Saos-2, L929, and MC3T3-E1 cell lineage were purchased from the National Collection of Authenticated Cell Cultures (CAS, China). Saos-2 and L929 cell lines were cultured in Dulbecco’s Modified Eagle Medium (DMEM, Corning, United States) supplemented with 10% fetal bovine serum (FBS, Gibco, United States) and 1% (v/v) anti-anti. Similarly, MC3T3-E1 was cultured in Minimum Essential Medium α (MEMα, Viva cell bioscience, China) supplemented with 10% fetal bovine serum (FBS, Gibco, United States) and 1% (v/v) anti-anti. The cells were incubated at 37°C and 5% CO₂ in a 95% humidified cell incubator until 70%–80% confluency. Cells were dissociated with 0.25% trypsin-EDTA and seeded in microwell plates (MWP) for subsequent biological analysis.

The extract of undoped TNA and 0.5 wt%, 5 wt%, and 22 wt% nano Se-doped TNA were obtained according to the international standard ISO 1099325. Briefly, samples were immersed in DMEM supplemented with 10% FBS under 37°C and 5% CO₂ for 72 h. L929 cells (1 × 10³ cell/well) were seeded in 96 MWP with 150 μL of above supplemented extract and incubated for 48 h under 37°C and 5% CO₂ atmosphere. DMEM medium with 10% FBS served as control group. 20 μL of 5 mg/mL of 3-(4,5-dimethylthiazol-2)-2,5-diphenyltetrazolium bromide (MTT, Sigma, United States) was added into each well, followed by incubation at 37°C for 3 h to determine succinate dehydrogenase activity.

After the mixture of medium and MTT were removed, the methanogenic product was solubilized in 100 µL of isopropanol solution containing 0.04 mol/L HCl. Absorption at 570 nm was measured using an automated plate reader (Perkin-Elmer) for quantitative detection.

Undoped TNA and nano Se-doped TNA (0.5 wt%, 5 wt%, 22 wt%) served as test groups, and Triton X-100 (1%) and 0.9% normal saline (NS) were employed as positive control group (PC) and negative control group (NC), respectively. Each sample was put into a centrifuge tube, and then red blood cell (RBC) saline resuspension was added into each tube. All specimens were incubated at 37°C for 3 h and then were centrifuged at 1,600 rpm for 8 min to obtain the supernatant. The absorbance of the supernatant was measured at 545 nm. The hemolysis rate was calculated according to formula (1). H e m o l y s i s   r a t e   % = A s − A n c A p c − A n c × 100 % (1) Where A_s, A_nc, and A_pc were the absorbance values of the experimental group, negative control group, and positive control group, respectively.

The effect of nano Se-doped TNA with different Se concentrations (0.5wt%, 5wt%, 22wt%) on the proliferation of Saos-2 cells was tested using the AlamarBlue reagent (Thermofisher, United States) according to the manufacturer’s instructions. Briefly, undoped TNA and 0.5 wt%, 5 wt%, and 22 wt% nano Se-doped TNA were placed in a 96 MWP. 20 μL of AlamarBlue was added into 1 × 10⁵ cells/mL Saos-2 cell suspension to mix well, and then 180 μL cell suspension was added into each well, followed by incubation for 24 h at 37°C and 5% CO₂ atmosphere. The reduction of AlamarBlue was measured by microspectrophotometer and the reduction rate of AlamarBlue was calculated using the formula (2) provided by the manufacturer’s protocol. R e d u c e d   % = ε o x λ 2 ⋅ A λ 1 − ε o x λ 1 ⋅ A λ 2 ε R e d λ 1 ⋅ A ’ λ 2 − ε R e d λ 2 ⋅ A ’ λ 1 × 100 % (2) Where λ₁ = 570 nm, λ₂ = 600 nm; ε_ox was the molar extinction coefficient of the oxidized form of AlamarBlue (blue), ε_ox 570 = 80586, ε_ox 600 = 117216; ε_red is the molar extinction coefficient of the reduced form of AlamarBlue (pink), ε_{Red 570nm} = 155677, ε_{Red 600nm} = 14652. A indicates the absorbance value, and A′ denotes the absorbance value of the negative control.

We also performed Cell Counting Kit-8 (CCK-8, Bimake, China) assay to investigate the effect of nano Se-doped TNA with different Se concentrations (undoped, 0.5 wt%, 5 wt%, 22 wt%) on the proliferation of MC3T3-E1 cells. MC3T3-E1 cells (2 × 10³ cell/well) were cultured on the surface of undoped, 0.5 wt%, 5 wt%, 22 wt% nano Se-doped TNA and blank well (control group) for 24 h.

MC3T3-E1 (3 × 10³cells/well) and Saos-2 (3 × 10³cells/well) were cultured on undoped TNA and 0.5 wt%, 5 wt%, and 22 wt% nano Se-doped TNA for 24 h. MC3T3-E1 and Saos-2 cells cultured on the samples were fixed with 2.5% glutaraldehyde and permeabilized with 1% Triton X-100. F-actin and nuclei were stained with Fluorescein isothiocyanate (FITC)-phalloidin and DAPI (Solarbio, Beijing, China), respectively, according to the manufacturer’s protocol. Cellular morphology was observed by confocal laser scanning microscopy (CLSM, LEICA TCS SP8, Germany).

The exact percentage of the apoptotic Saos-2 cells was analyzed with Annexin V- FITC/PI staining. Saos-2 cells were cultured onto the surface of undoped TNA and 0.5 wt%, 5 wt%, and 22 wt% nano Se-doped TNA (3 × 10⁵/well) for 24 h, and then the Saos-2 cells were collected and stained with Annexin V- FITC/PI (Solarbio, Beijing, China). Flow cytometer (FCM, Beckman Coulter, United States) was used to determine the fluorescence signal of Saos-2 cells.

Saos-2 cells (10⁴cells/well) were seeded on the surface of undoped TNA, 0.5 wt%, 5 wt%, and 22 wt% nano Se-doped TNA, incubated for 24 h, and then stained with Hoechst 33258 (Solarbio, China) and acridine orange/ethidium bromide (AO/EB) according to manufacturer’s protocols, respectively. Apoptotic cells were examined using a fluorescent microscope (Olympus, Tokyo, Japan) within 30 min.

ROS Assay Kit (Beyotime, Shanghai, China) was used to detect the levels of ROS production following the manufacturer’s protocol. Briefly, Saos-2 cells cultured on the surface of TNA groups for 24 h were collected by trypsin digestion and stained by 2,7-Dichlorodihydrofluorescein diacetate (DCFH-DA) ROS sensitive probe. The intracellular ROS content was evaluated using flow cytometer (FCM, Beckman Coulter, United States) and confocal laser scanning microscopy (CLSM, LEICA TCS SP8, Germany).

Saos-2 cells (4 × 10⁵/well) were cultured onto the surface of undoped TNA and 0.5 wt%, 5 wt%, and 22 wt% nano Se-doped TNA for 24 h. The total RNA of saos-2 cells was extracted using TRIzol Reagent (Thermofisher, United States), and the complementary DNA (cDNA) was synthesized using the total RNA with the TransScript^® All-in-One First-Strand cDNA Synthesis SuperMix for qPCR (TransGen, China) according to manufacturer instructions. The expression levels of the apoptosis-related genes (BAX, CYTC, CASP9, CASP8, CASP3) were detected by reverse transcription-polymerase chain reaction (RT-PCR) assay, and the primers’ sequences were shown in Table 1. GAPDH was chosen as the reference gene in this study.

All statistical analyses were performed using SPSS 20.0 (SPSS, Chicago, IL, United States). Data was expressed as mean ± standard deviation (SD) (n = 3). Differences among groups were analyzed and determined by Student’s t-test or one-way ANOVA followed by Tukey’s post hoc test or Bonferroni post hoc test. The p-value less than 0.05 was considered statistically significant.

Figure 1A shows the preparation process of nano Se-doped TiO₂ nanotube arrays: TiO₂ nanotube arrays are fabricated on the surface of pure titanium by anodization, and then, nano Se is deposited on TiO₂ nanotube arrays by electrochemical deposition. The surface morphology and composition of nanostructured materials is crucial to the actual function of the material. The surface morphology of undoped TNA and nano Se-doped TNA are shown in Figure 1B. The undoped TNA are highly ordered and uniform in size, and the diameter of nanotubes is about 120 nm and the length of nanotubes is about 400 nm. The surface of 0.5wt% nano Se-doped TNA with 2 min deposition is covered by some nano scaled fine deposits with sizes around 100 nm, and the structure of TNA remained intact. It can be observed that the surface of 5wt% nano Se-doped TNA with 4 min deposition is locally covered by nanoflakes of irregular shape and size about 100 nm, which evenly and individually disperse on the surface of the TNA, and the flakes are confirmed to be Se elements by EDX test. There are still some uncovered parts of TiO₂ nanotube array, and the morphological structure remained clear and consistent. When the electrochemical deposition time was extended to 7 min, a layer of reddish-brown deposit on the surface of the sample could be observed by naked eye, and the surface of 22 wt% nano Se-doped TNA was completely covered and the nanotubes were not visible. The deposited Se agglomerated into 300–400 nm lumpy particles and there were very obvious cracks among the lumpy particles. When the deposition time was 10 min, the surface morphology of the TiO₂ nanotube array was similar to 22 wt% nano Se-doped TNA.

Lowering the pH value of the electrolyte enables the hydrolysis reaction to proceed more rapidly and adequately. Moreover, we believe that the special porous structure of the TiO₂ nanotube array also facilitates the deposition of selenium. The porous structure of TiO₂ nanotube array provides a fixation point for the deposition and merger of Se. At the initial stage of electrochemical deposition, nano Se particles are attached and immobilised at certain points on the surface of the TiO₂ nanotube arrays, and these nanoscale particles gradually agglomerate to form blocks and sheets.

The EDX spectra of the surface of nano Se-doped TNA are shown in Figures 1C, D. In addition to the peaks of Ti, O, and F elements, there is Se peak in the spectrum, which proves that Se has been successfully deposited on the TiO₂ nanotube arrays. The content of various elements of nano Se-doped TNA fabricated in this work is shown in Table 2. Nano Se-doped TNA mainly contains Ti, O and Se elements. The selenium content is 0.53 ± 0.03wt% (0.5 wt% nano Se-doped TNA), 5.17 ± 0.11 wt% (5wt% nano Se-doped TNA), and 22.19 ± 0.19wt% (22 wt% nano Se-doped TNA), respectively. The XPS spectrum of 5 wt% nano Se-doped TNA are presented in Figure 1E. XPS spectra in Figure 1F show two peaks at 55.21 eV and 55.9 eV corresponding to Se 3d_5/2 and Se 3d_3/2, the binding energies are similar to the 3d_5/2 and Se 3d_3/2 peaks of monatomic Se at 55.2 eV and 55.9 eV, respectively, and the results suggest that Se is deposited as a monomer in the whole row of TiO₂ nanotube arrays (Shalvoy et al., 1977; Shenasa et al., 1986). Supplementary Figure S1 shows a typical X-ray diffraction pattern of 5 wt% nano Se doped TNA, which reveals that the polymorph of the obtained nano Se is amorphous Se (Li et al., 2010). The concentration and morphology of the deposited Se element on the TiO₂ nanotube arrays can be adjusted by changing the electrochemical deposition time and the concentration of the electrolyte. Figure 1G depicts the trend of Se concentration deposited on the TiO₂ nanotube arrays with increasing deposition time when the electrolyte concentration is fixed (taking 1.5 × 10⁻³ mol/L Na₂SeO₃, for example). It is clear that the Se concentrations are positively associated with increasing deposition time: Se content is around 0.5wt% when the deposition time is 2 min, and Se content is around 5 wt% when the deposition time is 4 min, and around 22 wt% of Se when the deposition time is 7 min, and around 24 wt% of Se when the deposition time is 10 min. It can also be seen from Figure 1G that the deposition rate of Se on the nanotubes at the beginning stage is higher than that at the later stage. Taking the results (Figures 1C, G) into consideration, we speculate that nano Se is scattered on the surface of the nanotubes at the beginning, and then the dispersed nano Se grow up gradually and interconnect when the deposition time exceeds a certain value. Eventually, Se agglomerates to cover the whole surface of TNA, and the surface of TNA shows obvious reddish-brown color. As the deposition time continues to extend, the reddish-brown particles floating in the electrolyte around the sample could be observed by nude eye, indicating the deposition of Se on the surface of nanotubes has reached saturation state. Figure 1H shows the trend of the deposition concentration of Se on the TiO₂ nanotube arrays with the electrolyte concentration when the deposition time is fixed at 7 min. The concentration of Se is around 1wt% when the electrolyte concentration is 0.25 × 10⁻³ mol/L, and 5wt% of Se when the electrolyte concentration is 0.9 × 10⁻³ mol/L, and 22wt% of Se when the electrolyte concentration is 1.5 × 10⁻³ mol/L, and Se concentration is as high as 47wt% when the electrolyte concentration is 3 × 10⁻³ mol/L, which indicates that the concentration of Se deposited on the surface of TNA increases as the electrolyte concentration increases. The deposition rate of Se in the high-concentration electrolyte is much higher than that in the low-concentration electrolyte. The floating reddish-brown particles are visible in the electrolyte around the sample, implying Se deposition on the surface of TNA reaches saturation.

A novel biomedical material must possess good biocompatibility (Shirzaei Sani et al., 2019). As an potential orthopedic implant material, the nano Se-doped TNA will inevitably contact with blood. Therefore, we conduct hemolysis tests to assess the hemocompatibility of nano Se-doped TNA. The standard specifies the materials with hemolysis rate (HR) of 2%–5% and >5% as slightly hemolytic and hemolytic, respectively, whereas the materials with 0%–2% of hemolysis are categorized as non-hemolytic (Leitão et al., 2013). As shown in Figure 2A, the bright red represent the positive control group (PC), whereas all TNA groups appear colorless, similar to the negative control group (NC). The hemolysis rates of all TNA groups are less than 2%, indicating that nano Se-doped TNA has excellent hemocompatibility. Moreover, the cytocompatibility of the nano Se-doped TNA is evaluated by MTT testing the enzymatic activity of mouse-derived fibroblasts (L929) cultivated into the extracts of all TNA samples for 48 h. There is no significant difference in the enzymatic activity among all TNA groups and the control group, indicating that the extracts of 0.5wt% Se, 5wt% Se, and 22wt% Se nano Se-doped TNA have no cytotoxic effect (Figure 2B). Overall, the increased content of deposited nano Se does not affect the hemocompatibility and cytocompatibility of TNA.

The mouse preosteoblast cell line (MC3T3-E1) and the human osteosarcoma cell line (Saos-2) are cultured on the surface of undoped TNA and nano Se-doped TNA to investigate the effect of nano Se-doped TNA on the proliferation of MC3T3-E1 and Saos-2. Figure 3A shows the proliferation of MC3T3-E1 cultured on the surface of nano Se-doped TNA (test groups) and DMEM cell culture media (control group). There is no any significant difference in the proliferation of MC3T3-E1 cells between all groups, which is consistent with the results of Figure 2B. As seen in Figure 3B, the proliferation of Saos-2 cell seeded on 0.5wt% and 5wt% nano Se-doped TNA is significantly lower than that of the undoped TNA and 22wt% Se-doped TNA. Among them, Saos-2 cells on the 5 wt% nano Se-doped TNA show the lowest active proliferation. Furthermore, fluorescence staining is carried out to observe the cell morphology cultured for 24 h on the surface of undoped TNA and nano Se-doped TNA (Figure 3C). MC3T3-E1 cells on all groups show very high cell confluence and intact morphology. However, Saos-2 cells on 0.5 wt% and 5 wt% nano Se-doped TiO₂ TNA exhibit extremely low cell confluence and cytoskeletal abnormalities, compared with undoped TNA and 22 wt% Se-doped TNA, demonstrating that 0.5 wt% and 5 wt% nano Se-doped TiO₂ TNA significantly inhibit the proliferation of Saos-2 cell lineage.

The surface morphology also alters the celluar response. For MC3T3-E1 cells, the unique structure of TiO₂ nanotubes facilitates proliferation, adhesion and differentiation of preosteoblasts (Brammer et al., 2012). Therefore, MC3T3-E1 adhered easily on the surface of the nanotube incompletely covered group (0.5 wt% and 5 wt% nano Se-TNA), while 22 wt% nano Se-doped TNA was a rough surface, and previous studies have shown that roughness is favourable for cell proliferation and adhesion (Faia-Torres et al., 2014; Han et al., 2020). Compared with the Control group, the number of cells, cellular integrity, and adhesion in the Se-doped TNA group did not differ significantly. In addition, nano-selenium had no significant effect on MC3T3-E1 proliferation, and adhesion. There was no significant difference in cell number, cell integrity and adhesion in the Se-doped TNA group compared with the Control group, nano Se had no significant inhibitory effect on MC3T3-E1 (Cheng et al., 2017). For Saos-2 cells, compared with the uniform surface of TiO₂ nanotube array, the 0.5wt and 5wt% nano Se-doped TNA surfaces were covered with un-interconnected tiny flakes of nano selenium, and osteosarcoma cells planted on the 0.5wt and 5wt% nano Se-doped TNA surfaces would be up taken by the cell in the cytosolic form of nano selenium and eventually induced a significant inhibitory effect on Saos-2 cells (Dai et al., 2021). Around the organelles to play a role, and eventually induced apoptosis in Saos-2 cells. On the other hand, the surface of 22 wt% nano Se-doped TNA was covered with interconnected block particles of nano selenium, which prevented the tumor cells from uptaking the block particles of nano selenium, so the inhibition of tumor cells by nano Se at this concentration was relatively small (Zhang et al., 2009).

Previous studies have shown that the mechanism of Se inhibiting tumor cell viability is mainly inducing the apoptosis of tumor cells (Wang et al., 2016; Li et al., 2020), particularly those mediated by ROS (Chen et al., 2012). Here, we hypothesize that nano Se-doped TNA may effectively inhibit tumor cells proliferation by promoting cancer cell apoptosis. Therefore, we used Hoechst 33258 staining, AO/EB double staining, and Annexin V-FITC/PI double staining flow cytometry to verify our hypothesis. The AO/EB double-stained fluorescence microscopy images (Figure 4A) show the effect of the nano Se-doped TiO₂ nanotube arrays on the viability and apoptosis of Saos-2 cells. The living cells cover the whole surface of undoped TNA group in Figure 4A, indicating that Saos-2 cells have high viability on the surface of the undoped TNA. The apoptotic and dead cells almost occupy the whole surface of 0.5wt% and 5wt% Se-doped TNA groups and the living cells are almost undetectable, indicating that the Saos-2 cells on these two groups are hard to getting survive, and these two groups significantly reduce the survival rate of Saos-2 cells (Figure 4A). The live, apoptotic and dead Saos-2 cells coexist on the surface of 22 wt% Se-doped TNA group, but the living cells are still in the majority, indicating that the survival rate of Saos-2 cells on the 22 wt% Se-doped TNA group is higher than that on the 0.5 wt% and 5 wt% Se-doped TNA groups, but still lower than that on the undoped TNA group. It can be concluded that the nano Se deposited on the surface of TNA possesses a strong inhibitory effect on the viability of Saos-2 cells, and however, higher Se doping contents do not exert a stronger inhibitory effect on the viability of Saos-2 cells. Only an appropriate concentration of doped nano selenium has a significant inhibitory effect. The results of Figure 4A are consistent with those shown in Figures 3B, C.

We further investigate the nuclear morphology of Saos-2 cells stained with Hoechst 33258 by fluorescence microscopy (Figure 4B). The nuclei of Saos-2 cells on all nano Se-doped TNA groups exhibit brighter blue staining and more concentrated morphology compared to that on the undoped TNA, indicating the deposited nano Se enhances the cell apoptosis. The cell nuclei on the surface of 0.5 wt% and 5 wt% nano Se-doped TNA groups show brighter blue staining and more concentrated morphology than that on 22 wt% Se-doped TNA, implying that an appropriate concentration of doped nano Se significantly promotes cell apoptosis.

Next, we implement Annexin V-FITC/PI double-staining assay in order to quantify the apoptosis rate of Saos-2 cells cultured on the TNA samples. The apoptosis rate of Saos-2 cells on undoped TNA, 0.5 wt%, 5 wt%, and 22 wt% nano Se-doped TNA are 9.81%, 26.9%, 52.3%, and 16.19%, respectively (Figure 4C). Obviously, the apoptosis rate of 0.5 wt% and 5 wt% nano Se-doped TNA is significantly higher than that of undoped TNA and 22 wt% nano Se-doped TNA. It is concluded that the inhibitory effect of nano Se-doped TNA on osteosarcoma does not follow a dose-dependent behavior.

From the above results of the biological experiments, we suggest that the mechanism of nano Se inhibiting tumor cell proliferation is mainly the induction of cell apoptosis. It can be seen that the proliferation of Saos-2 cells on the surface of the nano Se-doped TiO₂ nanotube arrays is significantly inhibited compared to that of the undoped TiO₂ nanotube arrays. In this study, 0.5 wt% and 5 wt% nano Se-doped TiO₂ nanotube arrays exhibit strong inhibition towards Saos-2 cell activity, especially for nano Se content of 5 wt%. The inhibitory effect of 22 wt% nano Se-doped TNA on osteosarcoma cells is less than that of lower concentration nano Se doped TNA. We believe that this inhibitory effect depends heavily on the content of the deposited nano Se and should be strongly related to the surface morphology and diameter of nano Se. We assume the selenium ions in the electrolyte are reduced to red elemental nano selenium, and the unique morphology of TNA guides the deposition of red elemental nano selenium. When the nano Se-doped concentration is 0.5 wt% and 5 wt%, the deposited nano Se is distributed on the surface of the nanotubes as nanoflakes with a diameter of about 100 nm, and these nanoflakes are not connected with each other, and EDS verifies that these nanoflakes are Se. Subsequently, Saos-2 cells ingested the nanoflakes into the cells to exert anti-tumor effects through endocytosis (Dai et al., 2021); when the doped nano Se concentration is 22 wt%, nano Se agglomerates into particles with a diameter of about 400 nm, covering the whole surface of the nanotube array (Figure 1B), due to the large size of the nanoparticles, tumor cells can only absorb trace amounts of nano-selenium on the surface of the nanotubes, resulting in relatively fewer apoptotic Saos-2 cells than the other two groups (Zhang et al., 2009). Therefore, we believe that nanoflake morphology of nano Se exhibit better anti-tumor effects than agglomerated Se particles. In addition, the effects of selenium on tumor cells varies within different types and numbers of tumor cells (Dennert and Horneber, 2006).

Previous studies have demonstrated that the antitumor effects of selenium compounds are mainly due to ROS produced by intracellular metabolism (Menon et al., 2018). Hence, we hypothesize that the reason why nano Se-doped TiO₂ nanotube arrays induce apoptosis of osteosarcoma Saos-2 cells may be the production of ROS and the subsequent activation of endogenous and exogenous caspase-dependent pathways.

To verify this hypothesis, we first measured the fluorescence intensity of 2′,7′-dichlorofluorescein (DCF) to reflect the intracellular ROS level after Saos-2 cells culturing on the surface of all TNA groups for 24 h. As shown in Figure 5A and Figure 5B, the ROS levels of the 0.5 wt% nano Se-doped TNA and 5wt% nano Se-doped TNA groups are higher than those of the undoped TNA and 22 wt% nano Se-doped TNA groups, and the ROS levels of the 5 wt% nano Se-doped TNA group are significantly higher than those of the other three groups.

It is well known that caspase-3, as a cell apoptotic executor, can be activated by both caspase-8 and caspase-9. Caspase-8 is the apoptosis promoter of the exogenous pathway (death receptor pathway), whereas caspase-9 is the apoptosis promoter of the endogenous pathway (mitochondrial pathway) (Mata et al., 2016). To demonstrate the potential caspase-dependent apoptosis mechanism, RT-PCR is used to detect the expression levels of pro-apoptotic genes of endogenous and exogenous caspase-dependent pathways after Saos-2 cells culturing on all TNA groups for 24 h. As shown in Figure 5C, the expression of promoting apoptosis genes (BAX, CYTC, CASP9, CASP8, CASP3) of 0.5 wt% nano Se-doped TNA group and 5%wt nano Se-doped TNA group is remarkably upregulated, compared with undoped TNA and 22 wt% nano Se-doped TNA group. In this study, we reveal a potential intrinsic mechanism, namely caspases-dependent apoptosis, regarding the inhibitory effect of nano selenium-doped TNA on tumor.

In summary, we can provide a clearer picture of the mechanism of nano Se-doped TNA inducing apoptosis of Saos-2 cells (Figure 6). The nano Se-doped TNA enters osteosarcoma cells via the endocytic pathway (Zhang et al., 2009; Wang et al., 2016; Dai et al., 2021). Subsequently, nano Se-doped TNA inhibits osteosarcoma cell proliferation by inducing the production of ROS, which subsequently activates two caspase-dependent apoptotic pathways, endogenous and exogenous, leading to apoptosis of osteosarcoma cells. On the one hand, ROS production activates caspase-8 and triggers the death receptor-mediated exogenous apoptotic pathway, while on the other hand, activation of the Bcl-2 family pro-apoptotic member (BAX) causes mitochondrial outer membrane permeabilization, resulting in the release of cytochrome C and effective activation of the endogenous apoptotic pathway, and ultimately resulting in the apoptosis of osteosarcoma cells.