1 Introduction

Aloe-emodin (1, 8-dihydroxy-3-(hydroxymethyl)-anthraquinone) is a natural active compound found in leaves of Aloe vera [1]. Aloe-emodin has been found to exhibit multiple activities including antiviral, antimicrobial, hepatoprotective and neuroprotective activities [2]. Emodin has also been reported to induce apoptosis in various tumor cells [3]. Anti-tumor effects of emodin have e.g., been observed in neuroectodermal tumors [4], lung squamous cell carcinomas [5, 6], hepatoma cells [7] and the human U373MG glioma-derived cell line [8, 9]. So far, however, the effect of emodin on human gynecological cancer cells has not been reported.

Previous studies have suggested that the anti-tumor effect of emodin may be exerted through inhibiting the proliferation of cancer cells [10, 11], arresting progression through the cell cycle [12, 13], promoting apoptosis, and inhibiting angiogenesis and metastasis [14, 15]. Emodin has also been found to enhance the sensitivity to chemotherapy and to reverse the multidrug resistance of tumor cells [16, 17]. Several studies have revealed inhibitory effects of emodin on human cervical cancer cells. The underlying mechanisms are possibly associated with emodin-mediated arrest at the G2/M phase of the cell cycle and differentiation induction [13], activation of the apoptotic pathway via caspase-9 [18] and sensitizing cancer cells to paclitaxel-induced apoptosis [19].

Apoptosis is an actively regulated process of cell death, and apoptosis-induced cell death of cancer cells has for many years been the focus for anti-cancer therapy. Apoptosis is mediated by either an extrinsic pathway, triggered by extracellular stimuli, or by an intrinsic pathway, activated by intracellular modulators [20–23]. Mitochondria play a critical role in the intrinsic pathway by releasing a series of molecules, including cytochrome c. Within the cytosol, the binding of cytochrome c to adaptor protein Apaf-1 and pro-caspasse-9 molecules leads to the formation of the apoptosome that is responsible for activation of a caspase-dependent apoptotic pathway [24]. It has e.g., been reported that emodin-induced apoptosis in rat hepatic stellate cells transformed by simian virus 40 (t-HSC/Cl-6) involves a mitochondria-mediated pathway [25].

Here, we investigated the effect of emodin on gynecological cancer-derived cells by evaluating their viability, invasion and apoptosis. To further explore the molecular and cellular mechanisms underlying the anti-proliferative effect of emodin, we measured the mitochondrial membrane potential (ΔΨM) using a JC-1 fluorescent probe and evaluated the expression of an autophagy marker (MAP LC-3). Our results indicate that emodin can inhibit gynecological cancer cell proliferation and invasion by inducing cell cycle arrest and, eventually, apoptosis through an intrinsic apoptotic pathway.

2 Materials and methods

2.1 Cell lines and reagents

The Hela, JAR and HO-8910 cells used were purchased from the Sun Yat-sen cell bank. Emodin, 3-(4-, 5-dimethylthiazol-2-yl)-2, 5-dyphenyl tetrazolium bromide (MTT), dimethylsulfoxide (DMSO) and Trypan Blue were purchased from Sigma Chemical Co. (USA). Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine serum (FBS) were purchased from Thermo (USA). The caspase-9 activity assay kit was purchased from R&D (USA) and the JC-1 probe was purchased from ATT Bioquest (USA). The ATP assay kit was purchased from the Beyotime Institute of Biotechnology (China). Antibodies directed against Bcl-2, Mcl-1, cleaved-caspase-3, Cyclin D, Cyclin E, Beclin-1, Atg12-Atg5, VEGF, VEGFR, GAPDH and β-actin were purchased from Cell Signaling Technology (USA). Fluorescence-conjugated secondary antibodies were purchased from Invitrogen (USA). Other chemicals were obtained in their commercially available highest purity grade.

2.2 Cell culture and emodin treatment

Frozen cells were recovered and cultured in DMEM medium supplemented with 10 % (v/v) FBS, 100 μg/ml streptomycin and 100U/ml penicillin. Cell cultures were maintained at 37 °C in a humidified incubator containing 5 % CO2. Cells were passaged when reaching 80 % confluence at a ratio of 1:3. Cells were used for the respective experiments when they were confluent. Cells were treated with vehicle (0), 5, 10 or 15 μM emodin for 72 h before the execution of further assays.

2.3 MTT assay

Cells were seeded in 96-well plates at a density of 1 × 104 cells/ml in a final volume of 200 μl and maintained in DMEM medium in an incubator containing 5 % CO2 at 37 °C for 72 h. After emodin treatment, cell viability was assessed by MTT assay. To this end, MTT solution (10 μl, 5 mg/ml) was added to each well and plates were incubated at 37 °C for 4 h. Supernatants were removed and 100 μl DMSO (99.0 %) was added to each well to dissolve the resultant formazan crystals. Optical densities were measured at 540 nm using an ELISA-Reader (THERMO, USA). The MTT assays were carried out in triplicate.

2.4 Transwell in vitro invasion assay

To assess cancer cell migration, emodin-treated cells were subjected to migration assays in a Transwell membrane chamber (Transwell®, Corning, NY, USA). Transwell plates (8 μm, Corning) coated with Matrigel (50 μg/ well) were placed in a 6-well plate. Culture medium (400 μl) obtained from 24-h serum-free tumor cell cultures in RPMI-1640 was added to the lower chamber of the Transwell plate. A total of 5 × 104 cells (in 100 μl) was added to the chamber with RPMI-1640 medium containing 10 g/L BSA and 10 ml/L FBS. After 12 h, the Transwell membranes were washed in phosphate buffered saline (PBS) and cells that did not migrate and, thus, remained above the membrane were removed using a cotton swab. Next, the migrated cells were fixed in 95 % ethanol and stained using 4 g/L trypan blue. For cell migration quantification, five fields (upper, lower, left, right and middle) from each membrane were examined under a bright-field microscope and the average cell number was determined.

2.5 Caspase-9 activity assay

Cells were seeded in 6-well plates and confluent cells were treated with 0, 5, 10 or 15 μM emodin for 72 h. Caspase-9 activity measurement was performed using a Caspase-9 Assay Kit (R&D) according to the manufacturer’s instruction. Briefly, cells were collected and resuspended in culture medium at 5 × 105/ml. The cells were centrifuged at 600 × g for 5 min at 4 °C and the resulting cell pellets were rinsed in PBS once, resuspended in lysis buffer at 2 × 106 cells per 100 μl, and incubated on ice for 15 min. After centrifugation at 16,000 × g at 4 °C for 15 min, the supernatant was collected and transferred immediately to a pre-cooled Eppendorf tube. For the colorimetric measurement, 10 μl sample was mixed with 80 μl assay buffer and 10 μl 2 mM Ac-LEHD-pNA after which the mixture was incubated at 37 °C in the dark for 1–2 h. The absorbance was read at 405 nm and sample A405 with blank control A405 deducted was used for data analysis.

2.6 Mitochondrial membrane potential (ΔΨM) assay

Cells in a logarithmic growth phase were seeded in 6-well plates and treated with emodin for 72 h. Next, the cells were collected, digested in 0.25 % EDTA for 1 min, washed in PBS, and centrifuged at 600 × g for 5 min. To prepare a JC-1 working solution, 500 μl 1× incubation buffer was pre-warmed at 37 °C and 1 μl JC-1 probe was added to the incubation buffer. The cells were resuspended in JC-1 working solution and incubated at 37 °C in 5 % CO2 for 20 min. After this, the cells were collected by centrifugation at 600 × g for 5 min and washed in 1× incubation buffer twice. Finally, the cells were resuspended in 500 μl 1× incubation buffer and subjected to flow-cytometry. JC-1 can spontaneously form complexes and emit intense red fluorescence in healthy cells that have a high mitochondrial ΔΨm. In contrast, in apoptotic or unhealthy cells that have a low ΔΨm, JC-1 remains in the monomeric form and only emits green fluorescence.

2.7 ATP assay

Cells in logarithmic growth phase were seeded in 6-well plates and treated with 0, 5, 10 or 15 μM emodin for 72 h. The ATP concentration was measured using an ATP assay kit according to the manufacturer’s instructions (Beyotime Institute of Biotechnology). The culture medium was removed from the 6-well plates and 200 μl cell lysis buffer was added to each well and mixed to lyse the cells adequately. Next, the lysates were centrifuged at 12,000 × g at 4 °C for 10 min and the supernatants were collected for the ATP assay. An ATP standard curve was generated through a serial dilution of ATP standard solution according to the manufacturer’s instructions. To prepare 10 ml standard reaction solution, 0.5 ml of 20× reaction buffer, 0.1 ml of 0.1 M DTT, 0.5 ml of 10 mM D-luciferin and 2.5 μl of 5 mg/ml firefly luciferase stock solution were added to 8.9 ml diH2O. The background luminescence was measured by adding 100 μl standard reaction solution to the luminometer. To measure sample ATPs, cell lysates were added to the reaction solution and luminescence was read. The ATP concentrations of the experimental samples were calculated from the standard curve.

2.8 RT-PCR assay

Total RNA was extracted from emodin-treated cells using a Trizol reagent kit (Invitrogen, USA). The quality of each RNA sample (including its concentration and purity) was checked by measuring the absorbance. One μg RNA from each sample was used to generate cDNA using M-MLV reverse transcriptase as per manufacturer’s specifications (Promega Corporation, USA). After an initial denaturation step at 95 °C for 10 min using a SYBR Green PCR Master Mix (Applied Biosystems, USA), real-time PCR (RT-PCR) was carried out during 40 cycles of 95 °C /15 s and 60 °C /1 min. The amplifications were performed using a 7500 Fast Real-Time PCR System (Applied Biosystems, USA) and the products were routinely checked using dissociation curve software. Transcript quantities were compared using the relative Ct method and the amount of MMP-9 was normalized to the endogenous control (GAPDH). The value in relation to the control sample was given by 2-∆∆CT. The RT-PCR primers used were: MMP-9 sense: 5’-GGAGACCTGAGAACCAATCTC-3’ and MMP-9 anti-sense 5’-TCCAATAGGTGATGTTGTCGT-3’; GAPDH sense 5’-AGAAGGCTGGGGCTCATTTG-3’ and GAPDH anti-sense 5’-AGGGGCCATCCACAGTCTTC-3’.

2.9 MAP LC3 immunofluorescence staining assay

For this assay, cells were washed 3 times in PBS and fixed in 4 % paraformaldehyde for 20 min. After 3 subsequent washes in PBS, the cells were incubated with 0.5 % TritonX-100 in PBS for 20 min, washed 3 times in PBS, blocked with 5 % normal goat serum (NGS), and incubated with a rabbit-anti-MAP LC3 antibody (1:100) at 4 °C overnight. After 3 washes in PBS, the cells were incubated with a PE-conjugated goat-anti-rabbit secondary antibody at 37 °C for 30 min, mounted in medium containing DAPI and examined using fluorescence microscopy.

2.10 Western blotting

Emodin-treated cells were washed in ice-cold PBS and collected in lysis buffer including 50 mM Tris, pH 7.4, 150 mM NaCl, 1 % NP-40, 0.25 % sodium deoxycholate, 0.1 % SDS, 1 mM Na3VO4, 1 mM NaF, 1 mM EDTA, 1 mM PMSF and 1 μg/ml leupeptin. Supernatants were obtained by centrifugation at 13,500 rpm for 20 min, after which total proteins were extracted and protein concentrations were determined by Bradford assay. For immunoblotting, 120 μg protein from each sample was loaded for electrophoresis on 12 % SDS-PAGE gels and the separated proteins were transferred onto PVDF membranes. These membranes were blocked with 5 % non-fat milk powder (w/v) at room temperature for 2 h, and next incubated with primary antibodies directed against Bcl-2 (1:500), Mcl-1 (1:500), Cleaved-caspase-3 (1:500), Cyclin D (1:500), Cyclin E (1:500), Beclin-1 (1:500), Atg12-Atg5 (1:500), VEGF (1:500), VEGFR (1:500), GAPDH (1:1000) and β-actin (1:500), respectively, at 4 °C overnight. After washing, the membranes were incubated with a fluorescence-conjugated secondary antibody (anti-rabbit or anti-mouse, 1:10000) at room temperature for 50 min. GAPDH or β-actin were used as internal controls to monitor equal protein loading and efficient transfer of the proteins from the gels to the membranes. The resulting protein bands were quantified using an Odyssey infrared imaging system (LI-COR, USA). All results obtained represent three independent experiments.

2.11 Statistical analyses

The average values from at least three independent experiments were expressed as Mean ± S.D. All statistical analyses were carried out using SPSS17.0. One-way ANOVA followed by post hoc test was used to compare differences among multiple groups. A p value < 0.05 was considered statistically significant.

3 Results

3.1 Emodin inhibits proliferation and invasion of gynecological cancer cells

To evaluate the effect of emodin on gynecological cancer-derived cells, we used a standard MTT assay to assess their proliferative capacities. Compared to the vehicle treated cells, we found that incubation with 5, 10 or 15 μM emodin significantly decreased the proliferation of HO-8910, Hela and JAR cells. Treatment with 15 μM emodin reduced the viability of HO-8910, Hela and JAR cells by 79 ± 6.4 %, 81 ± 6.8 % and 76 ± 5.2 %, respectively (Fig. 1a). The emodin-mediated inhibition of cell proliferation showed a dose-dependent pattern and the anti-proliferative effect of emodin on Hela cells was greater than that on HO-8910 or JAR cells.

Fig. 1
figure 1

Emodin inhibits proliferation and invasion of gynecological cancer cells. a MTT assay showing dose-dependent inhibition of gynecological cancer cell viability by emodin. b Transwell assay showing decrease in cell invasion by emodin in a dose-dependent manner. c Increase in mRNA MMP-9 level by emodin in a dose-dependent manner. *p < 0.05 vs. control

Full size image

The effect of emodin on the invasive capacity of the respective cancer cells was evaluated using a Transwell invasion assay (Fig. 1b). By doing so, we found that administration of 5 μM emodin significantly decreased the invasion of all cells tested. Incubation with 10 or 15 μM emodin almost completely inhibited the invasion of HO-8910, Hela and JAR cells. Treatment with 15 μM emodin inhibited the invasion of HO-8910, Hela and Jar cells by 97 ± 7.3 %, 94 ± 6.5 % and 95 ± 5.8 %, respectively.

Migration of cancer cells is a characteristic of tumor progression. Many studies on tumor invasion and metastases have focused on degradation of the extracellular matrix (ECM), in which matrix metalloproteinases play a central role [26]. MMP-9, an enzyme that belongs to the matrix metalloproteinase family, is involved in degradation of the ECM in both normal physiological and disease processes. To assess whether the expression of MMP-9 changes during emodin-mediated reduction of cell invasion, we used RT-PCR on emodin-treated cells. We found that treatment with as little as 5 μM emodin resulted in 1.61-, 1.72-, and 1.83-fold increases in MMP-9 mRNA expression in HO-8910, Hela and Jar cells, respectively (Fig. 1c). We also found that the increases in MMP-9 expression showed an emodin dose-dependent pattern. Treatment with 15 μM emodin resulted in an increase in MMP-9 expression of 5.32-fold in the JAR cells.

3.2 Emodin induces apoptosis in gynecological cancer cells

The caspase-9 protease is activated during apoptosis, which leads to the activation of cleaved-caspase-3, DNA damage and apoptosis induction through the mitochondrial death pathway. Here, we found that 15 μM emodin increased the caspase-9 activity by 3.90-, 4.98-, and 4.52-fold in HO-8910, Hela and Jar cells, respectively (Fig. 2a). The Hela cells showed the greatest increase in caspase-9 activity after emodin treatment. The cleaved-caspase-3 protein level was evaluated using Western blotting, and our result showed that 15 μM emodin resulted in 3.42-, 3.76-, and 4.21-fold increases in HO-8910, Hela and Jar cells, respectively (Fig. 2b). The effect of emodin on apoptosis showed a dose-dependent pattern.

Fig. 2
figure 2

Emodin induces apoptosis in gynecological cancer cells. a Emodin-induced increase in caspase-9 activity in gynecological cancer cells in a dose-dependent manner. b Emodin-induced decrease in cleaved-caspase-3 protein level in gynecological cancer cells in a dose-dependent manner. c JC-1 probe assay showing an emodin-induced increase in FL2/FL1 ratio in a dose-dependent manner. d Emodin-induced decrease in ATP release in a dose-dependent manner. *p < 0.05 vs. control

Full size image

Mitochondria use oxidizable substrates to produce a negative inside transmembrane electrical potential. A decrease in ΔΨM has been reported to be associated with apoptosis through exposure of cytochrome c to the inter-membrane space. To test whether emodin affects the cellular ΔΨM, we used a fluorescent probe (JC-1) that is capable of entering selectively into mitochondria and to reversibly change its color from green to orange as the membrane potential increases. The ratio of red/green (FL2/FL1) JC-1 fluorescence represents the ΔΨM. The data were normalized using vehicle treatment. Our results showed that 15 μM emodin significantly decreased the ratio of FL2/FL1 in HO-8910, Hela and Jar cells to 0.21-, 0.10- and 0.19-fold of the vehicle control group (Fig. 2c). This result suggests that emodin treatment induces tumor cell apoptosis. The emodin-induced apoptosis observed showed a dose-dependent pattern in all cell lines tested, and the emodin treatment exhibited a relatively greater effect on Hela cells.

Cellular ΔΨM is a driving force for mitochondrial ATP synthesis, and it declines during apoptosis. Here, we found that the ATP concentration was significantly decreased in cells in an emodin dose-dependent manner. This result is consistent with the increased caspase-9 activity and decreased ΔΨM observed in emodin-treated cells. Treatment with 15 μM emodin decreased the ATP levels to 0.34-, 0.29- and 0.33-fold of the vehicle control group in HO-8910, Hela and Jar cells, respectively (Fig. 2d). Taken together, these results indicate that emodin treatment induces apoptosis in gynecological cancer cells.

3.3 Emodin decreases the expressions of apoptosis regulators in gynecological cancer cells

Bcl-2 is a core member of the Bcl-2 family that comprises a group of anti-apoptotic proteins. High levels of Bcl-2 family proteins are encountered in cells with a failing apoptotic program, often resulting in disease, including cancer. Mcl-1, another protein belonging to the Bcl-2 family, is commonly up-regulated in human tumors and is associated with tumor relapse and chemoresistance. To further explore the molecular mechanism underlying emodin-induced apoptosis, we assessed the levels of Bcl-2 and Mcl-1 in the gynecological cancer-derived cell lines after emodin treatment. Our Western blotting results showed that 5 μM emodin treatment significantly decreased the levels of both the Bcl-2 and Mcl-1 proteins in all cell lines tested. Treatment with 15 μM emodin reduced the Bcl-2 protein levels to 0.42-, 0.36- and 0.41-fold of the control group in HO-8910, Hela and JAR cells, respectively (Fig. 3a). It also decreased the Mcl-1 protein levels to 0.42-, 0.36- and 0.28- fold of the control group in HO-8910, Hela and JAR cells, respectively (Fig. 3b). These results suggest that emodin-induced apoptosis in these cancer-derived cells may be exerted by down-regulation of anti-apoptotic proteins. The development of specific inhibitors for Bcl-2 family member proteins is considered as a potential strategy for cancer therapy. Therefore, emodin-mediated reduction of Bcl-2 family proteins in gynecological cancers may be of particular interest.

Fig. 3
figure 3

Emodin decreases expressions of apoptosis regulators Bcl-2 and Mcl-1. Western blot analysis showing that emodin decreases Bcl-2 protein (a) and Mcl-1 protein (b) levels in HO-8910, Hela and JAR cells in a dose-dependent manner. *p < 0.05 vs. control

Full size image

3.4 Emodin induces cell cycle arrest in gynecological cancer cells

Vehicle treated HO-8910, Hela and JAR cells exhibited 46.55 ± 2.47 %, 45.10 ± 2.42 % and 46.25 ± 2.07 % of the respective cell populations being in the G0/G1 phase, respectively (Table 1). After treatment with 15 μM emodin, we found that 77.30 ± 3.84 % HO-8910 cells, 74.20 ± 3.49 % Hela cells and 74.30 ± 3.98 % JAR cells were in the G0/G1phase. From our results we conclude that emodin treatment significantly increased cell cycle arrest at the G0/G1 phase in all cells tested, and that the effect showed a dose-dependent pattern. We conclude that emodin can inhibit cancer cell proliferation via arresting the cell cycle.

Table 1 Emodin-induced cell cycle distribution in HO-8910, Hela and JAR cells
Full size table

Cyclin D and Cyclin E are two central members of the cyclin protein family that regulates cell cycle progression. In this study, we also assessed protein levels of Cyclin D and Cyclin E in gynecological cancer-derived cells treated with emodin using Western blotting. The expressions of both Cyclin D and Cyclin E was significantly reduced by emodin treatment. Treatment with 15 μM emodin reduced Cyclin D expression to 0.22-, 0.26- and 0.18-fold of the control group (Fig. 4a) and reduced Cyclin E expression to 0.37-, 0.26- and 0.41-fold of the control group (Fig. 4b) in HO-8910, Hela and JAR cells, respectively. This result suggests that emodin treatment affects cell cycle progression in gynecological cancer-derived cells.

Fig. 4
figure 4

Emodin decreases cell cycle regulators in gynecological cancer cells. Western blot analysis showing that emodin decreases Cyclin D1 protein (a) and Cyclin E1 protein (b) levels in HO-8910, Hela and JAR cells in a dose-dependent manner. *p < 0.05 vs. control

Full size image

3.5 Emodin induces autophagy in gynecological cancer cells

Previous studies have shown that emodin can block the cell cycle, induce caspase-dependent apoptosis and lead to the formation of intracytoplasmic acidic vesicles in C6 glioma-derived cells, which is indicative of autophagic cell death [9]. To evaluate the effect of emodin on autophagy in gynecological cancer-derived cells, we used immunofluorescent staining with an antibody directed against MAP LC3, a major constituent of the autophagosome, to assess the expression level of MAP LC3 (red) in HO-8910, Hela and JAR cells in response to emodin treatment (Fig. 5). Compared to the vehicle treated cells, emodin increased MAP LC3 expression in HO-8910 and Hela cells in a dose-dependent manner. The JAR cells also showed increased MAP LC3 expression in response to 5 and 10 μM emodin treatment, but a decreased expression after 15 μM emodin treatment. A reduced viability of JAR cells after 15 μM emodin treatment may be responsible for the decrease in MAP LC3 expression observed.

Fig. 5
figure 5

MAP LC3 immunostaining in emodin-treated gynecological cancer cells. MAP LC expression (in red) in HO-8910, Hela and JAR cells treated with 0, 5, 10 or 15 μM emodin. DAPI (in blue) was used for nuclear staining. Scale bar = 100 μm

Full size image

We also evaluated the expression of two other important autophagy regulators, i.e., Beclin-1 and Atg12-Atg5, in emodin-treated cells. Beclin-1 regulates autophagy-associated Atg genes and is required for the formation of the autophagosome during autophagy. The Atg genes control autophagosome formation through Atg12-Atg5 and LC3-II (ATG8-II) complexes. Our Western blotting results showed that emodin increased Beclin-1 expression in the gynecological cancer-derived cells tested in a dose-dependent manner. Treatment with 15 μM emodin increased Beclin-1 expression to 3.37-, 3.26- and 3.12-fold of the control level in HO-8910, Hela and JAR cells, respectively (Fig. 6a). Additionally, Atg12-Atg5 complex formation was also found to be increased in emodin-treated cells in a dose-dependent manner. Treatment with 15 μM emodin increased the Atg12-Atg5 complex to 3.67-, 4.26- and 2.81-fold of the control level in HO-8910, Hela and JAR cells, respectively (Fig. 6b).

Fig. 6
figure 6

Emodin increases autophagy regulators Beclin-1 and Atg12-Atg5. Western blot analysis showing that emodin increases Beclin-1 (a) and Atg-12-Atg5 (b) protein levels in HO-8910, Hela and JAR cells in a dose-dependent manner. *p < 0.05 vs. control

Full size image

3.6 Emodin affects the expression of VEGF and VEGFR-2 in gynecological cancer cells

Over-expression of VEGF in primary tumor cells and serum is associated with a poor survival rate in patients with ovarian cancer. The association between a high tissue VEGF level and a poor prognosis has also been reported in early stage ovarian cancer patients [27]. To ask whether emodin treatment has an effect on the expression of the VEGF and VEGFR proteins in gynecological cancer-derived cells, we used Western blotting to assess their levels. We found that emodin treatment decreased both VEGF (Fig. 7a) and VEGFR-2 (Fig. 7b) protein levels in a dose-dependent manner in the gynecological cancer-derived cells tested. Our quantitative results showed that 15 μM emodin decreased the protein levels of VEGF to 0.42-, 0.37- and 0.35-fold of the control level, and VEGFR-2 to 0.32-, 0.36- and 0.38-fold of the control level in HO-8910, Hela and JAR cells, respectively. This result suggests that emodin may have an anti-angiogenic effect and may be used as a potential inhibitor for gynecological tumor angiogenesis. This observation is consistent with a previous report on the anti-angiogenic effect of emodin in human colon cancer cells [28].

Fig. 7
figure 7

Expression levels of VEGF and VEGFR-2 in gynecological cancer cells. Western blot analysis showing that emodin significantly decreases VEGF (a) and VEGFR-2 (b) protein levels in HO-8910, Hela and JAR cells in a dose-dependent manner. *p < 0.05 vs. control

Full size image

4 Discussion

Here, we report an anti-tumor effect of emodin on gynecological cancer-derived cells, i.e., HO-8910, Hela and JAR. The effect of emodin exhibited a dose-dependent pattern. We found that emodin-induced apoptosis and cell cycle arrest may explain the anti-tumor effect of emodin.

The mechanism of cell cycle regulation is closely related to tumorigenesis. Many tumor suppressors, such as p53, RB, BRCA1, p16 and p15, as well as their downstream regulators, such as p21 and Gadd45, are important components of cell cycle checkpoints [29–32]. Interactions between cell cycle-associated proteins, i.e., cyclins (Cyclin D1 and Cyclin E1) and cyclin-dependent kinases (CDKs; CDK4 and CDK2) regulate progression through the G1 phase. RB, an important tumor suppressor (or transcriptional repressor), can also inhibit cell cycle progression via interacting with Cyclin D, CDK4 and p16. Loss of RB function in tumor cells results in disruption of the cell cycle and, subsequently, malignant proliferation of tumor cells. In addition, phosphorylation of RB acts as an important regulatory step in cell cycle progression. At the initial stage of the G1 the phase, the Cyclin D/CDK4 complex promotes RB phosphorylation and the concomitant release of RB binding proteins, such as transcription factor E2F family proteins. E2F family proteins stimulate the expression of G1/S genes, including Cyclin E, and these molecules promote progression through the G1/S check point and initiation of the S phase. When DNA damage occurs, high p21 expression inhibits RB phosphorylation through Cyclin D/CDK4. Binding of unphosphorylated RB to E2F proteins prevents the expression of G1/S genes and, thus, arrests cells at the G1/S transition checkpoint [33]. Here, we found that the expression levels of Cyclin D1 and Cyclin E1 decreased in emodin-treated cells, concomitantly with G1/S cell cycle arrest. We speculate that a decreased level of Cyclin D expression interferes with RB phosphorylation and, by doing so, suppresses the expression of downstream factors, such as Cyclin E and the inhibition of progression through G1/S, resulting in cell cycle arrest.

The anti-proliferative effects of emodin have been reported to be regulated through different cellular signaling pathways. Emodin has e.g., been shown to target the proliferation of human colon cancer cells by inducing G2/M cell cycle arrest and apoptosis via activation of the caspase-9/6 pathway [12]. Emodin treatment has also been found to cause G2/M cell cycle arrest in liver cancer cells via increasing the expression levels of cell cycle regulatory proteins including Cyclin A, Cyclin B, Chk2, Cdk2 and p27, and decreasing the expression of Cdc25c and p21. Emodin-induced cell cycle arrest has also been reported in Hela cells and this effect was found to be associated with decreased expression levels of Cyclin A and CDK2, increased expression levels of Cyclin B1 and CDK1 and an increased activity of alkaline phosphatase [13]. Here, we found that the anti-proliferative effect of emodin was associated with decreased expression levels of the cell cycle regulators Cyclin D1 and Cyclin E1. Moreover, we found that the dose-dependent down-regulation of these cell cycle regulators by emodulin was temporally correlated with up-regulation of cleaved-caspase-3 in all three gynecological cancer-derived cell lines tested. Together, these results suggest that emodin can inhibit the proliferation of gynecological cancer cells and arrest their cell cycle through (de)regulating cyclin proteins.

We found that both apoptosis and autophagy may underlie the emodin-mediated anti-tumor effect on gynecological cancer cells. Autophagy and apoptosis, two different forms of programmed cell death, may be functionally correlated with each other in three ways. First, autophagy may be necessary for apoptosis and, thus, may initiate apoptosis [34–37]. Second, autophagy may protect cells through inhibition of apoptosis or necrosis. It can also protect cells from apoptosis through increasing hypoxia tolerance [38]. Third, autophagy and apoptosis may synergistically promote cell death. Both inhibition of autophagy and apoptosis can promote cell death and inhibition of autophagy may convert its death signal pathway to apoptotic cell death, while inhibition of apoptosis through caspase inhibitors may promote cell death through the autophagy pathway [39, 40]. How apoptosis and autophagy interact with each other in emodin-treated gynecological cancer cells remains, however, to be established.

Emodin-induced apoptosis has been reported to contribute to its anti-proliferative effect in various types of cancer cells. However, the exact mechanism underlying emodin-mediated apoptosis has so far not been elucidated. In human hepatoma Huh-7 cells, emodin-mediated apoptosis was found to be associated with down-regulation of calpain-2 and ubiquitin-protein ligase E3A [41]. Emodin has also been shown to enhance the activity of gemcitabine-mediated pro-apoptosis in pancreatic cancer cells via Akt inhibition and NF-kappaB activation, thus promoting the mitochondria-dependent apoptotic pathway [42]. In the pancreatic cancer cell line SW1990E, emodin was found to induce apoptosis by significantly down-regulating NF-kappaB DNA-binding activity, and survivin and MMP-9 expression in SW1990 cells [43–45]. Moreover, the expression of cleaved-caspase-3 was found to be up-regulated in SW1990 cells after emodin treatment [15]. In addition, it has been shown that emodin-induced apoptosis of human nasopharyngeal carcinoma cells involves caspase-8-mediated activation of the mitochondrial death pathway [46]. Here, we found that emodin treatment of gynecological cancer cells significantly induced caspase-9 activity in a dose-dependent manner. Coincident with the increased capase-9 activity, we observed decreased ATP levels and Bcl-2 and Mcl-1 expression levels, but increased cleaved-caspase-3 protein levels. Together, these results suggest that emodin may exert its pro-apoptotic effect though (de)regulating the expression levels of apoptotic proteins, which is consistent with previous reports.

We also wondered whether, in addition to the apoptotic pathway, emodin may affect autophagy. Although limited, it has been shown that emodin may exert an anti-glioma effect through ERK inhibition [9, 47]. Emodin has also been found to lead to the formation of intra-cytoplasmic acidic vesicles that are indicative of autophagic cell death. Here, we found that emodin significantly increased the expression of MAP LC3, an important component of the autophagosome, and up-regulated the expression of Beclin-1 and Atg12-Atg5. These results further support the notion that emodin may induce cancer cell apoptosis through initiating autophagy. Beclin-1 is also known as autophagy-related gene Atg6, and is required for initiation of the formation of the autophagosome through binding to phosphoinositide 3-kinase (PI3K). Extensive crosstalk is known to exist between autophagy and apoptosis, and these processes can be both negatively and positively related. Autophagy may act as a pro-survival mechanism when nutrients are lacking. However, excessive autophagy causes cell death, a process that is morphologically distinct from apoptosis. Some pro-apoptotic signal molecules, such as TNF, TRAIL and FADD, may also induce autophagy. Beclin-1-dependent autophagy can be inhibited by Bcl-2, which functions both as a pro-survival and as an anti-autophagic regulator. Here, we observed an increase in Beclin-1 expression in an emodin dose-dependent manner, suggesting initiation of autophagosome formation in emodin-treated cells. Increased Atg12-Atg5 levels that we observed in emodin-treated cancer cells further suggest an effect of emodin on autophagy through regulating autophagosome formation.

In summary, we report an anti-tumor effect of emodin on gynecological cancer-derived cells. This effect may be exerted through a combination of multiple functions including inhibition of proliferation, induction of apoptosis and autophagy, and inhibition of angiogenesis. Our study provides a basis for using emodin as a potential therapeutic tool for the treatment of gynecological tumors.