In a recent study, we showed that the proteasome inhibitor bortezomib sensitizes human bladder cancer cells to IFN-induced cell death. Here, we characterized the molecular mechanisms underlying the antitumoral effects of the combination in more detail. Bortezomib synergized with IFN-α to promote apoptosis via a tumor necrosis factor–related apoptosis-inducing ligand–associated mechanism but did not inhibit production of proangiogenic factors (vascular endothelial growth factor, basic fibroblast growth factor, and interleukin-8) in human UM-UC-5 cells. In contrast, exposure to the combination did not increase the levels of apoptosis in human UM-UC-3 cells but did inhibit the production of basic fibroblast growth factor and vascular endothelial growth factor. Studies with tumor xenografts confirmed that combination therapy with bortezomib plus IFN-α was effective in both models but that the effects were associated with differential effects on tumor necrosis factor–related apoptosis-inducing ligand–associated apoptosis (predominant in UM-UC-5) versus inhibition of angiogenesis (predominant in UM-UC-3). Together, our results show that combination therapy with IFN-α plus bortezomib is effective but can work via different mechanisms (apoptosis versus angiogenesis inhibition) in preclinical models of human bladder cancer. [Mol Cancer Ther 2006;5(12):3032–41]

Bladder cancer is the fifth most common cancer in the United States, with >50,000 cases diagnosed each year (1). The majority of patients have superficial disease that can be managed with surgery. However, superficial tumors recur in 60% to 70% of patients and ∼30% of these tumors progress to a higher grade or stage, necessitating adjuvant therapy (2). Immunomodulators display excellent activity in superficial bladder cancer, with intravesical Bacillus Calmette-Guerin administration displaying the highest efficacy (3, 4). The antitumoral efficacy of Bacillus Calmette-Guerin correlates directly with induction of tumor necrosis factor–related apoptosis-inducing ligand (TRAIL; ref. 5) and other inflammatory cytokines (tumor necrosis factor and IFNs; ref. 3) that can be measured in the urine. Therefore, it may be possible to optimize therapy by administering one or more of these cytokines directly, perhaps in combination with modulators of their antitumoral activity.

We recently showed that recombinant human IFN-α induces apoptosis by stimulating TRAIL production in ∼30% of human transitional cell carcinoma cell lines in vitro (6). Interestingly, many of the IFN-α-resistant cells produced TRAIL in response to IFN-α but were TRAIL resistant (6), suggesting that agents that promote TRAIL sensitivity might also enhance the effects of IFN-α on apoptosis. Recent work indicates that the proteasome inhibitor bortezomib (also known as PS-341 or Velcade) is a potent TRAIL-sensitizing agent (7), and our own previous studies have shown that bortezomib promotes IFN- and TRAIL-induced apoptosis in human bladder cancer cells (6, 8).

We therefore designed the present study to characterize the antitumoral effects of combination therapy with IFN-α plus bortezomib. Our results show that bortezomib augments IFN-α-induced apoptosis in cells that produce TRAIL in response to IFN-α exposure, but the data also indicate that these effects are strictly schedule dependent. Importantly, we also show that combination therapy can still be biologically active in tumors that do not produce TRAIL but that in this case tumor growth inhibition is more closely linked to angiogenesis inhibition. Our results provide the conceptual framework for evaluating the effects of this combination of clinically approved biological agents in patients with bladder cancer.

Animals, Cell Lines, and Reagents

Male nude mice (BALB/c background) were purchased from the Animal Production Area of the National Cancer Institute-Frederick Cancer Research and Development Center (Frederick, MD). The UM-UC-3 (TCC) and UM-UC-5 (squamous) cell lines were provided by H. Barton Grossman (Department of Urology, M.D. Anderson Cancer Center, Houston, TX) and maintained as described previously (6). IFN-α-2A (Roferon, Roche Applied Science, Indianapolis, IN) was purchased from the University of Texas M. D. Anderson Cancer Center Pharmacy. TRAIL, vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), and interleukin-8 (IL-8) ELISA kits were purchased from R&D Systems (Minneapolis, MN). Polyclonal anti-caspase-3 and caspase-8 antibodies were obtained from Cell Signaling (Beverly, MA). Bortezomib (Velcade, PS-341) was provided by Millennium Pharmaceuticals, Inc. (Cambridge, MA).

Quantification of Apoptosis In vitro

DNA fragmentation was measured by propidium iodide staining and fluorescence-activated cell sorting as described previously (6). Cells were stored for at least 1 h at 4°C in a solution containing 50 μg/mL propidium iodide, 0.1% Triton X-100, and 0.1% sodium citrate and subsequently analyzed by flow cytometry (FL3 channel). Cells that contained a subdiploid DNA content were considered apoptotic.

Clonogenic Survival Assays

Cells were incubated with or without 10,000 units/mL IFN-α for 24 h, washed, incubated with or without 10 nmol/L bortezomib for an additional 24 h, and washed. After drug treatment, cells were harvested and 200 cells per well were plated into 60-mm dishes with fresh medium for 10 days. The colonies were washed with PBS, fixed with methanol, and stained with crystal violet. The surviving fraction was determined by dividing the number of surviving colonies in the treated wells by the number of colonies in the untreated control groups.

Quantification of TRAIL mRNA and Protein Levels

Cells were preincubated in the presence or absence of 10,000 units/mL IFN-α for 12 h in MEM containing 1% serum. The cells were then incubated for an additional 12 h with or without 50 nmol/L bortezomib. Total RNA was isolated using RNeasy kits (Qiagen, Inc., Valencia, CA), and RNase protection assays were done as described previously (6) using a commercial probe set (BD PharMingen, Inc., San Diego, CA). In parallel, TRAIL protein levels were examined in total cell or tumor extracts and conditioned medium by ELISA using kits obtained from R&D Systems.

Quantification of Angiogenic Factor Production by ELISA

Cells were preincubated in the presence or absence of 10,000 units/mL IFN-α for 12 h and then incubated with or without 50 nmol/L bortezomib for an additional 12 h. Conditioned media were collected from the cells, and the levels of bFGF, VEGF, and IL-8 were measured in them by ELISA using commercial kits. Absorbance values were translated into protein concentrations using a standard curve and normalized to account for differences in cell numbers.

Effects on the Growth of Tumor Xenografts

Cells were harvested by exposure to trypsin and resuspended in serum-free HBSS. Viability was assessed by trypan blue exclusion, and only single-cell suspensions exhibiting >95% viability were used. Cells (1 × 106) in 200 μL Matrigel (Becton Dickinson Labware, Bedford, MA) were injected over the flanks of the mice. To prevent leakage, a cotton swab was held for 30 s over the site of injection. Mice were treated with 104 or 106 units IFN-α, 1 mg/kg bortezomib, or IFN-α plus bortezomib twice weekly for up to 3 weeks. In the combination studies, IFN was given 24 h before administration of bortezomib to mimic the optimal schedule identified in the in vitro studies. Therapy was limited to twice weekly to allow for complete recovery of 20S proteasome activity before subsequent dosing (9). Tumors were measured throughout the course of therapy (twice weekly) using the formula (L × W2) / 2, where L and W are tumor length and width, respectively. The xenograft studies were approved by the M. D. Anderson Institutional Animal Care and Use Committee and conformed to the usual guidelines for the ethical treatment of animals.

Quantification of Apoptosis in Tumor Sections

DNA fragmentation associated with apoptosis was quantified by fluorescent terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) using a commercial kit (DeadEnd Fluorometric TUNEL System, Promega, Madison, WI) as described previously (10, 11). TUNEL-positive cells were quantified manually. To confirm the TUNEL results, paraffin sections were also stained with an antibody specific for the activated (processed) form of caspase-3 (Cell Signaling). Antigen retrieval was accomplished by microwaving the sections in 0.01 mol/L citrate buffer (pH 6.0). Staining was amplified using a biotinylated anti-rabbit secondary antibody and Cy5-conjugated streptavidin. Nuclei were counterstained for 5 min with 1 μg/mL Hoechst dye in PBS. Percentages of caspase-3-positive cells were determined manually. For each group (control, IFN, bortezomib, and combination), at least 10 independent fields were selected at random from different tumor sections so that the comparison among groups would involve roughly equivalent numbers of cells.

Quantification of Tumor TRAIL Expression by Laser Scanning Cytometry

Tumor TRAIL expression was detected by immunofluorescent staining using a polyclonal rabbit anti-TRAIL primary antibody (Santa Cruz Biotechnology, Santa Cruz, CA), a biotinylated anti-rabbit secondary antibody, and Cy5-conjugated streptavidin. Antigen retrieval for TRAIL was done by incubating sections at 37°C for 20 min. We used fish gelatin (1:10 dilution) in PBS as a blocking solution and for diluting the primary antibody. Sections were subsequently counterstained with Hoechst dye in PBS for 5 min, and ProLong antifade mounting agent (Molecular Probes, Eugene, OR) and a cover slide were added to each slide. Images were captured as described above. TRAIL expression was quantified by laser scanning cytometry (Compucyte, Cambridge, MA) analysis using phantom contouring. The mean fluorescence intensity of the brightest population TRAIL-positive cells was determined. Areas of the same size were used for each condition, and background staining was subtracted out using secondary antibody–stained control sections.

Effects of Therapy on Angiogenesis

Paraffin sections (4–6 μm thick) were mounted on positively charged Superfrost slides (Fisher Scientific, Co., Houston, TX) and dried overnight. Sections were deparaffinized in xylene followed by treatment with a graded series of alcohol [100%, 95%, and 80% ethanol/double-distilled water (v/v)] and rehydrated in PBS (pH 7.5), treated with pepsin (Biomeda, Foster City, CA) for 15 min at 37°C, and washed with PBS. Immunohistochemical procedures for detection of CD31 and VEGF were done as described previously (10, 11). Positive reactions were visualized by incubating the slides with stable 3,3′-diaminobenzidine for 10 to 20 min. The sections were rinsed with distilled water, counterstained with Gill's hematoxylin (colorimetric development), and mounted with Universal Mount (Research Genetics, Birmingham, AL). Control samples exposed to secondary antibody alone showed no specific staining.

Statistical Analyses

The primary end point of this study was tumor size after treatment in this 2 × 2 factorial design. Based on prior data, the study with the sample size of 10 mice per treatment group was expected to have >90% power to detect a minimum difference of 24 mm3 in tumor size at a statistical significance level of 0.05%. Tumor weights, TUNEL, CD31, and propidium iodide/fluorescence-activated cell sorting percentages were compared by unpaired Student's t test. The treatment effects of IFN-α and bortezomib on tumor volumes were studied simultaneously in this 2 × 2 factorial design. Statistical significance for this study was set at two-sided P < 0.05. All statistical analyses were done using commercial software (InStat, GraphPad Software, San Diego, CA).

Effects of IFN plus Bortezomib on Apoptosis

We first compared the effects of different dosing sequences on apoptosis induced by IFN-α plus bortezomib in UM-UC-3 and UM-UC-5 cells in vitro. Cells were incubated with one of the agents for 24 h, washed, and exposed to the second agent for an additional 24 h. This approach was possible because bortezomib is a reversible inhibitor of the proteasome (12). Exposure to IFN-α followed by bortezomib led to substantial induction of apoptosis in the UM-UC-5 cells but not in the UM-UC-3 cells as measured by propidium iodide/fluorescence-activated cell sorting analysis (Fig. 1A). In contrast, cells pretreated with bortezomib followed by IFN-α displayed levels of DNA fragmentation that were indistinguishable from the levels observed in cells exposed to bortezomib alone (Fig. 1B). The levels of apoptosis observed in cells exposed to IFN-α plus bortezomib simultaneously for 48 h were identical to the levels observed in Fig. 1B (data not shown). Similar sequence-dependent effects were observed in UM-UC-7 and UM-UC-11, two other TRAIL-refractory lines that produce substantial amounts of TRAIL in response to IFN-α exposure (data not shown; ref. 6). Therefore, we conclude that bortezomib can augment the proapoptotic effects of IFN-α in some cell lines as long as it is added to them after IFN-α.

Figure 1.

Sequence-dependent effects of IFN-α plus bortezomib on cell death. A, effects of IFN-α followed by bortezomib (BZ). Cells were preincubated with 104 units/mL recombinant human IFN-α for 24 h, washed, and exposed to 100 nmol/L bortezomib for an additional 24 h. DNA fragmentation was quantified by propidium iodide/fluorescence-activated cell sorting. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. B, effects of bortezomib followed by IFN-α. Cells were incubated with bortezomib for 24 h, washed, and exposed to 104 units/mL recombinant human IFN-α for an additional 24 h. DNA fragmentation was quantified by propidium iodide/fluorescence-activated cell sorting. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. C, effects of IFN-α plus bortezomib on clonogenic survival. Cells were incubated with 104 units/mL IFN-α, washed, and exposed to 100 nmol/L bortezomib for an additional 24 h. Cells were then harvested, washed, and replated, and colony formation was measured 10 d later by crystal violet staining. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. **, P < 0.01 versus IFN-α or BZ alone.

Figure 1.

Sequence-dependent effects of IFN-α plus bortezomib on cell death. A, effects of IFN-α followed by bortezomib (BZ). Cells were preincubated with 104 units/mL recombinant human IFN-α for 24 h, washed, and exposed to 100 nmol/L bortezomib for an additional 24 h. DNA fragmentation was quantified by propidium iodide/fluorescence-activated cell sorting. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. B, effects of bortezomib followed by IFN-α. Cells were incubated with bortezomib for 24 h, washed, and exposed to 104 units/mL recombinant human IFN-α for an additional 24 h. DNA fragmentation was quantified by propidium iodide/fluorescence-activated cell sorting. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. C, effects of IFN-α plus bortezomib on clonogenic survival. Cells were incubated with 104 units/mL IFN-α, washed, and exposed to 100 nmol/L bortezomib for an additional 24 h. Cells were then harvested, washed, and replated, and colony formation was measured 10 d later by crystal violet staining. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. **, P < 0.01 versus IFN-α or BZ alone.

Close modal

We then compared the effects of IFN-α, bortezomib, or a combination of the two agents on clonogenic survival in the UM-UC-3 and UM-UC-5 cells. In these studies, we adopted the schedule that was found to be optimal for inducing apoptosis in the UM-UC-5 cells (IFN-α followed by bortezomib; see Materials and Methods). Exposure to either IFN-α or bortezomib alone resulted in good (>60%) inhibition of colony formation in both cell lines (Fig. 1C). However, and consistent with the apoptosis results, scheduled exposure to the combination resulted in further inhibition of colony growth in the UM-UC-5 cells (P < 0.01) but had no further effect in the UM-UC-3 cells (Fig. 1C).

Effects of Bortezomib on TRAIL Expression

We showed that IFN-α-induced apoptosis is dependent on induction of TRAIL in human bladder cancer cells (6). In preliminary experiments, we confirmed that apoptosis induced by IFN plus bortezomib in the UM-UC-5 cells was associated with caspase-8 activation and blocked by a peptide inhibitor of caspase-8 (data not shown). Furthermore, cell death induced by the combination was attenuated by preincubating the UM-UC-5 cells with a blocking anti-TRAIL antibody (6) or by transiently transfecting cells with a small interfering RNA construct specific for TRAIL (data not shown). Therefore, to better understand the schedule-dependent effects of the IFN-bortezomib combination on apoptosis, we assessed the effects of IFN-α, bortezomib, or both agents on TRAIL mRNA and protein expression in the UM-UC-3 and UM-UC-5 cells. Consistent with our previous studies, UM-UC-5 cells expressed TRAIL mRNA (Fig. 2A) and protein (Fig. 2C) at baseline but these levels increased further in response to IFN-α (Fig. 2A–C). In contrast, UM-UC-3 cells expressed no detectable TRAIL mRNA (Fig. 2A) or protein (6) at baseline and IFN-α had no further effect on TRAIL expression. Strikingly, bortezomib inhibited TRAIL mRNA and protein expression in a time-dependent fashion, with dramatic reductions in mRNA observed as early as 4 h and reductions in protein occurring at ∼12 h (Fig. 2).

Figure 2.

Effects of bortezomib on TRAIL expression. A, UM-UC-5 (left) or UM-UC-3 (right) cells were incubated with 104 units/mL IFN-α for 24 h and then 100 nmol/L bortezomib for an additional 24 h, and TRAIL mRNA expression was measured by RNase protection assay. Note the complete inhibition of baseline and IFN-α-induced TRAIL expression in the UM-UC-5 cells exposed to bortezomib (compare lanes 1 and 2 with lanes 3 and 4). Also note that UM-UC-3 cells produced no detectable TRAIL mRNA under any condition. B, time course analysis of bortezomib-mediated inhibition of TRAIL. Cells were preincubated with bortezomib for the times indicated. Cells were then incubated with 104 units/mL IFN-α for 8 h, mRNA was isolated, and levels of TRAIL expression were quantified by RNase protection assay. C, time-dependent effects of bortezomib on TRAIL protein expression. UM-UC-5 cells were incubated with IFN-α, bortezomib, or a combination of the two for the times indicated. TRAIL protein expression was then measured in cell lysates (L) and conditioned medium (CM) by ELISA.

Figure 2.

Effects of bortezomib on TRAIL expression. A, UM-UC-5 (left) or UM-UC-3 (right) cells were incubated with 104 units/mL IFN-α for 24 h and then 100 nmol/L bortezomib for an additional 24 h, and TRAIL mRNA expression was measured by RNase protection assay. Note the complete inhibition of baseline and IFN-α-induced TRAIL expression in the UM-UC-5 cells exposed to bortezomib (compare lanes 1 and 2 with lanes 3 and 4). Also note that UM-UC-3 cells produced no detectable TRAIL mRNA under any condition. B, time course analysis of bortezomib-mediated inhibition of TRAIL. Cells were preincubated with bortezomib for the times indicated. Cells were then incubated with 104 units/mL IFN-α for 8 h, mRNA was isolated, and levels of TRAIL expression were quantified by RNase protection assay. C, time-dependent effects of bortezomib on TRAIL protein expression. UM-UC-5 cells were incubated with IFN-α, bortezomib, or a combination of the two for the times indicated. TRAIL protein expression was then measured in cell lysates (L) and conditioned medium (CM) by ELISA.

Close modal

Effects of IFN-α and Bortezomib on Angiogenic Factor Production

Previous studies showed that IFN-α inhibits angiogenesis in human bladder cancer cells by blocking tumor cell production of proangiogenic factors (bFGF and IL-8; ref. 13). Similarly, we have shown that bortezomib can inhibit angiogenesis in several solid tumor models, including bladder cancer (10, 14, 15). Therefore, we characterized the effects of IFN, bortezomib, or a combination of the two on the production of the proangiogenic factors, bFGF, VEGF, and IL-8, by the UM-UC-3 and UM-UC-5 cells. Exposure to either IFN-α or bortezomib alone resulted in down-regulation of bFGF and VEGF in the UM-UC-3 cells, and exposure to the combination resulted in further inhibition of factor production (Fig. 3A and B). In contrast, neither single agent nor a combination of the two down-regulated bFGF or VEGF production in the UM-UC-5 cells (Fig. 3A and B). Bortezomib stimulated increases in IL-8 secretion in both cell lines whether it was combined with IFN-α or not (Fig. 3C).

Figure 3.

Effects of IFN and bortezomib on angiogenic factor production. A, effects of IFN-α and bortezomib on bFGF expression. Cells were preincubated with 104 units/mL IFN-α for 24 h and 100 nmol/L bortezomib for an additional 24 h, and bFGF levels were measured in the conditioned medium by ELISA. Results are expressed as percentages of controls. Note that control UM-UC-3 cells secreted 146 pg/mL bFGF per 106 cells, whereas control UM-UC-5 cells secreted 56 pg/mL per 106 cells. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. B, effects of IFN and bortezomib on VEGF expression. UM-UC-3 and UM-UC-5 cells were incubated with 104 units/mL IFN-α and 100 nmol/L bortezomib as described in (A), and VEGF levels in the conditioned medium were measured by ELISA. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. Results are expressed as percentages of controls. Note that control UM-UC-3 cells secreted 10.7 ng/mL VEGF per 106 cells, whereas UM-UC-5 cells secreted 54 ng/mL VEGF per 106 cells. C, effects on IL-8 expression. UM-UC-3 (top) and UM-UC-5 (bottom) cells were incubated with IFN and bortezomib as described in (A), and IL-8 levels in the conditioned medium were measured by ELISA. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. Results are expressed as percentages of controls. Note that control UM-UC-3 cells secreted 83.5 ng/mL IL-8 per 106 cells, whereas control UM-UC-5 cells secreted 6.9 ng/mL per 106 cells.

Figure 3.

Effects of IFN and bortezomib on angiogenic factor production. A, effects of IFN-α and bortezomib on bFGF expression. Cells were preincubated with 104 units/mL IFN-α for 24 h and 100 nmol/L bortezomib for an additional 24 h, and bFGF levels were measured in the conditioned medium by ELISA. Results are expressed as percentages of controls. Note that control UM-UC-3 cells secreted 146 pg/mL bFGF per 106 cells, whereas control UM-UC-5 cells secreted 56 pg/mL per 106 cells. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. B, effects of IFN and bortezomib on VEGF expression. UM-UC-3 and UM-UC-5 cells were incubated with 104 units/mL IFN-α and 100 nmol/L bortezomib as described in (A), and VEGF levels in the conditioned medium were measured by ELISA. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. Results are expressed as percentages of controls. Note that control UM-UC-3 cells secreted 10.7 ng/mL VEGF per 106 cells, whereas UM-UC-5 cells secreted 54 ng/mL VEGF per 106 cells. C, effects on IL-8 expression. UM-UC-3 (top) and UM-UC-5 (bottom) cells were incubated with IFN and bortezomib as described in (A), and IL-8 levels in the conditioned medium were measured by ELISA. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls. Results are expressed as percentages of controls. Note that control UM-UC-3 cells secreted 83.5 ng/mL IL-8 per 106 cells, whereas control UM-UC-5 cells secreted 6.9 ng/mL per 106 cells.

Close modal

Effects of IFN and Bortezomib on Tumor Growth

We treated mice bearing established UM-UC-3 or UM-UC-5 tumors with IFN-α, bortezomib, or the combination twice weekly for up to 3 weeks and monitored the effects of therapy on tumor growth and therapy-induced changes in relevant biomarkers. The s.c. UM-UC-5 tumors were allowed to grow for 14 days before randomization and treated with 104 units IFN-α with or without 1 mg/kg bortezomib for 3 weeks. The faster-growing s.c. UM-UC-3 tumors were established for 9 days before randomization, and therapy was done with 106 units IFN-α plus 1 mg/kg bortezomib for 2 weeks, at which point the control animals became moribund. To mimic the optimal in vitro schedule, we dosed the animals with IFN-α 24 h before dosing them with bortezomib. All of the tumor-bearing animals maintained their body weights throughout the course of therapy, and no other signs of systemic toxicity were observed (data not shown; also see discussion below). Single-agent therapy with either IFN-α or bortezomib had no significant effect on tumor growth, whereas combination therapy with IFN-α plus bortezomib resulted in significant growth inhibition in both models (P < 0.05 versus the other groups; Fig. 4A and B).

Figure 4.

Effects of therapy with IFN-α plus bortezomib on tumor growth. A, effects on s.c. UM-UC-3 xenografts. Mice bearing established (9 d) s.c. tumors were treated with 106 units/mL IFN-α, 1 mg/kg bortezomib, or both for 2 wks as described in Materials and Methods. Tumor volumes were measured weekly using a caliper. Points, mean (n = 10); bars, SD. B, effects on s.c. UM-UC-5 xenografts. Mice bearing established (14 d) s.c. tumors were treated with 104 units IFN-α, 1 mg/kg bortezomib, or both as described in Materials and Methods. Points, mean (n = 10); bars, SD.

Figure 4.

Effects of therapy with IFN-α plus bortezomib on tumor growth. A, effects on s.c. UM-UC-3 xenografts. Mice bearing established (9 d) s.c. tumors were treated with 106 units/mL IFN-α, 1 mg/kg bortezomib, or both for 2 wks as described in Materials and Methods. Tumor volumes were measured weekly using a caliper. Points, mean (n = 10); bars, SD. B, effects on s.c. UM-UC-5 xenografts. Mice bearing established (14 d) s.c. tumors were treated with 104 units IFN-α, 1 mg/kg bortezomib, or both as described in Materials and Methods. Points, mean (n = 10); bars, SD.

Close modal

Effects of Combination Therapy on TRAIL Expression and Apoptosis

We used immunofluorescence staining and laser scanning cytometry analysis to quantify TRAIL expression in tumor sections (Fig. 5A). These results were confirmed by measuring human TRAIL expression in isolated tumor extracts by commercial ELISA (Fig. 5B). For these experiments, we compared the effects of therapy in tumors that were harvested after only 1 week of therapy (i.e., 24 h after the second dose of bortezomib) with those observed in tumors harvested at the experimental end point so that direct and indirect effects of the drugs could be compared. Overall, the effects of therapy on TRAIL expression at both time points in vivo were very similar to the effects of IFN-α and bortezomib in vitro. Specifically, IFN-α stimulated significant increases in TRAIL expression in the UM-UC-5 tumors but had no effect in the UM-UC-3 tumors. TRAIL levels in the UM-UC-5 tumors exposed to IFN plus bortezomib were similar to controls, confirming that bortezomib rapidly (i.e., within 24 h) inhibited tumor TRAIL production in vivo.

Figure 5.

Effects of therapy on TRAIL expression. A, quantification by laser scanning cytometry. UM-UC-5 tumors were harvested after 1 or 3 wks of therapy. Tumor sections were stained with a polyclonal anti-TRAIL antibody or an isotype-matched irrelevant control, and TRAIL expression was measured by immunofluorescence and laser scanning cytometry as described in Materials and Methods. Mean fluorescence intensities for each condition are indicated. B, quantification by ELISA. Extracts prepared from snap-frozen tumors treated with IFN-α, bortezomib, or a combination of the two for 1 wk were used to quantify TRAIL protein expression using a commercial ELISA kit. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls.

Figure 5.

Effects of therapy on TRAIL expression. A, quantification by laser scanning cytometry. UM-UC-5 tumors were harvested after 1 or 3 wks of therapy. Tumor sections were stained with a polyclonal anti-TRAIL antibody or an isotype-matched irrelevant control, and TRAIL expression was measured by immunofluorescence and laser scanning cytometry as described in Materials and Methods. Mean fluorescence intensities for each condition are indicated. B, quantification by ELISA. Extracts prepared from snap-frozen tumors treated with IFN-α, bortezomib, or a combination of the two for 1 wk were used to quantify TRAIL protein expression using a commercial ELISA kit. Columns, mean (n = 3); bars, SD. *, P < 0.01 versus controls.

Close modal

The effects of therapy on apoptosis were also very similar in vitro and at 1 week after the initiation of therapy in vivo. In particular, therapy with either IFN-α or bortezomib alone had little to no effect on apoptosis in either tumor model as measured by TUNEL staining or immunofluorescent detection of active caspase-3 (Fig. 6A and B). However, therapy with a combination of the two resulted in dramatic increases in apoptosis in the UM-UC-5 tumors but had little effect in the UM-UC-3 tumors (Fig. 6A and B). Nonetheless, by the experimental end point, the effects of therapy on apoptosis in vivo no longer correlated as well with the in vitro findings, in that levels of apoptosis were significantly elevated in UM-UC-3 tumors exposed to IFN-α plus bortezomib as measured by TUNEL staining or immunofluorescent detection of active caspase-3 (Fig. 6C and D).

Figure 6.

Effects of therapy on apoptosis. A, acute effects on DNA fragmentation. Sections from tumors harvested after 1 wk of therapy were stained by fluorescent TUNEL, and percentages of apoptotic cells were quantified manually. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. B, acute effects on caspase activation. Sections from tumors harvested after 1 wk of therapy were stained with an antibody specific for activated caspase-3, and percentages of positive cells were quantified manually. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. C, chronic effects on DNA fragmentation. Sections from tumors harvested after 3 wks of therapy were stained by fluorescent TUNEL as described in Materials and Methods, and percentages of positive cells were quantified manually. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. D, chronic effects on caspase-3 activation. Sections from tumors harvested after 3 wks of therapy were stained with an antibody specific for activated caspase-3, and percentages of apoptotic cells were quantified manually. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls.

Figure 6.

Effects of therapy on apoptosis. A, acute effects on DNA fragmentation. Sections from tumors harvested after 1 wk of therapy were stained by fluorescent TUNEL, and percentages of apoptotic cells were quantified manually. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. B, acute effects on caspase activation. Sections from tumors harvested after 1 wk of therapy were stained with an antibody specific for activated caspase-3, and percentages of positive cells were quantified manually. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. C, chronic effects on DNA fragmentation. Sections from tumors harvested after 3 wks of therapy were stained by fluorescent TUNEL as described in Materials and Methods, and percentages of positive cells were quantified manually. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. D, chronic effects on caspase-3 activation. Sections from tumors harvested after 3 wks of therapy were stained with an antibody specific for activated caspase-3, and percentages of apoptotic cells were quantified manually. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls.

Close modal

Effects of Therapy on Angiogenesis

We stained tissue sections with an anti-CD31 antibody (Fig. 7A) and quantified tumor microvessel densities manually (Fig. 7B and C). Reductions in microvessel density were observed in tumors treated with IFN-α plus bortezomib in both models at the experimental end point (Fig. 7C). However, UM-UC-3 tumors treated with IFN-α plus bortezomib already displayed significant angiogenesis inhibition by 1 week after therapy, whereas UM-UC-5 tumors did not (Fig. 7B). In addition, therapy with IFN-α, bortezomib, or especially a combination of the two agents down-regulated VEGF expression in the UM-UC-3 tumors but had no detectable effect on VEGF production in the UM-UC-5 tumors (Fig. 7D). Furthermore, UM-UC-3 tumors exposed to therapy with either single agent alone and especially the combination displayed extensive central necrosis as measured by H&E staining and densitometry, whereas UM-UC-5 tumors did not (Fig. 7E). Because central necrosis is also a consistent feature of effective antiangiogenic therapy (10), our results strongly suggest that combination therapy with IFN-α plus bortezomib attenuates tumor growth via angiogenesis inhibition in the UM-UC-3 model. We suspect that the increased apoptosis observed in the UM-UC-3 tumors at the experimental end point occurred as a secondary consequence of angiogenesis inhibition.

Figure 7.

Effects of therapy on angiogenesis. A, representative anti-CD31 staining. Microvessels were detected by CD31 immunohistochemistry as described in Materials and Methods. B, acute effects of therapy on angiogenesis. Microvessel densities (MVD) were quantified manually in tumors exposed to therapy for 1 week as described in Materials and Methods. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. C, chronic effects of therapy on angiogenesis. Microvessel densities were quantified manually in tumors exposed to therapy for 3 weeks as described in Materials and Methods. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. D, analysis of VEGF expression. VEGF was detected by immunohistochemistry as described in Materials and Methods. Representative images from each condition. E, effects of therapy on necrosis. Necrotic areas were detected by H&E staining and measured by optical imaging. Columns, mean (n = 10); bars, SD. *, P < 0.01 versus controls.

Figure 7.

Effects of therapy on angiogenesis. A, representative anti-CD31 staining. Microvessels were detected by CD31 immunohistochemistry as described in Materials and Methods. B, acute effects of therapy on angiogenesis. Microvessel densities (MVD) were quantified manually in tumors exposed to therapy for 1 week as described in Materials and Methods. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. C, chronic effects of therapy on angiogenesis. Microvessel densities were quantified manually in tumors exposed to therapy for 3 weeks as described in Materials and Methods. Columns, mean (n = 10); bars, SD. *, P < 0.05 versus controls. D, analysis of VEGF expression. VEGF was detected by immunohistochemistry as described in Materials and Methods. Representative images from each condition. E, effects of therapy on necrosis. Necrotic areas were detected by H&E staining and measured by optical imaging. Columns, mean (n = 10); bars, SD. *, P < 0.01 versus controls.

Close modal

Evaluation of Systemic Toxicity

The effects of IFNs are species specific. Therefore, recombinant human IFN-α does not directly affect mouse tissues. To evaluate the potential systemic toxicity of combination therapy with IFN-α plus bortezomib, we treated mice with recombinant murine IFN-α (104 units/injection) plus bortezomib (1 mg/kg) for 3 weeks using the same staggered schedule we used in the tumor therapy studies. We observed no significant weight loss in any of the groups (data not shown). Furthermore, direct analysis of tissue histology by H&E staining and active caspase-3 and TUNEL analysis revealed no changes in any of the major organs (heart, lungs, kidney, and liver; data not shown).

In a previous study, we showed that IFN-α induces TRAIL mRNA and protein production in the majority of human transitional cell carcinoma cell lines (6). Here, we investigated whether a potent TRAIL-sensitizing agent would increase the antitumoral effects of IFN-α in cells that produce large amounts of TRAIL (UM-UC-5) or no TRAIL (UM-UC-3) in response to IFN-α exposure. Our results confirm that bortezomib enhances apoptosis and inhibits growth in UM-UC-5 tumors, but it also enhanced the growth-inhibitory effects of IFN-α in UM-UC-3 tumors, although the cells did not display much direct induction of apoptosis (Fig. 6A and B). Rather, growth inhibition in the UM-UC-3 tumors was associated with down-regulation of VEGF, inhibition of angiogenesis (as measured by quantifying microvessel densities), and the development of extensive necrosis. Therefore, our results strongly suggest that combination therapy with IFN-α plus bortezomib will have antitumor activity in bladder cancer irrespective of whether IFN-α up-regulates TRAIL expression and induces apoptosis within the tumor.

Another important finding was that the effects of bortezomib on IFN-α-induced cell death in the UM-UC-5 tumors were schedule dependent. Optimal induction of apoptosis was only observed when tumor cells were exposed to IFN-α before bortezomib. The sequence dependence seemed to be related to the ability of bortezomib to block IFN-induced TRAIL mRNA and protein expression. In a parallel study, we have found that IFN-α-induced TRAIL expression in human bladder cancer cells is dependent on combined binding of two well-known, IFN-α-stimulated transcription factors (signal transducers and activators of transcription 1 and IFN regulatory factor-1) to specific elements within the TRAIL promoter,4

4

A. Papageorgiou, C. Dinney, and D.J. McConkey, manuscript submitted.

and a previous study showed that IFN-α induces TRAIL via a signal transducers and activators of transcription 1–dependent and IFN regulatory factor-1–dependent mechanisms in other models (16, 17). Although we did observe some inhibition of IFN regulatory factor-1 induction in cells exposed to bortezomib (data not shown), the effects of bortezomib on IFN regulatory factor-1 were not nearly as pronounced as its effects on baseline and IFN-α-induced TRAIL mRNA levels, strongly suggesting that additional mechanisms are involved.

Although we predicted that combination therapy with IFN-α plus bortezomib would have different effects on apoptosis in the UM-UC-3 and UM-UC-5 cells based on our knowledge of their different TRAIL expression patterns, we were surprised that the combination had such divergent effects on bFGF and VEGF production and angiogenesis inhibition in the two models. In several previous studies, we and others have reported that bortezomib blocks VEGF production and angiogenesis in diverse tumor models (10, 15, 18, 19), including bladder cancer (14), and in ongoing studies, we and others have linked these effects to bortezomib-mediated inhibition of hypoxia-inducible factor-1 (20).5

5

K. Zhu et al., in preparation.

Similarly, previous studies in a variety of different models (bladder cancer, prostate cancer, melanoma, and childhood hemangioma; ref. 21) have shown that IFN-α is a potent inhibitor of angiogenesis. Thus, within this context, the resistance of the UM-UC-5 cells to angiogenesis inhibition is surprising. The cells must use different molecular pathways to drive constitutive VEGF and bFGF production that are insensitive to IFN-α or bortezomib. One excellent candidate is autocrine epidermal growth factor receptor activation because the epidermal growth factor receptor is highly expressed by the UM-UC-5 cells (22)6
6

Shrader et al., Mol Cancer Ther, in press.

and epidermal growth factor receptor antagonists are potent antagonists of VEGF production and angiogenesis in the cells (22).6

We also do not have an explanation for why bortezomib up-regulated tumor cell IL-8 production in both models examined here, but this too is surprising given the ability of bortezomib to block nuclear factor-κB activation (23) coupled with the critical role nuclear factor-κB plays in controlling IL-8 expression in bladder cancer (24). Indeed, bortezomib did block IL-8 production in another human bladder cancer xenograft model (253J B-V; ref. 14). We did not directly test the effect of IL-8 up-regulation on the growth of the UM-UC-3 or UM-UC-5 tumors, but given the important role of IL-8 in driving endothelial cell migration and angiogenesis, understanding the molecular basis for the effects of bortezomib could be important for optimization of therapeutic efficacy. For example, excellent blocking anti-IL-8 antibodies are available, which could be used to neutralize the effects of IL-8 up-regulation in this subset of tumors (25). With respect to clinical translation of our findings, it seems that there are clear strengths and weaknesses associated with combining bortezomib with IFN-α in patients. Given that IFNs induce apoptosis via induction of TRAIL, it might seem more straightforward to develop TRAIL as opposed to IFN-α for bladder cancer therapy because bortezomib suppresses TRAIL expression. However, IFN-α and bortezomib are both clinically approved, whereas TRAIL and antibodies that activate the receptors of TRAIL are just entering phase II clinical trials (7). Furthermore, by adopting an optimal scheduling protocol, the beneficial effects of IFN-α on both TRAIL production and angiogenesis inhibition can be exploited, and our preliminary toxicity studies indicate that combination therapy with IFN-α plus bortezomib is very well tolerated. We plan to do head-to-head comparisons of these combinations in appropriate models. We also plan to compare the effects of bortezomib with those of other candidate TRAIL-sensitizing agents (conventional chemotherapy, histone deacetylase inhibitors, and SMAC mimetics) because some of these agents probably do not share the negative effects of bortezomib on TRAIL expression.

Grant support: M.D. Anderson Specialized Program of Research Excellence in Bladder Cancer grant P50 CA91846, Project 3.

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.

1
Lamm DL. Bladder cancer: twenty years of progress and the challenges that remain.
CA Cancer J Clin
1998
;
48
:
263
–8.
2
Cote RJ, Datar RH. Therapeutic approaches to bladder cancer: identifying targets and mechanisms.
Crit Rev Oncol Hematol
2003
;
46
Suppl:
S67
–83.
3
Kamat AM, Lamm DL. Immunotherapy for bladder cancer.
Curr Urol Rep
2001
;
2
:
62
–9.
4
Dinney CP, McConkey DJ, Millikan RE, et al. Focus on bladder cancer.
Cancer Cell
2004
;
6
:
111
–6.
5
Ludwig AT, Moore JM, Luo Y, et al. Tumor necrosis factor-related apoptosis-inducing ligand: a novel mechanism for Bacillus Calmette-Guerin-induced antitumor activity.
Cancer Res
2004
;
64
:
3386
–90.
6
Papageorgiou A, Lashinger L, Millikan R, et al. Role of tumor necrosis factor-related apoptosis-inducing ligand in interferon-induced apoptosis in human bladder cancer cells.
Cancer Res
2004
;
64
:
8973
–9.
7
Cretney E, Shanker A, Yagita H, Smyth MJ, Sayers TJ. TNF-related apoptosis-inducing ligand as a therapeutic agent in autoimmunity and cancer.
Immunol Cell Biol
2006
;
84
:
87
–98.
8
Lashinger LM, Zhu K, Williams SA, Shrader M, Dinney CP, McConkey DJ. Bortezomib abolishes tumor necrosis factor-related apoptosis-inducing ligand resistance via a p21-dependent mechanism in human bladder and prostate cancer cells.
Cancer Res
2005
;
65
:
4902
–8.
9
Adams J, Palombella VJ, Sausville EA, et al. Proteasome inhibitors: a novel class of potent and effective antitumor agents.
Cancer Res
1999
;
59
:
2615
–22.
10
Nawrocki ST, Bruns CJ, Harbison MT, et al. Effects of the proteasome inhibitor PS-341 on apoptosis and angiogenesis in orthotopic human pancreatic tumor xenografts.
Mol Cancer Ther
2002
;
1
:
1243
–53.
11
Nawrocki ST, Sweeney-Gotsch B, Takamori R, McConkey DJ. The proteasome inhibitor bortezomib enhances the activity of docetaxel in orthotopic human pancreatic tumor xenografts.
Mol Cancer Ther
2004
;
3
:
59
–70.
12
Ruiz S, Krupnik Y, Keating M, Chandra J, Palladino M, McConkey D. The proteasome inhibitor NPI-0052 is a more effective inducer of apoptosis than bortezomib in lymphocytes from patients with chronic lymphocytic leukemia.
Mol Cancer Ther
2006
;
5
:
1836
–43.
13
Slaton JW, Perrotte P, Inoue K, Dinney CP, Fidler IJ. Interferon-α-mediated down-regulation of angiogenesis-related genes and therapy of bladder cancer are dependent on optimization of biological dose and schedule.
Clin Cancer Res
1999
;
5
:
2726
–34.
14
Kamat AM, Karashima T, Davis DW, et al. The proteasome inhibitor bortezomib synergizes with gemcitabine to block the growth of human 253JB-V bladder tumors in vivo.
Mol Cancer Ther
2004
;
3
:
279
–90.
15
Williams S, Pettaway C, Song R, Papandreou C, Logothetis C, McConkey DJ. Differential effects of the proteasome inhibitor bortezomib on apoptosis and angiogenesis in human prostate tumor xenografts.
Mol Cancer Ther
2003
;
2
:
835
–43.
16
Clarke N, Jimenez-Lara AM, Voltz E, Gronemeyer H. Tumor suppressor IRF-1 mediates retinoid and interferon anticancer signaling to death ligand TRAIL.
EMBO J
2004
;
23
:
3051
–60.
17
Choi EA, Lei H, Maron DJ, et al. Stat1-dependent induction of tumor necrosis factor-related apoptosis-inducing ligand and the cell-surface death signaling pathway by interferon β in human cancer cells.
Cancer Res
2003
;
63
:
5299
–307.
18
Sunwoo JB, Chen Z, Dong G, et al. Novel proteasome inhibitor PS-341 inhibits activation of nuclear factor-κB, cell survival, tumor growth, and angiogenesis in squamous cell carcinoma.
Clin Cancer Res
2001
;
7
:
1419
–28.
19
LeBlanc R, Catley LP, Hideshima T, et al. Proteasome inhibitor PS-341 inhibits human myeloma cell growth in vivo and prolongs survival in a murine model.
Cancer Res
2002
;
62
:
4996
–5000.
20
Kaluz S, Kaluzova M, Stanbridge EJ. Proteasomal inhibition attenuates transcriptional activity of hypoxia-inducible factor 1 (HIF-1) via specific effect on the HIF-1α C-terminal activation domain.
Mol Cell Biol
2006
;
26
:
5895
–907.
21
Fidler IJ. Regulation of neoplastic angiogenesis.
J Natl Cancer Inst Monogr
2001
;
28
:
10
–4.
22
Kassouf W, Dinney CP, Brown G, et al. Uncoupling between epidermal growth factor receptor and downstream signals defines resistance to the antiproliferative effect of Gefitinib in bladder cancer cells.
Cancer Res
2005
;
65
:
10524
–35.
23
Adams J. The development of proteasome inhibitors as anticancer drugs.
Cancer Cell
2004
;
5
:
417
–21.
24
Karashima T, Sweeney P, Kamat A, et al. Nuclear factor-κB mediates angiogenesis and metastasis of human bladder cancer through the regulation of interleukin-8.
Clin Cancer Res
2003
;
9
:
2786
–97.
25
Mian BM, Dinney CP, Bermejo CE, et al. Fully human anti-interleukin 8 antibody inhibits tumor growth in orthotopic bladder cancer xenografts via down-regulation of matrix metalloproteases and nuclear factor-κB.
Clin Cancer Res
2003
;
9
:
3167
–75.