Autonomous Feedback-Driven Engineered DNAzyme-Coated Trojan Horse-like Nanocapsules for On-Demand CRISPR/Cas9 Delivery

Abstract

Manipulating the expression of cellular genes through efficient CRISPR/Cas9 delivery is rapidly evolving into a desirable tumor therapeutics. The exposure of CRISPR/Cas9 to a complex external environment poses challenges for conventional delivery carriers in achieving responsive and accurate release. Here, we report a Trojan horse-like nanocapsule for the on-demand delivery of CRISPR/Cas9 in a microRNA-responsive manner, enabling precise tumor therapy. The nanocapsule comprises a nanoassembled, engineered DNAzyme shell encasing a Cas9/sgRNA complex core. The DNAzyme, functioning as a catalytic unit, undergoes a conformational change in the presence of tumor-associated microRNA, followed by activating a positive feedback-driven autonomous catabolic cycle of the nanocapsule shell. This catabolic cycle is accomplished through chain reactions of DNAzyme “cleavage–hybridization–cleavage”, which ensures sensitivity in microRNA recognition and effective release of Cas9/sgRNA. Utilizing this Trojan horse-like nanocapsule, as low as 1.7 pM microRNA-21 can trigger the on-demand release of Cas9/sgRNA, enabling the specific editing of the protumorigenic microRNA coding gene. The resulting upregulation of tumor suppressor genes induces apoptosis in tumor cells, leading to significant inhibition of tumor growth by up to 75.94%. The Trojan horse-like nanocapsule, with superior programmability and biocompatibility, is anticipated to serve as a promising carrier for tailoring responsive gene editing systems, achieving enhanced antitumor specificity and efficacy.

Introduction

The discovery of the clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) system has led to revolutionary advances in gene editing.¹⁻³ This system operates by introducing an engineered single guide RNA (sgRNA) to efficiently edit any target gene through either gene knockout (via insertion/deletion) or knock-in (via homology-directed repair).⁴ Despite the high efficiency of the Cas9/sgRNA system for gene editing, off-target effects caused by unpredictable cleavage events mediated by Cas9, potentially resulting in unforeseen side effects, cannot be overlooked.⁵ Another challenge is the effect delivery of this system into cells in order to cleave the specific gene sites.^6,7 Therefore, developing on-demand CRISPR/Cas9 delivery methods is crucial for efficient and specific gene editing therapeutics.

Among various delivery vehicles including liposomes,⁸ polyethylenimine,⁹ and nanoparticles,¹⁰⁻¹² DNA-based carriers have shown tremendous promise for Cas9/sgRNA delivery due to their excellent biocompatibility and high loading capability.¹³ Innovations such as Cas9/sgRNA complex-filled spherical nucleic acid nanogels and self-assembled DNA nanoclews have been developed to achieve efficient gene editing.^14,15 Although a DNA-based carrier can transport the Cas9/sgRNA complex in an efficient and nontoxic manner, the challenge of achieving on-demand Cas9/sgRNA release still remains. For tumor treatment, the stimuli-responsive delivery of CRISPR/Cas9 is expected to maximize the gene editing efficacy and minimize side effects. Yang et al. constructed an ATP-responsive Cas9/sgRNA release platform based on a zeolitic imidazole framework-90 for targeted gene editing.¹⁶ Wang et al. developed a thermo-triggered vehicle for the controllable release of Cas9-sgPlk-1 plasmids for efficient cancer therapy.¹⁷ Additionally, DNA sequences in DNA-based carriers have been programmed to respond to specific stimuli for controllable doxorubicin release.¹⁸ In this regard, programming the self-assembly of a DNA-based delivery system, particularly in response to endogenous stimuli, is highly desired for on-demand CRISPR/Cas9 delivery.

Considering its specific recognition ability and high catalytic activity, functionalized DNA represents an ideal building block for a DNA-based carrier.¹⁹⁻²¹ Among the multiple functionalized DNA, DNAzyme, a single-stranded nucleic acid, has been applied as a recognition element for endogenous stimuli due to its excellent thermal stability, ease of modification, and high specificity.²²⁻²⁴ To achieve stimuli recognition, an allosteric DNAzyme can be designed by introducing a stimuli inhibitor to alter its conformation and inhibit DNAzyme activity. In response to stimuli, the allosteric DNAzyme can be activated and cleave the substrate.²⁵⁻²⁸ To enhance the stimuli-response efficiency, self-driven cycle cleavage of allosteric DNAzyme can be achieved by designing the substrate as an analogue of the stimuli. Furthermore, a core–shell nanocapsule structure coated with layer-by-layer (LbL) assembled allosteric DNAzyme can be constructed to deliver Cas9/sgRNA. The layered nature of the LbL core–shell nanocapsules allows for the inclusion of multiple allosteric DNAzyme reaction units within a single film.²⁹⁻³¹ Through sequence programming, the allosteric DNAzyme in the nanocapsule shell can respond to endogenous stimuli and initiate autonomous cleavage to execute the self-driven catabolic process without introducing external fuel DNA strands or other enzymes.³²⁻³⁴ The nanocapsule-coated engineered self-driven DNAzyme can serve as a “Trojan horse” for the efficient delivery and on-demand release of Cas9/sgRNA. Consistent with the stimuli-dependent catabolic degree, the camouflage of the “Trojan horse” can be removed, releasing the encapsulated Cas9/sgRNA, or “soldiers”, to disrupt the targeted gene sequences for accurate and enhanced gene editing.^35,36

Herein, we have developed a Trojan horse-like nanocapsule with an engineered self-driven DNAzyme coating for on-demand Cas9/sgRNA delivery. This nanocapsule consists of an LbL assembly of allosteric DNAzyme as the shell and a Cas9/sgRNA complex as the core. The shell is composed of oligonucleotide layers cross-linked by substrate P1, DNAzyme P2, and backbone P3. The Cas9/sgRNA complex is encapsulated within the hollow space of the nanocapsules (Figure 1). Moreover, microRNA (miRNA) can function as a “switch” that triggers the self-driven catabolic process of the Trojan horse-like nanocapsule. MiRNA has been shown to play an essential role in regulating tumor-related genes.³⁷ MicroRNA-21 (miR-21) is overexpressed in human hepatocellular carcinoma and can inhibit the expression of tumor suppressor genes.³⁸ Therefore, for our Trojan horse scheme, miR-21 was selected as the intracellular stimuli model in this study. As the bridged DNA, the active region of DNAzyme P2 was inhibited by the inhibitor P4, which included the recognition sequence of miR-21. After the nanocapsule was taken up by the tumor cells, with the Mg²⁺ as a cofactor of DNAzyme, trace amounts of miR-21 specifically hybridized with P4 via base pairing and triggered allosteric DNAzyme reactivation. The activated DNAzyme, serving as the reaction unit, cleaved the substrate and released substrate fragments that can hybridize with P4, continuously triggering other reaction units. The cycle cleavage effects of the self-driven DNAzyme accelerated the nanocapsule catabolic process and Cas9/sgRNA release. The released Cas9/sgRNA complex can then be transported into the nucleus for subsequent gene editing and gene expression regulation.³⁹ Consequently, by recognizing miR-21 and silencing the MIR-21 gene that generates the pre-miRNA,⁴⁰ the miR-21-responsive DNAzyme-functionalized nanocapsule (R-DN) facilitated the upregulation of miR-21-induced downregulated tumor suppressor genes, ultimately inhibiting the proliferation and invasion of tumor cells.

Figure 1.

Schematic illustration of the self-assembly process and antitumor gene therapeutic process of miR-21-responsive DNAzyme-functionalized nanocapsule (R-DN). (A) Sequence design of R-DN shell and formation of R-DN composed of layer-by-layer assembled DNAzyme and Cas9/sgRNA complex. (B) Delivery of Cas9/sgRNA by the R-DN for gene editing and tumor therapy. (I) Endocytosis. (II) Recognition of miR-21. (III) Catabolic process. (IV) MiR-21-responsive Cas9/sgRNA release. (V) Gene editing.

Results

Synthesis and Characterization of R-DN

The R-DN construction process is shown in Figure 1A. For the Cas9/sgRNA complex, the Cas9 protein with a molecular weight of ∼160 kDa (Figure S1) was prebound with synthetic sgRNA to form a complex. Polyacrylamide gel electrophoresis (PAGE) analysis verified that the Cas9/sgRNA complex effectively cleaved the target MIR-21 gene (Figure S2). The R-DN template was synthesized using CaCO₃ particles precipitated from the solution of CaCl₂ and Na₂CO₃ (Figure S3). The nanocapsule shell skeleton comprised four nucleic acid chains, including substrate P1, DNAzyme P2, backbone P3, and inhibitor P4. To encapsulate the Cas9/sgRNA complex, template CaCO₃ particles were coprecipitated with Cas9/sgRNA, following a previously reported protocol.⁴¹ After coprecipitation, the particles were coated with the first layer of a positively charged polyelectrolyte, poly(allylamine hydrochloride) (PAH, average MW 50 kDa). Furthermore, nucleic acid hybrids P1/P2 as the first layer were immobilized onto PAH-adsorbed particles via electrostatic interactions. And the first layer did not include inhibitor P4 of the DNAzyme. Hybrids P3/P2/P4 were employed as the second layer of the particles via the LbL self-assembly. Similarly, hybrids P1/P2/P4 were coated on the outer layer of the particles. As a bridge chain, DNAzyme P2 cross-linked three layers through complementary base pairing. Finally, the resulting DNA-coated CaCO₃ particles were etched with ethylenediaminetetraacetic acid (EDTA) for CaCO₃ core dissolution and R-DN synthesis. The loading capacity of the Cas9/sgRNA complex in the synthetic R-DN was calculated to be 8 pg per capsule.

After synthesis, R-DN was first characterized by using scanning electron microscopy (SEM; Figure 2A). The shape of the prepared uncoated CaCO₃ core was almost spherical (part I image). The surface of the particles was relatively rough and composed of multitudinous carbonate nanoparticles with a specific morphology. Compared with the uncoated CaCO₃ core, the surface of the DNAzyme-coated CaCO₃ core (part II image) was extremely rough, similar to brush hair, suggesting the successful self-assembly of DNAzyme layers. The nanocapsule generated via CaCO₃ core dissolution is shown in part III image. After CaCO₃ core dissolution, a collapsed sphere was observed, indicating that the nanocapsule was hollow and that the core was completely removed, leaving only the Cas9/sgRNA complex. The collapse was caused by the SEM vacuum conditions; however, the capsule remained completely spherical in water (Figure S4).

Figure 2.

Characterization of the synthesis process of R-DN. (A) SEM images corresponding to (I) uncoated CaCO₃ particles, (II) DNAzyme-coated CaCO₃ particles before and (III) after EDTA treatment. All scale bars correspond to 1 μm. (B) Elemental mappings of P(I), Ca(II), C(III), and O(IV) and overlap of the four elements (V) corresponding to CaCO₃ core. Scale bar: 1 μm. (C) Elemental mappings of P(I), Ca(II), C(III), and O(IV) and overlap of the four elements (V) corresponding to the DNAzyme-coated CaCO₃ core. Scale bar: 1 μm. (D) Elemental mappings of P(I), Ca(II), C(III), O(IV) and overlap of the four elements (V) corresponding to R-DN in the final morphology. Scale bar: 1 μm. (E) Particle diameter distributions of R-DN performed by nanoparticle tracking analysis (NTA). (F) Measuring the ζ-potential during the R-DN assembly process. Bars represent the mean ± SD (n = 3). (G) Confocal fluorescence microscopy images of (I) DNAzyme-coated CaCO₃ particles before and (II) after EDTA treatment. Scale bar: 5 μm.

The R-DN synthesis process was further characterized by elemental mapping. Figure 2B depicts the multicolor patterns of the elemental mapping corresponding to the CaCO₃/Cas9/sgRNA core. After the core was modified with the DNAzyme shell, the multicolor patterns of elemental mapping showed more distinct phosphorus (part I of Figure 2C), followed by calcium (part II), carbon (part III), and oxygen (part IV). The final disappearance of calcium indicated removal of the CaCO₃ core and synthesis of R-DN (Figure 2D). Moreover, the particle diameter and ζ-potential of R-DN were evaluated using a laser particle size analyzer. The results of the nanoparticle tracking analysis (NTA) demonstrated a homogeneous size distribution of R-DN with an average diameter of 704.1 nm (Figure 2E). The determination of the ζ-potential at each synthetic step indicated that the negatively charged CaCO₃/Cas9/sgRNA core (−12.6 ± 3.71 mV) underwent a positive charge conversion after PAH coating (+8.7 ± 4.25 mV) and the ζ-potential reversed to a negative value upon the self-assembly of DNAzyme (−21.4 ± 3.21 mV; Figure 2F). Subsequently, the potential remained negative after CaCO₃ core dissolution (−22.4 ± 5.82 mV), which substantiated the successful synthesis of the R-DN.

In addition, the DNAzyme shell of the established nanocapsules was modified with Cy3 for the visualization of its morphology using confocal laser scanning microscopy (CLSM) imaging (Figure 2G). Before EDTA treatment, the Cy3-labeled shell showed high fluorescence, plainly displaying the round shape of R-DN with a diameter of approximately 3.5 μm (part I image). The same shape can be seen in Figure S5. Furthermore, the fluorescence image of the EDTA-treated nanocapsules, exhibiting a dark interior surrounded by a bright shell, can be seen in part II image. It was noteworthy that the removal of the CaCO₃ core by reaction with EDTA for a short period of time resulted in the shrinkage of the nanocapsule to a size of approximately 2.8 μm; however, the overall circular structure was intact, consistent with the trend of the CaCO₃ core dissolution. The further dissolution of the CaCO₃ core can lead to the shrinking of the nanocapsules to smaller size and the final formation of the R-DN.

In Vitro miR-21-Responsive Release of Cas9/sgRNA from R-DN

Figure 3A illustrates the catabolic process of the nanocapsule and the responsive release of the Cas9/sgRNA complex based on DNAzyme self-driven cleavage. Briefly, the addition of miR-21 activated the active region of the DNAzyme by forming the miR-21–P4 complex and initiated self-circulation cleavage, followed by releasing the encapsulated Cas9/sgRNA complex. Next, the fluorescence signal generated by the released Cas9/sgRNA complex cleaving the fluorescent probe was detected to assess the performance of the on-demand release of the R-DN.

Figure 3.

Performance of miR-21-responsive release of Cas9/sgRNA from the R-DN. (A) Schematic miR-21-induced release of Cas9/sgRNA from the R-DN. (B) The gel analysis of the R-DN self-assembly order. Lane 1: P1 hybridized with P2 at 37 °C for 1 h and then reacted with P4 at 37 °C for 1 h without Mg²⁺ and miR-21; lane 2: lane 1 after adding 20 mM Mg²⁺; lane 3: lane 2 after adding 1 μM miR-21. (C) The fluorescence curves of the on-demand release of Cas9/sgRNA from R-DN with different experimental conditions: without R-DN (a), with only double-stranded fluorescent probe (b), without Mg²⁺ (c), without miR-21 (d), and after the addition of R-DN, Mg²⁺, and miR-21 (100 nM) (e). (D) Fluorescence spectra generated by the released Cas9/sgRNA complex cleaving fluorescent probe observed upon the incubation of the Cas9/sgRNA-loaded R-DN with different concentrations of miR-21 corresponding to (a) 0, (b) 10, (c) 100 pM, (d) 1, (e) 10, and (f) 100 nM. (E) The histogram of peak fluorescence intensity (FI), and (F) the corresponding calibration curve for the peak FI vs lg C/pM. Bars represent the mean ± SD (n = 5). (G) Sequence selectivity of the R-DN for the miR-21 response. Comparison of the fluorescence changes upon the released Cas9/sgRNA from nanocapsules in the presence of different sequences at the same concentration of 1 μM. The mismatched bases were labeled in a red color. The nonresponsive DNAzyme-functionalized nanocapsule (nR-DN) group had the inhibitor P4 replaced with a scramble sequence and exhibited no response to miR-21. Bars represent the mean ± SD (n = 3), ****P < 0.0001 compared with blank.

The hybridization temperature and order of P1, P2, and P4 were determined using PAGE (Figures S6 and S7; P4 included the miR-21 recognition sequence). The results showed that P1 hybridized with P2 before hybridization with P4, which could achieve better inhibition and cleavage effects for the DNAzyme (Figure 3B). The Mg²⁺ concentration, which most likely affects the cleavage efficacy, was optimized. The substrate cleavage capability of the nanocapsule shell in the presence of Mg²⁺ was analyzed by PAGE. Substrates P1 and DNAzyme P2 were reacted with different concentrations of Mg²⁺ at 37 °C for 2 h. The results indicated that substrate P1 or DNAzyme P2 was excessive; the cleavage effect was the best with 20 mM Mg²⁺ (Figure S8). Therefore, 20 mM was selected as the ideal concentration of Mg²⁺ for miR-21-responsive fluorescence intensity detection.

Next, the feasibility of the miR-21-responsive load release of R-DN was verified by using confocal fluorescence microscopy (Figure S9). The part I image shows R-DN before the addition of Mg²⁺ and miR-21. No obvious change was observed in the fluorescence intensity of the nanocapsule shell after the addition of 20 mM Mg²⁺ alone (part II). However, a significant decrease in the fluorescence intensity of the nanocapsule shell was observed (part III) after the addition of 10 μM miR-21, suggesting that the shell underwent DNAzyme self-driven cleavage. These results demonstrate that the nanocapsules responded to the miR-21 stimulation.

Fluorescence spectrophotometry was used to further confirm the on-demand release of Cas9/sgRNA from the nanocapsules (Figure 3C). The solution with only the double-stranded fluorescent probe exhibited a very low fluorescence intensity (curve b). The fluorescence intensity was also low without nanocapsules in the reaction solution (curve a). A negligible increase in fluorescence intensity was observed without Mg²⁺ and miR-21 (curves c and d, respectively), indicating the weak nonspecific effects of DNAzyme. However, a sharp increase in fluorescence intensity was observed in the presence of the nanocapsules Mg²⁺ and miR-21, suggesting the initiation of DNAzyme reaction units and the successful release of Cas9/sgRNA (curve e).

To determine the sensitivity of the on-demand release of the nanocapsules, various concentrations of miR-21 were added to the reaction solution containing an equal amount of Cas9/sgRNA-loaded nanocapsules. The resulting solution was centrifuged to separate the unreacted nanocapsules. The supernatant was incubated with a double-stranded fluorescent probe for 40 min, and the fluorescence intensity of the solution was examined. The fluorescence spectra of the nanocapsules gradually increased as the miR-21 concentration increased (Figure 3D). A good linear relationship was observed between the peak intensity of each response and the logarithm of miR-21 concentration within a range of 10 pM to 100 nM (Figure 3E), with a correlation coefficient (R²) of 0.9944. The linear fitting equation is as follows: FI = 53.6 × lg C + 62.76 (Figure 3F), where FI and C represent the peak fluorescence intensity and concentration of the target, respectively. The limit of detection of this system in response to miR-21 was calculated as 1.7 pM based on the 3σ/k rule (where σ is the standard deviation of the blank signals and k is the slope of the obtained linear fitting equation). Thus, the nanocapsules responded sensitively to the trace miR-21 based on the DNAzyme self-driven cleavage strategy.

The specificity of R-DN in response to miR-21 was an important factor, as it indicated its ability to recognize only the molecule of interest. Consequently, several groups of similar molecules, including miR-27, miR-134, miR-128, two base-mismatched targets (TMTs), and single base-mismatched targets (SMTs), were used to analyze the selectivity of the nanocapsules. Additionally, the nonresponsive DNAzyme-functionalized nanocapsule (nR-DN) was used to assess the R-DN specificity. The nR-DN group had inhibitor P4 replaced with a scrambled sequence and exhibited no response to miR-21. The noncomplementary sequences of miR-27, miR-134, and miR-128 signaled a fluorescence intensity that was indistinguishable from that of the blank (Figure 3G). In addition, although the fluorescence intensities of SMT and TMT were slightly higher than those from the blank, they remained significantly lower than the intensity from the addition of miR-21. Therefore, R-DN exhibited excellent specificity for target recognition and was efficient for precise response to target cellular molecules. The R-DN synthesized in the same batch was also utilized for miR-21 detection at concentrations of 10, 100 pM, and 1 nM for five consecutive trials. The relative standard deviation (RSD) of these three concentrations was determined to be 1.52, 1.97, and 2.19%, respectively, thereby confirming the excellent reproducibility of R-DN (Figure S10).

Cellular Uptake Analysis and miR-21-Responsive Release of Cas9/sgRNA in HepG2 Cells

To investigate the cellular uptake of the nanocapsules, human hepatocellular carcinoma HepG2 cells were selected as a model. DNAzyme P2 of R-DN was labeled with Cy3 at both the 5′- and 3′-end. The HepG2 cells were incubated with R-DN for various time intervals, and the cellular uptake was visualized using CLSM imaging (Figure 4A). After 1 h of incubation, colocalization between R-DN (red) HepG2 cells (blue) was observed, indicating successful internalization of R-DN. The three-dimensional (3D) imaging of cells incubated for 2 h revealed that the R-DN (red fluorescence) did not translocate into the cell nucleus (blue fluorescence) but instead situated above the nucleus (Figure S11). With prolonged incubation time, an increasing number of nanocapsules colocalized with HepG2 cells and the fluorescence intensity became stronger until it gradually stabilized at 6 h. The ability of the nanocapsules to penetrate the cells was further verified using flow cytometry (FCM). The FCM results revealed that the fluorescence intensity in HepG2 cells gradually increased with increasing incubation time, indicating the highly efficient internalization of R-DN in HepG2 cells (Figure S12). To explore the R-DN internalization mechanism, HepG2 cells were treated with different inhibitors of the endocytosis pathways and incubated with Cy3-labeled R-DN. The average cell count analysis of HepG2 cells with fluorescence showed that the inhibitors, methyl-b-cyclodextrin (MCD) and chlorpromazine (CPZ), induced maximum reduction in the Cy3-labeled-R-DN intake (Figure 4C), suggesting that R-DN was primarily internalized via clathrin-mediated endocytosis and lipid rafts (Figure 4B). The individual cell count profiles of each sample in the groups treated with the different endocytosis inhibitors are shown in Figure S13. The mean fluorescence intensity recorded by FCM confirmed this conclusion (Figure S14). Moreover, the effect of R-DN on cell viability was examined. The Cell Counting Kit-8 (CCK-8) assay demonstrated that a higher cytotoxicity was observed with a higher concentration of R-DN (Figure S15). The cell proliferation reduction caused by the treatment of HepG2 cells with R-DN after 12 h was statistically significant, validating that the nanocapsules conferred enhanced therapeutic effects.

Figure 4.

Cellular internalization pathway and miR-21-responsive release of Cas9/sgRNA from R-DN in HepG2 cells. (A) Confocal laser scanning microscopy (CLSM) images of HepG2 cells incubated with R-DN for 1, 2, 4, 6, or 8 h (blue, Hoechst 33342-stained nucleus; red, Cy3-labeled R-DN). Scale bar: 10 μm. The spatial colocalization of cells and R-DN were statically analyzed by ZEN. (B) Schematic diagram of the endocytosis pathways of R-DN. (C) R-DN intracellular uptake by HepG2 cells in the presence of different endocytosis inhibitors (MCD: methyl-β-cyclodextrin, CPZ: chlorpromazine, GEN: genistein, AMI: amiloride). (D) CLSM images of the entry of Cas9/sgRNA from R-DN into the nucleus of HepG2 cells (blue, Hoechst 33342-stained nucleus; green, fluorescent protein-labeled Cas9/sgRNA; red, Cy3-labeled R-DN). Scale bar: 10 μm. Bars represent the mean ± SD (n = 3). ****P < 0.0001 as compared to the control group.

Furthermore, the release and import process of Cas9/sgRNA from R-DN into the nucleus were evaluated by labeling Cas9/sgRNA with green fluorescent protein (GFP) and DNAzyme P2 of R-DN with Cy3, followed by CLSM imaging. As shown in Figure 4D, for R-DN that responded to high levels of miR-21 in HepG2 cells, Cas9/sgRNA (green) delivered by R-DN gradually detached from Cy3-labeled R-DN (red) and a small portion of Cas9/sgRNA had penetrated into the nucleus (blue) after incubation for 4 h. The subcellular distribution of the carrier R-DN and cargo Cas9/sgRNA was also investigated through colocalization analysis. The internalized R-DN, after 6 h of incubation, exhibited significant dissociation between the Cas9/sgRNA and the R-DN shell in the cytoplasm, accompanied by enhanced translocation of Cas9/sgRNA into the nucleus. After 8 h of incubation, the majority of red and green fluorescent spots exhibited detached subcellular localization, while the nucleus displayed robust green fluorescence, indicating the effective dissociation of a substantial amount of Cas9/sgRNA from the carrier and its functional localization within the nucleus.

On-Demand Gene Editing of R-DN across Different Cell Lines

The miR-21 responsiveness of gene editing was further verified and compared across various cell lines exhibiting diverse levels of miR-21 expression, including LO2 cells (human normal hepatocytes) with minimal miR-21 expression profiles, MDA-MB-231 cells (human breast carcinoma cell lines) with low miR-21 expression profiles, and HeLa cells (human cervical cancer cell lines) with moderate miR-21 expression profiles (Figure 5G). To explore the optimal uptake duration of R-DN, three types of cell lines were incubated with Cy3-labeled R-DN for different periods, followed by CLSM imaging. As depicted in Figure 5A, a gradual increase in the mean fluorescence intensity was observed in LO2 cells with a prolonged incubation duration of up to 6 h. No significant difference in the mean fluorescence intensity was noted between the incubation times of 6 and 8 h (Figure 5B). Similarly, an increasing number of R-DN colocalized with MDA-MB-231 cells and the mean fluorescence intensity became stronger until it gradually stabilized at 6 h (Figure 5C,D). The mean fluorescence intensity also reached a plateau at 6 h for HeLa cells, surpassing that of LO2 and MDA-MB-231 cells, suggesting that R-DN was more easily internalized by HeLa cells with miR-21 overexpression than by LO2 and MDA-MB-231 cells with low miR-21 expression (Figure 5E,F).

Figure 5.

On-demand gene editing according to the level of miR-21 across different cell lines. (A) CLSM images of LO2 cells incubated with fluorescent dye-labeled R-DN for 1, 2, 4, 6, or 8 h (blue, Hoechst 33342-stained nucleus; red, Cy3-labeled R-DN). Scale bar: 10 μm. (B) Statistics of mean fluorescence intensity in panel (A). (C) CLSM images of fluorescent dye-labeled R-DN internalized into MDA-MB-231 cells with different incubation times. Scale bar: 10 μm. (D) Statistics of the mean fluorescence intensity in panel (C). (E) CLSM images of fluorescent dye-labeled R-DN internalized into HeLa cells with different incubation times. Scale bar: 10 μm. (F) Statistics of the mean fluorescence intensity in panel (E). (G) RT-qPCR analysis of relative miR-21 levels in LO2 cells, MDA-MB-231 cells, HeLa cells, and HepG2 cells before R-DN treatments. (H) Cleavage efficiency of the target gene detected by T7EI assay from different cell lines after R-DN treatments for 72 h. (I) RT-qPCR analysis of relative miR-21 levels in LO2 cells, MDA-MB-231 cells, HeLa cells, and HepG2 cells after R-DN treatments for 72 h. Bars represent the mean ± SD (n = 3). *P < 0.05, **P < 0.01, ***P < 0.001, NS, not significant.

Next, the miR-21-responsive editing efficiency of the target MIR-21 gene in different cells could be readily assessed by the T7EI assay (Figure 5H). The cleavage efficiencies of 32.22% in HeLa cells and 46.48% in HepG2 cells were significantly observed after incubation with R-DN, which aligned with the miR-21 expression profiles. In contrast to high-level genetic cleavage induced by R-DN in both HeLa cells and HepG2 cells, the absence of significant cleavage in R-DN-treated LO2 and MDA-MB-231 cells could be attributed to the low abundance of miR-21 in the LO2 and MDA-MB-231 cells, leading to insufficient unlocking of R-DN and restraining Cas9/sgRNA from being liberated and transported into the nucleus. To demonstrate the efficacy of on-demand gene editing, the expression levels of miR-21 in four cell lines were determined following incubation with R-DN. The expression level of miR-21 was significantly decreased in both HeLa and HepG2 cells, while no significant difference was observed between the two cell lines (Figure 5I). Overall, R-DN exhibits a pronounced cellular uptake and gene editing efficiency in cancer cells characterized by elevated miR-21 expression, demonstrating its excellent potential as a delivery system for Cas9/sgRNA and affirming the validity of the proposed miR-21-responsive mechanism underlying gene editing.

MiR-21-Responsive Gene Editing and Gene Regulation In Vitro

According to the design of the R-DN, miR-21-responsive released Cas9/sgRNA was expected to silence the MIR-21 gene and downregulate miR-21, thus promoting the upregulation of the corresponding tumor suppressor genes (Figure 6A). To validate the effectiveness of R-DN in releasing Cas9/sgRNA into the cytoplasm in a miR-21-responsive manner and to compare this with a widely used sustained release method, we used Lipo 8000 to encapsulate Cas9/sgRNA and tested its MIR-21 gene silencing efficiency. The cultured HepG2 cells were consequently divided into four groups. Each group was separately incubated with phosphate-buffered saline (PBS), nR-DN, liposome, and R-DN for 48 h for the Annexin V-FITC/PI apoptosis assay. As expected, the cellular apoptosis rates of the four groups were 8.12, 12.96, 28.8, and 31.02%, respectively, confirming that treatment with R-DN had an enhancing regulatory effect on tumor suppressor gene expression and increased HepG2 cell apoptosis (Figure 6B). The specific sums of apoptosis percentage in each group of the Q2 and Q3 quadrants of the FCM analysis are shown in Figure 6C. Meanwhile, the expression levels of miR-21 were quantified in HepG2 cells treated separately with phosphate-buffered saline (PBS), nR-DN, liposome, and R-DN using reverse transcription quantitative PCR (RT-qPCR) in the same experimental batch. The miR-21 levels in HepG2 cells in the liposome and R-DN groups also decreased correspondingly, consistent with the cellular apoptosis results (Figure 6D). However, miR-21 levels in HepG2 cells in the nR-DN group were almost unchanged, suggesting that the control nanocapsules did not respond to miR-21 and thus, as expected, failed to release the encapsulated Cas9/sgRNA. These data demonstrated that R-DN achieved the miR-21-responsive controlled release of Cas9/sgRNA with high-quality therapeutic performance. To further verify the Cas9/sgRNA cleavage of the target MIR-21 gene without significant off-target effect, the mutation frequency of insertions and deletions (indels) was detected using the T7 endonuclease I (T7EI) assay. The assay revealed similar mutation frequencies of 39.64, and 43.26% for HepG2 cells with the established liposome transfection method and R-DN (Figure 6E), respectively, closely paralleling the FCM analysis data. However, no mutations were detected in the other two controls.

Figure 6.

R-DN-induced gene editing and gene regulation in vitro. (A) Schematic illustration of miR-21-mediated gene regulation after gene editing in tumor cells. (B) Cell apoptosis analysis of HepG2 cells treated with PBS, nR-DN, liposome, and R-DN, respectively, analyzed by flow cytometry. (C) The cellular apoptosis rates (Q2 + Q3) of HepG2 cells with treatments of PBS, nR-DN, liposome, and R-DN for 48 h. (D) RT-qPCR analysis of relative miR-21 levels in HepG2 cells with different treatments. (E) T7EI assay for editing MIR-21 gene in HepG2 cells after the treatment of PBS, nR-DN, liposome, and R-DN. (F) RT-qPCR analysis of mRNA transcription of PTEN, PDCD4, and RECK proteins in HepG2 cells with different treatments for 48 h. (G) Western blotting results showing the levels of PTEN, PDCD4, and RECK proteins with different treatments for 72 h. Bars represent the mean ± SD (n = 3), *P < 0.05, **P < 0.01, NS, not significant.

Furthermore, to confirm the effect of MIR-21 gene silencing in HepG2 cells on the upregulation of the tumor suppressor genes (PTEN, PDCD4, and RECK) after downregulation by miR-21,⁴² the mRNA (mRNA) transcription and protein expression of PTEN, PDCD4, and RECK were analyzed using RT-qPCR and Western blot assays, respectively. PTEN mRNA increased by 77.62, and 76.69% after treatment with liposome and R-DN, respectively (Figure 6F). However, only a 19.07% increase in mRNA levels was observed in the control group treated with equivalent amounts of nR-DN, and the difference between the nR-DN and PBS groups was not statistically significant (P = 0.6125). This result further confirmed that nR-DN did not respond to miR-21 and thus failed to release the encapsulated Cas9/sgRNA. In addition, after incubation with liposome and R-DN, PDCD4 mRNA increased by 75.89, and 75.15%, respectively, and RECK mRNA increased by 80.37, and 78.19%, respectively. These data suggest that miR-21-related tumor suppressor genes could be upregulated by MIR-21 gene knockdown. Consistent with the mRNA results, a considerable increase in the expression of PTEN, PDCD4, and RECK proteins was observed only in HepG2 cells treated with liposome and R-DN (Figure 6G). No significant increase in the expression of PTEN, PDCD4, or RECK proteins was observed in any control groups. Overall, these results indicate that, compared to Cas9/sgRNA liposome, the R-DN had better bioresponsiveness and more efficient gene editing capabilities; it was also controllable and required no transfection reagent.

Evaluation of Animal Antitumor Efficacy

The therapeutic effect of R-DN in vivo was investigated by injecting HepG2 cells into the subcutaneous armpits of BALB/C nude mice to establish hepatocellular carcinoma models. The timeline and animal treatment experiments are shown in Figure 7A. When the tumor volume increased to 100 mm³, the mice were randomly divided into four treatment groups (n = 5) and intratumorally injected with PBS, nR-DN, liposome, or R-DN every 3 days for three consecutive times. During the 14 day treatment period, the tumor volume changes in the mice indicated that R-DN had the strongest tumor inhibition effect (Figure 7B,C). On the 14th day, the tumor volumes of the PBS, nR-DN, liposome, and R-DN groups were 7.39-, 6.16-, 2.27-, and 1.87-fold more than the initial volume, respectively. However, the difference in tumor volume between the PBS and nR-DN groups was not statistically significant (P = 0.27), suggesting that nR-DN treatment had no observable antitumor effect. The degree of body weight loss in the R-DN group exhibited a significantly lower magnitude compared to that in the PBS group (P = 0.0378) during the 14 day tumor-bearing experiment, and that was proof for demonstrating the antitumor efficacy of R-DN (Figure S16). The slight reduction in the weight of mice may be attributed to the tumor’s impact on their bodies. After sacrificing the mice, the tumor tissues were collected and weighed. The results indicated that the tumors from R-DN-treated mice were smaller and had the lightest tumor mass (Figure 7D,E). Consistently, R-DN had the highest tumor growth inhibition (TGI) rates (75.94%), whereas the TGI rates of nR-DN and liposome were only 21.41, and 56.94%, respectively, confirming that tumor growth can be significantly inhibited through miR-21-responsive gene therapy (Figure 7F). Finally, cell proliferation and apoptosis in the tumor sections were observed using hematoxylin and eosin (H&E), K_i-67, and terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end labeling (TUNEL) staining. H&E staining revealed that more necrotic regions were present in the tumors from the liposome- and R-DN-treated groups than in the other two groups (Figure 7I). Furthermore, TUNEL analysis revealed that the R-DN-treated group displayed the highest apoptotic ratio (green in the TUNEL images), suggesting that the upregulation of tumor suppressor genes substantially accelerated tumor apoptosis (Figure 7G,J). Moreover, R-DN-treated tumor cells exhibited the fewest brown nuclei in immunohistochemical staining with K_i-67, indicating that R-DN could suppress tumor cell proliferation (Figure 7H,K). In addition, after the 14 day treatment with PBS, nR-DN, liposome, and R-DN, H&E staining of the excised vital organs (heart, liver, spleen, lung, and kidney) in mice did not reveal any physiological pathological abnormalities in any treatment group (Figure S17). Routine animal blood and liver function analyses showed that R-DN exhibited nontoxic effects during the treatment period (Figures S18 and S19). Overall, these results confirmed that the established miR-21-responsive Cas9/sgRNA delivery nanocapsules possessed amplified antitumor efficacy and negligible toxicity and hold significant potential for clinical gene therapy.

Figure 7.

In vivo antitumor efficacy of R-DN packaging gene editing system. (A) Timeline and treatment of PBS, nR-DN, liposome, and R-DN. (B) Tumor growth curves after treatment of PBS, nR-DN, liposome, and R-DN. (C) Therapeutic efficacy of R-DN in hepatocellular carcinoma-bearing mice. (D) Tumor images collected from mice for the four groups on the 14th day (i, PBS; ii, nR-DN; iii, liposome; iv, R-DN). (E) Inhibitory effect of the formulations on tumor weight. (F) Tumor growth inhibition rate of different treatments. (G) The percentage of apoptotic cells quantified based on the TUNEL images (n = 3). (H) The percentage of K_i-67 positive cells quantified based on the K_i-67 staining images (n = 3). (I) H&E staining of tumor tissues collected from the tumor-bearing mice after the end of different treatments. (J) TUNEL assay of tumor tissues. (K) K_i-67 staining of tumor tissues. Scale bar of three staining images represents 100, 50, and 50 μm, respectively. Bars represent the mean ± SD (n = 5), *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, NS, not significant.

To investigate the genetic evidence supporting the antitumor effect of R-DN, Sanger sequencing and T–A cloning were performed to confirm the presence of the indel mutation. The disruption of both liposome and R-DN to the target gene was quantified as 90, and 83%, respectively (Figure S20), which corresponded to their therapeutic efficacy in inhibiting tumor growth. Moreover, we predicted five potential off-target sites (Figure S21) and subsequently examined the mutation status of these sites using the T7EI assay, which revealed no observable cleavage bands (Figure S22). Furthermore, the Sanger sequencing data exhibited intact and single peaks within the cleavage region of predicted off-target sgRNAs (Figure S23). Tracking of indels by decomposition (TIDE) analysis revealed that the five potential off-target sites displayed editing efficiencies below 3.4%, which fell within an acceptable range (Figure S24). The results collectively demonstrate the high biosafety of the R-DN-based Cas9 delivery system and its promising applicability in tumor gene therapy.

Conclusions

In this study, a Trojan horse-like DNAzyme-functionalized nanocapsule was developed as a versatile platform for the delivery of functional protein and tumor therapy. The allosteric DNAzyme of nanocapsule shell functions not only as a biocompatible carrier for the Cas9/sgRNA complex but also as a responder to intracellular trace miR-21 and an engine for the self-driven catabolic process of nanocapsule. The inner Cas9/sgRNA complex was released in a miR-21 concentration-controlled manner. The Trojan horse-like nanocapsule with an engineered self-driven DNAzyme coating can achieve precise gene therapy by identifying different cell types based on their distinct miR-21 abundance. In the future, nanocapsules are promising for enabling more precise identification of multiple stimuli and on-demand release of encapsulated proteins through the utilization of logic gates. Moreover, when combined with interventional therapy, the application of R-DN can be expanded to deep-tissue tumors. In this manner, R-DN is anticipated to become a dependable platform for the development of biomarker-responsive gene editing tools for targeted therapy, based on tumor heterogeneity, thereby advancing precision medicine.

Experimental Section

Materials and Reagents

Calcium chloride dehydrate, sodium carbonate, sodium chloride, ethylenediaminetetraacetic acid disodium salt dihydrate (EDTA), and poly(allylamine hydrochloride) (PAH, average MW 50 kDa) were obtained from Sigma-Aldrich (USA). The Cas9 protein was purchased from Inovogen Biotech. (Chongqing, China). The sgRNA was synthesized from Bio-Lifesci (Guangzhou, China). The 20 bp DNA Ladder (Dye Plus) and DL1,000 DNA Marker were obtained from Takara (Shiga, Japan). Magnesium chloride hexahydrate and RIPA lysis buffer were purchased from Solarbio Science & Technology Co., Ltd. (Beijing, China). 4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES acid) was acquired from Aladdin (Shanghai, China). Cell Counting Kit-8 (CCK-8) was obtained from BioGround Biotech Co., Ltd. (Chongqing, China). Hoechst 33258, Lipo 8000, phenylmethanesulfonyl fluoride, Annexin V-FITC/PI Apoptosis Kit, T7 endonuclease I, and BCA protein quantification kit were obtained from Beyotime Biotechnology (Shanghai, China). PTEN (D4.3) XP Rabbit mAb, PDCD4 (D29C6) XP Rabbit mAb, RECK (D8C7) XP Rabbit mAb, and antibeta actin antibody were acquired from Cell Signaling Technology (Danvers, MA). Dulbecco’s modified Eagle’s medium (DMEM), RPMI 1640 medium, 0.25% trypsin-EDTA (1×, with phenol red), fetal bovine serum (FBS), Opti-MEM medium (1×), and penicillin–streptomycin (penicillin 10,000 U/mL, streptomycin 10 mg/mL) were obtained from GIBCO BRL (Grand Island, NY).

All oligonucleotides in Table S1 were synthesized from Shanghai Sangon Biotech Co., Ltd. (Shanghai, China) and purified with high-performance liquid chromatography (HPLC). All chemicals and solvents used were analytically pure. Ultrapure water with a resistivity of 18.2 MΩ/cm was employed to prepare all solutions.

Characterizations

The size and morphology of the DNAzyme-functionalized nanocapsules were conducted with a Hitachi SU8020 scanning electron microscopy (SEM) (Hitachi Co. Ltd., Japan). Elemental mapping was implemented with an EX-350 (HORIBA, Japan). Nanoparticle tracking analysis (NTA) and ζ-potential were performed using a Zetasizer Nano ZS90 (Malvern Instrument Ltd., UK). Fluorescence spectra were performed with an F-7000 fluorescence spectrophotometer (Hitachi Co. Ltd., Japan). The reverse transcription quantitative polymerase chain reaction (RT-qPCR) analysis was recorded with a CFX96 real-time system (Bio-Rad). Cells were counted using LUNA-II (Logos Biosystems, Korea). Fluorescence images were performed using a ZISS 880 confocal microscope at 63× magnification (Zeiss Co. Ltd., German). The nanocapsules of cellular uptake were recorded with a BD LSRFortessa instrument (BD).

DNA Cleavage Validation of Cas9/sgRNA

The 100 nM Cas9 protein was preincubated with 100 nM sgRNA in reaction buffer (50 mM tris–HCl, pH 7.9, containing 10 and 100 mM Na⁺) at 25 °C for 10 min. Afterward, the mixture was reacted with the synthetic MIR-21 gene fragment (10 nM) at 37 °C for 1 h. Finally, the product was analyzed using 10% native polyacrylamide gel electrophoresis (PAGE) at 150 V for 40 min.

Synthesis of Cas9/sgRNA-Loaded CaCO₃ Core

The template of the miR-21-responsive nanocapsules was composed of CaCO₃ capsules, which were precipitated from an equal amount of supersaturated solution of Na₂CO₃ and CaCl₂ under vigorous stirring at ambient temperature. The Cas9 protein and sgRNA were mixed in a 1:1 molar ratio at 25 °C for 10 min to ensure the complete formation of the Cas9/sgRNA complex. Then, equal volumes (300 μL) of Na₂CO₃ and CaCl₂ solutions (0.33 M) were stirred with the Cas9/sgRNA complex (final concentration of 0.078 mg/mL) using a coprecipitation method in a round-bottom flask affixed to a magnetic stirrer operating at 900 rpm for 110 s. After stirring, the CaCO₃ core wrapped with the Cas9/sgRNA complex was left for 3 min at an ambient temperature to settle down. The obtained CaCO₃ core wrapped with the Cas9/sgRNA complex was centrifuged at 6000 rpm for 5 s and washed with pure water 3 times. The solution was centrifuged at 6000 rpm for 5 s after each washing step. After removal of the supernatant, 500 μL of PAH solution with a concentration of 1 mg/mL (in 10 mM HEPES, pH 7, including 0.5 M Na⁺ and 0.05 M Mg²⁺) was added to the CaCO₃ sediment, followed by placing on the rotary oscillator to vibrate for 20 min with a parameter of MODE F6, speed 70, and frequency 6 to acquire a PAH-coated CaCO₃ core with a positive charge. Afterward, the CaCO₃ core with a positive charge was centrifuged at 3000 rpm for 30 s and washed with 10 mM HEPES (pH 7, including 0.5 M Na⁺) 3 times, followed by resuspending in 10 mM HEPES for the layer-by-layer (LbL) DNAzyme assembly.

Synthesis of LbL DNAzyme-Functionalized Nanocapsule

The P1/P2 and P3/P2 were incubated in 200 μL of buffer (6 μM in 10 mM HEPES, pH 7, including 0.5 M Na⁺) at 37 °C for 1 h. The ratio of P1 to P2 and P3 to P2 was 2 to 1. Three layers of nucleic acids were prepared in total. The second layer and the third layer were additionally incubated with inhibitor P4 in 200 μL of 10 mM HEPES (pH 7, including 0.5 M Na⁺) at 37 °C for 1 h. The obtained PAH-coated CaCO₃ core above was coated by sequential incubation in the DNA hybridization products P1/P2 and P3/P2 solutions on the rotary oscillator to vibrate for 30 min at ambient temperature with a parameter of MODE F5, speed 70, and frequency 7. After adsorption of each layer, the capsules were washed twice with 10 mM HEPES (pH 7, including 0.5 M Na⁺) to remove nonadsorbed DNA. The first layer and the second layer were followed by centrifuging at 2000 rpm for 8 min after washing. And the third layer was followed by centrifuging at 2000 rpm for only 5 min after washing.

After the LbL assembly of the DNAzyme shell, 0.5 M EDTA solution (pH 6) was added to the obtained solution to dissolve the CaCO₃ core at a final concentration of 0.1 M EDTA. The mixture was vibrated mildly for 1 h. After the suspension became clear, the supernatant EDTA solution was removed by slow centrifugation at 1000 rpm for 20 min to avoid aggregation of the DNAzyme capsules. The capsules were washed by 10 mM HEPES (pH 7, including 0.5 M Na⁺) 3 times, followed by centrifuging at 1000 rpm for 20 min. Finally, the synthesized capsules were resuspended in pure water and stored at 4 °C.

MiR-21-Responsive DNAzyme-Functionalized Nanocapsule (R-DN) Unlocking and the Release of the Encapsulated Cas9/sgRNA

The sensitivity and specificity of miR-21 response were measured, employing an R-DN solution with a concentration of 1000 capsules/μL. Various concentrations (1, 10, 100 pM, 1, 10, 100 nM, and 1 μM) of miR-21 (2 μL) were mixed with the R-DN solution (5 μL). After the final volume was adjusted to 20 μL by adding HEPES buffer (10 mM, pH 7, including 0.5 M Na⁺ and 0.02 M Mg²⁺), the entire mixture was incubated at 37 °C for 80 min. The reaction samples were then centrifuged at 1000 rpm for 20 min to precipitate the unreacted capsules. The released Cas9/sgRNA complex in the supernatant was incubated with the target double-stranded DNA (dsDNA) of the sgRNA modified with FAM and BHQ1 at both ends at 37 °C for 40 min. The fluorescence intensity generated by the released Cas9/sgRNA cleaving target dsDNA was recorded by a fluorescence spectrophotometer.

Several similar molecules including miR-27, miR-134, miR-128, two base-mismatched target (TMT), and single base-mismatched target (SMT) were spiked in 10 mM HEPES and reacted with the R-DN solution. Meanwhile, 1 μM miR-21 was spiked in 10 mM HEPES and reacted with the nonresponsive DNAzyme-functionalized nanocapsule (nR-DN) solution. The concentration of nR-DN was 1000 capsules/μL. Following centrifugation and cleavage of dsDNA, the fluorescence intensity of the supernatant was measured by a fluorescence spectrometer.

SEM Imaging

50 μL of each prepared sample solution was dropped onto the silicon wafer and dried overnight. Then, the dry sample was sputtered with gold, and images were photographed with a Hitachi SU8020 SEM at an operation voltage of 3.0 kV.

Cleavage Efficacy of DNAzyme In Vitro

The Mg²⁺-dependent DNAzyme cleavage was examined by 12% PAGE. Briefly, substrate P1 and DNAzyme P2 were reacted at 37 °C for 2 h. Then, different concentrations of Mg²⁺ (0, 2, 4, 8, 16, and 20 mM) were incubated with the DNAzyme mixture at 37 °C for 2 h. Finally, 5 μL of samples were diluted with a 6× loading buffer (Promega) and analyzed by PAGE.

Cell Culture

The HepG2 cells (human hepatocellular carcinoma cells), MDA-MB-231 cells (human breast carcinoma cells), and HeLa cells (human cervical cancer cells) were cultured with DMEM high glucose medium added with 10% (v/v) FBS, 100 μg/mL streptomycin, and 100 U/mL penicillin at 37 °C in a moist atmosphere containing 5% CO₂. The LO2 cells (human normal hepatocytes) were cultured in RPMI 1640 medium supplemented with 10% (v/v) FBS, 100 μg/mL of streptomycin, and 100 U/mL of penicillin at 37 °C in a moist atmosphere containing 5% CO₂.

Cytotoxicity Assay In Vitro

The 100 μL aliquot of HepG2 cell suspension was planked in 96-well plates. Density of cells was 2 × 10⁴ cells per well. Cells were cultured for 24 h to reach a 70% density. The R-DN with two different concentrates was added into the cells for different times (2, 4, 6, 12, and 24 h). Afterward, 10 μL of CCK-8 reagent was added and incubated in the dark CO₂ incubator for 1 h. The absorbance was recorded at 450 nm with a microplate reader (Thermo Fisher Scientific). Cell viability was calculated with the equation: cell viability (%) = (average absorbance of treated cells–average absorbance of culture medium)/(average absorbance of control cells–absorbance of culture medium) ×100%.

Intracellular Distribution of R-DN

The 1 × 10⁵ HepG2 cells were seeded in a confocal dish and incubated for 12 h. Afterward, the R-DN was added into the confocal dish (MatTek) and incubated with cells for 1, 2, 4, 6, and 8 h. After washing with PBS 3 times, the cells were stained with 1 mL of Hoechst 33,342 for 25 min. After washing with PBS again, the cells were imaged with confocal microscopy. The fluorescence data was analyzed by ZEN.

The cellular uptake was also recorded with flow cytometry. After treatment with R-DN for different times (1, 2, 4, 6, and 8 h), the cells were washed with PBS and digested by trypsin-EDTA solution with 0.25% phenol red for 90 s. Then, the cells were centrifuged at 1200 rpm for 3 min, resuspended in PBS, and measured by flow cytometry.

Analysis of Endocytosis Pathway

The DNA shell of R-DN was fluorescently labeled with Cy3 to trace its cellular uptake. After seeding in 6-well plates (∼150,000 cells/well) and culturing for 24 h, the HepG2 cells were washed and afterward pretreated by different endocytosis inhibitors including chlorpromazine (CPZ, 10 μM), amiloride (AMI, 1 mM), methyl-β-cyclodextrin (MCD, 1 mM), and genistein (GEN, 20 μM) for 1 h. The Cy3-labeled R-DN was then added to the cells, and the cells were incubated for another 2 h. Then, cells were centrifuged and resuspended in PBS, and fluorescence signals were measured by flow cytometry.

T7EI Mutation Detection Assay

The 4 × 10⁵ HepG2 cells were planked in 6-well plates and cultured overnight. The next day, the cells were washed and afterward incubated with PBS, nR-DN, liposome, and R-DN in the CO₂ incubator for 48 h. The MiniBEST Universal Genomic DNA Extraction Kit (Takara) was used to extract the genomic DNA of cells. Subsequently, genomic DNA was quantified and diluted to 100 ng/μL. Next, the mix prepared by PrimeSTAR GXL DNA Polymerase (Takara) was analyzed by PCR according to the following thermal cycling procedure: 98 °C for 2 min; then 35 cycles of 98 °C for 10 s, 60 °C for 15 s, 68 °C for 60 s, and by a last extension at 68 °C for 5 min.

The mixture including 5 μL of PCR products (200 ng), 2 μL of 10× T7EI reaction buffer, and 12 μL of nuclease-free water was denatured at 95 °C for 5 min and decreased to 85 °C with a 2 °C/s ramp rate, followed by decreasing to 25 °C with a 0.1 °C/s ramp rate in a thermal cycler. Then, 1 μL of T7EI (10 U) was incubated with the mixture at 37 °C for 30 min. After reaction termination, the digested DNA was examined by 1% agarose electrophoresis at 130 V for 30 min, followed by staining with Super GelRed (US Evebright) and imaging using ChemDoc XRS (Bio-Rad). Finally, the density of bands was quantified by employing ImageJ.

RT-qPCR Analysis of mRNA

After being seeded in 6-well plates (4 × 10⁵ cells/well) and cultured for 24 h, the HepG2 cells were washed and incubated with PBS, nR-DN, liposome, and R-DN for 48 h at 37 °C. The total RNA of the treated cells was extracted with RNAiso Plus (Takara). After measurement of the quantitation of total RNA, the total mRNA was reverse-transcribed to cDNA employing a PrimeScript RT reagent kit (Takara). Subsequently, qPCR was executed according to TB Green Premix Ex TaqII (Takara) with the thermal cycling states: 95 °C for 30 s; then continued 40 cycles of 95 °C for 5 s, 60 °C for 30 s.

Cell Apoptosis Assay

After being seeded in 6-well plates (4 × 10⁵ cells/well) and cultured for 24 h, the HepG2 cells were divided into four groups. Each group was incubated with PBS, nR-DN, liposome, and R-DN (Cas9:8 ng/μL) for 48 h at 37 °C for both miR-21 RT-qPCR and apoptosis assay. The method of miR-21 RT-qPCR was the same as that of mRNA, except that the reverse transcription was finished, employing an miRNA first-strand cDNA synthesis kit (Sangon). As for the apoptosis assay, after being washed with PBS once, the cells were digested by 300 μL of trypsin per well and centrifuged at 1000 g for 5 min. Afterward, the 10⁵ cells were stained by Annexin V-FITC and propidium iodide (PI) for 20 min and examined by flow cytometry.

Western Blot Assay

HepG2 cell suspension was cultured in 6-well plates (4 × 10⁵ cells/well) for 24 h. The next day, the cells were washed and incubated with PBS, nR-DN, liposome, and R-DN for 72 h. The cells were then lysed with RIPA supplemented with 1% phenylmethanesulfonyl fluoride and quantified using a BCA protein assay kit. The cellular lysate was centrifuged at 12,000g for 10 min at 4 °C, and the protein was transferred to a poly(vinylidene fluoride) membrane after separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The blots obtained were blocked with 5% skimmed milk and incubated with the anti-PTEN antibody (1:1000), anti-PDCD4 antibody (1:1000), anti-RECK antibody (1:1000), and antibeta actin antibody (1:1000) overnight at 4 °C. Subsequently, all bolts were incubated with horseradish peroxidase-conjugated secondary antibody (1:2000; Proteintech, China) for 1 h at ambient temperature. Proteins were detected by employing the chemiluminescent system and quantified using ImageJ.

Animal Model

All animal experiments were approved by the Laboratory Animal Welfare and Ethics Committee of the Third Military Medical University (approval number: AMUWEC20225084). Male BALB/C nude mice (4–5 weeks old) were obtained from Hunan SJA Laboratory Animal Co., Ltd. (China) and fed in a sterile environment. The body weight of the mice was 19–22 g. HepG2 tumor models were set up by a subcutaneous armpit injection of 2 × 10⁶ cells in PBS/Matrigel (1/1, v/v).

Antitumor Efficacy in HepG2 Models In Vivo

When the tumors grew to 100 mm³ in volume, the mice were assigned randomly into four groups for the treatment with 100 μL of PBS, nR-DN, liposome, and R-DN (Cas9:16 ng/μL). The mice were intratumorally injected with different preparations every third day 3 times in total. The body weights and the tumor volumes were determined every other day. The tumor volumes (V) were quantified by the following formula: V = L × W²/2, where L and W represent the tumor length and width, respectively. On the 14th day, all of the mice were euthanized. Meanwhile, their tumors were collected, imaged, and weighed, followed by soaking in the 10% formalin for histological examinations by hematoxylin and eosin (H&E), TUNEL, and K_i-67 staining. The major organs including heart, liver, spleen, lung, and kidney were also collected for routine staining with H&E for histological analysis.

Examination of Animal Blood Routine and Blood Biochemistry

The whole blood and plasma of model mice were obtained on the 14th day of different treatments. The whole blood was counted with a BC-6800 (Mindray, China). Blood biochemistry was measured employing an automatic chemical analyzer AU5831 (Beckman).

In Vivo Gene Editing Analysis

For Sanger sequencing and T–A cloning, the genomic DNA from excised HepG2 tumors in the mice treated with liposome and R-DN was extracted using MiniBEST Universal Genomic DNA Extraction Kit (Takara) following the manufacturer’s instructions. Then, the genomic DNA was quantified and diluted to 300 ng/μL. Subsequently, 1 μL of the target DNA was amplified by PCR using PrimeSTAR GXL DNA Polymerase (Takara). The obtained PCR product was sent to Sangon Corp. for sequencing and cloning. The sequencing data was mapped using Chromas. And sequence alignment of T–A cloning was performed with SnapGene Viewer.

In Vivo Off-Target Analysis

The genomic DNA from excised HepG2 tumors in the mice treated with liposome and R-DN was extracted using the MiniBEST Universal Genomic DNA Extraction Kit (Takara). Next, the extracted DNA was diluted to 100 ng/μL and 1 μL of the diluted DNA was amplified by PCR using PrimeSTAR GXL DNA Polymerase (Takara). The obtained PCR product was sent to Sangon Corporation for Sanger sequencing. The off-target editing efficiency of sequencing data was calculated using the online tool TIDE.

Statistics

All data were analyzed using GraphPad Prism 9.0. A two-sided Student’s t test was employed to compare two groups in parametrical data, and one-way analysis of variance (ANOVA) was employed to compare multiple groups. A value of P < 0.05 was considered statistically significant.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsnano.4c04147.

• Experimental details and additional data as noted in the text; such as the nucleic acid sequence; gel electrophoresis analysis; optimization of synthesis conditions; the characterization and reproducibility of R-DN; the flow cytometry analysis on the cellular uptake of R-DN; cell viability, weight changes on HepG2 tumor-bearing mice; H&E assay, hematology and plasma biochemical analysis; Sanger sequencing; T–A cloning; and the off-target site analysis (DOCX)

Author Contributions

^# X.T. and Y.C. contributed equally to this work. X.T. and K.C. designed this study, interpreted the results, and drafted the manuscript. X.T. and Y.C. performed the experiments. X.T. and K.C. analyzed the data. Y.C. constructed the animal model. X.T. prepared figures. B.W., J.W., Y.H., L.F., Y.X., and S.X. provided critical protocols and advice for conducting experiments. D.L., M.C., and K.C. supervised the study.

The authors declare no competing financial interest.