Oxidatively activated DNA-modifying agents for selective cytotoxicity

Article abstract of DOI:10.1002/cmdc.201100014

DNA-modifying agents are stalwarts of chemotherapeutic cancer treatments, but require significant design improvements to improve selectivity, minimize side effects, and for their widespread use to continue. Herein we present a novel design strategy in which DNA-modifying agents contain an oxidizable leaving group and a nitrogen mustard. The agents form strong electrophiles specifically when oxidized. Activation, measured by hydrolysis, illustrates that oxidants increase reactivity 1700-fold. Reaction in the presence of 2'-deoxyguanosine leads to the formation of lesions. Cytotoxicity measured in HeLa cells showed that low IC₅₀ values require an oxidizable hydroquinone and a nitrogen mustard fragment. Cytotoxicity measurements in 15 cancer cell lines demonstrates that oxidatively activated DNA-modifying agents are highly selective, as the analogue tested has IC₅₀ values less than 10μM for only three of the 15 cell lines; in contrast, cisplatin is highly toxic to 13 of the 15 cell lines. The selective cytotoxicity of oxidatively activated DNA-damaging agents could be useful against kidney cancer cells, as the 786-O cell line model assay resulted in an IC₅₀ value of 5μM. DNA-modifying agents require design improvements to lower side effects. A novel design strategy in which agents that contain oxidizable leaving groups is shown. Oxidation increases reactivity 1700-fold. Reactions occur between the agent and 2'-deoxyguanosine, and compounds in this class selectively target the renal carcinoma model cell line 786-O.

Full text of DOI:10.1002/cmdc.201100014

DOI: 10.1002/cmdc.201100014
Oxidatively Activated DNA-Modifying Agents for Selective
Cytotoxicity
Guorui Li, Tiffany Bell, and Edward J. Merino*[a]
DNA-modifying agents are stalwarts of chemotherapeutic
cancer treatments, but require significant design improve-
ments to improve selectivity, minimize side effects, and for
their widespread use to continue. Herein we present a novel
design strategy in which DNA-modifying agents contain an ox-
idizable leaving group and a nitrogen mustard. The agents
form strong electrophiles specifically when oxidized. Activa-
tion, measured by hydrolysis, illustrates that oxidants increase
reactivity 1700-fold. Reaction in the presence of 2’-deoxygua-
nosine leads to the formation of lesions. Cytotoxicity measured
in HeLa cells showed that low IC50 values require an oxidizable
hydroquinone and a nitrogen mustard fragment. Cytotoxicity
measurements in 15 cancer cell lines demonstrates that oxida-
tively activated DNA-modifying agents are highly selective, as
the analogue tested has IC50 values less than 10 mm for only
three of the 15 cell lines; in contrast, cisplatin is highly toxic to
13 of the 15 cell lines. The selective cytotoxicity of oxidatively
activated DNA-damaging agents could be useful against
kidney cancer cells, as the 786-O cell line model assay resulted
in an IC50 value of 5 mm.
Introduction
DNA-modifying agents are stalwarts of chemotherapeutic
cancer treatments, but require vital design improvements to
lower side effects and to continue their widespread use.[1]
These agents react with DNA bases, usually at guanine, to halt
DNA replication at the reaction site.[2] This reaction, although
not selective, ultimately leads to cell death by one of several
pathways.[3–5] The two common reactive scaffolds used in alky-
lating agents, N,N-bis(2-chloroethyl)amino groups and platinu-
m(II) centers, have led to a variety of chemotherapeutic
agents. However, these suffer from reactivity in non-cancerous
cells which gives rise to many side effects.[6] Recent work has
focused on the synthesis of molecules that attach reactive scaf-
folds to motifs that enter cancer cells specifically.[7–9] Our re-
search takes an alternative design strategy wherein the reac-
tive scaffold is stable, but is activated by processes that occur
more frequently in cancer cells. Thus reactivity, and not uptake,
is controlled to induce cytotoxicity more specifically in cancer
cells and to lower off-target reactions. One process that has
been shown to occur in some cancers is the elevation of reac-
tive oxygen species (ROS). We hypothesize that the design of
highly selective DNA-modifying agents can be achieved by oxi-
dative activation.
mol involve oxidative-stress-dependent toxicity.[18,19] Several
new chemotherapy strategies are being examined to take ad-
vantage of increased reactive oxygen levels within cancerous
cells relative to healthy cells.[20–22] Thus, DNA-modifying agents
that are activated by ROS can be more specific than the cur-
rent agents if an appropriate activation strategy can be de-
signed (Figure 1).
Figure 1. Oxidative-stress-induced cytotoxicity: Many cancers have been
shown to possess elevated levels of reactive oxygen. We hypothesize that
oxidation-prone leaving groups coupled to DNA-modifying agents can lead
to specific control of cytotoxicity by reaction with cellular components.
Oxidative stress, as measured by levels of reactive oxygen,
oxidized biomolecules, and enzyme activity, is a hallmark of
cancer cells and excised tumors.[10–12] For instance, hypoxic
tumors have been shown to experience increases in NADPH
oxidases that give off elevated ROS as a byproduct of their cat-
alysis.[13] The mitochondrion is another source of reactive
oxygen in cells.[14,15] Mitochondria generate hydrogen peroxide
via complex I in oxidative phosphorylation.[16,17] Recently, oxida-
tive stress has been found to play a key role in the activation
of several agents either in or close to clinical trials. For in-
stance, mechanisms of both deoxynyboquinone and elesclo-
[a] G. Li, T. Bell, Prof. E. J. Merino
Department of Chemistry
University of Cincinnati, Cincinnati, OH 45221-0172 (USA)
Fax: (+1)513-556-9239
E-mail: merinoed@uc.edu
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201100014.
ChemMedChem 2011, 6, 869 – 875
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
869

E. J. Merino et al.
MED
Our design strategy focuses on selective activation that
occurs at the leaving group (Figure 1). Importantly, control of
reactivity at the leaving group maximizes flexibility because
the remainder of the agent is left intact for further structural
optimization. Traditional nitrogen mustards are not selective
because they are strong electrophiles, leading to off-target re-
activity; the chlorine leaving group has a pKa value of À4. We
instead synthesized agents that are structurally similar to nitro-
gen mustards. These agents do not have a chlorine leaving
group, but instead use a hydroquinone at the leaving group
position. The hydroquinone is a poor leaving group, with a pKa
value of 10.3. Owing to the hydroquinone, reactivity via a tradi-
tional mechanism is limited. Oxidation of the hydroquinone-
coupled DNA-modifying agent forms a strong electrophile and
activates reactivity (Figure 1). Therefore, the cytotoxic reactivity
of the agent is intimately coupled with the elevated cellular
oxidative stress present in many cancers. Several model com-
pounds were synthesized for the examination of our design
strategy. Aniline-based nitrogen mustards have been used in a
variety of prodrug strategies and clinically relevant alkylating
agents.[23] Thus, compounds with this scaffold were synthesized
(Scheme 1).
Ag/AgCl, making it an appropriate choice for the oxidation of
hydroquinone derivatives.[24] Cells produce ROS from both hy-
droxyl radical and singlet oxygen. Copper can be used to pro-
duce both oxygen species.[25,26] Oxidation of An-Hq by CuCl2–
H2O2 led to rapid hydrolysis with a half-life of 12.1Æ0.7 h. The
extent of hydrolysis depends on the concentration of hydro-
gen peroxide, and was slow due to the lack of ascorbate (data
not shown). Oxidases such as horseradish peroxidase reduce
molecular oxygen to either hydrogen peroxide or water and at
the same time oxidize a substrate with an appropriate poten-
tial. Thus, we determined the extent of reaction of An-Hq in
the presence of horseradish peroxidase as an oxidase. The re-
action was too rapid to measure by HPLC, as after 1 min ap-
proximately half (53Æ13%) of the An-Hq was hydrolyzed. The
hydrolysis half-life of An-Hq was increased >1700-fold upon
addition of oxidative equivalents.
Comparison with standards characterized by RP HPLC was
used to verify the hydrolysis reaction products of An-Hq (Fig-
ure 2C). In the presence of IrIV, An-OH (tR =19 min) and benzo-
quinone (tR =10 min) were observed. An-OH is the product of
hydrolysis, as water has displaced the hydroquinone leaving
group. Copper–hydrogen peroxide and peroxidase treatment
produced the same products. These data show that
once molecules with hydroquinone-based leaving
groups enter a cell, oxidases and several forms of
ROS will activate reactivity.
Reaction with 2’-deoxyguanosine
Scheme 1. General synthesis reagents and conditions: a) SOCl2, 408C, 1 h; b) K2CO3, KI,
808C, hydroquinone (5 equiv), 6 h. X=H, OH.
Reactivity toward DNA was investigated. By investi-
gating DNA reactivity the likely cytotoxic reaction
product will be elucidated (Figure 3B). DNA-modify-
ing agents induce cytotoxicity in cells by modifica-
Results and Discussion
Oxidative activation
The oxidative activation of the DNA-modifying
agents was investigated. Water, one of the simplest
nucleophiles that can serve as a substrate, was exam-
ined under various oxidative conditions (Figure 2). A
model modified aniline nitrogen mustard analogue,
An-Hq, was synthesized in two steps. An-Hq is similar
to a nitrogen mustard, except one of the alkylating
arms is nonreactive, whereas the second arm pos-
sesses a single hydroquinone replacing the chlorine
leaving group (Figure 2A).
The hydrolysis of An-Hq as a function of time was
quantified by separation of the reaction products by
Figure 2. Hydrolysis of an oxidatively activated agent: A) Model compound An-Hq was
reversed-phase HPLC (Figure 2B). In phosphate buffer synthesized. An-Hq possesses a hydroquinone leaving group that reacts slowly with nu-
cleophiles (reaction). In the presence of oxidative stress, An-Hq is activated, leading to
hydrolysis was slow, with a half-life of 693Æ42 h as-
rapid reactivity and formation of benzoquinone (Hqox) and An-OH. B) The hydrolysis of
suming first-order kinetics. To obtain quantifiable
An-Hq was monitored by HPLC at l 260 nm at the indicated times. In the absence of oxi-
data, time points were taken daily over the course of
dant (black, top), An-Hq is highly stable in buffered solutions. Addition of CuCl2–H2O2
one week. Thus, An-Hq is a poor DNA-modifying (green) or IrIV (orange) resulted in rapid hydrolysis. Cellular antioxidant proteins such as
peroxidases (blue), can also rapidly oxidize An-Hq. C) The reaction products were identi-
agent if not activated by oxidation. Addition of the
fied by comparison with standards. An-Hq elutes at tR 22 min (black). Addition of an oxi-
one-electron oxidant Na2Ir2Cl6 led to quantitative hy-
dant led to the formation of new products at tR 19 and 10 min (orange). Comparison
with standards of An-OH (top, blue) and Hqox (top, light green) revealed that hydrolysis
drolysis of An-Hq with a half-life of 0.4Æ0.2 h (Fig-
ure 2B). The oxidation potential of IrIV is 0.7 V versus occurs via elimination of benzoquinone.
870
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 869 – 875

Oxidatively Activated DNA-Modifying Agents
served at ~4 min. After several purifications, the
product was concentrated threefold (complete
drying leads to product degradation, as many modi-
fied nucleosides possess labile glycosidic bonds). In-
terestingly, the product possessed an absorbance at
l 330 nm, indicating the addition of an aromatic
group to the guanine base (Supporting Information).
The isolate was subjected to HRMS to identify the
product (Figure 3D). Notably, in all cases, errors in
ion masses are <750 ppb. Therefore, differences in
mass such as that between O and NH2, a difference
of 942 ppm, are easily distinguishable and permit as-
signment of the elemental composition even on dis-
sociated fragments in MS–MS. The spectrum showed
a product with a mass of 358.1146 Da and an ele-
+
mental composition of C16H16N5O5 (theoretical mass:
358.1146 Da, error: 126 ppb). A mass of 358 Da dem-
onstrates the addition of a benzoquinone group with
loss of water. The published structure for such a
lesion is shown in Figure 3B. Collision-induced disso-
ciation induced fragmentation of the lesion. MS–MS
led to a loss of C5H8O3 (error: 666 ppb), which is a
common fragmentation product of nucleosides at
the glycosidic bond, representing the loss of 2’-deox-
yribose (Figure 3E). MS–MS confirms that the adduct
contains 2’-deoxyguanosine and that modification
did not occur on the ribose ring. Thus, the adduct
was neither the Michael addition or alkylation prod-
uct, but rather a guanine lesion modified at two het-
eroatom positions. Notably, other isomers are possi-
ble with connectivity at ring positions, but these iso-
mers are only distinguishable by selective isotopic la-
beling. The structure listed corresponds to the isomer
identified in the literature.[27]
Figure 3. Oxidation-induced modification of 2’-deoxyguanosine: A) The reaction of An-
Hq with 2’-deoxyguanosine was analyzed. B) Three possible products with 2’-deoxygua-
nosine are (from left to right): the Michael addition product, the alkylation product, and
a product that has been reported in the literature. C) The reaction was monitored by
HPLC at l 260 nm. In the absence of 2’-deoxyguanosine, hydrolysis proceeds (grey). The
presence of the nucleoside leads to formation of a specific product (black). D) The prod-
uct was partially dried and analyzed by MS. Observed m/z ratios are in grey. Injection of
the sample showed the product has the same mass and elemental composition as one
of the possible products. Collision-induced dissociation led to a single fragment ob-
served. E) Theoretical mass in grey and bold; errors <750 ppb. Superimposing the pro-
posed structure shows correlation with the structure. The single fragmentation observed
is cleavage at the glycosidic bond.
Structural investigation into cytotoxicity
The cytotoxicity of An-Hq and its derivatives were
evaluated in HeLa cells (Figure 4). Cell viability was
monitored by MTT dye, which forms a purple forma-
zan product in viable cells that are undergoing me-
tion of either guanine or adenine bases (Figure 3A). We there-
fore probed the reactivity of An-Hq with 2’-deoxyguanosine.
An-Hq is a bifunctional molecule that can modify DNA via sev-
eral reaction pathways; we envisioned three likely routes. Qui-
none chemistry is well studied, and the predominant reaction
product arises via Michael addition. This reaction product
occurs after the formation of benzoquinone (Hqox, Figure 3) via
nucleophilic addition of N7 of the guanine base. Moreover, tra-
ditional N7-alkylation of the guanine base (Figure 3B) could
also be possible. These adducts have masses of Mr 376 and
415 Da, respectively. Finally, a survey of the literature found
that some isomers of hydroquinones form guanine lesions in
which N1 and N2 have added to produce a substituted phenol
(Figure 3B).[27] 2’-Deoxyguanosine was incubated with an iridi-
um oxidant and An-Hq for 7 days. Reaction mixture (5 mg) was
then purified by RP HPLC (Figure 3C). In comparison with a
control lacking 2’-deoxyguanosine, a reaction product was ob-
Figure 4. High cytotoxicity requires hydroquinone and a nitrogen mustard:
Structures of each compound tested are on the right with the correspond-
ing point. Hydroquinone (&) and the corresponding oxidized form, benzo-
^
quinone ( ), were only slightly toxic, with IC50 values of 161 and 227 mm, re-
&
spectively. An analogue equipped with a chlorine leaving group ( ) had an
~
estimated an IC50 value of 211 mm. The IC50 of An-Hq ( ) was 23 mm.)
ChemMedChem 2011, 6, 869 – 875
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
871

E. J. Merino et al.
MED
tabolism. Concentrations of An-Hq and its analogues ranged
from 0 to 100 mm, and IC50 values were calculated with a four-
parameter regression analysis of at least three replicates. Errors
reported are between two separate experiments; curve-fitting
errors were <25% (R>0.98). An-Hq was toxic to HeLa cells
both An-Hq-CH3 and An-Hq would have the same tendency
toward enzymatic degradation. Additionally, we synthesized
An-Hq2. An-Hq2 produces the same guanine base lesion as An-
Hq (Supporting Information), and has an IC50 value of 11Æ
4 mm. Importantly, An-Hq2 is twice as cytotoxic as An-Hq, as
upon activation two toxic groups are delivered. The therapeut-
ically useful agent cisplatin had an IC50 value similar to that of
An-Hq2, while chlorambucil had a lower cytotoxicity of 76Æ
10 mm. Thus, although An-Hq2 leads to modification of a single
DNA base and does not induce cross-links, it can still lead to
equivalent cytotoxicity.
~
with an IC50 value of 23Æ4 mm (Figure 4, ). In contrast, the
unmodified nitrogen mustard with a chlorine leaving group
&
(Figure 4, ) has much lower cytotoxicity, with an IC50 value of
211Æ17 mm. Cytotoxicity was not derived from the hydroqui-
none portion of An-Hq or from the benzoquinone oxidation
product, as IC50 values of these compounds were 161Æ14 (&)
^
and 227Æ21 mm ( ), respectively (Figure 4). The data taken to-
gether illustrate that the hydroquinone and the DNA-modifica-
tion portion of the molecule are necessary for low IC50 values.
The cytotoxicity of several derivatives was monitored in
HeLa cells (Table 1). To examine the relationship between oxi-
dation potential and cytotoxicity, several quinone derivatives
Cytotoxicity in 15 cell lines
These data prompted us to further investigate the cytotoxicity
of oxidatively activated DNA-modifying agents toward several
cell lines (Table 2). Cisplatin, a traditional alkylating agent, was
used for comparison against An-Hq2. Fif-
teen cancer cell lines were compared. The
Table 1. Comparison of the cytotoxicity of several synthesized derivatives to that of known DNA-
modifying agents in HeLa cells.
cell lines tested are a large set of lines
from 11 different cancer types giving high
diversity to the cytotoxicity measure-
ments. Importantly, five of the 15 cell
lines analyzed had IC50 values >200 mm.
The oxidative activation of An-Hq2
showed a much different cytotoxicity pro-
file than the profile obtained from cispla-
tin (Table 2). In contrast, IC50 values
>200 mm were observed for only one of
the 15 cell lines. Many cells seem to be re-
sistant to the cytotoxic mechanism of An-
Hq2. Furthermore, An-Hq2 showed impor-
tant differences in IC50 values of the re-
maining 10 cell lines. The IC50 values
ranged from 5 to 93 mm with an average
of 37 mm and a deviation of 33 mm. Only
Name
Structure
IC50 [mm][a]
Name
Structure
IC50 [mm][a]
>300
An-
Hq
23
An-Hq-CH3
An-Rs
44
71
An-Cl2
10
77
An-
Cat
chlorambucil
An-
Hq2
11
cisplatin
9
[a] Values determined by MTT assay, with errors from triplicate experiments <25% of the listed
value.
were synthesized. Resorcinol, with a 1,3-diol, is less readily oxi-
Table 2. Comparison of the cytotoxicity of several synthesized derivatives
to that of known DNA-modifying agents in HeLa cell.
dized than hydroquinone, which has hydroxy groups at the 1-
and 4-positions.[28] Resorcinol was attached to An-Cl to create
An-Rs. An-Rs has an IC50 value of 44Æ10 mm. Additionally, an
analogue was synthesized in which a catechol group is at-
tached to An-Cl. Catechol was recently used in the selective
modification of DNA in cancer cells.[29] Catechol, with a 1,2-diol,
also shows a decreased IC50 value of 71 mm. We synthesized a
derivative that is not prone to oxidation: An-Hq-CH3. An-Hq
was methylated with methyl iodide. Hydroquinones oxidize
from the phenoxide anion. Methylation does not allow forma-
tion of the required anion, and thus prohibits oxidation of hy-
droquinone. Addition of IrIV does not produce the oxidation-
dependent color change to brown (data not shown). An-Hq-
CH3 showed no detectable toxicity toward HeLa cells (Table 1).
These synthesized analogues illustrate that cytotoxicity corre-
lates with oxidation potential.
Cell line
IC50 [mm][a]
Cancer
An-Hq2
Cisplatin
786-O
5
4
36
2
10
>200
Kidney
Kidney
Breast
Prostate
Colorectal
Colorectal
Pancreatic
Pancreatic
Melanoma
Oral
Hek-293
MDA-MB-231
PC-3
DLD-1
LOVO
PANC-1
BX-PC-3
SK-MEL-5
KB
>200
27
>200
>200
93
8
5
26
6
6
4
20
99
9
SK-OV-3
OVCAR3
HeLa
33
>200
10
6
5
9
Ovarian
Ovarian
Cervix
Calu-6
NCI-N87
>200
42
2
4
Lung
Gastric
We examined the mechanism of cytotoxicity further. An-Hq-
CH3 shows that cytotoxicity is induced through oxidation, and
not some other enzymatic process such as hydrolysis, because
[a] Values determined by MTS assay, with errors from triplicate experi-
ments <25% of the listed value.
872
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 869 – 875

Oxidatively Activated DNA-Modifying Agents
conducted on silica gel (230–400 mesh). All synthesized com-
pounds were additionally analyzed by IR spectroscopy and HRMS.
three cell lines were observed with high cytotoxicity (IC50
values <10 mm). This level of discrimination of An-Hq2 is in
stark contrast to cisplatin cytotoxicity. Cisplatin gave average
IC50 values of 7 mm with a deviation of 6 mm. IC50 values
<10 mm were observed in most cells. Cisplatin was highly cyto-
toxic to 12 of the remaining 14 cell lines. An-Hq2 had high cy-
totoxicity against 786-O kidney cancer cells, which are used as
a model for renal cell carcinoma. Therefore, oxidation-activated
DNA-modifying agents showed highly selective cytotoxicity.
Hydrolysis studies
Hydrolysis reactions were performed in 1.5 mL total volume with
0.6 mm analogue, 5% CH3CN, 25 mm phosphate buffer (pH 7.4) in
a 2 mL glass HPLC tube. Time points were obtained by repetitive
injection of 5 mL into a Beckman Coulter System Gold instrument
with an autosampler equipped with a diode array detector. Solvent
A was 98% H2O and 2% CH3CN; solvent B was 95% CH3CN and
5% H2O. The column used was an Agilent Zorbax SB-C18 (5 mm,
4.6ꢁ150 mm). The gradient was 0% B for 1 min, 10% B over
Conclusions
8 min, 100%
B over 8 min. Absorbance was monitored at
In summary, our data demonstrate that oxidatively active leav-
ing groups can be used to control the cytotoxicity of DNA-
modifying agents. Oxidation-activated DNA-modifying agents
are unreactive until activated by oxidative stress. ROS and oxi-
dases can serve to activate these types of agents. The reactivi-
ty of a given agent is tuned by the oxidation potential of the
leaving group; the substitution pattern of the nitrogen is
intact and may be further modified. These agents likely modify
DNA by the addition of a phenol to guanine at two ring posi-
tions such as N1 and N2. Thus, these large structural modifica-
tions likely induce the observed cytotoxicity. Importantly, de-
spite not forming cross-links, oxidatively activated DNA-modi-
fying agents have cytotoxicity similar to that of clinically rele-
vant agents. Therefore, oxidatively activated DNA-modifying
agents represent a design improvement over traditional alky-
lating agents and may help address the problem of severe side
effects caused by off-target reactivity in noncancerous cells.
This conclusion is supported by data that show one third of
the cell lines tested exhibit limited cytotoxicity, whereas many
other cell lines have low (>50 mm) cytotoxicity.
When tested in a diverse panel of 15 cell lines, oxidatively
activated DNA-modifying agents induce high cytotoxicity in
786-O renal cell carcinoma, which requires more therapeutic
options.[30] Renal cell carcinoma is marked by substantial
changes to redox homeostasis such that oxidative stress has
an important role in the growth and development of this type
of cancer.[31,32] Reactive oxygen-activated DNA-modifying
agents directly contradict the paradigm that they cannot be
designed without high levels of nonspecific reactivity, as the
agents synthesized in this work do show selective cytotoxicity.
l 260 nm. CuCl2, Na2IrCl6, and peroxide concentrations were
0.1 mm, 1.2 mm, and 50 mm, respectively. Horseradish peroxidase
was used at a concentration of 0.02 UmLÀ1. Each experiment was
performed in at least duplicate; half-lives are reported. Curve fit-
ting for hydrolysis in H2O involved a first-order fit, while other half-
lives are empirical. Time points were taken as indicated. Purification
of the DNA reaction column was carried out by using an Alltech
C18 Rocket (7 ID, 53 mm, 3 mm) on a 5 mg reaction. Solvent A was
95% H2O and 5% CH3CN; solvent B was 95% CH3CN and 5% H2O.
The gradient was 0% B for 1 min and then 100% B over 20 min.
The collected fraction was vacuum dried to one third the volume.
MS identification was performed directly on an HPLC fraction that
contained 0.25% acetic acid and 15% CH3CN. Analyses involved in-
fusion directly into the instrument at 5 mLminÀ1. MS analysis was
performed on a Thermo-Fisher Scientific LTQ-FT, a hybrid instru-
ment consisting of a linear ion trap and a Fourier transform ion cy-
clotron resonance mass spectrometer. The entire eluent was intro-
duced into the LTQ-FT using the standard electrospray ionization
source for the instrument, with a spray voltage of 5 kV and a capil-
lary temperature of 2758C. Autogain control (AGC) was used and
set at 500000 with a maximum injection time of 1250 ms for FT-
ICR full scans. Collision-induced dissociation, MS–MS, was executed
in the linear trap with an AGC setting of 10000 and a maximum in-
jection time of 500 ms. FT-ICR full scans were acquired in the posi-
tive ion mode at 100000 resolving power at m/z 400.
Cell culture and cytotoxicity measurements
Human cervical carcinoma (HeLa) cells were obtained from the
ATCC for cell culture. The cells were cultured in DMEM supple-
mented with 10% fetal bovine serum, 100 UmLÀ1 penicillin, and
100 mgmLÀ1 streptomycin (Invitrogen) at 378C in a humidified at-
mosphere containing 5% CO2. MTT assays were performed by
seeding cells at a density of 2ꢁ105 cells per well in a 24-well plate
and incubation at 378C for 5 h. Cells were then treated for 65 h
with indicated concentrations of freshly dissolved compounds. The
medium containing compounds was discarded, and fresh medium
containing 20 mL MTT (5 mgmLÀ1) was added to each well and in-
cubated for an additional 4 h. The medium was removed. After
adding 200 mL DMSO to each well, the optical densities at
l 570 nm were determined. Cytotoxicity data are expressed as IC50
values obtained from the fit to a four-parameter sigmoid. All R
values were >0.98, and standard errors of the three replicates
were <20%. Experiments on 15 cell lines were performed by
CrownBio Inc. (Santa Clara, CA, USA) according to a similar MTS
assay (Promega), according to the manufacturer’s protocol. Briefly,
each cell line (obtained from ATCC) was plated on a 96-well plate
and grown. An-Hq2 was incubated with each cell line for 3 days,
and cell viability was monitored at l 490 nm.
Experimental Section
Most commercially supplied reagents were used without further
purification. Chemicals and enzymes were obtained from Sigma–
Aldrich or Thermo-Fisher unless otherwise noted. All synthesized
molecules were >95% pure for intermediates and 98% pure for
cytotoxicity measurements as analyzed by HPLC. 1H and 13C NMR
spectra are available in the Supporting Information. 1H and
13C NMR spectra were recorded on a Bruker AMX 400 MHz spec-
trometer. Chemical shifts (d) are reported in ppm, with 13C and re-
1
sidual H signals from deuterated solvents as references. High-reso-
lution mass spectra (ESI) were recorded on a Micromass Q-TOF 2
(Waters). Analytical thin-layer chromatography (TLC) was per-
formed on silica gel 60 GF254 (Merck). Column chromatography was
ChemMedChem 2011, 6, 869 – 875
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
873

E. J. Merino et al.
MED
2-{2-[Ethyl(phenyl)amino]ethoxy}phenol (An-Cat). A mixture of
An-Cl (500 mg, 2.7 mmol), K2CO3 (750 mg, 5.4 mmol), KI (90 mg,
0.54 mmol) in DMF (50 mL) was bubbled with argon for 10 min.
Catechol (600 mg, 5.4 mmol) was added to the mixture under
argon. The reaction mixture was heated and stirred at 808C for 6 h.
After cooling to room temperature, the solvent was evaporated,
and the residue was quenched with 100 mL H2O. The mixture was
adjusted to pH~7 with dilute HCl. The mixture was then extracted
with EtOAc, and the combined organic layer was washed with H2O
and then brine, and dried over Na2SO4. The solvent was evaporat-
ed, and the residue was purified by silica gel column chromatogra-
phy with CH2Cl2 to provide An-Cat (400 mg, 1.55 mmol, 57%) as a
Synthesis
N-(2-Chloroethyl)-N-ethylaniline (An-Cl). SOCl2 (8.8 mL, 121 mmol)
was added dropwise over 20 min to a solution of 2-(N-ethyl-N-phe-
nylamino)ethanol (10 g, 60 mmol) and CH2Cl2 (150 mL). The mix-
ture was held at reflux during addition, and the solution turned
from colorless to yellow and then to brown. After the addition was
complete, the mixture was heated at reflux for one hour. After
cooling to room temperature, the mixture was quenched carefully
with cold, saturated aqueous K2CO3 (150 mL). The mixture was
then extracted with CH2Cl2, and the combined organic layer was
washed with saturated aqueous K2CO3, washed with brine, and
dried over Na2SO4. The solvent was evaporated, and the residue
was purified by silica gel column chromatography with hexane/
CH2Cl2 (4:1) to provide An-Cl (5.54 g, 30 mmol, 50%) as a yellow
1
white solid. H NMR ([D6]DMSO, 400 MHz): d=8.86 (s, 1H), 7.15 (t,
J=8 Hz, 2H), 6.90 (d, J=8 Hz, 1H), 6.78 (m, 5H), 6.57 (t, J=7.2 Hz,
1H), 4.07 (t, J=6.0 Hz, 2H), 3.68 (t, J=6.0 Hz, 2H), 3.45 (q, J=
6.8 Hz, 2H), 1.10 ppm (t, J=6.8 Hz, 3H); 13C NMR (CDCl3, 100 MHz):
d=148.00, 146.35, 145.81, 129.47, 122.24, 120.06, 117.24, 115.04,
112.86, 67.44, 49.79, 45.65, 12.19 ppm; IR (KBr): n˜ =3462, 2926,
2869, 1594, 1500, 1258, 740, 695 cmÀ1; HRMS (ESI+) m/z calcd for
C16H20NO2 [M+H]+: 258.1494, found: 258.1465; calcd for
C16H19NO2Na [M+Na]+: 280.1313, found: 280.1345.
1
oil. H NMR (CDCl3, 400 MHz): d=7.28 (m, 2H), 6.74 (m, 3H), 3.65
(m, 4H), 3.46 (q, J=7.2 Hz, 2H), 1.22 ppm (t, J=7.2 Hz, 3H);
13C NMR (CDCl3, 100 MHz): d=147.02, 129.56, 116.61, 111.90, 52.45,
45.44, 40.56, 12.57 ppm; IR (KBr): n˜ =2971, 2891, 1599, 1505, 1353,
1269, 748, 693 cmÀ1; HRMS (ESI+) m/z calcd for C10H15ClN [M+H]+:
184.0893, found: 184.0836.
4-{2-[Ethyl(phenyl)amino]ethoxy}phenol (An-Hq). A mixture of
An-Cl (500 mg, 2.7 mmol), K2CO3 (750 mg, 5.4 mmol), and KI
(90 mg, 0.54 mmol) in DMF (50 mL) was bubbled with argon for
10 min. Hydroquinone (600 mg, 5.4 mmol) was added to the mix-
ture under argon. The reaction mixture was heated and stirred at
808C for 6 h. After cooling to room temperature, the solvent was
evaporated, and the residue was quenched with 100 mL H2O. The
mixture was adjusted to pH~7 with dilute HCl. The mixture was
then extracted with EtOAc, the combined organic layer was
washed with H2O and then brine, and was dried over Na2SO4. The
solvent was evaporated, and the residue was purified by silica gel
column chromatography with CH2Cl2 to provide An-Hq (400 mg,
1.55 mmol, 57%) as a brown oil. 1H NMR ([D6]DMSO, 400 MHz):
d=8.92 (s, 1H), 7.14 (m, 2H), 6.75–6.64 (m, 6H), 6.57 (m, 1H), 3.98
(t, J=5.8 Hz, 2H), 3.62 (t, J=5.8 Hz, 2H), 3.41 (q, J=7 Hz, 2H),
1.09 ppm (t, J=7 Hz, 3H); 13C NMR (CDCl3, 100 MHz): d=152.85,
149.74, 147.71, 129.49, 116.24, 115.73, 112.13, 66.20, 49.80, 45.69,
12.31 ppm; IR (KBr): n˜ =3400, 2970, 1598, 1508, 1230, 1035, 827,
749, 694 cmÀ1; HRMS (ESI+) m/z calcd for C16H20NO2 [M+H]+:
258.1494, found: 258.1436; HRMS (ESI+) m/z calcd for
C16H19NO2Na [M+Na]+: 280.1313, found: 280.1250.
N-Ethyl-N-[2-(4-methoxyphenoxy)ethyl]aniline (An-Hq-CH3). MeI
(1 mL, 16 mmol) was added to a mixture of An-Hq (78 mg,
0.3 mmol) and K2CO3 (95 mg, 0.69 mmol) in acetone (50 mL). The
reaction mixture was heated at reflux for 48 h. After cooling to
room temperature, the solvent was evaporated, and the residue
was quenched with 50 mL H2O. The mixture was then extracted
with EtOAc, and the combined organic layer was washed with H2O
then brine, and dried over Na2SO4. The solvent was evaporated,
and the residue was purified by silica gel column chromatography
with hexane/CH2Cl2 (2:1) to provide An-Hq-CH3 (49 mg, 0.18 mmol,
1
60%) as a yellow oil. H NMR (CDCl3, 400 MHz): d=7.23 (m, 2H),
6.83 (m, 4H), 6.73 (d, J=8.4 Hz, 2H), 6.68 (t, J=7 Hz, 1H), 4.07 (t,
J=6.4 Hz, 2H), 3.77 (s, 3H), 3.70 (t, J=6.4 Hz, 2H), 3.47 (q, J=7 Hz,
2H), 1.20 ppm (t, J=7 Hz, 3H); 13C NMR (CDCl3, 100 MHz): d=
153.93, 152.91, 147.62, 129.38, 115.95, 115.40, 114.67, 111.80, 66.01,
55.74, 49.76, 45.59, 12.30 ppm; HRMS (ESI+) m/z calcd for
C17H22NO2 [M+H]+: 272.1645, found: 272.1646.
3-{2-[Ethyl(phenyl)amino]ethoxy}phenol (An-Rs). A mixture of
An-Cl (500 mg, 2.7 mmol), K2CO3 (750 mg, 5.4 mmol), KI (90 mg,
0.54 mmol) in DMF (50 mL) was bubbled with argon for 10 min.
Resorcinol (600 mg, 5.4 mmol) was added to the mixture under
argon. The reaction mixture was heated and stirred at 808C for 6 h.
After cooling to room temperature, the solvent was evaporated,
and the residue was quenched with 100 mL H2O. The mixture was
adjusted to pH~7 with dilute HCl. The mixture was then extracted
with EtOAc, and the combined organic layer was washed with H2O
and then brine, and dried over Na2SO4. The solvent was evaporat-
ed, and the residue was purified by silica gel column chromatogra-
phy with CH2Cl2 to provide An-Rs (380 mg, 1.48 mmol, 55%) as a
light-yellow oil. 1H NMR ([D6]DMSO, 400 MHz): d=9.38 (s, 1H), 7.15
(t, J=8 Hz, 2H), 7.04 (t, J=8 Hz, 1H), 6.70 (d, J=8 Hz, 2H), 6.57 (t,
J=7.2 Hz, 1H), 6.32 (m, 3H), 4.02 (t, J=5.8 Hz, 2H), 3.64 (t, J=
5.8 Hz, 2H), 3.44 (q, J=7 Hz, 2H), 1.10 ppm (t, J=7 Hz, 3H);
13C NMR (CDCl3, 100 MHz): d=155.14, 151.95, 142.80, 125.44,
124.66, 111.56, 107.52, 103.46, 102.10, 97.43, 60.60, 44.77, 40.83,
4,4’-[2,2’-(Phenylazanediyl)bis(ethane-2,1-diyl)bis(oxy)]diphenol
(An-Hq2). A mixture of An-Cl2 (500 mg, 2.29 mmol), K2CO3 (634 mg,
4.59 mmol), and KI (76 mg, 0.46 mmol) in DMF (50 mL) was bub-
bled with argon for 10 min. Hydroquinone (1.32 g, 12 mmol) was
added to the mixture under argon. The reaction mixture was
heated and stirred at 808C for 6 h. After cooling to room tempera-
ture, the solvent was evaporated, and the residue was quenched
with 100 mL H2O. The mixture was adjusted to pH~7 with dilute
HCl. The mixture was then extracted with EtOAc, and the com-
bined organic layer was washed with H2O then brine, and dried
over Na2SO4. The solvent was evaporated, and the residue was pu-
rified by silica gel column chromatography with CH2Cl2/CH3OH
(15:1) to provide An-Hq2 (150 mg, 0.41 mmol, 18%) as a yellow oil.
1H NMR ([D6]DMSO, 400 MHz): d=8.91 (s, 2H), 7.16 (t, J=7.8 Hz,
2H), 6.78–6.72 (m, 6H), 6.66–6.60 (m, 5H), 4.02 (t, J=5.8 Hz, 2H),
3.75 ppm (t, J=5.8 Hz, 2H); 13C NMR (CDCl3, 100 MHz): d=152.87,
149.60, 147.31, 129.50, 116.55, 116.08, 115.56, 111.79, 65.91,
50.99 ppm; IR (KBr): n˜ =3366, 2926, 1598, 1508, 1229, 826,
750 cmÀ1; HRMS (ESI+) m/z calcd for C22H24NO4 [M+H]+: 366.1700,
found: 366.1701.
7.34 ppm; IR (KBr): n˜ =3390, 2971, 1598, 1505, 1149, 748, 687 cmÀ1
;
HRMS (ESI+) m/z calcd for C16H20NO2 [M+H]+: 258.1494, found:
258.1534.
874
www.chemmedchem.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2011, 6, 869 – 875

Oxidatively Activated DNA-Modifying Agents
[18] J. S. Bair, R. Palchaudhuri, P. J. Hergenrother, J. Am. Chem. Soc. 2010,
132, 5469–5478.
Acknowledgements
[19] M. Wangpaichitr, C. Wu, M. You, J. C. Maher, V. Dinh, L. G. Feun, N.
Savaraj, Cancer 2009, 1, 23–38.
[20] D. Trachootham, J. Alexandre, P. Huang, Nat. Rev. Drug Discovery 2009,
8, 579–591.
We thank the University of Cincinnati for startup funding. We
also thank the University of Cincinnati Faculty Research Grants
Program for funding E.J.M. and the University of Cincinnati Post-
doctoral Fellow Research Program for funding G.L.
[21] M. Doering, L. A. Ba, N. Lilienthal, C. Nicco, C. Scherer, M. Abbas, A. A.
Zada, R. Coriat, T. Burkholz, L. Wessjohann, M. Diederich, F. Batteux, M.
Herling, C. Jacob, J. Med. Chem. 2010, 53, 6954–6963.
[22] J. M. Laparra, D. Vꢂlez, R. Barberꢃ, R. Farrꢂ, R. Montoro, Toxicol. In Vitro
2008, 22, 444–449.
[23] A. V. Patterson, D. M. Ferry, S. J. Edmunds, Y. Gu, R. S. Singleton, K. Patel,
S. M. Pullen, K. O. Hicks, S. P. Syddall, G. J. Atwell, S. Yang, W. A. Denny,
W. R. Wilson, Clin. Cancer Res. 2007, 13, 3922–3932.
[24] J. G. Muller, V. Duarte, R. P. Hickerson, C. J. Burrows, Nucleic Acids Res.
1998, 26, 2247–2249.
Keywords: chemotherapy
· DNA modification · nitrogen
mustard · oxidative stress · reactive oxygen species
[1] B. A. Chabner, T. G. Roberts, Nat. Rev. Cancer 2005, 5, 65–72.
[2] A. J. Berdis, Chem. Rev. 2009, 109, 2862–2879.
[3] Z. H. Siddik, Oncogene 2003, 22, 7265–7279.
[4] L. P. Bignold, Anticancer Res. 2006, 26, 1327–1336.
[5] J. L. Guerriero, D. Ditsworth, Y. Fan, F. Zhao, H. C. Crawford, W. X. Zong,
Cancer Res. 2008, 68, 9595–9600.
[6] E. Kim, P. T. Rye, J. M. Essigmann, R. G. Croy, J. Inorg. Biochem. 2009, 103,
256–261.
[7] K. S. Lovejoy, S. J. Lippard, Dalton Trans. 2009, 48, 10651–10659.
[8] M. H. David-Cordonnier, W. Laine, T. Gaslonde, S. Michel, F. Tillequin, M.
Koch, S. Lꢂonce, A. Pierrꢂ, C. Bailly, Curr. Med. Chem. Anticancer Agents.
2004, 4, 83–92.
[9] C. A. Puckett, J. K. Barton, Bioorg. Med. Chem. 2010, 18, 3594–3599.
[10] S. Elchuri, T. D. Oberley, W. Qi, R. S. Eisenstein, L. Jackson-Roberts, H.
VanRemmen, C. J. Epstein, T. T. Huang, Oncogene 2005, 24, 367–380.
[11] B. Kumar, S. Koul, L. Khandrika, R. B. Meacham, H. K. Koul, Cancer Res.
2008, 68, 1777–1785.
[25] M. Valko, D. Leibfritz, J. Moncol, M. T. Cronin, M. Mazur, J. Telser, Int. J.
Chem. Kinet. Int. J. Biochem Cell. Biol. 2007, 39, 44–84.
[26] S. Frelon, T. Douki, T. Favier, J. Cadet, Chem. Res. Toxicol. 2003, 16, 191–
197.
[27] J. L. Bolton, M. A. Trush, T. M. Penning, G. Dryhurst, T. J. Monks, Chem.
Res. Toxicol. 2000, 13, 135–160.
[28] S. Steenken, P. Neta, J. Phys. Chem. 1982, 86, 3661–3667.
[29] M. Bai, J. Huang, X. Zheng, Z. Song, M. Tang, W. Mao, L. Yuan, J. Wu, X.
Weng, X. Zhou, J. Am. Chem. Soc. 2010, 132, 15321–15327.
[30] S. M. Wilhelm, L. Adnane, P. Newell, A. Villanueva, J. M. Llovet, M. Lynch,
Mol. Cancer Ther. 2008, 7, 3129–3140.
[31] M. Pljesa-Ercegovac, J. Mimic-Oka, D. Dragicevic, A. Savic-Radojevic, M.
Opacic, S. Pljesa, R. Radosavljevic, T. Simic, Urol. Oncol. 2008, 14, 175–
181.
[12] J. Cadet, T. Douki, J. L. Ravanat, Free Radical Biol. Med. 2010, 49, 9–21.
[13] K. Block, Y. Gorin, P. Hoover, P. Williams, T. Chelmicki, R. A. Clark, T.
Yoneda, H. E. Abboud, J. Biol. Chem. 2007, 282, 8019–8026.
[14] S. Orrenius, Drug Metab. Rev. 2007, 39, 443–455.
[15] E. J. Merino, A. K. Boal, J. K. Barton, Curr. Opin. Chem. Biol. 2008, 12,
229–237.
[32] W. M. Frederiks, K. S. Bosch, K. A. Hoeben, J. van Marle, S. Langbein,
Acta Histochem. 2010, 112, 364–371.
Received: January 14, 2011
Revised: January 24, 2011
[16] M. S. King, M. S. Sharpley, J. Hirst, Biochemistry 2009, 48, 2053–2062.
[17] M. J. Solivio, T. J. Joy, L. Sallans, E. J. Merino, J. Inorg. Biochem. 2010, 104,
1000–1005.
Published online on March 4, 2011
ChemMedChem 2011, 6, 869 – 875
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
875