Self-assembling pH-responsive peptide nanoparticles delivering pyrvinium pamoate induce osteosarcoma senescence through CBX4 ubiquitination–mediated inhibition of YAP1 SUMOylation

Abstract

Osteosarcoma is a highly malignant and metastasis-prone bone tumor, and current treatment options remain unsatisfactory, underscoring the urgent need for new targeted strategies. Induction of cellular senescence represents a promising non-apoptotic antitumor mechanism. Although CK1α is considered to have tumor-suppressive potential, its underlying mechanism and the limited bioavailability of its specific activator pyrvinium pamoate (PP) have hindered clinical translation. Here, we developed a pH-responsive biomineralization-induced peptide self-assembly nanodelivery system (NP@PP) to improve the druggability of PP. Through in vitro and in vivo experiments combined with transcriptomic sequencing, co-immunoprecipitation, and Western blot analysis, we found that PP released from NP@PP efficiently activated CK1α and specifically promoted the ubiquitin-dependent degradation of CBX4. The loss of CBX4 further suppressed YAP1 SUMOylation and blocked its nuclear translocation, thereby activating p16^INK4a and p21^Cip1 expression, inducing osteosarcoma cell senescence, and inhibiting proliferation and metastasis. In mouse models, NP@PP markedly suppressed xenograft growth and lung metastasis, prolonged survival, and showed no obvious toxicity. These findings reveal a critical role of the CK1α-CBX4-YAP1 signaling axis in senescence-based osteosarcoma therapy and provide a theoretical and practical basis for the development of new nanomedicines.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12951-026-04729-z.

Introduction

Osteosarcoma (OS) stands as a highly aggressive primary bone tumor [1] that predominantly occurs in adolescents and children [2]. It exhibits strong local invasiveness and a high propensity for distant metastasis, with lung metastasis being the most common [3]. Although current clinical strategies primarily involve a combination of surgery, radiotherapy, and chemotherapy [4], only a subset of patients achieve short-term survival benefits, and the overall five-year survival rate remains at 60%-70% [5]. Furthermore, outcomes for patients with recurrent or metastatic disease remain poor [6]. Due to the high heterogeneity of OS and its complex molecular regulatory networks [7], conventional treatment approaches often fail to intervene precisely in its progression [8]. Recent studies have increasingly highlighted dysregulation of the tumor cell cycle and associated signaling pathways [9], aiming to control tumor progression through targeted intervention of critical pathways [10]. However, targeted strategies that precisely control OS cell growth, migration, and senescence remain limited. Therefore, clarifying the molecular mechanisms of OS progression and developing novel therapies based on advanced drug delivery platforms are urgently needed.

Cellular senescence is a stable state of cell cycle arrest induced by endogenous or exogenous stress, characterized by irreversibility, persistence, and active metabolism [11, 12]. In antitumor research, inducing senescence has been recognized as a tumor-suppressive strategy through a non-apoptotic mechanism, effectively limiting the unlimited proliferative potential of tumor cells [13]. In various malignancies, including OS, inducing cellular senescence has been confirmed to inhibit tumor growth and metastasis significantly [14]. CBX4, a multifunctional nuclear protein [15], is extensively involved in chromatin remodeling, DNA repair, and ubiquitination regulation, and serves as a crucial factor in regulating cell fate [16]. Ubiquitination modification regulates protein stability and degradation processes [17], and CBX4 itself functions as an E3 ubiquitin ligase involved in the degradation of downstream proteins [18]. YAP1, a core transcriptional coactivator of the Hippo signaling pathway [14], is regulated by SUMOylation, which influences its subcellular localization and activity, thereby contributing to tumor cell proliferation, metastasis, and stemness [13]. Studies have suggested that the degradation of CBX4 may impact the transcriptional function of YAP1, thereby inducing cellular senescence [13]. However, no systematic research has yet clarified the specific role of the interplay between CBX4 ubiquitination and YAP1 SUMOylation in OS, particularly regarding their functions in regulating cellular senescence.

CK1α is a serine/threonine kinase involved in the regulation of multiple cellular signaling pathways, including Wnt, p53, and DNA damage repair [19, 20]. Studies have shown that CK1α can influence the stability of various proteins by modulating ubiquitination and SUMOylation processes, thereby exerting tumor-suppressive effects in multiple cancers [21, 22]. In OS, CK1α expression levels are closely associated with tumor differentiation and proliferation, and targeted activation of CK1α has been proposed as a promising therapeutic strategy [23]. However, the low bioavailability of CK1α itself limits its clinical application, highlighting the need for the development of efficient CK1α delivery systems. Pyrvinium pamoate, an FDA-approved anthelmintic drug, has recently been identified as a selective activator of CK1α. It significantly enhances CK1α activity, thereby inducing tumor cell death and suppressing cancer stem cell properties [24]. Nevertheless, the poor solubility and unfavorable pharmacokinetic profile of PP in vivo underscore the urgent need for suitable delivery carriers to improve its therapeutic efficacy.

Meanwhile, the development of nanodrug delivery systems has provided new opportunities for targeted cancer therapy [25]. Nanoplatforms have been widely applied in precision oncology due to their strong structural designability, high drug-loading efficiency, and excellent biocompatibility [26]. As intelligent drug delivery systems, pH-responsive nanomaterials could release therapeutic agents in response to the acidic microenvironment characteristic of tumor tissues, thereby significantly enhancing tumor targeting and reducing systemic toxicity [27]. Biomineralization-induced peptide self-assembly technology has recently emerged as a novel method for nanostructure fabrication. By mimicking the natural biomineralization process, it generated stable and well-ordered nanostructures with excellent stability and controllability [28]. Additionally, multi-omics approaches such as high-throughput transcriptomic sequencing (RNA-seq) and post-translational modification prediction technologies provided powerful tools for elucidating drug mechanisms and identifying key targets [29, 30]. Integrating precise nanodrug delivery platforms with in-depth mechanistic research strategies promised the construction of novel drug systems that combined targeted delivery and functional regulation, offering breakthroughs in OS therapy.

Based on this background, we developed pH-responsive, biomineralization-induced peptide self-assembled nanoparticles loaded with the CK1α activator pyrvinium pamoate (NP@PP). This system was designed to target CBX4 ubiquitination and YAP1 SUMOylation, thereby inducing OS cell senescence and suppressing tumor proliferation and metastasis. This study systematically identified the CBX4–YAP1 axis as a potential regulator of OS cell senescence and evaluated its feasibility in tumor microenvironment-responsive nanodrug delivery. These findings provide new mechanistic insights into OS progression and offer a promising nanotherapeutic strategy with potential clinical translational value.

Materials and methods

Ethical statement

All animal experiments were conducted in accordance with relevant ethical guidelines and regulations. The study protocol was approved by the Institutional Animal Care and Use Committee (IACUC) (Approval No.: CMUKT2025182). Animals were maintained under humane conditions, and all efforts were made to minimize suffering. At the end of the study, all mice were euthanized under ether anesthesia.

Synthesis of γ-dodecyl-L-glutamate N-carboxyanhydride (DLG NCA)

L-Glutamic acid (5.0 g, 1.0 equiv., W275601, Sigma-Aldrich, USA) was suspended in a mixed solution of tert-butanol (34.0 mL, 10.5 equiv., PHR2184, Sigma-Aldrich, USA) and 1-dodecanol (22.9 g, 3.6 equiv., 8.03462, Sigma-Aldrich, USA). The reaction system was heated at 40 °C, and 95.0 wt% sulfuric acid (2.9 mL, 30743, Sigma-Aldrich, USA) was added dropwise. The mixture was then heated to 65 °C and maintained for 1 h until a homogeneous solution was formed. After stirring for an additional 1 h, the reaction mixture was cooled to room temperature, and triethylamine (2.7 mL, PHR2883, Sigma-Aldrich, USA) was added within 1 min to terminate the reaction. Water (8.0 mL) and ethanol (119.0 mL) were then added, and the mixture was stirred for 30 min to yield a white precipitate. The precipitate was collected by filtration, dried, and recrystallized in 91.0 mL water at 65 °C for 20 min. It was subsequently washed with methanol (23.0 mL) and diethyl ether (23.0 mL) and finally dried under vacuum to obtain a white powder (5.7 g). The resulting solid was recrystallized again using a 2:1 isopropanol-water (160.0 mL) mixed solvent at 85 °C and dried overnight under vacuum at 20 °C to obtain DLG.

DLG NCA was synthesized using the triphosgene method. In a dried three-neck flask, DLG (10.0 g, 3.0 equiv.) was suspended in anhydrous tetrahydrofuran (THF, 100.0 mL, 900636, Sigma-Aldrich, USA), and triphosgene (3.8 g, 1.2 equiv., 330752, Sigma-Aldrich, USA) was added. Under nitrogen (N₂) protection, the mixture was stirred at 55 °C for 30 min until the solution became clear, indicating the complete reaction of DLG. After flushing the reaction system with N₂ for 30 min, the solution was concentrated to 30.0 mL and poured into 300.0 mL of ice-cold n-hexane to precipitate a crude white powder. The crude NCA product was dissolved in 150.0 mL of ice-cold ethyl acetate (EtAc, PHR1481, Sigma-Aldrich, USA) and washed three times with ice water. Anhydrous magnesium sulfate (MgSO₄, 208094, Sigma-Aldrich, USA) was then added to remove residual water, and the solution was stored at -80 °C for 12 h. The purified solution was filtered through a G4 sintered funnel, and EtAc was removed by vacuum evaporation to obtain pure DLG NCA.

Synthesis of N(τ)-Trityl-L-histidine N-Carboxyanhydride (TLH NCA)

Boc-His(Trt)-OH (20.0 g, 1.0 equiv., 8.53034, Sigma-Aldrich, USA) was placed in a 500 mL round-bottom flask and dried overnight under high vacuum. Anhydrous THF (150.0 mL) was then distilled into the flask to dissolve the compound, yielding a clear pale-yellow solution. The flask was cooled in an ice bath under argon, after which thionyl chloride (3.25 mL, 1.1 equiv., 8.08154, Sigma-Aldrich, USA) diluted in 20.0 mL THF was added dropwise over 10 min. The mixture was stirred at 0 °C for 2 h and then slowly transferred into 2.0 L ice-cold diethyl ether (24004, Sigma-Aldrich, USA), resulting in precipitation of TLH NCA·HCl. The precipitate was collected using a G3 sintered glass funnel, transferred to a 500 mL round-bottom flask, and dried under high vacuum to afford 17.1 g of crude solid. The crude product was dissolved in 300.0 mL anhydrous EtAc, followed by solvent removal under high-vacuum distillation. The mixture was heated at 45 °C for 1 h and then gradually cooled to 0 °C to reprecipitate TLH NCA·HCl. The solid was collected by filtration and dried overnight under high vacuum.

Subsequently, 200.0 mL anhydrous EtAc was distilled into the flask. The reaction system was purged with argon and cooled in an ice bath. At 0 °C, triethylamine (3.5 mL, 1.0 equiv.) dissolved in 50.0 mL EtAc was added dropwise over 1 h under vigorous stirring. After completion, the generated triethylamine hydrochloride was removed by filtration, and the filtrate was poured into 1.5 L ice-cold n-hexane to promote TLH NCA crystallization. The product was further purified by a second recrystallization using an EtAc/n-hexane mixture at a ratio of 1:5. The final TLH NCA product was obtained as a white powder after filtration, dried overnight under high vacuum, and stored in a glovebox.

Synthesis of γ-Benzyl-L-Glutamate N-Carboxyanhydride (BLG NCA)

In a dried three-neck flask, BLG (20.0 g, 3.0 equiv., 49510, Sigma-Aldrich, USA) was dissolved in 200.0 mL of anhydrous THF. The reaction system was stirred under a N₂ atmosphere. Subsequently, triphosgene (10.0 g, 1.2 equiv.) was added, and the mixture was stirred at 55 °C for 1 h until the solution became clear, indicating a complete reaction. After completion, the system was purged under N₂ for 1 h, and the solution was concentrated at 40.0 mL. The concentrated solution was poured into 400.0 mL of ice-cold n-hexane (PHR1555, Sigma-Aldrich, USA), and the crude product was collected by filtration.

The crude solid was redissolved in 200.0 mL ice-cold EtAc and washed three times with ice water. Anhydrous MgSO⁴ was added to remove residual water, and the solution was kept at − 80 °C for 12 h. After filtration through a G4 sintered glass funnel, the solvent was evaporated under reduced pressure to obtain crude BLG NCA.

The crude product was further purified by recrystallization from THF/n-hexane (1:5, v/v). Finally, purified BLG NCA was obtained after vacuum drying, with a yield of 80.0 wt%.

Synthesis of pH-responsive biomineralization-induced peptide self-assembly NP

First, N-dodecylamine (DDA, 325163, Sigma-Aldrich, USA) was dissolved in an appropriate anhydrous solvent such as toluene. Under a N₂ atmosphere, DLG NCA and TLH NCA were added, and a ring-opening copolymerization was conducted to obtain the DDA-P(DLG-co-TLH) segment. The reaction mixture was stirred at room temperature for 72 h to ensure complete polymerization. BLG NCA was added, and the reaction was continued under the same conditions for 72 h to form the block copolymer DDA-P(DLG-co-TLH)-b-PBLG. Activated alanine-derived NCA was subsequently introduced to graft alanine onto the PBLG segment, and the reaction was continued for another 48 h to yield the final product, DDA-P(DLG-co-LH)-b-P(LG-g-ALN). The product was purified by precipitating the reaction solution into excess cold methanol, followed by thorough washing with methanol to remove unreacted monomers, and then dried under vacuum.

Then, 10.0 mg polymer was dissolved in 2.0 mL anhydrous DMF and stirred at 25 °C for about 2 h until a clear solution was obtained. The solution was transferred into a pre-hydrated dialysis membrane with a molecular weight cutoff of 3500 Da (132592-1, Spectrum Laboratories, USA) and dialyzed against 10 mM PBS (pH 7.4, 70011044, Gibco, USA) for 24 h, with the external buffer replaced every 2 h. The obtained micellar dispersion was lyophilized using a Labconco FreeZone 2.5 freeze dryer to produce purified DDA-P(DLG-co-LH)-b-P(LG-g-ALN) nanoparticles, designated as NP.

Cy5.5-labeled NPs were prepared using a similar procedure. Briefly, EDC·HCl (3.49 mg; E6383, Sigma-Aldrich, USA) and NHS (2.09 mg; 130672, Sigma-Aldrich, USA) were added to a solution containing 50.0 mg of DDA-P(DLG-co-LH)-b-P(LG-g-ALN) and stirred for 30 min at room temperature to activate polymer carboxyl groups. Sulfo-Cy5.5 amine (6.54 mg; R-C-3583, Xi’an Ruixi Biological Technology Co., Ltd., China) was then introduced, and the reaction was allowed to proceed for 24 h. Free Cy5.5 was removed by dialysis, yielding Cy5.5-labeled NPs.

Loading of PP

PP (P0027, Sigma-Aldrich, USA) was dissolved in anhydrous DMSO (276855, Sigma-Aldrich, USA) at 2 mg/mL, while DDA-P(DLG-co-LH)-b-P(LG-g-ALN) was prepared in DMSO at 10 mg/mL. The two solutions were mixed thoroughly at a polymer-to-drug weight ratio of 10:1 to obtain a uniform precursor solution. The mixture was transferred into a dialysis membrane with a MWCO of 3,500 Da (MP1704-5 M, Viskase, USA) and dialyzed against deionized water for 48 h, with water replaced every 2 h to remove DMSO and promote nanoparticle self-assembly. CaCl₂ was then added to a final concentration of 2.5 mM to facilitate ALN–Ca²⁺ coordination and enhance drug retention within the nanoparticles. Afterward, the dispersion was lyophilized using a Labconco FreeZone 2.5 freeze dryer to obtain NP@PP. Drug loading content and encapsulation efficiency were quantified by HPLC (Agilent 1260 Infinity, USA) at 405 nm, using acetonitrile/water (60:40, v/v) as the mobile phase at 1.0 mL/min. The formulas were: Drug loading content = (mass of loaded drug / total mass of NPs) × 100%; Encapsulation efficiency = (mass of loaded drug / total mass of drug added) × 100%.

Dynamic light scattering (DLS) measurement

NPs were dispersed in deionized water at 1 mg/mL and analyzed at room temperature using a Zetasizer Nano ZS instrument (Malvern Instruments, UK) at a scattering angle of 173°. The viscosity setting was based on standard water parameters. To assess pH responsiveness, particle size and ζ-potential were measured after adjusting the NP suspension to pH 7.4, 6.5, or 5.5. Each sample was tested three times, and mean values were recorded.

Transmission electron microscopy (TEM) observation

NP morphology was examined by TEM. Briefly, NP suspensions were diluted to 0.1 mg/mL, and 10 µL of each sample was deposited onto a carbon-coated copper grid (931519, Electron Microscopy Sciences, USA) for 2 min. Excess liquid was removed with filter paper, followed by staining with 2% phosphotungstic acid (P4006, Sigma-Aldrich, USA). After air drying, samples were observed using a JEOL JEM-2100 TEM at 120 kV.

pH-Responsive drug release experiments

NP@PP was dispersed in PBS at 2 mg/mL, and the pH was adjusted to 7.4, 6.5, or 5.5. A 1 mL aliquot was sealed in a dialysis bag with a MWCO of 3.5 kDa and immersed in 20 mL PBS of the corresponding pH. The system was incubated in a shaking incubator at 37 °C and 100 rpm. At 0, 2, 4, 8, 12, 24, 48, and 72 h, 1 mL of release medium was collected and replaced with fresh PBS. PP concentration was measured by HPLC, and cumulative release was calculated. Each experiment was performed in triplicate.

Cell culture and grouping

MG-63 OS cells (CL-0157, Procell, China) were maintained in MEM (PM150410, Procell, China), U2OS cells (HTB-96, ATCC, USA) in McCoy’s 5 A medium (16600082, Gibco, USA), and human osteoblasts hFOB 1.19 (bio-73294, Beijing Beiaobo Biotechnology Co., Ltd., China) in DMEM/F12 (12634010, Gibco, USA). All media were supplemented with 10% fetal bovine serum (FBS, F8687, Gibco, USA) and 1% penicillin–streptomycin (TMS-AB2, Gibco, USA). Cells were cultured at 37 °C in a humidified incubator with 5% CO₂ and passaged at 80%–90% confluence before seeding into 6- or 96-well plates.

The experimental groups were as follows: Control group (treated with PBS), CK1α activator group (PP group, treated with 200 nM PP), NP@PP group (treated with NP@PP containing 200 nM PP), and, specifically for the senescence phenotype analysis, a Flavopiridol group (MCE, HY-10005, 200 nM) was additionally included. The overexpression groups included the oe-NC group (MG-63 cells transduced with empty control lentivirus and treated with 200 nM NP@PP), the oe-CBX4 group (CBX4-overexpressing MG-63 cells treated with 200 nM NP@PP), and the oe-YAP1 group (YAP1-overexpressing MG-63 cells treated with 200 nM NP@PP). All treatments were applied for 24 h, and each experiment was repeated three times [18]. Plasmid sources: His-tagged pcDNA3.1-SUMO1-AA/GG, Flag-tagged pcDNA3.1-UBC9, and pLKO.1-puro containing shRNA targeting YAP1 were all designed and synthesized by GeneCopoeia (Guangzhou, China) [13].

Lentiviruses were packaged by Sangon Biotech (Shanghai, China). pHAGE-puro vectors and helper plasmids pSPAX2 and pMD2.G (#118692, #12260, #12259; Addgene, USA) were co-transfected into HEK293T cells using Lipofectamine 2000 (11668030, Thermo Fisher, USA). Viral supernatants were collected at 48 and 72 h, filtered (0.45 μm), concentrated, combined, and titrated.

MG-63 cells in the logarithmic phase were seeded at 1 × 10⁵ cells/well in 6-well plates. After 24 h (≈ 75% confluence), cells were infected with lentivirus (MOI = 10, ~ 5 × 10⁶ TU/mL) in the presence of polybrene (5 µg/mL, TR-1003, Merck, USA). After 4 h, fresh medium was added, and the medium was replaced after 24 h. Stable cell lines were established by puromycin selection (2 µg/mL, E607054, Sangon Biotech, China), with stepwise increases (2–10 µg/mL). Resistant cells were collected, and overexpression efficiency was validated by Western blot (WB) and RT-qPCR [31].

Cancer cell membrane adhesion analysis

MG-63 cells were incubated with FITC-labeled NP@PP (1.0 mg/mL, HY-66019, MedChemExpress, USA) for 1 h, followed by fixation with 4% (w/v) paraformaldehyde in PBS for 30 min. Cell membranes were then stained with 10.0 µM DiD dye (λ_ex = 644 nm, λ_em = 665 nm, C1039, Beyotime, China) for 30 min. Subsequently, nuclei were labeled with Hoechst 33,342 (5.0 µg/mL, λ_ex = 345 nm, λ_em = 478 nm, C1022, Beyotime, China) for 5 min. Fluorescence images were acquired using a confocal laser scanning microscope (CLSM, LSM 780 NLO, Carl Zeiss AG, Oberkochen, Germany).

Cell membrane calcification analysis

MG-63 cells were seeded on glass coverslips in 3.5 cm dishes and cultured for 24 h. Nuclei and membranes were labeled with Hoechst 33,342 and DiD, respectively, for 10 min. After washing with saline, cells were treated with NP@PP (1.0 mg/mL) in MEM adjusted to pH 7.4 or 6.5 for 3 h.

Cells were then fixed with 4.0% paraformaldehyde (P0099, Beyotime, China) for 30 min and stained with Calcein (1.0 µg/mL; λ_ex = 495 nm, λ_em = 515 nm; C1367S, Beyotime, China) for 15 min to detect calcium phosphate deposition. After rinsing, images were captured by CLSM and analyzed using Olympus FV10-ASW 4.2 software (Olympus Co., Tokyo, Japan).

For scanning electron microscopy (SEM) observation, MG-63 cells were seeded on sterile 8 mm × 8 mm silicone sheets in 48-well plates at 5.0 × 10⁴ cells/well. After treatment with NP@PP (1.0 mg/mL) for 3 h, cells were washed, fixed with 2.5% glutaraldehyde for 2 h, dehydrated through graded ethanol solutions (10%–100%, 30 min each), dried at 4°C, sputter-coated with gold, and observed by SEM.

Detection of cell membrane fluidity using the TMA-DPH Method

MG-63 cells were cultured for 24 h and then treated with NP@PP (1.0 mg/mL) diluted in complete MEM at pH 7.4 or 6.5 for 12–24 h. Membrane fluidity was evaluated using a TMA-DPH assay kit (BB-48118, Beyotime Biotechnology, China). TMA-DPH solution was added directly to the wells and incubated for 10 min to allow membrane binding. Fluorescence was measured using a microplate reader (Spark M200, Tecan Austria GmbH, Grödig, Austria) at 365 nm excitation and 430 nm emission. Data were collected and analyzed with SparkControl V2.3 software (Tecan Group Ltd., Männedorf, Switzerland).

Atomic force microscopy (AFM) analysis

The Young’s modulus of treated MG-63 cells was analyzed using AFM (Nanosurf FlexAFM, Nanosurf AG, Switzerland). A probe equipped with a spherical cantilever (diameter 6.1 μm, stiffness k = 0.1 N/m) was used to apply a precise vertical compressive force to the cells. All images were acquired at room temperature in MEM medium. Force-displacement curves were analyzed using AtomicJ software (developed by Pawel Hermanowicz, Poland).

In vitro biosafety evaluation

The cytotoxicity of NP@PP toward hFOB 1.19 cells was evaluated using a CCK-8 assay kit (C0041, Beyotime, China). Cells were seeded in 96-well plates and treated with NP@PP at 0, 25, 50, or 100 µg/mL in complete medium containing 10% FBS and 1% penicillin–streptomycin. After incubation at 37 °C with 5% CO₂ for 12–24 h, CCK-8 reagent was added and incubated for another 2 h. Absorbance was then recorded at 450 nm using a microplate reader.

Cell viability was further examined with a Live/Dead Cell Staining Kit (PF00007, Proteintech, USA). hFOB 1.19 cells were seeded in 24-well plates at 1 × 10⁴ cells/cm² and cultured with NP@PP for 1, 3, or 7 days. At each time point, cells were washed with PBS, stained with 200 µL live/dead working solution for 30 min at room temperature, and observed under a fluorescence microscope.

Hemolysis test

NP@PP solutions were prepared at 6.25, 12.5, 25, 50, 100, and 200 µg/mL. PBS and deionized water served as negative and positive controls, respectively. EDTA-anticoagulated fresh whole blood from BALB/c nude mice (Mouse Whole Blood, Huizhi Heyuan Biotechnology, Suzhou, China) was mixed with NP@PP solutions or control solutions and incubated at 37 °C for 3 h. Samples were then centrifuged at 2000 g for 10 min, and the absorbance of the supernatant was measured at 540 nm to assess hemoglobin release.

CCK-8 assay

MG-63 cells were seeded into 96-well plates at 5 × 10³ cells/well. After 24 h, cells were treated according to the experimental design. CCK-8 reagent (10 µL; CK04, Dojindo, Japan) was added after treatment and incubated at 37 °C for 1 h. Absorbance was measured at 450 nm using a microplate reader, and cell viability was calculated for each group.

EdU staining assay

Cell proliferation was further examined using the BeyoClick™ EdU-594 kit (C0078, Beyotime, China). After treatment, MG-63 cells were cultured in 6-well plates for 24 h and incubated with 10 µM EdU for 2 h. Cells were then fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.5% Triton X-100 (X100PC, Sigma-Aldrich, Germany) for 5 min, and reacted with EdU staining solution in the dark for 30 min. Nuclei were counterstained with Hoechst 33,342, and EdU-positive cells were observed under a fluorescence microscope.

Colony formation assay

MG-63 cells were plated into 6-well plates at 500 cells/well and cultured until visible colonies appeared. Treatment-containing medium was refreshed every three days. Colonies were fixed with 4% paraformaldehyde for 10 min, stained with 0.1% crystal violet (V5265, Sigma-Aldrich, USA) for 15 min, washed, air-dried, and photographed. Colony numbers were quantified using ImageJ software (NIH, USA).

Wound healing assay

MG-63 cells were seeded into 6-well plates and grown to approximately 90% confluence. A linear wound was created using a sterile pipette tip, and detached cells were removed with PBS. Serum-free medium containing the indicated treatments was then added. Images were taken at 0 and 24 h using an inverted microscope, and wound closure was analyzed with ImageJ software.

Transwell assay

Cell migration and invasion were assessed using 24-well Transwell chambers with 8 μm pores (CLS3396, Corning, USA). Treated MG-63 cells were suspended in serum-free medium at 1 × 10⁵ cells/mL. For invasion assays, 200 µL cell suspension was added to the upper chamber pre-coated with 60 µL Matrigel (E1270, Sigma-Aldrich, USA), while 500 µL medium containing 10% FBS was added to the lower chamber. After 10 h, cells remaining on the upper membrane surface were removed. Migrated or invaded cells were fixed with 4% paraformaldehyde for 15 min, stained with 0.25% crystal violet for 30 min, imaged under an inverted microscope, and counted. For migration assays, the same procedure was used without Matrigel coating.

Flow cytometry analysis of apoptosis and cell cycle

Apoptosis was assessed using an Annexin V-FITC/PI Apoptosis Detection Kit (40302ES60, Yeasen Biotechnology, China). After treatment, MG-63 cells were trypsinized, collected, washed twice with PBS, and resuspended in 100 µL 1× Binding Buffer. Cells were then stained with 5 µL Annexin V-FITC and 5 µL PI for 15 min in the dark. Early and late apoptotic cells were quantified using a BD Accuri C6 flow cytometer (USA).

For cell cycle analysis, treated cells were fixed overnight at 4 °C in pre-cooled 70% ethanol. After ethanol removal, cells were incubated with 500 µL PI staining solution containing RNase A (50 µg/mL, R6513, Sigma-Aldrich, USA) at 37 °C for 30 min. DNA content was detected by flow cytometry, and the proportions of cells in G0/G1, S, and G2/M phases were analyzed using FlowJo software (Tree Star, USA).

For combined senescence and apoptosis detection, cells were incubated on ice for 30 min in 100 µL Annexin-V binding buffer containing Annexin-V-FITC (BioLegend #640945; 0.1 M Hepes, pH 7.4; 1.4 M NaCl; 25 mM CaCl₂). Cells were then stained with PE-conjugated anti-uPAR antibody (human uPAR, BioLegend #338906, clone VIM5; mouse uPAR, R&D Systems #FAB531P, clone 109801), resuspended in PBS containing 10% FBS and DAPI (1 µg/mL), and analyzed using an LSRIIW or BD Symphony flow cytometer (Becton Dickinson) [32].

RT-qPCR analysis of gene expression

Total RNA was isolated from tissues and cells using Trizol reagent. cDNA was synthesized with the RevertAid First Strand cDNA Synthesis Kit (K1621, Thermo Fisher, USA). RT-qPCR was performed using SYBR^® Premix Ex Taq™ II Kit (DRR081, Takara, Japan) on an ABI7500 real-time PCR system (Thermo Fisher, USA). The amplification program was: 95 °C for 30 s, followed by 40 cycles of 95 °C for 5 s and 60 °C for 30 s, with a final extension at 95 °C for 15 s, 60 °C for 60 s, and 90 °C for 15 s. GAPDH served as the internal control. Each reaction was run in triplicate, and all experiments were repeated three times. Relative gene expression was calculated using the 2^−ΔΔCt method. Primer sequences are provided in Table S1.

Western blot analysis

Total protein was extracted from treated MG-63 cells using RIPA lysis buffer (P0013B, Beyotime, China), and protein concentration was measured with a BCA Protein Assay Kit (A55860, Thermo Fisher, USA). Protein samples (20 µg) were separated by electrophoresis and transferred onto PVDF membranes (PFL00010, Millipore, USA). After blocking with 5% non-fat milk for 1 h, membranes were incubated overnight at 4 °C with primary antibodies listed in Table S2. The next day, membranes were incubated with HRP-conjugated secondary antibodies (1:10,000; ab6728 or ab6721, Abcam, UK) for 1 h. Bands were detected using ECL reagent (P0018S, Beyotime, China), imaged with a ChemiDoc system (Bio-Rad, USA), and quantified by grayscale analysis.

Detection of CBX4 ubiquitination

MG-63 cells were lysed in RIPA buffer containing protease inhibitor cocktail (P8340, Sigma-Aldrich, Germany) on ice for 30 min, followed by centrifugation at 12,000 × g for 15 min at 4 °C. Supernatants were incubated with CBX4 antibody (ab242149, Abcam, UK) and Protein A/G magnetic beads (ab286842, Abcam, UK) at 4 °C with rotation overnight. After washing, immunoprecipitates were boiled in SDS loading buffer, separated by SDS-PAGE, and transferred to PVDF membranes. Ubiquitination levels were detected by WB using anti-ubiquitin (ab134953, Abcam, UK) and anti-CBX4 antibodies. Signals were visualized with ECL reagent and captured using a ChemiDoc system.

Co-IP analysis

Cell proteins were collected using RIPA lysis buffer containing 1% protease inhibitor. Agarose beads (Santa Cruz) were added, and the mixture was rotated overnight at 4 °C. Samples were centrifuged at 3000 g for 1 min at 4 °C, washed three times with TBS buffer, denatured in loading buffer, and subjected to WB analysis using antibodies against CK1α (938201, BioLegend), HA-tag (C29F4), His-tag (D3I1O), and Flag-tag (D6W5B) (Cell Signaling Technology) [13].

Detection of YAP1 SUMOylation

Cell lysates were incubated with SUMO1 antibody (1:200, Abcam, UK) at 4 °C overnight, followed by incubation with Protein A/G magnetic beads for 4 h. After washing, samples were analyzed by SDS-PAGE and transferred to PVDF membranes. Membranes were probed with anti-YAP1 (1:1000, ab205270, Abcam, UK) and anti-SUMO1 (1:1000, ab32058, Abcam, UK) antibodies. Signals were detected using a chemiluminescence imaging system.

β-Galactosidase staining

Cell senescence was evaluated using a β-galactosidase staining kit (C0602, Beyotime, China). Treated cells were washed with PBS and incubated with staining solution at 37 °C in the dark for 12 h. Blue-stained cells were observed under an inverted microscope, and senescence rates were calculated.

Immunofluorescence (IF) staining

MG-63 cells were seeded on coverslips and cultured for 24 h. Cells were fixed with 4% paraformaldehyde for 15 min and permeabilized with 0.3% Triton X-100 for 10 min. Primary antibodies against CBX4 (ab4189, Abcam, UK) and YAP1 (66900-1-Ig, Proteintech, China) were applied and incubated at 4 °C overnight. After washing, cells were incubated with fluorescent secondary antibodies (ab150077, Abcam, UK; RGAM002, Proteintech, China) for 1 h in the dark. Nuclei were stained with DAPI (C1005, Beyotime, China), and images were captured using a confocal microscope to assess CBX4–YAP1 colocalization and YAP1 localization.

Nuclear and cytoplasmic protein extraction

Nuclear and cytoplasmic proteins were isolated using a Nuclear and Cytoplasmic Extraction Kit (MIT1000, Sigma-Aldrich, Germany). YAP1 expression in each fraction was analyzed by WB. PCNA (1:1000, ab92552, Abcam, UK) and GAPDH (1:10000, ab181602, Abcam, USA) were used as nuclear and cytoplasmic loading controls, respectively, to evaluate YAP1 distribution.

In vitro SUMOylation assay

The in vitro SUMOylation assay was performed using purified recombinant proteins as previously reported with minor modifications [33]. Each 20 µL reaction contained 50 mM Tris/HCl (pH 7.5), 5 mM MgCl₂, 2 mM ATP, 250 ng Aos1/Uba2, 100 ng Ubc9, 3 µg SUMO-1, and 1 µg GST-Dnmt3aN (amino acids 1–482). Reactions were incubated at 30 °C for 2 h and stopped by adding SDS loading buffer. Products were analyzed by SDS-PAGE and WB. To assess the effect of Cbx4, the reaction was performed under the same conditions, except that Ubc9 was reduced to 20 ng and 25–100 ng Cbx4 was added.

RNA-seq and data quality control

Control and NP@PP-treated MG-63 cells were collected as described in the “Cell Culture and Treatment” section, with three biological replicates per group. Total RNA was extracted using Trizol reagent (16096020, Thermo Fisher, USA). RNA concentration, purity, and integrity were evaluated using a Qubit^® RNA Assay Kit (HKR2106-01, Shanghai Baiji Biotechnology Co., Ltd., China) on a Qubit^® 2.0 Fluorometer^® (Life Technologies, Q33216, USA), a Nanodrop spectrophotometer (IMPLEN, USA), and an RNA Nano 6000 Assay Kit (5067 − 1511, Agilent, USA) on a Bioanalyzer 2100 system, respectively. For each sample, 3 µg RNA was used for library construction with the NEBNext^® Ultra™ RNA Library Prep Kit for Illumina^® (NEB, E7435L, China). Library quality was checked using an Agilent Bioanalyzer 2100 system. Indexed libraries were clustered using the TruSeq PE Cluster Kit v3-cBot-HS (PE-401-3001, Illumina, USA) on a cBot Cluster Generation System. Sequencing was then performed on an Illumina HiSeq 550 platform to generate 125/150 bp paired-end reads. Raw paired-end reads were assessed with FastQC v0.11.8. Cutadapt v1.18 was used to remove Illumina adapters and poly(A) tails. Reads containing > 5% N bases were discarded using a Perl script, and reads with ≥ 70% bases showing Q > 20 were retained using FASTX Toolkit v0.0.13. Paired-end reads were corrected with BBMap, and the resulting high-quality reads were mapped to the human reference genome using HISAT2 v0.7.12.

Differential expression analysis

Based on the RNA-seq results, differential expression analysis was performed using the R package “limma,” with the Control group samples used as a reference. Differentially expressed genes (DEGs) were identified according to |log₂FC| > 0.5 and P < 0.05 criteria. The results were visualized using heatmaps and volcano plots generated with the R package “pheatmap.”

Gene Ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) functional enrichment analysis

GO and KEGG functional enrichment analyses of the DEGs were performed using the R package “clusterProfiler.” p-values were adjusted using the Benjamini-Hochberg method, and terms with an adjusted p-value (adjp) < 0.05 were considered significantly enriched. Particular attention was given to enrichment results related to senescence, proliferation, migration, and the cell cycle.

Protein-protein interaction (PPI) network construction

Genes enriched in GO and KEGG terms associated with senescence, proliferation, migration, and the cell cycle were selected for PPI network construction. The STRING database was used with a confidence score threshold of 0.15.

Correlation analysis of key gene expression in the TARGET-OS osteosarcoma cohort

To investigate the relationship between the CK1α–CBX4–YAP1 regulatory axis and classical osteosarcoma-associated pathways, gene expression profile data from the TARGET-OS osteosarcoma cohort were used to analyze the expression correlations of CSNK1A1, CBX4, and YAP1 with key components of the PI3K/MAPK pathways, including PIK3CA, AKT1, MAPK1, and BRAF. After extracting the normalized expression data of the target genes from the TARGET-OS cohort, Pearson correlation analysis was performed to assess the linear relationships between gene pairs, and the correlation coefficient (r) and significance level (P value) were calculated.

A correlation matrix bubble plot was generated to present the overall correlation structure, in which both color intensity and bubble size represented the strength of correlation. In addition, scatter plots were drawn for representative gene pairs, with linear regression fitting curves and 95% confidence intervals overlaid to visualize the correlation trends. A two-sided P < 0.05 was considered statistically significant.

Prediction of Ubiquitination and SUMOylation Sites

The FASTA sequences of CBX4 (Gene ID: 8535) and YAP1 (Gene ID: 10413) proteins were retrieved from NCBI (https://www.ncbi.nlm.nih.gov/). Ubiquitination sites on CBX4 were predicted using the GPS-Uber (https://gpsuber.biocuckoo.cn/online.php) database (Threshold = Medium) and the MusiteDeep (https://www.musite.net/) database (PTMscores = 0.5). SUMOylation sites on YAP1 were predicted using GPS-SUMO (https://sumo.biocuckoo.cn/index.php) (SUMOylation = Medium), MusiteDeep (PTMscores = 0.5), and abcepta-SUMOplot™ (https://www.abcepta.com/sumoplot) (high probability). Predicted sites from different databases were intersected using the R package “VennDiagram” to identify key modification sites.

Establishment of OS Xenograft tumor and lung metastasis models

Female BALB/c nude mice aged 6–8 weeks and weighing 18–22 g (Nanjing Experimental Animal Center, China) were housed under SPF conditions at 22 ± 2 °C and 50%–60% humidity with a 12 h light/dark cycle. Food and water were provided freely. For the subcutaneous xenograft model, MG-63 cells (5 × 10⁶) suspended in 100 µL PBS were injected into the right flank. Tumor volume was measured every three days, and treatment began when tumors reached approximately 100 mm³.

For the lung metastasis model, MG-63 cells stably expressing firefly luciferase (D2102, Beyotime, Shanghai, China) were injected via the tail vein at 1 × 10⁶ cells/100 µL PBS. Treatment started three days later. Bioluminescence signals were detected after intraperitoneal injection of luciferin (150 mg/kg, ST196, Beyotime, Shanghai, China) using the IVIS Lumina II system (PerkinElmer, USA).

Mice that successfully established models were randomly divided into six groups (n = 6 per group): Control group (tail vein injection of equivalent volume saline), PP group (tail vein injection of 0.5 mg/kg PP), NP@PP group (tail vein injection of 0.5 mg/kg NP@PP, dosage based on PP concentration contained in NPs), oe-NC group (model established with MG-63 cells transduced with control lentivirus + tail vein injection of 0.5 mg/kg NP@PP), oe-CBX4 group (model established with oe-CBX4 MG-63 cells + tail vein injection of 0.5 mg/kg NP@PP), and oe-YAP1 group (model established with oe-YAP1 MG-63 cells + tail vein injection of 0.5 mg/kg NP@PP). Treatments were administered once every three days for a total of 27 days. Mouse body weight, activity, and survival status were monitored and recorded. Tumor size was measured every three days using a caliper, recording the longest diameter (L) and the shortest diameter (W), and tumor volume (V) was calculated using the formula: V = 0.5 × L × W². Tumor growth curves were plotted. At the end of the experiment, mice were euthanized, tumor tissues were harvested, weighed using an electronic balance, and photographed.

Pulmonary fluorescence signals were monitored in real time using the IVIS Lumina II system. At the endpoint, lung tissues were harvested, washed with PBS, and photographed. Visible metastatic nodules on the lung surface were counted manually, and ex vivo fluorescent images of lung tissues were captured with the IVIS system to quantify metastatic signal intensity [34].

Biodistribution

Tumor-bearing nude mice were divided into groups with six mice each. On day 27, mice received 100 µL Cy5.5-PP, Cy5.5-NP@PP, or Cy5.5-PBS via tail vein injection. At 1, 12, and 24 h after injection, mice were anesthetized with 1.5% isoflurane in an O₂/N₂ mixture at 1:2. Fluorescence accumulation in tumors and major organs, including the heart, liver, spleen, lungs, kidneys, and eyes, was detected using an IVIS^® Lumina II imaging system (Xenogen) [35].

Hematoxylin and Eosin (H&E) staining

Major organs, including the heart, liver, spleen, lungs, kidneys, and bone, were fixed in 10% neutral formalin (F5554, Sigma-Aldrich, USA) for 48 h, embedded in paraffin, and cut into 4 μm sections. After deparaffinization and rehydration, sections were stained with an H&E Staining Kit (C0105S, Beyotime, China). After mounting, pathological changes such as cell morphology, mitosis, and necrosis were observed under a light microscope.

Immunohistochemical (IHC) staining

IHC staining was performed to detect the expression of CK1α, p16^INK4a, and p21^Cip1 in tumor tissues. Following deparaffinization, tissue sections underwent antigen retrieval in sodium citrate buffer (P0083, pH 6.0, Beyotime, China) under high pressure for 10 min. Endogenous peroxidase activity was blocked by incubation with 3% H₂O₂ for 10 min. After blocking, sections were incubated overnight at 4 °C with the following primary antibodies: CK1α (ab185629, 1:200, Abcam, USA), p16^INK4a (ab270058, 1:500, Abcam, UK), and p21^Cip1 (ab109520, 1:200, Cell Signaling Technology, USA). The next day, sections were incubated with HRP-conjugated secondary antibody (ab6721, 1:1000, Abcam, USA) for 1 h. DAB substrate was used for color development, and nuclei were counterstained with hematoxylin. Images were captured using a light microscope, and protein expression levels were quantified using ImageJ software.

Bone mineral density measurement

Bone mineral density was measured by dual-energy X-ray absorptiometry (DEXA). Briefly, this method uses two X-ray sources to scan the bone. The higher the bone density, the greater the absorption of X-rays and the lower the amount of radiation reaching the detector. Computer-based analysis was used to calculate the radiation received by the detector and convert it into bone mineral density values expressed in g/cm².

Safety evaluation

Mouse health status was evaluated by monitoring body weight throughout the experiment. At the endpoint, major organs, including the heart, liver, spleen, lungs, and kidneys, were collected for H&E staining to assess tissue morphology. Serum levels of CK, LDH, ALT, AST, CREA, and urea were measured using an automated biochemical analyzer (Hitachi 7180, Japan) to evaluate hepatic and renal function.

Statistical analysis

All data were derived from at least three independent experiments and are presented as mean ± standard deviation (Mean ± SD). Comparisons between the two groups were conducted using independent-sample t-tests. Comparisons among three or more groups were performed using one-way analysis of variance (ANOVA). If ANOVA results indicated significant differences, Tukey’s HSD test was used for multiple comparisons. The Mann-Whitney U test or Kruskal-Wallis H test was applied for non-normally distributed or heteroscedastic data. Statistical analyses were performed using GraphPad Prism 9.5.0 (GraphPad Software, Inc.) and R version 4.2.1 (R Foundation for Statistical Computing). A two-tailed p-value < 0.05 was considered statistically significant; p ≥ 0.05 was considered not statistically significant.

Results

Successful preparation of pH-responsive peptide NP@PP

PP is an FDA-approved drug that selectively activates the CK1α isoform and exhibits antitumor activity in various cancer models [24]. Based on the acidic characteristics of the tumor microenvironment, we designed and synthesized pH-responsive self-assembling peptide NP@PP. These NPs were composed of the block copolymer DDA-P(DLG-co-LH)-b-P(LG-g-ALN), wherein the histidine fragments endowed the system with pH responsiveness by undergoing protonation under acidic conditions, leading to micelle swelling and enhanced drug release. Additionally, the ALN groups provided biomineralization capability, facilitating CaP deposition and enhancing adhesion and uptake by tumor cells. DDA-P(DLG-co-LH)-b-P(LG-g-ALN) was an amphiphilic polymer that self-assembled into core-shell structured NPs. During the self-assembly process, the hydrophobic segments (DLG, DLG-co-LH) formed the inner core, while the hydrophilic PLG-g-ALN chains formed the outer shell. Due to these properties, PP, a hydrophobic small-molecule drug, was embedded into the hydrophobic core during self-assembly to achieve efficient drug loading. This design not only improved the delivery efficiency of PP within the tumor microenvironment but also provided a novel carrier platform for the precise treatment of OS (Fig. 1A).

Fig. 1.

Preparation and characterization of peptide NPs.(A) Schematic diagram of the synthesis steps of NP@PP, created in BioRender; (B) Particle size distribution and morphology of NP under pH 5.5, pH 6.5, and pH 7.4 conditions detected by TEM and DLS (Scale bar = 500 nm); (C) Zeta potential of NP under pH 5.5, pH 6.5, and pH 7.4 conditions measured by DLS; (D) Zeta potential of NP and NP@PP measured by DLS; (E) Kinetics of PP release from NP@PP under pH 5.5, pH 6.5, and pH 7.4 conditions in the pH-responsive release experiment; (F) Particle size distribution and morphology of NP@PP under pH 5.5, pH 6.5, and pH 7.4 conditions detected by TEM and DLS (Scale bar = 500 nm); (G) FTIR spectra of NP, NP + ALN, and NP@PP, showing characteristic P = O/P–O stretching peaks in the 1100–900 cm⁻¹ region and their changes after Ca²⁺ coordination; (H) XRD patterns of NP and NP@PP, showing the predominantly amorphous diffraction profile of the crosslinked nanoparticles. Each experiment was repeated three times. * indicates comparison between two groups, *P < 0.05, **P < 0.01, ****P < 0.0001, ns indicates no statistically significant difference

DLS and TEM results revealed that under pH 7.4 conditions, the peptide NPs exhibited a uniform spherical morphology, with an average particle size of 89.08 ± 2.60 nm. Under pH 6.5 conditions, the particle size increased to 283.87 ± 16.49 nm and further expanded to 351.29 ± 18.03 nm at pH 5.5, indicating excellent pH-responsive characteristics of the NPs (Fig. 1B). Moreover, the zeta potential of the NPs showed minimal changes under different pH conditions (Fig. 1C). Further DLS analysis demonstrated that at pH 7.4, the zeta potential of NP was − 25.80 ± 1.75 mV, whereas, after PP loading, the zeta potential of NP@PP shifted to -31.80 ± 4.02 mV, suggesting that drug loading altered the surface charge (Fig. 1D).

HPLC analysis showed that the drug loading content of NP@PP was 12.7 ± 0.4%, with an encapsulation efficiency as high as 85.2 ± 2.3%, indicating high drug-loading capacity and stability of the NPs. NP@PP exhibited distinct drug release behaviors under different pH conditions in pH-responsive drug release experiments. Under the neutral condition (pH 7.4), NP@PP showed a slow drug release rate, whereas, under acidic conditions (pH 6.5), the release rate significantly increased and further accelerated under pH 5.5 (Fig. 1E). These results indicated that NP@PP could efficiently release the loaded drug PP in the mildly acidic tumor microenvironment while maintaining stability under normal physiological conditions.

Dynamic monitoring of particle size changes of NP@PP at different pH values by DLS and TEM further confirmed these findings. At pH 7.4, the particle size of NP@PP remained stable at 122.41 ± 5.43 nm. Under pH 6.5 conditions, the particle size increased to 312.90 ± 18.53 nm, and at pH 5.5, it further expanded to 382.01 ± 19.79 nm. Notably, under identical pH conditions, the particle size of NP@PP was larger than that of NP alone (Fig. 1F). These data demonstrated that PP was successfully loaded into the NPs and that NP@PP exhibited significant pH-responsive properties, facilitating accelerated drug release under the mildly acidic tumor environment. In the FTIR spectra, the introduction of the ALN side chains generated characteristic stretching vibration peaks of P = O/P–O in the 1100–900 cm⁻¹ region. After the addition of Ca²⁺, these characteristic peaks showed peak shifting and broadening, indicating strong coordination interactions and thereby confirming the occurrence of Ca²⁺ chelation (Fig. 1G).

In the XRD analysis, the crosslinked NP@PP mainly exhibited a broad diffraction pattern characteristic of an amorphous structure. However, owing to the enrichment of local calcium and phosphonate groups, the system may retain the potential to induce in situ calcium phosphate crystallization in the tumor microenvironment (Fig. 1H).

These results confirmed that we successfully prepared pH-responsive, biomineralization-induced self-assembling peptide NPs, achieving efficient loading and encapsulation of the PP.

NP@PP exhibited excellent cellular uptake and biocompatibility

We further evaluated the cellular uptake capacity of NP@PP. First, we assessed the differences in cell membrane adhesion of NP@PP under different pH conditions. The results showed that FITC-labeled NP@PP readily adhered to the membranes of MG-63 cells, with a markedly higher adhesion rate at pH 6.5 than at pH 7.4 (Fig. 2A-B).

Fig. 2.

Cellular uptake and biosafety evaluation of NP@PP.(A) Fluorescence microscopy showing cell membrane adhesion of NP@PP to MG-63 cells under different pH conditions (Scale bar = 12 μm); (B) Relative quantification of red and green fluorescence colocalization intensity; (C) IF staining showing cellular uptake of NP@PP under different pH conditions (Scale bar = 50 μm); (D) CCK-8 assay evaluating the cytotoxicity of NP@PP against hFOB 1.19 cells; (E) Live/Dead staining assessing the toxicity of NP@PP toward hFOB 1.19 cells (Scale bar = 50 μm); (F) Hemolysis assay evaluating the hemolysis rate of NP@PP at different concentrations. Each experiment was independently performed in triplicate. **** indicates p < 0.0001; ns indicates no statistically significant difference

Next, we assessed the ability of NP@PP to induce biomineralization deposition in MG-63 cells. SEM imaging and Ca element mapping revealed that CaP deposition was more pronounced under acidic conditions (Figure S1A-B). These results suggest that, after uptake by MG-63 cells, NP@PP accumulated on the tumor cell surface and initiated a biomineralization cascade. Calcein staining further confirmed this effect. Compared with the control group, NP@PP markedly enhanced Calcein fluorescence, with stronger signals at pH 6.5 than at pH 7.4 (Figure S1C–D).

Because biomineralization can alter cellular properties, we further examined its effects on MG-63 cell behavior. NP@PP treatment increased cell stiffness compared with the control group, and this effect was more pronounced at pH 6.5 than at pH 7.4 (Figure S1E). Since tumor cells generally display higher membrane fluidity than normal cells, and this feature is closely linked to malignancy [36], we next assessed membrane fluidity. NP@PP significantly reduced MG-63 membrane fluidity under acidic conditions compared with neutral conditions (Figure S1F). These findings indicate that NP@PP exerts stronger cytotoxic effects in an acidic microenvironment, supporting its potential for selective tumor cell killing in vivo.

Furthermore, CCK-8 assays demonstrated that NP@PP had no significant effect on the viability of human osteoblasts (hFOB 1.19) within a concentration range of 0–100 µg/mL (Fig. 2D). Live/dead cell staining further confirmed that nearly all hFOB 1.19 cells in the NP@PP group were viable, with no obvious dead cells observed (Fig. 2E). Hemolysis assays showed no significant hemolytic response for NP@PP at concentrations ranging from 0 to 200 µg/mL (Fig. 2F). These results indicated that NP@PP possessed high cellular uptake efficiency and excellent biocompatibility.

In conclusion, NP@PP demonstrated outstanding cellular uptake efficiency while maintaining good biocompatibility in vitro.

NP@PP significantly inhibited OS cell proliferation and migration in vitro

Next, we further investigated the effects of NP@PP nanoparticles on the biological behaviors of OS cells, including MG-63 cells (Fig. 3) and U2OS cells (Figure S2).

Fig. 3.

Effects of NP@PP on MG-63 cell proliferation, migration, and apoptosis.(A) Schematic diagram of the experimental workflow, created in BioRender; (B) CCK-8 assay assessing MG-63 cell viability in different treatment groups; (C) EdU staining detecting DNA synthesis levels in MG-63 cells under different treatments (Scale bar = 50 μm); (D) Colony formation assay evaluating the clonogenic ability of MG-63 cells (Scale bar = 100 μm); (E) Western blot analysis of the proliferation markers Ki67 and PCNA in MG-63 cells; (F) Wound healing assay assessing MG-63 cell migration ability (Scale bar = 100 μm); (G) Transwell assay evaluating MG-63 cell migration (Scale bar = 50 μm); (H) Transwell assay evaluating MG-63 cell invasion (Scale bar = 50 μm); (I) Western blot analysis of migration-related markers MMP-9 and E-cadherin in MG-63 cells; (J) Flow cytometry analysis of apoptosis rates in MG-63 cells; (K) Flow cytometry analysis of cell cycle distribution in MG-63 cells. All experiments were repeated three times. * indicates p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001

As shown in Fig. 3A, the experimental design was first outlined. CCK-8 assays revealed that both PP and NP@PP significantly reduced cell viability compared with the control group, with NP@PP showing a stronger inhibitory effect (Fig. 3B). Consistently, EdU staining demonstrated a marked decrease in EdU-positive cells after PP or NP@PP treatment, and this reduction was more evident in the NP@PP group (Fig. 3C). Colony formation assays further supported these results, showing the lowest colony number in the NP@PP group, followed by the PP group (Fig. 3D). WB analysis also showed that Ki67 and PCNA expression was decreased by PP and NP@PP, with a greater reduction observed after NP@PP treatment (Fig. 3E).

Migration and invasion assays further showed that NP@PP markedly suppressed the metastatic phenotype of MG-63 cells. In the wound healing assay, both PP and NP@PP significantly inhibited cell migration compared with the Control group, and NP@PP produced a stronger inhibitory effect than PP (Fig. 3F). Similarly, Transwell assays showed that both migration and invasion abilities were reduced after PP or NP@PP treatment, with NP@PP exhibiting a more prominent suppressive effect (Fig. 3G-H). In addition, WB results demonstrated that MMP-9 expression was significantly downregulated, whereas E-cadherin expression was significantly upregulated in the PP and NP@PP groups, with more pronounced changes in the NP@PP group (Fig. 3I). These findings indicate that NP@PP inhibits the migration and invasion of MG-63 cells more effectively than free PP.

Flow cytometric analysis further indicated that both PP and NP@PP significantly increased the rates of early and late apoptosis compared with the Control group, and NP@PP induced apoptosis more effectively than PP (Fig. 3J). Cell cycle analysis showed that the proportions of cells in the G0/G1 phase were significantly increased, whereas the proportion of S-phase cells was significantly decreased in both treatment groups, with a stronger effect observed in the NP@PP group (Fig. 3K). These results suggest that NP@PP induces apoptosis and cell cycle arrest more potently than free PP.

Similar inhibitory effects were also observed in U2OS cells (Figure S2). NP@PP treatment significantly suppressed cell proliferation and migration compared with the Control and PP groups, further supporting the inhibitory effect of NP@PP on OS cells in vitro.

In summary, NP@PP exhibited significant inhibitory effects on OS cell proliferation, migration, and invasion, with efficacy markedly superior to free PP. This was characterized by more efficient induction of apoptosis and cell cycle arrest. These findings suggest that NP@PP enhanced the therapeutic efficacy of the PP, leading to more potent suppression of OS cell growth.

RNA-seq analysis of the effects of NP@PP on OS cell senescence, proliferation, and migration

RNA-seq was performed to further explore how NP@PP affects the biological behavior of MG-63 cells. Using |log₂FC| > 0.5 and P < 0.05 as thresholds, 128 DEGs were identified between the control and NP@PP groups, including 77 upregulated and 51 downregulated genes in NP@PP-treated cells (Figure S3A-B). GO and KEGG analyses showed that these genes were mainly enriched in cellular senescence, proliferation, and cell-cycle-related processes. Several tumor-associated pathways were also enriched, including PI3K-Akt, FoxO, HIF-1, MAPK, and JAK-STAT signaling pathways, which are closely related to tumor growth, migration, and cell-cycle control (Fig. 4A-B).

Fig. 4.

Transcriptomic analysis of the effects of NP@PP on MG-63 cells.(A-B) GO and KEGG enrichment analysis of 128 DEGs associated with cellular senescence, proliferation, migration, and cell cycle; (C-D) RNA-seq expression profiles of 20 genes enriched in functions related to senescence, proliferation, migration, and cell cycle

To assess the relationship between the CK1α–CBX4–YAP1 axis and classical OS-related pathways, we analyzed correlations among CK1α/CSNK1A1, CBX4, YAP1, and PI3K/MAPK core kinases, including PIK3CA, AKT1, MAPK1, and BRAF, in the TARGET-OS cohort. YAP1 was positively correlated with PIK3CA, MAPK1, and BRAF, suggesting possible crosstalk between Hippo-YAP signaling and PI3K/MAPK pathways in OS. CSNK1A1 showed weak positive correlations with MAPK1 and BRAF, whereas CBX4 had generally weak associations with PI3K/MAPK components. In addition, CSNK1A1 showed a certain correlation with YAP1, while CBX4 was only weakly correlated with YAP1, implying that CK1α may influence YAP1 activity indirectly through regulation of CBX4 stability (Figure S4).

Overall, these results indicate that the CK1α–CBX4–YAP1 regulatory axis is associated to some extent with the classical PI3K/MAPK proliferative signaling pathway, but its core regulatory mechanism appears to rely more on post-translational modification–mediated Hippo-YAP regulation, particularly through ubiquitination and SUMOylation. These findings further support the potential uniqueness of this pathway in OS.

Furthermore, we analyzed the expression patterns of genes enriched in these functional categories. The results showed that 20 genes were enriched in senescence, proliferation, migration, and cell cycle-related functional categories, with their expression profiles presented in Figs. 4C-D. The PPI network indicated complex interactions among these genes (Figure S3C). Specifically, NP@PP treatment markedly reduced the expression of proliferation- and cell-cycle-related genes, including CDK2, CDK6, and PCNA, as well as migration- and invasion-associated genes such as VEGFA, CXCL8, MMP9, and EGFR. The pro-survival gene WT1 was also downregulated. In contrast, PTEN and the senescence-associated markers IL1B and IL6 were significantly increased after NP@PP treatment (Figs. 4C-D), suggesting that NP@PP treatment promoted senescence, inhibited proliferation and migration, and induced cell cycle arrest in MG-63 cells.

The above transcriptomic data analysis indicated that NP@PP significantly promoted OS MG-63 cell senescence, inhibited their proliferation and migration, and induced cell cycle arrest.

NP@PP Promoted CBX4 degradation and inhibited YAP1 SUMOylation to induce cellular senescence

We investigated how NP@PP regulated CBX4 ubiquitination and YAP1 SUMOylation (Fig. 5A). Co-immunoprecipitation and WB analyses showed that, compared with the Control group, CBX4 ubiquitination was increased in both the PP and NP@PP groups, accompanied by a marked decrease in CBX4 protein expression; this effect was more pronounced in the NP@PP group (Fig. 5B-C).

Fig. 5.

Regulation of CBX4 ubiquitination, YAP1 SUMOylation, and cellular senescence by NP@PP.(A) Schematic diagram of the experimental workflow, created in BioRender; (B) Co-IP analysis of CBX4 ubiquitination and Western blot analysis of CK1α and CBX4 expression in MG-63 cells; (C) Quantification of CK1α and CBX4 protein levels; (D) Co-IP analysis of YAP1 SUMOylation and Western blot analysis of YAP1 expression in the cytoplasm and nucleus; (E) Quantification of cytoplasmic and nuclear YAP1; (F) Immunofluorescence staining showing the nuclear-cytoplasmic distribution of YAP1; (G) Co-IP analysis of the interaction between CBX4 and CK1α in MG-63 cells; (H) β-galactosidase staining; (I) Western blot analysis of p16^INK4a and p21^Cip1; (J) Flow cytometric analysis. Each experiment was repeated three times. * indicates P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001

Further examination of YAP1 SUMOylation showed that, compared with the Control group, YAP1 SUMOylation was decreased in the PP group and further reduced in the NP@PP group (Fig. 5D). WB analysis also showed that nuclear YAP1 levels were significantly decreased, whereas cytoplasmic YAP1 levels were increased, in the PP and NP@PP groups, with more prominent changes in the NP@PP group (Fig. 5D-E). Immunofluorescence showed that YAP1 was mainly distributed in the nucleus in the Control group, whereas after CBX4 degradation in the NP@PP group, nuclear YAP1 localization was markedly reduced and cytoplasmic distribution was increased. In addition, CBX4 degradation in the PP and NP@PP groups reduced the colocalization of CBX4 and YAP1 (Fig. 5F). Co-IP analysis further showed a direct interaction between CK1α and CBX4 (Fig. 5G). These findings indicate that NP@PP enhances the ubiquitination-dependent degradation of CBX4, markedly lowers CBX4 protein levels, suppresses YAP1 SUMOylation, and reduces YAP1 nuclear localization.

To determine whether NP@PP induces cellular senescence, β-galactosidase staining and senescence marker detection were performed. Compared with the control group, PP and NP@PP increased the proportion of β-galactosidase-positive cells, with the strongest effect observed in the NP@PP group. In contrast, Flavopiridol showed no obvious difference from the control group (Fig. 5H). WB further showed that NP@PP markedly elevated p16^INK4a and p21^Cip1 levels compared with both control and PP treatment, whereas Flavopiridol had little effect (Fig. 5I). Flow cytometry also revealed an increased senescent cell population after NP@PP treatment, among which some cells were Annexin V-positive, indicating a combined senescence-apoptosis phenotype (Fig. 5J).

In summary, NP@PP significantly enhanced CBX4 ubiquitination and degradation, thereby reducing intracellular CBX4 protein levels, inhibiting YAP1 SUMOylation, decreasing YAP1 nuclear localization, and ultimately inducing cellular senescence while upregulating the senescence-associated markers p16^INK4a and p21^Cip1. These findings revealed how NP@PP regulated cellular senescence through the CBX4-YAP1 signaling axis.

To further verify the effects of NP@PP in another OS cell line, we performed parallel experiments in U2OS cells (Figure S5). Co-immunoprecipitation and WB analyses showed that, compared with the Control group, CBX4 ubiquitination was increased in both the PP and NP@PP groups, accompanied by decreased CBX4 expression and increased CK1α expression, with more pronounced changes in the NP@PP group (Figure S5A-B).

Further analysis showed that YAP1 SUMOylation was reduced after PP treatment and was further decreased in the NP@PP group (Figure S5C). At the same time, WB analysis of nuclear and cytoplasmic proteins showed that nuclear YAP1 expression was significantly decreased, whereas cytoplasmic YAP1 expression was increased in the PP and NP@PP groups, with a stronger effect in the NP@PP group (Figure S5C-D). Immunofluorescence staining further showed that the colocalization of CBX4 and YAP1 was reduced after PP or NP@PP treatment, especially in the NP@PP group (Figure S5E). Co-IP analysis also confirmed the interaction between CK1α and CBX4 in U2OS cells (Figure S5F).

Moreover, β-galactosidase staining showed that NP@PP significantly increased the proportion of senescent cells (Figure S5G). WB analysis demonstrated that the senescence-related markers p16^INK4a and p21^Cip1 were markedly upregulated in the NP@PP group (Figure S5H). Flow cytometry further showed that apoptosis was also increased after NP@PP treatment (Figure S5I).

These results further support that NP@PP promotes CBX4 ubiquitination and degradation, inhibits YAP1 SUMOylation, reduces YAP1 nuclear localization, and induces a senescence-associated phenotype in OS cells.

Prediction of CBX4 Ubiquitination Sites and YAP1 SUMOylation Sites

To further clarify how NP@PP promotes CBX4 ubiquitination and thereby suppresses YAP1 SUMOylation, we first predicted the potential ubiquitination sites of CBX4. As described in the Methods, intersection analysis of the GPS-Uber and MusiteDeep databases identified six common candidate sites, namely K106, K114, K178, K205, K280, and K352 (Figure S6A-B), and the corresponding scores are listed in Table S3. Specifically, 12 ubiquitination sites were predicted by the GPS-Uber database and 12 sites were predicted by the MusiteDeep database (Figure S7A). We then predicted YAP1 SUMOylation sites. GPS-SUMO predicted nine SUMOylation sites, whereas MusiteDeep (Figure S7B) and the abcepta-SUMOplot™ tool (Figure S7C) each predicted two sites. Intersection analysis showed that only K280 was shared by all three tools (Figure S6C-D), and the corresponding scores are listed in Table S4. In addition, we identified a SUMO-interacting motif (SIM) at residues 54–58 of YAP1 (Figure S6D), which may be associated with YAP1 and downstream Hippo signaling activation.

These database prediction results suggested that NP@PP might promote CBX4 degradation by enhancing ubiquitination at K106, K114, K178, K205, K280, and K352, thereby inhibiting SUMOylation of YAP1 at the K280 site.

oe-CBX4 and oe-YAP1 reversed NP@PP-Induced YAP1 inhibition and cellular senescence

To confirm whether NP@PP regulates senescence and cellular phenotypes through CBX4 ubiquitination and YAP1 SUMOylation, CBX4- or YAP1-overexpressing MG-63 cells were established. RT-qPCR and WB analyses verified successful overexpression of CBX4 or YAP1 in the corresponding groups (Figure S8).

After NP@PP treatment, cellular functions were assessed using PBS as the control. CCK-8 and EdU assays showed that CBX4 or YAP1 overexpression enhanced cell viability and DNA synthesis compared with oe-NC under both PBS and NP@PP conditions, whereas NP@PP still reduced these effects relative to PBS (Figure S9A–B). Colony formation and WB analyses further showed that oe-CBX4 and oe-YAP1 increased colony numbers and upregulated Ki67 and PCNA, while NP@PP decreased these proliferation-related indicators (Figure S9C–D).

Similarly, wound-healing and Transwell assays demonstrated that CBX4 or YAP1 overexpression promoted MG-63 cell migration and invasion compared with oe-NC, whereas NP@PP markedly suppressed these phenotypes (Figure S9E–G). Consistently, oe-CBX4 and oe-YAP1 reduced E-cadherin and increased MMP-9 expression, while NP@PP showed the opposite regulatory pattern (Figure S9H).

Mechanistically, Co-IP and WB results showed that CBX4 ubiquitination and CBX4 protein levels were decreased in the oe-CBX4 group compared with oe-NC, while no obvious change was observed in the oe-YAP1 group (Fig. 6A–B). In contrast, YAP1 SUMOylation and nuclear YAP1 expression were increased in both oe-CBX4 and oe-YAP1 groups (Fig. 6C–D). Immunofluorescence further confirmed enhanced YAP1 nuclear localization and increased CBX4–YAP1 colocalization after CBX4 or YAP1 overexpression (Fig. 6E).

Fig. 6.

NP@PP promotes CBX4 ubiquitination to regulate YAP1 SUMOylation and induce cellular senescence.(A-B) Co-IP detection of CBX4 ubiquitination and Western blot analysis of CK1α and CBX4 in MG-63 cells; (C-D) Co-IP detection of YAP1 SUMOylation and Western blot analysis of YAP1 in the cytoplasm and nucleus; (E) Immunofluorescence staining showing the nuclear-cytoplasmic distribution of YAP1 in MG-63 cells (Scale bar = 25 μm); (F) β-galactosidase staining detecting senescence in MG-63 cells (Scale bar = 50 μm); (G) Western blot analysis of p16^INK4a and p21^Cip1 in MG-63 cells; (H) in vitro SUMOylation assay; (I) Co-IP detection of YAP1 SUMOylation in MG-63 cells transfected with His-tagged SUMO1-GG and HA-tagged YAP1, including wild-type (WT) and the lysine-to-arginine mutant YAP1-K280R, with vector or CBX4 overexpression. Each experiment was repeated three times. *P < 0.05, **P < 0.01, ***P < 0.001

To evaluate the roles of CBX4 and YAP1 in NP@PP-induced cellular senescence, β-galactosidase staining was performed. Compared with the oe-NC group, the proportion of blue-stained senescent cells was significantly decreased in both the oe-CBX4 and oe-YAP1 groups (Fig. 6F). WB analysis further showed significant downregulation of p16^INK4a and p21^Cip1 in the oe-CBX4 and oe-YAP1 groups (Fig. 6G). In the in vitro SUMOylation assay, no covalent SUMO1-YAP1 conjugate at approximately 212 kDa was detected when His-SUMO1, His-Ubc9, and YAP1 were transfected alone or in combination (lanes 1–3). However, after Flag-CBX4 was added (lane 4), this high-molecular-weight band became clearly detectable, indicating that CBX4 can directly catalyze YAP1 SUMO1 modification in a Ubc9-dependent manner. These findings confirm the function of CBX4 as a SUMO E3 ligase for YAP1 (Fig. 6H). In both MG-63 cells, cotransfection of His-SUMO1 and Flag-CBX4 markedly enhanced SUMO1 modification of HA-YAP1-WT, whereas the K280R single mutation almost completely abolished the CBX4-induced high-molecular-weight SUMO1-YAP1 band, with quantitative levels close to those of the control group. These results indicate that K280 is the key lysine residue required for CBX4-mediated YAP1 SUMOylation (Fig. 6I).

Similar regulatory trends were further confirmed in U2OS cells (Figure S10A-H). Specifically, compared with the oe-NC group, CBX4 ubiquitination and CBX4 expression were significantly reduced in the oe-CBX4 group, whereas the oe-YAP1 group showed no obvious change (Figure S10A-B). In contrast, YAP1 SUMOylation and nuclear localization were both increased in the oe-CBX4 and oe-YAP1 groups, accompanied by enhanced CBX4-YAP1 colocalization (Figure S10C-E). In addition, β-galactosidase staining and WB analysis showed that senescence-associated phenotypes, including the proportion of senescent cells and the expression of p16^INK4a^ and p21^Cip1^, were significantly reduced in the oe-CBX4 and oe-YAP1 groups (Figure S10F-G). Consistent with the findings in MG-63 cells, CBX4-mediated SUMOylation of YAP1 was also validated in U2OS cells (Figure S10H).

In summary, these results further confirmed that NP@PP promoted CBX4 ubiquitination and degradation, reduced YAP1 SUMOylation, and inhibited YAP1 nuclear localization, thereby inducing cellular senescence and suppressing the proliferation, migration and invasion of OS cells. oe-CBX4 and oe-YAP1 reversed the effects of NP@PP, confirming the critical roles of CBX4 and YAP1 in regulating OS cell behavior mediated by NP@PP.

NP@PP significantly inhibited OS growth and lung metastasis

To evaluate the inhibitory effects of NP@PP on OS growth and lung metastasis, we established subcutaneous OS xenograft tumor models and tail vein injection lung metastasis models using MG-63 cells, followed by treatment with NP@PP via tail vein injection (Fig. 7A). The subcutaneous tumor growth results showed that compared with the Control group, tumor volume was significantly reduced in the PP group and decreased in the NP@PP group (Figs. 7B-C). Tumor weight measurements were consistent with these findings (Fig. 7D).

Fig. 7.

In vivo, effects and molecular mechanisms of NP@PP inhibit OS growth and lung metastasis.(A) Schematic diagram of the animal experimental procedures, Created in BioRender; (B) Tumor volume measurements showing the inhibitory effect of NP@PP on the growth of MG-63 subcutaneous xenograft tumors; (C) Dissection images of subcutaneous xenograft tumors in each group; (D) Tumor weight measurements assessing the inhibitory effect of NP@PP on MG-63 xenograft tumor growth; (E) Fluorescence imaging of lung metastases in MG-63 models treated with NP@PP using an in vivo imaging system; (F) Quantification of lung metastatic nodules to evaluate the inhibitory effect of NP@PP; (G) H&E staining showing necrotic areas in subcutaneous tumor tissues, Scale bar = 50 μm; (H) IHC analysis of CK1α, p16^INK4a, and p21^Cip1 expression in MG-63 tumor tissues, Scale bar = 50 μm; (I) IF staining showing CBX4 and YAP1 colocalization in tumor tissues after NP@PP treatment, Scale bar = 25 μm. ** indicates P < 0.01, *** indicates P < 0.001, **** indicates P < 0.0001

In the lung metastasis model, bioluminescence imaging revealed strong fluorescent signals in the lungs of the Control group, indicating a higher number of metastatic foci. In the PP group, the lung fluorescence signal was moderately reduced, while the NP@PP group exhibited the weakest fluorescence signal (Fig. 7E). After dissection and enumeration of pulmonary metastatic nodules, the number of lung metastases was significantly lower in the PP group compared to the Control group and was further reduced in the NP@PP group (Fig. 7F). H&E staining of lung tissues revealed widespread metastatic lesions in the Control group, fewer metastatic foci in the PP group, and only a few scattered tumor cells with largely intact lung architecture in the NP@PP group (Fig. 7G). These results demonstrated that NP@PP exhibited superior efficacy compared to free PP in inhibiting OS growth and lung metastasis.

IHC staining showed that PP and NP@PP treatment increased CK1α, p16^INK4a, and p21^Cip1 expression in tumor tissues compared with the control group, with the strongest upregulation observed in the NP@PP group (Fig. 7H). IF analysis further revealed marked CBX4–YAP1 colocalization in control tumors, whereas this colocalization was reduced after PP treatment and was further weakened by NP@PP (Fig. 7I).

NP@PP biosafety was assessed by body weight monitoring, organ histology, and serum biochemical tests. No significant body weight changes were observed among the control, PP, and NP@PP groups during treatment (Figure S11A). H&E staining showed no evident pathological damage in the heart, liver, spleen, lung, or kidney after PP or NP@PP administration, with tissue structures similar to those in controls (Figure S11B). Serum levels of CK, LDH, ALT, AST, CREA, and urea also showed no significant differences among groups (Figure S11C–E). In addition, H&E staining and bone-density analysis indicated that NP@PP did not exert obvious toxicity toward normal osteogenic cells or bone tissues (Figure S11F-G).

To examine in vivo distribution, Cy5.5-PBS, Cy5.5-PP, and Cy5.5-NP@PP were injected into xenograft-bearing mice, followed by fluorescence imaging at 1, 12, and 24 h (Figure S12A). Cy5.5-PBS and Cy5.5-PP signals were only detectable at 1 h, whereas Cy5.5-NP@PP showed maximal accumulation at 24 h. At this time point, fluorescence was mainly enriched in tumors and lungs compared with other major organs (Figure S12B-C).

In summary, NP@PP nanoparticles significantly inhibited OS growth and lung metastasis in vivo, with markedly greater efficacy than the PP group. This effect was closely associated with the promotion of CBX4 ubiquitin-mediated degradation, suppression of YAP1 SUMOylation, and induction of cellular senescence. Meanwhile, NP@PP exhibited favorable biosafety, with no significant adverse effects observed in major organs or blood parameters. These findings further confirm the high efficacy and safety of NP@PP in suppressing OS progression.

oe-CBX4 and oe-YAP1 partially reversed the inhibitory effect of NP@PP on OS metastasis

By constructing MG-63 cells oe-CBX4 and oe-YAP1, we successfully established subcutaneous xenograft tumor and lung metastasis models in nude mice, followed by treatment with NP@PP in each group (Fig. 8A). Evaluation of subcutaneous tumor growth showed that compared with the oe-NC group, tumor volumes were significantly increased in both the oe-CBX4 and oe-YAP1 groups (Figs. 8B-C). Tumor weight measurements were consistent with these findings (Fig. 8D).

Fig. 8.

Effects of oe-CBX4 and oe-YAP1 on NP@PP-mediated inhibition of OS growth and metastasis.(A) Schematic diagram of the animal experimental procedures, Created in BioRender; (B) Tumor volume measurements showing the inhibitory effect of NP@PP on MG-63 subcutaneous xenograft tumor growth; (C) Dissection images of subcutaneous xenograft tumors from each group; (D) Tumor weight measurements assessing the inhibitory effect of NP@PP on MG-63 xenograft tumor growth; (E) Fluorescence imaging of lung metastases in the MG-63 model treated with NP@PP using an in vivo imaging system; (F) Quantification of lung metastatic nodules to evaluate the inhibitory effect of NP@PP; (G) H&E staining showing necrotic areas in MG-63 subcutaneous tumor tissues, Scale bar = 50 μm; (H) IHC analysis of CK1α, p16^INK4a, and p21^Cip1 expression levels in MG-63 tumor tissues, Scale bar = 50 μm; (I) IF staining showing colocalization of CBX4 and YAP1 in MG-63 tumor tissues following NP@PP treatment, Scale bar = 25 μm. ** indicates P < 0.01, *** indicates P < 0.001, **** indicates P < 0.0001

In the lung metastasis model, in vivo imaging revealed that the oe-NC group exhibited the weakest pulmonary fluorescence signal, indicating the fewest metastatic lesions, whereas the oe-CBX4 and oe-YAP1 groups exhibited significantly stronger fluorescence signals (Fig. 8E). Quantification of lung metastatic nodules after dissection showed that, compared with the oe-NC group, the oe-CBX4 and oe-YAP1 groups had significantly increased lung metastases (Fig. 8F). H&E staining of lung tissues further confirmed these observations, with the oe-NC group displaying the fewest metastatic lesions, while the oe-CBX4 and oe-YAP1 groups showed markedly increased and more widespread metastases (Fig. 8G). These results indicated that oe-CBX4 and oe-YAP1 could partially reverse the inhibitory effects of NP@PP on OS growth and lung metastasis.

IHC staining revealed that CK1α expression was highest in the tumor tissues of the oe-NC group, whereas CK1α levels were decreased in both the oe-CBX4 and oe-YAP1 groups. Compared with the oe-NC group, the expression levels of p16^INK4a and p21^Cip1 were also reduced in the oe-CBX4 and oe-YAP1 groups (Fig. 8H). IF staining showed that, in the oe-NC group, CBX4 and YAP1 were largely separated within tumor tissues, whereas in the oe-CBX4 and oe-YAP1 groups, colocalization of CBX4 and YAP1 was enhanced, with a significantly increased proportion distributed within the nuclei (Fig. 8I).

In summary, oe-CBX4 and oe-YAP1 partially reversed the inhibitory effects of NP@PP on OS growth and lung metastasis, as evidenced by increased tumor volume, greater numbers of lung metastatic nodules, and reduced levels of cellular senescence. These findings further confirmed that NP@PP exerted its anti-OS effects primarily by promoting CBX4 ubiquitination-mediated degradation and inhibiting YAP1 SUMOylation, thereby regulating OS cell senescence and suppressing metastasis.

Discussion

This study developed a novel pH-responsive biomineralization-induced peptide NP@PP system and systematically evaluated its mechanism of inducing senescence in OS cells by regulating the CBX4 ubiquitination and YAP1 SUMOylation regulatory axis. Unlike previous studies that primarily focused on OS cell apoptosis, proliferation pathways, or immune modulation [37], this research innovatively targeted cellular senescence as an anti-tumor strategy and incorporated post-translational modification regulation, proposing the potential of the “CK1α-CBX4-YAP1” axis in OS treatment. This expanded the research direction of molecular intervention strategies for OS and provided new insights into combining targeted nanotherapy with cellular senescence mechanisms.

In terms of nanodrug system construction, NP@PP exhibited favorable stability, targeting capability, and drug-release responsiveness, indicating promising therapeutic potential. Drug loading was achieved under specific weakly acidic conditions in vitro, where the charge states of PP and the polymeric scaffold were tuned to promote efficient co-assembly through strong electrostatic attraction. Calcium ions were then introduced at physiological pH (7.4), allowing alendronate (ALN) to form a robust coordination-crosslinked network with Ca²⁺ that effectively ‘locked’ the drug within the nanoparticle core. During in vivo release, however, the acidic tumor microenvironment (pH 6.5) and endosomal/lysosomal compartments (pH 5.0) act as structural disruptors. On the one hand, high concentrations of H⁺ directly compete with Ca²⁺ for the phosphate groups of ALN, leading to rapid dissociation of the surface calcium shell; on the other hand, protonation of the imidazole rings in.

the L-histidine block (pKa ~ 6.0) drives a hydrophobic-to-hydrophilic phase transition, resulting in core swelling and electrostatic repulsion. The synergy between acid-triggered decoordination and protonation-induced swelling enables precise on-demand drug release at tumor sites [34, 38].

In addition, this system showed markedly enhanced cell adhesion in the acidic tumor microenvironment, mainly due to a charge-reversal mechanism. As the pH decreases, protonation of the imidazole rings in the L-histidine segment rapidly reverses the nanoparticle surface charge from near-neutral/negative to positive, thereby strengthening electrostatic interaction with negatively charged tumor cell membranes and promoting cellular binding and endocytosis [39].

CBX4, as a SUMO E3 ligase, has been reported to exhibit dual roles in various tumor types [40]. In some cancers, CBX4 promotes tumorigenesis by regulating factors such as HIF-1α; in others, it acts as a tumor suppressor [41]. Our study revealed that high CBX4 expression in OS was closely associated with enhanced cell proliferation and migration and that its protein levels could be reduced through CK1α activation-induced ubiquitination-mediated degradation. By combining ubiquitination site prediction and immunoprecipitation assays, we confirmed that ubiquitination is the primary mechanism driving CBX4 degradation and, for the first time, clarified its critical role in OS cell senescence induction, thereby providing clear experimental support for targeting CBX4 therapeutically.

YAP1, a key regulatory protein in the Hippo signaling pathway, is modulated by multiple post-translational modifications, including phosphorylation, ubiquitination, and SUMOylation [42, 43]. Although the phosphorylation mechanisms of YAP1 have been investigated in OS, the role of SUMOylation in regulating its nuclear translocation and activity has remained insufficiently explored [40]. Our experiments demonstrated that CBX4 degradation led to a significant decrease in YAP1 SUMOylation, accompanied by reduced nuclear localization, thereby diminishing its transcriptional activity and inducing the expression of p16^INK4a and p21^Cip1, suggesting a negative regulatory role in cellular senescence. Moreover, we identified K280 as a key SUMOylation site of YAP1 through predictive analysis, offering an important target for future precision drug design.

The application of multi-omics strategies represented a major technical advantage of this study. Through RNA-seq, we systematically characterized the DEGs and enriched signaling pathways in OS cells following NP@PP intervention, thereby strengthening the reliability of functional findings such as cell cycle arrest and senescence pathways. We achieved a closed loop between mechanistic inference and experimental confirmation by integrating protein post-translational modification site database analysis with experimental validation. This multidimensional and multilayered research model enhanced the persuasive power of mechanistic explanations and provided a reference paradigm for the future development of nanosystems.

In terms of therapeutic strategy, this study employed the induction of cellular senescence instead of traditional cytotoxic killing approaches, thereby avoiding the toxicity associated with the widespread apoptosis of normal cells. Cellular senescence, as a durable and regulatable state, offered a new intervention window for OS, a tumor type characterized by high recurrence rates and strong drug resistance. We demonstrated that NP@PP significantly inhibited OS xenograft tumor growth and lung metastasis in vivo, prolonged mouse survival, and did not induce apparent toxicity, thus verifying the practical feasibility of this strategy. NP@PP exhibited improved tumor targeting, favorable safety, and sustained therapeutic efficacy compared to conventional chemotherapy, providing a better therapeutic option for clinical application.

This study successfully developed a pH-responsive self-assembled NP@PP and systematically elucidated its mechanism in regulating OS cell senescence through CBX4 ubiquitination and YAP1 SUMOylation. The findings expand our understanding of epigenetic regulatory networks in tumors at the mechanistic level and highlight the translational potential of functional nanomaterials in cancer therapy. However, several limitations remain, including restricted cell model diversity, a relatively small sample size in animal studies, and the need for further mechanistic exploration. Future studies should further validate the generalizability and safety of this strategy in clinical tissues and orthotopic tumor models and explore its combination with immunotherapy or genetic intervention. In addition, it should be noted that the current characterization of the nanoplatform mainly focused on physicochemical properties and drug-delivery behaviors directly related to biological function, whereas deeper materials-level analyses of possible bond changes and mineralization-associated structural features after Ca²⁺ chelation have not yet been performed. Therefore, the Ca²⁺-mediated local structural remodeling and mineralization-related characteristics of this system are still inferred mainly from functional results and design logic. In future work, FTIR/Raman, XRD, and other complementary characterization methods will be used to systematically verify Ca²⁺-related interactions and potential mineralization processes, so as to more comprehensively reveal the structural basis of this nanoplatform and its relationship to biological effects.

Conclusion

This study elucidated the mechanism of pH-responsive biomineralization-induced peptide self-assembled nanoparticles (NP@PP) loaded with a CK1α activator. NP@PP promoted CBX4 ubiquitin-mediated degradation via CK1α activation, suppressed YAP1 SUMOylation, and reduced its nuclear localization, thereby inducing cellular senescence and inhibiting proliferation, migration, and invasion of OS cells. In vivo, NP@PP effectively suppressed tumor growth and lung metastasis, exhibiting superior antitumor efficacy compared with PP alone. Both in vitro and in vivo results highlighted the critical roles of the CBX4–YAP1 axis in mediating NP@PP function, as overexpression of CBX4 or YAP1 partially reversed these effects. Additionally, NP@PP demonstrated favorable biosafety with no evident toxicity to major organs.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AFM

Atomic Force Microscopy

ANOVA

Analysis Of Variance

BLG NCA

γ-Benzyl-L-Glutamate N-Carboxyanhydride

Ca

Calcium

CaP

Calcium-Phosphate

CCK-8

Cell Counting Kit-8

Co-IP

Co-Immunoprecipitation

DDA

N-Dodecylamine

DEGs

Differentially Expressed Genes

DLS

Dynamic Light Scattering

DLG NCA

γ-Dodecyl-L-Glutamate N-Carboxyanhydride

DMSO

Dimethyl Sulfoxide

EDTA

Ethylenediaminetetraacetic Acid

EtAc

Ethyl Acetate

FBS

Fetal Bovine Serum

FITC

Fluorescein Isothiocyanate

GO

Gene Ontology

H&E

Hematoxylin And Eosin

HPLC

High-Performance Liquid Chromatography

IF

Immunofluorescence

IHC

Immunohistochemical

KEGG

Kyoto Encyclopedia Of Genes And Genomes

Mean ± SD

Mean ± Standard Deviation

MgSO₄

Magnesium Sulfate

MWCO

Molecular Weight Cutoff

N₂

Nitrogen

NP@PP

Nanoparticles Loaded with CK1α Activator Pyrvinium Pamoate

NPs

Nanoparticles

oe-CBX4

CBX4-Overexpressing

oe-YAP1

YAP1-Overexpressing

OS

Osteosarcoma

PBS

Phosphate-Buffered Saline

PP

CK1α Activator Pyrvinium Pamoate

PPI

Protein-Protein Interaction

RNA-seq

High-Throughput Transcriptomic Sequencing

SEM

Scanning Electron Microscopy

TEM

Transmission Electron Microscopy

THF

Tetrahydrofuran

TLH NCA

N(τ)-Trityl-L-Histidine N-Carboxyanhydride

Author contributions

R.Z., and Y.M. conceived and designed the study. R.Z., and Y.Z. performed the experiments. K.C. and C.C. analyzed the data. R.Z. wrote the manuscript. X.L revised the manuscript. All authors reviewed and approved the final version of the manuscript.

Funding

This work was supported by the Natural Science Foundation of Liaoning Province (Grant No. 2024-MSLH-585).

Data availability

All data generated or analyzed during this study are included in this article and/or its supplementary material files. Further inquiries can be directed to the corresponding author.

Declarations

Ethics approval and consent to participate

All animal experiments were approved by the Animal Ethics Committee of The First Hospital of China Medical University (Approval No. CMUKT2025182).Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.