Composite nano copper carrier combining cuproptosis and photodynamic therapy for spatiotemporal synergistic anti-tumor therapy

Abstract

Cuproptosis is a copper-dependent programmed cell death triggered by mitochondrial dysfunction, which offers significant anti-tumor potential but requires tumor-specific copper delivery to avoid systemic toxicity. Here, we developed a synergistic nanoplatform (Cu₂O@SiO₂-Ce6, CSC) integrating cuproptosis induction with photodynamic therapy (PDT). A cuprous oxide (Cu₂O) core was encapsulated in silicon dioxide and covalently linked to the photosensitizer Ce6. After accumulating in tumor regions via the EPR effect, CSC releases a surge of copper ions upon encountering the acidic tumor microenvironment and lysosomes. These ions induce mitochondrial cuproptosis by aggregating lipoylated TCA cycle proteins and depleting Fe–S cluster proteins. Concurrently, laser irradiation activates Ce6 to generate cytotoxic reactive oxygen species, actively disrupting tumor membranes. The experiments showed that the nanoplatform significantly enhanced the tumor cell killing effect and reduced the toxicity to normal tissues through a dual mechanism, which provided a new strategy for overcoming the drug resistance of PDT and the precise regulation of cuproptosis.

---

Introduction

Programmed cell death such as apoptosis, necrosis and ferroptosis serves as a precise and effective therapeutic strategy for eliminating tumor cells during cancer treatment.^1,2 Each programmed cell death pathway exhibits distinct biological features that can be exploited for targeted treatment. For instance, apoptotic tumor cells can induce a bystander effect, sensitizing adjacent cells to chemotherapeutic agents.^3,4 Pyroptotic cells undergo rapid release of tumor-associated antigens, potentiating tumor-specific immune responses.^5,6 Cuproptosis is a newly identified form of programmed cell death triggered by mitochondrial dysfunction.⁷ In humans, copper serves as an essential trace element, with balanced copper homeostasis regulating numerous biochemical reactions to maintain normal physiological functions.^8,9 However, upon entering mitochondria, excess copper binds to ferredoxin 1 (FDX1) as a substrate, undergoing reduction to cuprous ions (Cu⁺) before being released.¹⁰ FDX1 and lipoic acid synthetase (LIAS) can catalyze lipoylated dihydrolipoamide S-acetyltransferase (DLAT) in the mitochondria. On the one hand, Cu⁺ can bind to DLAT, the “switch” that allows pyruvate to enter the mitochondria, which become dysfunctional, resulting in severe inhibition of the entire tricarboxylic acid (TCA) cycle due to the lack of substrate. On the other hand, it can reduce the synthesis of the Fe–S cluster, which plays various roles in mitochondria, including as cofactors for respiratory enzyme complexes (I–III), involvement in lipoic acid synthesis, and biosynthesis of heme.¹¹ The combined action of both mechanisms ultimately leads to cell death. This unique programmed cell death mechanism can further elicit robust immune responses, not only killing tumor cells but also activating immune cells for long-term eradication of residual cancer cells.^12,13 Notably, cuproptosis may occur in both normal and malignant cells, creating an urgent need to develop strategies that enhance drug accumulation specifically in tumor tissues while minimizing damage to healthy tissues.

Nano copper carriers can significantly enhance tumor-specific copper accumulation.¹⁴ Representative copper-containing nanoplatforms, such as those with copper sulfide (CuS) and cuprous oxide (Cu₂O),^15–20 serve as excellent delivery vehicles that not only carry substantial amounts of copper but also effectively mask its ionic toxicity through their stable nanostructures. The aberrant vascular architecture of tumor tissue enables such nanoplatforms (typically 10–200 nm in size) to extravasate through the vessel walls and accumulate preferentially in tumor sites via the enhanced permeability and retention (EPR) effect.^21,22 Under the acidic stimulation of the tumor microenvironment and lysosomes, these nanostructures undergo degradation, releasing a surge of copper ions that rapidly elevate intracellular concentrations, thereby inducing acute cuproptosis. Copper-based nanoplatforms represent a highly promising cuproptosis-mediated antitumor strategy by precisely disrupting copper ion homeostasis to induce tumor-specific cuproptosis.^23–25 Notably, these nanoplatforms exhibit versatile physicochemical properties, including adjustable nanostructures and high chemical modifiability. Through rational integration of therapeutic agents (e.g., chemotherapeutics, photosensitizers, and immune adjuvants), they can synergize cuproptosis with chemotherapy, photodynamic therapy (PDT), and immunotherapy, thereby significantly enhancing antitumor efficacy.

PDT is an anticancer modality that employs external light activation of photosensitizers to generate cytotoxic reactive oxygen species (ROS).^26–32 Chlorin e6 (Ce6), a widely used photosensitizer in PDT, demonstrates favorable therapeutic characteristics including reduced phototoxicity and high-efficiency singlet oxygen (¹O₂) generation under red-light irradiation, making it clinically valuable for oncological applications.^33–36 The therapeutic efficacy of PDT stems from direct ROS-mediated cytotoxicity. However, the clinical application of PDT faces significant challenges, particularly the development of resistance mechanisms that limit its long-term effectiveness. Tumor cells can develop PDT resistance through multiple pathways, including upregulation of antioxidant systems (e.g., activation of hypoxia-inducible factor-1α (HIF-1α) under low-oxygen conditions)³⁷ and enhanced DNA repair mechanisms.³⁸ To overcome these limitations, innovative strategies such as nanoparticle-based oxygen carriers,³⁹ ROS-amplifying agents,⁴⁰ and combination therapies with immune checkpoint inhibitors⁴¹ have been explored. Notably, the integration of PDT with cuproptosis—a novel copper-dependent cell death pathway—has shown promise in bypassing traditional resistance mechanisms by inducing mitochondrial protein aggregation and proteotoxic stress. The strategic combination of PDT with cuproptosis induction creates a dual “active-passive” therapeutic paradigm that synergistically enhances antitumor efficacy through complementary mechanisms. As demonstrated in Fig. 1, we developed a cuprous oxide (Cu₂O) nanoplatform-based theranostic probe (denoted as Cu₂O@SiO₂-Ce6, CSC) that synergistically integrates cuproptosis induction with photodynamic therapy. Specifically, we engineered a silica shell-coated cuprous oxide nanoplatform (Cu₂O@SiO₂), with the photosensitizer Ce6 conjugated to the shell surface via amide bonds. The CSC nanoplatform preferentially accumulates at tumor sites through the EPR effect. Upon external laser irradiation, Ce6 generates abundant ROS that disrupt tumor cell membrane integrity. Simultaneously, under the acidic stimulation of the tumor microenvironment and lysosomes, the Cu₂O nanoplatform undergoes structural degradation, releasing abundant copper ions. These ions subsequently enter mitochondria and bind to lipoylated proteins in the TCA cycle, inducing protein aggregation. This process leads to iron–sulfur cluster protein (Fe–S protein) depletion, triggering proteotoxic stress and ultimately resulting in cell death, thereby achieving synergistic tumor treatment through combined PDT and cuproptosis. Both in vitro and in vivo experimental results demonstrate that this integrated nanotherapeutic platform effectively inhibits tumor cell growth while maintaining excellent biosafety in normal tissues. This work provides a significant conceptual advance for multimodal synergistic cancer therapy.

Fig. 1 Synthesis and antitumor mechanism of CSC. (A) Schematic illustration of synthesis of CSC and (B) its application for photodynamic-enhanced cuproptosis-apoptosis synergistic tumor therapy.

Results and discussion

Synthesis and characterization of CSC

In this paper, we prepared the Cu₂O nanoplatform by reduction of copper hydroxide with ascorbic acid using polyvinylpyrrolidone (PVP) as surfactant.^42,43 Cu₂O is inherently unstable and prone to oxidation, but SiO₂ coating not only significantly enhances its stability⁴⁴ but also facilitates further modification.⁴⁵ Functionalization with mPEG-silane and APTES (γ-aminopropyltriethoxysilane) improves nanoparticle stability, reduces off-target toxicity, and promotes tumor accumulation through the EPR effect,⁴⁶ while providing anchoring sites for Ce6.⁴⁷ TEOS was hydrolyzed under alkaline environment to form SiO₂, which was uniformly coated on the surface of Cu₂O particles (Cu₂O@SiO₂, CS), and after surface modification with APTES, the surface of CS carries a large number of amino groups (Cu₂O@SiO₂-NH₂, CS-NH₂). The carboxyl groups in the photosensitizer Ce6 were covalently attached to the aminated silica surface (Cu₂O@SiO₂-Ce6, CSC) via amide bonds. To verify the successful preparation of the nanoparticles, we first characterized them using transmission electron microscopy (TEM). As shown in Fig. 2A, the TEM images demonstrate that Cu₂O, CS, and CSC are uniformly dispersed with homogeneous particle sizes. The Cu₂O particles exhibit a cubic morphology with an approximate size of 80 nm. Both CS and CSC nanoplatforms are around 100 nm in diameter and display a distinct core–shell structure, where the Cu₂O core is enveloped by an approximately 10 nm-thick outer SiO₂ shell. This confirms the successful coating of silica on the Cu₂O surface without altering its morphology or size. Compared with CS, CSC conjugated with Ce6 shows no significant changes in either morphology or particle size. These results indicate the successful preparation of Cu₂O, CS, and CSC nanoplatforms. Cu₂O exhibits pH-responsive characteristics, being prone to degradation under acidic conditions (pH = 6.5) while remaining relatively stable in a neutral environment (pH = 7.4).

Fig. 2 Characterization of the physicochemical properties of the probes. (A) TEM images of Cu₂O, CS and CSC. Scale bar: 100 nm. (B) TEM images of CSC at pH = 7.4 and pH = 6.5. Scale bar: 100 nm. (C) Element mapping image of CSC. (D) Hydrated particle sizes and (E) zeta potentials of Cu₂O, CS, CS-NH₂ and CSC. Data are presented as the mean ± SD (n = 3 independent samples). (F) UV-visible absorption spectra of Cu₂O, CS, CS-NH₂ and CSC. (G) XRD pattern of CSC. (H) Energy dispersive spectra (EDS) of CSC. (I) Copper ion release from CSC in PBS at different pH (n = 3 independent experiments).

To investigate whether CSC retains its pH-responsive properties after silica coating, we dispersed CSC in buffer solutions at different pH values (7.4 and 6.5) for 24 hours and subsequently characterized their morphological changes using TEM. As illustrated in Fig. 2B, CSC maintained its original morphology and size at physiological pH (7.4). In contrast, under acidic conditions (pH 6.5), only the approximately 10 nm-thick SiO₂ shell remained observable, indicating that coating did not prevent the degradation of the Cu₂O core under acidic conditions. SiO₂ exhibits minimal dissolution under weakly acidic conditions.⁴⁸ This ensures controlled Cu²⁺ release primarily through pH-triggered degradation of the Cu₂O core (not SiO₂ dissolution). These results confirm that CSC maintains its pH-responsive behavior even after SiO₂ encapsulation, confirming that the SiO₂ coating did not interfere with the pH-dependent degradation of the Cu₂O core. These findings verify that the pH-responsive property of Cu₂O was preserved in the CSC nanoplatform. To further characterize the size distribution of Cu₂O, CS, and CSC nanoplatforms, dynamic light scattering analysis was performed. As shown in Fig. 2D, the hydrodynamic diameters of all three nanoplatforms predominantly ranged between 100 and 200 nm, which were larger than the sizes obtained by TEM. This discrepancy arises because surface charges on nanoplatforms attract polar water molecules to form ordered hydration layers. As shown in Fig. S1, the CSC nanoplatform is very stable in both PBS and DMEM solutions over one week.

Furthermore, the observed variations in size among Cu₂O, CS, and CSC indicate that different surface modifications led to subtle differences in their respective hydration layers. To further characterize the differences among Cu₂O, CS, CS-NH₂, and CSC, we measured their surface charges. As shown in Fig. 2E, bare Cu₂O exhibited a surface potential of approximately +15 mV, while SiO₂-coated CS showed a negative potential of −21 mV due to the presence of abundant Si–OH groups on the silica surface. After modification with (3-aminopropyl)triethoxysilane (APTES), the surface charge became positive, confirming successful introduction of amino groups. Subsequent conjugation with Ce6 through activated carboxyl groups (–COOH) neutralized part of the positive charges, while the remaining exposed –COOH groups rendered CSC negatively charged. To verify successful Ce6 conjugation, we performed UV-visible spectroscopic analysis. As shown in Fig. 2F, Cu₂O, CS, and CS-NH₂ showed no characteristic peaks between 400 and 800 nm, with nearly identical absorption profiles. In contrast, CSC displayed a distinct absorption peak at 660 nm corresponding to Ce6, providing conclusive evidence for successful nanoplatform functionalization. To verify whether the Cu₂O composition was altered during CSC synthesis, we conducted X-ray diffraction (XRD) analysis. As shown in Fig. 2G, the XRD pattern of CSC exhibits excellent consistency with the standard Cu₂O powder diffraction file (PDF# 34-1354), confirming that the silica coating process did not modify the crystalline composition of Cu₂O. We further conducted energy-dispersive X-ray spectroscopy (EDS) analysis. As shown in Fig. 2H, the CSC nanoplatform primarily consisted of Cu, Si, O, and N elements, which is consistent with the elemental mapping results. As shown in Fig. S2, the XPS spectrum of Cu 2p confirmed that the main Cu species was Cu⁺ (Cu⁺ 2p_3/2: 932.55 eV and Cu⁺ 2p_1/2: 954.43 eV), though a few features of Cu²⁺ were also found at binding energies of 935.27 eV (Cu²⁺ 2p_3/2) and 955.06 eV (Cu²⁺ 2p_3/2), which could be ascribed to oxidation. The CSC nanoplatform demonstrated excellent pH-responsive properties. We quantitatively analyzed the copper ion (Cu²⁺) release capability under acidic conditions using atomic absorption spectroscopy. As shown in Fig. 2I, the Cu²⁺ release exhibited a time- and pH-dependent increase. At pH 5.5, the cumulative Cu²⁺ release reached 102.3 ± 0.6 μg mL⁻¹, while only 23.6 ± 0.2 μg mL⁻¹ was detected at physiological pH 7.4. These results clearly demonstrate the superior acid-triggered degradation capability of the CSC nanoplatform.

Cellular uptake and cuproptosis

The cellular uptake of a nanoplatform by tumor cells plays a pivotal role in determining therapeutic efficacy, as the internalization process directly influences the effectiveness of drug delivery systems. To verify whether the nanoplatform could be internalized by cancer cells, we conducted confocal fluorescence imaging and flow cytometry experiments. 4T1 cells were co-incubated with CSC probes for varying durations (0 h, 1 h, 2 h, 4 h, and 6 h), with cell nuclei counterstained using DAPI fluorescent dye.

As shown in Fig. 3A, confocal fluorescence imaging revealed distinct red fluorescence signals from Ce6-labeled CSC probes appearing in the cytoplasm after 1 h incubation. The fluorescence intensity exhibited a time-dependent enhancement, demonstrating progressive internalization of the CSC nanoplatform into cells. Notably, the fluorescence signal plateaued at 4–6 h incubation periods, indicating saturation of cellular uptake capacity for the CSC nanoplatform. To quantitatively assess the cellular uptake efficiency of CSC probes by 4T1 cells, flow cytometric analysis was performed. As illustrated in Fig. 3B and Fig. S3, the internalization rate reached 77.2% within the first hour of incubation. A time-dependent increase in nanoplatform uptake was observed, with the uptake efficiency peaking at 93.9% after 6 h of incubation, indicating near-maximal cellular internalization capacity had been achieved. Following cellular internalization, CSC released copper ions, triggering a cascade of biological effects. To validate the cuproptosis-inducing capability of CSC, we investigated the expression of cuproptosis-related proteins through western blot analysis. 4T1 cells were treated with PBS (control), CS, or CSC, and comparative protein expression profiling was conducted to examine CSC-mediated cuproptosis. As demonstrated in Fig. 3C, both CS and CSC treatments significantly downregulated the protein levels of DLAT, ferredoxin 1 (FDX1), and lipoyl synthase (LIAS). These results confirm that CS and CSC effectively induce cuproptosis in tumor cells.

Fig. 3 The uptake of CSC by tumor cells. (A) Fluorescence imaging. (B) Flow cytometry analysis. (C) DLAT, FDX1, and LIAS protein levels after treatment of 4T1 cells with different samples. The samples of the same group were derived from the same experiment, and the gels/blots were processed in parallel.

Cytotoxicity

To evaluate the cytotoxicity of CSC, NIH 3T3 and 4T1 cells were incubated with varying concentrations of CSC probes for 12 h, followed by cell viability assessment using CCK-8 assay. As shown in Fig. 4A, both cell types maintained viability above 90% at probe concentrations ranging from 0 to 50 μg mL⁻¹, demonstrating negligible toxicity. At 100 μg mL⁻¹, NIH 3T3 cells retained high viability (81.7 ± 0.2%), while 4T1 cell viability decreased to 61.2 ± 0.2%. These results indicate that CSC probes exhibit excellent biocompatibility with normal cells while demonstrating selective cytotoxicity toward tumor cells. We further evaluated the synergistic therapeutic effect of CSC (50 μg mL⁻¹) combined with PDT on 4T1 cells. As shown in Fig. 4B, laser irradiation alone showed negligible impact on cell viability, indicating minimal cellular damage from the laser treatment. Without laser irradiation, both CS and CSC probes at the same concentration induced approximately 10% cell mortality. Following laser treatment, the CS group showed no significant changes, whereas the CSC group exhibited markedly reduced cell viability. This phenomenon can be attributed to the abundant ROS generation by Ce6 in CSC probes upon laser excitation, which effectively killed tumor cells. In contrast, CS probes lacked the capacity to produce ROS under laser irradiation, resulting in lower cytotoxicity.

Fig. 4 Cytotoxicity of nanocarrier and the production of ROS. (A) CCK-8 cytotoxicity of 4T1 and 3T3 cells incubated with different concentrations of CSC. Data are presented as the mean ± SD (n = 5 independent experiments). (B) Survival rate of 4T1 cells incubated and treated with different probes; ***p < 0.01. Data are presented as the mean ± SD (n = 5 independent experiments). (C) The ultraviolet–visible absorption spectra of DPBF and CSC mixed solution with different irradiation times. (D) ROS production detected by fluorescence of DCFH-DA in 4T1 cells incubated with different probes. Scale bar: 50 μm. (E) Fluorescence images of Calcein-AM/PI-stained 4T1 cells incubated with different probes under different conditions. Scale bar: 50 μm.

To further confirm ROS generation by CSC under photoexcitation, we employed 1,3-diphenylisobenzofuran (DPBF) as a molecular probe. The reaction between DPBF and ROS leads to cleavage of the furan ring, producing colorless 1,2-dibenzoylbenzene and consequently decreasing the UV-visible absorption at 410 nm. As demonstrated in Fig. 4C, the absorption loss at 410 nm progressively increased with prolonged irradiation time, indicating continuous ROS generation that oxidized DPBF. These results provide direct evidence that CSC can efficiently produce substantial ROS upon laser irradiation. To further verify intracellular ROS generation by CSC probes, we utilized 2′,7′-dichlorodihydrofluorescein diacetate (DCFH-DA) as a fluorescent probe. This membrane-permeable reagent is hydrolyzed by intracellular esterases and subsequently oxidized by ROS to emit green fluorescence, with intensity proportional to ROS levels. As shown in Fig. 4D, the PBS control group showed no green fluorescence regardless of laser irradiation, confirming that laser treatment alone cannot generate ROS. Both CS and CSC groups exhibited weak baseline fluorescence, potentially attributable to endogenous H₂O₂ reacting with copper ions to form ROS. Notably, laser irradiation caused no significant fluorescence enhancement in the CS group, confirming its inability to produce ROS under illumination. In contrast, the CSC group demonstrated dramatically intensified green fluorescence upon laser exposure, unequivocally demonstrating the capacity for robust ROS generation. These results collectively establish that CSC probes can effectively elevate intracellular ROS levels under laser irradiation, thereby achieving therapeutic tumor cell ablation.

To further validate the synergistic therapeutic effect of CSC probes, we performed confocal microscopy using Calcein-AM and propidium iodide (PI) on treated cells. Calcein-AM penetrates live cells and is hydrolyzed by intracellular esterases to emit green fluorescence, while PI enters membrane-compromised dead cells and binds nucleic acids to produce red fluorescence. This staining strategy enables clear differentiation between viable and dead cells. As shown in Fig. 4E, PBS-treated control groups exhibited intense green fluorescence regardless of laser irradiation, confirming that laser treatment alone cannot induce cell death. Cells incubated with CS probes showed predominant green fluorescence with sporadic red signals, suggesting partial cell death potentially caused by copper ion release from internalized probes. Notably, laser irradiation failed to enhance the red fluorescence in CS groups, indicating its inability to potentiate CS-mediated cytotoxicity. In striking contrast, CSC probe-treated cells displayed dramatically different responses, laser-irradiated samples showing minimal green fluorescence but extensive red staining, indicating near-complete cell death, while non-irradiated counterparts maintained substantial viability with only scattered red signals. These results conclusively demonstrate the laser-dependent synergistic effect between copper ion release and PDT in CSC-mediated tumor cell killing.

Biosafety evaluation in vivo

The biocompatibility of nanoplatforms is one of the key characteristics for their application in the biomedical field, directly influencing their safety and effectiveness both in vivo and in vitro. The hemolysis assay is a critical method for evaluating their blood compatibility. Good biocompatibility requires that a nanoplatform does not induce hemolytic reactions when in contact with blood components while maintaining the integrity of red blood cell membranes. Therefore, we first investigated the in vitro hemocompatibility evaluation of the nanoplatform. A total of 1.5 mL of fresh blood was collected from two 6-week-old KM mice into anticoagulant tubes. After centrifugation at 700g for 5 min to remove the supernatant, the red blood cells (RBCs) were washed multiple times with PBS. Then, 10 μL of the RBC suspension was added to 400 μL of PBS containing the CSC nanoplatform at concentrations of 3.2, 6.3, 12.5, 25, 50, 100, and 200 μg mL⁻¹. PBS and ultrapure water were used as the negative control and positive control, respectively. The samples were incubated at 37 °C in a 5% CO₂ incubator for 2 h, followed by centrifugation at 4000 rpm for 5 min. Photographs were taken, and the absorbance of the supernatant was measured at 577 nm. As shown in Fig. S4, the hemolysis rate of RBCs in ultrapure water was approximately 100%, whereas in PBS, it was nearly 0%, which is attributed to the difference in osmotic pressure between the two solutions. When the nanoplatform was introduced into the system, the RBCs mostly settled, and the supernatant exhibited minimal changes compared to that before probe addition. Even at a high nanoplatform concentration of 200 μg mL⁻¹, the hemolysis rate remained low. These results indicate that the nanoplatform possesses excellent biocompatibility with blood cells, exerting only a negligible impact on RBCs. This finding provides a safety assurance for subsequent in vivo experiments.

The biocompatibility evaluation of nanoplatforms requires comprehensive investigation of their effects on blood biochemical indicators, which represents a critical component of preclinical safety assessment. To determine potential in vivo toxicity, healthy mice were randomly divided into five groups (n = 3 per group). The CSC nanoplatform was administered via tail vein injection on days 3, 7, 15, and 30, with one control group receiving an equivalent volume of PBS. Blood samples were collected at the endpoint of the experiment for complete biochemical analysis. As shown in Fig. 5A–H, no significant differences were observed in white blood cells (WBC), red blood cells (RBC), hemoglobin (HGB), mean corpuscular volume (MCV), or platelets (PLT) between mice injected with nanoplatform on days 3, 7, 15, and 30 and the control group, demonstrating excellent hematological safety of the nanoplatform. Furthermore, key biochemical markers including alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), creatinine (CREA), uric acid (UA), and blood urea nitrogen (BUN) all remained within normal ranges and showed no statistically significant differences compared to controls. These comprehensive results confirm that the nanoplatform caused no detectable hepatic or renal toxicity throughout the entire evaluation period. In addition to hematological analysis, we investigated the effects of the CSC nanoplatform on mouse organs to further evaluate the biosafety of the nanoplatform. Healthy mice were randomly divided into four groups (n = 3 per group). The CSC nanoplatform was administered via tail vein injection at 1, 2 and 4 weeks, with one control group receiving an equivalent volume of PBS. Upon completion of the experiment, hematoxylin–eosin staining (HE) was performed on the heart, liver, spleen, lungs and kidneys. As shown in Fig. 5I, all tissue samples from mice injected with the probes at 1, 2 and 4 weeks exhibited normal and well-defined cellular morphology without pathological alterations, the myocardial cells are arranged in an orderly manner with clearly visible striations, and no ruptured or necrotic structures are observed; the liver structure shows distinct characteristic features such as hepatic lobules, with no swelling or necrotic structures present; the spleen exhibits clear splenic corpuscles with visible central arteries; the lungs display well-defined alveoli without thickening of alveolar walls, collapse, or emphysematous changes; the kidney tissue demonstrates clear structural organization with no signs of atrophy or swelling. These results demonstrate that CSC probes induce no pathological lesions in major organs, confirming their excellent biocompatibility and long-term in vivo safety. This foundational safety profile strongly supports further investigation of CSC probes for tumor therapy applications.

Fig. 5 Blood analysis of mice after injection of CSC: (A) WBC, (B) RBC, (C) HGB, (D) MCV and (E) PLT. Liver function index analysis: (F) ALT, AST and ALP. Kidney function index analysis: (G) CREA and UA and (H) BUN. (I) HE staining of heart, liver, spleen, lungs, and kidneys after injection of CSC at 1, 2 and 4 weeks. Scale bar: 100 μm.

Tumor targeting and metabolism

Encouraged by the biocompatibility in vitro, the biodistribution in various organs and pharmacokinetics of the CSC nanoplatform were further evaluated in vivo. In this experiment, Cy5.5 was first covalently conjugated to SiO₂–NH₂via an amide bond, forming SiO₂-Cy5.5 nanoplatform (SC), similar to the method used for Ce6. Then SC was injected into 4T1 tumor-bearing mice through tail vein, and its distribution in tumor sites and various organs was monitored by fluorescence imaging. As shown in Fig. 6A and B, following intravenous administration of SC, the fluorescence signal at the tumor site peaked at 4 hours post-injection, then gradually declined before stabilizing between 8 and 12 hours. Notably, significant fluorescence intensity remained detectable even at 48 hours, demonstrating effective tumor-specific accumulation of SC. As shown in Fig. 6C, ex vivo imaging of organs harvested at the 12-hour time point demonstrated predominant SC accumulation in tumor tissue, while substantial amounts were also detected in hepatic and renal tissues. To evaluate the pharmacokinetic behavior of the CSC nanoplatform, blood samples were collected from treated mice at predetermined time intervals. The metabolic clearance of the nanoplatform was assessed by tracking the concentration of the copper element in circulation. Pharmacokinetic analysis revealed a blood circulation half-life (t_1/2) of 2.78 h for CSC (Fig. 6D), demonstrating its gradual elimination from the bloodstream. In summary, CSC achieves prolonged tumor accumulation through enhanced circulatory retention while maintaining efficient systemic clearance, thus providing a pharmacological foundation for synergistic antitumor therapy.

Fig. 6 Tumor targeting and metabolism. (A) Fluorescence imaging of tumor at different time points in 4T1 tumor-bearing mice after intravenous injection of SC. (B) Quantitative fluorescence analysis of tumor sites at different time points (n = 3). (C) The distribution of SC nanoplatform in major organs at 12 h. (D) The pharmacokinetics of copper ions during blood circulation over time in healthy mice after injection of CSC (n = 3).

In vivo tumor treatment

We next evaluated the in vivo therapeutic efficacy of the CSC nanoplatform in tumor-bearing mice. Female Balb/c mice (4 weeks old) were subcutaneously inoculated with 4T1 cells (10⁶ cells), and when tumor volumes reached 80–100 mm³, the mice were randomly divided into six groups (n = 5 per group) for different treatments: (I) PBS, (II) PBS + L, (III) CS, (IV) CS + L, (V) CSC, (VI) CSC + L. During the experiment, body weight and tumor volume were measured every two days to monitor therapeutic effects and potential toxicity.

As shown in Fig. 7A, after establishing the subcutaneous breast cancer model, different nanoplatforms were injected, followed by laser irradiation at the 12th hour post-injection to evaluate the therapeutic efficacy over 18 days. Fig. 7B demonstrates that the body weight trends of mice in all groups remained consistent with that of the blank control group throughout the 18-day period, indicating that the nanoplatform treatments caused no harm to the mice. Tumor volume monitoring (Fig. 7C) revealed that neither the PBS group nor the PBS + L group exhibited tumor growth inhibition. In the CS, CS + L, and CSC groups, tumors continued to grow progressively, suggesting no significant therapeutic effect. In contrast, the CSC + L group showed marked suppression of tumor growth, confirming that copper ions alone had limited antitumor efficacy, while their combination with PDT achieved pronounced therapeutic outcomes. Fig. 7D displays tumors excised from mice on day 18, highlighting the long-term therapeutic benefits of the nanoplatform. Furthermore, statistical analysis of tumor volume changes in individual mice over 18 days (Fig. 7E) and comparative images of tumors before and after treatment (Fig. 7F) provided direct visual evidence of the superior therapeutic performance of the CSC + L treatment compared to other treatments. These results collectively demonstrate that only the CSC nanoplatform combined with laser irradiation (CSC + L) could effectively inhibit tumor growth, exhibiting statistically significant differences compared to all other groups.

Fig. 7 In vivo tumor therapy with CSC. (A) Treatment process. (B) Body weight change. (C) Tumor volume after different treatments. (D) Photographs of representative tumors after different treatments. (E) Tumor volume change curves for each mouse after different treatments. (F) Representative photographs of mice from different groups after various treatments. (G) TUNEL staining of tumor. Scale bar: 50 μm. (H) HE staining of tumor after various treatments. Scale bar: 100 μm. Data are presented as the mean ± SD (n = 5 mice per group). *p < 0.05; **p < 0.01; ***p < 0.001.

To further evaluate tumor tissue damage post-treatment, TUNEL and HE staining were performed on tumor sections after different therapeutic interventions. As shown in Fig. 7G, tumor tissues from the PBS and PBS + L groups exhibited strong blue fluorescence but negligible green fluorescence, indicating minimal cellular damage. In contrast, the CS, CS + L, and CSC groups showed sporadic green fluorescence, suggesting limited apoptosis likely induced by copper ion overload. Notably, the CSC + L group (laser-irradiated CSC) displayed intense green fluorescence, demonstrating major apoptosis and confirming the synergistic effect between copper ions and PDT. HE staining results (Fig. 7H) revealed that control group tumors exhibited characteristic malignant features: densely packed cells with minimal intercellular space, enlarged nuclei, and irregular morphology due to rapid proliferation. In contrast, tumors treated with the CSC nanoplatform combined with laser irradiation showed markedly improved histopathology: increased intercellular spacing, reduced cell density, and normalized nuclear morphology, demonstrating effective suppression of tumor growth through combined cuproptosis and PDT.

Experimental

Materials

Copper chloride dihydrate (CuCl₂·2H₂O) and sodium hydroxide (NaOH) were purchased from Sinopharm Chemical Reagent Co. Ltd (Shanghai, China). PVP, L-ascorbic acid, tetraethyl orthosilicate (TEOS) and APTES were purchased from Shanghai Aladdin Biochemical Technology Co. Ltd. Ce6 was purchased from Shanghai Macklin Biochemical Co. Ltd. mPEG₅₀₀₀-silane was purchased from GuangZhou Tanshtech Co. Ltd. DCFH-DA, calcein acetoxymethyl ester (Calcein AM), PI, BCA protein assay kit, SDS-PAGE sample loading buffer, and hydrophobic polyvinylidene fluoride (PVDF) were purchased from Beyotime Biotechnology. All reagents were used directly without further purification. Deionized water (Millipore Milli-Q grade, 18.2 MΩ) was used in all the experiments. Female Balb/c mice were purchased from Beijing Vital River Laboratory Animal Technology Co. Ltd.

Instruments

The morphology of the nanoplatform was characterized with an HT770 transmission electron microscope (Hitachi, Japan). Hydrated particle size and zeta potential of the nanoplatform were measured using a Nano-ZS90 nanometer (Malvern, UK). The absorption spectrum was recorded with a UV-2550 UV–visible spectrophotometer (Shimadzu, Japan). The elemental composition of nanoplatform was analyzed using an AXISULTRA DLD-600 W X-ray photoelectron spectroscopy (XPS) instrument (Shimadzu, Japan). The concentration of copper was determined with a WFX-200 atomic absorption spectrophotometer (Beijing Beifen–Ruili Analytical Instrument Co. Ltd, China). The fluorescence imaging of cells was conducted with an FV3000 confocal microscope (Olympus, Japan). An MDL-III-660-2.3 W laser (Changchun New Industries Optoelectronics Tech. Co. Ltd, China) was used for laser irradiation.

Synthesis of Cu₂O nanoplatform

CuCl₂·2H₂O (0.10 mmol) and PVP (0.10 g) were dissolved in 50.0 mL of deionized water, followed by the dropwise addition of 2.0 mL of 0.2 M NaOH; the solution was stirred magnetically for 5 min. Finally, 2.0 mL of 0.10 M ascorbic acid was added dropwise, and the solution was further stirred for 5 min. The product was recovered by centrifugation and then washed with ethanol twice.

Synthesis of CS

Briefly, Cu₂O particles were redispersed in 40 mL of ethanol by sonication for 30 min. Then, 40 μL of TEOS was added to the mixture and stirred for 10 min. Later, 5 mL of NaOH (8 mM) was added to the above mixture and stirred thoroughly for another 10 min. The resultant mixture was further stirred at room temperature for 12 h before separating the solid phase via centrifugation. The process was carried out under argon shielding throughout. The product (CS) was washed twice with ethanol.

Synthesis of CSC

CS particles were redispersed in 40 mL of ethanol by sonication for 30 min. Then, 10 mg mPEG-Saline and 40 μL APTES were added to the above mixture and stirred thoroughly for another 10 min. Later, 5 mL of NaOH (8 mM) was added to the above mixture and stirred for another 6 hours. The product (CS-NH₂) was washed twice with ethanol. CS-NH₂ particles were redispersed in 10 mL of HEPES buffer, a mixed solvent (6 mg of EDC, 4 mg of NHS and 4 mg of Ce6) was added to the above mixture and stirred thoroughly for another 4 hours. The product (CSC) was washed twice with ethanol and then dispersed in water.

Release of copper ions

A solution of CSC (4 mg mL⁻¹) was placed in a dialysis bag. The pH of the dialysate was then adjusted to 5.5, 6.5, and 7.4. 200 μL of the solution collected at specific time intervals of 5, 10, 30 min and 1, 2, 6, 12, 24, 36 h was taken out for measurement.

Cellular uptake

4T1 cells were seeded in a confocal dish and cultured overnight, and 1 mL serum-free medium containing CSC (50 μg mL⁻¹) was added and incubated for 1, 2, 4, and 6 h, respectively. After washing with PBS three times, the cells were stained with DAPI for 10 min, and then observed under a laser confocal scanning microscopic imaging system. For flow cytometry assay, 4T1 cells was seeded in a 6-well plate and incubated overnight, and 1 mL serum-free medium containing CSC (50 μg mL⁻¹) was added and incubated for 1, 2, 4, 6 h respectively. Then, 10 000 cells were quantitatively analyzed via Ce6 fluorescence intensity with a flow cytometer.

Detection of intracellular protein expression

4T1 cells were seeded into 6-well plates at a density of 1.0 × 10⁵ cells per well and cultured in a 5% CO₂ incubator at 37 °C for 12 h. The cells were then treated with PBS, CS or CSC for 12 h. The cell pellet was collected and lysed using RIPA buffer to extract total proteins. Protein concentration was determined using a BCA Protein Concentration Assay Kit and samples were adjusted to ensure equal loading amounts. Subsequently, SDS loading buffer was added to the protein samples, followed by heat denaturation at 95 °C for 5–10 minutes. Electrophoresis was then performed using SDS-PAGE gels. After separation, proteins were transferred to PVDF or nitrocellulose membranes using wet transfer systems, with careful attention to proper “sandwich” assembly and transfer conditions. The membranes were subsequently blocked with 5% skim milk or BSA to prevent nonspecific binding before incubation with primary antibodies (typically overnight at 4 °C) and HRP-conjugated secondary antibodies. Following thorough washing steps to remove unbound antibodies, protein bands were visualized using enhanced chemiluminescence detection systems.

Cytotoxicity assay

4T1 cells and NIH 3T3 cells were separately inoculated into 96-well plates and cultured in a 5% CO₂ incubator at 37 °C overnight. Then the medium was removed and different concentrations of CSC (0, 3.1, 6.3, 12.5, 25, 50, 100 and 200 μg mL⁻¹) were added. After 12 h, the medium was removed, and the cells were washed with PBS three times. Then, 200 μL of CCK-8 solution was added to each well. Subsequently, the absorption at 450 nm of each well was measured using an enzyme marker to determine cell viability.

Detection of intracellular ROS

DCFH-DA was used to detect intracellular ROS. 4T1 cells were seeded in 6-well plates and cultured overnight, and 2 mL of serum-free medium containing the same concentration of Cu₂O, CS, and CSC (50 μg mL⁻¹) was added and incubated for 4 h. After washing three times with PBS, serum-free medium containing DCFH-DA (10 μM) was added and incubated for 30 min. Afterward, the cells in some groups were irradiated with a 660 nm laser (100 mW cm⁻²) for 3 min in the dark and then observed under a laser confocal scanning microscopic imaging system.

In vivo toxicity evaluation of CSC

Six-week-old female Balb/C mice were randomly divided into five groups. Four groups were injected with 200 μL of CSC (20 mg kg⁻¹) through the tail vein, and the last group was injected with the same amount of PBS as control. After 1, 7, 14 and 30 days of observation, the mice were sacrificed and blood was collected for routine blood and biochemical determination. In addition, the major organs of heart, liver, spleen, lungs, and kidneys of each group were fixed with 4% paraformaldehyde, embedded in paraffin, and sliced for HE staining. All animal surgeries were performed in accordance with the Guidelines of Assessment for Humane Endpoints in Animal Experiments (Certification and Accreditation Administration of P. R. China, RB/T 173-2018) and approved by the Ethics Committee for Animal Experimentation of Huazhong University of Science and Technology (no. 4208).

In vivo synergistic treatment effect evaluation of CSC

A subcutaneous tumor model was established by injecting about 1 × 10⁶ 4T1 cells per mouse into the lower right back of Balb/c mice. And the mice bearing 4T1 tumor were randomly divided into 6 groups with 5 mice in each group, including (I) PBS, (II) PBS + L, (III) CS, (IV) CS + L, (V) CSC, (VI) CS + L. After injecting different nanoplatforms (20 mg kg⁻¹) through tail vein for 12 h, groups II, IV and VI were irradiated with a 660 nm laser (1 W cm⁻²) for 6 min, and the temperature change was recorded by near-infrared thermal imaging. In addition, one mouse in each group was euthanized and the tumor was collected. After washing several times with PBS, it was immersed in 4% paraformaldehyde and fixed for HE and TUNEL staining. At the same time, the weight and tumor volume of each mouse in every group were recorded by digital vernier caliper and scale every other day. After 18 days, the mice were euthanized, and the tumor was dissected and weighed to calculate the tumor mass and tumor suppression rate.

Conclusions

In this study, we successfully developed a nanotherapeutic probe (CSC) by integrating Cu₂O and Ce6 photosensitizer for combined PDT and cuproptosis in breast cancer treatment. The nanoplatform accumulated at the tumor site through the EPR effect. Under external laser irradiation, the photosensitizer Ce6 generated substantial ROS. Simultaneously, the acidic tumor microenvironment triggered the degradation of the Cu₂O nanoplatform, releasing abundant copper ions. These ions entered cells, bound to lipoylated proteins, and induced their aggregation, thereby disrupting mitochondrial metabolism and triggering cuproptosis. Experimental results demonstrated significant synergistic therapeutic effects between PDT and cuproptosis.

Author contributions

Yong-Guo Hu and Fang Zhang designed the project, performed the experiments and wrote the original manuscript. Fang Zhang and Li-Li Chen analyzed the results. Xin-Yue Xu, Pei-Jie Zhang and Yong Li investigated relevant literature. Jin-Xuan Fan, Jia-Hua Zou and Yuan-Di Zhao provided useful advice for this work.

Conflicts of interest

The authors declare no competing interests.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article and its SI. See DOI: https://doi.org/10.1039/d5bm00954e.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (grant no. 62375093), Technology Innovation Program of Hubei Province (2024BCB058), the Natural Science Foundation of Hubei Province of China (2025AFB559), and China Ageing Development Foundation – Elderly Navigation Program Research Fund (193). We also thank the Analytical and Testing Center (HUST) and the Research Core Facilities for Life Science (HUST) for help with measurements.

References

1. Y. You, Z. Guo, T. Wolter and Q. Hu, Chem. Soc. Rev., 2025, 54, 1552–1582.
2. R. Joshi, B. Telang, G. Soni and A. Khalife, Oncol. Transl. Med., 2024, 10, 105–109.
3. I. Poon, C. Lucas, A. Rossi and K. Ravichandran, Nat. Rev. Immunol., 2014, 14, 166–180.
4. Y. Huang, C. Lee, P. Borgström and R. A. Gjerset, Mol. Ther., 2007, 15, 524–533.
5. J. Zhang, Y. Hu, X. Wen, Z. Yang, Z. Wang, Z. Feng, H. Bai, Q. Xue, Y. Miao, T. Tian, P. Zheng, J. Zhang, J. Li, L. Qiu, J. Xu and D. Ye, Nat. Nanotechnol., 2025, 20, 563–574.
6. J. Inkol, M. Westerveld, S. Verburg, S. Walsh, J. Morrison, K. Mossman, S. Worfolk, K. Kallio, N. Phippen, R. Burchett, Y. Wan, J. Bramson and S. Workenhe, J. Immunother. Cancer, 2024, 12, e006781.
7. P. Tsvetkov, S. Coy, B. Petrova, M. Dreishpoon, A. Verma, M. Abdusamad, J. Rossen, L. Joesch-Cohen, R. Humeidi, R. Spangler, J. Eaton, E. Frenkel, M. Kocak, S. Corsello, S. Lutsenko, N. Kanarek, S. Santagata and T. Golub, Science, 2022, 375, 1254–1261.
8. L. Chen, J. Min and F. Wang, Signal Transduction Targeted Ther., 2022, 7, 378.
9. W. Wang, W. Mo, Z. Hang, Y. Huang, H. Yi, Z. Sun and A. Lei, ACS Nano, 2023, 17, 19581–19599.
10. P. Tsvetkov, A. Detappe, K. Cai, H. Keys, Z. Brune, W. Ying, P. Thiru, M. Reidy, G. Kugener and J. Rossen, Nat. Chem. Biol., 2019, 15, 681–689.
11. R. Lill and S. Freibert, Annu. Rev. Biochem., 2020, 89, 471–499.
12. X. Zhang, B. Tang, J. Luo, Y. Yang, Q. Weng, S. Fang, Z. Zhao, J. Tu, M. Chen and J. Ji, Mol. Cancer, 2024, 23, 255.
13. F. Hu, J. Huang, T. Bing, W. Mou, D. Li, H. Zhang, Y. Chen, Q. Jin, Y. Yu and Z. Yang, Adv. Sci., 2024, 11, 2309388.
14. C. Wang, X. Yang, C. Dong, K. Chai, J. Ruan and S. Shi, Coord. Chem. Rev., 2023, 487, 215156.
15. Z. Li, J. Yang, B. Ren, Q. Fan, L. Huang, S. Guo, R. Zhou, S. Chen, J. Feng, C. Yan, X. Chen and Z. Shen, Adv. Mater., 2024, 36, 2313212.
16. S. Ning, M. Lyu, D. Zhu, J. W. Y. Lam, Q. Huang, T. Zhang and Z. Tang, ACS Nano, 2023, 17, 10206–10217.
17. Y. Wang, X. Yao, Y. Lu, J. Ruan, Z. Yang, C. Wang, N. Yang, Y. Gao and S. Shi, Adv. Mater., 2025, 21, 2501435.
18. M. Guan, K. Cheng, X. Xie, Y. Li, M. Ma, B. Zhang, S. Chen, W. Chen, B. Liu, J. Fan and Y. Zhao, Nat. Commun., 2024, 15, 10060.
19. H. Xiang, C. You, W. Liu, D. Wang, Y. Chen and C. Dong, Biomaterials, 2021, 277, 121071.
20. Y. Zhang, S. Ya, J. Huang, Y. Ju, X. Fang, X. Ouyang, Q. Zeng, X. Zhou, X. Yan and G. Nie, Exploration, 2025, 5, 20240275.
21. Y. Ding, Y. Xu, W. Yang, P. Niu, X. Li, Y. Chen, Z. Li, Y. Liu, Y. An, Y. Liu, W. Shen and L. Shi, Nano Today, 2020, 35, 100970.
22. S. Sindhwani, A. Syed, J. Ngai, B. Kingston, L. Maiorino, J. Rothschild, P. MacMillan, Y. Zhang, N. Rajesh, T. Hoang, J. Wu, S. Wilhelm, A. Zilman, S. Gadde, A. Sulaiman, B. Ouyang, Z. Lin, L. Wang, M. Egeblad and W. Chan, Nat. Mater., 2020, 19, 566–575.
23. L. Wang, J. Piao, H. Qi, C. Qiao, Q. Guo, M. Wei and Z. Li, Coord. Chem. Rev., 2025, 523, 216284.
24. C. Zhang, T. Huang and L. Li, J. Hematol. Oncol., 2024, 17, 68.
25. S. Gao, H. Ge, L. Gao, Y. Gao, S. Tang, Y. Li, Z. Yuan and W. Chen, Adv. Sci., 2025, 12, 2417676.
26. Y. Li, X. Lei, X. Zhang, B. Zhang, Y. Hu, M. Guan, K. Cheng, W. Chen, B. Liu, J. Fan and Y. Zhao, Angew. Chem., Int. Ed., 2024, 63, e202411598.
27. J. Fan, M. Peng, H. Wang, H. Zheng, Z. Liu, C. Li, X. Wang, X. Liu, S. Cheng and X. Zhang, Adv. Mater., 2019, 31, 1808278.
28. S. Sun, Q. Chen, Z. Tang, C. Liu, Z. Li, A. Wu and H. Lin, Angew. Chem., Int. Ed., 2020, 59, 21041–21048.
29. Z. Yang, X. Shen, J. Jin, X. Jiang, W. Pan, C. Wu, D. Yu, P. Li, W. Feng and Y. Chen, Adv. Sci., 2024, 11, 2400251.
30. T. Wu, X. Yan, R. Zhang and K. Fan, J. Inorg. Mater., 2023, 38, 43–54.
31. C. Wu, Y. You, D. Yu, Y. X. Zhu, H. Lin and J. Shi, Adv. Healthcare Mater., 2025, 14, 2501021.
32. B. Jia, Y. Liu, X. Geng, Y. Li, C. Zhang, Y. Qu, X. Liu, M. Zhao, Y. Yang and W. Li, Research, 2025, 8, 0732.
33. N. Song, H. Li, R. Tao, Y. Huang, R. Zhang, J. Guo, P. Liu, C. Yao and D. Yang, Adv. Mater., 2025, 21, 2416161.
34. L. Yu, W. Li, J. Cao, R. Miao, Y. Fu, X. Wang, J. Xie, W. Zhang, Z. Mao, H. Zhang, Y. Zhang, M. Ou and L. Mei, Adv. Funct. Mater., 2025, 21, 2425831.
35. J. Shi, X. Qi, Y. Ran, Q. Zhou, Y. Ding, L. Li, Y. Zeng, D. Qiu, Z. Cai, X. Cai and Y. Pan, Bioact. Mater., 2025, 47, 212–228.
36. X. Chen, L. Lu, Z. Liang, B. Yu, Z. Shao, J. Li, Y. Cen, Y. Liu, S. Li and X. Chen, Angew. Chem., Int. Ed., 2025, 21, e202502120.
37. R. Zhou, X. Zeng, H. Zhao, Q. Chen and P. Wu, Coord. Chem. Rev., 2022, 452, 214306.
38. Y. Liang, Q. Yuan, Q. Zheng, Z. Mei, Y. Song, H. Yan, J. Yang, S. Wu, J. Yuan and W. Wu, Nucleic Acids Res., 2024, 52, D1218–D1226.
39. K. Sun, X. Wei, S. Han, Y. Sun, H. Xiao and D. Wei, ACS Nano, 2025, 19, 13300–13313.
40. Q. Zhan, Y. Kuang, X. Chen, Y. Yang, L. Jiang, J. Chen, L. Li, J. Wang, S. Zhu and H. Huang, Bioact. Mater., 2025, 47, 327–342.
41. Y. Zhong, Z. Qiu, K. Zhang, Z. Lu, Z. Li, Y. Cen, S. Li and H. Cheng, Adv. Mater., 2025, 37, 2415078.
42. G. Zhan, L. Fan, S. Zhou and X. Yang, ACS Appl. Mater. Interfaces, 2018, 10, 35234–35243.
43. C. Yec and H. Zeng, Chem. Mater., 2012, 24, 1917–1929.
44. S. Dong, Y. Dong, T. Jia, S. Liu, J. Liu, D. Yang, F. He, S. Gai, P. Yang and J. Lin, Adv. Mater., 2020, 32, 2002439.
45. B. Li, J. Ma and P. Cheng, Angew. Chem., 2018, 130, 6950–6953.
46. H. Zhou, Z. Fan, J. Deng, P. Lemons, D. Arhontoulis, W. Bowne and H. Cheng, Nano Lett., 2016, 16, 3268–3277.
47. Y. Akiyama, T. Mori, Y. Katayama and T. Niidome, J. Controlled Release, 2009, 139, 81–84.
48. C. Xue, M. Li, Y. Zhao, J. Zhou, Y. Hu, K. Cai, Y. Zhao, S. Yu and Z. Luo, Sci. Adv., 2020, 6, eaax1346.

---

Footnote

† These authors contributed equally to this work.

---