Immunomodulators such as Bacillus Calmette-Guerin and interferon are clinically active in transitional cell carcinoma of the bladder, but their mechanisms of action remain unclear. Here we investigated the effects of IFNα on tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) expression and apoptosis in a panel of 20 human bladder cancer cell lines. Six (30%) displayed significant DNA fragmentation in response to increasing concentrations of IFNα (10–100,000 units/mL). In these lines IFNα induced early activation of caspase-8, and DNA fragmentation was blocked by a caspase-8-selective inhibitor (IETDfmk), consistent with the involvement of death receptor(s) in cell death. IFNα stimulated marked increases in TRAIL mRNA and protein in the majority of IFN-sensitive and IFN-resistant cell lines. A blocking anti-TRAIL antibody significantly inhibited IFN-induced DNA fragmentation in four of six IFN-sensitive cell lines, confirming that TRAIL played a direct role in cell death. Bortezomib (PS-341, Velcade), a potent TRAIL-sensitizing agent, increased sensitivity to IFNα in two of the IFN-resistant cell lines that produced large amounts of TRAIL in response to IFN treatment. Our data show that IFN-induced apoptosis in bladder cancer cells frequently involves autocrine TRAIL production. Combination therapy strategies aimed at overcoming TRAIL resistance may be very effective in restoring IFN sensitivity in a subset of human bladder tumors.
The immunomodulator Bacillus Calmette-Guerin (BCG) is the current frontline therapy for superficial transitional cell carcinoma (TCC) of the bladder and produces response rates 50 to 89% in previously untreated patients with locally invasive disease (1) . BCG seems to induce tumor regression by stimulating host cells to produce inflammatory cytokines, including tumor necrosis factor-α (2 , 3) , IFNs (4 , 5) , and tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), a death receptor ligand that triggers tumor cell apoptosis after binding its surface receptors (DR4, DR5; ref. 6 ). Recent work indicates that IFNs also display promising activity in TCC (7) and that they can augment the effects of local BCG (8) . Importantly, unlike BCG, IFNα can be delivered systemically, allowing it to be used in patients with disseminated cancer. Preclinical studies have shown that IFNs prevent the growth of orthotopic human bladder tumors in nude mice by inhibiting angiogenesis (9, 10, 11) . However, the possibility that IFNs can also directly induce apoptosis in human bladder cancer cells has not been systematically addressed. We therefore undertook the present study to characterize the effects of IFN on apoptosis within a panel of common TCC cell lines and identify the molecular mechanisms underlying cell death.
MATERIALS AND METHODS
Cell Lines and Reagents.
RT4, 253J-P, and T24 were purchased from American Type Culture Collection (Manassas, VA). The 253J B-V metastatic variant was isolated from the 253J-P cells by orthotopic “recycling” as described previously (12) . Cell lines in the UM-UC series were provided by Dr. Barton Grossman (Department of Urology, University of Texas M. D. Anderson Cancer Center). KU7 cells were provided by Dr. William Benedict. All cell lines are human bladder TCC with the exception of UC4, which is an adenocarcinoma, and UC5 and UC15, which are squamous cell carcinomas. Cells were grown in MEM (Life Technologies, Inc., Rockville, MD) supplemented with 10% fetal bovine serum and 1% each of MEM vitamin solution (Life Technologies, Inc.), sodium pyruvate (BioWhitaker, Walkersville, MD), l-glutamine (BioWhitaker), L-glutamine, penicillin/streptomycin solution, and nonessential amino acids (Life Technologies, Inc.) in a 5% CO2 incubator.
Interferon-α-2A (Roferon, Roche Applied Science, Indianapolis, IN) was purchased from the University of Texas M. D. Anderson Cancer Center Pharmacy. Recombinant TRAIL, the antihuman TRAIL neutralizing monoclonal antibody, and the TRAIL ELISA kit were purchased from R&D Systems (Minneapolis, MN). Other antibodies were obtained from the following commercial sources: IFN regulatory factor-1 (IRF-1, Santa Cruz Biotechnology, Santa Cruz, CA), phosphorylated STAT1 (Cell Signaling, Beverly, MA), caspase-8 (BD Pharmigen, San Diego, CA), Bortezomib (Velcade, PS-341) was provided by Millennium Pharmaceuticals, Inc (Cambridge, MA). Caspase-8-selective inhibitor (IETDfmk) and synthetic substrate (IETD-AFC) were obtained from Enzyme Systems Products, Inc. (Dublin, CA).
Quantification of Apoptosis by Propidium Iodide Staining and Fluorescence-Activated Cell Sorter Analysis.
DNA fragmentation was measured by propidium iodide staining and fluorescence-activated cell sorter (PI/FACS) as described previously (13 , 14) . Cells were stored for at least 1 at 4°C in PI solution before analysis by flow cytometry. Cells that contain a subdiploid DNA content are considered apoptotic (14) .
Measurement of STAT-1 Phosphorylation and IRF-1 Protein Accumulation.
Cells were incubated with 10,000 unit/mL IFNα for the times indicated and lysed for 4 hours at 4°C in 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 a protease inhibitor mixture (Complete Mini tablet, Roche Applied Science)]. Postnuclear extracts were obtained by centrifuging the lysates for 15 minutes at 14,000 rpm (4°C). Protein concentrations were determined by the Bradford method (Bio-Rad, Inc., Hercules, CA). Total lysates (20 μg of protein) were resolved on 12% SDS-PAGE gels and transferred to nitrocellulose membranes as described previously (14) . Blots were probed for 16 hours at 4°C with relevant primary antibodies diluted 1:1,000 in blocking buffer and developed with species-specific secondary antibodies (sheep antimouse horseradish peroxidase, donkey antirabbit horseradish peroxidase, 1:2,000, diluted in blocking buffer, obtained from Amersham, Arlington Heights, IL) for 2 hours at 4°C. Blots were developed by enhanced chemiluminescence (Renaissance, NEN, Boston, MA).
Electrophoretic Mobility Shift Assays.
Cells were incubated with 10,000 units/mL IFNα in MEM containing 1% serum, harvested by trypsinization, resuspended in 0.5 mL buffer A [10 mmol/L HEPES (pH 7.9), 50 mmol/L KCl, 1.5 mmol/L MgCl2, 1.0 mmol/L EDTA and 3% glycerol], lysed by addition of 0.5 mL buffer B (buffer A containing 10% NP40) and gently layered onto a cushion of 3 mL buffer C (containing 10 mmol/L Tris (pH 7.4), 1.5 mmol/L MgCl2, 25% glycerol). Nuclei were collected by centrifugation for 5 minutes at 3,000 rpm. The pellets were washed with 1 mL cold PBS, and nuclear protein was extracted by resuspending the nuclei and rotating them in a buffer containing 20 mmol/L HEPES (pH 7.9), 400 mmol/L NaCl, 1.0 mmol/L EDTA, 1 mmol/L DTT and 1 mmol/L phenylmethylsulfonyl fluoride at 4°C for 30 minutes. Insoluble material was collected by centrifugation at 12,000 × g for 15 minutes at 4°C, and supernatants were snap-frozen and stored at −80°C. Protein concentrations were determined by the Bradford method.
The cis-inducible element oligonucleotide (GTGCATTTGCCGTAAATCTTGTCTACA) containing a consensus binding site for Stat1 (15) was obtained from Santa Cruz Biotechnology. An aliquot of nuclear extract (10 μg of protein) in 19 μL of binding reaction mixture was incubated at room temperature for 20 minutes with 5X binding buffer (5 mmol/L MgCl2, 2.5 mmol/L EDTA, 250 mmol/L KCl, 50 mmol/L Tris (pH 7.5), and 20% glycerol) and 1 μg of poly(dI·dC). The [γ-32P]ATP-labeled cis-inducible element oligonucleotide (50,000 cpm) was incubated with the above binding reaction for 20 minutes at room temperature. For competition experiments, a 150-fold excess of specific unlabeled double-stranded oligonucleotide was added to the binding reaction. The identity of shifted complexes was confirmed by including an anti-Stat-1 E-23X (Santa Cruz Biotechnology) antibody (2 μg) in the reaction mixture (“supershift” analysis). Protein-DNA complexes were resolved by electrophoresis for 3.5 hours at 150V on 5% polyacrylamide gels, and protein-DNA complexes were detected by autoradiography.
Quantification of Caspase-8-Like Protease Activity.
Cells were then treated with 10,000 units/mL IFNα with or without 50 μmol/L IETDfmk for 48 hours. In control experiments, cells were treated with 50 ng/mL recombinant human TRAIL (R&D Systems) with or without 50 μmol/L IETDfmk for 3 hours. Caspase-8 activity was measured in cytosolic extracts as described for caspase-3 previously (16) . Liberated AFC fluorescence was determined at 400 nm excitation and 505 nm emission on a Shimadzu 1500 spectrofluorometer (Shimadzu, Kyoto, Japan).
RNase Protection Assays.
Cells were preincubated overnight in MEM medium containing 1% serum and then exposed to 10,000 units/mL IFNα for 8 hours. We isolated total RNA from cultured cells using an RNeasy kit (Qiagen, Valencia, CA), and RNase protection assay was done using a RiboQuant Multi-Probe kit and the Apo-3D probe set (BD Biosciences, San Diego, CA) according to the manufacturers’ instructions.
Quantification TRAIL Protein Expression.
Cells were incubated with 10,000 units/mL IFNα for 48 hours. The Centricon filtration system (10,000 kDa cutoff, Amicon, Bedford, MA) was used to concentrate conditioned media, and cell pellets were lysed for 30 minutes in a buffer supplied by the manufacturer. The concentrated conditioned media and cellular extracts were assayed for TRAIL content by ELISA according to the manufacturer’s instructions. TRAIL standard curves were generated for each experiment and were used to calculate sample TRAIL content by linear regression analysis. Surface TRAIL expression was measured by immunofluorescence staining and flow cytometry as described previously (17) .
Concentration-Dependent Effects of IFNα on Apoptosis.
We exposed a panel of 20 human bladder cancer cell lines to increasing concentrations of recombinant IFNα (Roferon) and measured apoptosis-associated DNA fragmentation 48 hours later by PI staining and FACS analysis. Six of the cell lines displayed statistically significant (P < 0.05) increases in apoptosis (Fig. 1 ⇓ ; Table 1 ⇓ ). Further analyses with representative IFN-sensitive and IFN-resistant cell lines revealed that resistance was not caused by major defects in IFN receptor-mediated signal transduction. Specifically, although the levels of IFNα-induced STAT-1 phosphorylation (Fig. 2A) ⇓ and IRF-1 protein accumulation (Fig. 2B) ⇓ seemed to be somewhat lower in IFN-resistant cells, IFNα stimulated comparable increases in STAT-1 DNA binding activity in all of the lines examined (Fig. 2C) ⇓ .
IFN-Induced Apoptosis Is Associated with Caspase-8-Like Protease Activation.
Recent studies suggest that death receptors are involved in IFN-mediated apoptosis in other model systems (18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29) . Therefore, we investigated the effects of IFNα caspase-8 activation, which is an early event associated with death receptor ligation. Consistent with the hypothesis, IFNα induced time-dependent (data not shown) caspase-8-like protease activation in both of the IFN-sensitive cell lines examined (Fig. 3A) ⇓ . The levels of caspase-8 activation were similar to those observed in cells treated with recombinant human TRAIL, and caspase-8 activation was blocked by the peptide caspase-8 inhibitor, IETDfmk (ref. 30 ; Fig. 3A ⇓ and data not shown). Immunoblotting confirmed that IFNα promoted the conversion of procaspase-8 into the processed/active form of the protease (Fig. 3B) ⇓ . IETDfmk also reduced the levels of IFN-induced DNA fragmentation to background in all of the cell lines tested (Fig. 3C) ⇓ .
Role of TRAIL in IFN-Induced Apoptosis.
We next investigated the effects of IFNα on the expression of death receptor pathway components using multiprobe RNase protection assays. IFN increased TRAIL mRNA levels in all of the IFN-sensitive lines (Fig. 4A) ⇓ and in most of the IFN-resistant lines (Fig. 4B) ⇓ . The TRAIL receptor, DR4, was also up-regulated by IFN in most of the cell lines (Fig. 4) ⇓ . Some of the cell lines also displayed increases in Fas expression (Fig. 4A and B) ⇓ , but these changes were less dramatic and were not observed in the majority of IFN-sensitive cells (Fig. 4A) ⇓ . A small subset of IFN-resistant cells (n = 3) failed to display any increase in TRAIL mRNA levels in response to IFN treatment (Fig. 4C) ⇓ . IFN also increased TRAIL protein production in most of the cell lines (15 of 20) as measured by ELISA (Table 1) ⇓ . Surface staining and FACS analysis confirmed that IFN increased surface TRAIL expression in IFN-sensitive (RT-4) as well as IFN-resistant (UM-UC5, UM-UC7, UM-UC11) bladder cancer cells (Fig. 4D) ⇓ .
We used a neutralizing anti-TRAIL antibody (25 , 26 , 29) to determine whether or not IFN-induced apoptosis was dependent on TRAIL production. The antibody significantly inhibited IFN-induced DNA fragmentation in four of six of the cell lines (Fig. 5 ⇓ and data not shown). The antibody also consistently reduced levels of IFN-induced DNA fragmentation in the UM-UC6 cells, but the differences observed did not reach statistical significance, and it had no effect in UM-UC-10 cells (data not shown), presumably because they do not express TRAIL (Table 1) ⇓ . In other experiments neither a blocking anti-Fas antibody nor an isotype-matched control antibody had any effect on IFN-induced DNA fragmentation (data not shown).
Effects of Exogenous TRAIL on Apoptosis.
Because most of the IFN-resistant cell lines produced TRAIL in response to IFNα, we wondered whether TRAIL resistance could account for IFN resistance. To address this possibility, we incubated the 20 bladder cancer cell lines in our panel for 24 hours in the absence or presence of 50 ng/mL recombinant human TRAIL and quantified the levels of DNA fragmentation by PI/FACS. Most of the cell lines (16 of 20) displayed significant increases in DNA fragmentation, but the magnitudes of the responses were heterogeneous (Table 1) ⇓ .
Effects of Bortezomib on IFN-Induced Apoptosis.
Some of the most impressive IFN-induced increases in TRAIL production were observed in cell lines that were resistant to TRAIL-induced apoptosis (i.e., UM-UC-7, Table 1 ⇓ ). We therefore wondered whether an agent that is capable of enhancing TRAIL sensitivity would also sensitize cells to IFNα. The proteasome inhibitor, bortezomib, functions as an extremely potent TRAIL sensitizing agent in TRAIL-resistant bladder cancer cells (Fig. 6A ⇓ 4 ). Furthermore, bortezomib synergized with IFN to promote apoptosis in cells that were completely resistant to IFN alone (Fig. 6A) ⇓ . The blocking anti-TRAIL antibody partially inhibited these effects and also reduced the levels of DNA fragmentation observed in the UM-UC5 cells treated with bortezomib alone (Fig. 6B) ⇓ , presumably because direct cytotoxic effects of bortezomib involved the TRAIL that was produced by the cells at baseline (Fig. 4D) ⇓ . These results strongly suggest that modulation of TRAIL sensitivity can enhance IFN-induced apoptosis.
Previous studies in human bladder xenografts have shown that IFNα is a strong inhibitor of bladder cancer growth in vitro and in vivo. The antiangiogenic effects of IFN have been linked to its ability to down-regulate tumor angiogenesis by inhibiting angiogenic factor production and stimulating the expression of antiangiogenic proteins (9, 10, 11 , 31) . Here we investigated effects of IFN on apoptosis within a diverse panel of 20 human bladder cancer cell lines. Thirty percent of the lines displayed statistically significant, IFN-induced increases in DNA fragmentation. Analysis of the molecular mechanisms involved revealed a central role for autocrine TRAIL production in the responses observed, consistent with observations made in other model systems (25 , 26 , 29 , 32 , 33) . The failure of IFNα to stimulate apoptosis in the other 14 lines was not caused by global defect(s) in IFN signal transduction or TRAIL expression. Assuming that this panel of cell lines reflects the spectrum of tumors found in patients, our results suggest that IFN-induced apoptosis will contribute to tumor growth inhibition in a subset of cases and that specific strategies to reverse baseline IFN resistance will have a major impact on its clinical activity.
Although TRAIL expression was important for IFN-mediated apoptosis, other mechanisms also contributed to cell death. The most obvious example of this was found in the UM-UC-10 cells, which were among the most IFN-sensitive lines in the panel but did not express TRAIL in response to IFN treatment (Fig. 1 ⇓ , Table 1 ⇓ ). Previous studies have implicated the transcription factor IRF-1 and the protein kinase regulated by RNA (PKR) in IFN-induced apoptosis in other model systems (28) , making them attractive candidate mediators of this TRAIL-independent cell death.
Most of the cell lines that produced the largest amounts of TRAIL (UC5, UC7, UC11, and UC17) were among the least sensitive to IFN, and some of them were cross-resistant to exogenous TRAIL. We found that these cells could be sensitized to IFN by treating them with the proteasome inhibitor, bortezomib, a potent TRAIL-sensitizing agent (Fig. 6A ⇓ ; ref. 34, 35, 36 ). The apoptosis induced by IFN plus bortezomib was inhibited by a neutralizing anti-TRAIL antibody (Fig. 6B) ⇓ , which suggests that cell death was at least partially TRAIL-dependent. Thus, other TRAIL sensitizers (flavopiridol, histone deacetylase inhibitors; ref. 37, 38, 39, 40 ) may also synergize with IFNα to promote the killing of this subset of cells.
IFNα and other immunomodulators are considered among the most active bladder cancer therapies, but it has been impossible to predict which patients will benefit from them. A very recent study found that urine TRAIL levels correlated with response in patients treated with BCG for superficial TCC (6) , and IFNs are among the cytokines implicated in the effects of this agent. In ongoing studies, we are using quantitative immunofluorescence-based methods to measure TRAIL expression and apoptosis in biopsies obtained from patients enrolled in a clinical trial designed to measure the biological effects of IFNα in patients with TCC in a neoadjuvant setting. It is possible that by monitoring TRAIL expression and apoptosis, we will be able to identify those patients who are benefiting from systemic IFN at an early point in the course of therapy.
Grant support: by a SPORE in Bladder Cancer (P50 CA91846, Project 3) to R. Millikan, H.B. Grossman, W. Benedict, C.P.N. Dinney, and D. 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.
Requests for reprints: David McConkey, Department of Cancer Biology -173, University of Texas M. D. Anderson Cancer Center 1515 Holcombe Boulevard, Houston, Texas 77030. Phone 713-792-8591; Fax: 713-792-8747; E-mail:
↵4 L.M. Lashinger, S.A. Williams, M. Schrader, C.P.N. Dinney, and D.J. McConkey, manuscript submitted.
- Received May 31, 2004.
- Revision received September 2, 2004.
- Accepted October 12, 2004.
- ©2004 American Association for Cancer Research.