Formulation and Characterization of Novel Ionizable and Cationic Lipid Nanoparticles for the Delivery of Splice-Switching Oligonucleotides

Abstract

Despite increasing knowledge about the mechanistic aspects of lipid nanoparticles (LNPs) as oligonucleotide carriers, the structure-function relationship in LNPs has been generally overlooked. Understanding this correlation is critical in the rational design of LNPs. Here, a materials characterization approach is utilized, applying structural information from small-angle X-ray scattering experiments to design novel LNPs focusing on distinct lipid organizations with a minimal compositional variation. The lipid phase structures are characterized in these LNPs and their corresponding bulk lipid mixtures with small-angle scattering techniques, and the LNP-cell interactions in vitro with respect to cytotoxicity, hemolysis, cargo delivery, cell uptake, and lysosomal swelling. An LNP is identified that outperforms Onpattro lipid composition using lipid components and molar ratios which differ from the gold standard clinical LNPs. The base structure of these LNPs has an inverse micellar phase organization, whereas the LNPs with inverted hexagonal phases are not functional, suggesting that this phase formation may not be needed for LNP-mediated oligonucleotide delivery. The importance of stabilizer choice for the LNP function is demonstrated and super-resolution microscopy highlights the complexity of the delivery mechanisms, where lysosomal swelling for the majority of LNPs is observed. This study highlights the importance of advanced characterization for the rational design of LNPs to enable the study of structure-function relationships.

Keywords: drug delivery; lipid nanoparticle; oligonucleotide; small angle scattering; stochastic optical reconstruction microscopy.

Figure 1

Schematic illustration of the study workflow. 16 novel LNPs, loaded with SSO, were formulated by tip sonication, followed by thorough structural characterization and evaluation of cell interactions in vitro. LNP = lipid nanoparticle, SSO = splice‐switching oligonucleotide, STORM = stochastic optical reconstruction microscopy. Figure created in BioRender (publication license acquired).

Figure 2

Bulk SAXS data (N = 1) collected at 25 °C for lipid formulations A) C1 B) C2 C) C1 + DOTAP / DODAP (0, 10, 20 mol‐%) D) C2 + DOTAP / DODAP (0, 10, 20 mol‐%). Note that the 0 mol‐% formulations plotted in C and D for DOTAP / DODAP are the same datasets, plotted twice for clarity. E) Fitted results from C1 SAXS data where error bars are of the order of the data points and reflect the error in each individual fit. F) Fitted results from C2 SAXS data where error bars are of the order of the data points and reflect the error in each individual fit.

Figure 3

Formulation of LNPs from the novel lipid mixtures leads to high‐quality nanoparticles with low polydispersity, good cargo loading, and distinct morphologies. A) Representative DLS intensity plots for the 16 novel LNPs, measured in PBS. B) Average LNP hydrodynamic diameter (Z‐average) and polydispersity (polydispersity index, PdI) in PBS, analyzed by DLS. N = 3. C) SSO cargo loading efficiency % into the novel LNPs, analyzed by RiboGreen assay. N = 3. In B and C error bars represent standard deviation. D, E) CryoEM visualization of the morphology of novel LNP series C1 and C2 at pH 7 (PBS), respectively. Representative images are shown. Scale bars 100 nm.

Figure 4

The novel LNPs demonstrate high cell viability in both HeLa and Huh7 cell models at 24 h, outperforming the MC3 LNPs (Onpattro composition). Cell viability in A,B) HeLa, and Huh7 cells treated with C1 LNPs, respectively. C,D) HeLa and Huh7 cells treated with C2 LNPs, respectively. Cell viability was evaluated with an alamarBlue assay. N = 3, n = 9 for the LNP samples, and n = 18 for the control cells, pH 3 control, and Lipofectamine (LF) control. * < 0.05 (compared to the control cells), # < 0.05 (compared to the pH 3 control), † < 0.05 (compared to the MC3 control), unless otherwise indicated. Kruskal‐Wallis test with Dunn's post hoc correction. Data is shown as means plus standard deviation.

Figure 5

C2 LNPs containing DOTAP and P80 stabilizer outperform MC3 LNPs (Onpattro composition) in functional SSO delivery into HeLa cells at 4 h, whereas none of the DODAP LNPs or the C1 LNPs can deliver the cargo. A,B) Delivery of C1 LNPs into HeLa cells at 4 and 24 h, respectively. C,D) Delivery of C1 LNPs into Huh7 cells at 4 and 24 h, respectively. E,F) Delivery of C2 LNPs into HeLa cells at 4 and 24 h, respectively. G,H) Delivery of C2 LNPs into Huh7 cells at 4 h and 24 h, respectively. LF = Lipofectamine. N = 3, n = 9 for the LNP samples, and n = 18 for the control cells, pH 3 control, and Lipofectamine (LF) control. * < 0.05 (compared to the control cells), # < 0.05 (compared to the pH3 control), † < 0.05 (compared to the MC3 control), unless otherwise indicated. Kruskal‐Wallis test with Dunn's post hoc correction. Data is shown as means plus standard deviation.

Figure 6

Small angle neutron scattering (SANS) data of LNP samples formulated with SSO and stabilized with P80 with lipid compositions A) C1 + DOTAP 20. B) C1 + DODAP 20. C) C2 + DOTAP 20. D) C2 + DODAP 20 measured at 37 °C in either pH 7.4 or pH 5. For A‐D. Data points are raw data with the power law fit over a truncated q range overlayed as a straight line (red). For E,F) each individual data point represents N = 1 with the error bars showing the error in the fit.

Figure 7

Small angle neutron scattering (SANS) data of LNP samples formulated with SSO and lipid composition C2 DOTAP 20 and stabilized with either A) P80. B) F127. C) DMPE‐PEG2000. D) DSPE‐PEG2000 and measured at 25, 37 °C in either pH 7.4 or pH 5. For A–D. Data points are raw data with the unified model fit over a truncated q range overlaid as a red line. For E–H) each individual data point represents an individual fit where N = 1 and the error bars represent the error in the fit.

Figure 8

DOTAP 20 LNPs in both C1 and C2 series are generally well uptaken into both HeLa and Huh7 cells, whereas DODAP 20 LNPs do not enter the cells. A,B) Representative images of LNP uptake into HeLa and Huh7 cells at 24 h time points, respectively. DOTAP 20 LNPs on the left side and DODAP 20 on the right side. Rhodamine B‐labelled LNPs = green, actin cytoskeleton labeled with phalloidin Alexa Fluor 647 = red, nuclei labeled with DAPI = blue. Scale bars 30 µm.

Figure 9

LNPs induce lysosome swelling in both HeLa and Huh7 cells treated with MC3, C1 DOTAP, and C2 DOTAP LNPs. A,B) Representative images of HeLa and Huh7 lysosomes, respectively after 24 h of LNP cell treatment. Anti‐LAMP1 staining (yellow), nuclei stained with DAPI (blue). Scale bars: diffraction‐limited and full view STORM 5 µm, zoom‐in STORM 1 µm. C,D) Quantification of lysosome sizes in HeLa and Huh7 cells, respectively, with a deep learning method. *p < 0.05 compared to the ctrl unless otherwise indicated. Kruskal‐Wallis test with Dunn's post hoc correction. The box extends from the 25^th to 75^th percentile, the line in the middle of the box represents the median and the whiskers show 5 and 95 percentiles. The number of images and identified objects (lysosomes) in the quantification are indicated in Tables S1 and S2 (Supporting Information).