In Vivo Targeted Delivery of Nucleic Acids and CRISPR Genome Editors Enabled by GSH-Responsive Silica Nanoparticles

Abstract

The rapid development of gene therapy and genome editing techniques brings up an urgent need to develop safe and efficient nanoplatforms for nucleic acids and CRISPR genome editors. Herein we report a stimulus-responsive silica nanoparticle (SNP) capable of encapsulating biomacromolecules in their active forms with a high loading content and loading efficiency as well as a well-controlled nanoparticle size (~50 nm). A disulfide crosslinker was integrated into the silica network, endowing SNP with glutathione (GSH)-responsive cargo release capability when internalized by target cells. An imidazole-containing component was incorporated into the SNP to enhance the endosomal escape capability. The SNP can deliver various cargos, including nucleic acids (e.g., DNA and mRNA) and CRISPR genome editors (e.g., Cas9/sgRNA ribonucleoprotein (RNP), and RNP with donor DNA) with excellent efficiency and biocompatibility. The SNP surface can be PEGylated and functionalized with different targeting ligands. In vivo studies showed that subretinally injected SNP conjugated with all-trans-retinoic acid (ATRA) and intravenously injected SNP conjugated with GalNAc can effectively deliver mRNA and RNP to murine retinal pigment epithelium (RPE) cells and liver cells, respectively, leading to efficient genome editing. Overall, the SNP is a promising nanoplatform for various applications including gene therapy and genome editing.

Graphical Abstract

INTRODUCTION

Safe and efficient delivery of biomacromolecules (e.g., nucleic acids and CRISPR ribonucleoproteins (RNPs)) to target cells is of great importance, yet it remains a challenge. Nucleic acids, including DNA and mRNA, are widely used for gene therapy because of their relatively rapid and safe protein production [1–3]. CRISPR-Cas9 RNPs can achieve genome editing by introducing gene deletion, correction, and/or insertion with high efficiency and specificity [4–6]. However, under physiological conditions, naked nucleic acids and RNPs are prone to enzymatic degradation. Moreover, the transfection/gene editing efficiency is negligible due to the lack of cellular uptake and endosomal escape capability [7–10]. In addition, efficient delivery of protein/nucleic acid complexes such as RNP or RNP together with single-stranded oligonucleotide DNA (i.e., RNP+ssODN) for genome editing is hindered by its heterogenous charges and complicated structures [10, 11].

Non-viral nanovectors, including lipid-based [12–16] nanoparticles (NPs), polymer-based NPs [17–20] and functionalized inorganic NPs [21–27] have been actively investigated for the delivery of biomacromolecules. However, current state-of-the-art non-viral nanovectors often suffer a number of limitations, including low payload encapsulation content/efficiency, high cytotoxicity. insufficient in vivo stability, and/or relatively large nanoparticle sizes (typically around 100 nm) [28–30]. To overcome these limitations, we came up with a set of design criteria of a desirable biomolecule nanocarrier including: (1) high loading content and loading efficiency, while maintaining the payload activity, (2) small NP size (e.g., hydrodynamic diameter < 50 nm), (3) versatile surface chemistry (e.g., ligand conjugation) to facilitate the payload delivery to target cells, (4) excellent biocompatibility, (5) efficient endo/lysosomal escape capability, (6) rapid payload release in the target cells, and (7) ease of handling, storage, and transport.

Silica-based delivery platforms have been extensively investigated due to their unique characteristics, such as high stability, multifunctionality, biocompatibility and versatility in chemistry [31, 32]. The most studied ones are mesoporous silica nanoparticles, which have been developed for biomacromolecule delivery, but the payloads are mostly exposed on the nanoparticle surface and subject to premature release or degradation [33]. Considering most biomacromolecules have good solubility in an aqueous solution but not in organic solvents, we sought to use a water-in-oil microemulsion method to synthesize a biodegradable silica nanoparticle (SNP), which can encapsulate and protect the payloads before reaching the target cells, but can also rapidly release the payloads intracellularly through stimulus-triggered NP degradation. Furthermore, when appropriate surfactants, oil phase, water phase and nanoparticle formulations are chosen, various hydrophilic payloads can be encapsulated into SNP with high loading content and efficiency via in situ SNP formulation [34, 35]. The size and polydispersity of the final SNP can be controlled by emulsification conditions (e.g., water to oil ratio and type/ratio of surfactant) [34, 36], while functional moieties can be integrated into the SNP by the addition of silica reactants, making the SNP a versatile nanoplatform [35, 37].

Herein, we report the facile fabrication of a sub-50 nm, multifunctional and stimuli-responsive SNP via a water-in-oil microemulsion method. A disulfide bond-containing crosslinker was integrated into the SNP network to achieve glutathione (GSH)-responsive degradation and cargo release in the cytosol. Functional moieties including imidazole capable of facilitating endosomal escape can be incorporated into SNPs during the fabrication process. The surface of SNP can be further tuned with charges from negative to positive, or conjugated with poly(ethylene glycol) (PEG). PEG can shield the surface charge of the SNP, enhance its in vivo stability and circulation time, and allow versatile ligand conjugations onto the SNP, which is necessary for targeted delivery [38]. This multifunctional SNP can encapsulate various biomacromolecules (i.e., DNA, mRNA, Cas9 RNP and Cas9 RNP with single-stranded oligonucleotide DNA (ssODN)) with superior loading content and loading efficiency, independent of payload types or surface charges. The gene transfection or editing efficiencies of the multifunctional SNP for the delivery of nucleic acids (i.e., DNA and mRNA) and CRISPR gene editors (e.g., Cas9 RNP and RNP+ssODN) were systematically investigated.

For in vivo study, we proved efficient delivery of mRNA and Cas9 RNP in the retinal pigmented epithelium (RPE) using SNP in Ai14 transgenic mice. As RPE plays an important role in neuroretina survival and visual function, the dysfunction of RPE will result in various eye diseases [39, 40]. To enhance the RPE-specific internalization, a targeting ligand, all-trans-retinoic acid (ATRA) was conjugated on SNP-PEG (i.e., SNP-PEG-ATRA). ATRA binds to the inter-photoreceptor retinoid-binding protein that selectively transports all-trans-retinol to the RPE [41, 42]. SNP-PEG-ATRA can efficiently deliver both mRNA and RNP into RPE in transgenic Ai14 mice, leading to high genome editing.

SNPs can also effectively deliver mRNA and genome editors in adult Ai14 mice through systemic administration routes. Without any targeting moieties, SNP-PEG showed a preferential delivery to liver. To further enhance the liver targeting capability, a targeting ligand, N-acetylgalactosamine (GalNAc) was conjugated onto SNP-PEG (i.e., SNP-PEG-GalNAc), as GalNAc is known for its ability to bind with higher selectivity to the asialoglycoprotein receptors (ASGPRs) overexpressed on hepatocytes [43, 44]. SNP-PEG-GalNAc showed a 2-fold higher mRNA and RNP delivery efficiency in liver than non-targeted SNP-PEG according to the tdTomato expression levels induced by genome editing. Additionally, no significant change in the blood biochemical profile of SNP treated mice was observed. Overall, this multifunctional SNP is an efficient, biocompatible, and versatile nanoplatform for targeted delivery of a broad range of biomacromolecule cargos both in vitro and in vivo.

EXPERIMENTAL SECTION

Materials

Tetraethyl orthosilicate (TEOS), 1H-imidazole-4-carboxylic acid, thionyl chloride (SOCl₂), Traut’s reagent (2-iminothiolane), Triton X-100, acetone, ethanol, glutathione (GSH), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC), N-hydroxysuccinimide (NHS), tris(2-carboxyethyl)phosphine hydrochloride (TCEP) and ammonia (30% in water) were purchased from Fisher Scientific, USA. Hexanol, cyclohexane, and (3-aminopropyl)triethoxysilane (APTES) were bought from Tokyo Chemical Industry Co., Ltd., USA. Triethylamine (TEA) and dimethyl sulfoxide (DMSO) were purchased from Alfa Aesar, USA. Bis[3-(triethoxysilyl)propyl]-disulfide (BTPD) was purchased from Gelest, Inc., USA. Methoxy-poly(ethylene glycol)-silane (mPEG-silane, M_n = 5000), amine-poly(ethylene glycol)-silane (NH₂-PEG-silane, M_n = 5000) and maleimide-poly(ethylene glycol)-silane (Mal-PEG-silane, M_n = 5000) were purchased from Biochempeg Scientific Inc., USA. All-trans-retinoic acid (ATRA) and 4-aminophenyl 2-acetamido-2-deoxy-β-D-glucopyranoside were purchased from Santa Cruz Biotechnology, USA. Nuclear localization signal (NLS)-tagged Streptococcus pyogenes Cas9 nuclease (sNLS-SpCas9-sNLS) was obtained from Aldevron, USA. In vitro transcribed single guide RNAs (sgRNAs) and ssODNs were purchased from Integrated DNA Technologies, Inc., USA.

Synthesis of N-(3-(Triethoxysilyl)propyl)-1H-Imidazole-4-Carboxamide (TESPIC)

A mixture of 1H-imidazole-4-carboxylic acid (250 mg, 1.9 mmol) and SOCl₂ (4 mL) was refluxed at 75 °C overnight. The reaction mixture was then cooled down to room temperature and added into anhydrous toluene (20 mL). The precipitate was collected by filtration and vacuum-dried to yield the intermediate, 1H-imidazole-4-carbonyl chloride. The as-prepared 1H-imidazole-4-carbonyl chloride was suspended in anhydrous THF (5 mL), followed by the addition of triethylamine (232 mg, 2.3 mmol) and APTES (420 mg, 1.9 mmol). The mixture was stirred at room temperature overnight under a nitrogen atmosphere, and then filtered. The solvent was removed by rotary evaporation to yield the final product TESPIC. Since the silica reactants have the tendency to undergo hydrolysis/polymerization during column purification, TESPIC was synthesized and used without purification [37, 45]. ¹H NMR (400 MHz, DMSO-D6): δ 0.62 (dd, 2 H, J = 14.6, 6.2 Hz), δ 1.12 (t, 9 H, J = 7.0 Hz), δ 1.60 (dt, 2 H, J = 15.9, 8.0 Hz), δ 2.70 (m, 2 H), δ 3.83 (q, 6 H, J = 6.0 Hz), δ 7.00 (s, 1 H), δ 7.40 (s, 1 H). ¹³C NMR (100 MHz, DMSO-D6): δ 166, 137, 134, 128, 58, 43, 23, 18, and 7.6.

Preparation of GalNAc-PEG-Silane

A mixture of 4-Aminophenyl 2-Acetamido-2-deoxy-β-D-glucopyranoside (37.5 mg, 0.12 mmol) and Traut’s reagent (2-iminothiolane, 13.8 mg, 0.1 mmol) was dissolved in anhydrous DMSO (2 mL) and stirred at room temperature for 24 h. Thereafter, a portion of Mal-PEG-silane (500 mg, 0.1 mmol) dissolved in 10 ml anhydrous DMSO was added to the above solution. After 24 h, the reaction was terminated by precipitation into diethyl ether (100 mL). The precipitate was washed twice with diethyl ether, and vacuum dried to obtain the product GalNAc-PEG-silane without further purification. ¹H NMR (400 MHz, DMSO-D6) spectrum of GalNAc-PEG-silane was shown in Figure S1.

Preparation of the GSH-Responsive Silica Nanoparticles (SNPs)

SNPs were synthesized by a water-in-oil microemulsion method. Triton X-100 (1.8 mL) and hexanol (1.8 mL) were dissolved in cyclohexane (7.4 mL) to form the oil phase. Separately, 30 μL of an aqueous solution containing the desired payload (e.g., DNA, mRNA, RNP or RNP+ssODN) (5 mg/mL) were mixed with TEOS (3.1 μL, 14 μmol), BTPD (6 μL, 13 μmol) and TESPIC (1 mg (i.e., 3 μmol) for imidazole incorporation at a 10% molar ratio, or 2 mg for 20% molar ratio). After shaking, this mixture was added to the oil phase (1.1 mL), and then the water-in-oil microemulsion was formed by vortex for 1 min. Under stirring (1500 rpm), an aliquot of 30% aqueous ammonia solution (5 μL) was added and the water-in-oil microemulsion was under stirring at room temperature for 4 h to obtain unmodified SNP with negative surface charge. To prepare positively charged SNP (SNP-NH₂), the as-prepared SNP was modified with amine groups by the addition of APTES to the microemulsion, and the mixture was stirred vigorously for another 4 h at room temperature. To purify unmodified SNP or SNP-NH₂, acetone (1.5 mL) was added in the microemulsion in order to precipitate the SNPs, and the precipitate was recovered by centrifugation and was subsequently washed twice with ethanol and three times with water. The purified SNP or SNP-NH₂ were finally collected by centrifugation.

Preparation of PEGylated SNP (SNP-PEG) and GalNAc-Conjugated SNP (SNP-PEG-GalNAc)

The as-prepared, unmodified SNP (2 mg) was re-dispersed in 2 mL water. An aliquot of mPEG-silane (for neutral surface charge, 200 μg), or NH₂-PEG-silane (40 μg) + mPEG-silane (160 μg) (for subsequent ATRA conjugation), or GalNAc-PEG-silane (80 μg) + mPEG-silane (120 μg) (for SNP-PEG-GalNAc) was added to the above mixture. The pH of the solution was adjusted to 8 using 0.1 M NaOH solution. The solution was stirred at room temperature for 4 h. The resulting SNP-PEG, SNP-PEG-NH₂, or SNP-PEG-GalNAc was purified by washing with water for three times and collected by centrifugation. To study the serum stability of SNP-PEG, freshly prepared SNP-PEG was dispersed in fetal bovine serum (FBS) at a concentration of 0.5 mg/ml. The nanoparticle dispersion was incubated at 37 °C. At each time point, a portion of the SNP-PEG dispersion (0.2 ml) was collected, the SNP-PEG was harvested by centrifugation and re-dispersed in DI water for nanoparticle size analysis by DLS. For stability tests, mRNA-encapsulated SNP-PEG were redispersed in DI water with SNP concentration of 1 mg/ml and stored at different temperatures (i.e., 4 °C, −20 °C and −80 °C); RNP encapsulated SNP-PEG were redispersed in RNP storage buffer (20 mM HEPES-NaOH pH 7.5,150 mM NaCl,10 % glycerol), flash frozen in liquid nitrogen, and stored at −80 °C.

Preparation of ATRA-Conjugated SNP (SNP-PEG-ATRA)

SNP-PEG-ATRA was synthesized via EDC/NHS catalyzed amidation. Payload-encapsulated SNP-PEG-NH₂ (1 mg) was re-dispersed in 0.5 mL DI water. EDC (15 μg), NHS (9 μg) and a DMSO solution of ATRA (12 μg in 10 μL DMSO) were added to the above solution. The solution was stirred at room temperature for 6 h, and then the resulting SNP-PEG-ATRA was washed by water three times and collected by centrifugation. The size and zeta-potential change before and after ATRA conjugation onto the surface of SNP is summarized in Table S1.

Characterization

The chemical structure of TESPIC was confirmed by nuclear magnetic resonance (NMR) spectroscopy (Avance 400, Bruker Corporation, USA). The Fourier transform infrared (FTIR) spectra were recorded on a Bruker Tensor 27 FTIR spectrometer. The hydrodynamic diameters and zeta potentials of the SNPs were characterized by a dynamic light scattering (DLS) spectrometer (Malvern Zetasizer Nano ZS) at a 90° detection angle with a sample concentration at 0.1 mg/mL. The morphology of DNA-loaded SNP-PEG was characterized by transmission electron microscopy (TEM, Tecnai 12, Thermo Fisher, USA).

To calculate the loading content and loading efficiency of the payloads in the SNPs, SNPs were re-suspended in water (1 mg/mL, 40 μL) and incubated with 0.1 M GSH aqueous solution (pH 7.4, 160 μL) with a pH of 7.4 for 1 h to allow for complete release of the payload. The RNP loading contents and loading efficiencies were measured via a bicinchoninic acid assay (BCA assay, Thermo Fisher, USA). DNA and mRNA loading contents and loading efficiencies were quantified using a NanoDrop One (Thermo Fisher, USA) by measuring OD₂₆₀.

Cell Culture for In Vitro Studies

Human embryonic kidney cells (i.e., HEK293 cells) were used for in vitro studies. HEK293 cells were purchased from ATCC. Green fluorescence protein (GFP)-expressing HEK 293 cells were bought from GenTarget Inc. Blue fluorescence protein (BFP)-expressing HEK 293 cells generated through lentiviral transduction of a BFP dest clone was obtained from Addgene [46]. All HEK 293 cells were cultured with DMEM medium (Gibco, USA) added with 10% (v/v) fetal bovine serum (FBS, Gibco, USA) and 1% (v/v) penicillin–streptomycin (Gibco, USA). Cells were cultured in an incubator (Thermo Fisher, USA) at 37 °C with 5% carbon dioxide at 100% humidity.

DNA and mRNA Transfection Efficiency Study

A red fluorescence protein (RFP)-expressing plasmid DNA (i.e., RFP-DNA, Addgene #40260, USA) and an RFP-mRNA (Trilink Biotechnologies #L-7203, USA) were used for DNA and mRNA transfection studies, respectively. HEK293 cells were placed into 96-well plates 24 h prior to treatment, at a density of 15,000 cells/well. Cells were incubated with either RFP-DNA-loaded SNPs, or RFP-mRNA-loaded SNPs. A commercially available transfection agent, Lipofectamine 2000 (Lipo 2000), was used as the positive control. The dosage of DNA or mRNA was 200 ng/well. The Lipo 2000-DNA (or Lipo 2000-mRNA) complex was prepared following the protocols provided by the manufacturer, with a final dosage of Lipo 2000 at 0.5 μL per well. An untreated group was used as the negative control. After 48 h, cells were harvested with 0.25% trypsin-EDTA, spun down and resuspended in 500 μL PBS. RFP expression efficiencies were obtained with a flow cytometer and analyzed with FlowJo 7.6.

To study the stability of DNA encapsulated SNPs in the presence of GSH, transfection experiments were carried out under similar conditions as described above except GSH was intentionally added to the cell culture media with a GSH concentration ranging from 0.001 to 10 mM.

RNP Genome-Editing Efficiency Study

For gene editing studies, GFP-expressing HEK 293 cells were used as an RNP delivery cell model. RNP was prepared by mixing sNLS-SpCas9-sNLS and sgRNA (GFP protospacer: 5’-GCACGGGCAGCTTGCCGG-3’) at 1:1 in molar ratio. Cells were seeded at a density of 5,000 cells per well onto a 96-well plate 24 h before treatment. Cells were treated with RNP-loaded SNP, RNP-complexed Lipo 2000 (0.5 μL/well) or RNP-complexed Lipofectamine CRISPRMAX (CRISPRMAX). Untransfected cells, as well as cells treated with negative control SNP (i.e., SNPs encapsulating an RNP with a negative control sgRNA (Alt-R CRISPR-Cas9 Negative Control crRNA #1, Integrated DNA Technologies, Inc., USA)) were used as controls. For each treatment, the RNP dosage was kept at 150 ng/well, with an equivalent Cas9 protein dosage at 125 ng/well.

The T7 Endonuclease I (T7E1) indel detection assay (#E3321, New England Biolabs, USA) was performed following the protocol from manufacturer. Briefly, genomic DNA of GFP-HEK cells was extracted using Monarch Genomic DNA Purification Kit (#T3010, New England Biolabs, USA) according to the manufacturer’s instructions. The sgRNA targeted genomic locus was amplified with Phusion Hot Start II High Fidelity DNA Polymerase (Thermo Fisher, USA) using the designed primers (forward primer: 5’-CGCCCCATTGACGCAAATGGGCGGTAG-3’; Reverse primer: 5’-GCTGTTGTAGTTGTACTCCAGCTTGTGCCCCAGGATG-3’). The amplicons were then treated with T7E1 and incubated at 7 °C for 30 min. The digested DNA was analyzed using 2% agarose gel electrophoresis. Indel formation efficiencies were calculated using Image J.

For gene correction studies, BFP-expressing HEK 293 cells were employed as a model cell line [46]. The RNP+ssODN mixture was prepared by simply mixing the as-prepared BFP gene-targeting RNP (BFP protospacer: 5’-GCTGAAGCACTGCACGCCAT-3’) and single-stranded oligonucleotide DNA (ssODN) (BFP to GFP ssODN sequence: 5’-CCTGAAGTTCATCTGCACCACCGGCAAGCTGCCCGTGCCCTGGCCCACCCTCGTG ACCACCCTGACGTACGGCGTGCAGTGCTTCAGCCGCTACCCCGACCACATGA-3’, changing BFP to GFP via alternation of histidine to tyrosine) donor template at 4 °C for 5 min at a 1:1 molar ratio [46, 47]. Gene correction will lead to the replacement of three nucleotides in the genome converting BFP to GFP, and thus the percentage of GFP positive cells can be used to evaluate the genome editing efficiency [47, 48]. BFP-expressing HEK 293 cells were seeded at a density of 5,000 cells per well onto a 96-well plate 24 h before treatment. Cells were treated with RNP+ssODN-loaded SNP or Lipo 2000 (0.5 μL/well) carrying RNP+ssODN as the positive control. For each treatment, the RNP+ dosage was kept at 150 ng/well (i.e., an equivalent Cas9 protein dosage of 125 ng/well), and the ssODN dosage was 25 ng/well. The precise genome editing efficiencies were quantified six days after treatment using flow cytometry by counting the percentage of green fluorescence positive cells. Data were analyzed with FlowJo 7.6.

Cell Viability Assay

The cytotoxicity of SNPs was studied by an MTT assay. Cells were treated with complete medium, DNA-complexed Lipo 2000 (0.5 μL/well), and DNA-loaded SNP-PEG with concentrations ranging from 10 to 1000 μg/mL. Cell viability was measured using a standard MTT assay 48 h after treatment (Thermo Fisher, USA). Briefly, cells were treated with media containing 500 μg/mL MTT and incubated for 4 h. Then, the MTT-containing media was aspirated, and the purple precipitate was dissolved in 150 μL of DMSO. The absorbance at 560 nm was obtained with a microplate reader (GloMax® Multi Detection System, Promega, USA).

Subretinal Injection

All animal research was approved by UW-Madison animal care and use committee. Ai14 reporter mice (obtained from The Jackson Laboratory) were used to assess the mRNA delivery/genome editing efficiency induced by mRNA- or RNP-encapsulated SNP-PEG-ATRA, respectively. Cre-mRNA was purchased from Trilink Biotechnologies, USA (#L-7211). RNPs were prepared using either a sgRNA targeting the stop cassette composed of 3x SV40 polyA blocks (protospacer: 5’-AAGTAAAACCTCTACAAATG-3’) in Ai14 mice, or a mouse negative control sgRNA (Alt-R CRISPR-Cas9 Negative Control crRNA #1, Integrated DNA Technologies, Inc., USA). Subretinal injection and subsequent RPE tissue collection were performed as reported previously [49]. Mice were maintained under tightly controlled temperature (23 ± 5 °C), humidity (40–50%), and light/dark (12/12 h) cycle conditions in a 200-lux light environment. The mice were anesthetized by intraperitoneal injection of ketamine (80 mg/kg), xylazine (16 mg/kg) and acepromazine (5 mg/kg) cocktail. For mRNA delivery studies, right eyes of mice were injected with mRNA-encapsulated SNP-PEG-ATRA (2 μL with 4 μg mRNA), and left eyes were injected with PBS. For RNP delivery studies, right eyes of mice were injected with SNP-PEG-ATRA encapsulating RNP with a sgRNA targeting the Ai14 stop cassette (i.e., Ai14 SNP), left eyes of mice were injected with SNP-PEG-ATRA encapsulating RNP with a negative control sgRNA (i.e., negative control SNP). The injection volume was 2 μL, containing 4 μg RNP. SNP-PEG-ATRA was injected into the subretinal space using a UMP3 ultramicro pump fitted with a NanoFil syringe, and the RPE-KIT (all from World Precision Instruments, Sarasota, FL) equipped with a 34-gauge beveled needle. The tip of the needle remained in the bleb for 10 s after bleb formation, then it was gently withdrawn.

Four days (for mRNA delivery studies) or fourteen days (for RNP delivery studies) post subretinal injection, mice were sacrificed, and eyes were collected for tdTomato expression evaluation. Collected eyes were rinsed twice with PBS and puncture was made at ora serrata with an 18-gauge needle. The eye was opened along the corneal incisions and the eyecup was incised radially to the center and flattened to give a final floret shape. The RPE layer was then separated and flat-mounted on a cover-glass slide (i.e., RPE floret). RPE florets were imaged with a NIS-Elements using a Nikon C2 confocal laser scanning microscope (CLSM).

Intravenous Injection

Ai14 mice (6–8 weeks; three mice in each group) were injected with Cre-mRNA (20 μg per mouse) or RNP (100 μg per mouse)-encapsulated SNP-PEG or SNP-PEG-GalNAc through retro-orbital injections; PBS injected Ai14 mice were used as controls. The SNP-injected and control mice were sacrificed 3 days (Cre mRNA) or 7 days (RNP) post-injection. Organs and tissues (liver, heart, lung, spleen, kidney and muscle) were then collected and analyzed.

Fresh organs/tissues were imaged using the in vivo imaging system (IVIS Lumina system, Perkin Elmer) for tdTomato expression. A portion of the liver samples were weighed and homogenized with cell lysis buffer as reported previously [50]. Briefly, the liver tissues were washed with chilled 1X PBS, cut into small pieces and accurately weighed. RIPA buffer with protease inhibitor (cOmplete^™ Protease Inhibitor Cocktail, Sigmaaldrich, USA) was added to the liver tissue with 1 ml RIPA buffer for every 50 mg tissue. The liver tissue samples were homogenized on ice by a sonicate homogenizer for 5 min at a power of 150 watts with 10 seconds sonication/10 seconds rest cycles, and kept on ice for 1 h with occasional vortex. The tissue lysate was then centrifuged at 15,000 x g for 30 minutes at 4°C to yield pellet of cell debris. The supernatant was transferred to a fresh microfuge tube without disturbing the pellet. The protein concentration of the lysate was determined by the BCA protein assay and kept the same prior to IVIS analysis. The homogenized liver samples were then added to a 96-well black/clear flat bottom Imaging Microplate (Corning Life Science, USA). Thereafter, the tdTomato fluorescence was measured and analyzed by the IVIS system.

Immunofluorescence Staining

Tissues were fixed in 4% paraformaldehyde (PFA) at RT for 24 hours, then switched to PBS solution containing 30% sucrose and stored at 4 °C for 72 h. Thereafter, the tissues were embedded in Tissue-Tek® Optimal Cutting Temperature Compound (Sakura Finetek, USA), and frozen in dry ice. The blocks were sectioned using a cryostat machine (CM1900, Leica Biosystems, USA) at 8 μm thickness and mounted on microscope slides. The sections were incubated in 10% goat serum and 0.3% Trixon X-100 in PBS at RT for 1h. For immunofluorescence staining, the sections were first incubated with a rabbit anti-tdTomato primary antibody (ab152123, 1:5000, Abcam, USA) for 1 h at RT. The primary antibody was then detected by a fluorescence-conjugated secondary antibody (goat anti-rabbit IgG H&L (Alexa Fluor® 594), ab150080, 1:1000, Abcam, USA). Finally, the slides were mounted with DAPI and covered with microscope cover glasses. All of the images were acquired using CLSM.

Blood Biochemical Profile

Blood samples were immediately collected from the orbital sinus of each mouse from the SNP-treated groups or PBS control groups and centrifuged at 1500 x g and 4°C for 10 min for serum preparation. Clinical biochemical assessment of levels of blood urea nitrogen (BUN), creatinine (CRE), alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), total bilirubin (TBIL), glucose (GLU), Calcium (CA), total protein (TP), albumin (ALB), globulin (GLOB), Na⁺, K⁺, Cl^- and total carbon dioxide (tCO2) was performed using VetScan Preventative Care Profile Plus rotors (Abaxis, USA) in a VetScan VS2 chemistry analyzer (Abaxis, USA).

Statistical Analysis

Results are presented as mean ± standard deviation (SD). One-way analysis of variance (ANOVA) with Tukey’s multiple comparisons was used to determine the difference between independent groups. Statistical analyses were conducted using GraphPad Prism software version 6.

RESULTS

Design and Characterization of GSH-Responsive SNPs

Silica nanoparticles (SNPs) were synthesized via a water-in-oil microemulsion method (Figure 1A and B). A mixture of payload aqueous solution and silica reagents was emulsified in a continuous oil phase. Tetraethyl orthosilicate (TEOS) was used as a basic building block that constructs the silica network. A disulfide bond-containing crosslinker, bis[3-(triethoxysilyl)propyl]-disulfide (BTPD), was incorporated into the silica network to yield the GSH-responsive SNP. To enhance endosomal escape, a imidazole-containing silica reagent, N-(3-(triethoxysilyl)propyl)-1H-imidazole-4-carboxamide (TESPIC), was added to the microemulsion. The Fourier transform infrared (FTIR) absorption spectra confirmed the successful incorporation of imidazole groups and disulfide crosslinkers (Figure S2) [51, 52]. The SNP surface can be modified with amine groups by addition of (3-aminopropyl)triethoxysilane (APTES) to yield positively-charged SNP (i.e., SNP-NH₂). The as-prepared, unmodified SNP or SNP-NH₂ were purified by precipitation in acetone, and washed by ethanol and deionized (DI) water. Unmodified SNP can be PEGylated to obtained neutral surface charge. The liver-targeting SNP (i.e., SNP-PEG-GalNAc) was prepared by simultaneous addition of silane-PEG-GalNAc and silane-mPEG. The RPE-targeting ligand, ATRA, was conjugated to the distal ends of the surface PEG through amidation.

Figure 1. Design and synthesis of SNPs conjugated with either ATRA or GalNAc.

(A) Illustration of the multifunctional SNP for the delivery of nucleic acids (e.g., DNA and mRNA) and CRISPR genome editor (e.g., RNP, RNP+ssODN). (B) Synthesis scheme for the ATRA- or GalNAc-conjugated SNPs. SNP: silica nanoparticle; TEOS: tetraethyl orthosilicate; TESPIC: N-(3-(triethoxysilyl)propyl)-1H-imidazole-2-carboxamide; BTPD: bis3-[(triethoxysilyl) propyl]-disulfide; PEG: polyethylene glycol; GalNAc: N-acetylgalactosamine; ATRA: all-trans-retinoic acid.

The morphology of the DNA-loaded SNP-PEG was characterized by transmission electron microscopy (TEM). Figure 3A shows a TEM image of the PEGylated SNP with a spherical structure and an average size of 40 nm. the GSH-responsiveness of SNP-PEG was further studied by TEM, SNP-PEG was dispersed in 2 mM GSH solution and incubated at 37 °C. The morphology of the SNP changed significantly after incubation in 2 mM GSH for 1 h, indicating the GSH-responsive cargo release capability of the SNPs. The hydrodynamic diameter of DNA-loaded SNP-PEG was 45 nm, as measured by dynamic light scattering (DLS) (Figure 3B). The zeta-potential of DNA-loaded SNP-PEG was 6.4 mV, indicating a nearly neutral surface charge after PEGylation.

Figure 3. Characterization of SNP and formulation optimization.

(A) Size distribution of SNPs measured by DLS with an average hydrodynamic diameter of 46 nm. (B) TEM image of SNP. (C) The effect of (1) molar ratio of TESPIC, and (2) surface charge in DNA-delivery by SNP. The transfection efficiency was quantified by the percent of RFP-positive HEK293 cells 48 h post treatment. (D) Effects of GSH concentration in cell culture medium on the DNA transfection efficiency of SNP-PEG. (E) mRNA delivery efficiency of SNP after storage at different conditions. NS: not significant; *: p < 0.05; **: p < 0.01; ****: p<0.0001; n = 3.

The SNP formulation was then optimized in HEK 293 cells to achieve high transfection efficiencies, using DNA and mRNA as payloads, separately. Imidazole groups can be quickly protonated in acidic endocytic compartments, and thus enhance the endo/lysosomal escape capability of the SNP-PEG due to the proton sponge effect (Figure 2) [53, 54]. Therefore, the ratio of imidazole in the SNP is an essential factor for efficient nucleic acid delivery. The optimal ratio of imidazole-containing reactant TESPIC in the SNP was investigated by fixing the feed molar ratio of TEOS and BTPD. As shown in Figure 3C, SNP-PEG with 10 mol% imidazole-containing TESPIC exhibited higher DNA transfection efficiency (1.3-fold) than the one without TESPIC, but further increase in the TESPIC molar ratio did not lead to higher DNA transfection efficiency. Therefore, the TESPIC ratio was set at 10 mol% for further studies. The TESPIC ratio in mRNA-encapsulated SNP-PEG was also optimized, as shown in Figure S4. Different from DNA delivery, the mRNA delivery efficiency was independent of the TESPIC ratio. This discrepancy may be attributed to the difference in cytoplasm stability of DNA and mRNA [55, 56]. One major limiting step for mRNA delivery is its poor cytosolic stability [55, 57, 58]. In other words, although imidazole-containing SNP can induce rapid endo/lysosomal escape and efficiently transport mRNA payload into the cytosol, due to mRNA’s poor cytosolic stability, it could result in less durable transfection. As a result, incorporation of imidazole did not significantly enhance mRNA delivery efficiency.

Figure 2.

Schematic illustration of the intracellular trafficking pathways of SNP.

To investigate the influence of SNP surface charges on nucleic acid delivery efficiencies, we prepared DNA- and mRNA-encapsulated SNPs with different surface charges (Figure 3C and Figure S4). The as-prepared, unmodified SNP had a strong negative zeta-potential; positively charged SNP (i.e., SNP-NH₂) and neutral PEGylated SNP (SNP-PEG) were prepared by APTES and mPEG-silane conjugation, respectively. As shown in Figure 3C, SNP-NH₂ exhibited a 1.6-fold higher DNA transfection efficiency and a 1.8-fold higher mRNA transfection efficiency than negatively charged SNP. This finding is consistent with previous reports [59–61]. In comparison with negatively charged NPs, positively charged NPs can induce higher cargo-delivery efficacy by promoting the rate of nanoparticle internalization, increasing the total number of internalized nanoparticles, and enhancing the endo/lysosomal escape capability. SNP-PEG with a neutral surface charge exhibited similar DNA and mRNA transfection efficiencies as SNP-NH₂, indicating that moderate surface PEGylation did not significantly affect the SNP uptake by HEK 293 cells.

Disulfide bonds were integrated into the SNP to facilitate payload release in the cytosol with a high GSH concentration (2−10 mM). The extracellular GSH (0.001–0.02 mM), although much lower, can still raise stability concerns or induce premature cargo release. Hence, to study the GSH-responsive behavior of SNP, DNA encapsulated SNP was incubated with HEK 293 cells in culture media containing intentionally added GSH with a GSH concentration ranging from 0–10 mM. As shown in Figure 3D, the DNA transfection efficiency was not affected at GSH concentrations equal to or lower than 0.1 mM, suggesting that the SNP is stable in the extracellular space. However, a significant decrease in the DNA transfection efficiency was observed at a GSH concentration of 1 mM or higher, suggesting that the SNP are not stable at high GSH concentrations, therefore, they can effectively break down in the cytosol to release the payload. The serum stability of SNP-PEG was investigated. SNP-PEG was dispersed in fetal serum albumin (FBS) and incubated at 37 °C, the hydrodynamic diameter was monitored by DLS at each time point. As shown in Figure S5, the size of SNP-PEG remained unchanged after 48 h and only slight size change was observed after 72 h, indicating excellent serum-stability of SNP-PEG for in vivo applications. The stability of mRNA-loaded SNP-PEG after long-term storage was studied. The mRNA transfection efficiency of SNP-PEG was intact after 60-day storage at −80 °C, or 25 days at 4 °C or −20 °C (Figure 3E), indicating SNP-PEG is desirable for future biomedical applications.

In Vitro Nucleic Acid and Genome Editor Delivery by SNPs

A variety of biomacromolecules were encapsulated into SNP, including plasmid DNA, mRNA, RNP and the mixture of RNP and donor oligonucleotide for gene correction (i.e., RNP+ssODN). The hydrodynamic diameter, zeta-potential, loading content and loading efficiency of PEGylated SNP with different payloads were summarized in Table 1. The size and zeta-potential of SNP-PEG remained relatively constant for the various payloads investigated. As shown in Table 1, the loading contents varied between 9.0–9.4 wt%, with an overall high loading efficiency of >90%. In particular, there was no significant difference in loading content and loading efficiency between payloads, indicating that the SNP is a versatile nanoplatform for encapsulation of nucleic acids, proteins, and gene editors.

Table 1.

Summary of the SNP-PEG size, zeta-potential, loading content and loading efficiency of different payloads. Hydrodynamic diameter and zeta-potential data are presented as mean ± SD.

Payload	Hydrodynamic diameter (nm)	Zeta-potential (mV)	Loading content (wt%)	Loading efficiency (%)
DNA	45 ± 2	6.4 ± 0.6	9.0	90
mRNA	46 ± 4	3.0 ± 0.3	9.2	91
RNP	52 ± 4	6.5 ± 0.4	9.1	90
RNP+ssODN	49 ± 3	5.9 ± 0.7	9.4	93

The intracellular trafficking of RNP-encapsulated SNP-PEG was studied by confocal laser scanning microscopy (CLSM) in HEK 293 cells (Figure 4). RNP was prepared by mixing the NLS-tagged Cas9 and ATTO-550-tagged guide RNA. After the cells were incubated with the RNP-loaded SNP-PEG for 0.5 hour, RNP was mainly co-localized with endo/lysosomes, indicating the internalization of SNP-PEG via endocytosis. Endo/lysosomal escape of the SNP-PEG assisted by imidazole was observed 2 h post-treatment, indicated by a decrease in the co-localized RNP and endo/lysosome signals. The RNP signal showed considerable overlap with the nucleus and further decreased co-localization with endo/lysosomes 6 h post-treatment, indicating the successful nuclear transportation of RNP induced by the NLS tags on the RNP.

Figure 4. Intracellular trafficking of SNP.

Colocalization of ATTO-550-tagged RNP and endo/lysosomes was studied at 0.5 h, 2 h, and 6 h post-treatment.

To investigate the versatility of the SNP for biomacromolecule delivery, HEK 293 cells were used for nucleic acid delivery/genome editing efficiency studies, and flow cytometry was used to quantify the delivery efficiency. The DNA and mRNA transfection efficiency by SNP-PEG were tested in HEK293 cells (Figure 5A and 5B). SNP-PEG exhibited statistically higher DNA and mRNA transfection efficiency (1.3-fold and 1.1-fold, respectively) than the commercially available transfection reagent, Lipofectamine 2000 (Lipo 2000), indicating the superior nucleic acid delivery capability of SNP.

Figure 5. Delivery efficiency of nucleic acids and CRISPR-Cas9 genome editors by SNP.

(A) and (B) The transfection efficiency of (A) DNA- and (B) mRNA-loaded SNP-PEG in HEK293 cells. (C) Gene editing efficiency of RNP-loaded SNP-PEG in GFP-expressing HEK 293 cells. (D) Illustration of HDR at a BFP reporter locus induced by the RNP+ssODN. Sequences of unedited (BFP) and edited (GFP) loci are shown. The protospacer adjacent motif sequence of RNP is underlined and the RNP cleavage site is marked by a red arrow. (E) Gene-correction efficiency of RNP+ssODN co-encapsulated SNP-PEG in BFP-expressing HEK 293 cells. NS: not significant *: p < 0.05; **: p < 0.01; n = 3. (F) Viability of HEK 293 cells treated with DNA-loaded SNP-PEG with different concentrations and DNA-complexed Lipo 2000. NS: not significant; ****: p < 0.0001; n = 7.

The CRISPR-Cas9 RNP is a fast, efficient and accurate genome editor. Cas9 as a nuclease can cause double-stranded DNA break in a specific genomic locus under the guidance of gRNA, achieving gene editing by the non-homologous end-joining (NHEJ) DNA repair pathway [6, 62]. Moreover, with a donor DNA template (e.g., single-stranded oligonucleotide DNA (ssODN)) delivered together with RNP, gene correction or insertion can be achieved through the homology-directed repair (HDR) pathway [46]. We first investigated the genome-editing efficiency of SNP-PEG by delivering the RNP targeting the GFP gene in a transgenic GFP-expressing HEK 293 cell line. As shown in Figure 5C, RNP-encapsulated SNP-PEG exhibited a significantly higher gene-knockout efficiency (1.3-fold) than Lipo 2000 and CRISPRMAX. T7E1 assay was performed to further investigate the indel formation (Figure S6). This assay revealed a mutation frequency of 72 % for cells treated with RNP-encapsulated SNP-PEG, which is higher than the mutation frequency of cells treated with Lipo 2000 (49 %) and CRISPRMAX (56 %). To investigate gene correction capability of SNP, a BFP-expressing HEK 293 cell line was used. Precise gene editing by HDR will lead to the replacement of three nucleotides in the genome, thereby altering one histidine to tyrosine (Figure 5D), which leads to the BFP to GFP conversion [47, 48]. RNP targeting the BFP gene and a donor ssODN were co-encapsulated into SNP-PEG. The genome-editing efficiency was evaluated by the percentage of GFP-positive cells. As shown in Figure 5D, SNP exhibited a statistically higher (1.1-fold) gene-correction efficiency than Lipo 2000. These results demonstrate the capability of SNP as an efficient nanoplatform for genome editor delivery.

We next evaluated the biocompatibility of SNP. HEK 293 cells were treated with DNA-encapsulated SNP-PEG at different SNP-PEG concentrations, and the cell viability was studied by an MTT assay. As shown in Figure 5E, SNP-PEG did not induce significant cytotoxicity in HEK293 cells with concentrations up to 1000 μg/mL, 45-times higher than the working concentration used for our delivery efficiency studies. However, at the working DNA concentration, DNA-complexed Lipo 2000 showed only 77% cell viability, indicating a significantly higher cytotoxicity than SNP-PEG. Together with the delivery efficiency results, we show that the SNP is a desirable nanoplatform for efficient delivery of various biomacromolecules.

Genome Editing in RPE via Subretinal Injection

Nucleic acid delivery/genome editing efficiency of SNP were further investigated in transgenic Ai14 mice (Figure 6). Ai14 mouse genome contains a CAGGS promoter and a LoxP-flanked stop cassette with three repeats of the SV40 polyA sequence that prevents the expression of the downstream tdTomato fluorescent protein gene. The gain-of-function fluorescence can be achieved by: 1) Cre-Lox combination via the delivery of Cre recombinase or Cre-encoding DNA/mRNA (Figure 5A), or 2) excision of 2 of the SV40 polyA blocks by Cas9 RNP (Figure 6C). The tdTomato fluorescence signal in edited cells provides a robust and quantitative readout of nucleic acid delivery/genome editing in Ai14 mice [49, 63, 64].

Figure 6. Nucleic acid and RNP delivery efficiency of SNP in Ai14 mice via subretinal injection.

(A) The tdTomato locus in the Ai14 reporter mouse. tdTomato expression can be achieved by Cre-Lox recombination. (B) Illustration of subretinal injection targeting the RPE tissue. (C) The stop cassette containing 3 Ai14 sgRNA target sites prevents downstream tdTomato expression. Excision of two SV40 polyA blocks by Ai14 RNP results in tdTomato expression. (D) Efficient delivery of Cre-mRNA by SNP-PEG-ATRA in mouse RPE. D1, RPE floret of eyes subretinally injected with Cre-mRNA-encapsulated SNP; D2, 20X magnification images of tdTomato+ RPE tissue; D3, RPE floret of PBS controls. (E) Efficient delivery of RNP by SNP-PEG-ATRA in mouse RPE. E1, RPE floret of mouse eyes subretinally injected with Ai14 RNP-encapsulated SNP; E2, 20X magnification images of tdTomato+ RPE tissue; E3, RPE floret of Ai14 mice injected with negative control SNP-PEG-ATRA (SNP-PEG-ATRA encapsulating RNP with negative control sgRNA). The whole RPE layer was outlined with a white dotted line.

To study the mRNA delivery efficiency by SNP, eyes of Ai14 mice were subretinally injected with a Cre-mRNA-encapsulated SNP-PEG-ATRA (Figure 6B); subretinal injection of PBS was used as a control. Four days post injection, the mice were sacrificed and eyes were enucleated. The eyes were then dissected to remove anterior region leaving the posterior eye cup of an eye. Retina was later removed from the RPE-choroid and flat-mounted, tdTomato expression in the flattened RPE tissue (i.e., RPE floret) was studied by CLSM. As shown in Figure 6D, strong tdTomato fluorescence was visualized in the RPE florets with SNP-PEG-ATRA injection, indicating efficient delivery of Cre-mRNA by SNP. The ratio of tdTomato positive area to the total area of the RPE floret was 6.9 % (Figure S7). Moreover, the genome editing efficiency of SNP was studied by subretinal injection of Cas9 RNP encapsulated SNP. Mice were subretinally injected with a SNP-PEG-ATRA encapsulating the RNP targeting the SV40 polyA block (i.e., Ai14 RNP), or a SNP-PEG-ATRA encapsulating the RNP with the negative control sgRNA (i.e., negative control). The tdTomato expression was evaluated 14 days post-injection. As shown in Figure 6E, Ai14 RNP-loaded SNP induced very strong tdTomato expression in the RPE surrounding the injection site. The ratio of tdTomato positive area to the total area of the RPE floret was calculated as 4.5% (Figure S8). No tdTomato signal was found in eyes injected with negative control SNP. These results suggest that SNP is a reliable nanoplatform for in vivo biomacromolecule delivery.

Genome Editing in Liver via Intravenous Injection

The nucleic acid and RNP delivery efficiency of intravenously injected SNP was also evaluated in vivo using Ai14 mice. Two types of SNPs were involved in this study: (1) SNP-PEG and (2) liver-targeting SNP-PEG-GalNAc. We chose liver as the target organ because liver is an important target for therapeutics development. Nanoplatforms capable of safe and efficiency gene/gene editor delivery to liver can be powerful tools for the treatment of liver diseases (e.g., nonalcoholic fatty liver disease, liver cancer and hereditary tyrosinemia) [65–67].

Cre-mRNA delivery was first investigated, with a mRNA dosage of 20 μg per mouse. Major organs were collected 3 days post injection, and the tdTomato fluorescence was analyzed by IVIS (Figure S9A). Although tdTomato signal was mainly detected in the liver for both non-targeted and targeted SNPs, the SNP-PEG-GalNAc injected mice exhibited a stronger liver tdTomato signal than SNP-PEG (Figure 7A). In fact, homogenized liver tissue showed a 2-fold increase of tdTomato signal in the liver of SNP-PEG-GlaNAc injected mice than the SNP-PEG group (Figure 7B), indicating GalNAc conjugation on the SNP surface can further enhance liver targeting efficiency. To confirm the tdTomato expression, liver sections were immunofluorescence stained with anti-tdTomato antibody and then fluorescein-tagged secondary antibody. The immunostained liver sections were examined using CLSM. As shown in Figure 7C and Figure S9B, tdTomato-positive cells were found in liver tissue, while tdTomato positive cells were not detected in the PBS-injected mice, indicating that SNPs, with or without GalNac, can deliver mRNA into liver via systemic administration.

Figure 7. SNP enabled nucleic acid (A-C) and RNP (D-F) delivery in vivo via systemic administration.

Ai14 mice were injected with Cre-mRNA or RNP encapsulated SNP-PEG or SNP-PEG-GalNAc, and the tdTomato fluorescence of the whole liver (A and D) and homogenized liver tissue (B and E) was detected and analyzed by IVIS imaging. Liver sections were studied by immunofluorescence staining and confocal laser scanning microscopy (C and F). *: p < 0.05; ***: p < 0.001; n = 3.

We further investigated the in vivo systemic RNP delivery efficiency using SNP. Adult Ai14 mice were retro-orbitally injected with RNP encapsulated SNP or SNP-PEG-GalNAc (100 μg RNP per mouse), and major organs were collected 7 days post-injection. Similar to Cre mRNA, tdTomato signal were mainly found in the liver (Figure 7D and Figure S10A), and SNP-PEG-GalNAc showed a 2-fold higher gene editing efficiency than SNP-PEG, as quantified by the fluorescence intensity of homogenized tissue (Figure 6E). Immunofluorescence staining of sectioned liver showed strong tdTomato expression induced by RNP delivery (Figure 7F and Figure S10B). To evaluate the potential systemic toxicity of SNP, blood biochemistry test was performed for all the injected mice (Figure S11). The key elements of the blood biochemical profile (e.g., total CO₂ (tCO₂), ALT, AST, BUN, etc.) showed no significant difference between SNP-injected groups and the PBS control group, indicating that the SNP possessed good biocompatibility. This proof-of-principle data indicates that intravenous administration of SNP can achieve gene delivery/gene editing in vivo. Furthermore, SNP conjugated with targeting moieties can further enhance the biomolecule delivery efficiency in targeted tissues/cells.

DISCUSSION

Intracellular delivery of nucleic acids (e.g., DNA, mRNA) and genome editors (e.g., RNP, RNP+ssODN) plays an important role in revolutionizing the treatment of genetic diseases. Although a number of non-viral vectors have been developed recently, safe and efficient delivery of these payloads in their active forms to the target site remains a challenge, The key properties of non-viral vectors including in vitro and in vivo stabilities, biocompatibility, payload loading content and encapsulation efficiency, payload types, payload release profile, and surface characteristics (e.g., zeta potential and ligand conjugation) are largely determined by the composition (formulation) and fabrication process of the non-viral vectors. Lipid- or polymer-based self-assembled nanocarriers with stimuli-responsive capabilities have been previously reported [20, 68, 69]; however, such self-assembled nanosystems often lack of sufficient in vivo and in vitro stabilities because the interactions between the carrier materials (i.e., lipids or polymers) and payloads are merely physical interactions (e.g., electrostatic interaction) [70, 71]. Moreover, biomacromolecules such as DNA, mRNA and RNP are very different payloads in terms of structure, charge, and functionality, therefore, it is not surprising to see some of the reports suggesting that minor changes in the chemical structure of the lipids or polymers can affect the delivery efficiency of different payloads in very distinctive, or even opposite manners [56, 72], creating potential obstacles for co-delivery of multiple types of biomolecules (e.g., RNP+ssODN for HDR) in vivo. Co-delivery of multiple types of biomolecules was reported using a functionalized gold NP core for payload absorption, followed by a cationic polymer surface coating [73]. However, this type of non-viral vector still encounters potential instability issue, the resulting ~500-nm nanoparticle with positive surface charges is not desirable for systemic administration. To address these challenges, we have developed a versatile strategy to encapsulate various types of biomacromolecules (e.g., DNA, mRNA, CRISPR RNP, and RNP+ssODN) in a GSH-responsive silica nanoparticle (SNP) via a microemulsion process. We demonstrated that a silica-based network is constructed around the payload molecule in the nano-sized water droplet, forming a covalently crosslinked, yet GSH-responsive SNP entrapping the payload inside. It is noted that our GSH-responsive SNP is distinctively different from previously reported mesoporous silica NP (MSN) systems [74, 75]. Macromolecular payloads were loaded onto the MSN surface via electrostatic interaction only, leading to poor NP stability (aggregation) and low payload loading content (<5 wt%) [76, 77]. Our GSH-responsive SNP possessed excellent stability due to its covalent nature. The SNP also offered a high payload loading content (>9 wt%), and high encapsulation efficiency (>90 %), regardless of payload type. It is expected that the SNP can also facilitate the delivery of other CRISPR variant payloads, such as epigenomic editors, RNA editors and base editors [78].

The size and surface chemistry of the nanocarrier can directly affect its circulation time, iv vivo biodistribution, and internalization by target cells [79]. For example, A study found that 60 nm PEGylated NPs exhibited the highest tumor accumulation among the various NPs studied with a particle size ranging from 10–200 nm [80, 81]. It is also known that PEGylated NPs exhibit a longer circulation time and better mucus-penetration efficiency [82, 83]. The PEGylated SNP used for the studies presented in this paper had a hydrodynamic diameter around 50 nm that allows for both local and systemic delivery. However, the simplicity of the SNP formulation (i.e., only silane reagents involved) and the microemulsion fabrication process allow us to conveniently change the size of the SNP in the range of 40 to 200 nm by varying the emulsification conditions (e.g., water to oil ratio and type/ratio of surfactant) [34, 73], making it desirable for various target tissues/cells in vivo.

Another advantage of the SNP is its versatile chemistry. In this study, different functional moieties were integrated into the SNP. Imidazole moieties were incorporated into the SNP to enhance the endo/lysosomal escape capability. Surface modification of the pristine negatively charged SNP with APTES (for positive charge), and PEG (for neutral surface charge) both increased the transfection efficiency of DNA by 1.6-fold in vitro. Moreover, various targeting ligands can be conveniently conjugated onto the surface of the SNP. In vivo studies in adult mice demonstrated the successful delivery of nucleic acid and RNP in the RPE (mediated by SNP-PEG-ATRA via subretinal injection), and liver (mediated by SNP-PEG-GalNAc via intravenous injection) with good biocompatibility, indicating that SNP is a highly desirable nanoplatform for a variety of biomedical applications. Other than functional moieties and surface modifications, the stimuli-responsive linker in the current SNP, i.e., GSH-responsive disulfide linker, can also be conveniently replaced by other types of stimuli-responsive linkers (e.g., hypoxia- or reactive oxygen species-responsive) according to the unique features of the target cells [84, 85], making it suitable for targeted nucleic acid delivery/genome editing in various diseases.