EPOX Inhibits Angiogenesis by Degradation of Mcl-1 through ERK Inactivation

Angiogenesis, namely neovascularization from preexisting vasculature, is necessary for continued growth and metastasis of tumors (1). Angiogenesis-related molecules, such as vascular endothelial growth factor and vascular endothelial growth factor receptor, have been suggested to be ideal targets, and a large number of inhibitors have been developed. Apart from angiogenesis inhibitors, so-called vascular disrupting agents targeted at the established tumor vasculature are currently being investigated in clinical trial. Recently, we revealed that an endostatin-cytosine deaminase fusion protein possesses the tumor vessel–targeting property of endostatin to allow selective tumor-killing effect of cytosine deaminase, producing potent antitumor activity (2).

Endothelial cell apoptosis is intimately involved in the sculpting of blood vessels (3). Natural inhibitors of angiogenesis, such as endostatin, angiostatin, and thrombospondin-1, act in part through selective induction of endothelial cell apoptosis to promote active vessel regression (4). Metronomic chemotherapy, defined as the frequent administration of chemotherapeutic drugs at doses significantly below the maximally tolerated doses, continuously exposes tumor endothelial cells to these drugs. During such therapy, endothelial cell apoptosis precedes tumor cell apoptosis (5). In preclinical studies, metronomic chemotherapy was found effective for managing tumors with cancer cells that had developed resistance to the same chemotherapeutic agents (6). According to current laboratory and clinical findings, inducing apoptosis in tumor endothelial cells contributes to the antitumor efficacy of antiangiogenic and cytotoxic drugs (7).

Apoptosis is characterized by a variety of biochemical and morphologic alterations, including caspase activation, mitochondrial depolarization, chromatin condensation, cell shrinkage, and formation of apoptotic bodies (8). Intracellular redox status is recognized to play an important role in the induction of cell apoptosis. Glutathione (GSH), which accounts for >90% of cellular nonprotein thiols, is critical to maintaining cellular redox balance (9). Therefore, compounds that block the effects of GSH possess the potential to induce cancer cell apoptosis or to reverse drug resistance (10).

Xanthone compounds, secondary metabolites derived from higher plants and microorganisms, have diverse biological properties, including tuberculostatic, antithrombotic, antiallergic, and anticancer activities (11). The emerged “hit” xanthone derivatives, such as 5,6-dimethylxanthenone-4-acetic acid (DMXAA), psorospermin, mangiferin, norathyriol, mangostins, and AH6809, constitute excellent starting materials to leads development in the field of cancer. DMXAA has completed phase II testing (12) and is currently undergoing phase III evaluation in non–small cell lung cancer (ATTRACT-1[Antivascular Targeted Therapy: Researching ASA404 in Cancer Treatment] trial). DMXAA induces apoptosis of tumor vascular endothelial cells and cytokine production, leading to tumor vascular collapse (13). As we previously reported, a series of oxygenated xanthones and [3-(dialkylamino)-2-hydroxypropoxy]xanthones have been prepared and tested for anticancer activity (14). 2,6-Di(2,3- epoxypropoxy)xanthone was chosen for further study, and the mechanism of action is proposed to involve a selective down- regulation of Ras (15). After examining the antiproliferation effects of xanthone derivatives on human umbilical vascular endothelial cells (HUVEC), structure-activity analysis indicated epoxidation of the hydroxyxanthone increased cytotoxicity, but ring opening of the epoxide group with dialkylamine did not enhance the antiangiogeneic activity (Supplementary Fig. S1A). 3,6-Di(2,3-epoxypropoxy)xanthone (EPOX; Supplementary Fig. S1B) was found to be the most effective. The present study was therefore conducted to elucidate the mechanism whereby EPOX exerts its apoptotic effects in endothelial cells.

Xanthone derivatives were provided by Chun-Nan Lin (Kaohsiung Medical University, Kaohsiung, Taiwan). EPOX was synthesized by a method described previously (14). Briefly, 3,6-dihydroxyxanthone was reacted with one equivalent of NaOH in aqueous 2-propanol and an excess of epichlorohydrin to yield the EPOX. The melting point of EPOX is 187°C to 188°C. Anti-Bax 6A7 and all chemicals, unless specifically mentioned, were ordered from Sigma. 4,6-Diamidino-2-phenylindole, anti-HA, were purchased from Roche Diagnostics Corp. Rhodamine 123, anti–complex IV, and MitoTracker were obtained from Molecular Probes. The caspase inhibitor Z-VAD-fmk was purchased from Alexis. Caspase-3 antibody was from Imgenex. Antibodies against caspase-8, caspase-9, cleaved caspase-3, cleaved poly(ADP-ribose) polymerase, phospho and total extracellular signal-regulated kinase (ERK), phospho and total GSK-3β, and phospho and total c-Jun-NH₂-kinase (JNK) were purchased from Cell Signaling Technologies. Cytochrome c and Mcl-1 antibodies came from BD Pharmingen. Antibodies against Bcl-2, Bcl-xL, Bax, Bak, Bim, actin, and α-tubulin were obtained from Santa Cruz Biotechnology. Antibodies against phosphorylated Thr 163 of Mcl-1 were generated by Bethyl Laboratories, Inc. (16). pHA-hMcl-1 was kindly provided by Hsin-Fang Yan-Yen (Academia Sinica, Taipei, Taiwan). Constitutively active mitogen-activated protein (MAP)/ERK kinase (MEK) was a gift from Pier Paolo Pandolfi (Harvard Medical School, Boston, MA).

HUVECs were obtained from human umbilical cord veins and cultured as previously described (17). The endothelial-like EAhy926 cells were maintained in DMEM supplemented with 100 μmol/L sodium hypoxanthine, 0.4 μmol/L aminopterin, and 16 μmol/L thymidine. These cells retain the properties of endothelial cells, including expression of factor VIII–related antigen and tube-forming activity (18). Cells were transfected with Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. Small interfering RNA (SiRNA)-mediated down-regulation of Bim was done by transfecting SMARTpool specific or nonspecific control double-stranded RNA oligonucleotides (Dharmacon). For the EPOX treatment, HUVECs were seeded and then incubated with 3 μmol/L EPOX unless specifically mentioned.

The percentage of surviving cells after treatment was assessed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay (MTT) (19). Cell proliferation was determined with the use of a bromodeoxyuridine labeling and detection kit (Amersham Biosciences) following manufacturer's guidelines. Cell cycle distribution was assessed by a flow cytometry assay as described (19). For the morphologic examination, cells were examined under a phase contrast microscope (Nikon). To identify apoptotic cells, the cells were inspected by 4′,6-diamidino-2-phenylindole staining (DAPI) as previously described (19). For a quantitative assessment of oligonucleosomal DNA fragmentation, apoptosis was detected through a Cell Death ELISAPLUS kit (Roche Molecular Biochemicals) according to the manufacturer's protocol. Measurement of mitochondrial membrane potential was determined as detailed elsewhere (20). The caspase colorimetric assay kits specific for caspase-3, caspase-8, or caspase-9 (Biovision) were used to detect caspase activation.

Tube formation assay was done as previously described (2). In brief, 3 × 10⁴ endothelial cells were seeded on a layer of polymerized Matrigel (BD Biosciences) in a 96-well plate. All assays were done in triplicate. Cells were incubated for 6 h and viewed under a microscope. Tube formation was quantified by measuring the long axis of each tube in three random fields per well with the use of Image-Pro Plus software.

In vivo Matrigel plug assay was done as previously described (17). In brief, C57BL/6J mice were given s.c. injections of 0.5 mL of Matrigel (BD Biosciences) mixed with endothelial growth medium 2 (EGM2; Clonetics) growth medium. Matrigel pellets were harvested on day 7 and were processed for Masson's trichrome staining. Hemoglobin was measured as an indication of blood vessel formation with the use of the Drabkin method (Sigma-Aldrich).

For HCT-116 cancer cell xenograft experiments, male athymic nude mice were inoculated s.c. with 10⁷ cells per mouse. When the tumors reached the average volume of 50 mm³, the mice were randomized into two groups (n = 5 each). EPOX oral treatment at a dose of 50 mg/kg was given daily. Tumors were measured with a caliper every 3 to 4 d, and their volumes were calculated with the formula: 0.5 L × W², where L and W are the long and short diameters of the tumor, respectively. For histologic analysis, the xenograft tissue samples were stained with CD31 antibody to permit determination of the number of blood vessels. All animal experiments were done in accordance with institutional guidelines for animal welfare.

Immunoblotting and immunoprecipitation were done essentially as previously described (21). For quantification of the western analysis, the density of each band was quantified by ImageQuant software and indicated below the blots. Enriched mitochondrial fraction was isolated according to protocols from Yamaguchi et al. (22). For detection of Bax conformational change, cells were lysed with Chaps lysis buffer (10 mmol/L HEPES, pH 7.4, 150 mmol/L NaCl, 1% Chaps) and subjected to immunoprecipitation with anti-Bax 6A7 as described (22).

The content of glutathione was determined with the use of monochlorobimane, a permeable probe that forms blue fluorescent adducts with the reduced form of GSH. To determine the GSH level, after treatment cells in 96-well plates were incubated with 100 μmol/L monochlorobimane for 30 min in the dark. Then, plates were read on a microplate reading fluorometer with excitation and emission wavelengths of 390 and 460 nm, respectively. For imaging the intracellular GSH, cells seeded into a 4-well slide chamber were incubated with monochlorobimane as described above. Cells were fixed for detecting the fluorescence images with a fluorescence microscope (Nikon).

Cells were treated with EPOX for 6 h. After incubation with 25 nmol/L MitoTracker Red CMXRos (Molecular Probes) for 30 min, cells were fixed with ice-cold methanol for 15 min at 4°C and blocked for 30 min with 3% bovine serum albumin. Cells were then incubated with anti-Bax (1:100) overnight at 4°C, followed by incubation at 37°C for 1 h with FITC-conjugated secondary antibody. The slide chamber was mounted onto glass slides for microscopic observation with a Leica TCS SP2 confocal spectral microscope.

Each experiment was done at least thrice, and representative data are shown. Data in the bar graph are given as the means ± SE. The means were checked for statistical difference with the use of a t test, and P values <0.05 were considered significant (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

To test the hypothesis that EPOX may inhibit angiogenesis, initially we determined the effect of EPOX with the Matrigel plug assay. The Matrigel plugs in the control groups were visually bloodier than the EPOX-treated ones, suggesting a higher level of angiogenesis (Fig. 1A, top). Histologic analysis of the Matrigel pellets identified more erythrocyte-containing vessels within the control than in the EPOX-treated (Fig. 1A, middle). Quantification of angiogenesis by hemoglobin content further showed that EPOX inhibited angiogenic response in a dose-dependent manner (Fig. 1A, bottom), suggesting that EPOX attenuated angiogenesis in vivo. We proceeded to examine whether EPOX inhibits tube formation, which is a well-accepted technique to measure in vitro angiogenesis. Treatment with EPOX significantly reduced this process (Fig. 1B).

In an attempt to discover the mechanism of EPOX to inhibit angiogenesis, we first did MTT and bromodeoxyuridine incorporation assays to measure the effect of EPOX on the growth of HUVECs. EPOX was found to decrease cell viability (Fig. 1C) and DNA synthesis (Supplementary Fig. S1C) in a concentration-dependent manner. The EAhy926 cell line has been suggested to be an alternative model for HUVECs to study the control of pathologic angiogenesis by endogenous factors and pharmacologic compounds (23). Consistently, EPOX inhibited EAhy926 cell growth (Supplementary Fig. S2A) and tube formation (Supplementary Fig. S2B).

To examine whether EPOX-induced growth inhibition was associated with cell cycle regulation, the cell cycle distributions of EPOX-treated cells were analyzed by flow cytometry. As illustrated in Fig. 1D, EPOX treatment increased the percentage of cells in the sub-G₁ phase. EPOX-treated HUVECs exhibited both rounding and blebbing during morphologic examination (Supplementary Fig. S2C, top). To verify the occurrence of apoptosis, cells were inspected after DAPI staining. In the absence of EPOX, cells exhibited nuclei with homogeneous chromatin distribution. However, EPOX treatment induced nuclear fragmentation and chromatin condensation (Supplementary Fig. S2C, bottom). DNA fragmentation was also increased in EPOX-treated cells in a concentration-dependent manner (Supplementary Fig. S2D). Taken together, the results suggest that EPOX treatment induces apoptosis of endothelial cells.

The activation of selected caspases serves to identify the involvement of specific apoptotic signaling pathways. To further characterize the mechanism through which EPOX triggers apoptosis, the activation states of caspase-3, caspase-8, and caspase-9 were therefore examined. Notably, EPOX treatment resulted in the cleavage and activation of caspase-9 and caspase-3 in HUVECs (Fig. 2A).

Caspase-9 has been proposed as the predominant initiator caspase in the mitochondria-dependent intrinsic apoptosis pathway. In response to apoptotic stimuli, dissipation of the mitochondrial membrane potential results in the release of cytochrome c from the mitochondrial intermembrane space. Upon release to the cytosol, cytochrome c binds Apaf-1, and then recruits and activates procaspase-9. The mitochondrial membrane potential and release of cytochrome c from the mitochondria to the cytosol were therefore measured in untreated and EPOX-treated HUVECs. Treatment with EPOX decreased the mitochondrial membrane potential in a time-dependent manner (Fig. 2B). As shown in Fig. 2C, the efflux of cytochrome c from mitochondria was detected in EPOX-treated cells simultaneously with the accumulation of cytochrome c in the cytosol. Taken together, these findings are consistent with the hypothesis that EPOX induces apoptosis in HUVECs through a mitochondria-mediated pathway.

Bcl-2 proteins, a family that includes both proapoptotic and antiapoptotic members, are critical regulators of the mitochondrial pathway of apoptosis. Proapoptotic members, such as Bax and Bak, promote the release of cytochrome c from the mitochondria, whereas antiapoptotic members, such as Bcl-2, Bcl-xL, and Mcl-1, prevent release. The balance of proapoptotic and antiapoptotic Bcl-2 proteins therefore decides the fate of the cell. Because the above findings reveal that EPOX activates the mitochondrial pathway of cell apoptosis, it was of interest to identify the Bcl-2 family members involved in the apoptosis triggered by EPOX. EPOX treatment induced the rapid down-regulation of Mcl-1, whereas no significant changes in expression of Bcl-2, Bcl-xL, Bax, Bak, or Bim were observed (Fig. 3A). If loss of Mcl-1 after EPOX treatment was responsible for the subsequent induction of apoptosis, then maintenance of Mcl-1 should prevent apoptosis. As shown in Fig. 3B, transient transfection of Mcl-1 partially overcomes the cleavages of caspase-3 and poly(ADP-ribose) polymerase, and the loss of viability in response to EPOX treatment. These observations support the hypothesis that down-regulation of Mcl-1 contributes to EPOX-induced apoptosis in HUVECs.

To determine how EPOX reduces Mcl-1 expression, we first measured Mcl-1 mRNA content by reverse transcription-PCR and found that Mcl-1 mRNA values were unaffected by EPOX (Supplementary Fig. S3A). We then investigated the degradation of Mcl-1. As shown in Fig. 3C, Mcl-1 degradation in EPOX-treated HUVECs was almost fully blocked by exposure to the proteasomal inhibitor MG132. An increase in ubiquitination of Mcl-1 was also observed for EPOX-treated compared with nontreated preparations. Consistently, in the presence of cycloheximide, which inhibits de novo protein synthesis, the rate of degradation of Mcl-1 was significantly greater in EPOX-treated compared with vehicle-treated cells (Supplementary Fig. S3B). These findings support the proposal that EPOX decreases Mcl-1 level through activation of the ubiquitin-dependent and proteasome-dependent pathways.

Mcl-1 is also described as a substrate for caspase-mediated degradation. In previous data, cleavage of caspase-3 was observed in EPOX-treated HUVECs. To ascertain whether caspase activation contributes to Mcl-1 down-regulation in these cells, experiments were also conducted in the presence of ZVAD-fmk, an established caspase inhibitor. In EPOX-treated HUVECs exposed to ZVAD-fmk, caspase-3 and poly(ADP-ribose) polymerase processing were significantly reduced, and viability was increased compared with preparations not treated with ZVAD-fmk (Fig. 3D). However, exposure to ZVAD-fmk did not notably influence Mcl-1 down-regulation in these cells, suggesting that this down-regulation occurred upstream of caspase activation.

Mcl-1 is known to sequester Bim in an inactive state. Experiments were therefore done to ascertain whether Mcl-1/Bim complexes are affected in EPOX-treated HUVECs. Endogenous Mcl-1/Bim complexes were detectable in extracts of nontreated HUVECs. After 6 hours of EPOX treatment, however, an increase in Mcl-1–free Bim was observed (Fig. 4A). To examine the effects of Bim on the apoptotic response to EPOX, Bim was knocked down by RNA interference (RNAi). Knockdown of Bim conferred partial resistance to EPOX-induced caspase-3 and poly(ADP-ribose) polymerase processing, and the subsequent apoptosis (Fig. 4B).

Once liberated from Mcl-1, Bim is able to activate Bax directly (24). A change in Bax conformation to expose its N-terminal and C-terminal regions is intimately associated with the mitochondrial targeting and proapoptotic activity of the protein (25). To ascertain whether Bax is activated in EPOX-treated HUVECs, the protein was immunoprecipitated from cell extracts with the use of an antibody (anti-Bax 6A7) that recognizes only the conformationally altered protein. As shown in Fig. 4B, Bax underwent a conformational change in EPOX-treated cells. However, knockdown of Bim attenuated Bax activation in EPOX-treated cells, implying that Bim is involved in the activation. Subcellular fractionation (Fig. 4C) and confocal microscopy (Fig. 4D) confirmed Bax translocation to mitochondria during EPOX treatment. In untreated cells, the cytoplasm displayed diffuse immunostaining for Bax, whereas a punctate distribution, colocalizing with MitoTracker staining, was observed after EPOX treatment.

To further characterize the upstream signals involved in the induction of apoptosis in EPOX-treated HUVECs, specific pharmacologic agents were tested for their capacity to ameliorate the apoptosis. Modulators of the “oxidant,” “calcium,” and “arachidonic acid” pathways associated with the induction of endothelial cell apoptosis in response to stimuli, including lipopolysaccharide, hypoxia, hydrogen peroxide, tumor necrosis factor α, IFN γ, activated neutrophils, or oxidized low-density lipoprotein (26), were tested. As shown in Fig. 5A, only glutathione ethyl ester, a cell-permeable analogue of GSH, significantly reduced EPOX-induced cell death.

Supplementation with glutathione ethyl ester was done for 6 hours followed by replacement with nonsupplemented medium to avoid nonspecific reactions between glutathione ethyl ester and EPOX. Supplementation was found to reverse the effect of EPOX on cell growth (Fig. 5B) and tube formation (Supplementary Fig. S3C). With regard to the importance of glutathione for the susceptibility of endothelial cells to EPOX, it was of interest to observe the changes in intracellular GSH levels after EPOX treatment. Cells were incubated with monochlorobimane, a well-established GSH-specific probe, and then analyzed for GSH content with a fluorescence microplate reader. EPOX treatment induced the depletion of GSH in a time-dependent manner (Fig. 5C). When fluorescence microscopy was used, consistently EPOX treatment was found to cause a striking decrease in monochlorobimane fluorescence (Supplementary Fig. S3D). These findings implicate GSH depletion in the induction of apoptosis by EPOX.

Reduced GSH levels are frequently associated with activation of apoptosis through mitochondrial pathways (27). Studies were therefore done to explore the role of GSH in the mitochondrial alterations observed in response to EPOX treatment. As shown in Supplementary Fig. S4A, glutathione ethyl ester supplementation significantly restored the collapse of mitochondria membrane potential in response to EPOX treatment. Additionally, pretreatment with glutathione ethyl ester reduced the down-regulation of Mcl-1, the conformational change in Bax, and the cleavages of caspase-3 and poly(ADP-ribose) polymerase after EPOX treatment (Fig. 5D). Depletion of GSH therefore seems to serve as an upstream event for induction of Mcl-1 degradation by EPOX.

The stability of Mcl-1 changes in response to phosphorylation by ERK, JNK, and GSK-3β (16, 28, 29).Among these three kinases, ERK was found to be the most significantly inactivated by EPOX treatment (Fig. 6A). When a constitutively active MEK that activated ERK was transiently transfected into HUVECs, the ERK inactivation by EPOX was partially restored. Interestingly, Mcl-1 down-regulation, downstream apoptotic pathway activation (Fig. 6B), and the loss of viability (Supplementary Fig. S4B) in response to EPOX were also attenuated, consistent with a role for ERK inactivation in EPOX-induced Mcl-1 degradation.

Phosphorylation of Mcl-1 at T₁₆₃, the conserved ERK site located within the proline, glutamic acid, serine, and threonine (PEST) region (16), was abolished by treatments with the MEK inhibitors PD98059 and U0126 (Supplementary Fig. S4C). Domina et al. (28) showed that phosphorylation of Mcl-1 at T₁₆₃ by ERK serves to prolong the half-life of Mcl-1. Half-life prolongation is attributable to the binding of Pin1 at the phosphorylated T₁₆₃ site, which serves to protect Mcl-1 from degradation by interfering with its ubiquitination (29, 30). Accordingly, treatments with MEK inhibitors promoted Mcl-1 ubiquitination (Supplementary Fig. S4C).

Experiments were done to examine T₁₆₃ phosphorylation of Mcl-1 after EPOX treatment and to ascertain whether ERK activation attenuates the effect of EPOX. As shown in Fig. 6C, overexpression of constitutively active MEK, which served to activate ERK, reduced the T₁₆₃ dephosphorylation and ubiquitination of Mcl-1 in response to EPOX treatment. Interestingly, glutathione ethyl ester supplementation was observed to reverse the effects of EPOX on ERK inactivation, on Mcl-1 dephosphorylation, and on Mcl-1 ubiquitination and degradation (Figs. 5D, and 6A and D). These observations, which are consistent with those of others (28, 31), highlight the involvement of the ERK in the regulation of Mcl-1 stability. These findings also reveal that ERK is involved in the down-regulation of Mcl-1 in EPOX-treated HUVECs.

The increasing use of antiangiogenic drugs for the treatment of cancer is the consequence of decades of extensive basic and clinical research. However, the clinical benefits of antiangiogenic therapy are relatively modest, with survival measured in months (32). Better understanding of the molecular mechanisms governing angiogenesis and of the molecular responses to antiangiogenic therapies are required for therapeutic improvements. The present study reveals that the xanthone derivative EPOX induces apoptosis in endothelial cells through pathways initiated by GSH depletion and Mcl-1 degradation. In this regard, EPOX was also examined for apoptotic effects with the use of a panel of normal cells and cancer cell lines (Supplementary Table S1). This agent produced prominent apoptotic effects in several types of cancer cells, and modest apoptotic effects in normal prostatic cells and fibroblasts. In a colon cancer xenograft model, treatment with EPOX caused a significant inhibition of tumor growth (Supplementary Fig. S5A) without overt toxicity (Supplementary Fig. S5B). We also observed that the EPOX-treated tumor had less CD31+ vessel structures than the control group (Supplementary Fig. S5C), supporting the potential of EPOX as an angiognenesis inhibitor for cancer therapy.

GSH effectively scavenges free radicals and other reactive oxygen species, both directly and indirectly through enzymatic reactions. Depletion of intracellular GSH content is frequently accompanied by a concomitant increase in reactive oxygen species. However, GSH depletion could lower the reducing capacity of the cells, thereby promoting oxidative stress independently of reactive oxygen species (33). In the present study, supplementation with glutathione ethyl ester, but not with other antioxidants, ameliorated EPOX-induced apoptosis. Furthermore, EPOX treatment did not promote accumulation of reactive oxygen species as monitored through use of 2′,7′-dichlorodihydrofluorecein diacetate or hydroxyethidine (data not shown). Reactive oxygen species have been shown to activate ERK in a variety of cell types (34), although the mechanisms responsible for the activation have not been identified clearly. In the current study, EPOX treatment was observed to suppress activation of ERK in endothelial cells. Taken together, these findings support the hypothesis that EPOX induces endothelial cell apoptosis independently of reactive oxygen species. Further studies are clearly required to determine the mechanism whereby GSH depletion leads to ERK inactivation after EPOX treatment.

Because most tumors are recognized to have defects in the p53 pathway or to overexpress a Bcl-2 homologue, the therapeutic potential of anticancer drugs that, like the BH3 domain, bind one or more Bcl-2 homologues and trigger apoptosis are of significant current interest. Such“BH3 mimetics” should prove more effective than antitumor agents acting far upstream (35). A structure-based approach recently resulted in the development of ABT-737, an agent that binds strongly to Bcl-2, Bcl-xL, and Bcl-w, but not to Mcl-1 (36). Because inactivation of Mcl-1 is critical for cell death to proceed (37), knockdown of Mcl-1 with the use of siRNA was found to sensitize cells to the effects of ABT-737 (38). Small molecules that target Mcl-1, such as obatoclax (39), should serve to complement the antitumor actions of ABT-737. However, it should be noted that the rapid and marked down-regulation of Mcl-1 through proteasome-dependent degradation, as observed in the present study, was found to contribute to EPOX-induced apoptosis. EPOX in combination with ABT-737 should therefore represent an effective therapy for induction of tumor cell apoptosis (40).

Current evidence supports the importance of Raf/MEK/ERK signaling in stimulation of angiogenesis. Raf-1 knockout mice die from vascular defects in the yolk sac in association with an increased propensity for apoptosis (40). In response to intratumor hypoxia, the proangiogenic factors vascular endothelial growth factor, basic fibroblast growth factor, and platelet-derived growth factor are up-regulated and secreted. Interestingly, all of these growth factors exert their signaling effects through the Raf/MEK/ERK pathway. Furthermore, sorafenib, a Raf kinase inhibitor, strongly inhibits angiogenesis in tumor xenograft models (41). Although sorafenib was recently approved by the Food and Drug Administration for treating advanced renal cell carcinoma (42), the molecular mechanism through which this agent promotes apoptosis is not completely characterized. However, sorafenib was recently reported to enhance proteasome-mediated Mcl-1 degradation (31) and to increase apoptosis induced by tumor necrosis factor–related apoptosis-inducing ligand through down-regulation of Mcl-1 (43). The findings of the present study highlight the involvement of the ERK in the regulation of Mcl-1 stability. Recent findings in our laboratory revealed that ERK could phosphorylate Mcl-1 at Thr 92 and Thr 163, which is required for the association of Mcl-1 and Pin1 (30). A novel mechanism was identified linking the ERK-Pin1 pathway with Mcl-1–mediated chemoresistance. Consistent with these findings, the inhibition of ERK activation in response to EPOX treatment was also found to promote the dissociation of Mcl-1 from Pin1.⁵

Supplementary Fig. S6 presents a model, based on the findings of the present report, for EPOX-induced apoptosis in endothelial cells. In response to EPOX treatment, the depletion of GSH signals ERK inactivation, which results in the dephosphorylation and down-regulation of Mcl-1 through the proteasome-mediated degradation. The liberation of Bim from Mcl-1 serves to induce the conformational change in Bax required for its activation. Upon translocation to the mitochondria, the activated Bax triggers apoptosis through the release of cytochrome c, and activations of caspase-9 and caspase-3. It is therefore proposed that therapies designed to manipulate GSH content, the ERK/Mcl-1 pathway, and crosstalk among Bcl-2 members could prove effective for suppression of angiogenesis.

No potential conflicts of interest were disclosed.

We thank Hsin-Fang Yang-Yen and Pier Paolo Pandolfi for providing reagents; Kuo-Shyan Lu for the EAhy926 cells; Chieh-Yu Peng, Ying-Cheng Chen, Yao-Ting Huang, Dong-Ming Huang, Dung-Fang Lee, Chun-Te Chen, Jung-Mao Hsu, and Jer-Yen Yang for technical support; and Chin-Yu Lai, Anita Shin-Yuan Chang, Margaret A. Brostrom, Jeng C. Cheng, and Stephanie A. Miller for editing the manuscript.