Codelivery of CRISPR-Cas9 and chlorin e6 for spatially controlled tumor-specific gene editing with synergistic drug effects

See allHide authors and affiliations

Science Advances  15 Jul 2020:
Vol. 6, no. 29, eabb4005
DOI: 10.1126/sciadv.abb4005


Controlled release of CRISPR-Cas9 ribonucleoprotein (RNP) and codelivery with other drugs remain a challenge. We demonstrate controlled release of CRISPR-Cas9 RNP and codelivery with antitumor photosensitizer chlorin e6 (Ce6) using near-infrared (NIR)– and reducing agent–responsive nanoparticles in a mouse tumor model. Nitrilotriacetic acid–decorated micelles can bind His-tagged Cas9 RNP. Lysosomal escape of nanoparticles was triggered by NIR-induced reactive oxygen species (ROS) generation by Ce6 in tumor cells. Cytoplasmic release of Cas9/single-guide RNA (sgRNA) was achieved by reduction of disulfide bond. Cas9/sgRNA targeted the antioxidant regulator Nrf2, enhancing tumor cell sensitivity to ROS. Without NIR irradiation, Cas9 was degraded in lysosomes and gene editing failed in normal tissues. The synergistic effects of Ce6 photodynamic therapy and Nrf2 gene editing were confirmed in vivo. Controlled release of CRISPR-Cas9 RNP and codelivery with Ce6 using stimuli-responsive nanoparticles represent a versatile strategy for gene editing with potentially synergistic drug effects.


Antitumor drugs include cytotoxic agents and inhibitors that target oncogenic or tumor-associated genes (1). RNA interference can be used to inhibit gene expression, but the resulting transient gene silencing limits its therapeutic use (2, 3). CRISPR-Cas9 gene editing provides a promising approach to treat diseases by reprogramming or activating specific genes. Antitumor applications of CRISPR-Cas9 that target oncogenic or antioncogenic pathways have been explored extensively (4, 5). Despite rapidly growing preclinical and clinical research, CRISPR-Cas9 technology faces substantial challenges. Uncontrolled chromosomal integration of viral vectors limits its clinical use (6). Delivery of CRISPR-Cas9 using nonviral vectors is an alternative approach (7) but has been hindered by unwanted genetic mutations and immunogenicity due to sustained Cas9 expression (8). Transient direct Cas9 ribonucleoprotein (RNP) delivery has advantages, including reduced off-target mutations, low cytotoxicity, and high gene editing efficiency (911). Several platforms have been developed to encapsulate Cas9/single-guide RNA (sgRNA) complexes into nanoparticles for gene editing, including cationic liposomes, micelles, DNA nanoclews, metal-organic frameworks (MOFs), and gold nanoparticles, with encapsulation achieved via nonspecific electrostatic interactions, base complementary, or MOF packing, respectively (1215). The specific interaction between nitrilotriacetic acid (NTA)–modified liposomes and His-tagged Cas9 RNP is an effective approach to encapsulate Cas9 (16). However, two issues with Cas9 RNP delivery remain: (i) lack of an effective approach to encapsulate Cas9 RNP with small-molecule drugs in a stable nanoparticle and (ii) uncontrolled release of Cas9 RNP from endocytosed nanoparticles (17, 18).

Combined therapy involving codelivery of two or more drugs is a common tumor treatment strategy (1922), but codelivery of Cas9 with small-molecular drugs has not been investigated thoroughly. For controlled release of Cas9, one approach is to induce release using a combination of internal and external stimuli (23). Controlled release of Cas9 using intracellular stimuli has been achieved by reduction of Au─S bonds in the cytoplasm and by dissolution of MOF structures in acidic lysosomal environments (2426). When Cas9 is released in lysosomes, it must escape into the cytoplasm quickly to avoid enzymatic degradation. Several pH-buffering molecules, including imidazole, peptides, and polyethylenimine (PEI), have been introduced into nanoparticles to enhance lysosomal escape of Cas9 (25, 27). However, degradation of Cas9 in normal cells is desired to prevent unwanted gene editing. Thus, spatial control of Cas9 delivery to tumor cells is desirable. For spatial control of delivery and gene editing, an external stimuli-responsive design is an effective approach (28).

Several methods have been developed to control Cas9 activation to reduce off-target gene editing. A DNA sequence encoding Cas9/dead Cas9 (dCas9) RNP was fused with DNA encoding light-responsive protein (29) and an unnatural amino acid–bearing azide (30) to control Cas9/dCas9 activity. However, these approaches involved the delivery of plasmids that encode Cas9, which limit their applications in vivo. An alternative strategy to achieve tunable activation of Cas9 is needed. Direct Cas9 modification using stimuli-responsive polymers or nanoparticles is a promising approach. Cas9 was linked to nanoparticles using a near-infrared (NIR) irradiation–responsive bond to allow spatial control of Cas9 activation (31). However, the PEI-coated nanoparticles were administered via intratumor injection, and no controlled release of Cas9 from lysosomes was attempted. Here, we developed a Cas9/sgRNA delivery system using NIR-responsive and reducing agent–responsive nanoparticles, allowing codelivery with chlorin e6 (Ce6) and controlled release of Cas9/sgRNA in the tumor cell cytoplasm to achieve synergistic antitumor activity (Fig. 1).

Fig. 1 Schematic of the preparation, delivery, and intracellular fate of NIR-sensitive and reducing agent–sensitive nanoparticles containing Nrf2-targeting Cas9/sgRNA and the antitumor photosensitizer Ce6.

Nanoparticles are formed by (i) self-assembly of anionic micelles using nitrilotriacetic acid–disulfanediyldipropionate–polyethyleneglycol–b-polycaprolactone (NTA-SS-PEG-PCL) copolymer encapsulating Ce6, (ii) binding of His-tagged Cas9/sgRNA, and (iii) coating with cationic iRGD-modified copolymer. Nanoparticles bind tumor cells via integrin-iRGD binding and are internalized. Upon NIR irradiation, Ce6 generates reactive oxygen species (ROS), and nanoparticles are released from lysosomes into the cytoplasm, where a disulfide bond is reduced, releasing Cas9/sgRNA. Cas9/sgRNA targets the antioxidant gene Nrf2, enhancing tumor cell sensitivity to ROS synergistically with Ce6. PDT, photodynamic therapy.


The copolymers NTA-disulfanediyldipropionate-polyethyleneglycol-b-polycaprolactone (NTA-SS-PEG-PCL) and iRGD-PEG-b-polyasparte-g-1,4-butanediamine [internalizing RGD-PEG-pAsp(DAB)], referred to here as iRGD-PD, were synthesized, as we have described previously (fig. S1) (32), and were characterized using 1H–nuclear magnetic resonance (NMR), Fourier transform infrared (FTIR) spectroscopy, gel permeation chromatography (GPC), and mass spectroscopy (figs. S2 and S3). The degrees of polymerization for PEG and PCL in the NTA-SS-PEG-PCL copolymer were 27 and 47, and those for PEG and pAsp(DAB) in iRGD-PD were 77 and 70, respectively. The molecular weights of the hydrophobic PCL and hydrophilic NTA-SS-PEG blocks were 5.4 and 1.7 kDa, respectively (fig. S2, B and C, and table S1). The shoulders in the GPC curves of PEG and NTA-SS-PEG-PCL polymers may be attributed to the by-products of the polymerization (fig.S2C), because the molecular weight (~551.5 kDa, calculated with a PEG standard curve) of the by-products was much higher than that of cross-linked NTA-SS-PEG-PCL. The presence of disulfide bonds in NTA-SS-PEG-PCL and precursor molecule tBuO3-NTA-SS-NHS was confirmed by lH-NMR or mass spectroscopy analyses (fig. S2, A and B). The -SS- signal detected in 509 nm of the Raman spectrum was weak, because the number of disulfide bond per NTA-SS-PEG-PCL molecule was only 0.93, as determined by the Ellman’s assay (fig. S2D). The presence of iRGD in iRGD-PEG-pAsp(DAB) was confirmed by 1H-NMR resonance peaks at 1.92 and 4.21 parts per million (fig. S3A). The reducing agent–insensitive copolymer NTA-PEG-PCL was synthesized and characterized (fig. S4).

NTA-SS-PEG-PCL assembles into micellar nanoparticles (NTA-NPs), with PCL blocks forming the micellar core, and can encapsulate hydrophobic photosensitizer Ce6; NTA-terminated PEG block on the surface of the micelle can then bind His-tagged Cas9 RNP. To optimize Cas9/sgRNA loading, we prepared NTA-NPs loaded with Ce6 and Cas9/sgRNA (CC-NPs) at various mass ratios of NTA-Ce6-NPs to Cas9 (NTA/Cas9 ratio), and SDS–polyacrylamide gel electrophoresis (PAGE) was performed to determine the Cas9 content in the nanoparticles (Fig. 2A). Loading efficiency was defined as the mass of Cas9 or Ce6 in the nanoparticles divided by the mass in the feed; loading content was defined as the mass of Cas9 or Ce6 in the nanoparticles divided by the mass of the nanoparticles. At an NTA/Cas9 ratio of 8, the Cas9 loading efficiency was 91.2%, and the loading content was 9.9% (Fig. 2B). At this NTA/Cas9 ratio, the Cas9 RNP–binding capacity of the NTA-NPs reached saturation, and further increase of this ratio reduced loading content and slightly increased loading efficiency. At an NTA/Cas9 ratio of 8, the Ce6 loading efficiency and loading content were 82.4 and 8.6% (table S1), respectively. NTA-NPs alone or loaded with Ce6 (NTA-Ce6-NPs) showed a negative surface charge of −13 mV because of the negatively charged NTA carboxyl groups. CC-NPs at NTA/Cas9 ratios of 8 to 10 showed a more negative surface charge (−26 mV) than NTA-Ce6-NPs (−13 mV) (Fig. 2C) because of the negative charge of the Cas9/sgRNA complexes. The size of CC-NPs at NTA/Cas9 ratios of 2 to 14 was relatively constant (84 to 92 nm) (Fig. 2C). Thus, the NTA/Cas9 ratio of 8 was selected for preparation of CC-NPs in subsequent studies.

Fig. 2 Characterization, tumor cell uptake, and in vivo biodistribution of nanoparticles.

(A) SDS-PAGE of Cas9- and Ce6-containing nanoparticles (CC-NPs) prepared at different NTA/Cas9 ratios. (B) Cas9 loading efficiency (LE) and loading content (LC) in CC-NPs. (C) Size and zeta potential of CC-NPs. (D) Size and zeta potential of iRGD-PD–coated CC-NPs (T-CC-NPs) using different amounts of iRGD-PD. Data are shown as means ± SD (n = 3). (E) Transmission electronic microscopy (TEM) images of NTA-Ce6-NPs (self-assembled from NTA-SS-PEG-PCL and Ce6), CC-NPs, and T-CC-NPs. Scale bars, 200 nm. Cas9/sgRNA complex was prepared at a molar ratio of 1. (F) TEM image of T-Au-NP (His-tagged Au nanoparticles conjugated to NTA-NP and then coated with iRGD). (G) Transfection efficiency of CNE-2 cells with T-CC-NPs or NT-CC-NPs. Mean ± SD (n = 3). ***P < 0.001. (H) Confocal laser scanning microscopy (CLSM) of single CNE-2 cells transfected with T-CC-NPs or NT-CC-NPs. Nuclei (blue) were stained using Hoechst 33324; lysosomes (purple) were labeled with LysoTracker Red; Ce6 emitted red fluorescence. Scale bar, 10 μm. (I) Flow cytometry of cells in (B) and (C). (J) Biodistribution of T-CD-NPs and NT-CD-NPs [containing the dye 1,1-dioctadecyl-3,3,3,3-tetramethylindotricarbocyanine iodide (DiR) rather than Ce6] in CNE-2 xenograft mice. (K) Quantification of DiR fluorescence in (E) (mean ± SD, n = 3). eGFP protein was fused with Cas9 to its tumor cell uptake. MFI, mean fluorescence intensity.

The cationic copolymer iRGD-PD was then used to neutralize the negative charge of the CC-NPs and to introduce the tumor-targeting ligand iRGD. The zeta potential and size of the iRGD-PD–coated CC-NPs (T-CC-NPs) increased with the amount of iRGD-PD used (Fig. 2D). At 15 μg of iRGD-PD, T-CC-NPs showed a lower negative charge (−3 mV) and a smaller size (110 nm) than at >15 μg of iRGD-PD (favorable for cell uptake and long blood circulation); thus, 15 μg of iRGD-PD was selected for subsequent experiments.

The nanoparticles were examined using transmission electronic microscopy (TEM). NTA-Ce6-NPs showed a core-shell structure with a diameter of ~65 nm. Binding of Cas9/sgRNA to form CC-NPs increased the radius by ~10 nm, roughly the diameter of Cas9 RNP (33). The iRGD-PD coating (T-CC-NPs) resulted in a sandwich-like structure and further increased the nanoparticle size to ~108 nm (Fig. 2E). His-tagged Au nanoparticles instead of Cas9 RNPs were used to enhance the contrast between the layers for TEM observation. The image revealed a sandwich-like structure for T-Au-NP (Fig. 2F), with a micellar core (PCL block) of ~63 nm in diameter, an interlay (Au nanoparticles conjugated to NTA-SS-PEG) of ~10 nm thick, and an outer layer (iRGD-PD polymer) of ~9 nm thick. To prepare nontargeting CC-NPs (NT-CC-NPs), we used the polymer Allyl-PEG-pAsp(DAB) with an allyl terminus instead of iRGD-PD. In addition, blank nanoparticles (iRGD-PD–coated nanoparticles without Cas9 or Ce6; T-B-NPs) showed low cytotoxicity (>90% cell viability) even at high concentrations of 300 μg/ml (fig. S5A). T-CC-NPs showed excellent colloidal stability in both phosphate-buffered saline (PBS) solution and mouse serum (fig. S5B).

On the basis of the above studies, we investigated the in vitro and in vivo delivery efficiency and biodistribution of the targeting nanoparticles (T-CC-NPs) and nontargeting nanoparticles (NT-CC-NPs). To track the delivery of Cas9 RNP, we fused enhanced green fluorescent protein (eGFP) to Cas9. Cells receiving T-CC-NPs showed stronger green fluorescence (Cas9-eGFP) than cells receiving NT-CC-NPs (2450 ± 32 versus 732 ± 15), indicating that the iRGD enhanced delivery efficiency (Fig. 2G). Flow cytometry data showed that 99.8% of tumor cells emitted both red (Ce6) and green (Cas9) fluorescence and that these fluorescence signals overlapped (Fig. 2, H and I), indicating that Ce6 and Cas9-eGFP were codelivered effectively into tumor cells. Cellular uptake experiments under various conditions were conducted to optimize the codelivery of Cas9 and Ce6 (fig. S5C). Compared with irradiation for 0.5 min, cells irradiated for 1 and 2 min showed stronger fluorescence intensity in all subcellular compartments. However, NIR irradiation for 2 min caused abnormal cell structures such as large vacuoles and rough membranes. A Ce6 concentration of 0.2 μg/ml generated a stronger Cas9-eGFP fluorescence than in a Ce6 concentration of 0.1 μg/ml; it also induced a lower cell damage than the 0.4 μg/ml concentration. On the other hand, a variation of Cas9-eGFP between 2 and 4 μg/ml showed no clear difference in cellular uptake and retention. Therefore, the conditions of NIR irradiation at 800 mW/cm2 for 1 min, Ce6 at 0.2 μg/ml, and Cas9 at 2 μg/ml were used in subsequent experiments.

To probe the in vivo distribution of intravenously injected nanoparticles, we used the NIR fluorescent dye 1,1-dioctadecyl-3,3,3,3-tetramethylindotricarbocyanine iodide (DiR) in place of Ce6. The fluorescence intensity in tumor tissue injected with targeting nanoparticles (T-CD-NPs) was 1.8 times greater than that in tumor tissue receiving NT-CD-NPs. In both cases, the fluorescence intensity reached a maximum at 12 hours after intravenous injection (Fig. 2, J and K). The biodistributions of the targeting and nontargeting nanoparticles were also analyzed on the basis of the fluorescence in excised organs (fig. S5D). The change of Ce6 fluorescence was similar to that of Cas9-eGFP fluorescence in sera of mice receiving T-CC-NPs or NT-CC-NPs (fig. S5E), indicating the codelivery of Ce6 and Cas9-eGFP in systemic circulation. Very few (<0.1%) Ce6-eGFP– and Cas9-eGFP–positive blood cells indicated the shielding effect of nanoparticles to blood cells (fig. S5F). The significant decrease of Ce6 and Cas9-eGFP in circulation with time was related to nanoparticle (T-CD-NPs) accumulation in tumor tissues (Fig. 2J and fig. S5E). The Ce6 fluorescence mainly distributed in cells from liver and tumor tissues, and Cas9-eGFP fluorescence was mostly localized in cells from tumor tissue (fig. S5G). The disappearance of Cas9-eGFP in the liver cells would be consistent with the proteolytic degradation of the protein.

Different from biologically catalyzed reactive oxygen species (ROS) generation in mitochondria and endoplasmic reticulum, photosensitizer-mediated intracellular ROS generation was used in this study to promote endo/lysosomal vesicle rupture for subsequent drug release. NIR irradiation would excite the electrons of the photosensitizer to jump from the lowest singlet excited state (S1) to the lowest triplet excited state (T1), which, in turn, would induce generation of 1O2 from ground-state O2 (3O2). This ROS-generation process can occur in the presence of photosensitizer, NIR irradiation, and O2 (34). Ce6 is a photodynamic therapeutic reagent that generates ROS under NIR irradiation to destabilize lysosome membranes, allowing release of Cas9/sgRNA into the cytoplasm (35). To investigate the intracellular fate of Cas9, we examined the intracellular location of Cas9 with and without Ce6 and NIR irradiation using fluorescence microscopy (Fig. 3). Ce6-loaded nanoparticles effectively generated ROS under NIR irradiation (fig. S5H). Without NIR irradiation and/or Ce6, Cas9-eGFP colocalized with lysosomes (labeled with LysoTracker DNN; purple fluorescence). In contrast, Cas9-eGFP and Ce6 were not colocalized with lysosomes in the presence of both NIR irradiation and Ce6 (Fig. 3B). The green fluorescence of Cas9-eGFP was observed in nuclei (Fig. 3C), indicating nuclear translocation of Cas9 RNP. These results demonstrated that both NIR irradiation and Ce6 were vital for Cas9 RNP lysosomal escape.

Fig. 3 Mechanism and characterization of lysosomal escape of nanoparticles and cytoplasmic release of Cas9/sgRNA.

(A) Schematic of NIR-triggered generation of ROS by Ce6 and lysosomal escape of nanoparticles, followed by release of Cas9/sgRNA in the reducing environment of the cytoplasm. (B) CLSM images of intracellular distribution of nanoparticles under NIR irradiation. T-Cas9/NIR, T-CC/NIR, and NT-CC/NIR indicate T-Cas9-NPs, T-CC-NPs, and NT-CC-NPs with NIR irradiation, respectively. Scale bar, 10 μm. (C) Fluorescence intensities in cells receiving T-Cas9/NIR or T-CC/NIR along the dotted yellow arrows in (B). (D) Flow cytometry and (E) CLSM imaging of ROS levels in cells receiving various treatments. Scale bar, 50 μm. (F) Schematic of glutathione (reduced form) (GSH)–responsive Cas9 RNP release in the cytoplasm. (G) SDS-PAGE of T-CC-NPs in PBS (pH 7.4) with or without 10 mM GSH (mimicking cytoplasmic conditions) and T-In-CC (reducing agent–insensitive nanoparticles) in PBS (pH 7.4) with 10 mM GSH for various times and then centrifuged with ultrafiltration to remove released Cas9/sgRNA. (H) Cas9 RNP release profile of T-CC-NPs in PBS (pH 7.4) with or without 10 mM GSH and T-In-CC-NPs in PBS (pH 7.4) with 10 mM GSH (mean ± SD, n = 3). (I) TEM images and (J) dynamic light scattering (DLS) analysis of T-CC-NPs with or without 10 mM GSH. Scale bar, 100 nm.

To confirm that Ce6 was generating ROS in response to NIR irradiation, we compared the ROS levels in cells with or without nanoparticles and NIR irradiation. Most (86.8%) cells treated with targeting T-Ce6-NPs plus NIR irradiation (T-Ce6/NIR) produced ROS; only 35.8% of cells treated with nontargeting NT-Ce6-NPs plus NIR (NT-Ce6/NIR) produced ROS (Fig. 3D); very few cells (0.41 and 0.11%) produced ROS when treated with T-Ce6-NPs without NIR irradiation or with NIR irradiation alone. In confocal laser scanning microscopy (CLSM) fluorescence imaging, we observed that ROS level appeared to be closely related to the intracellular Ce6 level (Fig. 3E). Cas9 RNP was degraded if trapped in lysosomes for >24 hours but escaped from lysosomes in the presence of NIR-irradiated Ce6 (fig. S5I). These results indicated that treatment of cells with iRGD-modified Ce6-containing nanoparticles and NIR irradiation was required for ROS production.

Upon Ce6/ROS-induced lysosomal escape, Cas9 RNP must be released from nanoparticles to perform gene editing. The breakage of a disulfide bond between NTA and PEG (Fig. 3F) allowed detachment of Cas9 RNP in response to the high glutathione (reduced form) (GSH) level in the cytoplasm. GSH concentration in cells treated with T-Ce6-NPs recovered to 90% of the control cells without NIR irradiation at 1 hour after NIR irradiation (fig. S6A), which was enough to trigger the breakage of disulfide bonds (36). As shown in Fig. 3 (G to J), in solution conditions mimicking the cytoplasm (pH 7.4, 10 mM GSH), most of the Cas9 (88.4 ± 3.8%) was released from the nanoparticles (T-CC-NPs). Detachment of Cas9 RNP from nanoparticles was confirmed on the basis of TEM images and dynamic light scattering (DLS) analysis, as indicated by a reduced particle diameter from 111.7 to 79.9 nm in 10 mM GSH (pH 7.4) solution versus the same solution without GSH (Fig. 3, I and J). A reducing agent–insensitive nanoparticle (T-In-CC-NPs) did not show Cas9 release in 10 mM GSH solution. Overlaying the fluorescence of Cas9-eGFP and nuclei indicated that released Cas9 RNP was relocated to the nucleus when T-CC-NPs were used (Fig. 3B). Inhibitors of different endocytic pathways were used to study the intracellular trafficking of nontargeting nanoparticles (NT-CC-NPs) and targeting nanoparticles (T-CC-NPs) in fig. S6 (B and C): (i) Cell uptake of nanoparticles was adenosine 5′-triphosphate (ATP) dependent, because inhibitors (NaN3 and 4°C) of ATP synthesis reduced the endocytic efficiency of nanoparticles. NaN3 reduced percentage of Cas9-eGFP–positive cells by 51.5% for T-CC-NPs and 59.8% for NT-CC-NPs; 4°C incubation reduced relative percentage of Cas9-eGFP–positive cells by 72.4% for T-CC-NPs and by 78.2% for NT-CC-NPs, respectively; (ii) inhibitors of caveolin-dependent endocytosis (nystatin and β-cyclodextrin), macropinocytosis (amiloride), microtubulin polymerization (nocodazole), and dynamin II (Dynasore) reduced relative percentages of Cas9-eGFP–positive cells by about 10 to 15% for both T-CC-NPs and NT-CC-NPs; (iii) clathrin-mediated endocytosis inhibitor (chlorpromazine) greatly reduced relative percentages of Cas9-eGFP–positive cells by 34.3% for T-CC-NPs and 54.4% for NT-CC-NPs, suggesting that clathrin-mediated endocytosis played a key role in internalization of these nanoparticles; and (iv) competitive inhibition with free iRGD reduced fluorescence intensity of cells transfected with T-CC-NPs to that with NT-CC-NPs. These results demonstrated that T-CC-NPs, upon binding to cell membrane via interactions with iRGD receptor, were internalized into cells via multiple pathways, among which ATP-dependent and clathrin-dependent endocytosis played primary roles. Moreover, colocalization of Golgi marker and Cas9-eGFP fluorescence indicated that Cas9-eGFP in cytoplasm was transported to Golgi apparatus (fig. S6B). Therefore, the combination of iRGD modification, NIR irradiation/Ce6-triggered lysosomal rupture, and GSH-responsive Cas9 release effectively transported Cas9 RNP into the nucleus, achieving lysosomal escape, controlled Cas9 release, and delivery to the nucleus—all important for highly efficient Cas9 gene editing (12).

Next, we determined the gene editing efficiency at the eGFP gene locus of a U2OS-eGFP cell line containing a single copy of the reporter gene eGFP and constitutively expressing a destabilized eGFP protein. Gene editing efficiency was quantified using CLSM, flow cytometry, and genomic cleavage detection (GCD) assays (Fig. 4). Under NIR irradiation, treatment with T-CC-NPs converted 38.9% of the eGFP-positive cells to eGFP-negative cells; treatment with nontargeting NT-CC-NPs converted 20.4% eGFP-positive cells to eGFP-negative cells, indicating a reduced tumor cell uptake of nanoparticles in the absence of iRGD (Fig. 4, A and B). To test whether eGFP gene disruption resulted from insertion/deletion (indel) mutation, we determine the frequency of indel events at the target eGFP locus using a GCD assay (Fig. 4C). We observed an indel event frequency of 42.6% in cells treated with T-CC-NPs and 26.1% in cells treated with NT-CC-NPs; cells treated with T-Cas9/NIR and control (Cas9/sgRNA alone, no nanoparticles) showed almost no indel events or eGFP gene disruption. These results showed that targeted codelivery of Cas9 RNP and Ce6 was necessary to achieve highly efficient gene editing.

Fig. 4 Gene editing efficiency in vitro and in vivo.

(A) CLSM images of U2OS-eGFP cells receiving nanoparticles under NIR irradiation. The U2OS-eGFP cell line contains a single copy of a reporter gene that constitutively expresses destabilized eGFP. Scale bar, 100 μm. (B) eGFP-negative cells quantified using flow cytometry. (C) Genomic cleavage detection (GCD) analysis of eGFP gene disruption in U2OS-eGFP cells treated with nanoparticles. (D) Bioluminescence and fluorescence imaging of luciferase (Luc) and red fluorescence protein (RFP) in mice bearing CNE-2-Luc-RFP recombinant cells at time points after tail vein administration of nanoparticles under NIR irradiation. (E) Quantitation of fluorescence signals of tumor tissues in (D). Curved green and red arrows indicate changes of Luc bioluminescence and RFP fluorescence, respectively. Data are means ± SD (n = 3). **P < 0.01 and ***P < 0.001. (F) Western blot and (G) immunohistochemistry analyses of luciferase and (H) immunofluorescence analyses of Cas9 and Ce6 in tumor tissue sections from CNE-2 xenograft mice receiving various nanoparticles. Luciferase proteins were stained brown. Scale bars, 50 μm.

We next evaluated gene editing in mice bearing CNE-2–luciferase (Luc)–red fluorescence protein (RFP) tumor cells, which constitutively expresses RFP and Luc. We codelivered Cas9/sgRNA complexes targeting the Luc gene and Ce6 into these mice using T-CC-NPs, NT-CC-NPs, and T-Cas9-NPs (iRGD NPs with Cas9 but lacking Ce6), respectively. The intensity of red fluorescence in all mice increased with time, indicating a sustained tumor growth (Fig. 4, D and E). At 3 days after the mice receiving T-CC/NIR or NT-CC/NIR treatments, the Luc bioluminescence intensity was reduced by 82.6 and 52.7% versus day 0, respectively, indicating Luc gene disruption. In contrast, the Luc bioluminescence intensity increased over a 3-day period in mice receiving T-Cas9/NIR, which lacked Ce6 (Fig. 4E). We confirmed luciferase gene disruption at the protein level (Fig. 4, F to H). All mice showed similar levels of Luc protein before nanoparticle treatments. T-CC/NIR and NT-CC/NIR treatments reduced the Luc content in tumor tissues by 70.3 and 42.9%, respectively, in comparison to that in the control group (Fig. 4F). The luciferase level was negatively correlated with the levels of both Cas9 protein and Ce6 in tumor tissue (Fig. 4, D to H). Together, these results indicated that iRGD-mediated delivery of both Cas9 RNP and Ce6 using NIR irradiation was necessary to achieve a high level of luciferase gene disruption.

We next investigated whether Ce6-mediated photodynamic therapy and CRISPR-Cas9–based Nrf2 gene editing could act synergistically to kill tumor cells. Nrf2 plays a critical role in the cell antioxidant response. Inhibiting Nrf2 expression has been shown to enhance tumor cell sensitivity to Ce6-generated ROS and suppress tumor growth (37). First, we confirmed that treatment with T-Ce6-NPs (lacking Cas9) in the absence of NIR caused few apoptotic cells (3.4%) (Fig. 5A). Treatment with T-Cas9-NPs (lacking Ce6) with NIR resulted in only a slight increase in the percentage of apoptotic cells (5.7%). In contrast, treatment with T-Ce6-NPs, NT-CC-NPs, and T-CC-NPs, together with NIR, resulted in 45.7, 53.9, and 83.0% apoptotic cells, respectively. We then tested the effect of treatments at various Ce6 concentrations with or without NIR irradiation on cell viability. The viability of cells treated with T-Ce6-NPs or NT-Ce6-NPs in the absence of NIR irradiation was >90% even at a high concentration of Ce6 (0.25 μg/ml), indicating that Ce6 alone (without NIR) causes little cytotoxicity to CEN-2 cells. Combined nanoparticle/NIR therapy (T-CC/NIR and NT-CC/NIR) reduced Ce6 half maximal inhibitory concentration from 0.28 to 0.17 μg/ml (by 39.3%) and from 0.49 to 0.25 μg/ml (by 49.0%) compared to the T-Ce6/NIR and NT-Ce6/NIR treatments, which lacked Cas9 (fig. S6, D and E). Similar results were observed by staining live and dead cells (green and red fluorescence) (fig. S6F). These results showed that Nrf2 gene editing enhanced the tumor sensitivity to ROS and that the targeted delivery strategy strengthened this effect. We also evaluated the in vivo antitumor effect of combining gene editing and Ce6/NIR photodynamic therapy. T-Ce6-NPs treatment without NIR and T-Cas9-NPs treatment with NIR did not significantly inhibit tumor growth (Fig. 5, B and C, and fig. S7, A to C). In contrast, treatment with T-Ce6/NIR or T-CC/NIR reduced tumor volume by 41.8 and 87.7% and reduced tumor weight by 43.8 and 92.5%, respectively, when compared with mice treated with T-Ce6-NPs in the absence of NIR at 18 days after the first drug administration (Fig. 5B and fig. S7, B and C). Targeting T-CC/NIR treatment showed greater antitumor growth effect than nontargeting NT-CC/NIR and improved the mouse survival rate and survival time (Fig. 5C). All mice treated with single drug (Ce6 or Cas9) treatments (T-Ce6-NPs or T-Cas9-NPs with NIR) or with nontargeting treatment (NT-CC-NPs) died at 45 days after the first drug administration; T-CC-NPs in the presence of NIR (T-CC/NIR) treatment prolonged survival time by 80% to >60 days. The body weight of mice receiving T-CC/NIR treatment was similar to that of control (PBS-treated) mice (fig. S7A), indicating a low cytotoxicity for the T-CC/NIR treatment. In addition, the T-CC/NIR treatment resulted in the greatest tumor cell apoptosis and necrosis based on analysis of hematoxylin and eosin (H&E)–stained and terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL)–stained tumor tissue sections (Fig. 5D and fig. S7D). These results demonstrated that the combined treatment involving both Ce6-based photodynamic therapy and Nrf2 gene editing had a greater antitumor effect than any single treatment alone, indicating additive or synergistic effects by comparative experiments (e.g., tumor growth, survival time, genome editing assays, etc.) of combination and single therapy. All liver tissues from mice treated with various treatments showed no ROS generation and normal level of Nrf2; on the contrary, tumor tissue from mice treated with T-CC/NIR showed significantly higher ROS fluorescence and lower Nrf2 level than that with T-Cas9/NIR (Fig. 5, F and G). Mice treated with T-CC/NIR showed similar tissue structures to those of control mice. Liver tissue sections from control mice are shown in Fig. 5E and fig. S7E. These results indicated that the spatially controlled ROS generation resulted in tumor-specific gene editing and prevent unwanted gene editing in normal tissues.

Fig. 5 Antitumor activity and spatial specificity of combined photodynamic therapy and Nrf2 gene editing.

(A) Flow cytometry analysis of cell apoptosis. A1, dead cells; A2, late apoptotic cells; A3, normal cells; and A4, early apoptotic cells. (B) Tumor growth in CEN-2 xenograft tumor-bearing mice after tail vein injection of nanoparticles with or without NIR irradiation. Data are means ± SD (n = 6). **P < 0.01 and ***P < 0.001. Black arrows in (C) Black arrows in (B) indicate intravenous injection times. (C) Survival of CEN-2 xenograft mice in (B). Mice were euthanized when tumor volume reached 2000 mm3. (D) Hematoxylin and eosin (H&E)–stained and terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL)–stained tumor tissue sections from CEN-2 xenograft mice receiving various treatments. (E) H&E staining of tissue sections from major organs. (F) ROS and (G) Nrf2 protein levels in tumor and liver tissues form CEN-2 xenograft tumor-bearing mice receiving T-CC/NIR or T-Cas9/NIR. NIR irradiation was only applied on tumor tissues. Nuclei were stained blue using 4′,6-diamidino-2-phenylindole (DAPI); ROS and Nrf2 protein were labeled green and red, respectively. Scale bars, 50 (D and E) and 100 μm (F and G). Doses of Ce6 and Cas9/sgRNA were 1 and 1.5 mg/kg, respectively. Cas9/sgRNA complex was prepared at a molar ratio of 1.

We investigated the mechanism of the combined therapy (Fig. 6A). First, we evaluated the effect of combining Ce6 and CRISPR-Cas9 on the ROS level. The ROS fluorescence intensity in cells treated with T-CC/NIR was 2.7 times that of cells treated with T-Ce6/NIR (Fig. 6B and fig. S8, A and B), indicating that Cas9-mediated Nrf2 gene editing enhanced the ROS level of tumor cells. Next, we analyzed the gene editing efficiency in vitro and in vivo. Tumor cells receiving T-CC/NIR showed greater Cas9 and lower Nrf2 protein expression than cells receiving NT-CC/NIR in vitro (fig. S8C) and in vivo (Fig. 6C and fig. S7D). However, in tumor cells receiving T-Cas9/NIR and T-CC-NPs without NIR, Cas9 was hardly detected, and the Nrf2 protein level was almost the same as the control group. These results indicated the importance of both Ce6 and NIR irradiation in Cas9 lysosomal escape (avoiding degradation) and Nrf2 gene disruption. We also quantified Nrf2 gene disruption efficiency using deep sequencing (Fig. 6, D to F, and fig. S8E), real-time quantitative polymerase chain reaction (PCR) (fig. S8F), and GCD assay (Fig. 6G). In CNE-2 cells, T-CC/NIR and NT-CC/NIR treatments resulted in 31.2 and 15.4% gene disruption efficiency, respectively, 67.7 and 35.71% reduction in mRNA, and 35.6 and 18.5% reduction in protein level in the tumor tissue, consistent with the 26.7% mutation rate following T-CC/NIR treatment, as determined by deep sequencing. In contrast, T-Cas9/NIR treatment (lacking Ce6) did not induce indel events at the Nrf2 locus.

Fig. 6 Mechanism of combined therapy of Ce6 photodynamic therapy and Nrf2 gene editing.

(A) Schematic of synergistic effects of CRISPR-Cas9 gene editing and NIR-triggered ROS release. (B) Flow cytometry of ROS levels and (C) immunohistochemistry analyses of Cas9 and Nrf2 proteins in tumor tissue sections from CNE-2 xenograft mice receiving nanoparticles. Scale bar, 50 μm. Cas9 and Nrf2 protein were stained brown, respectively. Deep sequencing analysis of indel percentage in (D) tumor tissue and (E) liver tissue from CNE-2 xenograft mice treated with T-CC/NIR. bp, base pair. (F) Comparison of on- and off-target of gene editing in tumor and liver tissues. (G) GCD analysis of tumor cells receiving various treatments. (H) Western blot and (I) immunofluorescence analyses of Ki67 (red), apoptotic protein caspase-3 (red), angiogenesis factors vascular endothelial growth factor–A (VEGF-A; green) and hypoxia-inducible factor 1α (HIFα; green), and vascular endothelial cell biomarker CD31 (red) in tumor tissue sections from CNE-2 xenograft mice treated with nanoparticles. Nuclei were stained blue. Scale bar, 100 μm. T-CC-NPs, NT-CC-NPs, and T-Cas9-NPs were prepared at an NTA/Cas9 ratio of 8 and a dose of 1.5 mg/kg Cas9 RNP.

As described above, tumor-specific gene editing was achieved via spatially controlled NIR irradiation–triggered lysosomal escape of Cas9-containing nanoparticles and reducing agent–sensitive drug release in the cytoplasm. Nanoparticle treatment without NIR irradiation caused little gene editing; thus, the treatment should avoid unwanted side effects in normal organs such as the liver and lungs. To confirm the lack of off-target gene editing, we measured indel events at off-target loci in the tumor and in liver tissues using deep sequencing. Tumor tissues in mice receiving T-CC/NIR treatment showed off-target/on-target indel event frequencies of 0.08 and 0.15% at two loci (Fig. 6, D and F)—lower than those in mice receiving lipid- or cell-penetrating peptide-mediated delivery of Cas9 expression plasmids (7). An indel event frequency of only 3.77% was observed in liver tissue without subtracting background single-nucleotide polymorphism (Fig. 6, E and F).

Since the Nrf2 protein is also associated with angiogenesis (38), we tested the expression of the angiogenetic factors hypoxia-inducible factor 1α (HIF1α) and vascular endothelial growth factor–A (VEGF-A) in the tumor tissue. Up-regulation of Nrf2 in mice treated with T-Ce6/NIR increased the expression of HIF1α and VEGF-A. T-CC/NIR treatment resulted in 50.0 and 44.6% reduction of HIF1α and VEGF-A levels compared to those following T-Ce6/NIR treatment and resulted in the lower vascular density (Fig. 6, H and I, and fig. S9). T-CC/NIR treatment also resulted in higher levels of apoptotic protein caspase-3 and lower levels of the tumor proliferative biomarker Ki67. In contrast, T-Cas9/NIR treatment did not inhibit HIF1α, VEGF-A, or Ki67 expression and showed a lower caspase-3 level and a high vascular tube density.


Photodynamic therapy is an effective clinical approach to treat superficial nasopharyngeal carcinoma by generating ROS to kill tumor cells. However, tumor cell develops antioxidant mechanism to escape from ROS toxicity (39). Combined gene editing with photodynamic therapy is a promising modality to overcome this limitation by disrupting antioxidant gene. Direct delivery of Cas9 RNP has emerged as one of most exciting approach of gene editing (10). Therefore, nanocarrier for codelivery of Cas9 RNP and photosensitizer is needed to enhance tumor cell’s sensitivity to ROS. Spatially controlled activation of Cas9 is an important strategy to achieve tumor-specific gene editing, thus preventing unwanted genetic mutation in normal tissue.

In this study, we report an NIR irradiation and reducing agent–dual sensitive nanocarrier for the codelivery of the photosensitizer Ce6 and Cas9/sgRNA complexes. Codelivery of Ce6 and Cas9 RNP targeting Nrf2 likely enhanced the antitumor effect via three mechanisms: (i) ROS generated by Ce6 following NIR irradiation induced cell apoptosis and disrupted endosomes/lysosomes, causing release of Cas9/sgRNA into the cytoplasm (35). (ii) In the cytoplasm, high GSH levels caused breakage of a disulfide bond between NTA and PEG, resulting in the detachment of Cas9/sgRNA from the nanoparticles, and disruption of Nrf2. Nrf2 is a critical regulator that prevents ROS accumulation and promotes angiogenesis (38); thus, disrupting Nrf2 prevents tumor cell escape from ROS-mediated killing and inhibits tumor progression. (iii) Spatially controlled ROS generation activated gene editing only in tumor tissues, thus preventing gene editing events in normal tissues.

In summary, spatial NIR irradiation allowed tumor-specific gene editing and cytoplasmic Cas9/sgRNA release that resulted in highly efficient gene editing. Nrf2 gene editing enhanced the tumor sensitivity to the NIR/Ce6-generated ROS for synergistic photodynamic therapy and gene editing. Ce6-based photodynamic therapy is workable for superficial tumor therapy but is limited for deep-seated tumor therapy due to lower tissue penetration of 670-nm laser (39). Theoretically, the described strategy based on NIR– and reducing agent–dual sensitive nanoparticles may be applied in combined therapy of gene editing with other photosensitizers, including protein-capped nanogolds, up-conversion nanoparticles, and Ag2S quantum dots (31, 40), which may be excited by NIR II light. Compared to Ce6, these photosensitizers can generate ROS in deep tumor tissues under excitation by NIR II light. Moreover, by using other stimuli-sensitive bonds (e.g., pH sensitive, enzyme sensitive, and NIR sensitive bonds) instead of reducing agent–sensitive disulfide bonds, one may also endow nanomedicines with more smart control features. Overall, this multifunctional nanocarrier has several advantages over existing vectors and is promising for use in topical antitumor therapy.


Chemical reagents

General chemical reagents were purchased from Guangzhou Chemical Reagent Company (Guangzhou, China). tert-Butyl bromoacetate, Nε-carbobenzyloxy-l-lysine-tert-butyl ester hydrochloride, 10% Pd/C catalyst, ε-caprolactone, N,N-diisopropylethylamine, stannous octoate, N-Boc-1,4-butanediamine, 2,2-azobisisobutyronitrile, cysteamine hydrochloride, N-Boc-1,4-butanediamine, and 3-maleimidopropionic acid N-hydroxysuccinimide (NHS) ester were purchased from Sigma-Aldrich (St. Louis, MO, USA). 3,3′-Disulfanediyldipropionic acid, 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride, NHS, and DiR were purchased from J&K Chemical Reagent Co. (Beijing, China). Ce6 and singlet oxygen sensor green (SOSG) were purchased from Dalian Meilun Biotech Co. (Dalian, China). α-Allyl-ω-hydroxyl PEG (APEG-OH; Mn, 3.4 kDa) was purchased from 3A Chemicals Co. (China). iRGD-SH peptide [Ac-C-c(CRGDKGPDC)] was synthesized by Haode Peptide Company (Wuhan, China). The human nasopharyngeal carcinoma cell line CNE-2 was obtained from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China). U2OS-eGFP cells were provided by K. W. Leong. Primers for sgRNA transcription and GCD assay were synthesized by Sangon Biotechnology Company (Shanghai, China). Cell culture media, penicillin-streptomycin, fetal bovine serum, and 0.25% trypsin were purchased from Gibco BRL (Carlsbad, CA, USA). 3-(4,5-Dimethyl-thiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) and the annexin V fluorescein isothiocyanate (FITC) and propidium iodide (PI) apoptosis kit were purchased from KeyGen Biotechnology Company (Nanjing, China). The LIVE/DEAD Viability/Cytotoxicity Kit was purchased from Invitrogen (Carlsbad, CA, USA).

Preparation of Ce6-containing nanoparticles (NTA-Ce6-NPs)

The detailed synthesis of the NTA-SS-PEG-PCL copolymer is described in the Supplementary Materials. The preparation of Ce6-containing NTA-NPs (NTA-Ce6-NPs) and blank anionic micelles (NTA-B-NPs) is summarized in Fig. 1. Briefly, 22 mg of NTA-SS-PEG-PCL and 2.2 mg of Ce6 were dissolved in 4 ml of dimethyl sulfoxide (DMSO) and tetrahydrofuran (THF) (DMSO:THF, 1:3). The solution was added to deionized water under sonication (60 Sonic Dismembrator, Fisher Scientific, USA) and dialyzed [molecular weight cutoff (MWCO), 14 kDa] against 20 mM Hepes, 150 mM KCl, and 1 mM NiSO4 (pH 7.4) and then washed three times with 20 mM Hepes and 150 mM KCl (pH 7.4) using ultrafiltration (MWCO, 100 kDa; Millipore Amicon Ultra) to remove excess nickel ions. The NTA-Ce6-NPs solution was filtered through a 220-nm syringe filter and stored at 4°C for experiments. DiR-loaded micelles (NTA-DiR-NPs) were prepared using the same procedure, but with DiR instead of Ce6.

Preparation of CC-NPs and T-CC-NPs nanoparticles containing Cas9/sgRNA and Ce6

The detailed preparation of Cas9/sgRNA and synthesis of the cationic iRGD-PD copolymer are described in the Supplementary Materials. The preparation of CC-NPs is summarized in Fig. 1. Briefly, Cas9 (30 μg) and sgRNA (7.5 μg) were mixed in 20 mM Hepes and 150 mM KCl (pH 7.4) at 37°C for 5 min. This solution was added to 240 μg of NTA-Ce6-NPs (5 μg/μl) and kept at 4°C for 5 hours. The solution was centrifuged three times using ultrafiltration (MWCO, 300 kDa; Sartorius Vivaspin 500) to remove free Cas9 RNP. A solution of anionic CC-NPs (50 μl, 5.5 μg/μl) was then mixed with 15 μg of cationic iRGD-PD (5 μg/μl) in 20 mM Hepes and 150 mM KCl (pH 7.4). The mixture was incubated at room temperature for 10 min to obtain a solution of iRGD-PD–coated CC-NPs (T-CC-NPs). APEG-pAsp(DAB)–coated CC-NPs (NT-CC-NPs) lacking iRGD were prepared using the same procedure except using APEG-pAsp(DAB) copolymer instead of iRGD-PD.

Copolymer and nanoparticle characterization

1H-NMR spectra were collected using a 400-MHz spectrometer (Bruker, Germany) at room temperature. FTIR spectra were recorded using an FTIR spectrometer (Nicolet/Nexus 670, USA) between 4000 and 500 cm−1, with a resolution of 2 cm−1. Copolymer molecular weights and molecular weight distributions were analyzed using a GPC system (Water Breeze, USA), consisting of a Waters 1515 pump, Ultrahydrogel 500 and 250 columns, and a Waters 2417 differential refractive index detector, with PEG as the calibration standard; a mobile phase of N,N′-dimethylformamide containing LiBr (1 g/liter) at a flow rate of 1.0 ml/min at 40°C was used. Fluorescence spectra were obtained using a spectrofluorophotometer (PerkinElmer Ltd., UK). Raman spectra were obtained using an FT Raman spectrometer (Nicolet NXR 9650, USA) with an excitation wavelength of 1064 nm. TEM of nanoparticles was performed using a JEM-1400 Plus (JEOL, Japan) system at 120 kV. Sample solution (5 μl, 0.5 mg/ml) was dropped onto a copper grid coated with amorphous carbon. After being dried at 40°C, the sample was stained with phosphotungstic acid solution (3 weight %). The grid was dried overnight. Nanoparticle sizes and zeta potentials were determined by DLS using a Zetasizer Nano ZS instrument (Malvern Instruments Ltd., UK) at 25°C.

Loading efficiency and loading content of Cas9 and Ce6

Lyophilized T-CC-NPs (2 mg) were dissolved in 50 ml of DMSO. Ultraviolet (UV) absorbance of Ce6 in DMSO at 404 nm was measured using a UV-visible spectrometer (PerkinElmer UV750, USA). Ce6 concentration was calculated using a calibration curve for Ce6 in DMSO. To determine the Cas9 RNP loading content and efficiency, a mixture of Cas9 (30 μg) and sgRNA (7.5 μg) was added to different volumes of NTA-Ce6-NPs solution (5 μg/μl) and incubated at 4°C for 5 hours. The solution was then centrifuged with ultrafiltration (MWCO, 300 kDa; Sartorius Vivaspin 500) to remove free Cas9/sgRNA and concentrated to 30 μl. The amount of Cas9 protein in the T-CC-NPs was measured using SDS-PAGE. Loading content (LC) and drug loading efficiency (LE) were determined using the following formulas: LC = (mass of Ce6 or Cas9 / mass of T-CC-NPs) × 100%, LE = (mass of Ce6 or Cas9 in T-CC-NPS / mass of Ce6 or Cas9 in feed) × 100%.

Cas9 RNP release from nanoparticles in vitro

T-CC-NPs (200 μg) were added to 400 μl of 50 mM PBS at pH 7.4 with or without 10 mM GSH (mimicking cytoplasmic conditions) and T-In-CC-NPs (reducing agent–insensitive nanoparticles, 200 μg) to 400 μl of 50 mM PBS at pH 7.4 with 10 mM GSH. The resultant solutions were incubated for different times at 37°C and then washed three times to remove released Cas9/sgRNA using ultrafiltration. The amount of Cas9 remaining in T-CC-NPs or T-In-CC-NPs was determined using SDS-PAGE. The amount of Cas9 released was calculated by subtracting the remaining amount of Cas9 in T-CC-NPs or T-In-CC-NPs from the original amount in the T-CC-NPs or T-In-CC-NPs.

Detection of singlet oxygen in vitro

Singlet oxygen production was measured, as described previously (40). Briefly, SOSG reagent was dissolved in methanol and mixed with a solution of T-CC-NPs at a final concentration of 5 μM SOSG and 1 μM Ce6. This solution was used to detect singlet oxygen generation by NIR (671 nm) irradiation for various times at a power density of 800 mW/cm2. Singlet oxygen production was measured according to the SOSG fluorescence at 530 nm under 494-nm excitation.

Cell uptake of Cas9-eGFP

CNE-2 cells were plated in a glass-bottom well at 1 × 104 cells per well. After incubation overnight at 37°C under a humidified atmosphere with 5% CO2, the cells were treated with various nanoparticle formulations with or without NIR irradiation. Cas9-eGFP fusion protein (GenScript, Nanjing, China) was used to track cell uptake of Cas9. The concentrations of Cas9-eGFP and Ce6 were 2 and 0.2 μg/ml, respectively. After incubation for 4 hours, the medium was replaced with a fresh medium without nanoparticles, and NIR irradiation was applied for 1 min. After further incubation for 24 hours, cell nuclei were stained with Hoechst 33342 (10 μg/ml) for 15 min. Fluorescence from Ce6, Cas9-eGFP, and nuclei was recorded using a CLSM (Leica SP8, Germany). Cell uptake of Cas9-eGFP was also quantified using a flow cytometer (Attune NxT, Invitrogen, America). Briefly, CNE-2 cells were seeded in a 12-well plate at 5 × 104 cells per well. After treatment with different nanoparticle formulations, the cells were rinsed three times with 1 × PBS and detached using 0.25% trypsin. The detached cells were resuspended in 300 μl of 1 × PBS for flow cytometry analysis.

Measurement of intracellular ROS

ROS levels in CNE-2 cells receiving NIR irradiation were evaluated using a 2′,7′-dichlorofluorescein diacetate (DCFH-DA; 10 μM) ROS-detection kit. Briefly, the cells were seeded on a 24-well plate at 5 × 104 cells per well and incubated overnight at 37°C under a humidified atmosphere with 5% CO2. After treatment with different nanoparticles with or without NIR irradiation for 4 hours, the ROS-detection agent DCFH-DA was added, and the cells were incubated for 30 min at 37°C. NIR irradiation was performed if required. The fluorescence of Hoechst 33342, Ce6, and DCFH was recorded by CLSM and flow cytometry, respectively. The green fluorescence intensity of DCFH indicated the ROS level. To evaluate the effect of Nrf2 gene disruption on the ROS level, CNE-2 cells treated with T-CC-NPs for 4 hours were incubated with fresh culture medium following with NIR irradiation. The cells were then incubated for 6 hours under hypoxic conditions with 5% CO2 and 94% N2 at 37°C. The ROS level in cells receiving Nrf2 gene disruption was evaluated, as described above.

Analysis of lysosomal escape of nanoparticles and cytoplasmic release of Cas9/sgRNA

To analyze the NIR-controlled lysosomal escape of nanoparticles and cytoplasmic release of Cas9, CNE-2 cells were incubated for 2 hours following NIR irradiation. Cas9-eGFP fusion protein was used to monitor the intracellular distribution of Cas9. Cell nuclei were stained blue using Hoechst 33342 (10 μg/ml), and lysosomes were stained red using LysoTracker Red (0.5 μg/ml). The intracellular distributions of Cas9-eGFP protein and Ce6 were recorded using CLSM.

Cell viability and apoptosis

Cell viability was quantified using an MTT assay and an annexin V FITC/PI apoptosis assay. Briefly, CNE-2 cells were seeded into a 96-well plate at a density of 5 × 103 cells per well and cultured for 24 hours at 37°C under a humidified atmosphere with 5% CO2. CNE-2 cells were treated with five formulations: T-Ce6-NPs, NT-Ce6-NPs, T-Cas9-NPs, NT-CC-NPs, and T-CC-NPs. The sgRNA directed Cas9 to the Nrf2 gene. Cytotoxicity of various Ce6 concentrations (0.05, 0.1, 0.15, 0.2, and 0.25 μg/ml) was investigated with or without NIR irradiation. Cas9/sgRNA (2.5 μg/ml) was added at a Cas9/sgRNA mass ratio of 4 (i.e., a Cas9/sgRNA molar ratio of 1). After the cells were incubated for 48 hours, MTT solution (0.5 mg/ml) was added to each well, the cells were incubated for another 4 hours, and the absorbance at 570 nm was measured using a microplate reader (BioTek). For the cell apoptosis assay, cells were plated in a 12-well plate at a density of 5 × 104 cells per well and were incubated overnight. The cells were then treated with T-Ce6-NPs, T-Cas9-NPs, NT-CC-NPs, or T-CC-NPs. The concentrations of Ce6 and Cas9/sgRNA were 0.2 and 2.5 μg/ml, respectively. At 48 hours after treatments, the cells were analyzed using flow cytometry. Cells were exposed to 671-nm NIR laser irradiation (800 mW/cm2, 1 min) and hypoxic conditions for 48 hours after irradiation.

Gene disruption assay

To determine gene editing efficiency in eGFP reporter gene and endogenous Nrf2 gene, sgRNA targeting eGFP gene and Nrf2 was prepared and complexed with Cas9, respectively. U2OS-eGFP cells were cultured overnight at a plating density of 5 × 104 cells per well at 37°C under a humidified atmosphere with 5% CO2. Cells were treated with nanoparticles for 4 hours followed by NIR irradiation if required. After further incubation for 48 hours, eGFP-positive cells were observed and quantified using CLSM and flow cytometry. Cells were also harvested to isolate genomic DNA using a Dzup Genomic DNA Isolation Reagent (Tiangen Biotech Co. Ltd., Beijing, China). The sgRNAs targeting the eGFP gene and Nrf2 were amplified using the High-Fidelity PCR Master Mix (Sangon Biotech Co. Ltd., Shanghai, China) and PCR primers (table S1). Indel events were detected using a T7 Endonuclease I kit (New England Biolabs, USA), and indel event frequencies were calculated using ImageJ (Image J2x, National Institutes of Health, USA). To image intracellular Cas9 RNP and Nrf2 protein levels, cells were incubated for 48 hours under hypoxic conditions after treatment with nanoparticles. Cells were then fixed in 4% paraformaldehyde and stained. Levels of Nrf2 protein in cells receiving various treatments were determined by Western blot. Detailed protocols for immunofluorescence staining and Western blot are in the Supplementary Materials.

Mouse tumor model

Female nu/nu nude mice (6 weeks old) were purchased from Guangdong Medical Laboratory Animal Center and were used in accordance with the regulations of the Animal Care and Use Committee of Sun Yat-sen University (Guangzhou, China). To establish subcutaneously implanted tumor model, 1 × 107 CNE-2 cells were injected subcutaneously in the arm pit. When tumor volume reached 100 mm3, mice were treated with various nanoparticles.

In vivo fluorescence imaging and luciferase gene disruption assay

NIR fluorescent dye DiR was used instead of Ce6 to prepare targeting (T-CD-NPs) and nontargeting (NT-CD-NPs) nanoparticles for in vivo imaging. Fluorescence of mice bearing CEN-2 tumor was imaged using a Carestream IS 4000 imaging system at different time points after nanoparticle administration. Excised organs (i.e., heart, liver, spleen, kidney, and lung) and tumors were imaged at the last time point after nanoparticle administration. To evaluate in vivo gene disruption efficiency, recombinant CEN-2 cells constitutively expressing RFP and luciferase proteins were constructed using the protocol described in the Supplementary Materials. Luciferase gene–targeted Cas9/sgRNA was used to prepare T-Cas9-NPs, NT-CC-NPs, and T-CC-NPs. The doses of Ce6 and Cas9/sgRNA were 1 and 1.5 mg/kg body weight, respectively. At 12 hours after intravenous administration via the tail vein, tumors were exposed to NIR irradiation. At various time points after NIR irradiation, bioluminescence of luciferase and fluorescence of RFP were recorded using a Carestream live-animal imaging system. Tumor tissues were also sliced and lysed to evaluate the levels of luciferase protein by Western blot. The in vivo gene disruption efficiency was also assessed using immunohistochemical staining. After tumor sections were fixed in 4% paraformaldehyde overnight, the levels of luciferase protein and Cas9 were measured using rabbit polyclonal anti-firefly luciferase antibody (Abcam, Cambridge, UK) and mouse monoclonal anti-Cas9 antibody (Cell Signaling Technology, Danvers, MA, USA) with CLSM.

In vivo therapeutic evaluation

Mice were randomly divided into six groups (n = 12) and were intravenously administrated every 2 days (at 0, 2, 4, 6, 8 days after injection) with 100 μl of PBS, T-Ce6-NPs, T-Cas9-NPs, NT-CC-NPs, or T-CC-NPs, respectively. The doses of Ce6 and Cas9/sgRNA were 1 and 1.5 mg/kg body weight, respectively. If required, then tumor sites were exposed to NIR irradiation at 12 hours after nanoparticle administration. Body weights and tumor volumes were recorded every 2 days. Half of the mice were sacrificed at 18 days after administration, and the remaining mice were maintained with feeding to record the survival time. Tumor volume was defined as V = ab2/2, where a is tumor length and b is tumor width. Tumor and major organs (heart, liver, spleen, lung, and kidney) were sliced, and the sliced tumor tissues were weighed. To determine ROS level in tumor tissue, at 6 hours after NIR irradiation, mice were injected intravenously via the tail with DCFH-DA at a dose of 1 mg/kg body weight. After 2 hours postinjection, mice were sacrificed, and tumor tissues were excised. ROS levels in cells isolated from tumor tissues were recorded by CLSM and flow cytometry, respectively. To evaluate off-target gene editing in tumor tissue and normal tissue, tumor and liver tissues were subjected to deep sequencing by Sangon Biotech Co. Ltd. (Shanghai, China). The levels of Nrf2, HIF-1α, and VEGF-A proteins were determined by Western blot. H&E stain and TUNEL assay were performed on the excised tumor tissue. Expression of Cas9, Nrf2, Ki67, and caspase-3 protein in excised tumor tissue sections was assessed using immunohistochemical staining. Tissue sections were observed using a Vectra Polaris system (PerkinElmer, USA). The detailed information about in vivo experiments were described in the Supplementary Materials.

Statistical analysis

Data were shown as means ± SD. Statistical significance was determined using a two-sided Student’s t test (*P < 0.05, **P < 0.01, and ***P < 0.001); ns, no significance.


Supplementary material for this article is available at

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: Funding: This work was supported by the National Key R&D Program of China (2016YFE0117100), the National Natural Science Foundation of China (21875289 and U1501243), the Guangdong–Hong Kong Joint Innovation Project (2016A050503026), the Major Project on the Integration of Industry, Education and Research of Guangzhou City (201704030123), the Science and Technology Program of Guangzhou (201704020016), the Guangdong Innovative and Entrepreneurial Research Team Program (2013S086), and the Guangdong Province Science and Technology Innovation Special Fund (International Scientific Cooperation, 2018A050506035). Author contributions: D.C. designed the research strategy and experiments; D.C., S.D., X.L., S.L., and J.C. performed experiments and/or analyzed the data; D.C., K.W.L., S.D., X.L., M.L., and S.Y.C. wrote the paper; and D.C. and K.W.L. supervised the whole process. Competing interests: D.C., S.D., and X.L. are inventors on a pending patent related to this work filed by Sun Yat-sen University (no. 2019108929428, filed on 20 September 2019). The other authors declare that they have no other competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Stay Connected to Science Advances

Navigate This Article