Artesunate-loaded bovine serum albumin nanoplatform with metal–polyphenol network coating for ferroptosis-driven breast cancer therapy

Abstract

Ferroptosis, an iron-dependent programmed cell death process driven by reactive oxygen species (ROS) accumulation, represents a promising therapeutic strategy for breast cancer. However, the efficacy of ferroptosis therapy in breast cancer is often compromised by insufficient intracellular levels of hydrogen peroxide (H₂O₂) and iron ions. To address these problems, we developed pH-responsive nanoparticles comprising a metal–phenolic network (MPN) shell and a bovine serum albumin (BSA) core for the delivery of artesunate (ART) (designated as BAM NPs). The designed BAM NPs aim to amplify oxidative stress and enhance ferroptosis-based therapy for breast cancer. Upon endocytosis by tumor cells, BAM NPs underwent degradation under acidic conditions to release ART and Fe³⁺. Subsequently, the reduction of Fe³⁺ to Fe²⁺ by glutathione (GSH) initiated the Fenton reaction, which led to aberrant accumulation of ROS. Meanwhile, the endoperoxide bridge of ART could be cleaved by Fe²⁺ to further generate carbon-centered radicals (·C). Furthermore, BAM NPs effectively deplete GSH via Fe³⁺/Fe²⁺ conversion to inactivate glutathione peroxidase 4 (GPX4), thereby disrupting redox homeostasis and increasing intracellular LPO levels. Both in vitro and in vivo experiments showed that BAM NPs significantly inhibited tumor cell proliferation by inducing robust ferroptosis in tumor cells, leading to a tumor inhibition rate of 83.11%. In summary, the constructed BAM NPs served as a promising tailored nanoplatform for augmenting ferroptosis therapy in breast cancer.

Graphical Abstract

1. Introduction

Breast cancer is a prevalent malignancy in clinical oncology among women, imposing a substantial global health burden. A total of 2.3 million new breast cancer cases were documented across the globe in 2022 (Zhang et al., 2025c). Numerous risk factors (e.g. age, hormonal disorders, alcohol intake, and unhealthy lifestyles) are closely associated with breast cancer incidence (Johansson et al., 2021). Although continuous progress in treatment approaches such as surgical interventions, radiation therapy, and chemotherapeutic regimens has been achieved, the treatment outcomes may remain suboptimal (Dixon and Olzmann, 2024). Hence, developing innovative therapeutic approaches is critical to overcoming these persistent challenges (Lei et al., 2022).

Ferroptosis is a distinct form of regulated cell death that relies on intracellular iron accumulation. Distinct from apoptosis, autophagy, and necrosis, this cell death modality is characterized by mitochondrial perturbations, excessive reactive oxygen species (ROS) generation, glutathione (GSH) depletion, and elevated lipid peroxidation levels (Li et al., 2020). Notably, ferroptosis offers novel insights into the advancement of cancer therapeutic strategies. It can selectively trigger tumor cell death by regulating organellar functions and antioxidant defense systems (Han et al., 2024). Despite substantial advances, insufficient intracellular iron levels remain a critical bottleneck in ferroptosis-mediated antitumor therapy (Chen et al., 2021). Previous research has primarily centered on the development of diverse iron-based nanoparticles (such as metal-phenolic networks (MPNs), ultrasmall superparamagnetic iron oxide (USPIO) and ferroferric oxide (Fe₃O₄)), owing to the need for high iron loading to enable efficient ROS generation (Li et al., 2025). Among these materials, MPNs are widely employed as iron sources due to their superior biocompatibility, low toxicity, and facile assembly process (Centurion et al., 2026). Notably, MPNs can bind to the surface of diverse materials owing to the adhesive capability of EGCG. They catalyze the conversion of endogenous hydrogen peroxide (H₂O₂) into highly toxic hydroxyl radicals (·OH) while concomitantly depleting intracellular GSH (Gong et al., 2025). This process promotes the occurrence of ferroptosis and thereby effectively suppresses tumor progression (Liang et al., 2022).

Besides, the inherent deficiency of endogenous H₂O₂ in tumors significantly impedes the efficient induction of ferroptosis. Recent studies have proposed that many H₂O₂-irrelevant nanosystems can also generate carbon-centered radicals (·C) (Zhang et al., 2025a). Artesunate (ART), a derivative of artemisinin, has been recently repurposed as a promising antitumor agent (Wang et al., 2025). Its mechanism of action hinges on the cleavage of the peroxide bridge in the ART molecular structure under the catalytic action of Fe²⁺ to generate cytotoxic ·C, which induces ferroptosis in tumor cells (Chen et al., 2023b). However, the clinical application of ART is impeded by several limitations such as poor water solubility (∼0.08 mg/mL in water at 25 °C) (Bezuidenhout et al., 2023), short half-life and low bioavailability (Yang et al., 2023). Therefore, it is urgent to construct a nanosystem to co-transport sufficient iron ions and ART for efficient induction of tumor ferroptosis (Xiang et al., 2022).

Herein, we successfully fabricated ART-loaded bovine serum albumin (BSA) nanoparticles coated with the MPNs, which were designated as BAM NPs (Scheme 1). By leveraging the EPR effect, BAM NPs accumulated efficiently at tumor sites following intravenous administration. After internalization by cancer cells, BAM NPs rapidly released Fe³⁺ and ART. Subsequently, in the intracellular environment, GSH reduced Fe³⁺ to Fe²⁺, which subsequently reacted with endogenous H₂O₂ through the Fenton reaction to generate toxic ·OH. Simultaneously, Fe²⁺ mediated the cleavage of the endoperoxide bridge in ART, yielding toxic ·C. This process not only augmented radical generation but also led to the depletion of intracellular GSH, which resulted in the inactivation of GPX4. Disruption of cellular redox homeostasis further induced the substantial accumulation of lipid peroxides, ultimately triggering ferroptosis. Finally, these findings demonstrated that the constructed BAM NPs served as a promising nanoplatform for ferroptosis-based antitumor therapy.

Scheme 1.

Schematic illustration of the formation of BAM NPs for ferroptosis-based cancer therapy.

2. Materials and methods

2.1. Materials

Artesunate (ART), 5,5′-dithiobis-(2-nitrobenzoic acid), iron chloride hexahydrate (FeCl₃·6H₂O), propidium iodide (PI), glutathione (GSH), phosphate buffered saline (PBS) and 3,3′,5,5′-tetramethylbenzidine (TMB) were purchased from Macklin (Shanghai, China). Methylene blue (MB), fluorescein isothiocyanate (FITC), bovine serum albumin (BSA), fluorescein diacetate (FDA) and hydrogen peroxide (H₂O₂) were provided by Aladdin (Shanghai, China). 2′,7’-Dichlorodihydrofluorescein diacetate (DCFH-DA) and Epigallocatechin gallate (EGCG) were obtained from Dalian Meilun Biotechnology (Dalian, China). Boron dipyrromethenes-C11 (C11-BODIPY) and trypsin were supplied by Thermo Fisher Scientific (Shanghai, China). 4′,6-Diamidino-2-phenylindole (DAPI) and cell counting kit-8 (CCK-8) were bought from Dalian Meilun Biotechnology (Dalian, China).

2.2. Preparation of BA NPs and BAM NPs

The synthesis of BAM NPs was conducted following previous literature. Specifically, BSA (1 mL, 10 mg/mL dissolved in ultrapure water) and ART (50 μL, 20 mg/mL dissolved in DMSO) were prepared. Then, the ART-containing DMSO solution was added dropwise into the BSA aqueous solution and stirred overnight. The unencapsulated ART and organic solvent were removed via dialysis against ultrapure water over a period of 48 h, yielding the BA NPs. Next, 0.5 mL of EGCG aqueous solution (10 mg/mL) and 1 mL of FeCl₃·6H₂O aqueous solution (8 mg/mL) were added to the BA NPs solution. The mixture was stirred for 5 min at room temperature and subsequently underwent centrifugation for 30 min (12,000 rpm) to obtain the BAM NPs.

2.3. Characterization of BAM NPs

The freshly prepared dilute solution of BAM NPs was deposited on the copper network. After drying, TEM was used to investigate the morphology of the BAM NPs. DLS was employed to measure zeta potential and hydrodynamic diameter of BAM NPs. The crystalline structure of ART and BAM NPs was determined by XRD.

2.4. Loading capacity (LC) and encapsulation efficiency (EE)

The different concentrations of ART solutions were incubated with 0.2% NaOH solution at 60 °C for 30 min respectively to convert ART into hydrolysate with ultraviolet absorption. A standard curve for ART was plotted based on the absorbance of the ART hydrolysate measured by a UV–Vis spectrophotometer. After centrifugation of the freshly prepared BAM NPs, the ART-containing supernatant was collected. Then, the supernatant was treated with 0.2% NaOH solution at 60 °C for 30 min. The UV absorption of ART hydrolysate was measured and the content of ART was calculated using the standard curve above. Lastly, LC and EE of BAM NPs were calculated by the following equations:

$$\mathrm{LC}\left(\%\right)=\frac{\text{total}\mathrm{ART}-\text{free}\mathrm{ART}}{\text{total amount of nanoparticles}}\times 100$$

$$\mathrm{EE}\left(\%\right)=\frac{\text{total}\mathrm{ART}-\text{free}\mathrm{ART}}{\text{total}\mathrm{ART}}\times 100$$

2.5. In vitro drug release study

The in vitro ART release profile from BAM NPs was studied as follows. BAM NPs (3 mL) were sealed in dialysis bags (MWCO 3500 Da) and immersed in 20 mL of PBS (pH 7.4 or pH 5.0) with stirring at 100 rpm. At the given time points (2, 4, 6, 10, 12, 24 and 36 h), the release medium was taken and treated with 0.2% NaOH solution at 60 °C for 30 min. Finally, UV–Vis spectrophotometry was employed to analyze the ART content in the release medium.

2.6. GSH depletion detection

To evaluate the GSH depletion capability of BAM NPs, DTNB was employed as an indicator. BAM NPs solutions (0, 5, 10, 25, 50, 75, 100 and 200 μg/mL) were incubated with 1 mM GSH at 37 °C for 30 min. Following the addition of 10 mM DTNB, the supernatant was obtained by centrifugation, and its absorbance at 412 nm was determined using UV–Vis spectrophotometry.

2.7. Measurement of ·OH generation

The methylene blue (MB) was employed as a probe to monitor the ·OH generation efficiency. Firstly, BAM NPs (50 μg/mL) were dispersed in PBS solution and cocultured with GSH for 30 min, followed by sequential addition of H₂O₂ oxidant and MB solution. Then, the solution was incubated at 37 °C for 30 min. Finally, UV–Vis spectroscopy was used to measure the absorbance of the various mixed solutions at 665 nm.

Next, the catalytic efficiency of ·OH generation was further evaluated using 3,3′,5,5′-tetramethylbenzidine (TMB). Briefly, BAM NPs (50 μg/mL) in PBS were mixed with GSH, H₂O₂ and TMB solution, and UV–Vis spectroscopy was employed to monitor absorbance at 652 nm.

2.8. Cellular uptake study

In preparation for uptake experiments, 4T1 cells were uniformly plated on confocal dishes and incubated until ∼80% confluence. Then, the cells were cultured with FITC/BAM NPs for 3, 6 and 9 h. Following fixation with 4% paraformaldehyde at 4 °C for 30 min, the 4T1 cells were subsequently counterstained with DAPI for nuclear visualization. After staining, cellular uptake was evaluated by confocal laser scanning microscopy (CLSM) (Ji et al., 2026). In addition, flow cytometry (FCM) was utilized for its quantitative assessment.

2.9. Intracellular GSH assay

To assess intracellular GSH depletion, 4T1 cells at an appropriate density were plated in 6-well plates and incubated overnight. Cells were treated with BSA, ART, BA NPs, or BAM NPs for 24 h. Then, the cells were collected to measure the intracellular GSH level by the GSH assay kit.

2.10. ROS assay

The intracellular ROS-generating capacity of the nanoparticles was determined using the DCFH-DA probe. After seeding and overnight culture in confocal dishes, 4T1 cells were treated with BSA, ART, BA NPs, or BAM NPs for 24 h. Then, the 4T1 cells were cultured with 1 mL of DCFH-DA at 37 °C for 30 min. Finally, fluorescence images were observed via CLSM (Liu et al., 2025). Furthermore, intracellular ROS levels were quantified by FCM.

2.11. Lipid peroxidation assay

LPO was assessed using the C11-BODIPY probe. 4T1 cells were first incubated in confocal culture systems for 24 h, and then subjected to distinct samples for a further 24 h. Then, cells were labeled with C11-BODIPY for 30 min at 37 °C. LPO fluorescence was observed using CLSM (An et al., 2026). Additionally, LPO levels of samples were quantified via FCM.

2.12. Mitochondrial membrane potential (MMP) measurement

JC-1 staining was employed to monitor the disruption of mitochondrial membrane potential (MMP) induced by nanoparticle treatment. In brief, 4T1 cells were plated on confocal dishes and exposed to various samples for 24 h. Subsequently, PBS was used to wash the cells, which were then stained with JC-1 probe for 20 min in darkness. Finally, The fluorescence was analyzed by CLSM.

2.13. Malondialdehyde (MDA) assay

Levels of MDA were quantified using the MDA detection kit (Gao et al., 2025). Different samples were used to treat the cells, and the cells were placed on ice to lyse. After centrifugation, the supernatant was gathered and subsequently mixed with thiobarbituric acid (TBA) at 100 °C for 15 min. Following cooling of the mixture to room temperature, it was centrifuged at 12,000 rpm. Finally, the content of MDA was analyzed at the wavelength of 532 nm.

2.14. Cytotoxicity studies

The viability of 4T1 cells after being treated with BSA, ART, BA NPs, and BAM NPs was estimated via cell counting kit-8 (CCK-8). Firstly, 4T1 cells were seeded into 96-well plates. Subsequently, various concentrations of prepared samples were added to the cultured cells. After 24 h, each well was treated with CCK8 solution for 1 h. Finally, the absorbance at 450 nm was recorded to calculate the cell viability (Qiu et al., 2023).

L929 cells were seeded in 96-well plates and cultured overnight. The cells were then treated with various concentrations of BAM NPs for 24 h. After incubation, CCK-8 solution was added and incubated for 1 h. Finally, the absorbance was measured using a microplate reader.

Initially, 4T1 cells were placed in confocal dishes, and incubated at 37 °C for 1 day. Next, cells were exposed to BSA, ART, BA NPs, and BAM NPs, respectively. After being rinsed twice with PBS, the live cells were stained with fluorescein diacetate (FDA) in the dark for 30 min. Then, the medium was discarded and propidium iodide (PI) was added to stain the dead cells for another 20 min. Lastly, the fluorescence images were obtained via CLSM.

The cellular death behavior was analyzed by Annexin V-APC/7-AAD apoptosis kit. Firstly, 4T1 cells were placed in 6-well plates for 1 day. After that, cells were incubated for a further 1 day with BSA, ART, BA NPs, and BAM NPs. Then, cells were exposed to 1 × binding buffer containing Annexin V-APC and 7-AAD. After incubation for 5 min, cells were detected by flow cytometry.

2.15. Biodistribution analysis in vivo

4T1 cells were injected subcutaneously into the right lateral dorsal region of female BALB/c mice. The mice were used for in vivo biodistribution analysis when the tumor volume reached 100mm³. The mice received intravenous injections of BAM@CY5.5 NPs or CY5.5, and imaged various times via the instrument. At 24 h post-injection, the mice were euthanized to obtain vital organs and tumors for fluorescence-based imaging.

2.16. In vivo antitumor study

Once the tumor volumes attained approximately 100 mm³, the mice were divided into five random groups (n = 5). The groups received intravenous injections of PBS, BSA, free ART, BA NPs and BAM NPs every 3 days (equivalent ART dose: 10 mg/kg) (Li et al., 2026). Tumor sizes and mouse body weights were documented every 2 days. Besides, the calculation of tumor volumes was performed with the following formula: tumor volume = (length × width²)/2. After 14 days of treatment, all mice were euthanized. At last, vital organs and tumor samples were gathered for H&E staining procedures (Li et al., 2026).

2.17. Hemolysis assay

Red blood cells (RBCs) were prepared and washed with PBS. The RBC suspension was mixed with water (positive control) or various concentrations of BAM NPs and incubated for 1 h. The samples were then centrifuged, and the supernatants were measured using a microplate reader.

2.18. Statistical analysis

Results are presented as mean ± SD of at least three separate experiments. Statistical significance was assessed via one-way ANOVA. All statistical analyses were performed using GraphPad Prism 10.0 software, with significance levels indicated as follows: *p < 0.05, **p < 0.01, ***p < 0.001.

3. Results and discussion

3.1. Fabrication and characterization of BAM NPs

The fabrication process of BAM NPs is illustrated in Scheme 1. Specifically, ART was first encapsulated into the hydrophobic pockets of BSA via hydrophobic interactions. Subsequently, BAM NPs were fabricated by wrapping the BA NPs with the MPNs coating, which were derived from the coordination reaction between Fe³⁺ and the phenolic hydroxyl groups of EGCG. As illustrated in Fig. 1A, the TEM image revealed that BAM NPs were spherical with the size of approximately 100 nm. Notably, compared with BA NPs, the hydrodynamic diameters of BAM NPs were increased from 83.55 ± 4.46 nm to 105.49 ± 10.25 nm (Fig. 1B and S1). At the same time, BAM NPs were confirmed to have a larger absolute value of negative zeta potential (−19.83 ± 2.15 mV) than that of BA NPs (−8.58 ± 0.82 mV) (Fig. 1C). Coupled with the above particle size and potential changes, these results indicated a successful coating of EGCG-Fe³⁺ MPN on the surface of BAM NPs. In addition, the stability tests of BAM NPs in water revealed no significant change in particle size, which indicated excellent stability. (Fig. S2). Subsequently, the composition of BAM NPs was further investigated. Energy dispersive spectroscopy (EDS) confirmed that BAM NPs contained C, O, N and Fe elements (Fig. 1D), further demonstrating the successful synthesis of the NPs. Furthermore, BAM NPs possessed relatively high EE (85.00 ± 3.77%) and LC (7.83 ± 0.32%), which were further ascribed to the coating of MPNs and the wrap by the BSA. Besides, as presented in Fig. 1E, XRD pattern of free ART exhibited numerous sharp peaks, which demonstrated that free ART maintained a highly crystalline state. In contrast, no distinct characteristic peaks were observed in the pattern of BAM NPs, indicating the amorphous nature of BAM NPs. These findings demonstrated the successful encapsulation of ART within BAM NPs.

Fig. 1.

(A) The TEM image of BAM NPs. (B) Size distribution and (C) zeta potential of the BA NPs and BAM NPs. (D) Element mapping of BAM NPs. (E) XRD patterns of ART and BAM NPs. (F) XPS analysis of Fe 2p. (G) Release profile of BAM NPs under different pH conditions. (H) GSH depleting ability of BAM NPs at different concentrations. (I) UV–Vis absorption spectra of MB after different treatment. (J) UV–Vis absorption spectra of TMB incubated with different samples.

Next, the release behavior of BAM NPs under pH 7.4 and pH 5.0 was explored. As depicted in Fig. 1G, BAM NPs showed a fast release rate at pH 5.0, and the cumulative release rate of ART reached 80.30 ± 3.03% after 36 h. In contrast, the release rate of BAM NPs at pH 7.4 was significantly slower than that at pH 5.0, and the cumulative release of ART merely reached 32.73 ± 4.01% after 36 h. These results revealed that BAM NPs exhibited distinct pH-sensitive release behavior. The occurrence of the pH-sensitive phenomenon might be attributed to the protonation of the hydroxyl groups in the MPNs at low pH, facilitating the collapse of the coated film. This release property could effectively reduce side effects and prompt rapid drug release of BAM NPs in tumor cells.

To investigate the reaction between GSH and BAM NPs, the valence state of Fe in BAM NPs was determined via XPS analysis. As presented in Fig. 1F, distinct characteristic peaks were observed at 712.95 eV and 726.19 eV, corresponding to the 2p_3/2 and 2p_1/2 orbitals of Fe³⁺. Moreover, the peaks at 708.84 eV and 722.42 eV were assigned to Fe²⁺ in BAM NPs. These results confirmed that Fe³⁺ in BAM NPs was reduced to Fe²⁺ by GSH, which would facilitate the occurrence of the Fenton reaction.

To evaluate the GSH-depleting capacity of BAM NPs, DTNB was employed to quantify GSH consumption. DTNB could react with GSH to generate yellow TNB, which could be detected by a UV spectrophotometer at 412 nm. As presented in Fig. 1H, with increasing BAM NPs concentration, the absorbance of TNB at 412 nm gradually declined. This result confirmed that BAM NPs exhibited a depletion effect on GSH, which was attributed to the redox reaction of Fe³⁺ with GSH. Furthermore, such efficient GSH consumption was conducive to disrupting intracellular redox homeostasis, which impaired the antioxidant defense systems of tumor cells.

Based on the above experimental results, the Fe³⁺ in BAM NPs reacted with GSH and was converted to Fe²⁺, leading to the generation of ·OH via the Fenton reaction. To confirm this, MB was selected as the probe to monitor the formation of ·OH (Dai et al., 2023). As shown in Fig. 1I, no MB degradation occurred in either the MB + H₂O₂ group or the MB + H₂O₂ + GSH group, where the absorbance at 665 nm remained almost unchanged. In contrast, both the MB + H₂O₂ + GSH + BAM NPs group and the MB + H₂O₂ + GSH + Fe²⁺ group exhibited significantly decreased absorbance at 665 nm, which was attributed to the degradation of MB. This result indicated that BAM NPs with intrinsic Fenton reactivity effectively promoted the generation of ·OH, which could result in the degradation of MB. Meanwhile, TMB was also employed to assess the ·OH production ability of BAM NPs. As shown in Fig. 1J, there was no characteristic absorption observed in the TMB + H₂O₂ and TMB + H₂O₂ + GSH groups. However, the TMB + H₂O₂ + GSH + BAM NPs group and the TMB + H₂O₂ + GSH + Fe²⁺ group showed distinct characteristic absorption at 652 nm. Overall, these results demonstrated that the newly generated Fe²⁺ promoted the production of ·OH.

3.2. The cellular uptake of BAM NPs

To explore the uptake efficiency of BAM NPs, the nanoparticles were labeled with FITC and analyzed using CLSM and FCM. As depicted in Fig. 2A, it was observed that the green fluorescence at 9 h was significantly higher than at 3 h, indicating that the uptake efficiency of FITC/BAM NPs exhibited time dependence. The internalization of FITC/BAM NPs could be attributed to the produced ·OH, which could enhance membrane permeability and facilitate their efficient uptake by 4T1 cells. Besides, particles with an average size of approximately 100 nm are conducive to endocytic uptake (Qin et al., 2023). Moreover, as illustrated in Fig. 2B-C and S3, the CLSM quantitative and FCM analyses also showed consistent results, suggesting that the BAM NPs possessed high uptake efficiency.

Fig. 2.

(A) CLSM images of 4T1 cells at 3、6、9 h after incubation with FITC/BAM NPs. (B) Fluorescence intensity values of FITC in 4T1 cells calculated using ImageJ software. (C) Flow cytometry analysis of FITC level in 4T1 cells after incubation at different time points. (D) Flow cytometer analysis of 4T1 cells pretreated with different endocytic inhibitors. (E) Fluorescence intensity statistics from flow cytometry. (*p < 0.05, **p < 0.01, ***p < 0.001).

To elucidate the cellular uptake pathway of BAM NPs in 4T1 cells, classic endocytosis inhibition assays were employed, including chlorpromazine, indomethacin and amiloride (Fig. 2D and E). Unlike the strong fluorescence in the FITC/BAM NPs group, 4 °C pretreatment significantly inhibited the signal in 4T1 cells, confirming energy-dependent uptake. Additionally, the cellular uptake of FITC/BAM NPs was significantly diminished by treatment with chlorpromazine or indomethacin. In contrast, amiloride treatment showed a weak inhibitory effect, and substantial FITC/BAM NPs were still taken up in 4T1 cells. Taken together, these results indicated that BAM NPs were internalized by 4T1 cells via an energy-dependent process, primarily through clathrin- and caveolae-mediated endocytosis.

It was essential that the nanoparticles could successfully escape lysosomal entrapment following endocytic uptake. To confirm this, the lysosomal escape assay was conducted. As shown in Fig. S4, most of the green fluorescence (FITC/BAM NPs) was overlapped with red fluorescence (Lysotracker red) after 2 h of incubation. However, the red and green fluorescence were distinctly separated by 8 h of incubation. These results demonstrated the lysosomal escape capability of nanoparticles, which was beneficial for enhancing delivery efficacy.

3.3. Ferroptosis induced by BAM NPs

Given that high GSH levels are critical for scavenging excess ROS, GSH depletion could serve as a potential strategy to disrupt redox balance. Consequently, GSH levels are generally recognized as an indicator of ferroptosis. As illustrated in Fig. 3A, the BAM NPs group exhibited the lowest GSH level among all groups, which was attributed to the reaction between Fe³⁺ in BAM NPs and GSH to generate Fe²⁺. Additionally, GSH depletion could downregulate the GPX4 expression (Zeng et al., 2023). More importantly, the GPX4 expression in the BAM NPs group was the lowest among all groups, which could be confirmed by WB analysis in Fig. 3B and C.

Fig. 3.

(A) The intracellular GSH levels in 4T1 cells following different treatments. (B) The expression levels of GPX4 in 4T1 cells were confirmed by Western blot analysis. (C) Corresponding normalized analysis of the expression of GPX4. (D) Flow cytometric analysis of ROS levels in 4T1 cells. (E) CLSM images of the ROS generation in 4T1 cells. (F) Fluorescent quantitation of ROS staining of 4T1 cells after different treatments. (G) CLSM images depicting the generation of LPO in 4T1 cells. (H) Flow cytometric analysis of LPO levels in 4T1 cells. (I) CLSM observation of changes in mitochondrial membrane potential after different treatments. (J) Relative MDA content of 4T1 cells after different treatments. (*p < 0.05, **p < 0.01, ***p < 0.001).

Generally, the reduction in GSH levels and the disruption of intracellular redox homeostasis could cause the accumulation of ROS. Later, the DCFH-DA probe was utilized to investigate intracellular ROS levels in 4T1 cells (Chen et al., 2023a). As depicted in Fig. 3E, the BSA group showed extremely weak green fluorescence intensity. In comparison, the ART group showed moderate green fluorescence. Notably, the BAM NPs group displayed the strongest green fluorescence, which indicated that BAM NPs could induce abundant intracellular ROS in 4T1 cells. This was because the BAM NPs not only produced ·OH via the Fenton reaction with H₂O₂, but also cleaved the peroxide bridge of ART, both of which finally enhanced ROS generation. Furthermore, FCM analysis also revealed that BAM NPs exhibited a higher capacity for ROS production, with a fluorescence intensity 10.44-fold greater than that of the Control group (Fig. 3D and F). These results aligned with the observations from CLSM.

The assessment of cellular LPO, a core characteristic of ferroptosis, was performed using C11-BODIPY staining (Qiu et al., 2023). As shown in Fig. 3G, the green fluorescence intensity was weak in the Control group and the BSA group, suggesting that the BSA group had almost no capacity to induce LPO in cells. It should be noted that the green fluorescence intensity in the BAM NPs group was significantly higher than that in the ART and BA NPs groups, indicating that the generation of LPO could be amplified by the joint action of ART and Fe²⁺. FCM results further indicated that the BAM NPs group exhibited the highest green fluorescence intensity, which was consistent with the CLSM results (Fig. 3H and S5). These results demonstrated that BAM NPs significantly increased the level of LPO in 4T1 cells. Moreover, the elevation in LPO was effectively inhibited by the ferroptosis inhibitors Fer-1 or DFO, which confirmed ferroptosis induction.

To determine whether excessive ROS induces oxidative stress, mitochondrial damage, and cell death, the mitochondrial membrane potential (MMP) in 4T1 cells was monitored using the JC-1 fluorescent probe following different treatments (Lv et al., 2023). As shown in Fig. 3I, 4T1 cells treated with BAM NPs displayed the weakest red fluorescence and the strongest green fluorescence, which signified the most severe mitochondrial damage. TEM observation of mitochondrial morphology revealed reduced volume, decreased number, and increased membrane density, which were features of ferroptosis (Fig. S6). Besides, malondialdehyde (MDA) was also a lipid peroxidation product and a common ferroptosis indicator. As depicted in Fig. 3J, MDA levels in the BAM NPs group were significantly increased. Collectively, BAM NPs effectively triggered tumor ferroptosis, which ultimately induced cellular damage in 4T1 cells.

3.4. In vitro cytotoxicity evaluation of BAM NPs

The cytotoxic activity of each formulation against 4T1 cells was detected via CCK-8 assay. As shown in Fig. 4A, 4T1 cells incubated with 100 μg/mL BSA exhibited over 80% viability, indicating that BSA had extremely low cytotoxicity and excellent biocompatibility. Notably, the BA NPs group exerted stronger cytotoxic effects on 4T1 cells than the ART group across all tested concentrations. After the coating of MPNs, a further significant increase in cytotoxicity was observed for the BAM NPs (IC₅₀ = 21.50 μg/mL). These results demonstrated that BAM NPs possessed the strongest tumor suppressive effect, which could be explained by the following reasons. Firstly, the improved water solubility and enhanced cellular uptake of BAM NPs facilitated an increase of intracellular ART concentration. Secondly, Fe³⁺ in MPNs could be converted to Fe²⁺ (Fig. S7). On one hand, Fe²⁺ reacted with intracellular H₂O₂ to generate large amounts of ROS. On the other hand, it mediated the cleavage of the endoperoxide bridge in ART, leading to additional ROS production. Collectively, these actions enhanced the antitumor activity of BAM NPs (Fig. 4B).

Fig. 4.

(A) The cell viability of 4T1 cells exposed to different samples for 24 h (n = 5). (B) The schematic diagram of BAM NPs for inducing ferroptosis in 4T1 cells. (C) Cell viability of 4Tl cells treated with BAM NPs and·Z-VAD-FMK, Necrostatin-1, 3-MA, Fer-1, DFO (n = 5). (D) CLSM fluorescence images of 4T1 cells co-stained with FDA (live cells, green) and PI (dead cells, red) after different treatments. (E) Flow cytometry analysis of apoptosis in 4T1 cells cultured on different samples. (*p < 0.05, **p < 0.01, ***p < 0.001). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The tumor cell-killing efficacy of different formulations was further evaluated by analyzing the distribution of green and red fluorescence via the FDA/PI staining assay. As illustrated in Fig. 4D, extensive green fluorescence signals were observed in the Control group, indicating that the 4T1 cells were in a viable state. In contrast, the BSA group showed extremely weak red fluorescence, which demonstrated that the BSA carrier had no significant cytotoxic effect on 4T1 cells. Compared with other groups, the BAM NPs group displayed the most pronounced red fluorescence intensity, confirming its superior tumor therapeutic efficacy. In addition, to assess the cellular death behavior in 4T1 cells, Annexin V-APC/7-AAD staining was performed (Fig. 4E and S8). The cell survival rate in the BSA group was approximately 91.27%, whereas the survival rates of cells treated with ART, BA NPs and BAM NPs were 58.17%, 48.73% and 36.30%, respectively. These findings demonstrated that BAM NPs efficiently induced 4T1 cell death through ferroptosis. In addition, the BAM NPs exhibited low toxicity and favorable biocompatibility in normal cells (Fig. S9).

To further validate the ferroptosis mechanisms, cell viability was assessed using pathway-specific inhibitors. As shown in Fig. 4C, different pathway-specific inhibitors were co-incubated with BAM NPs, including Fer-1 (a ferroptosis inhibitor), DFO (a ferroptosis inhibitor), 3-MA (an autophagy inhibitor), Necrostatin-1 (a necroptosis inhibitor) or Z-VAD-FMK (an apoptosis inhibitor). As illustrated in Fig. 4C-E, in the groups treated with Fer-1 or DFO, the 4T1 cells viability was increased. These results demonstrated that ferroptosis was the main pathway induced by BAM NPs.

3.5. In vivo antitumor efficacy

Based on the significant anticancer efficacy of BAM NPs in vitro, we further investigated the in vivo biodistribution. Initially, for real-time imaging, BAM NPs were labeled with CY5.5. As revealed in Fig. 5A-C, free CY5.5 group accumulated predominantly in the liver, indicating that free CY5.5 exhibited poor tumor enrichment capability in vivo. In contrast, the fluorescence intensity of BAM@CY5.5 NPs at the tumor site increased gradually and reached the maximum at 12 h. After 24 h, imaging and fluorescence analysis of excised tumors and vital organs were performed ex vivo. The results further confirmed that the BAM@CY5.5 NPs group exhibitedyhigher fluorescence intensity in tumor tissues than the free CY5.5 group. The efficient tumor accumulation of BAM@CY5.5 NPs could be attributed to the EPR effect. In addition, we observed that BAM@CY5.5 NPs exhibited a longer circulation time than free CY5.5 (Fig. S10). Overall, these findings demonstrated that BAM@CY5.5 NPs possessed excellent tumor retention ability, confirming their potential as a promising nanoplatform for breast cancer therapy (Wang et al., 2017).

Fig. 5.

(A) Fluorescence images of 4T1 tumor-bearing mice after intravenous injection of CY5.5 and BAM@CY5.5 NPs at 0, 3, 6, 12 and 24 h. (B) The fluorescence imaging of major organs and tumors harvested at 24 h after intravenous injection of CY5.5 and BAM@CY5.5 NPs. (C) Ex vivo fluorescence intensity of free CY5.5 and BAM@CY5.5 NPs distribution in tumor and main organs at 24 h post-injection. (D) Experimental design and timeline for assessing in vivo antitumor. (E) Images illustrating individual tumors from the indicated groups (n = 5). (F) Tumor growth curves of the mice after different treatments. (G) Tumor growth curve with time under various treatments (n = 5). (H) Tumor weights of mice subjected to various treatment formulations (n = 5). (I) Body weight curves of tumor-bearing mice with different treatments (n = 5). (*p < 0.05, **p < 0.01, ***p < 0.001).

Moreover, the 4T1 tumor model in mice was further established to assess the in vivo therapeutic effect of BAM NPs (Fig. 5D). Mice were randomly separated into five experimental groups (Control, BSA, ART, BA NPs, BAM NPs) once their tumor volume reached 100 mm³. As illustrated in Fig. 5E–G, the BSA group exhibited no tumor growth inhibition compared to the Control group. In addition, the ART group exhibited limited antitumor efficacy, which could be attributed to its poor tumor-enrichment capability and rapid clearance from the bloodstream. Especially, the highest antitumor efficiency and the smallest tumor volume were exhibited in the BAM NPs group, which could be attributed to the ability of BAM NPs to induce ferroptosis (Fig. S11). Tumor tissues were excised and weighed after all mice were euthanized on day 14. The BAM NPs group exhibited the smallest tumor weight, which was consistent with the tumor growth curve (Fig. 5H). Notably, the mice body weight remained stable in the BAM NPs group, implying minimal toxicity and high biocompatibility of the BAM NPs. (Fig. 5I).

Besides, H&E staining and TUNEL assays were performed on harvested tumor tissues to analyze cell death after different treatments (Zhang et al., 2025b). H&E staining demonstrated that BAM NPs induced extensive nuclear pyknosis and the highest proportion of dead cells (Fig. 6A). Moreover, the TUNEL assay showed that BAM NPs caused significantly greater DNA damage in tumor tissues compared with other groups, which was evidenced by enhanced red fluorescence (Fig. 6B). Specifically, the fluorescence images of GPX4 and LPO showed that treatment with BAM NPs resulted in the lowest GPX4 expression and the highest LPO level (Fig. 6C and D). These results confirmed that a remarkable tumor suppressive effect was induced in mice by BAM NPs through tumor ferroptosis.

Fig. 6.

(A) H&E staining in tumors from each treatment group. (B) TUNEL histological analysis photographs of tumor slices obtained from tumor-bearing mice treated with various treatments. (C) Immunofluorescence staining of GPX4 in tumor tissues after different treatments. (D) Fluorescence images of LPO in tumor tissues after various treatments. (E) H&E staining images of major organs from mice in the Control and BAM NPs group.

Next, the biosafety of BAM NPs was assessed in major organs. H&E staining of these organs in the BAM NPs group showed no significant pathological lesions compared with the Control group, indicating that BAM NPs exhibited low systemic toxicity and excellent in vivo biocompatibility (Fig. 6E). As shown in Fig. S12, the hemolysis rates of red blood cells were negligible. Overall, the results demonstrated that BAM NPs were promising nanoplatforms for breast cancer therapy, with prolonged tumor retention and low systemic toxicity.

4. Conclusion

In conclusion, this study successfully prepared a nano-delivery system (BAM NPs), which possessed favorable storage stability and displayed pH-responsive behavior. After being efficiently internalized by 4T1 cells, BAM NPs could release Fe³⁺ and ART in the acidic tumor environment. Both in vivo and in vitro experiments demonstrated that BAM NPs possessed significant antitumor effects by inducing tumor ferroptosis through generating abundant free radicals and exacerbating the accumulation of LPO. Overall, BAM NPs exhibited remarkable tumor therapeutic effects by triggering tumor ferroptosis, which provided a novel strategy for the treatment of breast cancer.

Abbreviation

ART	artesunate
BSA	bovine serum albumin
EGCG	epigallocatechin gallate
DFO	deferoxamine
Fer-1	ferrostatin-1
GSH	glutathione
FITC	fluorescein isothiocyanate
MB	methylene blue
ROS	reactive oxygen species
TMB	3,3′,5,5′-tetramethylbenzidine

CRediT authorship contribution statement

Yinghong Huang: Conceptualization, Formal analysis, Investigation, Methodology, Writing - original draft. Yuanyuan Zhao: Formal analysis, Investigation, Methodology, Writing - original draft. Yuhong Su: Conceptualization, Formal analysis, Investigation, Methodology, Writing - original draft. Chudong He: Investigation, Methodology. Mengdie Mo: Investigation, Methodology. Weijie Zhou: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition. Yang Lu: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization. Fei Yu: Writing – review & editing, Supervision, Resources, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.