Enhanced Conversion of DNA Radical Damage to Double Strand

Oct 21, 2003 - Targeting the anticancer compound tirapazamine (3-amino-1,2,4-benzotriazine 1,4-dioxide; TPZ) to DNA by appended binding units has been...
0 downloads 0 Views 133KB Size
Download PDF
Chem. Res. Toxicol. 2003, 16, 1477-1483

1477

Enhanced Conversion of DNA Radical Damage to Double Strand Breaks by 1,2,4-Benzotriazine 1,4-Dioxides Linked to a DNA Binder Compared to Tirapazamine Robert F. Anderson,*,†,‡ Tracy A. Harris,† Michael P. Hay,‡ and William A. Denny‡ Departments of Chemistry and Auckland Cancer Society Research Centre, The University of Auckland, Private Bag 92019, Auckland 1, New Zealand Received June 10, 2003

Targeting the anticancer compound tirapazamine (3-amino-1,2,4-benzotriazine 1,4-dioxide; TPZ) to DNA by appended binding units has been found to greatly increase the free radicalinduced production of both single and double strand breaks under hypoxia compared to TPZ itself. The •OH radical, produced upon the radiolysis of aqueous solutions, was used to damage plasmid DNA, and both types of strand breaks were quantified in the absence and presence of TPZ and analogues. Targeted analogues of TPZ show increases of 12-18-fold in single strand breaks, and 60-110-fold in double strand breaks, as compared with TPZ itself. The observed increased formation of double strand breaks under hypoxia is the likely mechanism for the large increase in potency previously demonstrated for a similarly targeted analogue of TPZ as a bioreductive drug (Delahoussaye et al. (2003) Biochem. Pharmacol. 65, 1807-1815). The one-electron reduction potential of the two-electron reduced metabolite of TPZ (the 1-oxide, SR 4317) has been measured as -568 ( 9 mV, which is sufficiently high to oxidize carboncentered radicals such as those formed on the sugar moiety of DNA. Targeting the 1-oxide moiety to DNA resulted in a ca. 50% increase in single strand breaks over that seen for TPZ without the dramatic increase in double strand breaks seen for the similarly targeted benzotriazine 1,4-dioxides. These studies support the mechanism by which the reduction of TPZ to an oxidizing radical leads to free radical damage on DNA that can be further oxidized by TPZ or SR4317 (and especially well by DNA-targeted analogues) to yield lesions resulting in strand breakage. The targeting of benzotriazine 1,4-dioxide analogues to DNA by appending binding units to the compounds thus represents an efficient system for inducing strand breaks in DNA.

Introduction Tirapazamine (3-amino-1,2,4-benzotriazine 1,4-dioxide; TPZ)1 (1) is the most advanced anticancer bioreductive drug, currently in Phase II/III clinical trials in combination with radiotherapy and also with cisplatin based chemotherapy (1-3). TPZ has a unique mechanism of action, mediated through DNA double strand breaks generated from the action of an oxidizing radical (4). TPZ is reduced by one-electron reductases (5-9) to initially form a radical anion (2), which in aerobic cells, is back oxidized by molecular oxygen (10) to generate the super* To whom correspondence should be addressed. Tel.: (64) 9 3737599 ext 88315 Fax: (64) 9 3737422. E-mail: [email protected]. † Department of Chemistry. ‡ Auckland Cancer Society Research Centre. 1 Abbreviations: compound 1, TPZ, 3-amino-1,2,4-benzotriazine 1,4dioxide; compound 4, SR 4317, 3-amino-1,2,4-benzotriazine 1-oxide; compound 6, SN 28141, N-{3-[{3-[1,4-dioxido-1,2,4-benzotriazin-3-yl)amino]propyl}(methylamino]propyl-4-acridine carboxamide; compound 7, SN 27149, N-{3-[{3-[1,4-dioxido-1,2,4-benzotriazin-3-yl)amino]propyl}(amino]propyl-4-acridine carboxamide; compound 8, SN 27140, N-{2-[(1,4-dioxido-1,2,4-benzotriazin-3-yl)amino]ethoxy}ethyl)4-acridine carboxamide; compound 9, SN 28578, N-[3-(methyl{3-[(1-oxido1,2,4-benzotriazin-3-yl)amino]propyl}amino)propyl]-4-acridinecarboxamide; compound 10, SN26955, N1-(9-acridinyl)-N6-(1,4-dioxido-1,2,4benzotriazin-3-yl)-1,6-hexanediamine; TAE buffer, Tris-acetate (0.4 M) EDTA (1 mM); E(1), one-electron reduction potential vs SHE; G value, radiation chemical yield (mol J-1; M Gy-1); G(ssb), yield of single strand breaks; G(dsb), yield of double strand breaks.

Scheme 1

oxide radical. This mediates some toxicity (11, 12), but far less than the one-electron reduced species, and is removed by superoxide dismutase. It is inferred that the one-electron reduced species, if not back-oxidized, is cytotoxic to cells, because the stable two-electron reduced species ultimately formed (3-amino-1,2,4-benzotriazine 1-oxide; 4) is nontoxic (13). Thus, the one-electron reduced species is the key radical intermediate in the cytotoxic action of TPZ. The cytotoxicity of TPZ increases with decreasing pH (14), pointing to the protonated radical (3) (radical pKa ) 6.1 (10)) as the cytotoxin, or as an obligate intermediate in forming a cytotoxin, Scheme 1. However, because the protonated radical 3 is a reducing radical, being readily oxidized by molecular oxygen (10, 15), it is difficult to understand how it could also oxidize DNA and cause strand breaks, as has been suggested (10, 15, 16). It has been proposed from work with radical scavengers (17) that the •OH radical is eliminated from 3, although spin-trap EPR experiments

10.1021/tx034116v CCC: $25.00 © 2003 American Chemical Society Published on Web 10/21/2003

1478 Chem. Res. Toxicol., Vol. 16, No. 11, 2003

Anderson et al.

Scheme 2

are inconclusive (18). We have recently proposed that the one-electron reduced species, 3, undergoes a unimolecular reaction (independent of both radical and substrate concentrations) to form a new radical species, 5 (19). For TPZ, this unimolecular reaction proceeds with a rate constant of 112 ( 24 s-1 at pH 7 and is accelerated in analogues with electron-donating substituents on the A-ring. Spectral studies with analogues have shown that the same intermediate is formed upon one-electron reduction of the 1-oxide analogues of 4, and formation of a common benzotriazinyl radical, 5, is proposed (19). Such a radical most likely arises from the elimination of water from 3, Scheme 2. Whether the radical 5 can induce DNA double strand breaks per se, or by effecting an initial radical damage, which is amplified by a subsequent reaction with a further benzotriazine 1,4-dioxide moiety as has been proposed (20-22), is under study. We have found that radical 5 does have oxidizing properties, with a one-electron reduction potential (E(1) 5,H+/4) of 1.32 V, sufficiently high enough to oxidize DNA constituents, including H-atom abstraction from 2-deoxyribose (19). Abstraction of an H-atom from the deoxyribose backbone of DNA is thought to be a prerequisite for strand breaks (23). A recent cytotoxicity study, using an analogue of TPZ where the benzotriazine 1,4-dioxide moiety is linked to a 9-aminoacridine DNA-binding unit via an alkyl chain that confers a strong association with DNA (compound 10, Figure 1), has found that this compound is more potent than TPZ in killing hypoxic cells by 1-2 orders of magnitude (24). This enhancement in cell killing is suggested to arise through enzymatic reduction in the nucleus of cells rather than in the cytoplasm where metabolism does not result in any DNA damage (25). However, it is unclear if the enhancement in cytotoxicity, as well as the observed increase in both single and double strand breaks, is due to increased metabolism by the nuclear enzymes per se or that it is a consequence of concentrating the benzotriazine 1,4-dioxide moiety in the vicinity of radical damage on the DNA. In the present in vitro study, we use the radiolysis of aqueous solutions to produce the oxidizing •OH radical to generate a range of free radical damages on DNA (i.e., replacing the enzymatic activation of TPZ), to examine the possible conversion of oxidative radical damage to strand breaks on DNA by TPZ under hypoxia. We report a substantial increase in double strand breaks when the benzotriazine 1,4dioxide moiety is targeted to DNA by attachment of a DNA-binding 4-acridine carboxamide chromophore.

Materials and Methods Plasmid DNA, pCMV. Sport-βgal, 7853 base pairs (bp) (Life Technologies), was prepared by expression in E. coli followed by extraction and purification using a Qiagen Plasmid Maxikit (Qiagen GmbH, Germany). All preparations contain greater than 95% of the intact closed circular, supercoiled form (form I), the balance being the relaxed coiled circular form (form II). Tirapazamine and analogues 6-9, linked by the 3-amine substituent to 4-acridinecarboxamide with three different in-

Figure 1. Structures of DNA-targeted compounds. tervening chains to afford compounds with a range of DNAbinding strengths (Figure 1), were synthesized as described in Supporting Information. Pulse radiolysis was used to determine the one-electron reduction potentials, E(1) of compounds 4, 6-9 at pH 7.0 from measurements of the equilibrium constants between the radicals of the compounds and reference compounds, as described in the literature (26). Irradiation of Plasmid DNA. Plasmid DNA was irradiated using a 1.0 × 104GBq 60Co γ-ray source delivering a dose rate of 18 Gy min-1. The irradiation solutions (70 µL), consisting of plasmid (1.6 nM bp), Tris buffer (e10 mM), phosphate buffer (1 mM, pH 7.4), EDTA (0.1 mM), and the test compounds, were deaerated in septum-sealed glass vials by bubbling with N2 (or N2O) gas before and after each radiation dose and serial sampling. The radiolysis of water produces three well-characterized reactive radical species used to initiate radical reactions, as well as molecular products (concentrations in µM per absorbed dose of 1 Gy (J kg-1) given in parentheses).

H2O ∧∧∧∧ f e-aq(0.28) + •OH(0.28) + H•(0.06) + H2(0.04) + H2O2(0.07) + H3O+(0.28) Of these radical species, the •OH radical causes significant damage to DNA through oxidative attack, to form DNA radicals. (Also, in the absence of the test compounds, the eaq- and H• atoms are scavenged by the phosphate to form the oxidizing phosphate radical.) The influence of the added test compounds on oxidative radical damage to the plasmid was studied at a constant radical scavenging capacity, k; the summation of the pseudo first-order rate constants kx for the reaction of •OH radicals with each component, x, in solution multiplied by its concentration, k ) Σni kx[x]n s-1, as previously described (27). The scavenging capacity (1.5 × 107 s-1) is maintained by reducing the concentration of Tris buffer (e10 mM) so that proportionate protection of the plasmid through direct scavenging of the •OH radicals by Tris buffer (plus carrier DNA when present) and the test compounds (rate constant determined as 5.0 × 109 M-1 s-1, Anderson unpublished data) is similar to controls. Hence, the subsequent effect of the test compounds on the DNA radicals, in terms of increasing or decreasing the yield of DNA damage, can be studied. By maintaining a low radical scavenging environment, strand breakage arising from the direct ionization of the DNA is only a very minor component compared to the damage arising from the •OH radical yield formed in close proximity to the DNA (28, 29) and can be ignored. The Do values for the loss of form I due to the formation of single strand breaks were obtained from the slopes of the regression lines of the percentage of form I as a function of radiation dose using Do ) ((log 37)-2)/slope. Do values were used to calculate the G values (radiation chemical yield) from triplicate experiments for single strand breaks formation (in pM ‚ Gy-1), where G(ssb) ) [DNA]/Do. The slopes of lines constructed for the production of linear DNA, form III, yield G(dsb) values for the production double strand breaks (in pM ‚ Gy-1).

Studies of DNA-Targeted Tirapazamine

Chem. Res. Toxicol., Vol. 16, No. 11, 2003 1479

Table 1. Physicochemical Parameters and the Yields of DNA Single and Double Strand Breaks of the Compounds Tested 104KDNA/M-1a

compound (conc) control N2 1, TPZ 50 µM 100 µM 200 µM 400 µM 6, SN 28141 50 µM 7, SN 27149 50 µM 8, SN 27140 50 µM 9, SN 28578 50 µM

ND

3.3 ( 0.6 6.4 ( 2.0 0.010 ( 0.008 ND

E(1)/mV -456 ( 8b

-444 ( 8 -421 ( 8 -466 ( 9 -580 ( 10

control (N2O) 1, TPZ 100 µM 6, SN 28141 50 µM a

Gssb/pM Gy-1

Gdsb/pM Gy-1

25.2 ( 2.8 31.0 ( 2.4 30.5 ( 7.3 33.4 ( 3.9 555 ( 134 369 ( 56 164 ( 25 48.5 ( 10.2

0.382 ( 0.037 0.357 ( 0.048 0.463 ( 0.046 0.562 ( 0.039 0.714 ( 0.026 39.1 ( 4.0 21.3 ( 2.4 0.397 ( 0.037 0.658 ( 0.036

(16.0 ( 2.0) (18.6 ( 1.9) (574 ( 90)

(0.349 ( 0.033) (0.387 ( 0.039 (36.9 ( 4.9)

Pruijn et al. unpublished data. b References (37, 38), ND ) not determined.

Single and Double Strand Break Assays. Forms I, II, and III of the plasmid were separated out on agarose gels (0.8%) by electrophoresis in TAE buffer. The DNA bands were stained with ethidium bromide (which binds more strongly than the targeted compounds) and the NIH image software package was used in their quantification. Southern blotting was performed on test samples of compounds 6, 7, and 9 as calf thymus DNA (0.5 mM in base pairs) was used as carrier DNA to ensure that the DNA-targeted compounds did not overload the target plasmid DNA. Nonradioactive DIG detection was used to identify the DNA bands.

Results The one-electron reduction potentials of the DNAtargeted compounds, Table 1, were determined against methyl viologen, E(1) ) -447 ( 7 mV, using three equilibrium mixtures for each compound. The e-aq was used as the primary reductant in deaerated solutions containing 2-methylpropan-2-ol (0.5 M) to aid solubility. Compound 8, which possesses a neutral linker chain, was found to have an E(1) similar to TPZ, whereas compounds 6 and 7 have raised E(1) values due to protonation of the amine moieties in their chains. The E(1) value of the two-electron reduced metabolite of TPZ, the 1-oxide SR 4317, 4, was determined as -568 ( 9 mV against triquat, E(1) ) -548 ( 7 mV, in deaerated solutions containing propan-2-ol (0.2 M), where both the propan-2-oxyl radical and the e-aq act as primary reductants. Similarly, the E(1) of the 1-oxide of 6, compound 9, was determined as -580 ( 10 mV. DNA single strand breaks, induced by •OH radical attack in a constant radical scavenging environment, were determined in the presence and absence of TPZ and the four DNA-targeted compounds. Radiation dose response curves, constructed from triplicate experiments, are presented in Figures 2 and 3 and represent the loss of supercoiled plasmid DNA upon the formation of single strand breaks in the DNA for small radiation doses. Adding the compounds to irradiated controls immediately after irradiation had no modifying effect. The D0 values from the plots were used to calculate the yield of single strand breaks, G(ssb), Table 1. Compounds 6, 7, and 9, where the benzotriazine 1,4-dioxide and 1-oxide rings are targeted to DNA, show greatly enhanced production of single strand breaks over that seen with TPZ in solution. The acridine ring, when bound to the DNA, has been shown previously not to effect radiosensitization (30), and we ascribe the observed radiosensitization effects in our present study to the benzotriazine 1,4-dioxide and 1-oxide rings. The production of DNA double strand breaks in the plasmid was also measured with and without the com-

Figure 2. Effect of different TPZ concentrations and compound 9 on the radiation dose-dependent loss of supercoiled fraction of plasmid DNA upon the γ-irradiation of N2-saturated solutions at pH 7.4 (1 mM phosphate) containing (i) Tris buffer (10 mM) b; (ii) TPZ (50 µM) + Tris (9.8 mM) O; (iii) TPZ (100 µM) + Tris (9.7 mM) 0; (iv) TPZ (200 µM) + Tris (9.3 mM) 3; and (v) 9 (50 µM) + Tris (9.8 mM) 1.

Figure 3. Effect of DNA-targeted compounds (50 µM) on the radiation dose-dependent loss of supercoiled fraction of plasmid DNA upon the γ-irradiation of N2-saturated solutions at pH 7.4 (1 mM phosphate) containing (i) Tris buffer (10 mM) b; (ii) 6 + Tris (9.8 mM) 9; (iii) 7 + Tris (9.8 mM) 2; (iv) 8 + Tris (9.8 mM) ∆; and (v) 6 + Tris (9.8 mM) but saturated with N2O gas ".

pounds being present. TPZ was tested at 50-400 µM concentrations, Figure 4, and the targeted compounds at 50 µM concentration, Figures 4 and 5. It is seen that concentrations of TPZ g 100 µM effect an increase in

1480 Chem. Res. Toxicol., Vol. 16, No. 11, 2003

Figure 4. Effect of different concentrations of TPZ, 1, and compounds 8 and 9 on the radiation dose-dependent appearance of linear fraction of plasmid DNA upon the γ-irradiation of N2-saturated solutions at pH 7.4 (1mM phosphate) containing (i) Tris buffer (10 mM) b; (ii) TPZ, 1 (50 µM) + Tris (9.8 mM) O; (iii) TPZ, 1 (100 µM) + Tris (9.7 mM) 0; (iv) TPZ, 1 (200 µM) + Tris (9.3 mM) 3; (v) TPZ, 1 (400 µM) + Tris (8.7 mM) ]; (vi) 8 (50 µM) + Tris (9.8 mM) ∆; and (vii) 9 (50 mM) + Tris (9.8 mM) 1.

Figure 5. Effect of compounds 6 and 7 (50 µM) on the radiation dose-dependent appearance of linear fraction of plasmid DNA upon the γ-irradiation of N2-saturated solutions containing (i) Tris buffer (10 mM) b; (ii) 6 + Tris (9.8 mM) 9; (iii) 7 + Tris (9.8 mM) 2; and (iv) 6 + Tris (9.8 mM) but saturated with N2O gas ".

DNA double strand breaks, whereas 50 µM of TPZ, as well as compound 8, are similar to controls, Figure 4. The 1-oxide analogue, compound 9, at 50 µM effects an increase in double strand breaks similar to 200-400 µM TPZ, Table 1. These results are in stark contrast to those seen for compounds 6 and 7 where large increases in the yield double strand breaks are observed, Figure 5. Similar results were obtained for compound 6 when the deaerating gas was changed from N2 to N2O, Figure 5. The results obtained are summarized in Table 1.

Discussion The one electron reduction of TPZ by endogenous enzymes has long been known to produce a free radical species that causes oxidative damage to DNA (5). Our recent studies have obtained evidence that this species is likely to be a benzotriazinyl radical, which reacts with both the bases and the ribose constituents of DNA in a

Anderson et al.

fashion similar to the •OH radical (19). The radiolysis of water provides a convenient method to produce quantitative concentrations of the •OH radical and thereby to initiate free radical damage on biological molecules. The present study was designed so that a constant, but small, proportion of the •OH radicals react with the plasmid DNA in the absence and presence of different concentrations of the compounds. Additional radical damage to the plasmid DNA might arise through the reduction of the compounds by the radiation-produced eaq- (and H• atoms) to form the oxidizing benzotriazinyl radical. The similar results obtained using N2O as the deaerating gas in place of N2 with 6, Figure 5, rules out the possibility that part of the observed increase in strand breaks arising from the targeted analogues could be due to direct reduction of the benzotriazine 1,4-dioxide moiety by the e-aq. The high solubility of N2O in solution and its reactivity in scavenging the eaq- effectively prevent any direct reduction of 6. The larger amount of DNA strand breaks in controls under N2 compared to N2O, Table 1 may be explained by the eaq- being scavenged by the phosphate buffer present to produce additional H• atoms, which in turn are scavenged by the phosphate ions to form the phosphate radical (H2PO4•-). This radical is known to abstract H• atoms from 2-deoxyribose (31) and may induce strand breaks through attack at the DNA sugar more efficiently than the more reactive •OH radical, which additionally reacts with the DNA bases. The increases in single strand breaks for TPZ and 8 are similar to that seen when misonidazole, a classic nitroimidazole radiosensitizer, is present (data not presented). Misonidazole possesses an E(1) value in the same region as that of TPZ but does not form a radical upon one-electron reduction, which is capable of inducing DNA strand breaks. The present study thus effectively addresses the conversion of •OH radical damage to DNA into single and double strand breaks in the presence of the targeted and untargeted benzotriazine 1,4-dioxide moiety. This enhancement of DNA damage has been proposed to proceed through the adduct formation between TPZ and sugar radicals on DNA (20, 22), which breaks down, resulting in strand breakage. Our data support this overall reaction and give further information on the formation of double strand breaks, which is considered to be a lethal event in cells (32). The yield of double strand breaks in plasmid DNA arising from •OH radical attack is approximately 1% of the observed yield of single strand breaks. Double strand breaks in dilute solution most likely arise from clustered damage in which two single strand breaks occur on adjacent strands in close proximity to each other. TPZ at 100 and 200 µM concentration effect increases in the yield of single strand breaks over controls and also a higher, but modest, percentage conversion of such breaks to double strand breaks. The data for the DNA-targeted compounds is more complex. Compounds 6-8 effect up to an order of magnitude increase in single strand breaks over that seen at an equal concentration of TPZ (50 µM) but only the targeted compounds 6 and 7 effect large increases in double strand breaks. Compounds 6 and 7 have recently been shown, Table 1, to bind to calf thymus DNA with moderate association constants (and also, presumably, compound 9), slightly less than that reported for 10 of 14.3 ( 1.5 × 104 M-1 (22); whereas TPZ is not expected to have an appreciable DNA binding constant, and compound 8 has a KDNA of 100 M-1, which is at the

Studies of DNA-Targeted Tirapazamine

Chem. Res. Toxicol., Vol. 16, No. 11, 2003 1481 Scheme 3

limit of the detection range. The yields of double strand breaks for compounds 6 and 7 are dramatically high compared to that for TPZ (and controls) and represent an increase of up to ca. 5-fold in the ratio of double to single strand breaks. This result implies that a higher local concentration of the benzotriazine 1,4-dioxide moiety is a requirement for an efficient conversion of •OH radical damage to DNA into both single and double strand breaks. The increased conversion of single strand breaks to double strand breaks by 6 and 7 indicate that both analogues are particularly active in forming clustered damage. This may arise from a multistep process dependent on the compounds possessing one-electron reduction potentials (Table 1) in the region required to induce electron transfer from carbon-centered sugar radicals. Such an electron transfer to the 1,4-dioxide moiety can be envisaged to result in the formation of a labile cation on the sugar in DNA leading to a strand break (23). The one-electron reduced 1,4-dioxide (similar to 3) subsequently undergoes water elimination to form an oxidizing benzotriazinyl radical, 5a (Scheme 3), which as a result of DNA binding, reacts locally to the first lesion on the adjacent DNA strand and by a series of similar reactions results in a double strand break. The reaction of 5 with 2-deoxyribose and the occurrence of short chain reactions has been recently reported (19). This mechanism is supported by the observation that compound 9, the 1-oxide analogue of compound 6, effects

a large increase in single strand breaks in DNA, which has been damaged by •OH radical attack, over TPZ, but only modest increase in double strand breaks is seen, because the active benzotriazinyl radical cannot be formed to perpetrate a short chain reaction. A recent in vitro study concluded that the increased potency of the similar targeted compound 10 is due to its one-electron reduction by DNA-associated enzymes (24). Our studies add to this finding in that we show that a high level of clustered damage occurs in the DNA once free radical damage has been sustained. Scheme 3 illustrates the possible difference in the reactivity of benzotriazine 1,4-dioxides (1) to DNA-targeted analogues (6). One-electron, enzymatic, reduction of the benzotriazine 1,4-dioxides leads to the formation of the proposed oxidizing benzotriazinyl radicals (5 and 5a, respectively) which react in both cases with DNA to produce a variety of free radical damage including the formation of sugar radicals (e.g., the 5′ radical is illustrated in Scheme 3, but other radical positions are also formed). This reaction reduces the benzotriazinyl radicals (5 and 5a) to the 1-oxides (4 and 9) which, in the case of the DNA-targeted compounds, are held in the vicinity of the DNA radical. The E(1) of 9 at -580 ( 10 mV is considerably higher than that measured for related quinoxaline N-oxides (-800 to -700 mV (33)) and this may well be a major factor in it acting to enhance radical damage to DNA through electron transfer in an analogous manner to

1482 Chem. Res. Toxicol., Vol. 16, No. 11, 2003

electron affinic (radiosensitizing) drugs (34). There is evidence that the enhancement of DNA radical damage by radiosensitizers proceeds through the transient formation of adducts (35), as do the benzotriazine 1,4dioxides (20, 22, 36), although the breakdown products have yet to be fully illucidated for the different compounds and radical sites on DNA. Such a process of 9 reacting with the DNA radical could lead to the oneelectron reduction of 9 (possibly the nonoxidizing nitroxide radical 11) and formation of hydrolyzable cations on the sugar leading to a strand break, Scheme 3. It is apparent from this study that upon targeting benzotriazine 1,4-dioxides to DNA, an increased level of clustered damage would be produced irrespective of a possible enhanced level of metabolism by DNA-associated enzymes. Increased loading of DNA with these bioreductive compounds increases the frequency of clustered single strand breaks on adjacent strands, resulting in the formation of lethal double strand breaks. In this study, the radiolytically produced •OH radical was used to react with DNA as the initial radical reaction. Only a small proportion of the •OH radicals react with DNA to give rise to DNA radicals that may reduce the benzotriazine 1,4-dioxide compounds, and only a subset of these lead to DNA strand breaks. The order of radical production is thus different in these radiolysis experiments than that for enzymatic reduction. This means that the benzotriazine 1,4-dioxide moiety of DNAtargeted compounds is able to oxidize the DNA radicals directly, and this may lead to increased yields in DNA strand breaks over that for enzymatic reduction. There should be little overall difference between the enzymatic and radiolytic reduction in the case of targeted benzotriazine 1,4-dioxides, as the pivotal reaction is between DNA radicals and benzotriazine 1,4-dioxides irrespective of how the oxidizable DNA radicals are formed. Another possible difference between the two systems is that in preliminary experiments, we find the 1-oxide, 4, acts as a radioprotector of DNA by decreasing the amount of double strand breaks relative to control experiments. The reason for this is presently unclear, but may be related to the radiolytic reduction of 4 to produce a diffusible radical, or product (such as the benzotriazine nor-N-oxide (SR4330)), that back reduces DNA radical damage. The objective of targeting TPZ to DNA is to increase the proportion of “productive” metabolism occurring sufficiently close to nuclear DNA targets to contribute to cytotoxicity. Such targeting is expected to reduce wasteful extranuclear consumption of drug and increase potency, thereby increasing therapeutic efficacy relative to TPZ. However, such an approach may have limitations, due to the compounds binding to DNA restricting diffusion through tumors to target hypoxic cells (25). Our study shows that high potency is achievable with compounds with moderate association constants, and further studies to fully explore this relationship are warranted. In summary, we show that benzotriazine 1,4-dioxides effect an increase in both single and double strand breaks on DNA that has sustained free radical damage. We demonstrate an enhancement of such free radical damage, which can arise through the formation of an oxidizing radical upon the one electron reduction of TPZ, by TPZ itself. Targeting benzotriazine 1,4-dioxides to DNA with an acridine chromophore results in a large increase in single strand breaks and also a higher percentage conversion to double strand breaks. The increased forma-

Anderson et al.

tion of double strand breaks may arise from clustered damage through radical chain reactions by the DNAtargeted compounds, rather than increased reduction by one-electron reductases in the nucleus.

Acknowledgment. This work was supported by Grant 00/462 from the Health Research Council of New Zealand and the NCI under Grant CA82566. We thank Dr S.A. Gamage for the synthesis of compound 9. Supporting Information Available: Experimental for compounds 6-9. This material is available free of charge via the Internet at http://pubs.acs.org.

References (1) Lee, D. J., Trotti, A., Spencer, S., Rostock, R., Fisher, C., von Roemeling, R., Harvey, E., and Groves, E. (1998) Concurrent tirapazamine and radiotherapy for advanced head and neck carcinomas: a Phase II study. Int. J. Radiat. Oncol. Biol. Phys. 42, 811-815. (2) Del Rowe, J., Scott, C., Werner-Wasik, M., Bahary, J. P., Curran, W. J., Urtasun, R. C., and Fisher, B. (2000) Single-arm, openlabel phase II study of intravenously administered tirapazamine and radiation therapy for glioblastoma multiforme. J. Clin. Oncol. 18, 1254-1259. (3) von Pawel, J., von Roemeling, R., Gatzemeier, U., Boyer, M., Elisson, L. O., Clark, P. T., 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. (2000) Tirapazamine plus cisplatin versus cisplatin in advanced nonsmallcell lung cancer: A report of the international CATAPULT I study group. Cisplatin and tirapazamine in Subjects with Advanced Previously Untreated Non-Small-Cell Lung Tumors. J. Clin. Oncol. 18, 1351-1359. (4) Brown, J. M., and Wang, L. H. (1998) Tirapazamine: laboratory data relevant to clinical activity. Anti-Cancer Drug Des. 13, 529539. (5) Walton, M. I., and Workman, P. (1990) Enzymology of the reductive bioactivation of SR 4233. A novel benzotriazine di-Noxide hypoxic cell cytotoxin. Biochem. Pharmacol. 39, 1735-1742. (6) Walton, M. I., Wolf, C. R., and Workman, P. (1992) The role of cytochrome P450 and cytochrome P450 reductase in the reductive bioactivation of the novel benzotriazine di-N-oxide hypoxic cytotoxin 3-amino-1,2,4-benzotriazine-1,4-dioxide (SR 4233, WIN 59075) by mouse liver. Biochem. Pharmacol. 44, 251-259. (7) Fitzsimmons, S. A., Lewis, A. D., Riley, R. J., and Workman, P. (1994) Reduction of 3-amino-1,2,4-benzotriazine-1,4-di-N-oxide (tirapazamine, WIN 59075, SR 4233) to a DNA-damaging species: a direct role for NADPH: cytochrome P450 oxidoreductase. Carcinogenesis 15, 1503-1510. (8) Patterson, A. V., Barham, H. M., Chinje, E. C., Adams, G. E., Harris, A. L., and Stratford, I. J. (1995) Importance of P450 reductase activity in determining sensitivity of breast tumour cells to the bioreductive drug, tirapazamine (SR 4233). Br. J. Cancer 72, 1144-1150. (9) Patterson, A. V., Saunders, M. P., Chinje, E. C., Patterson, L. H., and Stratford, I. J. (1998) Enzymology of tirapazamine metabolism: a review. Anti-Cancer Drug Des. 13, 541-573. (10) Laderoute, K., Wardman, P., and Rauth, A. M. (1988) Molecular mechanisms for the hypoxia-dependent activation of 3-amino1,2,4-benzotriazine-1,4-dioxide (SR 4233). Biochem. Pharmacol. 37, 1487-1495. (11) Tuttle, S. W., Hazard, L., Koch, C. J., Mitchell, J. B., Coleman, C. N., and Biaglow, J. E. (1994) Bioreductive metabolism of SR4233 (WIN 59075) by whole cell suspensions under aerobic and hypoxic conditions: Role of the pentose cycle and implications for the mechanism of cytotoxicity observed in air. Int J. Radiat. Oncol. Biol. Phys. 29, 357-362. (12) Elwell, J. H., Siim, B. G., Evans, J. W., and Brown, J. M. (1997) Adaptation of human cells to tirapazamine under aerobic conditions: Implications of increased antioxidant enzyme activity to mechanism of aerobic cytotoxicity. Biochem. Pharmacol. 54, 249257. (13) Baker, M. A., Zeman, E. M., Hirst, V. K., and Brown, J. M. (1988) Metabolism of SR 4233 by chinese hamster ovary cells: basis of selective hypoxic cytotoxicity. Cancer Res. 48, 5947-5952. (14) Teicher, B. A., Liu, J. T., Holden, S. A., and Herman, T. S. (1993) A carbonic-anhydrase inhibitor as a potential modulator of cancer therapies. Anticancer Res. 13, 1549-1556.

Studies of DNA-Targeted Tirapazamine (15) Lloyd, R. V., Duling, D. R., Rumyantseva, G. V., Mason, R. P., and Bridson, P. K. (1991) Microsomal reduction of 3-amino-1,2,4benzotriazine 1,4-dioxide to a free radical. Mol. Pharmacol. 40, 440-445. (16) Brown, J. M. (1993) SR 4233 (tirapazamine): a new anticancer drug exploiting hypoxia in solid tumours. Br. J. Cancer 67, 11631170. (17) Daniels, J. S., and Gates, K. S. (1996) 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. (18) Patterson, L. H., and Taiwo, F. A. (2000) Electron paramagnetic resonance spectrometry evidence for bioreduction of tirapazamine to oxidising free radicals under anaerobic conditions. Biochem. Pharmacol. 60, 1933-1935. (19) Anderson, R. F., Shinde, S. S., Hay, M. P., Gamage, S. A., and Denny, W. A. (2003) Activation of 3-amino-1,2,4-benzotriazine 1,4dioxide antitumor agents to oxidizing species following their oneelectron reduction. J. Am. Chem. Soc. 125, 748-756. (20) Jones, G. D. D., and Weinfeld, M. (1996) Dual action of tirapazamine in the induction of DNA strand breaks. Cancer Res. 56, 1584-1590. (21) Daniels, J. S., Gates, K. S., Tronche, C., and Greenberg, M. M. (1998) Direct evidence for bimodal DNA damage induced by tirapazamine. Chem. Res. Toxicol. 11, 1254-1257. (22) Hwang, J.-T., Greenberg, M. M., Fuchs, T., and Gates, K. S. (1999) 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. (23) von Sonntag, C. The Chemical Basis of Radiation Biology; Taylor & Francis Ltd.: London, 1987. (24) Delahoussaye, Y. M., Hay, M. P., Pruijn, F. B., Denny, W. A., and Brown, J. M. (2003) Improved potency of the hypoxic cytotoxin tirapazamine by DNA-targeting. Biochem. Pharmacol. 65, 18071815. (25) Evans, J. W., Yudoh, K., Delahoussaye, Y. M., and Brown, J. M. (1998) Tirapazamine is metabolised to its DNA damaging radical by intranuclear enzymes. Cancer Res. 58, 2098-2101. (26) Wardman, P. (1989) Reduction potentials of one-electron couples involving free radicals in aqueous solution. J. Phys. Chem. Ref. Data 18, 1637-1755. (27) Anderson, R. F., Amarasinghe, C., Fisher, L. J., Mak, W. B., and Packer, J. E. (2000) Reduction in free-radical-induced DNA strand

Chem. Res. Toxicol., Vol. 16, No. 11, 2003 1483

(28)

(29)

(30) (31) (32)

(33)

(34) (35) (36)

(37)

(38)

breaks and base damage through fast chemical repair by flavonoids. Free Radical Res. 33, 91-103. Siddiqi, M. A., and Bothe, E. (1987) Single- and double-strand break formation in DNA irradiated in aqueous solution: Dependence on dose and OH radical scavenger concentration. Radiat. Res. 112, 449-463. Hodgkins, P. S., Fairman, M. P., and O’Neill, P. (1996) Rejoining of gamma-radiation-induced single-strand breaks in plasmid DNA by human cell extracts: dependence on the concentration of the hydroxyl radical scavenger, Tris. Radiat. Res. 145, 24-30. Roberts, P. B., Denny, W. A., Wakelin, L. P. G., Anderson, R. F., and Wilson, W. R. (1990) Radiosensitization of mammalian cells in vitro by nitroacridines. Radiat. Res. 123, 153-164. Nakashima, M., and Hayon, E. (1970) Rates of reaction of inorganic phoshate radicals in solution. J. Phys. Chem. 74, 32903291. Biedermann, K. A., Wang, J., Graham, R. P., and Brown, J. M. (1991) SR 4233 cytotoxicity and metabolism in DNA repaircompetent and repair-deficient cell cultures. Br. J. Cancer 63, 358-362. Priyadarsini, K. I., Dennis, M. F., Naylor, M. A., Stratford, M. R. L., and Wardman, P. (1996) Free radical intermediates in the reduction of quinoxaline N-oxide antitumor drugs: redox and prototropic reactions. J. Am. Chem. Soc. 118, 5648-5654. Wardman, P. (1987) The mechanism of radiosensitization by electron-affinic compounds. Radiat. Phys. Chem. 30, 423-432. Greenstock, C. L., and Whitehouse, R. P. (1985) Radiosensitizers as probes of DNA damage and cell killing. Int. J. Radiat. Biol. 48, 701-710. Ban, F., Gauld, J. W., and Boyd, R. J. (2001) Modeling the action of an antitumor drug: A density functional theory study of the mechanism of tirapazamine. J. Am. Chem. Soc. 123, 73207325. Hay, M. P., Gamage, S. A., Kovacs, M., Pruijn, F. B., Anderson, R. F., Patterson, A. V., Wilson, W. R., Brown, J. M., and Denny, W. A. (2003) Structure-activity relationships of 1,2,4-Benzotriazine 1,4-dioxides as hypoxia-selective analogues of tirapazamine. J. Med. Chem. 46, 169-182. Priyadarsini, K. I., Tracy, M., and Wardman, P. (1996) The oneelectron reduction potential of 3-amino-1,2,4-benzotriazine 1,4dioxide (tirapazamine): a hypoxia-selective bioreductive drug. Free. Radical Res. 25, 393-400.

TX034116V