Axial Coordination and Defect Engineering of Copper‐Based Single‐Atom Nanozymes: Driving Synergistic Disulfidptosis and Pyroptosis for Efficient Prostate Cancer Immunotherapy

ABSTRACT

Immunotherapy for prostate cancer is severely restricted by the immunosuppressive tumor microenvironment (TME) and the intrinsic antioxidant defense systems of tumor cells. To address these challenges, we develop an efficient copper‐based single‐atom nanozyme platform (CuN₃Cl‐BAY@COD) via a facile axial coordination and defect engineering strategy to boost immunotherapy by driving synergistic disulfidptosis and pyroptosis. This dual‐engineering strategy effectively regulates the electronic and geometric structures of Cu sites, resulting in significantly improved multienzyme‐mimicking activities. Such enhanced multienzyme‐mimicking activities endow CuN₃Cl‐BAY@COD with the robust capacity to disrupt intracellular redox homeostasis and amplify the disulfide stress mediated by BAY‐876, inducing efficient tumor cell pyroptosis and disulfidptosis. The loaded cholesterol oxidase degrades cholesterol to inhibit tumor cell invasion while elevating intracellular hydrogen peroxide levels to exacerbate pyroptosis. In summary, by coupling metabolic reprogramming with dual immunogenic cell death, this engineered nanoplatform overcomes TME resistance barriers, providing a robust strategy for prostate cancer immunotherapy.

An efficient copper‐based single‐atom nanozyme platform (CuN₃Cl‐BAY@COD) was constructed to induce immunogenic disulfidptosis/pyroptosis and remodel the immunosuppressive tumor microenvironment through the disruption of redox homeostasis and the reprogramming of cholesterol and glucose metabolism, providing a viable therapeutic strategy for prostate cancer immunotherapy.

1. Introduction

Prostate cancer is the most common malignant tumor in the male genitourinary system, posing a significant threat to men's health worldwide [1, 2, 3]. However, the tumor microenvironment (TME) of prostate cancer, characterized by inadequate immune cell infiltration, weak immunogenicity, and a dominance of immunosuppressive components, severely limits the efficacy of immunotherapy [4, 5, 6]. Furthermore, abnormally elevated cholesterol levels in prostate cancer aggravate this immunosuppression [7, 8]. Thus, enhancing tumor immunogenicity and remodeling the immunosuppressive TME remain central challenges in current prostate cancer immunotherapy research. Disrupting intracellular redox homeostasis by generating excessive reactive oxygen species (ROS) can induce immunogenic pyroptosis, thereby boosting antitumor immunity [9, 10, 11, 12, 13]. However, tumor cells utilize high levels of glutathione (GSH) to maintain intracellular redox homeostasis, forming a ROS resistance barrier [14, 15, 16, 17, 18]. L‐cysteine (L‐Cys), a key precursor substrate for GSH synthesis generated through the nicotinamide adenine dinucleotide phosphate (NADPH)‐dependent reduction of intracellular cystine, plays a crucial role in the antioxidant defense systems of tumor cells [19, 20, 21, 22]. Therefore, blocking the NADPH/L‐Cys pathway could not only downregulate GSH levels to amplify intracellular redox dyshomeostasis, thereby driving a self‐amplifying cascade of pyroptosis [23, 24], but also lead to intracellular cystine accumulation to aggravate disulfidptosis induced by glucose metabolism inhibition [25, 26, 27]. Consequently, the regulation of the NADPH/L‐Cys pathway offers a novel target to efficiently induce pyroptosis and disulfidptosis for improving prostate cancer immunotherapy.

Copper‐based single‐atom nanozymes (Cu‐SANs) have demonstrated potential in disrupting intracellular redox homeostasis via targeting the NADPH/L‐Cys pathway owing to their multienzyme‐mimicking activities, including NADPH oxidase‐like, glutathione oxidase‐like (GSHOx‐like), and peroxidase‐like (POD‐like) activities [28]. However, the catalytic performance of conventional Cu‐SANs with Cu‐N₄ configurations (CuN₄) is limited by their planar symmetric electronic structures, which hinder the adsorption and activation of reaction intermediates [29]. Planar coordination regulation strategies have been shown to improve catalytic performance [30, 31]; for instance, defect engineering increases the exposure of active sites and promotes electron transfer between these sites and the support [32, 33, 34, 35]. In addition, axial coordination engineering effectively regulates the electronic and geometric structures of catalytic sites by introducing additional ligands perpendicular to the plane [36, 37, 38]. This approach mimics the coordination features of metal centers in natural enzymes, playing a crucial role in the specific affinity and cleavage of substrates [39, 40, 41, 42]. However, a single modulation strategy is often insufficient to significantly enhance catalytic performance [43, 44], and multi‐scale modulations are generally sophisticated. Therefore, developing a facile and efficient strategy to optimize enzyme‐like catalytic performance through the multi‐scale coupling of electronic structures, coordination environments, and spatial microenvironments remains a critical challenge.

In this work, we developed a facile molten chloride salt‐assisted pyrolysis strategy to construct novel Cu‐SANs (CuN₃Cl) with axial chlorination and abundant carbon defects. The synergistic effect of axial chlorination and defect engineering enhanced the enzyme‐mimicking activities of CuN₃Cl, including the catalase‐like (CAT‐like), POD‐like, GSHOx‐like, NADPH oxidase‐like, and L‐Cys oxidase‐like (LCO‐like) activities. Upon functionalization with the glucose transporter 1 (GLUT1) inhibitor BAY‐876 and cholesterol oxidase (COD), the resulting nanozyme platform (CuN₃Cl‐BAY@COD) efficiently disrupted intracellular redox homeostasis through a self‐amplifying cascade of ROS generation by its POD‐like and GSHOx‐like activities, triggering gasdermin D (GSDMD)‐mediated pyroptosis. Furthermore, its NADPH oxidase‐like and LCO‐like activities amplified the disulfide stress mediated by BAY‐876, inducing disulfidptosis. The loaded COD degraded cholesterol to inhibit tumor cell invasion while elevating intracellular hydrogen peroxide (H₂O₂) levels to further exacerbate tumor cell pyroptosis. Overall, CuN₃Cl‐BAY@COD efficiently induced immunogenic disulfidptosis/pyroptosis and remodeled the immunosuppressive TME through the disruption of redox homeostasis and the reprogramming of cholesterol and glucose metabolism, providing a viable therapeutic strategy for prostate cancer immunotherapy.

2. Results and Discussion

The detailed synthesis process of CuN₃Cl‐BAY@COD is illustrated in Scheme 1. First, a zeolitic imidazolate framework (ZIF‐8) was used as a precursor and mixed with sodium chloride (NaCl), followed by programmed high‐temperature pyrolysis to obtain a carbon support (NaCl‐NC, Figure S1). Cu ions were then adsorbed onto the surface of NaCl‐NC and subjected to a second high‐temperature pyrolysis to produce CuN₃Cl. For comparison analysis, we prepared CuN₄ using the same procedure without the addition of NaCl (Figure S2). Finally, BAY‐876 and COD were loaded onto CuN₃Cl to construct the nanozyme platform CuN₃Cl‐BAY@COD. Transmission electron microscopy (TEM) images showed that the CuN₃Cl nanoparticles were uniform with an average diameter of 100 nm (Figure 1a). Aberration‐corrected high‐angle annular dark‐field scanning transmission electron microscopy (HAADF‐STEM) images of CuN₃Cl revealed the atomic dispersion of Cu on the carbon support without any discernible Cu clusters or nanoparticles (Figure 1b,c). Elemental mapping images showed the uniform distribution of Cu, N, C, and Cl in CuN₃Cl (Figure 1d–h), while the corresponding maps for CuN₄ were shown in Figure S3. X‐ray diffraction patterns indicated that both CuN₃Cl and CuN₄ exhibited characteristics of amorphous carbon, and no metal or metal oxide diffraction peaks were observed (Figure S4). Inductively coupled plasma optical emission spectroscopy measurements showed that the Cu contents in CuN₄ and CuN₃Cl were 1.294% and 1.136%, respectively. The elemental composition and chemical states of CuN₃Cl and CuN₄ were further analyzed by x‐ray photoelectron spectroscopy. The high‐resolution N 1s spectrum confirmed the coexistence of graphitic N, pyrrolic N, oxidized N, and pyridinic N (Figure S5). A distinct peak corresponding to the Cu─Cl bond was observed in the Cl 2p spectrum of CuN₃Cl (Figure S6). Nitrogen adsorption‐desorption isotherm results indicated that the specific surface area of CuN₃Cl (1045.503 m² g⁻¹) was 65.6% higher than that of CuN₄ (631.218 m² g⁻¹), which could be attributed to the porosity‐generating capability of molten salts, facilitating efficient drug loading (Figure S7). Raman spectroscopy was employed to evaluate the degree of carbon defects in CuN₃Cl and CuN₄. As shown in Figure S8, CuN₃Cl exhibited a substantially higher I _D/I _G ratio (the intensity ratio of the defect‐associated D‐band to the sp²‐hybridized graphitic G‐band) than CuN₄, confirming the presence of abundant carbon defects in CuN₃Cl. The successful loading of BAY‐876 and COD was verified by zeta potential measurements (Figure S9). The drug‐loading capacities of BAY‐876 and COD were 0.18% and 15.04%, respectively. Furthermore, the release of BAY‐876 under different pH conditions was evaluated through high‐performance liquid chromatography, revealing that a weakly acidic environment promoted its release (Figure S10).

SCHEME 1.

Schematic illustration of the mechanism driving CuN₃Cl‐BAY@COD‐mediated antitumor immunotherapy.

FIGURE 1.

(a) Representative transmission electron microscopy (TEM) image of CuN₃Cl. (b, c) High‐angle annular dark‐field scanning transmission electron microscopy (HAADF‐STEM) images of CuN₃Cl, with red circles indicating single Cu atoms. (d–h) Elemental mapping images of CuN₃Cl (Cu, C, N, and Cl). (i) X‐ray absorption near‐edge structure (XANES) spectra at the Cu K‐edge of CuN₃Cl and reference samples. (j) Fourier‐transformed extended x‐ray absorption fine structure (FT‐EXAFS) spectra of CuN₃Cl and reference samples. (k, l) FT‐EXAFS fitting in R‐space and k‐space of CuN₃Cl. (m–p) Wavelet transform‐EXAFS plots of CuN₃Cl and reference samples.

To elucidate the chemical state and coordination environment of the Cu atoms in CuN₄ and CuN₃Cl, extended x‐ray absorption fine structure (EXAFS) and x‐ray absorption near‐edge structure (XANES) analyses were conducted. As shown in Figure 1i, the Cu K‐edge XANES spectra indicated that the absorption edges of CuN₃Cl and CuN₄ lay between those of Cu foil and CuO, suggesting that the valence states of Cu in both materials were between 0 and +2. Relative to CuN₄, CuN₃Cl exhibited a positive shift in the absorption edge energy, indicating a higher Cu valence state resulting from the formation of an axial Cu─Cl bond. The Fourier‐transformed EXAFS (FT‐EXAFS) spectra of CuN₄ and CuN₃Cl displayed a main peak at 1.59 Å and the absence of a Cu─Cu metallic bond peak at 2.24 Å, confirming the atomic dispersion of Cu species in both materials (Figure 1j). FT‐EXAFS curve fitting results for CuN₃Cl revealed the coexistence of Cu─N and Cu─Cl bonds, with coordination numbers of 3 and 1, respectively, confirming that the predominant coordination configuration of Cu atoms in CuN₃Cl was Cu‐N₃Cl (Figure 1k,l and Table S1). As shown in Figure S11 and Table S1, the Cu atoms in CuN₄ existed as Cu‐N₄ sites. Wavelet transform (WT) analysis was further performed at the Cu K‐edge for both materials (Figure 1m–p and S12). The WT contour plot of Cu foil showed a characteristic intensity maximum corresponding to the Cu‐Cu metallic bond at approximately 7.5 Å^{− 1} in k‐space and 2.2 Å in R‐space. In contrast, the intensity maxima for CuN₃Cl and CuN₄ were observed near 3.5 Å^{− 1} in k‐space and 1.5 Å in R‐space, attributable to Cu‐N/Cl and Cu‐N bonds, respectively. Compared with CuN_4, the main peak of CuN₃Cl was stronger and slightly shifted toward a larger R‐space distance, further confirming the presence of an additional Cu─Cl bond. Collectively, these results demonstrated that Cu species were atomically dispersed in both CuN₄ and CuN₃Cl, existing primarily as Cu‐N₄ and Cu‐N₃Cl sites, respectively.

We systematically evaluated the multienzyme‐mimicking activities of CuN₃Cl. Compared with CuN₄, CuN₃Cl exhibited an enhanced capacity to catalyze the decomposition of H₂O₂ to O₂ in an acidic environment (pH 6.5, 1 mM H₂O₂) (Figure S13). This excellent CAT‐like activity could effectively alleviate tumor hypoxia and provide an oxygen source for subsequent oxidative reactions. Next, we evaluated the POD‐like activity using H₂O₂ as the substrate and 3,3′,5,5′‐tetramethylbenzidine as the chromogenic reagent [45, 46, 47]. Under identical conditions, the ultraviolet‐visible (UV–vis) absorption spectrum of CuN₃Cl showed a stronger characteristic peak at 652 nm than that of CuN₄ (Figure 2a). Steady‐state kinetic parameters derived from Michaelis‐Menten plots indicated that CuN₃Cl had a lower Michaelis constant (K _m) and a higher maximum reaction rate (V _max) (Figure 2b,c). Furthermore, its catalytic efficiency (k _cat/K _m) was 3.029 times greater than that of CuN₄, outperforming most previously reported Cu‐based single‐atom nanozymes (Table S2). These findings confirmed the excellent POD‐like activity of CuN₃Cl, which catalyzed the conversion of intracellular H₂O₂ to cytotoxic ROS, thereby triggering GSDMD‐mediated pyroptosis. NADPH oxidase can catalyze the conversion of NADPH to NADP⁺ and H₂O₂. We further assessed the NADPH oxidase‐like activity by detecting the characteristic absorption of NADPH. After co‐incubation with CuN₃Cl, a more significant decrease at 340 nm and a more pronounced increase at 260 nm were observed, indicating more efficient NADPH consumption under acidic conditions compared with CuN₄ (Figure 2d). Kinetic analysis revealed that this reaction adhered to the Michaelis‐Menten model. Compared with CuN₄, CuN₃Cl exhibited an enhanced catalytic profile, characterized by a lower K _m, a higher V _max, and significantly enhanced catalytic rate constant (k _cat) and k _cat/K _m values (1.824 and 1.972 times higher, respectively) (Figure 2e,f). Upon the inhibition of glucose metabolism, CuN₃Cl could deplete the intracellular NADPH pool via its NADPH oxidase‐like activity, thereby intensifying disulfide stress and exacerbating disulfidptosis in prostate cancer cells.

FIGURE 2.

(a) Ultraviolet–visible (UV–vis) absorption spectra of 3,3′,5,5′‐tetramethylbenzidine oxidation catalyzed by CuN₃Cl and CuN₄. (b, c) Steady‐state kinetic curves and corresponding Lineweaver‐Burk plots for CuN₃Cl and CuN₄ using hydrogen peroxide (H₂O₂) as the substrate. (d) UV–vis absorption spectra of nicotinamide adenine dinucleotide phosphate (NADPH) oxidation catalyzed by CuN₃Cl and CuN₄. (e, f) Steady‐state kinetic curves and corresponding Lineweaver‐Burk plots for CuN₃Cl and CuN₄ with NADPH as the substrate. (g) UV–vis absorption spectra of glutathione (GSH) oxidation catalyzed by CuN₃Cl and CuN₄. (h, i) Steady‐state kinetic curves and corresponding Lineweaver‐Burk plots for CuN₃Cl and CuN₄ using GSH as the substrate. (j) UV–vis absorption spectra of L‐cysteine (L‐Cys) oxidation catalyzed by CuN₃Cl and CuN₄. (k, l) Steady‐state kinetic curves and corresponding Lineweaver‐Burk plots for CuN₃Cl and CuN₄ with L‐Cys as the substrate. (m) Comparison of kinetic parameters between CuN₄ and CuN₃Cl. (n, o) Electron spin resonance spectra showing the generation of hydroxyl radicals. All data were presented as mean ± standard deviation (SD).

Furthermore, the intrinsic antioxidant defense systems of tumor cells, characterized by high GSH levels, conferred protection against oxidative stress and diminished the therapeutic efficacy of ROS‐mediated pyroptosis. Notably, we discovered that CuN₃Cl exhibited excellent GSHOx‐like activity (Figure 2g). This activity was evaluated using GSH as the substrate and 5,5′‐dithiobis‐(2‐nitrobenzoic acid) (DTNB) as the chromogenic reagent. This catalytic process also followed Michaelis‐Menten kinetics. Compared with CuN₄, CuN₃Cl displayed a lower K _m alongside k _cat and k _cat/K _m values that were 1.422 and 2.32 times greater, respectively (Figure 2h,i). Interestingly, CuN₃Cl effectively oxidized L‐Cys to produce H₂O₂, thereby depleting GSH precursors and elevating intracellular ROS. We evaluated the LCO‐like activity using L‐Cys as the substrate and DTNB as the chromogenic reagent. As shown in Figure 2j, CuN₃Cl exhibited significant LCO‐like activity. Consistent with the other enzyme‐mimicking behaviors, the reaction kinetics aligned with the Michaelis‐Menten model, yielding k _cat and k _cat/K _m values for CuN₃Cl that were 2.066 and 2.426 times greater than those of CuN₄ (Figure 2k,l). Importantly, in addition to generating ROS, CuN₃Cl utilized its GSHOx‐like and LCO‐like activities to deplete intracellular reducing species (Figure 2m), thereby impairing the antioxidant defense systems of tumor cells.

We further detected H₂O₂ production during NADPH and L‐Cys oxidation, which revealed that CuN₃Cl generated more H₂O₂ than CuN₄ (Figure S14). Similarly, CuN₃Cl@COD promoted cholesterol oxidation in a concentration‐dependent manner, leading to increased H₂O₂ levels (Figure S15). The multienzyme‐mimicking activities of CuN₃Cl drove the generation of hydroxyl radicals (•OH) through cascading reactions, thereby inducing pyroptosis in tumor cells. This mechanism was verified using terephthalic acid as a specific probe. When H₂O₂, NADPH, and L‐Cys were used as substrates, the fluorescence intensity of CuN₃Cl was significantly stronger than that of CuN₄ (Figure S16). Electron spin resonance spectroscopy corroborated these results, with CuN₃Cl exhibiting stronger characteristic signal peaks under identical conditions (Figure 2n). Additionally, •OH generation was detected in three independent systems: CuN₃Cl + NADPH, CuN₃Cl + L‐Cys, and CuN₃Cl@COD + cholesterol (Figure 2o). These findings demonstrated that the excellent NADPH oxidase‐like and LCO‐like activities of CuN₃Cl directly disrupted the NADPH/L‐Cys antioxidant axis in tumor cells, effectively blocking the reduction of cystine. This severely exacerbated the disulfide stress induced by BAY‐876‐mediated glucose starvation, leading to cytoskeletal collapse and triggering disulfidptosis. Concurrently, the GSHOx‐like and POD‐like activities of CuN₃Cl profoundly depleted intracellular GSH while driving robust ROS generation. This cascade‐amplified oxidative stress ultimately induced GSDMD‐mediated pyroptosis in prostate cancer cells.

The catalytic mechanisms of CuN₃Cl and CuN₄ were further investigated through density functional theory (DFT) calculations. The density of states plots revealed that the d‐band center of CuN₃Cl (−3.066 eV) was closer to the Fermi level than that of CuN₄ (−3.241 eV), indicating a stronger interaction between the substrates and the CuN₃Cl surface, thereby facilitating the catalytic reaction (Figure S17). The charge density difference maps confirmed that the dual‐engineering strategy of axial chlorination and carbon support defects modulated the d‐band electronic structure of the Cu atom, enhancing the affinity of CuN₃Cl for reaction intermediates (Figure 3a–h). For the POD‐like activity (Figures 3i,j and S18), DFT results showed that CuN₃Cl exhibited a lower H₂O₂ adsorption free energy (0.20 eV) compared with CuN₄ (0.38 eV), reflecting a stronger binding affinity. The adsorbed H₂O₂ dissociated into two hydroxyl intermediates adsorbed on the Cu active site. Subsequently, the desorption of one adsorbed *OH to form a free •OH radical constituted the rate‐limiting step of the overall reaction. The energy barrier for CuN₃Cl (0.46 eV) was significantly lower than that of CuN₄ (0.73 eV), promoting •OH generation. The remaining *OH formed H₂O and desorbed, restoring the Cu active site to its initial state. For the NADPH oxidase‐like activity, the process was divided into two half‐reactions (Figures 3k–n and S19). Compared with CuN₄, CuN₃Cl exhibited lower dehydrogenation and desorption energy barriers in the NADPH dehydrogenation step. At an equilibrium potential of 0.69 V, the *OOH adsorption free energy of CuN₃Cl in the rate‐limiting oxygen reduction reaction was closer to the ideal value (3.52 eV), validating its optimal catalytic pathway. In summary, the dual‐engineering strategy of axial chlorination and carbon support defects broke the inherent planar symmetry of conventional SANs with Cu‐N₄ configurations, enhanced electron transfer, effectively optimized the reaction free energy, and lowered the energy barrier of the rate‐limiting step, thereby endowing CuN₃Cl with robust multienzyme‐mimicking activities.

FIGURE 3.

Optimized structures and charge density maps of (a, b) CuN₄ and (c, d) CuN₃Cl. Charge density difference maps for H₂O₂ adsorption on (e) CuN₄ and (f) CuN₃Cl. Charge density difference maps for NADPH adsorption on (g) CuN₄ and (h) CuN₃Cl, respectively. (i) Catalytic mechanism of the peroxidase‐like (POD‐like) reaction on CuN₃Cl. (j) Free energy profiles of the POD‐like reaction on CuN₄ and CuN₃Cl. (k, m) Catalytic mechanisms of the NADPH oxidase‐like reaction on CuN₃Cl. (l, n) Free energy profiles of the NADPH oxidase‐like reaction on CuN₄ and CuN₃Cl.

Encouraged by the excellent multienzyme‐mimicking activities of CuN₃Cl, we further investigated its therapeutic effect on prostate cancer. The biocompatibility of CuN₃Cl and CuN₄ was evaluated using a Cell Counting Kit‐8 assay. As shown in Figure 4a, cell viability in the CuN₃Cl group was significantly higher at 120 µg mL⁻¹, indicating the favorable biocompatibility of CuN₃Cl. Under simulated TME conditions, cell viability decreased markedly in the CuN₃Cl group at the same concentration, confirming its efficient tumor cell‐killing capacity (Figure 4b). We further assessed the effect of CuN₃Cl loaded with COD and/or BAY‐876 on RM‐1 cell viability. The highest antitumor efficacy was observed in the CuN₃Cl‐BAY@COD group, which was attributable to the synergistic enhancement of ROS generation among CuN₃Cl, BAY‐876, and COD, leading to suppressed tumor cell proliferation (Figure 4c). Intracellular ROS production in RM‐1 cells was then validated using the 2′,7′‐dichlorofluorescein diacetate (DCFH‐DA) probe (Figure 4d). The CuN₃Cl group exhibited stronger green fluorescence under simulated TME conditions than CuN₄, which was attributed to its enhanced enzyme‐mimicking activities. The fluorescence significantly increased in the CuN₃Cl@COD group, resulting from increased H₂O₂ generation via cholesterol oxidation, which enhanced the POD‐like activity. The CuN₃Cl‐BAY group showed elevated fluorescence, as reduced glucose uptake decreased NADPH and L‐Cys levels, thereby impairing GSH synthesis and weakening the antioxidant defense systems. The strongest fluorescence was observed in the CuN₃Cl‐BAY@COD group. The generation of H₂O₂ in RM‐1 cells was detected by an H₂O₂‐fluorescent probe. As shown in Figure S20, no green fluorescence was observed in the control group. In contrast, weak green fluorescence was detected in the CuN₃Cl group, indicating that CuN₃Cl oxidized NADPH and L‐Cys to generate H₂O₂. This signal was significantly enhanced in the CuN₃Cl@COD group, demonstrating that the loaded COD oxidized intracellular cholesterol, further increasing intracellular H₂O₂ levels. This was further confirmed through cholesterol content assays (Figure S21). Subsequently, the intracellular GSH levels of RM‐1 cells were measured. As shown in Figure 4e, compared with the CuN₄ group, the GSH levels in the CuN₃Cl group significantly decreased, which was attributed to the excellent GSHOx‐like activity of CuN₃Cl. The decline in the CuN₃Cl‐BAY group was more prominent, indicating that BAY‐876 could reduce the synthesis of GSH in tumor cells. The lowest GSH level was observed in the CuN₃Cl‐BAY@COD group. The inhibitory effects of BAY‐876 and COD on RM‐1 cell invasion were evaluated through Transwell invasion assays in normal culture medium (Figure S22). Compared with the control and CuN₃Cl groups, CuN₃Cl‐BAY and CuN₃Cl@COD significantly inhibited RM‐1 cell invasion. The most potent inhibition of cell invasion was observed in the CuN₃Cl‐BAY@COD group. Furthermore, hemolysis assays confirmed the high hemocompatibility of CuN₃Cl‐BAY@COD (Figure S23). These findings demonstrated that CuN₃Cl‐BAY@COD disrupted intracellular redox homeostasis by efficiently generating ROS and depleting reducing species, highlighting its potential for prostate cancer therapy.

FIGURE 4.

(a) Viability of RM‐1 cells following treatment with varying concentrations of CuN₃Cl and CuN₄ (n = 4). (b) Cell viability under simulated tumor microenvironment (TME) conditions after treatment with varying concentrations of CuN₃Cl and CuN₄ (n = 4). (c) Viability of RM‐1 cells across various treatment groups under simulated TME conditions (n = 4). (d) Representative fluorescence images of intracellular reactive oxygen species (ROS) generation obtained using the 2′,7′‐dichlorofluorescein diacetate (DCFH‐DA) probe (scale bar: 50 µm). Quantification of intracellular: (e) glutathione (GSH) level, (f) adenosine triphosphate (ATP) content, (g) NADPH level, and (h) cystine content across various treatment groups (n = 3). (i) Confocal fluorescence images of F‐actin staining in different groups (scale bar: 50 µm). (j) Western blot analysis of pyroptosis‐associated proteins. Release of (k) lactate dehydrogenase (LDH) and (l) interleukin‐1β (IL‐1β) in different treatment groups (n = 3). (m) High mobility group box 1 (HMGB1) release across various treatment groups (n = 3). (n) Representative immunofluorescence images of calreticulin (CRT) exposure on the surface of RM‐1 cells in various treatment groups (scale bar: 100 µm). (o, p) Flow cytometric analysis of dendritic cell (DC) maturation, as indicated by CD80 and CD86 expression (n = 3). (q, r) Flow cytometric analysis of macrophage repolarization (n = 3). All data were presented as mean ± SD. Statistical significance was determined using a two‐tailed Student's t‐test, with P‐values indicated in the respective figures.

We investigated the effects and the underlying mechanisms of CuN₃Cl‐BAY@COD on prostate cancer cells. First, we analyzed the effect of the loaded BAY‐876 on glucose uptake in RM‐1 cells by measuring the extracellular glucose content. As shown in Figure S24, no significant difference was observed between the control and CuN₃Cl groups, whereas a significant increase in extracellular glucose was detected in the BAY‐876 and CuN₃Cl‐BAY groups, indicating the successful inhibition of glucose uptake by BAY‐876. Consistently, a decrease in intracellular adenosine triphosphate content further confirmed this effect (Figure 4f). We further examined the impact of CuN₃Cl‐BAY on intracellular NADPH levels (Figure 4g). Compared with the CuN₄ group, the NADPH level decreased in the CuN₃Cl group, which was further exacerbated in the CuN₃Cl‐BAY group, indicating that the combination of CuN₃Cl and BAY‐876 not only reduced NADPH production but also increased its consumption. This caused a significant reduction in the reducing capacity of tumor cells. This finding was further validated by intracellular cystine measurements (Figure 4h). Compared with the CuN₄ group, cystine levels increased in the CuN₃Cl group, with an even greater accumulation in the CuN₃Cl‐BAY group. This accumulation was driven by the depletion of NADPH, which blocked cystine reduction, while the remaining L‐Cys was oxidized by CuN₃Cl, leading to excessive cystine accumulation. The staining of cytoskeletal proteins revealed cytoskeletal collapse in treated cells, confirming that CuN₃Cl‐BAY induced disulfidptosis in prostate cancer cells (Figure 4i). Notably, treated cells exhibited characteristic plasma membrane ballooning, suggesting that pyroptosis of tumor cells was triggered (Figure S25). We validated the expression of pyroptosis‐related proteins through Western blot analysis (Figure 4j). Compared with the control group, an upregulation of cleaved caspase‐1 and GSDMD‐N proteins was observed in the CuN₃Cl group, indicating the occurrence of pyroptosis. The expression levels further increased upon the addition of BAY‐876, indicating that the disulfidptosis induced by CuN₃Cl‐BAY enhanced pyroptosis. The highest expression levels were observed in the CuN₃Cl‐BAY@COD group. Additionally, lactate dehydrogenase (LDH) release and interleukin‐1β (IL‐1β) secretion corroborated these findings (Figure 4k,l). Collectively, these results demonstrated that CuN₃Cl‐BAY@COD could efficiently dismantle the antioxidant defense systems to induce efficient disulfidptosis and pyroptosis in prostate cancer cells.

We further evaluated the ability of CuN₃Cl‐BAY@COD to induce immunogenic cell death (ICD) in tumor cells in vitro. First, we detected the release and surface exposure of damage‐associated molecular patterns (DAMPs), such as high mobility group box 1 (HMGB1) and calreticulin (CRT) (Figure 4m,n). Relative to untreated cells, CRT exposure on the surface of RM‐1 cells and HMGB1 secretion significantly increased in all treatment groups, confirming that CuN₃Cl‐BAY@COD effectively induced ICD and enhanced tumor immunogenicity. Next, we evaluated the effect of CuN₃Cl‐BAY@COD on dendritic cell (DC) maturation in vitro (Figure S26). DC maturation was progressively enhanced across treatment groups, reaching the highest level in the CuN₃Cl‐BAY@COD group (Figure 4o,p). Concurrently, pro‐inflammatory cytokines were detected using an enzyme‐linked immunosorbent assay. Secretion levels of tumor necrosis factor‐α (TNF‐α), interleukin‐6 (IL‐6), and interleukin‐12p70 (IL‐12p70) gradually increased across the treatment groups, with the highest concentrations detected in the CuN₃Cl‐BAY@COD group (Figure S27). We further investigated the effect of CuN₃Cl‐BAY@COD on the macrophage phenotype in a simulated TME. The expression of CD86 (M1 phenotype) was significantly increased (49.8%) in the CuN₃Cl‐BAY@COD group, while that of CD206 (M2 phenotype) was downregulated to 14.4%, indicating that CuN₃Cl‐BAY@COD promoted the repolarization of M2 macrophages toward the M1 phenotype in the TME (Figures 4q,r and S28). These results demonstrated that CuN₃Cl‐BAY@COD enhanced the immunogenicity of prostate cancer cells and remodeled the immunosuppressive TME by inducing immunogenic disulfidptosis/pyroptosis, promoting DC maturation, and driving macrophage repolarization.

Based on the promising therapeutic effects in vitro, we further evaluated the antitumor efficacy of CuN₃Cl‐BAY@COD in vivo. Unilateral tumor‐bearing mice were randomized into five groups: control, CuN₃Cl, CuN₃Cl@COD, CuN₃Cl‐BAY, and CuN₃Cl‐BAY@COD. The mice received saline or the corresponding nanomaterials via tail vein injection on Days 1, 3, and 5 (Figure 5a). Tumor volume and body weight were monitored daily. When the tumor diameter in the control group reached 15 mm (Day 8), all mice were euthanized in accordance with institutional guidelines for animal care and use. Relative to the control group, the tumor inhibition rate in the CuN₃Cl group was 38.65%, indicating that CuN₃Cl with excellent enzyme‐mimicking activities significantly suppressed tumor growth in vivo. The highest inhibition rate was observed in the CuN₃Cl‐BAY@COD group (Figure 5b,c). Hematoxylin and eosin (H&E)‐stained sections of tumor tissues corroborated these findings (Figure S29). No significant differences in body weight were observed among the groups throughout the experimental period (Figure 5d). Furthermore, H&E‐stained sections of the major organs and tissues from treated mice revealed no histopathological abnormalities (Figure S30), demonstrating the favorable biocompatibility of CuN₃Cl‐BAY@COD in vivo. Because CuN₃Cl‐BAY@COD induced ICD and remodeled the TME in vitro, we further investigated its antitumor immunotherapeutic efficacy in vivo using a bilateral tumor‐bearing mouse model (Figure 5e). Mice were randomly divided into the same five groups. Compared with the control group, CuN₃Cl not only inhibited primary tumor growth through its enzyme‐mimicking activities but also elicited a tumor‐specific immune response to suppress distant tumor growth (Figures 5f–h and S31). The strongest inhibition of distant tumor growth was observed in the CuN₃Cl‐BAY@COD group. H&E‐stained sections of both primary and distant tumors corroborated this conclusion (Figure 5i). Consistent with the unilateral model, no significant difference in body weight change was observed among the groups throughout the study (Figure 5j), further confirming the excellent biocompatibility of CuN₃Cl‐BAY@COD in vivo.

FIGURE 5.

(a) Schematic illustration of the treatment schedule for unilateral RM‐1 tumor‐bearing mice. (b) Tumor volume changes during treatment (n = 5). (c) Tumor weights after treatment (n = 5). (d) Body weight changes of unilateral tumor‐bearing mice during treatment (n = 5). (e) Schematic illustration of the treatment schedule for bilateral RM‐1 tumor‐bearing mice. Tumor volume changes of (f) primary and (g) distant tumors during treatment (n = 5). (h) Distant tumor weights after treatment (n = 5). (i) Representative images of hematoxylin and eosin‐stained sections of primary and distant tumors (scale bar: 50 µm). (j) Body weight changes of bilateral tumor‐bearing mice during treatment (n = 5). All data were presented as mean ± SD. Statistical significance was determined using a two‐tailed Student's t‐test, with P‐values indicated in the respective figures.

Furthermore, we assessed in vivo antitumor immunity through immunofluorescent staining of CRT in primary tumors. As shown in Figure S32, the strongest fluorescence signal was observed in the CuN₃Cl‐BAY@COD group. We also examined the effect of CuN₃Cl‐BAY@COD on DC maturation in mouse lymph nodes (Figure 6a,b). Relative to untreated mice, the percentage of mature DCs gradually increased across treatment groups and reached its peak (38.63%) in the CuN₃Cl‐BAY@COD group. Because mature DCs activate T lymphocytes to initiate a tumor‐specific immune response, we subsequently evaluated changes in CD4⁺ and CD8⁺ T cell populations in mouse spleens (Figure 6c–f). The largest populations of activated CD4⁺ and CD8⁺ T cells were observed in the CuN₃Cl‐BAY@COD group. Similarly, analysis of the TME revealed extensive infiltration of these activated T cells, which was highest in the CuN₃Cl‐BAY@COD group (Figure 6g–j). These findings were further corroborated by immunofluorescent staining of tumor tissues (Figure 6k). Concurrently, the secretion of granzyme B and perforin was elevated in the treatment groups (Figure 6l). Systemic immune responses were evaluated by measuring pro‐inflammatory cytokine levels in mouse serum. Secretion levels of TNF‐α, IL‐6, and IL‐12p70 progressively increased in all treatment groups (Figure S33), with the highest concentrations detected in the CuN₃Cl‐BAY@COD group. Finally, macrophage phenotypes in tumor tissues were also assessed via flow cytometry. As shown in Figure S34, the proportion of M1‐like macrophages gradually increased while that of M2‐like macrophages concomitantly decreased across the treatment groups. The highest proportion of M1‐like and the lowest proportion of M2‐like macrophages were observed in the CuN₃Cl‐BAY@COD group. Immunofluorescent staining of tumor tissues further corroborated these findings (Figure S35).

FIGURE 6.

Flow cytometric analysis of (a, b) DC maturation in lymph nodes, (c–f) the proportions of splenic CD4⁺ and CD8⁺ T cells, and (g–j) tumor‐infiltrating CD4⁺ and CD8⁺ T cells across various treatment groups (n = 3). Representative immunofluorescence images of (k) CD4⁺ and CD8⁺ T cells and (l) perforin and granzyme B in tumor sections (scale bar: 50 µm). All data were presented as mean ± SD. Statistical significance was determined using a two‐tailed Student's t‐test, with P‐values indicated in the respective figures.

3. Conclusion

We successfully developed an efficient Cu‐based single‐atom nanozyme platform for prostate cancer immunotherapy via an axial coordination and defect engineering strategy. By breaking the inherent planar symmetry of conventional SANs with Cu‐N₄ configurations, this dual‐engineering strategy effectively regulated the electronic and geometric structures of the Cu sites, facilitated electron transfer, and significantly lowered the energy barrier for catalytic reactions, thereby endowing CuN₃Cl with robust multienzyme‐mimicking activities. Consequently, CuN₃Cl‐BAY@COD disrupted intracellular redox homeostasis and amplified BAY‐876‐mediated disulfide stress through ROS generation, GSH depletion, and cystine accumulation, thereby triggering disulfidptosis and pyroptosis. Furthermore, the loaded COD degraded cholesterol to inhibit tumor cell invasion while elevating intracellular H₂O₂ levels to exacerbate pyroptosis. Ultimately, by coupling metabolic reprogramming with dual immunogenic cell death, CuN₃Cl‐BAY@COD significantly enhanced tumor immunogenicity, remodeled the immunosuppressive TME, and promoted robust antitumor immune responses, offering a novel paradigm for prostate cancer immunotherapy.

Author Contributions

Bo Xu: conceptualization, investigation, methodology, validation, writing – original draft. Bin Wang: methodology. Wen Zhang: investigation. Bin Zhang: investigation, methodology. Xiaofeng Wang: investigation. Siqi Liu: methodology. Yinzhe Wang: methodology. Lianchao Yang: visualization. Rui Niu: conceptualization, investigation, supervision. Ying Tang: conceptualization, supervision. Chunxi Wang: supervision. Yinghui Wang: conceptualization, supervision, funding acquisition, writing – review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: anie73525‐sup‐0001‐SuppMat.docx.

Data Availability Statement

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