Curcumin mediates ceramide generation via the de novo pathway in colon cancer cells

Abstract

A wealth of evidence supports the notion that curcumin, a phytochemical present in turmeric, is a potent chemopreventive agent
for colon cancer. Its mechanism of action remains incompletely understood. Here we report that curcumin's apoptosis-inducing
effects in colon cancer cell lines are accompanied by robust ceramide generation. This occurs through de novo synthesis as the increase in ceramide could be attenuated by pre-incubation of the cells with myriocin, and no changes were
observed in sphingomyelin levels, or in either acidic or neutral sphingomyelinase activities. Furthermore, cell death could
in part be reversed by myriocin, indicating, for the first time, that endogenous ceramide generation by this agent contributes
towards its biological activity. We then investigated the role of reactive oxygen species (ROS) in this phenomenon and demonstrated
that curcumin induced robust oxidant generation in the cell lines tested, and its reversal by N-acetylcysteine, completely attenuated apoptosis. We next confirmed that curcumin could activate c-jun N-terminal kinase (JNK)
and that its modulation could reverse cell death; however, this intervention could not block ceramide generation, or ROS production.
Conversely, however, the inhibition of ROS using N-acetylcysteine led to an inhibition of JNK activation. Hence, we conclude that curcumin induces apoptosis via a ROS-associated
mechanism that converges on JNK activation, and to a lesser extent via a parallel ceramide-associated pathway.

Full text

Curcumin mediates ceramide generation via the de novo pathway in
colon cancer cells
Maryam Moussavi
1
, Kiran Assi
1
,
Antonio Go
´
mez-Mun
˜
oz
2
and Baljinder Salh
1,3,

1
The Jack Bell Research Centre, 2660 Oak Street, Vancouver, BC, Canada,
2
Department of Biochemistry and Molecular Biology, Faculty of Science,
University of the Basque Country, PO Box 644, 48080 Bilbao, Spain and
3
Division of Gastroenterology, University of British Columbia, 100-2647
Willow Street, Vancouver, BC, Canada V5Z 3P1

To whom correspondence should be addressed. Tel: +1 604 875 5287;
Fax: +1 604 875 5447;
Email: bsalh@interchange.ubc.ca
A wealth of evidence supports the notion that curcumin,
a phytochemical present in turmeric, is a potent chemo-
preventive agent for colon cancer. Its mechanism of
action remains incompletely understood. Here we report
that curcumin’s apoptosis-inducing effects in colon cancer
cell lines are accompanied by robust ceramide generation.
This occurs through de novo synthesis as the increase in
ceramide could be attenuated by pre-incubation of the cells
with myriocin, and no changes were observed in sphin-
gomyelin levels, or in either acidic or neutral sphing-
omyelinase activities. Furthermore, cell death could in
part be reversed by myriocin, indicating, for the first
time, that endogenous ceramide generation by this agent
contributes towards its biological activity. We then
investigated the role of reactive oxygen species (ROS) in
this phenomenon and demonstrated that curcumin induced
robust oxidant generation in the cell lines tested, and
its reversal by N-acetylcysteine, completely attenuated
apoptosis. We next confirmed that curcumin could activate
c-jun N-terminal kinase (JNK) and that its modulation
could reverse cell death; however, this intervention
could not block ceramide generation, or ROS production.
Conversely, however, the inhibition of ROS using
N-acetylcysteine led to an inhibition of JNK activation.
Hence, we conclude that curcumin induces apoptosis
via a ROS-associated mechanism that converges on JNK
activation, and to a lesser extent via a parallel ceramide-
associated pathway.
Introduction
Colon cancer is without doubt an important health concern
globally, and consumes considerable resources in terms of
disease management and screening strategies (1). Hence,
increasing attention has been focused on ways to reduce its
incidence. The roles of dietary fiber, folate, calcium and
non-steroidal anti-inflammatory drugs are receiving consider-
able attention (2,3). Other dietary agents, such as curcumin
(diferuloylmethane) [1,7-bis-(4-hydroxy-3-methoxyphenyl)-
1,6-heptadiene-3,5-dione], are believed to play a sign ificant
role in populations that have a significantly reduced incidence
of this disorder (4).
Curcumin is a yellow pigment that has been isolated from
the ground rhizome of Curcuma species. Besides its use as a
coloring and flavoring agent, curcumin has anti-carcinogenic
properties, induces apoptosis in numerous cancer cell lines,
and inhibits carcinogen-induced tumorigenesis in rodent
intestine, as well as at other sites (5). At a cellular level cur-
cumin inhibits NFkB, AP-1, Cox-2, angiogenesis, MMP9, and
suppresses cell proliferation by inhibiting protein kinase PKC
and the epidermal growth factor receptor. Its apoptosis-inducing
effect is mediated through caspase 8, Bid cleavage, Bax, react-
ive oxygen species (ROS) generation and c-ju n N-terminal
kinase (JNK) (6–9).
Ceramides are the lipid backbone of sphi ngomyelin and
glycolipids, and are capable of influencing cell growth, viab-
ility and apoptosis. They are generated via two main pathways:
the de novo and the sphingomyelinase pathways. Both of these
have been implicated in the induction of apoptosis by several
drugs, including camptothecin, irinoteca n and gemcitabine
(10–12). Additionally, ceramides activate stress-activated
protein kinases (SAPK) such as JNK, which have also been
implicated in the induction of apoptotic cell death (13).
Ceramides have also been shown to act on mitochondria,
promoting the release of ROS (generated by the respiratory
chain), as well as promoting the release of cytochrome c,
thereby inducing apoptosis.
Significantly, dietary sphingolipids (including ceramides)
have recently been shown to reduce both the number of aber-
rant colonic crypt foci and number of tumors present in mice
(14,15). Furthermore, modulation of this pathway has been
reported to induce apoptosis in human colonic tumor xeno-
grafts as well as in metastatic colonic cance r (16,17).
In this study we demonstrate that curcumin induces ceram-
ide generation in colonic cancer cells, and have examined its
relationship with ROS generation and JNK activation; these
factors having recently been shown to play a significant role in
curcumin-induced apoptosis in Caki and HCT116 cells,
respectively (7,8). The data indicate that both ROS generation
and JNK activation are important mechanisms in curcumin-
mediated apoptosis in colon cancer cell lines. Additional
findings are supportive of ROS, but not ceramide, being
upstream of curcumin-induced JNK activation.
Materials and methods
Cell culture
HCT116 colon carcinoma cells were cultured using 10% fetal bovine serum
(FBS) supplemented McCoy’s 5A (Gibco) media containing penicillin and
streptomycin. These cells were a gift from B. Vogelstein. DLD1 cells were
cultured in 10% FBS containing RPMI1640 media (Fisher), and HT29 cells
were cultured in M199 media (Fisher, Nepean, ON, Canada) containing 10%
Abbreviations: BSO,
L-buthionine-[S,R]-sulfoximine; FACS, fluorescence-
activated cell sorter; FBS, fetal bovine serum; JNK, c-jun N-terminal kinase;
NAC, N-acetylcysteine; PBS, phosphate-buffered saline; ROS, reactive
oxygen species; TCL, thin layer chromatography.
Carcinogenesis vol.27 no.8 pp.1636–1644, 2006
doi:10.1093/carcin/bgi371
Advance Access publication February 25, 2006
#
The Author 2006. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org 1636
by guest on June 1, 2013http://carcin.oxfordjournals.org/Downloaded from

FBS and supplemented with penicillin and streptomycin. DLD1 and HT29
cells were obtained from American Type Culture Collection.
Cell death assay (MTS assay)
HCT116 cells were plated into 96-well plates at a density of 1 · 10
4
cells/well.
Cells were grown for 24 h before being placed in 1% FBS containing media
for 3 h. Cells were then stimulated with dose-escalating concentrations of
compounds for 24 h. MTS [3-(4,5-dimethylthiaol-2-yl) 5-(3-carboxymethoxy-
phenyl)-2-(4-sulphophenyl)-2H-tetrazolium, (250 mg/ml) inner salt] (Promega,
Nepean, ON, Canada) plus 20 mg of PMS (Gibco, Burlington, ON, Canada)
solutions were mixed together and added to each well for 1.5 h. Colorimetric
analysis was carried out using an ELISA plate reader at an absorbance of 490
nm. Each condition was plated in quintuplicate.
Fluorescence-activated cell sorter (FACS) analysis for sub G1 cell
population determination
Cells were seeded in 12-well plates and grown to 80% confluence. After
incubation in 1% FBS containing media for 3 h, cells were then exposed to
curcumin (Sigma-Aldrich, Sigma-Aldrich, Oakville, ON, Canada), for a dura-
tion of 4 h. Cells were then mechanically lifted by repeated pipetting and
washed twice with ice-cold phosphate-buffered saline (PBS) (Sigma). After
centrifugation at 1200 r.p.m., cell pellets were re-suspended in ice-cold 70%
ethanol for 1 h. Cells were then washed twice with ice-cold PBS and re-
suspended in 50 mg/ml propidium iodide (Gibco) and 25 mg/ml RNAse
(Fisher). The sub-diploid (sub G1, apoptotic) cell population was measured
using FACS (Epics XL-MCL; Beckman Coulter, Fullerton, CA). At least 10
4
cellular events were counted.
Hoechst staining
Colon carcinoma cells were seeded on sterile cover slips in six-well plates.
Upon reaching 80% confluence, cells were placed in 1% FBS containing media
for 3 h, and then incubated with 50–100 mM curcumin for a 4 h period. Media
were then removed and the cells were washed twice with PBS. Cells were fixed
using 1 ml of 100% ice-cold ethanol for 1 h. One milliliter of diluted Hoechst
33258 (1 mg/ml in PBS) was added in each well for 1 h. The cover slips were
then placed on glass slides. Pictures were taken from at least three different
fields of view.
Mitochondrial isolation for cytochrome c immunoblotting
Cells were seeded in six-well plates. After incubation in 1% FBS containing
media for 3 h, the cells were treated with curcumin for 4 h. Cells were harvested
by centrifugation at 1000 g for 5 min at 4

C. Cells were washed with ice-cold
PBS, and then collected using mitochondrial buffer (20 mM HEPES–KOH,
pH 7.5, 10 mM KCl, 1.5 mM MgCl
2
, 1 mM sodium EDTA, 1 mM sodium
EGTA, 250 mM sucrose, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl
fluoride, 20 mg/ml leupeptin, 1 mg/ml aprotinin and 10 mg/ml pepstatin). After a
15 min incubation of cells on ice, they were lysed by repeatedly passaging
through a 25-gauge needle 10–15 times. After a further 15 min incubation on
ice, the homogenates were centrifuged for 5 min at 1000 g (4

C) and the
supernatants were then centrifuged for 5 min at 10 000 g (4

C). The supernatant
containing the cytosolic fraction was collected and the pellet was re-suspended
in 50 ml mitochondrial buffer, thus yielding the mitochondrial fraction.
Western blot analysis was carried out to detect cytochrome c (Transduction
Laboratories, Mississauga, ON, Canada) released from mitochondria into the
cytosol, using both fractions.
Western immunoblotting
Cell lysates containing equivalent amounts of protein were resolved using 12%
SDS–PAGE and transferred on to nitrocellulose membrane using Biorad trans-
blot apparatus at 300 mA for 90 min. The nitrocellulose membrane (Gibco) was
then blocked in 5% skimmed milk in Tris-buffered saline/Tween-20 (TBST;
20 mM Tris–HCl, pH 7.4, 250 mM NaCl and 0.05% Tween-20) for 1 h. The
primary antibody was applied to the membrane overnight (1 : 1000 dilution)
and the latter then washed with TBST. The secondary antibody (conjugated to
horseradish peroxidase) was added at a dilution of 1 : 5000 for 1 h and the
membrane washed three times in TBST. Enhanced chemiluminescence was
utilized to visualize the blots.
Measurement of ceramide generation
Cells were labeled with 5 mCi [
3
H]palmitate (Mandel Scientific, Guelph, ON,
Canada) overnight for the determination of sphingomyelin. The radioactive
medium was removed and the cells were washed with non-radioactive medium.
For the determination of ceramide generation the cells were left in [
3
H]palmit-
ate-containing medium (18). After 3 h of starvation in 1% FBS containing
media, curcumin at increasing concentrations was added. The lipids were then
extracted with chloroform/methanol. Cells were scraped into 1 ml of ice-cold
methanol. One milliliter of chloroform and 0.9 ml of 2 M KCl + 0.2 M H
3
PO
4
was added to each aliquot and the chloroform phases were dried under nitro-
gen. Ceramides were separated by thin layer chromatography (TLC) utilizing
Silica Gel 60-coated glass plates (Fisher). Fifty percent of the lengths of these
TLC plates were developed in chloroform/methanol/acetic acid (9 : 1 : 1) and
then dried. The plates were re-developed in petroleum ether/diethylether/acetic
acid (60 : 40 : 1) and then dried and stained with iodide vapor. The identity of
the ceramide was standardized by the addition of authentic ceramide standards.
Radioactive ceramides were than quantified by scraping from the TLC plates
followed by liquid scintillation counting.
ROS determination
Cells were seeded on 12-well plates and grown to 80% confluence. After 3 h
incubation in 1% FBS containing media, 2 mM DCFH-DA (Molecular probes)
was added. In the presence of ROS this dye was oxidized and the change is
detected by using flow cytometry. After 1 h incubation with the dye, cells were
then treated with curcumin with or without antioxidants for 4 h. The cells were
than manually lifted by repeated pipetting, and spun down to remove the media
at 1200 r.p.m. for 1 min. The pellets were then re-suspended in 1 ml PBS (after
washing in PBS) and fluorescence measured. Where inhibitor studies were
performed, ROS scavengers were added after the cells had been labeled with
2 mM DCHF-DA dye for 1 h.
Statistical analysis
This was performed using the Student’s t-test (two-tailed).
Results
Curcumin induces apoptosis in colon cancer cell lines
The MTS assay was utilized to assess the effect of curcumin on
the viability of colon cancer cells. Three cell lines HCT116,
DLD1 and HT29 were exposed to increasing concentrations of
curcumin for 24 h, which is shown to be capable of reducing
cell viability in a dose-dependent manner (Figure 1B). Sub-
sequently, HCT116 cells were treated with increasing concen-
trations of curcumin for 4 h, and then stained with propidium
iodide. Using FACS, HCT116-treated cells exhibited a dose-
dependent increase in the sub G1 cell population (Figure 1C).
The effects of exposure of the HCT116 cells to curcumin on
PARP and procaspase 3 were then examined. In accordance
with the previous observations we observed changes compat-
ible with an apoptosis-inducing effect (Figure 1D), as indicated
by the presence of the cleaved form of PARP and disappear-
ance of procaspase 3. Next, confirmation that curcumin leads
to mitochondrial release of cytochrome c was then obtained in
DLD1 cells (Figure 1E). Figure 1F illustrates chromatin frag-
mentation, another marker of apoptosis, in Hoechst 33258
stained HT29 cells (similar results were obtained using
HCT116 cells).
Curcumin exposure leads to ceramide gener ation
In order to assess the role of curcumin on ceramide generation
HCT116 cells were incubated with tritiated palmitate. Labeled
cells were then treated with increasing concentrations of cur-
cumin. Cellular lipids were isolated and separated using TLC
and the amount of radioactivity was analyzed on a scintillation
counter. Here for the first time we have demonstrated that the
treatment of HCT116 cells with curcumin leads to an increase
in relative ceramide generation both dose- and time-
dependently (Figure 2A and B). Over a 5-fold increase in
ceramide generation was observed with curcumin used at a
concentration of 100 mM, while peak induction appeared to
occur after 12 h of incubation. Interestingly, the amplitude of
ceramide generated appeared to peak at a curcumin concen-
tration associated with the induction of apoptosis (50 mM).
Ceramides are generated via two different pathways, the
de novo pathway and the sphingomyelin hydrolysis pathway.
In order to delineate which pathway curcumin utilizes
to increase relative ceramide generation, we initially
Ceramide generation by curcumin
1637
by guest on June 1, 2013http://carcin.oxfordjournals.org/Downloaded from

performed pre-labeling experiments and determined changes
in sphingomyelin content of curcumin-treated cells. No
changes were found in this parameter. To further establ ish
that apoptosis did not rely on this pathway the activities of
acidic and neutral sphingomyelinases were determined. No
changes were found in these parameters either (data not
shown). Subsequently, the specific de novo pathway inhibitor,
myriocin (Figure 3A) was used to determine the role of this
pathway. HCT116 cells were employed in continuous labeling
experiments with tritiated palmitate. The data indi cate that pre-
treatment of HCT116 cells with myriocin leads to an impress-
ive attenuation of curcumin-induced ceramide gener ation.
Collectively, these findings provide the first evidence that
curcumin effects robust ceramide generation through the
de novo path way.
Inhibition of curcumin-induced ceramide generation leads to
the attenuation of cell death
As ceramide generation has been linked with apoptotic cell
death, the role of curcumin-induced ceramide generation in the
induction of cell death was analyzed. Utilizing FACS, sub-
diploid (apoptotic) populations were determined in HCT116
colon cancer cells, following exposure to curcumin, plus myri-
ocin pre-treatment (25 nM). Figure 3B indicates that in
curcumin-treated cells, pre-treated with myriocin, apoptosis
is partially attenuated. Specifically there is a reduction in
the sub G1 fraction from 23.9 to 17.4%. As this represented
only a 50% reversal of cell death, it implied that other mech-
anisms likely operate concurrently, and are also important
contributors to curcumin-induced apoptosis.
Treatment of HCT116 cells with curcumin leads to ROS
generation
ROS have been implicated in cell death induced under a vari-
ety of circumstances. Previous work has also indicated a role
for ROS in curcumin-induced cell death in other cell systems
(7,19). Hence, we were keen to explore whether the same
mechanism contributed towards curcumin’s ability to effect
cell death in our syst em. HCT116 cells were pre-labeled
with DCHF-DA (a free radical recognizing dye). After the
Fig. 1. Effect of curcumin on cell survival, chromatin fragmentation and apoptotic cell death in colon cancer cells. (A). Chemical structure of curcumin.
(B) Viability in three colon cancer cell lines (HCT116, DLD1 and HT29) decreases in response to curcumin exposure in a dose-dependent manner. The
bar chart depicts an MTS assay conducted following a 24 h exposure period in HCT116 (gray), DLD1 (black) and HT29 (diagonal stripes) cell lines.
(C) Increasing the concentration of curcumin results in an increase in sub-diploid cell population as analyzed by flow cytometry in HCT116 cells.
(D) HCT116 cells were exposed to curcumin at the indicated concentrations and western blot analysis was performed as described in the Materials and
methods section. The appearance of the cleaved form of PARP, together with a reduction of procaspase 3, was observed, following incubation with curcumin.
(E) Treatment of DLD1 cells with 100 mM of curcumin leads to the release of cytochrome c from the mitochondrial fraction into the cytosolic fraction. Cells
were starved for 3 h in 1% FBS containing media, and were then treated with 100 mM of curcumin for 4 h. Mitochondrial and cytosolic fractions were then
separated. Western blot analysis indicates the leakage of cytochrome c from the mitochondrial to cytosolic compartment. Results are representative of three
independent experiments. (F) Pictorial representation of chromatin fragmentation in HT29 cells treated with 100 mM of curcumin after a 4 h period. HT29
cells were stained with Hoechst 33258.
M.Moussavi et al.
1638
by guest on June 1, 2013http://carcin.oxfordjournals.org/Downloaded from

treatment of cells with increasing concentrations of curcumin,
cells were analyzed using flow cyto metry. The data indicate
that curcumin leads to an impressive increase in ROS produc-
tion (Figure 4A and B).
To determine whether the underlying mechanism leading to
this ROS gener ation could be modulated, several agent s were
employed. These included
L-buthionine-[S,R]-sulfoximine
(BSO), an agent depleting glutathione, N-acetylcysteine
(NAC), an antioxidant and a glutat hione (GSH) precursor
and catalase (CAT). The data demonstrate that BSO
(25 mM), NAC (5 mM) and catalase (1000 U) were all able
to quench curcumin-generated ROS (Figure 4C).
Inhibition of ROS leads to the attenuation of cell death
We were interested in whether the observed increase in ROS
had any relevance for curcumin-induced cell death. Using
FACS analysis we demonstrate that NAC and BSO were
both able to almost completely attenuate apoptosis in
curcumin-treated HCT116 cells (Figure 5). These data indicate
that the curcumin-induced ROS observed in our system is
comparable with the previous findings in other systems
where a direct link with apoptosis has been established.
Curcumin-activated JNK is not upstream of ROS or ceramide
pathways
In a previous work, curcumin has been shown to induce apop-
tosis by activating the JNK pathway in HCT116 cells. We first
confirmed that JNK is activated in response to curcumin as
early as 1 h post-exposure (Figure 6A). It is interesting to note
that there appears to be a reduction in the JNK protein con-
comitant with the activation especially of the p 54 isoform(s).
Next we demonstrated that a specific JNK inhibitor, SP600125
(Anthra [1,9-cd] pyrazol-6 (2H)-one1, 9-pyrazoloanthrone),
was capable of significantly preventing curcumin-induced
apoptosis in both HCT116 and DLD cells (Figure 6B and
6C, respectively).
In order to determine the sequence of events initiated by
curcumin we performed additional experiments, which were
directed at determining whether or not JNK was an upstream
regulator of either ROS or ceramide generation. First, HCT116
cells were pre-labeled with DCHF-DA for 1 h and treated with
50 mM curcumin along with 20 mM SP600125. ROS generation
was measured using flow cytometry. The data indicate that
Fig. 2. Effect of curcumin on ceramide generation. (A) HCT116 colon
cancer cells were labeled with [H
3
]palmitate overnight and placed for 3 h
in 1% FBS containing McCoy’s 5A media. Cells were than treated with
increasing concentrations of curcumin for 4 h. Ceramide generation was
determined as described in the Materials and methods section. The data are
derived from three independent experiments performed in duplicate.
(B) Time course of ceramide generation utilizing pre-labeled HCT116
cells. Cells were placed for 3 h in 1% FBS containing media and treated
for various time points (up to 24 h) with 50 mM curcumin. The graph
shows a time-dependent increase in ceramide generation.
Fig. 3. Curcumin-induced ceramide generation occurs via the de novo
pathway and is partially responsible for cell death. (A) Curcumin elevates
ceramide generation via the de novo pathway. HCT116 cells were
continuously labeled with [H
3
]palmitate throughout the experiment.
Cells were treated with myriocin (25 nM) for 0.5 h and then treated with
50 mM of curcumin for 4 h, before the determination of ceramide levels
(**P < 0.001). (B) Flow cytometry analysis of the sub-diploid population
of HCT116 cells treated with the de novo ceramide generation inhibitor
myriocin and curcumin demonstrates a partial but significant reduction in
the apoptotic cell population in comparison with the curcumin treatment
alone (*P < 0.01). (This experiment was repeated in triplicate on three
separate occasions and the data are shown as the mean + SD).
Ceramide generation by curcumin
1639
by guest on June 1, 2013http://carcin.oxfordjournals.org/Downloaded from

JNK inhibition does not influence the curcumin-induced
change in ROS production (Figure 6D). Second, the effect
of SP600125 on ceramide generation was investigated.
There was no change in the level of relative ceramide genera-
tion in either of the two cell lines exam ined, consequent upon
the inhibition of JNK (data not shown).
Collectively, these observations indicated that JNK activa-
tion is downstream of ROS generation, and indeed, NAC is
able to attenua te the curcumin-induced JNK activation
(Figure 7A). This figure shows a correlation between the gen-
eration of ROS and JNK activity, with re-activation of JNK
occurring when the NAC concentration was reduced to below
1 mM. The data also indicate that a certain threshold of ROS
may be required for JNK activation to occur as concentrations
of NAC between 1 and 2.5 mM only partially attenuate ROS
generation but there is no corresponding JNK act ivation.
When the role of ceramide generation upon JNK activation
was examined (Figure 7B), there appeared to be no obvious
role for ceramide in generating ROS or JNK activation at
the concentration of myriocin used throughout this study.
Fig. 4. Curcumin leads to ROS generation in HCT116 cells. (A and B) HCT116 cells treated with increasing concentrations of curcumin (0–0.1 mM) exhibit
an increase in ROS generation dose-dependently. HCT116 were placed in 1% FBS containing McCoy’s 5A media. Cells were pre-labeled with DCHF-DA
for 0.5 h before the incubation with curcumin for 4 h. Fluorescence emitted was determined by using flow cytometry. (C) Interruption of ROS generation
reverses curcumin-induced ROS generation. HCT116 cells were pre-labeled with DCHF-DA, for 0.5 h before the incubation with inhibitors for another 0.5 h,
followed by curcumin for 4 h. The emitted fluorescence was determined as described above.
Fig. 5. Inhibition of curcumin-induced ROS generation leads to the
attenuation of cell death. HCT116 cells were placed in 1% FBS
containing McCoy’s 5A media for 3 h. Cells were then exposed to ROS
inhibitors for 30 min. Subsequently, HCT116 cells were exposed to
curcumin (100 mM) for 4 h. Flow cytometry was performed to determine
the sub G1 population.
M.Moussavi et al.
1640
by guest on June 1, 2013http://carcin.oxfordjournals.org/Downloaded from

However, a potential role for ceramide-mediated JNK activa-
tion (independently of ROS generation) cannot be completely
excluded as some inhibition of the p45 JNK isoform may be
seen at a myriocin conce ntration of 50 nM. Overall, the repor-
ted findings are compatible with the notion that JNK activation
is predominantly downstream of ROS generation.
Discussion
Chemopreventative agents, such as micronutrients and syn-
thetic pharmacological compounds, lead to apoptotic cell
death in pre-malignant cells. Curcumin has emerged as an
important compound in this arena. Previous studies have
shown that its administratio n to cells leads to apopt otic cell
death, characterized by caspase 8 activation, ROS generation
and JNK activation. Other important players include Bax
and Hsp 70 (20). Our results extend these findings, by
demonstrating that it leads to apoptosis in colon cancer cells
via a mechanism that includes ceramide generation for the first
time. In keeping with mechanisms utilized by some othe r drugs
we have confirmed that curcumin utilizes the de novo pathway.
Activation of sphingomyelinases has been thought to be the
main pathway leading to the generation of ceramides particip-
ating in apoptosis, following stress stimuli. However, this
appears not to be the case with curcumin as indicated by
the following observations. First, there was no change in sphin-
gomyelin levels in colon cancer cells following curcumin
exposure. Second, by employing desipramine, a widely used
sphingomyelinase inhibitor, levels of ceramide generation in
HCT116 cells treated with curcumin were determined. The
results demonstrated no significant difference in curcumin-
induced ceramide generation in HCT116 cells, which were
pre-treated with desipramine compared with those that were
not (data not shown). Finally, direct evaluation of sphingomy-
elinase activity failed to demonstrate any change consequent
upon curcumin exposu re. Collectively, these results argue
against acidic and neutral sphingomyelinases being involved
in curcumin-induced apoptosis.
Several important findings have implicated distur bances in
ceramide metabolism as being important in colon cancer.
First, ceramide levels have been reported to be decreased in
human colon cancer; treatment using ceramide analogs and
ceramidase inhibitors led to rapid cell death in SW403
Fig. 6. Inhibition of JNK attenuates apoptosis but not curcumin-induced ROS or ceramide generation. (A) Curcumin activates JNK in HCT116 cells.
After seeding on six-well plates, a time-course study was performed following curcumin exposure. Equivalent amounts of protein from the cell extracts were
resolved on 11% SDS–PAGE and transferred on to nitrocellulose membranes. These were probed with the phospho-JNK antibody and after stripping the
membranes were then probed with the protein (JNK) antibody. (B) HCT116 cells (and in C, DLD1 cells) were seeded on to 12-well plates, upon reaching
80% confluence, cells were placed in 1% FBS containing media for 1 h, then pre-incubated with 20 mM of SP600125 (JNK inhibitor) for 0.5 h before
exposure to curcumin (100 mM) for 4 h. Flow cytometry was performed as described in the Materials and methods section (*P < 0.01). (D) HCT116
cells were prepared as in B (i.e. incubated with 20 mM of SP600125 for 0.5 h) and incubated with DCDA-FA for 1 h, before curcumin exposure.
Absolute fluorescence emitted was then determined.
Ceramide generation by curcumin
1641
by guest on June 1, 2013http://carcin.oxfordjournals.org/Downloaded from

cells. Furthermore, the ceramidase inhibitor B13 prevented the
growth of metastatic lesions (16). Sphingomyelin supple-
mentation has not only been shown to reduce colonic aberrant
colonic crypts and adenocarcinomas in 1, 2-dimethylhydrazine
(DMH)-treated CF1 mice (14,21) but also enhances 5FU effi-
cacy in colonic tumor xenografts (17). Thus, the potential of
combining agents that increas e sphingomyelinase activity and
elevate intracellular ceramide through de novo generation will
warrant further attention in future work.
A number of studies have established a link between
ceramide elevation and ROS generation. Other reports have
established the role of ROS in the induction of apoptosis. Since
curcumin leads to apoptotic cell death utilizing the mitochon-
drial apoptotic pathway (i.e. cytochrome c release), we invest-
igated the effect of curcumin on ROS generation. In
accordance with the recent studies (19), we have demonstrated
that the exposure of HCT116 colon cancer cells to curcumin
leads to an increase in ROS generation. Furthermore, we
confirm findings in other cancer cell systems, specifically
that pre-treatment with NAC in curcumin-treated human breast
epithelial cells MCF10A, and human renal Caki cells, attenu-
ates apoptosis.
Notably, it was observed that the activity of curc umin could
be attenuated by BSO as well as NAC. This seems somewhat
counterintuitive since BSO inhibits g-glutamylcysteinyl syn-
thetase and would affect the redox status of the cell toward a
climate favoring apoptosis. However, curcumin has been
shown to affect the same enzyme also (22) as well as modu-
lating glutathione levels and forming glutathionylated pro-
ducts; hence, we cannot exclude an alternative mode of
action for BSO in this study. One possibility includes that
of glutathione acting as a possible cofactor for mediating
curcumin’s oxidant effect. Hence the depletion of glutathione
may reduce both the extent of ROS generation and thus
Fig. 7. Correlation between ROS generation and JNK activation, in response to modulation of ROS and ceramide in curcumin-treated HCT116 cells. For the
western blot analyses, HCT116 cells were seeded in six-well plates and following pre-treatment with concentrations of NAC (or myriocin) indicated, they
were exposed to curcumin for 4 h. Cells were then harvested and equivalent aliquots of cell lysate protein were subjected to 11% SDS–PAGE, and the
resulting membranes processed as described in the Materials and methods section. Blots were first probed with the P-JNK antibody, and then the membranes
were stripped and then re-probed with the protein JNK antibody. ROS were determined as described in the legend to Figure 4. The lanes in the western blot
analyses correspond to the bars (and hence the treatment conditions) in the charts below for each figure.
M.Moussavi et al.
1642
by guest on June 1, 2013http://carcin.oxfordjournals.org/Downloaded from

apoptosis. In this respect (albeit without a demonstrable
oxidant effect) BSO has been shown to attenuate
methylseleninic-acid-induced apoptosis in Hep G2 cells
(23). While NAC elevates glutathione levels it is likely that
its antioxidant activity is more important in the context of our
study, hence explaining its ability to reduce curcumin-induced
apoptosis.
Previously, it has been demonstrated that the stimulation of
HCT116 cells with exogenous C
2
-ceramide does not lead to
JNK activation (24). Our data indicate that JNK activation via
curcumin is likely to be independent of endogenous ceramide
generation as well . The picture is rendered more complex by
the fact that ROS are robustly generated by curcumin and that
these species could pote ntially saturate JNK signaling, and
thus making the pathway refractory to other inputs. Indeed,
the possibl e involvement of JNK-mediated signaling being
responsible for the small attenuation of curcumin-induced
apoptosis by myriocin cannot be completely excluded by
the findings reported in this study.
In conclusion, we have shown that curcumin’s apoptosis-
inducing effects on colon cancer cells incorporate a mechan-
ism that is predominantly dependent upon ROS production and
downstream JNK activation, and to a lesser degree by ceramide
generation (summarized in Figure 8).
Acknowledgements
This work was supported by funds from the Canadian Society of Intestinal
Research, and in part from the Crohn’s and Colitis Foundation of Canada.
Conflict of Interest Statement: None declared.
References
1. Parkin,D.M., Bray,F., Ferlay,J. and Pisani,P. (2005) Global cancer
statistics, 2002. CA Cancer J. Clin., 55, 74–108.
2. Giovannucci,E. and Willett,W.C. (1994) Dietary factors and risk of colon
cancer. Ann. Med., 26, 443–452.
3. Willett,W.C. (2000) Diet and cancer. Oncologist, 5, 393–404.
4. Mohandas,K.M. and Desai,D.C. (1999) Epidemiology of digestive tract
cancers in India. V. Large and small bowel. Ind. J. Gastroenterol., 18,
118–121.
5. Aggarwal,B.B., Kumar,A. and Bharti,A.C. (2003) Anticancer potential of
curcumin: preclinical and clinical studies. Anticancer Res., 23, 363–398.
6. Anto,R.J., Mukhopadhyay,A., Denning,K. and Aggarwal,B.B. (2002)
Curcumin (diferuloylmethane) induces apoptosis through activation of
caspase-8, BID cleavage and cytochrome c release: its suppression by
ectopic expression of Bcl-2 and Bcl-xl. Carcinogenesis, 23, 143–150.
7. Woo,J.H., Kim,Y.H., Choi,Y.J. et al. (2003) Molecular mechanisms of
curcumin-induced cytotoxicity: induction of apoptosis through
generation of reactive oxygen species, down-regulation of Bcl-XL and
IAP, the release of cytochrome c and inhibition of Akt. Carcinogenesis,
24, 1199–1208.
Fig. 8. Proposed effect of curcumin on colon cancer cells. Curcumin leads to ceramide generation via the de novo pathway. ROS are generated, which then
leads to the activation of JNK. Inhibition of each of the three components studied (ROS, JNK and ceramide) is able to variably reverse cell death observed
due to curcumin exposure in this system.
Ceramide generation by curcumin
1643
by guest on June 1, 2013http://carcin.oxfordjournals.org/Downloaded from

8. Collett,G.P. and Campbell,F.C. (2004) Curcumin induces c-jun N-terminal
kinase-dependent apoptosis in HCT116 human colon cancer cells.
Carcinogenesis, 25, 2183–2189.
9. Rashmi,R., Kumar,S. and Karunagaran,D. (2005) Human colon cancer
cells lacking Bax resist curcumin-induced apoptosis and Bax
requirement is dispensable with ectopic expression of Smac or
downregulation of Bcl-XL. Carcinogenesis, 26, 713–723.
10. Kolesnick,R. (2002) The therapeutic potential of modulating the ceramide/
sphingomyelin pathway. J. Clin. Invest., 110, 3–8.
11. Radin,N.S. (2003) Killing tumours by ceramide-induced apoptosis: a
critique of available drugs. Biochem. J., 371, 243–256.
12. Ogretmen,B. and Hannun,Y.A. (2004) Biologically active sphingolipids in
cancer pathogenesis and treatment. Nat. Rev. Cancer, 4, 604–616.
13. Mansat-de Mas,V., Bezombes,C., Quillet-Mary,A., Bettaieb,A.,
D’Orgeix,A.D., Laurent,G. and Jaffrezou,J.P. (1999) Implication of
radical oxygen species in ceramide generation, c-Jun N-terminal kinase
activation and apoptosis induced by daunorubicin. Mol. Pharmacol., 56,
867–874.
14. Dillehay,D.L., Webb,S.K., Schmelz,E.M. and Merrill,A.H.,Jr. (1994)
Dietary sphingomyelin inhibits 1,2-dimethylhydrazine-induced colon
cancer in CF1 mice. J. Nutr., 124, 615–620.
15. Schmelz,E.M., Bushnev,A.S., Dillehay,D.L., Sullards,M.C., Liotta,D.C.
and Merrill,A.H.,Jr. (1999) Ceramide-beta-
D-glucuronide: synthesis,
digestion, and suppression of early markers of colon carcinogenesis.
Cancer Res., 59, 5768–5772.
16. Selzner,M., Bielawska,A., Morse,M.A., Rudiger,H.A., Sindram,D.,
Hannun,Y.A. and Clavien,P.A. (2001) Induction of apoptotic cell death
and prevention of tumor growth by ceramide analogues in metastatic human
colon cancer. Cancer Res., 61, 1233–1240.
17. Modrak,D.E., Rodriguez,M.D., Goldenberg,D.M., Lew,W. and
Blumenthal,R.D. (2002) Sphingomyelin enhances chemotherapy efficacy
and increases apoptosis in human colonic tumor xenografts. Int. J. Oncol.,
20, 379–384.
18. Erdreich-Epstein,A., Tran,L.B., Bowman,N.N. et al. (2002) Ceramide
signaling in fenretinide-induced endothelial cell apoptosis. J. Biol.
Chem., 277, 49531–49537.
19. Bhaumik,S., Anjum,R., Rangaraj,N., Pardhasaradhi,B.V. and Khar,A.
(1999) Curcumin mediated apoptosis in AK-5 tumor cells involves
the production of reactive oxygen intermediates. FEBS Lett., 456,
311–314.
20. Rashmi,R., Kumar,S. and Karunagaran,D. (2004) Ectopic expression of
Hsp70 confers resistance and silencing its expression sensitizes human
colon cancer cells to curcumin-induced apoptosis. Carcinogenesis, 25,
179–187.
21. Schmelz,E.M., Dillehay,D.L., Webb,S.K., Reiter,A., Adams,J. and
Merrill,A.H. (1996) Sphingomyelin consumption suppresses aberrant
colonic crypt foci and increases the proportion of adenomas versus
adenocarcinomas in CF1 mice treated with 1,2-dimethylhydrazine:
implications for dietary sphingolipids and colon carcinogenesis. Cancer
Res., 56, 4936–4941.
22. Piwocka,K., Jaruga,E., Skierski,J., Gradzka,I. and Sikora,E. (2001)
Effect of glutathione depletion on caspase-3 independent apoptosis
pathway induced by curcumin in Jurkat cells. Free Radic. Biol. Med.,
31, 670–678.
23. Shen,H., Ding,W. and Ong,C. (2002) Intracellular glutathione is a cofactor
in methylseleninic acid-induced apoptotic cell death in HepG2 cells. Free
Radic. Biol. Med., 33, 552–561.
24. Ohmori,M., Shirasawa,S., Furuse,M., Okumura,K. and Sasazuki,T. (1997)
Activated Ki-ras enhances sensitivity of ceramide-induced apoptosis
without c-Jun NH2-terminal kinase/stress-activated protein kinase or
extracellular signal-regulated kinase activation in human colon cancer
cells. Cancer Res., 57, 4714–4717.
Received September 9, 2005; revised January 14, 2006;
accepted February 11, 2006
M.Moussavi et al.
1644
by guest on June 1, 2013http://carcin.oxfordjournals.org/Downloaded from