Visualized self-propelled nanorobots overcome tumor barriers for boosted photothermal-nanocatalytic synergistic therapy against triple-negative breast cancer Authors: Ze Wang, Hongyu Chu, Jincheng Du, Hao Liang, Yiyang Ding, Linlin Liu, Quan Lin, Ying Wang Published: 2026 ABSTRACT Conventional nanozyme-based therapeutics for triple-negative breast cancer (TNBC) are severely hindered by insufficient tumor penetration and lysosomal degradation-induced enzyme inactivation, crippling therapeutic efficacy. Herein, we rationally construct visualized dual-propelled Janus PDAPdAu nanorobots (PPA NBs) with cascade catalytic activity for synergistic photothermal-nanocatalytic therapy of TNBC. Powered by near-infrared light-triggered thermophoresis and nanozyme-fueled oxygen propulsion, PPA NBs realize deep tumor penetration and lysosomal escape. PPA NBs exhibit catalase-mimicking activity to alleviate tumor hypoxia through hydrogen peroxide decomposition, and glucose oxidase-mimicking activity to deplete glucose and downregulate heat shock proteins, thus sensitizing TNBC to photothermal therapy (PTT). Localized hyperthermia further enhances nanozyme catalysis, forming an amplified synergy between nanocatalytic therapy and PTT that affords an outstanding 97% tumor inhibition rate. Furthermore, PPA NBs act as excellent fluorescent biosensing probes for visual lesion localization. Overall, this work provides visual diagnosis and advanced TNBC phototherapy strategies, a reliable paradigm for cancer patients’ health monitoring and treatment.Graphical Abstract Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12951-026-05065-y. Introduction Triple-negative breast cancer (TNBC) is an exceptionally aggressive and rapidly spreading malignancy, marked by heightened tendencies for recurrence and fatal outcomes [1, 3]. Due to the negative expression of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2), it cannot benefit from endocrine therapy or anti-HER2 therapy, and current management remains largely reliant on conventional treatment modalities including surgery, chemotherapy, and radiotherapy [4, 6]. However, given such critical limitations as lack of actionable targets and severe systemic toxicity, TNBC typically demonstrates suboptimal responsiveness to these conventional therapeutic strategies [7, 8]. In addition, highly efficient and non-invasive visualization approaches for real-time monitoring of tumor location and size changes during therapeutic intervention remain lacking. This pressing challenge underscores the urgent need to develop innovative and promising therapeutic paradigms that transcend existing treatment frameworks. Driven by advances in nanotechnology, growing research attention has been directed toward the design and engineering of nanozymes for combating TNBC, owing to their unique nanoscale effects, favorable enzymatic catalytic stability, ease of functionalization, and cost-effectiveness [6, 9, 10]. Among them, noble metal nanozymes have garnered considerable attention due to their unique physicochemical properties and versatile enzyme-like capabilities. These nanozymes serve as catalytic cores for nanocatalytic therapy, such as mimicking glucose oxidase (GOx) activity to reduce the intratumoral glucose level, and mimicking catalase (CAT) activities to generate oxygen (O2) in the tumor microenvironment (TME), thereby alleviating hypoxia while inducing catalytic tumor suppression [11, 13]. Nevertheless, standalone nanozyme platforms suffer from two unresolved physical and intracellular barriers that markedly compromise therapeutic efficacy. First, the dense fibrotic stroma, hyperplastic connective tissues, and highly viscous extracellular matrix (ECM) of TNBC form rigid physical barriers, leading to passive diffusion-limited penetration; most nanozymes only accumulate at the tumor periphery and fail to reach deep hypoxic tumor parenchyma [14, 16]. Second, once internalized by tumor cells, freely dispersed nanozymes are readily trapped within lysosomes and degraded by hydrolases, resulting in irreversible loss of catalytic activity before exerting tumoricidal functions [17, 19]. Beyond delivery drawbacks, monofunctional nanozymes also fail to achieve synergistic therapy: single CAT-like nanozymes cannot cut off tumor energy supply, while single GOx-mimicking nanozymes aggravate intratumoral hypoxia without oxygen compensation. Furthermore, existing nanozyme systems rarely integrate real-time tumor imaging functions, making it impossible to dynamically track nanoparticle distribution and lesion response during treatment. These bottlenecks severely restrict the translational potential of nanozyme monotherapy for TNBC diagnosis and treatment. Endowing nanozymes with autonomous propulsion overcomes their passive diffusion limitation, giving rise to nanomotors driven by endogenous enzymatic forces. Such propulsion is fueled by substrate-specific reactions in the TME, where endogenous components such as hydrogen peroxide (H2O2), L-arginine and plasma amine oxidase act as substrates or catalysts [14, 17, 20, 23]. As intelligent nanorobots, these nanomotors theoretically achieve deeper tumor penetration and lysosomal escape than static nanozymes, prerequisites for effective biosensing, visualized monitoring and therapy. However, endogenous propulsion is limited by local substrate concentration, making a dual-drive mode that combines endogenous and exogenous forces desirable. Biocompatible near-infrared (NIR) lasers are ideal exogenous drivers, owing to their tissue penetration depth, tunable intensity and controllable directionality [19, 24]. Under NIR irradiation, asymmetrically structured nanorobots generate localized temperature gradients whose thermophoretic forces drive autonomous propulsion [25, 28]. NIR-induced photothermal therapy (PTT) is moreover inherently suited to TNBC [29, 30], since the generated heat both kills tumor cells and accelerates enzymatic kinetics. Integrating PTT with nanocatalytic therapy in intelligent nanorobots has therefore become a key research direction. Against this background, the visualized light/nanozyme dual-propelled Janus PDAPdAu nanorobots (PPA NBs) with cascade catalysis were rationally designed to enable deep tumor penetration, lysosome escape, real-time tumor visualization monitoring and synergistic photothermal-nanocatalytic therapy against TNBC (Scheme 1). Upon the NIR light irradiation, the asymmetrically polydopamine (PDA) and gold nanodots (AuNDs) absorbed the NIR light, forming a thermal gradient around PPA NBs, and the resulting autothermophoresis effect constitutes the first driving force for propulsion. Abundant H2O2 within the TME is catalytically decomposed into a large quantity of O2 through the CAT-mimicking activity of palladium nanoparticles (PdNPs), which constitutes the secondary driving force for the movement of PPA NBs. Powered by the light/nanozyme dual propulsion, the nanorobots achieve both deep intratumoral penetration and lysosomal escape within tumor cells, ensuring their sustained functionality for both therapy and tumor visualization monitoring. On the one hand, the AuNDs act as GOx-like to degrade glucose within the tumor cells. This process not only furnishes sufficient H2O2 as a substrate to fuel the CAT-like reaction of PdNPs, but also cuts off the energy supply for tumor cells. Energy-starved tumor cells exhibit an intrinsic hypersensitivity to PTT. Additionally, this energy deprivation serves as a key driver of the downregulated expression of HSPs, which are otherwise massively upregulated by local hyperthermia and strongly counteract PTT efficacy [31]. On the other hand, the in-situ O2 generation via the catalytic action of PdNPs alleviates the hypoxia within the tumor cells. An increase in temperature will enhance the catalytic activity of nanozymes. Through the above cascade of synergistic events, organically integrating nanocatalytic therapy with PTT can achieve a supra-additive therapeutic effect (“1 + 1>2”) for TNBC treatment. Furthermore, the excellent fluorescence properties of AuNDs endow the nanorobots with fluorescence imaging capability, making them excellent fluorescent biosensing probes for visual lesion localization and health monitoring. Scheme 1Preparation of PPA NBs and schematic illustration of the supra-additive therapeutic effect of nanocatalytic therapy synergized with photothermal therapy Preparation of PPA NBs and schematic illustration of the supra-additive therapeutic effect of nanocatalytic therapy synergized with photothermal therapy Results and discussion Design, synthesis and characterization of PPA NBs The dual-propelled Janus PPA NBs with cascade catalysis were designed against TNBC. Firstly, mesoporous PDA nanoparticles (PDANPs) were synthesized by the method of the emulsion-induced interface anisotropy assembly. The block copolymer TMB/F127/PDA composite micelles served as the structural unit, which could promote efficient island nucleation and anisotropic growth on the emulsion droplets, thereby forming an asymmetric mesoporous PDA. Subsequently, PdNPs were in situ reduced on PDA to obtain catalase-like activity. Finally, AuNDs with excellent photothermal, fluorescence and GOx-like activities were loaded on PDAPd to obtain PPA NBs. The morphologies of PDA, PDAPd, AuNDs, and PPA NBs were characterized by transmission electron microscopy (TEM) and scanning electron microscopy (SEM). TEM analysis revealed that PDANPs possess uniform mesopores and exhibit minimal aggregation, indicating excellent dispersibility (Fig. 1a). The asymmetric PDANPs featured abundant irregular pores (approximately 10–25 nm) on their external surface, facilitating the loading of ultrasmall AuNDs. The XRD pattern of PDA is consistent with that previously reported [32], showing a very broad peak at 2θ = 22.5° (Figure S1), which proves the synthesis of PDA. As shown in Fig. 1b, PDAPd displayed small black spherical particles (size ranges from 4 to 8 nm) uniformly distributed on the PDA substrate, confirming successful in situ synthesis of PdNPs. Ultrasmall AuNDs with lysozyme as the ligand have been successfully synthesized, with an average size of 2.4 nm (Figure S2-S3). Subsequent loading of these AuNDs onto PDAPd yielded PPA NBs, with its morphology remaining largely unchanged (Fig. 1c and f). The elemental composition of PPA NBs was determined by Energy-dispersive X-ray spectroscopy (EDX). The C, N, and O elements were highly enriched in the core region, while Pd and Au were asymmetrically distributed on the PDA (Fig. 1d-e), indicating the successful synthesis of the PPA NBs. Structural characterization of PPA NBs was systematically performed. Hydrodynamic particle size and zeta potential were determined by dynamic light scattering (DLS). PDA exhibited a hydrodynamic particle size of 599.3 nm (Fig. 1g), slightly larger than dimensions observed in TEM images. This discrepancy is attributed to the outward-expanding hydration layer on the nanoparticle surface. After in-situ growth of Pd, PDAPd showed an increased hydrodynamic particle size of 694.9 nm. The hydrodynamic particle size of AuNDs was 77.91 nm (Figure S4). Following AuNDs loading, PPA NBs displayed a hydrodynamic particle size of 754.2 nm. Notably, PDA and PDAPd carried negative charges of -5.42 mV and − 14.1 mV, respectively (Fig. 1h). After loading the highly negatively charged gold nanoparticles (-32.82 mV, Figure S5), PPA NBs acquired a stronger negative charge (-25.15 mV). Negatively charged nanomaterials are not easy to bind to negatively charged substances in the blood, which reduces the loss of their systemic administration and is more conducive to their enrichment in tumor tissues for effective release [33]. UV-Vis absorption spectroscopy revealed a distinct surface plasmon resonance (SPR) peak at 537 nm for AuNDs (Fig. 1i). In contrast, PDA and PDAPd exhibited no characteristic absorption peaks across the measured wavelength range. Following AuNDs loading, PPA NBs displayed a prominent SPR peak at the identical position (537 nm), confirming successful gold deposition. Fourier transform infrared (FTIR) spectroscopy verified the successful synthesis of both PDA and AuNDs with lysozyme as the ligand (Fig. 1j). The absorption peak at 1597 cm− 1 in the FTIR spectrum of PDA is attributed to the stretching vibration of the aromatic ring [34]. The peak at 1512 cm− 1 belongs to the bending vibration of N–H in the amide group [35]. The peaks at 1352 cm− 1 and 1281 cm− 1 correspond to the phenolic C–O–H bending and C–O stretching vibrations, respectively [36]. The FTIR spectrum of AuNDs shows a characteristic amide I band at 1650 cm− 1, which is attributed to the C = O stretching vibration of the peptide backbone. After loading AuNDs, the FTIR spectrum of PPA NBs exhibits the corresponding characteristic absorption peaks, thereby confirming the successful synthesis of PPA NBs. Fig. 1Synthesis and Characterization of PPA NBs. (a)-(c) TEM images of the asymmetric mesoporous PDANPs, PDAPd NPs and PPA NBs. (d)-(e) HAADF, and corresponding mapping images of PPA NBs. Scale bar: 200 nm. (f) SEM image of PPA NBs. Scale bar: 500 nm. (g) Hydrodynamic particle size, (h) zeta potential, (i) UV–vis absorption spectra, and (j) FTIR spectra of materials. (k) XPS survey spectrum of PPA NBs. High-resolution XPS spectra of (l) C 1s, (m) O 1s, (n) Pd 3d and (o) Au 4f of PPA NBs Synthesis and Characterization of PPA NBs. (a)-(c) TEM images of the asymmetric mesoporous PDANPs, PDAPd NPs and PPA NBs. (d)-(e) HAADF, and corresponding mapping images of PPA NBs. Scale bar: 200 nm. (f) SEM image of PPA NBs. Scale bar: 500 nm. (g) Hydrodynamic particle size, (h) zeta potential, (i) UV–vis absorption spectra, and (j) FTIR spectra of materials. (k) XPS survey spectrum of PPA NBs. High-resolution XPS spectra of (l) C 1s, (m) O 1s, (n) Pd 3d and (o) Au 4f of PPA NBs Further composition and valence state analysis of PPA NBs was performed via X-ray photoelectron spectroscopy (XPS). The survey spectrum confirmed the presence of C 1s, O 1s, N 1s, Au 4f, and Pd 3d signatures (Fig. 1k). In the high-resolution C 1s spectrum, peaks at 284.8 eV, 286.1 eV, and 288.4 eV were assigned to C = C/C-C, C-O/C-N, and O = C-O bonds, respectively (Fig. 1l). The O 1s spectrum exhibited two peaks at 531.9 eV (C = O) and 533.5 eV (C-O) (Fig. 1m). The N 1s characteristic peak appeared at 399.6 eV (Figure S6). For Pd 3d spectrum (Fig. 1n), peaks at 337.3 eV (3d5/2) and 342.7 eV (3d3/2) corresponded to Pd2⁺, while those at 335.0 eV (3d5/2) and 340.3 eV (3d3/2) indicated Pd (0), confirming successful synthesis of PdNPs and providing the basis for PPA NBs’ catalase-like activity. The Au 4f spectrum showed characteristic peaks at 83.6 eV (4f7/2) and 87.4 eV (4f5/2), further verifying AuNDs loading (Fig. 1o). Collectively, these data conclusively demonstrate the successful synthesis of PPA NBs. Optical properties Fluorescence imaging is very important in tumor diagnosis and visualization, which has significant advantages such as non-invasiveness, real-time and high sensitivity. Gold nanoparticles (AuNPs), as an outstanding type of fluorescent nanoprobes, offer powerful tools for precise tumor imaging due to their ultrasmall size and excellent biocompatibility. As shown in Fig. 2a, the AuNDs exhibit a distinct red fluorescence emission peak at 666 nm. Further, the AuNDs were loaded onto PDAPd to form PPA. PPA NBs still maintains excellent fluorescence properties (Figure S7). The stability of the fluorescent probe is a key indicator for evaluating its applicability in vivo. Therefore, we systematically evaluated the stability of PPA within the physiological-related pH range (4–10) (Figure S8). The results showed that the fluorescence intensity remained stable within this range, demonstrating excellent pH stability. Meanwhile, even in the presence of common cations (such as Ca2+, K⁺, Na⁺, NH4+) at high concentration interference, the fluorescence intensity did not show significant changes (Figure S9), confirming its good ionic stability. Moreover, after storage for 7 days, the fluorescence intensity of PPA NBs remained good (Figure S10). In conclusion, based on its excellent pH stability, ionic stability, and long-term stability, the prepared PPA NBs can be used as a high-performance fluorescent biosensing probe for imaging studies both inside and outside the biological system. The photothermal material designed in this study aims to achieve a triple synergistic effect: (1) The Janus nanorobots generate a thermal gradient under NIR light irradiation, driving their autonomous movement based on the thermophoretic kinetics; (2) Local temperature increase accelerates the enzymatic reaction kinetics, enhancing the catalytic efficiency of the nanozymes; (3) PTT is utilized to achieve tumor thermal ablation. To enhance the photothermal performance of PDA, we uniformly loaded small-sized AuNDs inside its pores, hoping to achieve a synergistic enhancement of catalysis and photothermal effects. The photothermal performance of the obtained PPA NBs was evaluated. PPA NBs exhibited significant absorbance at 808 nm (Fig. 1i), which is conducive to its absorption of near-infrared light and the improvement of its photothermal performance. After being irradiated with 808 nm laser (0.5 W/cm2) for 10 min, the temperature of the PPA NBs and PDAPd significantly increased, while the temperature change of the PBS control group showed only a limited increase (Fig. 2b). The heating effect of PPA NBs was significantly better than that of PDAPd, which was due to the plasmonic effect of the incorporated AuNDs. The infrared thermal imaging graph visually presented this difference (Fig. 2d). Furthermore, under different power lasers (0.25, 0.5, 1.0, 1.5 W/cm2) irradiation, the temperature of the PPA NBs increased from room temperature to 36.4 ℃, 44.8 ℃, 63.8 ℃, and 78.8 ℃, respectively (Fig. 2c), confirming its potential as an efficient photothermal agent for PTT. Further calculations of the photothermal conversion efficiency of the PPA NBs were conducted by the temperature rise-cooling curve, and the photothermal conversion efficiency (η) of the PPA NBs was determined to be 45.5% (Fig. 2e-f). Furthermore, robust photothermal stability is essential for PTT agents. As shown in Fig. 2g, PPA NBs maintained its photothermal capability with minimal attenuation over three laser on/off cycles, demonstrating excellent stability and potential for repeatable PTT applications. Fig. 2Optical properties and Multi-enzymatic catalytic activities of PPA NBs. (a) Fluorescence spectra of AuNDs. (b) Temperature elevation of PBS, PDAPd and the PPA NBs under the 808 nm laser (0.5 W/cm2). (c) Temperature profiles of the PPA under 808 nm laser irradiation at different power densities. (d) Infrared thermal images of PBS and the PPA NBs under the 808 nm laser (0.5 W/cm2). (e) Temperature change curve of the PPA NBs under the 808 nm laser (0.5 W/cm2) on and off. (f) The plot of linear fitting cool time versus -Ln θ of the PPA NBs. (g) Thermal stability of the PPA NBs. (h) Schematic illustration of the multi-enzymatic cascade catalytic activity of the PPA NBs. (i) Glucose depletion by the PPA NBs at different concentrations. (j) O2 generation after different concentrations of the PPA NBs incubating with H2O2. (k) O2 generation of the PPA NBs incubating with H2O2 under different power densities Optical properties and Multi-enzymatic catalytic activities of PPA NBs. (a) Fluorescence spectra of AuNDs. (b) Temperature elevation of PBS, PDAPd and the PPA NBs under the 808 nm laser (0.5 W/cm2). (c) Temperature profiles of the PPA under 808 nm laser irradiation at different power densities. (d) Infrared thermal images of PBS and the PPA NBs under the 808 nm laser (0.5 W/cm2). (e) Temperature change curve of the PPA NBs under the 808 nm laser (0.5 W/cm2) on and off. (f) The plot of linear fitting cool time versus -Ln θ of the PPA NBs. (g) Thermal stability of the PPA NBs. (h) Schematic illustration of the multi-enzymatic cascade catalytic activity of the PPA NBs. (i) Glucose depletion by the PPA NBs at different concentrations. (j) O2 generation after different concentrations of the PPA NBs incubating with H2O2. (k) O2 generation of the PPA NBs incubating with H2O2 under different power densities Multi-enzymatic catalytic activities PPA NBs have excellent cascade catalytic activity and are expected to play a key role in the treatment of TNBC (Fig. 2h). AuNDs can simulate the function of GOx, catalyzing the oxidation reaction of glucose within cells to produce gluconic acid and H2O2. The consumption of glucose not only enables the starvation therapy and regulates the energy metabolism, but also generates H2O2 to serve as a substrate for the CAT enzyme. The GOx-like activity of the PPA NBs was evaluated. The results showed that the decrease in glucose content was positively correlated with the concentration of the PPA NBs, indicating that it has great GOx-like activity (Fig. 2i). The Pd-based nanoparticles exhibit excellent CAT-like activity, which can decompose H2O2 to produce O2. We evaluated the CAT-like activity of PPA NBs and its influencing factors. The H2O2 scavenging (Figure S11) and O2 production rate (Fig. 2j) of PPA NBs increased with the increase of its concentration, confirming its concentration-dependent CAT-like activity. Further investigation of the effect of NIR light power on the activity revealed that the amount of O2 generated by PPA NBs increased with the increase of power (Fig. 2k), indicating that temperature rise can effectively enhance its enzymatic reaction kinetics and catalytic efficiency. Figure S12 further visually demonstrates the O2-producing capacity of PPA NBs. To quantitatively evaluate the catalytic performance of PPA NBs, the steady-state kinetics was further derived via the Michaelis-Menten equation (Figure S13). The maximum reaction velocities (Vmax) of the PPA NBs and the PPA NBs + NIR group were determined to be 3.40 and 8.03 mM min− 1, respectively. These results further indicated that increasing the temperature significantly enhances the CAT-like activity of the PPA NBs. Additionally, PPA NBs exhibits excellent cascade catalytic performance (Figure S14). After glucose addition, the Au component exhibits a GOx-like activity to produce H2O2 (accompanied by O2 consumption), while the Pd component regenerates O2 through CAT-like activity, forming a highly efficient catalytic cycle. Autonomous motion performance The dense and robust ECM characteristic of TNBC significantly impedes nanoparticle penetration into deep tumor regions, thereby limiting therapeutic efficacy [16]. Consequently, rationally designed nanorobots exhibiting autonomous motion hold promise for overcoming this barrier. However, nanorobots driven by a single gas drive face inherent limitations due to their dependence on specific enzyme activity and substrate availability, often proving inadequate for achieving both efficient self-propulsion and penetration through the formidable ECM. To address this challenge, the development of nanorobots endowed with robust dual-propulsion mechanisms is critical, enabling effective ECM penetration and access to deeper tumor tissue. The asymmetric Janus structure of PPA NBs imparted autonomous propulsion capability (Fig. 3a). The motion of the nanozyme/laser-driven PPA NBs was captured, and analyzed using ImageJ software. Representative trajectories of PPA NBs group and PPA NBs + NIR group at corresponding time points are depicted in Fig. 3b. Analysis revealed that the PPA NBs group’s movement had certain directionality and the trajectory was extended (Fig. 3c). This was due to the asymmetric release of O2, which provided the propulsive force of the nanorobots. Notably, the PPA NBs + NIR group exhibited significantly self-propelled with the typical directional trajectories (Fig. 3d). This amplification arises from the photothermal effect mediated by AuNDs and PDA. The resultant localized temperature gradient drives enhanced nanorobots diffusion via self-thermophoresis. To quantitatively assess motion capacity, mean-square displacement (MSD) analysis was performed. The MSD curve was derived from the two-dimensional (x, y) coordinate values extracted along the trajectory, as a function of the time interval (Δt). It was found that when the PPA NBs were exposed to NIR light, the MSD showed a significant parabolic increase over time, indicating that the NIR light enhanced the directionality of the movement of PPA NBs (Fig. 3e). The average speeds were further calculated to demonstrate the outstanding autonomous movement ability of the PPA NBs. The speed of the PPA NBs group was 4.98 ± 0.64 μm/s (Fig. 3f). Under the condition of NIR irradiation, the speed of the PPA NBs group was increased to 5.34 ± 1.21 μm/s. The above results indicate that NIR light effectively induces the directional movement of PPA NBs, and the speed and directionality of the nanorobots are both improved. In summary, the PPA NBs + NIR group can exhibit significant autonomous movement ability under the nanozyme/NIR dual-driven conditions, and is expected to penetrate deep tumor tissues. Fig. 3Autonomous motion ability assessment of PPA NBs. (a) The device and mechanism diagram of motion of PPA NBs. (b) Photos taken at different times of the movement of PPA NBs group and PPA NBs+NIR group in the hydrogen peroxide solution. Scale bar: 5 μm. (c)-(d) Trajectories of PPA NBs group and PPA NBs + NIR group. (e) The mean squared displacement and (f) the average speed of PPA NBs group and PPA NBs + NIR group Autonomous motion ability assessment of PPA NBs. (a) The device and mechanism diagram of motion of PPA NBs. (b) Photos taken at different times of the movement of PPA NBs group and PPA NBs+NIR group in the hydrogen peroxide solution. Scale bar: 5 μm. (c)-(d) Trajectories of PPA NBs group and PPA NBs + NIR group. (e) The mean squared displacement and (f) the average speed of PPA NBs group and PPA NBs + NIR group Enhanced tumor penetration and lysosomal escape To intuitively evaluate the deep tumor penetration capacity of PPA NBs, three-dimensional multicellular tumor spheroids (3D MTSs) derived from murine mammary carcinoma 4T1 cells were established (Fig. 4a). In the absence of any driving force, the red fluorescence of the PDA group was confined solely to the periphery of tumor spheroids. By comparison, the red fluorescence signals of the PPA NBs group and the PDA + NIR group were distributed relatively closer to the spheroid centers, which was attributed to nanozyme-driven diffusion and NIR light-driven diffusion, respectively. Notably, the PPA NBs + NIR group exhibited the most extensive and highest-density red fluorescence signals even in the core regions of the 3D MTSs. This observation directly demonstrates that the nanorobots successfully reached the central parts of the tumors, a phenomenon ascribed to the nanozyme/NIR dual-driven diffusion behavior of the nanorobots within the tumor spheroids. Figure 4b and c show the penetration percentage calculated based on the fluorescence intensity ratio between the internal and peripheral regions of the tumor spheroids. The penetration efficiency of PPA NBs + NIR was 72 ± 3.0%, which was significantly higher than those of all other groups. Collectively, the results demonstrate that the robust dual-propulsion capability of PPA NBs + NIR enables deep penetration through the dense tumor barrier of TNBC, holding great promise for achieving effective therapeutic outcomes. Once internalized by cells and subsequently engulfed by lysosomes, nanomedicines are at risk of degradation and inactivation, underscoring that lysosomal escape is pivotal for nanorobots to exert enzyme-like functions within tumor cells. Therefore, we detected the fluorescence co-localization between Cy5-labeled nanoparticles (red) and Lyso-Tracker-stained lysosomes (blue) using a confocal laser scanning microscope (CLSM), so as to evaluate the lysosomal escape capacity of PPA NBs in 4T1 cells (Fig. 4d). In the CLSM images, the PDA group showed extensive overlap between red and blue fluorescence signals, indicating that most nanoparticles were confined within lysosomes. In contrast, the PPA NBs + NIR group exhibited a clear separation of red and blue fluorescence, suggesting effective lysosomal escape. The Pearson’s correlation coefficients were further quantitatively analyzed to confirm these observations (Fig. 4e). The PDA group exhibited the highest co-localization coefficient (0.71 ± 0.04), while the PPA NBs group showed a slightly reduced value (0.64 ± 0.04), attributed to the O2 generated by CAT-like catalysis of PdNPs providing partial propulsive force for lysosomal membrane disruption. Under NIR irradiation, the co-localization coefficients decreased markedly, the PPA NBs + NIR group demonstrated the lowest co-localization coefficient (0.23 ± 0.02), significantly lower than all other groups. This is ascribed to the synergistic dual-propulsion mechanism, where NIR-induced self-thermophoresis and nanozyme-driven O2 generation collaboratively disrupt the lysosomal membrane, enabling the nanorobots to efficiently escape into the cytoplasm and exert their enzymatic functions. Fig. 4Deep tumor penetration and lysosomal escape of PPA NBs. (a) CLSM images of 3D 4T1 MTSs after being incubated with nanoparticles (n = 3, scale bar: 500 μm). (b) Penetration rates at 80 μm cross-section of different groups (n = 3). (c) Fluorescence area of tumor spheres at 80 μm cross-section of different group (n = 3). (d) The co-localization images of Cy5-labeled nanoparticles (red) and Lyso-Tracker-stained lysosomes (blue) in 4T1 cells treated with PDA, PPA NBs, PDA + NIR and PPA NBs + NIR (n = 3, scale bar: 5 μm). (e) Pearson’s correlation coefficient analysis of the co-localization between nanoparticles and lysosomes (n = 3). Data presented as mean ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 Deep tumor penetration and lysosomal escape of PPA NBs. (a) CLSM images of 3D 4T1 MTSs after being incubated with nanoparticles (n = 3, scale bar: 500 μm). (b) Penetration rates at 80 μm cross-section of different groups (n = 3). (c) Fluorescence area of tumor spheres at 80 μm cross-section of different group (n = 3). (d) The co-localization images of Cy5-labeled nanoparticles (red) and Lyso-Tracker-stained lysosomes (blue) in 4T1 cells treated with PDA, PPA NBs, PDA + NIR and PPA NBs + NIR (n = 3, scale bar: 5 μm). (e) Pearson’s correlation coefficient analysis of the co-localization between nanoparticles and lysosomes (n = 3). Data presented as mean ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 Anti-TNBC effects in vitro First, a systematic validation of the biocompatibility of PPA NBs was performed to evaluate their potential for biomedical applications. In the hemolytic assays, the supernatants from all groups were colorless even at the maximum tested concentration of 400 µg/mL, demonstrating their excellent blood compatibility (Fig. 5a-b, Figure S15). This finding is consistent with the results of the cell counting kit-8 (CCK-8) assay performed on the normal mammary epithelial cell line MCF-10 A, which fully confirms the favorable biosafety profile of the nanoparticles (Fig. 5c). Then, the in vitro cytotoxicity of PPA NBs against 4T1 cells was evaluated using the CCK-8 assay (Fig. 5d). In the absence of NIR irradiation, the PDA group exhibited only a slight decrease in cell viability at the maximum concentration of 400 µg/mL, with the cell survival rate reaching 85.51 ± 2.6%. The PPA NBs group exhibited a similar trend, whereas the cell survival rate dropped to 59.80 ± 0.85% at 400 µg/mL, which was significantly lower than that of the PDA group. Further, when exposed to NIR laser, PDA and PPA NBs exhibited a clear concentration-dependent cytotoxicity. At concentrations ≥ 100 µg/mL, cell viability decreased significantly in PPA NBs + NIR group, falling to 32.12 ± 0.98% at 200 µg/mL, and 11.1 ± 1.63% at 400 µg/mL. Although the PDA group also exhibited an analogous trend upon NIR irradiation, its cell survival rate remained at 48 ± 2.32% at a concentration of 200 µg/mL, and plummeted to 23.60 ± 1.74% at 400 µg/mL. Subsequently, we visually demonstrated the cytotoxicity of PPA NBs combined with NIR via live-dead cell staining and colony formation assays. As shown in Fig. 5e and f, the PPA NBs + NIR group induced the highest level of 4T1 cell death compared with all other groups. These results demonstrate that PPA NBs exert a potent synergistic anti-TNBC effect in combination with NIR irradiation. Fig. 5Biocompatibility and cytotoxicity in vitro. (a) Photograph of hemolysis test and hemolysis ratio of the PPA NBs (n = 3). (b) Hemolysis ratio of the PDA (n = 3). (c) Relative cell viability of MCF-10 A cells across different treatment groups. (d) Relative cell viability of 4T1 cells across different treatment groups. (e) Co-staining of 4T1 cells with calcein-AM (live cells, green) and propidium iodide (dead cells, red) after different treatments. Scale bar = 100 μm (f) The clonal formation assay of 4T1 cells under different treatment conditions. Data presented as mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 Biocompatibility and cytotoxicity in vitro. (a) Photograph of hemolysis test and hemolysis ratio of the PPA NBs (n = 3). (b) Hemolysis ratio of the PDA (n = 3). (c) Relative cell viability of MCF-10 A cells across different treatment groups. (d) Relative cell viability of 4T1 cells across different treatment groups. (e) Co-staining of 4T1 cells with calcein-AM (live cells, green) and propidium iodide (dead cells, red) after different treatments. Scale bar = 100 μm (f) The clonal formation assay of 4T1 cells under different treatment conditions. Data presented as mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 Tumor visualization monitoring and photothermal ability To evaluate the visualized tumor monitoring capability of the designed nanorobots, PPA NBs were intravenously injected via the tail vein of 4T1 tumor-bearing mice. The tumor tissues as well as major organs (heart, liver, spleen, lung and kidney) in each group were harvested at 4 h, 12 h, and 24 h post-injection for ex vivo fluorescence imaging. As illustrated in Fig. 6a and b, PPA NBs showed peak accumulation at the tumor site at 12 h post-administration, which was higher than that observed at 4 h post-administration. More importantly, the intense fluorescence signal at the tumor site clearly demonstrated that PPA NBs possess excellent precise tumor monitoring capability through fluorescence imaging, enabling accurate localization and real-time tracking of tumor lesions. Based on the above results, we selected 12 h post-injection as the time point for NIR irradiation to verify the photothermal performance of PPA NBs, considering the highest accumulation efficiency of nanorobots at the tumor site at this time point. The tumor sites were irradiated (0.5 W/cm2, 5 min) 12 h post-injection, and temperature variation of the tumor site was monitored. Infrared thermal imaging and quantitative analysis showed that the PPA NBs + NIR group reached 49.3 ℃ at the tumor site with a temperature increase of 13.13 ℃, significantly higher than the PBS + NIR group (39.2 ℃, < 5 ℃), demonstrating the superior photothermal performance of PPA NBs in vivo (Fig. 6c-d). Figure 6e presents the confocal imaging of the isolated tumor in mice subjected to different treatments, where the red fluorescence signals correspond to the nanoparticles. In the absence of any driving force, PDA was only able to reach the tumor periphery. In comparison, either nanozyme-driven or NIR-driven single modality slightly facilitated the infiltration of nanoparticles toward the tumor center. Most notably, the PPA NBs + NIR group achieved deep penetration to the core region of the tumor. These findings strongly confirm that the dual driving force of nanozyme and NIR endows the nanorobots with an excellent capacity for deep tumor penetration. Fig. 6Biodistribution and photothermal ability of PPA NBs. (a) Fluorescence imaging of PPA NBs in isolated organs and tumor tissues of 4T1 tumor-bearing mice at different time points (n = 3). (b) Fluorescence quantification of PPA NBs distribution in various organs and tumor tissues. (c) The temperature changes at the tumor sites during local NIR irradiation corresponding to the photothermal imaging of mice in (d). (d) Infrared thermal images of tumors in mice with different treatments (n = 3). (e) CLSM images of tumors in mice with different treatments, scale bar = 1 mm (n = 3) Biodistribution and photothermal ability of PPA NBs. (a) Fluorescence imaging of PPA NBs in isolated organs and tumor tissues of 4T1 tumor-bearing mice at different time points (n = 3). (b) Fluorescence quantification of PPA NBs distribution in various organs and tumor tissues. (c) The temperature changes at the tumor sites during local NIR irradiation corresponding to the photothermal imaging of mice in (d). (d) Infrared thermal images of tumors in mice with different treatments (n = 3). (e) CLSM images of tumors in mice with different treatments, scale bar = 1 mm (n = 3) Anti-TNBC effects in vivo Inspired by these encouraging findings, the antitumor effects of PPA NBs combined with NIR were evaluated in 4T1 tumor bearing mice. Previous studies have proposed that the overexpression of HSPs during hyperthermic reactions procedures attenuates the antitumor efficacy of PTT by promoting the correct folding of nascent proteins, restoring the native conformations of denatured proteins, clearing cytotoxic protein aggregates, and modulating immune responses [37, 39]. To further decipher the mechanism behind this synergistic therapeutic action, immunofluorescence staining was performed to evaluate the expression levels of HIF-1α and HSP 70 in tumor tissues across different treatment groups. As shown in Figure S16, the PDA + NIR group exhibited the highest HIF-1α expression among all groups, which is attributed to the photothermal-induced vascular damage and subsequent exacerbation of intratumoral hypoxia. Notably, the PPA NBs + NIR group showed significantly lower HIF-1α expression than the PDA + NIR group, indicating the CAT-like activity of Pd effectively catalyzed the decomposition of H2O2 to generate abundant O2, thereby alleviating intratumoral hypoxia. Furthermore, the expression of HSP 70 was evaluated (Figure S17). The PDA + NIR group exhibited the highest HSP 70 expression, as the NIR-induced hyperthermia strongly stimulated the upregulation of HSPs, while PDA alone lacks the capacity for glucose depletion to counteract this response. The PPA NBs + NIR group showed reduced HSP 70 expression compared with the PDA + NIR group. Although NIR irradiation similarly induced local hyperthermia that upregulated HSPs expression, the GOx-like activity of AuNDs partially depleted intracellular glucose, thereby attenuating the energy supply required for HSPs synthesis. These results collectively demonstrate that the cascade catalytic activity of PPA NBs synergizes with PTT by simultaneously alleviating tumor hypoxia and suppressing the self-protective heat shock response, thereby significantly sensitizing TNBC to photothermal therapy. Given that HSPs have been confirmed to impair the sensitivity of tumor cells to PTT, we subsequently evaluated the sensitizing effect of nanorobots on PTT via tumor inhibition assays. When tumor volumes reached 50–100 mm3, mice were treated with PBS, PDA, PPA NBs, PDA + NIR or PPA NBs + NIR. 12 h post-administration, all mice in the NIR groups were subjected to NIR irradiation (Fig. 7a). By day 14, the tumor volume in the PBS group reached 1224.78 mm3. Both PPA NBs and PDA + NIR suppressed tumor growth to different degrees, yielding tumor suppression rates (TSRs) of 28.5% and 75.02%, respectively (Fig. 7b). Remarkably, PPA NBs + NIR exerted outstanding anti‑tumor potency, attaining a TSR of 97%, and significantly reduced tumor size compared to other groups on day 14 (Fig. 7b-c). Bliss analysis showed that the combination index (CI) was 0.168, quantitatively confirming a synergistic interaction between nanocatalytic therapy and PTT. The tumor weights at the end of treatment were consistent with the tumor growth curves, with PPA NBs + NIR group showing significantly lower tumor weight than the PDA + NIR group (Fig. 7d). Fig. 7In vivo antitumor efficacy study. (a) Schematic diagram depicting the experimental workflow of the 4T1 tumor model. (b) Tumor progression in 4T1 tumor‑bearing mice following diverse therapeutic interventions. (c) Tumor images collected from mice with different treatment groups (G1: PBS; G2: PDA; G3: PPA NBs; G4: PDA + NIR; G5: PPA NBs + NIR). (d) Average tumor weight in each group at the end of the study. (e) Body‑weight variations throughout the therapeutic intervention period. (f) Individual tumor‑growth curves of 4T1 tumors under diverse therapeutic regimens. (g) H&E staining of tumors in each group after treatment, scale bar = 100 μm. Data presented as mean ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 In vivo antitumor efficacy study. (a) Schematic diagram depicting the experimental workflow of the 4T1 tumor model. (b) Tumor progression in 4T1 tumor‑bearing mice following diverse therapeutic interventions. (c) Tumor images collected from mice with different treatment groups (G1: PBS; G2: PDA; G3: PPA NBs; G4: PDA + NIR; G5: PPA NBs + NIR). (d) Average tumor weight in each group at the end of the study. (e) Body‑weight variations throughout the therapeutic intervention period. (f) Individual tumor‑growth curves of 4T1 tumors under diverse therapeutic regimens. (g) H&E staining of tumors in each group after treatment, scale bar = 100 μm. Data presented as mean ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 No obvious decline in body weight was detected across all cohorts over the course of therapy, demonstrating the favorable biocompatibility of these therapeutic modalities (Fig. 7e). Individual tumor growth curves for each mouse in different treatment groups are shown in Fig. 7f. Hematoxylin and eosin (H&E) staining revealed that PPA NBs + NIR induced apoptosis in 4T1 tumor tissues without causing significant damage to major organs (heart, liver, spleen, lung, and kidney; Figs. 7g and 8a). No significant treatment-related histopathological abnormality, such as inflammatory cell infiltration, necrosis, fibrosis or structural disorganization, was observed in the major organs of any group. Routine hematological assays performed at day 14 revealed negligible inter‑group discrepancies, verifying that these therapeutic interventions do not impair hematological safety (Fig. 8b). Collectively, these observations reveal that PPA NBs combined with PTT exerts a highly efficient inhibitory effect on the growth of TNBC with a satisfactory safety profile. To further verify the antitumor mechanism at the histological level, TUNEL and Ki67 staining were performed on tumor sections from each group. The PPA NBs + NIR group exhibited the strongest TUNEL-positive green fluorescence among all groups (Figure S18), indicating the most extensive apoptosis, whereas the Ki67-positive red fluorescence of this group was the weakest (Figure S19), indicating the most effective inhibition of tumor cell proliferation. These results are consistent with the tumor growth curves and tumor weights described above, and jointly confirm that PPA NBs combined with PTT simultaneously promote tumor cell apoptosis and suppress tumor cell proliferation. Fig. 8Security evaluation in vivo. (a) H&E-stained histological slices of major visceral organs (heart, liver, spleen, kidney, lung) derived from mice receiving distinct therapeutic regimens, scale bar = 200 μm. (b) Hematological indices acquired from mice at 14‑day post‑treatment. Data presented as mean ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 Security evaluation in vivo. (a) H&E-stained histological slices of major visceral organs (heart, liver, spleen, kidney, lung) derived from mice receiving distinct therapeutic regimens, scale bar = 200 μm. (b) Hematological indices acquired from mice at 14‑day post‑treatment. Data presented as mean ± SD. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 Conclusion In this study, we successfully synthesized the visualized dual-propelled Janus PPA NBs for tumor visualization monitoring and photothermal-nanocatalytic therapy against TNBC. The asymmetric structure enables the PPA NBs to undergo self-thermophoretic motion upon NIR light irradiation, endowing them with autonomous navigation capability. The nanozyme activity provides the nanorobots with a second driving force for locomotion. Propelled by dual driving forces of light/nanozyme, PPA NBs autonomously navigate and penetrate deep into the tumor, undergo lysosomal escape within tumor cells, and subsequently exert their enzymatic functions. The GOx-like activity mediates extensive glucose catabolism, which induces energy deprivation in tumor cells. This dual effect not only directly sensitizes tumor cells to PTT, but also attenuates the cellular repair ability via the downregulation of HSPs expression. The H2O2 generated during the aforementioned process further provides a more abundant substrate for CAT-like activity, leading to the production of a large amount of O2 and alleviation of tumor hypoxia. Meanwhile, the local hyperthermia of tumors induced by PTT further enhances the catalytic efficiency of nanozymes. Through the interplay of the above-mentioned synergistic effects, the integrated photothermal-nanocatalytic therapy exerts a highly efficient inhibitory effect on TNBC. Furthermore, the AuNDs component integrated into PPA NBs endows the nanorobots with stable fluorescence imaging functionality, which enables precise tumor visualization monitoring-facilitating unambiguous localization of tumor lesions, dynamic tracking of nanorobot distribution within tumors, and real-time monitoring of therapeutic efficacy of the tumor. Collectively, the present work offers a promising theranostic strategy that integrates efficient therapy, specific biosensing, and precise tumor visualization monitoring, providing a comprehensive solution for the clinical management of TNBC. Methods Preparation of mesoporous PDA 0.10 g F127 and 0.15 g dopamine hydrochloride were weighed into a 20 mL glass vial, followed by addition of 5 mL water, 5 mL ethanol and 0.3 mL TMB. After ultrasonication to form a milky emulsion, 0.4 mL ammonia solution was added dropwise under magnetic stirring. The reaction proceeded at 25 °C for 2 h with continuous stirring. The precipitated products were harvested via centrifugation, rinsed thoroughly with anhydrous ethanol, and finally redispersed in 1 mL of anhydrous ethanol for subsequent experiments. Preparation of PDAPd Add 3 mL of the above-prepared PDA solution and 50 mM Na2PdCl4 solution to 2 mL of water, and stir and heat until it reaches 90 ℃. Add 1 mL of ascorbic acid solution with a concentration of 50 mg/mL to the system, and then dropwise add 1 mL of PVP solution (10 mg/mL), and continue to stir for 30 min. Centrifuge at 9800 rpm (15 min) to collect the precipitate, disperse the precipitate in water, and obtain the PDAPd solution. Preparation of AuNDs Lysozyme-stabilized AuNDs were prepared with a modified aqueous synthesis method. Briefly, 1 mL of 4 mM HAuCl4 aqueous solution was added to 2 mL of 5 mg·mL− 1 lysozyme solution under vigorous room-temperature stirring. After 5 min incubation, 100 µL of 1 M NaOH solution was added dropwise, and the mixture was incubated overnight at 37 °C. The crude product was dialyzed against ultrapure water with 12 kDa MWCO dialysis membrane. The purified AuNDs were obtained and stored at 4 °C for follow-up experiments. Preparation of PPA NBs In 2 mL of water, added 1 mL of the prepared PDAPd and 2 mL of AuNDs. Under room temperature conditions, continuously stirred the reaction for 5 h. Then, the un-loaded AuNDs were removed by centrifugation and obtain PPA NBs. Cell Lines Both 4T1 and MCF-10 A cell lines were maintained in RPMI 1640 supplemented with 10% fetal bovine serum (FBS) and 1% (v/v) penicillin-streptomycin. Cells were maintained at 37 °C under humidified conditions with 5% CO2 and 95% air (approximately 20% O2). Penetration of PPA NBs in 3D MTSs 4T1 cells were plated in ultralow-attachment 96‑well plates and cultured over 48 h to generate tightly packed tumor spheroids. Uniform‑sized 3D MTSs were subsequently picked out and exposed to Cy5-labeled PDA (50 µg/mL) or PPA NBs (50 µg/mL) for 12 h. Following PBS washes, the spheroids were irradiated with NIR light for 5 min (0.5 W/cm2). Tumor sphere images were captured using CLSM, and the penetration of nanoparticles was quantified with ImageJ. Lysosome localization test Following 2 h of attachment after seeding 4T1 cells in confocal dishes, Cy5-labeled PDA (50 µg/mL) or PPA NBs (50 µg/mL) were supplemented for 12 h. After PBS rinsing to wash away residual nanoparticles, cells underwent 30 min LysoTracker (blue) staining, and fluorescence signals were acquired via CLSM. Hemolytic assays Erythrocytes were isolated from whole blood collected from BALB/c mice. Red blood cell suspensions were incubated with serially diluted PPA NBs or PDA solutions at a 1:1 volume ratio, followed by 1 h incubation at 37 °C. Normal saline and deionized water were set as negative and positive controls. Following centrifugation, the absorbance of supernatants was recorded at 540 nm. The hemolysis ratio (%) was calculated as: Hemolysis ratio (%) = (Asample -Anegative)/(Apositive -Anegative) × 100%, where A represents the absorbance value of each experimental group at 540 nm. CCK-8 assay 96‑well plates were plated with MCF‑10 A cells at 1 × 104 cells/well. The cells were exposed to different concentrations of PDA or PPA NBs in RPMI 1640 (10% FBS) and incubated at 37 °C with 5% CO2 over 24 h. Post‑incubation, each well received 100 µL complete medium plus 10 µL CCK‑8 solution, followed by 1 h incubation. Optical density at 450 nm was acquired via a microplate reader (Tecan Austria GmbH) to evaluate cell viability. 96‑well plates were plated with 4T1 cells at 1 × 104 cells/well, which were cultured for 24 h prior to medium exchange. Fresh medium containing serially diluted PDA or PPA NBs was added, and cells were maintained at 37 °C for 2 h. Cells assigned to the NIR‑treated groups were irradiated with an 808 nm laser (0.5 W/cm2) for 5 min. After an additional 24 h of incubation, cell viability was determined using the CCK‑8 assay. Live-dead cell staining 4T1 cells were plated into well plates and cultured for 24 h to achieve complete cell adhesion. After that, the cells were separately incubated with PDA and PPA NBs for 2 h. Following the co-cultivation process, the cell samples were irradiated with 808 nm NIR light at a power density of 0.5 W/cm2 for 5 min, and continuously cultured for another 2 h. Upon the completion of incubation, the original culture medium was discarded. A pre-configured mixed staining solution containing propidium iodide (PI) for dead cell labeling and Calcein-AM for live cell staining was supplemented to the plates, and the cells were incubated at 37 °C for 30 min. Ultimately, the stained cell specimens were visualized and analyzed via a Zeiss confocal laser scanning microscope (CLSM). Colony formation assays 4T1 cells were seeded into 96-well plates. The cells were then incubated with 400 µg/mL PDA or PPA NBs for 2 h, followed by NIR irradiation (808 nm, 0.5 W/cm2, 5 min). After treatment, the cells were cultured for 2 weeks, and cell colonies were imaged. Biodistribution analysis and ex vivo imaging Subcutaneous inoculation of 4T1 tumor cells into female BALB/c mice was implemented to construct the TNBC tumor model. Once tumor volumes reached approximately 50 mm3, PPA NBs were administered via the intravenous route. To perform ex vivo fluorescence imaging, key visceral organs (heart, liver, spleen, lung, kidney) together with tumor tissues were dissected and collected at 4 h, 12 h and 24 h following nanoparticle administration. Tissue specimens were then subjected to signal acquisition utilizing an IVIS Lumina LT Series III imaging platform. Infrared thermal imaging 4T1 tumor-bearing mice were treated with PBS or PPA NBs (5 mg/kg) intravenously. The tumor sites were irradiated (0.5 W/cm2, 5 min) at 12 h post-administration, while temperature fluctuations were documented using an IR thermal camera. Anti-tumor experiment Tumors were allowed to grow to a volume of 50 mm3, after which the experimental animals were randomly allocated into five independent groups with five mice in each subgroup. Thereafter, the mice received tail‑vein intravenous administration of 100 µL PBS, PDA (5 mg/kg), or PPA NBs (5 mg/kg), respectively. After 12 h, mice in the corresponding groups were subjected to 808 nm NIR irradiation (0.5 W/cm2) for 5 min. Tumor size (length and width) and body weight were monitored and recorded every 2 days. Tumor volume was computed using the formula: volume (mm3) = 1/2 × length × width2. All animals were euthanized 14 days post-administration. Key visceral organs (heart, liver, spleen, lung, kidney) together with tumor tissues were excised and preserved for subsequent H&E staining. Therapeutic synergy was evaluated on day 14 using the Bliss independence model [40]. Relative tumor burden was defined as SX = VX/VPBS, and the combination index was calculated as CIBliss = S(PPA NBs + NIR)/[SPPA NBs × S(PDA + NIR)]. CI values below, equal to, or above 1 indicate synergistic, additive, or antagonistic interactions, respectively. Immunofluorescence staining Tissue sections were deparaffinized and blocked with 4% bovine serum albumin (BSA), followed by incubation with primary antibodies or fluorophore-conjugated primary antibodies. For primary antibodies without fluorescent conjugation, sections were further incubated with the corresponding secondary antibodies. For immunofluorescence staining, tumor tissue sections were incubated with primary antibodies against HIF-1α or HSP 70, followed by Cy3-conjugated secondary antibody. For TUNEL and Ki67 staining, two serial sections were prepared from each tumor and stained separately: one section was treated with a FITC-labelled TUNEL assay kit to visualize apoptotic cells (green), and the other was incubated with an anti-Ki67 primary antibody followed by a Cy3-conjugated secondary antibody (red). Cell nuclei of both sections were counterstained with DAPI (blue). All stained sections were observed under a CLSM. H&E staining Resected tissue specimens underwent overnight fixation with 4% buffered paraformaldehyde prior to paraffin embedding. The paraffin‑embedded tissues were sliced into 5 μm‑thick slices, which were further deparaffinized and processed for hematoxylin‑eosin (H&E) staining. Histomorphological features of tissue sections were evaluated using a Nikon TE2000U optical microscope. Safety analysis At the 14‑day experimental endpoint, peripheral blood specimens were harvested from mice across all experimental groups for complete blood count (CBC) assays. Multiple hematological parameters, namely WBC, RBC, PLT, LYM, MCV and MCHC, were measured to assess possible myelosuppression induced by the therapeutic interventions. Statistical analysis All in vitro experiments were performed independently, and each animal model represented a distinct biological replicate. Quantitative data are expressed as mean ± standard deviation (SD). Statistical evaluation was carried out using two‑tailed unpaired Student’s t‑test and one‑way analysis of variance (ANOVA). Statistical significance was defined as: *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1. Supplementary Material 1.