The Anticancer Role of Cerium Oxide Nanoparticles by Inducing Antioxidant Activity in Esophageal Cancer and Cancer Stem-Like ESCC Spheres

Introduction Esophagus squamous cell carcinoma (ESCC) has a poor prognosis, a high rate of metastasis, and rapid clinical progression. One hypothesis is that therapeutic failure is due to the presence of cancer stem cells (CSC). Previous studies showed the anticancer effect of cerium oxide nanoparticles (CNP) in different cancer cells. In this study, we aim to evaluate the effect of cerium oxide nanoparticles on cell antioxidants, toxicity, as well as cell oxidant level in esophageal cancer (YM1) and cancer stem cell-like (CSC-LC) cell lines. Method YM1 and CSC-LC spheres were treated with CNP at different concentrations. The cell viability was assessed by using the MTT test. Antioxidant levels (SOD (superoxide dismutase, CAT (catalase), thiol, and TAC (total antioxidant capacity)), antioxidant genes expression (SOD and CAT), ROS (reactive oxygen species), and MDA (malondialdehyde) levels were assessed in both cell lines. Results CSC-LC had significantly elevated SOX4 and OCT4 pluripotent genes. The ROS and MDA levels were significantly reduced in both YM1 and CSC-LC spheres after treatment with CNP. Also, the antioxidant levels and expressions were elevated significantly in both cell lines after CNP treatment. Conclusion These results suggest the potential anticancer effect of CNP by elevating antioxidant levels and expressions, and reducing oxidant levels.


Introduction
Esophageal cancer (EC) is the eighth most common cancer and the sixth most common cause of cancer-related death worldwide [1]. Esophagus squamous cell carcinoma (ESCC), the most frequent type of EC, has a poor prognosis, a high rate of metastasis, and rapid clinical progression [2]. The global incidence of ESCC was reported to be 87% of all EC cases in 2012 [3]. Despite the progression of early detection, surgery, and chemotherapy in patients with ESCC, its prognosis remains poor and challenging [4]. One hypothesis is that therapeutic failure is due to the presence of cancer stem cells which can cause recurrence, distant metastasis, and therapy resistance [5]. These cells have the ability to maintain and induce malignancy proliferation and metastasis in different types of cancers [6]. Finding a novel treatment to eliminate these cancer stem cells can help find new diagnostic and treatment approaches [7]. Nanotechnology has become a main focus of biomedical research area in recent years and its applications include drug delivery systems, tissue engineering, and luminescent biomarkers, among others [8].
The free radicals play a critical role in killing bacteria and viruses, as well as activation of enzymes and hormones, and producing energy [9]. They also have an important role in cell homeostasis and cell signal transductions [10]. The levels of these free radicals and reactive oxygen species (ROS) are controlled by antioxidant agents in human cells. An imbalance between ROS and the antioxidant agent is defined as oxidative stress, which has been linked to cardiovascular disease, neurodegenerative disease, diabetes mellitus, and different types of cancers [11,12]. It was shown that cancer cells have elevated levels of ROS in comparison to normal nontransformed cells [13]. In addition to the impact of ROS on the genome, it can play a role in promoting cancer cell proliferation, angiogenesis, survival, and metastasis [14]. However, ROS can have a contrary effect on cancer cells. Excessive levels of ROS can induce cancer cell death by increasing cell oxidative stress [15]. To prevent cancer cell death, they increase the level of antioxidant capacity to scavenge excessive ROS. Therefore, cancer cells have elevated levels of both ROS and antioxidants and this feature can make cancer cells more sensitive to ROS levels alteration [16,17].
Cerium, as a lanthanide rare earth metal, has two oxide forms with crystalline fluorite lattice structure. In particular, cerium oxide nanoparticles (CNP) consist of a cerium core surrounded by an oxygen lattice [18]. It has been shown that CNP has several antioxidant roles including catalase mimetic activity, superoxide dismutase (SOD) activity, hydroxyl radical scavenging, and nitric oxide radical scavenging. However, other studies revealed that CNP has a cytotoxicity role for cancer cells, an antiinvasive role, sensitizing role to radiation for cancer cells, in addition to protecting other surrounding normal cells [19].
In this study, we aim to evaluate the effect of cerium oxide nanoparticles on cell antioxidant (SOD, TAC (total antioxidant capacity), thiol, and CAT (catalase)), toxicity, as well as cell oxidant (ROS and MDA (malondialdehyde)) levels in esophageal cancer (YM1) and cancer stem cell-like (CSC-LC) cell lines.

ESCC Cell Sphere
Formation. The polymer of poly HEMA (2-Hydroxyethyl methacrylate) coated Petri dishes were used for transferring single cell suspensions derived from adherent cells with a concentration of 100000 cells/ mL for ESCC cells. The cells were maintained in the following solution at 37°C with 5% CO 2 to form spheres: serumfree RPMI/F12 medium supplemented with 20 ng/mL basic fibroblast growth factor (Grand Island, NY, USA), 2% B-27 supplement (Grand Island, NY, USA), and 20 ng/mL epidermal growth factor (Sigma-Aldrich Company, USA). Every two days, the medium was refreshed to replenish nutrients. Following the separation of the spheres into single cells, after 6 days, they were cultured in the new nonadherent Petri dishes with the same before condition. The stem-like properties of sphere cells were ready for the following experiments after three passages.

Cell Viability Assay.
Resazurin assay was used to investigate the cell toxicity of CNPs. Briefly, 25000 cells/well were seeded in a 96-well plate. After 24 hours of incubation at 37°C, CNPs were inoculated into the grown cells with different concentrations (0, 15.6, 31.2, 62.5, 125, and 250 μg/mL). Then, each well received the resazurin solution (phosphate buffer saline, 0.01 mg/mL) every 24 and 48 hours of incubation. The medium was discarded after finishing the incubation. Afterward, the resazurin solution (phosphate buffer saline, 0.01 mg/mL) was added to each well. After shaking the plates for 3 minutes, the optimal absorbance in the subsequent 3 hours was recorded at 600 nm excitation and 570 nm emission on a Perkin Elmer fluorimeter, and the IC 50 value was evaluated by using the GraphPad Prism® 6 (GraphPad Software, San Diego, CA, USA) software.
2.4. ROS Level Assay. Intracellular ROS production level was measured by using a 2 ′ ,7 ′ -Dichlorodihydrofluorescein diacetate (DCFDA) cellular ROS detection assay based on the manufacturer's protocol. Accordingly, 20 μM DCFDA was exposed to cells and then DCFDA was washed after 24 hours of incubation. Afterward, the CNPs (100 μM, 300 μM, 500 μM, 700 μM, and 1000 μM) were added to rewashed cells for 24 hours. Also, the positive control consists of tert-butyl hydroperoxide (TBHP). Finally, by using the fluorescence plate reader Perkin Elmer, the relative fluorescence intensity was recorded for these groups.
2.5. MDA Level Assay. The commercial kit r (Teb Pazhouhan Razi, Tehran, Iran) was used to assess the MDA level as a marker of oxidative stress level based on the manufacturer's protocol. Briefly, the cells were treated with the CNPs (100 μM, 300 μM, 500 μM, 700 μM, and 1000 μM) and after 24 hours of incubation, 1× Butylated hydroxytoluene 2 BioMed Research International (BHT) was added to lyse the cells. Subsequently, after mixing with 500 μL TCA, the prepared sample was incubated at 95°C for 5 min. The mixture was centrifuged (14000 g) for 5 min. After adding thiobarbituric acid (TBA) to the supernatants, the mixture was incubated at 95°C for 30 min. Finally, a spectrophotometer was used to measure the absorption of mixtures at 532 nm. The MDA level was measured by using a standard curve from GraphPad Prism software.
2.8. Statistical Analysis. The experimental data are shown as mean ± standard error of the mean. The data were analyzed by ANOVA test followed by Bonferroni's t-test and the GraphPad Prism® 6.0 software (San Diego, CA, USA) for Windows. All the results were analyzed triplicate in comparison to the untreated control group. A p value lower than 0.05 was used for a statistically significant level.

Results
3.1. Tumor Spheres Showed CSC-LC Features. Firstly, the corresponding adherent cells were compared with the YM1 derived sphere in passage three in order to confirm the CSCs enrichment. As indicated in Figure 1, prominent wellshaped spheres can be seen after three passages. The level of SOX2 and OCT4 have assessed in both passage 3 spheres and the corresponding adherent cells to characterize the stemness of spheres and the qRT-PCR demonstrated that both SOX2 and OCT4 were significantly overexpressed in sphere cells. The upregulation of pluripotency genes characterized the sphere cells and CSC-LC was confirmed by anchorage independent growth characteristics in consequent experiments.

CSC-LC and YM1
Cell Lines Viability Assay. 3.3. The ROS Levels in CSC-LC and YM1 Cell Lines. The ROS level was assessed in both CSC-LC and YM1 cell lines to elucidate the effect of CNP on cell oxidative products with H 2 DCFDA staining. Our results showed that CNP at concentrations equal to or above 100 μM was significantly associated with decreased ROS levels in YM1 cell lines, as shown in Figure 3(a). Whereas, in CSC-LC spheres, significantly decreased ROS levels were observed at concentrations higher than 500 μM of CNP (Figure 3(b)).

CAT Expression and Activity in CSC-LC and YM1 Cell
Lines. Figures 4 and 5 summarized the results of CAT expression and activity in CSC-LC and YM1 cell lines, respectively. The qRT-PCR showed that CAT expression was significantly higher at concentrations of 700 and 1000 μM of CNP in YM1 cell lines (Figure 4(a)). Also, in CSC-LC treated with CNP with 1000 concentration was followed by significantly higher CAT expression (Figure 4(b)). Also, identical results

SOD Expression and Activity in CSC-LC and YM1 Cell
Lines. As demonstrated in Figure 6(a), SOD activity was significantly higher in YM1 cell lines treated with CNP at concentrations of 500 and 700 μM. Moreover, the CSC-LC spheres which were treated with CNP at concentrations of 700 and 1000 μM, have significantly higher SOD activity (Figure 6(b)). Also, SOD expression was measured in CSC-LC and YM1 cell lines with qRT-PCR which showed SOD was significantly overexpressed in YM1 cell lines treated with 500, 700, and 1000 μM of CNP (Figure 7(a)). However, in CSC-LC spheres, SOD expression differences were insignificant between different concentrations of CNP (Figure 7(b)).
3.6. MDA Activity in CSC-LC and YM1 Cell Lines. As shown in Figure 8(a), the YM1 cells which were treated with CNP at concentrations of 300, 500, and 700 μM had significantly lower MDA activity. Also, MDA activity was significantly reduced in CSC-LCs treated with CNP at concentrations of 700 and 1000 μM of CNP (Figure 8(b)).
3.7. TAC and Thiol in CSC-LC and YM1 Cell Lines. As shown in Figure 9(a), the TAC capacity level was significantly elevated in YM1 cells treated with 500, 700, and 1000 μM of CNP. Moreover, CSC-LCs treated with CNP at concentrations of 700 and 1000 μM had significantly higher levels of TAC capacity (Figure 9(b)). Furthermore, measuring thiol levels revealed that YM1 cells that were treated with 700 and 1000 μM of CNP, have significantly elevated levels of thiol, as shown in Figure 10(a). Also, in CSC-LC spheres, thiol increased significantly at 1000 μM of CNP (Figure 10(b)).

Discussion
As previously noted EC is the sixth leading cause of cancer death worldwide and cancer stem cells are a subpopulation of cancer cells that are responsible for metastasis and treatment resistance [1,5]. Surgery, chemotherapy, and radiotherapy are usually used for cancer treatment but they have limited efficacy and side effects and may damage normal tissues [25]. Recent studies showed the significant antitumoral effect of CNP nanoparticles in several cancer cell lines [26][27][28][29]. Contrary to conventional cancer treatments, CNPs did not have a toxic effect on healthy cells [30]. To the best of our knowledge, this is the first study that is aimed In this study, the resazurin cytotoxicity assay showed CNP could cause ESCC cell death in both YM1 (ESCC) and CSC-LC cell lines in a dose-and time-dependent manner. The ROS and MDA levels significantly decreased in both YM1 cell line and CSC-LC spheres after incubation with CNP. Also, the level of SOD, CAT, thiol, and TAC sig-nificantly increased in both YM1 and CSC-LC spheres after treatment with CNP. Further examinations revealed the gene expressions of SOD and CAT were significantly elevated in cancer cells treated with CNP. However, in CSC-LC spheres, SOD expression did not change significantly after CNP incubation. An explanation for this result is that posttranscriptional modifications such as glycation, sulfation, and phosphorylation can alter the protein behavior.  ( * shows a significant difference between each concentration and the control group and # shows a significant difference between different concentrations) ( * or # p value < 0.05, * * or ## p value < 0.01, * * * or ### p value < 0.001, * * * * or #### p value < 0.0001) (CNP: cerium oxide nanoparticle, ROS: reactive oxygen species).

BioMed Research International
Therefore, as a result of posttranscriptional modifications, some gene expression changes cannot be detected by RNA analysis [23,31].
ROS can initiate and progress cancer cell growth, as well as downregulate the antioxidant enzymes [32]. The healthy cells intensely control the level of ROS by using antioxidants including SOD, CAT, thiols, glutathione, and peroxidase [10]. It has been indicated that cancer cells including ESCC have elevated levels of ROS in comparison to healthy cells which may be the result of mitochondria dysfunction, increased metabolic activities, elevated peroxisome activity, increased cell signaling, and oncogenes activity [33]. Furthermore, ROS can induce genetic instability, proliferation, angiogenesis, and metastasis in cancer cells [34]. With regard to the double-edged sword character of ROS in the treatment of cancer cells, which will be explained later, lowering and elevating strategies have been suggested for cancer treatment. ( * shows a significant difference between each concentration and the control group and # shows a significant difference between different concentrations) ( * or # p value < 0.05, * * or ## p value < 0.01, * * * or ### p value < 0.001, * * * * or #### p value < 0.0001) (CNP: cerium oxide nanoparticle, CAT: catalase). ( * shows a significant difference between each concentration and the control group and # shows a significant difference between different concentrations) ( * or # p value < 0.05, * * or ## p value < 0.01, * * * or ### p value < 0.001, * * * * or #### p value < 0.0001) (CNP: cerium oxide nanoparticle, CAT: catalase).

BioMed Research International
Previous studies showed the antiapoptotic effect of high levels of ROS which is the result of redox-sensitive transcription activation including nuclear factor κ-light-chainenhancer of activated B cells (NF-κB) [35]. NF-κB as an important transcription factor can inhibit apoptosis by regulating antiapoptotic genes such as Bcl-2 and survivin [36,37]. The location of NF-κB is within the cytosol in healthy cells as inactive forms that is bond to IκBα. However, cancer cells have active forms of NF-κB due to IκBα phosphorylation. The active form of NF-κB can induce prosurvival gene expression including inhibitors of apoptosis and result in uncontrol cell growth [38]. Previous studies have shown the inhibition effect of CNP on NF-κB in different cell lines [39,40]. Thus, inhibiting the NF-κB by the downregulation ( * shows a significant difference between each concentration and the control group and # shows a significant difference between different concentrations) ( * or # p value < 0.05, * * or ## p value < 0.01, * * * or ### p value < 0.001, * * * * or #### p value < 0.0001) (CNP: cerium oxide nanoparticle, SOD: superoxide dismutase). effect of CNP on ROS levels can be a promising approach for cancer treatment.
Recent studies showed both antioxidant and oxidant roles of CNP in different cells. In agreement with our findings, Patel et al. showed the inhibitory effect of CNP on ROS levels and suggested that CNP has a potential therapeutic effect on human monocytic leukemia cells [26]. However, some previous studies demonstrated the antitumoral effect of CNP with increasing or even unchanged ROS levels. Contrary to our results Lin et al. demonstrated the dose-dependent and timedependent effect of CNP on human lung cancer cell lines by increasing the ROS level [27]. Also, Park et al., showed the cytotoxic effect of CNP on cultured human epithelial cells by increasing ROS levels and decreasing antioxidant levels which induce cell apoptosis [28]. However, in Xiao et al.'s study, although the CNP induces a cytotoxic effect on gastric cancer cells, the ROS level was unchanged after CNP treatment [29]. These discrepancies can be due to different doses of CNP and TAC capacity in CSC-LC. ( * shows a significant difference between each concentration and the control group and # shows a significant difference between different concentrations) ( * or # p value < 0.05, * * or ## p value < 0.01, * * * or ### p value < 0.001, * * * * or #### p value < 0.0001) (CNP: cerium oxide nanoparticle, TAC: total antioxidant capacity). 8 BioMed Research International also different cell lines. Also, it has been indicated that cell pH is an important key factor for the oxidant or antioxidant role of CNPs [41]. Interestingly, as previously noted, ROS has a doubleedged sword function. Both elevating and lowering oxidant level has been suggested as a treatment strategy for cancer cells [42,43]. In cancer cells, increasing levels of ROS as a result of signaling cascades and metabolic reactions may induce cellular antioxidant upregulation to maintain redox homeostasis. Therefore, exogenous ROS-producing agents can induce cancer cell deaths by triggering the ROS level [33]. On the other hand, as elevated levels of ROS play an important role in carcinogenesis, upregulation of cell antioxidants can deplete the ROS level and consequently cause growth inhabitation and cancer cell death [33].
It is worth mentioning that the least concentration of CNP that was needed to change the oxidant and antioxidant level was totally higher in the CSC-LC spheres in comparison to the YM1 cell line. Also, the IC 50 for CNP was higher in the CSC-LC spheres in comparison to the YM1 cell line (968 and 758 μM, respectively). These results can be explained by the potential treatment resistance features of CSC cells.
In summary, our study suggests the potential role of CNP as an effective anticancer treatment for EC and cancer stem cells. However, our study had some limitations. First, the signaling pathways of NF-κB were not investigated. Second, the study was an in vitro examination, and further in vivo studies are missing. Further in vivo and clinical studies are recommended to highlight the effect of CNP on EC and cancer stem cells.

Conclusion
In conclusion, the present study showed that cerium oxide nanoparticles have potential anticancer effects on esophageal and cancer stem cells by increasing the cell antioxidant levels (including SOD, CAT, thiol, and TAC) and decreasing the oxidant levels (including ROS and MDA) in YM1 and CSC-LC spheres.

Data Availability
The authors of this article will share all the data underlying the findings of their manuscripts with other researchers. This sharing of data allows researchers to replicate the results of an article and conduct secondary analyses. Therefore, I hereby declare the statement of "availability" for the data used in this manuscript. All the results were analyzed triplicate in comparison to the untreated control group. A p value lower than 0.05 was used for a statistically significant level. And researchers can communicate with the first author and the corresponding authors for the data by email.

Conflicts of Interest
The authors declare that there were no conflict of interest.

Authors' Contributions
All authors contributed to data gathering and manuscript writing. HJ, MFH, and AE contributed to the writing of the article. HJ, SGF, and SIH were responsible for coordinating the authors. SGF and SIH finalized and submitted the paper. Seyed Isaac Hashemy and Sattar Gorgani-Firuzjaee as co-corresponding authors. ( * shows a significant difference between each concentration and the control group and # shows a significant difference between different concentrations) ( * or # p value < 0.05, * * or ## p value < 0.01, * * * or ### p value < 0.001, * * * * or #### p value < 0.0001) (CNP: cerium oxide nanoparticle). 9 BioMed Research International