Methanol, hydrogen peroxide (H₂O₂, 30%), and zinc nitrate hexahydrate (≥98%) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd., Zn(NO₃)₂·6H₂O), 2-methylimidazole (98%, C₄H₆N₂, 2-MI), doxorubicin (C27H29NO11, DOX), Ce6, 4,6-diamino-2-phenyl indole (DAPI), and MTT assay, were purchased from Thermo Fisher Co., Ltd.; the hypoxia/oxidative stress detection kit was obtained from Enzo Life Sciences, and the annexin V-FITC/PI double staining apoptosis detection kit was purchased from Beyotime Biotechnology Co., Ltd.

The scanning electron microscope (SEM) [S-4800 (Hitachi)] was operated at an acceleration voltage of 10 kV with 8 mm working distance. The transmission electron microscope (TEM) was operated at 200 kV with 8 mm working distance. The automatic volumetric adsorption instrument (Micromeritics ASAP 2020; Zhao et al., 2012) was operated at 77 and 298 K for nitrogen and oxygen condition. The Vertex PerkinElmer 580BIR spectrophotometer (Bruker) was used for the FT-IR experiment. The X-ray photoelectron spectroscope used was ESCALAB 250. In the nitrogen stream, the TGA experiment was measured by using a thermal analyzer (SDT 2960, TA Instruments, New Castle, DE). In this study, SDT was set at 1.0 MHz. X-ray diffraction (XRD) measurements were taken at ALBA Synchrotron (Cerdanyola del Vallès, Barcelona, Spain) at the BL04-MSPD by using a curved primary beam monochromator at a wavelength of 1.5406 Å (Copper radiation), recorded between 5 and 90° with a step width of 0.02°/S.

An improved method was used to prepare ultra-small ZIF-8 nanoparticles (Xie et al., 2019b). In short, 2-MI was added into methanol with Zn(NO₃)₂·6H₂O. After 30-minute mixing at 25°C, the product was centrifuged. After resulting from the mixture of DOX and Zn(NO₃)₂·6H₂O, DOX/ZIF-8 was prepared in the similar method. The final product was collected by centrifugation. Then, a ZIF-8 shell layer loaded with PS-Ce6 was uniformly grown around DOX/ZIF-8. Ce6 (6 mg) and DOX/ZIF-8 (60 mg) were first mixed in 2-MI solution and added to Zn(NO₃)₂·6H₂O solution in a ratio of 1:1. The mixing was continued for another 2 h at room temperature. The product was obtained by centrifugation, followed by the re-dissolution in ethanol. PDA (15 mg) was dispersed in 1 ml of ethanol and mixed with DZC solution (3 ml) for 2 h. Finally, the product was obtained by centrifugation.

An ultraviolet spectrophotometer was used to analyze the released Ce6−visible spectrophotometer. DZCP was dissolved in PBS solution with a pH of 5.6, 6.5, or 7.4. The supernatant was collected at a predetermined time point and diluted in ethanol. According to the standard curve of free Ce6, the absorbance of Ce6 was read at 410 nm. In the DOX release study, DZCP was dissolved in PBS solution with pH 6.5 and 7.4. Both drug release detections were quantitatively analyzed.

The DPBF probe was used to measure the singlet oxygen (¹ O₂) produced by Ce6 under ultrasound. The DPBF solution, with pH 6.5 or 7.3, was inserted into DZCP, respectively. Next, the UV-DPBF analysis was carried out using a visible spectrophotometer, and extracellular hydroxyl radicals (•OH) produced were measured. After that, the resultant was added to PBS buffer, pH 6.5, containing MB and mixed well. For the ESR measurement, 100 μL of the sample aqueous solution (1 mg/L) was mixed with 500 μL of the TEMP solution. In addition to the use of DMPO (50 mM) as a spin capture agent, the •OH detection was also carried out in accordance with the aforementioned method.

With DMEM (2 ml), 4T1 cells were seeded at 1 × 10⁵ cells per well into a 12-well culture plate and incubated overnight at 37°C. The cells were incubated under normoxic conditions (21% oxygen) with 5% CO₂ or under an oxygen-deficient environment (1% oxygen) for 8 h. After 4 h of incubation with DZCP at 50 μg ml⁻¹, the ultrasound-treated group was subjected to ultrasonic waves at 1 MHz for 10 min, 1.5 W cm⁻². Then, the cyto-ID hypoxia/oxidative stress detection kit was used, followed by fluorescent imaging. After washing several times with DI water, the cells were stained with DAPI.

The MTT method was applied to detect cytotoxicity. L929 and 4T1 cells were seeded in a 96-well plate at a rate of 1×10⁴ cells each well and incubated under normoxia (21% oxygen) or hypoxia (1% oxygen) and 5% CO₂ condition for 8 h. Different doses of DZCP were applied to each well. The cells were incubated for another 20 h following the DZCP treatment before applying the MTT reagent. At 490 nm, the absorbance was recorded on the Bio-Rad 680 microplate reader.

Balb/c mice were purchased from the Experimental Animal Center of Shanghai Jiaotong University (Shanghai, China). Animal experiments were carried out in accordance with the standards of “China National Regulations on the Care and Use of Laboratory Animals.” The mice with about 20 g weight were injected with 1×10⁶ 4T1 cells by hypodermic injection to establish a tumor-bearing model, which were then divided into three groups (PBS, DOX/Ce6@ZIF-8@PDA (DZCP), and DOX/Ce6@ZIF-8@PDA treatment group with ultrasound (DZCP⁺)). For the in vivo experiment, the ultrasonic parameter was set as 1.0 MHz, 50% duty cycle, and 1.0 W cm⁻². All materials used Ce6 at a concentration of 4.0 mg kg⁻¹ body weight. The nanoparticle IVs were operated on 1 and 3 days. After 24 h of injection, ultrasound was used during the treatment. Both body weight and tumor size were measured every 2 days. After 15 days, the mice were euthanized, and blood was collected for hematology research.

To evaluate DZCP biodistribution, after the injection, the in vivo fluorescence of mice was obtained at fixed intervals. In addition, the signal intensity of the nanoparticles was measured at the same time point to point out the metabolism of nanoparticles in vivo.

To investigate the nanoparticle signal intensity in vivo, nude mice were divided into three groups and were intravenously injected with 100 μL of PBS, DZCP nanoparticles, and DZCP nanoparticles with SDT. Then, the T1-weighted images were acquired by in vivo imaging with a special animal imaging coil at room temperature. The T1-weighted signal intensities were acquired through in vivo imaging via manually drawn regions of interest.

Data were statistically analyzed by OriginLab (OriginLab Corporation, Northampton, MA) and GraphPad Prism 7 software (GraphPad Software, Inc., La Jolla, CA, United States). The t-test and one-way analysis of variance (ANOVA) were applied for significance analysis, followed by the Bonferroni test. p < 0.05 was considered to indicate a statistically significant difference. The open-source image analysis platform, ImageJ, was used in image acquisition and subsequent analysis.

DOX/ZIF-8 crystals doped with DOX were constructed via an ion doping technology from previous reports. In the FT-IR result shown in Figure 1, at 1,584 cm⁻¹, an absorption peak of C=N stretching vibrations in ZIF-8 (Figure 1A) was obtained (Tran et al., 2020). The DOX loading detection was shown by TGA, suggesting a sharp weight loss in DOX/ZIF-8 compared to ZIF-8 between 300 and 450°C, which was due to the DOX decomposition (Figure 1B). Additionally, to characterize the ZIF-8 with its MOF structure, the X-ray diffraction (XRD) results are shown in Figure 1C. The small-angled XRD in the right corner of Figure 1C clearly verified the MOF architecture. The images of TEM demonstrated the size of DOX/ZIF-8 to be about 78 nm (Figure 1D). It also implied that Ce6 was loaded into the ZIF-8 shell by the one-pot process at the beginning. Finally, a polymer PDA was coated onto the DZC particles to improve their biocompatibility. In this case, the TEM image, Figure 1E, showed that DOX/Ce6@ZIF-8@PDA with an average size of about 135 nm was obtained. The DLS was also applied to DZCP to evaluate the diameter distribution in Figure 1F and presented 138.27 ± 17.14 nm of NP diameter, which was consistent with the TEM result.

In this experiment, type I isotherms of DOX/ZIF-8 and ZIF-8 were shown by nitrogen adsorption/desorption experiments and proved the evidence of micropore presence (Figure 2A). Thus, the high amount of Ce6 and DOX loading and releasing could be expected in our study. To evaluate the degradability of DZCP with ultrasonic waves, nanoparticles were dispersed in PBS buffer of pH 7.4 and treated with ultrasound for 0, 5, and 10 min, respectively, as shown in Figure 2B. The ultrasonic parameter was set as 1.0 MHz and 1.5 W cm⁻². The complete destruction of DZCP could be shown by the different shades, confirming the sonodynamic sensitiveness of DZCP. Furthermore, the released Ce6 was detected via UV−Vis spectroscopy. With a decrease in pH to 5.6, almost 90% of Ce6 is released due to the decomposition of DZCP, as shown in Figure 2C. Thus, the result demonstrated that the drug release is negatively related to the pH value. A similar result was obtained in the in vitro experiment of DOX@ZIF-8 release behavior. In Figure 2D, the DOX release behaviors are studied under different conditions. At pH 7.4, the DOX@ZIF-8 proved a poor drug leakage because of its immobilization into the framework of ZIF-8 at 48 h. After pH decreased to 5.6, nanoparticles collapsed and released DOX particles. Finally, about 88% of DOX is released into the aqueous environment under acidic conditions.

For the cytotoxicity test, MTT assay was used to evaluate the biocompatibility of DZCP to L929 cells. As shown in Figure 3A, at a concentration of 200 μg ml⁻¹, the cells treated with DZCP still showed a high survival rate, indicating the biological stability of DZCP. Due to its good biocompatibility, DZCP was gradually endocytosed by 4T1 cells within 6 h. At a concentration of 160 μg/ml, DZCP showed significantly increased cytotoxicity, indicating the importance of hypoxia alleviation for combination treatment.

In order to measure the production of ROS, 4T1 cells were cultured in vitro and incubated with DZCP⁺ (DZCP combined with ultrasonic treatment) under hypoxia (1% oxygen) conditions for 3 and 6 h, respectively. The ROS/hypoxia detection kit was used (Figure 3B), the characteristics of which are the same as the reported method (Silva et al., 2021; Zhang et al., 2021). The generated ROS was detected in DZCP⁺-treated cells with the appearance of green fluorescence, demonstrating that ROS generation and tumor cells decrease under hypoxic conditions. This phenomenon was due to the enhanced ROS generation of nanoparticles in the SDT with time. Another TEM result shown in Figure 3C demonstrated the permeability and retention of DZCP in the 4T1 cells. After incubation with 160 μg/ml DZCP for 24 h, the 4T1 cells were sliced to exhibit the permeability of the nanoparticles. Benefited from the PDA layer, the amount of DZCP was promoted by enhanced endocytosis of the 4T1 cells. Most importantly, as shown in Figure 3D, the ROS production was additionally revealed by the fluorescence staining with the 4T1 cells, which were incubated with PBS, DZCP, and DZCP⁺, respectively. Compared with PBS- or DZCP-treated cells in Figures 3Da,b, an increased appearance of green fluorescence was observed in DZCP⁺-treated cells (shown in Figure 3Dc), demonstrating the success of ROS generation. In addition, the mean fluorescence intensity (MFI) results showed that the DZCP⁺ group had significantly increased fluorescence intensity than the other two groups, which was about 8-fold of that in the PBS group.

Balb/c mice with 4T1 tumor bearing were used to evaluate antitumor experiments in vivo. The experimental mice were randomly divided to receive PBS, DZCP, or DZCP⁺ treatments. By the IVs, the mice were treated with the drugs with a Ce6 concentration of 4.0 mg kg⁻¹. Then, the DZCP ⁺ group was exposed to an ultrasound of 1 MHz for 10 min⁻². Fifteen days after the drug treatment, each group was monitored for body weight and tumor size. As shown in Figure 4A, in the mice treated with DZCP, the slight inhibition of tumor growth indicated the improved effect of DOX released by DZCP. After ultrasonic treatment, the tumor-inhibiting ability of DZCP was significantly enhanced. On day 15, in the DZCP-treated group, the mice had about a one-third-size tumor of the DZCP group, suggesting the synergistic effect of SDT and DOX highly attributed to the antitumor effect. The H&E staining from the tumor of the three groups also reported tumor suppression, declaring that the tumors in the DZCP group showed more necrotic sections (Figure 4B). Additionally, histological analyses of mice in the PBS group and the DZCP group showed normal and similar cell shapes and tissue formation, which indicated that the side effect of DZCP was negligible and did not influence body health (Figure 4C). Also, almost no nanoparticles were detected on the slices, indicating that the DZCP was metabolized in vivo safely and there was no accumulation in the recipient mice. Those slices were obtained from important organs in the body, including the heart, liver, spleen, lung, and kidney.

To evaluate the DZCP biodistribution in vivo, the images of small animals were screened to the mice at intervals by IVs. The zinc levels were determined by the signal intensities. The quantitative analysis in Figure 5 clearly shows a decreased trend of the DZCP with time. It suggested that DZCP accumulated in the liver and moved to the spleen at the first 12 h after injection. From 12 h, the biodistribution result suggested that the DZCP accumulation amount began to reduce by the incremental time.

As the benefit of the five mismatched electrons of Zn²⁺, one of the possible applications of ZIF-8 nanoparticles is the effective contrast agents in in vivo imaging. Thus, after the IVs of DZCP nanoparticles into the tumor-bearing mice, it was obvious that T1-weighted MR signals could be clearly observed with enhancement at the tumor site compared with the PBS group. The results revealed that the DZCP group signal intensity after intravenous injection for 3 weeks showed a slight decrease, suggesting that a cumulative accumulation process of DZCP in the mice could last a long period of time (Figure 6). Due to the EPR effect, the time point of the maximum cumulative amount of DZCP in tumors is about 24 h (shown in Figure 6). In addition, with SDT treatment, the signal intensity of DZCP decreased from 2 weeks, demonstrating the faster metabolism of DZCP with SDT. Comparing to the PBS-treated group, the signal intensity of DZCP with SDT was similar after 2 weeks. Therefore, due to the complexity of an in vivo microenvironment, the DZCP nanoparticles with SDT were able to gradually collapse into small ions and molecules, followed by the movement of mice metabolic systems in a relatively rapid manner.