Low immunogenicity of LNP allows repeated administrations of CRISPR-Cas9 mRNA into skeletal muscle in mice

Abstract

Genome editing therapy for Duchenne muscular dystrophy (DMD) holds great promise, however, one major obstacle is delivery of the CRISPR-Cas9/sgRNA system to skeletal muscle tissues. In general, AAV vectors are used for in vivo delivery, but AAV injections cannot be repeated because of neutralization antibodies. Here we report a chemically defined lipid nanoparticle (LNP) system which is able to deliver Cas9 mRNA and sgRNA into skeletal muscle by repeated intramuscular injections. Although the expressions of Cas9 protein and sgRNA were transient, our LNP system could induce stable genomic exon skipping and restore dystrophin protein in a DMD mouse model that harbors a humanized exon sequence. Furthermore, administration of our LNP via limb perfusion method enables to target multiple muscle groups. The repeated administration and low immunogenicity of our LNP system are promising features for a delivery vehicle of CRISPR-Cas9 to treat skeletal muscle disorders.

Fig. 1. LNP-mediated Luc-mRNA or CRISPR-Cas9 mRNA/sgRNA delivery into muscle tissue.

a Schematic illustration of LNP-CRISPR. Either Luc mRNA or Cas9 mRNA/sgRNA is encapsulated into LNP that consists of TCL053, DPPC (Dipalmitoylphosphatidylcholine), PEG-DMG (Polyethylene glycol-dimyristoyl glycerol), and cholesterol. b Chemical structure of the newly synthesized ionizable lipid, TCL053. c Representative bioluminescence images of C57BL/6J mice after the intramuscular injection of AAV2-Luc (1 × 10⁸, 1 × 10⁹, or 1 × 10¹⁰ v.g., vector genomes) or LNP-Luc mRNA (1 or 10 μg mRNA). d, e Quantification of the bioluminescence signal in skeletal muscle of C57BL/6J mice treated with AAV2-Luc (d) or LNP-Luc mRNA (e). The same mice (n = 3 mice per group) were examined repeatedly over time. Total flux data (p s⁻¹, photons per second) are plotted as a single line per mouse.

Fig. 2. Exon 45 skipping of human dystrophin using dual sgRNAs in DMD-iPS cells.

a Two sgRNAs were designed to induce human exon 45 skipping. hEx45 sgRNA #1 and hEx45 sgRNA #23 target the splice acceptor and donor site, respectively. The human DMD exon 45 sequence is highlighted in red. b Therapeutic strategy of genomic exon skipping by the dual sgRNA approach. Exon 45 skipping results in restoration of the dystrophin protein by adjusting the protein reading frame. c DMD-iPS cells were derived from a DMD patient lacking exon 44. Myogenic differentiation of the DMD-iPS cells was induced by the overexpression of MYOD1 (DMD-Myoblast). LNP-CRISPR, which is a mixture of LNP-Cas9 mRNA (1 µg mRNA) and LNP-sgRNA (total 1 μg of sgRNA: either 1 μg of #1, 1 μg of #23, or 0.5 μg each of #1 and #23), was administrated for 72 h. d Exon skipping efficiency was measured by RT-PCR and TapeStation. n = 1 experiment is shown, but similar exon skipping were observed in more than 3 experiments. e Dystrophin restoration was measured by Western blot. n = 1 experiment is shown, but similar Dystrophin recovery were obtained in more than 3 experiments.

Fig. 3. Generation of an exon 45 humanized mouse model to assess sgRNAs that target human dystrophin.

a Schematic illustration showing the generation of humanized DMD exon45 KI-Dmd exon44 knock-out mice (hEx45KI-mdx44). mESCs mouse embryonic stem cells. b Confirmation of exon 44 (ΔEx44) deletion was assessed by RT-PCR in GC, heart, and liver from C57BL/6J, humanized DMD exon45 KI (hEx45KI) or hEx45KI-mdx44 mice. Upper band shows normal transcript and lower band indicates ΔEx44. n = 2 mice per each strain. Similar results were observed in more than 3 experiments. c The lower band from b was sequenced to confirm exon–exon junctions. The black line shows mouse Dmd, and red line shows humanized DMD sequences. d Lack of dystrophin protein expression in GC and heart were confirmed by Western blot analysis. n = 2 mice per each strain. Lack of dystrophin protein in hEx45KI-mdx44 mice was confirmed in more than 3 experiments. e Hematoxylin and eosin (H & E) staining and immunohistochemical staining of dystrophin in TA of hEx45KI (upper) and hEx45KI-mdx44 (lower) mice 9 weeks of age. n = 4 mice per each strain, but only a representative image from each strain is shown. Scale bars indicate 50 µm. f, g GC weight per body weight (no unit) (f) or plasma creatine kinase (CK) level (IU L⁻¹, international unit per liter) (g). White, black, and red columns show C57BL/6J, hEx45KI, and hEx45KI-mdx44 mice, respectively. NT not tested. Data are represented as means ± S.D. (n = 4 mice). p values by two-sided Dunnett’s test showed significantly different from same-age hEx45KI mice.

Fig. 4. LNP delivery of dual sgRNAs into humanized model mice restored dystrophin protein over 12 months.

a Effect of dual sgRNA combination on exon skipping efficiency. A week after the single administration of LNPs encapsulating 10 μg of Cas9 mRNA and the indicated amounts of sgRNA(s) and mRNA from TA in hEx45KI-mdx44 mice (n = 6 mice) was assessed by qRT-PCR. Data are represented as means ± S.D. b Long-term effect of exon skipping was measured by qRT-PCR in GC after a single injection of ASO (phosphorodiamidate morpholino oligomer) or LNP-CRISPR (sgRNA #1 + #23) into hEx45KI-mdx44 mice. Each dot represents data from an individual mouse, and average values from each time point (n = 4 mice) are connected by a solid line to show the trend of each group over time. c, d The recovered dystrophin level was assessed by Western blotting in TA after a single intramuscular administration of ASO (c) or LNP-CRISPR (d). Numbers below the WB images shows relative amount of dystrophin expression normalized by Gapdh amount at each time point. n = 4 mice per each time point, but only two mice of Western blot images are shown.

Fig. 5. LNP-Cas9/sgRNA delivery enables repeated administrations and persistent recovery of dystrophin protein.

a Schematic illustration of the detection of exon skipping efficacy in CAG-Luc2 hDMDEx45 KI reporter mice. Luciferase expression can be detected only if human dystrophin exon 45 skipping occurs. b LNP-CRISPR (5 μg of each sgRNA and 10 µg Cas9 mRNA) or AAV-CRISPR (1 × 10⁹, 1 × 10¹⁰, or 1 × 10¹¹ v.g.) was injected into the left leg of CAG-Luc2 hDMDEx45 KI mice first. Then after 28 days, LNP-CRISPR or AAV-CRISPR was injected into the right leg of the same mice at the same dose as the first injection. Another 28 days later, exon skipping was detected as luciferase activity by IVIS. c Representative bioluminescence images of CAG-Luc2 hDMDEx45 KI mice. LNP-CRISPR successfully induced exon skipping in the right leg after the secondary injection, whereas AAV-CRISPR failed to induce. d Administration frequency-dependent dystrophin recovery in TA one week after one, two, or three intramuscular injections of LNP with one-month intervals in hEx45KI-mdx44 mice. Data are represented as means ± S.D. (n = 3 mice for one injection group, n = 4 mice for other groups). e Immunohistochemical staining of dystrophin 8 weeks after 6 injections of LNP-CRISPR (5 µg of each sgRNA and 10 µg Cas9 mRNA, 6 shots in 2 weeks). Representative images in each group of 8 mice are shown. f Percentage of dystrophin positive fibers in TA. Data are represented as means ± S.D. (n = 8 mice). p values by two-sided Aspin–Welch’s t-test showed significantly different from hEx45KI mice. g Percentage of fibers with a central nucleus in TA. Data are represented as means ± S.D. (n = 8 mice). A significant difference was found among the three groups (two-sided Steel-Dwass test).

Fig. 6. Intravenous delivery of LNP-Cas9/sgRNA by limb perfusion.

a Schematic illustration of mouse hindlimb and relevant muscle groups. The 3D model of hindlimb is adapted from ref. . LNP-CRISPR was injected into the dorsal saphenous vein. b Effect of the injection volume (2.5–10 mL/kg) on exon skipping efficiency in each muscle group of the lower hindlimb following limb perfusion of LNP-CRISPR in hEx45-mdx44 mice. TA tibialis anterior, EDL extensor digitorum longus, GC gastrocnemius, PL plantaris, Sol soleus, PP planta pedis, QD quadriceps. Data are represented as means ± S.D. (n = 4 mice for limb perfusion injection groups, n = 3 mice for intramuscular injection group). p value by two-sided Dunnett’s test showed significantly different from intramuscular injection (2.5 mL/kg) in the same skeletal muscle. c Effect of the injection dose (1–10 mg/kg) of total RNA packaged in LNP under 5 mL/kg injection volume. Data are represented as means ± S.D. (n = 4 mice). p value by two-sided Williams’ test showed significantly different from limb perfusion-PBS of the same skeletal muscle. d Dose-dependent dystrophin restoration in TA and GC following 14 days of limb perfusion of LNP-CRISPR (n = 4 mice for each LNP injection group, n = 2 mice for PBS injection group).