Tungsten hexachloride (WCl₆), poly acrylic acid (PAA, Mw = 2000 D), diethylene glycol (DEG), ethyl dimethylaminopropyl carbodiimide (EDC) and sulfo-NHS were obtained from Sigma Aldrich (St. Louis, MO, United States). Methyl 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)benzoate, methylene blue, 9-bromo-10-phenylanthracene, and H₂N-PEG-OH (MW: 3,000 g/mol) were purchased from Sigma-Aldrich. All chemicals were analytically pure and used as purchased without further purification, unless otherwise noted.

W₁₈O₄₉ NPs were prepared according to a procedure previously reported (Huo et al., 2014). Briefly, 500 mg WCl₆ and 200 mg PAA were dissolved in 100 ml DEG. The mixture was stirred under the protection of N₂ to fully dissolve the substances, and then placed at 180°C for approximately 30 min to allow the reaction. When the solution turned colorless to dark green, it was allowed to cool to room temperature. Subsequently, the mixture was centrifuged, the supernatant was removed and the pellet was dissolved in 50 ml deionized water. This procedure was repeated three times.

Compound 3 was prepared according to a method previously reported (Mahmoud Asadirad et al., 2013; Kolemen et al., 2016). Next, compound 3 (50 mg, 0.12 mmol) and H₂N-PEG-OH (292 mg, 0.086 mmol, MW: 3,000 g/mol) were dissolved in 5 ml dry tetrahydrofuran. N,N′-dicyclohexylcarbodiimide (26 mg, 0.12 mmol) and 4-dimethylaminopyridine (11 mg, 0.15 mmol) were added to the solution. The reaction mixture was stirred for 2 h until the formation of a precipitate. The precipitate was filtered and removed, and cold diethyl ether was added to the solution to precipitate the product (EP). Pure EP was filtered (187.1 mg, 42%).

W₁₈O₄₉ (50 mg) NPs dissolved in PBS (10 ml) were activated with EDC (100 mg) and sulfo-NHS (50 mg). Subsequently, EP (75 mg) was added into the above solution. The mixture was allowed to react for 10 h at 4°C under vigorous stirring. Then, the as-obtained W₁₈O₄₉@EP NPs were harvested by three cycles of centrifugation and washing with PBS to remove the unreacted materials and excess of the reagents. Purified W₁₈O₄₉@EP NPs were re-suspended in PBS and stored at 4°C for further application.

Images from transmission electron microscopy (TEM) were obtained using JEOL JEM-2010 (HR). Absorption spectra were recorded using a UV-Vis spectrophotometer (Persee DU1900, Beijing, China). The hydrodynamic particle size was characterized using dynamic light scattering (DLS) by ZetaSizer Nano-ZS90 (Malvern Instrument). The X-ray diffraction (XRD) pattern was recorded in the 2θ range of 20–60°.

Human gastric carcinoma cells (MNK cells) were routinely cultured in Dulbecco’s Modified Eagle Medium (DMEM) at 37 and 5% CO₂. Subsequently, they were cultured in a six-well plate to evaluate the ability of W₁₈O₄₉@EP NPs to produce intracellular ROS. When cells reached a confluence over 80%, the medium was removed and fresh medium containing the following substances was added: 1) PBS, without any treatment, used as control; 2) H₂O₂ (0.1 mM) without any other treatment; 3) PBS and laser irradiation (808 nm, 0.48 W/cm²) for 5 min; 4) W₁₈O₄₉ NPs (0.1 mg/ml) and laser irradiation (808 nm, 0.48 W/cm²) for 5 min; 5) W₁₈O₄₉@EP NPs (0.1 mg/ml) and laser irradiation (808 nm, 0.48 W/cm²) for 5 min. The fluorescent probe H₂DCFDA was added after 12 h and incubated for 2 h. Confocal laser microscopy was used to evaluate the fluorescence of the cells after the treatments and images were collected.

The method used for this experiment was similar to the one above but using different groups. Cells were divided into six groups and treated as follows: 1) control: without any treatment; 2) W₁₈O₄₉ + RT: W₁₈O₄₉ NPs (0.1 mg/ml) and X-ray exposure (5 Gy) for 5 min; 3) W₁₈O₄₉ + Laser: W₁₈O₄₉ NPs (0.1 mg/ml) and laser irradiation (808 nm, 0.48 W/cm²) for 5 min; 4) W₁₈O₄₉@EP + Laser: W₁₈O₄₉@EP NPs (0.1 mg/ml) and laser irradiation (808 nm, 0.48 W/cm²) for 5 min; 5) W₁₈O₄₉ + Laser + RT: W₁₈O₄₉ NPs (0.1 mg/ml) and laser irradiation (808 nm, 0.48 W/cm²) simultaneous to X-ray exposure (5 Gy) for 5 min; 6) W₁₈O₄₉@EP + Laser + RT: W₁₈O₄₉@EP NPs (0.1 mg/ml) and laser irradiation (808 nm, 0.48 W/cm²) simultaneous to X-ray exposure (5 Gy) for 5 min. Different fluorescent probes were added after 12 h and incubated for 2 h. Confocal laser microscopy was used to evaluate the fluorescence of the cells after the treatments and images were collected.

A total of female BALB/c athymic nude mice, 6–8 week old, weighing 20 g were provided by the Comparative Medical Center of Yangzhou University (Yangzhou, China) and housed in a conventional clean, specific pathogen-free (SPF) facility, under a 12-h light/dark cycle with food and water at libitum, where food, water, bedding, and cages were irradiated before use. Room temperature and humidity at ∼25°C and ∼50% were continuously monitored. All procedures using animals were approved by the animal Protection Committee of Nanjing University (Nanjing, China). Next, 2 ×10⁶ hepatocellular carcinoma (HCC-4) cells suspended in 50 μl PBS were subcutaneously injected into the back of each mouse to establish the tumor model.

Mice were treated with 100 μL PBS, W₁₈O₄₉ (0.1 mg/ml), and W₁₈O₄₉@EP (0.1 mg/ml). The analysis was carried out at 6 h after injection using Hiscan XM Micro CT with X-ray tube settings at a current of 133 μA and a voltage of 60 kV (Suzhou Hiscan Information Technology, Jiangsu, China).

When the tumor size reached approximately 100 mm³, mice were randomly divided into six groups (n = 5 per group). The treatment of the mice in each group was identical as in the in vitro experiment described in the paragraph 2.7. The body weight of the mice was recorded, and the tumor size was measured every 2 days. The tumor volume was calculated according to the following Eq. 1: V = ( L × W 2 ) 2 (1) where L is the shortest diameter of the tumor and W is the shortest diameter of the tumor.

Normal mice were intravenously (i.v.) injected with W₁₈O₄₉@EP NPs, sacrificed, and blood and major organs (kidney, lung, spleen, liver, and heart) were harvested at day 1, 7, and 21. They were fixed, embedded into paraffin, cut into sections, stained with H&E using a standard protocol, and visualized under light microscopy. Blood was collected in anticoagulant tubes containing sodium EDTA and separation gel were analyzed.

W₁₈O₄₉ NPs were synthesized according to a method previously reported (Huo et al., 2014; Huo et al., 2017). W₁₈O₄₉@EP NPs were synthesized as shown in Scheme 2.

The introduction of a bridging oxygen structure (compound 2–3) was confirmed by UV-Vis spectra (Figure 1A). The disappearance of the absorption peaks at 358 nm, 376 nm, and 396 nm indicated the destruction of the anthracene ring as the introduction of the bridging oxygen (Chen et al., 2017). The morphology and size of W₁₈O₄₉ NPs and W₁₈O₄₉@EP NPs were analyzed by TEM (Figures 1B,C), DLS (Figures 1B,C insert), XRD (Figure 1D) and UV-Vis spectra (Figure 1E). W₁₈O₄₉ NPs had good mono-dispersibility in PBS with a size of approximately 10 nm according to TEM analysis, as shown in Figure 1B. DLS analysis indicated that W₁₈O₄₉ NPs were fairly uniform in size: their mean hydrodynamic diameter was 7.5 nm, with a poly-dispersity index of 0.132. W₁₈O₄₉ NPs were confirmed by XRD and UV-Vis-NIR spectrum analysis (Figures 1D,E) through the comparison with the data presented in a previous report (Kolemen et al., 2016). The diameter of W₁₈O₄₉@EP NPs increased after the modification of W₁₈O₄₉ NPs with EP and was approximately 25 nm as determined by the TEM images. The mean hydrodynamic diameter was 43.8 nm, with a poly-dispersity index of 0.108, which was bigger than that observed by TEM due to the influence of the surface organics. The XRD and UV-Vis-NIR spectrum of W₁₈O₄₉@EP NPs did not change (Figures 1D,E) compared with that of W₁₈O₄₉ NPs. EP had thermal response ability, thereby releasing its bridging oxygen, as shown in Figure 1F. A water bath at 65°C for 10 min resulted in the reappearance of the three absorption peaks that disappeared in Figure 1A, indicating that the bridging oxygen was released under the heating conditions and compound 3 was converted to compound 2. The three absorption peaks also reappeared in W₁₈O₄₉@EP NPs after 10 min of irradiation with an 808 nm laser, due to the photothermal conversion ability of W₁₈O₄₉ NPs (Kolemen et al., 2016).

The photothermal conversion effect of W₁₈O₄₉@EP NPs using an 808 nm laser was evaluated to ensure the photothermal effect of W₁₈O₄₉@EP NPs. The in vitro results showed that the temperature of the W₁₈O₄₉@EP solution (1 mg/ml) rapidly increased under laser irradiation (808 nm, 0.48 W/cm²) from 27.2°C to 35.1°C in 10°s, reaching 62.3°C after 5 min (Figure 2A). Next, 0.1 ml W₁₈O₄₉@EP solution (10 mg/ml) was injected into the tumor of tumor-bearing mice. The temperature of the tumor changed after 30 min after laser irradiation (808 nm, 0.48 W/cm²) as monitored by a thermal imaging camera. The temperature at the tumor site increased from 31.2°C to 55.7°C within 5°min, indicating that W₁₈O₄₉@EP NPs possessed a photothermal therapeutic effect in vivo (Figure 2B). These results suggested that W₁₈O₄₉@EP NPs released enough heat under laser irradiation both in vitro and in vivo through the photothermal conversion effect, which was sufficient to release ROS by EP (Chen et al., 2017).

The above-mentioned experimental analysis showed the successful synthesis of W₁₈O₄₉@EP NPs, and the presence of EP did not affect the basic performance and photothermal conversion capacity of W₁₈O₄₉ NPs. The photothermal effect generated by W₁₈O₄₉ NPs was sufficient to release the loaded bridging oxygen by EP.

Since the effectiveness of PDT relies on the action of ROS (Dougherty et al., 1998; Dolmans et al., 2003), the ROS producing ability plays a decisive role in the photodynamic effect. Therefore, the ROS production ability of W₁₈O₄₉@EP NPs was evaluated. The intracellular ROS producing ability of W₁₈O₄₉@EP NPs by H₂DCFDA revealed that the combination of EP and W₁₈O₄₉ NPs greatly increased the amount of ROS (green fluorescence) produced under laser irradiation, which enhanced the PDT effect (Figure 3). In addition, cells were not able to detoxify and eliminate ROS through their self-regulation because of the increase in ROS content, and the effect of RT on tumor killing through ROS was increased, thus greatly enhancing the RT effect.

Only RT (W₁₈O₄₉+X-ray) and PDT (W₁₈O₄₉ + Laser) could kill a small number of cells without any additional oxygen supply, and the ability to kill cells was enhanced when a combination of RT and PDT (W₁₈O₄₉ + Laser + X-ray) was used (Figure 4A). However, almost all cells died after RT and PDT when the sensitizer W₁₈O₄₉ was replaced by W₁₈O₄₉@EP (W₁₈O₄₉@EP + Laser + X-ray), suggesting that W₁₈O₄₉@EP exerted a very significant effect when used in a combined treatment. This result might be due to the additional effects of EP in the material. The above-mentioned comparison experiments revealed that the ROS released by EP significantly enhanced the therapeutic effect of PDT and RT.

A series of in vitro experiments was performed to further elucidate the enhanced therapeutic effect of W₁₈O₄₉@EP NPs and to explore the underlying mechanism of action for its enhanced therapeutic effect. The amount of ROS production in the cells was significantly enhanced after the treatment with W₁₈O₄₉@EP NPs. Since ROS production was the main reason for RDT, the increased ROS production exerted an enhanced effect, representing an important aspect for the successful effect of RT (Figure 4A). Moreover, the hypoxia status in tumor cells was greatly improved after the treatment with W₁₈O₄₉@EP NPs, and almost no hypoxia was observed in the “W₁₈O₄₉@EP + Laser” group and “W₁₈O₄₉@EP + Laser + X-ray” group (Figure 4B). The hypoxia status after different treatments was also evaluated by staining the important hypoxia marker hypoxia inducible factor-1α (HIF-1α) (Dachs and Tozer, 2000; Liu et al., 2017), which further indicated that the hypoxia status after the treatment with W₁₈O₄₉@EP NPs was greatly reduced (Figure 4C). Since hypoxia is an important limiting factor for the effect of PDT and RT, overcoming hypoxia was another factor to consider for improving the treatment effect. Taken together, the therapeutic effect of W₁₈O₄₉@EP NPs was greatly improved through the above two approaches.

Ki67 is an important marker of tumor progression (Scholzen and Gerdes, 2000) to prove that our treatment involves the ability to inhibit the proliferation of the tumor (Figure 4D), and the results further proved that the treatment with W18O49@EP + Laser + X-ray effectively inhibited tumor proliferation. The combined treatment group (W₁₈O₄₉@EP + Laser + X-ray) was the one that showed the most reduced Ki67 expression, which proved that the combination treatment group was effective in inhibiting tumor proliferation.

Next, NaN₃ was used to remove ROS to further confirm that the improvement of the treatment effect in the combination treatment group was mainly caused by an increase in ROS production during the treatment due to the introduction of EP. Our results showed that the treatment effect was greatly inhibited by adding NaN₃ (Figure 4E), which further proved the role of ROS production in the combination therapy of RT and PDT. Therefore, W₁₈O₄₉@EP NPs, which were capable of self-supplying ROS, had a significant advantage in the combination therapy of RT and PDT.

Since W₁₈O₄₉ NPs had a CT enhancement effect (Kolemen et al., 2016), both the W₁₈O₄₉ NPs and the W₁₈O₄₉@EP NPs were significantly enhanced at the tumor site, indicating that W₁₈O₄₉@EP NPs had enhanced CT angiography (Figure 5). These results also indicated that W₁₈O₄₉@EP NPs were enriched at the tumor site through passive targeting, thereby providing the possibility for subsequent in vivo treatment.

The anti-tumor effect in vivo was evaluated on HCC-4 tumor bearing mice. After the tumor volume reached 100 mm³, mice were subjected to different treatments as shown in Figure 6. The body weight of all mice did not significantly change during the treatment (Figure 6A), which indicated that the treatment was not toxic to the mice. Figure 6B shows that the tumor volume after different treatments was reduced in the combination treatment group, and the tumor almost disappeared after 2 weeks of treatment. The tumor growth was inhibited to some extent in the other treatment groups, but the effect was not significant. Figure 6C shows the morphological changes of tumor tissues in different treatment groups. We found that in the ‘‘W₁₈O₄₉@EP Laser + X-ray” group, there was a lot of necrosis of the tumor cells, the cells were no longer tightly connected and the nuclei were huddled together. Indicating that the combination therapy was the most effective and effectively kills tumor cells. These findings were consistent with the results of the in vitro experiments, which also indicated that W₁₈O₄₉@EP NPs exerted an excellent therapeutic effect through the combination therapy.

Although W₁₈O₄₉@EP NPs exerted a very good therapeutic effect both in vitro and in vivo, their long-term in vivo compatibility should be evaluated prior to clinical applications. Therefore, a pathological evaluation was performed on several major organs of normal mice that received W₁₈O₄₉@EP NPs to test their potential cytotoxicity. The morphology of cells in these five organs was hardly affected, indicating that the above NPs did not affect these organs (Figure 7A). In general, changes in blood cells are used to measure whether the various functions in the body are affected. The results showed that the treatment with these NPs hardly affected the content of various blood cells in the body of mice, as well as the function of some organs (Figures 7B,C). Thus, these results indicated that W18O49@EP was safe and feasible for in vivo treatment.