Baicalin induces apoptosis and autophagy in human osteosarcoma cells by increasing ROS to inhibit PI3K/Akt/mTOR, ERK1/2 and β-catenin signaling pathways

Highlights • Baicalin causes apoptosis and autophagy through accumulating ROS to suppress PI3K/Akt/mTOR, ERK1/2 and β-catenin pathways in OS cells.• Baicalin-induced autophagosome further triggers apoptosis.• Baicalin-induced ROS and Ca2+ interactions induce apoptosis.• Baicalin molecule targets PI3Kγ, inhibiting downstream effectors AKT and mTOR.


Introduction
Osteosarcoma (OS) is a common malignant bone cancer with the highest incidence in adolescents and occurs in the epiphysis of long bones [1,2]. The median age of osteosarcoma patients is 16 years, which appears to be associated with growth spurts [3,4]. At present, despite advancements in chemotherapy and surgery, the 5-year survival rates for primary and metastatic osteosarcoma remain<65% and 25%, respectively, with few treatment options [5,6]. Thus, it is essential to explore practical and safe therapeutic agents for osteosarcoma. Baicalin, a flavonoid derivative compound, is obtained from the root of Scutellaria. In recent reports, baicalin was found to possess antioxidant and antiinflammatory properties and is virtually nontoxic to healthy tissues [7,8]. Data indicate that baicalin inhibits tumor growth and metastasis in various cancer cells [9][10][11]. These effects are transduced by reactive oxygen species-mediated mitochondrial pathways and via inhibition of the AKT pathway [12][13][14][15]. Although the mechanism underlying baicalin's inhibition of osteosarcoma cells can partially be attributed to autophagy, apoptosis and the suppression of multiple signaling pathways, their interplay remains unclear. Therefore, understanding baicalin-mediated changes and their underlying mechanisms may lead to identification of an effective candidate for osteosarcoma treatment.
Disturbed regulation of the cell cycle can promote the occurrence and development of tumors [16,17]. Many cytotoxic agents have been found to cause cell cycle S-phase arrest, promoting cell death [18,19]. The characteristic manifestations of apoptosis are nuclear fragmentation, cell shrinkage and apoptotic vesicle development. [20,21]. In particular, a recent report showed that numerous chemotherapeutic agents regulate the family members of Bcl-2 to induce mitochondrial apoptosis [22]. The Bcl-2 protein family is considered to be the crucial factor controlling mitochondrial membrane permeabilization. Apoptosis has played a significant role in osteosarcoma treatment during the past three decades, and elemental signaling pathways have been extensively investigated. The cell death mechanism type II, known as autophagy, which is a ''self-feeding" process, includes the formation of autophagic vesicles and the decomposition of cytoplasmic components [23]. Autophagy plays a dual role in tumorigenesis and progression because it promotes cell survival to prevent apoptosis in some cellular environments, while it induces cell death in others [24,25]. The association between autophagy and apoptosis is complex due to differences in cell types and environmental stimuli. Autophagy is activated and regulated by reactive oxygen species and Ca 2+ concentration [26,27]. Recent research has shown that a significant number of apoptosis-inducing cancer chemotherapy drugs also activate autophagy [28,29]. Whether baicalin induces apoptosis or autophagy remains to be determined.
Endogenous and exogenous factors activate reactive oxygen species (ROS), which are involved in regulating biological activity [30]. In the presence of external stimuli beyond a certain degree, the production of excess ROS damages cellular integrity due to a disrupted balance between the generation and metabolism of ROS [31]. The damage of ROS to cells is primarily reflected in the damage to biomacromolecules, such as oxidative damage to DNA, proteins, and biofilms [32]. An increasing number of reports indicate that reactive oxygen species have a critical function in the induction of apoptosis and autophagy [33,34]. Moreover, reactive oxygen species affect multiple signaling pathways [35]. Recent studies have shown that ROS regulate the PI3K/Akt/mTOR, ERK1/2 and b-catenin signaling pathways. These pathways impact the cell cycle, apoptosis, invasion and autophagy induction [36][37][38]. Intracellular calcium concentration is intimately linked to reactive oxygen species generation [39]. Extensive investigations have revealed that elevated intercellular Ca 2+ levels induce endoplasmic reticulum stress and constitute a significant proapoptotic signal [40]. In addition, Ca 2+ is also a stimulator of apoptosis and autophagy [41,42]. In this study, we explored the antiosteosarcoma mechanism of baicalin in vitro. In addition, we investigated baicalin-induced cell cycle arrest, cell death and potential signaling pathways.

Cell culture
The HOS, MG63 U2OS and 143B human osteosarcoma cell lines (Chinese Academy of Sciences Cell Bank) were cultured in MEM containing 10% fetal bovine serum (Gibco, Carlsbad, USA) and 0.5% penicillin and streptomycin (Gibco, Sydney, Australia) and maintained in a cell incubator with 95% humidity at 37°C and 5% carbon dioxide.
Cell viability was measured using Cell Counting Kit-8 (CCK-8) reagent (Beyotime, Shanghai, China). After incubation for 2 h, the absorbance of the cells was measured at 450 nm.

Colony formation assay
HOS, MG63, U2OS, and 143B osteosarcoma cells (approximately 800 cells/well) were cultured in 6-well plates for two weeks with different concentrations of baicalin. At the end of cell culture, cells were fixed in methanol for 15 min and then stained with crystal violet staining solution at room temperature for 10 min. Colonies containing > 50 cells were counted under a microscope.

Invasion assay
A chamber was coated with Matrigel matrix (Corning) and placed in an incubator for 30 min. Then, 100 mL of MEM containing a total of 2 Â 10 4 cells was added. After incubation for 24 h, we removed cells that did not penetrate the membrane inside the chamber. Cells inside the membrane were immobilized with methanol for 15 min and then stained with crystal violet staining solution at room temperature for 10 min. The average cell numbers of six microscopic fields of view (100 Â ) were counted using Ima-geJ software.

Cell cycle analysis
HOS and 143B cells were treated with baicalin for 48 h. Next, cells were collected and fixed in frozen ethanol for 12 h. Next, cells were stained with propidium iodide (PI) (Beyotime, Shanghai, China) for 30 min at room temperature and then analyzed by flow cytometry.

Apoptosis analysis by flow cytometry
Apoptosis was evaluated using the Annexin V-FITC/PI Apoptosis Detection Kit (Beyotime, Shanghai, China). HOS, MG63, U2OS, and 143B cells were trypsinized after different treatments and collected. Then binding buffer was added to the cells, and they were incubated in the dark for 30 min before evaluation by flow cytometry.

Mitochondrial membrane potential (MMP) assay
Osteosarcoma cells were subjected to various treatments for 48 h. Subsequently, we collected and washed the cells. Next, the JC-1 working solution of the MMP Assay Kit (Beyotime, Shanghai, China) was used to treat the cells for 30 min, which were finally examined by flow cytometry. JC-1 is widely used as an ideal fluorescent probe for detecting mitochondrial membrane potential. The mitochondrial membrane potential of normal cells is high, so JC-1 accumulates in the matrix of mitochondria and form polymers (J-aggregates), resulting in the production of red fluorescence. In contrast, when the mitochondrial membrane potential of cells is low, JC-1 cannot accumulate in the matrix of mitochondria, resulting in increased green fluorescence, allowing convenient detection of changes in mitochondrial membrane potential through the transition of fluorescence color.

ROS assay
The ROS Assay kit (Beyotime, Shanghai, China) was used to assess intracellular reactive oxygen species production. Osteosarcoma cells were subjected to different treatments for 48 h. Subsequently, the cells were collected and stained with 10 lM DCFH-DA for 30 min while protected from light and eventually measured by flow cytometry.

Ca 2+ assay
To analyze intracellular Ca 2+ concentrations, approximately 2 lM of the Ca 2+ indicator Fluo-4 AM (Beyotime, Shanghai, China) was loaded into HOS and 143B cells, and the cells were then incubated with a fluorophore for 30 min. Finally, the Ca 2+ content was estimated using flow cytometry.

Western blot analysis
Total protein was extracted from cells in each group. Protein concentration was measured and subjected to gel electrophoresis. After membrane transfer and blocking, we added a primary antibody overnight at 4°C. Then, the membrane was washed with TBST and incubated with secondary antibody. Finally, immunoreactive bands were measured using a chemiluminescence and fluorescence imaging system. The intensity of the target protein bands was normalized to the intensity of GAPDH bands, and ImageJ software was used to perform band grayscale analysis.

Molecular docking
To investigate the interaction between baicalin and the PI3Kc active site, Discovery Studio software (DS 2019, Accelrys, CA, USA) and AutoDock Vina1.1.2 software were used to conduct automated molecular docking. The three-dimensional crystal structure of PI3K gamma (PI3Kc) was complexed with the inhibitor from PDB (PDB ID: 4XZ4). PyMOL 2.3.0 software was used to remove protein crystalline water, primitive ligands, etc. Subsequently, the PI3Kc structure was loaded into AutoDocktools (v1.5.6) software for hydrogenation, charge assignment and designation of atoms. Molecular docking was conducted using AutoDock Vina1.1.2 software, and finally, the molecular docking results were imported into Discovery Studio2019 for visualization.

Statistical analysis
All data are expressed as the mean ± SD of at least three independent experiments. GraphPad Prism 8.0 software (California, USA) was used for data analysis. Student's t test or one-way ANOVA was used to analyze differences between groups. P*＜ 0.05 were considered statistically significant.

Baicalin inhibits proliferation and induces S-phase blockade in osteosarcoma cells
To investigate the antiproliferative activity of baicalin on osteosarcoma, the human osteosarcoma cell lines HOS, MG-63, U2OS, and 143B were treated with different concentrations of baicalin for 24, 48 or 72 h. Then, cell viability assays revealed that baicalin suppressed osteosarcoma cell proliferation in a concentration-and time-dependent manner (Fig. 1A). Moreover, colony formation assays demonstrated that baicalin treatment decreased colony formation (Fig. 1B). The above data indicated that baicalin reduced the proliferation of osteosarcoma cells. To investigate the relationship between the anti-osteosarcoma effects of baicalin and cell cycle blockade, we analyzed the influence of baicalin on cell cycle progression. After treatment with different concentrations of baicalin (0, 40, or 80 lM) for 48 h, the percentages of S phase cells were 38.8 ± 1.0%, 44.9 ± 1.8%, and 51.9 ± 0.9% in HOS cells and 24.4 ± 2.6%, 28.1 ± 0.9%, and 34.7 ± 0.9% in 143B cells (Fig. 1C). These data indicate that the percentage of HOS and 143B cells in S phase was increased in response to baicalin treatment. Moreover, western blotting results revealed that baicalin treatment significantly decreased c-Myc, cyclin A2, Cdk2 and cyclin E1 expression levels. In summary, these results showed that baicalin causes cell cycle blockade at the S-phase and inhibits proliferation by interfering with cell cycle-and proliferation-related proteins.

Baicalin inhibits the invasive ability of osteosarcoma cells
Invasion assays were performed to determine the invasion of HOS and 143B cells. Osteosarcoma cells were treated with baicalin for 24 h. Then, cells penetrating the polycarbonate membrane were observed by microscopy at 100 Â magnification. The results of the cell invasion assay revealed that the number of invasive cells (% compared to the control group) of HOS and 143B cells was 31.3 ± 7.4% and 13.3 ± 3.2% after treatment with 80 mmol/L baicalin for 24 h (Fig. 2). Overall, our findings reveal that baicalin inhibits the invasive ability of osteosarcoma cells.

Baicalin induces mitochondria-mediated apoptosis in osteosarcoma cells
Commonly, cell cycle inhibition is closely associated with apoptosis. Hence, baicalin-induced apoptosis was further investigated using a flow cytometry assay. To quantify apoptosis, we used an Annexin V-PI double staining method. After treatment with different concentrations of baicalin for 48 h, the percentages of apoptotic HOS cells were 8.1 ± 2.0%, 10.4 ± 4.0%, 20.1 ± 3.5% and 40.9 ± 6.8%. The percentage of apoptosis was 4.3 ± 2.2%, 11.2 ± 7.2%, 18.3 ± 4.5% and 45.3 ± 6.1% in MG63 cells. The percentages of apoptosis were 5.8 ± 1.4%, 14.1 ± 3.1%, 18.5 ± 6.8%, and 27.8 ± 2.3% in U2OS cells. The percentages of apoptotic 143B cells were 8.0 ± 0.7, 8.9 ± 0.7, 14.5 ± 0.9 and 31.2 ± 1.3%. These results demonstrated that osteosarcoma cells displayed an increase in the apoptotic cell ratio in response to baicalin treatment for 48 h (Fig. 3A). A decrease in mitochondrial membrane potential is a hallmark event in the early stages of apoptosis. To explore the influence of baicalin on mitochondrial membrane potential, we measured MMP using JC-1 working solution and flow cytometry analysis. The transition of JC-1 from red fluorescence to green fluorescence indicates a decrease in mitochondrial membrane potential. At the same time, this transition can be used as a detection indicator of the early stage of apoptosis. After treatment with different concentrations of baicalin for 48 h in HOS, MG63, U2OS and 143B cells, the rate of green fluorescence increased to different extents, and the rate of green fluorescence in the 80 mmol/L baicalin treatment group was 57.4 ± 14.0%, 45.4 ± 4.8%, 40.0 ± 6.1% and 46.4 ± 4.8%, respectively. These results indicate that the green fluorescence intensity ratio increased after baicalin treatment for 48 h as determined by flow cytometry (Fig. 3B), suggesting that baicalin induces apoptosis in osteosarcoma cells by decreasing MMP and impairing the integrity of the mitochondrial membrane. Subsequently, we further assessed apoptosis-associated proteins. These results demonstrated that Bcl-2 was sharply reduced in cells after baicalin treatment, while Bax, cleaved caspase-3 and cleaved PARP gradually increased (Fig. 3C). These findings indicate that baicalin induces apoptosis through a mitochondria-mediated apoptotic pathway.

Baicalin triggers autophagy in osteosarcoma cells
To investigate whether autophagy participates in baicalininduced apoptosis, we analyzed the effect of baicalin on autophagy in osteosarcoma cells. LC3-II and P62 are markers of autophagosomes and autophagic activity, respectively. The results indicated that levels of LC3-II and p62 were increased by baicalin treatment (Fig. 4A). Autophagosomes observed by electron microscopy are critical evidence of autophagic activity. In our investigation, transmission electron microscopy revealed that autophagosomes in the cytoplasm were significantly increased after baicalin treatment (Fig. 4B). In summary, our outcomes indicated that baicalin induces autophagosome generation. To further investigate the interaction between baicalin-induced autophagy and apoptosis, we added 3-MA, an autophagy inhibitor, to suppress autophagic activity. A cell viability assay revealed that pretreatment with 3-MA restored baicalin-medicated inhibition of cell viability (Fig. 4C). This was further substantiated by results showing that pretreatment with 3-MA reduced the percentage of baicalin-induced apoptotic cells from 26.8 ± 1.7 to 20.1 ± 1.3% (Fig. 4D). In summary, these results indicate that inhibition of autophagy attenuates baicalin-induced apoptosis.
3.5. Baicalin induces ROS and Ca 2+ generation and blocks the PI3K/Akt/ mTOR, ERK1/2 and b-catenin signaling pathways in OS cells ROS are intracellular and intercellular second messengers that are essential regulators of apoptosis and autophagy. ROS generation is primarily derived from the respiratory chain of mitochondria [43]. Our results showed that baicalin treatment sharply decreased the mitochondrial membrane potential in osteosarcoma cells. Therefore, we further speculated that the release of ROS in mitochondria would be induced by baicalin. To demonstrate this, we performed a ROS assay, which revealed that OS cells exposed to baicalin exhibited augmented intracellular ROS levels in a concentration-dependent manner. The levels of intracellular ROS increased by 3.9 ± 0.6-fold, 2.7 ± 0.5-fold, 2.1 ± 0.1-fold, and 2.2 ± 0.1-fold in the 80 mmol/L baicalin-treated group in HOS, MG63, U2OS and 143B cells, respectively, compared to controls (Fig. 5A). ROS and calcium ions are closely linked to each other through multiple mechanisms, and both inhibit the occurrence and development of tumors by mutual regulation. We then employed Fluo-4 to determine the intracellular Ca 2+ concentration by flow cytometry. The results showed that intercellular Ca 2+ concentration increased 4.1 ± 1.2-fold and 2.8 ± 0.2-fold after 48 h of baicalin treatment in HOS and 143B cells, respectively (Fig. 5B). To further investigate the anti-osteosarcoma mechanism of baicalin, we investigated its effects on the PI3K/Akt/mTOR, ERK1/2 and bcatenin signaling pathways. These signaling pathways regulate proliferation, invasion, apoptosis and autophagy. Our results revealed that baicalin decreased expression levels of b-catenin (Fig. 5C). Then, the results of our experiments demonstrated that baicalin inhibited AKT, mTOR and ERK1/2 protein phosphorylation (Fig. 5C). In summary, these findings reveal that baicalin activates ROS and Ca 2+ and blocks the PI3K/Akt/mTOR, ERK1/2 and b-catenin pathways in osteosarcoma cells.

Baicalin induces apoptosis and autophagy in osteosarcoma cells by accumulating ROS to inhibit the PI3K/Akt/mTOR, ERK1/2 and b-catenin signaling pathways
To explore whether baicalin induces apoptosis and autophagy by accumulating ROS, we pretreated cells with 5 mM NAC, a ROS scavenger, to eliminate accumulating ROS induced by baicalin. The results showed that NAC pretreatment reduced baicalininduced intracellular ROS generation (Fig. 6A). Remarkably, these results demonstrated that NAC pretreatment restored the inhibitory cell viability of baicalin (Fig. 6B). In addition, we demonstrated that baicalin-induced apoptosis was significantly decreased after NAC pretreatment (Fig. 6C). NAC pretreatment also reversed the decrease in mitochondrial potential induced by baicalin (Fig. 6D). These results reveal that NAC pretreatment reverses baicalininduced expression of apoptosis-associated proteins (Fig. 6E), consistent with the above results. In terms of autophagy, the results showed that autophagy-related protein expression levels of LC3-Ⅱ and p62 were decreased after NAC pretreatment (Fig. 6F).  Moreover, these results indicated that pretreatment with NAC reversed the baicalin-induced p-Akt, p-mTOR, p-ERK and bcatenin expression levels (Fig. 6G). Overall, these findings demonstrate that ROS negatively regulate the PI3K/Akt/mTOR, ERK1/2 and b-catenin signaling pathways.

Baicalin-induced ROS and Ca 2+ interactions induce apoptosis in osteosarcoma
Intercellular Ca 2+ and ROS are stimulators of autophagy and apoptosis. However, whether Ca 2+ is a stimulator of baicalininduced apoptosis remains to be revealed. BAPTA-AM is designed to reduce Ca 2+ overload due to its ability to penetrate the cell membrane and chelate free Ca 2+ inside the cell. Our results revealed that pretreatment with BAPTA-AM (5 lM) decreased baicalin-induced ROS expression from 2.73 ± 0.4-fold to 2.09 ± 0.4-fold and from 2.1 ± 0.3-fold to 1.6 ± 0.1-fold in HOS and 143B cells, respectively (Fig. 7A). At the same time, pretreatment with NAC decreased intracellular Ca 2+ concentrations from 4.12 ± 1.3-fold to 1.6 ± 0.8 -fold and from 3.0 ± 0.4-fold to 1.5 ± 0.1-fold (Fig. 7B). Moreover, our observations revealed that BAPTA pretreatment significantly restored cell viability (Fig. 7C) and decreased the apoptotic ratio from 45.6 ± 4.7 to 28.2 ± 3.0% and from 25.9 ± 3.4% to 17.6 ± 2.0% in HOS and 143B cells, respectively, compared to baicalin alone (Fig. 7D). In summary, these findings reveal that baicalininduced ROS and Ca 2+ interactions cause apoptosis in OS cells.

The specific interactions between PI3Kc and baicalin
To explore the specific interactions between baicalin and PI3Kc, molecular docking was applied to simulate their binding. The results demonstrated that the binding energy of PI3Kc and baicalin was À9.7 kcal/mol. The binding energy of PI3Kc and a PI3Kc inhibitor (PDB ID: 4XZ4) was À7.5 kcal/mol. After importing software to simulate molecular docking, the three-dimensional structure of baicalin (Fig. 8A) was processed by hydrogenation, charge calculation, and charge assignment. Molecular docking of PI3Kc and baicalin revealed that baicalin was encapsulated by the internal  central cavity of PI3Kc and docked at the PI3Kc active site (Fig. 8B). The specific interactions between baicalin and amino acid residues of PI3Kc in the 3D schematic diagram (Fig. 8C) and 2D diagram of interactions (Fig. 8D) revealed that PI3Kc was linked to baicalin through four hydrogen bonds, indicating that baicalin binds human-derived PI3Kc with high affinity. Altogether, these results suggest that the specific interaction between PI3Kc and baicalin exerts an anti-osteosarcoma effect.

Discussion
Osteosarcoma is the most common form of bone malignancy and derives from primitive bone-forming mesenchymal cells. At present, surgical resection is the primary means of treating patients with osteosarcoma [44]. Remarkably, the active ingredients of natural compounds also have an essential function in osteosarcoma treatment [45]. The present work demonstrates that baicalin suppresses the proliferative capacity and invasive ability of human osteosarcoma cells. Cell cycle arrest can inhibit tumor cell proliferation. According to the results, baicalin increased the S-phase ratio in osteosarcoma cells. Furthermore, research revealed that baicalin inhibits c-Myc, cyclin A2, Cdk2 and cyclin E1 expression, which is crucial for proliferation and the cell cycle. These results suggest that baicalin induces S-phase accumulation and inhibits proliferation in osteosarcoma cells.
Cell cycle arrest can trigger cell death [16,46]. Apoptosis is triggered by multiple signaling pathways [47,48]. Most conventional chemotherapeutic agents depend on the activation of apoptotic pathways to exert their anticancer effects. Recent evidence indicates that endogenous and exogenous pathways are engaged in regulating apoptosis [49,50]. Moreover, mitochondria play a significant role in the endogenous pathway [51]. PARP is a marker of apoptosis that repairs damaged nuclear DNA and is inactivated by cleavage of caspase-3 [52]. Our results revealed that treatment with baicalin reduces mitochondrial membrane potential (MMP). By observing increased Bax, cleaved caspase-3 and cleaved PARP expression in osteosarcoma cells, we demonstrated that baicalin induced apoptosis. These results showed that baicalin treatment decreased the expression of Bcl-2, which also supported the induction of apoptosis. Since the activation of Bax protein can lead to a decrease in mitochondrial membrane potential [53], our findings indicated that baicalin triggers a cascade of apoptotic mitochondrial pathways. These outcomes demonstrated that baicalin disrupts the mitochondrial membrane integrity of osteosarcoma cells and activates apoptosis.
In addition to apoptosis, autophagy plays an essential role in regulating cell death. Throughout the autophagic process, cellular materials and organelles form autophagosomes, which are eventually digested in lysosomes [54,55]. When autophagy is triggered, LC3 transforms from its soluble LC3-I form to the lipolytic LC3-II form and binds to autophagic vesicles to form autophagosomes [56]. Our investigation revealed an increase in the LC3-II/LC3-I ratio in baicalin-treated cells, indicating that autophagy was triggered and autophagosome formation occurs. The dynamic process of autophagy includes the formation of autophagosomes, fusion of autophagosomes with lysosomes and lysosomal degradation. Dur- ing lysosomal degradation, substrate-bound p62 is degraded by proteolytic enzymes. At the same time, blockade of the downstream process of autophagy leads to accumulation of p62. Therefore, elevated p62 expression is generally considered a marker of downstream inhibition of autophagy [57,58]. These results indicated that baicalin treatment increased the expression of p62. Taken together, these results indicate an increase in the LC3-II/ LC3-I ratio and p62 expression in baicalin-treated cells, indicating that baicalin treatment induces autophagosome formation but blocks dynamic processes downstream of autophagy. Another reason for the accumulation of p62 may be correlated to the nature of the protein itself: p62 is an oxidative stress protein whose expression levels are markedly upregulated under stress conditions, and this process is primarily regulated by the transcription factor EB [59]. Accumulating evidence indicates that autophagy can both protect and damage cells in different cancer settings [60]. The above study indicates that the baicalin-induced autophagic process induces the formation of autophagosomes rather than the completion of autophagy. However, this incomplete process recruits caspase-8 to initiate the caspase-3 cascade reaction, causing cell death [61]. Furthermore, these results revealed that baicalininduced apoptosis was attenuated by 3-MA. We presumed that the primary reason for this is that baicalin-induced autophagosomes further triggered caspase-3-dependent apoptosis. Therefore, 3-MA suppressed the initial phase of autophagy, leading to a decrease in apoptosis and confirming the above speculation.
ROS are byproducts of mitochondrial metabolism. Cancer cells contain high basal ROS levels, which cause sensitizes cells to excessive oxidative stress. Therefore, excessive ROS are also deadly to osteosarcoma cells [62]. Research has indicated that multiple chemotherapeutic agents induce apoptosis and autophagy in tumor cells through ROS generation [30]. Interestingly, baicalin has antioxidant properties in human umbilical vein endothelial cells, fibroblasts, and glial cells [63][64][65]. However, it promotes ROS generation in a variety of tumor cells, inducing tumor cell apoptosis [66,67]. At the same time, such a phenomenon is not uncommon in anticancer drugs and may be attributed to different mechanisms or targets for increasing ROS in different types of cells.
Our previous data showed that baicalin dramatically increased ROS production in osteosarcoma cells. In this study, baicalin-induced apoptosis and autophagy were significantly reversed after ROS removal by NAC. Therefore, these results reveal that baicalin induces high ROS production to trigger apoptosis and autophagy. Increased intracellular Ca 2+ content aggravates endoplasmic reticulum stress and disturbs mitochondrial membrane potential, leading to mitochondria-dependent apoptosis [68]. Several investigations have suggested that cytoplasmic Ca 2+ significantly induces both autophagy and apoptosis [69]. Because ROS and Ca 2+ are tightly linked in many pathways, it is not surprising that increasing mitochondrial Ca 2+ uptake leads to enhanced mitochondrial ROS production in cancer cells [39]. In our work, we discov-ered that baicalin significantly increased intracellular Ca 2+ concentrations in osteosarcoma cells. Baicalin-induced apoptosis and ROS production were attenuated when cells were pretreated with BAPTA-AM. Notably, NAC pretreatment decreased the Ca 2+ concentration. However, our work did not focus on intercellular Ca 2+ in autophagy or the associated signaling pathways. An abundance of data indicates that the PI3K/Akt/mTOR, ERK1/2 and bcatenin signaling pathways are closely related to proliferation, apoptosis and autophagy [70][71][72]. In addition, extensive studies have reported that accumulating ROS inhibits the PI3K/Akt/mTOR, ERK1/2 and b-catenin pathways, leading to autophagy and apoptosis [36,37,73]. This research revealed a remarkable reduction in p-Akt, p-mTOR, p-ERK1/2, and b-catenin in response to baicalin treatment. Furthermore, NAC pretreatment restored these pathways inhibited by baicalin. In summary, these results suggest that baicalin induces apoptosis and autophagy in OS cells by inducing ROS to inhibit the PI3K/Akt/mTOR, ERK1/2 and b-catenin signaling pathways.
All of these results suggest that baicalin suppresses osteosarcoma cells. However, the relationship between baicalin's chemical components and targets remains unclear. PI3Kc is a class of PI3Ks [74]. Data have indicated that blocking PI3Kc improves tumor sensitivity to chemotherapeutic agents and enhances cancer immunotherapy to eradicate tumors [75]. Therefore, we speculated that baicalin may target PI3Kc. To investigate the specific interaction between PI3Kc and baicalin, we performed molecular docking simulations. The results showed that baicalin strongly interacts with the inner lumen of PI3Kc, suggesting tight binding of PI3Kc to baicalin. The underlying direct binding suggested that the combination of PI3Kc and baicalin may inhibit the downstream effectors AKT and mTOR. However, further experimental evidence is needed to verify our results. In summary, baicalin exerts an anti-osteosarcoma effect by targeting PI3Kc. This discovery provides a new candidate target drug for osteosarcoma treatment and proposes a new perspective on the mechanism underlying the anti-osteosarcoma effect of baicalin.

Conclusion
Our results demonstrated that apoptosis and autophagy are induced by baicalin in osteosarcoma cells. Next, we examined the relationship between baicalin-induced apoptosis and autophagy. Moreover, we further proposed that the anti-osteosarcoma effect of baicalin is mediated by accumulating ROS to inhibit the PI3K/ Akt/mTOR, ERK1/2 and b-catenin pathways. Baicalin also exerts an anti-osteosarcoma effect by targeting PI3Kc. Due to its excellent anticancer efficacy and relevant mechanism (Fig. 9), we recommend considering baicalin as a possible new anticancer agent against osteosarcoma.

Author contributions
BW, HP, and LS proposed the idea. HP, ZP, and QT explore the experiments. HP, TW, ZZ, XP, and ZP analyzed the results. BW, HP, LS, and TW finished the paper.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.