PMOZ-DM-amide was formulated at 1.0 mol%, 1.5 mol% or 2.5 mol% and the amount of cholesterol was adjusted to achieve a total of 100 mol% for all lipid components of the LNP (Figure 1B). LNPs were prepared with either unmodified eGFP mRNA or N¹-methylpseudouridine (m1Ψ)-modified RABV-G mRNA. Ribogreen assay analysis revealed that replacement of PEG by PMOZ does not influence mRNA encapsulation efficiency. All LNPs consistently showed encapsulation efficiencies over 95% (Figure 1C). Next, we assessed the physicochemical characteristics of the LNPs via dynamic light scattering (DLS) and this showed that the percentage of PMOZ influenced the hydrodynamic size (Figure 1D), polydispersity index (PDI) (Figure 1E) and zeta potential (Figure 1F) of the LNPs. Higher percentages of PMOZ correlated with decreases in hydrodynamic size and zeta potential. Conversely, the percentages of PMOZ correlated negatively with the PDI. Next, we evaluated how dilution influences changes in hydrodynamic size and PDI in response to freeze-thaw (Figure 1G). These data show that PMOZ-containing LNPs have physicochemical characteristics that are typical for LNPs used for mRNA delivery for medical applications and raise no significant concerns for further testing in biological models.

We next assessed the performance of PMOZ-containing LNPs in various cell lines. To this end, HeLa, 293T, HepG2 and THP-1 Null2 cells were treated with LNPs that encapsulate unmodified eGFP mRNA. The resulting cells were analyzed by flow cytometry to quantify the percentage eGFP-positive cells and the eGFP median fluorescence intensity (MFI) (Figure 2; Supplementary Figure S1). This revealed that a lower percentage of PMOZ correlated with an increase in transfection in HeLa, 293T and HepG2 cells. Moreover, LNP 1.0% PMOZ (for LNP nomenclature see Figure 1B) transfected these cell lines to a similar degree as LNP 1.5% PEG. This trend was not observed in the monocytic cell lines THP-1 Null2 and LNP 1.5% PMOZ more readily transfected cells in comparison to all other LNPs. However, this trend is not statistically significant. These results strongly suggest that LNPs can be formulated with PMOZ to successfully deliver mRNA to cell lines.

Having confirmed that PMOZ-containing LNPs successfully transfected cell lines, we aimed to test them in primary immune cells. Accordingly, we selected PBMCs as they comprise numerous cell types and enable the immunostimulatory properties of LNPs to be determined (Tahtinen et al., 2022). PBMCs from several donors were treated overnight with eGFP mRNA LNPs and the resulting supernatants and cells were analyzed (Figure 3). The immunostimulatory profile of the LNPs was determined by quantifying cytokines and chemokines in cell culture supernatants using a cytokine bead array (Figures 3A,B). This revealed that a select number of cytokines was induced by the tested LNPs (Figure 3A). Quantification data from TNF-α, IP-10, IFN-ʎ1, IFN-α, IFN-β and IFN-γ show that LNP 1.5% PEG and LNP 1.0% PMOZ were potent immune activators. In contrast LNP 1.5% PMOZ and LNP 2.5% PMOZ less readily induced cytokines and there is a general trend that suggests increasing PMOZ percentages have a negative impact on immune stimulation. Next, we performed detailed analysis on the various adaptive and innate immune cell populations that are transfected by LNPs that contain PEG or PMOZ lipid-polymers (Figure 3C; Supplementary Figure S2, S3). LNP 1.5% PEG most readily transfected the majority of immune cell types in comparison to LNP 1.0% PMOZ, LNP 1.5% PMOZ and LNP 2.5% PMOZ. The only observed exception are non-classical monocytes, where the type and percentage of polymer-lipid did not have a significant impact on the percentage of cells that contain the eGFP translation product. Lastly, we determined the degree of innate cell activation by quantifying the percentage of cells that express the costimulatory molecules CD80, CD83 or CD86 at their surface (Figure 3D; Supplementary Figure S4). This showed that there is a clear relationship between the amount of PMOZ and immunostimulatory potential, with increasing PMOZ percentages corresponding with decreased chemokine and cytokines values. Moreover, it is evident that LNP 1.5% PEG and LNP 1.0% PMOZ displayed a similar phenotype, and most readily induced activation in the majority of innate cells. These data demonstrate that the in vitro/ex vivo transfection efficiency and immunostimulatory potential of LNPs can be modulated by changing the amount of PMOZ.

To interrogate LNPs containing PMOZ in a vaccination setting, LNP-encapsulated m1Ψ-modified RABV-G mRNA was used to immunize BALB/cJRj mice. Animals were injected intramuscularly on day 0 and day 21 with a dose of 1 µg mRNA (Figure 4A). The immunostimulatory profiles of the LNPs were determined 14 h after the primary immunization by quantifying various chemokines and cytokines in the sera of animals (Figure 4B). LNP 1.5% PEG most robustly induced cytokine responses, while the tendency of PMOZ-containing LNPs to induce cytokines in vivo increased as the percentage of PMOZ was decreased. Next, we assessed RABV-G-specific CD8⁺ T-cell responses present in animals at day 28 (Figure 4C; Supplementary Figure S5). To this end, splenocytes were stained with a class I pentamer loaded with the RABV-G epitope LPNWGKYVL (453-461 aa). The percentage of double-positive CD8⁺class I pentamer⁺ cells were clearly higher in animals immunized with LNP 1.5% PEG relative to LNPs that contain PMOZ (1.0%, 1.5% or 2.5%). However, this observation is only statistically significant when comparing LNP 1.5% PEG and LNP 2.5% PMOZ. Lastly, sera from day 28 were analyzed for the presence of rabies virus neutralizing antibody titers (VNTs) (Figure 4D). This showed that all animals vaccinated with m1Ψ-modified RABV-G mRNA LNPs, irrespective of lipid-polymer type and percentage, had titers that exceed the minimal protection threshold (≥0.5 IU/mL) (World Health Organization, 2018). However, notable differences can be observed between treatment groups. Our results show that LNP 1.5% PMOZ induced the highest mean VNTs, a result that is statistically significant when compared with LNP 1.5% PEG. Interestingly, LNP 1.0% PMOZ also induced higher mean VNTs than LNP 1.5% PEG, although this finding is not statistically significant. Together these findings suggest that LNPs formulated with PMOZ (1 mol% and 1.5 mol%) give rise to a superior VNT response, but less readily induce antigen-specific CD8⁺ T-cell responses when compared to LNPs formulated with PEG.