Morpho-crystallo-electronic Co-optimization of PtAu/GOx nanozyme for enhanced NIR-II photothermal-cascade catalytic tumor therapy Authors: Xu Sun, Long Zhang, Jiaxin Niu, Gongxian Sun, Hangyang Ling, Shengyan Wang, Rui Zhang, Shuyu Chen, Hairui Deng, Xiangdong Kong, Xiaolu Zhuo, Han Zhang Published: 2026 ABSTRACT Cascade catalytic therapy offers a specific strategy by sequentially converting highly expressed endogenous substances (e.g., H2O2, glucose) in the tumor microenvironment into highly cytotoxic reactive oxygen species to kill cancer cells. Photothermal therapy has high precision and low side effects. The combination of photothermal therapy and cascade catalytic therapy holds promise in overcoming current limitations in tumor therapy. Although noble metal-based nanozymes exhibit both stable photothermal conversion and multiple catalytic activities, their intrinsic inefficiency and unclear structure-activity relationships hinder further clinical translation. This work proposed a “morphological-crystallographic-electronic” tripartite synergistic optimization strategy. A multifunctional PtAu/GOx composite nanozyme was synthesized through wet-chemical methods, leveraging high surface area for active site exposure, high/low-index hybrid facets for optimized substrate adsorption, and Au-Pt charge transfer to regulate d-band centers for enhancing the multi-enzyme-like catalytic activities. This enables an efficient cascade catalytic cycle (H2O2→O2→more H2O2→•OH) within tumor. Moreover, the PtAu/GOx composite nanozyme has superior photothermal conversion performance under the irradiation of second near-infrared laser and can synergistically achieve the photothermal-augmented tumor cascade catalytic therapy. Advanced characterizations and simulations were employed to decipher their structure-activity relationships. Both in vitro and in vivo experiments validate the enhanced antitumor efficacy of the PtAu/GOx nanozyme through the synergistic optimization. Introduction With the advancement of eras and the improvement of medical conditions, there is a growing demand for precision and non-invasiveness in tumor therapy. Therefore, the development of tumor treatment methods that are minimally invasive and have low side effects is extremely urgent. Specific enzyme-catalyzed therapeutic strategies utilizing the tumor microenvironment (TME) have become one of the hot research topics in the current field of tumor therapy [1]. Tumor cells have different metabolic pathways from normal cells due to the metabolic reprogramming [2,3], creating TME with distinctive biochemical features such as hypoxic, acidic and high concentrations of glucose, glutathione and hydrogen peroxide (H2O2) [[4], [5], [6]], which provide specific targets for tumor therapy [7]. Enzyme-catalyzed therapies lie in utilizing catalytic reactions to convert over-accumulated endogenous substances within the TME into highly cytotoxic reactive oxygen species (ROS), thereby inducing oxidative stress-mediated death of tumor cells. Among them, cascade catalysis therapy can sequentially catalyze multiple endogenous substances present in the TME and the product of an initial reaction serves as the substrate for the subsequent reaction, forming a self-sustaining catalytic cycle. This strategy overcomes the kinetic constraints and the substrate dependence inherent to single-enzyme reactions, significantly enhancing both catalytic efficiency and therapeutic outcome [8,9]. For example, catalase (CAT) first catalyzes endogenous H2O2 in the TME to generate oxygen (O2). Glucose oxidase (GOx) then utilizes the O2 produced in the previous step, along with the high concentration of glucose in the TME, to generate gluconic acid and H2O2, achieving starvation therapy while simultaneously supplying substrate for the subsequent POD catalysis. Finally, POD further converts H2O2 into highly toxic hydroxyl radicals (•OH), realizing a coordinated CAT-GOx-POD cascade catalytic therapy [10]. Nanozyme is a class of nanomaterials harboring enzymological characterizations, which have similar catalytic activities and kinetic behaviors to bioenzymes [11]. Compared with bioenzymes such as protein and RNA, nanozyme exhibits the advantages of simple synthesis, high stability and easy modification, which have been widely applied in nanocatalytic medicine [[12], [13], [14]]. Metal-based nanozyme consist of metals or metal oxides [15]. They are capable of mimicking multiple enzyme-like catalytic activities (such as CAT-like and POD-like) by integration of different metal components, constituting a cascade catalysis path for biomimetic self-cascade catalytic tumor therapy [10,16]. Common metal-based nanozymes include Fe-based, Au-based, Pt-based and Mn-based nanomaterials [17]. In addition, metal nanozymes containing noble metals (e.g., Au, Pt, Ag, Pd and Ru) possess localized surface plasmon resonance (LSPR) properties, which endow them outstanding photothermal conversion performance. Photothermal therapy (PTT) has high precision and low side effects. Researchers have found that the heat generated from photothermal conversion can enhance enzymatic activity. The integration of PTT with cascade catalysis therapy can achieve a synergistically enhanced therapeutic outcome [18,19]. Therefore, metal nanozymes provide the potential of developing photothermal-augmented tumor cascade catalytic therapy. The catalytic activity of metal nanozymes can be co-modulated by their intrinsic properties, such as composition, morphology (specific surface area, surface defects, etc.), crystal structure (exposed facets, atomic arrangement, etc.) and electronic structure (d-band center, charge transfer, etc.) [[20], [21], [22], [23]]. In recent years, researchers have proposed various strategies to enhance the therapeutic efficacy of metal nanozymes in tumor treatment. Deng et al. synthesized RuTe2 nanoparticles with POD-like activity and co-loaded them with natural GOx and the 3,3′,5,5′-tetramethylbenzidine (TMB) molecule, successfully constructing a RuTe2-GOx-TMB composite nanoreactor as a cascade reaction system. In this multicomponent catalyst, the RuTe2-GOx component can generate •OH via the GOx-POD cascade reaction. The TMB component then reacts with •OH to form ox-TMB, which exhibits photothermal conversion capability in the near-infrared region, thereby initiating photothermal therapy [24]. To enhance the catalytic activity of the nanozyme, this study combined metal nanozymes with natural enzymes and integrated photothermal therapy for synergistic tumor treatment. With the advancement of research, the construction of metal-based composite nanozymes is increasingly trending towards complexity and multifunctionality. Recently, Li et al. developed an “all-in-one” MnO2@PtAuRu nanoreactor for self-supplying cascade catalytic therapy of tumors. The researchers loaded ultrafine PtAuRu nanoclusters (∼2 nm) into hollow MnO2 spheres. The resulting construct exhibited multiple enzyme-like activities (GOx/CAT/POD), enabling cascade catalytic therapy. The MnO2 carrier degrades in the weakly acidic TME, releasing the PtAuRu nanoclusters and consuming glutathione, thereby further enhancing the ROS-mediated therapeutic effect [25]. The aforementioned studies demonstrate that efficient tumor therapy can be achieved by constructing complex composite nanozyme systems. However, there is a notable lack of studies aimed at optimizing nanozyme catalytic activity by strategically engineering the intrinsic morphological, crystallographic and electronic structures of metal nanozymes. Furthermore, the tripartite synergistic optimization mechanisms remain largely unexplored. A deeper investigation into the influence of their intrinsic structural properties on catalytic activity is therefore crucial and of significant scientific importance, as it would not only enhance the efficacy of cascade catalytic therapy but also help elucidate the underlying mechanisms of catalytic treatment [26]. Based on the above inspirations, we synthesized a porous PtAu nanoframe (PAN) with a rough surface through a growth-etching method and then modified with GOx to obtain a multifunctional PtAu/GOx composite nanozyme (PAGN) system (Fig. 1). By integrating CAT-like, GOx, and POD-like activities into a cascade system, PAGN enabled the continuous generation of •OH. This disrupts intracellular redox homeostasis, induces tumor cell apoptosis, and ultimately achieves effective cascade catalytic therapy. Furthermore, PAN exhibits superior photothermal conversion performance, which can be utilized to further enhance the efficiency of tumor therapy.Fig. 1Schematic illustration of the synthetic route of the PAGN and its application for NIR-II photothermal-augmented cascade catalytic tumor therapy. Schematic illustration of the synthetic route of the PAGN and its application for NIR-II photothermal-augmented cascade catalytic tumor therapy. Understanding the structure-activity relationships of the PAGN system is crucial for predicting and designing highly efficient metal composite nanozymes [27]. The catalytic activity of PAN was systematically optimized through the engineering of its morphological, crystallographic, and electronic structures. Morphologically, PAN features a framework structure with a high specific surface area, ensuring sufficient exposure of active sites. Crystallographically, the (111) plane exhibits moderate H2O2 adsorption strength, which is considered optimal for catalytic activity. Electronically, the d-band center is upshifted through electronic coupling between Pt and Au atoms, thereby optimizing the adsorption and desorption behavior of PAN toward H2O2 and related intermediates. PAN features a pentagonal bipyramid nanoframe structure with a high specific surface area, which increases the utilization of atoms and facilitates the exposure of more catalytic active sites [28,29]. PAN is a PtAu alloy (Pt:Au = 1:1) nanoframe with predominantly exposed (111), (200), (220), and (311) crystal planes. PAGN demonstrated high specific CAT-like and POD-like activities of 101.5 and 0.3813 U mg−1, respectively. Its catalytic behavior adhered to the Michaelis-Menten model of steady-state kinetics [30]. The Michaelis constant (Km) and maximum reaction rate (Vmax) were determined to be 518.20 mM and 0.3417 μM s−1 for the CAT-like reaction, and 28.81 mM and 3.24 × 10−8 M s−1 for the POD-like reaction, respectively. The catalytic mechanisms of PAN can be elaborated by density functional theory (DFT) computations from microkinetics and thermodynamics perspectives, with the adsorption energy, d-band center (εd) and energy barrier (Eb) of rate-determining step (RDS) serving as the key activity descriptors [31,32]. DFT computations identified the (111) plane as the most active facet according to the Sabatier principle. The optimal H2O2 adsorption energy (−0.66 eV), an upward-shifted εd (−2.52 eV), and the low Eb of RDS (0.48 and 0.05 eV for CAT-like and POD-like reactions, respectively) collectively account for the superior catalytic activities of (111) plane, suggesting that the PtAu alloying can modulate the electronic structure, optimize the adsorption of H2O2, and lower the Eb of RDS. PAN also possesses tunable LSPR properties [33]. The LSPR peak of PAN can be tuned to the second near-infrared (NIR-II) region by regulating its geometrical parameters. Finite-difference time-domain (FDTD) simulation shows that PAN exhibits pronounced electric field localization [34,35]. Under 1064 nm laser irradiation, PAN exhibits rapid photothermal response and achieves a high photothermal conversion efficiency of 74.77%. Antitumor evaluations confirmed the high specificity and efficacy of PAGN. It displayed good cytocompatibility with L929 cells (>80% viability at 50 μg mL−1) and significant cytotoxicity against 4T1 cells (5.59% viability at 40 μg mL−1 under H2O2 and laser irradiation) in vitro. PAGN triggers a cascade catalytic reaction that generates abundant •OH, provoking a ROS storm and inducing apoptosis in 4T1 cells through redox homeostasis disruption. Furthermore, the antitumor molecular mechanisms of PAGN were revealed in detail by transcriptome analysis. PAGN not only suppresses the proliferation and metastasis of 4T1 cells by downregulating the expression of related genes, but also activates oxidative stress and cell death pathways. Antitumor evaluation in vivo validated the efficacy and biosafety of PAGN. PAGN can effectively inhibit tumor growth with tumor growth inhibition (TGI) ratio as high as 85.11%. In addition, PAGN has reliable blood safety and metabolic safety. This work presents a promising strategy for enhancing the effectiveness and sustainability of tumor catalytic therapy. Experimental section Synthesis of PAN and PAGN CA (0.2 M, 1.5 mL) and HAuCl4 (0.01 M, 1.5 mL) were added into CTAC (0.05 M, 60 mL). NaBH4 (0.05 M, 0.75 mL) freshly prepared in ice ultrapure (UP) water was added rapidly after thorough mixing. PAN seed was synthesized at 100°C in oil bath for 2.5 h under vigorous stirring (900 r min−1). HCl (1 M, 800 μL), AgNO3 (0.01 M, 400 μL), HAuCl4 (0.01 M, 1.0 mL) and H2PtCl6 (0.01 M, 1.0 mL) were added into CTAB (0.1 M, 40 mL), then AA (0.1 M, 320 μL) freshly prepared and 300 μL of PAN seed were added sequentially after thorough mixing. PAN was synthesized at 70°C for 48 h. PAN (100 μg mL−1, 10 mL) was centrifuged at 7000 r min−1 for 8 min and redispersed in HS-PEG-NH2 (200 μg mL−1, 10 mL). PAPN was synthesized by stirring at 600 r min−1 for 12 h at room temperature (RT) after ultrasonic mixing. 500 μg of GOx was dissolved in 10 mL of PBS (0.01 M, pH = 6.5) containing 5 mg of NHS and 5 mg of EDC, after that the carboxyl groups of GOx were activated by vortex mixing at RT for 1 h. 10 mL of PAPN was centrifuged at 7000 r min−1 for 8 min and redispersed in 10 mL of activated GOx. Subsequently, PAGN was synthesized by stirring at 600 r min−1 for 12 h at RT after ultrasonic mixing. Photothermal conversion of PAN 200 μL of PAN at different concentrations (10, 20, 40, 60, 80 and 100 μg mL−1) and UP water were irradiated under a 1064 nm laser at power density of 1.00 W cm−2. PAN (200 μL, 80 μg mL−1) was irradiated under 1064 nm laser at different power densities (0.50, 0.75 and 1.00 W cm−2). PAN (200 μL, 80 μg mL−1) was irradiated under 1064 nm laser at power density of 1 W cm−2, the laser was switched off when the system temperature rose to maximum steady-state value and switched on when it fell to initial value. The process was repeated for four cycles. The infrared thermography camera was employed to record the temperature in real time throughout aforementioned processes. The cooling phase of the first photothermal cycle was applied as heat transfer process to calculate the photothermal conversion efficiency (η) by a reported method [36]. η can be expressed as follows:(1)η=hS(Tmax−Tsur)−QdisI(1−10−A1064)where h is the heat transfer coefficient, S is the irradiated area, Tmax represents the maximum steady-state temperature of system (79.0°C), Tsur represents the temperature of the surroundings (19.1°C), I is the power of 1064 nm laser (0.5 W) and A1064 is the absorbance of PAN at a wavelength of 1064 nm (1.425). Qdis is derived from:(2)Qdis=hS(Tmax,water−Tsur) where Tmax,water is the maximum steady-state temperature of water (30.1°C). and hS is given by:(3)hS=mCpτwhere m and Cp are considered as the mass (0.2 g) and heat capacity (4.2 J (g °C)−1) of water, τ is the thermal time constant of the process. τ can be calculated as 114.16 s by linear fitting t versus −lnθ, corresponding equation goes here:(4)t=−τlnθwhere t and θ are the cooling time and the temperature driving force. θ is defined as:(5)θ=T−TsurTmax−Tsur Consequently, η was solved as 74.77% by considering the above equations. CAT-like catalytic activity assay of PAGN PAGN (20 μg mL−1) and H2O2 (1 M) were added to 3 mL of PBS (0.01 M, pH = 6.5), then thoroughly mixed via vortex. The dissolved O2 concentration was continuously monitored for 30 min using the dissolved oxygen meter with gentle stirring at RT. PAGN (10, 20, 30, 50 and 100 μg mL−1) and H2O2 (1 M) were added to 3 mL of PBS (0.01 M, pH = 6.5), then thoroughly mixed via vortex. The dissolved O2 concentration was continuously monitored for 60 s using the dissolved oxygen meter with gentle stirring at RT. The reaction from 0 to 60 s was defined as zero-order reaction, used to calculate initial rate of the CAT-like catalytic reaction. The following equations were used to calculate the CAT-like catalytic activity (ACAT) and specific activity (SACAT) of PAGN.(6)ACAT=V×31.25×60=(d[O2]dt)×31.25×60(7)SACAT=d[ACAT]dmwhere V is the catalytic reaction initial rate (mg L−1 s−1), [O2] is the dissolved O2 concentration (mg L−1), t is the time (s) and m is the PAGN mass (mg). PAGN (20 μg mL−1) and H2O2 (100, 200, 300, 400, 600, 800, 1200 and 1600 mM) were added to 3 mL of PBS (0.01 M, pH = 6.5), then thoroughly mixed via vortex. The dissolved O2 concentration was continuously monitored for 60 s using the dissolved oxygen meter with gentle stirring at RT. GOx catalytic activity assay of PAGN PAGN (100 μg mL−1) and glucose (1 mg mL−1) reacted in 3 mL of PBS (0.01 M, pH = 6.5) at 37°C in oil bath with gentle stirring. The glucose depletion, O2 consumption and pH reduction were measured by the dissolved oxygen meter, blood glucose meter and pH meter, respectively. POD-like catalytic activity assay of PAGN PAGN (20 μg mL−1), H2O2 (1.5 M) and TMB (1 mM in DMF) were added to 1 mL of PBS (0.01 M, pH = 6.5). After reacting for 5 min at RT, the UV-vis spectra were measured by the UV-vis spectrophotometer. PAGN (1, 2, 4, 8 and 16 μg mL−1), H2O2 (1 M) and TMB (1 mM in DMF) were added to 1 mL of PBS (0.01 M, pH 6.5). After mixing quickly, the absorbance at 652 nm was measured directly for 60 s by the UV-vis spectrophotometer. The reaction from 0 to 60 s was defined as zero-order reaction, used to calculate initial rate of the POD-like catalytic reaction. The POD-like catalytic activity (APOD) and specific activity (SAPOD) of PAGN were calculated by the following equations:(8)APOD=Vε×l×dA652dt(9)SAPOD=d[APOD]dmwhere V represents the volume of reaction solution (μL), ε represents the molar absorption coefficient of ox-TMB (39000 M−1 cm−1), l represents the path length of cuvette (cm) and A652 represents the absorbance at 652 nm. PAGN (10 μg mL−1), H2O2 (10, 20, 40, 60 and 80 mM) and TMB (1 mM in DMF) were added to 1 mL of PBS (0.01 M, pH = 6.5), then thoroughly mixed via vortex. Following quick mixing, the absorbance at 652 nm was measured immediately for 60 s by the UV-vis spectrophotometer. CAT-like and POD-like catalytic kinetics analysis of PAGN The Michaelis-Menten equation (Equation (10)) is a steady-state kinetics model which describes the relationship between the substrate concentration and the rate of enzyme-catalyzed reaction. It can be linearized by the Hanes-Woolf equation (Equation (11)), plotting [S]V versus [S]. The line fitted by Hanes-Woolf equation has a slope equal to 1Vmax and an x-intercept equal to −Km.(10)V=Vmax[S]Km+[S](11)[S]V=[S]Vmax+KmVmaxwhere V represents the catalytic reaction rate, Vmax represents the maximum catalytic reaction rate, [S] represents the substrate (H2O2) concentration and Km represents the Michaelis constant. DFT computations DFT computations were used to interpret the CAT-like and POD-like catalytic mechanisms of PAGN. All the DFT computations were executed on the Vienna Ab initio Simulation Package [37,38]. The exchange-correlation potential was described using the Perdew-Burke-Ernzerhof generalized gradient approximation [39]. The electron-ion interactions were accounted by the projector augmented wave [40]. All the DFT computations were performed at a cut-off energy of 400 eV, and the 2 × 2 × 1 Gamma centered Monkhorst-Pack grid k-points were applied to sample the Brillouin zone integration. The energy and force convergence criteria of the self-consistent iteration were set to 10−4 eV and 0.05 eV Å−1, respectively. The van der Waals interactions were described by the DFT-D3 method [41]. The adsorption energy (Eads) was defined as follows:(12)Eads=Etotal−EH2O2−EPANwhere Etotal, EH2O2 and EPAN represent the energies of adsorption system (H2O2 adsorbed on the PAN plane), isolated H2O2 and isolated PAN plane, respectively. The computational hydrogen electrode (CHE) model was employed to compute the Gibbs free energy change (ΔG) of the proton-coupled electron transfer (PCET) step. The one half of the chemical potential of hydrogen was set as the chemical potential of the proton-electron pair in the CHE model. The ΔG was determined using the following equation:(13)ΔG=ΔE+ΔZPE−TΔSwhere ΔE represents the electronic energy difference between the initial and final states of the PCET step, ΔZPE is the zero-point energy difference between the initial and final states and TΔS represents the entropy difference between the initial and final states. T is the absolute temperature. Cytocompatibility assay of PAGN L929 cells were seeded at density of 1 × 104 cells well−1 in a 96-well plate and incubated for 12 h. After washing twice with PBS, the cells were treated with PAN, PAPN and PAGN separately at varying concentrations (10, 20, 30, 40 and 50 μg mL−1) and incubated for 24 h. After washing twice with PBS, 100 μL of DMEM containing 10 μL of CCK-8 reagent was added to each well, followed by incubation at 37°C for 1 h. The absorbance at 450 nm was measured using the microplate reader. Cell viability was calculated using the following formula:(14)Cellviability=At−AbcAnc−Abc×100%where At, Abc and Anc represent the absorbance of the treatment, blank control and negative control group, respectively. Cytotoxicity assessment of PAGN in 4T1 cells 4T1 cells were typically seeded at 1.5 × 105 cells per dish or well (unless specified for the CCK-8 assay at 1 × 104 cells well−1 in a 96-well plate), incubated for 12 h, and washed twice with PBS before treatment. The treatment conditions for cytotoxicity evaluation consisted of 40 μg mL−1 PAGN, 300 μM H2O2, and 5 min of 1064 nm laser irradiation (1 W cm−2), followed by an 8-12 h incubation. Specific assays were then conducted: 1) Cellular uptake of FITC-labeled PAGN was analyzed by flow cytometry (FCM) at various time points and visualized by confocal laser scanning microscopy (CLSM) at 4 h using DiI and DAPI counterstains. 2) Cell viability was quantified using the CCK-8 assay. 3) Apoptosis was assessed by FCM after staining with Annexin V-FITC and PI. 4) Live/dead cell staining was performed using calcein AM and PI and observed via CLSM. 5) Intracellular O2 and ROS levels were detected by CLSM after staining with RDPP and DCFH-DA, respectively. Extracting and sequencing of RNA 4T1 cells were seeded density of 3 × 105 cells dish−1 in 6 cm culture dish and incubated for 12 h. The cells treated with 300 mM of H2O2, 40 μg mL−1 of PAGN and 5 min of 1064 nm laser irradiation were set as treatment group, meanwhile untreated cells were set as control group. Total RNA was extracted using trizol reagent. After 6 h of the treatment, the cells including floating cells from each dish were collected by trypsin digestion into RNase-free centrifuge tubes, 1.5 mL of trizol reagent was added to each tube. After thorough mixing, left the mixture at RT for 5 min to ensure adequate lysis. RNA sequencing analysis was performed on the high-throughput sequencing platform (NovaSeq X Plus, Illumina, USA). Antitumor evaluation of PGAN in vivo All animal procedures were performed by the Guidelines for Care and Use of Laboratory Animals of Zhejiang Sci-Tech University and approved by the Animal Ethics Committee of Zhejiang Sci-Tech University (assigned study No.20250220-01). Twenty-four 4T1 tumor-bearing mice were randomly divided into four groups, when the tumor volume reached approximately 100 mm3. On the 1st and 7th days of the treatment, 4T1 tumor-bearing mice were subjected to different treatments (PBS, PAGN, Laser and PAGN + Laser). Mice in groups B and D were administered PAGN (100 μL, 10 mg kg−1) while mice in group A were received 100 μL of PBS, via intravenous injection. Twelve hours post-injection, mice in groups C and D were irradiated with 1064 nm laser (1 W cm−2, 5 min) on tumors. During the treatment, tumor volume and body weight of mice were recorded daily. Treatment was terminated when the tumor volume reached approximately 1000 mm3. Tumor volume was calculated using the formula:(15)V=lw22where V, l and w represent the volume, length and width of tumor, respectively. After 14 days of treatment, blood was collected from the mice for complete blood count and comprehensive metabolic panel, and then major organs (heart, liver, spleen, lung and kidney) and tumors were harvested from the mice for H&E staining and TUNEL staining. The TGI ratio is defined as follows:(16)TGIratio=(1−ΔVtΔVnc)×100%Where Vt and Vnc are the tumor volume of the treatment and control groups, respectively. Statistics analysis The Shapiro-Wilk test was utilized to assess normality of continuous variables. Normally distributed variables are reported as mean ± standard deviation. The F-test was applied to confirm the homogeneity of variance of two groups of continuous variables. For significance test, two groups of normally distributed continuous variables with homogeneous variances were compared using the Student's t-test, whereas those with heterogeneous variances were compared using the Welch's t-test, and the Mann-Whitney U test was employed to make a comparison between two groups of continuous variables which were not normally distributed. All of the above hypothesis tests are unpaired two-tailed and are considered statistically to be extremely significant (***), very significant (**) and significant (*) at p < 0.001, p < 0.01 and p < 0.05, respectively. The aforementioned statistics analysis was performed using the Prism (10.1.2, GraphPad Software, USA). Results and discussion Characterizations of PAN The transmission electron microscopy (TEM) and high-angle angular dark-field scanning transmission electron microscopy (HAADF-STEM) images show that PAN exhibits pentagonal bipyramid nanoframe structure (Fig. 2a and b). The nanoframe structure of PAN yield a high BET surface area of 7.50 m2 g−1 (Fig. S1), which improves the utilization of Pt and Au atoms and facilitates the exposure of more catalytic active sites. The geometrical parameters of PAN are highly tunable, offering the possibility of controlling its LSPR properties. The particle size distribution (PSD) was obtained by measuring the structural parameters of PAN in TEM images, revealing an average width of 51 nm and an average length of 145 nm (n = 100) (Fig. 2c). The narrow PSD suggests the uniformity of PAN. The X-ray diffraction (XRD) pattern shows the prominent characteristic diffraction peaks of 38.97°, 45.29°, 65.97°, 79.16° and 83.98° (Fig. 2d). By matching the powder diffraction file (PDF), the peak positions and relative intensities of PAN were highly consistent with those of Pt0.5Au0.5 (ICDD No. 04-003-7035). The peaks can be indexed to (111), (200), (220), (311) and (222) planes in the face-centered cubic (FCC) structure. Along the zone axis [01-1], the lattice fringes were observed by high resolution transmission electron microscopy (HRTEM) (Fig. 2e). Fast Fourier transform (FFT) analysis of the selected area (dashed rectangle) revealed diffraction spots corresponding to the (111), (200), and (311) planes. Inverse FFT (IFFT) processing was then employed to measure the interplanar spacings, yielding values of 2.29 Å for the (111) plane and 2.02 Å for the (200) plane, with an interplanar angle of 52.03° (Fig. 2f and S2). The Pt and Au atomic columns are ordered without obvious lattice distortion, confirming the formation of a substitutional solid solution (Fig. 2f). The selected area electron diffraction (SAED) pattern of PAN consists of a series of well-defined concentric diffraction rings that were indexed as (111), (200), (220), (311) and (222) planes (Fig. 2g). In the above plane indexing, the measured d-spacing was compared with the calculated d-spacing from the PDF to get indices of planes. Combining the above crystallographic analysis, the unit cell of PAN was resolved into FCC structure (a = b = c = 3.995 Å, α = β = γ = 90°) (Fig. 2h).Fig. 2Synthesis and characterizations of PAN. (a,b) TEM and HAADF-STEM images of PAN. (c) PSD of PAN. (d) XRD pattern of PAN. (e) HRTEM image of PAN (inset: FFT pattern within the dashed rectangle. (f) Filtered HRTEM image of the PAN within the dashed rectangle. (g) SAED pattern of PAN. (h) A unit cell of PAN. (i,j) TEM elemental mapping and line scanning of PAN. (k,l) XPS Pt 4f and Au 4f narrow scans of PAN. (m,n) Regulation the morphology of PAN by seed amount and etching time, respectively. Synthesis and characterizations of PAN. (a,b) TEM and HAADF-STEM images of PAN. (c) PSD of PAN. (d) XRD pattern of PAN. (e) HRTEM image of PAN (inset: FFT pattern within the dashed rectangle. (f) Filtered HRTEM image of the PAN within the dashed rectangle. (g) SAED pattern of PAN. (h) A unit cell of PAN. (i,j) TEM elemental mapping and line scanning of PAN. (k,l) XPS Pt 4f and Au 4f narrow scans of PAN. (m,n) Regulation the morphology of PAN by seed amount and etching time, respectively. The TEM elemental maps, line scanning and HAADF-STEM-energy dispersive spectroscopy (HAADF-STEM-EDS) spectrum confirm that the PAN only consists of Pt and Au elements (Fig. 2i,j and S3). Quantitative analysis determined a near 1:1 Pt/Au stoichiometry (weight ratio = 49.83:50.17; atom ratio = 50.07:49.93) in PAN, which is consistent with both the inductively coupled plasma-optical emission spectroscopy (ICP-OES) results (Pt: 26.45 μg mL−1, Au: 26.38 μg mL−1) and the XRD phase identification. X-ray photoelectron spectroscopy (XPS) scans further determined the chemical states of Pt and Au on the surface of PAN (Fig. 2k,l and S4). XPS Pt 4f narrow scan shows a spin-orbit splitting doublet peak attributed to metallic Pt (Pt0) consists of Pt0 4f7/2 (70.81 eV) and Pt0 4f5/2 (74.15 eV), and a spin-orbit splitting doublet peak attributed to the oxidized Pt (Pt2+) consists of Pt2+ 4f7/2 (71.74 eV) and Pt0 4f5/2 (75.16 eV) (Fig. 2k). XPS Au 4f narrow scan shows a spin-orbit splitting doublet peak identified as metallic Au (Au0) consists of Au0 4f7/2 (84.53 eV) and Au0 4f5/2 (88.20 eV) (Fig. 2l). XPS narrow scan analysis evidences that the chemical state of Au is Au0, and the chemical states of Pt are Pt0 and Pt2+, on the surface of PAN. The relative atom percentages of Pt0, Pt2+ and Au0 on surface are 30.9%, 15.14% and 53.96%, respectively. The seed amount and etching time are two key factors in regulating the structure of PAN. The size of PAN reduces as the seed amount increases (Fig. 2m) and the cavity of PAN becomes larger with the increase of seed amount and etching time (Fig. 2m and n). Based on the above characterizations, PAN possesses a framework structure with a high specific surface area and hybrid high/low-index facets. We further modified GOx onto PAN to develop a multifunctional PtAu/GOx composite nanozyme. The grafting efficiency of the GOx on PAN was calculated as 56.4% using BCA assay (Fig. S5). The zeta potentials of PAN and PAGN were 5.74 and −8.39 mV, respectively, indicating that the modification of GOx on the surface of PAN can change its zeta potential from positive to negative, which was favorable for improving the biocompatibility and prolonging the circulation time in blood (Fig. S6a). PAN and PAGN were dispersed in water, PBS and DMEM, respectively. After 48 h, PAN agglomerated severely, while PAGN had no obvious agglomeration, indicating the modification of GOx on the surface of PAN can improve its dispersion stability (Fig. S6b). Furthermore, PAN and PAGN were dispersed in PBS and DMEM, and their hydrodynamic sizes were measured at various time points (6, 12, 24 and 48 h) by DLS method to evaluate dispersion stability. Within 48 h, the hydrodynamic size of PAGN shows no obvious fluctuation in either PBS or DMEM, whereas the hydrodynamic size of PAN increases sharply under both conditions (Fig. S6c,d). These results confirm that GOx modification significantly improves the dispersion stability of PAN. LSPR properties and photothermal performance of PAN The ultraviolet-visible (UV-vis) spectrum of PAN shows a prominent LSPR peak in the NIR-II region at 1025 nm (Fig. 3a), which agrees well with the simulated extinction spectrum (Fig. 3b). The electric field enhancement distributions show the field enhancement is largely concentrated at two ends of the PAN, consistent with the classical “tip effect” in anisotropic plasmonic nanoparticles. Additionally, the rough surface of PAN features numerous protrusions that exhibit locally enhanced electric fields (Fig. 3c and d), further contributing to the overall field confinement. The combination of strong NIR-II extinction and electric field localization endows PAN with excellent photothermal performance.Fig. 3FDTD simulations and photothermal performance of PAN. (a) UV-vis extinction spectrum of PAN. (b) The simulated extinction cross-section spectrum of PAN. (c,d) 2D and 3D electric field enhancement contours (log |E/E0|2) of a single PAN at its longitudinal LSPR wavelength. (e,f) Infrared thermography images and corresponding temperature of PAN under 1064 nm laser irradiation (1.00 W cm−2). (g) The temperature change of PAN (80 μg mL−1) under 1064 nm laser irradiation with different intensities. (h) Photothermal cycles of PAN (80 μg mL−1) under 1064 nm laser (1.00 W cm−2). (i) Calculation of the thermal time constant in the cooling phase of the first photothermal cycle. FDTD simulations and photothermal performance of PAN. (a) UV-vis extinction spectrum of PAN. (b) The simulated extinction cross-section spectrum of PAN. (c,d) 2D and 3D electric field enhancement contours (log |E/E0|2) of a single PAN at its longitudinal LSPR wavelength. (e,f) Infrared thermography images and corresponding temperature of PAN under 1064 nm laser irradiation (1.00 W cm−2). (g) The temperature change of PAN (80 μg mL−1) under 1064 nm laser irradiation with different intensities. (h) Photothermal cycles of PAN (80 μg mL−1) under 1064 nm laser (1.00 W cm−2). (i) Calculation of the thermal time constant in the cooling phase of the first photothermal cycle. The photothermal performance of PAN was investigated under 1064 nm laser irradiation. PAN at concentrations of 10, 20, 40, 60, 80 and 100 μg mL−1 were heated to 49.2, 60.3, 74.6, 76.3, 77.6 and 80.4°C under 1064 nm laser at 1.00 W cm−2 within 5 min, respectively (Fig. 3e and f). PAN at concentration of 80 μg mL−1 was heated to 57.9, 72.4 and 80.0°C under 1064 nm laser at power densities of 0.50, 0.75 and 1.00 W cm−2 within 6 min, respectively (Fig. 3g). These demonstrate the photothermal conversion of PAN is positively correlated with the concentration of PAN, power density of laser and duration, which enables precise regulation of temperature. Notably, PAN at a mere concentration of just 40 μg mL−1 surged to a high temperature of 46.2°C within the first minute, indicating PAN exhibits a responsive and efficient photothermal performance. Furthermore, there is no significant attenuation in the maximum steady-state temperature after undergoing four photothermal cycles, which indicates the strong photothermal stability of PAN (Fig. 3h). The thermal time constant during the cooling phase of the first photothermal cycle was calculated to be 114.16 s (Fig. 3i) and the photothermal conversion efficiency of PAN was ultimately solved to be 74.77%. The photothermal performance of PAN is not significantly affected by its size or cavity (Fig. S7). These results prove the excellent photothermal performance of PAN, which can be utilized in the PTT applications. Enzyme-like catalytic activities of PAGN The dissolved O2 concentration continuously climbed to 11.32 mg L−1 within 30 min following the introduction of PAGN into H2O2, whereas the dissolved O2 levels in either H2O2 or PAGN alone remained stable (Fig. 4a). As expected, PAGN exhibits CAT-like catalytic activity at pH of 6.5, which can rapidly catalyze the decomposition of H2O2 to produce O2. The CAT-like catalytic activity increases with decreasing PAN size and increasing cavity volume due to higher specific surface area and more interior active sites (Fig. S8). To further quantitatively assess the CAT-like catalytic activity of PAGN, different masses (0.03, 0.06, 0.09, 0.15 and 0.30 mg) of PAGN were introduced into the same concentration of H2O2. The CAT-like catalytic activity of PAGN exhibits mass dependence (Fig. S9a). The CAT-like catalytic specific activity of PAGN was calculated to be 101.50 U mg−1, based on the initial rate of the CAT-like catalytic reaction (Fig. 4b). The initial rates of the CAT-like catalytic reaction were measured by varying the concentration (100, 200, 300, 400, 600, 800, 1200 and 1600 mM) of H2O2 and fixing the concentration of PAGN to reveal kinetics behavior of the CAT-like catalytic reaction (Fig. 4c). When H2O2 concentration (100, 200 and 300 mM) is low, the catalytic rate is proportional to H2O2 concentration, following first-order reaction. As the concentration of H2O2 (400, 600 and 800 mM) increases, the growth of catalytic rate becomes slow, following mixed-order reaction. When H2O2 reaches a certain concentration (1200 mM), the catalytic rate tends to be constant, following zero-order reaction (Fig. 4d). The CAT-like activity of PAGN followed Michaelis-Menten kinetics, with Km of 546.40 mM and Vmax of 0.34 μM s−1. The Km value, which represents the substrate concentration at half of Vmax, indicates the affinity for H2O2, while Vmax reflects the maximum catalytic capacity. These kinetic parameters are summarized in Fig. 4e.Fig. 4Enzyme-like catalytic activities of PAGN. (a) O2 concentration over time under distinct conditions. (b) CAT-like catalytic specific activity of PAGN. (c) O2 increment by CAT-like catalytic activity of PAGN at different concentrations of H2O2. (d) Michaelis-Menten fitting for describing kinetics behavior in CAT-like catalytic reaction of PAGN. (e) Primary kinetic parameters of CAT-like catalytic reaction of PAGN. (f-h) Glucose depletion, O2 consumption and H2O2 generation by GOx catalytic activity of PAGN at different times. (i) ESR spectra of •OH trapped by DMPO with and without 1064 nm laser irradiation. (j,k) Absorbance increment at 652 nm with different concentrations of PAGN and H2O2. (l) Michaelis-Menten fitting for describing kinetics behavior in POD-like catalytic reaction of PAGN. Enzyme-like catalytic activities of PAGN. (a) O2 concentration over time under distinct conditions. (b) CAT-like catalytic specific activity of PAGN. (c) O2 increment by CAT-like catalytic activity of PAGN at different concentrations of H2O2. (d) Michaelis-Menten fitting for describing kinetics behavior in CAT-like catalytic reaction of PAGN. (e) Primary kinetic parameters of CAT-like catalytic reaction of PAGN. (f-h) Glucose depletion, O2 consumption and H2O2 generation by GOx catalytic activity of PAGN at different times. (i) ESR spectra of •OH trapped by DMPO with and without 1064 nm laser irradiation. (j,k) Absorbance increment at 652 nm with different concentrations of PAGN and H2O2. (l) Michaelis-Menten fitting for describing kinetics behavior in POD-like catalytic reaction of PAGN. Glucose can be oxidized to gluconic acid and H2O2 by O2 under the catalysis of GOx. The H2O2 generation was measured by UV-vis spectrophotometer using Ti(SO4)2 as the chromogenic agent. This measurement was based on the formation of a yellow peroxo-titanium complex with Ti(SO4)2, which has a characteristic absorption at 415 nm. Within 5 h, the glucose was depleted from 6.13 to 4.27 mM (Fig. 4f). The dissolved O2 was consumed by the rate of 4.15 mg L−1 within 60 min (Fig. 4g). The H2O2 was generated from 0 to 1.62 mM (Fig. 4h) and the pH decreased from 6.50 to 3.55 (Fig. S9b). These results confirm that PAGN retains the GOx catalytic activity. POD-like catalytic activity of PAGN was evaluated by UV-vis spectrophotometry using TMB as the chromogenic substrate. The POD-like catalytic activity can catalyze the one-electron oxidation of TMB in the presence of H2O2, generating blue dimer charge-transfer complex (ox-TMB) that exhibits a characteristic absorption peak at 652 nm. A distinct absorption peak at 652 nm emerges in the presence of both H2O2 and PAGN, whereas no peak emerges in the presence of only H2O2 or PAGN (Fig. S9c). It is demonstrated that PAGN has POD-like catalytic activity at pH of 6.5. The electron spin resonance (ESR) spectra exhibit characteristic quartet peaks of DMPO-•OH adduct with intensity ratio of 1:2:2:1, confirming the production of •OH (Fig. 4i). It is worth noting that the ESR intensity with laser irradiation is 2.08-fold higher than that without laser irradiation, which suggests that laser irradiation can promote the production of •OH. This phenomenon can be attributed to the photothermal enhancement of the POD-like activity in PAGN. The elevated temperature enhances the POD-like activity of PAGN through two mechanisms. First, it increases the frequency of effective collisions between H2O2 molecules and the surface Pt/Au atoms. Second, it induces surface reconstruction by enhanced lattice vibrations, thereby generating more active sites and thus improving the intrinsic catalytic activity. To determine the POD-like catalytic specific activity of PAGN, the increments of absorbance at 652 nm with different PAGN masses were measured by UV-vis spectrophotometer. In the POD-like catalytic reaction of PAGN, the catalytic rate is positively correlated with PAGN mass and the specific activity of PAGN is 0.3813 U mg−1 (Fig. 4j). The POD-like catalytic kinetics behavior of PAGN was investigated by determining initial catalytic rates under varying H2O2 concentrations (10, 20, 40, 60 and 80 mM) while maintaining constant PAGN concentration (Fig. 4k). At low H2O2 concentration, the catalytic reaction is considered to be first-order. At intermediate H2O2 concentration, the catalytic reaction is regarded as mixed-order. At high H2O2 concentration, the catalytic reaction constitutes zero-order (Fig. 4l). The POD-like catalytic kinetics behavior of PAGN follows steady-state kinetics Michaelis-Menten equation with Km and Vmax of 28.81 mM and 3.24 × 10−2 μM s−1. Equally, it is consistent with the Hanes-Woolf equation (Fig. S9d). PAGN has multiple catalytic activities and can constitute cascade catalytic reaction to improve the catalytic efficiency. The absorbance of ox-TMB with PAGN group was noticeably higher than that with PAN group (Fig. S10), proving that the introduction of GOx has enabled an enhanced cascade catalysis (H2O2→O2 →H2O2→•OH), thereby increasing the yield of •OH. Catalytic microscopic mechanisms of PAGN Although the apparent kinetic analysis demonstrated that the CAT-like and POD-like catalytic behaviors of PAGN conformed to the classical steady-state kinetic equations, this analysis only reflected the macroscopic relationship between the catalytic reaction rate and substrate concentration but failed to provide insights into the microscopic mechanism of the CAT-like and POD-like catalysis of PAGN. However, DFT allows the computation of potential energy surfaces along the reaction coordinates in CAT-like and POD-like catalytic reactions, providing atomic-scale interpretations into understanding the relationship between the surface structure and catalytic activity of PAGN. The adsorption configuration and adsorption energy were used to describe the adsorption behavior of H2O2 on PAGN planes. The adsorption energies of H2O2 molecule on PAGN (111), (200) and (311) planes are −0.66, −0.81 and −0.83 eV, indicating the possible formation of chemisorption (Eads < −0.50 eV) between H2O2 and these PAGN planes (Fig. 5a). According to the Sabatier principle, the adsorption between PAGN plane and H2O2 should be moderate, which facilitates the desorption and dissociation of H2O2 and avoids the prolonged occupation of the catalytic active sites, thus accelerating the CAT-like and POD-like catalytic reaction. Therefore, the adsorption of H2O2 on PAGN (111) plane is considered to be the most conducive to the CAT-like and POD-like catalytic reaction.Fig. 5Catalytic mechanisms of PAGN by DFT computations. (a) The adsorption configurations, adsorption energies and bond lengths of H2O2 on PAGN (111), (200) and (311) planes. (b) The CDD and Bader charge transfer of H2O2 adsorbed on PAGN (111) plane. Yellow and cyan express charge accumulation and depletion, respectively. The isosurface value is set to 8 × 10−4 e Bohr−3. (c) The ELF of H2O2 adsorbed on PAGN (111) plane. (d) The pCOHP of O-Au bond. The Ef is shifted to 0 eV. (e) The PDOS of Pt and Au atoms in s and d orbitals on PAGN (111) plane. The Ef is shifted to 0 eV. (f,g) The free energy and relative free energy of CAT-like and POD-like catalytic reaction on PAGN (111) plane. IS, TS and FS represent initial state, transition state and final state, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Catalytic mechanisms of PAGN by DFT computations. (a) The adsorption configurations, adsorption energies and bond lengths of H2O2 on PAGN (111), (200) and (311) planes. (b) The CDD and Bader charge transfer of H2O2 adsorbed on PAGN (111) plane. Yellow and cyan express charge accumulation and depletion, respectively. The isosurface value is set to 8 × 10−4 e Bohr−3. (c) The ELF of H2O2 adsorbed on PAGN (111) plane. (d) The pCOHP of O-Au bond. The Ef is shifted to 0 eV. (e) The PDOS of Pt and Au atoms in s and d orbitals on PAGN (111) plane. The Ef is shifted to 0 eV. (f,g) The free energy and relative free energy of CAT-like and POD-like catalytic reaction on PAGN (111) plane. IS, TS and FS represent initial state, transition state and final state, respectively. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) The charge density difference (CDD) was executed to analyze the charge transfer between H2O2 and PAGN (111) plane. There is an explicit charge transfer between H2O2 and PAGN (111) plane, with significant charge density depletion around the Au atom of PAGN (111) and significant charge density accumulation around the O atom of the H2O2. Further, the Bader charge population quantitatively determined the Bader charge transfer of 0.07 e from PAGN (111) plane to H2O2 (Fig. 5b). The electron localization function (ELF) was employed to characterize the localization distribution of electron and determine the type of bond in the adsorption of H2O2 on PAGN (111) plane. The even electron delocalization distribution with ELF value about 0.15 between Pt and Au atoms agrees with the properties of metallic bond. Obviously, the high degree of electron localization distribution with ELF value about 0.80 between the O and Au atoms demonstrates the formation of covalent bond in the adsorption (Fig. 5c). The crystal orbital Hamilton population (COHP) reveals the O-Au bond strength by quantifying the bonding and antibonding contributions. The bonding orbitals exhibit stronger overlap and contribute more to bonding interactions near the energy of −5.56 eV, whereas the antibonding orbitals exhibit stronger overlap and contribute more to antibonding interactions near the energy of −0.85 eV. The integrated COHP (ICOHP) up to the Fermi energy (Ef) quantitatively characterized the strength of the O-Au bond, the ICOHP was computed as −0.25 (Fig. 5d). The adsorption behavior of noble metal nanozymes for H2O2 is closely related to εd. The εd of PAGN (111) plane was computed to be −2.52 eV by projected density of state (PDOS) (Fig. 5e), while the εd of Au (111) plane is −3.50 eV [42]. The εd of Au (111) plane is relatively low, but the introduction of Pt can significantly upshift the εd through alloying, thereby optimizing the adsorption of H2O2 [[43], [44], [45]]. In catalytic reactions, the RDS limits the overall reaction rate, therefore the Eb of RDS can be used as a descriptor of the catalytic activity of the PAGN. It has been confirmed that noble metal nanozymes usually exhibit CAT-like catalytic activity under basic conditions, while they usually exhibit POD-like catalytic activity under acidic conditions [46]. Therefore, in the study of the CAT-like and POD-like catalytic reaction mechanisms on PAGN (111) plane, OH− and H+ were pre-adsorbed on PAGN (111) plane to simulate basic and acidic chemical environments, respectively. Under basic conditions, the H2O2 reacts: 2H2O2* = 2H2O* + O2, where the homolytic cleavage of H2O2* is the RDS with Eb of 0.48 eV. Under acidic conditions, the H2O2 reacts: H2O2* + H* = H2O* + OH*, where the homolytic cleavage of H2O2* is the RDS with Eb of 0.05 eV (Fig. 5f and g). On the (111) plane of PAGN, the Eb values for the RDS of the CAT-like and POD-like reactions are reduced to 0.48 eV and 0.05 eV, compared to 0.80 eV and 0.61 eV on pure Au, respectively [46]. The PtAu alloying lowers the Eb of RDS, thus kinetically favors the CAT-like and POD-like catalytic reactions occurring on PAGN (111) plane. According to the Brønsted-Evans-Polanyi (BEP) relationship, the reaction energy (Er) is linearly correlated with the Eb of RDS in catalytic reactions, thus the Er can also be used as a descriptor of the catalytic activity of PAGN. The Er of the CAT-like and POD-like catalytic reaction of PAGN (111) plane are −1.46 and −1.70 eV, which conform to the BEP relationship (Fig. 5g). The initial state (IS), transition state (TS) and final state (FS) structures of CAT-like and POD-like catalytic reaction are arranged (Fig. S11). Antitumor evaluations of PAGN in vitro The viabilities of L929 cells were not significantly reduced remaining above 80% after treatment with PAN, PAPN and PAGN for 24 h in the concentration range of 10-50 μg mL−1, which indicates that PAN, PAPN and PAGN have good cytocompatibility to normal cells (Fig. 6a). The viabilities of 4T1 cells treated for 12 h with PAGN, PAGN + H2O2, PAGN + Laser and PAGN + H2O2 + Laser are 85.79%, 28.71%, 14.44% and 5.59%, respectively (Fig. 6b). After treatment with PAGN, PAGN + H2O2, PAGN + Laser and PAGN + H2O2 + Laser, intracellular ATP levels of 4T1 cells are reduced by 10.74%, 28.43%, 73.62% and 86.71%, respectively (Fig. 6c). The cellular uptake behavior of PAGN by 4T1 cells was quantitatively analyzed by FCM, and the uptake ratio can reach 71.4% after 4 h and 96.6% after 8 h (Fig. S12). CLSM images visualized the intracellular distribution of PAGN in 4T1 cells after 4 h of uptake (Fig. 6d). Those results suggest that PAGN can be efficiently internalized by 4T1 cells, providing a favorable prerequisite for PAGN to disrupt intracellular homeostasis of 4T1 cells. PAGN exhibits significant cytotoxicity to 4T1 cells by inducing oxidative and thermal stress in the presence of H2O2 and 1064 nm laser, and the dual stress can act synergistically to further reinforce the cytotoxicity to 4T1 cells.Fig. 6Antitumor evaluations of PAGN in vitro. (a) Cytocompatibility of PAN, PAPN and PAGN to L929 cells at different concentrations. (b) Cytotoxicity of PAGN to 4T1 cells under distinct conditions. (c) Intracellular ATP level in the 4T1 cells treated with distinct conditions. (d) Cellular uptake of PAGN by 4T1 cells after 4h of incubation. (e) Live/dead staining of 4T1 cells treated with distinct conditions. (f,g) Intracellular O2 and ROS in 4T1 cells under distinct treatments. (h) The apoptosis of 4T1 cells under distinct treatments. Antitumor evaluations of PAGN in vitro. (a) Cytocompatibility of PAN, PAPN and PAGN to L929 cells at different concentrations. (b) Cytotoxicity of PAGN to 4T1 cells under distinct conditions. (c) Intracellular ATP level in the 4T1 cells treated with distinct conditions. (d) Cellular uptake of PAGN by 4T1 cells after 4h of incubation. (e) Live/dead staining of 4T1 cells treated with distinct conditions. (f,g) Intracellular O2 and ROS in 4T1 cells under distinct treatments. (h) The apoptosis of 4T1 cells under distinct treatments. CLSM was employed to visually analyze the PAGN-mediated cytotoxicity and dysregulation of intracellular metabolic homeostasis in 4T1 cells. In the PAGN + H2O2, PAGN + Laser and PAGN + H2O2 + Laser groups, the fluorescence signal of dead cells gradually increases, indicating that H2O2 and 1064 nm laser can cooperatively augment the cytotoxicity of PAGN to 4T1 cells (Fig. 6e). To clarify the cause of this cytotoxicity, the effects of PAGN on the intracellular O2 and •OH levels in 4T1 cells under different conditions were examined. Compared with the control and H2O2 groups, the fluorescence signals of O2 in the PAGN + H2O2 and PAGN + H2O2 + Laser groups are significantly weakened, and the fluorescence signal in the PAGN + H2O2 + Laser group is even weaker (Fig. 6f). It was demonstrated that PAGN still has CAT-like catalytic activity in 4T1 cells to catalyze the generation of O2 from endogenous H2O2, and photothermal can promote the generation of O2. In stark contrast to the control and H2O2 groups, the fluorescence signals of •OH in the PAGN + H2O2 and PAGN + H2O2 + Laser groups are obviously strengthened, and the fluorescence signal of •OH in the PAGN + H2O2 + Laser group is even stronger (Fig. 6g). That demonstrates PAGN still has POD-like catalytic activity in 4T1 cells and is able to catalyze the production of •OH from intracellular H2O2, while photothermal facilitates the production of •OH. Notably, •OH production is positively correlated with O2 production, reflecting to some extent the cascade catalysis of PAGN in 4T1 cells. The above results indicate that PAGN generates large amount of •OH in 4T1 cells through cascade catalysis, causing a ROS storm, breaking the intracellular redox homeostasis, leading to a highly oxidative stress state and ultimately inducing apoptosis. In addition, PAGN can also induce apoptosis by causing heat stress through photothermal. To further quantify the induction of apoptosis in 4T1 cells by PAGN, apoptosis assay was performed by FCM using Annexin V/PI staining. The apoptosis ratios in PAGN + H2O2, PAGN + Laser and PAGN + H2O2 + Laser groups are 27.33%, 42.60% and 58.70%, respectively, whereas the apoptosis ratio in PAGN group is only 8.65% (Fig. 6h). It was validated that PAGN can significantly induce apoptosis in 4T1 cells in the presence of H2O2 or 1064 nm laser irradiation. In addition, the apoptosis ratio in PAGN + H2O2 + Laser group is increased by 16.10% compared with PAGN + H2O2 group, indicating that the photothermal effect can augment the induction of apoptosis in 4T1 cells by PAGN. Antitumor transcriptome analysis of PAGN To reveal the underlying molecular mechanisms of PAGN in antitumor, differential expressed genes (DEG) analysis was performed on 4T1 cells in control and treatment groups (Fig. 7a). Compared to control group, the treatment of PAGN led to the dysregulation of approximately 4576 genes in 4T1 cells, including 3348 upregulated and 1228 downregulated genes, as identified by differential expression analysis using the absolute fold change (FC) ≥ 2.0 (Fig. 7b). Notably, several biomarkers (Mki67, CCND1, CDK6, HER2, MET, BCL2 and BCL-XL) associated with tumor cell proliferation and metastasis are downregulated, suggesting that PAGN may exert antitumor efficacy by suppressing malignant biological processes such as tumor growth and dissemination [[47], [48], [49], [50], [51], [52], [53]]. Concurrently, the treatment group exhibits increased expression of canonical antioxidant enzymes and key genes HMOX1, DUOX1 and DUOX2 involved in ROS detoxification [[54], [55], [56]].Fig. 7Antitumor transcriptome analysis of PAGN. (a) Heat map displaying the expression levels of the DEG between control and treatment groups, which red and green represent significantly upregulated and downregulated genes, respectively. (|FC| ≥ 2.0, p < 0.05 and n = 3). (b) Volcano plot depicting DEG between the control and treatment groups. (c) GO enrichment analysis of the DEG. The significantly enriched gene categories were marked. (d) KEGG pathway enrichment analysis between control and treatment groups. (e) GSEA analysis of DEG in GOBP including the hydrogen peroxide biosynthetic process, positive regulation of apoptotic signaling pathway and positive regulation of reactive oxygen species metabolic process and response to reactive oxygen species. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Antitumor transcriptome analysis of PAGN. (a) Heat map displaying the expression levels of the DEG between control and treatment groups, which red and green represent significantly upregulated and downregulated genes, respectively. (|FC| ≥ 2.0, p < 0.05 and n = 3). (b) Volcano plot depicting DEG between the control and treatment groups. (c) GO enrichment analysis of the DEG. The significantly enriched gene categories were marked. (d) KEGG pathway enrichment analysis between control and treatment groups. (e) GSEA analysis of DEG in GOBP including the hydrogen peroxide biosynthetic process, positive regulation of apoptotic signaling pathway and positive regulation of reactive oxygen species metabolic process and response to reactive oxygen species. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.) Furthermore, gene ontology (GO) enrichment analysis was conducted to reveal the biological processes after treatment with PAGN. The dysregulated genes are significantly involved in various biological processes, predominantly including cellular metal ion homeostasis, positive regulation of cell cycle, glucose homeostasis, positive regulation of ROS metabolic process, stress response to metal ion. These biological processes are related to the regulating of PAGN mentioned above, such as inhibiting overactive tumor cell cycle, promoting ROS production and regulating metal ion homeostasis and metabolism (Fig. 7c). The Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis showed that the PAGN treatment elicited an up-regulation in the expression of PI3K-Akt, MAPK, NF-κB and IL-17 signaling pathways (Fig. 7d) [[57], [58], [59]]. Most of the above signaling pathways are closely related to cell death, oxidative stress and inflammatory regulation. Subsequently, changes in genetic information patterns were examined in detail. The PAGN treatment significantly involved the regulation of ROS, and Gene Set Enrichment Analysis (GSEA) analysis confirmed that the gene set of gene ontology biological processes (GOBP) including the H2O2 biosynthetic process, positive regulation of apoptotic signaling pathway and positive regulation of ROS metabolic process and response to ROS (Fig. 7e). Antitumor evaluations of PAGN in vivo PAGN does not induce significant hemolysis, with hemolysis ratio of only 2.84% at 100 μg mL−1, which demonstrates PAGN has reliable hemocompatibility and allows PAGN to be administered intravenously (Fig. S13). To evaluate the tumor therapeutic efficacy in vivo, PAGN was administered into 4T1 tumor-bearing mice by intravenous injection (Fig. 8a). 4T1 tumor-bearing mice were subjected to different treatments (Control, Laser, PAGN and PAGN + Laser), when the tumor volume reached approximately 100 mm3.Fig. 8Tumor therapy evaluations of PAGN in vivo. (a) The establishment of cell line-derived allograft model and the administration regimen of the PAGN. (b,c) Tumor volume, body weight of the 4T1 tumor-bearing mice throughout distinct treatments. (d-g) The tumor growth inhibition ratio, tumor images, tumor weight and TUNEL mean fluorescence intensity of the 4T1 tumor-bearing mice after distinct treatments. (h,i) H&E staining and TUNEL staining of tumors from the 4T1 tumor-bearing mice after distinct treatments. Tumor therapy evaluations of PAGN in vivo. (a) The establishment of cell line-derived allograft model and the administration regimen of the PAGN. (b,c) Tumor volume, body weight of the 4T1 tumor-bearing mice throughout distinct treatments. (d-g) The tumor growth inhibition ratio, tumor images, tumor weight and TUNEL mean fluorescence intensity of the 4T1 tumor-bearing mice after distinct treatments. (h,i) H&E staining and TUNEL staining of tumors from the 4T1 tumor-bearing mice after distinct treatments. After 14 days of treatment, the tumor volumes of PAGN and PAGN + Laser groups are 868 and 798 mm3, as the tumor volumes of control and laser groups are 452 and 188 mm3 (Fig. 8b). There are no noticeable fluctuations in the body weights of the mice in PAGN and PAGN + Laser groups throughout the treatment (Fig. 8c), indicating that PAGN has no significant side effects. For further quantifying the tumor therapeutic efficacy in vivo, tumor volumes at the endpoint of the treatment were used to calculate the tumor growth inhibition (TGI) ratio (Fig. 8d). In contrast to control group, the TGI ratio in PAGN group is 56.21%, and the TGI ratio in PAGN + Laser group achieves 85.11%. The tumor sizes in PAGN and PAGN + Laser groups are clearly smaller than the control group (Fig. 8e). The tumor weights in the Laser, PAGN and PAGN + Laser groups are 1.16, 0.99, 0.57 and 0.11 g, and the tumor weights in PAGN and PAGN + Laser groups are reduced by 51% and 91%, respectively, compared with the control group (Fig. 8f). The results proved that PAGN can effectively inhibit tumor growth. It is because PAGN selectively accumulates in tumor tissue through the EPR effect. After being internalized by tumor cells, PAGN catalyzes endogenous H2O2 into highly oxidative •OH via cascade catalytic reactions, severely disrupts the redox homeostasis in tumor cells and ultimately inducing apoptosis under oxidative stress. Histopathological examination of the tumors in PAGN and the PAGN + Laser groups confirmed tumor therapeutic efficacy of PAGN in vivo. H&E staining shows morphological changes in tumor such as reduced cell density, karyopyknosis and karyorrhexis, which are the characteristics of necrosis. TUNEL staining shows apoptosis in tumor, TUNEL-positive cells are significantly increased. Semi-quantitative analysis suggests the mean fluorescence intensity (MFI) of TUNEL in PAGN and PAGN + Laser groups are 14.66 and 26.75 times higher than the control group, respectively (Fig. 8g). H&E and TUNEL staining demonstrate that PAGN produces excellent tumor therapeutic efficacy by inducing necrosis and apoptosis in tumor cells (Fig. 8h and i). In order to further evaluate the biosafety of PAGN, hematological examinations were performed on mice in PAGN and PAGN + Laser groups. The complete blood count (CBC) indexes are within the normal range, which confirmed the hemocompatibility of PAGN (Fig. S14a). In addition, comprehensive metabolic panel (CMP) shows normal results, indicating that PAGN has no metabolic toxicity and the liver and kidney functions of the mice remain normal (Fig. S14b). H&E staining of major organs (heart, liver, spleen, lung and kidney) of the mice are not found evident pathological changes in PAGN and PAGN + Laser groups, which proves that PAGN has good histocompatibility (Fig. S14c). Conclusion In this study, GOx was introduced into PtAu nanoframe to develop a multifunctional PtAu/GOx composite nanozyme for photothermal-augmented catalytic therapy of tumor. PAN exhibits excellent photothermal conversion performance with an efficiency of 74.77%. FDTD simulations elucidate the underlying photothermal conversion mechanism. PAGN exhibits outstanding CAT-like, GOx and POD-like catalytic activities, where the CAT-like and POD-like catalytic behaviors are in accordance with the Michaelis-Menten equation of steady-state kinetics. DFT computations have elucidated in depth that the PAN (111) plane is the most favorable plane for CAT-like and POD-like catalytic reactions. The Eb of RDS in the CAT-like and POD-like catalytic reactions are 0.48 and 0.05 eV, respectively. PAGN can generate a large number of •OH through the cascade catalytic reaction, which can disrupt the intracellular redox homeostasis and effectively reduce the viability of tumor cells. Transcriptome analysis has revealed that PAGN can inhibit proliferation and metastasis of tumor cells by downregulating the expression of Mki67, CCND1, CDK6, HER2, MET, BCL2 and BCL-XL, and cause death, oxidative stress and inflammatory of tumor cells through PI3K-Akt, MAPK, NF-κB and IL-17 signaling pathways. Treatment with PAGN led to the upregulation of genes involved in ROS metabolism, response to ROS, apoptosis, and H2O2 biosynthesis. A significant tumor growth inhibition (TGI = 85.11%) was observed, with no notable signs of toxicity. The excellent therapeutic performance is due to the synergistic effect of photothermal and cascade catalytic actions. This study offers new insights into the design of multifunctional nanozymes for tumor therapy. CRediT authorship contribution statement Xu Sun: Data curation, Formal analysis, Investigation, Methodology, Validation, Writing – original draft. Long Zhang: Methodology, Software, Validation. Jiaxin Niu: Formal analysis, Validation. Gongxian Sun: Formal analysis, Validation. Hangyang Ling: Data curation, Validation. Shengyan Wang: Methodology. Rui Zhang: Methodology. Shuyu Chen: Validation. Hairui Deng: Validation. Xiangdong Kong: Funding acquisition, Resources, Supervision, Writing – review & editing. Xiaolu Zhuo: Funding acquisition, Methodology, Supervision. Han Zhang: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing. 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.