Co-delivery of TRAIL DNA and salinomycin for cancer therapy via enhanced apoptosis and cancer stemness inhibition

Abstract

Cancer treatment faces challenges including drug resistance to intrinsic apoptosis pathways and the persistence of cancer stem cells (CSCs). The tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) selectively induces extrinsic apoptosis in tumor cells by binding to death receptors while sparing normal cells. TRAIL-based gene therapy has emerged as a promising solution to address the short half-life limitation of recombinant TRAIL proteins. To further enhance the therapeutic efficacy, combination strategies with TRAIL sensitizer salinomycin (Sali) appear to be particularly effective. Sali not only upregulates death receptor expression and reactive oxygen species (ROS) levels but also modulates CSC-related gene expression, contributing to its multifaceted antitumor activity. In this study, a micelle system was developed for co-delivery of TRAIL plasmids (pTRAIL) and Sali to tumor sites. γ-Polyglutamic acid (γ-PGA)-coating improved the in vivo stability profile of the co-delivery nanocomposites. The resulting co-delivery system (γ-PGA/Sali@P–S/pTRAIL) effectively induced apoptosis and suppressed CSC-associated gene expression in cervical cancer cells. Both in vitro and in vivo evaluations demonstrated the system's potent antitumor efficacy, highlighting its potential as a promising strategy for cervical cancer therapy.

---

Introduction

Cervical cancer is the second most common gynecological malignancy after breast cancer and has become one of the major global health concerns for women.¹ Its asymptomatic progression makes it highly insidious and often leads to diagnosis at advanced stages.² Current treatments primarily involve surgery combined with adjuvant therapies such as radiotherapy and chemotherapy.³ However, challenges such as drug resistance, poor response rates, recurrence, and metastasis remain unresolved.^4,5 Developing alternative therapeutic strategies beyond current treatment options is of significant clinical importance for effective cervical cancer management.⁶

Induction of tumor cell apoptosis is a key therapeutic mechanism of most treatments, including chemotherapy and radiotherapy. However, conventional treatments face two major limitations: systemic toxicity and resistance to intrinsic apoptosis pathways.^7,8 Therefore, there is an urgent need to identify safer and more effective therapeutic approaches. The tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL), a member of the TNF family, selectively induces extrinsic apoptosis in tumor cells while sparing normal cells.^9,10 The TRAIL exerts its effects by binding to death receptors (DRs), such as DR5, which are mainly expressed on the tumor cell surface, thereby activating downstream apoptotic pathways.¹¹ Thus, effective TRAIL therapy relies on two critical factors: efficient delivery of the TRAIL and sufficient DR expression. While the development of TRAIL receptor agonists remains challenging, current clinical trials primarily focus on the delivery of recombinant TRAIL proteins,¹² which are limited by their short half-life and ineffective delivery.^13,14 With the advancement of interdisciplinary fields, particularly materials science and nanobiotechnology, combination therapies have provided targeted approaches for cancer treatment. Nanoparticle-mediated TRAIL gene delivery represents a promising alternative, with non-viral DNA vectors offering a safer and more convenient option.^15–17 Various cationic polymers have been developed, including polyethyleneimine (PEI), poly(β-amino ester)s, poly[(2-dimethylamino)ethyl methacrylate], and dendrimers.^18,19 PEI is one of the most widely used polymeric vectors for gene delivery due to its ability to facilitate endosomal escape through the “proton sponge” effect.^20,21 For instance, PEI-grafted cationic liposomes demonstrate potent transfection efficacy.²² However, tumor cells can evade TRAIL-induced apoptosis by downregulating DR expression.²³ Polyether ionophore antibiotics have been reported to sensitize TRAIL therapy by upregulating DR5 and inhibiting FLIP.²⁴ Therefore, combining TRAIL delivery with effective sensitization strategies may enhance its antitumor efficacy and help overcome resistance, thereby improving therapeutic outcomes.

In addition, cancer stem cells (CSCs) are another major contributor to poor prognosis and are closely associated with therapy resistance.²⁵ CSCs represent a small subpopulation of tumor cells with self-renewal and differentiation capabilities, capable of generating heterogeneous tumor cell populations.^26,27 These cells play a pivotal role in tumor initiation, progression, recurrence, metastasis, and drug resistance.^28,29 Thus, targeting CSCs in cancer therapy is of great clinical significance.^30,31 Although our understanding of CSCs is still evolving, current evidence shows that their regulation involves multiple signaling pathways. The activation of stemness-related genes and oncogenes has been shown to transform original cells into tumorigenic CSCs.³² Targeted inhibition of aberrant Wnt/β-catenin signaling and c-Myc activation can effectively suppress proliferation and induce apoptosis in CSCs.³³ Moreover, additional pathways, such as STAT3 signaling, contribute significantly to CSC maintenance and function.^34,35 Precise modulation of these critical signaling networks offers a promising therapeutic strategy to impair CSC viability and potentially inhibit tumor progression and metastasis. These signaling pathways are also closely associated with TRAIL-based therapies.^36–38 Salinomycin (Sali), a polyether ionophore antibiotic, has been reported as one of the most potent inhibitors of CSC fate regulators.^30,39 The anti-CSC mechanisms of Sali involve the modulation of oncogenes and stemness-associated factors.⁴⁰ Although Sali may serve as an effective sensitizer for TRAIL-based cancer therapy,⁴¹ the mechanistic basis underlying their combinatorial action remains incompletely defined.

Based on these considerations, we constructed a PEI-based amphiphilic material for the co-delivery of Sali and TRAIL plasmids (pTRAIL) (Scheme 1). Surface modification with poly-γ-glutamic acid (γ-PGA) reduced the zeta potential and improved stability for in vivo delivery.^42–44 This study focuses on developing a dual-targeted therapeutic strategy that combines enhanced TRAIL-mediated apoptosis with the inhibition of CSC-related signaling pathways. While the expressed TRAIL activates the extrinsic apoptotic pathway in HeLa cells, Sali effectively sensitizes this process by upregulating DR5 expression. Beyond death receptor modulation, Sali exerts complementary anticancer effects by elevating reactive oxygen species (ROS) levels and coordinately downregulating CSC-associated signaling pathways, including Wnt/β-catenin, STAT3, and c-Myc, thereby synergizing with TRAIL therapy. Our co-delivery nanoplatform demonstrates remarkable therapeutic efficacy in murine HeLa xenograft models, achieving approximately 80% tumor growth inhibition.

Scheme 1 Schematic illustration of co-delivery of the pTRAIL and salinomycin for inducing apoptosis of cancer cells and CSC-associated gene downregulation.

Methods

Synthesis and characterization

Stearic acid (0.5 g, 1.76 mmol) was dissolved in 5 mL of N,N-dimethylformamide (DMF), followed by the addition of NHS (0.22 g, 1.9 mmol) and EDC (0.37 g, 1.9 mmol). The mixture was stirred at room temperature for 6 h. The resulting reaction solution was then added dropwise to a 10 mL DMF solution containing PEI_5k (4.4 g, 0.88 mmol). After stirring at room temperature for 12 h, the reaction mixture was dialyzed against methanol for one day (MWCO: 1 kDa). Finally, the dialysate was rotary evaporated and vacuum-dried to obtain the product PEI–stearyl copolymer (P–S). NMR spectroscopy was employed to obtain the 1H spectra of P–S using a Bruker 400 MHz instrument (Germany) under ambient conditions.

Preparation of nanoparticles

The lyophilized P–S material was reconstituted in absolute ethanol to prepare a 100 mg mL⁻¹ stock solution. A 50 mg mL⁻¹ ethanolic solution of DOPE and a 40 mg mL⁻¹ stock solution of salinomycin in absolute ethanol were similarly prepared. For nanoparticle formulation, 80 μL of P–S stock solution, 40 μL of DOPE solution, and 12 μL of salinomycin stock were combined and mixed. This organic phase was then slowly added dropwise into 2 mL of HEPES buffer under continuous magnetic stirring at room temperature. The ethanol was removed by rotary evaporation. P–S self-assembled into nanomicelles, with DOPE facilitating micelle formation and encapsulating salinomycin within the hydrophobic core, yielding Sali@P–S nanomicelles at a final concentration of 4 mg mL⁻¹. Unencapsulated salinomycin was removed by filtration through a 0.45 μm aqueous membrane filter.

The gel retardation of the P–S/pDNA nanocomplexes was examined. In brief, the nanocomplexes with different mass ratios were freshly prepared before use as mentioned above. The electrophoresis experiment was performed in 1% agarose gel containing 0.1 mL mL⁻¹ Gelred at room temperature in 1 × TAE buffer at 80 V for 45 min. Gel imaging was visualized using a chemiDOX system (Bio-Rad, USA). Particle sizes and zeta potentials were evaluated using a Zetasizer Nano ZS (Malvern Instruments Ltd, UK).

Drug loading capacity analysis

Vanillin solution was prepared by dissolving 2 g of vanillin in 40 mL of 95% ethanol under stirring, followed by the addition of 10 mL of concentrated hydrochloric acid. Salinomycin solution was mixed with vanillin solution at a 6 : 4 (v/v) ratio and reacted at 72 °C for 40 min, generating a salinomycin–vanillin derivative with characteristic absorption at 526 nm. 1 mg mL⁻¹ salinomycin stock solution (in absolute ethanol) was diluted to prepare standard solutions at concentrations of 0, 6.25, 12.5, 25, 50, 100, and 200 μg mL⁻¹. For the standard curve determination, 600 μL of each salinomycin standard solution was mixed with 400 μL of vanillin derivatization solution, followed by incubation at a constant temperature of 72 °C in a water bath for 40 minutes. After rapid cooling to room temperature, the absorbance of the reaction mixture was measured at 526 nm using a multimode microplate reader. For quantification of encapsulated salinomycin in Sali@P–S nanomicelles, 200 μL of the nanoparticle suspension was mixed with 800 μL of absolute ethanol, vortexed for 30 seconds, and subsequently sonicated for 10 minutes to ensure complete release of the encapsulated drug. From this solution, 600 μL was combined with 400 μL of vanillin solution and subjected to identical derivatization conditions (72 °C for 40 minutes in a water bath). Following rapid cooling to ambient temperature, the absorbance of the resulting derivative was similarly determined at 526 nm using the microplate reader. The drug-loading capacity was calculated using the following formula:

Drug-loading capacity (%) = weight of encapsulated drug/weight of nanoparticles × 100%

Transmission electron microscopy (TEM) analysis

For TEM observation, 10 μL of the nanoparticle suspension (0.5 mg mL⁻¹) was dropped onto a carbon-coated copper grid and allowed to air dry at room temperature. Subsequently, 10 μL of 1% (w/v) uranyl acetate solution was applied to the grid for negative staining. After 1 minute, excess stain was removed using filter paper. The grid was then rinsed with a drop of deionized water to remove residual stain and air dried completely before imaging.

Stability of nanocomposites and drug release in vitro

The stability of liposomes was investigated in PBS with 10% FBS. With shaking at 37 °C, the size of liposomes was measured at different times. The drug release profile in vitro was evaluated with a dialysis membrane (10–14 kDa) in PBS with 0.5% Tween 80. With shaking at 37 °C, the released Sali was quantified by vanillin derivatization at successive time points, and the cumulative release was calculated.

Cell uptake assay

The cellular uptake of the nanodrug was investigated by flow cytometry and confocal laser scanning microscope (CLSM). Coumarin 6 (C6) was loaded into nanomicelles (1% C6 in P–S micelles), forming γ-PGA/C6@P–S/pDNA (12 : 5 : 1), utilized for cellular uptake analysis. The cells were seeded on glass bottom dishes. After being incubated with γ-PGA/C6@P–S/pDNA in different pH conditions, the cells were observed by CLSM. For flow cytometry analysis, the HeLa cells were seeded in a 24-well plate at 5 × 10⁴ cells per well. After being incubated with γ-PGA/C6@P–S/pDNA in different pH conditions, the cells were detected by flow cytometry. The cells were stained with LysoTracker (Thermo) to investigate the uptake process of nanocomposites.

Transfection efficacy assay

The transfection in vitro was observed utilizing the enhanced green fluorescent protein (EGFP) plasmid (pEGFP) reporter gene. Briefly, the HeLa cells were seeded on the glass coverslips in a 12-well plate at 1 × 10⁵ cells per well. When the cells reached about 70% confluence, the different ratios of P–S/pEGFP were added to the medium and incubated for 48 h. Then the cells were observed by CLSM.

Reactive oxygen species (ROS) detection

The ROS generation was detected by a reactive oxygen species assay kit (Beyotime). Briefly, the HeLa cells were seeded in 6-well plates at 2 × 10⁵ cells per well and incubated for 24 h. Then, the cells were incubated with PBS, γ-PGA/Sali@P–S, γ-PGA/Sali@P–S/pTRAIL. The cells were treated with DCFH-DA according to the manufacturer's protocol. Then, the ROS generation was observed by flow cytometry.

Cell toxicity assay

The HeLa cells were seeded in 96-well plates and incubated for 24 h at 5 × 10⁴ cells per well before adding the nanodrugs. The five groups, control group, γ-PGA/P–S, γ-PGA/P–S/pTRAIL, γ-PGA/Sali@P–S, γ-PGA/Sali@P–S/pTRAIL were set up to study the cytotoxicity of nanodrugs. After incubated with the nanodrugs for 48 h, CCK-8 was added to the medium for a further 2 h incubation. The absorbance at 450 nm was detected by a microplate reader (Tecan).

The cell viability (%) was calculated by the following formula:

Cell viability (%) = (A_E − A_B)/(A_C − A_B) × 100%

A _E presented the absorbance of the samples with nanodrugs; A_C was the absorbance of the samples without nanodrugs; A_B was the absorbance of the blank control.

The HeLa cells were seeded at 2 × 10⁵ cells per well in 6-well plates and incubated for 24 h. After the confluence reached about 70%, the nanodrugs were added to incubate for another 48 h. Afterwards, the cells were harvested and stained with annexin V-FITC and propidium iodide (PI) according to the manufacturer's protocol. Finally, the cells were detected by flow cytometry.

Western blotting

Western blotting (WB) was utilized to determine the expression level of TRAIL, DR5, cleaved caspase-3, β-catenin, c-Myc, and STAT3. HeLa cells were seeded in 6-well plates at 2 × 10⁵ cells per well. After being cultured for 24 h, the cells were treated with different groups for 48 h. The cells were obtained for WB analysis.

Transwell migration assay and flow cytometry assay of CSCs

The metastatic potential of HeLa cells in response to Sali was evaluated using a Transwell assay (24-well format, 8 μm pore polycarbonate membrane). Specifically, 600 μL of complete medium containing predetermined concentrations of Sali was added to the lower chamber, while 100 μL of serum-free medium containing 5 × 10⁴ HeLa cells was seeded in the upper chamber. After 24 h incubation at 37 °C with 5% CO₂, the upper chambers were washed twice with Ca²⁺-free PBS, fixed with 4% paraformaldehyde for 30 min at room temperature, and air-dried. Migrated cells were stained with 0.1% crystal violet for 20 min, after which non-migratory cells on the upper membrane surface were mechanically removed using cotton swabs. Following three PBS washes, migrated cells were imaged under a microscope.

To assess the impact of Sali and TRAIL on CSC subpopulations, HeLa cells were seeded in 6-well plates at 2 × 10⁵ cells per well and cultured for 24 h. Cells were then treated with either Sali alone or Sali + γ-PGA/Sali@P–S/pTRAIL nanocomplex (containing 2 μg pTRAIL per well) for an additional 24 h. Harvested cells were trypsinized, washed with PBS, and stained with anti-CD44-FITC and anti-CD133-APC antibodies for 30 min at 4 °C in the dark. After resuspension in PBS supplemented with 2% FBS, CD44⁺CD133⁺ dual-positive subpopulations were analyzed using a flow cytometer.

Animal modelling and in vivo nanoparticle distribution

The xenografted tumor model was established via subcutaneous injection of the HeLa cells (1 × 10⁶ per mouse) in 100 μL serum-free RPMI-1640 culture medium. The IR780 loaded nanocomposites (2% IR780 in P–S micelles) including IR780@P–S/pTRAIL (5 : 1) and γ-PGA/IR780@P–S/pTRAIL (12 : 5 : 1) were prepared for in vivo imaging study. After being injected intravenously into the tumor-bearing mice with IR780@P–S/pTRAIL and γ-PGA/IR780@P–S/pTRAIL, the mice were imaged by in vivo imaging system under isoflurane anesthesia at different time intervals (0.5 h, 1 h, 2 h, 6 h, 4 h, 8 h, 24 h, and 48 h). The mice were euthanized at the last time point, and the tumors were isolated for ex vivo imaging. The fluorescence intensities were analyzed by Carestream MI software.

In vivo anti-tumor efficacy

The xenografted tumor model was established. When the tumor volume reached around 50 mm³, the mice were divided into four groups randomly when the tumors were palpable, which were the PBS control group, Sali@P–S/pTRAIL group, γ-PGA/Sali@P–S group, γ-PGA/P–S/pTRAIL group, and γ-PGA/Sali@P–S/pTRAIL group. Treatments were carried out 5 times intravenously every two days. Besides, the tumor size and body weight were recorded. The volume of the tumor was calculated by the following formula: V = L × W²/2, where L represented the length of the tumor and W represented the width of the tumor. The mice were euthanized when the tumor volume of the control group reached 2000 mm³ for further investigation. At the study endpoint, animals were euthanized via CO₂ asphyxiation, followed by dissection of dorsal transplanted tumors and major organs (heart, liver, spleen, lungs, and kidneys). Residual blood was removed by PBS rinsing, and tissues were photographed for documentation.

Tumor inhibition (%) = (PBS group tumor weight − Experimental group tumor weight)/PBS group tumor weight × 100%

Organ coefficient (%) = Organ weight/body weight × 100%

The tissues were fixed with 4% paraformaldehyde for 24 h for immunohistochemical staining. The paraffin-embedded tissues were cut into 4 μm thick sections and dewaxed, and rehydrated. The epitope retrieval was performed by boiling in Tris/EDTA buffer (pH 9.0), followed by treated with 3% hydrogen peroxide and 5% BSA solution. The tissues were incubated with primary antibodies at 4 °C overnight. After washing three times with PBS, the HRP-conjugated secondary antibodies were incubated at room temperature for 1 h. Tissues were finally stained with a 3,3′-diaminobenzidine (DAB, brown) kit according to the manufacturer's instruction and counterstained with hematoxylin (blue). After dehydration and mounted with coverslips, tissues were observed for the proportion and brown intensity of the protein-positive cells under a bright-field microscope.

Statistical analysis

Data are expressed as mean ± SD. Statistical differences were performed using one-way analysis of variance (ANOVA) (*p < 0.05, **p < 0.01, and ***p < 0.001).

Results and discussion

Preparation and characterization of nanocomplexes

The hydrophilic PEI_5K was conjugated with hydrophobic C18 chains to create an amphiphilic material capable of co-delivering genetic drugs and small-molecule hydrophobic drugs. This was achieved by first activating the carboxyl group of stearic acid, followed by its reaction with the primary amines of PEI, yielding the target product, PEI_5K-C18. The synthetic procedure is illustrated in Fig. S1, and the resulting product was characterized by ¹H NMR spectroscopy, with key functional groups identified (Fig. S2). Additionally, FT-IR spectroscopy was conducted to further confirm the successful conjugation. The characteristic amide bond peak at 1660 cm⁻¹ in the FT-IR spectrum of PEI_5K-C18 indicates that stearic acid was grafted onto PEI_5Kvia an amide linkage, confirming the successful synthesis of PEI_5K-C18 (Fig. S3). The PEI_5K-C18 conjugate spontaneously self-assembled with DOPE at a 2 : 1 (w/w) ratio to form nanomicellar structures, designated as P–S nanomicelles. Agarose gel electrophoresis demonstrated that P–S could effectively condense plasmid DNA (pDNA) at a mass ratio of 2 : 1 (Fig. 1A and Fig. S4). When P–S was complexed with pEGFP at various mass ratios and transfected into HeLa cells under serum-free conditions (with PEI_25K serving as the positive control), optimal GFP transfection efficiency was achieved at a 5 : 1 mass ratio, showing comparable performance to PEI_25K-mediated transfection (Fig. 1B).

Fig. 1 Transfection efficacy of P–S and characterization of nanocomposites. (A) The pDNA loading capacity of the nanocarrier was observed by agarose gel electrophoresis. (B) The plasmid EGFP (green) transfection efficacy of P–S into HeLa cells was observed by CLSM (scale bar = 50 μm). (C) The particle sizes of Sali@P–S/pDNA and γ-PGA/Sali@P–S/pDNA. (D) TEM images of Sali@P–S/pDNA and (E) γ-PGA/Sali@P–S/pDNA. (F) The zeta potentials of Sali@P–S/pDNA and γ-PGA/Sali@P–S/pDNA (mean ± SD; n = 3).

Based on the optimal ratio determined from transfection experiments, Sali@P–S was complexed with pTRAIL to prepare the nanocomposites Sali@P–S/pTRAIL. To shield the positive surface charge of the nanocomposites and improve in vivo applicability, the anionic polymer γ-PGA was coated onto the surface at a mass ratio of 2 : 1 (γ-PGA : Sali@P–S/pTRAIL), yielding γ-PGA/Sali@P–S/pTRAIL (12 : 5 : 1). Dynamic light scattering analysis revealed that Sali@P–S/pTRAIL exhibited a uniform particle size distribution centered at approximately 130 nm (Fig. 1C), which was corroborated by transmission electron microscopy (TEM) observations (Fig. 1D). Following γ-PGA coating, the particle size of γ-PGA/Sali@P–S/pTRAIL decreased to about 100 nm (Fig. 1C), likely due to charge compression induced by the negatively charged γ-PGA wrapping. TEM examination of γ-PGA/Sali@P–S/pTRAIL confirmed the formation of spherical nanoparticles (Fig. 1E). Zeta potential measurements demonstrated a complete surface charge reversal from highly positive (+45 mV) to negative (−30 mV) after γ-PGA coating (Fig. 1F), a modification expected to enhance colloidal stability for in vivo applications.

To evaluate the protective effect of γ-PGA coating on nanoparticles under serum-containing conditions, we investigated the P–S/pEGFP system at an optimal mass ratio of 5 : 1 (carrier to plasmid). Under these conditions, the γ-PGA-coated formulation demonstrated significantly improved performance. While naked P–S/pEGFP nanoparticles exhibited compromised transfection efficiency in the presence of serum, the γ-PGA-coated nanoparticles maintained enhanced transfection efficiency in HeLa cells (Fig. 2A). These results clearly indicate that the anionic γ-PGA coating effectively protects the cationic P–S/pEGFP nanocomposites from serum protein interference, thereby preserving their gene delivery capability. We next evaluated the drug loading capacity and stability of the nanocomposites. The P–S copolymer was first self-assembled with salinomycin (Sali) to form Sali@P–S nanomicelles, with the Sali encapsulation efficiency determined to be approximately 4.35% using the vanillin derivatization method (Fig. S5). Stability testing revealed that γ-PGA-coated nanoparticles maintained excellent colloidal stability in serum-containing media, whereas uncoated nanoparticles rapidly aggregated under serum conditions due to their strongly positive surface charge (Fig. 2B). We further investigated the pH-dependent drug release profile of the nanocomposites under different pH conditions (Fig. 2C). The free salinomycin control group released approximately 80% of the drug within 12 hours. In contrast, the γ-PGA/Sali@P–S nanoparticle system demonstrated pH-responsive release characteristics—sustained release under physiological pH (7.4) and accelerated release in acidic environments. This pH-sensitive behavior originates from the conformational changes of γ-PGA.⁴⁵ At neutral pH, γ-PGA maintains a compact random coil structure that forms a dense protective layer around the micelles. Under acidic conditions, however, it undergoes a conformational transition to an extended helical structure, which facilitates its detachment from the micelle surface. This structural rearrangement promotes drug release from the hydrophobic core into the aqueous medium after the nanocomposites are internalized by tumor cells.

Fig. 2 The behavior and cell uptake of nanocomposites in vitro. (A) The plasmid EGFP (green) transfection efficacy of P–S and γ-PGA coated nanocomposites into HeLa cells was observed by CLSM (scale bar = 50 μm). (B) The stabilities of Sali@P–S/pDNA and γ-PGA/Sali@P–S/pDNA in PBS with 10% FBS (mean ± SD; n = 3). (C) The release behavior of Sali from γ-PGA/Sali@P–S in different pH conditions. (D) CLSM images of HeLa cells incubated with nanocomposites in different pH conditions for 4 h. Green: nanocomposites were labeled by C6; blue: nucleus was stained with DAPI (scale bar = 20 μm). (E) Statistical analyses of cell uptake of nanocomposites in different pH conditions by flow cytometry (mean ± SD; n = 3). The p values were calculated by ANOVA with Tukey's test. *p < 0.05, **p < 0.01, ***p < 0.001.

Cell uptake and anti-tumor mechanism exploration

Due to the pH-responsive properties of γ-PGA, we investigated nanoparticle cellular uptake under different pH conditions using coumarin-6 (C6) as a fluorescent probe, substituting salinomycin in the nanocomposites. Under physiological conditions (pH 7.4), uncoated nanoparticles demonstrated high cellular uptake efficiency, attributed to their positive surface charge. In contrast, γ-PGA coating significantly reduced uptake due to charge shielding. Under acidic conditions (pH 6.0), which mimic the tumor microenvironment, enhanced cellular uptake was observed (Fig. 2D and E). This phenomenon correlates with the pH-dependent conformational transition of γ-PGA from a random coil to a helical structure, leading to partial detachment from the nanoparticle surface. The consequent exposure of underlying positive charges facilitates increased internalization by tumor cells.

Lysosomal escape is crucial for effective intracellular drug delivery, as lysosomes contain hydrolytic enzymes that can degrade encapsulated plasmids if the nanocomposites remain trapped, thereby compromising their antitumor efficacy. To monitor this process, γ-PGA/C6@P–S/pTRAIL nanoparticles were used for uptake tracking, with lysosomes labeled using LysoTracker Red and nuclei stained with DAPI for cellular localization. When green fluorescent nanoparticles colocalize with red fluorescent lysosomes, merged yellow fluorescence is observed; isolated green fluorescence, in contrast, indicates successful lysosomal escape. Three time points (2, 4, and 6 hours) were examined to assess nanoparticle–lysosome colocalization. At 2 hours, nanoparticles were predominantly internalized and colocalized with lysosomes, confirming entry via the lysosomal pathway. By 4 hours, cellular uptake increased while lysosomal colocalization was still evident. However, by 6 hours, most nanoparticles had separated from lysosomes, with only minimal residual colocalization remaining (Fig. 3A). These results demonstrate that the formulated nanoparticles can effectively escape from lysosomes following endocytosis, thereby enabling their intended antitumor function. The temporal progression from initial uptake (2 h) to lysosomal accumulation (4 h), and subsequent escape (6 h), confirms the system's capability to overcome this critical intracellular barrier to gene delivery.

Fig. 3 (A) CLSM images of HeLa cells incubated with nanocomposites for various time points. Red: lysosome was stained by LysoTracker; green: nanocomposites were labeled by C6; blue: nucleus was stained with DAPI. (B) Viabilities of HeLa cells receiving treatments of Sali and combination of Sali with γ-PGA/P–S/pTRAIL (mean ± SD; n = 3). (C) Western blot assay indicating the expression of DR5 and (D) cleaved caspase-3 in HeLa cells treated with Sali and γ-PGA/P–S/pTRAIL. GAPDH was used as a control. (E) Viabilities of HeLa cells receiving treatments of PBS, γ-PGA/P–S, γ-PGA/P–S/pTRAIL, γ-PGA/Sali@P–S and γ-PGA/Sali@P–S/pTRAIL (mean ± SD; n = 3). (F) Western blot assay indicating the expression of TRAIL in HeLa cells treated with PBS, γ-PGA/Sali@P–S, γ-PGA/P–S/pTRAIL and γ-PGA/Sali@P–S/pTRAIL. Actin was used as a control. (G) Flow cytometry analysis of ROS generation in HeLa cells treated with PBS, γ-PGA/Sali@P–S and γ-PGA/Sali@P–S/pTRAIL. (H) Statistical analyses of ROS generation in HeLa cells treated with PBS, γ-PGA/Sali@P–S and γ-PGA/Sali@P–S/pTRAIL by flow cytometry (mean ± SD; n = 3). (I) Western blot assay indicating the expression of β-catenin, c-Myc and (J) STAT3 in HeLa cells treated with Sali. GAPDH was used as a control. (K) Apoptosis levels determined by flow cytometry of HeLa cells treated with PBS, γ-PGA/P–S/pTRAIL, γ-PGA/Sali@P–S and γ-PGA/Sali@P–S/pTRAIL. (L) Statistical analyses of apoptosis levels of HeLa cells treated with PBS, γ-PGA/Sali@P–S, γ-PGA/P–S/pTRAIL and γ-PGA/Sali@P–S/pTRAIL by flow cytometry (mean ± SD; n = 3). The p values were calculated by ANOVA with Tukey's test. *p < 0.05, **p < 0.01, ***p < 0.001.

We further evaluated the cytotoxic effects of the nanoparticles on tumor cells, beginning with a comparative assessment of salinomycin monotherapy versus salinomycin/TRAIL combination therapy in HeLa cells. With the pTRAIL dosage fixed at 1 μg mL⁻¹, the results demonstrated that salinomycin significantly potentiated TRAIL-induced cytotoxicity (Fig. 3B). Western blot analysis revealed that salinomycin upregulated DR5 expression in HeLa cells (Fig. 3C and Fig. S6A), and the combination treatment triggered caspase-3 cleavage, thereby inducing tumor cell apoptosis (Fig. 3D and Fig. S6B). The cytotoxic effects of the nanodelivery system were subsequently examined. After 24 hours of treatment, the dual-drug-loaded γ-PGA/Sali@P–S/pTRAIL nanocomposites exhibited superior tumor cell-killing efficacy, whereas blank nanoparticles showed minimal toxicity, indicating good biocompatibility (Fig. 3E). Furthermore, the nanoparticle system successfully mediated TRAIL protein expression in HeLa cells through efficient gene transfection (Fig. 3F and Fig. S6C).

Previous studies have reported that salinomycin exerts antitumor effects by inducing intracellular reactive oxygen species (ROS) generation. In our study, intracellular ROS levels were measured using a ROS detection kit after 24-hour treatment with either γ-PGA/Sali@P–S or γ-PGA/Sali@P–S/pTRAIL nanoparticles. The results demonstrated that both drug-loaded nanoparticles effectively induced ROS production, with the combination therapy group exhibiting significantly higher ROS levels (Fig. 3G and H). Consistent with prior findings on salinomycin's ability to downregulate cancer stem cell (CSC)-related pathways, our western blot analysis revealed significant suppression of Wnt/β-catenin signaling, accompanied by reduced expression of the downstream oncogene c-Myc (Fig. 3I and Fig. S6D) and the transcription factor STAT3 (Fig. 3J and Fig. S6), both of which are key regulators of tumor progression. Subsequent apoptosis assays showed that γ-PGA/Sali@P–S/pTRAIL induced the highest proportion of apoptotic cell death, confirming its superior therapeutic efficacy (Fig. 3K and L).

Given the established correlation between tumor cell stemness and metastatic potential, we performed Transwell migration assays demonstrating that Sali significantly inhibits cell migration at concentrations exceeding 200 nM (Fig. 4A). Subsequent flow cytometric analysis revealed a marked reduction in the CD44⁺CD133⁺ dual-positive fraction upon Sali treatment, indicating decreased CSCs prevalence. Notably, combinatorial treatment with TRAIL further diminished this subpopulation, suggesting that Sali sensitizes CSCs to TRAIL-mediated cytotoxicity (Fig. 4B and C). As downstream effectors co-regulated by Wnt/β-catenin signaling, transcription factor c-Myc and STAT3 frequently exhibit overexpression in advanced malignancies, where their synergistic activation drives tumor proliferation and metastasis. Thus, concurrent suppression of both molecules may hold therapeutic significance. Therefore, simultaneous suppression of these oncogenic factors may offer therapeutic benefit. Importantly, our data demonstrate that Sali concurrently inhibits the Wnt/β-catenin, c-Myc, and STAT3 pathways, while augmenting TRAIL-mediated CSC eradication.

Fig. 4 Drug delivery and distribution in vivo. (A) Transwell migration assay of HeLa cells treated by Sali. (B) Flow cytometry analysis of CD44 and CD133 expression on HeLa cells treated with Sali and Sali + γ-PGA/Sali@P–S/pTRAIL (pTRAIL). (C) Statistical analyses of CD44⁺CD133⁺ dual-positive subpopulations after flow cytometry (mean ± SD; n = 3). (D) Distribution of IR-780 labelled nanocarriers with and without γ-PGA coating (the red arrows denote the tumor sites). (E) Analysis of fluorescence intensities of tumors at different time points (mean ± SD; n = 3). (F) Fluorescence images of ex vivo tumors of γ-PGA/IR780@P–S/pTRAIL and IR780@P–S/pTRAIL. (G) Analysis of fluorescence intensities of ex-vivo tumors (mean ± SD; n = 3). The p values were calculated by ANOVA with Tukey's test. *p < 0.05, **p < 0.01, ***p < 0.001.

Drug delivery and distribution in vivo

In vivo imaging was employed to evaluate the tumor-targeting capability and systemic stability of the nanocomposites in tumor-bearing mice, using the near-infrared fluorescent dye IR-780 to label two formulations (IR-780@P–S/pDNA and γ-PGA/IR-780@P–S/pDNA) for comparative analysis. The results demonstrated that the γ-PGA-uncoated nanoparticle group reached peak fluorescence intensity at the tumor site approximately 6 hours post-injection, followed by a gradual signal decline. In contrast, the γ-PGA-coated nanoparticles exhibited comparable initial fluorescence intensity but showed markedly improved retention—sustaining strong fluorescence beyond 6 hours and remaining detectable even at 48 hours. Meanwhile, the IR-780@P–S/pDNA group displayed progressive signal attenuation after 6 hours (Fig. 4D and E). At the 48-hour endpoint, following euthanasia and ex vivo tissue analysis, the γ-PGA/IR-780@P–S/pDNA group exhibited approximately twofold higher fluorescence intensity in tumor tissues compared to the uncoated IR-780@P–S/pDNA group (Fig. 4F and G). These findings indicate that the γ-PGA/IR-780@P–S/pDNA nanoparticles achieve enhanced tumor-specific accumulation. The γ-PGA coating significantly improves the systemic stability of the nanocarrier, effectively protecting the nanoparticles from rapid clearance. Consequently, it prolongs circulation half-life and enhances passive targeting efficiency through improved pharmacokinetic behavior.

Anti-tumor efficacy

The results revealed rapid tumor progression in the PBS control group. The Sali@P–S/pTRAIL group showed limited efficacy, likely due to its unshielded positive surface charge, which facilitated plasma protein adsorption and rapid clearance by the reticuloendothelial system. Both monotherapy groups (γ-PGA/Sali@P–S and γ-PGA/P–S/pTRAIL) exhibited modest antitumor effects, without statistical significance. In contrast, the co-delivery nanocomposites achieved pronounced tumor suppression, demonstrating superior therapeutic performance (Fig. 5A and B). Ex vivo tumor analysis revealed approximately 80% tumor growth inhibition in the γ-PGA/Sali@P–S/pTRAIL group compared to the PBS control (Fig. S7 and Fig. 5C). Furthermore, this group showed the longest survival duration (Fig. 5D), highlighting the synergistic therapeutic potential of the co-delivery system. The protective role of the γ-PGA coating against nonspecific clearance appears critical in enabling this in vivo efficacy.

Fig. 5 Anti-tumor of nanocomposites in HeLa xenograft model. (A) The sizes of HeLa xenograft tumors during treatments (mean ± SD; n = 5). The applied Sali was 0.5 mg kg⁻¹, and pTRAIL was 50 μg per mouse. (B) Images of HeLa tumors after treatments. (C) Tumor inhibition rates of treatments (mean ± SD; n = 5). (D) Survival curves of HeLa tumor-bearing mice during treatments (n = 8). (E) TUNEL assay of apoptosis in tumors after treatments (scale bar = 100 μm). (F) TRAIL, (G) IFN-γ, and (H) Ki67 immunohistochemical staining of HeLa tumor tissues after treatments (scale bar = 100 μm). The p values were calculated by ANOVA with Tukey's test. *p < 0.05, **p < 0.01, ***p < 0.001.

Toxicity was preliminarily assessed through body weight monitoring and organ coefficient analysis at the study endpoint. No significant differences in body weight were observed among treatment groups (Fig. S8). However, organ coefficient analysis revealed splenomegaly in the Sali@P–S/pTRAIL group (Fig. S9), likely due to the unshielded positive charge of these nanoparticles, which may cause spleen-specific toxicity via electrostatic interactions with reticuloendothelial cells. This finding aligns with the known tendency of cationic nanoparticles to accumulate in and potentially damage splenic tissue. In contrast, γ-PGA-coated formulations exhibited no such adverse effects, confirming the importance of surface charge shielding in reducing organ-specific toxicity. Histopathological examination of major organs (heart, liver, spleen, lungs, kidneys) revealed no significant abnormalities across all treatment groups, indicating acceptable systemic tolerability (Fig. S10).

TUNEL assays conducted on tumor sections showed that the γ-PGA/Sali@P–S/pTRAIL group induced significantly higher levels of apoptosis compared to single-agent treatments, which induced only moderate apoptosis (Fig. 5E). Immunohistochemical analysis revealed differential TRAIL expression among the groups: while Sali@P–S/pTRAIL showed partial expression, both γ-PGA/P–S/pTRAIL and γ-PGA/Sali@P–S/pTRAIL groups demonstrated substantially higher TRAIL levels in tumor tissues (Fig. 5F). Evaluation of angiogenesis via CD31 staining showed the lowest expression levels in the γ-PGA/Sali@P–S/pTRAIL group (Fig. 5G), suggesting that extensive tumor cell apoptosis may suppress tumor vascularization. Similarly, Ki67 immunostaining indicated markedly reduced tumor cell proliferation in this group relative to all others (Fig. 5H). Collectively, these histopathological analyses confirm that the co-delivery nanocomposites enhance apoptosis, inhibit angiogenesis, and suppress proliferation through coordinated and synergistic mechanisms.

Conclusions

This study employed dendrimeric PEI as the structural backbone, with hydrophobic stearic acid chains conjugated via amide bonds to synthesize the amphiphilic polymer P–S. In the presence of DOPE as a helper lipid, P–S self-assembled with salinomycin into nanomicelles (Sali@P–S), comprising hydrophilic PEI coronas and hydrophobic cores encapsulating the drug. Subsequent complexation with TRAIL plasmid (pTRAIL) via electrostatic interaction with surface PEI yielded the dual-functional nanoplatform Sali@P–S/pTRAIL. A final surface coating with γ-PGA enhanced colloidal stability, reduced particle size, and improved size distribution uniformity. Comprehensive in vitro and in vivo studies demonstrated that this co-delivery system effectively suppressed cervical cancer in murine models through synergistic mechanisms—namely, the induction of tumor cell apoptosis and the downregulation of cancer stem cell-associated signaling pathways.

Despite promising results, translating such nanotherapeutics from bench to bedside remains a formidable challenge. TRAIL-based therapies, in particular, have encountered persistent limitations in clinical translation. Overcoming these barriers requires the convergence of multiple scientific disciplines. When such interdisciplinary synergies mature, they may ultimately unlock the full therapeutic potential of TRAIL-associated combination strategies.

Author contributions

Huihai Zhong: conceptualization, writing – original draft; Yuefei Fang: writing – review & editing; Pengfei Zhao: writing – review & editing; Aihua Wu: writing – review & editing, validation; Guohui Nie: writing – review & editing, validation; Yongzhuo Huang: conceptualization, funding acquisition, writing – review & editing; Bin Zhang: conceptualization, funding acquisition, writing – review & editing.

Conflicts of interest

There are no conflicts to declare.

Data availability

All relevant data of this study are available within the paper and its SI files. Experimental details and additional characterization results including NMR, FTIR and related raw data for the biological tests are provided in the Supplemental Information. See DOI: https://doi.org/10.1039/d5tb01203a

Acknowledgements

We are grateful for the support from the National Key Research and Development Program of China (2024YFA1210200, China), the National Natural Science Foundation of China (82341232, 52203335, and 82192865), the Natural Science Foundation of Shanghai Municipality (24ZR1477500), the High-level Innovative Research Institute (2021B0909050003) from the Department of Science and Technology of Guangdong Province, the Shenzhen Science and Technology Innovation Committee (JCYJ20240813115805008 and ZDSYS201707281114196) and the Zhongshan Municipal Bureau of Science and Technology (LJ2021001 and CXTD2022011). This work is also supported by the Sanming Project of Medicine in Shenzhen (NO. SZSM202211022) and Shenzhen Portion of Shenzhen-Hong Kong Science and Technology Innovation Cooperation Zone (No. HTHZQSWS-KCCYB-2023060). We thank the staff members of the Large-scale Protein Preparation System (https://cstr.cn/31129.02.NFPS.LSPS) at the National Facility for Protein Science in Shanghai (https://cstr.cn/31129.02.NFPS), for providing technical support and assistance in data collection and analysis in the transmission electron microscope and fluorescence microscope.

Notes and references

1. K. S. Tewari, N. Engl. J. Med., 2025, 392, 56–71.
2. P. Tsikouras, S. Zervoudis, B. Manav, E. Tomara, G. Iatrakis, C. Romanidis, A. Bothou and G. Galazios, J. BUON, 2016, 21, 320–325.
3. W. J. Koh, N. R. Abu-Rustum, S. Bean, K. Bradley, S. M. Campos, K. R. Cho, H. S. Chon, C. Chu, R. Clark, D. Cohn, M. A. Crispens, S. Damast, O. Dorigo, P. J. Eifel, C. M. Fisher, P. Frederick, D. K. Gaffney, E. Han, W. K. Huh, J. R. Lurain, A. Mariani, D. Mutch, C. Nagel, L. Nekhlyudov, A. N. Fader, S. W. Remmenga, R. K. Reynolds, T. Tillmanns, S. Ueda, E. Wyse, C. M. Yashar, N. R. McMillian and J. L. Scavone, J. Natl. Compr. Cancer Network, 2019, 17, 64–84.
4. R. Sun, C. Chen, X. Deng, F. Wang, S. Song, Q. Cai, J. Wang, T. Zhang, M. Shi, Q. Ke and Z. Luo, J. Cancer, 2021, 12, 4638–4647.
5. S. Das, A. Babu, T. Medha, G. Ramanathan, A. G. Mukherjee, U. R. Wanjari, R. Murali, S. Kannampuzha, A. V. Gopalakrishnan, K. Renu, D. Sinha and C. George Priya Doss, Med. Oncol., 2023, 40, 149.
6. D. Gu, X. Liang, Z. Shao, H. He, L. Zhu, S. Qiu, Z. Liu, S. Wang, H. Chen, S. Ci, H. Wu and Y. Wang, J. Mater. Chem. B, 2025, 13, 6128–6143.
7. U. Anand, A. Dey, A. K. S. Chandel, R. Sanyal, A. Mishra, D. K. Pandey, V. De Falco, A. Upadhyay, R. Kandimalla, A. Chaudhary, J. K. Dhanjal, S. Dewanjee, J. Vallamkondu and J. M. Perez de la Lastra, Genes Dis., 2023, 10, 1367–1401.
8. H. C. Reinhardt and B. Schumacher, Trends Genet., 2012, 28, 128–136.
9. A. Ashkenazi, R. C. Pai, S. Fong, S. Leung, D. A. Lawrence, S. A. Marsters, C. Blackie, L. Chang, A. E. McMurtrey, A. Hebert, L. DeForge, I. L. Koumenis, D. Lewis, L. Harris, J. Bussiere, H. Koeppen, Z. Shahrokh and R. H. Schwall, J. Clin. Invest., 1999, 104, 155–162.
10. S. Wang and W. S. El-Deiry, Oncogene, 2003, 22, 8628–8633.
11. D. Singh, M. Tewari, S. Singh and G. Narayan, Future Oncol., 2021, 17, 581–596.
12. H. H. Zhong, H. Y. Wang, J. Li and Y. Z. Huang, Acta Pharmacol. Sin., 2019, 40, 1373–1385.
13. B. Thapa, R. Kc and H. Uludag, J. Controlled Release, 2020, 326, 335–349.
14. R. S. Herbst, S. G. Eckhardt, R. Kurzrock, S. Ebbinghaus, P. J. O'Dwyer, M. S. Gordon, W. Novotny, M. A. Goldwasser, T. M. Tohnya, B. L. Lum, A. Ashkenazi, A. M. Jubb and D. S. Mendelson, J. Clin. Oncol., 2010, 28, 2839–2846.
15. Y. Wang, M. Wang, H. Chen, H. Liu, Q. Zhang and Y. Cheng, J. Mater. Chem. B, 2016, 4, 1354–1360.
16. S. Goklany, P. Lu, S. Godeshala, A. Hall, E. Garrett-Mayer, C. Voelkel-Johnson and K. Rege, J. Mater. Chem. B, 2019, 7, 7014–7025.
17. N. Qiu, X. Liu, Y. Zhong, Z. Zhou, Y. Piao, L. Miao, Q. Zhang, J. Tang, L. Huang and Y. Shen, Adv. Mater., 2016, 28, 10613–10622.
18. H. Yin, R. L. Kanasty, A. A. Eltoukhy, A. J. Vegas, J. R. Dorkin and D. G. Anderson, Nat. Rev. Genet., 2014, 15, 541–555.
19. M. Taghdiri and C. Mussolino, Int. J. Mol. Sci., 2024, 25, 7333.
20. O. Boussif, F. Lezoualc'h, M. A. Zanta, M. D. Mergny, D. Scherman, B. Demeneix and J. P. Behr, Proc. Natl. Acad. Sci. U. S. A., 1995, 92, 7297–7301.
21. C. Qiu, F. Xia, J. Zhang, Q. Shi, Y. Meng, C. Wang, H. Pang, L. Gu, C. Xu, Q. Guo and J. Wang, Research, 2023, 6, 0148.
22. Y. Yamazaki, M. Nango, M. Matsuura, Y. Hasegawa, M. Hasegawa and N. Oku, Gene Ther., 2000, 7, 1148–1155.
23. S. A. J. Buck, S. L. W. Koolen, R. H. J. Mathijssen, R. de Wit and R. J. van Soest, Drug Resistance Updates, 2021, 56, 100761.
24. M. J. Yoon, Y. J. Kang, I. Y. Kim, E. H. Kim, J. A. Lee, J. H. Lim, T. K. Kwon and K. S. Choi, Carcinogenesis, 2013, 34, 1918–1928.
25. E. Batlle and H. Clevers, Nat. Med., 2017, 23, 1124–1134.
26. V. Plaks, N. Kong and Z. Werb, Cell Stem Cell, 2015, 16, 225–238.
27. C. E. Meacham and S. J. Morrison, Nature, 2013, 501, 328–337.
28. X. Chu, W. Tian, J. Ning, G. Xiao, Y. Zhou, Z. Wang, Z. Zhai, G. Tanzhu, J. Yang and R. Zhou, Signal Transduction Targeted Ther., 2024, 9, 170.
29. T. Shibue and R. A. Weinberg, Nat. Rev. Clin. Oncol., 2017, 14, 611–629.
30. X. Suo, J. Zhang, Y. Zhang, X. J. Liang, J. Zhang and D. Liu, J. Mater. Chem. B, 2020, 8, 3985–4001.
31. J. S. Dosch, E. K. Ziemke, A. Shettigar, A. Rehemtulla and J. S. Sebolt-Leopold, Cancer Res., 2015, 75, 4582–4592.
32. M. Nishi, Y. Sakai, H. Akutsu, Y. Nagashima, G. Quinn, S. Masui, H. Kimura, K. Perrem, A. Umezawa, N. Yamamoto, S. W. Lee and A. Ryo, Oncogene, 2014, 33, 643–652.
33. N. Takebe, P. J. Harris, R. Q. Warren and S. P. Ivy, Nat. Rev. Clin. Oncol., 2011, 8, 97–106.
34. W. Jin, Cells, 2020, 9, 217.
35. P. C. Shih and K. C. Mei, Drug Discovery Today, 2021, 26, 1450–1458.
36. M. Loi, P. Becherini, L. Emionite, A. Giacomini, I. Cossu, E. Destefanis, C. Brignole, D. Di Paolo, F. Piaggio, P. Perri, M. Cilli, F. Pastorino and M. Ponzoni, J. Controlled Release, 2014, 192, 157–166.
37. H. Y. Lee, J. Cha, S. K. Kim, J. H. Park, K. H. Song, P. Kim and M. Y. Kim, Mol. Cancer Res., 2019, 17, 544–554.
38. H. Fuke, K. Shiraki, K. Sugimoto, J. Tanaka, T. Beppu, K. Yoneda, N. Yamamoto, K. Ito, M. Masuya and Y. Takei, Biochem. Biophys. Res. Commun., 2007, 363, 738–744.
39. A. Arfaoui, C. Rioualen, V. Azzoni, G. Pinna, P. Finetti, J. Wicinski, E. Josselin, M. Macario, R. Castellano, C. Leonard-Stumpf, A. Bal, A. Gros, S. Lossy, M. Kharrat, Y. Collette, F. Bertucci, D. Birnbaum, H. Douik, G. Bidaut, E. Charafe-Jauffret and C. Ginestier, EMBO Mol. Med., 2019, 11, e9930.
40. C. Naujokat and R. Steinhart, J. Biomed. Biotechnol., 2012, 2012, 950658.
41. S. Shen, S. Lin, Y. Chen, Y. Zhang, Y. He, X. Xu, Y. Feng, Y. Lu and R. Mo, Nano Lett., 2022, 22, 2419–2428.
42. Z. Zhang, N. Qiu, S. Wu, X. Liu, Z. Zhou, J. Tang, Y. Liu, R. Zhou and Y. Shen, Adv. Mater., 2021, 33, e2102219.
43. T. T. Smith, S. B. Stephan, H. F. Moffett, L. E. McKnight, W. Ji, D. Reiman, E. Bonagofski, M. E. Wohlfahrt, S. P. S. Pillai and M. T. Stephan, Nat. Nanotechnol., 2017, 12, 813–820.
44. T. Kurosaki, T. Kitahara, S. Fumoto, K. Nishida, J. Nakamura, T. Niidome, Y. Kodama, H. Nakagawa, H. To and H. Sasaki, Biomaterials, 2009, 30, 2846–2853.
45. F. Xu, H. Zhong, Y. Chang, D. Li, H. Jin, M. Zhang, H. Wang, C. Jiang, Y. Shen and Y. Huang, Biomaterials, 2018, 158, 56–73.

---