Malignant melanoma remains the most life-threatening form of cutaneous malignancy, associated with high invasiveness and early metastasis. Current management strategies encompass surgical resection for localized disease, alongside systemic therapies such as BRAF/MEK inhibitors and PD-1/PD-L1 checkpoint blockade to address advanced cases [1–3]. Despite these approaches, acquired resistance substantially limits long-term survival outcomes, particularly in metastatic patients, resulting in a poor 5-year prognosis [4,5]. To address this critical issue, many potent therapeutic modalities such as photothermal therapy (PTT) [6,7], chemodynamic therapy (CDT) [8,9], sonodynamic therapy (SDT) [10], photodynamic therapy (PDT) [11], photoacoustic therapy (PAT) [12], magnetic hyperthermia (MHT) [13] and high-intensity focused ultrasound (HIFU) [14] have been developed to surmount traditional treatment limitations in melanoma management.

CDT emerges as an attractive modality due to its self-sustaining reactive oxygen species (ROS) generation mechanism (which obviates the need for additional external stimuli), cost-effectiveness, and non-specific cytotoxicity [15,16]. The therapeutic principle relies on transition metal ions (e.g., Fe²⁺, Cu⁺, Mn²⁺) catalyzing the Fenton-like reaction in the tumor microenvironment (TME), where hydrogen peroxide (H₂O₂) is decomposed into highly reactive hydroxyl radicals (^•OH) to kill cancer cells directly [17–19]. However, the inherent limitation lies in the insufficient endogenous H₂O₂ concentration (~100 × 10⁻⁶ mol/L) within the TME to sustain effective ^•OH production for therapeutic outcomes [8]. Nanomaterials designed to enhance intracellular H₂O₂ generation have emerged as a promising avenue. Notably, acid-responsive metal peroxides (e.g., CaO₂, CuO₂, BaO₂, MgO₂) have garnered attention as innovative candidates. Their pH-triggered decomposition at acidic pH (~6.0) within the TME enables selective H₂O₂ release while minimizing systemic exposure [20,21].

The elevated glutathione (GSH) levels (up to 1000-fold higher than normal tissues) in malignant cells create a robust antioxidant defense network that significantly compromises CDT efficacy through ^•OH scavenging [22]. This intrinsic redox imbalance creates a therapeutic resistance barrier, as evidenced by the compromised ROS generation capacity during CDT. Conversely, transition metal ions (e.g., Cu²⁺) serve as catalytic redox mediators that disrupt this defense system by oxidizing GSH, thereby creating a GSH-deficient TME that synergistically enhances CDT outcomes through the inhibition of ROS detoxification [23]. Continuous GSH depletion triggers glutathione peroxidase 4 (GPX4) inactivation, a pivotal event that compromises the lipid peroxide repair machinery and activates ferroptosis [24]. Ferroptosis, a distinctive programmed cell death mechanism, is driven by iron-catalyzed lipid peroxidation, ultimately leading to catastrophic membrane destabilization. Extensive research highlights its therapeutic potential, particularly as a targeted strategy for refractory cancers that exhibit resistance to conventional treatments [25]. In certain circumstances, ferroptosis and CDT could be intricately linked. High GSH levels in malignant cells impede CDT-induced ROS production. Thus, triggering ferroptosis can boost ROS generation. Conversely, CDT can also induce the lipid-peroxide-mediated cell death process of ferroptosis, enabling the two to surmount the shortcomings of single-modality treatments and augment anti-cancer therapeutic efficacy. However, conventional CDT administration modalities, such as oral and intravenous routes, are often associated with significant off-target toxicity due to non-specific biodistribution. Even local CDT application faces limitations from rapid drug release and high-concentration-mediated cytotoxicity, which restricts its clinical utility.

In recent years, injectable hydrogels have emerged as promising functional agents/drug delivery platforms owing to their rapid sol-gel transitions enabled by dynamic covalent bonds, including Schiff base cross-linking [26], borate-diol interactions [27], and Diels-Alder reactions [28]. These mechanisms not only facilitate controlled drug release but also enhance spatiotemporal specificity, overcoming the limitations of conventional systemic therapies [29,30]. Notably, Schiff-based hydrogels exhibit pH-responsive drug release behavior: They degrade and swell efficiently in the acidic TME while maintaining structural integrity at physiological pH, thereby minimizing premature drug leakage and ensuring targeted therapy [31].

Herein, we propose a GSH-depleting strategy utilizing adipic dihydrazide (ADH)-modified HA (HA-ADH)/aldehyde terminated polyethylene glycol (PEG-CHO) (HP) to deliver pH-responsive cupric peroxide (CuO₂) nanoparticles into the TME of melanoma cells. This self-healing hydrogel system (HPC) exhibits acid-triggered kinetics and maintains structural integrity under complex mechanical stress, enabling sustained GSH depletion to induce ferroptosis. Notably, the HPC hydrogel system self-supplies H₂O₂ through its redox-active components, compensating for the inherent H₂O₂ deficiency within the TME of melanoma cells. The integration of injectable hydrogels with functional CuO₂ nanoparticles addressed the limitations of conventional CDT. Both in vitro and in vivo results demonstrated HPC’s potent capabilities in eradicating malignant melanoma cells and inhibiting distant metastasis. Therefore, this study offers novel insights for the development of highly efficient and safe melanoma therapy.

The CuO₂ nanoparticles were synthesized as shown in Fig. 1A. By sequentially adding PVP, CuSO₄, NaOH, and H₂O₂, the dispersion color gradually transitioned from light blue to green and subsequently to yellow, indicating successful CuO₂ nanoparticles formation. The ultraviolet (UV) spectrophotometric analysis was conducted to characterize the synthesis process of CuO₂ nanoparticles (Fig. 1B). The appearance of a new absorption band at 400–600 nm indicated the successful preparation of CuO₂ nanoparticles [32]. The atomic force microscopy (AFM) image demonstrated that the size of CuO₂ nanoparticles was homogeneous with a thickness of about 96.80 nm (Fig. 1C and Fig. S1 in Supporting information).

In addition, KMnO₄ was utilized to confirm the successful synthesis of CuO₂ nanoparticles. Upon acidification, CuO₂ nanoparticles decomposed to release Cu²⁺ and H₂O₂. Subsequently, KMnO₄ reacted with H₂O₂ to generate yellow Mn²⁺ (Fig. 1D). And the UV absorption spectrum exhibited distinct changes after KMnO₄ treatment (Fig. 1E). The particle size of CuO₂ nanoparticles was quantitatively analyzed by dynamic light scattering (DLS), revealing an average diameter of 121.84 nm and a low polydispersity index (PDI) of 0.062. This low PDI indicates a narrow size distribution and favorable colloidal stability (Fig. 1F). The result is consistent with the AFM detection.

The proton nuclear magnetic resonance (¹H NMR) spectrum and Fourier-transform infrared (FTIR) spectrum were employed to detect the preparation of HA-ADH (Figs. 1G and H, Fig. S2 in Supporting information). HA-ADH showed a new characteristic absorption peak at 1700 cm⁻¹ compared to HA, which demonstrated the successful introduction of ADH groups. The emerging peak at 1.65 and 2.39 ppm in the ¹H NMR spectrum indicates the successful preparation of HA-ADH (Fig. S2). The grafting rate of HA-ADH was about 34.5%. Importantly, the successful modification of HA with ADH provides a functionalized backbone for the hydrogel system. A grafting rate of 34.5% can potentially optimize the hydrogel's cross-linking density and responsiveness.

Next, we explored the functional properties of CuO₂ nanoparticles that were central to the CDT mechanism, specifically for its ^•OH-generating and GSH-depleting capabilities. ^•OH can be generated through a Fenton-like reaction, primarily mediated by low-valence metal ions (e.g., Fe²⁺ and Cu⁺) reacting with H₂O₂ [33]. In the case of CuO₂ nanoparticles, acidity induces Cu²⁺ release. Typically, TME contains abundant GSH, which can reduce high-valence metal ions (such as Fe³⁺/Cu²⁺) to generate Fenton reagents like Fe²⁺/Cu⁺ [34]. 3,3′,5,5′-Tetramethylbenzidine (TMB) undergoes oxidation by ^•OH, resulting in a blue color change, and is thus widely used for ^•OH detection. The mechanism of TMB color development involves electron loss in the benzidine structure, leading to dimer charge-transfer complex formation (Fig. 2A). As shown in Figs. 2B and C, Cu²⁺ would catalyze H₂O₂ decomposition into ^•OH under GSH presence, producing a characteristic absorption peak at 655 nm and inducing a distinct blue color shift in solution. Critically, CuO₂ nanoparticles exhibited exceptional ^•OH-generating capacity under acidic conditions. The kinetics of ^•OH production by CuO₂ nanoparticles were investigated through systematic optimization of reaction time and pH, with ^•OH levels quantified as shown in Figs. 2D and E. These results confirmed the sustained and pH-responsive ^•OH production of CuO₂ nanoparticles over time. The efficient ^•OH generation by CuO₂ nanoparticles under TME-mimicking environmental conditions (pH 6.0) is highly promising for enhancing CDT during anti-tumor treatments. Furthermore, the continuous production of ^•OH over time may cause cumulative damage to tumor cells.

Next, the GSH scavenging capacity of CuO₂ nanoparticles was evaluated using the DTNB-based method. In this method, DTNB reacted with the sulfhydryl groups of GSH to produce yellow 2-nitro-5-thiobenzoic acid (TNB) (Fig. S3 in Supporting information), which exhibited a characteristic absorption peak at 412 nm, while unreacted DTNB showed negligible absorbance above 400 nm (Figs. 2F and G). A standard curve for GSH consumption was established based on absorbance measurements at 412 nm (Fig. S4 in Supporting information) Subsequently, the quantitative GSH scavenging efficiency of CuO₂ nanoparticles was determined. Notably, CuO₂ nanoparticles exhibited approximately 34.8% higher GSH scavenging efficiency compared to Cu²⁺ (Fig. 2H). Moreover, this capacity is enhanced continuously with increasing CuO₂ nanoparticle concentrations (Fig. 2I), which may be attributed to the cyclic regeneration of Cu²⁺ via the following mechanism: Cu²⁺ consumes GSH to generate Cu⁺, which then catalyzes H₂O₂ decomposition to regenerate Cu²⁺. Interestingly, under physiological pH conditions, CuO₂ nanoparticles exhibited negligible GSH scavenging activity. However, this effect was significantly amplified upon exposure to an acidic environment (Fig. 2J). The superior GSH scavenging efficiency of CuO₂ nanoparticles, along with its pH-dependent activation, is in line with the concept of disrupting the tumor cells' antioxidant defense system. Therefore, the functional nanoparticles could provide a TME-targeted approach to enhancing CDT efficacy in the acidic TME.

Motivated by the unique ^•OH-generating and GSH-depleting capabilities of CuO₂ nanoparticles, comprehensive characterization experiments were subsequently conducted on the synthesized hydrogel carrier systems. The features revealed by SEM-mapping and energy dispersive spectrometry (EDS) analyses further validated the successful preparation of the HPC hydrogel (Fig. S5 in Supporting information). The SEM images (Figs. S5 and S6 in Supporting information) clearly demonstrate that CuO₂ nanoparticles are uniformly dispersed within the hydrogel matrix. The HPC hydrogel could well maintain the morphological structure of the encapsulated nanoparticles without further aggregation. This anti-aggregation property facilitates homogeneous nanoparticle distribution, which is essential for achieving reliable anti-tumor treatment. Previous studies have established that the integration of nanomaterials within hydrogel systems can enhance their mechanical properties [35]. Herein, HP hydrogel without CuO₂ nanoparticles was prepared as the control. As shown in Figs. 3A and D, both HP and HPC hydrogels exhibited three-dimensional network structures. The mechanical property of the hydrogels was improved upon CuO₂ nanoparticles integration (Fig. S7 in Supporting information). At shear strains exceeding 200%, the storage modulus of both HP and HPP hydrogels still outstripped their loss modulus, indicating the good preservation of their gel state (Figs. 3B and E). The capacity of HPC hydrogel to preserve its gel state under high-shear strains is crucial for its in vivo stability. Such an outstanding ability allows them to endure the intricate and dynamic mechanical forces within the complex physiological environment of the body.

The self-healing behavior of hydrogels was assessed through rheological and dye staining techniques. The hydrogels were subjected to alternating low (1%) and high (500%) shear strains. After three complete gel-sol transitions, the storage and loss moduli of HPC hydrogel remained stable compared to HP hydrogel (Figs. 3C and F). Furthermore, the dyed hydrogels exhibited reversible stretchability after fusion (Figs. 3G and H). The self-healing property of HPC hydrogel is highly promising for biomedical applications. In complex physiological and pathological settings, tissues face diverse mechanical stresses like stretching, compression, and torsion. Our self-healing hydrogel system could maintain structural integrity, repair mechanical damage, and ensure continuous release of therapeutic agents during in vivo anti-tumor therapies. We further studied the release profile of CuO₂ nanoparticles from HPC under different pH conditions. As showed in Fig. S8 (Supporting information), the cumulative release reached 71.4% at pH 6 within 24 h (compared to 30.3% at pH 7), with almost complete release of CuO₂ nanoparticles observed at 72 h at pH 6. This confirms the sustained and pH-dependent release behavior of CuO₂ nanoparticles from HPC, which is consistent with the pH-responsive mechanism observed in similar systems [36]. Such pH-responsive CuO₂ nanoparticle release kinetics, together with the TME-triggered ^•OH production (as demonstrated in Fig. 2E) establish a promising application of the selective anti-tumor treatment with minimized systemic toxicity.

Notably, the synthesized HPC hydrogel exhibits outstanding injectability. HPC hydrogel samples loaded into a syringe could be injected smoothly for writing applications, and the extruded gel retained its structural integrity (Fig. 3I). The mechanical properties of HPC hydrogel are specifically engineered to balance appropriate mechanical performance while maintaining excellent injectability, allowing for minimally invasive needle delivery. After in vivo injection, the self-healing ability of the hydrogel system helps restore its structural integrity and forming a stable in situ depot around tumors. This depot ensures controlled release of therapeutic agents into the TME while maintaining spatial stability to prevent off-target distribution. Notably, the incorporation of functional nanoparticles enhances the in vivo stability of the hydrogel system, thereby supporting sustained antitumor activity (Fig. S9 in Supporting information).

We further evaluate the biocompatibility of the hydrogels and their impact on tumor cells, which directly relates to the therapeutic potential of the HPC hydrogel system. To evaluate the cellular biocompatibility of HP hydrogel, we assessed the effects of HP hydrogel leachate on NIH-3T3 and B16 cell proliferation under neutral pH condition. As shown in Fig. S10 (Supporting information), when cultured in DMEM/RPMI 1640 medium containing different concentrations of HP hydrogel leachates, no significant differences in cell growth kinetics were observed among all experimental groups (Figs. S10A and D). These results collectively demonstrate that HP hydrogel exhibited excellent biocompatibility, suggesting minimal cytotoxicity and potential safety for normal tissues under physiological pH environments.

TME is characterized by acidic pH and high GSH levels, which may reduce CDT efficiency. Within this context, the glutathione/glutathione disulfide (GSH/GSSG) redox system plays a critical role in scavenging ^•OH and alleviating oxidative stress. After co-culture with free CuO₂ nanoparticles and HPC hydrogel, B16 cells exhibited a significant decline in GSH content (Fig. S10B). The GSH/GSSG ratio, reflecting intracellular redox homeostasis, was notably decreased in tumor cells due to elevated oxidative stress (Fig. S10C) [37]. This ratio was further reduced upon free CuO₂ nanoparticles or HPC hydrogel treatment, indicating that CDT disrupts redox balance in B16 cells and enhances ^•OH-mediated tumor-killing effects. Therefore, the significant reduction in GSH content and the GSH/GSSG ratio validates the disruption of the tumor cells' antioxidant defense system.

Results of the cell counting kit-8 (CCK-8) and live-dead assays demonstrated that both free CuO₂ and HPC hydrogel induced marked tumor cell death, with a positive correlation between red fluorescence intensity (indicative of dead cells) and cell death proportion observed in these groups (Figs. S10E and F). Notably, the free CuO₂ group exhibited significantly higher cytotoxic effects compared to the HPC group, suggesting that CuO₂ nanoparticles may act as a critical factor in augmenting CDT-induced toxicity in vitro.

Next, we delve deeper into the underlying mechanism of how the CDT system in the HPC hydrogel elicits anti-tumor effects, specifically through ROS generation and ferroptosis induction. This ferroptosis-mediated process is a crucial element of our proposed therapeutic strategy. To investigate the ferroptosis induction capacity of the CDT system, B16 cells were treated with different treatment modalities (control, HP, free CuO₂ or HPC) for 6 h. 2′,7′-Dichlorodihydrofluorescein diacetate (DCFH-DA) staining confirmed that CuO₂ nanoparticles significantly elevated intracellular ROS levels in B16 cells. As shown in Fig. S11 (Supporting information), due to the direct absorption by tumor cells, the free CuO₂ group exhibited significantly higher ROS production compared to the HPC group (Fig. S11A). Further aminophenyl fluorescein (APF) detection verified that the generated ROS was ^•OH (Fig. S11B). Subsequently, ferroptosis-related parameters including lipid peroxidation levels and GPX4 expression were evaluated via Liperfluo staining and Western blot analysis. Notably, comparative fluorescence analysis demonstrated that both free CuO₂ and HPC formulations generated substantially stronger green fluorescence signals compared to control groups, with free CuO₂ showing greater intensity than HPC (Fig. S11C). Western blot analysis consistently revealed a significant downregulation of GPX4 expression in both the free CuO₂ group and the HPC group. Remarkably, free CuO₂ nanoparticles induced a more substantial suppression of GPX4 expression than the HPC hydrogel did (Fig. S11D). These findings collectively indicate enhanced lipid peroxidation in both free CuO₂ and HPC groups. The notable rise in intracellular ROS levels, lipid peroxidation, and downregulation of GPX4 are hallmarks of intracellular ^•OH generation and ferroptosis induction. These results establish that HPC hydrogel triggers ferroptosis in B16 cells through redox dysregulation mechanisms. Importantly, HPC hydrogel exhibited attenuated lipid peroxidation levels compared to direct drug administration, indicating a more controlled and safer means of inducing ferroptosis by preventing excessive oxidative stress and off-target effects.

Inspired by the comprehensive characterizations, in vitro investigations, and elucidated mechanisms of ferroptosis-mediated antitumor effects, we were motivated to pursue in vivo studies, which were approved by the Animal Ethics Committee of the State Key Laboratory of Oral Diseases, West China Hospital of Stomatology, Sichuan University. To investigate the antitumor efficacy of CDT under clinically relevant conditions, a B16 tumor-bearing mouse model was established. Mice were respectively treated with 100 µL saline, HP hydrogel, free CuO₂, or HPC hydrogel to evaluate their treatment responses (Fig. 4A). The results of comparative photography revealed that the HPC group exhibited superior tumor growth inhibition compared to other groups after 10 days of treatment (Fig. 4B). This visual trend was corroborated by quantitative analysis of tumor volume and weight (Figs. 4C and D), where the HPC group displayed the smallest tumor volume at the experimental endpoint. Histopathological evaluation via hematoxylin and eosin (HE) staining demonstrated that CuO₂- and HPC-treated tumors exhibited significantly larger necrotic areas compared to the Control and HP groups. Furthermore, immunohistochemical (IHC) analysis of tumor tissues confirmed that the treatment of both CuO₂ nanoparticles and HPC hydrogel induced marked downregulation of GPX4 expression in B16 cells (Fig. 4E). Previous studies have shown that copper nanoparticles exhibit long-term stability under physiological conditions and are primarily metabolized and cleared through hepatic and renal pathways [38,39]. It was reported that excessive systemic administration might be associated with hepatic, renal, neuronal, and immunological toxicity in animal models [40–42]. However, our localized delivery strategy employed a significantly lower dose, which is under the previously reported safe dose [42]. To comprehensively evaluate the local and systemic biocompatibility of CuO₂ nanoparticles within the HPC hydrogel system, we performed subcutaneous injections in mice and collected skin tissue samples on days 3, 5, 7, and 10 post-injection for histological analysis (Fig. S12 in Supporting information). HE staining revealed no signs of inflammation or tissue damage in the surrounding skin, indicating favorable local tolerance. Furthermore, systemic toxicity was assessed by HE staining of major organs, including the heart, liver, spleen, lungs, and kidneys, with no significant histopathological abnormalities observed across all treatment groups (Fig. S13 in Supporting information). Throughout the 10-day observation period, no signs of systemic toxicity or abnormal behaviors were detected, further confirming the excellent biocompatibility of the HPC hydrogel. Further histopathological evaluation revealed distinct metastatic lesions in pulmonary and renal tissues of the Control and HP groups, with scattered metastatic cells detected in the free CuO₂ group. However, no significant metastatic lesions were detected in the HPC group, confirming the potent anti-metastatic efficacy of HPC hydrogel against B16 melanoma progression (Fig. S13). These results demonstrate that CuO₂ nanoparticles exhibit potent anti-tumor activity. Once encapsulated within the HP hydrogel system, CuO₂ nanoparticles give rise to a sustained tumor-killing profile. This inherent property enables CuO₂ nanoparticles to continuously exert tumor-suppression effects on tumors over a 10-day treatment regimen. Consequently, the HPC hydrogel system not only elicits potent tumor-killing effects but also effectively impedes potential distant metastasis. We further evaluated the in vitro and in vivo antitumor efficacy of our hydrogel system using the 4T1 breast cancer model. The results consistently demonstrated that HPC hydrogel exerted potent therapeutic effects, comparable to those observed in the B16 melanoma model (Figs. S14 and S15 in Supporting information). These findings highlight the potential of our hydrogel to achieve effective antitumor outcomes across different tumor types, underscoring its adaptability to heterogeneous TME.

In conclusion, we engineered a self-healing and injectable HPC hydrogel system to potentiate ferroptosis in malignant melanoma via CDT. Mechanistically, the CuO₂ nanoparticles within the hydrogel endogenously generated H₂O₂, effectively compensating for the endogenous H₂O₂ deficiency in TME. Moreover, the hydrogel system was capable of sustaining GSH depletion through dual Cu²⁺/^•OH-cycling mechanisms. The depletion of GSH led to the inactivation of GPX4, thereby triggering the ferroptotic cascade. The induced ferroptotic cascade exhibited remarkable anti-tumor efficacy and significantly curtailed metastatic dissemination. Our collective findings demonstrate that the self-healing hydrogel system possesses dynamic ^•OH-generating and GSH-scavenging capabilities. It shows significant potential as a viable platform for enhancing the susceptibility of melanoma cells to both CDT and ferroptosis induction, thus offering a novel therapeutic approach for melanoma treatment.

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.

CRediT authorship contribution statement

Yongzhi Wu: Writing – original draft, Software, Methodology, Investigation, Conceptualization. Rong He: Writing – original draft, Methodology, Investigation, Data curation. Bowen Tan: Methodology, Conceptualization. Longjiang Li: Writing – review & editing, Supervision, Funding acquisition, Conceptualization. Jinfeng Liao: Writing – review & editing, Supervision, Funding acquisition, Conceptualization.

Acknowledgments

This work was financially supported by the National Natural Science Foundation of China (Nos. 32171354, 82372735). The authors thank BioRender.com (agreement Nos. VN280FZSDG, SZ280FZZZC, and PA280FZN1X) for providing the graphical tools used to create the schematic figures, including the graphical abstract, Fig. 1, and Fig. 4.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2025.111339.