Crystallinity‐Gradient Etched Hollow Mesoporous Zinc Nanoparticle for Dual‐Source H₂O₂‐Driven Tumor Ferroptosis Therapy

ABSTRACT

Zn²⁺ offers significant advantages for cancer therapy via mitochondrial disruption, immune activation, and metal homeostasis perturbation, but faces challenges of rapid clearance and systemic toxicity. Existing zinc‑based (e.g., zinc oxide (ZnO), and zinc peroxide) nanoparticles often lack internal voids and sufficient surface area, restricting drug loading and functionalization. In this study, we developed a crystallinity‑gradient selective etching method to tailor biodegradable hollow mesoporous zinc nanoparticle (HMZN). Typically, zinc peroxide nanoparticles with low‑crystallinity core and high‑crystallinity shell were synthesized, and ammonia water was used to selectively etch the inner low‑crystallinity core, producing HMZN. Hemin (HEM) and Gd‑poly(acrylic acid) macrochelates (GP) were then loaded to construct HEM@HMZN@GP for dual‐source H₂O₂ surge‐driven tumor ferroptosis therapy. In acidic tumor microenvironment, Zn²⁺, H₂O₂, HEM, and GP can be released from HEM@HMZN@GP. The released Zn²⁺ damages mitochondrial electron transport chain (ETC) and generates superoxide anions (•O₂ ⁻), which can be rapidly catalyzed by superoxide dismutase into H₂O₂. The dual‐source of H₂O₂ markedly elevates local oxidative stress. HEM can be enzymatically converted by heme oxygenase‑1 to Fe²⁺, which reacts with the abundant H₂O₂ via Fenton reaction to produce hydroxyl radicals, initiating lipid peroxidation and ferroptosis. GP enhances T ₁‑weighted magnetic resonance imaging signals, facilitating drug delivery visualization in tumors.

A crystallinity‐gradient selective etching method was developed to tailor biodegradable hollow mesoporous zinc nanoparticle (HMZN), which was used to load hemin (HEM) and Gd‐poly(acrylic acid) macrochelates (GP) to construct HEM@HMZN@GP for dual‐source H₂O₂ surge‐driven tumor ferroptosis therapy.

1. Introduction

Zinc (Zn), an essential trace element in the human body, has demonstrated immense potential in the field of cancer therapy [1, 2, 3]. High concentration of zinc ions (Zn²⁺) can induce mitochondrial dysfunction by disrupting the electron transport chain (ETC), activate the cGAS‐STING signaling pathway to bolster anti‐tumor immunity, and trigger intracellular metal homeostasis imbalance, thereby exhibiting potent anti‐tumor activity [4, 5, 6, 7]. However, direct injection of excessive Zn²⁺ for tumor treatment leads to rapid metabolic clearance and poses risks of systemic toxicity [8, 9]. In contrast, zinc‐based nanoparticles can serve as “micro‐ion reservoirs” to achieve the sustained release of Zn²⁺ in tumors, significantly enhancing therapeutic efficacy while minimizing side effects [10, 11, 12]. Therefore, the construction of zinc‐based nanoparticles with high anti‐tumor potency is a pivotal strategy for translating the biological activities of Zn²⁺ into effective clinical therapies [13].

Recently, Zhang et al. reported that orally administered zinc supplements can spontaneously assemble into zinc oxide (ZnO) nanoparticles under physiological conditions. These nanoparticles can extravasate through tumor vasculature, and enter into tumor cells. The released Zn²⁺ specifically binds to metallothionein‐1X (MT‐1X), which is highly expressed in Caki‐2 renal cancer cells. This binding triggers the ERK‐METTL3‐m⁶A signaling pathway, further upregulating intracellular MT‐1X expression. The accumulated MT‐1X acts as a molecular “magnet”, attracting and sequestering additional nanoparticles to create a self‐amplifying enrichment effect. Furthermore, the accumulated ZnO nanoparticles downregulate METTL3 expression, thereby activating the IFN‐γ‐Stat1‐Irf1 signaling pathway and bolstering anti‐tumor immunity [14]. Additionally, Tu et al. developed a zinc peroxide (ZnO₂)‐based nanomotor driven by near‐infrared (NIR) light. By asymmetrically modifying the ZnO₂ core with polydopamine (PDA) and chelating Fe²⁺, the photosensitive PDA coating generates a localized thermal gradient under NIR irradiation. This asymmetric temperature field drives the nanomotor via self‐thermophoresis, enabling it to actively penetrate deep tumor tissues like a “drill” and achieve efficient lysosomal escape, ultimately inducing tumor ferroptosis [15]. However, these zinc‐based nanoparticles lack effective internal cavities, resulting in limited specific surface areas, which severely constrains drug loading efficiency and material functionalization. Furthermore, drugs can only be adsorbed on the surface or embedded in shallow layers of zinc‐based nanoparticles due to the dense architectures, which results in highly susceptible to leakage or deactivation in complex physiological environments, generating off‐target side effects [16, 17, 18]. Therefore, the development of a zinc‐based nanoparticle that provides ample internal space and facilitates controlled drug release has become pivotal for enhancing the functionality of such materials [19, 20, 21].

In this study, we developed a crystallinity‐gradient selective etching method to tailor biodegradable hollow mesoporous zinc nanoparticle (HMZN). Typically, zinc peroxide nanoparticle (ZPN) with low‐crystallinity core and high‐crystallinity shell was synthesized, and ammonia water was used to selectively etch the inner low‐crystallinity core, producing HMZN. The HMZN integrates high drug‐loading capacity with exceptional biocompatibility. Importantly, the hollow core serves as a vast internal reservoir, while the mesoporous shell enables tumor microenvironment (TME)‐responsive controlled release, thereby minimizing off‐target effects and enhancing circulatory stability. After that, Hemin (HEM) and Gd‐poly(acrylic acid) macrochelates (GP) were then loaded to construct HEM@HMZN@GP for dual‐source hydrogen peroxide (H₂O₂) surge‐driven tumor ferroptosis therapy (Scheme 1a). Within the acidic TME, HEM@HMZN@GP undergoes stepwise degradation, releasing Zn²⁺, H₂O₂, HEM, and GP. Notably, HMZN itself decomposes in acidic TME to generate the first wave of H₂O₂ (chemical source). Concurrently, the released Zn²⁺ disrupts mitochondrial electron transport chain, causing electron leakage and formation of superoxide anions (•O₂ ⁻), which are rapidly catalyzed by superoxide dismutase (SOD) into a second wave of H₂O₂ (biological source). The dual‑source H₂O₂ surge substantially elevates local oxidant levels, creating a highly oxidative TME. HEM can be enzymatically converted by heme oxygenase‑1 (HMOX1) to Fe²⁺, which reacts with the abundant H₂O₂ via Fenton reaction to produce hydroxyl radicals (•OH), initiating lipid peroxidation and ferroptosis. Furthermore, GP can enhance T ₁‑weighted magnetic resonance imaging (MRI) signals, facilitating visualization of nanoparticle distribution in tumors (Scheme 1b).

SCHEME 1.

Schematic illustration for the synthesis of HEM@HMZN@GP (a), and the mechanism of dual‑source H₂O₂ surge‑driven tumor ferroptosis therapy by HEM@HMZN@GP (b).

2. Experimental Section

2.1. Materials and Reagents

Hydrogen peroxide (H₂O₂, 30%), ammonium hydroxide (NH₃·H₂O, 28%), polyvinylpyrrolidone (PVP, M_w = 10000), Rhodamine (R6G) and 3‐(4,5‐dimethylthiazol‐2‐yl)‐2,5‐diphenyltetrazolium bromide (MTT) were purchased from Sigma‐Aldrich (USA). Zinc acetate dihydrate (Zn(Ac)₂·2H₂O) and hemin were acquired from Shanghai yuanye Bio‐Technology Co., Ltd. Titanic sulfate (Ti(SO₄)₂) and 3,3’,5,5’‐Tetramethylbenzidine (TMB) were sourced from Aladdin (China) and Shanghai Macklin Biochemical Co., Ltd., respectively. Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum (FBS), trypsin and penicillin‐streptomycin were obtained from Invitrogen. The commercial kits and fluorescent probes, including DCFH‐DA, FITC‐Phalloidin, Endo/Lyso‐Tracker Green, cell plasma membrane staining probe (DIO), C11 BODIPY^581/591, Calcein‐AM/Propidium iodide (PI) Assay Kit, JC‐1 Assay Kit, and Annexin V‐FITC/PI Apoptosis Detection Kit, were supplied by Beyotime Biotechnology Co. (Shanghai, China).

2.2. Preparation of GP

First, 40.0 mL of a 4.0 mg/mL poly(sodium acrylate) (PAANa, Mw = 5100) solution was purged with nitrogen for at least 50 min to remove dissolved oxygen. Next, the polymer solution was refluxed at 100°C, followed by addition of a 125 mM Gd(NO₃)₃ solution (0.8 mL). After magnetic stirring at 100°C for 90 min, GP were obtained. After cooling to room temperature, the GP solution was dialyzed in Milli‑Q water for three days using a 6‐8 kDa cutoff membrane, with the dialysis water replaced twice daily. The purified GP was subjected to rotary evaporation for further concentration, and the resulting sample was quantified for Gd content by inductively coupled plasma optical emission spectrometry (ICP‐OES) [22, 23].

2.3. Synthesis of HMZN

10.0 mL of polyvinylpyrrolidone (PVP, Mw = 10000) solution (80.0 mg/mL) was mixed with 10.0 mL of zinc acetate dihydrate (Zn(Ac)₂·2H₂O) solution (48.0 mg/mL) in a 50.0 mL round‐bottom flask under magnetic stirring at 800 rpm. Subsequently, 2.0 mL of H₂O₂ (30%) was added dropwise to the mixture, followed by continuous stirring at room temperature for 24 h to yield a colloidal solution of ZPN. The reaction solution was then centrifuged at 10000 × g for 10 min. After washing and centrifugation, the resulting precipitate was redispersed in 25 mL of ultrapure water and ultrasonicated for 10 min at 40 kHz to obtain a uniform dispersion. To fabricate HMZN, 1.0, 2.0, 4.0, or 8.0 mL of NH₃·H₂O (28%) was directly added to the aforementioned dispersion (Table S1), followed by stirring at 500 rpm for 5.0 h. The products were collected by centrifugation at 10000 × g for 8.0 min at room temperature, washed three times with pure water under the same centrifugation conditions, yielding HMZN1‐4.

2.4. Synthesis of HEM@HMZN@GP

A homogeneous dispersion of HMZN3 (2.0 mL, 2.0 mg/mL) was mixed with 2.0 mL of HEM solution at concentrations of 0.13, 0.25, 0.5 or 1.0 mg/mL, followed by magnetic stirring at 500 rpm for 24 h (Table S2). After centrifugation at 10000 × g for 15 min, the precipitate was washed with pure water and centrifuged to obtain HEM1‐4@HMZN3.

Subsequently, HEM3@HMZN3 dispersion (2.0 mL, 2.0 mg/mL) was mixed with GP solution (4.0 mL) at Gd concentrations of 0.1, 0.2, 0.4, 0.8 or 1.0 mM, followed by magnetic stirring at 500 rpm for 24 h (Table S3). After centrifugation (10000 × g, 15 min) and washing, the precipitate was obtained as HEM3@HMZN3@GP1‐5.

2.5. Characterizations

Transmission electron microscopy (TEM, JEM‐2100F, JEOL) was employed to examine the morphologies and elemental mapping of the nanoparticles. The hydrodynamic diameters and surface zeta potentials were evaluated using a dynamic light scattering (DLS) instrument (Nano‐Brook 90PlusZata, Brookhaven). To determine the pore size distributions and specific surface areas, N₂ adsorption‐desorption isotherms were recorded on a Micromeritics ASAP 2020 M automated sorption analyzer. Optical absorption properties were monitored via a UV‐vis spectrophotometer (Evolution 300, Thermo Fisher). The elemental valence states and crystalline phases of the nanoplatforms were characterized by x‐ray photoelectron spectroscopy (XPS) and x‐ray diffraction (XRD, operated with Cu Kα radiation at 40 mA and 40 kV), respectively. The Zn content of nanoparticles was verified through inductively coupled plasma (ICP) analysis.

2.6. H₂O₂ Monitoring for HEM3@HMZN3@GP4

A colorimetric assay based on Ti(SO₄)₂ was utilized for H₂O₂ detection, wherein a distinctive color transition from colorless to yellow served as the indicator of their specific reaction. Briefly, 2.0 mL of HEM3@HMZN3@GP4 dispersion (1.0 mg/mL) was mixed with 20.0 mM of Ti(SO₄)₂ at pH 6.7, and the mixture was allowed to react for 0, 1.0, 2.0, 4.0, 8.0 or 16.0 h, respectively. After that, removal of the HEM3@HMZN3@GP4 nanoparticles was achieved via a 10‐min centrifugation process at 10000 × g, and the supernatant was collected for UV‑vis spectroscopic analysis.

A potassium permanganate (KMnO₄) colorimetric method was used to monitor H₂O₂ generation. Briefly, 2.0 mL of HEM3@HMZN3@GP4 dispersion (1.0 mg/mL) was incubated with 10 mM of KMnO₄ at pH 5.5, 6.7 or 7.4 for 0, 1.0, 2.0, 4.0, 8.0 or 16.0 h. After that, removal of the HEM3@HMZN3@GP4 nanoparticles was achieved via a 10‐min centrifugation process at 10000 × g, and the supernatant was collected for UV‑vis spectroscopic analysis.

2.7. Fenton Reaction Monitoring for HEM3@HMZN3@GP4

The progression of the Fenton reaction was tracked through a colorimetric assay employing 3,3’,5,5’‐tetramethylbenzidine (TMB). Briefly, 2.0 mL of HEM3@HMZN3@GP4 dispersion (1.0 mg/mL) was supplemented with a mixture of 5.0 mM TMB and 20 mM Fe²⁺ at pH 5.5, 6.7 or 7.4 for 0, 1.0, 2.0, 4.0, 8.0 or 16.0 h, respectively. Subsequently, the supernatant was harvested via centrifugation (10000 × g, 10 min) to exclude HEM3@HMZN3@GP4 and then subjected to UV‑vis spectroscopic analysis.

Moreover, spin‐trapping electron spin resonance (ESR) spectra were recorded to verify •OH generation using 5,5‐dimethyl‐1‐pyrroline N‐oxide (DMPO). Briefly, 2.0 mL of HEM3@HMZN3@GP4 dispersion (1.0 mg/mL) was incubated with 20 mM of Fe²⁺ at pH 5.5, 6.7 or 7.4 for 4.0 h. After that, the ESR spectra were subsequently gathered using the ESR spectrometer after blending 100 µL of DMPO into the reaction system.

2.8. Assessment of Biodegradability Performance

To evaluate the degradation behavior, 50 µL of HEM3@HMZN3@GP4 dispersion (20 mg/mL) was dispersed in 2.0 mL of PBS under varying pH conditions (5.5, 6.7 or 7.4) at 37°C with constant shaking. At predefined intervals (6.0, 12.0 or 24.0 h), 10 µL of the mixture was sampled and monitored via TEM.

2.9. In Vitro Drug Release

After enclosing 2.0 mL of HEM3@HMZN3@GP4 dispersion (1.0 mg/mL) inside a dialysis membrane (MW: 20 kDa), the setup was submerged in 10.0 mL of PBS (pH 5.5, 6.7, or 7.4) and maintained at 37°C in a shaking incubator (100 rpm). At designated timepoints (2.0, 4.0, 8.0, 16.0, 24.0, 48.0, and 72.0 h), aliquots of 0.5 mL release medium were sampled, followed by immediate replenishment with an equivalent volume of isothermal fresh PBS. The cumulative amounts of liberated HEM and metal ions (Zn²⁺ and Gd³⁺) were subsequently determined via UV‐vis spectrophotometry and ICP‐OES analysis, respectively.

2.10. In Vitro MRI Performance

A 3.0 T MRI scanner system (Philips, Ingenia, NL) was employed to evaluate the magnetic resonance imaging capability of HEM3@HMZN3@GP4. For the vitro MRI assay, HEM3@HMZN3@GP4 was dispersed in PBS solutions (pH 5.5, 6.7 or 7.4) to achieve gradient Gd concentrations (6.25 to 200 µM). Following a 24‐h incubation at 37°C, the mixtures were transferred to Eppendorf tubes (1.0 cm diameter) for magnetic resonance evaluation. The longitudinal (r ₁) or transverse (r ₂) relaxivity values were extracted from the linear slope by plotting the relaxation rate (1/T ₁ or 1/T ₂, s⁻¹) as a function of Gd concentration (mM). Under a 3.0 T magnetic field, the repetition time (TR) and echo time (TE) were configured at 200 ms and 8.2 ms for T ₁ mapping, and 5000 ms and 80 ms for T ₂ mapping, respectively.

2.11. MTT Assay

Cell viability of 4T1 cells was determined via a standard methyl thiazolyl tetrazolium (MTT) colorimetric assay. Typically, 4T1 cells were plated into 96‐well microplates at an initial concentration of 1.0 × 10⁴ cells/well and allowed to adhere overnight. Subsequently, the spent medium was replaced with 100 µL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III) or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). After 24.0 h of incubation, the medium was replaced with 100 µL of fresh one, followed by adding 10 µL of PBS‐based MTT (5.0 mg/mL). After 4.0 h of exposure, 150 µL of DMSO was introduced into each well to fully solubilize the purple formazan crystals. Ultimately, the optical density (OD) at 495 nm was recorded using a multi‐mode microplate reader (Synergy H1, BioTek Instruments, USA).

2.12. R6G Labeling of Nanoparticles

4.0 mL of HEM3@HMZN3@GP4 dispersion (C _Zn = 20.0 µg/mL) was co‐incubated with 0.50 mL of Rhodamine 6G (10 µM) under continuous magnetic stirring (300 rpm) at ambient temperature. Following a 24.0‐h reaction, the resulting HEM3@HMZN3@GP4@R6G was harvested via centrifugation (15000 × g, 10 min) and thoroughly rinsed with deionized water for subsequent applications.

2.13. Cellular Uptake

For confocal laser scanning microscope (CLSM) analysis, 4T1 cells were plated into confocal dishes (1.0 × 10⁵ cells/dish) and allowed to attach overnight. The spent medium was subsequently replaced with 1.0 mL of fresh one, or that containing HEM3@HMZN3@GP4@R6G (C _Zn = 20.0 µg/mL). Following a 4.0‐h incubation, the cells underwent standard rinsing with PBS, fixation using 4.0% paraformaldehyde (30 min), and subsequent permeabilization with 0.10% Triton X‐100 for 5.0 min. After being blocked in a 1.0% BSA solution for 30 min, the cells were co‐stained with DAPI and FITC‐Phalloidin for another 30 min prior to Nikon ECLIPSE Ti2 CLSM examinations.

For flow cytometry evaluation, 4T1 cells were plated into 6‐well microplates at an initial density of 5.0 × 10⁵ cells/well and left overnight for cell attachment. Subsequently, the spent medium was replaced with 2.0 mL of fresh one, or that containing HEM3@HMZN3@GP4@R6G (C _Zn = 20.0 µg/mL). After incubation for further 1.0, 2.0 or 4.0 h, the cells were rinsed twice using PBS and isolated via trypsin digestion and spinning (1000 × g, 5.0 min). The collected cell pellets were then resuspended in a 1.0 mL PBS for subsequent flow cytometric interrogation.

2.14. Observation of Lysosomal Escape

4T1 cells were plated into the confocal dishes (1.0 × 10⁵ cells/dish) and allowed to attach overnight. Subsequently, the spent medium was replaced with 1.0 mL of fresh one, or that containing HEM3@HMZN3@GP4@R6G (C _Zn = 20.0 µg/mL). Following a 4.0‐h incubation, the cells underwent standard rinsing with PBS, fixation using 4.0% paraformaldehyde (30 min), and subsequent permeabilization with 0.10% Triton X‐100 for 5.0 min. After being blocked in a 1.0% BSA solution for 30 min, the cells were stained with DAPI (30 min) and LysoTracker Green (1.0 h) sequentially, and ultimately subjected to Nikon ECLIPSE Ti2 CLSM examinations.

2.15. Calcein‐AM/PI Assay

4T1 cells were plated into the confocal dishes (1.0 × 10⁵ cells/dish) and allowed to attach overnight. Subsequently, the spent medium was replaced with 1.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III), HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV) or HEM3@HMZN3@GP4 (C_Zn = 160 µM) supplemented with NAC (a ROS scavenger, 5.0 mM). After 24.0 h of cultivation, the cells were double‐rinsed with PBS and co‐incubated with Calcein‐AM (2.0 µM) and PI (3.0 µg/mL) at 37°C for 30 min. Ultimately, the fluorescently labeled cellular samples were subjected to CLSM screening.

2.16. Cell Apoptosis Assay

4T1 cells (5.0 × 10⁵ cells/well in 6‐well plates) were cultured for 24.0 h to achieve adherence. The spent medium was then replaced with 2.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III), or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). Following a 24.0‐h incubation, the cells were washed twice with PBS and harvested via centrifugation (2000 × g, 3.0 min). The collected cell pellets were dispersed in 185 µL of chilled PBS, followed by the introduction of 5.0 µL of Annexin V‐FITC and 10 µL of PI. After a 15‐min light‐shielded co‐incubation, the mixtures were subjected to flow cytometric verification.

2.17. Evaluation of Intracellular ROS

To assess intracellular ROS production, the fluorescent probe DCFH‐DA was introduced. 4T1 cells were plated into confocal dishes (1.0 × 10⁵ cells/dish) and allowed to attach overnight. The spent medium was then replaced with 1.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III), or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). At 8.0 h post‐incubation, the cells were rinsed with PBS, and the fluid was changed to a DCFH‐DA‐containing culture medium (10 µM). Following an additional 30 min cultivation, the cells were rinsed with PBS and subsequently subjected to CLSM examinations.

4T1 cells were plated into 6‐well plates (5.0 × 10⁵ cells/well) and allowed to attach overnight for subsequent flow cytometric analysis. The spent medium was then replaced with 2.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III), or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). After 8.0 h of treatment, the cells were washed with PBS and incubated with DCFH‐DA (10 µM) for 30 min. Following a second chilling PBS rinsing, the cells were trypsinized, centrifuged (1000 × g, 5.0 min), and verified by flow cytometry.

2.18. Detection of Intracellular Fe2⁺

To track intracellular Fe²⁺ fluctuations, the commercial fluorescent probe RhoNox‐6 was introduced. Briefly, 4T1 cells were inoculated into confocal dishes (1.0 × 10⁵ cells/dish) and sustained overnight. The culture medium was then replaced with 1.0 mL of: (I) fresh medium (Control); (II) medium containing HEM (5.616 µg/mL); (III) HMZN3 (C_Zn = 160 µM); (IV) HEM3@HMZN3@GP4 (C_Zn = 160 µM); or (V) HEM3@HMZN3@GP4 (C_Zn = 160 µM) supplemented with ZnPP (an HMOX‐1 inhibitor, 10 µM). Following a 24.0‐h incubation, the PBS‐rinsed cells were co‐incubated with RhoNox‐6 (5.0 µM) and Hoechst 33342 at 37°C for 30 min. The cells were then thrice‐washed with PBS and inspected via CLSM. The red channel (excitation: 540 nm) was used to detect Fe²⁺‐activated RhoNox‐6 fluorescence, and the blue channel (excitation: 405 nm) was used to visualize cell nuclei. The inclusion of the ZnPP‐treated group served to confirm that the increase in labile iron pool was attributed to the enzymatic degradation of HEM mediated by HMOX‐1.

2.19. Assessment of Mitochondrial Membrane Potential (MMP)

The MMP was assessed with a commercial JC‐1 staining kit (C2006‐2, Beyotime, China). Briefly, 4T1 cells were inoculated into confocal dishes (1.0 × 10⁵ cells/dish) and sustained overnight The spent medium was then replaced with 1.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III), HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV), or HEM3@HMZN3@GP4 (C_Zn = 160 µM) supplemented with TPEN (a specific Zn²⁺ chelator, 10 µM). At 24.0 h post‐incubation, the cells were washed with PBS, stained with JC‐1 (10 µg/mL), thrice‐rinsed with PBS, and inspected via CLSM. The rescue effect of TPEN was evaluated to confirm the role of released Zn²⁺ in inducing mitochondrial dysfunction.

2.20. Evaluation of Intracellular Lipid Peroxides (LPO) Level

For CLSM analysis, 4T1 cells (1.0 × 10⁵ cells/dish in confocal dishes) were cultured overnight to ensure adherence. The spent medium was then replaced with 1.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III), or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). At 24.0 h post‐incubation, the cells were PBS‐rinsed, fixed (4.0% paraformaldehyde, 30 min), permeabilized (0.10% Triton X‐100, 5.0 min), and blocked (1.0% BSA, 30 min). After being stained via DAPI and BODIPY‐C11 581/591 (5.0 µM) for 30 min, the cells were verified using CLSM.

For flow cytometry analysis, 4T1 cells (5.0 × 10⁵ cells/well in 6‐well plates) were cultured overnight to ensure adherence. The spent medium was then replaced with 2.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III), or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). Following a 24.0‐h exposure, the cells were stained with the BODIPY‐C11 581/591 probe (5.0 µM) for 30 min. After dual washings with PBS, the targets were collected via trypsin digestion and spinning (1000 × g, 5.0 min), and the recovered cell pellets were re‐suspended in 1.0 mL of cold PBS for flow cytometric evaluation.

2.21. Measurement of Intracellular MDA Expression

4T1 cells (5.0 × 10⁵ cells/well in 6‐well plates) were cultured overnight to ensure adherence.

The spent medium was then replaced with 2.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III) or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). After 24 h of treatment, cells were DPBS‐washed, detached, sonicated, and centrifuged (8000 × g, 4.0°C, 10 min). The collected clear extract was examined at 450, 532, and 600 nm using a microplate reader for MDA level calculations.

2.22. Observation of Cell Membrane Integrity

4T1 cells were inoculated into confocal dishes (1.0 × 10⁵ cells/dish) and sustained for a 24‐h maintenance. The spent medium was then replaced with 1.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III) or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). At 24 h post‐treatment, cells were PBS‐rinsed, fixed (4.0% paraformaldehyde, 30 min), permeabilized (0.10% Triton X‐100, 5.0 min), and blocked (1.0% BSA, 30 min). After being stained via DIO (1.0 mM) and DAPI for 30 min, cells were verified using CLSM.

For flow cytometry analysis, 4T1 cells (5.0 × 10⁵ cells/well in 6‐well plates) were cultured overnight to ensure adherence. The spent medium was then replaced with 2.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III) or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). At 24 h post‐treatment, cells were cultured with DIO (1.0 mM), twice‐rinsed, trypsinized, and centrifuged (1000 × g, 5.0 min). The cells were re‐suspended within PBS and verified by flow cytometry.

2.23. Scratch Wound Assay for Evaluation of Cell Migration

4T1 cells (5.0 × 10⁵ cells/well in 6‐well plates) were cultured overnight to ensure adherence. Briefly, monolayers were straightly scratched using a sterile 200 µL tip, and nonadherent cells were cleared by washing twice with PBS. The spent medium was then replaced with 2.0 mL of fresh one (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C _Zn = 160 µM) (III) or HEM3@HMZN3@GP4 (C _Zn = 160 µM) (IV). The cells were maintained under standard conditions (37°C, 5% CO₂). Micrographs of the wounds were recorded at 0, 12, and 24 h via a conventional optical microscope. The denuded area was measured via ImageJ. And the migration rate was defined as the percentage of wound closure by comparing the time‐dependent gap areas to the initial scratch area (0 h).

2.24. Colony Formation Assay for Evaluation of Cell Proliferation

4T1 cells (1.0 × 10³ cells/well in 6‐well plates) were cultured under standard conditions (37°C, 5% CO₂) for 14 days to yield distinct cell colonies. Every 3 days, the spent medium was evacuated and substituted with an equal volume of fresh medium. (I), or that containing HEM (5.616 µg/mL) (II), HMZN3 (C_Zn = 160 µM) (III), or HEM3@HMZN3@GP4 (C_Zn = 160 µM) (IV). Upon completion of the incubation, the harvested colonies were fixed (4% paraformaldehyde, 15 min), stained (0.1% crystal violet, 20 min), extensively PBS‐washed, and quantified via a conventional optical microscope.

2.25. In Vivo Animal Experiments

All in vivo protocols complied strictly with the institutional Guidelines for Care and Use of Laboratory Animals at Southern Medical University (Guangzhou, China) and received formal sanction from its Animal Ethics Committee under license number SYXK(YUE)2021‐0167. Male BALB/c mice (4 to 5 weeks, ∼20 g) were purchased from the Animal Biosafety Lab (Level III) of Southern Medical University. 4T1 tumor‐bearing models were established via right‐flank subcutaneous injection (5 × 10⁶ cells in 100 µL of PBS) under isoflurane anesthesia (1.0%–2.0%). Tumors were measured at fixed intervals to extract the length (a) and width (b). Volume was defined as V = a × b² / 2, and the relative volume variation was derived from the V/V ₀ scaling.

2.26. Pharmacokinetics Study

HEM3@HMZN3@GP4 dispersion was intravenously injected into BALB/c mice (6 to 7 weeks, male, 20–25 g) with 5.0 mg/kg of Zn dosage. Briefly, blood samples were collected from the orbital sinus via a capillary tube at scheduled intervals (0, 1.0, 2.0, 4.0, 8.0, 12, 24, 48, or 72 h). The blood samples were then weighted, digested by 1.0 mL of 68% of nitric acid, and dissolved in 1.0% of nitric acid solution (8.0 mL). Finally, the Zn concentration in each sample was detected by the ICP‐OES.

2.27. Biodistribution of HEM3@HMZN3@GP4

Once the tumor volume developed to a range of 100–150 mm³, HEM3@HMZN3@GP4 was intravenously administered into the tumor‐bearing mice at a Zn equivalent dose of 5.0 mg/kg.

Then, the mice were sacrificed at 1.0, 6.0, 12, 24, or 48 h after injection. Briefly, the tumors and major organs were isolated and sliced into three functional segments. After weighting, the samples were oxidized within concentrated HNO₃ (5.0 mL, 68%), and dissolved with 8.0 mL of diluted nitric acid solution (1.0%). Finally, the Zn and Gd concentration in each sample was detected by the ICP‐OES.

2.28. In Vivo MR Imaging

Once the tumor volume developed to ∼200 mm³ in the 4T1‐bearing BALB/c mice, HEM3@HMZN3@GP4 was intravenously injected (Gd dosage: 5.0 mg/kg). Briefly, at 6.0∼48.0 h post‐injection, the mice were anesthetized (1.0%‐2.0% isoflurane) and monitored using a 3.0 T MRI apparatus equipped with a specialized animal body coil. For the control group, clinical Gd‐DTPA was intravenously administered at the same Gd dosage (5.0 mg/kg). Due to the rapid clearance of small molecules, MR images for the Gd‐DTPA group were acquired pre‐injection and at 10, 20, 30, and 40 min post‐injection. Mice were sustained at 37°C in a prone pose under respiratory gating during scanning. T ₁‐weighted imaging was executed pre‐ or post‐intravenous injection of HEM3@HMZN3@GP4 using an MSME pulse sequence (TR/TE = 400/8.0 ms, flip angle = 180°, matrix = 256 × 256, FOV = 40 × 40 mm², 16 slices, thickness = 1.0 mm). Grayscale data were segmented via ImageJ to deduce the destination SNR and ∆SNR variations.

2.29. Tumor Therapy Performance

To investigate the efficacy of HEM3@HMZN3@GP4 mediated ferroptosis therapy of tumors, the 4T1‐bearing mice were randomly allocated into four therapeutic cohorts once the tumor volume approached ∼100 mm³: (I) normal saline (NS); (II) HEM; (III) HMZN3; (IV) HEM3@HMZN3@GP4. 100 µL of NS, HEM (0.75 mg/kg), HMZN3 (Zn dosage = 5.0 mg/kg), HEM3@HMZN3@GP4 (Zn dosage = 5.0 mg/kg) were intravenously administered (i.v.) into the 4T1 tumor‐bearing mice on days 1, 4, and 7, respectively. Briefly, body weights and tumor volumes were recorded every 2 days. Mice were humanely sacrificed upon their tumor burden escalating to 2000 mm³.

The tumor weight was weighed and tumor histological analysis was carried out. The tumor tissues of each group were excised, fixed in 4.0% of paraformaldehyde, embedded in paraffin, sliced and stained with hematoxylin and eosin (H&E). Tumor sections were also stained with TUNEL apoptosis, Ki67 detection kit assay, and C11 BODIPY^581/591, respectively.

The excised tumor tissues were weighed, fixed (4% paraformaldehyde), paraffin‐sectioned, and stained with H&E for structural analysis. Additionally, cellular apoptosis, proliferative activities, and lipid peroxidation within the tumor matrices were systematically probed via TUNEL, Ki67 immunoflourescence, and C11 BODIPY^581/591 fluorescent staining, respectively.

2.30. Hemolysis Assay

Hemolysis experiment was performed to examine the hemocompatibility of the HEM3@HMZN3@GP4. To isolate pure RBCs, a 1.5‐mL blood specimen drawn from the orbital venous plexus was diluted with 3.5 mL of NS. After being centrifuged (400 × g, 10 min) and washed repeatedly, the pellet was re‐dispersed in a 5.0‐mL NS vehicle. Thereafter, 100 µL of RBCs was treated with 900 µL of NS, water, or HEM3@HMZN3@GP4 (12‐200 µg/mL) at 37°C for 2.0 h. After spinning (400 × g, 10 min), samples were photographed, and supernatant absorbances (A) were monitored at 540 nm. The hemolysis rate was extracted using the equation: Rate = [(A_sample—A_negative)/(A_positive—A_negative)] × 100%.

2.31. Biosafety Evaluation of HEM3@HMZN3@GP4 on Tumor‐Bearing Mice

The healthy Balb/c mice were intravenously administered with 0.10 mL of PBS or HEM3@HMZN3@GP4 (Zn dosage = 5.0 or 10 mg/kg). The mice were sacrificed at 24 h post‐injection for terminal blood collection. Centrifuged at 1500 × g for 15 min, the supernatant serum (300 µL) was isolated to quantify standard biochemical and hematological indicators. Target indicators encompassed biochemical variables (ALT, AST, BUN, CR, and CK‐MB) and routine blood vectors (WBC, RBC, PLT, MCV, MCHC, MCH, HGB, and HCT).

For the major organ toxicity detection, the healthy Balb/c mice were intravenously injected with 100 µL of saline (I), HEM (0.75 mg/kg) (II), HMZN3 (Zn dosage = 5.0 mg/kg) (III), HEM3@HMZN3@GP4 (Zn dosage = 5.0 mg/kg) (IV). Major organs (heart, liver, spleen, lung, and kidney) were collected at 72 h post‐injection. Organ slices were then analyzed via routine H&E staining for structural evaluation.

2.32. Statistical Analysis

Quantitative metrics are expressed as mean ± standard deviation (SD). Inter‐group variations were resolved via Student's t‐test (for dual cohorts) or one‐way ANOVA (for multiple groups). Statistical thresholds were assigned as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

3. Results and Discussion

3.1. Preparation and Characterization of HMZN

Transmission electron microscopy (TEM) images demonstrate that the ZPN exhibits a uniform particle size, excellent dispersibility and negligible agglomeration, thereby providing an ideal “precursor template” for subsequent etching processes (Figure S1). As shown in Figure S2, the core region (a) exhibits a diffuse halo in the FFT pattern, indicating its predominantly amorphous or low‐crystallinity nature. In contrast, the shell region (b) displays distinct diffraction spots/rings in the FFT pattern, confirming its higher crystallinity. This direct visual and diffraction evidence confirms the existence of a crystallinity gradient within the ZPN, which is the structural basis for our selective etching strategy.

As shown in the TEM images (Figure 1a–d) of HMZN1‐4 obtained after etching with NH₃·H₂O at varying concentrations (Table S1), it can be seen that HMZN1 and HMZN2 retained a substantial solid core, resulting in an incompletely hollow morphology. Notably, HMZN3 displays a hollow mesoporous architecture, with the solid core effectively removed and its spherical morphology well‐preserved. In contrast, HMZN4 displays severe structural collapse due to over‑etching with NH₃·H₂O. Furthermore, the maximum drug loading content of HMZN3 reaches approximately 10%, which is significantly higher than that of ZPN (∼2.9%) (Table S4), indicating that the hollow mesoporous structure of HMZN3 is the key factor for its markedly higher drug loading capacity than ZPN.

FIGURE 1.

(a‐e) TEM images of HMZN1‐4 (a‐d) and HEM3@HMZN3@GP4 (e). (f, g) SEM image (f), and Zn, O, Gd and Fe elemental mapping of HEM3@HMZN3@GP4 (g). (h‐k) XPS survey spectrum (h) and high‐resolution spectra of O 1s (i), Zn 2p (j), and Gd 4d (k) for HEM3@HMZN3@GP4. (l) XRD patterns of HMZN3 and HEM3@HMZN3@GP4. (m, n) Nitrogen adsorption‐desorption isotherms (m), and corresponding pore size distribution curves (n) of HMZN3 and HEM3@HMZN3@GP4.

The synthesis of HMZN3 involves two key steps. First, Zn(Ac)₂·2H₂O, H₂O₂, and PVP are reacted at a high initial concentration, giving a fast reaction that forms an inner core with low crystallinity. As the reaction continues, the concentration drops and the rate slows, producing a high‑crystallinity shell on the core surface [24, 25, 26]. The resulting ZPN shows high crystallinity outside and low crystallinity inside. Second, adding NH₃·H₂O to the ZPN dispersion causes ammonia molecules to undergo complexation with ZPN, preferentially attacking its low‑crystallinity inner core [27]. The following reaction occurs when the ZPN dispersion is treated with NH₃·H₂O:

$$\left(Zn{O}_2+4N{H}_3·H_{2}O\to \left[Zn{\left(N{H}_3\right)}_4\right]{\left(OH\right)}_2+H_{2}O\right)$$

During this process, ammonia coordinates with Zn²⁺ to form [Zn(NH₃)₄]²⁺ complexes, disrupting the inner core of ZPN. Due to the outer shell is highly crystalline and compact, NH₃·H₂O attacks it only weakly. This facilitates a more controllable etching process compared to simple inorganic bases, preserving the structural integrity of the hollow framework. As a result, a hollow mesoporous structure is generated. At low ammonia concentrations, NH₄ ⁺ migrates and reacts too slowly to fully penetrate and decompose the inner core of ZPN, whereas at excessively high concentrations, the reaction is too aggressive, causing non‑selective etching and structural collapse [28].

Unlike hard‐templating which requires the separate synthesis, coating, and subsequent removal of templates (e.g., SiO₂ or polystyrene beads), our method achieves the hollow structure through a “self‐etching” process within a single precursor. This significantly reduces the complexity and time of the synthesis. Hard‐templating often leaves trace amounts of template residues or etching reagents, which may impair the biocompatibility of the nanoplatform. Our approach utilizes the volatile and mild NH₃•H₂O as the etching agent, ensuring a cleaner product with improved safety for tumor‐targeted nanomedicine applications.

The x‑ray photoelectron spectroscopy (XPS) survey spectrum (Figure S3) shows the presence of Zn, O, and C elements (the C 1s signal primarily originates from adventitious carbon contamination or surface‐adsorbed organic species), which aligns with the elemental constituents of zinc peroxide. The high‐resolution O 1s spectrum (Figure S4) exhibits two distinct peaks: the peak at ∼531.8 eV is attributed to surface‐adsorbed oxygen species (including O₂ ⁻, OH⁻, or physisorbed H₂O/O₂), while the peak at 532.5 eV corresponds to peroxide ions (O₂ ²⁻). The high‐resolution Zn 2p spectrum (Figure S5) displays characteristic peaks at 1021.88 eV (Zn 2p^3/2) and 1044.88 eV (Zn 2p^1/2), which are well‐defined with no observable impurity peaks, indicating that Zn is stably present in the +2 oxidation state without significant valence shifts or secondary phase formation.

3.2. Characterization of HEM3@HMZN3@GP4

TEM (Figure 1e) and scanning electron microscopy (SEM) (Figure 1f) images reveal that HEM3@HMZN3@GP4 exhibits a typical hollow mesoporous structure, with uniform particle size and excellent dispersibility. As shown in Figure S6, the average diameter of HEM3@HMZN3@GP4 is approximately 73.9 ± 8.5 nm. The narrow distribution curve further confirms the excellent structural uniformity of our nanoplatform. Figure 1g displays the elemental mapping of HEM3@HMZN3@GP4, in which Zn and O (from HMZN), Gd (from GP), as well as Fe (from HEM) are clearly observed. As shown in Figure S7, the EDX spectrum clearly confirms the presence and co‐existence of Zn, O, Gd, and Fe in the nanoplatform, which is consistent with the elemental mapping results in Figure 1g, demonstrating that HEM3@HMZN3@GP4 is successfully fabricated.

The XPS survey spectrum (Figure 1h) shows characteristic peaks of Zn, O, C, and Gd, thereby confirming the elemental composition of HEM3@HMZN3@GP4. In the high‐resolution XPS spectra, the O 1s peak (Figure 1i) is deconvoluted into two components, one of which is assigned to lattice oxygen (Zn─O bond, ∼530 eV) and the other to surface‑adsorbed oxygen (O─O bond, ∼532 eV); the spin‐orbit doublet peaks of Zn 2p (Figure 1j) exhibit binding energies consistent with those of HMZN3 (∼1022.1 eV and ∼1045.1 eV), indicating that the oxidation state of Zn remains essentially unchanged upon drug loading; the fitted peaks of Gd 4d (Figure 1k) are assigned to Gd 4d^5/2 (∼140.5 eV) and Gd 4d^3/2 (∼148.1 eV), indicating the presence of Gd in HEM3@HMZN3@GP4.

The X‑ray diffraction (XRD) patterns (Figure 1l) show that the diffraction peaks of HMZN3 and HEM3@HMZN3@GP4 match the standard reference pattern (PDF#77‑2414), indicating that the loading process retains the crystalline structure of HMZN3. The Brunauer‐Emmett‐Teller (BET) analysis (Figure 1m,n) indicates that HMZN3 possesses an average pore size of 3.8 nm, which decreases to 3.3 nm after the loading process. The pore size distribution curves for both samples exhibit a single‐modal mesoporous structure. This reduction in pore size is presumably due to the partial occupation of the mesopores by HEM and GP.

The dynamic light scattering (DLS) analysis shows that HMZN3, HEM3@HMZN3, and HEM3@HMZN3@GP4 exhibit hydrodynamic diameters of 99.3, 122.1, and 131.8 nm as well as polydispersity indices (PDI) of 0.173, 0.203, and 0.212, respectively. (Figure S8). This gradual increase in particle size reflects the stepwise functionalization process. All formulations show PDI values below 0.22, indicating excellent colloidal dispersion and minimal nanoparticle aggregation. The zeta potential measurements (Figure S9) reveal that HMZN3 carries a positive surface charge of +20.3 mV, which likely arises from unneutralized surface hydroxyl groups or residual cationic sites on HMZN. HEM3@HMZN3 shows a small decrease in its positive zeta potential to +14.7 mV, a change attributed to the deprotonation of carboxyl groups within the HEM structure. Following loading, these negatively charged carboxylate moieties electrostatically interact with the positively charged HMZN3 surface, partially neutralizing the surface positive charge density and consequently reducing the overall zeta potential [29]. In contrast, HEM3@HMZN3@GP4 displays a strongly negative zeta potential of ‐25.2 mV, which can be directly ascribed to the incorporation of GP bearing hydrophilic functional groups such as carboxylates [30]. After 70‐hour of incubation in PBS, DMEM or DMEM + 10% FBS (simulating physiological environments), the hydrodynamic diameter of HEM3@HMZN3@GP4 remains remarkably stable (Figure S10). These findings confirm that HEM3@HMZN3@GP4 can exhibit robust colloidal stability in complex biological media.

3.3. TME‐Responsive Release and Degradation of HEM3@HMZN3@GP4 for Enhanced Chemodynamic Activity

As illustrated in Figure 2a, the detection strategy for chemodynamic activity is based on two coupled analytical steps: (1) In acidic medium, HEM3@HMZN3@GP4 degrades to produce H₂O₂, which subsequently reacts with titanium sulfate (Ti(SO₄)₂) to form a yellow chromogenic product, the amount of product is quantified by UV‐Vis absorbance spectroscopy [31]. (2) •OH detection: The generated H₂O₂ further reacts with Fe²⁺ to yield •OH, which oxidizes 3,3’,5,5’‐tetramethylbenzidine (TMB) to form oxTMB, displaying a characteristic absorption maximum at approximately 652 nm [32, 33].

FIGURE 2.

(a): Schematic illustration for detection mechanisms of H₂O₂ by titanium sulfate assay and •OH by TMB assay. In acidic environments, HEM3@HMZN3@GP4 degrades to release H₂O₂, which reacts with titanium sulfate to form a yellow product (for H₂O₂ quantification), or with Fe2+ to generate •OH. The •OH further oxidizes TMB to oxTMB, producing a blue color (for •OH quantification). (b): UV‐vis absorption spectra of HEM3@HMZN3@GP4 solutions at pH 6.7 with Ti(SO4)2 to measure the H₂O₂ generation at different time points (0, 1.0, 2.0, 4.0, 8.0, 16 h). (c): Time‐dependent UV‐vis absorption spectra of oxTMB (blue product) at pH 6.7 to testify the •OH generation. (d‐f): pH‐responsive drug release profiles of Zn2⁺ (d), Gd3⁺ (e), and HEM (f) from HEM3@HMZN3@GP4. (g): TEM images of nanoparticle morphology after 0, 6, 12, 24 h of incubation in pH 6.7 or 7.4 of buffers.

)The Ti(SO₄)₂ assay exhibits a time dependence H₂O₂ self‐supplying capacity (Figure 2b), reflecting the effective H₂O₂‑generating activity of HEM3@HMZN3@GP4. Moreover, this temporal trend is consistent with potassium permanganate titration results presented in Figure S11, demonstrating that acidic conditions (particularly at pH 5.5) promote the degradation of HEM3@HMZN3@GP4 and the subsequent release of H₂O₂.

In the •OH generation analysis (Figure 2c), the absorbance of oxTMB at 652 nm shows a time‑dependent increase at pH 6.7 over 0–16.0 h, indicating continuous •OH generation and reaching a maximum at 16.0 h. Furthermore, comparative UV‐Vis absorption spectra recorded at pH 7.4, 6.7, and 5.5 show a gradual increase in oxTMB absorbance as the pH decreased from 7.4 to 5.5 (Figure S12), suggesting that the •OH generation and Fenton reaction rate are pH‐dependent. The electron spin resonance (ESR) analysis (Figure S13) further corroborates this trend: well‑defined •OH signals appear in the magnetic field range of 324.4‐325.0 G for the pH 5.5 and pH 6.7 groups, with the signal intensity being highest at pH 5.5. Conversely, the pH 7.4 group shows only negligible signal intensity. These results indicate that HEM3@HMZN3@GP4 can effectively trigger the Fenton reaction under acidic conditions, leading to •OH generation.

Figure 2d reveals that the release rate of Zn stays under 0.50% at pH 7.4 and rises sharply to 60.1% and 84.9% at pH 6.7 and 5.5 following 72 h incubation at 37°C. As shown in Figure 2e,f, Gd and HEM display comparable release behaviors, with minimal release at pH 7.4 and maximal release at pH 5.5. Under acidic conditions, HEM3@HMZN3@GP4 undergoes preferential degradation, promoting the release of Zn²⁺ together with the encapsulated HEM and GP. Conversely, under neutral or mildly alkaline conditions, the structural integrity of HEM3@HMZN3@GP4 remains preserved, leading to slowed drug release. This pH‐sensitive release property ensures the stability of HEM3@HMZN3@GP4 within the bloodstream, minimizing off‐target release.

TEM observations (Figure 2g) reveal that after 24 h of incubation in PBS at pH 7.4, HEM3@HMZN3@GP4 maintains a well‐dispersed spherical morphology, demonstrating excellent stability under physiological conditions. In contrast, at pH 6.7, HEM3@HMZN3@GP4 exhibits slight aggregation and surface roughening. This phenomenon may arise from surface charge imbalance or localized structural collapse induced by HEM3@HMZN3@GP4 degradation [34].

3.4. Cell Uptake and Lysosome Escape

As shown in Figure 3a,b, red fluorescence signals derived from R6G are evident in 4T1 cells treated with HEM3@HMZN3@GP4, in contrast to the control group, indicating the successful internalization of HEM3@HMZN3@GP4. Flow cytometry results (Figure 3c,d) shows that the fluorescence intensity of 4T1 cells after 4.0 h of incubation with HEM3@HMZN3@GP4@R6G is significantly increased relative to that at 2.0, 1.0 or 0 h, suggesting that the cellular uptake of HEM3@HMZN3@GP4 is time dependent.

FIGURE 3.

(a, b) CLSM images of 4T1 cells pre‐ (a) or post‐treatment (b) with HEM3@HMZN3@GP4@R6G (red fluorescence from R6G) after staining by FITC‐phalloidin (green for cytoskeleton) and DAPI (blue for nuclei). (c, d) R6G fluorescence distributions (c) and quantitative analysis (d) of 4T1 cells after incubation with HEM3@HMZN3@GP4@R6G for different times (0, 1.0, 2.0, 4.0 h), measured by flow cytometry. (e) CLSM imaging of 4T1 cells after treatment by HEM3@HMZN3@GP4@R6G, and staining by DAPI and Lyso‐Tracker (green for lysosomes), showing the lysosomal escape. (f‐h) CLSM images of 4T1 cells pre‐ (I) or post‐treatment with HEM (II), HMZN3 (III) or HEM3@HMZN3@GP4 (IV) after staining by DHE (a •O2–specific fluorescent probe, red fluorescence) (f), BBoxiProbeTMA (a H₂O₂‐specific fluorescent probe, red fluorescence) (g) or HPF (a •OH ‐specific fluorescent probe, green fluorescence) (h). (i‐l) Fluorescence distributions (i, k) and the corresponding quantitative analysis (j, l) of 4T1 cells pre‐ (I) or post‐treatment with the II‐IV groups after staining by DHE (i, j), or by HPF (k, l), measured by flow cytometry. (m): The intracellular H₂O₂ levels in 4T1 cells pre‐ (I) or post‐treatment with the II‐IV groups for 12 h, tested by H₂O₂ assay kit.

The confocal laser scanning microscopy (CLSM) images of 4T1 cells incubated with HEM3@HMZN3@GP4@R6G (Figure 3e) show that most of the red fluorescence signal is largely localized outside the green fluorescence regions (lysosomes), suggesting successful lysosomal escape of HEM3@HMZN3@GP4. It has been reported that lysosome functions as an endogenous “iron reservoirs”, characterized by the presence of abundant acidic hydrolases and a labile iron pool (Fe²⁺/Fe³⁺) [35, 36]. Here, H₂O₂ reacts with Fe²⁺ via the Fenton reaction to generate •OH, leading to lysosomal membrane peroxidation and subsequent rupture [37].

3.5. Mechanism of HEM3@HMZN3@GP4‐Induced Ferroptosis in Tumor Cells

Figure 3f,g show the CLSM images of dihydroethidium (DHE, •O₂ ⁻ probe) or BBoxiProbeTMA (H₂O₂ probe)‐stained 4T1 cells pre‐ (I) or post‐treatment with HEM (II), HMZN3 (III) or HEM3@HMZN3@GP4 (IV) to monitor the intracellular levels of •O₂ ⁻ and H₂O₂. As shown, the 4T1 cells stained with dihydroethidium and BBoxiProbeTMA exhibit minimal red fluorescence pre‐ or post‐treatment with HEM, a moderate increase in fluorescence post‐treatment with HMZN3 and the highest fluorescence intensity post‐treatment with HEM3@HMZN3@GP4. In addition, the CLSM images of HPF‐stained 4T1 cells (HPF, •OH probe) exhibit the most intense green fluorescence after exposure to HEM3@HMZN3@GP4. (Figure 3h).

The corresponding flow cytometry measurements (Figure 3i,j) confirm that the 4T1 cells present the most intense DHE fluorescence when treated with HEM3@HMZN3@GP4. That's because excessive Zn²⁺ can primarily inhibit mitochondrial Complex III (ubiquinol‐cytochrome c oxidoreductase). This inhibition typically occurs at the Qi site, which blocks electron transfer and leads to the premature leakage of electrons to oxygen, subsequently forming •O₂ ⁻ [38]. Moreover, the flow cytometry measurements (Figure 3k,l) further confirm the highest HPF fluorescence after treatment with HEM3@HMZN3@GP4. This result stems from the reaction between Fe²⁺ and H₂O₂ (the Fenton reaction), which directly drives the catalytic generation of excess •OH [39]. As demonstrated in Figure S14, while the nanoplatform (Group IV) significantly increased intracellular Fe²⁺ levels, the pre‐treatment with ZnPP (Group V) effectively suppressed this iron elevation. This inhibition study confirms that the catalytic release of Fe²⁺ is indeed mediated by the upregulation of HMOX1 activity. The quantitative results measured by H₂O₂ assay kit (Figure 3m) demonstrate that the 4T1 cells treated with HMZN3 or HEM3@HMZN3@GP4 exhibit markedly higher H₂O₂ levels than those in groups I and II. This can be attributed to the fact that over‐produced •O₂ ⁻ is partially converted into H₂O₂ through the catalytic action of SOD [40]. As shown in Figure S15, treatment with HMZN leads to a dramatic increase in intracellular H₂O₂ levels (∼780% of control). Upon the addition of Mito‐TEMPO, the H₂O₂ level is significantly suppressed to ∼530%. This significant reduction (∼250% decrease) directly quantifies the contribution of mitochondrial‐derived H₂O₂, which originates from the Zn²⁺‐induced inhibition of the electron transport chain. The remaining elevation (∼430% above control) is primarily attributed to the chemical catalytic activity of the nanoscaffold itself. These data confirm that both chemical and biological pathways synergistically contribute to the H₂O₂ pool, with the chemical pathway serving as the dominant initial source that subsequently triggers mitochondrial dysfunction.

As shown in Figure 4a, the CLSM images of C11‐BODIPY^{5 8 1 / 5 9 1}‐stained 4T1 cells present red fluorescence dominance (reduced form) pre‐ or post‐treatment with HEM, a marked increase in green fluorescence (oxidized form) post‐treatment with HEM, and maximal green signal post‐treatment with HEM3@HMZN3@GP4. Flow cytometry data (Figure 4b,c) demonstrate that, among all tested groups, HEM3@HMZN3@GP4 treatment resulted in the strongest green fluorescence., which is attributable to •OH generated by HEM3@HMZN3@GP4. Consistent with expectations, malondialdehyde (MDA) is quantified at 11.3, 23.6, 57.4, or 121.5 nmol per 10⁴ cells pre‐ or post‐treatment with groups II‐IV, suggesting the strongest cellular LPO accumulation attributable to HEM3@HMZN3@GP4 (Figure S16). As shown in Figures S17 and S18, treatment with HEM3@HMZN3@GP4 (Group IV) leads to a significant downregulation of SLC7A11 (∼40% of control) and a dramatic decrease in GPX4 expression (∼25% of control). These molecular changes confirm the impairment of the glutathione‐dependent antioxidant system, a hallmark of ferroptosis. Notably, as a hallmark of ferroptosis, excessive LPO generation can directly assault cell membrane lipids, inducing membrane structural damage [41, 42, 43].

FIGURE 4.

(a) CLSM images of 4T1 cells pre‐ (I) or post‐treatment with HEM (II), HMZN3 (III) or HEM3@HMZN3@GP4 (IV) after staining by DAPI and C11‐BODIPY581/591 (red for reduced form of LPO; green for oxidized form of LPO) (b, c) Fluorescence distributions (b) and the corresponding quantitative analysis (c) of 4T1 cells pre‐ (I) or post‐treatment with the II‐IV groups after staining by C11‐BODIPY581/591, measured by flow cytometry. (d) CLSM images of 4T1 cells pre‐ (I) or post‐treatment with the II‐IV groups after staining by DAPI and DIO (green for cell membrane) showing the integrity of the cell membrane. (e, f) Fluorescence distributions (e) and the corresponding quantitative analysis (f) of 4T1 cells pre‐ (I) or post‐treatment with the II‐IV groups after staining by DIO, measured by flow cytometry. (g, h) CLSM images (g), or JC‐1 fluorescence quantitative analysis (h) of 4T1 cells pre‐ (I) or post‐treatment with the II‐IV groups after staining by JC‐1 (red for aggregates, green for monomers). (i) TEM images of mitochondrial structure in the group I and IV.

The CLSM images of 4T1 cells stained with DIO (a cytomembrane probe) reveal a well‐preserved membrane pre‐ or post‐treatment with HEM, a partially disrupted membrane post‐treatment with HMZN3, and an almost completely disintegrated membrane post‐treatment with HEM3@HMZN3@GP4 (Figure 4d). The corresponding flow cytometry analysis (Figure 4e,f) further confirms the lowest green fluorescence in the HEM3@HMZN3@GP4 group, indicating its potent ferroptosis‑inducing effect.

Figure 4g presents the CLSM images of 4T1 cells labeled with JC‑1 pre‐ or post‐treatment with groups II‐IV for assessment of intracellular mitochondrial membrane potential (MMP). As shown, the JC‐1 displays red fluorescence pre‐ or post‐treatment with HEM, reflecting the formation of JC‐1 aggregates within mitochondria, whereas treatment with HMZN3 results in a gradual enhancement of green fluorescence, culminating in maximal intensity after HEM3@HMZN3@GP4 exposure, indicating a shift toward monomeric JC‐1 within mitochondria. Flow cytometry analysis (Figure 4h) reveals that the proportion of JC‐1 in monomeric form is 6.52%, 10.5%, 54.5%, and 87.7% before or after treatment with groups II‐IV, with the highest value in group IV, indicating potent •OH generating capacity of HEM3@HMZN3@GP4. This occurs because excessive Zn²⁺ or •OH can impair mitochondrial integrity, consequently resulting in decreased MMP [44, 45, 46]. As shown in the JC‐1 staining results in Figure S19, the treatment with HEM3@HMZN3@GP4 leads to a dramatic loss of mitochondrial membrane potential. Notably, the addition of TPEN significantly reverses this effect. This result directly confirms that the liberated Zn²⁺ from HMZN is the primary trigger for mitochondrial impairment. Consequently, the 4T1 cells show the lowest level of adenosine triphosphate (ATP) following treatment with HEM3@HMZN3@GP4, which is consistent with the declined MMP owing to its vital role in mitochondrial ATP production (Figure S20). Consistent with expectations, mitochondria of 4T1 cells are observed to retain complete cristae structure pre‐treatment, but display disappeared cristae and broken membrane structure following treatment with HEM3@HMZN3@GP4, which aligns with the typical morphological characteristics of mitochondrial damage in ferroptosis (Figure 4i) [47].

3.6. In Vitro Antitumor Potency

Figure 5a presents the CLSM images of calcein‐acetoxymethyl (Calcein‐AM, green fluorescence) and propidium iodide (PI, red fluorescence) co‐stained 4T1 cells pre‐ (I) or post‐treatment with HEM (II), HMZN3 (III), or HEM3@HMZN3@GP4 (IV). The green or red fluorescence means live or dead cells [48]. As shown in Figure 5a, the 4T1 cells exhibit intense green fluorescence and minimal red fluorescence pre‐ or post‐treatment with HEM. In contrast, the 4T1 cells show a marked increase in red fluorescence after treatment with HMZN3 and the maximum red fluorescence after treatment with HEM3@HMZN3@GP4. The corresponding flow cytometry analysis (Figure 5b) shows that the proportions of necrotic (Q1) and late apoptotic (Q2) cells in group IV were significantly elevated, accompanied by a marked increase in early apoptotic (Q3) cells. In contrast, the proportions of apoptotic cells in groups I‐III were relatively low, demonstrating the potent cytotoxicity of HEM3@HMZN3@GP4 against tumor cells. As demonstrated by the CLSM images in Figure S21, the massive cell death induced by HEM3@HMZN3@GP4 is significantly rescued by the addition of NAC. This rescue effect confirms that ROS‐dependent pathways are the primary mode of action and provides strong evidence for the specificity of the induced cell death.

FIGURE 5.

(a) CLSM images of 4T1 cells pre‐ (I) or post‐treatment with HEM (II), HMZN3 (III) or HEM3@HMZN3@GP4 (IV) after staining by AM (green for live cells) and PI (red for dead cells). (b) Quantitative analysis of apoptosis for 4T1 cells pre‐ (I) or post‐treatment with the II‐IV groups after staining Annexin V‐FITC/PI, measured by flow cytometry. Q1: necrosis; Q2: late apoptosis; Q3: early apoptotic; Q4: viable cells. (c, d) Representative images from the scratch wound healing assay at 0, 12, and 24 h post‐scratch (c), and quantification of cell migration rates at 12 and 24 h (d) in the I‐IV groups. (e, f) Representative images of colony formation assays in 4T1 cells pre‐ (I) or post‐treatment by the II‐IV groups (e), and quantification of colony numbers (f). (g) Cell viabilities of 4T1 cells after the corresponding treatments (groups II‐IV) for 24 h. (Mean ± SD, n = 3. ****p < 0.0001).

The wound healing assay and quantitative analysis (Figure 5c,d) reveal that, compared with pre‐ or post‐treatment with HEM, the migration rates of 4T1 cells are significantly reduced after treatment with HMZN3 or HEM3@HMZN3@GP4 for 12 or 24 h. Notably, the 4T1 cells post‐treatment with HEM3@HMZN3@GP4 exhibit the most pronounced inhibitory effect on cell migration, suggesting that HEM3@HMZN3@GP4 can markedly impairs the migratory capacity of 4T1 cells. This phenomenon may be attributed to disruptions in energy metabolism or cytoskeleton remodeling following cellular injury [49].

The colony formation assay (Figure 5e,f) assessed the proliferative potential of cells. The results demonstrate that the 4T1 cells pre‐ or post‐treatment with HEM form a large number of colonies, post‐treatment with HMZN3 results in moderately reduced colony counts, and in contrast, the 4T1 cells post‐treatment with HEM3@HMZN3@GP4 form the fewest colonies, indicating that HEM3@HMZN3@GP4 exerts the strongest inhibitory effect on the proliferation of 4T1 cells. Integrating the wound healing assay results, it can be inferred that HEM3@HMZN3@GP4 can not only induces cell death but also elicits anti‐tumor effects through synergistic inhibition of proliferation and migration.

Figure 5g shows the cell viabilities of 4T1 cells after 24 h of treatments with groups II‐IV, which are determined by MTT assay. The cell viability of HEM3@HMZN3@GP4 (group IV) at the equivalent HEM concentration of 17.28 µM is 19.8% ± 3.6%, which is significantly higher than free HEM (group II) (72.3% ± 3.3%) and HMZN3 (group III) (45.6% ± 3.7%). The IC50 value of HEM3@HMZN3@GP4 was calculated to be approximately 82.4 µM (based on Zn concentration). This relatively low IC50 value demonstrates the potent anti‐tumor efficacy of our nanoplatform.

The above results of LSCM, flow cytometry analysis and MTT assay all demonstrate that the HEM3@HMZN3@GP4 possesses exceptional ferroptosis therapy efficacy on cancer cells.

3.7. Biodistribution, MRI Performance, and Pharmacokinetic Properties of HEM3@HMZN3@GP4

Figure 6a indicates that the circulating half‐life (t_1/2) of HEM3@HMZN3@GP4 in the bloodstream is calculated to be 7.1 h, which facilitates prolonged systemic circulation and promotes greater accumulation at the tumor site. Figure 6b,c shows the biodistribution of Zn and Gd in major organs and tumors for tracking the in vivo metabolic behavior of HEM3@HMZN3@GP4. The Zn and Gd elements have the highest enrichment in tumors at 12 h following injection of HEM3@HMZN3@GP4, which is attributed to the enhanced permeability and retention (EPR) effect [50]. In addition, the Zn and Gd elements also show the highest accumulation in spleen and liver at 12 h, likely due to the high phagocytic activity of macrophages in these tissues [51]. The released Zn²⁺ ions can be integrated into the body's endogenous zinc pool and are primarily excreted via the hepatobiliary‐fecal route and, to a lesser extent, the renal pathway. The Gd³⁺ ions are stabilized by the macrocyclic GP chelates, which prevent ion leakage. These stable complexes are expected to be gradually redistributed from the reticuloendothelial system (RES) into the bloodstream and ultimately eliminated from the body through renal filtration (urine). Over time, the levels of Zn and Gd progressively declined in all tissues., suggesting the favorable biosafety of HEM3@HMZN3@GP4.

FIGURE 6.

(a) Blood half‐life of HEM3@HMZN3@GP4 in mice (n=3). (b, c) Zn (b) and Gd (c) distribution in heart, liver, spleen, lung, kidney, and tumor tissues at 1, 6, 12, 24, and 48 h post‐injection (Gd dose: 5.0 mg/kg). (d) 3T MRI signal intensity of HEM3@HMZN3@GP4 at varying Gd concentrations (0‐200 µM) and pH levels (7.4, 6.7, 5.5). (e) T1 relaxation rate (1/T1) plotted as a function of CGd for HEM3@HMZN3@GP4 at pH 5.5, 6.7, or 7.4. (f) 3T MRI images of mice pre‐injection or at 6, 12, 24, or 48 h post‐injection (i.v.) of HEM3@HMZN3@GP4 (Gd: 5.0 mg/kg).

The T ₁‑weighted signal measured on a 3.0 T MRI scanner is significantly enhanced after incubation with acidic solutions (Figure 6d). This can be attributed to the release of GP from HEM3@HMZN3@GP4 under acidic environment, leading to increased longitudinal relaxivity (r₁) values. As the pH decreased from 7.4 to 5.5, HEM3@HMZN3@GP4 exhibits an increase in r₁ value (Figure 6e) from 17.6 to 41.1 mM⁻¹·s⁻¹ and a decrease in transverse relaxivity (r₂) value (Figure S22) from 71.2 to 52.7 mM⁻¹·s⁻¹, measured on a 3.0 T MRI scanner. Moreover, the r₂/r₁ ratio of HEM3@HMZN3@GP4 after incubation at pH 6.7 is calculated to be 2.19, which is sufficiently low for the T ₁‐Weighted MRI.

Figure 6f presents T ₁‑weighted MR images of 4T1 tumor‑bearing mice post‐injection (i.v.) of HEM3@HMZN3@GP4, acquired on a 3.0 T MRI scanner to monitor its in vivo distribution and metabolism. Notably, the T ₁‑MRI signal at the tumor site gradually increased over the first 12 h and then progressively decreased thereafter. Figure S23 shows the T ₁‐weighted MRI images at different time points after i.v. injection of Gd‐DTPA. As shown in Figure S24, a quantitative comparison of the tumor signal‐to‐noise ratio (ΔSNR) is provided. The results demonstrate that HEM3@HMZN3@GP4 achieves a significantly higher peak ΔSNR (∼145 at 12 h) compared to the maximum ΔSNR of Gd‐DTPA (∼65 at 30 min) at the same Gd dosage (5.0 mg/kg). Furthermore, while Gd‐DTPA is rapidly cleared, the HEM3@HMZN3@GP4 exhibits prolonged tumor retention, offering a much broader diagnostic window. The corresponding quantitative analysis (Figure S24) of tumor images based on ΔSNR presents a similar trend, indicating that HEM3@HMZN3@GP4 possesses excellent T ₁ imaging capability at the tumor site. It is worth noting that the Gd³⁺ ions in our nanoplatform are stabilized by macrocyclic chelates (GP). Compared to free Gd³⁺ ions, which are known to be highly toxic and can cause nephrogenic systemic fibrosis (NSF), macrochelates offer significantly higher thermodynamic stability and kinetic inertness. This strategic design minimizes the risk of Gd³⁺ leakage, thereby ensuring the superior biocompatibility of the HMZN platform for in vivo applications.

3.8. In Vivo Anti‐Tumor Efficacy of HEM3@HMZN3@GP4

Figure 7a illustrates the experimental therapy design for 4T1‑bearing mice, including intravenous injection of Saline (I), HEM (II), HMZN3 (III) or HEM3@HMZN3@GP4 (IV) on days 1, 4, and 7. The tumors exhibit rapid growth after injection of Saline or HEM, suppressed growth after injection of HMZN3, and the slowest growth after injection of HEM3@HMZN3@GP4 (Figure 7b). As shown in Figure 7c, the group IV exhibits a survival rate of 100% throughout the 36‐day period, surpassing those of groups I‐III. Figure 7d presents the gross morphology of excised tumor tissues. During the experimental period, mice in groups I‐IV show no significant decrease in body weight., suggesting that HEM3@HMZN3@GP4 exhibits excellent biocompatibility, as it effectively suppresses tumor growth without inducing obvious systemic toxicity (Figure 7e). As shown in Figure 7f, tumors in group I grow rapidly, while HEM treatment only moderately delays progression. In contrast, the groups III and IV demonstrate marked tumor suppression. Notably, the group IV exhibits the most potent antitumor effect, achieving a tumor inhibition rate of 79.8%, significantly higher than that achieved by groups I‐III (Figure 7g).

FIGURE 7.

(a): Schematic diagram of the experimental timeline. Tumor cells were subcutaneously inoculated into mice on Day ‐7. The 4T1 tumor‐bearing mice were treated with Saline (I), HEM (II), HMZN3 (III), or HEM3@HMZN3@GP4 (IV) via intravenous injection on day 1, 4 and 7. HEM dosage: 0.75 mg/kg; Zn dosage: 5.0 mg/kg; Gd dosage: 0.12 mg/kg. Mice were euthanized when tumor volume reached ∼2000 mm³. Body weight and tumor volume were measured every other day. (b‐g): Average tumor growth curves (b), survival percentages (c), photographs of excised tumors (d), changes in body weight (e), Individual tumor growth curves (f), and tumor inhibition rates (g) of 4T1 tumor‐bearing mice. (h): Representative histological images (H&E staining) and immunofluorescence staining for Ki67 (proliferation), TUNEL (apoptosis), and LPO (lipid peroxidation) in tumor tissues from each group.

Figure 7h shows representative light microscopy images taken from tumor tissues of groups I‐IV stained with hematoxylin and eosin (H&E), Ki67, TdT‑mediated dUTP nick‑end labeling (TUNEL), and C11‑BODIPY 581/591 on day 14. The H&E staining reveals that tumor cells exhibit pronounced nuclear condensation after treatment with HEM3@HMZN3@GP4, a hallmark of ferroptosis. The Ki67 immunofluorescence staining reveals that the tumor tissue exhibits no Ki67‐positive cells after treatment with HEM3@HMZN3@GP4, indicating that cell proliferation is effectively suppressed. Analysis of TUNEL immunofluorescence staining shows that the tumor tissue displays the greatest number of TUNEL‑positive nuclei following treatment with HEM3@HMZN3@GP4, reflecting the most extensive cell death. The C11‑BODIPY^581/591 fluorescence staining revealed that the tumor tissue exhibits the highest level LPO accumulation after treatment with HEM3@HMZN3@GP4, which indicates that tumor cell death induced by HEM3@HMZN3@GP4 corresponds to ferroptosis.

3.9. In Vivo Biosafety of HEM3@HMZN3@GP4

Figure S25 evaluates the blood compatibility of HEM3@HMZN3@GP4 through an in vitro hemolysis assay. The results show that the positive control group (H₂O) yielded a hemolysis percentage of approximately 100%, whereas the negative control group (saline) and treatment group (HEM3@HMZN3@GP4) at concentrations of 12.5‐200 µg/mL exhibit very low hemolysis rates (< 5.0%). This indicates that HEM3@HMZN3@GP4 does not induce red blood cell lysis and exhibits excellent in vitro blood compatibility.

Figure S26 shows the changes in white blood cells (WBC), red blood cells (RBC), platelets (PLT), and hemoglobin (HGB) in healthy mice after intravenous injection with HEM3@HMZN3@GP4 at doses of 5.0 mg/kg or 10 mg/kg. The scatter data of all dose groups remain within the normal reference range and show no significant difference from the saline group, suggesting that HEM3@HMZN3@GP4 does not induce abnormal changes in the counts of blood cells (such as RBCs, WBCs, and PLTs), further confirming its systemic blood safety. Figure S27 assesses the potential toxicity of HEM3@HMZN3@GP4 to major organs by measuring key enzymes of liver and kidney function (ALT, AST, BUN, CRE) and the cardiac injury marker (CK‐MB) in serum. The results show that the biochemical indicator scatter points of all dose groups (5.0 mg/kg and 10 mg/kg) remain within the normal range, and the group means do not significantly deviate from the saline group, indicating that HEM3@HMZN3@GP4 does not cause significant toxic effects on important organs such as the liver, kidneys and heart. Figure S28 shows the representative optical microscopy images of heart, liver, spleen, lung, and kidney tissues from healthy mice at 72 h after intravenous injection with 100 µL of saline (I), HEM (0.49 mg/kg) (II), HMZN3 (Zn dosage = 5.0 mg/kg) (III) or HEM3@HMZN3@GP4 (Zn dosage = 5.0 mg/kg) (IV), as assessed by H&E staining. Under the microscope, the cell structures of organs in all groups remain intact: myocardial fibers are regularly arranged (heart), hepatocyte boundaries are clear (liver), splenic lymphoid tissue appears normal (spleen), alveolar structures show no abnormalities (lung), and renal tubules and glomeruli maintain regular morphology (kidney). Compared with the group I, the group IV (HEM3@HMZN3@GP4) exhibits no pathological changes such as cell necrosis, inflammatory infiltration, or tissue structure destruction. The results indicate that HEM3@HMZN3@GP4 exhibits excellent biosafety while effectively exerting antitumor effects.

4. Conclusions

In summary, we have successfully developed biodegradable HMZN via a crystallinity‑gradient selective etching method, which overcomes the intrinsic limitations of conventional zinc‑based (e.g., ZnO and zinc peroxide) nanomaterials in drug loading capacity and functionalization. By precisely tuning the ammonia‑water etching process, ZPN was transformed into well‐defined HMZN3, exhibiting a markedly increased drug loading capacity (∼10%) compared to ZPN (∼2.9%). The constructed HEM@HMZN@GP displays excellent tumor microenvironment (TME)‑responsive degradation behavior, enabling controlled release of Zn²⁺, H₂O₂, HEM, and GP under acidic conditions. Notably, degradation of HEM3@HMZN3@GP4 in the TME not only provides a chemical source of H₂O₂ but also synergistically promotes a biological source of H₂O₂ through Zn²⁺‐induced mitochondrial dysfunction. This dual‑source H₂O₂ surge significantly elevates intratumoral oxidative stress. Released HEM is enzymatically converted by HMOX1 to Fe²⁺, which reacts with the abundant H₂O₂ via the Fenton reaction to generate •OH, triggering lipid peroxidation and ferroptosis. Moreover, GP can enhance T ₁‑weighted MRI signals, facilitating drug delivery visualization in tumors. In vitro and in vivo experiments show that HEM3@HMZN3@GP4 efficiently induces ferroptosis in 4T1 cells (evidenced by LPO accumulation, reduced MMP and decreased cell viability). Moreover, it suppresses tumor growth (79.8% inhibition rate), prolongs survival of tumor‑bearing mice (100% over 36 days) and causes no pathological damage to major organs (heart, liver, spleen, lung, kidney), validating its excellent antitumor efficacy and biosafety.

Author Contributions

H.C.: data curation, investigation, formal analysis, software, methodology, validation, visualization, and writing – original draft. J.Y.: data curation, investigation, formal analysis, software, methodology, visualization, and writing – original draft. Z.W.: investigation, formal analysis, software, methodology, validation. X.C.: formal analysis, software, methodology, visualization. Q.F.: investigation, formal analysis, and methodology. L.H.: methodology, validation, and visualization. F.Q.: software and validation. H.M.: formal analysis and visualization. T.B.: validation and visualization. B.R.: software and validation. X.Y.: formal analysis and visualization. Y.L.: conceptualization, investigation, formal analysis, resources, and writing – review & editing. S.N.: conceptualization, investigation, formal analysis, funding acquisition, project administration, supervision, and writing – review & editing. Z.S.: conceptualization, investigation, formal analysis, funding acquisition, project administration, resources, supervision, and writing – review & editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting Information is available from the online library or the authors.Supporting File: adhm71443‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.