Retrochalcone Echinatin Triggers Apoptosis of Esophageal Squamous Cell Carcinoma via ROS- and ER Stress-Mediated Signaling Pathways

Esophageal squamous cell carcinoma (ESCC) is a poor prognostic cancer with a low five-year survival rate. Echinatin (Ech) is a retrochalone from licorice. It has been used as a herbal medicine due to its anti-inflammatory and anti-oxidative effects. However, its anticancer activity or underlying mechanism has not been elucidated yet. Thus, the objective of this study was to investigate the anti-tumor activity of Ech on ESCC by inducing ROS and ER stress dependent apoptosis. Ech inhibited ESCC cell growth in anchorage-dependent and independent analysis. Treatment with Ech induced G2/M phase of cell cycle and apoptosis of ESCC cells. It also regulated their related protein markers including p21, p27, cyclin B1, and cdc2. Ech also led to phosphorylation of JNK and p38. Regarding ROS and ER stress formation associated with apoptosis, we found that Ech increased ROS production, whereas its increase was diminished by NAC treatment. In addition, ER stress proteins were induced by treatment with Ech. Moreover, Ech enhanced MMP dysfunction and caspases activity. Furthermore, it regulated related biomarkers. Taken together, our results suggest that Ech can induce apoptosis in human ESCC cells via ROS/ER stress generation and p38 MAPK/JNK activation.


Introduction
Esophageal cancer is the one of most common cancer types. In 2018, approximately 321,670 patients in the USA suffer from this cancer, and the total estimated number of deaths from the disease was 13,020 (4%). Its global mortality rate is currently ranked fifth for men and ninth for women [1,2]. There are different types of esophageal cancer, including esophageal squamous cell carcinoma (ESCC:~90%) and esophageal adenocarcinoma (EA:~10%) [3]. In particular, ESCC shows a high incidence rate Colonies were counted using a microscope. The graph shows the percentage compared to the control group. Data represent mean ± SD. Asterisk (*) denotes p < 0.05 compared to the control.

Ech Arrests Cell Cycle of ESCC Cells at G2/M Phase and Induces Apoptosis
Cell growth processes contain the cell cycle's promotion [16]. Thus, Ech may affect the cell cycle and cause ESCC cell growth inhibition. When we treated KYSE 30 and KYSE 450 ESCC cells with Ech at 0, 5, 10, or 15 µM, cell cycles were accumulated at G2/M phase compared to control (Figure 2a). Sub-G1 population was dose-dependently increased by Ech (increase after treatment with Ech at 0, 5, 10, or 15 µM: 8.17 ± 0.99, 11.83 ± 1.78, 11.87 ± 0.55, and 36.53 ± 2.02% in KYSE 30 cells; 7.57 ± 0.47, 15.97 ± 0.25, 23.80 ± 1.15, and 36.47 ± 0.93% in KYSE 450 cells, respectively) ( Figure 2b). Sub-G1 death cells can be caused by apoptosis or necrosis [17]. Thus, we stained cells with Annexin V for apoptosis or 7-Aminoactinomycin D (7-AAD) for necrosis ( Figure 2c). Early apoptosis percentage of Annexin V+/7-AAD-gating was increased to 9.69 ± 0.17% or 16.79 ± 1.12%, while the late apoptosis percentage of Annexin V+/7-AAD+ gating was increased to 27.68 ± 1.53 or 19.02 ± 0.83% in KYSE 30 or KYSE 450 ESCC cells after treatment with 15 µM Ech, respectively (Figure 2c). To verify the effects of Ech on cell cycle and apoptosis, we conducted Western blot to examine expression of the cell cycle at G2/M phase and apoptosis signaling markers (Figure 3a,b). After KYSE 30 and KYSE 450, cells were treated with Ech at 5, 10, or 15 µM for 48 h, expression levels of cell cycle markers p21 and p27 were increased while those of cyclin B1 and cdc2 were decreased compared the control (Figure 3a). For apoptosis signaling markers, Ech induced expression levels of p-JNK and p-p38 mitogen-activated protein Colonies were counted using a microscope. The graph shows the percentage compared to the control group. Data represent mean ± SD. Asterisk (*) denotes p < 0.05 compared to the control.

Ech Induces Apoptosis by Increasing ROS Levels and ER Stress
To determine the increase of p-p38 and p-JNK expression via induction of ROS, we detected ROS levels after treatment with dimethyl sulfoxide (DMSO) as a control and Ech (5, 10, 15 µM) for 48 h (Figure 4a). Ech at 0, 5, 10, and 15 µM induced ROS levels by 6.71 ± 0.57, 12.06 ± 0.38, 14.84 ± 0.76, and 37.17 ± 1.01% in KYSE 30 cells, as well as 49.98 ± 1.28, 56.07 ± 1.68, 63.02 ± 0.54, and 70.27 ± 2.99% in KYSE 450 cells, respectively. To confirm the involvement of ROS in apoptosis induction, we measured viabilities of KYSE 30 and KYSE 450 cells treated with a combination of ROS scavenger Nacetyl-L-cysteine (NAC, 6 mM) and Ech (15 µM) (Figure 4b). Results revealed that treatment with NAC only did not significantly affect the viabilities of either cell line. However, Ech significantly decreased the viabilities of KYSE 30 and KYSE 450 cells. Its decreases were recovered by NAC treatment (Figure 4b). To further determine whether Ech-induced ROS could activate ER stress cascades, thereby inducing apoptosis of ESCC cells, we examined expression levels of ER stress related proteins (Figure 4c). Ech induced DR4 and DR5 expression in KYSE 30 and KYSE 450 cells in a dose-dependent manner (Figure 4c). Expression levels of GRP78 and CHOP, down-stream targets of DR4 and DR5 proteins, were also increased by Ech in a dose-dependent manner compared to those in the control (DMSO treated).

Ech Induces Apoptosis by Increasing ROS Levels and ER Stress
To determine the increase of p-p38 and p-JNK expression via induction of ROS, we detected ROS levels after treatment with dimethyl sulfoxide (DMSO) as a control and Ech (5, 10, 15 µM) for 48 h (Figure 4a). Ech at 0, 5, 10, and 15 µM induced ROS levels by 6.71 ± 0.57, 12.06 ± 0.38, 14.84 ± 0.76, and 37.17 ± 1.01% in KYSE 30 cells, as well as 49.98 ± 1.28, 56.07 ± 1.68, 63.02 ± 0.54, and 70.27 ± 2.99% in KYSE 450 cells, respectively. To confirm the involvement of ROS in apoptosis induction, we measured viabilities of KYSE 30 and KYSE 450 cells treated with a combination of ROS scavenger N-acetyl-l-cysteine (NAC, 6 mM) and Ech (15 µM) (Figure 4b). Results revealed that treatment with NAC only did not significantly affect the viabilities of either cell line. However, Ech significantly decreased the viabilities of KYSE 30 and KYSE 450 cells. Its decreases were recovered by NAC treatment (Figure 4b). To further determine whether Ech-induced ROS could activate ER stress cascades, thereby inducing apoptosis of ESCC cells, we examined expression levels of ER stress related proteins (Figure 4c). Ech induced DR4 and DR5 expression in KYSE 30 and KYSE 450 cells in a dose-dependent manner (Figure 4c). Expression levels of GRP78 and CHOP, down-stream targets of DR4 and DR5 proteins, were also increased by Ech in a dose-dependent manner compared to those in the control (DMSO treated).

Ech Provokes Mitochondrial Dysfunction and Caspases Activation
Next, we determined whether the induction of ROS level and ER stress caused by Ech treatment influenced mitochondrial membrane potential (MMP) dysfunction ( Figure 5). Ech at 15 µM obviously induced depolarization of MMP in KYSE 30 and KYSE 450 ESCC cells by 31.01 ± 1.72 and 44.05 ± 0.43%, respectively (Figure 5a). Mitochondria-mediated apoptosis markers were also regulated by Ech treatment (Figure 5b). Ech decreased expression of Bid, and Bcl-2 but increased Bax expression. In addition, Ech resulted in the release of cyto C from mitochondria to the cytosol in a dose-dependent

Ech Provokes Mitochondrial Dysfunction and Caspases Activation
Next, we determined whether the induction of ROS level and ER stress caused by Ech treatment influenced mitochondrial membrane potential (MMP) dysfunction ( Figure 5). Ech at 15 µM obviously induced depolarization of MMP in KYSE 30 and KYSE 450 ESCC cells by 31.01 ± 1.72 and 44.05 ± 0.43%, respectively (Figure 5a). Mitochondria-mediated apoptosis markers were also regulated by Ech treatment (Figure 5b). Ech decreased expression of Bid, and Bcl-2 but increased Bax expression. In addition, Ech resulted in the release of cyto C from mitochondria to the cytosol in a dose-dependent manner compared to the DMSO control and α-tubulin and COX-4 control fraction proteins (Figure 5b). Expression levels of the following signal proteins of apoptosis, Apaf-1 and cleaved Poly (ADP-Ribose) Polymerase (c-PARP), were increased by treatment with Ech in a dose-dependent manner compared to those in the control. We also measured activities of caspases, an upstream protein of PARP, and a downstream protein of Apaf-1 in apoptosis signaling pathway ( Figure 6).
Molecules 2019, 24, x 7 of 14 manner compared to the DMSO control and α-tubulin and COX-4 control fraction proteins ( Figure  5b). Expression levels of the following signal proteins of apoptosis, Apaf-1 and cleaved Poly (ADP-Ribose) Polymerase (c-PARP), were increased by treatment with Ech in a dose-dependent manner compared to those in the control. We also measured activities of caspases, an upstream protein of PARP, and a downstream protein of Apaf-1 in apoptosis signaling pathway ( Figure 6).

Discussion
Chemotherapy of ESCC still needs further investigation to obtain proper efficiency with low side effects [18]. Natural medicines such as aspirin from willow bark and eupatilin from artemisia have low side effects. They have been used for treating cancers [19]. Here, we examined the anticancer effect of Ech from Glycyrrhiza inflata (licorice) on ESCC and found that it had inhibitory activities against ESSC cells. Although other components of licorice such as licochalcones A, B, and C have shown anticancer effects on colon, skin, and oral cancers [20][21][22], the anticancer effect of Ech has not been well elucidated yet. Many anticancer agents can inhibit cancer cell proliferation by arresting cell cycle at G1-or G2/M-phase [23]. The G2 checkpoint can prevent cells from entering mitosis when DNA is damaged. It ensures the propagation of error-free copies of the genome to each daughter cell. Cdk1/cyclin B1 complex controls the cell cycle progression from the G2 phase to the M phase by regulating phosphorylation or dephosphorylation of proteins [24]. In addition, actin remodeling in coordination can ensure proper execution of G2/M checkpoint arrest. It is crucial for entry into mitosis. Flow-cytometry analysis results of the present study revealed that Ech induced G2/M phase arrest of cell cycle (Figure 2a). We then detected expression of related markers, including cyclin B1, cdc2 (Cdk1), and p27 (Figure 3a). Increase of p27 expression regulates cell cycle at G2/M and suppresses cdc2/cyclin B1 expression [25]. Apoptosis is a type I cell death. It physiologically shows plasma membrane shrinkage and nuclear fragmentation following death ligands (extrinsic pathway) and DNA damage/cell stresses (intrinsic pathway) [26]. Cell stresses caused by chemotherapeutic agents can stimulate stress activated MAPKs including p38 MAPK and JNK [15]. Ech induced activation of p38 MAPK and JNK based on the detection of their phosphorylation forms by Western blotting (Figure 3b). Accumulating evidence suggests that protein folding and generation of ROS as a byproduct of protein oxidation in the ER are closely linked to each other [27]. ROS have emerged as crucial regulators of ER function in several diseases. Induction of ER stress and ROS production occur concurrently. GRP78 and CHOP are commonly used as markers of ER stress. GRP78, the master regulator of the unfolded protein response (UPR), plays a role in proliferation, invasion, and metastasis in cancer [28]. GRP78 represents both a regulator and a target of the UPR. It is associated with pro-survival responses. Conversely, GRP78 can interact with components of ER related proapoptotic pathways [29]. In a previous study, only GRP78 negative colon cancer cells were found to be highly proliferative. They induced significant growth in tumor size and metastasized to the liver [30]. In contrast, GRP78 positive cells manifested reduced proliferation, colony formation, tumor

Discussion
Chemotherapy of ESCC still needs further investigation to obtain proper efficiency with low side effects [18]. Natural medicines such as aspirin from willow bark and eupatilin from artemisia have low side effects. They have been used for treating cancers [19]. Here, we examined the anticancer effect of Ech from Glycyrrhiza inflata (licorice) on ESCC and found that it had inhibitory activities against ESSC cells. Although other components of licorice such as licochalcones A, B, and C have shown anticancer effects on colon, skin, and oral cancers [20][21][22], the anticancer effect of Ech has not been well elucidated yet. Many anticancer agents can inhibit cancer cell proliferation by arresting cell cycle at G1-or G2/M-phase [23]. The G2 checkpoint can prevent cells from entering mitosis when DNA is damaged. It ensures the propagation of error-free copies of the genome to each daughter cell. Cdk1/cyclin B1 complex controls the cell cycle progression from the G2 phase to the M phase by regulating phosphorylation or dephosphorylation of proteins [24]. In addition, actin remodeling in coordination can ensure proper execution of G2/M checkpoint arrest. It is crucial for entry into mitosis. Flow-cytometry analysis results of the present study revealed that Ech induced G2/M phase arrest of cell cycle (Figure 2a). We then detected expression of related markers, including cyclin B1, cdc2 (Cdk1), and p27 (Figure 3a). Increase of p27 expression regulates cell cycle at G2/M and suppresses cdc2/cyclin B1 expression [25]. Apoptosis is a type I cell death. It physiologically shows plasma membrane shrinkage and nuclear fragmentation following death ligands (extrinsic pathway) and DNA damage/cell stresses (intrinsic pathway) [26]. Cell stresses caused by chemotherapeutic agents can stimulate stress activated MAPKs including p38 MAPK and JNK [15]. Ech induced activation of p38 MAPK and JNK based on the detection of their phosphorylation forms by Western blotting (Figure 3b). Accumulating evidence suggests that protein folding and generation of ROS as a byproduct of protein oxidation in the ER are closely linked to each other [27]. ROS have emerged as crucial regulators of ER function in several diseases. Induction of ER stress and ROS production occur concurrently. GRP78 and CHOP are commonly used as markers of ER stress. GRP78, the master regulator of the unfolded protein response (UPR), plays a role in proliferation, invasion, and metastasis in cancer [28]. GRP78 represents both a regulator and a target of the UPR. It is associated with pro-survival responses. Conversely, GRP78 can interact with components of ER related pro-apoptotic pathways [29]. In a previous study, only GRP78 negative colon cancer cells were found to be highly proliferative. They induced significant growth in tumor size and metastasized to the liver [30]. In contrast, GRP78 positive cells manifested reduced proliferation, colony formation, tumor growth, and liver metastases [30]. CHOP as a transcription factor is also involved in ER stress-induced apoptosis [31]. Ech induced ROS formation, ER stress, and expression of DR4, DR5, GRP78, and CHOP biomarkers (Figure 4). We also confirmed Ech-induced cell death via ROS through treatment with NAC (Figure 4b). ER stress and ROS formation can induce mitochondrial potential disruption and result in apoptosis [32,33]. Treatment of ESCC cells with Ech induced MMP dysfunction and released cyto C from mitochondria to cytosol, thereby activating caspases (Figures 5 and 6). In summary, Ech inhibited ESCC cell growth by inducing intrinsic and extrinsic apoptosis pathways through ROS-and ER-stress mediated signaling (Figure 7). At present, phase I or phase II clinical trials using natural compounds such as resveratrol and grape powder (NCT 00256334, NCT 01370889), ginger root extract (NCT 01344538), and sulforaphane (NCT 01228084) for colorectal cancer or prostate cancer have been conducted or are being conducted. A successful natural medicine with clear evaluation and prediction in clinic is needed. Thus, it is necessary to investigate the action mechanisms of these compounds in detail.
Molecules 2019, 24, x 9 of 14 growth, and liver metastases [30]. CHOP as a transcription factor is also involved in ER stressinduced apoptosis [31]. Ech induced ROS formation, ER stress, and expression of DR4, DR5, GRP78, and CHOP biomarkers ( Figure 4). We also confirmed Ech-induced cell death via ROS through treatment with NAC ( Figure 4b). ER stress and ROS formation can induce mitochondrial potential disruption and result in apoptosis [32,33]. Treatment of ESCC cells with Ech induced MMP dysfunction and released cyto C from mitochondria to cytosol, thereby activating caspases ( Figures  5 and 6). In summary, Ech inhibited ESCC cell growth by inducing intrinsic and extrinsic apoptosis pathways through ROS-and ER-stress mediated signaling (Figure 7). At present, phase I or phase II clinical trials using natural compounds such as resveratrol and grape powder (NCT 00256334, NCT 01370889), ginger root extract (NCT 01344538), and sulforaphane (NCT 01228084) for colorectal cancer or prostate cancer have been conducted or are being conducted. A successful natural medicine with clear evaluation and prediction in clinic is needed. Thus, it is necessary to investigate the action mechanisms of these compounds in detail.  and JNK MAPK kinases, or stimulates ER stress. p38 and JNK MAPK kinases signaling cascades increase death receptor expression and caspase-8/Bid activation, BAX expression, Bcl2 reduction, and cytochrome c release (cytosol), which then induces Apaf-1/caspase9 or caspase-3 activation, thus resulting in cell apoptosis. On the other hand, ER stress with increasing GRP78 and CHOP expression will induce cancer cell apoptosis.

Cell Cycle Analysis
KYSE 30 (7.5 × 10 4 cells/well) and KYSE 450 (10.5 × 10 4 cells/well) cells were seeded into 6-well plates and exposed to DMSO or Ech (5, 10, 15 µM) for 48 h. Cells were harvested and washed with 1× phosphate-buffered saline (PBS) three times and then fixed in 70% ethanol at −20 • C overnight. Cells were centrifuged at 4000 rpm for 10 min, washed with 1× PBS three times, and then suspended with 200 µL of Muse™ Cell Cycle Reagent (EMD Millipore, Billerica, MA, USA) for 30 min in the dark. The cell cycle status of cells was analyzed with a Muse™ Cell Analyzer (EMD Millipore, Billerica, MA, USA).

Western Blotting
Cells were lysed in RIPA buffer (iNtRON, Gyeonggi-do, Korea) for 30 min on ice. Proteins were separated by SDS-PAGE and transferred onto polyvinylidene fluoride membranes (EMD Millipore, Billerica, MA, USA). These membranes were blocked with 3% or 5% skim milk at RT for 2 h and incubated with primary antibodies (dilution, 1:1000) at 4 • C overnight. After removing non-specific binding antibodies with 1× washing buffer (PBS with 0.1% Tween-20) three times, blots were probed with HRP-conjugated secondary antibody (dilution of 1:7000) in skim milk at RT for 2 h. Specific bands were detected using an ImageQuant LAS 500 (GE Healthcare, Uppsala, Sweden). .01 mg/mL aprotinin, 0.01 mg/mL leupeptin) and 0.05% digitonin at RT for 1 min. Cytosolic fraction (supernatant) was harvested after centrifugation at 13,000 rpm for 5 min at 4 • C. For the mitochondrial fraction, cell pellets were further suspended with plasma membrane extraction buffer and 0.5% Triton X-100. Suspended pellets were tapped and incubated on ice for 10 min. The mitochondrial fraction was harvested after centrifugation at 13,000 rpm for 30 min. Cyto C expression in the cytosolic or membrane fraction was detected by Western blotting.

Multi-Caspase Activity Analysis
ESCC cells were treated with DMSO or Ech (5, 10, 15 µM) and harvested to measure multi-caspase activity. Samples were mixed with 1× caspase buffer and 50 µL of Muse™ Multi-Caspase reagent working solution (Muse™ Multi-Caspase kit, EMD Millipore), incubated at 37 • C for 30 min, and then added with 125 µL of Muse™ caspase and 7-AAD for each sample. Multi-caspase stained cells were then analyzed with the Muse™ Cell Analyzer.

Statistical Analysis
Data are presented as means ± standard deviation (SD). All statistical analyses of data were performed using Prism 5.0 statistical package. Statistical significance of differences among groups was analyzed using analysis of variance (ANOVA). Mean values were considered to be significantly different at p < 0.05.