Clinical trials have shown that chemotherapy with docetaxel combined with prednisone can improve survival of patients with androgen-independent prostate cancer. It is likely that the combination of docetaxel with other novel chemotherapeutic agents would also improve the survival of androgen-independent prostate cancer patients. We investigated whether the combination of docetaxel and flavopiridol, a broad cyclin-dependent kinase inhibitor, can increase apoptotic cell death in prostate cancer cells. Treatment of DU 145 prostate cancer cells with 500 nmol/L flavopiridol and 10 nmol/L docetaxel inhibited apoptosis probably because of their opposing effects on cyclin B1–dependent kinase activity. In contrast, when LNCaP prostate cancer cells were treated with flavopiridol for 24 hours followed by docetaxel for another 24 hours (FD), there was a maximal induction of apoptosis. However, there was greater induction of apoptosis in DU 145 cells when docetaxel was followed by flavopiridol or docetaxel. These findings indicate a heterogeneous response depending on the type of prostate cancer cell. Substantial decreases in X-linked inhibitor of apoptosis (XIAP) protein but not survivin, both being members of the IAP family, were required for FD enhanced apoptosis in LNCaP cells. Androgen ablation in androgen-independent LNCaP cells increased activated AKT and chemoresistance to apoptosis after treatment with FD. The proteasome inhibitor MG-132 blocked FD-mediated reduction of XIAP and AKT and antagonized apoptosis, suggesting that the activation of the proteasome pathway is one of the mechanisms involved. Overall, our data suggest that the docetaxel and flavopiridol combination requires a maximal effect on cyclin B1–dependent kinase activity and a reduction of XIAP and AKT prosurvival proteins for augmentation of apoptosis in LNCaP cells. [Mol Cancer Ther 2006;5(5):1216–26]
Prostate cancer is the most frequently diagnosed noncutaneous malignancy and the second leading cause of cancer-related deaths among men in the United States (1). The principal therapy for men with advanced disease is androgen ablation, but most of these patients eventually progress to an androgen-independent disease (2). Docetaxel (Taxotere), a semisynthetic derivative of paclitaxel (Taxol) originally derived from the yew tree (3), is a promising anticancer drug shown to inhibit a wide variety of tumor cells, including prostate cancer cells by diverse mechanisms that include cell cycle arrest, induction of apoptosis, stabilization of microtubules, and inhibition of angiogenesis (4, 5). Studies showing that treatment with docetaxel combined with prednisone can improve survival of patients with androgen-independent prostate cancer have been recently reported (6). It is likely that docetaxel combined with other novel chemotherapeutic drugs would also result in improved patient survival. The ability of chemotherapeutic drugs, such as docetaxel, to induce apoptotic cell death in prostate cancer cells is probably one of the chief mechanisms involved in improved survival. However, the precise mechanisms of how docetaxel in combination with other drugs induces apoptosis in prostate cancer cells are not known.
Flavopiridol, a semisynthetic flavonoid derived from an indigenous plant from India, is a broad inhibitor of cyclin-dependent kinases (cdk) and is being tested in clinical trials (7, 8). Treatment of cancer cells with flavopiridol results in a decrease in cyclins D1 and B1 leading to a cell cycle arrest in G1 and G2-M phases (9, 10). Flavopiridol also reduces the levels of the antiapoptotic proteins Bcl-2, Bcl-xL, Mcl-1, and X-linked inhibitor of apoptosis (XIAP) and sensitizes cancer cells to apoptosis after subsequent treatment with other chemotherapeutic agents (10–13). However, results from phase II clinical trials of flavopiridol as a single agent have been reported to be unsatisfactory, indicating that flavopiridol may work best as an anticancer agent when combined with other agents (14, 15). Whether flavopiridol and docetaxel can enhance apoptotic cell death in prostate cancer cells has not been investigated previously.
One of the proposed mechanisms for the anticancer effect of docetaxel is the stabilization of microtubules, activation of the mitotic checkpoint, and blockade of the degradation of cyclin B1, which leads to a prolonged activation of its associated kinase cdk1, mitotic arrest, and induction of apoptosis (4, 5, 16, 17). An increase in cdk1 activity results in phosphorylation and stabilization of survivin, a member of the IAP family and a substrate for cdk1 (18). Therefore, it has been proposed that the subsequent decrease in cyclin B1-cdk1 activity results in a decrease in the levels of survivin and an increase in sensitivity to induction of apoptosis (19). Preclinical studies in human gastric and breast cancer cell lines have shown that the greatest increase in apoptosis occurs when docetaxel is followed by flavopiridol (16). The hypothesis of this regimen is that flavopiridol treatment after the docetaxel-mediated mitotic block results in the inhibition of cyclin B1-cdk activity, a decrease in phosphorylated survivin, a more rapid exit from mitosis, and an increase in apoptosis. Whether this mechanism is generally applicable to androgen-dependent and androgen-independent prostate cancer cells is not known.
Treatment with single chemotherapeutic agents, such as docetaxel and flavopiridol, will not cure most cancers, including androgen-independent prostate cancer. Therefore, the purpose of the present study was to determine whether the combination of docetaxel and flavopiridol could increase apoptotic cell death in prostate cancer cells. Our results show that in androgen-dependent and androgen-independent LNCaP prostate cancer cells sequential treatment with flavopiridol followed by docetaxel (FD) produces the greatest enhancement of apoptotic cell death. Our data suggest that substantial decreases in XIAP (member of the IAP family; refs. 20, 21) and AKT (prosurvival factor; ref. 22) proteins are important mediators for increased apoptosis in FD-treated LNCaP cells. Because XIAP and AKT activity is up-regulated in most types of cancer, including androgen-independent prostate cancer (22–26), and confers resistance to chemotherapeutic drugs (22, 27, 28), these results suggest that drug combinations that substantially reduce XIAP and AKT proteins will provide the greatest extent of apoptosis.
Materials and Methods
Flavopiridol and docetaxel were obtained from Aventis Pharmaceuticals (Bridgewater, NJ). Propidium iodide (PI) and DMSO were purchased from Sigma (St. Louis, MO). Histone H1 protein was purchased from Roche Applied Sciences (Indianapolis, IN). 4′,6-Diamidino-2-phenylindole (DAPI), MG-132, LY 294002, 5,6-dichloro-1-β-d-ribofuranosyl-benzimidazole, alsterpaullone, epoxomicin, and parthenolide were purchased from Calbiochem (San Diego, CA). Annexin V-FITC was purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
Human prostate carcinoma cell lines LNCaP (29) and DU 145 (30) were obtained from the American Type Culture Collection (Rockville, MD). LN-AI is an androgen-independent derivative of the human prostate cancer cell line LNCaP, which was spontaneously derived in our laboratory (31). These cells express androgen receptor and prostate-specific antigen similar to LNCaP. LNCaP, LN-AI, and DU 145 were maintained in RPMI 1640 (Invitrogen, Carlsbad, CA) with 5% fetal bovine serum (Hyclone, Logan, UT), 100 units/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL amphotericin (Invitrogen). Unlike androgen-dependent LNCaP, the LN-AI cells are able to grow for long-term in RPMI 1640 with 5% charcoal-stripped fetal bovine serum and are called LN-AI/CSS. The normal rat prostate basal epithelial cell line NRP-152 (provided by Dr. David Danielpour, Case Western Reserve University, Cleveland, OH) was maintained in HEPES-free DMEM/F12 (1:1, v/v) with 5% fetal bovine serum, antibiotic/antimycotic, 20 ng/mL epidermal growth factor, 10 ng/mL cholera toxin, 5 μg/mL insulin, and 0.1 μmol/L dexamethasone (32). Human mesenchymal stromal cells derived from bone marrow were obtained from Gianluca D'Ippolito (University of Miami, Miami, FL) and cultured in DMEM (low glucose) with 5% fetal bovine serum and antibiotic/antimycotic (33).
Treatment with Flavopiridol and Docetaxel
For treatment with flavopiridol or docetaxel, 7 × 105 LNCaP cells, 5 × 105 LN-AI, 10 × 105 LN-AI/CSS, 3 × 105 DU 145 cells, 1 × 105 NRP-152 cells, and 2 × 105 mesenchymal stromal cells were seeded per 6-cm dish and allowed to attach overnight. The next day, fresh medium containing different doses of flavopiridol (10–500 nmol/L), docetaxel (2–50 nmol/L), or DMSO (0.1%) control was added and the cells were cultured for varying times (24–72 hours). For the sequential combinations of flavopiridol and docetaxel, after 24 hours treatment with 500 nmol/L flavopiridol or 10 nmol/L docetaxel, floating cells were removed, centrifuged, resuspended in the appropriate medium containing flavopiridol, docetaxel, or DMSO, added back to the attached cells, and incubated for an additional 24 hours. In all the experiments, floating and trypsinized attached cells were pooled for further analysis. Similar experiments were conducted using alsterpaullone (5 μmol/L), LY 294002 (20 μmol/L), MG-132 (5 μmol/L), and epoxomicin (1 μmol/L).
Flow Cytometric Analysis
Propidium/hypotonic citrate method (34) was used to study cell cycle distribution of flavopiridol- and docetaxel-treated prostate cancer cells. After harvesting and washing cells with PBS, the cell pellets were resuspended in 0.5 mL PI staining solution (0.1% sodium citrate, 0.03% NP40, 50 μg/mL PI) and vortexed to release nuclei, and DNA distribution histograms were generated by analysis of 10,000 nuclei in a Coulter (Miami, FL) XL flow cytometer. The percentage of cells in the G1, S, and G2-M DNA content was determined by the ModFit program (Verity Software House, Topsham, ME) from six to eight samples analyzed from at least three independent experiments.
Western Blot Analysis
Cell pellets were resuspended in NP40 cell lysis buffer [1% NP40, 50 mmol/L Tris (pH 8.0), 150 mmol/L NaCl, 2 mmol/L EGTA, 2 mmol/L EDTA, protease inhibitor tablet, 50 mmol/L NaF, 0.1 mmol/L NaVO4], lysed by vortex, left on ice for 30 minutes, and centrifuged, and the protein concentrations of the supernatant were determined with the Bio-Rad (Hercules, CA) protein assay. After separation of 25 to 50 μg protein by SDS-PAGE, proteins were transferred by electrophoresis to Immobilon-P membrane and incubated in 5% nonfat dry milk, PBS, and 0.25% Tween 20 for 1 hour. Antibodies specific for cyclin B1 (GNS1), cyclin A (H-432), survivin (FL-142), Mcl-1 (S-19; Santa Cruz Biotechnology), poly(ADP-ribose polymerase) (PARP; C2-10), Bcl-xL (polyclonal; BD Biosciences PharMingen, San Diego, CA), XIAP, cleaved caspase-3, phosphorylated AKT (Ser473; 587F11), and AKT (9272; Cell Signaling Technology, Beverly, MA) were diluted 1:1,000 to 1:3,000 in 5% nonfat dry milk, PBS, and 0.25% Tween 20 and incubated overnight at 4°C. Membranes were washed in PBS and 0.25% Tween 20 and incubated with the appropriate horseradish peroxidase–conjugated secondary antibody (1:1,000 dilution; Santa Cruz Biotechnology) for 1 hour, washed in PBS and 0.25% Tween 20, and analyzed by exposure to X-ray film using Enhanced Chemiluminescence Plus (Amersham Pharmacia Biotech, Piscataway, NJ). Phosphorylated AKT blots were stripped and reprobed with AKT antibody. Antibodies specific for α-tubulin (TU-02; 1:5,000 dilution; Santa Cruz Biotechnology) were used as protein loading controls. Total proteins were stained with Coomassie blue for an additional protein loading control. Changes in protein levels were determined as described previously (35) from at least four different samples analyzed from two to three independent experiments.
Cyclins B1– and A–Dependent Kinase Assay
Total protein (400 μg) was incubated with 2 μg anti–cyclin B1 (H-433; Santa Cruz Biotechnology) or cyclin A antibody for 3 hours on ice followed by the addition of 20 μL protein A/G-agarose (Santa Cruz Biotechnology) and incubation overnight at 4°C. with agitation. Immune complexes were collected by centrifugation, washed thrice with NP40 cell lysis buffer, thrice with kinase buffer [10 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 10 mmol/L MgCl2, 0.5 mmol/L DTT], resuspended in kinase buffer containing 2 μg histone H1 substrate protein, 25 μmol/L ATP, 5 μCi [γ32P]ATP, and incubated for 30 minutes at 30°C. Reactions were stopped with SDS gel loading buffer, samples were electrophoresed on SDS-PAGE, electroblotted to Immobilon P membranes, and analyzed by autoradiography. Coomassie blue staining of membranes revealed similar loading of histone proteins. Changes in kinase activity were determined as described previously (35) from at least four different samples analyzed from two independent experiments.
For the DAPI staining apoptosis assay, cells were resuspended in 0.6 mL of 4% paraformaldehyde/PBS for 15 minutes, washed with PBS, and resuspended in 0.5 mL DAPI (1 μg/mL)/PBS for 10 minutes. Cells were washed with PBS and 10 μL concentrated cells were added on a microscope slide followed by placement of a coverslip. Cells containing densely stained and fragmented chromatin were identified as apoptotic using a Nikon (Melville, NY) fluorescence microscope with a DAPI filter. The number of apoptotic cells in at least 200 total cells was determined from at least four random microscope fields. Changes in apoptosis from flavopiridol- and docetaxel-treated cells were determined as percentage of apoptotic cells in at least five different samples from three independent experiments. Only minimal apoptosis was detected in control-treated cells (<0.5%). For the Annexin V apoptosis assay, LN-AI and DU 145 prostate cancer cells were resuspended in 100 μL Annexin V binding buffer [10 mmol/L HEPES (pH 7.9), 140 mmol/L NaCl, 2.5 mmol/L CaCl2] followed by the addition of 2.5 μL Annexin V-FITC and 2 μL PI (50 μg/mL) and incubated for 20 minutes at room temperature. After the addition of 400 μL Annexin V binding buffer, the cells were read by flow cytometry and the percentage of early apoptotic cells was determined by measuring the Annexin V-FITC–positive/PI-negative quadrant using WinMDI version 2.8.
RNase Protection Assay
RNA from prostate cancer cells sequentially treated with flavopiridol and docetaxel was isolated using QIAshredder and RNeasy miniprep kit (Qiagen, Inc., Valencia, CA). The hAPO-5c human apoptosis multiprobe template set (BD Biosciences PharMingen) was used for T7 RNA polymerase (Ambion, Austin, TX) synthesis of 32P-labeled antisense RNA probes specific for XIAP, survivin, and glyceraldehyde-3-phosphate dehydrogenase. Total RNA (10 μg) was hybridized to antisense RNA probes at 56°C overnight followed by digestion with RNase mixture (Ambion) at 30°C for 45 minutes. RNase digestion products were analyzed by electrophoresis on 5% polyacrylamide-urea gels followed by autoradiography.
Transfection of XIAP Small Interfering RNA
LN-AI/CSS cells (2 × 105) were seeded in 12-well plates and transfected the next day with 200 nmol/L small interfering RNA (siRNA) SMART pool specific for XIAP and siCONTROL nontargeting pool (Dharmacon, Lafayette, CO) using Oligofectamine (Invitrogen) following the manufacturer's instructions. After 72 hours, cells were harvested and analyzed for expression of XIAP by Western blot as described above. Subsequently, LN-AI/CSS cells were transfected with XIAP and control siRNA for 24 hours followed by treatment with FD with or without 20 μmol/L LY 294002 for an additional 48 hours. The numbers of apoptotic cells were determined by DAPI staining as described above. Changes in apoptosis were determined in at least eight different samples analyzed from three independent experiments.
Statistical differences between drug-treated and control cells were determined by two-tailed Student's t test with P < 0.05 considered significant.
To evaluate the effect of flavopiridol and docetaxel as single drugs and in combination, we used various human prostate cancer cell lines (LNCaP, LN-AI, LN-AI/CSS, and DU 145), a nontransformed rat prostate cell line (NRP-152), and a primary human mesenchymal stromal cell line. These cells provide useful in vitro models of the different stages of progression of human prostate cancer, from normal nontransformed to androgen-independent prostate cancer compared with a primary nontransformed nonprostate cell.
G1 and G2-M Cell Cycle Effects of Flavopiridol- and Docetaxel-Treated Prostate Cancer Cells
To evaluate the cell cycle effects of flavopiridol and docetaxel on LNCaP and DU 145 prostate cancer cells, we did flow cytometric analysis after treatment with varying doses of flavopiridol (10–500 nmol/L) and docetaxel (0.5–50 nmol/L) for 24 hours (Fig. 1). The treatment of LNCaP and DU 145 cells with increasing doses of flavopiridol (≥50–100 nmol/L) results in an increase in cells in G1 and G2-M with a decrease in S phase (Fig. 1). These results reflect the ability of flavopiridol to inhibit multiple cdks important in the G1 and G2-M phases of the cell cycle. Flow cytometric analysis of LNCaP and DU 145 cells treated with varying doses of docetaxel (2–50 nmol/L) shows that ≥10 nmol/L docetaxel results in an increase in G2-M and a decrease in G1. As expected, these results are similar to those after treatment with 10 nmol/L paclitaxel (35) and indicate that docetaxel blocks prostate cancer cells in the G2-M phase of the cell cycle.
Opposite Effects of Flavopiridol and Docetaxel on Cyclin B1–Dependent Kinase Activity
To evaluate molecular changes involved in the mediated effects on the cell cycle in LNCaP and DU 145 cells by flavopiridol and docetaxel, we analyzed the expression of cyclins B1 and A proteins by Western blot and kinase analysis (Fig. 2). Cyclin A protein increases during the S and G2 phase of the cell cycle and is believed to be important for DNA replication (36). The transition from the G2 to the M phase of the cell cycle requires an accumulation of cyclin B1 and activation of its associated kinase, cdk1. The end of the G2-M transition and exit from mitosis requires the proteolysis of cyclin B1 and a reduction of cdk1 activity (37). Increased doses of flavopiridol (≥250 nmol/L) decrease both cyclins B1 and A proteins and their associated kinase activities in DU 145 cells (Fig. 2). In contrast, the treatment of DU 145 cells with the G2-M-promoting doses of docetaxel (≥10 nmol/L) for 24 hours results in an increase in cyclin B1 protein and kinase activity but not cyclin A protein and kinase activity (Fig. 2). Similar results were obtained in flavopiridol- and docetaxel-treated LNCaP cells (data not shown). These results indicate that (a) flavopiridol inhibits both cyclins B1– and A–dependent cdk activity probably explaining the G1 and G2-M effects on the cell cycle and (b) docetaxel increases cyclin B1–dependent kinase activity, which correlates with increased G2-M.
Differential Induction of Apoptosis by Flavopiridol and Docetaxel in Prostate Cancer Cells
The induction of apoptosis is a requirement for the effect of chemotherapeutic drugs on prostate cancer cells (38). To measure the induction of apoptosis by flavopiridol and docetaxel, we did a DAPI staining assay in LNCaP, LN-AI, LN-AI/CSS, DU 145, NRP-152, and mesenchymal stromal cells treated with 500 nmol/L flavopiridol and 10 nmol/L docetaxel for 72 hours (Fig. 3A and B). We selected these doses because of the increased effect on cyclin B1–dependent kinase (Fig. 2), which was shown previously to be important in the induction of apoptosis (35). The results indicate that both flavopiridol- and docetaxel-mediated apoptosis was greatest in LN-AI (38–49%) followed by LNCaP (14–25%) cells. A removal of androgens results in a decrease in flavopiridol- and docetaxel-mediated apoptosis in LN-AI/CSS (8–10%) compared with LN-AI cells. The levels of apoptosis in DU 145 (5–12%) are similar to those in LN-AI/CSS cells. Flavopiridol induces a similar degree of apoptosis in nontumorigenic NRP-152 and mesenchymal stromal cells (9%) compared with LN-AI/CSS and DU 145 cells, whereas docetaxel induces less apoptosis in NRP-152 and mesenchymal stromal cells (2–5%) compared with prostate cancer cells. These results indicate that flavopiridol and docetaxel are most effective in producing apoptosis in LN-AI cells and that the removal of androgens reduces the ability of flavopiridol and docetaxel to produce apoptosis in LN-AI/CSS cells.
Treatment with Flavopiridol and Docetaxel Antagonizes Induction of Apoptosis
If the effect of flavopiridol and docetaxel on cyclin B1–dependent kinase activity is important for induction of apoptosis, then it can be predicted from the previous results (Fig. 2) that treatment with flavopiridol (inhibits cyclin B1–dependent kinase) and docetaxel (increases cyclin B1–dependent kinase) should antagonize each other and result in less than additive induction of apoptosis. The treatment of DU 145 cells with 500 nmol/L flavopiridol and 10 nmol/L docetaxel for 72 hours results in only 2% apoptosis, which is a decrease from flavopiridol (5%) and docetaxel (12%) alone (Fig. 3C). This result was confirmed by Western blot showing less cleaved PARP (signifying apoptosis) in the flavopiridol and docetaxel combination compared with docetaxel alone (data not shown). In LNCaP cells, treatment with flavopiridol and docetaxel results in no additional apoptosis above each drug alone (data not shown). These results indicate that simultaneous treatment of prostate cancer cells with flavopiridol and docetaxel results in antagonism with respect to induction of apoptosis probably due to opposing effects on cyclin B1–dependent kinase activity.
Flavopiridol followed by Docetaxel Produces the Greatest Induction of Apoptosis in LNCaP cells
We used the DAPI apoptosis assay in LNCaP cells to identify the sequential combination of flavopiridol and docetaxel that can induce apoptosis to a greater extent than each drug alone (Fig. 4). Previous studies have shown that docetaxel followed by flavopiridol (DF) was more effective in inducing apoptosis in gastric cancer cells (16). Our results indicate that the best sequence for induction of apoptosis is 500 nmol/L flavopiridol for 24 hours followed by 10 nmol/L docetaxel for 24 hours. Similar results with the FD sequence were obtained in LN-AI and LN-AI/CSS cells, although there was less apoptosis in LN-AI/CSS (10%) compared with LN-AI (46%) cells (data not shown). In DU 145 cells, the DF sequence (9%) induces greater apoptosis compared with the FD sequence (2%). However, there was no additional apoptosis in DF-treated DU 145 cells compared with docetaxel followed by docetaxel (DD)–treated cells (Fig. 4). In LN-AI cells, a greater apoptosis in the FD sequence compared with the DF sequence was confirmed by flow cytometry using the Annexin V-FITC/PI assay, whereas in DU 145 cells there was no difference in any sequence of flavopiridol and docetaxel (Fig. 4B).
Interestingly, when flavopiridol was followed by vehicle control in LNCaP, LN-AI, and LN-AI/CSS cells, there was a greater induction of apoptosis compared with sequential use of flavopiridol (FF; Fig. 4C). Similar results were obtained when LNCaP cells were treated with the cdk inhibitor alsterpaullone (5 μmol/L; ref. 39) for 24 hours followed by vehicle control for 24 hours (data not shown), suggesting that increased apoptosis was due to inhibition of cdk activity by flavopiridol. In contrast, the treatment of LNCaP cells with the cdk9 inhibitor 5,6-dichloro-1-β-d-ribofuranosyl-benzimidazole (75 μmol/L; required for transcription elongation; ref. 40) for 24 hours followed by vehicle control or 5,6-dichloro-1-β-d-ribofuranosyl-benzimidazole resulted in low apoptosis (data not shown). These results indicate that the optimal sequence combination in LNCaP cells is flavopiridol followed by docetaxel and that a continuous treatment of LNCaP cells with flavopiridol for 48 hours is less effective in inducing apoptosis compared with a 24-hour treatment followed by no drug.
Increased Apoptosis Is Correlated with Decreased XIAP but not Survivin Protein
To investigate in LNCaP cells why the FD sequence induced greater apoptosis than the DF, FF, and DD sequences, we sought to identify the differences in the levels of proteins important for apoptosis by Western blot analysis. The results in LNCaP and LN-AI cells show that FD treatment contained more cleaved (activated) caspase-3 and PARP (Fig. 5) compared with the DF, FF, and DD sequences. There was much less cleaved caspase-3 in LN-AI/CSS compared with LN-AI cells using the FD sequence. In DU 145 cells, there was a greater cleaved capsase-3 in the DD and DF sequence compared with the FF and FD sequence. These results correlate with those obtained with the apoptosis assays (Fig. 4). There were no significant differences in the levels of the antiapoptotic protein Bcl-xL in any drug sequence.
In contrast, all of the FF- and DF-treated cells show a substantial reduction in the antiapoptotic protein Mcl-1 compared with FD and DD sequence (Fig. 6); however, these results did not correlate with greater apoptosis in the FD sequence. The FD sequence reduces the levels of XIAP and survivin, members of the IAP family (20, 21), in LNCaP, LN-AI, and LN-AI/CSS compared with DU 145 cells (Fig. 6). However, the FD sequence decreases XIAP to a greater extent in LNCaP (7-fold) and LN-AI (10-fold) compared with LN-AI/CSS (3-fold) cells and this correlates with increased apoptosis. In DU 145, there was a 2-fold decrease in XIAP protein in DD- and DF-treated cells. These results suggest that the greatest levels of apoptosis induced by the FD sequence may require a substantial decrease in XIAP protein.
Reduction of XIAP Protein by siRNA Increases Apoptosis in FD-Treated LN-AI/CSS Cells
To determine if a reduction of XIAP protein can increase FD-mediated apoptosis in LN-AI/CSS cells, we used siRNA specific for XIAP. The results show that XIAP siRNA caused a 7-fold decrease in XIAP protein in LN-AI/CSS cells compared with cells transfected with control siRNA (Fig. 7A). Subsequently, LN-AI/CSS cells were transfected with XIAP and control siRNA for 24 hours followed by treatment with FD for an additional 48 hours and the effect on apoptosis was determined by DAPI staining. The results indicate that in the presence of XIAP siRNA there was a small but significant increase in apoptotic LN-AI/CSS cells compared with control siRNA transfected cells (Fig. 7B). This suggests that lowering XIAP protein levels sensitizes prostate cancer cells to FD treatment.
Decrease in XIAP Protein Is Not Due to Decreased mRNA
One of the mechanisms proposed for flavopiridol inhibition of cancer cells is its ability to inhibit cdk7 and cdk9, which are required for phosphorylation of RNA polymerase II and transcription elongation (40). We therefore did RNase protection assay using the hAPO-5c human apoptosis multiprobe template set to determine whether the FD-mediated decrease in XIAP protein was due to a decrease in mRNA. The results show that this treatment of LNCaP, LN-AI, LN-AI/CSS, and DU 145 cells with FD did not reduce the levels of XIAP mRNA compared with control-treated cells (Fig. 8). In contrast, FD treatment of LNCaP, LN-AI, and LN-AI/CSS but not DU 145 cells results in a decrease in survivin mRNA, which may explain the decrease in survivin protein. In addition, there was a decrease in mRNA for XIAP in FF- and DF-treated cells, which possibly corresponds to a decrease in XIAP protein. These results indicate that decreased XIAP protein in FD-treated LNCaP and LN-AI cells was due to post-transcriptional mechanisms and not to decreased XIAP mRNA.
FD Treatment Decreases Total AKT Protein in LN-AI Cells
To investigate molecular changes possibly involved in the reduction of XIAP protein in FD-treated LNCaP cells, we analyzed the expression of activated and total AKT by Western blot. In addition to phosphorylation and inactivation of the proapoptotic protein Bad (22), AKT has recently been shown to phosphorylate and stabilize XIAP (41). Results indicate that LN-AI/CSS cells contain a higher level of activated (phosphorylated at Ser473) AKT compared with LN-AI cells (Fig. 9A). Because activated AKT has been shown to protect cells from apoptosis (22), this may explain why LN-AI/CSS cells are more resistant to apoptosis induced by FD treatment compared with LN-AI cells. We then analyzed the levels of activated and total AKT in FD-treated LN-AI compared with LN-AI/CSS cells (Fig. 9B). The results show that FD treatment greatly reduces the levels of total AKT protein (and therefore activated AKT) in LN-AI cells with few changes in LN-AI/CSS cells. Interestingly, the treatment of LN-AI but not LN-AI/CSS cells with FF results in a 5-fold increase in activated AKT compared with control-treated cells. Inhibition of AKT activity with the phosphatidylinositol 3-kinase inhibitor LY 294002 (20 μmol/L inhibits Ser473 phosphorylated AKT) increases apoptosis (Fig. 9C) and decreases XIAP protein levels (data not shown) in FF- and FD-treated LN-AI/CSS cells. Finally, the combination of a reduction in XIAP protein with siRNA and an inhibition of AKT activity with LY 294002 cause a greater increase in apoptosis than a reduction of XIAP or a decrease in AKT activity alone (Fig. 9D). Overall, these results suggest that (a) an increase in activated AKT in LN-AI/CSS cells may explain its greater resistance to apoptosis; (b) FD treatment reduces total AKT protein levels in LN-AI but not LN-AI/CSS cells, possibly explaining differentially decreased XIAP and increased apoptosis; and (c) a reduction of XIAP protein and an inhibition of AKT activity greatly increases apoptosis in FD-treated LN-AI/CSS cells.
Inhibition of the Proteasome Pathway Antagonizes FD-Induced Apoptosis
A possible mechanism explaining FD decrease in XIAP protein without a reduction in XIAP mRNA may be the increased degradation by the ubiquitin-proteasome pathway (42). To address this hypothesis, we treated LNCaP cells with flavopiridol (500 nmol/L), proteasome inhibitor MG-132 (5 μmol/L), and combination of flavopiridol and MG-132 for 24 hours. The results show that MG-132 blocks the ability of flavopiridol to reduce the protein levels of cyclin B1 and Mcl-1 (Fig. 10A). This suggests that flavopiridol activates the proteasome pathway to decrease these proteins. The addition of MG-132 (5 μmol/L) in the FD sequence combination results in lesser cleavage of PARP in LNCaP cells, which correlates with a block in FD-mediated decrease in XIAP and AKT (Fig. 10B). Conversely, FD lowers cleavage of PARP mediated by MG-132, suggesting antagonism relative to proteasome degradation activity. MG-132 also blocks FD-mediated apoptosis as determined by DAPI staining (data not shown). Similar results were obtained using the proteasome inhibitor epoxomicin (1 μmol/L; data not shown). In addition, treatment of LNCaP cells with FD and the nuclear factor-κB inhibitor parthenolide (20 μmol/L) had no effect on apoptosis, suggesting that an inhibition of nuclear factor-κB activity was not important for the ability of MG-132 to block FD-mediated apoptosis (data not shown). These results suggest that FD increases the degradation of cyclin B1, XIAP, and AKT proteins by stimulating the proteasome pathway and this augments apoptosis in LNCaP cells.
We analyzed in human prostate cancer cells whether the combination of flavopiridol and docetaxel can enhance apoptosis more than either drug alone. Our results indicate that the sequential addition of flavopiridol followed by docetaxel was required for maximal induction of apoptosis in LNCaP cells. The opposing effects on cyclin B1–dependent kinase activity by flavopiridol (a decrease) and docetaxel (an increase) likely plays an important role in the requirement for a sequential combination regimen and an induction of apoptosis. Our data also suggest that the FD-mediated decrease in XIAP and AKT proteins, both being inhibitors of apoptosis, is an important factor for increased apoptosis in LNCaP and LN-AI compared with LN-AI/CSS and DU 145 prostate cancer cells. A possible mechanism for flavopiridol-mediated decrease in XIAP, AKT, cyclin B1, and Mcl-1 proteins may be the activation of the proteasome pathway of protein degradation. Overall, our data suggest that the sequential regimen of flavopiridol and docetaxel that leads to the greatest decrease in XIAP and AKT protein also results in the greatest increase in apoptosis.
The deregulated increase of cyclin B1–dependent kinase activity by docetaxel is an important mechanism for the induction of apoptosis in prostate cancer cells (35, 43). However, it is not clear if an inhibition of cyclin B1–dependent kinase activity by flavopiridol is also important for induction of apoptosis. Our data show that when flavopiridol and docetaxel were added simultaneously there was an antagonism of apoptosis compared with either drug alone (Fig. 3C). This suggests that inhibition by flavopiridol of cyclin B1–dependent kinase activity is important for the induction of apoptosis in prostate cancer cells. Therefore, to maximize apoptosis, flavopiridol and docetaxel should not be added simultaneously because of their opposing effects on cyclin B1–dependent kinase activity.
In contrast to results in gastric cancer cell lines (16), our results show that in LNCaP cells treatment with FD is more effective than that with DF in inducing apoptosis (Fig. 4). Flavopiridol can decrease proteins that inhibit apoptosis, such as Bcl-2, Bcl-xL, Mcl-1, and XIAP (10–13). Therefore, it is possible that in LNCaP an initial treatment with flavopiridol decreases inhibitors of apoptosis and sensitizes cells for subsequent treatment with docetaxel. In contrast to LNCaP cells, DF is more effective for induction of apoptosis than FD in DU 145 cells as determined by DAPI and greater cleavage of caspase-3 and PARP (Figs. 4 and 5). However, overall apoptosis in DF-treated DU 145 cells were not different from that in DD-treated cells (Fig. 4). Therefore, the FD or DF sequence that optimally induces apoptosis may depend on the type of prostate cancer cells (i.e., androgen-dependent or androgen-independent) as well as on the expression of androgen receptor or p53 proteins. Because prostate cancers growing in vivo consist of a heterogeneous mixture of different cell types (44, 45), our results suggest that either the FD or the DF sequence regimen may work in inhibiting tumor growth and inducing apoptosis. In fact, we have shown in the Gγ/T-15 transgenic mouse model of androgen-independent prostate cancer (31) that DF or FD treatment can inhibit the growth of primary and metastatic prostate tumors more effectively than either drug alone by increasing apoptotic cell death (46).
In addition to inhibition of cyclin B1–dependent kinase activity and a decrease in inhibitors of apoptosis, another potential mechanism for flavopiridol is its ability to inhibit transcription elongation (40). It is thought that blocking transcription results in a loss of mRNAs with short half-lives (e.g., Mcl-1 and cyclin D1) and therefore a loss of protein (9, 47). Our results show that FD treatment of LNCaP cells did not decrease XIAP mRNA (Fig. 8) and therefore cannot be an explanation for decreased XIAP protein (Fig. 6). FD treatment results in decreased mRNA for survivin and therefore is a likely explanation for decreased survivin protein in LNCaP cells (Figs. 6 and 8). However, decreased survivin protein does not seem to play a major role in FD-mediated induction of apoptosis because there is less apoptosis in LN-AI/CSS compared with LN-AI cells despite that both have decreased survivin protein and mRNA (Figs. 6 and 8). XIAP protein as higher in FD-treated LN-AI/CSS compared with LN-AI and LNCaP cells and a reduction of XIAP protein with siRNA increased apoptosis (Fig. 7), suggesting a more important role for XIAP. A recent report shows that lowering XIAP protein can sensitize prostate cancer cells to a variety of chemotherapeutic agents (27). In addition, treatment of LN-AI cells with the transcription inhibitor 5,6-dichloro-1-β-d-ribofuranosyl-benzimidazole (75 μmol/L) did not result in increased apoptosis (data not shown), indicating that the ability of flavopiridol to inhibit transcription is not required for induction of apoptosis in FD-treated LN-AI cells.
Another potential mechanism for FD-mediated apoptosis is the differential effect on the prosurvival AKT protein in LN-AI compared with LN-AI/CSS cells (Fig. 9). Our results show that the greater resistance to apoptosis in LN-AI/CSS compared with LN-AI cells was possibly due to higher levels of activated AKT (Fig. 9A). It has been shown previously that androgen ablation can increase AKT activation in LNCaP cells and support survival and proliferation in conditions of androgen deprivation (48). Treatment with FD reduces total AKT protein levels in LN-AI but not LN-AI/CSS cells possibly by increased proteasome degradation of AKT. This may be an explanation why FD treatment induced greater apoptosis in LN-AI compared with LN-AI/CSS cells. Because activated AKT has recently been shown to phosphorylate and stabilize XIAP (41), lower AKT protein (and therefore reduced activated AKT) in FD-treated LN-AI and LNCaP cells may be another explanation for lower XIAP protein. In addition, our results suggest that a 24-hour treatment with flavopiridol followed by no drug is more effective in inducing apoptosis than a continuous 48-hour treatment (Fig. 4C) possibly because of the effect on total AKT protein, which is higher in continuous compared with noncontinuous treatment (Fig. 9B; data not shown). The addition of the phosphatidylinositol 3-kinase inhibitor LY 294002 enhanced apoptosis in FD-treated LN-AI/CSS cells, especially when XIAP protein was reduced with siRNA (Fig. 9C and D), further suggesting the importance of decreasing AKT activity to augment apoptotic cell death. Drug combinations that reduce XIAP protein and activated AKT are more likely to increase apoptosis (22), especially in prostate cancer cells, such as LNCaP, which has a mutated PTEN and a constitutively active AKT (49). However, this strategy may not be effective in prostate cancer cells, such as DU 145, in which PTEN is not mutated and AKT is not constitutively active (50).
A potential explanation for decreased XIAP protein without decreased mRNA in FD-treated LNCaP and LN-AI cells is the activation of the proteasome degradation pathway by flavopiridol. Our results show that the proteasome inhibitor MG-132 can block the ability of flavopiridol to decrease cyclin B1 and Mcl-1 proteins in LNCaP cells (Fig. 10A). This suggests that flavopiridol can decrease these proteins using the proteasome pathway. In FD-treated LNCaP cells containing MG-132, there was less PARP cleavage and higher levels of XIAP and AKT proteins, implying an antagonism between proteasome inhibition and FD treatment (Fig. 10B). These results are in contrast to synergistic induction of apoptosis in leukemic cells treated with flavopiridol and MG-132, which is associated with disruption of the nuclear factor-κB pathway (51). Thus, treatment strategies that are synergistic in one type of tumor may be antagonistic in another type of tumor.
In summary, our studies suggest that the FD sequence combination in LNCaP and LN-AI prostate cancer cells results in the activation of the proteasome pathway, degradation of prosurvival proteins XIAP and AKT, and induction of apoptosis. The FD sequence induces less apoptosis in the LN-AI/CSS cells possibly because of the failure to substantially decrease XIAP and AKT proteins. However, in the DU 145 cells, the DF and DD sequence combinations are more effective than the FD and FF regimens, suggesting a heterogeneous response depending on the type of prostate cancer cell. Phase I clinical trials on the DF sequence in patients with lung cancer have shown few responses, although there was some disease stabilization (8). Because of the heterogeneous nature of prostate cancers, we suggest that the efficacy of the DF or FD sequence may depend on the relative levels of the different types of prostate cancer cells, such as LNCaP (sensitive to FD) and DU 145 (sensitive to DF), present in tumors.
Grant support: Aventis Pharmaceuticals grant GIA 60025, Veterans Affairs Merit Review grant 026901, and Department of Defense grant DAMD17-03-1-0179 (C. Perez-Stable).
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.
We thank Dr. Andrew Schally (Veterans Affairs Medical Center, Miami, FL) and Dr. Diana Brassard (Aventis Pharmaceuticals) for review of this article, Dr. Gianluca D'Ippolito for mesenchymal stromal cells, and Dr. David Danielpour for NRP-152 cells.