Tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) is a member of the tumor necrosis factor family of cytokines that induces apoptosis in some tumor cells but not in normal cells. Unfortunately, many human cancer cell lines are refractory to TRAIL-induced cell death, and the molecular mechanisms underlying resistance are unclear. Here we report that TRAIL resistance was reversed in human bladder and prostate cancer cell lines by the proteasome inhibitor bortezomib (PS-341, Velcade). Synergistic induction of apoptosis occurred within 4 to 6 hours in cells treated with TRAIL plus bortezomib and was associated with accumulation of p21WAF-1/Cip-1 (p21) and inhibition of cyclin-dependent kinase (cdk) activity. Roscovitine, a specific cdk1/2 inhibitor, also sensitized cells to TRAIL. Silencing p21 expression reduced levels of DNA fragmentation by 50% in cells treated with bortezomib and TRAIL, confirming that p21 was required for the response. Analysis of the TRAIL pathway revealed that caspase-8 processing was enhanced in a p21-dependent fashion in cells exposed to TRAIL and bortezomib as compared with cells treated with TRAIL alone. Thus, all downstream components of the pathway (Bid cleavage, cytochrome c release, and caspase-3 activation) were amplified. These data strongly suggest that p21-mediated cdk inhibition promotes TRAIL sensitivity via caspase-8 activation and that TRAIL and bortezomib should be combined in appropriate in vivo models as a possible approach to solid tumor therapy.
Tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) is a homotrimeric cytokine that induces cell death in a variety of different cancer cell types but not in normal cells (1–3). TRAIL promotes apoptosis via binding to two surface receptors (DR4 and DR5) that contain homologous death domains within their cytoplasmic tails, resulting in receptor trimerization and recruitment of the cytosolic death domain–containing protein, Fas-associated death domain (FADD; refs. 4–7). This stimulated conformation of the TRAIL receptor, known as the death-inducing signaling complex (DISC; ref. 8), allows FADD to recruit and activate procaspase-8, which undergoes autocatalytic activation (9). Once fully activated, caspase-8 can either directly cleave and activate downstream effector caspases (3, 7) or it can stimulate a mitochondrial amplification loop by cleaving Bid, a BH3-only member of the Bcl-2 family (10–12). Studies in animal models indicate that systemic therapy with TRAIL is safe, and phase I clinical trials designed to evaluate TRAIL toxicity and antitumor efficacy are being opened this year (13). However, in vitro data show that up to 50% of tumor cell lines do not undergo apoptosis in response to TRAIL. Thus, understanding the molecular mechanisms underlying TRAIL resistance and identifying strategies to reverse it are high priorities for ongoing research.
The 26S proteasome is a multicatalytic enzyme expressed in the nucleus and cytoplasm of all eukaryotic cells that degrades proteins targeted by ubiquitin conjugation (14). The proteasome is responsible for maintaining homeostasis by controlling intracellular levels of cell cycle regulatory proteins (p21, p27, and p53), transcription factors, and certain tumor suppressor genes/oncogenes, making it an attractive therapeutic target in cancer (15–17). Bortezomib is a peptide boronate inhibitor of the proteasome that was developed as an anticancer agent several years ago and was the first such agent approved by the Food and Drug Administration for the treatment of a human cancer (multiple myeloma; ref. 18). It selectively binds to and inhibits the chymotryptic-like activity of the proteasome at nanomolar concentrations, and in the National Cancer Institute's 60 cancer cell line screen, bortezomib displayed a mean IC50 of 7 nmol/L with a unique spectrum of anticancer activity (19). Cellular responses depend on tumor type and range from cell cycle inhibition to apoptosis, and in vivo studies have shown that bortezomib inhibits the growth of a variety of different solid tumors without significant toxicity (19–23).
Here we report that TRAIL-resistant human prostate and bladder cancer cell lines can be rapidly sensitized to TRAIL-induced apoptosis by treating them with bortezomib. The molecular mechanisms underlying the effects of bortezomib involve p21 accumulation and enhanced activation of caspases 8 and 3.
Materials and Methods
Cell culture and reagents. The LNCaP-derived cell line, LNCaP-Pro5 (24), was generously provided by Dr. Curtis Pettaway (Department of Urology, University of Texas M.D. Anderson Cancer Center). The 253J B-V cells were derived from the 253J parental line by orthotopic “recycling” through the mouse bladder as described previously (25). The UM-UC3 cells were obtained from H. Barton Grossman (Department of Urology, University of Texas M.D. Anderson Cancer Center). Human PC-3 and DU-145 prostate cancer cells were obtained from American Type Culture Collection (Rockville, MD). The prostate cancer cells were grown in RPMI 1640 (Life Technologies, Inc., Gaithersburg, MD) supplemented with 10% fetal bovine serum (Life Technologies) and 1% MEM vitamin solution (Life Technologies), sodium pyruvate (Bio Whittaker, Rockland, ME), l-glutamine (Bio Whittaker), penicillin/streptomycin solution (Bio Whittaker), and nonessential amino acids (Life Technologies) under an atmosphere of 5% CO2 in an incubator. The bladder cancer cells were cultured in MEM containing the same supplements. Bortezomib was kindly supplied by Millenium Pharmaceuticals (Cambridge, MA), and recombinant human TRAIL (rhTRAIL) was purchased from R&D Systems, Inc. (Minneapolis, MN).
Quantification of apoptosis by propidium iodide/fluorescence-activated cell sorting. Cells were treated with 10 ng/mL of rhTRAIL and/or 100 nmol/L bortezomib for the times indicated. Both growth and wash medium were saved and cells were harvested with trypsin. Supernatants were removed and pellets were resuspended in 400 μL of propidium iodide (PI) solution (50 μg/mL PI, 0.1% Triton X-100, and 0.1% sodium citrate in PBS). Samples were then incubated overnight at 4°C in the dark before analysis by flow cytometry. The cells with subdiploid DNA content were quantified to determine the percentage of cells containing apoptotic, fragmented DNA (26).
Quantitative analysis of phosphatidylserine exposure. Cells were treated with 10 ng/mL of rhTRAIL and/or 100 nmol/L bortezomib for the times indicated before harvest with trypsin. Exposure of phosphatidylserine was measured by Annexin V binding as described previously (27) using a commercial kit (Annexin V/PE Apoptosis Detection kit, BD Biosciences, San Diego, CA) according to manufacturer's protocol. Cell pellets were washed twice with cold PBS and resuspended in 1× binding buffer [10 mmol/L HEPES/NaOH (pH 7.4), 140 mmol/L NaCl, 2.5 mmol/L CaCl2] at a concentration of 1 × 106 cells/mL. Aliquots of 100 μL were transferred to separate tubes and 5 μL of Annexin V/PE plus 5 μL of 7-AAD (7-amino-actinomycin D) were added to each. After vortexing, cells were incubated at room temperature for 15 minutes in the dark. Samples were diluted with 400 μL of 1× binding buffer, and surface Annexin V immunofluorescence was quantified immediately by flow cytometry.
Immunoblot analyses. Cells were lysed by incubation for 1h at 4°C in 100 μL of Triton lysis buffer [1% Triton X-100, 150 mmol/L NaCl, 25 mmol/L Tris (pH 7.5), 1 mmol/L glycerol phosphate, 1 mmol/L sodium orthovanadate, 1 mmol/L sodium fluoride, and one Complete Mini Protease Inhibitor Cocktail tablet (Roche, Indianapolis, IN)]. Lysates were centrifuged for 5 minutes at 12,000 × g (4°C), and 20 μg of the postnuclear supernatants were mixed with equal volumes of 2× SDS-PAGE sample buffer (50 mmol/L Tris-HCl, 2% SDS, 0.1% bromophenol blue, 10% glycerol, and 5% β-mercaptoethanol). Samples were then boiled, for 5 minutes at 100°C and resolved by 15% SDS-PAGE at 100 V for 90 minutes. Polypeptides were transferred to nitrocellulose membranes for 90 minutes at 100 V in a transfer buffer containing 39 mmol/L glycine, 48 mmol/l Tris, and 20% methanol. Membranes were blocked for 1 hour in 5% milk diluted in TBS containing 0.1% Tween 20 (TBS-T). Membranes were incubated overnight at 4°C with primary antibodies specific for caspase-8 (Cell Signaling Technology, Beverly, MA; 1:1,000 dilution), caspase-3, cytochrome c, p21, or p27 (PharMingen, San Diego, CA; 1:1,000 dilution), or Bid (R&D Biosystems, Minneapolis, MN; 1:1,000 dilution). Blots were washed 3× 5 minutes in TBS-T before incubation with secondary antibodies (horseradish peroxidase–conjugated sheep antimouse or donkey anti-rabbit antibody; Amersham Biosciences, Piscataway, NJ; 1:1,000 dilution) for 2 hours at 4°C. Blots were washed 3× 10 minutes in TBS-T and developed by enhanced chemiluminescence (Renaissance; New England Nuclear, Boston, MA).
Caspase-3 assay. Cells were treated with 100 nmol/L bortezomib and/or 10 ng/mL rhTRAIL for the times indicated and harvested with trypsin. Growth and wash medium were saved and cell pellets were washed once with PBS. Supernatants were removed and pellets were lysed with 200 μL cold lysis buffer [100 mmol/L HEPES (pH 7.4), 1% sucrose, 0.1% CHAPS, 1 mmol/L EDTA, 100 mmol/L DTT] containing a protease inhibitor cocktail (“Complete Mini” Protease Inhibitor Tablet, Boehringer, Indianapolis, IN). Cells were lysed at 4°C for 1 hour and centrifuged, and 800 μL of caspase buffer plus 2 μL of 20 mmol/L DEVD-AFC fluorogenic substrate (AFC 138, Enzyme Systems Products, Livermore, CA) was added to each supernatant. Samples were incubated for 1 hour at 37°C in the dark and diluted with 1 mL caspase buffer, and released AFC fluorescence was quantified using a Shimadzu spectrofluorimeter (Model RF-1501).
Immune complex cdk2 kinase assays. Cells were cultured to 60% confluency in 10-cm dishes and treated with various concentrations of bortezomib or roscovitine for 24 hours. Cells were then harvested with trypsin and lysed by rotating them for 1 hour at 4°C in 1 mL of the Triton X-100 lysis buffer described above. Lysates were cleared by centrifugation for 10 minutes at 12,000 × g (4°C). Supernatants containing 400 μg of protein were then incubated with an anti-cdk2 antibody for 2 hours followed by overnight incubation with 50 μL protein A/G-Sepharose beads (Santa Cruz Biotechnology, Santa Cruz, CA) at 4°C. The beads were then washed twice with lysis buffer and twice more with kinase buffer [25 mmol/L Tris (pH 7.2) and 10 mmol/L MgCl2]. Immunoprecipitates were incubated with 1 μg histone H1, 150 μmol/L ATP, and 20 μCi [γ-32P] ATP in 50 μL of kinase buffer for 15 minutes at 30°C. SDS sample buffer was used to terminate the reaction and the mixture was boiled for 5 minutes at 100°C. Finally, the mixture was loaded onto 12% SDS-PAGE gels and resolved at 100 V for 90 minutes. The gels were stained with Coomassie blue, destained, dried, and analyzed by autoradiography.
Small interfering RNA–mediated silencing of p21. Cells were grown to 60% confluency in 6-well plates and transfected with specific or nonspecific small interfering RNA (siRNA) constructs for 48 hours according to the manufacturer's protocols. The constructs used were the siRNA SMARTpool cdk-N-1A (p21WAF-1/Cip-1) and cdk-N-1B (p27Kip-1) (Upstate Cell Signaling Solutions, Lake Placid, NY) or the siRNA Nonspecific Control IV (Dharmacon RNA Technologies, Lafayette, CO), all at 200 nmol/L. Liposome-mediated transfection was accomplished with Oligofectamine reagent (Invitrogen Life Technologies, Carlsbad, CA) diluted 1:100 in serum-free MEM. Following silencing cells were treated with rhTRAIL (10 ng/mL) and bortezomib (100 nmol/L) for 8 hours and DNA fragmentation was quantified by PI/fluorescence-activated cell sorting (FACS). The efficiency of p21 or p27 silencing was verified in each experiment by immunoblotting.
Effects of bortezomib on tumor necrosis factor–related apoptosis-inducing ligand–induced apoptosis. Previous studies have implicated the nuclear factor κB (NF-κB) pathway in the regulation of TRAIL resistance (28, 29). In preliminary experiments, we found that many human bladder and prostate cancer cell lines are refractory to TRAIL-induced apoptosis at baseline. Hypothesizing that NF-κB activation maintains the resistant phenotype in these cell lines, we treated them simultaneously with TRAIL plus bortezomib (a potent NF-κB antagonist; ref. 30) and measured DNA fragmentation by PI/FACS analysis 24 hours later. The results revealed a dramatic synergistic interaction between bortezomib and TRAIL in all of the cell lines (Fig. 1A). We confirmed these results using an independent measure of apoptosis (Annexin V staining) for detection of phosphatidylserine externalization (Fig. 1B). In contrast, a more selective inhibitor of NF-κB (the IKK antagonist PS-1145; ref. 31) had no effect on TRAIL-induced apoptosis (Fig. 1C), strongly suggesting that NF-κB inhibition did not account for the effects of bortezomib on TRAIL sensitivity.
In subsequent experiments, we characterized the effects of bortezomib on critical components of the TRAIL cell death pathway. Kinetic analyses showed that TRAIL sensitization occurred as early as 4 to 6 hours in the LNCaP Pro5 and 253JB-V cells (60.0 ± 8.56, P < 0.001 and 60.8 ± 10.5, P < 0.001, respectively; Fig. 2A). Immunoblotting studies showed that bortezomib had no effect on proteolytic processing and activation of caspase-8, whereas incubation with TRAIL resulted in partial proteolytic processing of caspase-8 to form a 43- to 41-kDa intermediate by 8 hours (Fig. 2B). In cells treated with bortezomib plus TRAIL, this intermediate formed with much more rapid kinetics (<30 minutes), and it was accompanied by the formation of a smaller fragment (18 kDa) characteristic of the active, large subunit of caspase-8 (Fig. 2B). Cells treated with TRAIL plus bortezomib in the presence of a caspase-8–selective peptide antagonist, IETDfmk (10 μmol/L), displayed no DNA fragmentation above controls (data not shown), consistent with previous studies that showed that caspase-8 is required for TRAIL-induced cell death (32). Cells treated with TRAIL plus bortezomib also displayed enhanced cleavage of Bid and release of cytochrome c (Fig. 2C). Finally, exposure to either bortezomib or TRAIL alone had little effect on procaspase-3, whereas treatment with the combination promoted rapid proteolytic processing of procaspase-3 and its enzymatic activation (Fig. 2D). Together, these data show that bortezomib interacts with the TRAIL pathway at the level of caspase-8 to promote the initiation of mitochondrial events (cytochrome c release) that dramatically amplify caspase-3 activation. These effects probably account for the synergistic induction of DNA fragmentation and phosphatidylserine exposure observed in cells treated with the combination.
Role of p21 in bortezomib-induced tumor necrosis factor–related apoptosis-inducing ligand sensitization. Although treatment with bortezomib alone failed to induce significant increases in apoptosis in the tested cell lines, previous work from our laboratory showed that it blocks DNA synthesis at low nanomolar concentrations in bladder cancer cells irrespective of whether or not it induces cell death (33). The effects on DNA synthesis are associated with accumulation of cyclin-dependent kinase (cdk) inhibitors, p21 and p27, and inhibition of cdk2 and cdc2 activity (30, 33, 34). Furthermore, p21 accumulation is considered a marker for effective inhibition of the proteasome (35). Consistent with the previous studies, bortezomib induced a time-dependent accumulation of p21 in all of the TRAIL-resistant cells examined here (Fig. 3A; data not shown). Bortezomib also stimulated increases in p27 expression with similar kinetics (Fig. 4B; data not shown). Immune complex kinase assays confirmed that accumulation of p21 and p27 was associated with inhibition of cdk2 activity (Fig. 2B).
To determine whether or not cdk inhibition was sufficient to promote TRAIL sensitization, we examined the effects of the broad-spectrum cdk inhibitor, roscovitine, on TRAIL-induced apoptosis. Roscovitine had no effect on apoptosis at the concentration and time point studied in the 253J B-V cells but did induce DNA fragmentation in LNCaP-Pro5 cells (Fig. 3C). Combined treatment with roscovitine plus TRAIL resulted in synergistic induction of DNA fragmentation in both cell lines as measured by PI/FACS (Fig. 2C). Similar results were obtained with another, structurally unrelated cdk inhibitor (olomoucine) but not with an inactive structural analogue of the compound (iso-olomoucine; data not shown). Together, these results suggest that inhibition of cdk activation is sufficient to explain the effects of bortezomib on TRAIL sensitization. However, cdk inhibitors (i.e., flavopiridol) can also interfere with transcription (36, 37), and these off-target effects may contribute to the TRAIL sensitization observed in cells treated with roscovitine or olomoucine as well.
To more directly assess the involvement of p21 and p27 in bortzomib-mediated TRAIL sensitization, we compared the levels of DNA fragmentation observed in LNCaP-Pro5 cells exposed to a control siRNA construct with those observed in LNCaP-Pro5 cells exposed to siRNA specific for p21 or p27. Immunoblotting confirmed that silencing was efficient in cells exposed to either of the specific constructs but not in the controls (Fig. 4A and B). Levels of DNA fragmentation in the p21-silenced cells were significantly lower than those observed in controls (32% versus 66%, or a 50% reduction, P < 0.001), confirming that the bortezomib-induced accumulation of p21 contributed directly to TRAIL sensitization. Levels in the p27-silenced cells also seemed consistently lower than in controls (Fig. 4B), but the effects did not reach statistical significance, and our attempts to simultaneously silence both p21 and p27 were unsuccessful. Silencing p21 inhibited procaspase-8 activation as measured by immunoblotting (Fig. 4C) or using a fluorigenic caspase-8 peptide substrate (data not shown), demonstrating that p21 acted at the level of procaspase-8 to promote cell death. Thus, processing of procaspase-3 was also reduced in cells depleted of p21 (Fig. 4C).
Bortezomib and TRAIL are undergoing evaluation in clinical trials in a variety of different malignancies. Here we report that they can be combined to induce synergistic cell death in genitourinary cancer cells in vitro and in vivo. Characterization of the molecular mechanisms involved link the effects of bortezomib to increased caspase-8 activation, indicating that the drug affects TRAIL sensitivity at one of the earliest steps in the pathway. Cell death occurred with strikingly rapid kinetics (4-8 hours) as compared with responses to single or combined conventional chemotherapeutic agents, which in our hands require 24 to 48 hours in these cells. In fact, the kinetics of cell death observed here were more rapid than any we have observed in a solid tumor model exposed to any agent, including pharmacologic agents (staurosporine and thapsigargin) that are considered the most potent triggers of cell death.
The transcription factor NF-κB has received considerable attention for its role in cancer cell survival pathways (38). Bortezomib is a potent inhibitor of NF-κB activation via stabilization of NF-κB's physiologic inhibitor, IκBα, and its effects as a NF-κB inhibitor have been used to sensitize cancer cells to other death stimuli (38). Although inhibition of NF-κB was an attractive explanation for bortezomib's effects on TRAIL sensitivity, we were unable to mimic them with a more specific inhibitor of the pathway (the IKK inhibitor PS-1145). Rather, TRAIL sensitization was associated with the accumulation of p21 and inhibition of cdk2 activity, and it was reversed in cells transfected with an siRNA construct specific for p21. Although it is possible that p21 promotes TRAIL sensitivity via a direct mechanism, the observation that chemical cdk inhibitors like roscovitine (Fig. 3C; ref. 39) and flavopiridol (40–42) can also synergistically sensitize cells to TRAIL strongly suggests that p21's effects are mediated by cdk inhibition. Based on these results, we would predict that any stimulus that directly or indirectly causes cdk inhibition would sensitize cancer cells to TRAIL-mediated cell death. Support for this concept comes from the observation that tumor cells are most sensitive to TRAIL in the G1 phase of the cell cycle (43), and DNA damaging agents synergize with TRAIL to promote apoptosis in cells that retain wild-type p53 (44), where p53-mediated p21 expression and cell cycle arrest should occur. Accumulation of p21 also underlies TRAIL sensitization induced by resveratrol (45) and probably contributes to the synergistic increases in apoptosis observed in cells treated with TRAIL plus histone deacetylase (HDAC) inhibitors (46, 47).
Although our data suggest that p21-mediated cdk inhibition is responsible for the increased caspase-8 activation observed in cells treated with bortezomib plus TRAIL, further study is required to elucidate the specific mechanisms involved. One issue is our observation that enhanced caspase-8 processing was detected as early as 30 minutes after treatment with TRAIL plus bortezomib, which was somewhat faster than the kinetics of p21 accumulation measured by immunoblotting. This observation coupled with the incomplete suppression of DNA fragmentation observed in the p21- or p27-silenced cells suggest that additional bortezomib-sensitive mechanism(s) are involved. Studies in other models concluded that bortezomib enhanced surface DR5 expression (48) and decreased levels of c-FLIP (48, 49), both of which could contribute to the increased caspase-8 activation observed. We have confirmed that bortezomib and roscovitine increase surface DR5 expression in the LNCaP-Pro5 and 253J B-V cells, but their effects are delayed (>12 hours) relative to the rapid kinetics of caspase activation and DNA fragmentation (4-8 hours).3
L. Lashinger and M. Shrader, unpublished observations.
L. Lashinger, unpublished observations.
S. Williams, unpublished observations.
Accumulating evidence indicates that cell cycle progression and cell death are mechanistically interrelated (53). Specifically, alterations that promote cell cycle progression often sensitize cells to death, whereas processes that inhibit cell cycle progression block cell death (53). Most of the investigational agents being studied at present (i.e., growth factor receptor antagonists, kinase inhibitors, HDAC inhibitors, bortezomib, etc) arrest cells in G1 (54), which may enable them to reinforce the growth inhibitory/cytostatic effects of conventional chemotherapy but probably does not make them particularly effective in promoting cell killing. Coupled with the other studies described above, our data strongly suggest that cell cycle arrest at the G1-S checkpoint promotes sensitivity to TRAIL-mediated apoptosis in cancer cells, which places it in a unique category relative to other death-inducing stimuli. Thus, TRAIL-based combination therapy seems qualitatively different from other combinations of biological and cytotoxic agents because it is most active in cells that have been growth arrested. The data provide a compelling rationale for doing more extensive studies to optimize the antitumor activities of these combinations in appropriate preclinical models in preparation for clinical studies in patients. Our preliminary studies4 indicate that biologically active doses of bortezomib and recombinant human TRAIL can be delivered to nude mice without generating systemic toxicity.
Grant support: National Research Service Award grant ZRG 1F0920 (L.M. Lashinger), Department of Defense Prostate Cancer Research Program grant W81XWH-04-1-0182 (D.J. McConkey), and M.D. Anderson Specialized Programs of Research Excellence in Bladder Cancer grant P50 CA91846 (C.P.N. Dinney and D.J. McConkey).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.