Improved potency of the hypoxic cytotoxin tirapazamine by DNA-targeting

Article abstract of DOI:10.1016/S0006-2952(03)00199-0

To improve the potency of the hypoxic cytotoxin tirapazamine (TPZ), we have constructed an analog, SN26955, with the TPZ moiety attached to an acridine chromophore to target the drug to DNA. The underlying reason for this is our previous finding that the hypoxic cytotoxicity of TPZ is a result of its ability to produce DNA double-strand breaks, whereas many of the toxicities of the drug in clinical use are likely the result of its metabolism in the cytoplasm and effects on mitochondria. We found that the DNA-targeted TPZ analog was more potent than TPZ in killing hypoxic cells by 1-2 orders of magnitude, yet it retained the hypoxic selectivity for cell killing of TPZ. We show that SN26955 is only active in producing DNA damage when it is enzymatically reduced while bound to, or in close association with, the DNA. We also show that it has a different cofactor dependence than TPZ for reduction leading to DNA double-strand breaks, suggesting the involvement of a different reductase for production of the lethal lesion than for TPZ. These results show the promise of DNA-targeting of TPZ to produce a DNA compound with greater clinical efficacy than TPZ itself.

Full text of DOI:10.1016/S0006-2952(03)00199-0

Biochemical Pharmacology 65 (2003) 1807–1815
Improved potency of the hypoxic cytotoxin tirapazamine
by DNA-targeting
Yvette M. Delahoussayea, Michael P. Hayb, Frederik B. Pruijnb,
William A. Dennyb, J. Martin Browna,*
aDepartment of Radiation Oncology, Stanford University School of Medicine, 269 Campus Drive, CCSR 1255, Stanford, CA 94305-5152, USA
bAuckland Cancer Society Research Centre, Faculty of Medical and Health Sciences, The University of Auckland,
Private Bag 92019, Auckland, New Zealand
Received 6 September 2002; accepted 3 December 2002
Abstract
To improve the potency of the hypoxic cytotoxin tirapazamine (TPZ), we have constructed an analog, SN26955, with the TPZ moiety
attached to an acridine chromophore to target the drug to DNA. The underlying reason for this is our previous finding that the hypoxic
cytotoxicity of TPZ is a result of its ability to produce DNA double-strand breaks, whereas many of the toxicities of the drug in clinical use
are likely the result of its metabolism in the cytoplasm and effects on mitochondria. We found that the DNA-targeted TPZ analog was more
potent than TPZ in killing hypoxic cells by 1–2 orders of magnitude, yet it retained the hypoxic selectivity for cell killing of TPZ. We show
that SN26955 is only active in producing DNA damage when it is enzymatically reduced while bound to, or in close association with, the
DNA. We also show that it has a different cofactor dependence than TPZ for reduction leading to DNA double-strand breaks, suggesting
the involvement of a different reductase for production of the lethal lesion than for TPZ. These results show the promise of DNA-targeting
of TPZ to produce a DNA compound with greater clinical efficacy than TPZ itself.
# 2003 Elsevier Science Inc. All rights reserved.
Keywords: Tirapazamine; SN26955; Bioreductive drugs; Hypoxia; Tumor; DNA-targeting
1. Introduction
TPZ is converted by intracellular reductase(s) to a
cytotoxic radical that under hypoxic conditions generates
base damage, SSBs and DSBs [10,11]. However, when
oxygen is present, the TPZ radical is back-oxidized to the
non-toxic parent compound, thereby decreasing the cyto-
toxicity [12]. Studies with cells in vitro have indicated that
it is the DSB that is responsible for the toxicity of TPZ
under hypoxic conditions [10]. While many intracellular
enzymes can convert TPZ to its cytotoxic radical [13,14], it
is the nuclear reductase(s) that is responsible for the DNA
damage caused by TPZ [15]. However, like the cytoplasm,
the nucleus is highly compartmentalized with the nuclear
matrix, a non-chromatin protein network, dictating this
subnuclear organization [16]. Based on recent studies
showing that TPZ can be selectively reduced by nuclear
matrix-associated reductases [17], we hypothesized that
the TPZ metabolized at, or close to, the nuclear matrix
would affect metabolic activities associated with the
nuclear matrix. Consistent with this, we have shown
recently that TPZ, under hypoxic conditions, produces a
marked inhibition both of DNA replication [18], one of the
TPZ (SR 4233, 3-amino-1,2,4-benzotriazine-1,4-diox-
ide) is the first member of a new class of tumor-targeted
drugs to enter clinical trials. Its tumor specificity rests on
the fact that TPZ is a prodrug that becomes activated to its
cytotoxic form only in the hypoxic microenvironment that
is characteristic of solid tumors [1,2]. As hypoxic regions
in tumors confer resistance to standard radiotherapy and
chemotherapy [3,4], a drug selectively toxic to hypoxic
cells would be both tumor specific and kill the cells
resistant to conventional therapy [5,6]. TPZ has already
demonstrated significant activity in Phase II and III clinical
trials in combination with radiotherapy and cisplatin-based
chemotherapy [7–9].
* Corresponding author. Tel.: þ1-650-723-5881; fax: þ1-650-723-7382.
E-mail address: mbrown@stanford.edu (J.M. Brown).
Abbreviations: TPZ, tirapazamine; SSB, single-strand DNA break; DSB,
double-strand DNA break; P450 reductase, NADPH:cytochrome P450
reductase; HCR, hypoxic cytotoxicity ratio; DHR123, dihydrorhodamine.
0006-2952/03/$ – see front matter # 2003 Elsevier Science Inc. All rights reserved.
doi:10.1016/S0006-2952(03)00199-0

1808
Y.M. Delahoussaye et al. / Biochemical Pharmacology 65 (2003) 1807–1815
activities that occur at the nuclear matrix, and of the
activity of topoisomerase II (topo II) [19], an enzyme that
is highly enriched in the nuclear matrix and which is
involved in the cytotoxicity of several chemotherapeutic
agents, such as doxorubicin and etoposide [20]. Indeed,
much, if not all, of the cytotoxicity of TPZ under hypoxia
can be attributed to the interaction of the TPZ radical with
topo II [19].
vated fetal bovine serum (FBS) plus 200 units/mL of peni-
cillin and 0.2 mg/mL of streptomycin. HeLa cells were also
growninspinnerflasksinRPMImediumsupplemented with
10% FBS andantibiotics. HT29cellswereobtainedfromDr.
R.M. Sutherland (SRI International) and were cultured in
McCoy’s 5A medium supplemented with 10% FBS and
antibioticsasforHeLacells. Allexperimentsutilizedcellsin
logarithmic-phase growth.
The conclusion of the above studies is that much of the
metabolism of TPZ, namely that which occurs in the
cytoplasm and in the nucleus at sites away from the nuclear
matrix, is unproductive in that it does not contribute to
damage leading to DSBs. Indeed, it is likely that such
damage may be contributing to some of the toxic side-
effects of TPZ including damage to mitochondria, which
we have shown to be a target for TPZ-induced damage in
aerobic cells [21]. In addition, any metabolism that does
not directly contribute to hypoxic cytotoxicity has the
potential to compromise extravascular diffusion of the
drug into hypoxic zones [22,23]. A possible way in which
this unproductive metabolism could be minimized would
be to target TPZ directly to the DNA. To test this possi-
bility, we have synthesized a TPZ analog attached to a
DNA-intercalating acridine chromophore. We show that
this compound, SN26955 (Fig. 1), is considerably more
potent than TPZ in killing hypoxic cells, but retains a
similar differential killing of hypoxic to aerobic cells. This
drug, which establishes a proof of principle that the
potency of TPZ can be increased considerably by targeting
it to DNA, opens up a promising new area for further drug
development of this compound.
2.2. Drugs and chemicals
TPZ was obtained from Sanofi-Synthelabo. Dihydro-
rhodamine 123 and rhodamine 123 were from Molecular
Probes. All other chemicals were obtained from Sigma and
were of the highest analytical quality.
2.3. Synthesis of SN26955
SN26955 was prepared in five steps from 3-chloro-
1,2,4-benzotriazine 1-oxide, 6-t-butyloxycarbamoylhexy-
lamine (Aldrich) and 9-methoxyacridine (Fig. 1) as
detailed below.
2.3.1. 3-[(6-t-Butyloxycarbamoylhexyl)amino]-1,2,4-
benzotriazine 1-oxide (3)
A solution of 6-t-butyloxycarbamoylhexylamine 2
(12.8 g, 61.1 mmol) (Aldrich) in dichloromethane (DCM)
was added to a stirred solution of 3-chloro-1,2,4-benzotria-
zine 1-oxide 1 (3.70 g, 20.4 mmol) [24] and Et3N (5.7 mL,
40.8 mmol) in DCM (100 mL), and the solution was stirred
at 208 for 96 hr. The solvent was evaporated and the residue
chromatographed, eluting with a gradient (30–100%) of
EtOAc/pet. ether, to give 1-oxide 3 (4.77 g, 65%) as a
yellow powder, m.p. (EtOAc/pet. ether) 154–1568; 1H
NMR d 8.26 (d, J ¼ 8:6 Hz, 1 H, H 8), 7.70 (dd,
J ¼ 8:2, 7.2 Hz, 1 H, H 6), 7.59 (d, J ¼ 8:5 Hz, 1 H, H
5), 7.27 (dd, J ¼ 8:0, 7.5 Hz, 1 H, H 7), 5.34 (br s, 1 H, NH),
4.55 (br s, 1 H, OCONH), 3.51 (dd, J ¼ 6:8, 6.6 Hz, 2 H,
CH2N), 3.10–3.13(m,2H, CH2N),1.64–1.72(m,2H, CH2),
1.48–1.54 (m, 2 H, CH2), 1.44 [s, 9 H, C(CH3)3], 1.38–1.43
(m, 4 H, 2 Â CH2). Anal. calc. for C18H27N5O3: C, 59.8; H,
7.5; N, 19.4; found: C, 59.6; H, 7.7; N, 19.2%.
2. Materials and methods
2.1. Cell lines
Cell lines were maintained in a humidified atmosphere of
95%airand 5%CO2 at378. HeLacells, a cervicalcarcinoma
cell line, were obtained from the American Type Culture
Collection and maintained as monolayers in a-MEM (Life
Technologies, Inc.) supplemented with 10% heat-inacti-
Fig. 1. Structure of SN26955 showing the TPZ moiety on the left attached via a linker onto a DNA-intercalating acridine ring.

Y.M. Delahoussaye et al. / Biochemical Pharmacology 65 (2003) 1807–1815
1809
2.3.2. 3-[(6-t-Butyloxycarbamoylhexyl)amino]-1,2,4-
benzotriazine 1,4-dioxide (4)
(ddd, J ¼ 8:5, 7.2, 1.2 Hz, 1 H, H 700), 7.31–7.35 (m, 2 H, H
20, H 70), 7.15 (br s, 1 H, NH), 3.84 (dd, J ¼ 7:2, 7.1 Hz, 2
H, CH2N), 3.57 (dt, J ¼ 6:7, 6.5 Hz, 2 H, CH2N), 1.78–
1.85 (m, 2 H, CH2), 1.67–1.74 (m, 2 H, CH2), 1.43–1.53
(m, 4 H, 2 Â CH2), NH not observed; MS (FABþ) m/z 455
(MHþ, 20%), 439 (10%); HRMS (FABþ) calc. for
C26H27N6O2 (MHþ) m/z 455.2196, found 455.2182. The
compound was dissolved in MeOH and treated with HCl
gas and the solvent evaporated. The residue was crystal-
lized from MeOH/EtOAc to give the hydrochloride of 5,
m.p. (MeOH/EtOAc) 118–1198. Anal. calc. for
C26H26N6O2Á2HClÁ1/2H2O: C, 58.2; H, 5.5; N, 15.7;
found: C, 57.8; H, 5.5; N, 15.3%.
A solution of meta-chloroperbenzoic acid (MCPBA;
1.48 g, 6.02 mmol) in DCM (20 mL) was added dropwise
to a stirred solution of 1-oxide 3 (1.45 g, 4.01 mmol) in
DCM (100 mL) at 208, and the solution was stirred for 4 hr.
The solution was partitioned between DCM (200 mL) and
saturated KHCO3 solution (200 mL). The organic fraction
was dried and the solvent evaporated. The residue was
chromatographed on neutral alumina, eluting with 50%
EtOAc/DCM and then a gradient (0–10%) MeOH/CHCl3,
to give (i) starting material 3 (0.73 g, 50%), and (ii) 1,4-
dioxide 4 (0.55 g, 37%) as a yellow powder, m.p. (EtOAc/
DCM) 132–1348; 1H NMR [(CD3)2SO] d 8.30 (dd, J ¼ 6:3,
6.1 Hz, 1 H, OCONH), 8.19 (d, J ¼ 8:5 Hz, 1 H, H 8), 8.12
(d, J ¼ 8:5 Hz, 1H, H5), 7.91–7.95(m, 1H, H6), 7.53–7.57
(m, 1 H, H 7), 6.76 (br s, 1 H, NH), 3.32–3.39 (m, 2 H,
CH2N),2.87–2.92(m,2H, CH2N),1.56–1.61(m,2H, CH2),
1.32–1.40 [m, 13 H, 2 Â CH2, C(CH3)3], 1.25–1.31 (m, 2 H,
CH2). Anal. calc. for C18H27N5O4Á1/4H2O: C, 56.6; H, 7.3;
N, 18.3; found: C, 56.8; H, 7.3; N, 16.8%.
2.4. Clonogenic assay
For exposure to TPZ or SN26955, cells were seeded onto
60-mm diameter glass dishes in complete medium 2 days
before the treatment such that each dish would have
1 Â 106 cells at the time of drug exposure. TPZ or
SN26955 was added at the appropriate concentration in
medium at a final volume of 2 mL. The cells were then
added to prewarmed, leak-proof aluminum jigs, placed on
a shaking platform at room temperature and treated with
five rapid evacuations and filling with either 95% nitrogen–
5% CO2 for hypoxia or with 95% air–5% CO2 for aerobic
controls. The jigs were then placed in a 378 box for 1 hr
with shaking. Cells were washed with PBS, trypsinized,
counted, and plated in serial dilution in complete medium
for cloning. After 10–14 days, the plates were stained with
crystal violet, and colonies with more than 50 cells were
counted as survivors.
2.3.3. N1-(1,4-Dioxido-1,2,4-benzotriazin-3-yl)-1,6-
hexanediamine (5)
HCl gas was bubbled through a solution of carbamate 4
(204 mg, 0.54 mmol) in MeOH (20 mL) for 2 min, and the
solution was stirred at 208 for 16 hr. The solvent was
evaporated and the residue partitioned between CHCl3
(100 mL) and saturated KHCO3 solution (100 mL). The
aqueous fraction was further extracted with CHCl3
(3 Â 30 mL), the combined organic extracts were dried,
and the solvent was evaporated to give amine 5 (127 mg,
85%) as a red powder, m.p. 120–1228; 1H NMR d 8.34 (d,
J ¼ 8:5 Hz, 1 H, H 80), 8.29 (d, J ¼ 8:6 Hz, 1 H, H 50),
7.87–7.90 (m, 1 H, H 60), 7.48–7.52 (m, 1 H, H 70), 7.13 (s,
1 H, NH), 3.60 (t, J ¼ 7:1 Hz, 2 H, CH2N), 2.70 (t,
J ¼ 6:8 Hz, 2 H, CH2N), 1.70–1.76 (m, 2 H, CH2),
2.5. Radical detection with DHR123
TPZ or SN26955 radicals were detected by using
DHR123 oxidation as an indicator of free radical formation.
DHR123 is a non-fluorescent compound that is oxidized by
the TPZ radical to form the highly fluorescent rhodamine
123 compound [25]. DHR123 (25 mM) was added to the
metabolism system described above in a total volume of
2 mL. The reaction was placed in 60-mm diameter glass
dishes, and hypoxia was created in the aluminum jigs with
alternate flushing and filling of 95% N2, 5% CO2. After a 1-
hr drug exposure at 378, the solution was transferred to a
cuvette, and the fluorescence was measured in a Perkin-
Elmer LS-3 fluorescence spectrophotometer at 509-nm
absorbance, 529-nm emission, with a slit-width of 10 nm.
A blank containing 1Â metabolism buffer and cofactors was
used to zero the spectrophotometer.
1.35–1.50 (m,
6
H, 3 Â CH2). Anal. calc. for
C13H19N5O2: C, 56.3; H, 6.9; N, 25.3; found: C, 56.3;
H, 6.8; N, 22.2%. The compound was dissolved in MeOH,
treated with HCl gas, and the solvent evaporated. The
residue was crystallized to give the dihydrochloride of 5
as a red powder, m.p. (MeOH/EtOAc) 1508 (dec.).
2.3.4. N1-(9-Acridinyl)-N6-(1,4-dioxido-1,2,4-
benzotriazin-3-yl)-1,6-hexanediamine (SN26955)
A solution of amine 5 (64 mg, 0.23 mmol) and 9-meth-
oxyacridine (53 mg, 0.25 mmol) in MeOH (10 mL) was
stirred at reflux temperature for 10 hr. The solvent was
evaporated and the residue chromatographed on neutral
alumina, eluting with a gradient (0–5%) of MeOH/CHCl3,
to give SN26955 (63 mg, 60%) as a red solid; 1H NMR d
8.30 (d, J ¼ 8:5 Hz, 1 H, H 800), 8.29 (d, J ¼ 8:5 Hz, 1 H, H
500), 8.11 (d, J ¼ 8:6 Hz, 2 H, H 10, H 80), 8.04 (d,
J ¼ 8:6 Hz, 2 H, H 40, H 50), 7.84 (ddd, J ¼ 8:5, 7.2,
1.2 Hz, 1 H, H 600), 7.59–7.64 (m, 2 H, H 30, H 60), 7.57
2.6. Nuclei isolation
Intact nuclei were isolated from HeLa cells as previously
described with minor modifications [26]. Briefly, cells
were harvested in log phase, washed twice with cold

1810
Y.M. Delahoussaye et al. / Biochemical Pharmacology 65 (2003) 1807–1815
PBS, and resuspended at 12.5 mL/106 cells in NI buffer
(115 mM KCl, 5 mM NaCl, 1 mM KH2PO4, 20 mM
HEPES, 0.3 mM MgCl2, 0.5 mM phenylmethylsulfonyl
fluoride, and 0.5 mM dithiothreitol at pH 7.4). An equiva-
lent volume of NI buffer with 1% IGEPAL CA-630 was
added slowly, and the reaction was held on ice for 15 min.
The permeable cells were disrupted by pipetting 25–30
times with a 1-mL pipette tip. The isolated nuclei were
diluted with a 10Â volume of NI buffer, centrifuged at 48,
and washed with 10 mL of NI buffer.
lished protocol. The slides were rinsed in neutral buffer,
electrophoresed, and visualized as described above.
2.9. DNA binding assay
DNA binding was measured by equilibrium dialysis
using a Dianorm dialyzer (Diachema AG) with 20 two-
cell dialysis chambers (1 mL each). SN26955 (15 mM)
with calf thymus DNA (Sigma type I) (25–1000 mM) in
KHE buffer (10 mM KCl, 2 mM NaHEPES, 10 mM
Na2EDTA, pH 7.0) was dialyzed (2 hr at 378) against
SN26955 in KHE buffer using a Cuprophan dialysis
membrane (Medicell International Ltd.) with a MW cutoff
of 10,000 Da. After equilibrium conditions were reached,
the contents of the dialysis chambers with KHE buffer
were collected and analyzed directly by HPLC. To the
samples containing DNA, 9 vol. of ice-cold ethanol was
added and kept on ice for 1 hr followed by a brief spin to
precipitate the DNA. Subsequently, the volume of the
aqueous ethanol was reduced using a Speed-Vac concen-
trator and reconstituted with mobile phase followed by
injection onto a high performance liquid chromatogram.
Concentrations of SN26955 were calculated using an
internal standard.
2.7. Alkaline comet assay
SSBs were measured using the alkaline comet assay.
Isolated nuclei and whole cells were treated with TPZ or
SN26955 in glass dishes under hypoxia in aluminum jigs
as described above for 1 hr. Cells were treated with drug
dissolved in a-medium containing 10% serum. Isolated
nuclei were treated with drug in nuclei treatment buffer
(100 mM KCl, 1 mM KH2PO4, 25 mM NaHCO3, 20 mM
HEPES, 10 mM EDTA, 5 mM glutathione, and 0.9 mM
spermine at pH 7.4). After treatment, cells and nuclei
were suspended in cold PBS without Mg and CaCl2 at a
concentration of 3 Â 104 cells/mL. Slides for the comet
assay were prepared following the protocol of Olive [27].
A 500-mL aliquot of cell or nuclei suspension was placed
in a labeled 5-mL glass tube, and 1.5 mL of 1% low-
gelling temperature agarose (Sigma type VII dissolved in
deionized water and held in a 458 water bath) was added
to the suspension. The agarose solution was pipetted onto
a microscope slide and allowed to gel for 1 min at 48 on a
cold metal block. The slides were then placed in an
alkaline lysis solution containing 1 M sodium chloride,
0.03 M sodium hydroxide, and 0.1% N-laurylsarcosine
for 1 hr at room temperature. After lysis, the slides were
washed three times for 20 min in 2 mM EDTA with
0.03 M sodium hydroxide, placed in a gel box, and
electrophoresed in rinse buffer at 0.6 V/cm for 20 min.
The slides were then removed from the gel box and placed
in water for 15 min, followed by staining in propidium
iodide (2.5 mg/mL in water) for 15 min. The individual
comets were scored as described by Olive [27] and the
images analyzed with Loats comet assay analysis soft-
ware program. The tail moment is calculated as the
product of the percentage of DNA in the comet tail
multiplied by the length between the means of the head
and the tail distributions. Mean tail moments were deter-
mined from 150 or more individual comets for each data
point.
3. Results
3.1. SN26955 binding to DNA
We measured the binding of SN26955 (15 mM at 378) to
calf thymus DNA by equilibrium dialysis with quantitation
of the drug by HPLC. The results are shown in Fig. 2. Also
shown is the fit of the binding data using the neighboring
site exclusion model of McGhee and von Hippel [28]. The
derived association constant [ð14:3 Æ 1:5Þ Â 104 Mꢀ1] and
2.8. Neutral comet assay
DSBs were measured in cells and isolated nuclei by the
neutral comet assay [27]. Cells and nuclei were treated as
described above, placed in agarose, pipetted onto slides,
and lysed in a neutral lysis buffer according to the pub-
Fig. 2. Data showing the binding of SN26955 to calf thymus DNA. The
line shows the fit of the binding data using the neighboring site exclusion
model of McGhee and von Hippel [28].

Y.M. Delahoussaye et al. / Biochemical Pharmacology 65 (2003) 1807–1815
1811
Fig. 3. Sensitivity of HT29 human colon carcinoma cells (a and b) and HeLa human cervix carcinoma cells (c and d) to SN26955 under aerobic and hypoxic
conditions (a and c), and in comparison to TPZ under hypoxic conditions (b and d). In all cases, the cells were exposed for 1 hr to drugs under either aerobic
or hypoxic conditions, and cell viability was determined using clonogenic survival. The increased potency of both drugs under hypoxic compared to aerobic
conditions was calculated by the ratios of the C10 values, as were the relative potencies of SN26955 and TPZ under hypoxic conditions. Results were pooled
from two experiments for the HT29 cells and from one experiment for the HeLa cells.
binding site size (1:52 Æ 0:09 bp) indicate that SN26955 is
moiety of SN26955. We found that the presence of 9-
(N-ethylamino)acridine made no difference to the cyto-
toxicity of TPZ (data not shown). This suggests that the
enhanced cytotoxicity of SN26955 compared to TPZ
results from the targeting of TPZ to DNA rather than a
non-specific sensitization by intercalation of the acridine
into DNA.
a moderately strong DNA binder.
3.2. Activity of SN26955 in vitro
To determine whether SN26955 retained the selective
hypoxic cytotoxicity of TPZ, we exposed HT29 human
colon carcinoma cells and HeLa human cervix carcinoma
cells to various concentrations of SN26955 under hypoxic
and aerobic conditions. Panels a and c of Fig. 3 show that
SN26955 was much more toxic to hypoxic cells than to
aerobic cells with HCR values (defined as the ratio of
concentrations under aerobic to hypoxic conditions to give
90% cell kill) of 83 and 78 for HT29 and HeLa cells,
respectively. Under the same conditions the HCR for these
cell lines for TPZ was 82 and 55, respectively (data not
shown). Panels b and d of Fig. 3 show a comparison of the
cytotoxicities of SN26955 and TPZ under hypoxic condi-
tions and demonstrate the increased potency of the analog
by a factor of 415- and 20-fold for HT29 and HeLa cells,
respectively.
3.3. Free radical formation by SN26955
TPZ is reduced under hypoxic conditions to a free
radical [29]. This can be detected by oxidation of the
non-fluorescent dihydrorhodamine 123 (DHR123) to the
highly fluorescent rhodamine 123 (Rh123) [17]. We used
this assay to compare the extent of radical formation of
TPZ and SN26955. As shown in Fig. 4, the extent of radical
formation as measured by DHR fluorescence induced by
enzymatic reduction by purified NADPH:cytochrome
P450 reductase (0.3 mg/mL) was the same for TPZ and
for SN26955. This suggests that the increased toxicity of
SN26955 compared to TPZ is not the result of an improved
ability to reduce SN26955 compared to TPZ at least for
cytochrome P450 reductase. As an additional demonstra-
tion of the free radical mechanism for the DHR fluores-
cence, we added different concentrations of the radical
scavenger DMSO to a solution of 50 mM SN26955 with
To determine whether the acridine moiety in SN26955
itself was sensitizing cells to TPZ (rather than targeting the
TPZ to DNA), we incubated HT29 cells under hypoxic
conditions with various concentrations of TPZ with 0, 1, or
5 mM 9-(N-ethylamino)acridine, the DNA-intercalating

1812
Y.M. Delahoussaye et al. / Biochemical Pharmacology 65 (2003) 1807–1815
conditions after the aerobic incubation and washing. This
suggests that some of the SN26955 is retained in the cell,
presumably intercalated into DNA, and is present during
the hypoxic exposures in fresh medium. The fact that the
cytotoxicity of SN26955 following the washout was not as
great as that produced by hypoxic exposure with the drug in
the medium is presumably the result of loss of some of the
drug from the DNA during the washing procedure and
reintroduction of fresh medium. This was confirmed by
leaving the cells that had been exposed in air for 1 hr to
5 mM SN26955 for 30 min, 1, and 2 hr in fresh medium
prior to the hypoxic exposure. We found that the 1- and 2-
hr incubations in fresh medium exhibited no significant
additional toxicity on hypoxic incubation, with the 30-min
exposure being intermediate between no and 1-hr incuba-
tion in drug-free medium following the aerobic incubation
with SN26955.
Fig. 4. Radical production by P450 reductase (0.3 mg/mL) in the presence
of TPZ and SN26955 under hypoxic conditions in vitro. TPZ and SN26955
were incubated under hypoxia with 25 mM DHR123 for 1 hr with different
concentrations of drug and saturating levels of cofactors (500 mM each of
NADH and NADPH). Results pooled from three experiments with the
SEM limits are shown.
P450 reductase (0.3 mg/mL) and found that 100 and
200 mM DMSO inhibited DHR oxidation to Rh123 by
approximately 25 and 50%, respectively (data not shown).
3.5. Enzymatic requirement for DNA damage by SN26955
To determine whether enzymatic reduction was neces-
sary for DNA damage produced by SN26955, we incubated
isolated nuclei from HeLa cells with the drug for 1 hr with
or without saturating concentrations of the two cofactors
NADH and NADPH (500 mM of each). As shown in Fig. 6,
SN26955 produced very few SSBs, as assayed by the
alkaline comet assay, in the absence of cofactors but
produced extensive DNA damage in the presence of the
cofactors. This indicates that the SSBs are a result of
enzymatic reduction of the drug.
To compare the DNA damage produced by SN26955
with that of TPZ, we incubated isolated nuclei with
saturating concentrations of the two cofactors, NADH
and NADPH, with various concentrations of SN26955
and TPZ for 1 hr. Fig. 7 shows that SN26955 was more
potent than TPZ by a factor of approximately 10 in
producing SSBs and was more potent than TPZ in produ-
cing DSBs by a factor of approximately 100.
3.4. Activation of SN26955 to a cytotoxic product when
bound to DNA
We investigated the ability of SN26955 to become
activated while intercalated into DNA by incubating HeLa
cells with the drug under aerobic conditions for 1 hr
followed by two rinses of the cells in PBS, addition of
fresh medium, and then incubation under hypoxic condi-
tions. Fig. 5a shows that when this experiment was per-
formed with TPZ, there was no additional cytotoxicity
produced by the 1-hr exposure under hypoxic conditions in
fresh medium compared to the 1-hr exposure in air alone.
This demonstrates that the washing and addition of fresh
medium remove essentially all TPZ from the cells and that
the 1-hr exposure to hypoxia produces no cytotoxicity.
However, as shown in Fig. 5b, there was increased toxicity
by SN26955 produced by the 1-hr exposure under hypoxic
Fig. 5. Washout experiments examining residual TPZ and SN26955 in cells following extensive washing after a 1-hr exposure of the cells under aerobic
conditions to the respective drug. SN26955 or TPZ was added to HeLa cells for 1 hr at the indicated concentrations for the aerobic and hypoxic killing curves.
In the washout experiment, cells were exposed to the drugs for 1 hr under aerobic conditions. The drug was then removed, and the cells were washed and then
exposed in fresh medium under hypoxic conditions for 1 hr. Panel b shows that this hypoxic exposure produced additional killing over the 1-hr exposure of
the cells to SN26955 under aerobic conditions. Pooled data from two experiments are shown.

Y.M. Delahoussaye et al. / Biochemical Pharmacology 65 (2003) 1807–1815
1813
NADPH as cofactors. We have performed similar experi-
ments with TPZ previously and found a similar cofactor
dependence for SSBs. However, the enzyme(s) metaboliz-
ing TPZ to produce DSBs could only use NADPH as a
cofactor [17]. This different cofactor dependence for the
production of DSBs between TPZ and SN26955 suggests
that the DSBs are produced by different enzymes in the
nucleus.
4. Discussion
Fig. 6. SSBs produced by SN26955 in nuclei with or without cofactors
(500 mM NADH and NADPH). The isolated nuclei were incubated with
various concentrations of SN26955 under hypoxic conditions for 1 hr in
the presence of HeLa cell nuclei with or without cofactors, and the SSBs
were measured using the alkaline comet assay. A few breaks were
produced in the absence of cofactors. Data from three experiments pooled
with SEM limits are shown.
The present series of experiments was conducted to
determine whether attaching a DNA-targeting chromo-
phore to the TPZ moiety would increase the potency of
TPZ yet retain its selective toxicity towards hypoxic cells.
Although a number of investigators have shown that the
potency of some bioreductive drugs can be increased by
targeting them to DNA [30], these examples do not guar-
antee the efficacy of a TPZ moiety targeted to DNA
because, unlike the prior examples, the active reduction
species of TPZ is a highly reactive 1-electron reduction
radical. Thus, because TPZ metabolized to its active
radical in the cytoplasm does not produce any DNA
We also measured the cofactor dependence for DSBs
and SSBs using isolated nuclei (Fig. 8). It can be seen that
SSBs were produced by an enzyme that can use either
NADH or NADPH as a cofactor although there was a
strong preference for NADPH. Likewise, DSBs were
produced by an enzyme that can use both NADH and
Fig. 7. SSBs and DSBs produced in isolated nuclei by TPZ and SN26955. TPZ or SN26955 was incubated with HeLa cell nuclei under hypoxic conditions
with saturating levels of cofactors (500 mM each of NADH and NADPH) for 1 hr. The SSBs and DSBs were assayed by the alkaline and neutral comet assays,
respectively. SN26955 was approximately 10- and 100-fold more efficient than TPZ in producing SSBs and DSBs, respectively. Pooled data from three
experiments with SEM limits are shown.
Fig. 8. Cofactor dependencies for SSBs and DSBs produced in isolated nuclei by SN26955. HeLa cell nuclei were incubated under hypoxic conditions for
1 hr with 0.25 mM SN26955 with various concentrations of NADH or NADPH or both cofactors, and the SSBs and DSBs were measured using the alkaline
and neutral comet assays, respectively. Data pooled from three experiments with the SEM limits are shown.

1814
Y.M. Delahoussaye et al. / Biochemical Pharmacology 65 (2003) 1807–1815
damage [15], it is possible that a DNA-targeted TPZ
molecule would be metabolized to its reactive radical prior
to entry into the nucleus and binding to DNA, thus
rendering it inactive. Also, the efficacy of a TPZ tethered
to DNA would depend on the accessibility of the reducing
enzymes to the TPZ moiety on the DNA.
are metabolized to the radical that eventually gives rise to
DSBs by a different enzyme. As there are different TPZ-
metabolizing enzymes in different locations within the
nucleus [17], this suggests that the DSBs produced by
TPZ and SN26955 occur in different intranuclear loca-
tions. We have shown recently that some, if not all, of the
cytotoxicity of TPZ is produced by radical-stimulated
damage to, or activation of, topo II [19]. This (along with
the present data) is consistent with a model in which TPZ
can be metabolized throughout the nucleus, but it is only
the radicals produced close to topo II (located at the nuclear
matrix) that induce DSBs, whereas the DSBs induced by
SN26955 are produced throughout the DNA, not just in the
short stretches attached to the nuclear matrix. This would
imply that DSBs induced by TPZ are not randomly dis-
tributed along the DNA, but rather produce fragments
consistent with the 60- to 100-kb loops of DNA between
matrix attachment sites [16]. On the other hand, we would
expect that DNA fragments produced by the DSBs induced
by SN26955 would be more randomly distributed through-
out the DNA. These predictions are currently being tested.
In summary, we have demonstrated that TPZ attached by
a linker to an acridine chromophore binds to DNA and is
metabolized at the DNA to produce both SSBs and DSBs.
Further, this targeted construct is considerably more potent
than TPZ, yet retains its selective hypoxic cytotoxicity.
This increased potency and selective accumulation in the
nucleus can potentially provide two important advantages
in the clinical use of TPZ. First, it can greatly diminish the
metabolism that occurs within mitochondria that produces
loss of the transmembrane potential of mitochondria [21].
This is potentially responsible for many of the unique
toxicities of TPZ in the clinic including reversible hearing
loss, muscle cramps, and fatigue, all of which could be due
to this effect on mitochondria in aerobic tissue. Second, by
reduction of the cytoplasmic metabolism of TPZ, much of
the unproductive metabolism can be eliminated, thereby
allowing the drug to diffuse further in the tumor [22].
However, if the DNA binding is too great, this is likely to
restrict diffusion through the tumor and reduce access to
the distant hypoxic cells. Indeed, this is likely to be the case
with SN26955 since it is not active in potentiating tumor
cell kill by irradiation in vivo.1 Nonetheless, the present
data provide proof of principle for further drug develop-
ment with drugs with lower DNA-binding affinities. Such
studies are in progress.
The present data, however, show that targeting of TPZ
by attaching it via a linker to an acridine chromophore to
create the compound SN26955 increases the potency of
TPZ by 1–2 orders of magnitude, depending on the cell
line, while retaining its selective toxicity towards hypoxic
cells (Fig. 3). These data strongly suggest that the TPZ
moiety is reduced to its reactive radical while bound by the
acridine chromophore to DNA. This is consistent with our
recent finding that there are TPZ-metabolizing enzymes in
close association with the nuclear matrix [17], which is the
organizing structure for the intranuclear DNA [16]. Further
evidence for the cytotoxicity of SN26955 bound to DNA
was provided by the washout experiment, in which cells
incubated with the drug under aerobic conditions, washed,
and then immediately exposed to hypoxia showed
increased cytotoxicity compared to the aerobic exposure
alone. This did not occur for TPZ itself, which is not bound
to DNA and, therefore, not retained in the cell following
washing and medium change (Fig. 5).
Our finding that almost all of the DNA damage assessed
by SSBs in isolated nuclei produced by SN26955 depends
on the presence of reduced pyridine nucleotides as cofac-
tors (Fig. 6) demonstrates that the DNA damage, and
presumably cytotoxicity, is dependent upon enzymatic
reduction. We also show that the increased potency of
SN26955 compared to TPZ is probably not the result of an
increased rate of reduction of the compound to its oxidiz-
ing radical because the rates of radical reduction by the
drug and TPZ are identical, at least for NADPH cyto-
chrome P450 reductase (Fig. 4). Despite this similarity in
the rate of formation of the oxidizing radical, we found that
SN26955 in isolated nuclei produced a greater number of
DNA SSBs than did TPZ (by a factor of approximately 10),
and an even greater increase in the number of DSBs (by a
factor of approximately 100) compared to TPZ (Fig. 7).
This suggests that a larger proportion of the SSBs produced
by SN26955 is converted into DSBs, presumably because
of the presence of the intercalating chromophore.
When we investigated the cofactor dependence of SSBs
and DSBs produced by SN26955, we found a qualitative
difference between this drug and TPZ. As shown in Fig. 8,
the production of both SSBs and DSBs in isolated nuclei by
SN26955 required an enzyme(s) that can use either NADH
or NADPH cofactor although, at least for SSBs, the
NADPH cofactor is favored. Our previous work with
TPZ [17] showed a similar pattern for SSBs (ability to
use both cofactors but a preference for NADPH), but a
different cofactor dependence with DSBs with only
NADPH useable. This difference in the cofactor require-
ments for TPZ and SN26955 suggests that the two drugs
Acknowledgments
This work was funded by Grant CA 82566 from the US
National Cancer Institute. We thank H.D. Sarath Liyanage
for measuring the DNA binding and Dr. William R. Wilson
for helpful comments.
1 Dorie MJ and Brown JM, unpublished data.

Y.M. Delahoussaye et al. / Biochemical Pharmacology 65 (2003) 1807–1815
1815
[15] Evans JE, Yudoh K, Delahoussaye YM, Brown JM. Tirapazamine is
metabolized to its DNA damaging radical by intranuclear enzymes.
Cancer Res 1998;58:2098–101.
References
[1] Brown JM. SR 4233 (tirapazamine): a new anticancer drug exploiting
hypoxia in solid tumours. Br J Cancer 1993;67:1163–70.
[2] Denny WA, Wilson WR. Tirapazamine: a bioreductive anticancer drug
that exploits tumour hypoxia. Expert Opin Investig Drugs 2000;9:
2889–901.
[16] Berezney R, Wei X. The new paradigm: integrating genomic func-
tion and nuclear architecture. J Cell Biochem 1998;30-31(Suppl):
238–42.
[17] Delahoussaye YM, Evans JE, Brown JM. Metabolism of tirapazamine
by multiple reductases in the nucleus. Biochem Pharmacol 2001;62:
1201–9.
[3] Nordsmark M, Overgaard M, Overgaard J. Pretreatment oxygenation
predicts radiation response in advanced squamous cell carcinoma of
the head and neck. Radiother Oncol 1996;41:31–40.
[18] Peters KB, Wang H, Brown JM, Iliakis G, Inhibition of DNA
replication by tirapazamine. Cancer Res 2001;61:5425–31.
[19] Peters KB, Brown JM. Tirapazamine: a hypoxia-activated topoisome-
rase II poison. Cancer Res 2002;62:5248–53.
[4] Brizel DM, Sibley GS, Prosnitz LR, Scher RL, Dewhirst MW. Tumor
hypoxia adversely affects the prognosis of carcinoma of the head and
neck. Int J Radiat Oncol Biol Phys 1997;38:285–9.
[20] Fortune JM, Osheroff N. Topoisomerase II as a target for anticancer
drugs: when enzymes stop being nice. Prog Nucleic Acid Res Mol Biol
2000;64:221–53.
[5] Sartorelli AC. Therapeutic attack of hypoxic cells of solid tumors:
presidential address. Cancer Res 1988;48:775–8.
[6] Brown JM, The hypoxic cell: a target for selective cancer therapy—
eighteenth Bruce F. Cain Memorial Award lecture. Cancer Res 1999;
59:5863–70.
[21] Wouters BG, Delahoussaye YM, Evans JW, Birrell GW, Dorie MJ,
Wang J, MacDermed D, Chiu RK, Brown JM. Mitochondrial dysfunc-
tion after aerobic exposure to the hypoxic cytotoxin tirapazamine.
Cancer Res 2001;61:145–52.
[7] Craighead PS, Pearcey R, Stuart G. A phase I/II evaluation of
tirapazamine administered intravenously concurrent with cisplatin
and radiotherapy in women with locally advanced cervical cancer.
Int J Radiat Oncol Biol Phys 2000;48:791–5.
[22] Hicks KO, Fleming Y, Siim BG, Koch CJ, Wilson WR. Extravascular
diffusion of tirapazamine: effect of metabolic consumption assessed
using the multicellular layer model. Int J Radiat Oncol Biol Phys
1998;42:641–9.
[8] von Pawel J, von Roemeling R, Gatzemeier U, Boyer M, Elisson LO,
Clark P, Talbot D, Rey A, Butler TW, Hirsh V, Olver I, Bergman B,
Ayoub J, Richardson G, Dunlop D, Arcenas A, Vescio R, Viallet J,
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 2000;18:1351–9.
[23] Kyle AH, Minchinton AI. Measurement of delivery and metabolism of
tirapazamine to tumour tissue using the multilayered cell culture
model. Cancer Chemother Pharmacol 1999;43:213–20.
[24] Robbins RF, Schofield K. Polyazabicyclic compounds. Part II. Further
derivatives of benzo-1,2,4-triazine. J Chem Soc 1957:3186–94.
[25] Henderson LM, Chappell JB. Dihydrorhodamine 123: a fluorescent
probe for superoxide generation? Eur J Biochem 1993;217:973–80.
[26] Bump EA, Cerce BA, Al-Sarraf R, Pierce SM, Koch CJ. Radio-
protection of DNA in isolated nuclei by naturally occurring thiols at
intermediate oxygen tension. Radiat Res 1992;132:94–104.
[27] Olive PL. Detection of hypoxia by measurement of DNA damage in
individual cells from spheroids and murine tumours exposed to
bioreductive drugs. I. Tirapazamine. Br J Cancer 1995;71:529–36.
[28] McGhee JD, von Hippel PH. Theoretical aspects of DNA-protein
interactions: co-operative and non-co-operative binding of large li-
gands to a one-dimensional homogeneous lattice. J Mol Biol 1974;
86:469–89.
[9] Rischin D, Peters L, Hicks R, Hughes P, Fisher R, Hart R, Sexton M,
D’Costa I, von Roemeling R. Phase I trial of concurrent tirapazamine,
cisplatin, and radiotherapy in patients with advanced head and neck
cancer. J Clin Oncol 2001;19:535–42.
[10] Wang J, Biedermann KA, Brown JM. Repair of DNA and chromosome
breaks in cells exposed to SR 4233 under hypoxia or to ionizing
radiation. Cancer Res 1992;52:4473–7.
[11] Daniels JS, Gates KS, Tronche C, Greenberg MM. Direct evidence for
bimodal DNA damage induced by tirapazamine. Chem Res Toxicol
1998;11:1254–7.
[12] Baker MA, Zeman EM, Hirst VK, Brown JM. Metabolism of SR 4233
by Chinese hamster ovary cells: basis of selective hypoxic cytotoxi-
city. Cancer Res 1988;48:5947–52.
[29] Lloyd RV, Duling DR, Rumyantseva GV, Mason RP, Bridson PK.
Microsomal reduction of 3-amino-1,2,4-benzotriazine 1,4-dioxide to a
free radical. Mol Pharmacol 1991;40:440–5.
[13] Wang J, Biedermann KA, Wolf CR, Brown JM. Metabolism of the
bioreductive cytotoxin SR 4233 by tumour cells: enzymatic studies. Br
J Cancer 1993;67:321–5.
[30] Rauth AM, Melo T, Misra V. Bioreductive therapies: an overview of
drugs and their mechanisms of action. Int J Radiat Oncol Biol Phys
1998;42:755–62.
[14] Patterson AV, Saunders MP, Chinje EC, Patterson LH, Stratford IJ.
Enzymology of tirapazamine metabolism: a review. Anticancer Drug
Des 1998;13:541–73.