Tirapazamine Cytotoxicity for Neuroblastoma Is p53 Dependent

Bioreductive agents, such as Tirapazamine (TPZ), are preferentially cytotoxic in hypoxia and show activity against a variety of tumor cell lines (1–4). The toxic species has been inferred to be a hydroxyl radical that is produced by a cofactor-dependent, one-electron reduction of TPZ, catalyzed by various cellular reductases, including the flavoenzyme NADPH reductase (2, 5). This highly reactive radical can be detoxified by oxygen, but in hypoxia it is converted to a stable two-electron reduction product (known as SR4317) by reaction with cellular constituents. Exposure to TPZ under hypoxic conditions leads to increased reactive oxygen species (ROS), DNA strand breaks (single and double), chromosome aberrations, and cell death (3, 6). In hypoxia, TPZ enhanced the in vitro activity of radiation and alkylating agents (7–12). In clinical trials TPZ showed activity in combination with radiotherapy and cisplatin-based chemotherapy (13, 14).

Neuroblastoma is a common childhood tumor of the sympathetic nervous system (15–18). Treatment of high-risk neuroblastoma (stage IV patients >1 year old at diagnosis, and stage III disease with MYCN amplification and/or unfavorable histopathology) with myeloablative, multiagent chemoradiotherapy, supported with purged autologous bone marrow transplantation and followed by 13-cis-retinoic acid, has improved the outcome for high-risk neuroblastoma (16). However, >50% of high-risk neuroblastoma patients still ultimately die from progressive disease that is refractory to further therapy and is often associated with acquired loss of p53 function in the recurrent tumor cells (19). Because relapse of neuroblastoma commonly occurs in hypoxic sites (15), hypoxic tissue may serve as a “sanctuary” site for neuroblastoma (20), consistent with the hypoxia-mediated decrease in chemotherapy and radiotherapy activity reported for other tumor types (21, 22). Therefore, it is important to identify agents that both retain cytotoxicity in reduced-oxygen environments and are p53-independent for use in the treatment of neuroblastoma.

Because previous reports showed that TPZ was a p53-independent agent in non–small-cell lung cancer (23) and human squamous cell carcinoma of the tongue (24), we have examined the cytotoxic properties in hypoxia of TPZ against human neuroblastoma cell lines with and without p53 function. Here we report that TPZ only induced apoptosis, multi-log cytotoxicity, p53 and p21 overexpression, and loss of mitochondrial membrane potential (ΔΨ_m) in p53-functional neuroblastoma cell lines.

Human neuroblastoma cell lines. Five human neuroblastoma cell lines (25) were obtained from patients at various points of disease [three at diagnosis: SMS-SAN, SK-N-BE(1), and CHLA-15, all have functional p53 (19); two cell lines were derived at progressive disease during intensive multiagent chemotherapy: SK-N-BE(2) and CHLA-171, both p53-nonfunctiononal] and one cell line was derived at relapse after bone marrow transplantation (CHLA-90, p53-nonfunctional). SK-N-BE(1), SK-N-BE(2), and SMS-SAN were cultured in RPMI 1640 (Irvine Scientific, Santa Ana, CA) with 10% fetal bovine serum (FBS; Gemini Bio-Products, Inc., Calabasas, CA); CHLA-15, CHLA-171, and CHLA-90 were cultured in Iscove's modified Dulbecco's medium (BioWhittaker, Walkersville, MD) supplemented with 3 mmol/L l-glutamine, insulin, transferrin 5 μg/mL each, 5 ng/mL selenous acid (ITS Culture Supplement, Collaborative Biomedical Products, Bedford, MA), and 20% FBS (complete medium). All cell lines were cultured at 37°C in 5% CO₂ and were cultured without antibiotics to avoid drug interactions and to facilitate detection of Mycoplasma. Experiments were carried out at passages 15 to 30. Cells were detached from culture plates or flasks with the use of a modified Puck's Solution A plus EDTA (Puck's EDTA), which contains 140 mmol/L NaCl, 5 mmol/L KCl, 5.5 mmol/L glucose, 4 mmol/L NaHCO₃, 0.8 mmol/L EDTA (EDTA), 13 μmol/L phenol red, and 9 mmol/L HEPES buffer (pH 7.3; ref. 26).

The SMS-SAN EV cell line is a clone of SMS-SAN that was transduced with the LXSN retrovirus as an “empty vector” control and the SMS-SAN E6 cell line is a clone of SMS-SAN transduced with the HPV16E6 gene carried in the LXSN retrovirus (19). Both lines are cultured in RPMI 1640 with 250 μg/mL G418 to select for gene incorporation; G418 is removed just before any drug testing (19).

Reduced oxygen conditions. For hypoxia assays, cells were cultured in plates or flasks and incubated in a sealed chamber with a mixture of 2% O₂, 5% CO₂, and 93% N (referred to as 2% O₂ mixture) at 37°C. Under these conditions, medium in plates and flasks attains a pO₂ of ∼15 mm Hg (27). This level of oxygen is below the degree of hypoxia found in bone marrow (17) and in the range of hypoxia found in tumor tissue (28). TPZ stock solution was diluted in whole medium that had been allowed to equilibrate overnight in a loosely capped flask in the 2% O₂ chamber.

Chemicals. TPZ was provided by Sanofi-Winthrop, Inc. (Malvern, PA) and was dissolved just before use in sterile water to make a 1 mg/mL stock solution. l-Buthionine sulfoximine was obtained as a 50 mg/mL solution from the Investigational Drug Branch, National Cancer Institute (Rockville MD) and was diluted to various concentrations in complete medium. Eosin Y, Ploybrene (hexadimethrine bromide), and the thiol antioxidant N-acetylcysteine were purchased from Sigma Chemical Co. (St. Louis, MO). N-Acetylcysteine was freshly dissolved in medium at a final concentration of 500 μmol/L, the pH adjusted to 7.4 with NaOH, and sterilized by 0.22 μmol/L filtration. Fluorescein diacetate was from the Eastman Kodak Company (Rochester, NY). The mitochondrial fluorescent probe 5,5′, 6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazol-carbocyanine iodide (JC-1) was purchased from Molecular Probes (Eugene, OR). Boc-d-fmk was purchased from Enzyme System Products (Livemore, CA). Stock solutions of fluorescein diacetate (1 mg/mL), Boc-d-fmk (40 mmol/L) and JC-1 (2 mg/mL) were dissolved in dimethyl sulfoxide and stored at −20°C. Antibodies to p53 (DO-1) mouse monoclonal, p21 (C-19) rabbit polyclonal, Bcl-2 (100) mouse monoclonal, Bcl-xL (H-62) rabbit polyclonal, Bax (p-19) rabbit polyclonal, α-tubulin mouse monoclonal, and horseradish peroxidase (HRP)–labeled secondary antimouse and anti-rabbit antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Enhanced chemiluminescence (ECL) Western blotting detection reagents were purchased from Amersham Pharmacia Biotech (Piscataway, NJ).

Cytotoxicity assay. All cytotoxicity assays are done in 96-well plates using a semiautomated digital image microscopy (DIMSCAN) system that has a dynamic range of 4 logs of cell kill (29). CHLA-15, CHLA-90, CHLA-171, and SK-N-BE(2) cells (fast growing) were plated at 5,000 cells per well; SK-N-BE(1), SMS-SAN, SMS-SAN EV (empty vector), and SMS-SAN E6 (HPV 16 E6 transduction) cells (slower growing) were plated at 15,000 cells per well; all cell lines were seeded in 100 μL of complete medium per well. Cells were cultured in hypoxia (2% O₂) and were allowed to attach 1 day before the addition of TPZ (0-2 μg/mL) in complete medium (to various final concentrations in 200 μL of complete medium) in replicates of 12 wells per condition. For some assays, N-acetylcysteine was added to a final concentration of 500 μmol/L, 3 hours before the addition of TPZ. Plates were assayed at 5 days after initiation of drug exposure.

To measure cytotoxicity, fluorescein diacetate was added to the 96-well plate (final concentration, 10 μg/mL) and incubated for 20 minutes. Afterwards 30 μL of eosin-Y (0.5% in normal saline) were added to quench background fluorescence of fluorescein diacetate in the medium and in nonviable cells (25). Total fluorescence per well (after digital thresholding to further eliminate background fluorescence) was then measured using a DIMSCAN system and results were expressed as the fractional survival of treated cells compared with control cells. The concentration of drug lethal for 99% of cells (LC₉₉) was calculated using the software “Dose-Effect Analysis with Microcomputers” (30).

Apoptosis. To quantify apoptotic cells with sub-G₀ DNA content, cells were cultured in complete medium in 25-cm² flasks, with or without 2 μg/mL TPZ for 24 hours in 2% O₂. Cells were then harvested, washed in PBS, centrifuged, and resuspended in 1 mL of 0.1% sodium citrate containing 0.05 mg propidium iodide and 50 μg RNase for 30 minutes at room temperature in the dark. RNase was dissolved in 10 mmol/L Tris-Cl (pH 7.5) and 15 mmol/L NaCl to a concentration of 10 mg/mL, and boiled at 100°C for 15 minutes, then stored in −20°C. DNA content was measured with a Coulter Epics Elite flow cytometer using a 488 nm argon laser and a 610 ± 10 nm band pass filter (20, 31).

Glutathione. Intracellular glutathione (GSH) was measured in triplicate by culturing cells in 25-cm² tissue culture flasks (2 × 10² cells) with or without 2 μg/mL TPZ for 24 hours in 2% O₂. For some assays, N-acetylcysteine was added to a final concentration of 500 μmol/L, 3 hours before the addition of TPZ. Cells were analyzed for total GSH content within 48 hours by the 5,5-dithiobis(2-nitrobenzoic acid)-oxidized GSH reductase method (32) as previously described (20, 33).

Assessment of mitochondrial potential transition. Loss of mitochondrial membrane potential (ΔΨ_m) leads to release of cytochrome C and other factors that trigger apoptosis (34, 35). The ΔΨ_m was determined (36) in the SMS-SAN, SMS-SAN EV, SMS-SAN E6, CHLA-15, and CHLA-90 cell lines after a 24-hour treatment with 2 μg/mL TPZ in 2% O₂. For some assays, N-acetylcysteine was added to a final concentration of 500 μmol/L, 3 hours before the addition of TPZ. After collection with Puck's EDTA, cells were resuspended in 1 mL of complete medium containing 10 μg/mL of JC-1 for 10 minutes at 37°C. JC-1 is a cationic dye that exhibits potential-dependent accumulation in mitochondria (36), indicated by a fluorescence emission shift from green (525 ± 10 nm) to red (610 ± 10 nm). Cells (15,000 per sample) were analyzed by flow cytometry using an argon laser (488 nm). Mitochondria depolarization is specifically indicated by a decrease in the red to green fluorescence intensity ratio. The ratio in the control group was considered as 1.0.

Protein expression. Proteins were extracted in radioimmunoprecipitation assay buffer (50 mmol/L NaCl, 50 mmol/L Tris, 1% Triton X-100, 1% sodium deoxycholate, and 0.1% SDS) and 50 μg of total protein were loaded per lane. Proteins were fractionated on 12% Tris-glycine precast gels (Novex, San Diego, CA), transferred to nitrocellulose membrane (Protran, Keene, NH), and probed with primary antibodies and then with HRP-labeled secondary antibodies. Antibody-positive bands were visualized using ECL Western blotting detection reagents.

Statistics. Significance (unpaired two-sided Student's t test) was determined by Microsoft Excel 2000 software.

Cytotoxicity assays of tirapazamine under normoxia and hypoxia. The dose response of human neuroblastoma cell lines to TPZ (0-2 μg/mL) as a single agent in hypoxia (2% O₂) is shown in Fig. 1 and LC₉₉ values are shown in Table 1. TPZ achieved >1 log (0.25 μg/mL) and >3 logs (2 μg/mL) of cell kill in all four p53-functional cell lines [SMS-SAN, CHLA-15, and SK-N-BE(1), and in SMS-SAN transduced with the LXSN empty vector (SMS-SAN EV)]. TPZ LC₉₉ values for p53-functional cell lines in hypoxia (0.3-0.7 μg/mL; mean, 0.5 ± 0.2 μg/mL) were below clinically achievable steady-state plasma concentrations (2.88 ± 0.37 μg/mL; ref. 37). However, TPZ induced <1 log cell kill in four p53-nonfunctional cell lines: SK-N-BE(2),CHLA-90 (TP53 mutated), CHLA-171 (TP53 wild-type but nonfunctional), and SMS-SAN E6 (in which p53 function was abrogated by transduction of 16E6 HPV). The LC₉₉ values for TPZ in all p53-nonfunctional cell lines (mean, >42.6 μg/mL; range, 20.6 to > 50 μg /mL) were greater than the reported clinically achievable steady-state plasma concentrations for TPZ of 2.88 ± 0.37 μg/mL (37).

Apoptosis following treatment with tirapazamine. The SMS-SAN cell line was treated with 2 μg/mL of TPZ for 24 hours and examined for apoptosis by propidium iodide staining and flow cytometry. Spontaneous apoptosis (control) was seen in 3% to 11% of cells in hypoxia (Fig. 2A). In 2% O₂, TPZ induced apoptosis in 65% (SMS-SAN) and 46% (SMS-SAN EV) of cells at 24 hours, but TPZ did not increase apoptosis (relative to controls) in the p53 nonfunctional cell lines SMS-SAN E6 and CHLA-90.

Expression of p53, p21, Bcl-2, Bcl-xL, and Bax. The basal and TPZ-treated expression of p53, p21, Bcl-2, Bcl-xL, and Bax was measured by immunoblotting. As shown in Fig. 2B, 2 μg/mL TPZ increased p53 and p21 protein levels in SMS-SAN cells after 24 hours of exposure in 2% O₂. In CHLA-90 (a TP53-mutated cell line with high basal p53 expression), TPZ (2 μg/mL) failed to increase p53 and p21 expression in 2% O₂. Increased expression of Bcl-2, Bcl-xL, or Bax was not detected in either the CHLA-90 or SMS-SAN cell lines after 2 μg/mL TPZ treatment for 24 hours.

Effect of tirapazamine on glutathione. Cell lines were incubated for 24 hours with 2 μg/mL TPZ in hypoxia and total GSH was assayed. TPZ induced a significant decrease in GSH to <70% baseline (Fig. 3A, P < 0.05) in four neuroblastoma cell lines with p53 function [SMS-SAN, SMS-SAN EV, CHLA-15, and SK-N-BE(1)], but not in the p53 nonfunctional, highly drug-resistant cell lines [SMS-SAN E6, SK-A-BE(2), CHLA-90, and CHLA-171; P > 0.05]. N-Acetylcysteine (500 μmol/L) significantly increased (P < 0.05) the basal level of GSH and restored TPZ-mediated GSH depletion in the CHLA-15 cell line (Fig. 3B), but failed to inhibit the cytotoxicity of TPZ (Fig. 3C).

Loss of mitochondrial membrane potential (ΔΨ_m). We determined ΔΨ_m in SMS-SAN, SMS-SAN EV, SMS-SAN E6, CHLA-15, and CHLA-90 cell lines treated with 2 μg/mL of TPZ for 24 hours in hypoxia. Using flow cytometry, we measured ΔΨ_m by JC-1 staining. TPZ decreased ΔΨ_m in SMS-SAN, SMS-SAN EV, and CHLA-15 at 24-hour incubation (Fig. 4A), but failed to induce a loss of ΔΨ_m in SMS-SAN E6 and CHLA-90. Consistent with the inability of N-acetylcysteine to inhibit TPZ cytotoxicity (Fig. 3C), N-acetylcysteine (500 μmol/L) failed to restore the ΔΨ_m decrease caused by TPZ in CHLA-15 (Fig. 4B). The time course of apoptosis and ΔΨ_m loss induced by TPZ showed that the ΔΨ_m loss preceded the induction of apoptosis (Fig. 5A).

Effect of Boc-d-fmk on tirapazamine-induced apoptosis and loss of ΔΨ_m. To determine if TPZ-mediated apoptosis and loss of ΔΨ_m in hypoxia are caspase dependent, we incubated CHLA-15 cells with the pan-caspase inhibitor Boc-d-fmk (40 μmol/L) for 1 hour before adding 2 μg/mL TPZ for 24 hours. Cells were treated with 500 μmol/L l-buthionine sulfoximine (20, 38) for 48 hours in normoxia (20% O₂) as a positive control for the ability of Boc-d-fmk to inhibit apoptosis. As shown in Fig 5B, Boc-d-fmk significantly (P < 0.05) inhibited l-buthionine sulfoximine–induced apoptosis in normoxia but failed (P > 0.05) to block TPZ-mediated apoptosis in hypoxia. Both l-buthionine sulfoximine and TPZ induced ΔΨ_m decrease with or without Boc-d-fmk.

Neuroblastoma initially responds to chemotherapy but recurs in most high-risk patients as a chemotherapy-insensitive disease (15). Tumor hypoxia (found in common sites of neuroblastoma relapse, such as bone and bone marrow) could contribute to enhanced survival of tumor cells and facilitate selection for drug resistance (16–18). Because some chemotherapeutic agents are antagonized by hypoxia (1, 20, 22, 39), identifying regimens that are active in reduced oxygen environments could improve therapy for neuroblastoma.

TPZ is an anticancer drug that is activated in hypoxic cells to a highly damaging free radical, resulting in selective cytotoxicity of cells grown in hypoxia (5, 40). Preclinical models suggested that TPZ can enhance both irradiation and chemotherapy in hypoxic cancer cells (8–10), and previous reports indicated that TPZ is a p53-independent cytotoxic agent (23, 24). We showed that TPZ exhibited cytotoxicity only in p53-functional neuroblastoma cell lines. Clinically obtainable average levels of TPZ achieved 3 to >4 logs of cell kill in all p53-functional cell lines with LC₉₉ values ≤0.7 μg/mL, whereas all p53 nonfunctional cell lines showed high-level resistance to TPZ in hypoxia. The expression of p21 is transcriptionally activated by p53, and p21 mediates a p53-dependent G₁ arrest following DNA damage, also potentially regulating chemosensitivity of tumor cells as a cell cycle checkpoint (41, 42). We found that TPZ induced expression of p53 and p21 proteins in SMS-SAN, a cell line established at diagnosis with functional p53 (19). However, in CHLA-90, a multidrug resistant cell line established after myeloablative therapy with mutated and nonfunctional TP53 (19), basal and constitutive overexpression of p53 was observed, and p21 expression and apoptosis could not be induced by TPZ. These data suggested that TPZ is p53 dependent in neuroblastoma cell lines.

Because the p53-nonfunctional neuroblastoma cell lines are all derived after intensive chemotherapy and may possess drug resistance mechanisms beyond loss of p53 function, we studied the SMS-SAN E6 cell line in which p53 function was abrogated via HPV 16 E6 transduction (19). Unlike the SMS-SAN parental line, or empty vector controls, SMS-SAN E6 showed no response to TPZ in terms of apoptosis, ΔΨ_m loss, or cytotoxicity. These data confirmed that TPZ cytotoxicity in neuroblastoma cell lines results from p53-dependent apoptosis. Our data contrast with previous reports that TPZ is a p53-independent bioreductive agent in non–small-cell lung cancer (23) and human squamous cell carcinoma of the tongue (24). The differences observed in TPZ cytotoxic behavior may be due to the different tumor cell types or the various dosages studied. The maximum dosage of TPZ we used is 2 μg/mL (11.2 μmol/L), which is below clinically achievable steady-state plasma concentrations (2.88 ± 0.37 μg/mL; ref. 37), but as hypoxic tumors are very likely to be only exposed to drug levels lower those achieved in plasma, we felt that 2 μg/mL was an appropriate maximum concentration for these experiments.

It has been shown that a change in mitochondrial membrane potential (ΔΨ_m) plays a pivotal role in transducing a variety of proapoptotic stimuli (43–45). Opening of permeability transition pores in mitochondria induces the release of cell death–promoting factors, including cytochrome C and apoptosis-inducing factor (35, 46). Cytochrome C has been shown to be involved in the activation of a caspase cascade (47, 48), whereas apoptosis-inducing factor has been shown to directly trigger apoptosis (49). Our time-course data showed that the TPZ-induced ΔΨ_m loss preceded induction of apoptosis in hypoxic CHLA-15 cells, suggesting that the ΔΨ_m loss may result in apoptosis in these p53-dependent cell lines. To show whether ΔΨ_m-mediated apoptosis is involved in caspase activation, we determined the effect of the pan-caspase inhibitor Boc-d-fmk on TPZ-induced ΔΨ_m loss and apoptosis. The results showed that Boc-d-fmk significantly inhibited apoptosis induced by l-buthionine sulfoximine (a specific inhibitor of GSH synthesis; ref. 50) without an effect on the ΔΨ_m decrease mediated by l-buthionine sulfoximine. Boc-d-fmk did not block TPZ-induced apoptosis or loss of ΔΨ_m. Taken together, the caspase inhibitor and time-course studies suggest that TPZ-induced apoptosis is mediated by mitochondrial permeability changes that lead to caspase-independent cytotoxicity.

It is reported that the release of mitochondrial apoptogenic factor is regulated by the pro- and anti-apoptotic Bcl-2 family proteins, which either induce or prevent the permeabilization of the outer mitochondrial membrane (51). Bax gene expression can be regulated in vitro by p53, and the ratio of Bax to Bcl-2 within a cell is critical for the cellular response to apoptotic stimuli (52). We showed that TPZ-induced apoptosis and ΔΨ_m loss occurred only in p53-functional cell lines (SMS-SAN, SMS SAN EV, and CHLA-15) but not in p53-nonfunctional cell lines (CHLA-90 and SMS-SAN E6), suggesting that p53 activation may be essential for loss of ΔΨ_m, which then triggers apoptosis. However, TPZ-mediated p53 overexpression failed to change the expression of Bcl-2, Bcl-xL, and Bax proteins. Thus, further studies will be needed to address the actual link between TPZ-triggered p53 activation and the loss of ΔΨ_m and the Bcl/Bax family of proteins.

Intracellular redox potential plays a role in the commitment to cell death, and a major determinant of the cellular redox state is GSH (43). GSH protects cells by detoxifying ROS, which are powerful activators of a decrease in ΔΨ_m loss and apoptosis (43, 47, 53). Depletion of GSH is highly cytotoxic for neuroblastoma cell lines in vitro in standard culture conditions, likely due to unopposed ROS produced during catecholamine synthesis (38), and the protective effect of GSH in neuroblastoma can be replaced by antioxidants (20, 38), such as N-acetylcysteine. TPZ has been reported to induce ROS and cause DNA damage (3, 6, 20), and the elimination of ROS by N-acetylcysteine can inhibit p53-mediated apoptosis in some systems, suggesting a role for ROS in the p53 death response (54). Our data show that in p53-functional cell lines, TPZ significantly diminished GSH levels, ΔΨ_m, and cell survival. We also showed that N-acetylcysteine inhibited TPZ-mediated GSH depletion but not the loss of ΔΨ_m or cytotoxicity induced by TPZ. N-Acetylcysteine could only partially decrease cytotoxicity of TPZ at low doses (<1 μg/mL, data not shown). Therefore, GSH depletion and ROS production do not seem to play essential roles in the p53-dependent cytotoxicity of TPZ for neuroblastoma cell lines.

Curiously, TPZ-mediated GSH depletion was only observed in p53-functional cell lines, and not in p53-nonfunctional lines, indicating that ΔΨ_m loss and/or cytotoxicity may play a role in the decrease of GSH levels by TPZ. The ability of N-acetylcysteine to restore GSH without inhibiting ΔΨ_m loss, cytotoxicity, or apoptosis suggests that the decreased GSH seen in p53-functional cell lines treated with TPZ was not simply due to a loss of cell viability. However, as p53-functional cell lines underwent significant apoptosis after 24 hours of TPZ treatment (Fig. 2), we cannot exclude death-related events as an explanation for the difference in GSH decrease seen in p53-functional compared with p53-nonfunctional cell lines.

Nearly half of the cell lines established from high-risk neuroblastomas recurring after intensive chemotherapy have TP53 mutations and/or p53 loss of function (LOF; ref. 19). Whereas our data show that TPZ as a single agent is not effective against p53-LOF neuroblastoma cell lines, it is possible that combining TPZ with hypoxia-sensitive drugs that act in a p53-independent manner [examples include buthionine sulfoximine + melphalan (20) or fenretinide (27)] may result in a synergistic combination that is not limited by p53 function. However, our data suggest that use of TPZ may be best avoided in tumors in which loss of p53 function is common.

We thank Paul Alfaro for outstanding technical assistance.