Mycobacterium bovis Bacillus Calmette-Guérin (BCG) use in the treatment of bladder cancer was first reported in 1976, but the mechanism of the induced antitumor activity has still not been fully explained. BCG is a potent immunostimulant, normally producing a Th1 cytokine response, including IFN. Recent studies have shown CpG oligodeoxynucleotide induce tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) expression via IFN production. Given that Mycobacterial DNA contains high amounts of CpG motifs, we hypothesized that BCG’s antitumor properties are akin to CpG oligodeoxynucleotide, where the cytokine response to BCG induces TRAIL up-regulation. Using ELISA, urine IFN-γ, and TRAIL levels were initially undetectable in BCG therapy patients but were high after later induction treatments. More importantly, patients that responded to BCG therapy had significantly higher urine TRAIL levels, which killed bladder tumor cells in vitro versus nonresponders. Flow cytometry of fresh urine revealed TRAIL-expressing neutrophils. Given these data, we propose TRAIL plays a role in BCG-induced antitumor effects.
Annually, ∼13,000 deaths result from bladder cancer, and >60,000 new cases will appear in 2004, making it the fourth most common cancer among men and tenth in women (1). Among these newly diagnosed cases, the majority will be before muscle invasion, thus superficial in stage and potentially completely curable. In 1976, Morales et al. (2) first described the anti-tumor effects of intravesical administration of Mycobacterium bovis Bacillus Calmette-Guérin (BCG) in bladder cancer. Today, BCG therapy remains at the forefront of treating patients with superficial bladder cancer and carcinoma in situ and continues to be one of the most successful immunotherapies for any solid human malignancy. BCG instillation results in an early granulocytic influx into the bladder wall followed by mononuclear cells where CD4+ cells predominate (3). A proinflammatory Th1 cytokine response predominates [interleukin (IL)-2, IL-12, IFN-γ] after BCG stimulation (4), and the Th1 response is often associated with a favorable response (5). Moreover, BCG combined with IFN-α2B additionally polarizes the Th1 cytokine response (increasing IFN-γ, IL-12, tumor necrosis factor production; decreasing IL-6, IL-10 production; Ref. 6). However, ∼30 years after its inception as a treatment for bladder cancer, the exact mechanism for BCG’s antitumor properties has not been fully elucidated.
Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL/Apo-2L) induces apoptotic cell death in a variety of tumor cells but has no cytotoxic activity against normal cells or tissues (7). Expression can be induced after IFN-α and IFN-γ stimulation on T cells, natural killer cells, dendritic cells, and Mφ (8, 9, 10, 11). Oligodeoxynucleotides containing CpG motifs have profound immunostimulatory effects, including antitumor immunity (12), and recognition of CpG DNA by the immune system has evolved as a defense mechanism against bacterial (intracellular) pathogens (13). Oligodeoxynucleotides containing CpG motifs activate numerous immune cell populations and drives the immune response toward a Th1 phenotype (12). Recently, we showed that human peripheral blood mononuclear cell stimulated by oligodeoxynucleotides containing CpG motifs induces an anti-tumor response by up-regulating TRAIL expression via IFN-α production (14). Of similar interest, Mycobacterial species contain among the highest CpG nucleotide content of all prokaryotes, and much of the pioneering work on immunostimulatory DNA was performed on BCG-mediated tumor resistance studies (12).
Given the Th1 response after BCG stimulation, we proposed that the mechanism for the antitumor activity of BCG therapy was analogous to that of oligodeoxynucleotides containing CpG motifs, where the cytokine response, specifically IFN, to BCG leads to the up-regulation of TRAIL. Herein, we demonstrate functional TRAIL in the urine of patients undergoing BCG therapy and observed increased urine TRAIL levels in those patients who responded to BCG therapy compared with nonresponders. We also demonstrate the presence of TRAIL on neutrophils in voided urine after BCG instillation, suggesting a significant, clinically relevant antitumor role for neutrophils.
Materials and Methods
Monoclonal Antibody (mAb).
The following mAb for used for flow cytometry: UCHT1, FITC-conjugated IgG1 antihuman CD3; M5E2, FITC-conjugated IgG2a antihuman CD14; HI98, phycoerythrin (PE)-conjugated IgM antihuman CD15; HIB19, FITC-conjugated IgG1 antihuman CD19; B159, PE-conjugated IgG1 antihuman CD56 (BD Biosciences, San Diego, CA); RIK-2, biotinylated IgG1 antihuman TRAIL [provided by Dr. Hideo Yagita, Juntendo University (Tokyo, Japan)]; and IgG1-FITC, IgG2a-FITC, IgG1-PE, and IgG1-biotin isotype controls (Caltag Laboratories, Inc., Burlingame, CA).
Tumor Cell Line.
The human bladder tumor cell line RT-4 was obtained from Dr. Scott Crist (University of Iowa) and cultured in DMEM supplemented with 10% FBS, penicillin, streptomycin, sodium pyruvate, nonessential amino acids, and HEPES.
Cell analysis was performed on a FACScan (Becton Dickinson, Franklin Lakes, NJ) with >104 cells analyzed/sample. For multicolor analysis, 20 μl of cells were combined in a 96-well round-bottomed plate with 20 μl each of the directly PE- or FITC-labeled mAb and biotinylated anti-TRAIL mAb and incubated at 4°C for 30 min. After three washes with PBS containing 2 mg/ml BSA and 0.02% NaN3, 40 μl of FITC- or PE-labeled streptavidin (1:100 dilution; Caltag Laboratories, Inc.) was added for 30 min. Cells were analyzed immediately after staining or fixed in 1% paraformaldehyde until analysis.
TRAIL, IFN-γ, and IFN-Inducible Protein-10 (IP-10) ELISA.
Human TRAIL and IFN-γ levels in the urine from BCG patients were quantitated using a sandwich ELISA purchased from Diaclone Research (Besancon, France) and BD Biosciences, respectively. An ELISA specific for human IP-10 was developed using, as a capture reagent, protein A-purified rabbit immunoglobulin and a biotinylated affinity-purified rabbit antihuman IP-10 as the detecting antibody (15). Recombinant human IFN-γ and recombinant human IP-10 (BD Biosciences) were used as concentration controls.
Immunoprecipitation and Western Blotting.
The anti-TRAIL mAb M181 (20 μg; Amgen, Seattle, WA) was added to equal volumes of urine from BCG patients for 2 h at 4°C, followed by the addition of protein G-Sepharose beads (Sigma Chemical Co.) for 2 h at 4°C. The beads were pelleted by centrifugation, washed three times with PBS/0.1% Triton X-100, and resuspended in SDS-PAGE sample buffer. Precipitates were released from the protein G beads by boiling for 2–3 min and then separated by SDS-PAGE, transferred to nitrocellulose, and blocked overnight at 4°C in 5% nonfat dry milk in PBS-Tween-20 (0.05% v/v). The membrane was incubated with an anti-TRAIL polyclonal antiserum (Peprotech, Rocky Hill, NJ), washed, and then incubated with an antirabbit-horseradish peroxidase antibody and developed by chemiluminescence (SuperSignal West Pico Chemiluminescence Substrate; Pierce, Rockford, IL).
Voided urine was collected separately after each void for 24 h after BCG instillation or 2–12 h after therapy and pooled for later analysis. Samples were stabilized during patient collection with a concentrated buffer containing 2 m Tris-HCl (pH 7.6), 5% BSA, 0.1% sodium azide, and four protease inhibitors [aprotinin, pepstatin, and leupeptin at 0.01 mg/ml for each and 3-(2-aminoethyl) benzenesulfonyl fluoride at 0.1 mg/ml (all from Sigma Chemical Co.)]. A 10-ml sample was further preserved by the addition of a protease inhibitor mixture tablet (Boehringer Mannheim, Mannheim, Germany) and stored at −70°C before analysis. The urine used for Western blotting and flow cytometry was collected 3–5 h after BCG instillation and was stabilized during collection with FBS and HEPES (×10). After collection, the urine was centrifuged (1200 rpm/5 min), the supernatant removed and stored at −20°C until analysis. The cells were additionally separated with Ficoll and washed three times with PBS before flow cytometry staining.
Urine TRAIL-Mediated Killing of Human Tumor Cells.
Urine specimens from BCG responders, nonresponders, and normal healthy donors were dialyzed in PBS for 36 h using Membra-Cel dialysis tubing (Applied Technologies Group, Willowbrook, IL; molecular weight cutoff Mr 14,000). After dialysis, samples were concentrated using Centriplus spin concentrators (Millipore Corporation, Billerica, MA). After concentration, some of the BCG responder urine was immunodepleted of TRAIL. The anti-TRAIL mAb M181 (20 μg) was added to equal volumes of urine for 2 h at 4°C, followed by the addition of protein G-Sepharose beads (Sigma Chemical Co., St. Louis, MO) for 2 h at 4°C. The beads were pelleted by centrifugation, and the TRAIL-depleted urine or the concentrated urine from BCG responders, BCG nonresponders, and normal healthy donors was then incubated with RT-4 cells for 18 h, after which, cell death was determined by crystal violet staining (16).
Presence of TRAIL in the Urine of Patients Undergoing BCG Therapy.
TRAIL expression can be induced on numerous inflammatory cell types after IFN-α and IFN-γ stimulation (8, 9, 10, 11). Given that many proinflammatory cytokines, including IFN, are present in the urine of patients undergoing BCG therapy, we examined the urinary levels of TRAIL in patients after BCG instillation. Urine samples were collected from patients from hours 2 to 12 after BCG instillation during their induction course of therapy and pooled. The 10-h pooled urine samples were then analyzed for urinary TRAIL using ELISA. Interestingly, BCG responders (>12 months tumor free) had a significantly higher amount of urinary TRAIL (n = 11; mean, 299 pg/ml) versus BCG nonresponders (recurrence < 12 months; n = 6; mean, 105 pg/ml; P < 0.05; Fig. 1 A). Analysis of urine from patients with urinary tract infections revealed undetectable TRAIL levels (data not shown), thus the TRAIL response appears to be specific to BCG administration.
Although Fig. 1,A shows the presence of urinary TRAIL in pooled 10-h urine samples after BCG instillation, we also analyzed the kinetics of urinary TRAIL production during a specific treatment, as well as how the levels varied throughout the induction course. To do this, urine was collected from each void during the 24 h immediately after BCG instillation for each induction treatment and then measured TRAIL levels. BCG nonresponders had low TRAIL levels for all induction courses studied (Fig. 1,B). In contrast, BCG responders showed low TRAIL levels during the first treatment but had an increased and more sustained TRAIL response with later treatments (Fig. 1 C).
Because TRAIL can be induced after IFN-γ stimulation, we examined the relationship between urinary levels of IFN-γ, TRAIL, and the IFN-induced chemokine, IP-10. As in Fig. 1, B and C, each void from the first 24 h after BCG instillation was examined for IFN-γ, TRAIL, and IP-10 using ELISA. The data show IFN-γ levels increased first, followed by increases in TRAIL and IP-10 levels (Fig. 1 D). Collectively, these results demonstrate increased TRAIL levels in the urine of patients who responded favorably to BCG immunotherapy compared with those who did not and increasing TRAIL levels as patients progress through their induction course.
Urinary TRAIL-Dependent Killing of RT-4 Bladder Tumor Cells.
TRAIL induces apoptotic cell death in numerous malignancies, including bladder cancer, although not affecting normal tissues or cells (7, 17). Both soluble TRAIL and membrane-bound TRAIL can induce apoptosis through binding with the death-inducing TRAIL receptor 1 and TRAIL receptor 2 (18). Having established the presence of TRAIL in the urine of patients undergoing BCG therapy, we wished to identify the form of urinary TRAIL, demonstrate TRAIL-dependent killing of bladder cancer cells by voided urine of patients after BCG instillation, and determine which inflammatory cells expressed TRAIL when recruited to the bladder after BCG immunotherapy.
Full-length cell surface TRAIL can be cleaved by cysteine proteases to form soluble TRAIL (19). To determine the form of TRAIL present in the urine of patients after BCG instillation, we performed an immunoprecipitation and Western blot of the voided urine. The form of urinary TRAIL is the soluble cleaved form with prominent bands at Mr 19,000 (cleaved monomer) and Mr ∼45,000 (cleaved dimer; Fig. 2,A). Next, we tested whether the soluble urinary TRAIL was functional against bladder tumor cells. In Fig. 2 B, we demonstrate that the urine of a patient that responded favorably to BCG therapy kills RT-4 bladder tumor cells, whereas the urine of BCG nonresponders or urine from normal healthy donors shows little to no killing. This killing was observed using urine TRAIL from three different BCG responders. Moreover, immunodepletion of TRAIL from the urine of BCG responders resulted in baseline levels of killing. Thus, killing of bladder cancer cells by voided urine of patients after BCG instillation occurs in a TRAIL-dependent manner.
After BCG instillation, there is a large influx of leukocytes into the bladder and bladder wall (3). We analyzed urine of patients undergoing BCG therapy for TRAIL expression on the voided leukocytes. During the time period examined (3–5 h after instillation), there was high expression of TRAIL on voided neutrophils (Fig. 2 C). We also examined CD3+, CD14+, CD19+, and CD56+ cells from this time point; however, no TRAIL expression was detected (data not shown). The lack of TRAIL expression on these other populations may be secondary to the limited time frame in which the urine sample was collected. Collectively, the data presented herein demonstrate the presence of functional TRAIL in the urine of responding BCG patients, suggesting a potential mechanistic role in BCG immunotherapy. With additional investigation, these results may give clinicians the potential to clinically aid the surveillance and treatment of bladder cancer.
Nearly 30 years ago, Morales et al. (2) first described the successful administration of BCG for the treatment of bladder cancer. These observations were later confirmed in a larger study (20), but today, the exact mechanism of BCG’s antitumor properties remain unknown. Upon instillation into the bladder, BCG attachment to the urothelium, via fibronectin, is essential for effective therapy (21). Viable BCG must be used to induce the antitumor response because heat-killed, nonviable BCG does not bind to fibronectin (T. L. Ratliff, personal communication). An early influx of granulocytes occurs, followed by the influx of mononuclear cells (3), after BCG is instilled into the bladder, and CD4+ and CD8+ cells are required for BCG-mediated antitumor activity (22). Interestingly, we observed high TRAIL expression on the neutrophils present in the urine of patients immediately after BCG instillation (3–5 h), suggesting a new and very profound role for components (i.e., the massive granulocyte influx) of the innate immune system in the BCG immunotherapeutic scheme. Recently, IFN-α-stimulated neutrophils and monocytes were found to rapidly produce and release a functional, soluble form of TRAIL (23). Thus, in addition to recruiting and activating other cells of the immune system, granulocytes elicited by BCG instillation may also have a direct antitumor effect through the TRAIL pathway (both the soluble and surface bound forms). Although we only observed the presence of TRAIL on voided neutrophils, it is possible that the lack of TRAIL expression on the other inflammatory cells present was a product of the limited time course evaluated (3–5 h after instillation). It is probable that our collection time coincided with the early inflammatory response and limited the detection of TRAIL to those early inflammatory granulocytic cells while not entirely examining all of the recruited inflammatory cells that are present after BCG administration. Given that peripheral blood mononuclear cell stimulated with BCG induces TRAIL expression on T cells, B cells, natural killer cells, and Mφ (A. T. Ludwig and T. S. Griffith, unpublished observation), it could be expected that later in the inflammatory reaction, TRAIL expression would also be present on these peripheral blood mononuclear cell populations because they are recruited later into the bladder wall. Additional investigation is needed to evaluate the presence of TRAIL during later time points after BCG instillation.
The high recurrence rate of bladder cancer makes tumor surveillance extremely important, and the ability to identify those individuals most likely to recur after BCG therapy would be a tremendous advantage. Many markers have been examined to provide some detail as to which patients are likely to have a successful course of BCG therapy and which are more likely to fail. Many studies have focused on the cytokine response to BCG therapy, specifically the Th1 response, IL-2, IL-8, and IL-18 (24, 25). Recently, TRAIL has been examined as a potential response marker for IFN-β treatment in multiple sclerosis (26), and our data suggests TRAIL may also serve as a marker for patient outcomes after BCG instillation. Rather than the intermediaries that are currently being investigated, the exciting aspect of using TRAIL as a therapeutic marker lies in the ability to track an effector molecule. For example, the fresh urine supernatant from a patient currently undergoing BCG therapy revealed high soluble TRAIL levels (1500 pg/ml; unpublished observation) in addition to the expression of TRAIL on the surface of neutrophils, thus, we would predict this individual will have a favorable response to their BCG therapy. Although our results are very encouraging, a thorough prospective analysis is necessary to additionally examine the potential of TRAIL as a therapeutic success marker.
In addition to the potential predictive value of monitoring the TRAIL response after BCG instillation, the observation of TRAIL-dependent killing of bladder cancer cells after BCG instillation has many potential clinical ramifications. The ability to alter the immune response and increase the TRAIL expression after BCG instillation could have significant benefits in increasing patients response to BCG therapy and decreasing disease recurrence after BCG immunotherapy. Current strategies such as combination therapy of BCG and IFN-α may, in fact, act on these observations and alter the TRAIL response. Other strategies that could potentially be beneficial in increasing the TRAIL response include increasing the recruitment of inflammatory cells into the bladder after BCG instillation and/or increasing the Th1 inflammatory response to BCG therapy, namely increasing IFN production. Ultimately, the direct application of recombinant TRAIL intravesically could be formulated into the treatment of superficial bladder cancer. The ability to maximize the TRAIL response may allow for the decrease in BCG dosage and thus decrease the potential for the hazardous side effects of BCG therapy.
We introduce TRAIL as novel mechanism for the antitumor effects of BCG immunotherapy; however, it is likely that the antitumor effects of BCG are many fold. For example, TRAIL may play a critical role during the acute inflammatory state, yet secondary events such as BAK (CD8+/CD56+) cells (27) may similarly influence the response during the chronic inflammatory state. This may explain the observed late resolution of carcinoma in situ 4–5 months after the completion of BCG induction cycle (28) in which TRAIL killing accounts for the initial response to therapy, and the secondary events are responsible for the added 11% late response.
In summation, we demonstrate the presence of soluble TRAIL/Apo-2L in the urine of patients after BCG immunotherapy and observed increased levels of TRAIL/Apo-2L in those patients that responded favorably to BCG therapy. In addition, we found TRAIL/Apo-2L expression on voided neutrophils, which may signal a new and significant role for the massive granulocyte influx after BCG instillation. Collectively, this data suggests TRAIL/Apo-2L has a mechanistic role in BCG immunotherapy.
Grant support: Department of Defense Prostate Cancer Research Program New Investigator Award PC010599.
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: Thomas S. Griffith, Department of Urology, 3204 MERF, University of Iowa, 375 Newton Road, Iowa City, IA 52242-1089. Phone: (319) 335-7581; Fax: (319) 335-6971; E-mail: email@example.com
We thank Drs. Richard Williams and Timothy Ratliff for critical review of the manuscript.