mediated mechanism by b-phenylethyl isothiocyanate
Dunyaporn Trachootham,1,2,3Yan Zhou,1Hui Zhang,1,2Yusuke Demizu,1Zhao Chen,1,2Helene Pelicano,1
Paul J. Chiao,2,4Geetha Achanta,1Ralph B. Arlinghaus,1,2Jinsong Liu,2,5and Peng Huang1,2,*
1Department of Molecular Pathology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas 77030
2The University of Texas Graduate School of Biomedical Sciences at Houston, Houston, Texas 77030
3Faculty of Dentistry, Thammasat University, Rangsit Campus, Pathum-thani, Thailand 12121
4Department of Surgical Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas 77030
5Department of Pathology, The University of Texas M.D. Anderson Cancer Center, Houston, Texas 77030
Reactive oxygen species (ROS) stimulate cell proliferation and induce genetic instability, and their increase in cancer cells is
cer cells using b-phenylethyl isothiocyanate (PEITC). Oncogenic transformation of ovarian epithelial cells with H-RasV12
or expression of Bcr-Abl in hematopoietic cells causes elevated ROS generation and renders the malignant cells highly
sensitive to PEITC,whicheffectively disablesthe glutathione antioxidantsystem andcauses severe ROS accumulation pref-
erentially in the transformed cells due to their active ROS output. Excessive ROS causes oxidative mitochondrial damage,
inactivation of redox-sensitive molecules, and massive cell death. In vivo, PEITC exhibits therapeutic activity and prolongs
An ideal anticancer agent should be toxic to malignant cells with
minimum toxicity in normal cells. However, currently there are
limited numbers of such agents available for clinical use. Thus,
development of novel selective drugs is an important and chal-
lenging task, and understanding the biological differences
between normal and cancer cells is essential for achieving this
goal. A primeexample of successful design ofselective antican-
cer drugs is the development of Gleevec, which targets the
oncogenic tyrosine kinase BCR-ABL responsible for human
chronic myeloid leukemia (Ren, 2005). While identifying specific
oncogenes responsible for cancer development and designing
targeting agents represent a very important research area,
such tasks have proven to be challenging, due to multiple ge-
netic and epigenetic alterations in cancer cells (Couzin, 2002;
Frantz, 2005). Mutations or overexpression of the target mole-
cules often leadto drug resistance andimpose further challenge
in treatment of cancer using such target-specific agents. This
situation has prompted the consideration of alternative
approaches. Instead of targeting specific oncogenic molecules,
it may be possible to exploit the biochemical alterations in can-
cer cells as a basis for developing selective therapeutic agents.
If a biochemical alteration is common in cancer cells, targeting
such change would have broad therapeutic applications.
One common biochemical change in cancer cells is the in-
crease in reactive oxygen species (ROS) generation. Emerging
evidences suggest that most cancer cells are under oxidative
stress associated with increased metabolic activity and produc-
tion of ROS (Szatrowski and Nathan, 1991), although in some
cases the observed ROS increase could be due to sample han-
dling and analytical artifacts (Swartz and Gutierrez, 1977). The
ROS increase is thought to play an important role in maintaining
cancer phenotype due to their stimulating effects on cell growth
and proliferation (Hu et al., 2005), genetic instability (Radisky
et al., 2005), and senescence evasion (Chen et al., 2005). Owing
to their cancer-promoting effect, increased ROS in cancer cells
isoftenconsideredasanadversefactor. However,highlevels of
ROS can also cause cellular damage, depending on the levels
and duration of ROS stress (Pelicano et al., 2004). These
dose- and time-dependent effects may provide an opportunity
S I G N I F I C A N C E
Most cancer cells exhibit elevated ROS generation associated with active metabolism and oncogenic stimulation. Although this phe-
nomenon has long been recognized, its therapeutic implications remain unclear. The current study demonstrated that it is possible to
takeadvantageof thisROSstressincanceranduseROS-mediatedmechanismsto preferentially killthemalignantcells.Thistherapeu-
tic strategy can be achieved both in vitro and in vivo, using a natural compound, PEITC. Thus, ROS stress associated with oncogenic
transformation may serve as a biochemical basis for developing cancer therapeutics. Furthermore, because PEITC has low toxicity
A R T I C L E
CANCER CELL 10, 241–252, SEPTEMBER 2006 ª2006 ELSEVIER INC. DOI 10.1016/j.ccr.2006.08.009 241
to exploit the cell killing potential of ROS by using exogenous
level, or the threshold that triggers cell death. For instance, hu-
man leukemia cells with intrinsic oxidative stress are highly sen-
sitive to ROS stress induced by 2-methoxyestradiol and arsenic
trioxide (Huang et al., 2000; Zhou et al., 2003). Promoting ROS
generation in mitochondria was shown to effectively kill cancer
cells (Pelicano et al., 2003). Although induction of ROS genera-
tion by anticancer drugs including doxorubicin, arsenic trioxide,
and taxol has been observed, it is still unclear whether the intrin-
sic oxidative stress in cancer cells could provide a selective
response to ROS-generating agents.
Because increased ROS generation is common in cancer
cells with active metabolism under the influence of oncogenic
signalsincluding Ras,Bcr-Abl,andc-Myc (Iraniet al.,1997;Sat-
tler etal., 2000; Vafa et al.,2002),wepostulated thatthe intrinsic
ROS stress associated with oncogenic transformation would
make the cells highly dependent on their antioxidant systems
to counteract the damaging effect of ROS and to maintain redox
balance in a dynamic state (increased ROS generation and ac-
tive ROS scavenging). This situation would render them highly
vulnerable to further oxidative insults by exogenous agents.
On the other hand, normal cells may better tolerate such an in-
tervention owing to their low basal ROS output and normal met-
abolic regulation. As such, the differences in redox states be-
tween normal and cancer cells may provide a biological basis
for selective killing of malignant cells, using agents that cause
further ROS stress.
b-phenylethyl isothiocyanate (PEITC) is a natural compound
found in consumable cruciferous vegetables with chemopre-
ventive activity (Yu et al., 1998). Recent studies suggest that,
in contrast to the view of PEITC as an antioxidant, this com-
pound could increase ROS generation and induce apoptosis
in cancer cell lines (Yu et al., 1998; Zhang et al., 2003; Wu
et al., 2005). However, whether PEITC has a selective effect
against cancer cells andwhether the intrinsic redox statuscould
affect the fate of cellular response to ROS-modulating agents
remain unclear. The goal of this study was to test the concept
that increased ROS generation associated with oncogenic
transformation may serve as a biochemical basis to selectively
kill cancer cells using redox-modulating agents.
Oncogenic transformation by H-Ras or Bcr-Abl leads
to increased ROS generation
To test the hypothesis that oncogenic transformation causes in-
creased ROS generation and renders the transformed cells vul-
nerable to further ROS stress, we first used in vitro systems to
evaluate the effects of two oncogenic signals, Ras and Bcr-
Abl, on ROS generation and cellular sensitivity to exogenous
ROS stress under isogenic conditions. A previously immortal-
ized normal ovarian epithelial cell line (T72 cells harboring SV-
40 T/t and hTERT) was further transfected with H-RasV12(Liu
et al., 2004). The stably transfected cells (T72Ras) exhibited ma-
lignant behaviors capable of forming foci in soft agar and grow-
ing tumor in vivo. As illustrated in Figure 1A, T72Ras cells readily
formed tumor in all 20 mice tested and exhibited pathological
morphology resembling human ovarian clear cell carcinoma.
Western blot analysis showed that T72Ras cells expressed in-
creased H-RAS, which was functionally active, as evidenced
by constitutive RAS-GTP activity (Figure 1B). Importantly, com-
pared to the parental T72 cells, the Ras-transformed cells ex-
hibited a significant increase (200%) in basal ROS content, as
quantified by flow cytometry using CM-H2DCF-DA as a fluores-
cent probe (Figure 1C, fluorescence in log-scale). We also used
hydroethidine (HEt), a relatively specific probe for superoxide
(O22) (Zhao et al., 2003), to compare the basal levels of O22in
the two cell lines. T72Ras cells exhibited significantly higher
HEt fluorescence (11 units) than T72 cells (6 units; p < 0.001),
suggesting an increase in basal O22in the Ras-transformed
cells (Figure 1C, right panel). Further analyses revealed that
Figure 1. Oncogenic transformation by H-Ras or Bcr-Abl causes increased
A: T72Ras cells derived from T72 cells by Ras transformation readily formed
tumors in all 20 mice tested. Tumor tissue stained with hematoxylin-eosin ex-
hibited pathological morphology resembling human ovarian clear cell car-
B: Basal H-RAS protein expression and RAS activation in T72Ras cells, mea-
sured by Western blot and immunoprecipitation assays.
C: Increase of ROS in T72Ras cells detected by flow cytometry using DCF-DA
(left panel) and HEt (right panel). Each histogram is representative of three
experiments (p < 0.001, T72 versus T72Ras cells).
D: Comparison of total cellular GSH in T72 and T72Ras cells (mean 6 SD of
three experiments; p < 0.05).
E: Basal protein expression of catalase, glutathione peroxidase 1 (GPX1),
and g glutamyl-cysteine synthetase enzyme (GSH1) in T72 and T72Ras cells.
F: Time-dependent induction of BCR-ABL expression by doxycycline
(1 mg/ml) in TonB210 cells.
G: Increase of ROS in TonB210 cells after induction of BCR-ABL expression.
A R T I C L E
CANCER CELL SEPTEMBER 2006
the elevated ROS was not due to a decrease in antioxidant ca-
pacity, since there was a significant increase in cellular glutathi-
one (GSH) and an upregulation of catalase (Figures 1D and 1E).
There was no change in expression of glutathione peroxidase
(GPX-1) and g-glutamyl-cysteine synthetase (GSH1). These
data together suggest that transformation of ovarian epithelial
satory increase in cellular antioxidant activity.
cycline-inducible Bcr-Abl expression vector (TonB210 cells;
Klucher et al., 1998) was then used to test if induced expression
of BCR-ABL would also cause increased ROS generation. As
illustrated in Figure 1F, addition of doxycycline induced BCR-
ABL expression and a time-dependent increase in cellular
ROS detected by DCF-DA (Figure 1G). In control experiments,
an increase of DCF-DA fluorescence in the parental cells (data
not shown), suggesting that the ROS increase in TonB210 cells
was due to BCR-ABL expression.
Increased ROS generation renders the oncogenically
transformed cells highly sensitive to PEITC
Evaluation of cellular response to ROS-modulating agents re-
vealed a striking preferential activity of PEITC against the trans-
formed cells. Treatment of T72Ras cells with 10 mM PEITC
caused a substantial increase in DCF-DA-reactive ROS in
a time-dependent manner and reached a 16-fold increase at
5 hr (Figure 2A). In contrast, the parental T72 cells were less
sensitive to PEITC, showing a moderate ROS elevation without
further increase as the incubation prolonged (Figure 2B), sug-
gesting that the nonmalignant cells could better cope with
PEITC-induced ROS accumulation, likely due to their low basal
ROS output (Figure 1C). Quantitative analysis further confirmed
the significant difference between T72Ras and T72 cells in ROS
induction by 5 mM PEITC for various times (Figure 2C). This dif-
ference also held true when both cell lines were exposed to var-
ious concentrations of PEITC (Figure S1 in the Supplemental
Data available with this article online). Interestingly, there was
no significant change in HEt fluorescence after treatment with
sistently exhibited a higher basal O22(Figure S2A), suggesting
that ROS induced by PEITC were mainly DCF-DA-reactive spe-
cies such ashydrogen peroxide (H2O2) andnitric oxide (NO),but
not O22. Using 4-amino-5-methylamino-20,70-difluorofluores-
cein (DAF-FM), a relatively specific probe for NO (Balcerczyk
et al., 2005), we showed that treatment of T72Ras cells with
was reversed by the antioxidant N-acetyl-L-cysteine (NAC) but
not by the H2O2-scavenging enzyme catalase (Figure S2B). In
contrast, the PEITC-induced increase in DCF-DA fluorescence
could be readily reversed by either NAC or catalase.
The significant difference between T72 and T72Ras cells in
their ROS accumulation in response to PEITC prompted us to
compare the cytotoxic effect of this compound in the two cell
lines. Flowcytometric analysisshowedthatPEITCeffectively in-
duced cell death in T72Ras cells in a time-dependent manner,
trast, T72 cells were significantly less sensitive to PEITC at all
time points tested (Figure 2D). The preferential killing of Ras-
transformed cells by PEITC was further demonstrated with
multiple drug concentrations (Figure 2E) and various times
(Figure 2F). Similarly, the highly aggressive mouse leukemia
cells (DA1-3b) constitutively expressing BCR-ABL were more
sensitive to PEITC than the parental cells (DA1, myeloid progen-
itor cells). As shown in Figure 2G, incubation of DA1-3b cells
with 5 mM PEITC resulted in 50% cell death but only caused
10% cell death in DA1. This differential sensitivity was also ob-
served at 10 mM of PEITC (Figure 2H) and with various incuba-
tion times (Figure S3).
Because the nonmalignant T72 cells could not form colo-
nies, we adapted the MTT assay to compare PEITC’s effect
Figure 2. Preferential induction of ROS accumulation and cell death by
PEITC in oncogenically transformed cells
A: Time-dependent ROS accumulation induced by 10 mM PEITC in T72Ras
cells. ROS were measured by flow cytometry using DCF-DA.
B: Effect of PEITC (10 mM) on cellular ROS content in T72 cells.
C: Comparison of ROS in T72Ras cells (black circles) and T72 cells (white cir-
cles) induced by 5 mM PEITC (mean 6 SD of three experiments).
D: Time-dependent cell death induced by 10 mM PEITC in T72Ras and T72
cells. Cell death was measured by annexin-V/PI assay. The dot plots are
representative of three experiments.
E: Dose-dependent killing by PEITC (5 hr) in T72Ras and T72 cells (mean 6 SD
of three experiments; *p < 0.05).
F: Time-dependent killing of T72Ras and T72 cells by 10 mM PEITC (mean 6 SD
from three experiments; *p < 0.05).
G: Comparison of PEITC-induced cell death (5 mM, 24 hr) in Bcr-Abl trans-
formed cells (DA1-3b) and the parental DA1 cells.
H: Dose-dependent killing by PEITC (24 hr) in DA1-3b cells and DA1 cells
(mean 6 SD of three experiments; *p < 0.05).
A R T I C L E
CANCER CELL SEPTEMBER 2006243
on long-term cell proliferation in T72 and T72Ras cells. As
shown in Figure 3A, this compound exhibited greater inhibition
on T72Ras cells during 10 day incubation. We then compared
this selectivity with cisplatin (CDDP), a common drug for ovarian
cancer treatment. Both the long-term cell proliferation assay (10
day MTT; Figure 3B) and acute cell death analysis by annexin-V/
PI staining (Figure 3C) showed that CDDP exerted a similar ef-
fect on T72 and T72Ras cells. The ratio of IC50values for T72
cells and T72Ras cells was 1.0 for CDDP and 3.9 for PEITC (Fig-
ure 3D), suggesting that PEITC has a better selectivity. Since
CDDP was shown to increase ROS in certain cells, we com-
pared the ROS levels before and after CDDP exposure. As
nochange inROSafter 4 hrincubation andonly aslight increase
that the cytotoxic effect of CDDP in these cells was unlikely
triggered by ROS.
ROS-mediated damage is a critical mechanism
for PEITC-induced cell death
ticancer activity of PEITC. To further test if PEITC could pre-
ferentially induce ROS-mediated lipid peroxidation in Ras-trans-
formedcells,wetreatedT72andT72Rascells with 10mMPEITC
forvarioustimesandusednonyl acridine orange(NAO) todetect
oxidation of cardiolipin, a mitochondrial membrane lipid compo-
nent (Nomuraet al.,2000).As showninFigure 4A,PEITC caused
88% of T72Ras cells lost their cardiolipin within 4 hr and 8 hr, re-
spectively. Addition of antioxidant NAC almost completely re-
versed the PEITC-induced loss of NAO staining (Figure 4A, right
panel), suggesting that the decrease of NAO fluorescence was
due to oxidative damage. In contrast, only a small percent of
T72 cells showed a loss of NAO signal (Figure 4A).
of mitochondrial membrane damage by PEITC. The functional
integrity of mitochondrial membranes was assessed by flow cy-
tometry after cells were labeled with rhodamine-123 (Rho-123),
a potential-sensitive dye that accumulates in intact mitochon-
indicates a loss of transmembrane potential (Sureda et al.,
1997). As shown in Figure 4B, PEITC caused a time-dependent
loss of Rho-123 signal in T72Ras cells starting at 4 hr, and 64%
of cells lost their mitochondrial integrity by 8 hr. In contrast, only
a small portion of the nontransformed cells exhibited a loss of
Rho-123 signal. Interestingly, when the time courses for cardio-
lipin oxidation (Figure 4A) and loss of transmembrane potential
(Figure 4B) in T72Ras cells were compared, it appeared that
membrane oxidation occurred first, followed by the loss of
membrane integrity with a delay of approximately 1–2 hr.
The relationship between ROS accumulation and PEITC-in-
was a strong correlation between these two parameters (r =
0.9727; p < 0.001). To verify the cause-effect relationship be-
tween ROS increase and cell death, we tested the effect of
H2O2-scavenging enzymecatalase on PEITC-induced celldeath
in T72Ras cells. Since H2O2is relatively stable and can diffuse
Figure 3. Anticancer selectivity of PEITC is superior to that of cisplatin in vitro
A: Inhibition of cell proliferation by PEITC in T72 and T27Ras cells. Cell growth
inhibition was measured by long-term MTT assay (mean 6 SD of three exper-
iments; *p < 0.05).
B: Dose-dependent inhibition of cell proliferation by cisplatin in T72Ras and
T72 cells, measured by long-term MTT assay (mean 6 SD of three experi-
C: Cytotoxicity of cisplatin (10 mM) in T72Ras and T72 cells measured by
annexin-V/PI assay (mean 6 SD of three experiments).
D: Comparison of in vitro selectivity of PEITC and cisplatin (CDDP). The con-
centrations required to inhibit 50% cell proliferation (IC50) were determined
using the 10 day MTT assay, and the selectivity index (S.I.) was calculated as
the IC50ratio of T72/T72Ras cells.
E: Effect of CDDP (10 mM) on cellular ROS in T72Ras and T72 cells.
A R T I C L E
CANCER CELL SEPTEMBER 2006
membranes. Indeed, exogenous catalase decreased the basal
ROS in T72Ras cells (Figure 4D) and significantly offset the
ROS increase induced by PEITC (Figure 4E). Quantitative analy-
sis showed that catalase significantly reduced PEITC-induced
ROS accumulation in a dose-dependent manner (Figure 4F) and
proportionally abrogated PEITC-induced cell death (Figure 4G),
suggesting that H2O2was critical in mediating cytotoxicity.
We further tested if PEITC could cause cytochrome c release
from mitochondria and activate caspases. PEITC (5–10 mM)
caused a rapid release of cytochrome c from mitochondria to
cytosol within 5 hr in T72Ras cells, but this did not result in a sig-
nificant activation of caspase-3 (Figure S4). This is consistent
with oxidation of caspase-3 by ROS, a process known to abro-
gate caspase activation (Hampton et al., 2002). Thus, it ap-
peared that PEITC caused a severe damage to mitochondria
membranes and cytochrome c release, but the increased ROS
prevented caspase activation, leading to necrotic cell death.
This was consistent with the cell death profiles revealed by an-
nexin-V/PI staining showing a large population of necrotic cells
in the up-left quarter of the flow cytometry plots (Figure 2D;
T72Ras cells). Interestingly, cells overexpressing Bcl-2, a mole-
cule with antioxidant function capable of protecting the mito-
chondrial integrity (Hockenbery et al., 1993), were less sensitive
to PEITC (Figure S5), suggesting the important role of mito-
chondrial damage in PEITC-induced cell death.
PEITC causes severe ROS accumulation in
Ras-transformed cells by disabling the GSH
Since PEITC exerted cytotoxicity by causing excessive ROS
accumulation, we then investigated the mechanisms by which
PEITC caused ROS increase. Based on the important role of
Figure4. Selectivekilling ofRas-transformed cells
by PEITC through ROS-mediated damage
A: Comparison of oxidative damage to cardioli-
pin by 10 mM PEITC in T72 and T72Ras cells mea-
suredbyflow cytometry using NAO.M1indicates
cardiolipin oxidation. The right panel shows that
NAC (1 mM) reversed the loss of NAO signal in
T72Ras cells treated with PEITC (10 mM, 6 hr).
B: Preferential induction of mitochondrial trans-
membrane potential loss by 10 mM PEITC in
T72Ras cells detected by flow cytometry using
Rho-123. M1 indicates subpopulation of cells
that lost transmembrane potential. The right
panel shows the effect of the uncoupling agent
CCCP (100 mM) as a positive control for mem-
brane potential collapse.
C: Correlation between ROS increase and PEITC-
induced cell death in T72Ras cells. ROS and cell
death were measured by DCF-DA and an-
nexin-V/PI assays, respectively. Data points
from three experiments using 10 mM PEITC for
various times were analyzed by linear regression.
D: Effect ofexogenouscatalase(CAT, 2000U/ml,
6 hr) on the basal ROS level in T72Ras cells.
by catalase. T72Ras cells were pretreated with
catalase (2000 U/ml, 1 hr), followed by 5 mM
PEITC for 5 hr.
for 1 hr) and PEITC (5 mM, 5 hr) on ROS levels in
T72Ras cells detected by DCF-DA (mean 6 SD
of three experiments; *p < 0.05).
G: Effect of catalase on PEITC-induced cell
death. T72Ras cells were pretreated with cata-
lase for 1 hr, followed by 5 mM PEITC for 5 hr.
Cell death was detected by annexin-V/PI assay
(mean 6 SD of three experiments; *p < 0.05).
A R T I C L E
CANCER CELL SEPTEMBER 2006245
GSH as a major cellular antioxidant and the observation that
PEITC can conjugate with GSH for export from leukemia cells
(Xu and Thornalley, 2001), we postulated that the active ROS
generation in Ras-transformed cells would render them highly
dependent on GSH to maintain redox balance, and that a deple-
tion ofGSH by PEITC would result in an excessive accumulation
of ROS to a threshold that triggers cell death. To test this possi-
bility,wefirst examinedthe effectofPEITC onGSHcontents.As
shown in Figure 5A, incubation of T72Ras cells with 5 mM PEITC
led to a depletion of cellular GSH by 75% in 1 hr, and almost
complete depletion in 3 hr. In T72 cells treated with 5 mM PEITC,
approximately 40% and 20% GSH remained at 1 and 3 hr, re-
spectively. Analysis of GSH in the medium showed that GSH
was rapidly exported from the cells (Figure 5B). The amount of
GSH in the medium was quantitatively accountable for the
GSH loss in the cells (calculated based on cell number and vol-
ume), suggesting that PEITC-induced GSH export was a key
mechanismforGSHdepletion. Interestingly, catalaseeffectively
abolished PEITC-induced ROS increase (Figure 5C) but did not
prevent GSH depletion (Figure 5D), suggesting that the deple-
tion of GSH was not secondary to ROS stress.
In contrast, pretreatment of T72Ras cells with 3 mM NAC,
a precursor of GSH and a potent antioxidant (Deneke, 2000), ef-
fectively prevented PEITC-induced GSH depletion (Figure 5E)
and ROS accumulation (Figure 5F). Importantly, NAC also sup-
pressed the cytotoxicity of PEITC (Figure 5G), suggesting that
depletion of cellular GSH is an important mechanism responsi-
ble for PEITC-induced ROS accumulation and cell death. To fur-
ther demonstrate the important role of GSH in the survival of
Ras-transformed cells, we used buthionine sulfoximine (BSO),
an inhibitor of GSH synthesis, to test if this agent could also
cause preferential killing of the Ras-transformed cells. We
showed that BSO indeed exhibited greater cytotoxic effect on
T72Ras cells compared to T72 cells, although a longer incuba-
tion (24 hr) and higher concentrations were required to achieve
significant cell killing (Figure S6).
Because GPX is the major enzyme that uses GSH as the sub-
strate to scavenge peroxides, we tested if PEITC might affect
ing GSH. Incubation of purified human GPX with PEITC in vitro
resulted in a concentration-dependent inhibition of GPX activity
(Figure 6A). In the cell-free enzyme assay, 100 and 500 mM
PEITC inhibited GPX by approximately 50% and 90%, respec-
tively. These concentrations could be achieved intracellularly
when cells were incubated with 5–10 mM PEITC. HPLC analysis
showed that incubation of T72Ras cells with 5 and 10 mM PEITC
for 2 hr resulted in an accumulation of intracellular PEITC of 252
and 590 mM, respectively (Figure S7). Analysis of GPX activity in
protein extracts from cells pretreated with 10 mM PEITC ex-
without a loss of GPX protein (Figure 6C). These data suggest
that PEITC not only depletes cellular GSH pool but can also
inhibit GPX enzyme as dual mechanisms to disable the GSH
antioxidant system, leading to severe ROS accumulation in
malignant cells that are active in ROS generation.
Interestingly, we also found that PEITC significantly abolished
the H-RAS-GTP activity, as demonstrated in a pulldown assay
showing an almost complete loss of H-RAS binding with RAF-
1 after cells were incubated with 5 mM PEITC for 3 hr (Figure 6D).
This loss of H-RAS-GTP activity was not due to a decrease of H-
RAS protein (Figure 6D, middle panel). The antioxidant NAC
Figure 5. PEITC causes depletion of cellular GSH and its reversion by NAC
A: Time-dependent depletion of GSH by PEITC (5 mM, 1–3 hr) in T72 and
T72Ras cells. Cellular GSH was measured by spectrophotometric analysis
(mean 6 SD of three experiments).
B: PEITC-induced export of GSH from T72Ras and T72 cells to the culture
medium (mean 6 SD of three experiments).
C: Reversion of the PEITC-induced ROS by catalase (CAT). T72Ras cells were
preincubated with 2000 U/ml CAT for 1 hr followed by 5 mM PEITC (P) for
D: Effect of catalase on PEITC-induced GSH depletion. GSH levels were de-
termined after T72Ras cells were treated with catalase and PEITC as in C.
Error bars, mean 6 SD of four measurements.
E: Effect of NAC on PEITC-induced depletion of GSH. T72Ras cells were pre-
treated with 3 mM NAC for 1 hr, followed by 5 mM PEITC for 1 or 3 hr (mean 6
SD of three experiments).
F: Reversion of PEITC-induced ROS accumulation by NAC. T72Ras cells were
treated with 3 mM NAC for 1 hr, followed by 5 mM PEITC for 5 hr. A represen-
tative histogram (P+N indicates PEITC + NAC) and quantitative bar graph
(mean 6 SD of three experiments) are shown.
G: Reversion of PEITC-induced cell death by NAC. T72Ras cells were treated
with 3 mM NAC for 1 hr, followed by 10 mM PEITC for 5 hr. Representative dot
plots and quantitative bar graph (mean 6 SD of three experiments) are
A R T I C L E
CANCER CELL SEPTEMBER 2006
partially reversed the inactivation of H-RAS-GTP by PEITC (data
not shown). Basedon this observation, wethen tested the effect
of PEITC on another redox-sensitive molecule NF-kB, a tran-
scription factor that responds to oxidative stress and promotes
cell survival (Karin and Greten, 2005). Electrophoresis mobility
shift assay showed that PEITC inhibited the basal and TNFa-
stimulated NF-kB (p65/50) DNA binding activity (Figure S8).
NAC decreased the PEITC-induced NF-kB inhibition. These
data together suggest that the excessive ROS accumulation
induced by PEITC may abrogate certain redox-sensitive mole-
cules such as H-RAS and NF-kB, and thus impair cell survival
signals. Furthermore, PEITC caused a rapid activation of the
stress-activated protein kinase JNK in T72Ras cells, evidenced
by a significant increase of c-Jun phosphorylation (Figure 6E).
Interestingly, the JNK inhibitor SP600125 (10 mM) abolished
PEITC-induced c-Jun phosphorylation (Figure 6E, upper panel)
but did not significantly alter cell death (Figure 6E, lower panel),
suggesting that JNK activation was likely a response to ROS
stress, but not required to trigger cell death in T72Ras cells.
PEITC is effective in killing naturally occurring cancer
cells and exhibits significant therapeutic activity in vivo
The ability of PEITC to cause severe ROS-mediated damage in
the Ras-transformed cells prompted us to test its effectiveness
in killing naturally occurring cancer cells. As shown in Figure 7A,
PEITC was very effective in inhibiting proliferation of SKOV3
cells in both long-term MTT and colony formation assays, with
the IC50 values of approximately 0.6 mM. About 95% of cell pro-
liferating capacity was inhibited at 3 mM. Interestingly, the
growth-inhibitory curve obtained by long-term MTT assay was
that this MTT assay may be used to estimate the drug effect on
proliferation of cells that can not readily form colonies, such as
nontransformed cells. Flow cytometric analysis showed that
a short-term incubation with PEITC (10 mM) caused acute cell
death in more than half of the cells (Figures 7B and 7C). This
cytotoxic effect was largely abrogated by NAC. Consistently,
by 3 mM NAC (Figure 7D).
We then further compared the effect of PEITC on two other
ovarian cancer lines, A2008 (cisplatin-sensitive) and HEY (cis-
platin-resistant) cells. As shown in Figures 7E–7G, A2008 cells
were sensitive to PEITC, which exhibited significant cytotoxicity
in a ROS-dependent manner (reversed by NAC; Figure 7E).
Interestingly, although HEY cells were less sensitive to cisplatin,
they were highly sensitive to PEITC (Figure 7H), suggesting that
PEITC and cisplatin have different mechanisms of action. We
also used a colony formation assay to examine the effect of
PEITC on H1299 cells (lung cancer), and showed that this com-
pound was similarly effective, with more than 90% of the colony
forming capacity being abolished by 3 mM of PEITC (Figure S9).
The selectivity of PEITC against oncogenically transformed
cells and its effectiveness in killing naturally occurring cancer
cells in vitro led us to evaluate its activity in vivo. Nude mice
were inoculated intraperitoneally (i.p.) with T72Ras cells, which
grew tumor masses in peritoneal cavity and caused severe as-
cites (Figure 8A). Examination of the tumor sections revealed
the histological morphology resembling that of human ovarian
clearcell carcinoma (Figure8B).Interestingly, thetumoralsoex-
hibited active angiogenesis with extensive newly formed blood
Figure 6. Effect of PEITC on GPX, H-RAS-GTP, and JNK activity
A: Dose-dependent inhibition of human GPX activity by PEITC in vitro (mean
6 SD of three experiments).
B: Effect of PEITC on cellular GPX activity. T72Ras cells were incubated with
5–10 mM PEITC for 5 hr, and cellular protein extracts were assayed for GPX
activity without addition of PEITC in vitro (mean 6 SD of three experiments).
C: No decrease of GPX1 protein expression in T72Ras cells after PEITC treat-
ment. GPX expression was determined by Western blot analysis. D, dimer;
D: Inhibition of RAS activation by PEITC. T72Ras cells were incubated with
5 mM PEITC for 1–3 hr, followed by RAS-GTP activity assay. The input/GDP
band density of three experiments (mean 6 SD).
E: Activation of JNK by PEITC and the effect of JNK inhibitor on PEITC-
induced cell death. T72Ras cells were treated with the SP600125 for 1 hr
followed by 5 mM PEITC. JNK activity was assayed by analysis of phospho
c-Jun. Cell death was analyzed by flow cytometry after treatment with
PEITC 6 SP600125 (5 hr).
A R T I C L E
CANCER CELL SEPTEMBER 2006247
to die about 18 days after tumor inoculation, and most of the an-
imals died within 2 months. The median survival time for the un-
treated group (n = 20) was approximately 25 days (Figure 8C).
In contrast, treatment with PEITC by i.p. injection (50 mg/kg,
five times per week) significantly prolonged the animals’ sur-
vival. The median survival time for the PEITC-treated animals
was 48 days (n = 20), representing a 90% increase over that
of the untreated group (p < 0.05; Figure 8C). No severe toxicity
was observed, although some mice exhibited transient agitation
level of reversible body weight loss.
Increase of ROS stress incancer cells has long been recognized
and often viewed as an unfavorable event associated with car-
cinogenesis and cancer progression. However, recent studies
showed that it is possible to use agents that promote cellular
ROS accumulation to effectively kill cancer cells in vitro (Huang
et al., 2000; Pelicano et al., 2003). Since oncogenes have been
Figure 7. Cytotoxicity and ROS increase induced byPEITC in human ovarian
cancer cells and the protective effect of NAC
A: Dose-dependent inhibition of cell proliferation and colony formation by
PEITC in SKOV3 cells. Cell proliferation was measured by 10 day MTT assay
and colony formation assay (mean 6 SD of three experiments).
B: Acute cell death induced by PEITC. SKOV3 cells were preincubated with
3 mM NAC (1 hr) and then with 10 mM PEITC as indicated for 23 hr. Cell death
was measured by flow cytometry.
C: Effect of NAC on PEITC-induced cell death in SKOV3 cells under the same
conditions as in B (mean 6 SD of three experiments).
D: Effect of PEITC and NAC on ROS levels in SKOV3 cells. Cells were incu-
bated with 10 mM PEITC 6 3 mM NAC for 6 hr, and ROS were measured by
flow cytometry using DCF-DA.
or absence of 1 mM NAC. Cell death was detected by annexin-V/PI assay.
F: Quantitation of PEITC-induced cell death in A2008 cells in the presence
and absence of NAC under the same conditions as in E (mean 6 SDof three
G: Cytotoxicity of PEITC and CDDP in A2008 cells. Cells were incubated with
Cells were incubated with the indicated concentrations of each drug for
3 days followed by MTT assay (mean 6 SD of three measurements).
Figure 8. In vivo therapeutic activity of PEITC in mice bearing Ras-trans-
formed ovarian cancer cells
Forty nude mice were inoculated with T72Ras cells (2 3 106cells/mouse, i.p.)
and randomly divided into two groups (20/group) for treatment with PEITC
(50 mg/kg, i.p., daily, five times per week) or with control solvent as
described in the Experimental Procedures.
A: Gross appearance of two mice with severe ascites.
B: Histological morphology of hematoxylin-eosin-stained tissue slices of the
C: Comparison of animal survival of the PEITC-treated group and untreated
group (p < 0.05).
A R T I C L E
CANCER CELL SEPTEMBER 2006
shown to cause elevated ROS generation (Irani et al., 1997; Sat-
tler et al., 2000; Vafa et al., 2002), the intrinsic oxidative stress
associated with oncogenic transformation may render cancer
cells highly dependent on their antioxidant systems to maintain
redox balance and thus vulnerable to agents that impair antiox-
idant capacity. Inhibition of the antioxidant system in cancer
cells would lead to a severe accumulation of ROS leading to
cell death. In contrast, it is less likely to induce such severe
ROS stress in normal cells, due to their low basal ROS output.
This biochemical difference between normal and cancer cells
may constitute a basis for modulating cellular ROS as a strategy
to selectively kill cancer cells.
Using isogenic cell lines, here we showed that oncogenically
transformed cells are more sensitive to ROS-mediated damage
due to their high basal ROS generation, and that PEITC can be
a selective anticancer agent based on this biological basis.
formed with H-RasV12or Bcr-Abl exhibited a significant increase
of ROS compared to their parental cells. (2) This intrinsic oxida-
tive stress rendered them highly dependent on the GSH antiox-
idant system to maintain redox balance. Abolishing this system
by PEITC through depletion of GSH and inhibition of GPX en-
zyme activity led to a preferential ROS increase in the trans-
cells caused oxidative damage to the mitochondria membranes
and impaired the membrane integrity, leading to massive cell
death. In contrast, PEITC caused only a modest increase of
ROS insufficient to cause significant cell death in nontrans-
formed cells. (4) The strong correlation between ROS accumu-
lation and cell death induced by PEITC in T72Ras cells and
the suppression of cytotoxicity by catalase or NAC suggest
the critical role of ROS in PEITC-induced cell death. (5) The ob-
servations that BSO selectively killed transformed cells further
support the role of GSH in maintaining redox balance and can-
cer cell survival. Based on these findings, it is likely that the ther-
apeutic selectivity of PEITC is dependent on two important fac-
tors: the biological difference in redox regulation between the
oncogenically transformed cells and normal cells, and the ability
of PEITC to effectively abolish the GSH antioxidant system.
The elevated basal ROS in Ras-transformed cells seemed to
include multiple reactive species detected by HEt (O22) and
DCF-DA (H2O2, NO, etc.). This increase of ROS is not surprising,
reactive species due to their intracellular conversion. The rela-
tively low specificity of the chemical probes makes it difficult
to accurately estimate the proportion of each reactive species
in the cells. However, it is interesting to note that PEITC caused
a further increase of reactive species that oxidized DCF-DA and
DAF-FM but not HEt, suggesting that PEITC mainly induced ac-
cumulation of H2O2(and other species such as NO) but not O22.
PEITC has been shown to cause various degrees of ROS in-
crease in hepatoma cells. A study in HepG2 cells showed only
a minor role of ROS in PEITC-induced cell death (Rose et al.,
2003), whereas a study in PLC/PRF/5 hepatoma cell line dem-
onstrated that PEITC caused a significant ROS increase and in-
duced cell death reversible by antioxidants (Wu et al., 2005).
This variation may reflect different basal ROS levels in these
cells. In our study, the degrees of ROS accumulation and cell
death induced by PEITC were dependent on endogenous
ROS generation. The high basal ROS in Ras- or Bcr-Abl-trans-
formed cells seemresponsible for their highsensitivity toPEITC.
Because the transformed cells depend on GSH to counteract
the active ROS output, abrogation of the GSH antioxidant sys-
tem by PEITC would severely affect these cells, leading to oxi-
dative damage. Our study showed that PEITC induced ROS ac-
cumulation by twomechanisms: depletion ofGSH by promoting
its export and inhibition of GPX enzyme activity, which together
effectively disable the GSH antioxidant system.
Interestingly, incubation of cells with 5–10 mM PEITC led to
adepletion ofcellular GSH,whichisinthemMrange.Theexpla-
nation for this stoichiometric discrepancy is that PEITC can be
concentrated in the cells. A previous study showed that incuba-
tion of HL-60 cells with 5 mM [14C]-PEITC for 3 hr resulted in an
accumulation of PEITC at 0.5 nmole/106cells (Xu and Thornal-
ley, 2001), which is approximately in the mM range estimated
based on the cell number and cell size. In our study, incubation
of T72Ras cells with 5–10 mM PEITC for 2 hr led to intracellular
concentrations of 0.25–0.59 mM. Furthermore, 0.5 mM PEITC
was able to suppress more than 90% of GPX activity in the
GSH and inhibit GPX, along with the active ROS production in
cancer cells, may explain why this compound can cause a lethal
accumulation of ROS in the malignant cells within several hours.
The PEITC-induced ROS accumulation in Ras-transformed
cells seems to inactivate the redox-sensitive NF-kB and H-
RAS. Both molecules contain cysteine residues that are essen-
tial for their functions but sensitive to oxidative modification due
to the presence of thiol groups (Matthews et al., 1993; Mallis
et al., 2001). Incubation of T72Ras cells with PEITC significantly
inhibited the function of H-RAS to bind RAF-1 protein and sup-
pressed the ability of NF-kB to bind its consensus DNA. Since
this inactivation can be partially reversed by antioxidant NAC,
it is likely that ROS play a role in mediating PEITC-induced inac-
tivation of these two molecules. Oxidative inactivation of NF-kB
has been observed in other experimental systems, and PEITC
was recently shown to inhibit NF-kB in PC-3 cells (Khor et al.,
2006). While activation of RAS by S-glutathionylation is known
(Adachi et al., 2004), inactivation of RAS-GTP by ROS has not
tant roles in maintaining transformed phenotype and promoting
cell survival, and NF-kB seems required for the survival of the
Ras-transformed cells (Mayo et al., 1997), inactivation of these
molecules by PEITC likely contributes to its anticancer activity.
The stress-activated protein kinase (SAPK)/JNK pathway
plays an important role in cellular response to various stimuli, in-
cluding oxidative stress. We observed that treatment of T72Ras
tent with the observations in other cancer cell lines (Hu et al.,
2003; Xu et al., 2006). However, in PC-3 cells (prostate cancer)
triggering apoptosis (Hu et al., 2003; Xu et al., 2006), whereas in
Ras-transformed cells inhibition of JNK by SP600125 did not
suppress PEITC-induced cell death. It is possible that the high
level of ROS induced by PEITC in T72Ras cells caused lethal
damage to mitochondria and other cellular components and
important molecule for cell survival. Recent studies suggest that
PEITC is able to inhibit Akt activation, which may also contribute
to its cytotoxic action (Khor et al., 2006; Satyan et al., 2006).
Cisplatin is a common drug for ovarian cancer treatment. Our
study showed that PEITC has a superior selectivity compared
to cisplatin. The ability to preferentially kill malignant cells is
A R T I C L E
CANCER CELL SEPTEMBER 2006249
a promising feature of PEITC. Interestingly, PEITC seems more
nalley, 2000). We demonstrated that PEITC preferentially killed
oncogenically transformed cells by causing severe ROS accu-
mulation and oxidative inactivation of H-RAS and NF-kB. This
compound is also effective in naturally occurring cancer cells,
showedthat PEITC has therapeutic activity in mice bearing Ras-
transformed cells and prolongs median survival time by 90%.
Fora single agent, such invivo activity seems promising. The ef-
fective concentrations of PEITC (0.5–10 mM) may be achievable
in human. A pharmacokinetic study showed that an oral dose of
40 mg PEITC resulted in a plasma concentration of 1–2 mM
(Liebes et al., 2001). Interestingly, a recent work showed that
i.p. injection of PEITC (5 mmole, three times per week) retarded
the growth of prostate cancer xenografts when given 1 day be-
fore tumor implantation (Khor et al., 2006). Although this dosage
ofPEITC with curcumin waseffective, suggesting that combina-
In conclusion, our study suggests that the intrinsic oxidative
stress in cancer cells associated with oncogenic transformation
provides a basis for developing strategies to preferentially kill
cancer cells through ROS-mediated mechanism, and com-
pounds such as PEITC can be used to achieve such activity
in vitro and in vivo. Importantly, cancer cells in advanced
disease stage usually exhibit genetic instability and show signif-
icant increase in ROS generation due in part to the ‘‘vicious
cycle’’ in which ROS induce mutations leading to further meta-
bolic malfunction and more ROS generation (Pelicano et al.,
2004). Such highly malignant cells are often resistant to conven-
nant cells warrants further testing in preclinical and clinical
PEITC, 3-(4,5 dimethylthiazol-2-yl)-2,5-diphenyltetrazolium (MTT), NAC,
BSO, bovine catalase, and human GPX were purchased from Sigma-Aldrich
(St. Louis, MO). Cisplatin (PLATINOL-AQ) was a product of Bristol-Myers
Squibb (New York, NY). CM-H2DCF-DA, HEt, NAO, carbony cyanide 3-
chlorophenylhydrazone (CCCP), and Rho-123 were purchased from Invitro-
gen/Molecular Probes (Carlsbad, CA). SP600125 was acquired from EMD
biosciences (Calbiochem, San Diego, CA). PEITC was dissolved in DMSO
and freshly diluted in culture media before used. The final DMSO concentra-
tion was less than 0.1% (v/v).
Cell lines and cell culture
Both T72 (immortalized, nontumorigenic) and T72Ras (H-RasV12-trans-
formed, tumorigenic) cell lines (Liu et al., 2004) were cultured in medium con-
fetal bovine serum and 10 ng/ml epidermal growth factor. TonB210 cells har-
boring a tetracycline-inducible Bcr-Abl expression vector (Klucher et al.,
1998) were maintained in RPMI1640 medium with 10% FBS supplemented
with IL-3 (supplied as 10% conditioned medium from Wehi3B cell culture).
The IL3-dependent murine myeloid progenitor cell line DA1 and its stable
bcr-abl transfectant line DA1-3b (Vereecque et al., 1999) were cultured in
RPMI1640 medium. SKOV3, HEY, and H1299 were cultured in RPMI1640
Assays for cytotoxicity
Cell death was determined by flow cytometry after cells were double stained
with annexin-V-FITC and propidium iodide (PI), using an assay kit from BD
PharMingen (San Diego, CA) as described (Pelicano et al., 2003). To deter-
mine the drug effect on cell viability and long-term proliferation, we adapted
the MTT assay by seeding cells in 24-well plates for drug treatment, followed
by 10 day incubation. Cells were then incubated with MTT reagent for 4 hr,
lysed with DMSO, and transferred to 96-well plate for quantitation by a plate
reader. Since T72 cells could not form colonies, this 10 day MTT assay was
used as an alternative assay for drug effect on long-term cell proliferation.
adherence overnight, replacing the culture medium with 3 ml of fresh media
The cells were fixed and stained with 0.2% crystal violet (Sigma-Aldrich,
St. Louis, MO), and colonies of >50 cells were counted.
Determination of cellular ROS
Cellular ROS contents were measured by incubating the control or drug-
treated T72, T72Ras, or SKOV3 cells with 3 mM CM-H2DCF-DA for 60 min,
followed by flow cytometry using a FACSCalibur equipped with CellQuest
Pro software. For TonB210 cells, 5 mM of CM-H2DCF-DA was used in
a 60 min labeling to obtained sufficient fluorescence signal. O22was mea-
sured by flow cytometry using HEt (100 ng/ml) as described (Pelicano
et al., 2003).
Analysis of cellular GSH and its export to the culture medium
A GSH assay kit (Cayman Chemical Co. Ann Arbor, MI) was used to measure
total cellular glutathione. Cell extracts were prepared by sonication and de-
proteination using the conditions recommended by the manufacturer. Total
GSH was detected by measuring the product of glutathionylated DTNB by
using the standard curve generated in parallel experiments. Cells number
and median cell volume were quantified using a Coulter Z2particle count
and sizeanalyzer(Beckman Coulter, Inc., Fullerton, CA).Todeterminetheef-
fect of PEITC on GSH exported from cells to the culture medium, cells were
plated in equal density and cultured for 24 hr.The medium was then replaced
with serum-free medium with or without PEITC as indicated and incubated
for 3 hr, and the culture medium was removed for GSH assay as described
Determination of oxidative damage to mitochondrial membranes
Mitochondrial membrane lipid peroxidation was detected by measuring the
oxidation of cardiolipin, using NAO as a fluorescence dye (Nomura et al.,
2000). After cells were incubated with or without PEITC, the samples were la-
mitochondrial membrane potential were monitored by incubating cells with
1 mM Rho-123 for 1 hr, followed by flow cytometry. The decoupling agent
CCCP was used as a positive control to induce membrane depolarization.
GPX enzyme activity assay
The GPX enzyme activity was assayed using continuous spectrophotometric
rate determination as described (Wen et al., 2004). Human GPX (0.5 mU) pu-
rified from erythrocyte (Sigma-Aldrich) was premixed with or without PEITC
for 10 min and added to a reaction mixture containing 2 mM GSH, 0.1 U glu-
tathione reductase, and 0.2 mg/ml NADPH. Reaction was initiated by adding
H2O2(0.001%), andthekineticsofNADPH oxidation wasmonitored for 5min
at 340 nm. GPX activity (DA340 nm/min) was calculated by subtracting the
slope of the reaction from that of spontaneous oxidation in the control sam-
ple. Cellular GPX activity was determined using an equal amount of protein
extracts from control and PEITC-treated cells without further adding PEITC
in the reactions.
Assay of H-RAS-GTP activity
RAS-GTP activity was determined based on its specific binding to the down-
stream effector RAF-1 as described (Marais et al., 1995). Protein extracts
were prepared using the buffer provided in a RAS-GTP assay kit (Upstate
Cell Signaling Solution, Charlottesville, VA). RAS-GTP activity in the cell
extracts was detected by immunoprecipitation with 10 mg of RAF-1 binding
peptide conjugated to agarose beads, followed by Western blot with
1:1000 anti-H-RAS antibody (Santa Cruz Biotechnology, Inc., Santa Cruz,
CA). b-actin in whole-cell extracts was blotted as loading control. To verify
the specificity of the GTP bound activity, the control extracts were preincu-
bated with excess GDP to abolish the RAS-GTP ability to bind RAF-1
A R T I C L E
CANCER CELL SEPTEMBER 2006
peptide, which yielded negative signal after immunoprecipitation with
Assay of in vivo antitumor activity of PEITC
Animal experiments were performed under federal guidelines and approved
bythe Institutional AnimalCare andUseCommittee(IACUC) ofthe University
of Texas M.D. Anderson Cancer Center. Forty nude mice were inoculated
with T72Ras cells (2 3 106cells/mice, i.p.) and randomly divided into two
groups (20 mice each). Three days after tumor inoculation, the treatment
group received PEITC (50mg/kg, i.p., five times per week).The controlgroup
received an equal volume of solvent control. The mice were monitored daily
for signs of tumor growth, ascites, and body weight. Moribund animals were
sacrificed as mandated by the IACUC protocol, and the time of death was
recorded. Tumor tissues from representative mice were sectioned, embed-
ded in paraffin, and stained with hematoxylin and eosin for histopathologic
The statistical significance of the difference in cytotoxicity and ROS genera-
tion between transformed and nontransformed cells was evaluated using
Student’s t test. The statistical difference in animal survival curves between
the control and PEITC-treated animal groups was analyzed using the log-
rank test (two-tailed). These data analyses were performed using the Prism
software (GraphPad, San Diego, CA). A p value of less than 0.05 was con-
sidered statistically significant.
The Supplemental Data include nine supplemental figures and can be found
with this article online at http://www.cancercell.org/cgi/content/full/10/3/
The authors thank Tian-ai Wu, Yumin Hu, and Min Du for technical assis-
tance. This work was supported in part by grants CA085563, CA100428,
CA109041, and CA16672 from the National Institutes of Health. D.T. is a re-
cipient of a scholarship from the Anandamahidol Foundation under the royal
patronage of His Majesty the King of Thailand.
Received: January 22, 2006
Revised: June 3, 2006
Accepted: August 1, 2006
Published: September 11, 2006
Adachi, T., Pimentel, D.R., Heibeck, T., Hou, X., Lee, Y.J., Jiang, B., Ido, Y.,
and Cohen, R.A. (2004). S-glutathiolation of Ras mediates redox-sensitive
signaling by angiotensin II in vascular smooth muscle cells. J. Biol. Chem.
Balcerczyk, A., Soszynski, M., and Bartosz, G. (2005). On the specificity of
4-amino-5-methylamino-20,70-difluorofluorescein as a probe for nitric oxide.
Free Radic. Biol. Med. 39, 327–335.
J.A., Scher, H.I., Ludwig, T., Gerald, W., et al. (2005). Crucial role of p53-de-
pendent cellular senescence in suppression of Pten-deficient tumorigenesis.
Nature 436, 725–730.
Couzin, J. (2002). Cancer drugs Smart weapons prove tough to design. Sci-
ence 298, 522–525.
Deneke, S.M. (2000). Thiol-based antioxidants. Curr. Top. Cell. Regul. 36,
Frantz, S. (2005). Drug discovery: Playing dirty. Nature 437, 942–943.
Hampton, M.B., Stamenkovic, I., and Winterbourn, C.C. (2002). Interaction
with substrate sensitises caspase-3 to inactivation by hydrogen peroxide.
FEBS Lett. 517, 229–232.
Hockenbery, D.M., Oltvai, Z.N., Yin, X.M., Milliman, C.L., and Korsmeyer,
S.J. (1993). Bcl-2 functions in an antioxidant pathway to prevent apoptosis.
Cell 75, 241–251.
Hu, R., Kim, B.R., Chen, C., Hebbar, V., and Kong, A.N. (2003). The roles of
JNK and apoptotic signaling pathways in PEITC-mediated responses in hu-
man HT-29 colon adenocarcinoma cells. Carcinogenesis 24, 1361–1367.
Hu, Y., Rosen, D.G., Zhou, Y., Feng, L., Yang, G., Liu, J., and Huang, P.
(2005). Mitochondrial manganese-superoxide dismutase expression in ovar-
ian cancer: Role in cell proliferation and response to oxidative stress. J. Biol.
Chem. 280, 39485–39492.
Huang, P., Feng, L., Oldham, E.A., Keating, M.J., and Plunkett, W. (2000).
Superoxide dismutase as a target for the selective killing of cancer cells.
Nature 407, 390–395.
Irani, K., Xia, Y., Zweier, J.L., Sollott, S.J., Der, C.J., Fearon, E.R., Sundare-
san, M., Finkel, T., and Goldschmidt-Clermont, P.J. (1997). Mitogenic signal-
ing mediated by oxidants in Ras-transformed fibroblasts. Science 275,
to cancer development and progression. Nat. Rev. Immunol. 5, 749–759.
Khor, T.O., Keum, Y.S., Lin, W., Kim, J.H., Hu, R., Shen, G., Xu, C., Gopalak-
rishnan, A., Reddy, B., and Zheng, X. (2006). Combined inhibitory effects of
curcumin and phenethyl isothiocyanate on the growth of human PC-3 pros-
tate xenografts in immunodeficient mice. Cancer Res. 66, 613–621.
Klucher, K.M., Lopez, D.V., and Daley, G.Q. (1998). Secondary mutation
maintains the transformed state in BaF3 cells with inducible BCR/ABL
expression. Blood 91, 3927–3934.
Liebes, L., Conaway, C.C., Hochster, H., Mendoza, S., Hecht, S.S., Crowell,
J., and Chung, F.L. (2001). High-performance liquid chromatography-based
determination of total isothiocyanate levels in human plasma: Application to
studies with 2-phenethyl isothiocyanate. Anal. Biochem. 291, 279–289.
Liu, J., Yang, G., Thompson-Lanza, J.A., Glassman,A., Hayes, K.,Patterson,
defined model for human ovarian cancer. Cancer Res. 64, 1655–1663.
Mallis, R.J., Buss, J.E., and Thomas, J.A. (2001). Oxidative modification of
H-ras: S-thiolation and S-nitrosylation of reactive cysteines. Biochem. J.
Marais, R., Light, Y., Paterson, H.F., and Marshall, C.J. (1995). Ras recruits
Raf-1 to the plasma membrane for activation by tyrosine phosphorylation.
EMBO J. 14, 3136–3145.
Matthews, J.R., Kaszudska, W., Turcatti, G., Wells, T.N.C., and Hay, R.T.
(1993). Role of cysteine 62 in DNA recognition by the P50 subunit of
NF-kB. Nucleic Acids Res. 21, 1727–1734.
Mayo,M.W.,Wang,C.Y.,Cogswell, P.C.,Rogers-Graham, K.S.,Lowe,S.W.,
Der, C.J., and Baldwin, A.S., Jr. (1997). Requirement of NF-kB activation to
suppress p53-independent apoptosis induced by oncogenic Ras. Science
Nomura, K., Imai, H., Koumura, T., Kobayashi, T., and Nakagawa, Y. (2000).
Mitochondrial phospholipid hydroperoxide glutathione peroxidase inhibits
the release of cytochrome c from mitochondria by suppressing the peroxida-
tion of cardiolipin in hypoglycaemia-induced apoptosis. Biochem. J. 351,
Pelicano, H., Feng, L., Zhou, Y., Carew, J.S., Hileman, E.O., Plunkett, W.,
Keating, M.J., and Huang, P. (2003). Inhibition of mitochondrial respiration:
Anovelstrategy toenhance drug-induced apoptosisinhuman leukemia cells
by a reactive oxygen species-mediated mechanism. J. Biol. Chem. 278,
Pelicano, H., Carney, D., and Huang, P. (2004). ROS stress in cancer cells
and therapeutic implications. Drug Resist. Updat. 7, 97–110.
Radisky, D.C., Levy, D.D., Littlepage, L.E., Liu, H., Nelson, C.M., Fata, J.E.,
Leake, D., Godden, E.L., Albertson, D.G., Nieto, M.A., et al. (2005). Rac1b
and reactive oxygen species mediate MMP-3-induced EMT and genomic
instability. Nature 436, 123–127.
Ren, R. (2005). Mechanisms of BCR-ABL in the pathogenesis of chronic
myelogenous leukaemia. Nat. Rev. Cancer 5, 172–183.
A R T I C L E
CANCER CELL SEPTEMBER 2006 251
Rose, P., Whiteman, M., Huang, S.H., Halliwell, B., and Ong, C.N. (2003). Download full-text
Beta-Phenylethyl isothiocyanate-mediated apoptosis in hepatoma HepG2
cells. Cell. Mol. Life Sci. 60, 1489–1503.
Sattler, M., Verma, S., Shrikhande, G., Byrne, C.H., Pride, Y.B., Winkler, T.,
Greenfield, E.A., Salgia, R., and Griffin, J.D. (2000). The BCR/ABL tyrosine ki-
nase induces production of reactive oxygen species in hematopoietic cells.
J. Biol. Chem. 275, 24273–24278.
Satyan, K.S., Swamy, N., Dizon, D.S., Singh, R., Granai, C.O., and Brard, L.
(2006). Phenethyl isothiocyanate (PEITC) inhibits growth of ovarian cancer
cells by inducing apoptosis: Role of caspase and MAPK activation. Gynecol
Oncol. Published online April 17, 2006. 10.1016/j.ygyno.2006.03.002.
Sureda, F.X., Escubedo, E., Gabriel, C., Comas, J., Camarasa, J., and Ca-
mins, A. (1997). Mitochondrial membrane potential measurement in rat
cerebellar neurons by flow cytometry. Cytometry 28, 74–80.
Swartz, H.M., and Gutierrez, P.L. (1977). Free radical increases in cancer:
Evidence that there is not a real increase. Science 198, 936–938.
Szatrowski, T.P., and Nathan, C.F. (1991). Production of large amounts of
hydrogen peroxide by human tumor cells. Cancer Res. 51, 794–798.
Vafa, O., Wade, M., Kern, S., Beeche, M., Pandita, T.K., Hampton, G.M., and
Wahl, G.M. (2002). c-Myc can induce DNA damage, increase reactive oxy-
gen species, and mitigate p53 function: A mechanism for oncogene-induced
genetic instability. Mol. Cell 9, 1031–1044.
Quesnel, B. (1999). A new murine aggressive leukemic model. Leuk. Res. 23,
Wen, J.J., Vyatkina, G., and Garg, N. (2004). Oxidative damage during cha-
gasic cardiomyopathy development: Role of mitochondrial oxidant release
and inefficient antioxidant defense. Free Radic. Biol. Med. 37, 1821–1833.
ways in human PLC/PRF/5 cells. Eur. J. Pharmacol. 518, 96–106.
of human leukaemia cell growth by dietary isothiocyanatesand their cysteine
adducts in vitro. Biochem. Pharmacol. 60, 221–231.
Xu, K., and Thornalley, P.J. (2001). Involvement of glutathione metabolism in
the cytotoxicity of the phenethyl isothiocyanate and its cysteine conjugate to
human leukemia cells in vitro. Biochem. Pharmacol. 61, 165–177.
Xu, C., Shen, G., Yuan, X., Kim, J.H., Gopalkrishnan, A., Keum, Y.S., Nair, S.,
and Kong, A.N. (2006). ERK and JNK signaling pathways are involved in the
regulation of activator protein 1 and cell death elicited by three isothiocya-
nates in human prostate cancer PC-3 cells. Carcinogenesis 27, 437–445.
Yu, R., Mandlekar, S., Harvey, K.J., Ucker, D.S., and Kong, A.N. (1998). Che-
mopreventive isothiocyanates induce apoptosis and caspase-3-like activity.
Cancer Res. 58, 402–408.
Zhang, Y., Tang, L., and Gonzalez, V. (2003). Selected isothiocyanates rap-
idlyinduce growth inhibitionof cancer cells. Mol. Cancer Ther. 2, 1045–1052.
Zhao, H., Kalivendi, S., Zhang, H., Joseph, J., Nithipatikom, K., Vasquez-Vi-
var, J., and Kalyanaraman, B. (2003). Superoxide reacts with hydroethidine
but forms a fluorescence product that is distinctly different from ethidium:
Potential implications in intracellular fluorescence detection of superoxide.
Free Radic. Biol. Med. 34, 1359–1368.
Zhou, Y., Hileman, E.O., Plunkett, W., Keating, M.J., and Huang, P. (2003).
Free radical stress in chronic lymphocytic leukemia cells and its role in cellu-
lar sensitivity to ROS-generating anticancer agents. Blood 101, 4098–4104.
A R T I C L E
CANCER CELL SEPTEMBER 2006