Celastrol, a Triterpene, Enhances TRAIL-induced Apoptosis through the Down-regulation of Cell Survival Proteins and Up-regulation of Death Receptors*

Whether celastrol, a triterpene from traditional Chinese medicine, can modulate the anticancer effects of TRAIL, the cytokine that is currently in clinical trial, was investigated. As indicated by assays that measure plasma membrane integrity, phosphatidylserine exposure, mitochondrial activity, and activation of caspase-8, caspase-9, and caspase-3, celastrol potentiated the TRAIL-induced apoptosis in human breast cancer cells, and converted TRAIL-resistant cells to TRAIL-sensitive cells. When examined for its mechanism, we found that the triterpene down-regulated the expression of cell survival proteins including cFLIP, IAP-1, Bcl-2, Bcl-xL, survivin, and XIAP and up-regulated Bax expression. In addition, we found that celastrol induced the cell surface expression of both the TRAIL receptors DR4 and DR5. This increase in receptors was noted in a wide variety of cancer cells including breast, lung, colorectal, prostate, esophageal, and pancreatic cancer cells, and myeloid and leukemia cells. Gene silencing of the death receptor abolished the effect of celastrol on TRAIL-induced apoptosis. Induction of the death receptor by the triterpenoid was found to be p53-independent but required the induction of CAAT/enhancer-binding protein homologous protein (CHOP), inasmuch as gene silencing of CHOP abolished the induction of DR5 expression by celastrol and associated enhancement of TRAIL-induced apoptosis. We found that celastrol also induced reactive oxygen species (ROS) generation, and ROS sequestration inhibited celastrol-induced expression of CHOP and DR5, and consequent sensitization to TRAIL. Overall, our results demonstrate that celastrol can potentiate the apoptotic effects of TRAIL through down-regulation of cell survival proteins and up-regulation of death receptors via the ROS-mediated up-regulation of CHOP pathway.

toxic and not effective; the tumor gradually develops resistance. In recent years, the development of more selective, tumor biology-driven agents and targets for tumor-derived molecules has become of great interest in cancer research. Furthermore, it has been suggested that a combination of conventional chemotherapy with the biological agents may result in more effective therapy.
Tumor necrosis factor (TNF)-related apoptosis-inducing ligand (TRAIL) 2 is a member of the TNF superfamily (1, 2) that has drawn considerable attention as an attractive molecule for cancer therapy because it preferentially induces apoptosis of transformed or malignant cells. To date, five receptors have been identified for TRAIL, of which only two receptors have cytoplasmic death domains, DR4 (TRAIL-R1) and DR5 (TRAIL-R2), that engage the apoptotic machinery upon TRAIL binding (3). Antibodies that mimic TRAIL ability to trigger death receptor-mediated apoptosis are now in phase I or II clinical trials for treatment of various types of cancer (4). However, resistance of cancer cells to TRAIL remains a major problem. Why or how cancer cells develop resistance to TRAIL is not fully understood, but numerous mechanisms have been described including alteration of death receptors (5); decoy receptors that bind TRAIL but do not signal (6,7) such as osteoprotegin, a TRAILbinding nonsignaling protein that lacks transmembrane domain (8); and overexpression of cFLIP, an inhibitor of caspase-8 (9), X-linked inhibitor of apoptosis (XIAP) (10), Bcl-xL (11), survivin (12), Mcl-1 (13), and nuclear factor (NF)-B activation (14,15). Thus agents that can modulate some of these mechanisms of resistance to TRAIL have a potential in improving the cytokine therapy.
One such agent, celastrol, is a triterpenoid isolated from the traditional Chinese medicine "Thunder of God Vine" (Tripterygium wilfordii Hook F.) that has been used in the treatment of autoimmune diseases, asthma, chronic inflammation and neurodegenerative diseases (16 -19). It was also found to have anticancer activity against a variety of tumor cells, including prostate cancer (20), leukemia (21), and melanoma cells (22). The antitumor effects of celastrol are not limited to in vitro systems, as this triterpenoid was found to suppress the growth and metastasis of melanoma in syngeneic and xenograft mouse models (22), of human prostate tumor xenografts in mice (20), and of human glioma xenografts in nude mice (23), and it has exhibited antiangiogenic effects in zebra fish (24). Studies to define its therapeutic mechanism showed that it suppresses the nuclear factor (NF)-B signaling pathway (21,25) and VEGFR expression (23); inhibits heat shock protein (HSP) 90 (26,27), ERK (28), and proteasome (20,29), and activates caspase-8 (22).
These factors led us to investigate whether celastrol can modulate TRAIL-induced apoptosis and if so, through what mechanism. We found that celastrol can indeed enhance TRAIL-induced apoptosis through the down-regulation of various cell survival proteins and via up-regulation of TRAIL receptors. The up-regulation of death receptors by celestrol was mediated through production of reactive oxygen species (ROS) and expression of CHOP (C/EBP homologous protein).
Cell Lines-Human breast adenocarcinoma MDA-MB-231, MCF7, and T47D, colon adenocarcinoma HCT116, lung adenocarcinoma H1299, embryonic kidney carcinoma A293, prostate adenocarcinoma PC3, esophageal adenocarcinoma TT, pancreatic adenocarcinoma AsPC1, chronic myelogenous leukemia K-562, and acute T cell leukemia Jurkat cells were obtained from American Type Culture Collection. Human myeloid KBM-5 cells were kindly supplied by Dr. Nicholas Donato (University of Michigan Comprehensive Cancer Center, Ann Arbor, MI). HCT116 variants with deletions in p53, p21, and bax were kindly supplied by Dr. Bert Vogelstein (John Hopkins University, Baltimore, MD). The human colon cancer cell lines HCT116 and its derivatives were cultured in McCoy's 5A medium supplemented with 10% fetal calf serum and penicillin/streptomycin (Invitrogen). KBM-5 cells were cultured in Iscove's modified Dulbecco's medium with 15% fetal bovine serum. A293, MDA-MB-231 and TT were cultured in Dulbecco's modified Eagle's medium, and other cells lines were cultured in RPMI 1640 with 10% fetal bovine serum, 100 units/ml penicillin, and 100 mg/ml streptomycin.
Live/Dead Assay-To measure apoptosis, we used the Live/ Dead Assay (Invitrogen), which assesses intracellular esterase activity and plasma membrane integrity. This assay was performed as described previously (30).
Propidium Iodide (PI) Staining for DNA Fragmentation-Cells were pretreated with celastrol (2 mol/liters) for 6 h and then exposed to TRAIL (10 ng/ml) for 24 h. PI staining for DNA content analysis was performed as described elsewhere (30). A total of 10,000 events were analyzed by flow cytometry using an excitation wavelength set at 488 nm and emission set at 610 nm.
RNA Analysis and Reverse Transcription-Polymerase Chain Reaction (RT-PCR)-DR5 mRNA was detected using RT-PCR as follows. Total RNA was isolated from cells using TRIzol reagent (Invitrogen) as instructed by the manufacturer. One microgram of total RNA was converted to cDNA using Superscript reverse transcriptase and then amplified by platinum Taq polymerase using the Superscript One Step RT-PCR kit (Invitrogen). The total RNAs were then amplified by PCR using the following primers: DR5 sense 5Ј-AAGACCCTTGTGCTCGTTGTC-3Ј, DR5 antisense 5Ј-GACACATTCGATGTCACTCCA-3Ј, DR4 sense 5Ј-CTGAGCAACGCAGACTCGCTGTCCAC-3Ј, DR4 antisense 5Ј-TCCAAGGACACGGCAGAGCCTGTGCCAT-3Ј, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) sense, 5Ј-GTCT-TCACCACCATGGAG-3Ј, and GAPDH antisense 5Ј-CCACCC-TGTTGCTGTAGC-3Ј. The reaction sequence consisted of 50°C for 30 min, 94°C for 2 min, and 94°C for 35 cycles of 15 s each; 50°C for 30 s; and 72°C for 45 s with an extension at 72°C for 10 min. PCR products were run on 2% agarose gel and then stained with ethidium bromide. Stained bands were visualized under UV light and photographed.
Transfection with siRNA-High purity control (scrambled RNA), DR5, DR4, and CHOP small interfering RNA (siRNA) oligos were described previously (31,32) and synthesized by Qiagen. Briefly, MDA-MB-231 cells were plated in each well of 6-well plates and allowed to adhere for 24 h. On the day of transfection, 12 l of Hiperfect transfection reagent (Qiagen) was added to 50 nmol/liter siRNA in a final volume of 100 l of culture medium. After 24 h of transfection, cells were treated with celastrol for 6 h and then exposed to TRAIL for 24 h. Whole cell extracts were prepared for relevant protein analysis by Western blotting.
JNK Assay-To determine the effect of celastrol on the kinase activity of JNK, JNK complex from whole cell extracts was precipitated with antibody against JNK1, followed by treatment with protein A/G-agarose beads (Pierce). After 2 h of incubation, the beads were washed with lysis buffer and then assayed in kinase assay mixture containing 50 mmol/liter HEPES (pH 7.4), 20 mmol/liter MgCl 2 , 2 mmol/liter dithiothreitol, 20 Ci of [ 32 P]ATP, 10 mol/liter unlabeled ATP, and 2 g of substrate glutathione S-transferase (GST)-c-Jun (amino acids 1-79). The immunocomplex was incubated at 30°C for 30 min and then boiled with SDS sample buffer for 5 min. Finally, the protein was resolved on 10% SDS-PAGE, the gel was dried, and the radioactive bands were visualized using the PhosphorImager. To determine the total amount of JNK1 in each sample, whole cell extracts were subjected to Western blot analysis using anti-JNK1 antibody.
Western Blot Analysis-To determine the levels of protein expression, we prepared whole cell extracts as described previously (30). The proteins were separated by SDS-polyacrylamide gel electrophoresis. After electrophoresis, the proteins were electrotransferred onto nitrocellulose membranes, blotted with each antibody, and detected by an ECL reagent (GE Healthcare).

Analysis of DR4 and DR5 Surface
Expression-MDA-MB-231 cells (3 ϫ 10 5 ) were treated with celastrol and washed with 1ϫ PBS supplemented with 0.5% bovine serum albumin (BSA) after detachment with EDTA. Cells were then stained with phycoerythrin (PE)conjugated mouse monoclonal antihuman DR5 or DR4 (clone 71908 and 69036, respectively, R&D Systems) for 45 min at 4°C according to manufacturer's instructions before washing and resuspension in a fluorescence-activated cell sorting buffer (1ϫ PBS ϩ 0.5% BSA) for flow cytometric analysis using an excitation wavelength of 488. PEconjugated mouse IgG2B was used as an isotype control.

RESULTS
The goal of this study was to determine whether and how celastrol modulates the TRAIL apoptotic effects. Breast cancer MDA-MB-231 cells were used for most studies, but other cell types were also used to determine the specificity of this effect.
Celastrol Enhances TRAIL-induced Apoptosis-We examined first the sensitivity of various breast cancer cell lines to TRAIL. Three breast cancer cell lines, MDA-MB-231, MCF7, and T47D, were treated for 24 h with TRAIL and then assessed for cell viability using the MTT method. The MDA-MB-231 cells were found to be highly sensitive to TRAIL whereas T47D was highly resistant and MCF-7 was inbetween ( Fig. 1A, left panel). Whether celastrol could sensitize the resistant (T47D) cells to TRAIL, was examined. Celastrol alone had minimal effect on cell viability, but it significantly potentiated the cytotoxic effects of TRAIL in a dose-dependent manner (Fig. 1A, right panel). Next we determined whether celastrol could potentiate TRAIL-induced apoptosis, as indicated by plasma membrane integrity, in human breast cancer cells. To reveal celastrol enhancement of response, we used suboptimal dose of TRAIL for subsequent experiments. As shown in Fig. 1B, celastrol enhanced TRAIL-induced apoptosis from 9 to 51% in MDA-MB-231 cells and from 8 to 38% in T47D cells. We also examined apoptosis by determining the size of the sub-G1 cell pool. We found that TRAIL-induced apoptosis was increased from 10.7 to 46.8% in MDA-MB-231 cells and from 12.8 to 47.1% in T47D (Fig. 1C). Thus our results indicate that celastrol converted the TRAIL-resistant T47D cells to TRAIL-sensitive cells.
Because TRAIL mediates apoptosis through the activation of caspase-8, caspase-9, and caspase-3, we next examined the effect of celastrol on activation of these caspases and poly (ADP-ribose) polymerase (PARP) cleavage as induced by TRAIL in MDA-MB-231 and T47D breast cancer cells. As shown in Fig. 1D, although celastrol and TRAIL alone had a minimal effect on the activation of these caspases and on cleavage of PARP, the combination of the two was highly effective in activation of all caspases and consequent PARP cleavage. These results indicate that celastrol can enhance TRAIL-induced apoptosis. In case of T47D, the resistance to TRAIL was reversed by celastrol.
Celastrol Down-modulates the Expression of Antiapoptotic Proteins Linked to TRAIL Resistance-How celastrol enhances TRAIL-induced apoptosis was investigated next. First, cells were treated with different concentrations of celastrol for 24 h and then examined for expression of antiapoptotic proteins using relevant antibodies. The results indicated that celastrol down-regulated the expression of cFLIP (both long and short form), Bcl-2, Bcl-xL, and cIAP-1, XIAP, and survivin ( Fig.  2A). Whereas the down-regulation of survivin and XIAP was less pronounced, the down-regulation of cFLIP L was quite dramatic and was dose-dependent. Whether this downmodulation of antiapoptotic proteins was time-dependent, was also examined. Celastrol down-regulated cFLIP, Bcl-2, Bcl-xL, and cIAP-1, XIAP, and survivin in a time-dependent manner (Fig. 2B).
Celastrol Up-modulates the Expression of Bax-Whether celastrol can modulate the expression of pro-apoptotic proteins was also examined. We found that celastrol up-regulated the expression of Bax in a dose-and time-dependent manner but had little effect on Bid at the dose and time employed (Fig. 2, C and D).
Celastrol Induces the Expression of Death Receptor-To further explore the underlying mechanism that may be responsible for enhancement of TRAIL-induced apoptosis by celastrol, we examined the effect of celastrol on the expression of death receptors. Celastrol significantly increased the expression of DR5 proteins in a dose-dependent manner, with optimum induction occurring at around 2-3 M in human breast cancer MDA-MB-231, MCF-7, and T47D cells (Fig. 3A, upper panel). Celastrol also induced the expression of DR4 in MDA-MB-231 and MCF-7 cells. No induction of DR4 in TRAIL-resistant breast cancer T47D cells was observed (Fig. 3A, middle panel).
Whether this induction of the DRs was time-dependent was also examined. In brief, MDA-MB-231 cells were treated with 3 mol/liters celastrol for indicated times, and then analyzed the levels of DRs by Western blotting analysis. Celastrol induced both DR5 and DR4 in a time-dependent manner, with optimum induction occurring after 12 h (Fig. 3A, lower panel).
Whether TRAIL receptors are induced by celastrol at the transcriptional level was investigated by reverse-transcriptase PCR. Celastrol substantially up-regulated both DR4 and DR5 mRNA expression in a time-dependent manner (Fig. 3B). Whether celastrol induces cell surface expression of TRAIL receptors, was also examined. As shown in Fig. 3C, celastrol also FIGURE 1. Celastrol sensitizes breast cancer cell to TRAIL. A, left, cells (3000 cells/well) were incubated various concentrations of TRAIL. After 24 h, cell viability was determined by the MTT assay. Right, cells were pretreated with celastrol for 6 h, washed with PBS to remove celastrol, and then were exposed to the indicated concentrations of soluble TRAIL for 24 h. The cell viability was determined by the MTT assay. Points, mean percentage relative to control-treated cells (n ϭ 5); bars, standard deviation. B, MDA-MB-231 and T74D cells were treated with 2 mol/liter celastrol for 6 h and washed with PBS to remove celastrol. Then cells were treated with the indicated concentration of TRAIL for 24 h. Apoptosis was determined by the Live/Dead Assay. C, cells were exposed to 2 mol/liter celastrol for 6 h, and then the celastrol was removed. Then cells were treated with TRAIL (10 ng/ml) for 24 h. Cells were stained with PI, and the sub-G1 fraction was analyzed using flow cytometry. D, cells were pretreated with celastrol for 6 h and washed out. Then cells were treated with TRAIL for 24 h. Whole cell extracts were prepared and analyzed by Western blotting using antibodies against caspase-3, caspase-8, caspase-9, and PARP.   (Fig. 3D). Together, these findings suggest that the up-regulation of DR5 and DR4 by celastrol is not cell type-specific.
Gene Silencing of TRAIL Receptor Abolished the Effect of Celastrol on TRAIL-To verify that the up-regulation of DR5 by celastrol is needed for sensitization to TRAIL, we used siRNA targeting DR5 and DR4 and thus abolished celastrol-induced expression of these receptors. As shown in Fig. 4A, transfection of cells with DR5 siRNA but not with control siRNA (scrambled RNA; scRNA) reduced the celastrol-induced DR5 up-regulation. Similarly, transfection of cells with siRNA for DR4 reduced the celastrol-induced DR4 expression (Fig. 4A). However, DR4 siRNA had minimal effect on the expression of celastrolinduced DR5 expression.
We next examined whether the suppression of DR5 or DR4 by siRNA could abolish the sensitizing effects of celastrol on TRAIL-induced apoptosis using an esterase staining assay (the Live/Dead Assay; Fig. 4B) and the size of the sub-G1 cell pool (Fig. 4C). The results indicated that the effect of celastrol on TRAIL-mediated apoptosis was effectively blocked in cells transfected with either DR5 or DR4 siRNA (Fig. 4B). On the other hand, treatment with control siRNA had no effect (Fig. 4B). Silencing of DR5 had more dramatic effect on TRAIL-induced apoptosis than that of DR4. The silencing of both receptors abolished the apoptosis as much as silencing of DR5 alone, thus suggesting that DR5 is a major player in TRAIL-induced apoptosis.
Celastrol-induced Up-regulation of DR5 Up-regulation Is p53-independent-Because p53 has been reported to regulate DR5 expression (33), we next examined the role of p53 in celastrol-induced DR5 up-regulation. Because celastrol induced DR4 and DR5 in MDA-MB-231 cells, which are known to express mutant p53 protein (Fig. 5A), p53 may not be needed for celastrol to induce TRAIL receptors. Furthermore, when we treated HCT116 colon cancer cells, which do not express p53 with celastrol, we again found that celastrol induced DR5, thus suggesting that p53 is not needed for celastrol-induced up-regulation of TRAIL receptors (data not shown).
Celastrol-induced Up-regulation of DR5 Up-regulation Is Mediated through Induction of CHOP-Because CHOP has been linked with the up-regulation of DR5 expression (34,35), we next examined the role of CHOP in celastrol-induced DR5 up-regulation. Our results showed that celastrol induced the expression of CHOP, with optimum induction occurring at around 12 h (Fig. 5A).
Activation of JNK by Celastrol Is Not Needed for DR5 Up-regulation-Other triterpenoids, such as methyl-2-cyano-3,12-dioxooleana-1, 9-dien-28-oate (CDDO-Me). and acetylketo-␤-boswellic acid, have been reported to induce DR5 expression through JNK activation (36,37). We further examined the possibility that celastrol activates the JNK pathway in breast cancer cells. An immunoprecipitated kinase complex assay indicated that celastrol could induce JNK activation in a time-dependent manner (Fig. 5B, upper panel), but under the same conditions, it has no effect on total JNK protein level. These results are similar to previous results in melanoma cells (22).
Next, we examined whether inhibition of JNK could block DRs expression using specific inhibitor for JNK. As shown in Fig. 5B (lower panel), pretreatment of cells with a JNK inhibitor (SP600125) suppressed celastrol-induced up-regulation of DR5 partially, but had a greater effect on DR4 expression. These results suggested that JNK is involved in celastrol-induced DRs up-regulation in a part.

Gene Silencing of CHOP Abolishes Celastrol-induced Up-regulation of DR5 Expression and Its Effect on TRAIL-induced
Apoptosis-To clarify the functional role of CHOP in celastrolinduced up-regulation of death receptors, CHOP siRNA was used. Whereas DR5 was up-regulated by celastrol in non-transfected and control-transfected (scRNA) cells, transfection with CHOP siRNA significantly abrogated the up-regulation of DR5 (Fig. 5C, top panel). CHOP siRNA had no effect on celastrolinduced DR4 expression (Fig. 5C, middle panel).
We next examined whether the suppression of CHOP by siRNA could abrogate the effects of celastrol on TRAIL-induced apoptosis using an esterase-staining assay (the Live/ Dead Assay). Celastrol had significantly reduced effects (from 54 to 24%) on TRAIL-induced apoptosis in cells transfected with CHOP siRNA (Fig. 5D, middle panel), whereas treatment with control siRNA had no effect (Fig. 5D, bottom panel). These results indicate that CHOP-dependent DR5 up-regulation contributes to the sensitizing effect of celastrol on TRAIL-induced apoptosis.
Celastrol-induced Up-regulation of CHOP and DR5 Requires ROS-How celastrol induces CHOP-DR5 was further investigated. That ROS plays a critical role in TRAIL-induced apoptosis and in induction of DR5 by various agents has been demonstrated (38,39). To understand the roles of ROS in celastrol-induced DR5 expression, we examined whether celastrol can induce ROS production. As shown in Fig. 6A (left panel), treatment with celastrol increased the production of ROS. This induction of ROS generation was markedly reversed by pretreatment with ROS scavenger NAC (Fig. 6A, right).
Whether ROS production is needed for expression of DR5 by celastrol, was determined next. For this, we pretreated cells with various concentrations of NAC and a thiol antioxidant GSH for 1 h and then exposed them to celastrol for 24 h. We found that celastrol significantly up-regulated DR5 expression, whereas pretreatment of NAC blocked celastrol-induced DR5 protein expressions completely (Fig. 6B). Glutathione also abrogated the effect of celastrol on DR5 expression (Fig. 6B). MDA-MB-231 cells were cultured in 6-well plates and the next day transfected with DR5 siRNA, DR4 siRNA, and control siRNA alone or combined. Twenty-four hours after the transfection, cells were re-seeded in 6-well plates (A and C) or chamber slides (B) and treated with 3 mol/liter celastrol (A). After 24 h, the cells were subjected to preparation of whole cell lysates and Western blotting analysis. B and C, cells were exposed to 2 mol/liter celastrol for 6 h, washed with PBS to remove celastrol, and then treated with 10 ng/ml TRAIL. Cell death was determined by Live/Dead assay (B) and flow cytometry to measure sub-G1 (C).
We also investigated whether ROS production mediates celastrol-induced CHOP induction. For this, cells were exposed to various concentration of NAC or GSH for 1 h, and then treated with celastrol for a further 12 h. As shown in Fig. 6C, celastrol markedly up-regulated CHOP expression, whereas pretreatment with NAC or GSH blocked the celastrol-induced expression of CHOP, thus indicating that ROS generation is critical for celastrol-induced up-regulation of DR5.
Blockage of ROS Production Reversed Effect of Celastol on TRAIL-induced Apoptosis-We next examined whether scavenging of ROS could attenuate the TRAIL-induced cell death potentiated by celastrol. As shown in Fig. 6D, celastrol significantly enhanced TRAIL-induced apoptosis in breast cancer cells, and pretreatment of cells with NAC markedly reduced this celastrol-induced enhancement from 49 to 18%. These results suggest that ROS is needed for the sensitization of cells to TRAIL by celastrol.

DISCUSSION
Among all the apoptosis-inducing cytokines, TRAIL is the only one still being actively pursued for its anticancer properties in the clinic. Many human cancer cell types, however, are resistant to TRAIL-induced apoptosis including chronic lymphocytic leukemia, astrocytoma, meningioma, and medulloblastoma (40,41). Agents that can sensitize tumor cells to TRAIL have a great potential for making cancer therapy more effective.
In the present report we explored the role of celastrol, derived from the traditional Chinese medicine "Thunder of God vine," for its role in sensitizing tumor cells to TRAIL. Using human breast cancer cells, we demonstrate that celastrol significantly enhanced the TRAIL-induced apoptosis. When examined for the mechanism, we found celastrol down-regulated the expression of cFLIP, an inhibitor of caspase-8 (9). Both the short and long forms of cFLIP were equally modulated. Similarly, the down-regulation of cFLIP by the peroxisome proliferator-activated receptor (PPAR)-␥ agonist rosiglitazone has been shown to sensitize cells to TRAIL (42). We found that celastrol also significantly down-regulated the expression of Bcl-2. The latter FIGURE 5. Up-regulation of DR5 by celastrol requires CHOP. A, MDA-MB-231 cells (5 ϫ 10 5 ) were incubated with 3 mol/liter celastrol for the indicated times, and whole cell lysates were subjected to Western blotting analysis using relevant antibodies. B, upper panel, cells were incubated with 3 mol/liter celastrol for indicated times, and whole cell extracts were immunoprecipitated with anti-JNK1 antibody and subjected to kinase assay as described under "Experimental Procedures." The same protein extracts were subjected to Western blotting analysis using anti-JNK1 antibody. Lower panel, cells were pretreated with JNK inhibitor (SP600125) for 1 h and then exposed to 3 mol/liter celastrol for 24 h. Whole cell extracts were prepared and analyzed for the expression of DR4 and DR5 using relevant antibodies. C, MDA-MB-231 cells (3 ϫ 10 6 /well) were transfected with either CHOP siRNA or control siRNA. Twenty-four hours after the transfection, cells were re-seeded in 6-well plates or chamber slides. Cells were treated with 3 mol/liter celastrol for 24 h, and whole cell lysates were analyzed by Western blotting (C). The same blots were stripped and reprobed with ␤-actin antibody to verify equal protein loading. D, cells were exposed to 2 mol/liter celastrol for 6 h, washed with PBS to remove celastrol, and then treated with 10 ng/ml TRAIL. Cell death was determined by Live/Dead assay.
has been linked with suppression of apoptosis by TRAIL (43). In addition we found down-regulation of Bcl-xL, IAP-1, survivin, and XIAP, though to a lesser degree. All these proteins have been linked to TRAIL resistance (10 -12, 15). How these survival proteins are down-regulated by celastrol is less clear. Most of the antiapoptotic proteins described above are regulated by NF-B. Because celastrol has been shown to suppress NF-B activation (21), it is possible that down-regulation of expression of these proteins is through the downregulation of NF-B. When investigated for other potential mechanisms, we found that celastrol significantly up-regulated the expression of Bax but had no effect on Bid expression. The former has been shown to be critical for the TRAIL-induced apoptosis (14). Thus up-regulation of Bax by celastrol could also contribute to TRAIL-induced apoptosis.
We further found that celastrol significantly up-regulated the expression of both the TRAIL receptors, DR4 and DR5. The increase in expression occurred at the transcriptional level and was noted in cancer cells of the breast, lung, colon, prostate, esophageal, and pancreas and in myeloid and leukemia cells. Celastrol not only induced the expression of protein but also mRNA for the death receptor. We also demonstrated the up-regulation of cell surface receptors by celastrol. A large number of agents have been shown to induce death receptors including triterpenoid such as boswellic acid (37) and CDDO (36). We further found that silencing of death receptor induction abolished the effect of celastrol on TRAIL-induced apoptosis, thus suggesting that these receptors play a critical role. A recent report indicated that repeated treatment with subtoxic doses of TRAIL induces resistance to apoptosis in MDA-MB-231 cells (44). This resistance to TRAIL was reported to be mediated through the down-regulation of expression of cell surface death receptors and up-regulation of cFLIP. Because celastrol up-regulates death receptors and down-regulates cFLIP as described here, it should reverse the resistance to subtoxic doses of TRAIL. Our studies further revealed that induction of TRAIL receptors is mediated through the induction of CHOP. Celastrol induces CHOP and the silencing of CHOP gene abolished the effect of celastrol on the induction of death receptors and on the TRAIL-induced apoptosis. Other terpenoids, such as CDDO-Me and acetyl-keto-␤-boswellic acid, were also reported to induce expression of DR5 through CHOP induction (37, 45).

FIGURE 6. Celstrol-induced ROS is involved in CHOP induction leading to DR5 up-regulation.
A, MDA-MB-231 cells were treated with 3 mol/liter celastrol with or without 10 mmol/liter NAC. Twelve hours later, intracellular ROS levels were measured by flow cytometry using CM-H 2 DCFDA, as described under "Experimental Procedures." B, cells (5 ϫ 10 5 cells) were pretreated with various concentrations of NAC or GSH for 1 h and then treated with 3 mol/liter celastrol for 24 h. Whole cell extracts were prepared and analyzed by Western blotting using DR5 antibody. C, cells were treated with either NAC or GSH for 1 h and exposed to 3 mol/liter celastrol for 12 h, and then whole cell extracts were subjected to Western blotting for CHOP. The same blot was stripped and reprobed with ␤-actin antibody to verify equal protein loading. D, MDA-MB-231 cells were pretreated with NAC for 1 h and then treated with celastrol for 6 h. Next cells were washed with PBS to remove celastrol and treated with TRAIL for 24 h. Cell death was determined by the Live/Dead Assay.
Similarly celastrol induced the activation of JNK, an effect that was needed for induction of death receptors. The expression of DR5 by CDDO and boswellic acid also require JNK activation (37,45). Numerous reports indicate that CHOP binds to the DR5 promoter and up-regulates the receptor expression (34,35,37).
We found that perhaps the most important upstream signal linked to modulation of TRAIL signaling is ROS. Celastrol was found to induce ROS, and quenching of ROS abolished the effect of celastrol on induction of CHOP and of death receptors and thus abolished the potentiation of TRAIL-induced apoptosis by celastrol. These results are similar to that reported with sulforaphane for induction of DR5 (39). Our results clearly show that celastrol induces ROS which leads to the up-regulation of CHOP through the activation of JNK (see Fig. 7). Although p53 induction has been linked with induction of death receptors (33), we found that induction of TRAIL receptors was independent of expression of p53.
Celastrol has been shown to be a potent inhibitor of the proteasome activity (20). Several proteasome inhibitors such as MG-132 (46), PS-341 (47), and NPI-0052 (48) have been shown to induce TRAIL receptors. Like celastrol, MG132 also induces DR5 through induction of CHOP (35). Unlike celastrol, however, PS-341 sensitized the cells to TRAIL despite the up-regulation of the expression of cFLIP in lung cancer cells (47). These results suggest that different proteasome inhibitors may mediate their effects through different mechanisms. Celastrol is also a potent activator of heat shock protein (HSP)-90 (26,29,49). Inhibitors of the HSP-90 such as geldanamycin have been shown to sensitize cells to TRAIL (50). This effect of HSP90, however, was mediated through the inhibition of NF-B. Celastrol has been shown to suppress NF-B activation, which may also contribute to TRAIL sensitivity.
Overall our studies provide strong evidence that celastrol could potentiate TRAIL-induced apoptosis through downmodulation of cell survival gene products and up-regulation of death receptors. Among all the triterpenoids tested till now, celastrol was found to be most active. Further studies in animals are needed to investigate the anticancer potential of celastrol in combination with TRAIL.