Hierarchical self-uncloaking CRISPR-Cas13a–customized RNA nanococoons for spatial-controlled genome editing and precise cancer therapy

Abstract

CRISPR-Cas13a holds enormous potential for developing precise RNA editing. However, spatial manipulation of CRISPR-Cas13a activity remains a daunting challenge for elaborately regulating localized RNase function. Here, we designed hierarchical self-uncloaking CRISPR-Cas13a–customized RNA nanococoons (RNCOs-D), featuring tumor-specific recognition and spatial-controlled activation of Cas13a, for precise cancer synergistic therapy. RNCOs-D consists of programmable RNA nanosponges (RNSs) capable of targeted delivery and caging chemotherapeutic drug, and nanocapsules (NCs) anchored on RNSs for cloaking Cas13a/crRNA ribonucleoprotein (Cas13a RNP) activity. The acidic endo/lysosomal microenvironment stimulates the outer decomposition of NCs with concomitant Cas13a RNP activity revitalization, while the inner disassembly through trans-cleavage of RNSs initiated by cis-recognition and cleavage of EGFR variant III (EGFRvIII) mRNA. RNCOs-D demonstrates the effective EGFRvIII mRNA silencing for synergistic therapy of glioblastoma cancer cells in vitro and in vivo. The engineering of RNSs, together with efficient Cas13a activity regulation, holds immense prospect for multimodal and synergistic cancer therapy.

A hierarchical self-uncloaking strategy is explored for spatial-controlled Cas13a genome editing and precise synergistic therapy.

INTRODUCTION

Orchestrated regulation of gene expression at the RNA level is indispensable for maintaining biological homeostasis and improving disease states (1–7). CRISPR-Cas enzyme, Cas13a (formerly C2c2), is leveraged as an RNA-guided RNA-targeting CRISPR effector (8–10). Once assembled with mature CRISPR RNA (crRNA), CRISPR-Cas13a exhibits targeted recognition of single-stranded RNA (ssRNA; cis-recognition and cleavage) and trans-cleavage ribonuclease (RNase) function (trans-cleavage/collateral effect), holding promise for RNA gene silencing (11). CRISPR-Cas13a has proven to be a potential therapeutic agent for cancer treatment in mammalian cells because of the unique “collateral effect,” which may resist the complex escape mechanism of tumors in most cases (12). However, despite notable developments in CRISPR-Cas13a, contributing toward multifarious cancer therapeutics, the excessive and constantly active Cas13a may exacerbate undesirable safety concerns (13). To cope with this puzzler, recent efforts have been dedicated to seeking various artifices to manipulate its RNase activity. For example, a catalytically dead variant (dLwaCas13a) was created by mutating catalytic arginine residues to regulate Cas13a function, which retained RNA binding ability but lost cleavage ability (14). Meanwhile, dual-locking nanoparticle was proposed to restrict the activation of CRISPR-Cas13a in tumor tissues by responding to both pH and H₂O₂ in the tumor microenvironment (12). Besides, the anti-CRISPR (acr) gene molecules encoded by the bacteriophage or prophage region were identified as potential off-switches for regulating Cas13a endonuclease activity in human cells with controllable effectiveness (13). However, these approaches involving gene editing and activity regulation of Cas13a in vivo mainly relied on the plasmid, and no direct regulation of Cas13a/crRNA ribonucleoprotein (Cas13a RNP) activity was attempted. Moreover, conventional encapsulation or complexation of the Cas13a plasmid with liposome or lentivirus is often impeded by their potential insertional mutagenesis, off-target side effects, immunogenicity, and exposure to complex matrices (15–16). Considering the undesired biosafety issues after the systemic administration of the CRISPR-Cas13a system, it is imperative to design a feasible strategy to regulate the RNase activity of Cas13a across spatial dimensions, guaranteeing their safe application in cancer therapy.

Emerging therapeutics using synergistic gene therapy/chemotherapy open appealing avenues to treat solid tumors because of their inhibition of tumor-related gene expression, further improving the efficacy of chemotherapy drugs (17–18). However, the co-delivery of RNP with chemotherapy drugs has not been investigated thoroughly. Rolling circle transcription (RCT)–assembled RNA nanosponges (RNSs) have emerged as favorable carriers for in vivo delivery of (bio)chemical molecules, such as genes (19), enzymes (20), or drugs (21), by virtue of their intrinsic biocompatibility, programmability, serum/nuclease resistance, and template customizability (22–25). However, the densely packaged and unique hierarchical nanostructures debilitate the release rate and bioavailability of cargoes, thus attenuating the in vivo therapeutic efficiency (26). Furthermore, by resorting to specially designed exogenous stimuli (e.g., metal ions, polycationic reagents, or nonspecific enzymes) as self-excising initiators, which seriously ignores the cargo peculiarities, unexpected biotoxicity issues may be imposed (27–29). Considering the unique collateral effect of CRISPR-Cas13a after targeted recognition of ssRNA, we envisioned that the collateral activity of Cas13a regulated by tumor-associated gene, which caused architectural transformation of exogenous RNA carriers, could potentially activate the tumor-specific synergistic gene therapy/chemotherapy.

Given these issues, we designed hierarchical self-uncloaking RNA nanococoons (RNCOs-D), with efficient Cas13a activity regulation and multisite self-excising property for spatial-controlled genome editing and precise synergistic gene therapy/chemotherapy (Fig. 1). To implement tumor targetability, a DNA linear template is encoded with two complementary sequences, i.e., aptamer (Apt) and cholesterol-labeled single-stranded DNA (ssDNA) (C-Chol), and a Cas13a trans-cleavage substrate sequence. Through RCT, multivalent Apt is generated to identify epidermal growth factor receptor (EGFR) overexpressed on glioblastoma (GBM) cancer cell membranes, thus guaranteeing the precise delivery of RNCOs-D via receptor-mediated endocytosis. Multiple hybridizations with C-Chol not only condense the RCT-assembled RNA microsponges (RMSs) into the compact RNSs but also increase the cellular uptake kinetics (30–31). Tandem trans-cleavage substrate sequences enable the multisite self-excising property of RNSs. For the proof of concept, the EGFR variant III (EGFRvIII), a unique and persistently activated EGFR mutant subtype featuring a junction of exons 1 to 8, was selected as the silencing target (32). Aberrant EGFRvIII expression, invariably observed in GBM cancer cells, enables DNA-repair acceleration and chemoresistance, implying that targeting EGFRvIII mRNA and combining chemotherapy can prove advantageous in GBM treatment (33–35). To restrict the Cas13a activity, Cas13a RNP was initially concealed with positively charged acid-degradable nanocapsules (NCs) (36) and then embedded into RNSs by electrostatic interactions to yield RNCOs (NCs anchored on RNSs) for subsequent chemotherapeutic doxorubicin (DOX) encapsulation (RNCOs-D; fig. S1). For the prototype, the self-uncloaking hierarchy from the outer to inner order starts with endo/lysosomal response to the spatially controlled release of Cas13a/crRNA activity at the first self-uncloaking, followed by its monosite cis-recognition and cleavage of EGFRvIII mRNA, and ends with the multisite trans-cleavage of RNSs to effectively release DOX at the second self-uncloaking. Cargo-customized release in succession realizes synergistic gene/chemo-therapy via spatial-controlled EGFRvIII gene silencing to enhance sensitivity to chemotherapeutics, providing a distinct avenue to achieve multimodal and precise cancer therapy.

Fig. 1. Schematic illustration of hierarchical self-uncloaking RNCOs-D assembly/disassembly for synergistic gene therapy/chemotherapy after specific cellular uptake.

(A) Sequence design of circular template and formation of RNCOs-D composed of RCT-assembled RNSs and acid-degradable NCs. (B) Intracellular self-uncloaking hierarchy of RNCOs-D for cancer therapy. (I) Outer decomposition of NCs induced by acidic endo/lysosomal microenvironment with concomitant Cas13a/crRNA release as the first self-uncloaking. (II) Inner dissociation of RNSs through multisite trans-cleavage initiated by monosite cis-recognition and cleavage of EGFRvIII mRNA to release DOX as the second self-uncloaking. (III) Synergistic gene therapy/chemotherapy.

RESULTS

Assembly and characterization of RNCO architecture

The layer-by-layer assembly of the RNCOs architecture commences with the programming of RNSs as the core framework (fig. S1). In realizing the specific tissue- and/or cell-targeting delivery of Cas13a/crRNA, EGFR (an integral transmembrane protein receptor) plays a pivotal role in defining characteristic signals of GBM cancer cells and was thus selected as the recognition unit (37). Therefore, the complementary sequence of EGFR Apt was modularly encoded into the linear template (L_Apt) (table S1) to generate tandem multivalent Apt through RCT. T7 RNA polymerase moved around L_Apt to produce long concatemeric ssRNA, further assembling into RMSs via nucleic acid–driven Mg₂PP_i crystallization. During the lengthy process of delivering cargoes to cancer cells, excellent vectors ought to be compacted on the nanoscale to prolong blood circulation and to improve accumulation in tumors through enhanced permeability and retention effect (38). Therefore, along with the complementary sequence of C-Chol integrated into L_Apt, RMSs are hybridized with C-Chol maintaining Apt motifs (fig. S2) to be condensed into nanoscale (attributed to the hydrophobic interactions between hydrophobic molecules) (39). The particle size variation during RMSs evolving into RNSs was easily observed from 1 μm to 210 nm by transmission electron microscopy (TEM; Fig. 2, A and B); however, the three-dimensional hierarchical structure remained immutable. Meanwhile, RNSs (184.9 ± 15.2 nm; Fig. 2E) exhibited a smaller hydrodynamic diameter compared to RMSs (987.2 ± 10.2 nm; fig. S3), as reflected by dynamic light scattering (DLS).

Fig. 2. Formation and characterization of RNCOs-D.

TEM of (A) RMSs, (B) RNSs, (C) NCs encapsulating Cas13a/crRNA, and (D) RNCOs. Scale bars, 500, 100, 100, and 100 nm, respectively. (E) DLS analysis of NCs, RNSs, and RNCOs. (F) Native PAGE (12%) analysis of the RCT process: (i) primer (P), (ii) linear ssDNA template (L_Apt), (iii) circular template, (iv) RCT product RMSs, (v) RMSs/C-Chol hybrid RNSs, and (vi) RNCOs. (G) Changes in zeta potential during the gradual formation of RNCOs. Data are means ± SD (n = 3). (H) UV-vis spectra of NCs, RNCs, and RNCOs. (i) CD spectra of NCs, RNCs, and RNCOs. a.u., arbitrary units.

Cas13a from Leptotrichia wadei (LwaCas13a) was expressed with pET-28a-LwaCas13a plasmids (fig. S4) and purified (fig. S5). As a proof of principle, the EGFRvIII crRNA was engineered to demonstrate the fundamental working principle in vitro. After incubation with EGFRvIII crRNA, the bulge-containing stem loop within the crRNA was anchored on the cleft of the Cas13a recognition (REC) lobe to preassemble the Cas13a RNP complex (40). To promote escape from endo/lysosomes and achieve spatial-controlled Cas13a RNP activation, a thin acid-degradable covalently cross-linked polymer with a positive charge was synthesized for concealing Cas13a RNP by in situ free-radical polymerization (41, 42). The formation of NCs was observed via TEM with uniform size and an average diameter of 25 nm (Fig. 2C), consistent with the DLS results (Fig. 2E). The zeta potential of Cas13a RNP changed from −8.12 to +7.8 mV with cloaking of the NCs—indicating the prerequisite for binding to the negatively charged RNSs (Fig. 2G). The cationic feature was expected to promote lysosome escape of Cas13a RNP via the “proton-sponge” effect (25).

The homogeneous RNCOs were gradually assembled by electrostatic interactions between the NCs and RNSs, as verified by polyacrylamide gel electrophoresis (PAGE; Fig. 2F). To intuitively evaluate the payloads of Cas13a RNP, the morphology of RNCOs was observed through TEM (Fig. 2D) using gold nanoparticle (AuNP)–labeled Cas13a (fig. S6). Surface decoration increased the diameter of RNSs from 210 to 260 nm and the zeta potential from −12 to +5 mV (Fig. 2, E and G). Confocal laser scanning microscopy (CLSM) analysis of the colocalization of DOX-encapsulated RNSs (RNSs-D) (red for DOX) and fluorescein amidites (FAM)–N-hydroxysuccinimide (NHS)–labeled Cas13a (green for Cas13a) identified the intricate decoration of NCs onto RNSs (fig. S7). The efficient loading of Cas13a RNP in the RNCO nanoplatform was further demonstrated by ultraviolet-visible (UV-vis) spectra and circular dichroism (CD; Fig. 2, H and I). Moreover, the densely packed hierarchical pores and high surface area were anticipated features in RNCOs as a potential drug carrier. The effective DOX payload capability of RNCOs was calculated to be 4.16 mmol/g, which is 1.2 times higher than that of the RNSs (3.29 mmol/g); this can be attributed to the additional electrostatic interaction between the NCs and DOX (fig. S8).

In vitro hierarchical self-uncloaking properties of RNCOs

With the compact RNCOs assembled, we proceeded to the in vitro validation of the working principle of hierarchical self-uncloaking. A crucial factor in determining the efficacy of the hierarchical self-uncloaking process is the feasibility of acid-stimulated degradation of NCs at the first self-uncloaking, as evaluated by monitoring the cloaking/uncloaking activity variation of Cas13a RNP in the caging/decaging state. As a control, an acid-nondegradable NCs encapsulating Cas13a RNP (nNCs) were prepared with a nondegradable cross-linker, methylenebisacrylamide, instead of a pH-degradable cross-linker. Here, Cas13a RNP activity variation was reflected in terms of fluorescence intensity (FI) for a system of NCs, EGFRvIII mRNA target (T), and FAM-R-BHQ1 (F-R-Q). As expected, acid-triggered disassembly of NCs was elucidated by the trans-cleavage activity recovery of Cas13a RNP released from NCs at pH 5.4, comparable to that of its wild state, while no activity recovery occurred at pH 7.4 (Fig. 3, A and B). Moreover, at pH 5.4 and pH 7.4, Cas13a showed negligible trans-cleavage activity in the cloaking state of nNCs. These results demonstrated that the encapsulation of NCs had no impact on the RNase activity of Cas13a RNP and confirmed the feasibility of NCs enabling acid-responsive first self-uncloaking.

Fig. 3. Hierarchical self-uncloaking property of RNCOs in vitro.

(A) Cas13a RNP activities of NCs and nNCs at pH 7.4 and 5.4 via the trans-cleavage of F-R-Q as fluorescent signal (λ_ex 492 nm, λ_em 518 nm). (B) Fluorescence spectra analysis of Cas13a RNP recognition specificity via the trans-cleavage of F-R-Q as fluorescent signal (λ_ex 492 nm, λ_em 518 nm). (C) Fluorescence spectra analysis of Cas13a/crRNA responding to T (from i to vi: 0, 50, 100, 200, 300, and 400 nM). (D) DOX release from RNCOs and nRNCOs at pH 7.4 and 5.4. (i) RNCOs + T (pH 5.4), (ii) nRNCOs + T (pH 5.4), (iii) RNCOs + mT (pH 7.4), (iv) RNCOs + T (pH 7.4), and (v) nRNCOs + T (pH 7.4). Data are means ± SD (n = 3). (E) TEM characterization of hierarchical self-uncloaking of RNCOs. Scale bars, 100 nm. (F) PAGE analysis of RNCO dissociation efficiency. (G) Schematic illustration of multisite trans-cleavage of RNSs (left) and DOX release profiles of RNCOs with incremental NC payloads (right) (from i to iv: 100 nM, 500 nM, 1 μM, and 2 μM).

With the first self-uncloaking feasibility explicitly verified, we then investigated the Cas13a RNP multiturnover trans-cleavage capacity of RNSs activated by mono-site cis-recognition of mRNA at the second self-uncloaking. T was the prerequisite cofactor to initiate Cas13a RNP collateral effect–mediated nonspecific digestion of the RNA substrate, similar to that within living cells. To explore the recognition specificity of Cas13a RNP toward T, we compared the Cas13a RNP trans-cleavage effect of F-R-Q in the presence of either T or mismatched T (mT), and only observed the collateral effect when T was present (Fig. 3B). Inspired by this, we envisaged that the specific monosite cis-recognition target event would initiate the second self-uncloaking of RNSs. As anticipated, PAGE analysis (fig. S9A) revealed an effective RNA trans-cleavage behavior, as evidenced by the appearance of ssRNA-smeared bands of various lengths over time. Correspondingly, RNSs retained their integrity structure without T incubation, while almost complete degradation occurred with T as observed through TEM (fig. S9B). In the presence of dissociative Cas13a RNP, the second self-uncloaking process was further verified by DOX release from RNSs-D, and the intercalated DOX showed a higher or lower release rate with or without T (fig. S9C). The fluorescence recovery of DOX proceeded rather quickly with increasing T concentration, confirming that drug release occurred in a target-dependent manner (Fig. 3C). In particular, efficient drug release could be executed even at a moderate concentration of T (200 nM), contributing to the in vivo application. Above all, the trans-cleavage capacity of RNSs activated by cis-recognition of T ensured the foundation for efficient second self-uncloaking.

Rational validations for acid degradation and the collateral effect as initiators of the first and second self-uncloaking encouraged further investigation into the hierarchical self-uncloaking properties of our intelligent nanoplatform. In general, a two-stage RNCOs-D self-uncloaking was achieved: (i) An acidic microenvironment induced the disassembly of NCs to recover the activity of Cas13a RNP in a spatially controllable manner, and (ii) RNSs served as substrates for trans-cleavage by Cas13a after cis-recognition of the mRNA target to release drugs. Outer degradation of NCs and inner cleavage of RNSs as the hierarchical self-uncloaking stages were verified by the DOX release behavior of RNCOs-D and DOX-encapsulated nNCs anchored on RNSs (nRNCOs-D) under different pH conditions. The optimal DOX release over 95.0% of RNCOs-D occurred at pH 5.4, while a fairly low release below 14.3% appeared at pH 7.4 (Fig. 3D). To further substantiate the drug release capability through hierarchical self-uncloaking, nRNCOs-D was introduced to investigate drug release kinetics. Mild DOX leakage of nRNCOs-D at pH 5.4 was observed compared to that at pH 7.4, probably from a partial collapse of the RNS Mg₂PP_i skeleton in the acidic environment (43). Despite this, the DOX release efficiency of nRNCOs-D was far less than that of RNCOs-D at pH 5.4, as the first self-uncloaking stage was impeded. Similarly, little DOX release was observed with mT instead of T, attributed to the hampered second self-uncloaking stage. Accordingly, TEM (Fig. 3E) also showed that, in the presence of T, the almost full digestion of RNCOs was observed at pH 5.4, while RNCOs remained intact at the physiological pH 7.4. Then, the hierarchical self-uncloaking efficiency was explored by PAGE analysis, where the initial disintegration of RNCOs occurred within 60 min and more pronounced destruction over time (Fig. 3F). Compared with fig. S9A, RNSs showed faster and more complete dissociation because of the partial collapse in the acidic environment and multisite trans-cleavage function. Specifically, the collateral effect of Cas13a was dominated in RNS disassembly and drug release; therefore, the Cas13a-dependent drug release was extensively explored by monitoring the FI of DOX (fig. S10). The molar ratio between the RNSs and Cas13a RNP was determined as 5 μM:0.1 mM, which is sufficient to guarantee an efficient second self-uncloaking. As for RNCOs, multisite trans-cleavage of RNSs was confirmed by the incremental release rate of DOX with increasing NC payloads (Fig. 3G). DOX (37.4%) was released through 100 nM NC assembly, while 81.9% DOX was released at 2 μM NCs within 30 min, which strongly demonstrated that the multisite trans-cleavage of RNSs could facilitate DOX release. To further examine the stability of RNCOs in intracellular delivery, the serum stability of RNCOs was assessed by gel electrophoresis (fig. S11) after treatment with 10% fetal bovine serum (FBS) at 37°C. After 10 hours of incubation, the surface coating RNCOs showed no obvious degradation—indicating stronger resistance to nuclease attack than RNSs. The improved stability of RNCOs could be attributed to their condensed and hierarchical structures, implying the favorable suitability for in vivo applications.

Live cell hierarchical self-uncloaking properties of RNCOs

With the fundamental working principle of hierarchical self-uncloaking demonstrated in vitro, we next examined its utility for live cell synergistic treatment. U87 glioma cell lines, U87-EGFP (enhanced green fluorescent protein), U87-EGFRvIII, and U87-EGFP-EGFRvIII cells with stable expression of EGFP, EGFRvIII, and simultaneous expression of EGFP and EGFRvIII, respectively, were constructed by lentivirus infection; the cytomegalovirus promoter was used for driving gene expression in all these stably transfected lines (fig. S12). To demonstrate the feasibility of using RNCOs for subsequent tumor-targeting delivery, we systematically compared the fluorescence signal of RNCOs and NCs anchored on nontargeted RNSs (rRNCOs) with random sequences in EGFR-positive U87-EGFRvIII cells and EGFR-negative MCF-7 cells using CLSM and flow cytometry (FCM; fig. S13) (44). As anticipated, RNCOs presented strong fluorescence signals in U87-EGFRvIII cells (72%), whereas only a faint fluorescence was observed in MCF-7 cells (21%). U87-EGFRvIII and MCF-7 cells showed almost similar fluorescence signals after coincubation with rRNCOs (15 and 12%, respectively), indicating that the internalization of RNCOs relied on EGFR-mediated endocytosis. At the same time, we designed siEGFR to knock down EGFR expression in U87-EGFRvIII cells to observe the uptake of RNCOs (fig. S14). In the siEGFR group, the fluorescence signal was notably reduced, while the mismatch sequence treatment group (siNC) had almost no difference from the control group, confirming the efficient identification of EGFR Apt toward EGFR. In addition, cell counting kit-8 (CCK-8) analysis revealed that the therapeutic effect of RNSs-D was markedly enhanced compared with DOX-encapsulated nontargeted RNSs (rRNSs-D) on U87-EGFRvIII cells (fig. S15), while they showed no obvious therapeutic effect on MCF-7 cells, further demonstrating the potential of Apt in tumor-targeted delivery for precision oncology. To verify whether RNCOs were internalized into cells through endocytosis, confocal imaging was conducted to explore subcellular colocalization. RNCOs colocalized well with lysosomes, as revealed by the large overlap of RNCOs (green for FAM) and lysosomes (red for LysoTracker Red; fig. S16). The kinetic images describing the cellular uptake of RNCOs showed that the green fluorescence from the RNCOs largely overlapped with the red fluorescence of lysosomes in the first 30 min. Notably, however, there was no more overlap after 60 min of incubation. The reduction of Pearson’s colocalization coefficient from 0.90 to 0.27 showed that RNCOs gradually separated from lysosomes over time, demonstrating successful lysosomal escape of RNCOs (fig. S17).

Encouraged by the efficient cellular internalization and programmable hierarchical self-uncloaking of RNCOs-D in vitro, we next explored the spatial-controlled Cas13a cis-cleavage activation by examining the gene silencing effect of RNCOs in live cells. We selected EGFP crRNA and U87-EGFP cells as a model and then quantified the mean FI (MFI) of U87-EGFP cells incubated with RNCOs and nRNCOs by CLSM (fig. S18, A and B), observing notable fluorescence quenching after treatment with the RNCOs. Next, we measured EGFP mRNA levels in the RNCO-transported U87-EGFP cells via quantitative reverse transcription polymerase chain reaction (qRT-PCR), and an expected decrease in EGFP mRNA levels after RNCO treatment, but not with nRNCO treatment, was observed, further confirming the efficient gene silencing of RNCOs without the first self-uncloaking blockage (fig. S18C). The FCM analysis also confirmed that RNCOs reduced the fluorescence signal of EGFP by ~66.7%, whereas nRNCOs had a negligible quenching effect compared with the control group (fig. S18D), attributing to the locked first self-uncloaking. Continuous expression of EGFRvIII in GBM cells may result in tumor initiation and recurrence. Therefore, EGFRvIII crRNA and U87-EGFP-EGFRvIII cells were chosen to investigate the gene silencing capacity through cis-cleavage of EGFRvIII mRNA (Fig. 4A). Immunofluorescence microscopy (Fig. 4B) and MFI analysis (Fig. 4C) confirmed that RNCOs, not nRNCOs, inhibited EGFRvIII expression in U87-EGFP-EGFRvIII cells, consistent with the FCM results (Fig. 4D). To further illustrate this point, we collected total RNA from U87-EGFRvIII cells after various treatments with RNCOs and nRNCOs for qRT-PCR analysis (Fig. 4E), indicating the significant inhibition of EGFRvIII at the mRNA level. The efficient mRNA silencing caused by RNCOs strongly demonstrated that the spatial-controlled first self-uncloaking of RNCOs with the activity regulation of Cas13a RNP was efficient, while the second self-uncloaking had been initiated.

Fig. 4. Validation of hierarchical self-uncloaking property of RNCOs in live cell.

(A) Schematic illustration of target-specific cis-recognition and cleavage of EGFRvIII mRNA for gene silencing and trans-cleavage of RNSs for burst release of DOX. (B) Immunofluorescence (20×) of EGFP and EGFRvIII expressions in U87-EGFRvIII-EGFP cells treated with PBS (control), nRNCOs, and RNCOs, respectively. Scale bar = 100 μm. (C) Quantitative analysis of EGFRvIII expression from (B). (D) FCM and (E) qRT-PCR analysis of EGFRvIII expression after the same treatment with (B). (F) Quantitative analysis of DOX release from (G) and (H). (G and H) CLSM of DOX release profiles from RNCOs-D in U87 and U87-EGFRvIII cells. Scale bar = 25 μm. Pearson correlation coefficients are shown in the co-localization scatterplots, indicating the co-localization of Cas13a with DOX. Data are means ± SD (n = 3). Statistical significance was determined by unpaired two-tailed Student’s t-test (***P < 0.001).

After demonstrating the spatial-controlled Cas13a cis-cleavage properties of EGFP and EGFRvIII mRNA in live cells, the cis-recognition and trans-cleavage of RNSs at the second self-uncloaking was further surveyed in U87 and U87-EGFRvIII cells after incubation with RNCOs-D (Fig. 4A). For U87 cells, RNSs in RNCOs-D could not be trans-cleaved because of the lack of EGFRvIII mRNA–activated cis-recognition. DOX was encapsuled in RNSs and overlapped with Cas13a during the whole incubation, as reflected by the Pearson correlation coefficient from 0.83 to 0.61 (Fig. 4G). As for U87-EGFRvIII cells, DOX overlapped with Cas13a greatly in the first 10 min, with a Pearson correlation coefficient of 0.94. After a prolonged incubation of 30 min, the fluorescent signal of DOX was gradually separated from that of Cas13a with a reduced coefficient of 0.42, attributing to the trans-cleavage of RNSs and DOX release from RNCOs-D. Last, DOX absolutely separated from Cas13a into cell nucleus after 60-min incubation with a low coefficient of 0.18 (Fig. 4H). Above all, an obvious colocalization of DOX and cell nucleus occurred in U87-EGFRvIII cells instead of U87 cells after 60-min incubation, indicating that the trans-cleavage of RNSs stimulated by EGFRvIII mRNA–activated cis-recognition caused an efficient burst release of the chemotherapeutic agents (Fig. 4F).

Therapeutic effect and antitumor mechanism in vitro

After demonstrating the hierarchical self-uncloaking of RNCOs-D in various U87 glioma cells, we conducted a CCK-8 assay to assess the therapeutic performance of RNCOs-D for potent chemotherapy and gene silencing in U87-EGFRvIII cells (Fig. 5A). RNCOs presented an obviously higher dose-dependent therapeutic effect than bare RNSs (34.8% in RNCOs versus 97.6% in RNSs at RNSs of 5.6 μg/μl), indicating that Cas13a RNP payloads significantly suppressed U87-EGFRvIII cell proliferation, whereas RNSs had superior biocompatibility. To verify the therapeutic efficacy of DOX encapsulation into RNCOs, the cellular viability was evaluated after incubation with RNCOs-D of different concentrations. Less than 58.0% of cells survived at low concentration [RNSs (0.35 μg/μl), 1 μM DOX, and 0.5 μM Cas13a RNP]. Furthermore, less than 18.6% of cells survived at high concentration [RNSs (5.6 μg/μl), 8 μM DOX, and 4 μM Cas13a RNP], suggesting that DOX encapsulation enhanced the therapeutic efficacy compared to RNCOs (64.7 or 34.8% accordingly) and exhibited remarkably high cytotoxicity toward U87-EGFRvIII cells. The ascendancy of Cas13a gene silencing was confirmed by the superior treatment effect of RNCOs than RNSs-D. At the low concentration of RNSs, the tumor inhibition rate of RNSs-D [RNSs (0.35 μg/μl) and 1 μM DOX] reached ~8%, while that of RNCOs [RNSs (0.35 μg/μl) and 0.5 μM Cas13a RNP] reached ~35% (RNCOs achieved pure gene editing). At the high concentration of RNSs, the tumor inhibition rate of RNSs-D [RNSs (5.6 μg/μl) and 8 μM DOX] reached ~58%, while that of RNCOs [RNSs (5.6 μg/μl) and 4 μM Cas13a RNP] reached ~66%, showing that the antitumor effect of gene silencing of Cas13a RNP in RNCOs was better than that of chemotherapy of DOX in RNSs-D. nRNCOs-D and RNSs-D exhibited comparable antitumor effects, probably because they both lack first self-uncloaking; therefore, the subsequent accelerated release of DOX cannot be induced, resulting in good cell viability (40.6 and 42.0%, respectively). This was further validated by the higher therapeutic performance of U87-EGFRvIII cells treated with RNCOs-D (18.6%) compared to nRNCOs-D (40.6%), in which the first self-uncloaking was lacking. Moreover, to determine whether cell apoptosis was caused by hierarchical self-uncloaking of RNCOs-D resulting in step-wise cargo release, the cytotoxicity toward the U87 cells was also evaluated by CCK-8 (fig. S19). Unlike the phenomena in U87-EGFRvIII cells, RNCOs exhibited negligible damage to U87 cells, while the same cytotoxicity was observed against U87 cells in RNSs-D, nRNCOs-D, and RNCOs-D treatments, which was attributed to the blocked second self-uncloaking stage. In addition, the calculated half-maximal inhibitory concentration of RNCOs-D was 1.6 μM equivalent DOX, significantly lower than 6.1 μM equivalent DOX value of RNSs-D. Generally speaking, RNCOs-D exhibited the strongest cytotoxicity to U87-EGFRvIII cells compared to RNSs, RNCOs, RNSs-D, and nRNCOs-D, attributed to the decaging Cas13a-mediated gene silencing and the concomitant trans-cleavage of Cas13a-accelerated drug release.

Fig. 5. Therapeutic effect and antitumor mechanism in vitro.

(A) Cell viability of U87-EGFRvIII cells after incubation with RNSs, RNCOs, RNSs-D, nRNCOs-D, and RNCOs-D. The concentration of DOX (1, 2, 4, 6, and 8 μM) and NCs (0.5, 1, 2, 3, and 4 μM) increased as the RNS concentration increased. Data are means ± SD (n = 3). Statistical significance was determined by unpaired two-tailed Student’s t test (***P < 0.001). (B) Western blot (WB) analysis of apoptosis protein levels in U87-EGFRvIII cells. (C) CLSM of U87-EGFRvIII cells costained with calcein AM/PI after treatment with PBS, DOX, rRNCOs-D, RNSs-D, nRNCOs-D, and RNCOs-D. Scale bar, 100 μm. (D) Cell apoptosis and necrosis analyzed by annexin V–APC/DAPI (4′,6-diamidino-2-phenylindole) staining after treatment of PBS, RNSs, RNCOs, RNSs-D, nRNCOs-D, and RNCOs-D, respectively (from left to right). (E) Schematic illustration of antitumor mechanism. (F) WB analysis of EGFRvIII protein levels. RNCOs-1 refers to 500 nM equivalent Cas13a RNP; RNCOs-2 refers to 1 μM equivalent Cas13a RNP. (G) RNA denaturing gel electrophoresis examining total RNA integrity.

Moreover, the levels of apoptosis-related proteins were investigated to assess the therapeutic effects of RNCOs-D (Fig. 5B). With our hierarchical self-uncloaking strategy, RNCOs-D could activate caspase family proteins and induce the up-regulation of caspase-3 and caspase-9, down-regulation of Bcl-2, and up-regulation of Bax in U87-EGFRvIII cells, exhibiting optimal therapeutic performance among treatments with DOX, rRNCOs-D, or nRNCOs-D. In addition, live/dead cell double staining with calcein AM and propidium iodide (PI) further demonstrated the detailed therapeutic effects (Fig. 5C). In CLSM, the red fluorescence of PI was observed after treatment with RNCOs-D compared to DOX, rRNCOs-D, RNSs-D, and nRNCOs-D, validating that U87-EGFRvIII cells were effectively eradicated via hierarchical self-uncloaking property of RNCOs-D. Different indices of cell apoptosis were observed in U87 and U87-EGFRvIII cells by FCM (Fig. 5D), further demonstrating the precise operation of the hierarchical self-uncloaking strategy. In addition, the combination index (CI) of RNCOs (gene therapy) and DOX (chemotherapy) was 0.53, verifying the synergistic effect (45). These results support our hypothesis that Cas13a plays an indispensable role in synergistically enhanced combined therapy.

In view of the therapeutic effects on U87-EGFRvIII cells, we further explored the antitumor mechanism observed in RNCOs-D treatment. Cas13a/crRNA cis-cleaved the EGFRvIII mRNA and trans-cleaved ribosomal RNA, thereby inhibiting EGFRvIII and related protein expression, and finally enhancing the sensitivity to chemotherapeutics (DOX). We surmised that these three aspects worked together to induce GBM cell apoptosis (Fig. 5E). EGFRvIII protein down-regulation and knockdown of mRNA were confirmed by Western blot (WB; Fig. 5F) and qRT-PCR (Fig. 4E and fig. S20) analyses. The protein and mRNA levels of EGFRvIII were reduced by ~70% and ~75% compared to the control group, respectively. In qRT-PCR analysis, the C_t value of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) lost consistency accompanied by the increasing C_t value of EGFRvIII and GAPDH, indicating that RNCOs presented potent EGFRvIII gene expression inhibition and collateral effect in U87-EGFRvIII cells. In addition, gene silencing with obvious collateral efficiency at 500 nM Cas13a/crRNA was achieved (Fig. 5F and fig. S21), far lower than the currently required concentration for gene knockdown (46), implying that the present intelligent RNCOs-D has excellent bioresponsiveness and highly efficient gene silencing capabilities. RNA-denaturing gel electrophoresis confirmed the occurrence of collateral effects in U87-EGFRvIII cells, and ribosomal RNA cleavage was observed after treatment with RNCOs (Fig. 5G). As a control, nRNCOs showed neither collateral RNA cleavage nor target gene silencing, implying that the efficient tumor suppression ability of RNCOs was mainly due to the gene silencing and collateral cleavage of host RNAs by CRISPR-Cas13a system. Considering that overexpression of EGFRvIII induces severe chemoresistance, Cas13a-mediated EGFRvIII gene silencing undoubtedly enhances sensitivity to chemotherapeutics, as verified by CI value (0.53). These results indicated that the inhibition of EGFRvIII gene expression had notable antitumor effects and enhanced sensitivity to chemotherapy, leading to an effective synergistic gene therapy/chemotherapy for GBM.

Therapeutic efficacy in vivo

Inspired by the excellent therapeutic effects of the RNCOs-D nanoplatform in vitro, we further explored its tumor-specific ablation capacity in vivo by constructing a U87-EGFRvIII xenograft animal model (Fig. 6A). The in vivo distribution and tumor accumulation of this nanoplatform were verified using U87-EGFRvIII tumor-bearing Balb/c nude mice after tail vein injection with either Cy5-labeled RNCOs or rRNCOs. Cy5 fluorescence was observed throughout the whole body 2 hours after intravenous injection of RNCOs, and the maximum signal was observed at the tumor site after 6 hours (Fig. 6B). Quantitative analysis showed that the amount of RNCOs at the tumor site was ~2.9-fold higher than that in rRNCOs 6 hours after injection (Fig. 6C). The intratumoral fluorescence signal was maintained even at 24 hours after injection; however, no apparent fluorescence signal was observed in the tumor after rRNCO treatment, implying tumor-specific accumulation and retention of RNCOs with Apt incorporation. After 28 hours of monitoring, the mice were euthanized, and ex vivo fluorescence imaging was performed on the tumor and main organs, including the heart, spleen, liver, lung, and kidney. Mice administered with RNCOs showed Cy5 fluorescence in the tumor site, which was ~2.2-fold higher than that in rRNCOs, further demonstrating efficient tumor targetability without off-target side effects (fig. S22). Distinct fluorescence was observed in the liver and kidneys, which may be attributable to absorption by the mononuclear phagocyte system and renal excretion (47). Before performing the in vivo therapeutic evaluation, it was first necessary to investigate the safety profile of RNCOs through hematoxylin and eosin (H&E) staining. After 18-day treatments with phosphate-buffered saline (PBS), DOX, rRNCOs-D, RNSs-D, nRNCOs-D, and RNCOs-D, histological samples of the excised main organs (heart, liver, spleen, lung, and kidney) were collected for pathological evaluation. With Apt targeting, no obvious organ damage and pathological changes were observed in each treatment group (fig. S23). Besides, U87-EGFRvIII tumor-bearing Balb/c nude mice were sacrificed at 0, 6, 12, and 18 days after injection to collect blood samples for hematology and blood biochemical analyses. Notably, the liver and kidney function indices of the mice treated with RNCOs-D were normal, and a routine blood examination also demonstrated that RNCOs-D exhibited nontoxic effects on the Balb/c mice during the treatment period (fig. S24). These results confirmed the negligible toxicity of RNCOs-D toward normal tissues and the promise for practical application.

Fig. 6. Therapeutic efficacy in vivo.

(A) Administration schedule of RNCOs-D in the subcutaneous U87-EGFRvIII tumor model. (B) Dynamic biodistributions and (C) semiquantitative biodistributions of RNCOs and rRNCOs at different time intervals. (D) Body weight profile, (E) tumor growth curve, and (F) tumor weight after treatment with PBS, DOX, rRNCOs-D, RNSs-D, nRNCOs-D, and RNCOs-D, respectively. Data are means ± SD (n = 4). Statistical significance was determined by unpaired two-tailed Student’s t-test (*P < 0.05 and ***P < 0.001). (G) Immunohistochemical analysis for changes of EGFRvIII expression in harvested tumors followed by identical treatment in (D). Brown indicates the target protein, and blue is from the nucleus staining. Scale bar, 50 μm. (H) H&E, TUNEL, and Ki-67 analysis of tumor after the same treatments in (D). Scale bars, 100, 100, and 50 μm, respectively.

To test the in vivo therapeutic performance of RNCOs in U87-EGFRvIII xenograft animal models, tumor-bearing mice were randomly divided into six groups (n = 4 per group) and intravenously injected with PBS, DOX, rRNCOs-D, RNSs-D, nRNCOs-D, and RNCOs-D. During the 18-day treatment period, the body weight and tumor volume of the mice were monitored every 3 days (Fig. 6, D and E). No conspicuous body weight variations and abnormalities were found in any of the treatment groups, indicating the negligible biological toxicity of these systems. DOX administration slightly inhibited tumor growth compared to PBS, owing to its chemotherapeutic properties. A mildly higher suppression was observed in RNSs-D than in DOX and rRNCOs-D, indicating that Apt ensured specific aggregation of therapeutic agents and thus enhanced the therapeutic effect. Furthermore, nRNCOs-D had tumor suppression capacity comparable to RNSs-D because of the first self-uncloaking impediment, limiting spatial-controlled gene silencing and drug release. As expected, RNCOs-D presented with the maximum tumor inhibition effect, attributed to the propitious hierarchical self-uncloaking to achieve the synergistic effects of chemotherapy and gene silencing. In addition, monitoring the location of the excised tumors throughout the treatment period confirmed that the mice with RNCOs-D treatment had the smallest tumor size (fig. S25) and the lightest tumor weight (Fig. 6F). Consistently, RNCOs-D caused the highest tumor growth inhibition (TGI) rates (~78%), intuitively proving that tumor growth could be significantly restrained through synergistic gene therapy/chemotherapy (fig. S26). After the indicated treatments, we evaluated the relative EGFRvIII protein levels in tumors by immunohistochemical (IHC) staining. As anticipated, EGFRvIII expression was moderately inhibited in nRNCO-treated tumors because Cas13a was cloaked inside NCs. A potent down-regulation of EGFRvIII expression was detected after treatment with RNCOs-D, because the intratumorally uncloaked Cas13a mediated the effective silencing of the EGFRvIII gene (Fig. 6G).

Last, we conducted histopathological analysis via H&E staining, immunofluorescence via terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labeling (TUNEL) assay, and Ki-67 staining of the tumor sections harvested from each treatment group (Fig. 6H). H&E staining showed much more necrotic regions in the tumors from the RNCOs-D group than from other groups. The apoptotic index was the highest (green in the TUNEL images), and the proliferative index was the lowest (fewest brown nuclei in the IHC staining with Ki-67) in the RNCOs-D group, further validating that RNCOs-D enhanced tumor apoptosis with synergistic gene therapy/chemotherapy effects. Together, the multifunctional RNCOs-D–mediated therapeutic Cas13a delivery and hierarchical self-uncloaking strategy enabled Apt-directed tumor-specific accumulation and efficient on-site Cas13a activation, and consequently enhanced the anticancer effect.

DISCUSSION

In summary, we constructed a hierarchical self-uncloaking RNCOs-D with high-loading, self-excising, and tumor-targeting properties for spatial-controlled genome editing and precise cancer therapy. The RNS frame was built through a versatile RCT strategy, while pH-responsive NCs were used to encapsulate the Cas13a RNP. Briefly, the intelligent bioinspired RNCOs-D was formed via electrostatic force–induced RNSs combined with NCs. For targeted delivery, the powerful CRISPR-Cas13a system was temporarily shielded by acid-degradable thin polymers until specific cellular internalization with the help of a tandemly assembled EGFR Apt sequence. The tumor intracellular acidic microenvironment then induced self-uncloaking of NCs to allow spatial-controlled gene silencing (by the intact CRISPR-Cas13a system) and drug release (by Cas13a trans-cleavage after cis-recognition of the mRNA target). Thus, RNCOs-D could be disassembled in a controlled manner via the collateral effect in specific cancer cells, which is of great significance for precision medicine. In this design, the dual roles of Cas13a (in structural transformation and therapeutic functions) were fully exploited to maximize the therapeutic effects of the nanocomplex. Meanwhile, RNSs also played dual roles: a gene silencing/chemotherapeutic delivery platform and a random digestion substrate for Cas13a collateral reaction for programmable drug release. To the best of our knowledge, the programmable RCT-based structures as a substrate for Cas13a trans-cleavage nuclease realized for in vivo on-demand drug administrate and ultimately achieving synergistic therapy have not yet been infrequently reported. As an intelligent and self-excising delivery platform, the present hierarchical self-uncloaking system could be engineered with other therapeutic sequences as well as agents, and was anticipated to show attractive application prospects in personalized biomedicine and bioengineering.

One of the reasons why tumors are refractory is that cancer cells have the development of notable genomic instability, which may generate random mutations that allow cancer cells to actively evade attack and elimination by immune cells (48). Although CRISPR technology has made progress in cancer treatment, the tumor could evade Cas13 targeting by acquiring point mutations in the sequence targeted by the crRNA protospacer. This is where the CRISPR-Cas technology needs to be considered with caution.

Apart from this, in current studies, the collateral effect of Cas13 variants is complex and potentially multifactorial. For example, the collateral effect occurred in U87 and Hela cells, but not in human embryonic kidney (HEK) 293T, A549, and Huh7 cells, indicating that the extents of collateral effects differed by cell types. The degree of collateral effects is positively correlated with the target RNA expression levels. Higher mRNA target levels mean that more Cas13 effectors become activated, thereby giving more opportunity for trans-cleavage events. Nonetheless, it was also found that the sensitivity of some transcripts toward collateral effects was different, which might be attributed to the differences in transcript lengths, subcellular localization, or accessibility (49). The reason for these contradictory reports on whether Cas13 effectors have (12, 15, 49) or have no (8, 16, 50–51) obvious collateral effects in eukaryotic cells is unclear. Therefore, further research is necessary to clarify the cleavage pattern and mechanism of CRISPR-Cas13a in eukaryotic cells and reveal the underlying reasons for the transcript-specific differences.

In our design, RNCOs-D exploited the collateral effect of CRISPR-Cas13a system in EGFRvIII-overexpressing glioma cells, implementing the hierarchical self-uncloaking strategy and inhibiting tumor proliferation in vitro and in vivo. Nevertheless, regarding the feasibility of our strategy harnessed for other genes that are not as strongly expressed, further studies are also needed to elucidate the mechanism by which trans-cleavage is induced, making our delivery platform more applicable.

MATERIALS AND METHODS

Materials

Dulbecco’s modified Eagle’s medium (DMEM), RPMI 1640 medium, FBS, and bicinchoninic acid (BCA) protein assay kit were obtained from Thermo Fisher Scientific Inc. (USA). 11-Mercaptoundecanoic acid (11-MUA) was purchased from Sigma-Aldrich Inc. (St. Louis, MO, USA). 6× Loading buffer and 20–base pair DNA Marker were obtained from Takara Biotech. Inc. (Dalian, China). DOX, linear ssDNA template (L_Apt and L_Ran), T7 promoter primer (P), cholesterol-labeled complement strand to RMSs (C-Chol), EGFRvIII mRNA target (T), mismatched EGFRvIII mRNA target (mT), EGFP mRNA, FAM and BHQ1 group-labeled reporter strand (F-R-Q), and primers for qRT-PCR were all synthesized and purified from Sangon Biotechnology Co. Ltd. (Shanghai, China). All DNA and RNA oligonucleotide sequences are summarized in table S1. Bax rabbit antibody, Bcl-2 rabbit antibody, and EGFRvIII rabbit antibody were purchased from Cell Signaling Technology (Danvers, MA, USA). EGFR antibody (DH8.3)–EGFRvIII Mutant was purchased from Novus Biological (USA). T7 RNA polymerase, T4 DNA ligase, RNase inhibitor, ribonucleoside triphosphates (rNTPs) mix, and 10× NEBuffer 3.0 [1 M NaCl, 0.5 M tris-HCl, 0.1 M MgCl₂, 0.01 M dithiothreitol (DTT); pH 7.9] were obtained from New England Biolabs (Ipswich, MA, USA). The crRNAs used in vitro were designed and synthesized by Integrated Biotech Solutions (Shanghai, China). Human breast cancer cell line MCF-7 and Escherichia coli Rosetta2 (DE3) cells were purchased from the American Type Culture Collection (Manassas, VA, USA). Methylenebisacrylamide (bis), N,N,N′,N′-tetramethylethane-1,2-diamine (TEMED), acrylamide (AAm), ethylene glycol dimethacrylate (EGDMA), NHS, and N-(3-aminopropyl)-methacrylamide (APMAAm) were purchased from Macklin Biochemical Co. Ltd. (Shanghai, China). Puromycin, lentivirus containing EGFRvIII, and EGFP were purchased from GeneCopoeia (Chongqing, China). Human glioma cell line U87 cells and pET-28a-LwCas13a plasmid were obtained by Fenghui Biological Co. Ltd. (Hunan, China). CCK-8 assay kit was purchased from Dojindo Laboratories (Kumamoto, Japan). Annexin V–fluorescein isothiocyanate (FITC)/PI double staining kit, LysoTracker Red, and 4(PI er Redining kit, Lysoom (DAPI) were purchased from Beyotime Institute of Biotechnology (Shanghai, China). A PAGE gel fast preparation kit was purchased from Epizyme Biotech (Shanghai, China). Anti-rabbit immunoglobulin G (IgG) (H+L) DyLight 594 and Anti-mouse IgG (H+L)-FITC were purchased from EarthOx (USA). FAM-NHS ester was obtained from Xi’an ruixi Biological Technology Co. Ltd. (China). Paraformaldehyde universal tissue fixative (4%) was obtained from Biosharp Life Science (Beijing, China). Ultrafiltration centrifugal tube Amicon Ultra-0.5 [Molecular Weight Cut Off (MWCO) 30 kDa] and Amicon Ultra-15 (MWCO 50 kDa) were purchased from Millipore (USA). D-Tube Dialyzer Mini [Molecular Weight Cut Off (MWCO) 12 to 14 kDa] was obtained from G-Biosciences Technology Inc. (USA). Isopropyl-β-d-thiogalactopyranoside (IPTG) was purchased from Solarbio Life Science (Beijing, China). A calcein AM/PI double staining kit was purchased from Yeasen Biotechnology Co. Ltd. (Shanghai, China). Ni Sepharose column was supplied by GE Healthcare Life Sciences Inc. (USA). Female BALB/c nude mice were obtained from SiPeiFu Biological Co. Ltd. (Beijing, China). PBS consisted of 10 mM Na₂HPO₄, 1.75 mM KH₂PO₄, 137 mM NaCl, and 2.65 mM KCl (pH 7.2 to 7.6). Dulbecco’s PBS (D-PBS) consisted of 8.10 mM Na₂HPO₄, 1.47 mM KH₂PO₄, 136.89 mM NaCl, and 2.67 mM KCl. PBST buffer was prepared by adding Tween 20 (1 mg/ml) in PBS buffer. All aqueous solutions were prepared using ultrapure water (≥18 megohms, Milli-Q, Millipore).

Apparatus

UV-vis absorption spectra were recorded using a UV-3600 spectrophotometer (Shimadzu, Kyoto, Japan). CD spectra were recorded using the Chirascan Circular Dichroism Spectrometer (Applied Photophysics, London, UK). Features and morphologies of nanomaterials were imaged on JEM1200EX TEM (JEOL, Tokyo, Japan). qRT-PCR was performed and analyzed using the CFX connect software (Bio-Rad, USA). Hydrodynamic size and zeta potential of nanomaterials were measured by Zetasizer Nano ZS (Malvern, UK). Fluorescence spectra were monitored by a Cary Eclipse fluorescence spectrophotometer (Agilent Technologies, Palo Alto, CA). Confocal fluorescence micrographs of cells were acquired on a TCS SP8 CLSM (Leica, Germany). CCK-8 assay was measured with a reader (Thermo Fisher Scientific, USA). Concentrations of all RNA and DNA oligonucleotides were measured by a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, USA). WB images were performed on a ChemiDoc XRS+ imaging system (Bio-Rad, USA). PAGE, SDS-PAGE, and RNA-Denaturing Gel Electrophoresis analysis were imaged on a Bio-Rad ChemiDoc XRS (Bio-Rad, USA). The biodistribution of nanoconjugates was imaged on an IVIS Spectrum in vivo imaging system (PerkinElmer, USA). The FCM analysis was performed on a BD Influx flow cytometer (BD, USA).

Expression and purification of Cas13a

The LwaCas13a protein was overexpressed in E. coli Rosetta2 (DE3) cells by electrotransfection of the pET-28a-LwCas13a plasmid (fig. S4). Then, DE3 cells were cultured overnight in 4 ml of Terrific Broth (TB) growth media [tryptone (12 g/liter), yeast extract (24 g/liter), K₂HPO₄ (9.4 g/liter), KH₂PO₄ (2.2 g/liter)] at 37°C. The starter culture was transferred into 1 liter of TB for growth at 37°C and shaken at 220 rpm until an OD₆₀₀ (optical density at 600 nm) of 0.6. After treatment with IPTG (0.2 mM), the cells were incubated at 18°C for 20 hours to continuously produce Cas13a. After centrifugation (4000g, 30 min) at 4°C, the cells were resuspended in lysis buffer [tris-HCl (50 mM), NaCl (500 mM), imidazole (20 mM), and 5% glycerol] and sonicated for 1 hour at 25°C. The lysates were collected by high-speed centrifugation (12,000g, 30 min) at 4°C, and the supernatant was loaded into a Ni Sepharose column. Last, after elution with buffer [tris-HCl (50 mM), NaCl (500 mM), imidazole (120 mM), and 5% glycerol], Cas13a was dialyzed overnight in a store buffer [NaCl (600 mM), tris (1 M), and 5% glycerol; pH 7.5] and concentrated in an ultracentrifuge unit (MWCO 50 kDa). Purified Cas13a was quantified by BCA colorimetric protein assay and analyzed by 7.5% SDS-PAGE.

Self-assembly of RMSs via RCT

The phosphorylated L_Apt (0.6 μM) and P (1.2 μM) were mixed in DNA ligation buffer [tris-HCl (50 mM), MgCl₂ (10 mM), DTT (10 mM), and adenosine triphosphate (1 mM)] at a molar ratio of 1:2, and the mixed solution was annealed in a thermocycler at a temperature gradient from 95° to 25°C in ~2 hours. Then, the circular template was produced by incubating annealed products with T4 DNA ligase (1 U/μl) at 16°C for 16 hours. The circular template (0.6 μM) was mixed with T7 RNA polymerase (4 U/μl) and rNTP mix (2 mM) in the reaction buffer [tris-HCl (4 mM), MgCl₂ (10 mM), DTT (1 mM), and spermidine (0.2 mM); pH 7.9] containing RNase inhibitor (1 U/μl) and supplemented with additional MgCl₂ at a final concentration of 10 mM until a final volume of 40 μl was reached. RMSs gradually self-assembled via RCT when the mixture was incubated at 37°C for 10 hours. Then, the RCT process was inactivated by heating at 65°C for 10 min. The product was washed with nuclease-free water three times, followed by centrifugation (8000g, 10 min) to remove excess enzyme and RNA strands. The concentration of the RMSs was measured using a NanoDrop 2000 spectrophotometer.

Preparation of RNSs

RNSs were synthesized by hybridization of RMSs with C-Chol. C-Chol (370 μM in 50 mM tris-HCl buffer) was mixed with RMSs (5 μg) and then supplemented with MgCl₂ at a final concentration of 10 mM until a final volume of 20 μl was reached. NuPack software was used to predict the secondary structure of Apt after RMS condensation by cholesterol. The mixture underwent the same annealing conditions as those used for the preparation of RMSs, except for cooling to 4°C. The prepared RNSs were dispersed in PBS buffer (1×, 40 μl) and stored at −20°C until use. To encapsulate DOX, 20 μg of RNSs was mixed with 100 μl of DOX (1 mM) in D-PBS buffer and incubated at 25°C for 24 hours. Then, RNSs-D was pelleted at 7200g for 15 min. Last, the as-prepared RNSs-D was dispersed in D-PBS buffer (100 μl) and stored at −20°C until use.

Synthesis of NCs

The acid-responsive degradable NCs were synthesized by in situ free-radical polymerization. First, LwaCas13a (20 μM) was incubated with EGFRvIII crRNA (10 μM) in NaHCO₃ buffer for 10 min to preassemble the Cas13a RNP complex. Cas13a RNP (1 mg/ml, 200 μl) was mixed with AAm (100 mg/ml, 33.4 μl) and stirred at 4°C for 10 min. Then, APMAAm (100 mg/ml, 9.3 μl), a positively charged monomer, was added to the mixture and stirred for 10 min at 4°C. Subsequently, a pH-responsive cross-linking agent EGDMA (100 mg/ml, 21 μl) was added. Last, Ammonium persulphate (APS) (10 mg/ml in deoxygenated and deionized water) and TEMED were added to the mixture at 4°C for 2 hours to initiate free radical polymerization. The molar ratio of AAm/APMAAm/EGDMA was 27:3:3. The reaction solution was transferred into D-Tube Dialyzer Mini (MWCO 12 to 14 kDa) after the polymerization reaction and then dialyzed overnight in PBS buffer (10 mM, pH 7.4) to remove excess monomers, cross-linkers, and initiators. For the nondegradable nNCs, a nondegradable cross-linker bis was used instead of EGDMA, while other conditions remained identical. The prepared NCs were stored at −20°C until use. The Cas13a RNP cloaking level of NCs was determined by the BCA assay.

Assembly of RNCOs-D

RNSs and NCs were combined through electrostatic interactions to obtain the RNCOs. Briefly, RNSs (10 μg/ml) were incubated with NCs (20 μg/ml equivalent Cas13a) at 25°C for 1 hour. To encapsulate DOX, RNCOs (20 μg) were mixed with DOX (1 mM, 100 μl) in D-PBS buffer and incubated at 25°C for 24 hours. Then, RNCOs-D was pelleted at 7200g for 15 min, and the supernatant was aspirated to detect DOX fluorescence (λ_ex 470 nm, λ_em 593 nm). The amount of DOX payload into RNCOs was calculated by subtracting the amount of DOX in the supernatant and dividing it by the total amount of DOX. Last, RNCOs-D was dispersed in D-PBS buffer (100 μl) and stored at −20°C until use.

Characterization of RNCOs in vitro

To verify the successful synthesis of these conjugates (RNSs, RNSs-D, NCs, and RNCOs-D), eight sets of experiments were performed, including PAGE, TEM, SEM (scanning electron microscopy), DLS, Zeta, UV-vis, CD, and CLSM. For PAGE analysis, the mixture of each step of RNCO formation (P, L_Apt, circular template, RMSs, RNSs, and RNCOs) was mixed with 6× loading buffer, and electrophoresis was performed in 1× TBE buffer (2 mM EDTA and 89 mM tris-boric acid, pH 8.3) at 110 V for 50 min. Last, the gel was stained with Goldview and visualized using a Bio-Rad ChemiDoc XRS imaging system. To confirm that Cas13a RNP was encapsulated in NCs, the NCs were well decorated onto the RNS surface, as confirmed by TEM images. We used AuNP-labeled Cas13a for observing. Briefly, 11-MUA (100 μM, 25 μl) was slowly added to the AuNP solution (200 μl) and gently shaken at 25°C for 12 hours. NHS was then added for activation at 25°C for 12 hours to obtain 11-MUA–functionalized AuNPs (11-MUA-AuNPs). Then, 11-MUA-AuNPs (200 μl) were incubated with Cas13a (5 μM, 10 μl) for 12 hours; therefore, the surface-exposed amino groups of Cas13a first reacted with the NHS ester group of 11-MUA–functionalized AuNPs to afford AuNP-labeled Cas13a, which was encapsulated in NCs to further form AuNP-labeled RNCOs. For TEM imaging, NCs (5 μM, 5 μl) were dropped onto a TEM copper mesh sprayed with carbon film and dried for 30 min. After dyeing with 2% phosphotungstic acid solution (pH 7.0) for 2 min, the staining agent was removed by deionized water and observed after drying. To verify whether Cas13a was successfully loaded on RNSs, CD spectra of RNSs, NCs, and RNCOs were performed with nucleic acid concentration of 0.1 mg/ml and protein concentration of 0.2 mg/ml. RNSs, NCs, and RNCOs were diluted three times for UV-vis absorption spectra analysis. The hydrodynamic size and potential change of different nanoparticles were determined using Zetasizer (Malvern Panalytical, UK). To visualize the colocalization of RNSs and NCs, FAM-labeled Cas13a was used to represent the NCs, and DOX fluorescence encapsulated into RNSs was applied to denote the RNSs. First, Cas13a was adjusted to 1 mg/ml with 5 mM bicarbonate buffer (pH 8.3) and then FAM-NHS (5 mg/ml, dissolved in anhydrous dimethyl sulfoxide; 5 μl) was slowly added to the solution while stirring at 25°C for 12 hours. Excess FAM-NHS was removed by dialysis in PBS (10 mM, pH 7.4). RNSs-D was incubated with NCs as mentioned above. The mixture was dropped onto a glass slide and immediately observed via CLSM. For the FAM channel, the emission signal from 505 to 565 nm was collected under 488-nm excitation. For the DOX channel, the emission signal from 562 to 620 nm was collected under 552-nm excitation.

The first self-uncloaking process of NCs in vitro

Acid-stimulated NC degradation at the first self-uncloaking was evaluated by monitoring the cloaking/uncloaking activity variation of Cas13a in the caging/decaging state. Degradable NCs and nondegradable nNCs were preincubated in CH₃COONa buffer (50 mM, pH 5.4) or Hepes buffer (50 mM, pH 7.4) at 25°C for 1 hour and then resuspended in NEBuffer 3.0 (1×) after ultrafiltration purification using an ultracentrifuge unit (MWCO 50 kDa). The enzymatic activity of Cas13a was determined as previously described (8). In brief, preincubation products (1 μM), F-R-Q (125 nM), RNase inhibitor (1 U/μl), and the amounts of T (0, 50, 100, 200, 300, and 400 nM) were mixed in NEBuffer 3.0 (1×) at 37°C for 1 hour. The FAM emission signal at 518 nm was determined under 494-nm excitation.

The second self-uncloaking process of RNSs-D in vitro

The target-specific cis-recognition of LwaCas13a was evaluated by fluorescence signal recovery. Subsequently, LwaCas13a (20 nM) was incubated with crRNA (10 nM) against EGFRvIII in NEBuffer 3.0 (1×), followed by the addition of T or mT (1 μM), RNase inhibitor (1 U/μl), and F-R-Q (200 nM) at 37°C for 1 hour. The FAM emission signal at 518 nm was determined under 494-nm excitation. The degradation of RNSs caused by multiturnover trans-cleavage of Cas13a was then assessed by PAGE and TEM. Cas13a was preincubated with crRNA at a molar ratio of 1:2 to form Cas13a RNP, and then the Cas13a RNP complex was added into RNS (0.2 mg/ml, 5 μl) solution with or without T. TEM was performed after 12-hour incubation at 37°C. For PAGE analysis, RNSs (0.2 mg/ml, 5 μl) were incubated with Cas13a RNP (1 μM) and T (1 μM) in NEBuffer 3.0 (1×) at 37°C for different time intervals, and the incubation products were taken out at different intervals for PAGE analysis. Multisite trans-cleavage of RNSs activated by monosite cis-recognition of Cas13a RNP at the second self-uncloaking was verified by DOX release in the absence of NCs. Briefly, RNSs-D (0.2 mg/ml, 5 μl) was mixed with various concentrations of Cas13a RNP and T (1 μM) in NEBuffer 3.0 (1×). After incubation at 37°C, the mixture was centrifuged at 7200g for different intervals (0, 50, 100, 150, 200, 250, and 300 min), and the FI of the supernatant was measured to calculate the DOX release efficiency. The DOX emission signal at 593 nm was determined under 470-nm excitation.

Hierarchical self-uncloaking of RNCOs-D in vitro

Outer degradation of NCs and inner cleavage of RNSs as part of the hierarchical self-uncloaking strategy were validated by the DOX release behavior of RNCOs-D and nRNCOs-D under different pH conditions. Specifically, in the first self-uncloaking process, RNCOs-D and nRNCOs-D were preincubated in CH₃COONa buffer (50 mM, pH 5.4) or Hepes buffer (50 mM, pH 7.4) at 25°C for 1 hour, before changing the buffer with NEBuffer 3.0 (1×) by ultrafiltration purification using Amicon Ultra-0.5 (MWCO 30 kDa). In the second self-uncloaking process, the cleavage of RNSs (0.2 mg/ml, 5 μl) was initiated by adding T or mT (1 μM) in NEBuffer 3.0 (1×) at 37°C for different time intervals, and the DOX release kinetics were measured using a Cary Eclipse fluorescence spectrophotometer. Briefly, the mixture was removed and centrifuged at different intervals (0, 50, 100, 150, 200, 250, and 300 min), and the FI of the supernatant was measured to calculate the DOX release efficiency. The DOX emission signal at 593 nm was determined under 470-nm excitation. For PAGE analysis, RNCOs-D and nRNCOs-D were preincubated in CH₃COONa buffer (50 mM, pH 5.4) or Hepes buffer (50 mM, pH 7.4) at 25°C for 1 hour and then the buffer was changed with NEBuffer 3.0 (1×) by ultrafiltration purification using Amicon Ultra-0.5 (MWCO 50 kDa). Last, T (1 μM) was added to initiate the cleavage of RNSs at 37°C for different durations (30, 60, 90, 120, and 150 min). The mixture was verified by gel electrophoresis. RNase-free water was used for this assay.

Serum stability comparison of RNSs and RNCOs

RNSs and RNCOs (500 μg/ml) were dispersed in PBS buffer (1×, 20 μl, pH 7.4) containing 10% FBS and incubated at 37°C for 0, 2, 4, 6, 8, and 10 hours, respectively. The samples at each time gradient were removed, and the integrity of the conjugate was verified by 12% PAGE electrophoresis.

Preparation of stable U87-EGFP/U87-EGFRvIII/U87-EGFP-EGFRvIII cells

Lentivirus (fig. S12) transfection was performed according to the manufacturer’s protocol. Briefly, U87 cells were seeded at a density of 5 × 10⁵ cells/ml in 24-well dishes and cultured overnight. Then, the corresponding lentivirus [3.02 × 10⁸ Transduction Units (TU)/ml, 1.8 μl] was added into each well in line with multiplicity of infection (MOI) of U87 cells (MOI = 5) and incubated with cells at 37°C for 10 hours. Culture medium was then replaced with fresh medium and cultivated for 24 hours. The positive transfected cells were selected by puromycin for 1 week at 2 μg/ml and maintained at 1 μg/ml.

Cell culture

The human glioma cell lines U87, U87-EGFP, U87-EGFRvIII, and U87-EGFP-EGFRvIII and the human breast cancer cell line MCF-7 were cultured in DMEM supplemented with 10% (v/v) FBS, penicillin (100 U/ml), and streptomycin (100 μg/ml) in a 37°C incubator under 5% CO₂ and 90% humidity and maintained at densities between 5 × 10⁵ and 2 × 10⁶ cells/ml.

Validation of Apt binding specificity to EGFR

U87-EGFRvIII (EGFR-positive), MCF-7 (EGFR-negative), and U87-EGFRvIII cells (treated with siEGFR or siNC) were seeded in 24-well dishes (1 × 10⁵ cells per well) and cultured overnight. For specific cellular uptake analysis, FAM-modified C-Chol was used to hybridize with the RMSs. Cells were incubated with RNCOs (5 μg/ml) and rRNCOs (5 μg/ml) in culture medium for 2 hours. The cells were then harvested for FCM analysis or fixed with paraformaldehyde for CLSM observation after washing three times with PBS. The FAM emission signal from 505 to 565 nm was collected under 488-nm excitation.

Visualization of intracellular behavior of RNCOs

U87-EGFRvIII cells were seeded in 24-well dishes (1 × 10⁵ cells per well) and grown to 70% confluence before treatment. The cells were treated with RNCOs (5 μg/ml) for predetermined time intervals. The cells were then stained with LysoTracker Red (50 nM) at 37°C for 30 min. The cells were then washed twice with PBS buffer, fixed with paraformaldehyde, and stained with DAPI (10 μg/ml) for 5 min before visualization with CLSM. For the FAM channel, the emission signal from 505 to 565 nm was collected under 488-nm excitation. For the lysosome channel, the emission signal from 562 to 620 nm was collected under 552-nm excitation.

Monosite cis-recognition and cleavage of EGFRvIII mRNA

To evaluate EGFP gene silencing, U87-EGFP cells were seeded in 24-well dishes at a density of 5 × 10⁴ cells per well overnight before treatment. The cells were then treated with PBS, nRNCOs, and RNCOs (500 nM equivalent Cas13a) and incubated for 12 hours at 37°C. EGFP gene disruption was assayed using CLSM and FCM. qRT-PCR analysis of EGFP mRNA knockdown was performed together with EGFRvIII gene interruption.

U87-EGFRvIII-EGFP cells were seeded in 12-well dishes at a density of 1 × 10⁵ cells per well. PBS, nRNCOs, and RNCOs (500 nM equivalent Cas13a) were transfected into cells at 37°C for 12 hours. The cells were then incubated with EGFRvIII antibody and gently shaken overnight at 4°C, followed by incubation with a secondary antibody [anti-rabbit IgG (H+L) DyLight 594] for 1 hour. The cells were then washed three times with PBST buffer and fixed with paraformaldehyde for 15 min. Last, after washing with PBST three times, the cells were stained with DAPI for 5 min and visualized with CLSM immediately. The emission signal from 520 to 570 nm was collected by CLSM under 550-nm excitation. To assay the gene down-regulation effects by qRT-PCR, the treated cells were collected at 6 hours and the total RNA was extracted with TRIzol reagent. Then, the first-strand complementary DNA (cDNA) was synthesized from total RNA (2 μg) using a PrimeScript RT reagent kit. qRT-PCR was performed using SYBR Green Master Mix.

The following primers were used (15): EGFP, CCCGACAACCACTACCTGAG (forward) and GTCCATGCCGAGAGTGATCC (reverse); EGFRvIII, GGCTCTGGAGGAAAAGAAAGGTAAT (forward) and TCCTCCATCTCATAGCTGTCG (reverse); EGFR, CCATGCCTTTGAGAACCTAGAA (forward) and GAGCGTAATCCCAAGGATGTTA (reverse); GAPDH, TGCACCACCAACTGCTTAGC (forward) and GGCATGGACTGTGGTCATGAG (reverse).

Verification of hierarchical self-uncloaking of RNCOs-D

To validate the hierarchical self-uncloaking behavior of RNCOs-D in living cells, U87-EGFRvIII and U87 cells were treated with RNCOs-D (2 μM equivalent DOX) at 37°C for 15, 30, and 60 min to observe the intracellular DOX release. After washing twice with PBS and staining with DAPI for 5 min, the cells were immediately observed via CLSM. For the FAM channel, the emission signal from 505 to 565 nm was collected under 488-nm excitation. For the DOX channel, the emission signal from 562 to 620 nm was collected under 552-nm excitation. To evaluate the cleavage efficiency of the different treatment groups, the cells were seeded into six-well dishes at a density of 5 × 10⁵ cells per well overnight before treatment with PBS, nRNCOs, and RNCOs. After incubation for 3 hours, the culture medium was replaced with 2 ml of fresh medium containing 10% (v/v) FBS and cultured for another 5 hours. The cells were then collected and lysed using TRIzol reagent for total RNA analysis. The concentration of extracted RNA was adjusted to 500 ng/μl and determined using a NanoDrop 2000 spectrophotometer. For RNA denaturing gel electrophoresis analysis, the RNA samples (2 μl) were mixed with 6× loading buffer (1 μl), and RNase-free H₂O was added to a total volume of 10 μl. Agarose gel (1%) was prepared with agarose (0.24 g), 10× Mops buffer (2.4 ml), 37% formalin (1.26 ml), and distilled water (20.4 ml). After polymerization, the mixture was loaded onto each well, and the gel was run in 1× Mops buffer for 25 min at 125 V.

Evaluation of cell proliferation and apoptosis

MCF-7, U87, and U87-EGFRvIII cells were seeded in 96-well dishes (1 × 10⁴ cells per well) and cultured for 24 hours to reach 70% confluency. The cells were treated with a series of RNSs, RNCOs, RNSs-D, nRNCOs-D, and RNCOs-D for 24 hours. After the cells were washed with PBS buffer, culture medium (100 μl) containing CCK-8 reagent (10 μl) was added and incubated in the dark for another 2 hours. Absorbance was measured at 450 nm using a microplate reader. Cell viability was calculated using the following equation: cell viability (%) = (average OD value for treated cells/average OD value for control cells) × 100%. The synergistic effect of gene-silencing chemotherapy was evaluated using a CI analysis. The value of CI represents different meanings: CI = 1 indicates an addictive effect; CI > 1 indicates an antagonistic effect; and CI < 1 indicates a synergistic effect.

According to the standard procedure of an annexin V– Allophycocyanin (APC)/DAPI double staining kit, U87 and U87-EGFRvIII cells were seeded in six-well dishes (5 × 10⁵ cells per well) and treated with PBS, RNSs, RNCOs, RNSs-D, nRNCOs-D, and RNCOs-D. Then, the cells were harvested and stained with an annexin V–APC/DAPI double staining kit. Last, the results were obtained using FCM.

According to the standard procedure of the calcein AM/PI double staining kit, U87-EGFRvIII cells were seeded in six-well dishes (5 × 10⁵ cells per well) and treated with PBS, RNSs, RNCOs, RNSs-D, nRNCOs-D, and RNCOs-D for the live/dead staining analysis. The cells were incubated at 37°C for 24 hours, stained with calcein AM/PI double staining reagents, and imaged immediately via CLSM. For the calcein AM channel, the emission signal from 495 to 545 nm was collected by CLSM under 488-nm excitation. For the PI channel, the emission signal from 562 to 620 nm was collected by CLSM under 530-nm excitation.

For WB analysis, U87-EGFRvIII cells were seeded in six-well dishes (5 × 10⁵ cells per well) and cultured for 24 hours to reach 70% confluency. The cells were treated with PBS, DOX, RNCOs-D, rRNCOs-D, and nRNCOs-D for 24 hours and then lysed in a mixture of cold neutral radioimmunoprecipitation assay lysis and extraction buffer, and protease and phosphatase inhibitor cocktail (1×). After centrifugation (12,000g, 15 min) at 4°C, the supernatant was collected and the protein concentration was determined using a BCA protein assay kit. Each supernatant was mixed with 6× protein loading buffer, and the mixture was boiled at 95°C for 8 min. After electrophoresis on an SDS-PAGE gel, the protein bands on the separation gel were transferred to a polyvinylidene difluoride (PVDF) membrane at 300 mA for 60 min. The PVDF membrane was washed once with 1× TBST buffer. Tris buffered saline (TBS) consisted of 50 mM Tris-HCl, 3 mM KCl and 150 mM NaCl (pH 7.2 to 7.6). TBST buffer was prepared by adding Tween 20 (1 mg/mL) in TBS buffer. The membranes were then placed in skim milk and gently shaken at 25°C for 60 min, followed by washing twice with 1× TBST buffer. The bands were then incubated with primary antibody overnight at 4°C. The membranes were shaken at 25°C for 30 min and washed six times with 1× TBST (5 min each time). Last, it was incubated with secondary horseradish peroxidase–conjugated antibody, washed three times with 1× TBST buffer (10 min each time), and detected using a ChemiDoc XRS+ imaging system.

Animal model

A U87-EGFRvIII xenograft animal model was established by subcutaneously inoculating 2 × 10⁶ U87-EGFRvIII cells resuspended in PBS buffer (100 μl) into female BALB/c nude mice at the age of 4 to 8 weeks old. The body weight of the mice was 20 to 22 g. All procedures were in accordance with the Institutional Animal Use and Care Regulations authorized by the Model Animal Research Center of Chongqing Medical University (MARC). Animal ethics was approved by the Human Resources Committee of the First Affiliated Hospital of Chongqing Medical University (2022-K54).

Biodistribution of RNCOs in vivo

When the tumor volume reached 60 mm³, the mice were used to investigate the tumor targetability. U87-EGFRvIII tumor-bearing mice were injected with Cy5-labeled rRNCOs and RNCOs through the tail vein. At 0, 2, 4, 6, 8, 12, 24, and 28 hours, in vivo fluorescence microscopy was performed using an IVIS Lumina II in vivo imaging system. After in vivo imaging, the mice were euthanized, and the major organs (tumor, liver, heart, lung, spleen, kidneys, and intestines) were dissected and imaged.

Antitumor efficacy in U87-EGFRvIII models in vivo

When the tumor size reached ~100 mm³, the mice were randomly divided into six groups (n = 4 per group) and treated as follows: (i) PBS, (ii) DOX, (iii) rRNCOs-D, (iv) RNSs-D, (v) nRNCOs-D, and (vi) RNCOs-D. The injection volume of each solution was 100 μl at a DOX concentration of 1 mg/kg, except for the control group. Tumor volumes and body weights were monitored every 3 days after the first injection. The tumor volume was calculated according to the formula V = 0.5ab², where a and b are the long and short axes of the tumor (in millimeters), respectively. Relative tumor volume (RTV) = V_t/V₀, where V_t is the tumor volume at the end of one experimental period (treatment group) and V₀ is the tumor volume at the start of the experiment. TGI was calculated by the following formula: TGI = [1 − RTV (experimental group)/RTV (control group)] × 100%. After 18-day treatment, the mice were sacrificed, and the tumor tissues were imaged and weighed. The sections of major organs and tumors were stained with H&E, TUNEL, and Ki-67 for histopathological evaluation. The intratumoral EGFRvIII levels were assessed using IHC staining. Briefly, the harvested tumors were first placed in 4% paraformaldehyde. After the fixed tumor was dehydrated, embedded, and sliced, the sections were deparaffinized and soaked in 3% methanol hydrogen peroxide for 10 min at 25°C. After three washes with PBS buffer, the slices were submerged in citric acid buffer (0.01 M, pH 6.0) and then microwave-heated to boiling. After washing twice with PBS, the slices were blocked by incubation with goat serum for 20 min at 25°C. An anti-EGFR monoclonal antibody (DH8.3) was diluted 100 times with universal antibody diluent and added to the samples overnight at 4°C, followed by incubation with a secondary antibody (goat anti-mouse IgG/biotin) at 37°C for 30 min. Last, diaminobenzidine tetrahydrochloride was used to visualize immunoreactivities.

Statistical analysis

All data are represented as means ± SD (n = 3), and statistical differences between two groups were determined using unpaired two-tailed Student’s t test (*P < 0.05, **P < 0.01, and ***P < 0.001). Statistical significance was set at P < 0.05.

Supplementary Materials

This PDF file includes:

Figs. S1 to S26

Tables S1 and S2

View/request a protocol for this paper from Bio-protocol.