A purpurin-based metalloimmunostimulant amplifies ROS and modulates STING/TNF-α axis to potentiate melanoma immunotherapy

Abstract

The low infiltration of pro-inflammatory immune cells and the sustained activation of multiple immunosuppressive signaling pathways in melanoma significantly limit the efficacy of clinical immunotherapy. Therefore, developing an effective immunostimulant with reversing the immunosuppressive tumor microenvironment (TME) is of great significance for improving melanoma immunotherapy. Herein, a degradable metalloimmunostimulant (PurpN/Mn@PEG) is developed for immunotherapy targeting immunosuppressive melanoma. The PurpN/Mn@PEG NPs are fabricated by coordination-driven self-assembly of purpurin and Mn²⁺, followed by polyethylene glycol (PEG) modification. PurpN/Mn@PEG dissociates in acidic pH and high glutathione TME, releasing PurpN and Mn²⁺. The nanoparticle exhibits peroxidase-/oxidase-like activity, generating a reactive oxygen species (ROS) storm that induces immunogenic cell death. PurpN/Mn@PEG amplifies ROS via H₂O₂ production through phenolic oxidation, enhances TNF-α secretion via CCAAT/enhancer-binding protein beta (CEBPB) upregulation, and sensitizes cGAS-STING pathway, synergistically boosting melanoma immunotherapy. In vivo experiments demonstrated that this purpurin-based metalloimmunostimulant exhibits remarkable therapeutic efficacy with an 87.8 % tumor growth inhibition rate in B16-F10 melanoma-bearing mice by activating multiple immune pathways, thereby effectively augmenting melanoma immunotherapy. This study provides an innovative therapeutic strategy that effectively reprograms the immunosuppressive TME to potentiate melanoma immunotherapy.

Graphical abstract

Our work introduces a pH/GSH-responsive metalloimmunostimulant, PurpN/Mn@PEG, fabricated by coordination-driven self-assembly of purpurin and Mn²⁺ followed by PEGylation. Once internalized, the nanocomplex dissociates to release purpurin and Mn²⁺, triggering a ROS storm that induces immunogenic cell death, upregulates TNF-α via CEBPB, and sensitizes the cGAS-STING pathway, thereby reprogramming the immunosuppressive melanoma microenvironment and markedly enhancing antitumor immunity in B16-F10 mice.

1. Introduction

Melanoma is one of the most aggressive malignant cutaneous neoplasms, comprising less than 5 % of all cutaneous malignancies, yet contributing to approximately 75 % of skin cancer-related mortality [1,2]. Surgery was the gold standard of care for resectable melanoma owing to the lack of effective therapeutic agents for melanoma prior to the emergence of novel treatments [3], such as immunotherapy. Immunotherapy has significantly improved the median survival of patients with advanced melanoma, extending it from 6 to 9 months to approximately 6 years [4,5]. Despite the clinical progress achieved with immunotherapy, many patients currently treated with these agents still experience disease progression [6,7]. A significant portion of melanoma patients will not benefit from immune checkpoint blockade (ICB), as many non-responders exhibit primary resistance, while a smaller fraction acquires resistance during treatment [8]. This resistance is primarily driven by poor anti-tumor immunity in “cold” tumors, a low frequency of pro-inflammatory immune cells, and an immunosuppressive network in the tumor microenvironment (TME), highlighting the continued need for novel therapies to improve outcomes for melanoma patients [9,10]. Therefore, finding a new immunostimulant that can modulate the tumor environment, including enhancing immune cell infiltration and relieving tumor-induced immune suppression, to reduce resistance is critically essential, yet remains challenging.

Purpurin (PurpN), a bioactive compound first extracted from the roots of Rubia cordifolia (madder plants) with anticancer activity, is a natural reactive oxygen species (ROS) inducer [11,12]. Additionally, PurpN exhibits multifaceted pharmacological activities, including anti-inflammatory, anti-mutagenic, antimicrobial, and neuro-protective potential associated with antioxidant action [[13], [14], [15]]. ROS are involved in all aspects of immune cell biology and function, and play indispensable roles as signaling molecules in T cell activation, which enhances antitumor immunity [16]. Nonetheless, the application of PurpN is impeded due to its short half-life, poor water solubility, and low chemical stability [17,18]. To address these limitations, coordinated assembly of PurpN with metal ions—via self-assembly of phenolic ligands and metal ions into nanoparticle, is an effective way to prolong blood circulation time and enhance tumor accumulation [[19], [20], [21]], thereby enhancing the immunotherapy efficiency. Additionally, metalloimmunotherapy may regulate immune modulation using metal ions, such as calcium (Ca²⁺), iron (Fe²⁺/Fe³⁺), manganese (Mn²⁺), zinc (Zn²⁺), copper (Cu²⁺), cobalt (Co²⁺), tungsten (W⁶⁺), etc [[22], [23], [24], [25], [26]]. Mn²⁺, the classical cGAS-STING pathway stimulator, plays a pivotal role in initiating antitumor immunity and converting “cold” tumors into “hot” ones [25]. Thus, PurpN/Mn²⁺ assemblies, formed through self-assembly between PurpN's phenolic hydroxyl groups and Mn²⁺, induce ROS amplification and facilitate co-delivery of both agents into tumors and modulate immunity, enable tumor-targeted co-delivery of both components and robust immunomodulation. This strategy offers a concise mechanism, multifunctional integration, and high translational potential, providing clear advantages over existing approaches for cancer immunotherapy.

Herein, a degradable purpurin-based metalloimmunostimulant (PurpN/Mn@PEG) is reported with amplifying ROS and modulating the STING/TNF-α axis to boost melanoma immunotherapy (Scheme 1). The PurpN/Mn@PEG is synthesized through coordination-driven self-assembly of Mn²⁺ and PurpN, a natural product-derived ROS inducer, followed by surface functionalization with polyethylene glycol (PEG) to augment colloidal stability and biocompatibility. Upon encountering the TME characterized by low pH and elevated glutathione (GSH) levels, PurpN/Mn@PEG undergoes dissociation facilitating the targeted release of its active components, PurpN and Mn²⁺. This nanoparticle exhibits intrinsic dual-enzymatic activities, including peroxidase- and oxidase-like activity. Both PurpN and Mn²⁺ generate ROS storms, enhancing antigen presentation and stimulating dendritic cell (DC) maturation, thereby amplifying the immunostimulatory effect of PurpN/Mn@PEG. Additionally, PurpN amplifies ROS via oxidation of its phenolic hydroxyl group to a quinone moiety, concurrently generating hydrogen peroxide (H₂O₂)—a key precursor for subsequent ROS cascades [27]. Furthermore, PurpN upregulates the transcription factor CCAAT/enhancer-binding protein beta (CEBPB), thereby increasing TNF-α levels, while Mn²⁺ acts as a potent sensitizer of the cGAS-STING pathway, synergistically augmenting anti-tumor immune responses. In summary, this work presents a novel immunotherapeutic approach utilizing PurpN/Mn@PEG, a coordination metalloimmunostimulant composed of immune-active metal ions and immune agonists. This agent reprograms the immunosuppressive TME by amplifying the ROS storm and co-modulating the STING–TNF-α axis, thereby augmenting melanoma immunotherapy.

Scheme 1.

Purpurin-based metalloimmunostimulants amplify ROS and modulate the STING/TNF-α axis to augment melanoma immunotherapy. (Top) Synthesis process of PurpN/Mn@PEG NPs: Schematic illustration of the coordination-driven self-assembly of Mn²⁺ and the ROS inducer purpurin, followed by surface modification with amine ester-terminated PEG. (Middle left) ROS amplification mechanism of PurpN/Mn@PEG NPs: Under high GSH and acidic conditions, PurpN/Mn@PEG NPs undergo lysis, releasing Mn²⁺ and purpurin. These components utilize H₂O₂ to generate ·OH, while purpurin further produces H₂O₂ by consuming O₂, thereby sustaining ·OH generation. (Middle right) The TNF-α pathway activation by PurpN/Mn@PEG NPs is mediated through the release of free purpurin. (Bottom) Therapeutic mechanism of PurpN/Mn@PEG NPs: The NPs enter cells and undergo degradation, releasing Mn²⁺ and PurpN. The synergistic effect of PurpN-induced ROS amplification and STING–TNF-α axis modulation significantly enhances melanoma immunotherapy efficacy.

2. Results and discussion

2.1. Synthesis, characterization, and molecular dynamics of PurpN/Mn@PEG NPs

Leveraging the potent antioxidant activity and excellent metal-chelating capacity of the natural polyphenol purpurin (PurpN), hollow nanoparticles (PurpN/Mn NPs) were fabricated via one-step coordination-driven self-assembly by simply mixing PurpN with Mn²⁺. Subsequent surface functionalization with mPEG₂₀₀₀–NH₂ significantly improved aqueous dispersibility, yielding PurpN/Mn@PEG NPs (Fig. 1a). SEM and TEM images revealed monodisperse, hollow nanospheres with an average diameter of 126.7 ± 14.9 nm and a shell thickness of 20.9 ± 2.9 nm (Fig. 1c–e). HAADF-STEM elemental mapping clearly resolved the distributions of C, O, and Mn (Fig. 1b), and these assignments were corroborated by EDS results (Figure S1). The zeta potential shifted from −24.3 ± 0.7 mV for PurpN/Mn NPs to −17.4 ± 2.2 mV after PEGylation, confirming successful grafting of PEG (Fig. 1f). XPS analysis of PurpN/Mn@PEG NPs exhibited characteristic Mn 2p, O 1s, and C 1s signals, whereas pure PurpN lacked Mn peaks (Fig. 1g, Figure S2, Table S1). The high-resolution Mn 2p spectrum displayed binding energies at 641.76 eV (Mn 2p3) and 653.56 eV (Mn 2p1), accompanied by a shake-up satellite, verifying the presence of Mn²⁺ (Fig. 1h–Table S2). ICP-MS determined the Mn content to be ≈ 18.5 wt %. UV–vis spectroscopy showed that free PurpN possesses absorption maxima at 205, 255, and 484 nm. Upon coordination with Mn²⁺, the 484 nm peak red-shifted markedly, while the 205 and 255 nm bands disappeared (Fig. 1i); concomitantly, the color changed from orange-red to modena (Fig. 1i). Quantitative UV–vis analysis indicated a PurpN loading efficiency of 43.8 % (Figure S3). FTIR spectra further supported coordination: the C–O stretching band at ∼1170 cm⁻¹ vanished, and the C=O stretch shifted from 1617 cm⁻¹ in PurpN to 1555 cm⁻¹ in PurpN/Mn NPs, consistent with Mn²⁺ binding to both hydroxyl and carbonyl groups of PurpN (Fig. 1j).

Fig. 1.

Synthesis and characterization of PurpN/Mn NPs. (a) Schematic illustration and (b) EDX elemental mappings of PurpN/Mn NPs. (c) SEM image and (d) TEM image of PurpN/Mn NPs. (e) Size statistical results of PurpN/Mn NPs. (f) Zeta potentials of PurpN/Mn and PurpN/Mn@PEG NPs in water. (g–h) XPS patterns of PurpN/Mn NPs (g) and high-resolution XPS patterns of Mn 2p (h). (i) UV–vis absorption spectra of PurpN, and PurpN/Mn NPs (Photographs of PurpN in DMF, and PurpN/Mn NPs in water). (j) FTIR patterns of PurpN and PurpN/Mn NPs. Results of RMSD (k) and SASA (l) of PurpN/Mn NPs from molecular dynamics simulations. (m) Molecular cluster changes of PurpN and Mn²⁺ during 100 ns simulation (green balls: PurpN; violet balls: Mn²⁺). (n) Detailed illustrations of the hydrogen bonds, coordination bonds, π-π stacking, and electrostatic interactions within the PurpN/Mn NPs. (o) Analysis of intermolecular hydrogen bonds in the PurpN/Mn NPs nano-cluster. (p) Binding energy analysis of the PurpN/Mn NPs. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Furthermore, molecular-dynamics (MD) simulations were used to elucidate the self-assembly mechanism of PurpN and Mn(CH₃COO)₂ in methanol. Previous experiments have demonstrated that PurpN compounds and Mn²⁺ ions can self-assemble into nanoclusters in methanol. To characterise the final architecture and the principal driving forces involved, we performed MD simulations of the PurpN/Mn system at a PurpN:Mn²⁺ ratio of 2:1, hereafter referred to as PurpN/Mn. The preferred binding poses between PurpN and Mn²⁺ obtained by AutoDock were used to build the simulation box. The root-mean-square deviation (RMSD) of all atoms relative to the initial structure is a standard metric for system stability. Fig. 1k shows the RMSD of the PurpN/Mn system as a function of simulation time. After initial fluctuations, the RMSD stabilised after 80 ns, reaching an average value of 2.700 ± 0.317 nm. Because self-assembly reduces solvent exposure, the solvent-accessible surface area (SASA) was used to monitor compaction. Fig. 1l displays the SASA trajectory; a pronounced decrease in the first 20 ns followed by a plateau yielded an average SASA of 59.754 ± 7.629 nm², confirming the formation of compact nanoclusters. To examine structural convergence, we clustered the 70–100 ns segments of all four systems (RMSD cut-off = 1.2 nm) with the GROMACS cluster module. The dominant cluster of each system (Figure S4) was taken as the final MD configuration. The PurpN/Mn system adopted a nanotubular morphology. To visualise the assembly pathway, representative snapshots extracted every 20 ns (Fig. 1m) reveal progressive compaction into a stable nanoparticle. Interaction-energy analysis of the central cluster shows strong intermolecular contacts. Specifically, hydroxyl groups on PurpN coordinate Mn²⁺, while oxygen atoms participate in extensive hydrogen bonding. π–π stacking among the aromatic rings and favourable electrostatics further stabilise the assembly. A hydrogen-bond inventory yielded an average of 49.850 H-bonds per frame (Fig. 1o). Energy decomposition with gmx energy (Fig. 1p) gave average Coulombic (Coul-SR) and Lennard-Jones (LJ-SR) energies of −14206.98 and −708.395 kJ mol⁻¹, respectively, indicating that electrostatic interactions dominate self-assembly. In summary, 100-ns MD simulations show that PurpN/Mn molecules spontaneously form stable spherical nanoclusters driven by hydrogen bonding, π–π interactions, and metal coordination.

2.2. PurpN/Mn@PEG nanoparticles exhibit dual responsiveness to both acidic pH and GSH

Given the weakly acidic microenvironment and elevated GSH levels in the TME, the release kinetics of PurpN from PurpN/Mn@PEG NPs were systematically evaluated in buffers at pH 7.4 and 5.2, with or without 10 mM GSH. TEM images confirm these trends: at pH 7.4 the hollow nanospheres retained their intact architecture, attesting to high physiological stability, whereas after 24 h at pH 5.5—or pH 5.5 plus 10 mM GSH—the particles gradually disintegrated into small fragments (Fig. 2a). At pH 7.4, the framework of PurpN/Mn@PEG NPs remained intact, preserving their hollow nanoscale spherical morphology and demonstrating excellent physiological stability. After 48 h of incubation in pH 7.4 supplemented with 10 mM GSH, only 13.4 % of PurpN was released. In contrast, at pH 5.2 or pH 5.2+ 10 mM GSH, the nanoparticles gradually disintegrated into debris within 24 h (Fig. 2b). Following 48 h incubation at pH 5.2, 78.4 % of PurpN was released; this value increased to approximately 82.6 % in the presence of 10 mM GSH, significantly higher than those observed under the other two conditions. This enhanced release is attributed to the weakening of coordination bonds resulting from the reductive dissociation of Mn²⁺ by GSH. Consistently, Mn²⁺ release—quantified by ICP-MS—exhibited a profile like that of PurpN (Fig. 2c). Collectively, these data demonstrate that PurpN/Mn@PEG NPs remain stable under normal physiological conditions, yet exhibit dual responsiveness to acidic pH and GSH, enabling precise and controlled intratumoral release of both PurpN and Mn²⁺. Furthermore, absorption peaks of PurpN/Mn@PEG NPs at approximately 484 nm and 255 nm appear after treatment with GSH (Fig. 2d), and the solution color changes from magenta to orange-red after incubation with acidic conditions/GSH (Figure S5). To assess the GSH-depleting capability of PurpN/Mn@PEG NPs, UV–vis absorption spectra were recorded during incubation with GSH solution at different time points and GSH concentrations. These spectra display a gradual enhancement of absorption intensity at 484 nm due to the reduction of Mn²⁺ by GSH (Fig. 2e and. f). Concurrently, GSH was rapidly oxidized to oxidized glutathione (GSSG) (Fig. 2g), indicating oxidation-reduction reactions between PurpN/Mn@PEG NPs and GSH. In summary, these observations confirm that PurpN/Mn@PEG NPs exhibit dual pH/GSH-responsive behavior, enabling controlled drug release and GSH depletion in the TME.

Fig. 2.

(a) Schematic illustration and TEM images showing the pH-responsive degradation and GSH-passivation of PurpN/Mn@PEG NPs. (b) The release behaviors of PurpN under different conditions. Data are presented as mean ± S.D. (n = 3). (c) The release behaviors of Mn element under different conditions. Data are presented as mean ± S.D. (n = 3). (d) UV–vis absorption spectra of GSH as well as PurpN/Mn@PEG incubated with or without GSH (10 mM). (e) UV–vis absorption spectra of PurpN/Mn@PEG NPs at different times. (f) UV–vis absorption spectra of PurpN/Mn@PEG NPs at different GSH concentrations. (g) Changes in GSH and GSSG contents in the presence of PurpN/Mn@PEG over time. (h) PurpN/Mn@PEG degradation, (i) H₂O₂ production, and •OH generation processes activated by PurpN/Mn@PEG, PurpN, and Mn²⁺. (j) H₂O₂ contents produced by PurpN and PurpN/Mn@PEG after 24 h incubation in PBS solution. (k) Schematic illustration of •OH detection using MB as indicator. (l) Time-dependent absorbance of MB in the presence of H₂O₂ and PurpN/Mn@PEG NPs. (m) UV–vis absorption spectra of MB solution treated with different groups. (n) ESR spectra of ·OH generation by DMPO as the trapping agent.

In recent years, delivering transition metal ions (e.g., Fe²⁺, Cu²⁺, or Mn²⁺) to tumor sites to convert endogenous H₂O₂ into cytotoxic hydroxyl radicals (•OH) via Fenton/Fenton-like reactions has attracted extensive research for skin cancer therapy [[28], [29], [30]]. However, limited endogenous H₂O₂ at tumor sites restricts •OH generation, which is further scavenged by elevated antioxidant GSH levels, enabling drug-resistant cancer cell survival [31]. Thus, developing nanoplatforms that generate intratumoral H₂O₂ could enhance melanoma treatment efficacy. Next, H₂O₂ production by PurpN/Mn@PEG was quantified. After 24 h incubation in PBS, PurpN/Mn@PEG generated H₂O₂ dose-dependently (>300 μM at 200 μg mL⁻¹; Fig. 2j), likely due to catechol oxidation in the presence of oxygen (Fig. 2i) [27]. However, PurpN alone produced negligible H₂O₂ even at 200 μg mL⁻¹ (Fig. 2j), indicating that nanoparticle disassembly into free PurpN is essential for ROS generation. Upon cellular internalization, acidic TME and high GSH trigger rapid reduction of PurpN/Mn@PEG into Mn²⁺ and PurpN. The released Mn²⁺ then undergoes Fenton-like reactions with H₂O₂, producing ROS and Mn⁴⁺. Subsequent reduction of Mn⁴⁺ by GSH generates GSSG and regenerates Mn²⁺, establishing a catalytic cycle that amplifies oxidative stress in tumor cells (Fig. 2h and. i). The catalytic activity of PurpN/Mn@PEG was assessed using methylene blue (MB) as a •OH indicator [32]. In the presence of GSH and H₂O₂, PurpN/Mn@PEG catalyzed MB oxidation (blue → colorless oxidized methylene blue (oxMB)), decreasing absorption at 660 nm and indirectly confirming •OH generation (Fig. 2k). Significantly reduced MB absorption (660 nm) and a color shift from lake blue to colorless were observed for PurpN/Mn@PEG + H₂O₂ + GSH, indicating enhanced •OH production (Fig. 2m and. l, S7). This effect surpassed that of PurpN/Mn@PEG + H₂O₂ alone (Figure S6d), while control groups (MB, MB + H₂O₂, MB + H₂O₂ + GSH) showed minimal changes (Figure S6a–c). These results confirm high Fenton-like activity by PurpN/Mn@PEG, augmented by GSH.Finally, •OH generation was directly verified via ESR spectroscopy using DMPO as a spin trap [24]. A characteristic 1:2:2:1 quartet signal for •OH was observed in the PurpN/Mn@PEG + GSH + H₂O₂ group (Fig. 2n), with stronger intensity than in PurpN/Mn@PEG + H₂O₂. No interpretable signals appeared in H₂O₂-only or GSH + H₂O₂ control groups. Collectively, these data demonstrate that PurpN/Mn@PEG initiates cascade catalytic reactions to amplify ROS for cancer therapy.

2.3. ROS-amplifying of PurpN/Mn@PEG for antitumor effect

To explore the cellular uptake efficiency, we constructed ^FITCPurpN/Mn@PEG NPs labelled with the fluorophore FITC, and then incubated with B16-F10 cells. As shown in Fig. 3b, cells treated with ^FITCPurpN/Mn@PEG exhibit much stronger fluorescence than those treated with PBS, and their fluorescence progressively increased during the co-culture period, indicating efficient cellular uptake (Fig. 3c). Next, the cytotoxicity in vitro, the anticancer effect of PurpN/Mn@PEG NPs was assessed using the Cell Counting Kit-8 (CCK-8) assay against B16-F10 cells. As shown in Fig. 3d, the cell viability of the PurpN/Mn@PEG-treated cancer cells decreased to ≈10.3 % after 24 h-treatment at 100 μg mL⁻¹, much lower than that of the free PurpN or Mn²⁺-treated cells, showing PurpN/Mn@PEG exerted the highest therapeutic effect on cancer cell growth compared to free PurpN and Mn²⁺. The live/dead fluorescence assay further confirmed the cytotoxicity of PurpN/Mn@PEG. Cells treated with PurpN/Mn@PEG displayed intense red fluorescence, showing that 62.9 % of the cells were dead, which highlights the synergistic killing effect of free PurpN and Mn²⁺ (Fig. 3e, Figure S8). Furthermore, the long-term effects of various treatments on B16-F10 cell proliferation were analyzed through clone formation assays. Compared to the PBS control group (≈213 clones), cells treated with free PurpN and Mn²⁺ reduced the clone number to roughly 50, whereas the PurpN/Mn@PEG group showed a sharp decrease to fewer than 10 clones, indicating a pronounced inhibition of cell proliferation (Fig. 3f and. g). Additionally, BCL-2 (B-cell lymphoma 2) functions as a potent anti-apoptotic protein, helping prevent cells from undergoing programmed cell death [[33], [34], [35]]. Compared with the PBS group, BCL-2 protein expression in PurpN/Mn@PEG-treated cells decreased markedly (Figure S9). These findings indicate the synergistic mechanism between PurpN and Mn²⁺ in PurpN/Mn@PEG, confirming its superior cytotoxicity.

Fig. 3.

ROS Amplification by PurpN/Mn@PEG for Enhanced Antitumor Activity. (a) Schematic illustration of ROS amplification via the oxidase-mimetic activity of purpurin for enhanced anticancer therapy. (b) Fluorescence images of B16-F10 cells after 24 h incubation with 5 μg mL⁻¹ FITC–PurpN/Mn@PEG. (c) Intracellular Mn content in B16-F10 cells treated with 5 μg mL⁻¹ PurpN/Mn@PEG for 0, 2, 6, 12, or 24 h. (d) Cell viability of B16-F10 cells treated with PurpN/Mn@PEG NPs, free PurpN, or Mn²⁺ at the various concentrations (Mn²⁺: 0, 3.1, 6.2, 12.5, 25, 50 μg mL⁻¹; PurpN: 0, 5.7, 11.5, 23, 46, 92 μg mL⁻¹) for 24 h (n = 3). (e) Representative live/dead-stained fluorescence images of B16-F10 cells after the different treatments. (f) Photographs and (g) quantification of B16-F10 colonies after the indicated treatments. (h) Fluorescence images and (i) relative ROS generation intensities in B16-F10 cells after the various treatments (n = 3). (j) Representative JC-1 fluorescence images and corresponding JC-1 aggregate/monomer ratios in B16-F10 cells after the indicated treatments. Data are presented as mean ± standard deviation (SD) (∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

Given the excellent •OH generation performance and brilliant anticancer effect in vitro of PurpN/Mn@PEG, the ROS levels in B16-F10 cells was detected using confocal laser scanning microscopy (CLSM) and flow cytometry (Fig. 3h and. i). Compared with PBS, cells incubated with free PurpN or Mn²⁺ displayed weak green fluorescence, whereas those in the PurpN/Mn@PEG group showed the most significant ROS generation, with fluorescence intensities approximately 30- and 70-fold higher than those of the free PurpN and Mn²⁺ groups, respectively, confirming that PurpN exerts oxidase-like activity, amplifying ROS production. High intracellular ROS levels are recognized to trigger mitochondrial damage [[36], [37], [38]]. We next evaluated the mitochondrial integrity in B16-F10 cells under various treatments using a JC-1 staining probe (Fig. 3j and. k). This probe shows red fluorescence from J-aggregates in normal mitochondria with higher membrane potential and green fluorescence from the JC-1 monomer in damaged mitochondria with lower membrane potential. The JC-1 value, which is the ratio of red to green fluorescence intensity, is used to evaluate the mitochondrial membrane potential and is highly positively correlated with it [39]. As shown in Fig. 3k, the PBS group exhibits intense red fluorescence, with a JC-1 value of 102. Cells treated with free PurpN or Mn²⁺ displayed reduced red fluorescence and enhanced green fluorescence, with JC-1 values of 73.9 and 63.9, respectively, indicating some mitochondrial damage in those groups’ cells. ROS amplification exacerbates this effect, with the maximal mitochondrial damage observed in cells after PurpN/Mn@PEG treatment, which exhibited the lowest JC-1 value of 26. These results demonstrate that PurpN/Mn@PEG, via its oxidase-like activity in an H₂O₂ cascade reaction, amplifies ROS generation, damages mitochondrial function, and achieves optimal therapeutic efficacy (Fig. 3a).

2.4. Immune-activating efficacy of PurpN/Mn@PEG NPs in vitro

It has been reported that dying cancer cells can induce immune responses through the release of tumor-associated antigens (TAAs), including high mobility group box 1 (HMGB1), calreticulin (CRT), and ATP. Subsequently, antigen-presenting cells (APCs), such as dendritic cells (DCs), are activated to promote immune responses [[40], [41], [42]]. We evaluated HMGB1 levels under various treatments using fluorescence microscopy. As shown in Fig. 4b, after incubation with PurpN/Mn@PEG, cancer cells exhibited weaker green fluorescence in the nucleus compared to cells treated with PBS, free PurpN, and Mn²⁺. This indicates that HMGB1 was released from the nucleus to the cytoplasm and further extruded from the cells. CRT expression was investigated through Western blot analysis (Fig. 4c and. d). Cells treated with PurpN/Mn@PEG displayed the highest CRT exposure compared to other treatments. The CRT expression levels in cells treated with free PurpN, Mn²⁺, and PurpN/Mn@PEG increased by 1.8-, 2.7-, and 4.9-fold, respectively, compared to the control group. Additionally, cells co-cultured with free PurpN, Mn²⁺, or PurpN/Mn@PEG exhibited significant reductions in cellular ATP levels, which were 2.1-, 2.0-, and 19.5-fold reduced by those in PBS-treated cells, respectively.

Fig. 4.

Immune-enhancing mechanisms of PurpN/Mn@PEG in vitro. (a) Schematic illustration of immune enhancement via ICD induction and cGAS-STING pathway activation by purpurin in vitro. (b) CLSM images of HMGB1 release from B16-F10 cells after the various treatments. (c) Intracellular CRT expression levels and (d) its corresponding protein level in B16-F10 cells under the indicated treatments. (e) Western blot analysis of STING, p-STING, p-TBK, and TBK protein levels; (h) relative expression levels of p-STING/STING and (j) p-TBK/TBK after the different treatments (n = 4). (f) CLSM images of cGAS expression (top) and fluorescence intensity profiles along the red line (bottom) in RAW264.7 cells after the various treatments. (g) Cellular ATP levels in B16-F10 cells after the indicated treatments (n = 3). (i) CLSM images of STING expression (top) and fluorescence intensity profiles along the red line (bottom) in RAW264.7 cells after the different treatments. (k) Flow cytometry analysis of DC infiltration (CD80⁺CD86⁺CD11c⁺) in B16-F10 tumor tissues after the indicated treatments. Data are presented as mean ± SD (∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

It is widely known that Mn²⁺ can enhance cGAS-STING activity, thereby initiating DC maturation and promoting immune responses. The levels of cGAS and STING in cells after various treatments were evaluated using immunofluorescence. As shown in Fig. 4f, cells treated with free PurpN, Mn²⁺, and PurpN/Mn@PEG exhibited significantly stronger green fluorescence compared to the PBS-treated group, confirming that these treatments activated cGAS and upregulated its expression. Additionally, the phosphorylation of STING is an important factor in promoting its aggregation, and the aggregation of STING is a key step in activating its downstream signaling pathways [43,44]. It was observed that green fluorescence aggregated in cells treated with free PurpN, Mn²⁺, and PurpN/Mn@PEG, while it was dispersed throughout the PBS-treated cells, suggesting that STING was activated in those groups (Fig. 4i). Subsequently, the expression levels of proteins associated with the cGAS-STING pathway was further analyzed by Western blotting. The ratio of non-phosphorylated protein to phosphorylated protein can be used to measure the relative extent of phosphorylation modification, and this ratio can reflect the dynamic changes in phosphorylation modification. As shown in Fig. 4e, h and 4j, the values of p-STING/STING and p-TBK/TBK in cells treated with free PurpN, Mn²⁺, and PurpN/Mn@PEG were decreased, confirming that both leached Mn²⁺ and free PurpN could cause strong cGAS-STING pathway activation. Specifically, the ratios of p-STING/STING and p-TBK/TBK in cells treated with PurpN/Mn@PEG both were the highest, at 5.3-fold and 119.8-fold higher, respectively, than those in PBS-treated cells. Additionally, we quantified type I interferon (IFN-β) transcription and three canonical interferon-stimulated genes (Ifit1, Cxcl10, and Isg15). All those mRNAs were significantly up-regulated in cells treated with Mn²⁺ or PurpN/Mn@PEG, confirming robust STING pathway activation. Flow cytometric analysis of bone marrow-derived dendritic cells (BMDCs) revealed that the contents of mature DCs (CD11c⁺CD80⁺CD86⁺ cells) in the four groups were 0.73 %, 7.8 %, 19.0 %, and 21.4 %, respectively (Fig. 4k). Therefore, the PurpN/Mn@PEG group exhibited the highest proportion of mature DCs compared to the PBS, free PurpN, and Mn²⁺ groups. These results confirmed that PurpN/Mn@PEG significantly activates cellular immunity through dual pathways: the immunogenic cell death (ICD) and cGAS-STING pathway, triggered by the synergistic effect of released free PurpN and Mn²⁺.

2.5. Potential antitumor mechanism of PurpN/Mn@PEG NPs

To systematically investigate the antitumor mechanism of PurpN/Mn@PEG, the gene transcriptomic analysis was performed on B16F10 melanoma cells co-incubated with free PurpN, Mn²⁺, and PurpN/Mn@PEG. Differentially expressed genes (DEGs) were screened, and pathway enrichment analysis was conducted to explore the molecular basis of the synergistic anticancer effects (Fig. 5a). Compared with the control group, volcano plot analysis revealed 4853 and 6167 DEGs in the free PurpN and PurpN/Mn@PEG groups, respectively, whereas the Mn²⁺ group exhibited only 2438 DEGs (Fig. 5b–d), indicating that PurpN/Mn@PEG treatment exerted a more pronounced regulatory effect on gene expression. To assess the roles of DEGs in response to various treatments, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were further performed to investigate functional pathway alterations in B16F10 cells (Fig. 5e and f, Figure S10). The top 20 most significantly enriched KEGG pathways are presented as bubble plots. The results demonstrated the differentially regulated pathways were mainly concentrated in key biological modules such as glutathione metabolism and tumor necrosis factor (TNF) signaling, indicating that the anticancer effects of PurpN/Mn@PEG NPs are closely associated with these pathways.

Fig. 5.

Potential antitumor mechanism of PurpN/Mn@PEG NPs. (a) Schematic illustration of the gene-transcriptome assay protocol. (b) Volcano plot of DEGs in B16-F10 cells treated with free PurpN versus control. (c) Volcano plot of DEGs in B16-F10 cells treated with Mn²⁺ versus control. (d) Volcano plot of DEGs in B16-F10 cells treated with PurpN/Mn@PEG versus control. (e) KEGG enrichment analysis of DEGs in B16-F10 cells (PurpN vs control). (f) KEGG enrichment analysis of DEGs in B16-F10 cells (PurpN/Mn@PEG vs control). (g) Heatmap of key DEGs changes. The heatmap shows the downregulation of DEGs associated with TNF-α pathway. (h) Relative mRNA expression of CEBPB in B16-F10 cells after various treatments. (i) Western blot analysis of CEBPB protein levels. (j) relative expression levels after the different treatments (n = 3). (k) CLSM images of CEBPB expression in B16-F10 cells after the various treatments. (l) The qPCR analysis of TNF-α expression in RAW264.7 cells after incubation with culture supernatants from B16-F10 cells subjected to various treatments. (m) ELISA analysis of TNF-α expression in RAW264.7 cells was performed following the same incubation conditions. Data are presented as mean ± SD (n = 3) (∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

Further hierarchical clustering analysis of the heatmap revealed significant alterations in key genes within the TNF signaling pathway (Fig. 5g). Specifically, in the TNF production pathway, the expression levels of mitogen-activated protein kinases (Mapk12 and Mapk11) and interleukin-17D (Il17d) was downregulated. Given that these treatments all showed significant enrichment in the TNF signaling pathway in KEGG analysis, and considering that CEBPB— a markedly upregulated gene in this pathway—has recently been reported to be closely associated with melanoma immune microenvironment activation and improved prognosis [45,46], we validated its expression using qPCR, western blot, and immunofluorescence. As shown in Fig. 5k, CEBPB was activated in the cells treated with free PurpN and PurpN/Mn@PEG. The main manifestations were not only an increase in the total amount in the nucleus but also the phosphorylation-dependent assembly into observable green spots, compared with the PBS and Mn²⁺ treatment group—a trend that aligns with the western blot and qPCR results (Fig. 5h–k). Furthermore, TNF-α expression in RAW264.7 cells after incubation with culture supernatants from B16-F10 cells subjected to various treatments was detected using an ELISA kit and qPCR. The free PurpN and PurpN/Mn@PEG treatment groups exhibited significant up-regulation relative to the other groups (Fig. 5l and m). These results were fully consistent with the transcriptomic data, further confirming the reliability of our sequencing results. In summary, PurpN/Mn@PEG exerts its antitumor effects by synergistically modulating the TNF production pathway and activating immune-related signaling, providing a solid molecular mechanistic foundation for traditional Chinese medicine-based interventions in melanoma.

2.6. In vivo anticancer efficacy

We first investigated the biosafety of PurpN/Mn@PEG NPs before conducting in vivo anticancer treatment. A hemolytic assay using red blood cells was carried out to assess the hemocompatibility of PurpN/Mn@PEG NPs. After 4 h of incubation, the hemolysis rate was less than 3.3 % within the PurpN/Mn@PEG NPs concentration range of 3.12–100 μgmL⁻¹, expressed as equivalent Mn content, indicating good hemocompatibility (Figure S11). Additionally, blood biochemical analyses revealed that the hepatic (aspartate aminotransferase (AST), alanine aminotransferase (ALT), albumin (ALB), alkaline phosphatase (ALP)), and renal (blood urea nitrogen (BUN), creatinine (CREA)) functional markers remained within physiological reference ranges (Figure S14), suggesting negligible liver and kidney toxicity. Subsequently, we explored the pharmacokinetic characteristics of PurpN/Mn@PEG NPs in B16-F10 tumor-bearing mice using ICP-MS. The mice were sacrificed 24 h after intravenous administration of PurpN/Mn@PEG NPs, and the major organs (heart, liver, spleen, lung, and kidney) and tumors were harvested. The Mn content in these tissues was then detected, showing that PurpN/Mn@PEG NPs were predominantly accumulated in the liver, spleen, and tumor (Fig. 6h). These findings reveal that PurpN/Mn@PEG NPs exhibited good biocompatibility and were suitable for in vivo applications in B16-F10 tumor-bearing mice, demonstrating enhanced tumor accumulation via the enhanced permeability and retention (EPR) effect.

Fig. 6.

In vivo anticancer efficacy of PurpN/Mn@PEG NPs in B16-F10 tumor xenograft murine models. (a) Schematic illustration of the therapeutic protocol for treating B16-F10 tumor-bearing mice. (b) Individual tumor volume growth curves and (f) average tumor volume growth curves of B16-F10 tumor-bearing mice after various treatments ((1) PBS; (2) Free PurpN; (3) Mn²⁺; (4) PurpN/Mn@PEG). (c) Digital photograph and (d) tumor weights of tumors excised from B16-F10 tumor-bearing mice on day 8 after the indicated treatments (n = 8). (e) Tumor growth inhibition rates after various treatments for 8 days. (g) H&E staining, Ki67 immunohistochemistry staining, and TUNEL staining of tumor sections excised from B16-F10 melanoma xenograft mice on day 8 after the indicated treatments. (h) Mn content in major organs and tumors excised from B16-F10 melanoma xenograft mice on day 8 after the indicated treatments. (i) Average Ki-67 percentage determined from panel i (n = 3). Data are presented as mean ± SD (∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001).

Encouraged by the superior ability of PurpN/Mn@PEG to effectively combine with immunotherapy in vitro, we further evaluated its anti-tumor effects in vivo. The therapeutic protocol was schematically outlined in Fig. 6a. The mice bearing B16-F10 melanoma model were randomly divided into four groups when the tumor volumes reached ∼50 mm³: (i) PBS, (ii) free PurpN, (iii) Mn²⁺, and (iv) PurpN/Mn@PEG at the PurpN concentration of 27 mg kg⁻¹ or Mn²⁺ concentration of 30 mg kg⁻¹ (n = 8 per group). Each group was intravenously injected on days 0, 2, 4, and 6 with these drugs. The tumor volume and body weight of the mice were measured every other day during the 8-day treatment period. As shown in the tumor growth curves, the tumor volume of mice in the PBS group increased sharply over the treatment period, reaching approximately 2000 mm³ by day 8. In contrast, treatment with free PurpN and Mn²⁺ resulted in moderate therapeutic outcomes, with tumor growth inhibition rates of 34.2 % and 57.4 %, respectively, highlighting their positive antitumor efficacy. Encouragingly, PurpN/Mn@PEG exhibited the most potent antitumor efficiency with tumor growth inhibition rates of 87.8 %, suppressing tumor volumes to approximately 85.1 mm³, and achieving the lowest tumor weight of 0.23 g on day 8 (Fig. 6b–f). This superior therapeutic efficacy stems from the multi-pathway immune synergy within the PurpN/Mn@PEG NPs group. Moreover, no noticeable weight loss was observed in the treated mice throughout the entire treatment period (Figure S12), indicating negligible side effects induced by those treatments. Subsequently, the tumor-bearing mice were sacrificed after various treatments, and the major organs were collected and subjected to hematoxylin and eosin (H&E) staining. The results showed no adverse damage to the major organs (heart, spleen, liver, lung, and kidney), further confirming the negligible systemic toxicity associated with all treatments (Figure S13).

To further validate the therapeutic efficacy, hematoxylin and eosin (H&E) staining, Ki-67 staining, and terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining of tumors after various treatments were conducted (Fig. 6g). Compared with tumors in the control groups, tumors treated with PurpN/Mn@PEG displayed the most severe cell necrosis and considerable nuclear condensation. Additionally, the expression of Ki-67, a cell proliferation marker, was significantly decreased with PurpN/Mn@PEG, suggesting its inhibitory effects on tumor proliferation (Fig. 6i). Furthermore, the fluorescence signal of the apoptotic marker TUNEL was markedly increased, indicating the highest tumor apoptosis rate induced by PurpN/Mn@PEG. These results further demonstrated the maximum antitumor efficacy of PurpN/Mn@PEG in vivo through synergistic immunotherapy.

2.7. In vivo evaluation of antitumor immune responses

Based on the immune-activating efficacy observed in vitro and the satisfactory antitumor effects of PurpN/Mn@PEG, we further explored its immune activation in vivo. The spleen is the largest secondary lymphoid organ in the body and serves as an important site for antigen recognition and the initiation of immune responses [47,48]. Therefore, spleens and tumor cells were harvested and processed into single-cell suspensions to evaluate tumor-specific immune activation using flow cytometry analysis (Fig. 7d and. e). As displayed in Fig. 7d, mice treated with free PurpN or Mn²⁺ showed substantially improved DC maturation rates in the spleen, with mature DC ratios of 30.7 % and 39.6 %, respectively, compared to 22.9 % in the PBS group. Notably, mice treated with PurpN/Mn@PEG exhibited the most robust DC maturation (69.9 %), confirming that free PurpN induces ICD and activates the TNF-α pathway, whereas Mn²⁺ activates the STING pathway and triggers ICD, thereby synergistically promoting DC maturation. We further investigated the intertumoral infiltration of T cells, focusing on CD3⁺CD4⁺ (helper T cells) and CD3⁺CD8⁺ (effector T cells), which are two critical subsets involved in antitumor immunity [[49], [50], [51], [52]]. Flow cytometry analysis revealed that mice treated with free PurpN and Mn²⁺ showed a 1.5-fold and 1.6-fold increase in helper T cell infiltration, respectively, and a 1.1-fold and 1.3-fold increase in effector T cell infiltration, respectively, compared to PBS-treated mice. In contrast, 1.9-fold and 1.3-fold increases in helper and effector T cell infiltration were observed in PurpN/Mn@PEG-treated mice (Fig. 7f–i). Consistently, immunofluorescence staining of tumor sections showed a significant increase in the abundance of both helper and effector T cells, aligning with the flow cytometry results (Fig. 7j). These findings suggest that the synergistic immune activation induced by PurpN/Mn@PEG NPs could trigger a more robust adaptive immune response. Besides immune cells, we also examined the levels of various cytokines in the serum. Compared with the PBS group, mice treated with free PurpN and Mn²⁺ exhibited a significant upregulation in the levels of TNF-α and IL-6 (Fig. 7b and. c), whereas those in the PurpN/Mn@PEG group demonstrated the highest secretion of these cytokines. Collectively, these findings mechanistically support that PurpN/Mn@PEG NPs modulate the STING/TNF-α axis to enhance antitumor immunity through three synergistic mechanisms: (1) PurpN- and Mn²⁺-induced ICD pathway, (2) Mn²⁺-mediated activation of the cGAS-STING pathway, and (3) PurpN-triggered activation of the TNF-α signaling pathway, so the nanostimulator is expected to synergize with immune-checkpoint inhibitors. Additionally, PurpN/Mn@PEG NPs do not require local radio or ultrasound exposure; they can therefore be administered to patients with deep-seated or disseminated tumors that are inaccessible to external energy sources, offering a readily translatable route for clinical use.

Fig. 7.

Immune activation study of PurpN/Mn@PEG NP-enhanced anticancer therapy. (a) Schematic illustration of immune activation by PurpN/Mn@PEG NPs in B16-F10 tumor-bearing mice. (b) Serum TNF-α and (c) IL-6 levels in B16-F10 tumor-bearing mice after the indicated treatments (n = 4). (d) Abundance and (e) relative proportion of mature DCs in the spleen after the various treatments (n = 4). (f) Abundance and (g) relative proportion of CD3⁺CD4⁺ T cells, and (h) abundance and (i) relative proportion of CD3⁺CD8⁺ T cells in tumors after the various treatments (n = 4). (j) Representative immunofluorescence images showing CD8 (green) and CD4 (red) staining in tumors after the indicated treatments. Data are presented as mean ± SD (∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001). (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Certainly, although the self-assembly protocol is operationally straightforward, a post-synthesis purification step is indispensable to eliminate unreacted precursors and process-related impurities; scale-up to GMP-compliant 10-L batches with reproducible critical quality attributes has yet to be validated. In addition, the nanomaterial's hitherto uncharacterized intrinsic toxicity and its nano–bio interfacial interactions constitute a major translational hurdle. These limitations will be systematically de-risked through forthcoming GLP-compliant toxicology studies and Quality-by-Design-driven process-development programs.

In summary, this study reports on a degradable purpurin-based metalloimmunostimulant (PurpN/Mn@PEG) that amplifies ROS and modulates STING/TNF-α axis to boost melanoma immunotherapy. After entering tumor, PurpN/Mn@PEG degrades and specifically releases the Mn²⁺ and purpurin under acidic pH and high GSH levels in TME. Mechanistically, PurpN in PurpN/Mn@PEG amplifies ROS through phenolic hydroxyl oxidation to quinone moieties, generating H₂O₂ for ROS cascades, while both PurpN and Mn²⁺ collectively fuel a ROS storm that enhances antigen presentation and drives DC maturation. Simultaneously, PurpN up-regulates CEBPB to elevate TNF-α, while Mn²⁺ activates the cGAS-STING pathway together, synergistically augment immunotherapy. In vivo experiments demonstrate that the remarkable therapeutic efficacy of the purpurin-based metalloimmunostimulant against melanoma arises from its reversal of the immunosuppressive environment through ROS amplification and modulation of the STING/TNF-α axis. Thus, our degradable metalloimmunostimulant reprogramming the immunosuppressive TME represents a promising synergistic approach for enhancing melanoma immunotherapy.

3. Materials and methods

3.1. Chemicals and materials

Amine ester-terminated PEG (NH₂-mPEG, average molecular weight, 2 kDa) was commercially available and ordered from Hunan Hua Teng Pharmaceutical Co., Ltd. 1,2,4-trihydroxyanthraquinone (PurpN), manganese (II) acetate tetrahydrate (Mn (CH₃COO)₂ ·4H₂O), GSH, methylene blue (MB), catalase assay kit, hydrogen peroxide assay kit, methanol, N, N-dimethylformamide (DMF), ethanol, hydrogen peroxide (H₂O₂, 30 %), were obtained from Shanghai Titan Scientific Co. Fetal bovine serum (FBS), Roswell Park Memorial Institute 1640 medium, penicillin−streptomycin, and were all purchased from Shanghai ExCell Biotech Co., Ltd. trypsin-EDTA, Hoechst 33342, 2′,7′-Dichlorofluorescin diacetate (DCFH-DA) kit, JC-1 staining kit, Calcein-AM and propidium iodide (PI) were provided from Beyotime Biotech Co., Ltd. The Cell Counting Kit-8 (CCK-8) assay kit was supplied from Kumamoto Dojindo Laboratories. Recombinant anti-HMGB1 polyclonal antibody, recombinant anti-calreticulin polyclonal antibody, BCL2 monoclonal antibody, Beta-actin mouse monoclonal antibody, GAPDH mouse monoclonal antibody, Vinculin mouse monoclonal antibody, Alexa Fluor 488-conjugated goat anti-rabbit secondary antibody were all ordered from Proteintech Group, Inc. InVivoMab anti-mouse PD-1 (CD279) was obtained from BioXcell. APC-Cy7-labelled CD45 monoclonal antibody (APC-Cy7-CD45 mAb, catalog No. 557659), BB700-labelled CD4 monoclonal antibody (BB700-CD4 mAb, catalog No. 566407), PE-Cy7-labelled CD8a monoclonal antibody (PE-Cy7-CD8a mAb, catalog No. 552877), PE-Cy7-labelled CD86 monoclonal antibody (PE-Cy7-CD86 mAb, catalog No. 560582) were all purchased from Becton, Dickinson and Company. Zombie Aqua Fixable Viability Kit, APC-labelled CD11c monoclonal antibody (APC-CD11c mAb, catalog No. 117310), APC-labelled CD3 monoclonal antibody (APC-CD3 mAb, catalog No. 100236), BV650-labelled CD80 monoclonal antibody (BV650-CD80 mAb, catalog No. 104731) were ordered from BioLegend, Inc.

4. Characterization

The morphology of samples was also studied by using a scanning electron microscope (SEM, Zeiss Sigma 300). A transmission electron microscope (TEM, JEM-2100) was used to image samples at an acceleration voltage of 200 kV. The zeta protential of samples were measured on a Malvern Zetasizer Nano ZSP (Malvern Instruments, UK). The UV–visible spectra were obtained on a Shimadzu UV-3600 spectrophotometer. Fourier transform infrared (FT-IR) spectra were obtained on a Perkin Elmer Spectrum 100 spectrometer (Perkin Elmer) using the KBr pellet method. X-ray photoelectron spectroscopy (XPS) was collected using a multifunctional photoelectron spectrometer Axis Ultra DLD (Kratos, UK). MD calculations were performed using Gromacs 2022.2 software. Electron spin resonance (ESR) spectra were acquired using a Bruker EMXplus ESR spectrometer. The specific atomic ratio of the three metal compositions was determined by inductively coupled plasma-mass spectrometry (ICP-MS) (Agilent 8900, USA).

4.1. Synthesis of PurpN/Mn@PEG NPs

PurpN (51.2 mg, 0.2 mmol) and MnCH₃COO)₂ • 4H₂O (24.5 mg, 0.1 mmol) were dissolved in a 40 mL methanol in a 50 mL vial. This mixture was heated at 110 °C for 20 h. After cooling to room temperature, the powder was collected by centrifugation and washed three times with DMF and ethanol to remove unreacted materials. 5 mg of PurpN/Mn solution was mixed with 50 mg of mPEG-NH₂ for another 24 h at room temperature, followed by centrifugation at 4500 rpm and washing three times with ultrapure water to obtain PurpN/Mn@PEG NPs, which were stored at 4 °C.

4.2. Molecular dynamics (MD) simulation calculations of PurpN/Mn@PEG NPs

First, a 6 nm × 6 nm × 6 nm simulation box was constructed, and 30 purpurin molecules and 10 Mn ions were randomly filled into it to form the simulation system. The MD simulation was carried out using the Gromacs 2022.2 program [53] under constant temperature and pressure conditions with periodic boundary conditions. The GAFF all-atom force field and the TIP3P water model were applied [54]. During the MD simulation, all hydrogen bonds involved were constrained using the LINCS algorithm [55], and the integration step size was 2 fs. Electrostatic interactions were calculated using the ParticleA–Mesh Ewald (PME) methods [56]. The cutoff value for non-bonded interactions was set at 10 Å, and the configuration was updated every 10 steps. The temperature was controlled using the V-rescale method at 298.15 K, and the pressure was controlled using the Parrinello-Rahman method [57,58]. Firstly, the energy of the four systems was minimized using the steepest descent method to eliminate overly close contacts between atoms; then, a 100 ps NVT equilibrium simulation was conducted at 298.15 K; finally, 100 ns MD simulations were performed on two different systems, with a configuration saved every 10 ps. The simulation results were visualized using the GROMACS built-in program and VMD.

4.3. The calculation for encapsulation efficiency and drug loading content of PurpN

To evaluate the encapsulation efficiency and drug loading content of PurpN, we measured the concentrations of PurpN in the digested samples solution. Samples of PurpN/Mn NPs were digested and dissolved by sonication in concentrated nitric acid. The concentration of PurpN was determined by a UV–vis spectrophotometer (at 484 nm). The concentration of Mn ions was determined by ICP-MS. The drug loading content was calculated by using the following equation: Drug loading content = mass of the PurpN in the PurpN/Mn NPs/mass of PurpN/Mn NPs. Therefore, the drug loading content of PurpN in PurpN/Mn NPs was reached to 43.8 %. The encapsulation efficiency was calculated by using the following equation: Encapsulation efficiency = the drug loading content of PurpN in PurpN/initial mass of the PurpN × 100 %. The encapsulation efficiency of PurpN in PurpN/Mn@PEG is as high as 88.6 %.

4.4. The test of PurpN release behavior from PurpN/Mn@PEG NPs

5 mg of PurpN/Mn@PEG NPs were dispersed in 10 mL of PBS buffer containing 20 % methanol (v/v) with different pH values (pH = 7.4, pH = 5.2, and pH = 5.2 with 10 mM GSH) and stirred for different times (1, 6, 12, 24, 36 and 48 h) at 37 °C, 200 rpm. Next, the supernates were collected by centrifugation and the concentrations of PurpN were determined by UV–vis s spectrophotometry, and the medium was replaced with an equal volume of fresh medium.

4.5. Measurement of Mn²⁺ release

PurpN/Mn@PEG NPs were incubated in 10 mL of PBS buffer containing 20 % methanol (v/v) solution with different pH values (pH = 7.4, pH = 5.2, and pH = 5.2 with 10 mM GSH) at the different designated times (1, 6, 12, 24, 36 and 48 h), the released Mn²⁺ was collected after the centrifugation to evaluate by using ICP-MS.

4.6. Degradation of PurpN/Mn NPs by GSH

PurpN/Mn NPs were dispersed in 20 % methanol (v/v) containing different concentrations of GSH (0, 2, 5, 10, 20 mM), and incubated for 30 min at room temperature. Subsequently, the supernatants were obtained and measured by a UV–Vis spectrophotometer at 484 nm.

4.7. GSH oxidation by PurpN/Mn NPs

PurpN/Mn NPs dispersion was mixed with GSH (10 mM), and then shaken at room temperature. At different time points (10 min, 20 min, 30 min, 60 min, and 120 min), the supernatants were collected and measured using GSH and GSSG Assay Kit (Beyotime biotechnology).

4.8. H₂O₂ generation assay

PurpN and PurpN/Mn NPs were dispersed in PBS, respectively. The dispersions were shaken at room temperature for 24 h. After that, the supernatants were collected, and the H₂O₂ concentrations were measured using Hydrogen Peroxide Assay Kit.

4.9. OH generation assay

Methylene blue (MB) was used to detect •OH. Firstly, PurpN/Mn NPs (100 μg mL⁻¹) were incubated with solutions for 30 min. Next, MB aqueous solution, H₂O₂ (100 μM), and purpN/Mn NPs aqueous solutions were fully mixed, and then, the supernatant was collected by centrifugation at designated time intervals, and the absorption wavelength around 660 nm was monitored by UV–vis spectroscopy. As controls, the UV–vis spectra of pure MB, MB + H₂O₂, MB + H₂O₂ + GSH, purpN/Mn NPs + MB + H₂O, and purpN/Mn NPs + MB + H₂O + GSH were also monitored for comparison. Moreover, ESR spectra of different groups were detected with the assistance of 5,5-dimethyl-1-pyrroline N-oxide (DMPO) as an ·OH trapping agent.

4.10. Cell culture and cytotoxicity test

The B16-F10 mouse melanoma cancer cell line was obtained from the China Center for Type Culture Collection. B16-F10 cells were cultured in 1640 mediums supplemented with 10 % fetal bovine serum and 1 % penicillin-streptomycin at 37 °C in a 5 % CO₂ humidified incubator.

The cytotoxicity was evaluated against B16-F10 melanoma cells using CCK-8 assay kit. B16-F10 cells were seeded in a 96-well plate at a density of 1 × 10⁴ cells per well, and cultured overnight for cell adhesion. The culture medium was removed and replaced with fresh culture medium containing various concentrations of free PurpN, Mn²⁺, and PurpN/Mn@PEG (0, 6.25, 12.5, 25, 50, and 100 μg mL⁻¹). After incubation for 24 h, the cell viability was detected using CCK-8 assay kit. In addition, the treated cells were also evaluated by fluorescence microscopy after live and dead cells were co-stained with calcein-AM (green) and PI (red).

Additionally, the cytotoxicity was carried out against B16-F10 melanoma cells by colony formation assay. In brief, cells were seeded in 6-well plate with density of 5 × 10² cells/well, and further incubated 2 days for cell adhesion. Cells then were washed with PBS and treated with (1) PBS, (2) free PurpN, (3) Mn²⁺ and (4) PurpN/Mn@PEG for 24 h. Subsequently, the cells were replaced with fresh medium and further cultured for 8 days. Afterwards, the treated cells were washed with PBS, fixed with methyl alcohol, stained using crystal violet solution (0.05 %) and imaged by microscope.

4.11. Analysis of cellular uptake behaviors

To investigate the cellular uptake behaviors, the ^FITCPurpN/Mn@PEG was made through labeling by FITC. In brief, the B16-F10 cells were seeded in confocal dishes at a cell density of 3 × 10⁵ cells per well, and cultured overnight for adhesion. The medium was then replaced with fresh medium containing 5 μg mL^{−1 FITC}PurpN/Mn@PEG for various time-intervals (0, 2, 6, 12, and 24 h). Then, the cells were collected, and detected using flow cytometry. Finally, cells were co-stained with Hoechst 33342 (nucleus, blue), and imaged by fluorescence microscope.

4.12. Cellular ROS assessment in vitro

The intracellular ROS was detected by DCFH-DA probes. B16-F10 cells were initially seeded in a 6-well plate at a density of 3 × 10⁵ cells per well and cultured overnight to facilitate cell attachment. Subsequently, the medium was removed, and the cells were incubated with various concentrations of (1) PBS, (2) free PurpN, (3) Mn²⁺, and (4) PurpN/Mn@PEG for an additional 6 h. After incubation, the cells were washed with PBS to eliminate residual drugs, and then treated with fresh medium containing 10 μM DCFH-DA for 30 min. For flow cytometric analysis, the cells were dissociated from the plate, resuspended in PBS, and further tested using flow cytometry. In addition, following the flow cytometric analysis, the cells were stained with Hoechst 33342 for 5 min, and further observed under a fluorescence microscope.

4.13. Mitochondrial membrane potential assay

The cellular mitochondrial membrane potential was detected using a JC-1 assay kit. B16-F10 cells were seeded in a 6-well plate at a cell density of 3 × 10⁵ cells per well and incubated overnight for cell adhesion. Then, the cells were further co-cultured with medium containing (1) PBS, (2) free PurpN, (3) Mn²⁺, and (4) PurpN/Mn@PEG for 12 h. Afterwards, the cells were washed three times, stained with the JC-1 assay kit according to the manufacturer's instructions, incubated for 30 min, and visualized by CLSM.

4.14. Immunogenic cell death analysis in vitro

Cells were seeded in confocal dishes at a density with 3 × 10⁵ cells per dish, and cultured 12 h for cell adhesion. Then, cells were washed with PBS and treated with (1) PBS, (2) free PurpN, (3) Mn²⁺ and (4) PurpN/Mn@PEG for 12 h. Afterwards, the culture supernatant was collected and assessed using ATP detection kits.

Cells were fixed, permeabilized, blocked, and then treated with anti-HMGB1 antobodies, stained with Alexa Fluor 488-conjugated goat anti-rabbit secondary antibody. The cells were further stained with Hoechst 33342 (nuclei, blue), observed by CLSM. Additionally, the cell proteins were lysed using RIPA lysis buffer supplemented with PMSF and Phosphatase inhibitor cocktail A, and separated with SDS-polyacrylamide gel electrophoresis, and then transferred onto a 0.22 μm PVDF membrane. The membrane was blocked with a TBST solution containing 5 % non-fat powdered milk and incubated with primary antibodies at 4 °C overnight, including β-actin and CRT antibodies. The membrane was washed three times, and incubated with goat anti-rabbit IgG secondary antibody solution for another 1 h at room temperature. Finally, protein bands were visualized using ECL substrate western blotting by chemiluminescence detection.

4.15. cGAS-STING activation assay

RAW264.7 cells were seeded in 6-well plates at a cell density of 3 × 10⁵ cells/well, and incubated 12 h for cell adhesion. Then, cells were treated with (1) PBS, (2) free PurpN, (3) Mn²⁺ and (4) PurpN/Mn@PEG for 12 h. Subsequently, proteins were lysed using RIPA lysis buffer supplemented with PMSF and Phosphatase inhibitor cocktail A, and separated with SDS-polyacrylamide gel electrophoresis, and then transferred onto a 0.22 μm PVDF membrane. The membrane was blocked with a TBST solution containing 5 % non-fat powdered milk and incubated with primary antibodies at 4 °C overnight, including GAPDH, β-actin, cGAS, STING, p-STING, TBK, p-TBK rabbit antibodies. The membrane was washed three times, and incubated with goat anti-rabbit IgG secondary antibody solution for another 1 h at room temperature. Finally, protein bands were visualized using ECL substrate western blotting by chemiluminescence detection. While another portion protein was used for qRT-PCR (quantitative reverse transcription polymerase chain reaction) analysis of the expression levels of the Ifnb1, Cxcl10, Isg15, and Ifit1 genes. The primer sequences used are as follows.

Ifnb1	Forward (F): GTCCTCAACTGCTCTCCACTR Reverse (R): CCTGCAACCACCACTCATTC
Cxcl10	Forward (F): ATCATCCCTGCGAGCCTATCCT Reverse (R): GACCTTTTTTGGCTAAACGCTTTC
Isg15	Forward (F): GACGGTCTTACCCTTTCCAGT Reverse (R): CCTTTCGTTCCTCACCAGGAT
Ifit1	Forward (F): CTGAGATGTCACTTCACATGGAA Reverse (R): GTGCATCCCCAATGGGTTCT
Rps17	Forward(F): CGCCATTATCCCCAGCAAG Reverse(R): TGTCGGGATCCACCTCAATG

mRNA quantification was conducted using a Bio-Rad real-time quantitative PCR instrument; all gene expression data obtained from qRT-PCR were normalized with RPS17(Ribosomal Protein S17) as the internal control.

4.16. Gene transcriptome detection and evaluation of CEBPB-TNF-α axis expression

B16-F10 cells were seeded in 100 cm² culture dishes and treated with (1) PBS buffer, (2) free PurpN, (3) Mn²⁺ solution, and (4) PurpN/Mn@PEG NPs for 12 h. After washing the cells three times with PBS, they were harvested and sent to Tissuebank Genetic Technology Co., Ltd. (Shanghai, China) to measure the gene expression changes. Total RNA was then extracted and purified using the QIAGEN RNeasy Mini Kit according to the manufacturer's instructions to analyze the expression level of the CEBPB gene using real-time quantitative reverse transcription polymerase chain reaction (qRT-PCR). The following primer sequences were used for CEBPB expression analysis:CEBPB primer sequences:Forward (F): CAACCTGGAGACGCAGCACAAG and Reverse (R): GCTTGAACAAGTTCCGCAGGGT; GAPDH primer sequences (for normalization):

Forward (F): TGCACCACCAACTGCTTAGC and Reverse (R): GGCATGGACTGTGGTCATGAG.

Additionally, the collected supernatant was used to treat RAW264.7 cells for 12 h, and these RAW264.7 cells were subsequently collected. For these cells, the Mouse Tumor Necrosis Factor α (TNF-α) ELISA KIT was used according to the manufacturer's instructions to measure the TNF-α expression level. Furthermore, qRT-PCR was employed to analyze the expression level of the TNF-α gene, with the following primer sequences: TNF-α primer sequences: Forward (F): TGTAGCCCACGTCGTAGCAAA and Reverse (R): CTGGCACCACTAGTTGGTTGT. Finally, mRNA quantitative analysis was conducted using a Bio-Rad real-time quantitative PCR instrument; all gene expression data obtained from qRT-PCR were normalized using GAPDH and Rps17 as an internal control.

4.17. BMDCs maturation assay

The BMDCs were extracted from the femur and tibia of mice, and cultured with RPMI 1640 medium containing GM-CSF (20 ng mL⁻¹). The culture medium was replaced every two days and cells were harvested on the 8th day for the subsequent experiments. To evaluate the maturation of BMDCs, BMDCs were co-cultured with medium from B16-F10 cells treated with (1) PBS, (2) free PurpN, (3) Mn²⁺, and (4) PurpN/Mn@PEG in a 6-well plate for 24 h. Subsequently, mature DCs (CD11c⁺, CD80⁺, CD86⁺) were evaluated using flow cytometry.

4.18. Animal model

Female C57BL/6 mice (6–7 weeks, 18–20 g) were purchased from Hunan SJA Laboratory Animal Co., Ltd. All animal experimental protocols and care in this work were approved by the Institutional Animal Care and Use Committee of the Central South University (approval No. XMXH-2025-0850). To establish the B16-F10 melanoma mouse model, 25 μL of cell suspension containing 2.5 × 10⁵ B16-F10 cells dispersed in PBS were injected into the right flank of each mouse. Then, when the tumor reached ∼50 mm³, the B16-F10 tumor models were established and used for in vivo experiments. The tumor size, weight, and survival of the mice were monitored every other day.

4.19. In vivo anticancer efficiency assessment

B16-F10 tumor bearing-mice were randomly divided into four groups (n = 8), which were intravenously administered on days 0, 2, 4, and 6 with (1) PBS (200 μL), (2) free PurpN (27 mg kg⁻¹, 200 μL), (3) Mn²⁺ (30 mg kg⁻¹, 200 μL) and (4) PurpN/Mn@PEG (equivalent to 27 mg kg⁻¹ free PurpN and 30 mg kg⁻¹ Mn²⁺, 200 μL). The tumor volumes and body weight were measured every other day over the 18-day treatment period. On day 18 post-administration, the mice were sacrificed, and major organs (heart, liver, spleen, lung, and kidney), tumors, and serum were harvested. These tissues were then subjected to hematoxylin and eosin (H&E) staining, TUNEL immunofluorescence staining, and immunohistochemistry staining for Ki67 proteins.

4.20. In vivo antitumor immune response assay

To investigate the immune response induced by different treatments, the collected tumors and spleens were dissociated into single-cell suspensions and filtered through 100 μm diameter cell strainers. The cells were stained with the Zombie Aqua Fixable Viability Kit for 20 min at 4 °C to differentiate live cells from dead cells. All single-cell suspensions were further co-stained with APC-Cy7-anti-CD45, BB700-anti-CD4, PE-Cy7-anti-CD8a, APC-anti-CD3, APC-anti-CD11c, BV650-anti-CD80, and PE-Cy7-anti-CD86 for flow cytometry analysis. Additionally, TNF-α and IL-6 levels in serum samples collected from the mice were measured using an ELISA kit (Servicebio) according to the manufacturer's instructions.

4.21. In vivo biosafety analysis

Healthy C57BL/6 mice were intravenously injected with PBS (control, 200 μL) or PurpN/Mn@PEG NPs (equivalent to 27 mg kg⁻¹ free PurpN and 30 mg kg⁻¹ Mn²⁺ per mouse in 200 μL PBS) (n = 3). Blood was collected at 24 h post-treatment, and levels of aspartate transferase (AST), alanine transaminase (ALT), albumin (ALB), alkaline phosphatase (ALP), urea nitrogen (BUN), and creatinine (CREA) were measured.

5. Statistical analysis

All experiments were repeated at least three times, and the data are presented as the mean ± SD. Statistical analysis was performed using one-way analysis of variance (ANOVA) and two-way ANOVA with Tukey's post hoc test using SPSS v.26.0 software. p < 0.05 was statistically significant.

CRediT authorship contribution statement

Xiaoxin Yang: Validation, Supervision, Methodology, Data curation. Mi Huang: Validation, Software, Methodology, Data curation. Tianyi Pang: Software, Resources, Methodology, Formal analysis, Data curation. Dong Zhong: Visualization, Data curation. Hong Liu: Supervision, Project administration, Funding acquisition. Xiang Chen: Supervision, Funding acquisition. Jun Liu: Supervision, Project administration, Funding acquisition. Yu Wen: Writing – review & editing, Project administration, Investigation, Conceptualization.

Declaration of competing interest

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

Appendix A. Supplementary data

The following is/are the supplementary data to this article.

Data availability

Data will be made available on request.