MiR-657/ATF2 Signaling Pathway Has a Critical Role in Spatholobus suberectus Dunn Extract-Induced Apoptosis in U266 and U937 Cells

Though Spatholobus suberectus Dunn (SSD) has been reported to have anti-virus, anti-osteoclastogenesis, and anti-inflammation activities, its underlying anti-cancer mechanism has never been elucidated in association with the role of miR-657 in endoplasmic reticulum (ER) stress-related apoptosis to date. SSD treatment exerted cytotoxicity in U266 and U937 cells in a dose-dependent manner. Also, apoptosis-related proteins such as PARP, procaspase-3, and Bax were regulated by SSD treatment. Furthermore, Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay revealed that a number of apoptotic bodies were increased by SSD. Interestingly, the ER stress-related proteins such as p-ATF2 and CHOP were elevated by SSD. Interestingly, reactive oxygen species (ROS) generation and cytotoxicity by SSD treatment were significantly reduced by N-Acetyl-L-cysteine (NAC). Among the microRNAs (miRNAs) regulated by SSD treatment, miR-657 was most significantly reduced by SSD treatment. However, an miR-657 mimic reversed SSD-induced apoptosis by the attenuation of the expression of p-ATF2, CHOP, and PARP cleavage. Overall, these findings provide scientific evidence that miR657 is an onco-miRNA targeting the ER stress signal pathway and SSD induces apoptosis via the inhibition of miR-657, ROS, and the activation of p-ATF2 and CHOP as a potent anti-cancer agent for myeloid-originated hematological cancer.


Introduction
Myeloid-originated hematological malignancies, including multiple myeloma (MM) and myeloid leukemia (ML), have been constantly increasing worldwide [1,2]. MM is a tumor of long-lived malignant plasma cells in the bone marrow, leading to infection, osteolytic bone lesions, hypercalcemia, or renal insufficiency [3]. ML is the abnormal differentiation of myeloid cells in the bone marrow, leading to fatigue, hemorrhage, infection, and organ infiltration, such as hepatomegaly and splenomegaly [4,5]. Myeloid-originated hematological malignancies are capable of escaping . In order to obtain efficient chromatographic methods for the analysis of (c) SSD compared with (a) epigallocatechin (EGC) or (b) genistein in accordance with the retention time and UV-Vis wavelength, an Agilent series 1290 system was used. To exhibit efficient separation and reasonable results, intra-day and the inter-day analyses were conducted in the same day and on consecutive three days, respectively. The gradient eluent condition resulted in a satisfactory separation. The relative standard derivations (RSDs) were deliberated to indicate the level of precision. Table 1. Measurement repeatability of intra-day precision and inter-day precision was performed against reference standards; the percentage of RSD of six assay results was calculated.

Inter-day
Standards SSD 1 RT 2 (min) RSD 3 (%) RT 2 (min) SD 4 Figure 1. HPLC analysis of Spatholobus suberectus Dunn (SSD). In order to obtain efficient chromatographic methods for the analysis of (c) SSD compared with (a) epigallocatechin (EGC) or (b) genistein in accordance with the retention time and UV-Vis wavelength, an Agilent series 1290 system was used. To exhibit efficient separation and reasonable results, intra-day and the inter-day analyses were conducted in the same day and on consecutive three days, respectively. The gradient eluent condition resulted in a satisfactory separation. The relative standard derivations (RSDs) were deliberated to indicate the level of precision.

SSD Exerted Cytotoxicity in Hematological Cancers but Less in Normal Cells
To evaluate the cytotoxic effect of SSD against hematological cancers, EZ-CYTOX cell viability assay was performed. Cells were treated with various concentrations (0, 12.5, 25, 50, 100, 200, 400 µg/mL) of SSD for 24 h. SSD exerted significant cytotoxicity in cancer cells, including U266, U937, THP-1, and K562 cells, as well as in normal cells, including CPAE, CCD-18Co, and MDBK cells. SSD treatment exerted cytotoxicity in multiple myeloma U266 cells and myeloid leukemia U937, THP-1, K562 cells. However, CPAE cells (normal pulmonary artery endothelial cells), CCD-18Co cells (normal colon epithelial cells), and MDBK cells (normal kidney cells) were not affected by up to 50 µg/mL of SSD treatment (Figure 2a). The doses of SSD used in this study were 10 and 20 µg/mL, which are toxic to hematological cancer cells but harmless to normal cells. To compare the cytotoxic effect of SSD and its components, genistein and EGC were also administered to U266, U937, and MDBK cells. As shown in Figure 2b, genistein or EGC tends to exert higher cytotoxicity in U266 from 25 µg/mL and similar cytotoxicity in U937 cells. However, up to 25 µg/mL the cytotoxicity of SSD was similar or even higher than that of genistein and EGC. Also, in the normal MDBK cells, genistein and EGC showed higher cytotoxicity compared to SSD nearly from 50 µg/mL (Figure 2c).

SSD Exerted Cytotoxicity in Hematological Cancers but Less in Normal Cells
To evaluate the cytotoxic effect of SSD against hematological cancers, EZ-CYTOX cell viability assay was performed. Cells were treated with various concentrations (0, 12.5, 25, 50, 100, 200, 400 µg/mL) of SSD for 24 h. SSD exerted significant cytotoxicity in cancer cells, including U266, U937, THP-1, and K562 cells, as well as in normal cells, including CPAE, CCD-18Co, and MDBK cells. SSD treatment exerted cytotoxicity in multiple myeloma U266 cells and myeloid leukemia U937, THP-1, K562 cells. However, CPAE cells (normal pulmonary artery endothelial cells), CCD-18Co cells (normal colon epithelial cells), and MDBK cells (normal kidney cells) were not affected by up to 50 µg/mL of SSD treatment (Figure 2a). The doses of SSD used in this study were 10 and 20 µg/mL, which are toxic to hematological cancer cells but harmless to normal cells. To compare the cytotoxic effect of SSD and its components, genistein and EGC were also administered to U266, U937, and MDBK cells. As shown in Figure 2b, genistein or EGC tends to exert higher cytotoxicity in U266 from 25 µg/mL and similar cytotoxicity in U937 cells. However, up to 25 µg/mL the cytotoxicity of SSD was similar or even higher than that of genistein and EGC. Also, in the normal MDBK cells, genistein and EGC showed higher cytotoxicity compared to SSD nearly from 50 µg/mL (Figure 2c).

SSD Induced Apoptosis in U266 and U937 Cells
To identify whether the cytotoxic effect of SSD was due to apoptosis induction, U266 and U937 cells were treated with 10 and 20 µg/mL of SSD for Western blot analysis and transferase dUTP nick end labeling (TUNEL) assay. As shown in Figure 3A, SSD cleaved PARP and capase-3. Bax, a proapoptotic Bcl-2 family, was increased by SSD treatment. Also, TUNEL-positive cells were significantly increased in SSD-treated U266 and U937 cells ( Figure 3B). Along with the Western blot analysis results of apoptosis-related proteins, TUNEL assay data revealed the apoptotic effect of SSD in U266 and U937 cells.

SSD Induced Apoptosis in U266 and U937 Cells
To identify whether the cytotoxic effect of SSD was due to apoptosis induction, U266 and U937 cells were treated with 10 and 20 µg/mL of SSD for Western blot analysis and transferase dUTP nick end labeling (TUNEL) assay. As shown in Figure 3A, SSD cleaved PARP and capase-3. Bax, a pro-apoptotic Bcl-2 family, was increased by SSD treatment. Also, TUNEL-positive cells were significantly increased in SSD-treated U266 and U937 cells ( Figure 3B). Along with the Western blot analysis results of apoptosis-related proteins, TUNEL assay data revealed the apoptotic effect of SSD in U266 and U937 cells.

SSD Induced Apoptosis in U266 and U937 Cells
To identify whether the cytotoxic effect of SSD was due to apoptosis induction, U266 and U937 cells were treated with 10 and 20 µg/mL of SSD for Western blot analysis and transferase dUTP nick end labeling (TUNEL) assay. As shown in Figure 3A, SSD cleaved PARP and capase-3. Bax, a proapoptotic Bcl-2 family, was increased by SSD treatment. Also, TUNEL-positive cells were significantly increased in SSD-treated U266 and U937 cells ( Figure 3B). Along with the Western blot analysis results of apoptosis-related proteins, TUNEL assay data revealed the apoptotic effect of SSD in U266 and U937 cells.

SSD Upregulated ER Stress-Related Proteins in U266 and U937 Cells
Next, to reveal whether the apoptosis induction by SSD is associated with ER stress, Western blotting for ER stress-related proteins such as CHOP, p-eIF2α, ATF4, p-ATF2, and PERK was conducted. The result showed that SSD treatment upregulated the expression of ER stress proteins such as CHOP, p-eIF2α, ATF4, p-ATF2, and PERK in U266 and U937 cells (Figure 4).
-positive cell population in U266 and U937 cells. (A) Cells were exposed to SSD (10 or 20 µg/mL) for 24 h. The lysates were subjected to Western blot analysis for pro-PARP, pro-caspase-3, Bax, and βactin. (B) Cells were treated with 20 µg/mL of SSD and used for TdT-mediated dUTP nick end labeling (TUNEL) staining (sample number n = 3). Scale bar: 50 µm. Cells were visualized using an Olympus FLUOVIEW FV10i confocal microscope. Results show a representative of three independent experiments.

SSD Upregulated ER Stress-Related Proteins in U266 and U937 Cells
Next, to reveal whether the apoptosis induction by SSD is associated with ER stress, Western blotting for ER stress-related proteins such as CHOP, p-eIF2α, ATF4, p-ATF2, and PERK was conducted. The result showed that SSD treatment upregulated the expression of ER stress proteins such as CHOP, p-eIF2α, ATF4, p-ATF2, and PERK in U266 and U937 cells ( Figure 4).

SSD Increased ROS Generation in U266 and U937 Cells
ROS has pivotal role in ER stress-related apoptosis induction [32]. To confirm the effect of SSD on ROS generation, ROS was quantified by 2′,7′-Dichlorofluorescein diacetate (DCFDA) cellular ROS detection assay in SSD-treated hematologic cancer cells. As shown in Figure 5, SSD treatment significantly increased ROS generation in both U266 and U937 cells in a dose-dependent manner.

SSD Increased ROS Generation in U266 and U937 Cells
ROS has pivotal role in ER stress-related apoptosis induction [32]. To confirm the effect of SSD on ROS generation, ROS was quantified by 2 ,7 -Dichlorofluorescein diacetate (DCFDA) cellular ROS detection assay in SSD-treated hematologic cancer cells. As shown in Figure 5, SSD treatment significantly increased ROS generation in both U266 and U937 cells in a dose-dependent manner.
-positive cell population in U266 and U937 cells. (A) Cells were exposed to SSD (10 or 20 µg/mL) for 24 h. The lysates were subjected to Western blot analysis for pro-PARP, pro-caspase-3, Bax, and βactin. (B) Cells were treated with 20 µg/mL of SSD and used for TdT-mediated dUTP nick end labeling (TUNEL) staining (sample number n = 3). Scale bar: 50 µm. Cells were visualized using an Olympus FLUOVIEW FV10i confocal microscope. Results show a representative of three independent experiments.

SSD Upregulated ER Stress-Related Proteins in U266 and U937 Cells
Next, to reveal whether the apoptosis induction by SSD is associated with ER stress, Western blotting for ER stress-related proteins such as CHOP, p-eIF2α, ATF4, p-ATF2, and PERK was conducted. The result showed that SSD treatment upregulated the expression of ER stress proteins such as CHOP, p-eIF2α, ATF4, p-ATF2, and PERK in U266 and U937 cells (Figure 4).

SSD Increased ROS Generation in U266 and U937 Cells
ROS has pivotal role in ER stress-related apoptosis induction [32]. To confirm the effect of SSD on ROS generation, ROS was quantified by 2′,7′-Dichlorofluorescein diacetate (DCFDA) cellular ROS detection assay in SSD-treated hematologic cancer cells. As shown in Figure 5, SSD treatment significantly increased ROS generation in both U266 and U937 cells in a dose-dependent manner.

ROS Scavenger Reversed SSD-Induced Apoptosis in U266 and U937 Cells
To confirm that ROS generation is associated with SSD-induced apoptosis, an ROS measurement assay and a cytotoxicity assay were conducted with N-Acetyl-L-cysteine (NAC)-treated cells. As shown in Figure 6a, increased ROS generation by SSD treatment was significantly reduced by NAC pretreatment. Also, the cytotoxic effect of SSD was significantly downregulated by NAC pretreatment in both U266 and U937 cells (Figure 6b). (a) and U937 (b) cells were exposed to SSD (10 or 20 µg/mL) for 24 h (sample number n = 3). ROS production was analyzed using 2′,7′-Dichlorofluorescein diacetate (DCFDA) by cellular reactive oxygen species detection assay kit. Results are presented as the means ± SD of three independent experiments. ** p < 0.01; *** p < 0.001 versus the control group.

ROS Scavenger Reversed SSD-Induced Apoptosis in U266 and U937 Cells
To confirm that ROS generation is associated with SSD-induced apoptosis, an ROS measurement assay and a cytotoxicity assay were conducted with N-Acetyl-L-cysteine (NAC)-treated cells. As shown in Figure 6a, increased ROS generation by SSD treatment was significantly reduced by NAC pretreatment. Also, the cytotoxic effect of SSD was significantly downregulated by NAC pretreatment in both U266 and U937 cells (Figure 6b).
(a) (b) Figure 6. Effect of ROS scavenger on SSD-induced apoptosis in U266 and U937 cells. (a) Cells were treated with N-Acetyl-L-cysteine (NAC) for 2 h prior to the exposure to SSD (10 or 20 µg/mL) for 24 h. ROS production was analyzed using oxidation sensitive fluorescent dye (DCFDA) by a cellular reactive oxygen species detection assay kit. (b) Cells were seeded into 96-well microplates and preexposed to NAC (5 mM) for 2 h, then treated with the indicated concentrations of SSD for 24 h. EZ-CYTOX assay was used to measure cell viability in U266 and U937 cells (sample number n = 3). Results are presented as the means ± SD of three independent experiments. * p < 0.05; ** p < 0.01; *** p < 0.001 between the two groups.

SSD Reduced the Expression of miR-657 in U266 and U937 Cells
To elucidate the role of miRNAs in SSD-induced apoptosis, microRNA array and qRT-PCR were conducted. The miRNAs array included a set of miRNAs associated with ER stress and apoptosis. miRNA expression data were obtained and their altered expression was visualized by a heat map (Figure 7a). The results revealed significantly altered expression levels of various miRNAs (p < 0.05). Among the miRNAs, the expression of miR-657 was significantly reduced by SSD treatment in both U266 and U937 cells, which is reported to have role in tumorigenesis [27]. To confirm the negative Figure 6. Effect of ROS scavenger on SSD-induced apoptosis in U266 and U937 cells. (a) Cells were treated with N-Acetyl-L-cysteine (NAC) for 2 h prior to the exposure to SSD (10 or 20 µg/mL) for 24 h. ROS production was analyzed using oxidation sensitive fluorescent dye (DCFDA) by a cellular reactive oxygen species detection assay kit. (b) Cells were seeded into 96-well microplates and pre-exposed to NAC (5 mM) for 2 h, then treated with the indicated concentrations of SSD for 24 h. EZ-CYTOX assay was used to measure cell viability in U266 and U937 cells (sample number n = 3). Results are presented as the means ± SD of three independent experiments. * p < 0.05; ** p < 0.01; *** p < 0.001 between the two groups.

SSD Reduced the Expression of miR-657 in U266 and U937 Cells
To elucidate the role of miRNAs in SSD-induced apoptosis, microRNA array and qRT-PCR were conducted. The miRNAs array included a set of miRNAs associated with ER stress and apoptosis. miRNA expression data were obtained and their altered expression was visualized by a heat map (Figure 7a). The results revealed significantly altered expression levels of various miRNAs (p < 0.05). Among the miRNAs, the expression of miR-657 was significantly reduced by SSD treatment in both U266 and U937 cells, which is reported to have role in tumorigenesis [27]. To confirm the negative regulation of SSD on miR-657, qRT-PCR for miR-657 was performed. As shown in Figure 7b, the expression of miR-657 in SSD-treated U266 and U937 cells was decreased compared to untreated control groups. regulation of SSD on miR-657, qRT-PCR for miR-657 was performed. As shown in Figure 7b, the expression of miR-657 in SSD-treated U266 and U937 cells was decreased compared to untreated control groups.

MiR-657 Plays a Role in SSD-Induced Apoptosis in U266 and U937 Cells
To reveal the further role of miR-657 in SSD-induced apoptosis, U266 and U937 cells were transfected with an miR-657 mimic for 48 h and exposed to 20 µg/mL of SSD. Subsequently, qRT-PCR, viability assay, and Western blot analysis were implemented. As shown in Figure 8a, reduced expression of miR-657 by SSD treatment was reversed in the miR-657 mimic-transfected cells. Also, significantly downregulated cell viability was increased in the miR-657 mimic-transfected cells. MiR-657 mimic-transfected cells showed a significant increase in cell viability compared to the control group (Figure 8b). Western blot analysis data also revealed that the increased expression of p-ATF2, CHOP, and cleaved PARP induced by SSD treatment was attenuated in miR-657 mimic-transfected cells (Figure 8c). To reveal the further role of miR-657 in SSD-induced apoptosis, U266 and U937 cells were transfected with an miR-657 mimic for 48 h and exposed to 20 µg/mL of SSD. Subsequently, qRT-PCR, viability assay, and Western blot analysis were implemented. As shown in Figure 8a, reduced expression of miR-657 by SSD treatment was reversed in the miR-657 mimic-transfected cells. Also, significantly downregulated cell viability was increased in the miR-657 mimic-transfected cells. MiR-657 mimic-transfected cells showed a significant increase in cell viability compared to the control group (Figure 8b). Western blot analysis data also revealed that the increased expression of p-ATF2, CHOP, and cleaved PARP induced by SSD treatment was attenuated in miR-657 mimic-transfected cells (Figure 8c).

Discussion
MM is the most advanced plasma cell disease. MM is characterized by the excessive monoclonal proliferation of plasma cells, which secrete monoclonal myeloma proteins (M-proteins) composed of two heavy polypeptide chains of the same class and two light polypeptide chains of the same type [33]. ML is a cancer of the myeloid line of blood cells, characterized by a marrow stem cell disorder in which the accumulation of immature granulocytes such as neutrophils, eosinophils, and basophils is found [34].
Due to the development of proteasome inhibitors (PIs) including Bortezomib and Carfilzomib as well as immunomodulatory drugs (IMiDs) including Elotuzumab and Lorvotuzumab, the median survival rate of patients with MM or ML has been improved [35,36]. However, most patients with MM experience relapse and treatment with anti-hematologic cancer agents leads to side effects such as peripheral neuropathy [36,37]. Patients with ML suffer from anemia as well as easy bruising or bleeding. Moreover, older patients who are unable to receive intensive chemotherapy have a typical survival of 5-10 months [38].
Currently, MM and ML are considered incurable diseases. As such, new approaches using novel materials and mechanisms are needed. Thus, in this study, a new anti-cancer mechanism of SSD against hematological cancer was first elucidated, regarding ROS/ER stress and miRNA regulation. SSD significantly exerted cytotoxicity in a dose-dependent manner in U266 and U937 hematologic cancer cells. The cytotoxic effect of SSD leading to cell death was examined by cytotoxic assay, Western blot analysis, TUNEL assay, and qRT-PCR-all of which demonstrated that various factors that are correlated to ER stress influence the apoptosis mechanism. Of note, PARP, caspase-3 cleavage, and the difference in the level of Bax protein were observed. These results indicate that SSD displays apoptotic effects. Apoptosis is defense mechanism that is employed, for example, in immune reactions when cells are damaged by disease or toxic agents [39]. PARP is a family of proteins involved in the repair of single-strand DNA breaks, subsequently shown to be cleaved into 89-and 24-kDa fragments during drug-induced apoptosis in a variety of cells [40]. Such cleavage disables its ability to respond to DNA strand breaks and inactivates the enzyme [41]. SSD treatment cleaved PARP in both U937 and U266 cells. Caspase-3, a member of the caspase family that plays a critical role in the process of the apoptotic program, is primarily responsible for the cleavage of PARP during cell death [42]. Procaspase-3 is reduced by SSD treatment. Our results also showed that SSD treatment markedly elevated the protein expression of the pro-apoptotic protein Bax. Accordingly, TUNEL-positive cells were increased in SSD-treated U266 and U937 cells, indicating that SSD induces apoptosis.
The ER, an organelle related to Ca 2+ storage and protein folding/maturation, plays a pivotal role in protein synthesis and transport. Furthermore, ER stress is associated with various diseases including inflammation, diabetes, and cancer [43]. ER stress-related proteins such as CHOP, ATF4, p-ATF2, PERK, and p-eIF2α are regulated by SSD treatment, demonstrating that SSD induces ER stress.
ROS formation is known to trigger misfolded or unfolded proteins, inducing ER stress and apoptosis [44,45]. To identify that ER stress induced by SSD was due to ROS generation, ROS was measured by a cellular reactive oxygen species detection assay kit. SSD treatment significantly increased ROS generation in U266 and U937 cells. NAC, an aminothiol and synthetic precursor of intracellular cysteine and GSH, has an important role as a ROS scavenger [46]. Therefore, NAC has been used in apoptosis research to investigate the role of ROS in the induction of apoptosis. Notably, the increased ROS generation and cytotoxicity induced by SSD was reversed by NAC pretreatment, indicating that SSD-induced apoptosis was due to ROS generation. U266 cells were more susceptible to SSD treatment than U937 cells. The difference might be based on a different ROS generation level and a different decreased level of miR-657 after SSD treatment.
MiR-657 is known by its oncogenic properties in several cancer types. It plays a critical role in lung cancer [47] and liver cancer [27,48]. Also, the overexpression of miR-657 has been found in larynx carcinoma [49] and metastasis of cervical squamous cancer cells [50]. However, the role of miR-657 in hematologic cancer has not yet been reported. We reported miRNA-related apoptosis brought on by some herbal medicines used to treat MM [51,52]. In this study, it was identified that treatment with SSD decreased the expression of miR-657 and the deregulation of miR-657 led to the inhibition of ER stress and apoptosis by the regulation of p-ATF2, CHOP, and PARP. However, when miR-657 mimic transfection was applied, cell viability was significantly increased and the apoptotic effect of SSD was attenuated via ER stress reduction. These results suggest that miR-657 may act as an onco-miRNA via ER stress regulation and that SSD could induce apoptosis by the inhibition of this onco-miRNA and ER stress induction. Taken together, this study reports for the first time that SSD could be a potent anti-hematologic cancer agent inducing apoptosis via the modification of miR-657/ER stress pathways.

Chemicals and Reagents
Spatholobus suberectus Dunn (200 g) was harvested in Hongchungun, Gangwondo, Korea. A voucher specimen (no. KH-00086) was stored at the herbarium of the Cancer Molecular Targeted Herbal Research Center of Kyung Hee University. The preparation and extraction were carried out as previously described [53,54]. Briefly, SSD was extracted twice in 100% ethanol (EtOH, 1 L × 2) for 3 days each. The extracted solutions were filtered and evaporated to produce an EtOH extract (9.4 g, percent yield = 4.7%).

Cytotoxicity Assay
The cytotoxic effect of SSD against U266, U937, THP-1, K562, CPAE, CCD-18Co, and MDBK cells was assayed using an EZ-CYTOX cell viability assay kit (Daeil Lab Service, Seoul, Korea) according to the manufacturer's instruction. Briefly, cells were seeded onto a 96-well microplate and treated with various concentrations of SSD, genistein, and EGC (0, 12.5, 25, 50, 100, 200, 400 µg/mL) for 24 h. To measure the optical density, a microplate reader (Bio-Rad, Hercules, CA, USA) was used at 450 nm. Cell viability was determined as a percentage of viable cells in the drug-treated group versus the untreated control.

TUNEL Assay
To detect cell death, a DeadEndTM Fluorometric terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) system kit was used according to the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA). In brief, U266 or U937 cells (2 × 10 5 cells/mL) were treated with SSD for 24 h and plated onto a poly-L-lysine-coated slide. Cells were fixed with 4% paraformaldehyde for 30 min and washed twice with phosphate-buffered saline (PBS) for 2 min. Fixed cells were exposed to permeabilization solution (0.1% Triton X-100 and 0.1% Sodium citrate) and incubated with terminal deoxyribonucleotidyl transferase (TdT) enzyme buffer containing fluorescein-12-dUTP for 60 min at room temperature (RT) in the dark. The slides were mounted with mounting medium containing 4 ,6-diamidino-2-phenylindole (DAPI) (VECTOR, Burlingame, CA, USA). The TUNEL-stained cells were visualized by FLUOVIEW FV10i confocal microscopy (Olympus, Tokyo, Japan).

Quantitative Real-Time PCR Analyses
RNA was isolated from cells using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and cleaned using a RNeasy Mini kit (Qiagen, Seoul, Republic of Korea). The total RNA was reverse transcribed using mRQ enzyme (Takara, Tokyo, Japan), performed on 1 µg of total RNA using Mir-X TM miRNA First-Strand Synthesis and a SYBR qRT-PCR Kit (Takara) as per the manufacturer's protocol. In brief, for the mature-miRNA reverse transcription of miRNA, 3 -Primer and U6 primers supplied by Mir-X TM miRNA First-Strand Synthesis and SYBR ® qRT-PCR Kit (Takara) and an miRNA-specific 5 sequence were applied by Bioneer (Bioneer Corporation, AccuTarget TM Human miRNA Mimic and Inhibitor Library, Daejeon, Republic of Korea) with the following sequences: has-miR-657 5-forward-5 -GGCAGGUUCUCACCCUCUCUAGGATGAC-3 . PCR was initially performed at 95 • C for 10 s, 95 • C for 5 s, and 60 • C for 20 s, followed by 40 cycles at 95 • C for 60 s, 55 • C for 30 s, and 95 • C for 30 s. Relative miRNA changes in contents were normalized using the standard C t level of U6 snRNA. Three individual miR-specific values were calculated and averaged to obtained means ± standard deviations (SDs). RT-qPCR was performed using a LightCycler TM instrument (Roche Applied Sciences, Indianapolis, IN, USA).

Microarray
The total RNA quality and quantity were assessed by Agilent bioanalyzer 2100 analysis. Human microRNA expression was analyzed with a miRCURY LNA TM microRNA array, 7th gen-has, mmu and rno array (EXIQON, Vedbaek, Denmark), covering 1918 well-characterized human microRNA among 3100 capture probes for human, mouse, and rat miRNAs. In this procedure, 5 -phosphates from 1 µg of total RNA was removed by treating Calf Intestinal Alkaline Phosphatase (CIP) followed by labeling with Hy3 green fluorescent dye. Labeled samples were subsequently hybridized by loading onto a microarray slide using a Hybridization Chamber Kit part # G2534A (Agilent Technologies, Santa Clara, CA, USA) and a Hybridization Gasket Slide Kit part # G2534-60003 (Agilent Technologies). Hybridization was performed over 16 h at 56 • C followed by washing the microarray slide as recommended by the manufacturer. Processed microarray slides were then scanned with an Agilent G2565CA Microarray Scanner System (Agilent Technologies, Santa Clara, CA, USA). Scanned images were imported by Agilent Feature Extraction software version 10.7.3.1 (Agilent Technologies) and the fluorescence intensities of each image were quantified using the modified Exiqon protocol and corresponding GAL files.

Data Analysis for Microarray
The data were analyzed with the quantile normalization method. The normalized and log-transformed intensity values were then analyzed using GeneSpring GX 13.1.1 (Agilent Technologies, Santa Clara, CA, USA). This normalization method aims to make the distribution of intensities for each array in a set of arrays the same. The normalized and log-transformed intensity values were then analyzed using GeneSpring GX 13.1.1 (Agilent Technologies, Santa Clara, CA, USA). Fold change filters included the requirement that the genes presented in at least 150% of controls for upregulated genes and less than 66% of controls for downregulated genes. In the Cluster 3.0 program, using the Euclidean distance and average linkage algorithm gave a hierarchical clustering analysis.

Statistical Analysis
Statistical analyses of the data were conducted using Sigmaplot ® version 12 software (Systat Software Inc., San Jose, CA, USA). All data were expressed as means ± standard deviations (SDs). The statistically significant differences between the control and treatment groups were calculated by the Student's t-test and one-way ANOVA test.

Conclusions
The results indicate that SSD exerted cytotoxicity against U266, U937, THP-1, and K562 cancer cells while CPAE, CCD-18Co, and MDBK normal cells were less affected by SSD. The treatment of SSD facilitated ROS-dependent ER stress and triggered apoptosis by regulating CHOP, cleaved PARP, cleaved caspase-3, and Bax in both U266 and U937 cells. The mimic of miR-657 increased the viability of SSD-exposed U266 and U937 cells, indicating that miR-657 inhibition plays a role in SSD-induced apoptosis. Also, SSD treatment decreased the expression of miR-657, leading to the elevation of p-ATF2, CHOP, and cleaved PARP expression levels. In conclusion, SSD induced apoptosis via a novel miR-657/ATF2 signal pathway modification.