Erastin Disrupts Mitochondrial Permeability Transition Pore (mPTP) and Induces Apoptotic Death of Colorectal Cancer Cells

We here evaluated the potential anti-colorectal cancer activity by erastin, a voltage-dependent anion channel (VDAC)-binding compound. Our in vitro studies showed that erastin exerted potent cytotoxic effects against multiple human colorectal cancer cell lines, possibly via inducing oxidative stress and caspase-9 dependent cell apoptosis. Further, mitochondrial permeability transition pore (mPTP) opening was observed in erastin-treated cancer cells, which was evidenced by VDAC-1 and cyclophilin-D (Cyp-D) association, mitochondrial depolarization, and cytochrome C release. Caspase inhibitors, the ROS scavenger MnTBAP, and mPTP blockers (sanglifehrin A, cyclosporin A and bongkrekic acid), as well as shRNA-mediated knockdown of VDAC-1, all significantly attenuated erastin-induced cytotoxicity and apoptosis in colorectal cancer cells. On the other hand, over-expression of VDAC-1 augmented erastin-induced ROS production, mPTP opening, and colorectal cancer cell apoptosis. In vivo studies showed that intraperitoneal injection of erastin at well-tolerated doses dramatically inhibited HT-29 xenograft growth in severe combined immunodeficient (SCID) mice. Together, these results demonstrate that erastin is cytotoxic and pro-apoptotic to colorectal cancer cells. Erastin may be further investigated as a novel anti-colorectal cancer agent.


Introduction
The colorectal cancer is the major contributor of cancer-related mortality both in China [1] and around the world [2,3]. It is estimated that over 100,000 new cases of colorectal cancer are diagnosed each year, which cause over 50,000 deaths annually [4]. Chemotherapy has been widely-utilized for treatment of colorectal cancer, however drug resistance and/or off-target toxicity limit the efficiency of current chemo-drugs [5,6,7]. Thus, our group [8,9] and others [10,11] have been focusing on the development of novel and more efficient anti-colorectal cancer agents.
Mitochondrial permeability transition pore (mPTP) is a multi-protein channel complex lying in the mitochondria, whose main function is to maintain the balance of mitochondrial respiratory chain [12]. mPTP is primarily composed of three proteins: including voltage-dependent anion channel (VDAC) in the out mitochondrial membrane (OMM), adenine nucleotide translocator 1 (ANT-1) in the inner mitochondrial membrane (IMM) and matrix locating cyclophilin-D (Cyp-D) [12]. It has been shown that multiple stimuli will induce ANT-1 and Cyp-D association and mPTP opening, thus leading to reactive oxygen species (ROS) production, ATP depletion and pro-apoptotic molecule (i.e. cytochrome c) release [12,13]. Thereafter, caspases (mainly caspase-9) and cell apoptosis will be activated [12,13].
Recent studies have identified a first-in-class VDAC-binding small molecule, namely erastin [14,15]. It has been shown that erastin is selectively cytotoxic to certain cancer cell lines [14,15,16,17]. For example, Yagoda et al., showed that erastin binds to VDAC, resulting in lethal oxidative damage to cancer cells [18]. The potential role of erastin in colorectal cancer cells, and underlying signaling mechanisms have not been studied. In the current study, we showed that erastin was cytotoxic and pro-apoptotic to colorectal cancer cells, possibly via disrupting mPTP.

Cell culture
As described [8,9], colorectal cancer cell lines, including HT-29, DLD-1 and Caco-2, were purchased from the cell bank of Chinese Academy of Science (CAS) Shanghai Biological Institute (Shanghai, China). Cells were maintained in FBS-containing RPMI/DMEM medium. Human NCM460 colon epithelial cell line was provided by Fudan IBS Cell center (Shanghai, China). Cells were cultured in Ham's F12 nutrient medium (Gibco) [19].

Reagents and chemicals
Erastin was purchased from Selleck (Shanghai, China). The mPTP blockers including sanglifehrin A, cyclosporin A and bongkrekic acid were purchased from Sigma (Shanghai, China). The anti-oxidant MnTBAP was also from Sigma. The caspase-3 specific inhibitor z-DEVDfmk and the caspase-9 specific inhibitor z-LEHD-fmk were purchased from Calbiochem (Darmstadt, Germany). All antibodies applied in this study were obtained from Santa Cruz Biotech (Shanghai, China).

Trypan blue staining assay
Trypan blue assay was described in our previous studies [9]. Briefly, after applied erastin treatment, the number of dead (trypan blue positive) cells was counted. The death ratio (%) was calculated by the number of the trypan blue stained cells divided by the total number of the cells.

Colony formation assay
Following the erastin treatment, cells (2 × 10 3 ) were initially suspended in culture medium with 0.25% agar (Sigma). The cell suspension was then planked on top of a pre-solidified 0.25% agar onto a 100-mm culture dish. The medium was replaced every two days. After 10 days of incubation, survival colonies were stained and manually counted.

BrdU incorporation assay
Cells (3×10 3 per well) were seeded onto 96-well plates, after applied treatment, cell proliferation was assessed via the BrdU incorporation ELISA colorimetric assay (Roche, Indianapolis, IN) with the manufacturer's protocol. The ELISA OD value of treatment group was normalized to that of untreated control group.

Cell apoptosis assay through Annexin V staining
After treatment, cell apoptosis was detected by the Annexin V FACS assay as previously reported [9]. The number of Annexin V stained cells was recorded.
2.8. Quantification of apoptosis by enzyme-linked immunosorbent assay (ELISA) As described [8,9], the Cell Apoptosis ELISA Detection Kit Plus (Roche, Palo Alto, CA) was utilized to quantify cell apoptosis according to the manufacturer's protocol.

Caspase activity assay
After treatment, cytosolic proteins were extracted in the buffer described [8,9]. Twenty μg of cytosolic extracts per sample were added to caspase assay buffer with substrates of caspase-3/-8/-9 (Roche, Shanghai, China). The release of 7-amido-4-(trifluoromethyl)coumarin (AFC) was quantified via a Fluoroskan system set to an excitation value of 355 nm [8,9]. The results were expressed as relative fluorescence units/μg of protein.

Western blotting
Briefly, the aliquots of 30 μg of lysed proteins of each sample were separated by 10% SDS polyacrylamide gel electrophoresis and transferred onto PVDF membranes (Millipore, Bedford, MA). After blocking, the membranes were incubated with the primary antibody overnight at 4°C, followed by incubation with secondary antibody for one hour at room temperature. The bolt was visualized by ECL (enhanced chemiluminescence) machine. Each band was quantified via ImageJ software, and the value was normalized to each loading control band.

Reactive oxygen species (ROS) detection
Intracellular ROS was measured by flow cytometry via dichlorofluorescin (DCF) oxidation assay. DCFH-DA enters passively into cells and is cleaved by nonspecific cellular esterases and oxidized in the presence of ROS. After treatment, cells (3×10 5 per sample) were incubated with DCFH-DA (5 μM) for one hour at 37°C. Thereafter, cells were washed with PBS and kept in 1 mL of PBS, ROS fluorescence was analyzed using the above Fluoroskan system.

Detection of mitochondrial potential reduction (ΔΨ m )
As described [20], the ΔC m was measured through JC-10 fluorescence dye. When the mitochondrial potential is decreasing, monomeric JC-10 will form in the cytosol, which exhibits green fluorescence [20]. Briefly, after applied erastin treatment, cells were stained with 5 μg/mL of JC-10 (Invitrogen) for 10 min, and detected immediately on a Fluoroskan system set to an excitation value of 485 nm [20].

Mitochondrial immunoprecipitation (mito-IP)
Cells were trypsinized, and mitochondrial fractions were prepared via a Mitochondria/Cytosol Fractionation Kit (BioVision, Shanghai, China) according to the manufacturer's instructions. Two hundred μg of cell lysates from mitochondrial fractions were pre-cleared with 20 μL of protein A/G PLUS-agarose (Santa Cruz) for 1 hour. The supernatant was then rotated overnight with 0.25 μg of anti-ANT-1 (Santa Cruz Biotech). Next, the lysates were centrifuged for 5 min at 4°C in a micro-centrifuge to remove nonspecific aggregates. The protein A/G PLUS-agarose (35 μL) was then added to the supernatants for 4 hours at 4°C. Pellets were washed six times with PBS, resuspended in lysis buffer, and then assayed by Western blotting [20].

Stable knockdown of VDAC-1 by lentiviral shRNA
The two sets of lentivirus-packaged VDAC-1 short hairpin RNAs (shRNA-1 and shRNA-2, non-overlapping sequences) were designed, synthesized and verified by Genechem (Shanghai, China). Ten μL/mL of lentiviral particles were added to HT-29 cells for 12 hours. Afterwards, lentivirus containing medium was replaced by complete medium, and cells were cultured for another 24 hours. Afterwards, puromycin (5.0 μg/mL, Sigma) was added to select resistant stably colonies for 2-3 weeks. Expression of VDAC-1 was detected by Western blotting. Control cells were treated with scramble non-sense shRNA lentiviral particles (Santa Cruz Biotech).

In vivo antitumor efficacy evaluation
Tumor growth studies were performed in severe combined immunodeficient (SCID) mice xenograft model. All mice were purchased from the Animal Facility of Shanghai Jiao-tong University School of Medicine (Shanghai, China). Briefly, 2×10 6 viable HT-29 cells in 100 μL of growth medium (per mouse) were subcutaneously inoculated, and mice bearing~100 mm 3 tumors were randomly divided into three groups with 10 mice per group. Mice were treated daily with 10 or 30 mg/kg body weight of erastin (intraperitoneal injection, for 4 weeks) or vehicle control (Saline). Tumor volumes were calculated by the modified ellipsoid formula: (π / 6) ×AB 2 , where A is the longest and B is the shortest perpendicular axis of a tumor mass [22,23]. Mice body weights were also recorded every week. Humane endpoints were always utilized to minimize mice suffering. Animals were observed on daily bases. Signs such as significant-reduced locomotion, severe diarrhea, severe piloerection or a sudden weight loss (> 20%) were recorded. If animals reached these endpoints they were euthanized by exsanguination under 2,2,2-tribromoethanol anesthesia (4 mg/10 g body weight, Sigma). All injections were performed under the 2,2,2-tribromoethanol anesthesia method. The animal studies have been approved by the Shanghai Jiao-tong University School of Medicine's Institutional Animal Care and Use Committee (IACUC) and Ethics committee (Contact person: Dr. Jun Wang, 2014126).

Statistical analysis
All data were normalized to control values of each assay and were presented as mean ± standard deviation (SD). Data were analyzed by one-way ANOVA followed by a Scheffe's f-test by using SPSS 16.0 software (SPSS Inc., Chicago, IL). Significance was chosen as p < 0.05.
Interestingly, erastin (1-30 μM) appeared ineffective in inhibiting HT-29 cell proliferation, and the BrdU incorporation was not changed in HT-29 cells after cytotoxic erastin (1-30 μM) treatment ( Fig 1D). Therefore, the cytotoxic effect by erastin is unlikely due to proliferation inhibition. MTT results in Fig 1E showed that erastin (1-30 μM) was also cytotoxic to two other colorectal cancer cell lines: DLD-1 and CaCo2. Yet, same erastin treatment was generally safe to the non-cancerous NCM460 colon epithelial cells (Fig 1E). Once again, BrdU incorporation, the indicator of cell proliferation, was not affected by erastin (10 μM) in the DLD-1 and CaCo2 cells, nor in NCM460 epithelial cells (Fig 1F). Based on these results, we show that erastin exerts cytotoxic, but not cytostatic, activity to cultured colorectal cancer cells.

Erastin induces ROS production and caspase-dependent apoptosis in cultured colorectal cancer cells
Next, we evaluated the potential activity of erastin on cell apoptosis. As described in our previous studies [8,9], various apoptosis assays were performed. The caspase assay results showed that the activity of caspase-3 and caspae-9 was significantly increased in HT-29 cells after cytotoxic erastin (1-30 μM) treatment (Fig 2A), indicating mitochondrial apoptosis pathway activation [24]. On the other hand, the activity of caspase-8, an indicator of extrinsic apoptotic pathway activation [25,26], was unchanged in erastin-treated HT-29 cells (Fig 2A). Cell apoptosis activation by erastin was also confirmed by the Annexin V FACS assay ( Fig 2B) and Histone DNA apoptosis ELISA assay (Fig 2C). Erastin dose-dependently increased Annexin V percentage and Histone DNA ELISA OD in HT-29 cells (Fig 2B and 2C).
Since erastin is a VDAC-binding compound, which may disrupt mitochondrial respiratory chain and cause ROS production [18]. We next tested the oxidative stress level in erastintreated HT-29 cells. Results in Fig 2D demonstrated clearly that erastin increased the level of ROS in HT-29 cells. To study the role of apoptosis and ROS production in erastin-induced cytotoxicity, various pharmacological inhibitors were applied. As shown in Fig 2E and 2F, the caspase-3 specific inhibitor z-DEVD-fmk, the caspase-9 specific inhibitor z-LEHD-fmk, or the superoxide scavenger MnTBAP [27] all alleviated erastin-induced cytotoxicity in HT-29 cells (Fig 2E and 2F). In two other colorectal cancer cell lines, erastin (10 μM) also induced caspase-9 ( Fig 2G) and apoptosis (Fig 2H) activation as well as ROS production (Fig 2I). Such effects by erastin were again not seen in colon epithelial NCM460 cells (Fig 2G-2I). Therefore, erastin induces ROS production and caspase-dependent apoptosis in colorectal cancer cells.

Erastin induces mPTP opening in cultured colorectal cancer cells
Since erastin is a VDAC-binding compound [18], the status of mPTP in erastin-treated colorectal cancer cells was then assessed. First, mitochondrial immunoprecipitation (Mito-IP) assay [28,29] demonstrated that ANT-1 and Cyp-D formed a complex in erastin-treated HT-29 cells (Fig 3A), which is known as the initial step of mPTP opening [12,13]. Second, the level of cytosol cytochrome C was also increased in HT-29 cells after erastin treatment (Fig 3B), which is a known event following mPTP opening [12,13]. Further, the increase of JC-10 green fluorescence intensity indicated loss of mitochondrial potential (ΔCm) (Fig 3C). All these results clearly indicated that mPTP opening following erastin treatment in HT-29 cells. Note that similar results by erastin were also obtained in other two colorectal cancer cell lines (Data not shown).

Erastin administration suppresses HT-29 xenograft growth in SCID mice
The in vivo activity by erastin was also tested. SCID mice HT-29 xenograft model was applied. The weekly tumor growth curve results in Fig 5A demonstrated that erastin intraperitoneal injection dramatically inhibited HT-29 xenograft growth in SCID mice. Erastin's in vivo activity was again concentration dependent. Erastin at 30 mg/kg was clearly more potent than 10 mg/kg in suppressing HT-29 xenografts (Fig 5A). It should be noted that the mice body weight was not significant different between each groups (Fig 5B). Neither did we notice any signs of apparent toxicities in these mice. These results indicated that these animals were well-tolerated to the erastin regimens here. Further, tumor daily growth, calculated as mm 3 /day, was decreased with erastin administration (Fig 5C). At the end of experiments, the weights of erastin-administrated xenografts were also much lower than that of vehicle control mice (Fig 5D). These results showed that erastin administration inhibits HT-29 tumor growth in vivo.

Discussion
In the present study, we demonstrated that erastin exerted potent cytotoxic effect against multiple human colorectal cancer cells possibly via inducing oxidative stress and caspase-9 dependent cell apoptosis. mPTP opening was observed in erastin-treated cancer cells, which was evidenced by VDAC-1 and Cyp-D association, mitochondrial depolarization and cytochrome C release. Caspase inhibitors, the ROS scavenger MnTBAP, and mPTP blockers (sanglifehrin A, cyclosporin A and bongkrekic acid), as well as shRNA-mediated knockdown of VDAC-1, all significantly attenuated erastin-induced cytotoxicity and apoptosis in colorectal cancer cells. On the other hand, over-expression of VDAC1 potentiated erastin's cytotoxicity. Based on these results, we propose that erastin binds to VDAC1 to disrupt normal mitochondrial function, causing mPTP opening, and eventually leading to caspase-9-dependent apoptosis activation in colorectal cancer cells.
It should be noted that erastin was non-cytotoxic to NCM460 colon epithelial cells. We also failed to detect any significant caspase-9 activation nor mPTP dysfunction in erastin-treated NCM460 cells. One possible reason could be that these non-cancerous epithelial cells express very low level of VDAC (-1), therefore cells were not targeted by erastin (see Fig 4). As a matter of fact, when we exogenously over-expressed VDAC-1 in NCM460 cells, these cells, just like cancerous cells, became vulnerable to erastin (see Fig 4). Another possibility is that the cancerous cells are all rapid growing cells, who possibly require high level of mitochondrial respiratory chain to produce ATP. These cells might be more sensitive to VDAC or mPTP disruption. The fact that erastin only targets cancerous cells indicates that it could be an ideal candidate for anti-cancer treatment.
Although several studies have investigated the potential anti-cancer activity by erastin in vitro [14,16,33], the in vivo evidences are still lacking. In the present study, we showed that intraperitoneal injection of erastin at well-tolerated doses (10 or 30 mg/kg, daily) dramatically inhibited Stably HT-29 cells or NCM460 colon epithelial cells expressing empty vector (pSuper-puro, "Vec") or VDAC-1 cDNA ("VDAC-1") were treated with designed erastin for applied time, VDAC-1 expression, cell survival and apoptosis were tested by Western blotting assay (A and F), MTT assay (B and G) and histone DNA ELISA assay (C), respectively; ROS production (D) and JC-10 intensity (E) were also analyzed. For each assay, n = 4. The data presented were mean ± SD. Experiments were repeated three times with similar results obtained. VDAC-1 expression was quantified (F).* p < 0.05 vs. Ctrl group of "Vec" cells (B-E). # p < 0.05 vs. erastin group of "Vec" cells (B-E). "Trans" stands for transfection control (D and E). HT-29 xenograft growth in SCID mice. Importantly, the in vivo erastin regimens didn't affect mice weights, nor did it induce any significant toxicities to the experimental animals. These preclinical results suggest that erastin could be a promising anti-colorectal cancer agent.

Conclusions
In summary, erastin disrupts mPTP and induces apoptotic death of colorectal cancer cells. Erastin may be further investigated as a novel anti-colorectal cancer agent. Erastin administration inhibits HT-29 xenograft growth in SCID mice. HT-29 tumor bearing SCID mice were intraperitoneally administrated with erastin (10/30 mg/kg body weight or "bw", daily, for 4 weeks) or vehicle (Saline) control, tumor volumes (A) and mice body weights (B) were recorded weekly for five weeks; Daily tumor growth was also calculated (C). At the termination of experiments, all the xenografts were isolated and weighted (D). For each assay, n = 10. The data presented were mean ± SD. Experiments were repeated twice with similar results obtained. * p < 0.05 vs. group of "Vehicle". # p < 0.05 vs. group of erastin at 10 mg/kg of body weight. doi:10.1371/journal.pone.0154605.g005