Defining Composition–Cytokine Relationships Enables the Design of Lipid Nanoparticles with Programmed Immunogenicity

Figures

• 

  FIGURE 1. LNPs deliver mRNA to primary monocytes and a monocytic cell line, resulting in ionizable lipid‐dependent cytotoxicity and cytokine release. (A) Schematic of the standard LNP formulation used in the study (B) Human PBMCs were incubated with GFP LNPs (MC3, SM‐102, or LP01) for 48 h, stained for viability and major immune cell markers, including CD4+ T cells (CD4+, CD3+), CD8+ T cells (CD8+, CD3+), NK cells (CD56+, CD3−), B cells (CD19+, CD3−), and monocytes (CD14+), and analyzed by flow cytometry; GFP+ cells are presented as a percentage of each cell subset. Bars represent mean ± SD; n = 5 donors; individual data points represent individual donors, with one donor analyzed in technical duplicate. Data were analyzed using Student's t‐test. (C) THP‐1 cells (50 000/well) were treated with 1 µg/mL LNPs formulated with MC3, SM‐102, or LP01. After 48 h, GFP expression was quantified using Incucyte image analysis and plotted as total integrated green fluorescence per well. Bars represent mean ± SD; n = 3 independent experiments; individual data points represent technical replicates pooled across experiments. Data were analyzed using one‐way ANOVA. (D) Cell viability after 48 h of LNP…

• 

  FIGURE 2. LNPs synergize with LPS and IFN to induce IL‐1β and IP‐10, respectively. (A) To model context‐specific cytokine release, THP‐1 cells (50 000/well) were exposed to media, LPS (1 µg/mL, 4 h), or IFN (IFNβ or IFNγ, 20 ng/mL, 1 h) before addition of eGFP LNPs (1 µg/mL) formulated with MC3, SM‐102, or LP01. (B) Following a further 48 h incubation, GFP expression was quantified using Incucyte and plotted as total integrated green fluorescence as a percentage of the media‐only condition. Bars represent mean ± SD; n = 3 independent experiments; individual data points represent technical replicates pooled across experiments. Data were analyzed using two‐way ANOVA. (C) Cell viability was assessed using CellTiter‐Glo and is shown as the percentage of viable cells relative to the media‐treated control. Bars represent mean ± SD; n ≥ 3 independent experiments; individual data points represent technical replicates pooled across experiments. Data were analyzed using two‐way ANOVA. Aluminum hydroxide (Alum, 100 µg/mL) served as a non‐lipid‐based vaccine adjuvant. (D, E) To determine cytokine levels, supernatants were assessed using MSD U‐Plex or DuoSet ELISA kits, with measurements acquired using…

• 

  FIGURE 3. A design of experiments (DoE) approach identifies the contribution of non‐ionizable lipids to immunogenicity. (A) A D‐optimal DoE panel of 15 LNP formulations varying SM‐102, DSPC, cholesterol, and PEG‐DMPE ratios was applied to THP‐1 Dual cells pretreated with LPS, IFNγ, or media. Parameters measured included GFP expression, IRF, NF‐κB, IL‐1β, IP‐10, toxicity, particle size (PS), and polydispersity index (PDI). Principal component analysis (PCA) loading plots were used to visualize relationships among formulation variables, particle properties, and biological readouts. Colors indicate the pre‐stimulation condition associated with each readout (red, LPS; blue, IFNγ; green, and media), while black vectors denote formulation and particle parameters. Vectors pointing in similar directions indicate positive association, whereas near‐perpendicular vectors indicate weaker relationships. (B) Ternary plots illustrate the relationship between LNP composition and cytokine responses (IP‐10, IL‐1β, and MCP‐1) at fixed PEG‐DMPE levels of 0.5%, 1.5%, and 3 mol%. Within each ternary plot, each side of the triangle corresponds to the mol% of one formulation component: ionizable lipid, DSPC, or…

• 

  FIGURE 4. LNP vaccine immunogenicity is driven by the ionizable/PEGylated lipid combination and antigen potency. (A) BALB/c mice (n = 4 per group) were injected intramuscularly with luciferase‐expressing mRNA (5 µg, 50 µL) formulated in LNPs containing MC3, SM‐102, or LP01 as the ionizable lipid. A buffer control, as well as a predicted low‐immunogenicity LNP formulation (SM‐102/DoE), were included. Serum was collected 6 h post‐injection, and cytokines were quantified by MSD multiplex ELISA. Individual data points represent individual mice. Data were analyzed using one‐way ANOVA.*p < 0.05, ** p < 0.01. (B) Luciferase expression was quantified by injection of D‐luciferin (100 µL, 15 mg/mL) 10 min before imaging using an IVIS system. Representative images show luciferase signal at the injection site; regions of interest (ROIs) were drawn manually to quantify local photon flux. (C) Quantification of luciferase expression is shown as total photon flux per mouse (C). Individual data points represent individual mice. Data were analyzed using one‐way ANOVA. *** p < 0.001, **** p < 0.0001.

• 

  FIGURE 5. All LNPs elicit robust adaptive responses. (A) Schematic of the vaccination protocol: BALB/c mice (n = 5) received intramuscular injections of GFP mRNA LNPs formulated with MC3, SM‐102, or LP01 on days 0 and 21. Tissues were collected on day 35. (B) Serum anti‐GFP IgG titers at day 35 following BALB/c vaccination were quantified by ELISA using anti‐mouse IgG‐HRP and expressed by interpolation from a monoclonal anti‐GFP standard curve. Individual data points represent individual mice. Data were analyzed using one‐way ANOVA. (C) Serum IgG1 and IgG2a isotypes were assessed by ELISA; IgG1/IgG2a signal ratios are presented. Individual data points represent individual mice. (D) Spleens were processed into single‐cell suspensions; splenocytes were stimulated with media or GFP peptide (5 µg/mL) in the presence of brefeldin A, stained for viability, CD3, CD4, and CD8, and then stained intracellularly for IFNγ, TNF‐α, and IL‐2. Flow cytometry was used to identify cytokine‐positive CD8+ T cells. Individual data points represent single mice, and the median is shown. Data were analyzed using Student's t‐test. * p < 0.05, ** p < 0.01; ns, not significant.

• 

  FIGURE 6. Protein expression in vivo and in vitro shows the strongest correlation with antibody responses. In vivo cytokine data from Figure 4A, in vivo protein expression data from Figure 4C, and in vitro protein expression data from Figure 1C were plotted against serum anti‐GFP antibody titers from Figure 5B. Linear regression was performed using GraphPad Prism software, and the fitted trendline is shown. Error bars represent standard deviation, and R2 values are inset.