Selective Potentiation of the Hypoxic Cytotoxicity of ... - Cancer Research

[CANCER RESEARCH 64, 736 –742, January 15, 2004]

Selective Potentiation of the Hypoxic Cytotoxicity of Tirapazamine by Its 1-N-Oxide Metabolite SR 4317 Bronwyn G. Siim,1,2 Frederik B. Pruijn,1 Joanna R. Sturman,1 Alison Hogg,1 Michael P. Hay,1 J. Martin Brown,3 and William R. Wilson1 1 Auckland Cancer Society Research Centre, and 2Department of Molecular Medicine and Pathology, The University of Auckland, Auckland, New Zealand, and 3Cancer Biology Research Laboratory, Department of Radiation Oncology, Stanford University School of Medicine, Stanford, California

ABSTRACT Tirapazamine (TPZ), a bioreductive drug with selective toxicity for hypoxic cells in tumors, is currently in Phase III clinical trials. It has been suggested to have a dual mechanism of action, both generating DNA radicals and oxidizing these radicals to form DNA breaks; whether the second (radical oxidation) step is rate-limiting in cells is not known. In this study we exploit the DNA radical oxidizing ability of the 1-N-oxide metabolite of TPZ, SR 4317, to address this question. SR 4317 at high, but nontoxic, concentrations potentiated the hypoxic (but not aerobic) cytotoxicity of TPZ in all four of the human tumor cell lines tested (HT29, SiHa, FaDu, and A549), thus providing a 2–3-fold increase in the hypoxic cytotoxicity ratio. In potentiating TPZ, SR 4317 was 20-fold more potent than the hypoxic cell radiosensitizers misonidazole and metronidazole but was less potent than misonidazole as a radiosensitizer, suggesting that the initial DNA radicals from TPZ and radiation are different. SR 4317 had favorable pharmacokinetic properties in CD-1 nude mice; coadministration with TPZ provided a large increase in the SR 4317 plasma concentrations relative to that for endogenous SR 4317 from TPZ. It also showed excellent extravascular transport properties in oxic and anoxic HT29 multicellular layers (diffusion coefficient 3 ⴛ 10ⴚ6 cm2sⴚ1, with no metabolic consumption). Coadministration of SR 4317 (1 mmol/kg) with TPZ at a subtherapeutic dose (0.133 mmol/kg) significantly enhanced hypoxic cell killing in HT29 tumor xenografts without causing oxic cell killing, and the combination at its maximum tolerated dose was less toxic to hypoxic cells in the retina than was TPZ alone at its maximum tolerated dose. This study demonstrates that benzotriazine mono-N-oxides have potential use for improving the therapeutic utility of TPZ as a hypoxic cytotoxin in cancer treatment.

INTRODUCTION Hypoxia contributes to tumor progression, and limits the response of tumors to radiotherapy and chemotherapy (1, 2). However, tumor hypoxia can be turned to advantage by exploiting it to activate bioreductive prodrugs within tumors. The most advanced of the hypoxia-selective bioreductive drugs in clinical evaluation is TPZ (Tirapazamine; 1,2,4-benzotriazin-3-amine 1,4-dioxide, SR 4233), which showed significant activity in combination with cisplatin in a Phase III trial with non-small cell lung cancer (3), and an interim analysis also suggests activity in a cisplatin-based chemoradiation randomized Phase II trial in advanced head and neck cancers (4). However, there are significant toxicities associated with the clinical use of TPZ (including neutropenia, thrombocytopenia, diarrhea, nausea, vomiting, cramping, and fatigue) that preclude its administration at high enough doses to fully exploit tumor hypoxia throughout a treatment regimen. In addition, the differential toxicity of TPZ toward hypoxic cells in tumors (5) is much lower than in conventional (low cell Received 8/10/03; revised 10/21/03; accepted 11/13/03. Grant support: NIH National Cancer Institute Grant PO1-CA82566 and Cancer Society of New Zealand (Grant 02-03/25). 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: William R. Wilson, Auckland Cancer Society Research Centre, The University of Auckland, Private Bag 92019, Auckland, New Zealand. Phone: 64-9373-7599, extension 86883; Fax: 64-9-373-7571; E-mail: [email protected].

density) cultures, in part because rapid metabolic consumption of TPZ compromises its extravascular diffusion into hypoxic tissue (6 –9). The therapeutic selectivity of TPZ toward hypoxic tumor cells could be enhanced by improving its extravascular transport or by selectively increasing its cytotoxicity under hypoxic, but not aerobic, conditions. The mechanism of the selective toxicity of TPZ under hypoxia involves enzymatic reduction by the addition of a single electron to form a radical anion (TPZ⫺•), which ultimately results in DNA strand cleavage by a pathway that is only partially understood (Fig. 1). In the presence of oxygen TPZ⫺• is back-oxidized to regenerate TPZ, thus suppressing cytotoxicity in oxygenated cells. Under hypoxia the protonated TPZ radical disproportionates to form the nontoxic twoelectron reduction product 1,2,4-benzotriazin-3-amine 1-oxide (SR 4317), but also decays by a first-order process to a DNA-damaging oxidizing radical. The identity of the latter species is controversial, with evidence supporting decay of the protonated TPZ radical to form either OH• (10 –12) or loss of H2O to form the benzotriazinyl radical, BTZ• (13). Abstraction of H by OH• or BTZ• produces DNA radicals, which lead to the double-strand breaks considered to be the probable cytotoxic lesion (14 –16). Jones and Weinfeld (17) first suggested that TPZ is also involved in a second reaction in the pathway to DNA breakage (Step 2 in Fig. 1) by “fixing” (oxidizing) the initial DNA radicals, thus filling an analogous role to that of O2 in fixing radiation-induced DNA radicals. This “dual action” model was originally based on demonstration of oxidized phosphoglycolate residues in DNA incubated with a TPZ• generating system (TPZ/xanthine/xanthine oxidase) under anoxic conditions, and was subsequently supported by isotope transfer and kinetic studies showing oxidation of photolytically generated DNA radicals by TPZ in the absence of O2 (18, 19). Density function theory modeling supports the view that TPZ is capable of transferring an oxygen to a DNA sugar radical (20). The above studies with cell-free systems demonstrate the ability of TPZ to act as a radical oxidant, but it is not known whether this second step in the dual action of TPZ limits its cytotoxicity in cells. In this study we investigate whether DNA radical oxidation by TPZ limits its hypoxic cytotoxicity by taking advantage of the observation (19) that the nontoxic 2-electron reduction product from TPZ, SR 4317, oxidizes C1⬘-deoxyribosyl DNA radicals 3-fold more rapidly than does TPZ. Thus we investigate whether SR 4317 can potentiate the cytotoxicity of TPZ selectively under hypoxia and whether this might be a useful strategy for enhancing its therapeutic utility as a bioreductive drug. MATERIALS AND METHODS Compounds. TPZ and SR 4317 were synthesized in the Auckland Cancer Society Research Centre (21) and had a purity of ⬎99% by high performance liquid chromatography. Synthesis of SN 29051 will be reported elsewhere. Stock solutions of TPZ, SR 4317, misonidazole [1-(2-nitroimidazol-1-yl)-3methoxypropan-2-ol; MISO; Warner-Lambert, Ann Arbor, MI], and metronidazole (2-methyl-5-nitroimidazole-1-ethanol; METRO; Sigma, St. Louis, MO) at 300 mM in DMSO were stored at ⫺80°C and used for IC50 assays (final DMSO concentration ⱕ0.5%). Fresh solutions of SR 4317, SN 29051, MISO,

736

POTENTIATION OF TIRAPAZAMINE HYPOXIC CYTOTOXICITY

Fig. 1. Mechanism of the proposed dual action of tirapazamine (TPZ). Enzymatic one-electron reduction provides an O2-sensitive TPZ radical (TPZ•), which decays to an oxidizing OH• or benzotriazinyl radical (BTZ•). Oxidative damage to DNA (step 1) results in DNA breaks only after the initial DNA radicals are additionally oxidized by TPZ or SR 4317 (step 2).

and METRO in ␣MEM ⫹10% fetal bovine serum (FBS) ⫹100 IU/ml peniDiffusion through Multicellular Layers (MCLs). HT29 MCLs were inicillin and 100 ␮g/ml streptomycin were prepared for each clonogenic survival tiated by seeding 1 ⫻ 106 cells onto collagen-coated Teflon microporous experiment, with concentrations checked by spectrophotometry. 9(10H)-acri- support membranes (Biopore; average thickness 30 ␮m) in Millicell-CM cell done (Sigma-Aldrich, Castle Hill, NSW, Australia) was dissolved in DMSO. culture inserts (Millipore Corporation, Bedford, MA) and were grown for 3 [14C]Urea (specific activity 2.11 GBq/mmol) was from Amersham Pharmacia days submerged in stirred ␣MEM containing 10% FBS plus antibiotics as Biotech United Kingdom. detailed previously (23, 24). The kinetics of diffusion through HT29 MCLs Cell Lines. Human colon carcinoma HT29 cells, human cervical carcinoma was investigated in a diffusion chamber (25), using a gas phase of 95% O2/5% SiHa cells, human lung adenocarcinoma A549 cells, and human pharyngeal CO2 (oxic) or 95% N2/5% CO2 (anoxic) with TPZ (25 ␮M), SR 4317 (50 ␮M), carcinoma FaDu cells were from American Type Culture Collection. Cell lines or 9(10H)-acridone (10 ␮M) added to the donor compartment along with were maintained in ␣MEM with 5% v/v heat-inactivated FBS, without anti- 14C-urea (2 ␮M) as an internal standard for determination of MCL thickness. biotics, and were confirmed Mycoplasma free by PCR-ELISA (Roche Diag- The donor and receiver compartments were sampled hourly for 5 h. Radioactivity was determined by scintillation counting. TPZ, SR 4317, and 9(10H)nostics GmbH). Cytotoxicity: IC50 Assays. Drug exposures were performed in 96-well acridone were assayed by high performance liquid chromatography as above. plates (Nunc) as described previously (21). Aerobic exposures were in a 37°C The effective thickness of each MCL was determined from the urea flux using humidified incubator, and anoxic exposures in the incubator compartment the known DMCL for urea in HT29 MCLs (9). DMCL for TPZ, SR 4317, and (37°C) of a Pd-catalyst anaerobic chamber (⬍0.001% O2; Shell Lab). For each 9(10H)-acridone were determined by fitting the concentration (c)-time (t) experiment 600 FaDu cells and 1000 SiHa, A549, and HT29 cells were seeded profiles in both the donor and receiver compartment to Fick’s second law as in ␣MEM ⫹10% FBS ⫹100 IU/ml penicillin ⫹100 ␮g/ml streptomycin ⫹10 described previously (Ref. 9; Eq. A). For TPZ under anoxia, metabolic conmM added glucose ⫹100 ␮M 2⬘-dCyd, and allowed to attach for 2 h. Cells were sumption in the MCL was approximated as a first-order process with rate then exposed for 4 h to TPZ plus SR 4317 (0.75 mM) using serial 3-fold constant kmet (Eq. A). dilutions of TPZ in triplicate. Cells were then washed free of drug using the ⭸c ⭸2c same medium without glucose/2⬘-dCyd and with FBS lowered to 5%, and ⫽ D ⫺ kmetc (A) TPZ allowed to grow in air at 37°C for 5 (aerobic exposures) or 6 (anoxic ⭸t ⭸x2 exposures) days. Plates were stained with sulforhodamine B (22), and IC50 Pharmacokinetics and Toxicity in Mice. TPZ or SR 4317 was dissolved values determined as the TPZ concentration required to inhibit cell growth to in DMSO, diluted in 19 volumes saline, and final concentrations determined by 50% of that of untreated controls. Cytotoxicity: Clonogenic Assays. HT29 cells were obtained by trypsin spectrophotometry. TPZ and SR 4317 were administered i.p. to CD-1 nude dissociation of multicellular spheroids grown in spinner flasks in ␣MEM mice (0.02 ml/g body weight), individually or by coadministration, using ⫹10% FBS ⫹100 IU/ml penicillin and 100 ␮g/ml streptomycin. Continuously 1.33-fold dose increments. The maximum tolerated dose (MTD) was deterstirred drug solutions (9 ml at 1.11 ⫻ final concentration) and single-cell mined as the highest dose that did not cause lethality or unacceptable morbidity suspensions (5 ⫻ 106 cells/ml) were pregassed for 60 min under aerobic (with no weight losses ⬎15%) in a group of 6 mice, over an observation period (air/5% CO2), or hypoxic (N2/5% CO2, ⬍0.01% O2) conditions before addition of 28 days. On termination, eyeballs were collected from animals treated at the of 1 ml cells to drug solutions to initiate exposure. Cells (5 ⫻ 105) were MTD, fixed in 10% neutral-buffered formalin for 24 h, and embedded in sampled at various times, washed by centrifugation, and plated to determine paraffin. Retinal toxicity was assessed in H&E sections as described previously colony formation. Plates were stained with methylene blue (0.2% in 50% (26), except that toxicity was quantified using the outer nuclear layer (ONL): inner nuclear layer (INL) thickness ratio, which was measured at intervals of ethanol) 14 days later, and colonies containing ⬎50 cells were counted. TPZ Metabolism in Cell Suspensions. Metabolism of TPZ to SR 4317 ⬃200 ␮m along both retinas from each animal. For pharmacokinetic studies serial blood samples were taken from the tail was assessed in HT29 cell suspensions under the same conditions as for the clonogenic assays, except the cell density was 2 ⫻ 106 cells/ml and FBS was vein of restrained animals using a small incision to produce a droplet of blood not used. SR 4317 was assayed in samples of extracellular medium using high (10 –20 ␮l), which was collected in a heparinized hematocrit tube. The tube performance liquid chromatography, with an Alltima C8 5 ␮m column was sealed with hematocrit putty, held on ice for up to 0.5 h, and centrifuged. (150 ⫻ 2.1 mm; Alltech, Deerfield, IL) and an Agilent HP1100 equipped with Plasma samples (5 ␮l) were transferred to a microcentrifuge tube, and 400 ␮l a diode-array detector. Mobile phases were gradients of 80% acetonitrile (v/v) of ice-cold acetonitrile (80%; v/v) containing an internal standard [bis(2in 0.45 M ammonium formate (pH 4.5) at 0.3 ml/min. Quantitation was based chloroethyl)-N-methyl-N-(4-methyl-2-nitrobenzyl)-ammonium chloride] was added to deproteinize the sample. The samples were centrifuged briefly, and on calibration curves in culture medium buffer. Radiosensitizing Efficiency. Stirred suspensions of HT29 cells and drug the supernatant was reduced to approximately 30 –50 ␮l in a Speed-Vac solutions were gassed in a 37°C waterbath for 90 min under hypoxia as above. concentrator (Savant Instruments, Farmingdale, NY). Ammonium formate (pH After the addition of 1-ml cells to drug solutions to give a final drug concen- 4.5; 0.45 M) was added to give a final volume of ⬃200 ␮l, and 100 ␮l was tration of 0.6 mM, hypoxic cell suspensions were irradiated with an external analyzed by high performance liquid chromatography as above. Concentrabeam cobalt-60 unit (dose rate 1.2 Gy/min). Cells were sampled after each tions were calculated using a calibration curve of plasma samples spiked with dose increment (6 Gy), washed by centrifugation, and plated to determine cell known drug concentrations and assayed under the same conditions. On the survival by colony formation as above. Radiation survival curves were fitted basis of comparison of plasma calibration curves with standards in formate buffer, the recovery from plasma was ⬎95%. Plasma concentrations of TPZ using the linear-quadratic model. 737


Fig. 2. Rate of killing of (A) aerobic and (B) anoxic HT29 cell suspensions (5 ⫻ 105 cells/ml) exposed to tirapazamine (TPZ) alone, SR 4317 alone, or TPZ ⫹ SR 4317: A, F, control; E, 0.9 mM SR 4317; f, 2 mM TPZ; 䡺, 2 mM TPZ ⫹0.9 mM SR 4317. B, F, control; E, 300 ␮M SR 4317; ƒ, 1000 ␮M SR 4317; f, 30 ␮M TPZ; 䡺, 30 ␮M TPZ ⫹30 ␮M SR 4317; ‚, 30 ␮M TPZ ⫹300 ␮M SR 4317; 〫, 30 ␮M TPZ ⫹1000 ␮M SR 4317.

and SR 4317 were analyzed using a noncompartmental model with WinNonlin (v. 3.1; Pharsight Corporation, Mountain View, CA), and the area under the concentration-time curve was calculated for each animal by using the linear/ log trapezoidal rule and extrapolation to infinity. Tumor Excision Assay. HT29 xenografts were grown in CD-1 nude mice by s.c. inoculation of 107 cells (prepared from spheroids as above). Mice were treated with single i.p. drug doses when tumors reached ⬃0.3 g, either alone or 5 min after whole body irradiation (20 Gy; Co-60) without restraint or anesthesia. Tumors were excised 18 h after treatment, minced, and dissociated by stirring in a 37°C waterbath for 45 min using an enzyme mixture (Pronase; Sigma, 2.5 mg/ml), collagenase (Sigma; 1 mg/ml) and DNAase I (Sigma; 0.2 mg/ml)) in culture medium, added at 1 ml/50 mg tumor. Cells were enumerated using an electronic particle counter (Beckman Coulter; model Z2), diluted and plated to determined clonogenic survival as above.

(Table 1). In these experiments, cells were exposed to TPZ at a range of concentrations for 4 h under aerobic or hypoxic conditions, with or without SR 4317 at 0.75 mM (which caused no detectable inhibition of proliferation under either aerobic or hypoxic conditions in the absence of TPZ; data not shown). With all four of the lines, SR 4317 had no effect on the TPZ IC50 under aerobic conditions, but in each case gave a decrease in the TPZ IC50 under hypoxia. The ratio of hypoxic IC50 values with and without SR 4317 (tirapazamine potentiation ratio values) ranged from 2.3 for FaDu cells to 4.7 for SiHa. As a consequence of this selective potentiation under hypoxia, the hypoxic cytotoxicity ratio values for three of the four cell lines (HT29, SiHa, and FaDu) increased significantly, from 52–112 in the absence of SR 4317 to 133–331 in its presence. For the fourth cell line, A549, addition of SR 4317 increased the hypoxic cytotoxicity ratio from 80 to 288, although this did not quite reach statistical significance. The dose-response for enhancement of the hypoxic toxicity of TPZ, assessed by clonogenic assay of HT29 cell suspensions, confirmed this potentiation and showed the similar activity of SN 29051, an SR 4317 analogue with a basic dimethylaminoethyl sidechain (Fig. 3A). The enhancement of the hypoxic cytotoxicity of TPZ by these benzotriazine mono-N-oxides could, in principle, be due to an enhancement of its reduction to the TPZ radical. To test this possibility, we examined the effect of SN 29051 on reduction of TPZ by hypoxic HT29 cells, quantifying SR 4317 as the major stable end product from one-electron reduction to the TPZ radical (Fig. 3B). Using SN 29051 rather than SR 4317 as potentiator made it possible to avoid the

RESULTS To evaluate whether the hypoxic cytotoxicity of TPZ can be potentiated by its metabolite SR 4317, HT29 cells were exposed to TPZ in the presence of SR 4317 at concentrations up to its solubility limit of ⬃0.9 mM, and killing was assessed by clonogenic assay (Fig. 2). SR 4317 itself was nontoxic at this concentration under either aerobic or hypoxic conditions. Under aerobic conditions SR 4317 (0.9 mM) had no effect on the rate of killing by 2 mM TPZ (Fig. 2A), but under hypoxia it caused a concentration-dependent increase in killing (Fig. 2B). To confirm this hypoxia-selective potentiation of TPZ cytotoxicity, the ability of SR 4317 to enhance inhibition of cell proliferation by TPZ was evaluated in four human tumor cell lines using IC50 assays

Fig. 3. A, potentiator dose dependence for killing of anoxic HT29 cells (5 ⫻ 105 cells/ml) after a 1-h exposure to E, SR 4317 alone; ‚, misonidazole [1-(2-nitroimidazol1-yl)-3-methoxypropan-2-ol] (MISO) alone; 䡺, metronidazole (2-methyl-5-nitroimidazole-1-ethanol (METRO) alone; 〫, SN 29051 alone; F, 30 ␮M tirapazamine (TPZ) ⫹ SR 4317; Œ, 30 ␮M TPZ ⫹ MISO; f, 30 ␮M TPZ ⫹ METRO; ⽧, 30 ␮M TPZ ⫹ SN 29051. B, metabolism of TPZ (30 ␮M) to SR 4317 by anoxic HT29 cells (2 ⫻ 106 cells/ml) in the presence of 0 (E), 0.03 (F), 0.1 (‚), 0.3 (Œ), or 1.0 mM (f) SN 29051. Values are mean and range for duplicate determinations; bars, ⫾SD. Lines are regression lines for firstorder metabolism.

Table 1 Potentiation of tirapazamine (TPZ) cytotoxicity by SR 4317 (0.75 mM) as determined by IC50 assay

Cell line HT29 SiHa FaDu A549

Drug TPZ TPZ TPZ TPZ TPZ TPZ TPZ TPZ

⫹ SR 4317 ⫹ SR 4317 ⫹ SR 4317 ⫹ SR 4317

Aerobic IC50 (␮M)a 318 ⫾ 141 493 ⫾ 69f 337 ⫾ 49 243 ⫾ 49f 314 ⫾ 96 320 ⫾ 50f 329 ⫾ 132 444 ⫾ 164f

Aerobic tirapazamine potentiation ratio (TPR)b

Hypoxic IC50 (␮M)c

Hypoxic TPRb

6.2 ⫾ 1.0 2.3 ⫾ 0.4 4.3 ⫾ 1.8 0.9 ⫾ 0.4b 6.2 ⫾ 2.1 2.9 ⫾ 1.2b 4.7 ⫾ 0.4 1.7 ⫾ 0.4b

0.8 ⫾ 0.3 1.4 ⫾ 0.3 1.0 ⫾ 0.1 0.8 ⫾ 0.2

a

TPZ concentration required to reduce cell density to 50% of that of the controls, for a 4-h aerobic drug exposure. b TPZ potentiation ratio (IC50 for TPZ alone: IC50 for TPZ ⫹ SR 4317). c TPZ concentration required to reduce cell density to 50% of that of the controls, for a 4-h hypoxic drug exposure. d Hypoxic cytotoxicity ratio (aerobic IC50: hypoxic IC50). e Significance of difference between HCR with and without SR 4317. f IC50 value for TPZ in the presence of a fixed concentration of SR 4317 (0.75 mM).

738

2.6 ⫾ 0.2 4.7 ⫾ 0.2 2.3 ⫾ 0.2 2.8 ⫾ 0.4

Hypoxic cytotoxicity ratio (HCR)d 66 ⫾ 10 242 ⫾ 35 112 ⫾ 45 331 ⫾ 74 52 ⫾ 7 133 ⫾ 30 80 ⫾ 48 288 ⫾ 123

Pe 0.017 0.016 0.034 0.091


Fig. 4. Radiation survival curves for anoxic HT29 cell suspensions (5 ⫻ 105 cells/ml) irradiated with cobalt-60 ␥-rays in the presence of E, no radiosensitizer; 〫,⽧, 0.6 mM SR 4317 (dashed line); 䡺,f, 0.6 mM metronidazole (2-methyl-5-nitroimidazole-1-ethanol; METRO); ‚,Œ, 0.6 mM misonidazole [1-(2-nitroimidazol-1-yl)-3-methoxypropan-2-ol] (MISO). Curves are fitted using a linear-quadratic model.

problem of high background SR 4317 concentrations, facilitating measurement of SR 4317 formation. The fitted first order rate constant for SR 4317 formation was increased by 40% (0.17 versus 0.12 h-1) at the highest concentration of SN 29051 (1 mM), but there was no significant increase in SR 4317 formation at 0.03, 0.1, or 0.3 mM despite clear potentiation of cytotoxicity (Fig. 3A) at the latter concentration. Given that nitroimidazole radiosensitizers are considered to enhance radiation damage by oxidizing DNA radicals (27), we tested whether the 2-nitroimidazole MISO and the 5-nitroimidazole METRO also potentiate the hypoxic cytotoxicity of TPZ, using the HT29 clonogenic assay (Fig. 3A). The nitroimidazoles caused potentiation at nontoxic concentrations, but with an ⬃20-fold lower molar potency than the benzotriazine mono-N-oxides. Investigation of higher concentrations of the nitroimidazoles was precluded by their toxicity under hypoxic conditions. The much greater effect of SR 4317 as a potentiator of TPZ cytotoxicity was unexpected given that MISO is at least as effective as SR 4317 in oxidizing photolytically generated C1⬘ DNA radicals (19). Therefore, we compared the ability of all three of the radical oxidants (at 0.6 mM) to radiosensitize hypoxic HT29 cells, using identical experimental conditions to those for the above cytotoxicity (clono-

genic) assays (Fig. 4). This confirmed the known greater radiosensitizing efficiency of MISO than METRO (28) and showed that SR 4317 is similar in potency to METRO as a hypoxic cell radiosensitizer. These findings suggested the possibly utility of exogenous SR 4317 as a selective potentiator of the hypoxic cytotoxicity of TPZ in vivo. To evaluate whether SR 4317 has suitable pharmacological properties, we determined its plasma pharmacokinetics in CD-1 nude mice. First, we established the MTD of i.p. SR 4317 alone (MTD ⫽ 1.33 mmol/ kg), TPZ alone (MTD ⫽ 0.316 mmol/kg), and SR 4317 coadministered with a subeffective dose of TPZ (0.133 mmol/kg). The MTD of SR 4317 in the latter combination was 1.0 mmol/kg. The pharmacokinetics of SR 4317 (1.0 mmol/kg), TPZ (0.133 mmol/kg), and the combination is shown in Fig. 5. I.p administration of SR 4317 alone gave plasma concentrations that reached ⬃80 ␮M by 30 min and remained ⬎50 ␮M for 2 h. TPZ was identified as a minor metabolite of SR 4317, based on its retention time and absorbance spectrum, reaching ⬃1.6 ␮M. TPZ alone was cleared rapidly and resulted in average SR 4317 plasma concentrations of up to 25 ␮M. Coadministration of both compounds gave enhanced SR 4317 concentrations (still ⬎100 ␮M at 2 h) but had little if any effect on TPZ pharmacokinetics; the TPZ area under the concentration-time curve for TPZ alone (46.2 ⫾ 2.1 ␮M.h; mean ⫾ SE, n ⫽ 5) was not significantly different (P ⫽ 0.330, two-tailed t test) from that for TPZ ⫹ SR 4317 (55.1 ⫾ 9.3 ␮M.h; n ⫽ 4). Thus, it is possible to achieve concentrations of SR 4317 in plasma that are well above the levels seen with TPZ alone and that are consistent with significant potentiation of the hypoxic cytotoxicity of TPZ (Fig. 3). We investigated the extravascular transport properties of SR 4317 in the MCL model to assess whether it is likely to reach hypoxic cells in tumor tissue after systemic administration (Fig. 6). The flux of TPZ through oxic MCLs gave an estimated diffusion coefficient (DMCL) of (0.41 ⫾ 0.02) ⫻ 10⫺6 cm2s⫺1 (n ⫽ 3). Transport of TPZ through MCLs was much slower under anoxic conditions, which could be accounted for by assuming metabolic consumption under anoxia with an apparent first order rate constant kmet of (0.75 ⫾ 0.06) ⫻ min⫺1 in the MCL. These values are in good agreement with previous determinations (9). Flux of SR 4317 through HT29 MCLs was the same under oxic and anoxic conditions, with a fitted value of DMCL of (2.96 ⫾ 0.12) ⫻ 10⫺6 cm2s⫺1 (n ⫽ 6). This was 7.5-fold higher than the aerobic TPZ value, and similar to the lipophilic internal standard 9(10H)-acridone [DMCL ⫽ (3.62 ⫾ 0.13) ⫻ 10⫺6 cm2s⫺1 (n ⫽ 6)]. Given these favorable pharmacokinetic properties, the ability of exogenous SR 4317 to potentiate hypoxic cell killing by TPZ in HT29 tumors was evaluated by clonogenic assay 18 h after treatment of

Fig. 5. Plasma concentrations of A, tirapazamine (TPZ) and B, SR 4317, as measured by high-performance liquid chromatography, after i.p. administration of TPZ alone (0.133 mmol/kg), SR 4317 alone (1.0 mmol/kg), or TPZ ⫹ SR 4317 coadministered. Points are mean for 4 –5 mice; bars, ⫾SE.

739


Fig. 6. Transport of SR 4317 and tirapazamine (TPZ) through HT29 multicellular layers. Concentrations are shown in the “receiver” (distal) compartment of a diffusion chamber, normalized against the initial concentrations in the donor compartment (C0). The abscissa is the flux of the internal standard, 14C-urea; this provides a time axis corrected for differences in multicellular layer (MCL) thickness. Open symbols, oxic MCLs (95% O2/5% CO2); filled symbols, anoxic MCLs (95% N2/5% CO2). Lines are fits to the diffusion-reaction model used to determine DMCL and kmet. A, TPZ (C0 ⫽ 25 ␮M). B, SR 4317 (C0 ⫽ 50 ␮M). C, 9(10H)-acridone (C0 ⫽ 10 ␮M).

initial DNA radicals. There is solid evidence in cell-free systems that TPZ or SR 4317 can effect this second oxidation by oxygen transfer to deoxyribosyl radicals (19). The present findings strongly suggest that the DNA radical oxidation reaction is a limiting step in the development of the ultimate cytotoxic lesion. This is consistent with our recent demonstration that the rate of anoxic cell killing by TPZ is proportional to the square of its concentration, which is also compatible with the dual action model (9). A possible alternative explanation for the observed potentiation would be that SR 4317 enhances the one-electron reduction of TPZ. There is no obvious mechanism for such an enhancement, but surprisingly we did find a small increase in metabolism of TPZ to SR 4317 in the presence of a high concentration of the SR 4317 analogue SN 29051. However, this appears not to be the main mechanism of the interaction because potentiation of cytotoxicity is seen at 0.3 mM SN 29051 (Fig. 3A) without enhancement of TPZ metabolism (Fig. 3B). In addition, the ability of SR 4317 to potentiate cell killing by radiation (Fig. 4) confirms its activity as a DNA radical oxidant. Our findings leave open the question as to whether metabolic formation of SR 4317 from TPZ contributes to DNA radical fixation

mice at the same drug doses as used in the pharmacokinetic study (Fig. 7). In this experiment, a single large dose of ionizing radiation (20 Gy) was used to sterilize oxic cells, and the ability of the drugs to kill the hypoxic survivors was assessed by i.p. injection immediately after irradiation. At this suboptimal dose, TPZ alone did not cause statistically significant hypoxic cell killing. SR 4317 was also without activity, either with TPZ or with radiation, but showed marked activity when combined with TPZ postirradiation. Thus, SR 4317 is able to enhance hypoxic cell killing by TPZ in HT29 human tumor xenografts. Previous studies have shown that a single i.p. dose of TPZ at the MTD causes extensive loss of photoreceptors in C57Bl6 mice as a result of hypoxia in the retina (26). We confirmed this toxicity of TPZ (0.316 mmol/kg) in CD-1 nude mice, with marked thinning of the ONL in some animals and complete loss in others (Fig. 8). In contrast, the combination of TPZ with SR 4317 (at the MTD of the combination, 0.133 mmol/kg TPZ, 1.0 mmol/kg SR 4317) did not cause retinal toxicity in these animals (Fig. 8). The ratio of thickness of ONL/INL for mice treated with TPZ ⫹ SR 4317 (1.36 ⫾ 0.11, mean ⫾ SD) was not significantly different (P ⫽ 0.2, t test) from controls (1.49 ⫾ 0.14). Thus, the combination of TPZ with SR 4317 permits use of a lower dose of TPZ that gives activity against hypoxic cells in HT29 tumors, while reducing toxicity against hypoxic cells in a critical normal tissue.

Fig. 7. Tumor excision assay of HT29 xenografts treated with radiation alone (RAD, 20 Gy whole body); RAD ⫹ tirapazamine (TPZ; 0.133 mmol/kg); RAD ⫹ SR 4317 (1 mmol/kg); TPZ ⫹ SR 4317; or RAD ⫹ TPZ ⫹ SR 4317. Tumors were excised 18 h after treatment and clonogenic survival determined by staining colonies 14 days later. Points are mean for 4 –5 mice; bars, ⫾SE. * P ⬍ 0.05 versus Rad ⫹ SR 4317; ** P ⬍ 0.01 versus RAD ⫹ TPZ. Horizontal lines are the historical means (solid) and 95% confidence limits (dashed) for untreated controls (upper set) and 20 Gy radiation only (lower set).

DISCUSSION The present study demonstrates that the stable 2-electron reduced metabolite of TPZ, SR 4317, enhances the cytotoxicity of TPZ selectively under hypoxic conditions when added exogenously to cell cultures. This potentiation was seen in all four of the human tumor lines examined and resulted in hypoxic cytotoxicity ratio values 2–3-fold higher than in the absence of SR 4317. This finding is consistent with the dual action model for TPZ, which proposes that two distinct steps are required to generate cytotoxic DNA lesions (17, 18). In this model the first step is H-abstraction from DNA by an oxidizing radical derived from the initial one-electron reduced TPZ radical, and the second step is the additional oxidation (fixation) of the 740

Fig. 8. Retinal toxicity of tirapazamine (TPZ) alone and in combination with SR 4317 after single i.p. doses to CD-1 nude mice. A, control retina. B, 28 days after treatment with TPZ at its maximum tolerated dose (0.316 mmol/kg). C, 28 days after treatment with the combination at its maximum tolerated dose (TPZ 0.133 mmol/kg ⫹ SR 4317 1.0 mmol/kg).


in cell culture and tumors, or whether TPZ itself is the main mediator of the second step. On the basis of recent studies of SR 4317 formation from TPZ in HT29 cultures (9), the SR 4317 concentrations from TPZ under the conditions of Fig. 3 (30 ␮M TPZ, 5 ⫻ 105 cells/ml) would be ⬃1 ␮M at the end of the 1-h exposure period. The dose-response curve for exogenous SR 4317 (Fig. 3) shows only a marginal effect on addition of 30 ␮M SR 4317, so it seems unlikely that metabolic formation of SR 4317 makes a major contribution in vitro, although it may play a larger role in tumors when the SR 4317:TPZ ratio is higher (29). The demonstration that exogenous SR 4317 augments hypoxic cytotoxicity of TPZ selectively in vitro suggests the potential of relatively nontoxic benzotriazine mono-N-oxides as therapeutic agents in combination with TPZ. We show that SR 4317 itself has suitable pharmacological properties in this context; it strongly augments SR 4317 plasma concentrations at well-tolerated doses, relative to endogenous levels, without modifying the pharmacokinetics of TPZ significantly (Fig. 5). We observed that SR 4317 is oxidized to TPZ in mice (Fig. 5), which has not been reported previously, but this appears to be a minor pathway, and no other SR 4317 metabolites were observed. Importantly, SR 4317 diffused rapidly through HT29 MCLs under both aerobic and hypoxic conditions (Fig. 6). In fact, the diffusion coefficient of SR 4317 (2.96 ⫻ 10⫺6 cm2s⫺1) was similar to the very lipophilic marker 9(10H)-acridone (3.62 ⫻ 10⫺6 cm2s⫺1) and probably close to the theoretical limit for a molecule of this size in HT29 tissue.4 Thus, SR 4317 has excellent extravascular transport properties, and exogenously administered SR 4317 can be expected to considerably elevate SR 4317 concentrations in hypoxic regions of tumors. This expectation is borne out by the potentiation of hypoxic cell killing in HT29 tumors when SR 4317 is coadministered with a subtherapeutic dose of TPZ, whereas the lack of effect of SR 4317 ⫹ TPZ in the absence of radiation indicates that this combination does not cause oxic cell killing (Fig. 7). Clearly, these results will require extension to other xenograft models, and the toxicology of the combination will require more detailed evaluation to assess effects on TPZ toxicity to normal tissues. In this regard, it is interesting to find that the combination of TPZ ⫹ SR 4317 lacks the characteristic retinal toxicity of TPZ when both the combination and single agent are compared at their respective MTDs (Fig. 8). Thus, addition of SR 4317 to TPZ makes it possible to achieve hypoxic cell killing in HT29 tumors without damaging a critical hypoxic normal tissue. Despite these encouraging results, SR 4317 itself is not appropriate for clinical investigation because of its low aqueous solubility (approximately 0.9 –1 mM); in the present studies it was administered as a fine suspension in 5% DMSO/saline. In seeking a clinical candidate, the question then arises as to what structural/physicochemical features are required for efficient “step 2” potentiation of the hypoxic cytotoxicity of TPZ. Given that nitroimidazoles are well-characterized DNA radical oxidants (which underlies their use as hypoxic cell radiosensitizers) and have already received extensive clinical evaluation with radiotherapy (30), we investigated whether MISO or METRO would act as efficient TPZ potentiators. On the basis of its 2-fold higher rate constant for reaction with a C1⬘ radical in DNA (19), we expected MISO to be more efficient than SR 4317 as a TPZ potentiator, but both the nitroimidazoles were ⬃20-fold less potent that SR 4317 in this assay (Fig. 3). This surprising result led us to compare radiosensitization of hypoxic HT29 cells under identical experimental conditions (Fig. 4). The structure-activity relationships 4 F. B. Pruijn, J. R. Sturman, H. D. S. Liyanage, K. O. Hicks, M. P. Hay, and W. R. Wilson. Extravascular transport of drugs in tumor tissue: effect of lipophilicity on diffusion of tirapazamine analogs in multicellular layer cultures. J. Med. Chem., submitted for publication.

for radiosensitization (MISO ⬎ METRO ⬇ SR 4317) were different from TPZ potentiation (SR 4317 ⬎ MISO ⬇ METRO). The corollary of this observation would seem to be that the primary DNA damage by radiation (which is due primarily to clustered OH radicals; Ref. 31) and metabolically reduced TPZ must be different. This supports the hypothesis that the key DNA-damaging radical from TPZ is the benzotriazinyl radical BTZ• as proposed by Anderson et al. (13) rather than the OH radical (Fig. 1). An alternative possibility is that the key target for TPZ radical damage is not DNA itself but an associated molecule such as topoisomerase II, which has been implicated in TPZinduced DNA breakage and cytotoxicity under hypoxia (32). We are currently screening benzotriazine-N-oxides with improved formulation properties and have identified recently analogs such as SN 29051, which provide higher tirapazamine potentiation ratios at their solubility limits. The mechanism of potentiation, therapeutic activity, and toxicology of these analogs are under investigation to define more clearly the therapeutic potential of this class in combination with TPZ for radiotherapy and chemoradiation. ACKNOWLEDGMENTS We thank Elaine Wong and Anna Chappell for assistance with the IC50 and retinal toxicity assays, Jane Botting for assistance with the excision assay and toxicity testing in mice, and Alfred Degenkolbe and Sarath Liyanage for assistance with the pharmacokinetic study.

REFERENCES 1. Brown, J. M. The hypoxic cell: a target for selective cancer therapy - eighteenth Bruce F. Cain Memorial Award lecture. Cancer Res., 59: 5863–5870, 1999. 2. Harris, A. L. Hypoxia - a key regulatory factor in tumour growth. Nature Rev. Cancer (Phila.), 2: 38 – 47, 2002. 3. von Pawel, J., von Roemeling, R., Gatzemeier, U., Boyer, M., Elisson, L. O., Clark, P., Talbot, D., Rey, A., Butler, T. W., Hirsh, V., Olver, I., Bergman, B., Ayoub, J., Richardson, G., Dunlop, D., Arcenas, A., Vescio, R., Viallet, J., and Treat, J. Tirapazamine plus cisplatin versus cisplatin in advanced non-small-cell lung cancer: A report of the international CATAPULT I study group. J. Clin. Oncol., 18: 1351– 1359, 2000. 4. Rischin, D., Peters, L., Smith, J., Macann, A., Denham, J., Poulsen, M., Jackson, M., Kenny, L., Penniment, M., and Fisher, R. Preliminary results of TROG 98.02 - a randomized phase II study of 5-fluorouracil, cisplatin and radiation versus tirapazamine, cisplatin and radiation for advanced squamous cell carcinoma of the head and neck. Proc. Am. Soc. Clin. Oncol., 22(Suppl.): 495– 499, 2003. 5. Durand, R. E., and Olive, P. L. Physiologic and cytotoxic effects of tirapazamine in tumor-bearing mice. Radiat. Oncol. Investig., 5: 213–219, 1997. 6. Durand, R. E., and Olive, P. L. Evaluation of bioreductive drugs in multicell spheroids. Int. J. Radiat. Oncol. Biol. Phys., 22: 689 – 692, 1992. 7. Hicks, K. O., Fleming, Y., Siim, B. G., Koch, C. J., and Wilson, W. R. Extravascular diffusion of tirapazamine: effect of metabolic consumption assessed using the multicellular layer model. Int. J. Radiat. Oncol. Biol. Phys., 42: 641– 649, 1998. 8. Kyle, A. H., and Minchinton, A. I. Measurement of delivery and metabolism of tirapazamine to tumor tissue using the multilayered cell culture model. Cancer Chemother. Pharmacol., 43: 213–220, 1999. 9. Hicks, K. O., Pruijn, F. B., Sturman, J. R., Denny, W. A., and Wilson, W. R. Multicellular resistance to tirapazamine is due to restricted extravascular transport: a pharmacokinetic/pharmacodynamic study in HT29 multicellular layers. Cancer Res., 63: 5970 –5977, 2003. 10. Daniels, J. S., and Gates, K. S. DNA Cleavage by the Antitumor agent 3-amino-1, 2, 4-benzotriazine 1, 4-dioxide (SR4233): Evidence for involvement of hydroxyl radical. J. Am. Chem. Soc., 118: 3380 –3385, 1996. 11. Kotandeniya, D., Ganley, B., and Gates, K. Oxidative DNA base damage by the antitumor agent 3-amino-1, 2, 4- benzotriazine 1, 4-Dioxide (Tirapazamine). Bioorg. Med. Chem. Lett., 12: 2325, 2002. 12. Zagorevskii, D., Song, M., Breneman, C., Yuan, Y., Fuchs, T., Gates, K. S., and Greenlief, C. M. A mass spectrometry study of tirapazamine and its metabolites: Insights into the mechanism of metabolic transformations and the characterization of reaction intermediates. J. Am. Soc. Mass Spectrom., 14: 881– 892, 2003. 13. Anderson, R. F., Shinde, S. S., Hay, M. P., Gamage, S. A., and Denny, W. A. Activation of 3-amino-1, 2, 4-benzotriazine 1, 4-dioxide antitumor agents to oxidizing species following their one-electron reduction. J. Am. Chem. Soc., 125: 748 –756, 2003. 14. Biedermann, K. A., Wang, J., Graham, R. P., and Brown, J. M. SR 4233 cytotoxicity and metabolism in DNA repair-competent and repair- deficient cell cultures. Br. J. Cancer, 63: 358 –362, 1991. 15. Wang, J., Biedermann, K. A., and Brown, J. M. Repair of DNA and chromosome breaks in cells exposed to SR 4233 under hypoxia or to ionizing radiation. Cancer Res., 52: 4473– 4477, 1992.

741


16. Delahoussaye, Y. M., Evans, J. W., and Brown, J. M. Metabolism of tirapazamine by multiple reductases in the nucleus. Biochem. Pharmacol., 62: 1201–1209, 2001. 17. Jones, G. D., and Weinfeld, M. Dual action of tirapazamine in the induction of DNA strand breaks. Cancer Res., 56: 1584 –1590, 1996. 18. Daniels, J. S., Gates, K. S., Tronche, C., and Greenberg, M. M. Direct evidence for bimodal DNA damage induced by tirapazamine. Chem. Res. Toxicol., 11: 1254 – 1257, 1998. 19. Hwang, J. T., Greenberg, M. M., Fuchs, T., and Gates, K. S. Reaction of the hypoxia-selective antitumor agent tirapazamine with a C1⬘-radical in single-stranded and double-stranded DNA: the drug and its metabolites can serve as surrogates for molecular oxygen in radical- mediated DNA damage reactions. Biochemistry, 38: 14248 –14255, 1999. 20. Ban, F., Gauld, J. W., and Boyd, R. J. Modeling the action of an antitumor drug: a density functional theory study of the mechanism of tirapazamine. J. Am. Chem. Soc., 123: 7320 –7325, 2001. 21. Hay, M. P., Gamage, S. A., Kovacs, M. S., Pruijn, F. B., Anderson, R. F., Patterson, A. V., Wilson, W. R., Brown, J. M., and Denny, W. A. Structure-activity relationships of 1, 2, 4-benzotriazine 1, 4-dioxides as hypoxia-selective analogues of tirapazamine. J. Med. Chem., 46: 169 –182, 2003. 22. Skehan, P., Storeng, R., Scudiero, D., Monks, A., McMahon, J., Vistica, D., Warren, J. T., Bokesch, H., Kenney, S., and Boyd, M. R. New colorimetric cytotoxicity assay for anticancer-drug screening. J. Natl. Cancer Inst., 82: 1107–1112, 1990. 23. Hicks, K. O., Ohms, S. J., van Zijl, P. L., Denny, W. A., Hunter, P. J., and Wilson, W. R. An experimental and mathematical model for the extravascular transport of a DNA intercalator in tumours. Br. J. Cancer, 76: 894 –903, 1997.

24. Wilson, W. R., Pullen, S. M., Hogg, A., Hobbs, S. M., Pruijn, F. B., and Hicks, K. O. In vitro and in vivo models for evaluation of GDEPT: quantifying bystander killing in cell cultures and tumors. In: C. J. Springer (ed.), Suicide Gene Therapy: Methods and Reviews, pp. 403– 432. Totowa: Humana Press, 2003. 25. Hicks, K. O., Pruijn, F. B., Baguley, B. C., and Wilson, W. R. Extravascular transport of the DNA intercalator and topoisomerase poison N-[2-(dimethylamino)ethyl]acridine-4-carboxamide (DACA): diffusion and metabolism in multicellular layers of tumor cells. J. Pharmacol. Exp. Ther., 297: 1088 –1098, 2001. 26. Lee, A. E., and Wilson, W. R. Hypoxia-dependent retinal toxicity of bioreductive anticancer prodrugs in mice. Toxicol. Appl. Pharmacol., 163: 50 –59, 2000. 27. Wardman, P. The mechanism of radiosensitization by electron-affinic compounds. Radiat. Phys. Chem., 30: 423– 432, 1987. 28. Denekamp, J., and Fowler, J. F. Radiosensitization of solid tumors by nitroimidazoles. Int. J. Radiat. Oncol. Biol. Phys., 4: 143–151, 1978. 29. Walton, M. I., and Workman, P. Pharmacokinetics and bioreductive metabolism of the novel benzotriazine di-N-oxide hypoxic cell cytotoxin tirapazamine (WIN 59075; SR 4233; NSC 130181) in mice. J. Pharmacol. Exp. Ther., 265: 938 –947, 1993. 30. Overgaard, J., and Horsman, M. R. Modification of hypoxia-induced radioresistance in tumors by the use of oxygen and sensitizers. Semin. Radiat. Oncol., 6: 10 –21, 1996. 31. Ward, J. F. Biochemistry of DNA lesions. Radiat. Res., 104: S103–S111, 1985. 32. Peters, K. B., and Brown, J. M. Tirapazamine: A hypoxia-activated topoisomerase II poison. Cancer Res., 62: 5248 –5253, 2002.

742