pH-Responsive Polymer Nanoparticles for Efficient Delivery of Cas9 Ribonucleoprotein With or Without Donor DNA

Abstract

Clustered regularly interspaced short palindromic repeat (CRISPR)-associated protein 9 (Cas9) may offer new therapeutics for genetic diseases through gene disruption via nonhomologous end joining (NHEJ) or gene correction via homology-directed repair (HDR). However, clinical translation of CRISPR technology is limited by the lack of safe and efficient delivery systems. Here, facilely fabricated pH-responsive polymer nanoparticles capable of safely and efficiently delivering Cas9 ribonucleoprotein alone (termed NHEJ-NP, diameter = 29.4 nm), or together with donor DNA (termed HDR-NP, diameter = 33.3 nm) are reported. Moreover, intravenously, intratracheally, and intramuscularly injected NHEJ-NP induces efficient gene editing in mouse liver, lung, and skeletal muscle, respectively. Intramuscularly injected HDR-NP also leads to muscle strength recovery in a Duchenne muscular dystrophy mouse model. NHEJ-NP and HDR-NP possess many desirable properties including high payload loading content, small and uniform sizes, high editing efficiency, good biocompatibility, low immunogenicity, and ease of production, storage, and transport, making them great interest for various genome editing applications with clinical potentials.

Keywords: CRISPR-Cas9; genome editing; nanomedicine.

Figure 1 |. Design and characterization of NHEJ-NP and HDR-NP.

(A) The pH-sensitive mPEG-PC7A polymer forms nanoparticle (NP), through a self-assembly process, with either Cas9 RNP alone to enable genome editing via NHEJ (termed NHEJ-NP), or with both Cas9 RNP and ssODN (a donor DNA) to enable genome editing via HDR (termed HDR-NP). (B) NPs are taken up by cells typically via an endocytosis process. The PC7A polymer segments become protonated in the acidic endosomal compartments, thereby triggering the disassembly of the NPs, and enabling the release and endosomal escape of the payload, and ultimately its delivery to the cytoplasm. Thereafter, the Cas9 RNP or Cas9 RNP with ssODN enters the nucleus, facilitated by NLS presented on Cas9, for genome editing. (C) The hydrodynamic diameters of Cas9 RNP, ssODN, empty-NP, NHEJ-NP, and HDR-NP were measured by DLS. (D) A representative TEM image of NHEJ-NP. Scale bar: 50 nm. (E) An electrophoresis assay in 2% agarose gel for sgRNA, ssODN, Cas9 RNP, Cas9 RNP with ssODN, NHEJ-NP, and HDR-NP indicates efficient complexation between the payloads (i.e., Cas9 RNP for Cas9 RNP with ssODN) and mPEG-PC7A. (F) Zeta-potentials of the Cas9 RNP, ssODN, empty-NP, NHEJ-NP, and HDR-NP indicate that the anionic charges of Cas9 RNP and ssODN were neutralized after being encapsulated inside the NPs. (G) The sizes of NHEJ-NP dispersed in PBS, serum-containing cell culture medium, and serum (40 mg/mL bovine serum albumin in PBS) and stored at 4 °C and 37 °C were measured via DLS after 24 h. (H) The size of NHEJ-NP varied in the PBS solution containing either Tween 20 (red) or NaCl (light blue), but remained stable in the PBS solution (dark blue) or the PBS solution containing urea (green) at 37 °C. This study demonstrates that NHEJ-NP was formed by both hydrophobic and electrostatic interactions between the polymer and the Cas9 RNP. (I) The hydrodynamic diameters of NHEJ-NP at different pH conditions were measured by DLS. Data are presented as mean ± s.d. Statistical significance with the size of fresh NHEJ-NP was calculated via one-way ANOVA with Tukey’s post hoc test. *P < 0.05, **P < 0.01, ****P < 0.0001. ns, not significant. NLS, nuclear localization signal.

Figure 2 |. In vitro studies for NHEJ-NPs and HDR-NPs.

(A) Optimization of the pH value for preparation of NHEJ-NP for gene editing in GFP-expressing HEK293 cells. Four days after the treatment, the loss of GFP fluorescence was measured by flow cytometry to assay the editing efficiency via NHEJ. Data are presented as mean ± s.d. (n=3). (B) Optimization of the pH value for preparation of the HDR-NP for gene editing in BFP-expressing HEK293 cells. Gene correction via HDR (correcting CAT to TAC) leads to GFP expression, and gene disruption via NHEJ leads to silence of BFP expression. Four days after the treatment, the gain of GFP and loss of BFP fluorescence were measured by flow cytometry to assay the editing efficiency via HDR and NHEJ, respectively. The HDR efficiency and the HDR/NHEJ ratio are shown. Data are presented as mean ± s.d. (n=3). (C) HDR efficiency of cell-penetrating peptide-conjugated HDR-NPs (i.e., CPP-HDR-NP) in BFP-expressing H9 hESCs. Four days after the treatment, the gain of GFP fluorescence was measured by flow cytometry to assay the editing efficiency via HDR. Data are presented as mean ± s.d. (n=3). (D) Cell viability study of NHEJ-NP and HDR-NP in HEK293 cells via CCK-8 assay. Data are presented as mean ± s.d. (n=3). (E) Study of the endocytosis of NHEJ-NP and CPP-NHEJ-NP. HEK293 cells were first treated with or without various endocytosis inhibitors at 37 °C or were incubated at 4 °C. Cas9 RNPs were labeled with Atto 550-gRNAs and then encapsulated in NHEJ-NP or CPP-NHEJ-NP to treat HEK293 cells. Four hours after treatments, the cellular uptake of Cas9 RNP was quantified by flow cytometry measuring Atto 550⁺ cells. Data are presented as mean ± s.d. (n=4). (F) Intracellular distribution of NHEJ-NP in HEK293 cells was observed by confocal laser scanning microscopy at different time points after treatments. The Cas9 RNP was labeled with Atto 550-gRNA (red). Cells were stained with LysoTracker Green DND-26 (green) and Hoechst 33342 (blue) for endosomes/lysosomes and nuclei, respectively. Experiments were repeated three times and representative images are shown. Scale bar: 50 μm. Data were analyzed by ImageJ. Statistical significance was calculated via one-way ANOVA with Tukey’s post hoc test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. ns, not significant. n.d., not detected. “Lipo2000” represents Lipofectamine 2000. “CRISPRMAX” represents Lipofectamine CRISPRMAX. CPZ, chlorpromazine. MBCD, methyl-β-cyclodextrin.

Figure 3 |. In vivo gene editing with NHEJ-NP in Ai14 reporter mice via intravenous (i.v.) injections.

(A) Ai14 mice employ a STOP cassette that consists of three SV40 polyA sequences to prevent transcription of the downstream tdTomato, a red fluorescent protein. Successful gene editing with the Cas9 RNP that targets SV40 polyA sequences leads to excision of the STOP cassette and then the expression of tdTomato, although tdTomato activation requires removal of at least two of the three SV40 polyA repeat sequences and thus underreports the gene editing efficiency. (B) Ai14 mice were i.v. injected with PBS (n = 3) or NHEJ-NP (n = 3) on Day 0, and the organs and tissues were collected on Day 7 for analysis. (C) NHEJ-NP selectively edited and induced tdTomato expression in livers after intravenous injection, observed by IVIS. Images acquired from one PBS-injected and three NHEJ-NP-injected mice are presented. (D) The total flux of tdTomato fluorescence in major organs. Significant tdTomato fluorescence enhancement was only found in liver, not other organs. (E, F) Immunofluorescence staining of the liver sections from the PBS-injected and NHEJ-NP-injected mice suggests considerable gene editing in liver. Liver sections were stained with anti-RFP antibody for tdTomato (red), anti-hepatocyte specific antigen (green), anti-CD31 antibody for endothelial cells (magenta, pseudo-color in (E)), anti-F4/80 antibody for Kupffer cells (magenta, pseudo-color in (F)), and DAPI for nuclei (blue), respectively. Representative images are shown. Scale bar: 100 μm. Statistical significance was calculated via two-way ANOVA with Tukey’s post hoc test with PBS group as control. ***P < 0.001. ns, not significant. IVIS, in vivo imaging systems.

Figure 4 |. In vivo gene editing with NHEJ-NP in Ai14 reporter mice via intratracheal (i.t.) injections.

(A) Ai14 mice were i.t. injected with PBS (n = 3), isotonic NHEJ-NP (n = 3), and hypotonic NHEJ-NP (n = 3) on Day 0, and the organs and tissues were collected on Day 14 for analysis. (B) NHEJ-NP induced tdTomato expression in lung after intratracheal injections, observed by IVIS. Images acquired from all mice are presented. (C) The total flux of tdTomato fluorescence in lungs were measured. Significant tdTomato fluorescence enhancement was achieved after gene editing. (D) Immunofluorescence staining of the lung sections from the PBS-injected and NHEJ-NP-injected mice suggests efficacious gene editing in lungs. Sections were stained with anti-RFP antibody for tdTomato (red), anti-EpCAM antibody for airway epithelial cells (green), and DAPI for nuclei (blue), respectively. Representative images are shown. Scale bars: 200 μm (10 x) and 50 μm (40 x). Statistical significance was calculated via two-way ANOVA with Tukey’s post hoc test. **P < 0.01, ****P < 0.0001.

Figure 5 |. In vivo gene editing with NHEJ-NP in Ai14 reporter mice via intramuscular (i.m.) injections.

(A) The Ai14 mouse was i.m. injected with NHEJ-NP in the right side of tibialis anterior (TA) muscle on Day 0. The left side of TA muscle was injected with PBS as control. The muscles were collected on Day 7 for analysis. (B) NHEJ-NP efficiently edited and induced tdTomato expression in TA muscles after injections. tdTomato fluorescence in TA muscles was measured by IVIS. Images acquired with two magnifications from both the PBS-injected and NHEJ-NP-injected mice are presented (n=3). Scale bar: 500 μm for 4x images and 100 μm for 20x images. (C) Immunofluorescence staining of the muscle sections from the PBS-injected and NHEJ-NP-injected mice. The muscle sections were stained with anti-RFP antibody for tdTomato (red) and DAPI for nuclei (blue), respectively. Experiments were repeated three times, and representative images with two magnifications are shown. (D) Gene editing efficiency was quantified by the percentage area of the whole muscle section with a genome editing reporter (tdTomato+) or nuclei (control), analyzed by ImageJ. Statistical significance was calculated via one-way ANOVA with Tukey’s post hoc test. **P < 0.01. ns, not significant. IVIS, in vivo imaging systems.

Figure 6 |. In vivo gene editing with HDR-NP via intramuscular (i.m.) injections for Duchenne muscular dystrophy treatment.

(A) The mdx mouse was i.m. injected with HDR-NP in triceps brachii, gastrocnemius, and tibialis anterior muscles on Day 0, respectively. Muscles were harvested on Day 28 for analysis. (B) The four-limb hanging time assay of the PBS-injected mdx mice (n = 7, negative control), HDR-NP-injected mdx mice (n = 11), and untreated wild-type mice (n = 6, positive control) demonstrates the restoration of muscle strength of mdx mice after HDR-NP-mediated gene editing. Data were acquired 28 days after treatments. (C, D) The gene editing efficiency through (C) NHEJ and (D) HDR induced by HDR-NP in the tibialis anterior muscle assayed by Sanger sequencing. Data were acquired 7 days after treatments and presented as mean ± s.d. (n=3). (E) Immunogenicity of PBS (as reference) or HDR-NP injection in tibialis anterior muscles of mdx mice. Expression levels of cytokines were quantified by RT-PCR (n = 3). (F-H) Immunofluorescence staining of (F) triceps brachii, (G) gastrocnemius, and (H) tibialis anterior muscle sections from the PBS-injected and HDR-NP-injected mdx mice and untreated wild-type mice. Muscle sections were stained with anti-dystrophin antibody for dystrophin (red) and DAPI for nuclei (blue), respectively. Results indicate significant restoration of dystrophin expression after HDR-NP treatments. Representative images are presented. Scale bar: 100 μm. (I) The histology of muscle sections was investigated by Masson’s trichrome staining. Reduced muscular dystrophy was observed in HDR-NP-treated mice. Scale bar: 100 μm. Statistical significance was calculated via (B-D) one-way ANOVA or (E) two-way ANOVA with Tukey’s post hoc test. *P < 0.05, ***P < 0.001, ****P < 0.0001. ns, not significant. NC RNP, Cas9 RNP with negative control sgRNA. Dmd RNP, Cas9 RNP with sgRNA that targets Dmd exon 23.