Bladder cancer is the 4th most common type of cancer in men and the 10th most common in women and accounted for approximately 12,500 deaths in the United States in the year 2003.1 Whereas 70% of newly diagnosed bladder cancer is superficial, the remaining 30% are invasive or metastatic (Sternberg et al., “Bladder Cancer,” www.emedicine.com). Invasive, locally advanced and metastatic bladder cancer remain therapeutic challenges where cystectomy, systemic chemotherapy, radiation treatment, or a combination of these comprise the spectrum of interventions.
A review of current treatment modalities2, 3 highlights deficiencies where alternative therapies may hold promise. Radical cystectomy is the treatment of choice for muscle-invasive bladder cancer; however, only half of these patients will be cured by cystectomy alone.4 Metastatic bladder cancer is usually treated with systemic chemotherapy such as MVAC (methotrexate, vinblastine, adriamycin, cisplatin), which has been the standard combination regimen for nearly 2 decades. For locally advanced and metastatic bladder cancer, long-term survival is rare. On average, survival for patients receiving MVAC is less than 14 months, and a rare minority (< 4%) remains disease-free at 6 years.5 More recently, newer agents, such as gemcitabine, have emerged with a single-agent response rate of approximately 25%.6 Clearly, other agents are needed to decrease toxicity and augment response rates of existing regimens.
Identification of the molecular events underlying urothelial neoplastic progression raises the possibility of targeted therapy in late-stage bladder cancer. In this regard, the tumor suppressor genes p53 and Rb represent potential targets since they are frequently found altered in this patient cohort.7, 8 Accompanying such molecular changes are alterations associated with the cadherin/catenin complex, including the reduction or loss of E-cadherin and plakoglobin expression. Each of these events has been reported as an indicator of poor survival in bladder cancer patients.9, 10 Loss of gene expression or function can occur via multiple pathways, including mutation, deletion, gene loss, gene silencing and protein phosphorylation events. Alterations associated with the expression of different members of cadherin/catenin complex in bladder cancer can also occur as a result of epigenetic events affecting gene regulation.11, 12
To date, evidence for epigenetic events associated with tumorigenesis centers around DNA methylation and histone acetylation. Histone deacetylase inhibitors (HDACis) cause hyperacetylation of histones, resulting in transcriptional activation and deactivation of a subset of genes.13, 14, 15, 16, 17 Sodium butyrate (NaB) is a relatively nonspecific short-chain fatty acid produced by anaerobic bacterial fermentation of dietary fibers18 and is a noncompetitive inhibitor of HDAC whose exact mechanism is not known. Trichostatin A (TSA), originally identified as an antifungal agent, also acts as a noncompetitive inhibitor of HDAC by mimicking the lysine substrate as well as chelating a zinc atom crucial for enzymatic activity. Different members of this class of agents have been found to inhibit proliferation, cause cell cycle arrest, induce differentiation and/or apoptosis in transformed cells.15, 19, 20, 21, 22 Numerous HDACis have demonstrated antitumor activity in vivo in a variety of solid tumors.23, 24 A number of these agents have undergone evaluation in phase 1 and 2 clinical trials.25, 26, 27, 28, 29 In this report, we have characterized the therapeutic potential of 2 HDACis within a panel of bladder carcinoma cell lines in vitro and demonstrated the antitumor activity of TSA in a bladder cancer xenograft model. Following exposure of cells to TSA and NaB, we have identified upregulation of plakoglobin expression, a putative tumor suppressor gene, in bladder carcinoma cell lines representing late-stage disease.
The human bladder carcinoma cell lines J82, T24, UM-UC-3, EJ, TCCSUP and KK47 were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 7.5% fetal bovine serum and penicillin/streptomycin.
A stock solution of trichostatin A (1 mg/ml; Sigma, St. Louis, MO), dissolved in dimethyl sulfoxide (DMSO), was aliquoted and stored at −20°C. A stock solution of sodium butyrate (1 M; Sigma) was constituted fresh in sterile water at each experimental time point. The effect of each of these HDACis on the viability of bladder carcinoma cells was determined by cell counting using a hemocytometer. Cells were seeded at a density of 5 × 105 cells per 60 mm dish in DMEM supplemented with 7.5% serum and allowed to attach overnight. After 24 hr, the medium was replaced daily with fresh medium containing an appropriate concentration of the HDACi. Cells were counted at various intervals after the start of treatment. At each time point, cell viability was assessed by Trypan blue exclusion. All experiments were performed in triplicate.
Subconfluent dishes of cells were washed in phosphate-buffered saline (PBS) followed by lysis in hot sample buffer (2 × ESB: 0.08 M Tris, pH 6.8, 0.07 M SDS, 10% glycerol, 0.001% bromophenol blue) and sheared through a 26 gauge needle. Lysates were then assayed for protein concentration using the BSA method (Pierce, Rockford, IL). After determination of protein content, β-mercaptoethanol (1%) was added to each sample. Samples were boiled for 5 min and protein was loaded in each lane of a 7.5% or 12.5% polyacrylamide gel. Proteins were transferred overnight onto nitrocellulose. Membranes were blocked in 10% milk in TBS with 0.05% Tween (TBST) and placed on primary antibody overnight at 4°C. Antibodies to plakoglobin, p120ctn, α- and β-catenin (Transduction Labs, Lexington, KY), acetylated histone H4 (Cell Signaling Technology, Beverly, MA), p21waf1 (Novus Biologicals, Littleton, CO), p14ARF and p27kip (Oncogene Research, Cambridge, MA) were used in Western blot analysis of cell and tumor lysates. Following incubation with the primary antibody, blots were washed in TBST 3 times for 15 min each, and secondary antibody linked to horseradish peroxidase was incubated with the blots for 1 hr at room temperature. Blots were then washed as described above and developed with an ECL kit (Amersham, Arlington Heights, IL). Films were scanned and densitometric analysis was performed with the Scion Image Program.
Total RNA from bladder carcinoma cell lines treated with sodium butyrate (5 mM) or TSA (0.33 μm) was isolated at hourly intervals following initial exposure. Cells were lysed in UltraSpec RNA reagent (Biotecx Labs, Houston, TX) and extracted with chloroform. RNA was precipitated using isopropyl alcohol, air-dried and resuspended in DEPC water. cDNA was prepared from 3 μg of total RNA by incubating with 100 ng/ml oligo dT at 70°C for 10 min followed by incubation at 42°C for 1 hr with 0.5 mmol deoxyribonucleoside triphosphate, 1,000 U RNasin, 0.02 mmol dithiothreitol and 2,500 U Superscript II (Invitrogen, Carlsbad, CA). Multiplex PCR was performed using the following primers: p21waf1 forward 5′-CTGGAGACTCTCAGGGTCAAA-3′, reverse 5′-AGGGTATGTACATGAGGAGGT-3′; p27kip forward 5′-CGCTTTGTTTTGTTCGGTTT-3′, reverse 5′-TCTCTGCAGTGCTTCTCCAA-3′; p14ARF forward 5′-GGTTTTCGTGGTTCACATCCCGCG-3′, reverse 5′-CAGGAAGCCCTCCCGGGCAGC-3′; glyceraldehyde-3-phosphate dehydrogenase (GAPDH) forward 5′-CAGCCGAGCCACATCG-3′, reverse 5′-TGAGGCTGTTGTCATACTTCTC-3′. After initial denaturation at 95°C for 5 min, 35 cycles were performed at 94°C for 60 sec, 52°C for 60 sec and 72°C for 60 sec. The last cycle was followed by a 5-min extension at 72°C.
Establishment of bladder cancer xenografts and treatment with HDACis
The human bladder carcinoma cell lines EJ and UM-UC-3 have both been shown to be tumorigenic in this model and were considered representative of late-stage bladder lesions. Cells were harvested and resuspended in PBS. Six-week-old male BALB/c nude (nu/nu) mice were injected subcutaneously with inoculums of 1 × 106 cells per site for UM-UC-3 and 5 × 106 for EJ cells. Mice were kept under barrier conditions and maintained on a 12-hr day/night light cycle with food and water ad libitum. Mice were randomly assigned to 4 groups of 10 animals and tumors were allowed to grow to approximately 3 mm3. At this stage, tumor-bearing mice were treated with TSA (500 μg/kg) on a daily basis via subcutaneous injection. Care was taken to institute a rotation for the site of inoculation for each mouse throughout the study. Control animals followed the same regimen using DMSO vehicle. Daily inoculums of TSA and DMSO were delivered in a volume of 50 μl (stock solution 0.83 mM TSA). Mice were weighed weekly and animal behavior and food intake was monitored throughout the course of the experiment to assess the toxicity of the treatment. Tumors were measured at the same time using calipers and tumor volume was calculated using the following formula: tumor volume = length × width × height × π/6. The treatment period was 5 weeks for EJ tumors and 3 weeks for UM-UC-3 tumors, reflecting the different growth rates of the cell lines in vivo. The study period was determined by the size of the tumors in the control group, 6 weeks for EJ and 4 weeks for UM-UC-3. In vivo data were presented as mean tumor volume ± standard error of the mean. p-values at the end of the experiment were established using the Student's t-test.
Immunocytochemical staining was performed on an automated immunocytochemical processor (Ventana ES, Medical Systems, Tucson, AZ). Plakoglobin and acetylated histone 4 antibody were used at dilutions of 1:500 and 1:50, respectively. Cells were plated at subconfluent levels on sterile glass slides and allowed to attach and spread over a 24-hr period. TSA or DMSO vehicle was added to the medium, and cells were maintained for a further 18 hr. Cells were fixed in methanol (−20°C) for 10 min, washed in 3 changes of PBS and processed on the Ventana ES processor.
Effect of TSA and NaB on cancer cell growth in vitro
Figure 1 shows that NaB and TSA cause a concentration-dependent growth inhibition of all the bladder carcinoma cells included in this study panel. Fifty percent growth inhibition (IC50) was recorded within a range of 1.5–5.0 mM for NaB and 0.06–0.33 μM for TSA. Trypan blue exclusion, performed on floating cells and cells remaining attached to dishes after drug exposure, revealed > 99% loss of viability in the former, with only 60% of attached cells retaining the ability to exclude the dye. Cell viability from these experiments was further confirmed in colony formation assays (data not shown).
Having established concentration-dependent growth inhibition across a range of drug concentrations, we exposed bladder carcinoma cells to a continuous single concentration of drug representing the IC50 of the most resistant cell line as determined by Figure 1. Cells were harvested and counted at days 1, 2 and 3, with the results presented as a percentage of the control cell population (Fig. 2). Exposure to TSA and to NaB produced at least 50% cell death in all cell lines as assessed by Trypan blue exclusion. Fifty percent killing was attained by 48 hr in most, and by 72 hr in all cell lines exposed to either drug. Both NaB and TSA caused a progressive decline in cell number that was sustained across the experimental period in all cell lines.
Effect of TSA and NaB on cell cycle regulators
HDAC is have been reported to modulate the expression of cell cycle-regulatory proteins, including the cyclin-dependent kinase inhibitor p21waf1. To establish similar events in bladder carcinoma cell lines, we investigated the effects of TSA (0.33 μM) and NaB (5 mM) on the transcription and translation of a few such regulatory proteins. Using semiquantitative RT-PCR, we investigated the p21waf1 transcript expression levels at different intervals following the addition of HDACis. In the 6 bladder carcinoma cell lines tested, all displayed increased p21waf1 transcripts over time. In each case, multiplex analysis with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served to standardize mRNA levels over the course of the experiment. In replica experiments, we recorded no modulation of p27kip or p14ARF expression over the same time period (data not shown).
Paralleling these experiments with Western blot analysis corroborated RT-PCR findings and revealed no change in p27kip or p14ARF protein expression throughout the course of the experiment. Upregulation of p21waf1 protein was detected in cell lines UM-UC-3, T24 and J82 following exposure to HDACis (Fig. 3a). No increase in p21waf1 protein levels was detected in cell lines EJ, TCCSUP and KK47 in repeat experiments, despite detection of elevated levels of p21waf1 transcript in RT-PCR following exposure of cells to HDACis (Fig. 3b). To confirm elevated histone acetylation associated with exposure of cells to HDACis, we performed staining for acetylated histone 4 following an 18-hr exposure to TSA or NaB. Figure 4 shows immunohistochemical staining of acetylated H4 demonstrating increased nuclear staining in bladder carcinoma cell lines EJ and KK47 following exposure to TSA. A range of morphologic changes was recorded in all cell lines in response to exposure to HDACis.
Modulation of cadherin/catenin expression
Of the 6 bladder carcinoma cell lines used in the experimental panel, all lack E-cadherin expression and 5 display loss or reduced plakoglobin expression. Each of the cell lines is derived from a late-stage tumor and harbors molecular changes indicative of invasive bladder cancer. Figure 5 shows levels of plakoglobin expression following 48-hr exposure to different concentrations of either NaB or TSA. Both HDACis increased plakoglobin expression in a concentration-dependent manner. Some of the cell lines exhibited a steep decline in detectable plakoglobin levels at higher concentrations (e.g., T24, J82 treated with 0.99 and 1.65 μm TSA). Results from earlier experiments suggest cell death accompanied by increased protein degradation explains this observation. Analysis of E-cadherin, α-, β- and p120 catenins (data not shown) in the same experiment revealed no change in expression levels. In contrast, plakoglobin was found upregulated in 5 bladder cell lines in response to concentration levels approaching the IC50 for each agent. With the exception of the cell line KK47, baseline expression of plakoglobin was low in all cell lines. Plakoglobin expression in HU456, the control cell line, was representative of levels recorded in E-cadherin expressing bladder carcinoma cell lines and normal urothelium.
Timing and duration of plakoglobin expression
With plakoglobin modulation by TSA and NaB established, we used expression levels of this catenin as a surrogate marker to explore some of the pharmacokinetic properties of HDACis in bladder carcinoma cells. To determine the timing of plakoglobin expression following the addition of HDACis, we exposed cells to either 0.33 μM TSA or 5 mM NaB for increasing intervals up to 24 hr and performed immunoblot analysis to determine the levels of plakoglobin expression (Fig. 6). Treatment with either NaB or TSA resulted in increased expression of plakoglobin following 12–24 hr of drug exposure, peaking between 12 and 18 hr.
Since the EJ bladder carcinoma cell line displayed the greatest increase in plakoglobin expression following exposure to HDACis, this line was used to study the duration of high-level plakoglobin expression following removal of TSA. Cells were exposed to TSA for 18 hr and chased in drug-free medium for different time periods before plakoglobin levels were assessed (Fig. 7a). Following drug removal, plakoglobin expression progressively decreased to baseline levels after 24 hr, displaying a decline in plakoglobin protein levels starting between 6 and 12 hr in the drug-free chase period.
We next varied the duration of HDACi exposure to EJ cells to determine the minimal exposure time necessary to demonstrate a drug effect using plakoglobin expression as an indicator. EJ cells were exposed to TSA for between 1 and 6 hr, drug was removed and cells were lysed 12 hr later (Fig. 7b). TSA exposure for as little as 2 hr was sufficient to trigger increased plakoglobin expression.
Efficacy and toxicity of TSA in bladder xenograft models
Two bladder carcinoma xenograft models were established using inoculums of EJ and UM-UC-3 cell lines. Tumors were established before animals were treated and the dose of TSA selected for treatment was within the range successfully used by others.23, 24 The necessity for a daily dosing schedule was corroborated by the aforementioned in vitro findings. Daily administration of TSA in animals bearing EJ and UM-UC-3 tumors resulted in 87% and 72% reduction (Fig. 8), respectively, in mean tumor volume compared with vehicle control-treated animals. The toxicity of TSA was assessed by mouse survival and careful monitoring of body weight, behavior and food intake throughout the experiment. No deaths occurred in the TSA or vehicle-treated arm of the study and, on average, mice gained 5.0 and 3.4 g in body weight, respectively, relative to their weights at the start of the experiment.
In this study, we have shown the potential efficacy of HDACis against bladder cancer, demonstrating the growth-inhibitory effect of TSA and NaB across a panel of bladder cell lines. Although other groups have examined the effects of HDACis on the bladder carcinoma cell line T24,16, 17 such studies have not explored the therapeutic efficacy of this class of agents in bladder cancer. Initial experiments showed that millimolar concentrations of NaB were required to inhibit cellular proliferation, compared to micromolar concentrations of TSA, consistent with findings in the literature.15, 17 We also found that although both TSA and NaB caused concentration-dependent cell death, cell lines displayed disparate susceptibility to each agent. This finding is not surprising, considering that each belongs to a separate class of HDACis.21, 30 Indeed, alternative selective inhibitors for specific HDAC subtypes have been described31, 32, 33 that may bring important specificity to the treatment of different tumor types.
Consistent with their effects on cellular proliferation, HDACis are known to induce the expression of cell cycle-regulatory proteins.15, 16, 17 Previous studies have suggested that while the degree of specificity for any particular HDAC enzyme target varies between agents, the end result of drug action is to alter the expression of a subset of genes selectively.16, 17 Estimates of the percentage of the entire genome modulated by HDACis is reported to range from 2–5%14 to 8–10%,17 depending on the agent in question and the cell line under investigation. To establish the bladder cell panel as a suitable model for studying the potential anticancer action of HDACis, we looked at a series of cell cycle-regulatory genes previously reported to be modulated by the action of HDACis. We confirmed upregulation of p21waf1 in some bladder cell lines but no change in p27kip, findings similar to previously published data.16, 17 Upregulation of p21waf1 in response to HDACis has been commonly observed in a variety of human tumor models.16, 17, 34 Our results looking at message levels were consistent with this observation. Interestingly, upregulation of p21waf1 protein expression was not detected in all bladder carcinoma cell lines following exposure to NaB or TSA. This may represent a sensitivity issue, since in subsequent experiments we have found that by using large amounts of KK47 total cell lysate in Western blots analysis, differential expression of p21waf1 can be detected following HDACi exposure, albeit at minimal levels. However, under the same conditions we have not been able to detect modulation of p21waf1 expression in EJ or TCCSUP. This suggests that induction of p21waf1 expression is not an absolute requirement for growth inhibition in bladder carcinoma cells, and a wide range of HDACi-induced p21waf1 expression levels can be encountered in different cell lines. The absence of p21 expression has been similarly shown not to impede the antiproliferative activity of HDACis in cells lacking the p21waf1 gene.35 This observation does not appear to hold true in all tumor systems, for instance, p21waf1 expression is required for NaB-induced growth inhibition in colon cancer.36
There is extensive evidence that a compromised cadherin/catenin complex contributes to neoplastic progression in many cancers, where the changes incurred can involve different cellular events resulting in loss of cadherin function. In bladder cancer, loss or reduced expression of E-cadherin and plakoglobin has been linked to poor survival.9, 10 It is noteworthy that within a panel of bladder carcinoma cells, with the exception of KK47, those expressing N-cadherin in the absence of E-cadherin display low levels of plakoglobin. A survey of the literature illustrates this is consistent with observations in other cancers where representative cell lines display a more aggressive phenotype.37, 38, 39 E-cadherin has been shown to be an invasive-suppressor gene and plakoglobin has been reported to display both oncogenic and tumor suppressor activity, dependent on the model assayed.40, 41 During preparation of this article, Shim et al.42 reported induction of plakoglobin expression by HDACi in fibrosarcoma cells. Presently, however, the role of plakoglobin in bladder tumorigenesis is unknown. We know that loss of E-cadherin expression in bladder tumors rarely results from gene loss or mutation (data not shown) and expression is likely regulated by methylation events associated with the promoter region of the E-cadherin gene11, 12 or silencing by alternative transcriptional regulators.43, 44, 45 Restoration of plakoglobin expression in bladder carcinoma cells exposed to HDACis suggests that the silencing of this gene in bladder tumorigenesis is at least in part a result of acetylation events. Following extended exposure of these cells to HDACis, no restoration of E-cadherin protein was detected. However, one previous study has shown that N-cadherin can act to augment the tumor suppressor activity of plakoglobin.40 In addition to its importance in cell-cell adhesion and cellular morphology, plakoglobin possesses growth-regulatory functions by virtue of its involvement in signal transduction and its association with transcription factors, as well as other tumor suppressor proteins.46 Winn et al.39 reported that plakoglobin expression was notably absent in certain lung tumors, specifically a poorly differentiated subset with higher metastatic potential. Importantly, reexpression of plakoglobin had a growth-inhibitory effect. Similarly, renal carcinoma cells transfected with plakoglobin do not form tumors in nude mice as readily as clones lacking the protein.40
In this study, we have used plakoglobin as a convenient marker to track the timing and duration of HDACi drug effect. Western blot analysis demonstrated that exposure to TSA for as little as 2 hr was sufficient to trigger increased protein expression, that this deregulation can be demonstrated after 12–18 hr of drug exposure and finally that protein levels revert to baseline levels 12–24 hr following drug removal. Similar kinetics have been reported by other groups using acetylated histones as a marker of HDACi-induced change. As reported, Richon et al.16 studied the effect of suberoylanilide hydroxamic acid (SAHA) treatment on T24 bladder carcinoma cells. While one would not necessarily expect that the pharmacokinetic properties between 2 distinct classes of HDACi are comparable, our drug kinetics data bear numerous similarities. Richon et al.16 showed maximum p21waf1 protein expression after 15 hr (9-fold increase), comparable to the timing of peak plakoglobin expression observed after 12–18 hr in this study. They also showed that exposure of cells to HDACis for 2 hr was sufficient to trigger upregulation of acetylated histones H3 and H4. Our findings were the same with plakoglobin in bladder carcinoma cells with respect to the time course of peak protein expression and the minimum duration of drug exposure necessary to produce a measurable effect. Such information may prove important should HDACis be considered as intravesical agents, in which case urothelial exposure would be under clinical control.
Administration of TSA at a dose of 500 μg/kg/day in EJ and UM-UC-3 xenograft models of bladder cancer produced a relative reduction of 87% and 72%, respectively, in the mean final tumor volume. Over the course of the treatment period, EJ mean tumor volume increased 8-fold in the control group and decreased 0.7-fold in the TSA-treated animals. Mean final tumor volume was 7.7-fold greater in the control tumors than the TSA-treated group. Similarly, the increase in control versus TSA-treated mean final tumor volume in the UM-UC-3 model at the end of the experiment was 3.6-fold. Given that no toxicity was observed at the dose used in this study, as gauged by body weight, death, behavior and food intake, it is possible that increased doses would have produced greater reductions in tumor volume. Histologic assessment of bladders from animals treated with TSA (500 μg/kg/day) revealed no detectable change in the morphology or differentiation status of the bladder mucosa compared to controls. Others have reported that use of TSA at 5 mg/kg23 elicits no discernible toxicity in a rat model, suggesting that there remains considerable latitude in the dose range that can be used.
In summary, we present the first study of HDACis on a panel of bladder cell lines and establish preliminary evidence for in vivo efficacy of TSA. We demonstrate that TSA and NaB selectively restore the expression of plakoglobin, a putative tumor suppressor gene. Studies are underway using plakoglobin-transfected cell lines to determine whether this protein plays a role in the antitumor drug effect. In this study, we provide preclinical evidence for consideration of HDACis in a therapeutic setting for late-stage bladder cancer. In a recent publication, the potential efficacy of HDACis in combination therapy was shown, where pretreatment of cells with TSA or SAHA increased the killing efficiency of anticancer drugs targeting DNA.47 We are currently evaluating the use of HDACis, in combination with standard therapeutic agents used in the treatment of bladder cancer.