A transdermal CRISPR/Cas9 strategy using biodegradable microneedles to target YB-1 in melanoma Authors: Xiazhou Fu, Zhaoyi Chen, Ziru Zhang, Yufei Jiang, Xiaomeng Jiang, Yuxin Yang, Hongsheng Zhang, Xinyun Zou, Jingwen Luo Published: 2026 ABSTRACT Melanoma's aggressiveness and therapeutic resistance highlight the need for innovative strategies. This study developed a CRISPR/Cas9 ribonucleoprotein (RNP) system targeting the oncogenic YB-1 gene, combined with doxorubicin (DOX), delivered by chondroitin sulfate (CS)-based microneedles (MN) incorporating designed nanoparticles (CNPs). YB-1, a critical oncogene driving melanoma progression, metastasis, and drug resistance, was targeted to enhance therapy efficacy. In vitro experiments demonstrated that CNPs enhanced cellular uptake efficiency, achieving a dual-drug delivery rate of 74.4% and mediating 65.32% gene editing efficiency. The application of CNPs@DOX@RNP reduced tumor cell viability to just 3.51%, while simultaneously lowering off-target toxicity. In vivo C57BL/6 melanoma models showed significant tumor growth inhibition, increased apoptosis, and suppressed proliferation/angiogenesis. Following treatment with CNPs@DOX@RNP/MN, tumor weight exhibited a 92.56% reduction relative to the untreated control. Biocompatibility assays confirmed safety with minimal organ damage. This MN platform offers a synergistic, safe, and effective strategy for combined YB-1 gene editing and chemotherapy, addressing melanoma's therapeutic challenges. Introduction Clustered regularly interspaced short palindromic repeat - associated nuclease 9 (CRISPR/Cas9) system is composed of single guide RNA (sgRNA) and the Cas9 nuclease. [1,2]. Ribonucleoprotein (RNP) is a pre-assembled complex of Cas9 nuclease and sgRNA that directly executes DNA cleavage within cells, bypassing the need for cellular transcription and translation processes [3,4]. This complex produces the fastest genome-editing effect; moreover, it avoids the problems of mRNA degradation and plasmid DNA integration into the host genome [5]. However, the lack of efficient delivery vehicles for transporting RNP to specific cells or tissues hinders the translation of this biotechnology into clinical applications [6]. Polymeric nanoparticles (NPs) have been extensively studied [7,8] as attractive platforms for non-viral delivery of CRISPR/Cas9, owing to their unique properties such as high stability, versatility, and sustained release [9]. Lo et al. used porous NPs for RNP delivery and achieved efficient knockdown of target genes in vitro and in vivo [10]. Chae et al. designed multifunctional NPs with pH-sensitive PEG derivatives and targeting peptides to achieve efficient delivery of CRISPR/Cas9 and docetaxel to human TNBC MDA-MB-231 cells and tumor-bearing mice for the treatment of triple negative breast cancer [11]. However, NPs are also limiteded by finite delivery efficiency and susceptibility to off-target effects [12]. Therefore, enhancing the delivery efficiency and local concentration of NPs within tumor cells while reducing their off-target toxicity represents a critical challenge that urgently needs to be addressed. Biodegradable microneedle (MN) technology offers a compelling solution to these delivery barriers. As a painless transdermal system, MNs bypass the stratum corneum and hepatic first-pass metabolism, delivering drugs directly to the lesion site to maximize local concentration and patient compliance [[13], [14], [15]]. Although MNs are widely used for skin tumors like melanoma [16], their broader application has been restricted by limited drug-loading capacity [17]. To overcome this, integrating NPs into MNs has proven effective in enhancing chemotherapeutic delivery [18]. Recent studies have demonstrated that the combination of MNs with NPs achieves higher chemotherapeutic delivery efficiency than MNs alone for effective treatment of inflammatory skin diseases (ISD) [19]. Previously, We designed and synthesized a novel polymer polyethyleneimine-oleic acid (PEI-OA) for the combined delivery of monoclonal antibodies and chemotherapeutic drugs. The PEI-OA complex significantly enhanced the drug-loading capacity and encapsulation efficiency of the nanoplatform [20]. Similarly, Teng et al. utilized the PEI derivatives to effectively deliver siRNA into SK-HEP-1 cells, successfully inducing downregulation of luciferase reporter gene expression [21]. Building on this, we employ PEI-OA as the core nanocarrier, encapsulated within MNs to achieve the coordinated delivery of RNP complexes, which exerts potent therapeutic effects at an extremely low dose. We selected chondroitin sulfate (CS) over the commonly used hyaluronic acid (HA) [22]. CS is a sulfated glycosaminoglycan [23,24] and a fundamental component of the extracellular matrix in connective tissues [25,26]. Beyond its exceptional biocompatibility and controlled biodegradability, CS exhibits intrinsic anti-inflammatory, antioxidant, and antitumor properties [27]. These biological activities, combined with minimal immunogenicity, make CS a superior candidate for fabricating biodegradable MNs tailored for tumor therapy. Simultaneously, we modified the NP surface with CS to form CNPs, enabling enhanced targeting of the CD44 receptor on tumor cells. Y- box binding protein 1 (YB-1) [28] has been identified as a critical regulator of melanoma cell growth, differentiation, and oncogenic transformation [[29], [30], [31]]. As one of the most conserved proteins in evolution, YB-1 has long been considered a potential cancer target [32]. In normal cells, YB-1 is typically poorly expressed and performs cellular functions such as DNA repair, pre-mRNA splicing, and mRNA stability [33]. However, in cancer cells, YB-1 is phosphorylated by kinases like protein kinase B, 70 kDa ribosomal protein S6 kinase, and 90 kDa Ribosomal S6 kinase, leading to increased expression and nuclear translocation [34,35]. Evidence suggests that the nuclear localization and overexpression of YB-1 can predict poor outcomes in patients with over 20 different types of tumors. Li et al. reported a polyamine-modified thermosensitive hydrogel for transdermal delivery of CRISPR/Cas9 targeting YB-1 for effective gene therapy of melanoma. It was confirmed that the knockout of YB-1 could have a therapeutic effect on melanoma [36]. Therefore, we selected the YB-1 gene as the therapeutic target to evaluate its feasibility for melanoma treatment. Since chemotherapy remains the first-line treatment for inoperable cancers [37], combination therapy strategies are currently showing great promise [38]. In this study, we propose a combination therapy integrating CRISPR/Cas9-based YB-1 gene editing with DOX [39] based chemotherapy to target highly metastatic melanoma. Collectively, this study reports a biodegradable MN patch for the transdermal co-delivery of RNP and DOX (Fig. 1). The system utilizes a CS-based MN matrix reinforced with PEG6000, embedded with PEI-OA-based core-shell nanoparticles (CNPs). In this design, DOX is encapsulated in the hydrophobic core, while RNP complexes are adsorbed onto the cationic surface and stabilized by an outer CS coating. Upon application, the MNs mechanically penetrate the skin barrier and dissolve at physiological temperature, releasing CNPs. The surface CS facilitates CD44 receptor-mediated targeting and intracellular uptake in melanoma cells. Once internalized, the CNPs disassemble to release their payload: the RNP complex mediates YB-1 gene knockout in the nucleus, while DOX induces apoptosis via DNA intercalation. This novel MN-based strategy aims to overcome current delivery bottlenecks, offering a potent and precise therapeutic approach for melanoma.Fig. 1Schematic diagram of CNPs@DOX@RNP/MN, illustrating its preparation, in vivo delivery, and intracellular drug delivery mechanism. Schematic diagram of CNPs@DOX@RNP/MN, illustrating its preparation, in vivo delivery, and intracellular drug delivery mechanism. Materials & methods Expression and purification of Cas9 and Cas9-EGFP fusion proteins Primers were designed to construct the nuclear localization signal (NLS) contained pET28a-Cas9-6×His-2NLS and pET28a-Cas9-EGFP-6×His-2NLS recombinant plasmids to enhance the ability of protein to enter the nucleus [40]. The plasmids were transformed into E. coli BL21(DE3) competent cells. Protein expression was induced with 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) when the bacterial culture reached OD600 ≈ 0.8. Cells were collected by centrifugation, resuspended in phosphate-buffered saline (PBS), and lysed using a low-temperature ultra-high-pressure continuous flow cell disruptor (JNBIO, Guangzhou, China). The supernatant was collected via centrifugation and loaded onto Ni-NTA affinity chromatography columns. Cas9 proteins were purified by stepwise elution with imidazole solutions (100 mM, 200 mM, 300 mM). RNP preparation Two YB-1-targeting single-guide RNA (sgYB-1) sequences were designed. sgYB-1 sequences containing T7 promoters were obtained by PCR amplification followed by in vitro transcription using the Thermo Scientific™ TranscriptAid™ T7 High Yield Transcription Kit. The Cas9 protein was mixed with sgYB-1 in a molar ratio of 1:3, incubated at 37 °C for 10 min to form an RNP complex. NP preparation and characterization NPs were prepared via the thin-film hydration method. PEI-OA, Solutol® HS 15 (HS15) and DOX were mixed in ethanol at a 6.2:30:1 mass ratio. Organic solvents were removed by rotary evaporation at 37 °C, followed by hydration with 60 °C reverse osmosis (RO) water to obtain DOX loaded NPs (NPs@DOX). Particle size and zeta potential were measured using a Zetasizer Nano ZSP (Malvern Instruments, Worcestershire, UK). NPs@DOX and RNP were mixed at a 10:1 mass ratio and incubated at 37 °C for 10 min to form DOX and RNP loaded NPs (NPs@DOX@RNP). NPs@DOX@RNP and CS (M.W.: 50-80 kDa) were then combined at a 1:2 mass ratio and incubated for 10 min to yield CS-coated NPs@DOX@RNP (CNPs@DOX@RNP). HA, DOX, and RNP were mixed and incubated at room temperature for 10 min to prepare DOX and RNP loaded HA (HA@DOX@RNP). CS, DOX, and RNP were combined and incubated under identical conditions to obtain DOX and RNP loaded CS (CS@DOX@RNP). To confirm dual loading of DOX and RNP, RNP loaded NPs (NPs@RNP), CS-coated NPs@RNP (CNPs@RNP), CNPs@DOX@RNP, and HA@DOX@RNP were prepared. Particle size and zeta potential of all six formulations were analyzed using the Zetasizer Nano ZSP. Morphology of PEI-OA, CS@DOX@RNP, and CNPs@DOX@RNP was examined by 200 kV transmission electron microscope (TEM, 200 kV, Thermo, USA). In vitro release study Fluorescein isothiocyanate-labeled bovine serum albumin (FITC-BSA) replaced RNP for fluorescence quantification. PBS containing FITC-BSA (PBS@FITC-BSA), HA loaded with FITC-BSA (HA@FITC-BSA), CS loaded with FITC-BSA (CS@FITC-BSA), NPs loaded with FITC-BSA (NPs@FITC-BSA) and CNPs loaded with FITC-BSA (CNPs@FITC-BSA) (150 μL each, containing 40 μg FITC-BSA) were loaded onto 8-layer gauze. PBS containing DOX (PBS@DOX), HA loaded with DOX (HA@DOX), CS loaded with DOX (CS@DOX), NPs loaded with DOX (NPs@DOX) and CNPs loaded with DOX (CNPs@DOX) (1 mL each, containing 20 μg DOX) were placed in dialysis bags (MWCO 1 kDa). Samples were immersed in 40 mL PBS (pH 7.4) in amber wide-mouth bottles and incubated at 37 °C with shaking (100 rpm). At 1, 4, 8, 12, 24, 36, 48 and 72 h, 2 mL aliquots were collected and replaced with fresh PBS. FITC-BSA release was quantified fluorometrically at 470 nm excitation wavelength, while DOX release was measured at 525 nm excitation wavelength using SpectraMax M2e (Molecular Devices, USA). Cumulative release (%) was calculated based on standard curves:Cumulativerelease(%)=(Cn×40+∑i=1n‐1Ci×2)A×100Note: Cn: Drug concentration at time point n; Ci: Drug concentration of the sample i; A: The total amount of drug in the hydrogel. MN patch preparation and characterization MN patches were fabricated using silicone molds with pyramidal cavities (410 × 410 μm base, 800 μm height) arranged in an 11 × 11 array. Three different types of MNs were prepared: MNs made with HA as the main needle tip material (HA/MN), MNs made with CS as the main needle tip material (CS/MN), and MNs made with CNPs as the main needle tip material (CNPs/MN). In the absence of drug loading, the needle tip composition of HA/MN consists of 10 wt% HA and 1 wt‰ PEG6000, whereas the tips of CS/MN and CNPs/MN are composed of 50 wt% CS and 5 wt‰ PEG6000. The rest of the MN is ultra-pure water, and the drug in the drug-loaded MN is also dissolved in water. The fabrication process of MN patch involves the following steps. 150 μL MN solution was evenly spread over the mold. Vacuum degassing ensured cavity penetration, followed by centrifugation at 2000 × g for 2 min. The mold was rotated 180° and centrifuged again. After drying at 4 °C for 1 h, 500 μL 10% HA backing solution was added. Complete drying occurred at 4 °C for 48 h before demolding. For mechanical strength testing, depilated BALB/c mouse dorsal skin was pressed with HA/MN without PEG6000 (HA/MN(−)), CS/MN without PEG6000 (CS/MN(−)), HA/MN and CS/MN for 3 min. After euthanasia, skin samples were stained with 0.4% trypan blue for 20 min, washed, and penetration sites were examined under an IX73 inverted microscope (Olympus, Japan). HA/MN and CS/MN were imaged using a high-resolution smartphone (Huawei Mate60 Pro). CS/MN microstructure was characterized by field emission scanning electron microscopy (FE-SEM; OPTON, Beijing, China). All animal protocols were in accordance with institutional and local ethical regulations and were approved by the Institutional Animal Care and Use Committee of Hubei University (No.20240102, Hubei, China). Fluorescence imaging of drug-loaded MNs Based on the requirements of visualization research, we employed DiD dye instead of DOX and used FITC-BSA to replace RNP. To visualize drug loading within MN needle tips, DiD and FITC-BSA loaded HA/MN (HA@DiD@FITC-BSA/MN), and DiD and FITC-BSA loaded CNPs/MN (CNPs@DiD@FITC) formulations were prepared (containing 40 μg FITC-BSA and 20 μg DiD per MN). After fabrication, fluorescence images were acquired using a two-photon high-speed ultra-high resolution imaging system (LSM980) (ZEISS, Germany). Evaluation of transdermal drug delivery behavior To track drug release dynamics post-MN administration, CNPs@DiD@FITC-BSA/MN patches, 20 μg DiD and 40 μg FITC-BSA per MN, were applied to BALB/c mouse skin. BALB/c mice were sacrificed at 0.5, 4, 8 and 24 h post-application. Skin sections were prepared and imaged using LSM980. Drug delivery and tumor targeting ability of MNs in vivo To evaluate the in vivo drug delivery capacity and tumor targeting ability of different MNs, three groups of MN formulations with two fluorescent labels were prepared: HA@DiD@FITC-BSA/MN, DiD and FITC-BSA loaded CS/MN (CS@DiD@FITC-BSA/MN), and CNPs@DiD@FITC-BSA/MN. The depilated BALB/c mouse back was divided into three areas, each treated with one of these formulations. Fluorescence intensity was monitored at 4, 24, 48, and 72 h using an IVIS Lumina III imaging system (PerkinElmer, MA, USA). Skin healing assessment Following depilation, MN patches were applied to BALB/c mouse dorsa. Skin samples collected at 0, 4, and 12 h were sectioned, stained with hematoxylin and eosin (H&E), and examined under an IX73 inverted microscope. Cell transfection studies To visually track the intracellular distribution of RNP, we expressed and purified Cas9 protein fused with green fluorescent protein (Cas9-EGFP) in E.coli. B16F10 cells were seeded into 24-well plates at a density of 5 × 104 per well and incubated for 24 h. Subsequently, B16F10 cells were treated with DiD and Cas9-EGFP PBS solution (PBS@DiD@Cas9-EGFP), DiD and Cas9-EGFP loaded HA matrix solution (HA@DiD@Cas9-EGFP), DiD and Cas9-EGFP loaded CS matrix solution (CS@DiD@Cas9-EGFP), DiD and Cas9-EGFP loaded CS and PEI matrix solution (CP@DiD@Cas9-EGFP, the amount of PEI is the same as the amount of PEI contained in NP), DiD and Cas9-EGFP loaded NPs (NPs@DiD@Cas9-EGFP), DiD and Cas9-EGFP loaded CNPs (CNPs@DiD@Cas9-EGFP), and DiD and Cas9-EGFP loaded Lipofectamine CRISPRMax™ (LIPO, Thermo Fisher, USA) (LIPO@DiD@Cas9-EGFP, the amount of LIPO is 5 μL/well). Each formulation contained 0.5 μg DiD and 4 μg Cas9-EGFP in DMEM. After 6 h incubation, cells were fixed with paraformaldehyde, stained with DAPI, and imaged by LSM980. Additionally, the aforementioned formulation was applied to B16F10 cells seeded in a 6-well plate. After incubation for 6 h, the cells were washed three times with PBS, then digested with trypsin, centrifuged at 900×g for 3 min, resuspended in 500 μL of PBS, and finally analyzed using a flow cytometer (Becton Dickinson, NJ, USA). Lysosomal escape assay B16F10 cells were seeded into 24-well plates at a density of 5 × 104 per well and incubated for 24 h. Subsequently, B16F10 cells were treated with Cas9-EGFP contained PBS (PBS@Cas9-EGFP), Cas9-EGFP loaded HA matrix solution (HA@Cas9-EGFP), Cas9-EGFP loaded CS matrix solution (CS@Cas9-EGFP) and Cas9-EGFP loaded CNPs matrix solution (CNPs@Cas9-EGFP) (4 μg Cas9-EGFP per group in lysosome-tracker-containing DMEM). After 2 h and 6 h, cells were fixed, DAPI-stained, and analyzed by LSM980. GFP knockout assessment To confirm the gene-editing capability of the delivered RNP, we designed and prepared a GFP-targeting single-guide RNA (sgGFP). Cas9 protein was assembled with sgGFP to form RNPGFP. 293T-GFP cells were seeded into 24-well plates at a density of 5 × 104 per well and incubated for 24 h. Subsequently, 293T-GFP cells were treated with PBS@RNPGFP, HA@RNPGFP, CS@RNPGFP, CNPs@RNPGFP and LIPO@RNPGFP in fresh DMEM. GFP knockout efficiency was detected at 0, 8, 24, and 48 h using IX73 inverted microscope. T7 endonuclease I (T7EI) assay B16F10 cells were seeded into 24-well plates at a density of 5 × 104 per well and incubated for 24 h. Subsequently, B16F10 cells were treated with RNP contained PBS (PBS@RNP), RNP loaded HA matrix solution (HA@RNP), RNP loaded CS matrix solution (CS@RNP), and CNPs@RNP for 24 h at 37 °C. Genomic DNA was extracted by Genome-Editing Mutation Detection Kit. Target regions were amplified by PCR. 200 ng PCR products were digested with T7EI at 37 °C for 30 min. Cleaved fragments were separated on 2% agarose gels using a GelDoc XR + system (Bio-Rad, USA). Mutation frequency was quantified by ImageLab 5.0 software. Cytotoxicity assay The cytotoxicity of the blank carrier against B16F10 and 293T cells was evaluated, while the cytotoxicity of the DOX-loaded formulation and the formulation containing both DOX and RNP against B16F10 cells was also measured. B16F10 and 293T cells were seeded into 6-well plates at a density of 1 × 104 per well and incubated for 24 h. Subsequently, cells in 96-well plates were treated with above formulations in 2% fetal bovine serum and 0.2% penicillin-streptomycin containing DMEM for 24 h. Methylthiazolyldiphenyl-tetrazolium bromide (MTT) solution (20 μL, 5 mg/mL) was added for 4 h. Formazan crystals were dissolved in DMSO (150 μL/well). The absorbance at 570 nm was measured using SpectraMax M2, with cell viability (%) calculated relative to 2% DMEM-treated controls using the following formula.CellViability(%)=(ODs‐ODb)/(ODc‐ODb)×100Note: ODs: Optical density (OD) of experimental group; ODc: OD of control group; ODb: OD of blank group. Colony formation assay B16F10 cells were seeded into 12-well plates at a density of 1 × 105 per well and incubated for 24 h. Subsequently, mixtures of PBS containing DOX and RNP (PBS@DOX@RNP), HA@DOX@RNP, CS@DOX@RNP, and CNPs@DOX@RNP in fresh DMEM were added. After 6 h incubation at 37 °C, cells were trypsinized and reseeded into 12-well plates. Culture was terminated when colonies in the PBS@DOX@RNP group reached ≥ 50 cells per clone. Cells were washed with PBS, stained with 0.5% crystal violet for 20 min, imaged, and colony areas were quantified using ImageJ 1.53c software. Apoptosis assay B16F10 cells were seeded in a 12-well plate at 1 × 105 cells/well. After 24 h, the cells were treated for 6 h with culture media containing PBS@DOX@RNP, HA@DOX@RNP, CNPs, CNPs@DOX, or CNPs@DOX@RNP. Cells were then washed twice with PBS, trypsinized, and collected (5 × 104–1 × 105 cells per group) into 2 mL EP tubes. After centrifugation at 1000×g for 5 min, supernatants were removed, and cell pellets were resuspended in 195 μL of binding buffer. Annexin V-FITC (5 μL) and PI (10 μL) were added, mixed, and incubated for 10-20 min at room temperature in the dark. Aliquots of stained cells were examined under an inverted fluorescence microscope for visualization, while the remainder was analyzed by flow cytometry. The PBS@DOX@RNP, HA@DOX@RNP, CS@DOX@RNP, CNPs@DOX@RNP, and DOX and RNP loaded LIPO (LIPO@DOX@RNP) groups were set up and subjected to the same procedures as described above. Finally, the cells in the EP tubes were transferred to culture dishes and resuspended by pipetting. Cell apoptosis was then analyzed using an IX73 inverted microscope. Anti-tumor efficacy Melanoma murine models were established by shaving dorsal hair and subcutaneously injecting 5 × 106 B16F10 cells to C57BL/6 mice. Twenty-five C57BL/6 mice were divided into control group, DOX loaded HA/MN (HA@DOX/MN), DOX loaded CS/MN (CS@DOX/MN), DOX loaded CNPs/MN (CNPs@DOX/MN), DOX and RNP loaded CNPs/MN (CNPs@DOX@RNP/MN). Each CNPs@DOX@RNP/MN delivered 20 μg Cas9 protein and 12 μg sgRNA. Each treatment group encapsulated 50 μg DOX per MN. The first treatment was given on the fifth day after tumor cell inoculation. During administration, the MN patch was pressed on the surface of the tumor for 3 min, and the MN patch was further fixed with medical tape to ensure continuous drug release. The mice were treated on day 5, 8, and 11. Body weight and tumor volume (V = 0.5×length×width2) were recorded every 2 days. On day 23, C57BL/6 mice were euthanized after tumors were photographed. Tumor weight was recorded. Blood were collected for hematological analysis. Tumors and major organs were collected for histological analysis. Tumor sections were subjected to immunohistochemical (IHC) staining for Ki67, CD31 and PD-L1, along with Terminal Deoxynucleotidyl Transferase mediated dUTP Nick-End Labeling (TUNEL), YB-1 Immunofluorescence staining and H&E staining. Stained sections were imaged by LSM710 confocal microscope (LSM710) (ZEISS, Germany). The main tissues and organs were sectioned and stained with H&E and T7EI assay to evaluate the toxicity of the preparation on normal tissues. Statistical analysis Difference between two groups was calculated by using student's t-test. When the comparison was performed among three or more groups, one-way analysis of variance was performed. The p value was denoted by * for p < 0.05, ** for p < 0.01, *** for p < 0.001 and **** for p < 0.0001. When there are two independent variables, two-way ANOVA with Tukey's multiple comparison test was performed. GraphPad Prism (Version8.02, GraphPad Software, Inc.) was used. Results and discussion Preparation and characterization of NPs PEI-OA conjugates were synthesized to form NPs (Fig. 2A), and confirmed successful synthesis via 1H NMR spectroscopy (Fig. S1). Subsequently, PEI-OA was combined with HS15 to encapsulate DOX, yielding NPs@DOX. The positively charged NPs@DOX electrostatically adsorbed negatively charged RNP and CS in a stepwise manner to form CNPs@DOX@RNP (Fig. 2B). Based on zeta potential (Fig. S2A) and particle size (Fig. S2B) measurements across different NPs:DOX mass ratios, an optimal ratio of 6.2:1 was selected for cationic NP preparation.Fig. 2Preparation and characterization of NPs. (A) Structure of PEI-OA used in the study. (B) Schematic diagram of NP structure. Zeta potential (C) and size measurement (D) of NPs@RNP, CNPs@RNP, CNPs@DOX@RNP, and HA@DOX@RNP. (E) The loading capacity of NPs and CNPs on RNP. (F) In vitro release profile of FITC-BSA from PBS@FITC-BSA, HA@FITC-BSA, CS@FITC-BSA, NPs@FITC-BSA, CNPs@FITC-BSA. (G) In vitro release profile of DOX from PBS@DOX, HA@DOX, CS@DOX, NPs@DOX, CNPs@DOX. (H) TEM images of PEI-OA, CS@DOX@RNP, and CNPs@DOX@RNP. All data are presented as the mean ± s.d (n = 3 biological replicates per group). Preparation and characterization of NPs. (A) Structure of PEI-OA used in the study. (B) Schematic diagram of NP structure. Zeta potential (C) and size measurement (D) of NPs@RNP, CNPs@RNP, CNPs@DOX@RNP, and HA@DOX@RNP. (E) The loading capacity of NPs and CNPs on RNP. (F) In vitro release profile of FITC-BSA from PBS@FITC-BSA, HA@FITC-BSA, CS@FITC-BSA, NPs@FITC-BSA, CNPs@FITC-BSA. (G) In vitro release profile of DOX from PBS@DOX, HA@DOX, CS@DOX, NPs@DOX, CNPs@DOX. (H) TEM images of PEI-OA, CS@DOX@RNP, and CNPs@DOX@RNP. All data are presented as the mean ± s.d (n = 3 biological replicates per group). Cas9 and Cas9-EGFP proteins were expressed and purified. Primers (Table S1) were designed to construct pET28a-Cas9-6×His-2NLS (Fig. S3) and pET28a-Cas9-EGFP-6×His-2NLS recombinant plasmids (Fig. S4). Protein expression was carried out in E.coli BL21 competent cells, followed by bacterial harvesting and protein purification. After extracting cDNA from B16F10 cells, the YBX1 (Table S2) fragment was amplified and ligated into the pUC18 plasmid (Fig. S5) for subsequent in vitro restriction enzyme digestion assays to screen sgYB-1. Two sgYB-1s were designed (Table S2), with T7 promoter-containing sequences obtained via PCR and in vitro transcription. RNP complexes were formed by incubating Cas9 protein with sgYB-1 at 37 °C for 10 min. In vitro cleavage assays (Fig. S6) confirmed the superior efficiency of sgYB-1-1, which was thus selected for subsequent experiments. Zeta potential and particle size of NPs@RNP, CNPs@RNP, CNPs@DOX@RNP, and HA@DOX@RNP are presented in Fig. 2C and D. CNPs@DOX@RNP exhibited a zeta potential of −5.37 mV. All formulations had diameters < 200 nm, with NPs@RNP, CNPs@RNP, and CNPs@DOX@RNP demonstrating progressively increasing sizes, confirming successful electrostatic binding of RNP and CS to NPs. The NPs were incubated in 10% serum-containing medium at 37 °C to evaluate their serum stability. As shown in Fig. S7, there was no significant change in the zeta potential, and the particle size remained stable (<200 nm) over 48 h. This indicates good colloidal stability, despite the relatively neutral zeta. The RNP encapsulation efficiencies of different NPs were determined using a Coomassie G-250 protein standard curve (Fig. S8A). The encapsulation efficiency of CNPs reached 78.88%, showing a significant difference compared to the 57.23% observed for NPs (Fig. 2E). This indicates that the incorporation of CS enhances the RNP encapsulation capacity of NPs. Consequently, this allows for a reduction in NP dosage during drug delivery, thereby minimizing potential toxicity to normal cells. We employed FITC-BSA, a commonly used model protein in research, as a surrogate for Cas9 to investigate the in vitro release rate. The physicochemical properties of Cas9 and BSA are presented in Table S3. To investigate the in vitro drug release behaviors of the NPs, we first established standard curves for FITC-BSA (Fig. S8B) and DOX (Fig. S8C) using data obtained from a microplate reader. Subsequently, the cumulative release rates of PBS@FITC-BSA, HA@FITC-BSA, CS@FITC-BSA, CNPs@FITC-BSA, PBS@DOX, HA@DOX, CS@DOX, and CNPs@DOX were calculated at various time points. The release of FITC-BSA from CNPs@FITC-BSA exhibited a gradual upward trend, reaching 72.83% at 72 h (Fig. 2F). Similarly, the release of DOX from CNPs@DOX showed a slow and sustained increase, achieving a release rate of 66.31% at 72 h (Fig. 2G). These results demonstrate that CNPs possess a certain sustained-release effect. We hypothesize that the CS coating on the NP surface likely retards the release of FITC-BSA and DOX from the NP interior to some extent. The study of in vitro release kinetics theoretically confirms that CNPs exert a sustained-release effect in delivering both DOX and RNP. By prolonging the drug release duration and achieving release amounts exceeding 65% for both agents, CNPs can maintain stable drug concentrations to enhance therapeutic efficacy. This sustained release profile reduces dosing frequency, improves patient compliance, and optimizes drug utilization, thereby lowering the total administered dose and drug load, which ultimately alleviates the body's overall metabolic burden and the risk of long-term accumulation. TEM imaging (Fig. 2H) revealed that CNPs displayed uniform spherical morphology and smaller dimensions. In contrast, CS@DOX@RNP exhibited ill-defined boundaries and aggregation tendencies, with no distinct NPs observed. A clear interface was evident between the NPs and the MN matrix in the CNPs@DOX@RNP group, suggesting structural integrity within the MN material. It was an advantageous feature for drug delivery applications. The amphiphilic, cationic NPs were selected as the nanoplatform core due to their favorable charge characteristics, biocompatibility [41], and capacity for RNP integration via electrostatic interactions. CS coating was applied to mitigate off-target toxicity and enhance in vivo safety. The fabrication of CNPs combines simplicity, environmental compatibility, and excellent scalability for industrial production. Preparation and characterization of MNs The backing material for MNs was selected as 10% HA. MN needle tips require sufficient mechanical strength and good biocompatibility. Monocomponent HA and CS MNs exhibit poor formability and brittleness, necessitating the addition of excipients. Based on existing research, PEG6000 possesses excellent biocompatibility and enhances sustained drug release, is widely used as a thickener [42]. We prepared HA/MN, CS/MN, HA/MN(−) and CS/MN(−). Mechanical strength testing on BALB/c mouse skin compared to HA/MN(−) and CS/MN(−) showed that HA/MN and CS/MN demonstrated adequate mechanical strength, uniform needle tip shape, and excellent transdermal delivery capability (Fig. S9). The morphology of the MN was observed after 1 min of dissolution at 32 °C (simulating skin surface temperature [43]) and 37 °C. As shown in Fig. S10, the MN tips exhibited significant dissolution at both temperatures, with only minor differences observed. This suggests that temperature has a limited effect on the release rate. Therefore, an application time of 2 min on mouse skin was set to ensure sufficient drug delivery from the MN tips. Photographs of HA/MN and CS/MN (Fig. 3A) show intact, conical needle tips conforming to mold specifications. For detailed needle tip observation, SEM imaging was performed (Fig. 3B). To visualize MN loading capacity of RNP and DOX, We replaced DOX with DiD and RNP with FITC-BSA. HA@DiD@FITC-BSA/MN (Fig. S11) and CNPs@DiD@FITC-BSA/MN (Fig. 3C) were prepared and imaged using LSM980. In HA@DiD@FITC-BSA/MN group, MN showed uneven red fluorescence distribution and faint signals for both fluorophores. CNPs@DiD@FITC-BSA/MN exhibited uniform, intense green and red fluorescence filling the needle tips. The heterogeneous distribution of red fluorescence (DiD) may be attributed to poor solubility of the lipophilic drug in the hydrophilic MN matrix in HA@DiD@FITC-BSA/MN group. The dim green fluorescence (FITC-BSA) in HA@DiD@FITC-BSA/MN may result from electrostatic repulsion between the negatively charged MN matrix and the anionic protein. The superior fluorescence performance of CNPs@DiD@FITC-BSA/MN serves as evidence for its efficient co-loading of small-molecule chemical drugs (DOX) and macromolecular protein drugs (RNP).Fig. 3Characterization of MNs. (A) Photographs of HA/MN and CS/MN captured by camera. (B) SEM images of MNs. (C) LSM980 images of CNPs@DiD@FITC-BSA/MN. (D) In vivo fluorescence delivery of CNPs@DiD@FITC-BSA/MN. In vivo imaging of DiD fluorescence signals (E) and their quantitative analysis (F). In vivo imaging of FITC fluorescence signals (G) and their quantitative analysis (H). (I) Skin healing assessment in BALB/c mice after application of CS/MN. G1: HA@DiD@FITC-BSA/MN, G2: CS@DiD@FITC-BSA/MN, G3: CNPs@DiD@FITC-BSA/MN. All data are presented as the mean ± s.d (n = 2 biological replicates per group). Characterization of MNs. (A) Photographs of HA/MN and CS/MN captured by camera. (B) SEM images of MNs. (C) LSM980 images of CNPs@DiD@FITC-BSA/MN. (D) In vivo fluorescence delivery of CNPs@DiD@FITC-BSA/MN. In vivo imaging of DiD fluorescence signals (E) and their quantitative analysis (F). In vivo imaging of FITC fluorescence signals (G) and their quantitative analysis (H). (I) Skin healing assessment in BALB/c mice after application of CS/MN. G1: HA@DiD@FITC-BSA/MN, G2: CS@DiD@FITC-BSA/MN, G3: CNPs@DiD@FITC-BSA/MN. All data are presented as the mean ± s.d (n = 2 biological replicates per group). To validate whether CNPs@DiD@FITC-BSA/MN is capable of effectively delivering both drugs subcutaneously, BALB/c mice treated with CNPs@DiD@FITC-BSA/MN were sacrificed at different time points. Skin sections were prepared and imaged. In Fig. 3D, the region of strongest influence intensity migrating deeper over time, indicating progressive drug penetration into the epidermis. This phenomenon demonstrates that CNPs/MNs can effectively deliver diverse drug species into the deep skin layers.The fluorescence was the strongest at 4 h, and the obvious fluorescence could still be observed at 24 h, indicating that the MNs had a slow release effect. To directly observe the release kinetics of DOX and RNP (represented by DiD and FITC-BSA) from different MNs in vivo, live animal imaging was performed on BALB/c mice treated with HA@DiD@FITC-BSA/MN, CS@DiD@FITC-BSA/MN, and CNPs@DiD@FITC-BSA/MN at 4, 24, 48, and 72 h (Fig. 3E). DiD signal intensity of CNPs@DiD@FITC-BSA/MN group (Fig. 3F) was significantly higher than other groups at every time point, demonstrating CNPs/MN can stably and sustainably release chemotherapeutic drugs. Due to the close proximity between the emission wavelength of FITC and the autofluorescence of skin, the overall FITC fluorescence signal was relatively weak in all groups (Fig. 3G). FITC intensities for all three groups were similar (Fig. 3H), with CNPs@DiD@FITC-BSA/MN slightly higher. The in vivo imaging results demonstrate that among these groups, CNPs/MN is the most suitable for co-delivering RNP and DOX. Furthermore, the results show that CS@DiD@FITC-BSA/MN consistently exhibited lower fluorescence intensity than CNPs@DiD@FITC-BSA/MN at all time points, and its intensity decreased more rapidly over time. This indicates that, compared to a single MN delivery system, the combined NP-MN delivery system exhibits superior drug delivery efficiency and enhanced sustained-release capability. In addition, it can be observed from the image that the fluorescence hardly diffuses and transfers to the non-administered site, which proves that all three groups of MNs have excellent tumor targeting ability. To evaluate the biocompatibility of MN, we examined the skin condition at the CS/MN application site and analyzed the corresponding H&E-stained sections at 0, 4, and 24 h post-application (Fig. 3I). At the beginning, visible puncture marks and distinct needle tip tracks were observed in the tissue sections. By 4 h, the skin marks had faded, and the needle tip tracks appeared flattened. At 24 h, the skin had almost completely recovered with no signs of inflammation. Previous studies have reported that CS can mitigate inflammatory responses in vivo [44]. These histological findings demonstrate the excellent biocompatibility of CS/MN and CNPs/MN, supporting its suitability for local transdermal delivery. Subsequent in vitro cell experiments would further assess the delivery efficiency of different materials for DOX and RNP. Cellular uptake efficiency and in vitro anti-tumor efficacy Having confirmed MN-mediated transdermal drug delivery, we investigated the impact of the developed NPs on cellular uptake of chemotherapeutic and macromolecular drugs, and assessed the differential intracellular delivery capabilities between HA, CS, CP, NPs, CNPs and LIPO. We prepared PBS@DiD@Cas9-EGFP, HA@DiD@Cas9-EGFP, CS@DiD@Cas9-EGFP, CP@DiD@Cas9-EGFP, NPs@DiD@Cas9-EGFP, CNPs@DiD@Cas9-EGFP, and LIPO@DiD@Cas9-EGFP for application to B16F10 cells. To facilitate the visualization of intracellular drug transport and localization, we substituted the model drug DOX with the highly fluorescent dye DiD and replaced non-fluorescent Cas9 with a Cas9-EGFP fusion protein. After 6 h incubation, nuclei were stained with DAPI, and cells were imaged by LSM980 (Fig. 4A). Results showed significantly higher red (DiD) and green (Cas9-EGFP) fluorescence intensity in the CNPs@DiD@Cas9-EGFP group compared to PBS@DiD@Cas9-EGFP, HA@DiD@Cas9-EGFP, CS@DiD@Cas9-EGFP, CP@DiD@Cas9-EGFP, and LIPO@DiD@Cas9-EGFP groups, demonstrating CNPs' superior drug delivery capability over HA, CS, CP, and LIPO. The presence of green fluorescence within nuclei in the CNPs@DiD@Cas9-EGFP treatment group demonstrated successful nuclear localization of the Cas9-EGFP RNP, confirming its functional delivery for genome editing applications.Fig. 4Cellular uptake mediated by different preparations. (A) LSM980 images of B16F10 cells after 6 h incubation with PBS@DiD@Cas9-EGFP, HA@DiD@Cas9-EGFP, CS@DiD@Cas9-EGFP, CP@DiD@Cas9-EGFP, NPs@DiD@Cas9-EGFP, CNPs@DiD@Cas9-EGFP, and LIPO@DiD@Cas9-EGFP. (B) Flow cytometry analysis of endocytosis in B16F10 cells mediated by these formulations. Quantification of DiD (C) and (D) FITC fluorescence in B16F10 cells. All data are presented as the mean ± s.d (n = 3 biological replicates per group). Cellular uptake mediated by different preparations. (A) LSM980 images of B16F10 cells after 6 h incubation with PBS@DiD@Cas9-EGFP, HA@DiD@Cas9-EGFP, CS@DiD@Cas9-EGFP, CP@DiD@Cas9-EGFP, NPs@DiD@Cas9-EGFP, CNPs@DiD@Cas9-EGFP, and LIPO@DiD@Cas9-EGFP. (B) Flow cytometry analysis of endocytosis in B16F10 cells mediated by these formulations. Quantification of DiD (C) and (D) FITC fluorescence in B16F10 cells. All data are presented as the mean ± s.d (n = 3 biological replicates per group). To quantify transfection rates, cells treated with PBS@DiD@FITC-BSA, HA@DiD@FITC-BSA, CS@DiD@FITC-BSA, CP@DiD@FITC-BSA, NPs@DiD@FITC-BSA, CNPs@DiD@FITC-BSA, and LIPO@DiD@FITC-BSA were detected by flow cytometry (Fig. 4B). After 6 h incubation, the cell group treated with CNPs@DiD@FITC-BSA showed dual uptake rate of both DiD and FITC-BSA, which was comparable to NPs@DiD@FITC-BSA and significantly higher than other groups. DiD uptake efficiency was shown in Fig. 4C. The percentage of DiD-positive cells in CNPs@DiD@FITC-BSA group was 86.13%, which was 2.10 times that of LIPO@DiD@FITC-BSA (40.92%), 6.51 times of HA@DiD@FITC-BSA group (13.23%) and 6.26 times of CP@DiD@FITC-BSA group (13.75%). As shown in Fig. 4D, the percentage of FITC-BSA-positive cells in CNPs@DiD@FITC-BSA group was 74.40%, while it was 1.30% in HA@DiD@FITC-BSA group, 0.91% in CS@DiD@FITC-BSA group, 7.41% in CP@DiD@FITC-BSA group, 74.00% in NPs@DiD@FITC-BSA group, and 4.04% in LIPO@DiD@FITC-BSA group. Flow cytometry results aligned with LSM980 images of cellular uptake, confirming the outstanding transfection efficiency of CNPs. These results demonstrate that CNPs possess optimal delivery pathways facilitating concurrent transfection of both DOX chemotherapeutic agents and RNPs, as evidenced by their superior cellular internalization efficiency and carrier compatibility. Lysosomal escape represents a major hurdle in the delivery of the Cas9 protein. Therefore, it is crucial to determine whether the CNPs delivery system can protect Cas9 from lysosomal enzymes and achieve successful lysosomal escape. We treated B16F10 cells with different formulations and, following incubation for 2 and 6 h, stained the lysosomes with a green fluorescent probe and observed them using confocal microscopy (Fig. 5A). At 2 h, green fluorescence was nearly absent in the PBS@Cas9-EGFP, HA@Cas9-EGFP, and CS@Cas9-EGFP groups. In contrast, obvious green fluorescence was observed in the CNPs@Cas9-EGFP group, which co-localized with lysosomes, indicating that CNPs@Cas9-EGFP had successfully delivered the protein into the cells at this stage. At 6 h, green fluorescence remained almost undetectable in the PBS@Cas9-EGFP group. In the HA@Cas9-EGFP and CS@Cas9-EGFP groups, a small amount of green fluorescence (Cas9-EGFP) was observed, and nearly all of it co-localized with the lysosomal marker (red fluorescence). In the CNPs@Cas9-EGFP group, however, distinct green fluorescence (representing Cas9-EGFP) was clearly observed, with the majority of green signals showing substantial separation from lysosomes. Quantitative analysis of lysosomal escape was performed using Pearson's correlation coefficient. At 2 h post-treatment (Fig. S12), the Pearson coefficient between green fluorescence (Cas9 protein) and blue fluorescence (nuclei) in the CNPs@Cas9-EGFP group was 0.487, significantly higher than that of the other groups. By 6 h, this value increased to 0.787, indicating strong co-localization of the protein with the nucleus. These results demonstrate that CNPs efficiently deliver the protein into the nucleus, and the successful lysosomal escape mediated by this platform ensures the therapeutic efficacy of the RNP.Fig. 5Efficiency of RNP-mediated gene editing delivered by different formulations. (A) Lysosomal escape capability of PBS@Cas9-EGFP, HA@Cas9-EGFP, CS@Cas9-EGFP and CNPs@Cas9-EGFP. (B) GFP knockout efficiency in 293T-GFP cells treated with PBS@RNPGFP, HA@RNPGFP, CS@RNPGFP, CNPs@RNPGFP, LIPO@RNPGFP. (C) Western blot analysis of Cas9 protein delivery into B16F10 cells using different preparations. (D) Western blot analysis of YB-1 protein downregulation in B16F10 cells mediated by RNP delivery using different formulations. (E) T7EI assessment of YB-1 knockout efficiency in B16F10 cells. Efficiency of RNP-mediated gene editing delivered by different formulations. (A) Lysosomal escape capability of PBS@Cas9-EGFP, HA@Cas9-EGFP, CS@Cas9-EGFP and CNPs@Cas9-EGFP. (B) GFP knockout efficiency in 293T-GFP cells treated with PBS@RNPGFP, HA@RNPGFP, CS@RNPGFP, CNPs@RNPGFP, LIPO@RNPGFP. (C) Western blot analysis of Cas9 protein delivery into B16F10 cells using different preparations. (D) Western blot analysis of YB-1 protein downregulation in B16F10 cells mediated by RNP delivery using different formulations. (E) T7EI assessment of YB-1 knockout efficiency in B16F10 cells. The superior performance of the CNPs can be attributed to two distinct physicochemical mechanisms: enhanced stability and efficient endosomal escape. The core-shell architecture of the CNPs provides a robust protective barrier that effectively shields the encapsulated RNP from enzymatic degradation in the serum. Effective genome editing by CNPs@RNP To assess RNP delivery efficiency, we conducted a GFP knockout assay using sgGFP (Table S2) in 293T-GFP cells treated with various formulations (PBS@RNPGFP, HA@RNPGFP, CS@RNPGFP, CNPs@RNPGFP, LIPO@RNPGFP). Time-course fluorescence imaging (Fig. 5B) showed striking differences in editing efficacy: control groups (PBS@RNPGFP, HA@RNPGFP, CS@RNPGFP) displayed minimal fluorescence reduction with full signal recovery by 48 h, while LIPO@RNPGFP mediated only partial knockout. In contrast, CNPs@RNPGFP achieved rapid and durable gene editing. This superior performance stems from our multi-step targeting mechanism: (1) CNPs-facilitated cellular RNP delivery, (2) NLS-guided Cas9 transport, and (3) sgRNA-directed target recognition and binding [40]. Having validated this precision editing platform, we next applied structure-optimized RNPs targeting therapeutically relevant YB-1 for comprehensive functional characterization. We investigated the effects of different material-delivered RNP on YB-1 expression in B16F10 cells using Western blotting. The results showed that the Cas9 protein transfection in cells treated with CNPs@RNP was significantly higher than other groups (Fig. 5C). At the same time, the expression of YB-1 in B16F10 cells treated with CNPs@RNP was significantly reduced compared with other groups (Fig. 5D). It was shown that CNPs successfully transfected RNP into B16F10 cells and effectively reduced the expression of YB-1 gene. The T7EI assay was conducted to assess the gene editing efficiency of CNPs@RNP (Fig. 5E). Genomic DNA extracted from B16F10 cells treated with PBS@RNP, HA@RNP, CS@RNP and CNPs@RNP was analyzed using the T7EI kit. CRISPR/Cas9-mediated YB-1 gene disruption was most efficient with CNPs@RNP. Gel electrophoresis analysis showed that CNPs@RNP yielded the faintest uncut band intensity alongside a prominent cleaved product at 103 bp, confirming efficient target DNA cleavage. Quantitative analysis revealed significantly higher editing efficiency for CNPs@RNP (65.32%) versus control groups: CS@RNP (7.95%), HA@RNP (1.25%), and PBS@RNP (normalized to 1%) (p < 0.001). Moreover, CNPs@RNP successfully disrupted the YB-1 gene. In summary, the CNPs delivery platform exhibited superior performance by: (1) enabling efficient RNP delivery targeting both reporter (GFP) and therapeutic (YB-1) genes in multiple cell models (B16F10 and 293T-GFP); (2) promoting effective lysosomal escape; and (3) accomplishing high-fidelity genome editing in vitro. These findings collectively demonstrate CNPs as a versatile and potent vehicle for CRISPR/Cas9 RNP delivery and genome editing applications. In vitro antitumor efficacy of CNP@DOX@RNP Cytotoxicity assays were performed to evaluate the cytotoxicity of various preparations in human embryonic kidney 293T cells and murine B16F10 melanoma cells. In the absence of DOX (Fig. 6A and B), both 293T and B16F10 cells treated with different concentrations of HA, CS, and CP maintained high cell viability (>70%) across all tested concentrations. Cells treated with NPs and CNPs also showed high viability (>75%) at low concentrations. However, at higher concentrations (50, 100, 150, 200 μg/mL), CNPs demonstrated lower toxicity than NPs in 293T cells. In B16F10 cells, CNPs and NPs exhibited comparable toxicity at 150 and 200 μg/mL. When DOX was incorporated into the preparations at various concentrations (Fig. 6C), the CNPs@DOX group showed enhanced therapeutic efficacy. At 50 μg/mL, B16F10 cell viability in the CNPs@DOX group was significantly reduced to 10.19%, lower than all other groups. These results demonstrate that CNPs and CNPs@DOX exhibit selective toxicity, showing high potency against melanoma cells while maintaining lower toxicity toward normal cells. This result is consistent with published literature, indirectly demonstrating that CS enables tumor cell targeting and mediates enhanced selective cytotoxicity toward cancer cells [42,45].Fig. 6Cytotoxicity of different formulations on human embryonic kidney 293T cells and murine B16F10 melanoma cells. (A) Cytotoxicity of blank carriers on 293T cells at varying concentrations. (B) Cytotoxicity of blank carriers on B16F10 cells. (C) Cytotoxicity of DOX-loaded preparations on B16F10 cells. (D) Comparative cytotoxicity of HA@DOX@RNP versus CNPs@DOX@RNP in B16F10 cells. The above data are presented as the mean ± s.d. (n = 6 biological replicates per group). Colony-forming assay of B16F10 cells treated with PBS@DOX@RNP, HA@DOX@RNP, CS@DOX@RNP, and CNPs@DOX@RNP. Quantitative analysis (E) and representative images (F) are shown. (G, H) Flow cytometric analysis of apoptosis in B16F10 cells. Cells were treated with PBS@DOX@RNP, HA@DOX@RNP, CNPs, CNPs@DOX, and CNPs@DOX@RNP. Apoptosis rates are presented in (G) and quantified in (H). Data are presented as the mean ± s.d. (n = 3 biological replicates per group). In Fig. A-C, the same letter means no significant difference, and different letters mean significant difference. A group can be labeled with multiple letters (e.g., a, b), which means that it is not significantly different from the group labeled as a, nor is it significantly different from the group labeled as b, but there is a significant difference between the group labeled as a and the group labeled as b. Cytotoxicity of different formulations on human embryonic kidney 293T cells and murine B16F10 melanoma cells. (A) Cytotoxicity of blank carriers on 293T cells at varying concentrations. (B) Cytotoxicity of blank carriers on B16F10 cells. (C) Cytotoxicity of DOX-loaded preparations on B16F10 cells. (D) Comparative cytotoxicity of HA@DOX@RNP versus CNPs@DOX@RNP in B16F10 cells. The above data are presented as the mean ± s.d. (n = 6 biological replicates per group). Colony-forming assay of B16F10 cells treated with PBS@DOX@RNP, HA@DOX@RNP, CS@DOX@RNP, and CNPs@DOX@RNP. Quantitative analysis (E) and representative images (F) are shown. (G, H) Flow cytometric analysis of apoptosis in B16F10 cells. Cells were treated with PBS@DOX@RNP, HA@DOX@RNP, CNPs, CNPs@DOX, and CNPs@DOX@RNP. Apoptosis rates are presented in (G) and quantified in (H). Data are presented as the mean ± s.d. (n = 3 biological replicates per group). In Fig. A-C, the same letter means no significant difference, and different letters mean significant difference. A group can be labeled with multiple letters (e.g., a, b), which means that it is not significantly different from the group labeled as a, nor is it significantly different from the group labeled as b, but there is a significant difference between the group labeled as a and the group labeled as b. Based on these results, 50 μg/mL DOX was chosen as the optimal concentration for subsequent experiments. HA@DOX@RNP and CNPs@DOX@RNP were prepared and their toxicity were assessed in B16F10 cells (Fig. 6D). In vitro cytotoxicity analysis demonstrated that CNPs@DOX@RNP induced a 96.49% reduction in tumor cell viability, significantly higher than other groups. These results demonstrate that YB-1 knockout potently promotes tumor cell death. Colony formation assays (Fig. 6E and F) further confirmed the differential inhibitory effects of various formulations on cellular proliferation and survival. Cells treated with PBS@DOX@RNP served as the baseline control. Quantitative analysis (Fig. 6E) demonstrated the following potency hierarchy in clonogenic suppression: CNPs@DOX@RNP > HA@DOX@RNP > CS@DOX@RNP. The clonogenic assays revealed significantly stronger inhibition by CNPs@DOX@RNP (0.71% colony formation area) compared to HA@DOX@RNP (14.28%, p < 0.01) and CS@DOX@RNP (36.93%, p < 0.0001). Fluorescence imaging analysis was used to evaluate the induction of apoptosis by various preparations (Fig. S13). PBS@DOX@RNP, HA@DOX@RNP, CS@DOX@RNP, CNPs@DOX@RNP and LIPO@DOX@RNP were administrated to B16F10 cells for 6 h. Fluorescence imaging showed that CNPs@DOX@RNP induced the highest intensity of both red (PI) and green (Annexin V-FITC) fluorescence, indicating superior apoptosis induction. This result was consistent with colony formation results. Quantitative analysis of apoptosis study was conducted by flow cytometry. Flow cytometry results were analyzed by plotting PI (red fluorescence) on the Y-axis and Annexin V-FITC (green fluorescence) on the X-axis. Based on gating, the plots were divided into four quadrants: the upper left quadrant (Q1) represents mechanically damaged cells; the upper right quadrant (Q2) indicates early apoptotic cells; the lower right quadrant (Q3) represents late apoptotic cells; and the lower left quadrant (Q4) represents viable cells. The proportions of cells in Q2 and Q3 were statistically analyzed for quantitative evaluation. Cells treated with PBS@DOX@RNP served as the baseline control. The apoptosis rates of cells treated with HA@DOX@RNP, CNPs, CNPs@DOX, and CNPs@DOX@RNP were 16.05%, 10.54%, 32.71%, 40.70% respectively. As illustrated in Fig. 6H, CNPs@DOX@RNP group displayed maximal apoptosis rate among all experimental groups, significantly higher than HA@DOX@RNP (p < 0.001) and CNPs (p < 0.0001) groups. A progressive increase in apoptotic cells was observed: CNPs < CNPs@DOX < CNPs@DOX@RNP. These results validate the therapeutic efficacy of CNPs as a drug delivery platform for inducing cytotoxic effects in melanoma cells. In vivo antitumor analysis of CNP @DOX@ RNP/MN To evaluate in vivo antitumor efficacy of CNPs@DOX@RNP/MN, C57BL/6 melanoma models were established. Five days post-tumor inoculation, tumor-bearing C57BL/6 mice were randomized into five groups: Control, HA@DOX/MN, CS@DOX/MN, CNPs@DOX/MN, CNPs@DOX@RNP/MN. Treatments were administered on day 5, 8, and 11 (Fig. 7A), with body weight and tumor volume recorded every two days. Mice were euthanized by cervical dislocation under anesthesia and dissected on day 23. HA@DOX/MN, CS@DOX/MN, CNPs@DOX/MN, CNPs@DOX@RNP/MN showed significant tumor suppression, and CNPs@DOX@RNP/MN was most effective (Fig. 7B). Quantitative analysis (Fig. 7C) revealed a statistically significant gradient in tumor volumes: CNPs@DOX@RNP/MN (49.182 mm3) < CNPs@DOX/MN (245.842 mm3) < HA@DOX/MN (266.776 mm3) < CS@DOX/MN (335.852 mm3) < Control (1201.860 mm3) (p < 0.001), demonstrating CNPs@DOX@RNP/MN's potent tumor inhibition. The mean tumor weights in control group, HA@DOX/MN, CS@DOX/MN, CNPs@DOX/MN and CNPs@DOX@RNP/MN treated group were 0.981 g, 0.221 g, 0.319 g, 0.198 g and 0.073 g respectively (Fig. 7D). Additionally, no significant body weight differences were observed among all groups (Fig. 7E).Fig. 7In vivo antitumor efficacy of CNPs@DOX@RNP/MN. (A) Therapeutic regimen schematic. (B) Representative tumor sizes post-treatment. (C) Tumor volumes. (D) Tumor weights. (E) Body weight changes. The above data are presented as the mean ± s.d. (n = 5 biological replicates per group). (F) Ki67, CD31, PD-L1 immunohistochemical staining, and H&E staining of tumor sections. (G) Quantitative analysis of Ki67-positive cells. (H) Quantitative analysis of CD31-positive areas. (I) Quantitative analysis of PD-L1-positive cells. Data are presented as the mean ± s.d. (n = 3 biological replicates per group). In vivo antitumor efficacy of CNPs@DOX@RNP/MN. (A) Therapeutic regimen schematic. (B) Representative tumor sizes post-treatment. (C) Tumor volumes. (D) Tumor weights. (E) Body weight changes. The above data are presented as the mean ± s.d. (n = 5 biological replicates per group). (F) Ki67, CD31, PD-L1 immunohistochemical staining, and H&E staining of tumor sections. (G) Quantitative analysis of Ki67-positive cells. (H) Quantitative analysis of CD31-positive areas. (I) Quantitative analysis of PD-L1-positive cells. Data are presented as the mean ± s.d. (n = 3 biological replicates per group). Ki67 is a nuclear protein serving as a cell proliferation marker, identifying cells in active phases of the cell cycle. A higher Ki67-positive rate correlates with increased tumor growth velocity. Ki67 IHC staining (Fig. 7F and G) showed that CNPs@DOX@RNP/MN group had lowest positive cell percentage (14.137% vs 68.521% in control group, p < 0.0001). As a canonical endothelial marker, CD31 identifies tumor vasculature. Its expression level quantitatively correlates with angiogenic capacity, reflecting the tumor's ability to establish vascular networks for nutrient/waste exchange essential for malignant growth. CD31 IHC staining (Fig. 7F and H) showed near-absence of positive cells in CNPs@DOX@RNP/MN group (1.048% vs 4.423% in control group, p < 0.0001). YB-1 acts as an upstream regulator of PD-L1. Theoretically, downregulating YB-1 expression should reduce PD-L1 levels, thereby making it more difficult for tumor cells to evade immune surveillance. Therefore, we investigated the expression of PD-L1 in tumors and found that the PD-L1 expression in the CNPs@DOX@RNP/MN group was reduced by 72.27% compared to the control group (Fig. 7F and I). This significant downregulation is one of the key reasons for the superior therapeutic efficacy of CNPs@DOX@RNP/MN. H&E staining of tumor sections from CNPs@DOX@RNP/MN-treated C57BL/6 mice displayed extensive necrosis, characterized by nuclear shrinkage (pyknosis), fragmentation (karyorrhexis), and loss of cellular architecture (Fig. 7F). These histological findings, combined with a significant reduction in Ki67-positive cells and CD31 expression, confirm that CNPs@DOX@RNP/MN effectively induces tumor cell death while simultaneously suppressing proliferation and metastasis. Meanwhile, the treatment effectively downregulates PD-L1 expression to disrupt tumor immune evasion, thereby enhancing the cytotoxic activity of T cells against the malignancy. Survivin, a member of the inhibitor of apoptosis protein (IAP) family, blocks caspase activity to prevent chemotherapy-induced apoptosis and contributes to tumor drug resistance [46]. YB-1 upregulates Survivin via the PI3K/Akt pathway [47]. Immunohistochemical analysis (Fig. S14) showed that CNPs@DOX@RNP/MN reduced Survivin expression to 9.24% of the control level, significantly lower than the CNPs@DOX/MN group, in which Survivin was 4.62-fold higher. These findings, together with the apoptosis data, confirm that YB-1 knockdown mitigates drug resistance and sensitizes cancer cells to DOX-induced apoptosis. Moreover, given Survivin's role in promoting migration and metastasis via cytoskeletal regulation [48], its downregulation further suggests that this system suppresses not only proliferation but also invasive potential. Conversely, E-cadherin, the key epithelial adhesion molecule, maintains tissue integrity; its loss marks epithelial-mesenchymal transition (EMT), enabling detachment and metastasis [49]. CNPs@DOX@RNP/MN elevated E-cadherin expression to 6.57 times that of controls. Since YB-1 represses CDH1 (E-cadherin) transcription through EMT-related transcription factors, YB-1 knockdown enhances E-cadherin expression [50], thereby reducing invasiveness and metastatic potential, and potentially improving prognosis. TUNEL staining of tumor sections treated with CNPs@DOX@ RNP/MN (Fig. 8A) exhibited the highest fluorescence intensity, and quantitative results (Fig. 8B) confirmed the strongest apoptosis-inducing effect, with significant differences observed compared to all other groups (p < 0.0001). As shown in Fig. 8C and D, the CNPs@DOX@RNP/MN group exhibited the lowest YB-1 expression, which was only 27.48% of that in the control group. A correlation analysis between YB-1 expression and tumor weight yielded a curve with R2 value greater than 0.8. This demonstrates a strong positive correlation between YB-1 expression and tumor weight, indicating that downregulating YB-1 expression can enhance the tumor inhibition rate.Fig. 8In vivo antitumor efficacy of CNPs@DOX@RNP/MN. (A) TUNEL staining of tumor sections. (B) Quantitative analysis of MFI of TUNEL. (C) YB-1 Immunofluorescence staining of tumor sections. (D) Quantitative analysis of MFI of YB-1. (E) Correlation analysis between YB-1 expression level and tumor weight. Data are presented as the mean ± s.d. (n = 5 biological replicates per group). In vivo antitumor efficacy of CNPs@DOX@RNP/MN. (A) TUNEL staining of tumor sections. (B) Quantitative analysis of MFI of TUNEL. (C) YB-1 Immunofluorescence staining of tumor sections. (D) Quantitative analysis of MFI of YB-1. (E) Correlation analysis between YB-1 expression level and tumor weight. Data are presented as the mean ± s.d. (n = 5 biological replicates per group). We next evaluated the biosafety profile of CNPs@DOX@RNP/MN through multiparametric toxicity assessment. H&E staining of major organs (Fig. S15A) from C57BL/6 mice treated with CNPs@DOX@RNP/MN showed no pathological abnormalities. Serum biochemistry was analyzed for hepatotoxicity markers (ALT/AST) and nephrotoxicity indicators (UREA/CREA) (Fig. S15B and C and D and E). The results show that there was no significant difference in the levels of these enzymes in the blood of the CNPs@DOX@RNP/MN groups, and all values were within the normal range, indicating that the hepatotoxicity and nephrotoxicity of CNPs@DOX@RNP/MN were negligible. Nevertheless, the measured AST values in the remaining four groups exceeded the normal limits. This elevation is presumably attributed to the impact of excessive tumor growth. We also tested the editing efficiency of YB-1 in major organs of mice treated with CNPs@DOX@RNP/MN by T7EI. The results showed that CNPs@DOX@RNP/MN did not affect major organs (Fig. S16). CS modification reduced NP toxicity while enhancing biocompatibility, thereby preventing organ damage and inflammation. These results collectively demonstrate the high safety and biocompatibility profile of our delivery system. Even with repeated administrations, no significant adverse effects were observed, including: (1) minimal body weight fluctuations (<5% change vs. controls); (2) no histopathological abnormalities (as assessed by H&E staining of major organs); (3) stable serum biochemical parameters (ALT, AST, UREA, and CREA remaining within normal ranges). In conclusion, the CNPs@DOX@RNP/MN platform achieves potent anti-melanoma effects through transdermal co-delivery of YB-1-targeting RNPs and DOX, demonstrating synergistic therapeutic efficacy alongside excellent biosafety. This dual-action system simultaneously disrupts oncogenic pathways via CRISPR/Cas9-mediated YB-1 knockout and induces chemotoxic DNA damage, while the MN array ensures localized delivery that minimizes systemic exposure. The combined pharmacological and genetic intervention represents a promising strategy for melanoma treatment, particularly for inoperable lesions or as a neoadjuvant approach prior to surgical resection. However, further studies, including survival analysis, immune profiling, and metastasis assessment, are required to fully evaluate its translational potential. Conclusion In summary, this study successfully developed a multifunctional MN delivery system for the transdermal co-delivery of YB-1-targeting RNP and DOX for the treatment of murine melanoma. We engineered a two-stage platform in which RNP- and DOX-loaded CNPs were depot-loaded into MNs composed of CS and PEG6000. Mechanistically, the CNPs@DOX@RNP/MN system demonstrated strong internalization by melanoma cells. Upon skin insertion, the highly biocompatible MNs created micrometer-scale pores, facilitating the efficient delivery of drug-loaded CNPs directly to the melanoma tissue, thereby enabling intracellular gene editing and chemotherapy. This approach effectively suppressed the proliferation and metastasis of B16F10 cells. In vitro results collectively indicated that CNPs offered superior delivery performance, achieving 74.4% cellular uptake of both drugs and over 60% gene editing efficiency. This translated into potent antitumor activity, with CNPs@DOX@RNP suppressing cell survival to as low as 3.51%, alongside reduced off-target effects. Crucially, in vivo application of the CNPs@DOX@RNP/MN system proved highly effective, resulting in a 92.56% decrease in final tumor weights as compared with untreated controls. Overall, this novel multifunctional MN system provides a promising synergistic therapeutic strategy for melanoma and serves as a valuable blueprint for the development of safe and effective in vivo RNP delivery platforms. Looking forward, future iterations of this system could incorporate stimuli-responsive mechanisms, such as pH or photothermal sensitivity, to achieve more precise and controllable drug release, thereby further enhancing antitumor efficacy. This nanoparticle - microneedle delivery system can be extended to other cutaneous indications, including psoriasis and skin senescence. Additionally, the tumor model employed in this study has certain limitations; future investigations using clinically more relevant models are warranted to further validate these findings. CRediT authorship contribution statement Xiazhou Fu: Writing – review & editing, Project administration, Methodology, Funding acquisition, Conceptualization. Zhaoyi Chen: Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation. Ziru Zhang: Validation, Investigation. Yufei Jiang: Validation, Data curation. Xiaomeng Jiang: Investigation, Formal analysis. Yuxin Yang: Validation, Supervision. Hongsheng Zhang: Methodology, Investigation. Xinyun Zou: Visualization, Investigation. Jingwen Luo: Writing – review & editing, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization. Funding This work was financially supported by the 10.13039/501100012166National Key Research and Development (R&D) Program of China (No. 2025YFA0922400), 10.13039/501100001809National Natural Science Foundation of China (No. 81900987), and 10.13039/501100003819Natural Science Foundation of Hubei Province (No. 2025AFB729). Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.