Corilagin inhibits breast cancer growth via reactive oxygen species‐dependent apoptosis and autophagy

Abstract Corilagin is a component of Phyllanthus urinaria extract and has been found of possessing anti‐inflammatory, anti‐oxidative, and anti‐tumour properties in clinic treatments. However, the underlying mechanisms in anti‐cancer particularly of its induction of cell death in human breast cancer remain undefined. Our research found that corilagin‐induced apoptotic and autophagic cell death depending on reactive oxygen species (ROS) in human breast cancer cell, and it occurred in human breast cancer cell (MCF‐7) only comparing with normal cells. The expression of procaspase‐8, procaspase‐3, PARP, Bcl‐2 and procaspase‐9 was down‐regulated while caspase‐8, cleaved PARP, caspase‐9 and Bax were up‐regulated after corilagin treatment, indicating apoptosis mediated by extrinsic and mitochondrial pathways occurred in MCF‐7 cell. Meanwhile, autophagy mediated by suppressing Akt/mTOR/p70S6K pathway was detected with an increase in autophagic vacuoles and LC3‐II conversion. More significantly, inhibition of autophagy by chloroquine diphosphate salt (CQ) remarkably enhanced apoptosis, while the caspase inhibitor z‐VAD‐fmk failed in affecting autophagy, suggesting that corilagin‐induced autophagy functioned as a survival mechanism in MCF‐7 cells. In addition, corilagin induced intracellular reactive oxygen species (ROS) generation, when reduced by ROS scavenger NAC, apoptosis and autophagy were both down‐regulated. Nevertheless, in SK‐BR3 cell which expressed RIP3, necroptosis inhibitor Nec‐1 could not alleviate cell death induced by corilagin, indicating necroptosis was not triggered. Subcutaneous tumour growth in nude mice was attenuated by corilagin, consisting with the results in vitro. These results imply that corilagin inhibits cancer cell proliferation through inducing apoptosis and autophagy which regulated by ROS release.


| INTRODUCTION
Breast cancer is a primary carcinoma in threatening women's health, the incidence of which is rising in many countries. 1 In addition to its hereditary, the increasing risk factors are mainly late age of pregnancies, shorter or no periods of breastfeeding caused by pressure of competition and works with the development of economy. Obesity or inactivity caused by bad living habits can also increase the incidence. Although great advances have been made in treating breast cancer through using various therapies, which include chemotherapy, endocrine therapy and radiotherapy, 2,3 many problems such metastasis, tumour recurrence or drug resistance still exist after receiving treatments above mentioned. Consequently, it is urgent to expand treatments and find more impactful drugs for treating breast cancer.
Corilagin, a polyphenol tannic acid, extracting from Phyllanthus urinaria, possesses antiviral, anti-inflammatory, anti-oxidant, hypolipemic, hypotensive and anti-coagulation activities. 4,5 Many recent studies show that corilagin has effect in tumour therapy as a potential agent candidate. These researches demonstrated that corilagin could control cancer cell growth through regulating various cell signal pathways in different cancers,such as glioblastoma, 6 hepatocellular carcinoma 7,8 and ovarian cancer. 9,10 Corilagin could significantly diminish inflammation through reducing the secretion of IL-8, MCP-1 and RANTES in cystic fibrosis bronchial IB3-1 cells. 11 Besides, corilagin shows inhibition of cell growth through blocking the activation of both canonical Smad and non-canonical ERK/AKT pathways in ovarian cancer. 9 Another study shows that in colitis, corilagin could suppress the secretion of inflammatory cytokines, inhibit the degradation of IjB a, 12 and besides, it is worth noting that corilagin meanwhile inhibits apoptosis of intestinal epithelial cells. 12 Despite the fact that corilagin has biological activities in anti-inflammatory and anti-tumour, the possibility of its function or even mechanism in inhibiting breast cancer is still uncertain.
Programmed cell death (PCD) plays an important role in the growth and development of multi-cellular organisms and in regulating homeostasis. Apoptosis, autophagy and necroptosis are the three main forms of PCD, easily distinguished by morphological differences. 13 Apoptosis was mediated through four main pathways, among them the death receptor-mediated extrinsic pathway and the intrinsic mitochondrial pathway are two core pathways and intensively studied especially in tumour research. 13 Caspases,adapter proteins (caspase activators), Bcl-2 family proteins and inhibitor of apoptosis proteins (IAPs) are main components in participating cascades reactions. 14 Autophagy is activated under pathological conditions such as nutrient deprivations and unfavourable stress 15 and is a regulated cell survival process of degradation and recycling of cellular constituents within autophagosome. 16 Nevertheless, autophagic cell death may be evoked after excessive autophagy. 17 Necroptosis, or programmed necrosis, belonging to necrosis as a special form, is defined as RIP1/RIP3-dependent cell death and can be specifically inhibited by an RIP1 inhibitor Nec-1. 18,19 Existing studies show that autophagy and apoptosis cooperate to regulate cell survival and death. However, the relationship between autophagy and apoptosis in breast cancer cell is uncertain.
Reactive oxygen species are generally small and short-lived reactive molecules that are formed by incomplete one-electron reduction of oxygen. 20 As natural byproducts of normal metabolism, ROS play important roles in cell signalling and homoeostasis, including autophagy, by modulating theirs level. Studies indicate that salinomycininduced ROS may result in abortive autophagy and lead to regulated necrosis in glioblastoma. 21 Moreover, ROS are proved to be critical in linking autophagy with apoptosis through mitochondria. 22 In this study, we demonstrated that corilagin simultaneously induced apoptosis and autophagy in MCF-7 cells, inhibition of autophagy enhanced corilagin-mediated cell death. Furthermore, cell death introduced by corilagin was proved of relation to ROS formation.
Meanwhile, we found corilagin could suppress proliferation but could not induce necroptosis in SK-BR3 cells. Finally, cell death induced by corilagin in vitro was indeed observed in vivo.

| Cell viability assay
Cells were seeded at a density of 5 9 10 3 cells per well in a 96-well plate and then treated with corilagin. Added 20 lL MTT (5 mg/mL) for 4 hours and 150 lL DMSO. The absorbance was measured at 490 nm using a plate reader.

| EdU proliferation assay
Cell proliferation was detected using the EdU Cell Proliferation Assay Kit (Ruibo, Guangzhou, China). 23 The number of cells that incorporated EdU was determined by fluorescence microscopy.

| Colony formation assay
Cells were seeded in 6-well plate (500 cells/well) overnight and treated with corilagin. The plates were incubated for another 5 days.
Colonies were fixed with methanol, stained with crystal violet, washed with PBS and then counted.

| Hoechst 33342 Staining
Cells were fixed with paraformaldehyde, washed and stained with 10 mg/L Hoechst33258 (Promega) for 15 minutes at 37°C. The nuclear morphology was observed under fluorescence microscope.

| Lactate dehydrogenase Release assay
Cell death was attested by determining LDH released using a LDH assay kit (Beyotime biotechnology).

| Caspase activity measurement
Caspase-Glos3/8 Assay Kit (Promega) was used for caspase activity measurement following the manufacturer's instructions.

| Transmission electron microscopy imaging
Cells were fixed with 2.5% glutaraldehyde in phosphate-buffered saline, followed by 2% OsO4. After dehydration, thin sections were stained with uranyl acetate and lead citrate for observation under a SEM (JEM-2100hc, Hitachi, Tokyo, Japan).

| Immunofluorescence
Immunofluorescence was performed as in our previous publications. 24 Cells were fixed and subsequently incubated with anti-LC3 antibody (1:300) at 4°C overnight, stained with secondary antibody (Boster) (1:300). The image was observed under confocal microscopy.

| Flow cytometry
We used propidium iodide (PI, 5 mg/mL) for cell viability analysis and DCFH-DA (10 lmol/L) for detecting ROS production. The samples were analysed by flow cytometry (Becton Dickinson FACScan).

| Western blot analysis
Western blotting was performed as previous described. 25 In short, samples were collected by lysing cells in RIPA lysis buffer. Each sample was size fractionated using SDS-polyacrylamide gel electrophoresis (PAGE) and electro transferred onto polyvinylidene difluoride transfer membranes (Dupont, Boston, MA, USA). After bolted with milk, the membranes incubated with primary antibodies overnight at 4°C and then blotted with horseradish peroxidase conjugated secondary antibodies. The immunoblots were visualized using ECL (GE Healthcare, Bucks, UK).

| Xenograft assays in nude mice
For xenograft experiments, MCF-7 cells were implanted by subcutaneous injection into the right foreleg of the female Balb/c nude mice. 26 A known value of 5, 15, 25 mg/kg of corilagin were intraperitoneally injected every 3 days. The mice were sacrificed after four weeks. Tumour volume was calculated as 1/2ab2.

| Statistic analysis
Statistical analyses were performed to compare the differences between the groups using the unpaired Student's t test with Prism 5 software. All data are expressed as mean AE standard deviation (SD) or standard error of mean (SEM), and P value less than.05 was considered statistically significant.   Figure S1F).

| Corilagin suppress growth in MCF-7 cells but not in normal cells
In addition, experiments on corilagin-treated normal cells were performed to investigate whether corilagin has targeting property.
MTT assay revealed that cell viability was not decreased in corilagintreated mammary epithelial cells MCF-10A, hepatic epithelial cells L02 and gastric epithelial cells GES-1( Figure 1C, E and F). Besides, EdU assay showed that corilagin treatment group had no difference with control group in GES-1 cells ( Figure S1A) and L02 cells ( Figure S1B).
These data demonstrate that corilagin can specifically inhibit the growth of breast cancer cells MCF-7 and barely suppress normal cells.

| Corilagin activate extrinsic and intrinsic mitochondrial apoptosis pathways in MCF-7 cells
Research showed that corilagin treatment activated apoptosis in ovarian cancer cells, which significantly increased the number of apoptotic cells. 9 The total protein was extracted and the expression of caspase-3, caspase-8, caspase-9, PARP and the cleaved forms were analysed by Western blot assay. MCF-7 cells were treated with corilagin for 24 h. Caspase activities were determined with colorimetric assays using (D) caspase-8 IETDase assay kits and (E) caspase-3 DEVDase assay kits. (F) MCF-7 cells were treated with different concentrations of corilagin for 24 h. The total protein was extracted, and the expression of Bax, Bcl-2, procaspase-9 and cleaved form was analysed by Western blot assay. (G) MCF-7 cells were pretreated with z-VAD-fmk (20 lmol/L) for 2 h, then corilagin (60 lmol/L) treatment for 24 h. z-VAD-fmk efficiency was analysed by the LDH release assay. (H) MCF-7 cells were pretreated with z-VAD-fmk (20 lmol/L) for 2 h, then corilagin (60 lmol/L) treatment for 24 h and then dyed with PI. Cell death rate was analysed by flow cytometry. Data are expressed as means (n ≥ 3) AE SD over controls, *P < .05, **P < .01, ***P < .001, ****P < .0001. ns: no significant. Abbreviations: LDH, lactate dehydrogenase; PARP, poly(ADP-ribose) polymerase; PI, propidium iodide that z-VAD-fmk had slight effect on corilagin-induced cell damage ( Figure 2G) and reduction in cell viability ( Figure 2H), suggesting the existence of caspase-independent mechanism. Thus, corilagin induced cell death through both caspase-dependent and caspaseindependent pathway in MCF-7 cells.

| Corilagin activate autophagy in MCF-7 cells and autophagy inhibition enhance apoptosis
To ascertain whether corilagin activate autophagy, immunofluorescence was applied to verify the morphologic change of the corilagintreated MCF-7 cells, and autophagosomes were detected (Figure 3A). To further confirm autophagy, we checked detailed cellular images with TEM and observed an extensive intracellular autolysosome-like vesicles and cellular debris ( Figure 3B). Then we checked LC3 conversion by Western blot and found that conversion of LC3 exhibited on account of corilagin treatment ( Figure 3C). Results revealed that corilagin induced the formation of LC3 puncta and autophagosomes. These findings indicated that corilagin induced autophagic cell death.
To better apprehend the underlying mechanisms of corilagininduced autophagy, we detected the Akt/mTOR pathway, which was closely related to autophagy. 29 Corilagin dramatically downregulated the phosphorylation of Akt and mTOR, as well as that of substrates p70S6kinase and 4EBP1 ( Figure 3D and F), indicating that corilagin induce autophagy through Akt/mTOR/p70S6K pathway.
Several studies revealed autophagy may act as a protective mechanism in tumour cells and autophagy inhibition could augment apoptosis. 30 To test whether autophagy protect cancer cell from death, we used chloroquine (CQ), an inhibitor for late autophagy, and corilagin to coprocessing MCF-7 cells. The conversion of LC3 and viability was then investigated by Western blot and MTT assay, respectively. The results showed that corilagin co-treatment with CQ increased the level of LC3B-II, Beclin1 ( Figure 3E) and further suppressed cell viability ( Figure 3G) compared with corilagin-treated alone.
To further confirm that the decrease in viability was evoked by apoptosis after CQ co-exposure, we examined expression levels of apoptosis-related proteins by Western blot. Corilagin co-treatment with CQ further decreased the level of procaspase-3, PARP and Bcl-2, and increased cleaved caspase-8, cleaved PARP, Bax and cleaved caspase-9 comparing with corilagin treatment alone ( Figure 3H).
These data indicated that blocking autophagy enhanced corilagininduced apoptosis.
To check the interplay between autophagy and apoptosis, the effect of apoptosis inhibition on autophagy was investigated. We first examined the effect of z-VAD-fmk by detecting apoptosisrelated proteins in corilagin and z-VAD-fmk pretreated MCF-7 cells ( Figure 3I), indicating that z-VAD-fmk was effective. To detect whether autophagy also had a change, we tested LC3 and Akt/ mTOR pathway-related protein. The results showed that z-VAD-fmk had no effect on autophagy ( Figure 3J). Thus, we concluded that corilagin-induced autophagy which inhibited apoptosis may act as a protective role in MCF-7 cells.

| Autophagy and apoptosis were correlated with the production of ROS in MCF-7 cells
Reactive oxygen species have important roles in intracellular signal transduction and redox homoeostasis. 31 In our model, we observed that corilagin increased intracellular ROS generation ( Figure 4A). To further reveal the role of ROS, we blocked production of ROS using ROS scavenger N-acetylcysteine (NAC) (Figure 4B), and NAC shows no damage to cells ( Figure 4C). We examined cell viability by MTT assay, and the results revealed NAC completely relieved corilagin's effect on cell viability ( Figure 4D).
To further investigate whether ROS play a key role in corilagininduced apoptosis and autophagy. On the one hand, we checked the changes of apoptosis-related protein. Corilagin and NAC co-treatment significantly decreased cleaved caspase-8, caspase-9, PARP and increased Bcl-2 compared with corilagin treated alone (Figure 4E). Additionally, caspase-8 and caspase-3 activities were F I G U R E 3 Corilagin activate autophagy in MCF-7 cells, and autophagy inhibition enhances apoptosis. (A) Cells were treated with corilagin for various concentrations for 24 h, stained with LC3 antibody (Red) and appropriate secondary antibody and observed under a fluorescence microscope, the nuclear was stained with DAPI (Blue). EBSS as a positive control, and LC3 puncta (white arrows) was indicated. (B) Cells were treated as described above. After treatment, cells were harvested and subjected to transmission electron microscopy as described in Materials and Methods. Autophagosomes with typical double-layer membranes containing organelle remnants were pointed out by black arrows.  Figure 4F and G). On the other, to detect the effect on autophagy, we examined the expression changes of autophagyrelated protein. Results displayed that LC3 puncta was inhibited ( Figure 4H), and the level of LC3II along with the proteins such as p-mTOR and p-p70S6K related to Akt/mTOR pathway ( Figure 4I) was restored after NAC treatment. Taken together, ROS had influence on corilagin-induced cell death and maybe participate in regulating autophagy and apoptosis in MCF-7 cells. Besides, the increase in the level of autophagy induced by NAC treatment was consistent with the conclusion aforementioned that autophagy protected MCF-7 cells from death.

| Corilagin cannot induce necroptosis in SK-BR3 cells
Recently, programmed necrosis is found in cancer cell death by anticancer agents. 29,32 Activation of the canonical programmed necrosis includes the formation of a complex containing RIP3 and RIP1 TONG ET AL.  Figure 5C), but not in MCF-7 and SKBR3 cells ( Figure 5D and E). To further confirm the data, we performed Western blot to detect whether the level of RIP1 and RIP3 was increased. Results showed that corilagin up-regulated expression of RIP1, RIP3 in HT-29 cells ( Figure 5F). However, RIP1 and RIP3 were decreased in SK-BR3 cells ( Figure 5G, Figure S2). These results demonstrated that corilagin cannot induce necroptosis in breast cancer SK-BR3 cells.

| Corilagin inhibits the growth of MCF-7 xenograft tumours
It is reported that corilagin had anti-tumour activity on hepatocellular carcinoma 35  As shown in Figure 6A and B, corilagin prevented tumour growth and the mean volume of tumour and bodyweight of mice were obviously lower than the control group. In order to clarify whether corilagin induced apoptosis in vivo, the expression levels of caspase-3, PARP were tested. Comparing with control group, caspase-3 and cleaved PARP increased evidently in corilagin-treated groups (Figure 6C). As shown in Figure 6D, HE staining indicated greater cancer areas in control group. Besides, the PCNA-positive cells which exhibited brown punctate granules in the nucleus showed higher expression in control group. These results revealed that corilagin can suppress tumour growth in vivo. Although the fact corilagin can induce apoptosis in many kinds of cancer has been proved in many research, the specific mechanism is not so clear. In our study, we observed that the activity of caspase-8, caspase-3 and cleaved caspase-8, caspase-9, PARP was upregulated after corilagin treatment, inferring that corilagin induced caspase-dependent apoptosis in MCF-7 cells. Meanwhile, corilagin up-regulated the expression of Bax and down-regulated of Bcl-2, means intrinsic mitochondrial apoptosis pathway was activated either. Furthermore, we checked whether caspase-independent pathway is also involved. As expected, we found that z-VAD-fmk, a pan inhibitor of caspase, did not markedly block cell death upon corilagin treatment in MCF-7 cells, indicating that instead of caspase-dependent pathway, another way of cell death exists, namely caspaseindependent pathway. Taken together, corilagin induced caspasedependent and caspase-independent apoptosis in MCF-7 cells.

| DISCUSSION
In cancer therapy, autophagy seems to be playing a conflicting role, probably promoting or inhibiting cancer cell survival. The role of autophagy in different therapies is cell line and stimulus were warranted to elucidate the mechanism underlying autophagy induced by corilagin. Previous research has been shown that AMPK activation inhibits the rapamycin (mTOR) signalling pathway. 39,40 Depending on the binding partners and sensitivities to rapamycin, mTOR resides in at least two distinct complexes, termed mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2). 41 Among the multiprotein components, Raptor (regulatory associated protein of mTOR, complex (i) and Rictor (Raptor independent companion of mTOR, complex (ii) are characteristic components of mTORC1 and mTORC2, respectively. 42 Most studies targeting the mTOR pathway in cancer therapy mainly focus on rapamycin and its derivatives, which are inhibitors of mTORC1 but not of mTORC2. 43,44 mTORC1 is well-known to inhibit autophagy by phosphorylating p70S6K and 4EBP1, and mTORC2 is related to cytoskeleton construction and cell movement. The main factors regulating cell proliferation, apoptosis and autophagy are mTORC1 45 . In our study, we certified that the activation of autophagy was triggered by corilagin via inhibition of Akt/mTOR pathway. Whereas autophagy induction may also be mediated by other factors or not only depend on Akt/mTOR signalling, which are still need to be further examined.
Some recent data indicated there is a link between the apoptosis and autophagy. 46 Meanwhile, it should be noted that inhibition of autophagy induces apoptosis. 47 Numerous evidences suggested that autophagy was a survival mechanism which provided energy during metabolic stress and could protect cancer cells from apoptosis or necrosis after various anticancer treatments. 48,49 Here, we adopted the autophagy inhibitor CQ and the apoptosis suppressor z-VADfmk. Consistent with these findings, in our research, inhibition of autophagy, particularly at a late stage, facilitated cancer cells to apoptosis in corilagin-treated MCF-7 cells. Nevertheless, no changes were observed after z-VAD-fmk treatment on autophagy. Besides, it is found that chloroquine has antitumour and anti-metastatic activities in a murine model of breast cancer. 50 These data provide a new clue to understand that combined corilagin and CQ treatment may be a promising therapeutic strategy. Nevertheless, the exact mechanisms underlying the pro-apoptotic action of autophagy in corilagin exposed cancer cells remain unclear. Therefore, further study should be performed to explore the mechanism of the relationship between apoptosis and autophagy. Moreover, the safety and efficacy of the combination of corilagin and CQ therapy should be examined in vivo studies.
ROS are critical signalling molecules. ROS-mediated apoptosis and autophagy have been observed in various cancer cell types. 51 In the present study, in virtue of corilagin markedly increased | 3805 intracellular ROS generation in MCF-7 cells, and addition of NAC could significantly decrease ROS generation and also markedly restrain cell death. Hence, we suspected corilagin probably activated apoptosis and autophagy through inducing ROS generation. As expected, NAC pretreatment protected MCF-7 cells from corilagin induced cell apoptosis, including recover the level of caspase. Furthermore, we also found that NAC obviously reduced LC3II expression, restored Akt/mTOR pathway. Therefore, our results suggest that ROS may play a crucial role in corilagin-induced apoptosis and autophagy in MCF-7 cells.
Recently, programmed necrosis, namely necroptosis, is found as a new kind of programmed cell death in cancer cell death which was resulted by anticancer agents. 32 Receptor-interacting protein kinase-3 (RIP3 or RIPK3) is an essential part of the cellular machinery that executes the programmed necrotic cell death. [52][53][54] In our study, we utilized SK-BR3 cells which express high level of RIP3 to confirm whether necroptosis can be induced by corilagin. Meanwhile, we applied HT-29 cells as positive control. As expected, corilagin increased the level of RIP3 as well as RIP1, and employing Necrostatin-1 (Nec-1), a necrotizing apoptosis inhibitor which could competitively inhibit the activity of RIP1 kinase, can recover cell viability in HT-29 cells. However, the results displayed that corilagin cannot augment the level of RIP3 as well as RIP1, and Nec-1 cannot restore cell viability in SKBR3 cells. Therefore, we speculated that corilagin cannot induce necroptosis in breast cancer cells, but further research is needed to verify this conclusion.
Collectively, our research proved orilagin could inhibit breast cancer growth via ROS-dependent apoptosis and autophagy, and we consider that corilagin is a new promising candidate as a potential antitumour drug for treating breast cancer. In consideration of its high dose, further research could be performed on the area of reducing dose and enhancing effectiveness through modifying structure of corilagin.