Mitochondria-targeted gene delivery using fluorinated lipid nanoparticles to alleviate Leber's hereditary optic neuropathy

Abstract

Mutations in mitochondrial DNA (mtDNA) lead to various mitochondrial diseases for which no cure is currently available. Despite the promising potential of mtDNA correction to treat these disorders, the double mitochondrial membranes have proven to be a tough barrier to overcome. Here, we develop fluorinated lipid nanoparticles with a mitochondrial targeting sequence (F-M-LNP) to overcome the mitochondrial barrier by virtue of their high affinity for mitochondrial membranes, thereby effectively introducing gene into mitochondria. Through the rational design of ionizable lipid structures, we synthesize 16 lipid nanoparticles (LNPs) with varying degrees of fluorination and investigate the key structural features required for efficient mitochondria-targeted gene delivery. As fluorinated ionizable lipid-mediated mitochondrial transport is independent of mitochondrial membrane potential (MMP), F-M-LNPs deliver gene to mitochondria under pathological conditions where MMP is impaired, resulting in a 3.8-fold increase in functional protein expression compared to non-fluorinated LNPs. In a male mouse model of genetically induced mitochondrial disease, F-M-LNP demonstrate functional complementation of mutant mtDNA, alleviating disease symptoms. Together, our results show that modifying vectors with fluorinated groups offers valuable tools for correcting mitochondrial genome defects.

Fig. 1. Design and evaluation of fluorinated lipid for mitochondrial gene delivery.

A The synthesis and chemical structure of different fluorinated lipids. B, C Schematic showing in vitro screening to identify LNP library with different fluorinated lipids for pDNA delivery to mitochondria. Five colors in pie charts represent five lipids used in LNP formulations. The area percentage of each color in the pie charts represents the molar ratio of the lipid in the formulation. D Cellular uptake capacity of LNP library with different fluorinated lipids (n  =  3 biologically independent samples). E Mitochondrial uptake capacity of LNP library with different fluorinated lipids (n  =  3 biologically independent samples). F, G Mitochondrial GFP transfection efficiency of LNP library with different fluorinated lipids (n  =  3 biologically independent samples). H The mass ratio of fluorine atoms to total lipids (F wt%) in 16 LNPs. I The influence of the mitochondrial depolarization on the binding to 6F-LNPs and isolated mitochondria (n  =  3 biologically independent samples). B was created with BioRender.com. Data are presented as mean ± SD. Two-tailed unpaired Student’s t test (I) was used to calculate the statistical significance.

Fig. 2. Mechanistic framework for mitochondria-targeted gene delivery mediated by 6F-LNPs.

A The schematic illustration of mechanistic framework for mitochondria-targeted gene delivery mediated by 6F-LNPs. B The interaction between fluorinated molecules and mitochondrial membrane lipids by non-targeted lipidomic analysis (n  =  3 biologically independent samples). CL cardiolipin, PC phosphatidylcholine, PA phosphatidic acid, PE phospha-tidylethanolamine, PS phosphatidylserine, PI phosphatidylinositol. C MST detection between 6F-LNPs/0F-LNPs and mitochondrial membrane mimicking liposomes (n  =  3 biologically independent samples). D The molecular dynamics simulation of 6F-LNPs across mitochondrial membrane. 6F lipid in 6F-LNP is depicted in red and all other lipids are depicted in yellow. The inner mitochondrial membrane is depicted in green and gray. E Binding affinity of different fluorinated lipids with isolated mitochondria (n  =  3 biologically independent samples). F CLSM images showing 6F-LNP/Cy3-pDNA localization inside the mitochondria (representative images from n = 3 independent experiments). Outer mitochondrial membranes (OMMs, mouse mAb to TOMM20): green; mitochondrial matrix proteins (rabbit pAb to SDHA): red; Cy3-pDNA: blue. G The intracellular trafficking of 6F-LNPs in living cells observed by transmission electron microscopy (TEM), representative images from n = 3 independent experiments. Red arrow: gold colloid in 6F-LNPs. Scale bar: 200 nm. A was created with BioRender.com. Data are presented as mean ± SD. Two-tailed unpaired Student’s t test (C) or One-way ANOVA with Tukey’s multiple comparisons test (two-tailed; E) was used to calculate the statistical significance.

Fig. 3. Preparation, characterization and mitochondrial targeting behavior of 6F-M-LNPs.

A Formulation of 6F-M-LNPs via microfluidic mixing and thiol-maleimide coupling. B Cellular uptake capacity of 6F-M-LNPs with different proportions of MTS (n  =  3 biologically independent samples). C Mitochondrial GFP transfection efficiency of 6F-M-LNPs with different proportions of MTS (n  =  3 biologically independent samples). D Physicochemical characterization of 6F-LNPs before and after MTS grafting (n  =  3 biologically independent samples), EE encapsulation efficiency. E TEM image of 6F-M-LNPs. Scale bar: 200 nm. F Mitochondrial gene delivery of different F-LNPs detected by CLSM. Scale bar: 10 μm. Mitochondria (Mitotracker Green probe): green; gene cargo (Cy5-pDNA): red. White arrows: Cy5-pDNA in the cytosol. G Pearson’s correlation coefficient between Cy5-pDNA and mitochondria (n  =  3 biologically independent samples). H The mitochondrial targeting of different F-LNPs detected by MFI of Cy5-pDNA in extracted mitochondria (n  =  3 biologically independent samples). I Detection on key mitochondrial membrane proteins (representative images from n = 3 independent experiments). J The mtGFP transfection after gene silencing of Tomm20 and Tomm22 (n  =  3 biologically independent samples). K Mitochondrial colocalization of 6F-M-LNP/Cy5-pDNA observed by Multi-SIM (representative images from n = 3 independent experiments). Outer mitochondrial membranes were labeled by Tomm20-mEmerald transfection. Inner mitochondrial membranes were labeled by Mitotracker Red probe. L CLSM images showing 6F-M-LNP/Cy3-pDNA localization inside the mitochondria (representative images from n = 3 independent samples). Outer mitochondrial membranes (mouse mAb to TOMM20): green; mitochondrial matrix proteins (rabbit pAb to SDHA): red; Cy3-pDNA: blue. M Schematic illustration of mitochondrial targeting behavior of 6F-M-LNPs. A and M was created with BioRender.com. Data are presented as mean ± SD. One-way ANOVA with Tukey’s multiple comparisons test (two-tailed; B, C, G, H, J) was used to calculate the statistical significance.

Fig. 4. Therapeutic effect of 6F-M-LNP-medited mitochondrial gene therapy in LHON disease cells GM10742.

A Schematic illustration of hND4 plasmid (hND4-3xFLAG-mtLuc) transfection in LHON disease cells GM10742 with different LNP formulations. B Luciferase transfection of different LNP formulations in LHON disease cells GM10742 (n  =  3 biologically independent samples). C, D The hND4 protein expression of different LNP formulations in LHON disease cells GM10742 as detected by WB (n  =  3 biologically independent samples). E MMP change in LHON disease cells GM10742 by detection of JC-1 probe (n  =  3 biologically independent samples). F Examination of the ATP contents of different LNP formulations in LHON disease cells GM10742 (n  =  3 biologically independent samples). G Mitochondrial ROS level detection (n = 3 biologically independent samples). H, I OCR parameters (n  =  3 biologically independent samples). J Schematic illustration of the mitochondrial function restoration in the 6F-M-LNPs. A and J was created with BioRender.com. Data are presented as mean ± SD. One-way ANOVA with Tukey’s multiple comparisons test (two-tailed; B, D-G, I) was used to calculate the statistical significance.

Fig. 5. Therapeutic effect of 6F-M-LNP-mediated mitochondrial gene therapy in mutant ND4 mtTg LHON model male mice.

A Flowchart of various treatments in the mutant ND4 mtTg LHON model male mice. B, C Total ND4 protein expression of various treatments as detected by WB (n  =  3 biologically independent samples). D Examination of the ATP contents in retina under various treatments (n  =  3 biologically independent samples). E The number of head movements in various treatments (n  =  12 eyes per group). F, G Amplitude variation of a-wave and b-wave in ffERG test (n  =  6 eyes per group). H The ffERG waveforms of mice. I, J H&E images of LHON eyes in different groups (n  =  3 biologically independent samples), the black arrows indicate RGCs loss. A was created with BioRender.com. Data are presented as mean ± SD. One-way ANOVA with Tukey’s multiple comparisons test (two-tailed; C–G, J) was used to calculate the statistical significance.