Effect of 4-Methoxyindole-3-Carbinol on the proliferation of colon cancer cells in vitro, when treated alone or in combination with indole-3-Carbinol

Article abstract of DOI:10.1021/jf101806t

Consumption of cruciferous vegetables and cancer prevention seem to be positively associated. We present an easy two-step synthesis for 4-methoxyindole-3-carbinol (4MeOI3C), the expected breakdown product of 4-methoxyglucobrassicin during ingestion. 4MeOI3C inhibited the proliferation of human colon cancer cells DLD-1 and HCT 116 with IC₅₀ values of 116 μM and 96 μM, respectively, after 48 h in vitro, and is therefore a more potent inhibitor than indole-3-carbinol (I3C). 4MeOI3C and I3C combined in different molar ratios inhibited proliferation in a nearly additive to slightly synergistic manner. Proliferation was inhibited by 100 μM 4MeOI3C after 48 h without affecting cell cycle phase distribution, indicating an overall-slowdown effect on the cell cycle. However, 200 μM 4MeOI3C caused a very high level of cell death and an accumulation of living cells in the G₀/G ₁ phase, indicating a concentration-dependent mode of action. We conclude that 4MeOI3C might play a role in the cancer preventive effect of cruciferous vegetables.

Full text of DOI:10.1021/jf101806t

J. Agric. Food Chem. 2010, 58, 8453–8459 8453
DOI:10.1021/jf101806t
Effect of 4-Methoxyindole-3-carbinol on the Proliferation
of Colon Cancer Cells in Vitro, When Treated Alone
or in Combination with Indole-3-carbinol
R
EMY
KRONBAK, FRITZ
D
UUS
,
AND
O
LE
VANG
*
Department of Science, Systems and Models, Roskilde University, Universitetsvej 1,
000 Roskilde, Denmark
4
Consumption of cruciferous vegetables and cancer prevention seem to be positively associated.
We present an easy two-step synthesis for 4-methoxyindole-3-carbinol (4MeOI3C), the expected
breakdown product of 4-methoxyglucobrassicin during ingestion. 4MeOI3C inhibited the prolifera-
tion of human colon cancer cells DLD-1 and HCT 116 with IC50 values of 116 μM and 96 μM,
respectively, after 48 h in vitro, and is therefore a more potent inhibitor than indole-3-carbinol (I3C).
4
MeOI3C and I3C combined in different molar ratios inhibited proliferation in a nearly additive to
slightly synergistic manner. Proliferation was inhibited by 100 μM 4MeOI3C after 48 h without
affecting cell cycle phase distribution, indicating an overall-slowdown effect on the cell cycle.
However, 200 μM 4MeOI3C caused a very high level of cell death and an accumulation of living
cells in the G
0 1
/G phase, indicating a concentration-dependent mode of action. We conclude that
4
MeOI3C might play a role in the cancer preventive effect of cruciferous vegetables.
KEYWORDS: 4-Methoxyindole-3-carbinol; indole-3-carbinol; glucosinolates; cruciferous vegetables;
colorectal cancer; combination index analysis
INTRODUCTION
IC50 value around 250 μM in human colon cancer cells after 48 h
of treatment (17). However, 3-hydroxymethyl-1-methoxyindole
For years, a positive relationship between the high intake
of cruciferous vegetables and prevention of cancer has been
observed (1-3). The content of glucosinolates in such vegetables
is believed to cause this positive relationship. When disrupting the
plant cells, e.g., during food preparation or chewing, the enzyme
myrosinase is introduced to the glucosinolates and cleaves off
glucose, leaving a compound that rapidly converts to a thiocya-
nate, an isothiocyanate or a nitrile. The isothiocyanate products
formed from indolic glucosinolates (Figure 1) are unstable in
aqueous solutions, and further conversion rapidly takes place,
leaving hydroxymethyl (carbinol) compounds as the main initial
products of indolic glucosinolate breakdown (4). These products
might further react in the acidic milieu of the stomach, forming a
range of polyaromatic indolic metabolites including several
dimers and oligomers (2, 5).
So far, the focus in cancer prevention research of the indoles
has almost exclusively been on 3-hydroxymethylindole (indole-3-
carbinol, I3C), the breakdown product of the glucosinolate
glucobrassicin (GB) (Figure 1). The compound has been shown
to inhibit proliferating cancer cells from a range of human tissues
in vitro, including breast (6-8), prostate (9-11), skin (12), and
pancreas (13) among others. As reviewed elsewhere (14-16), the
inhibitory effect of I3C can be ascribed to the involvement of the
compound in multiple critical pathways of, e.g., the cell cycle and/or
apoptosis. I3C was previously shown to have a relatively high
(
(
NI3C), the breakdown product of 1-methoxy-GB (NeoGB)
Figure 1), was about 10-fold more potent, when comparing the
IC50 value to that of I3C in these cells (17).
Using a rat model studying the modulation of xenobiotic
metabolism, it was reported that a mixture of indolic glucosino-
lates, i.e., 48% GB, 36% NeoGB, and 16% 4-methoxy-GB
(
4MeOGB) (Figure 1) myrosinase-treated or untreated, was
significantly a more powerful inducer of microsomal hepatic
CYP1A1 protein than solutions of GB or NeoGB alone (18),
and the presence of 4MeOGB was ascribed as being possibly
responsible for this effect. This observation has stimulated our
interest to test the biological activity of the 4MeOGB breakdown
product, 3-hydroxymethyl-4-methoxyindole (4-methoxyindole-
3
-carbinol, 4MeOI3C) (Figure 1). We present here an easy
two-step synthesis of 4MeOI3C from commercially available
-methoxyindole. Additionally, we demonstrate the effect of the
4
compound, when exposed to proliferating DLD-1 and HCT 116
human colon cancer cells in vitro, either alone or combined with
I3C. Finally, the effect of 4MeOI3C on cell cycle phase distribu-
tion and the ability of the compound to cause cell death are
elucidated.
MATERIALS AND METHODS
Synthesis. For the synthesis and structure elucidation, highest grade
solvents and chemicals were used exclusively. 4MeOI3C was synthesized in
a two-step synthesis. First, a formyl group was introduced to 4-methoxy-
indole (99%, Aldrich, USA) through a Vilsmeier-Haack reaction inspired
by others (19,20). Phosphorus oxychloride (0.8 mL (8.7 mmol)) was mixed
*
Corresponding author. Phone: þ45 4674 2552. Fax: þ45 4674 3011.
E.-mail: ov@ruc.dk.
©
2010 American Chemical Society
Published on Web 07/01/2010
pubs.acs.org/JAFC

8454 J. Agric. Food Chem., Vol. 58, No. 14, 2010
Kronbak et al.
medium were seeded in each well. After an overnight incubation, the cells
of one plate were trypsinized for 20 min and counted on Coulter Particle
2
Counter and Size Analyzer Z (Beckmann-Coulter Inc.) operated by
Multisizer AccuComp 3.01 software. On the remaining plates, 4MeOI3C
(
(
0-200 μM in dimethyl sulphoxide (DMSO, Sigma-Aldrich Chemical Co.
USA)), 4 replicates per concentration) was introduced in new medium.
DMSO exposure never exceeded 0.1%. After 24, 48, 72, or 96 h of incuba-
tion, cells were counted using a Coulter counter. Cells with a diameter
smaller than 10.6 μm were considered dead cells and were subtracted from
the cell number.
Combinatorial Treatment. For the combinatorial treatment analysis,
another source of suppliers was used. McCoy’s 5A medium with L-gluta-
2þ
2þ
Figure 1. Structures of Indolic Glucosinolates and Carbinols of Relevance
for This Study.
mine, PBS without Ca and Mg , and gentamicin 50 mg/mL were pur-
chased from Lonza (Belgium) and FCS (lot 0739 L) from Biochrom AG,
Germany. EDTA solution was made as 0.02% EDTA (Sigma-Aldrich
Chemical Co.) in PBS. The affected cell number was estimated by Neutral
Red (NR) analysis. Similarly distributed on four 96-well plates (Black
Nunc F96 MicroWell 96-well plates (Nunc, Denmark)), five series of test
compound concentrations (8 replicates per concentration, 2 on each plate)
were prepared. In one series, 4MeOI3C (80-200 μM) was tested alone and
in another I3C (140-450 μM) alone (I3C from Sigma-Aldrich Chemical Co.
with 3.5 mL of ice-cold dimethylformamide. To this mixture, 1.0 g
(6.8 mmol) of 4-methoxyindole dissolved in 23 mL of ice-cold dimethyl-
formamide was added under magnetic stirring. The compounds reacted
for 1 h at 40 °C. To the pale yellow reaction mixture, 35 mL of ice-cold
water was added, and the color changed to burgundy. After basification
with 5% aqueous sodium hydroxide, the color of the solution again
changed to yellow, and an additional 30 mL of water was added. The
aqueous solution was extracted three times with a 5% (v) solution of
methanol in dichloromethane. Then the combined organic extracts
were washed with 150 mL of brine and dried over sodium sulfate.
After evaporating the organic layer under reduced pressure into a small
volume of about 10 mL, pure 3-formyl-4-methoxyindole was subse-
quently obtained by preparative TLC (Silica gel 40 F254, layer thickness
(
USA) was recrystallized before use). In addition, three molar 4MeOI3C/
I3C ratios, i.e., 1:1 (60-135 μM 4MeOI3C), 1:2 (40-100 μM 4MeOI3C)
and 2:1 (60-165 μM 4MeOI3C), were examined. Each plate contained
1
0 control wells and 12 background wells. The experiment was performed
as follows: 8000 DLD-1 or HCT 116 cells in 100 μL of medium were
seeded in all wells except background wells, which received 100 μL of
medium without cells. After an overnight incubation, 50 μL of medium
containingDMSO and testcompounds in concentrations 3-fold thedesired
final concentrations was added to each well. Control and background
wells received 50 μL of medium containing 0.3% DMSO only. After 48 h
of incubation, all wells were washed with 100 μL of PBS, and there was
added 100 μL of medium containing 50 μg/mL NR (Sigma-Aldrich
Chemical Co. (USA)) (freshly made from a stock solution of 1 mg/mL
NR in Milli-Q water), followed by incubation for three hours. All wells
were washed twice in 100 μL of PBS. Then, 100 μL of 96% ethanol was
added, and the plates were covered with parafilm and placed on a rocking
table for 20 min before scanning in a fluorescence-reader (excitation
0.25 mm) using as eluent a solution of 5% (v) methanol in diethylether/
dichloromethane (1:5, v/v). The yield of product was 745 mg (63%).
Mp=153-155 °C.
A subsequent reduction reaction led to the formation of 4MeOI3C.
Seven hundred milligrams of 3-formyl-4-methoxyindole (4.0 mmol) re-
acted with 150 mg of sodium borohydride in 12.5 mL of a solvent mixture
of methanol/ethanol/chloroform (1:8:5, v/v/v) for 6 h under magnetic
stirring. Afterward, the solvent was evaporated under reduced pressure.
The solid residue was dissolved in 7 mL of 0.1 M aqueous sodium
hydroxide, and the solution was extracted three times with 35 mL of
diethylether. The combined organic extracts were dried over sodium
sulfate, and the solvent was evaporated under reduced pressure. The yield
of the crude product was 662 mg (93%). Recrystallization from methanol
gave 432 mg of beige crystals. The purity was acceptable for biological
testing (>98%) with the main impurity being water shown by NMR.
Mp = 91-92 °C.
5
30 nm and emission 590 nm) (BioTek Synergy HT, Biotek Instruments
Inc., operated by BioTek KC4 3.4 software). On each plate, the back-
ground mean value was subtracted from all other values. The NR uptake
of cells treated with 4MeOI3C and/or I3C was estimated relative to that of
control cells (nontreated) as being the surviving fraction (SF) of cells. The
affected fraction (AF) was the fraction of living cells missing, when
compared to the control; thus, AF þ SF = 1. From the concentration-
effect relationship within the AF interval 0.2 to 0.8, the Combination
Index (CI) introduced by Chou and Talalay (21, 22) was found. The
following equation was used under the assumption of a mutually exclusive
compound interaction: CI = C4MeOI3C/ICx 4MeOI3C þ CI3C/ICx I3C, where
ICx 4MeOI3C and ICx I3C are the inhibitory concentrations of 4MeOI3C and
I3C, respectively, affecting a certain fraction (x %) of the cells, when
treated separately, and C4MeOI3C and CI3C are the concentrations of the
two compounds also affecting x % of cells, when used in combination.
CI < 1, CI = 1, and CI > 1 indicate synergistic, additive, and antagonistic
effects, respectively. In a review by Chou (23), the terms nearly additive
(CI: 1.10-0.90), slight synergism (CI: 0.90-0.85), and moderate syner-
gism (CI: 0.85-0.70) are used before plain synergism (CI: < 0.70). For
each of the series of 1:1, 1:2, and 2:1 ratios of 4MeOI3C and I3C of each
cell line, an AF-CI relationship was described using these terms.
Structure Determination. Different methods of NMR were used for
structure determination. 3-Formyl-4-methoxyindole was dissolved in
1
acetone-d
in CDCl
meter operated by VNMR 6.1B software). 4MeOI3C was dissolved in
6
for H NMR performed at 300 MHz or in 10% (v) CD
3
OD
1
3
3
for C NMR at 75 MHz (Varian Mercury 300 NMR spectro-
1
3
CDCl and similarly examined at 300 and 75 MHz. In the H NMR
spectrum obtained, the compound purity was found from integrals. In
addition, a NOE-dif experiment was performed in order to find the
chemical shift of 4MeOI3C H-5 (the proton closest to the methoxy group).
In order to determine the chemical shift of 4MeOI3C C-5, a HMQC 2D
NMR experiment was performed (Varian Inova 600 NMR spectrometer
operated by VNMR 6.1B software) in CDCl . 2D NMR data were
3
analyzed using Mestrelab Research MestRe-C software.
Cell Growing. DLD-1 and HCT 116 human colon cancer cells
(obtained from American Type Culture Collection, ATCC) were grown
under similar conditions. Except when handled during experimental
performance or when stated otherwise, cells were kept in an incubator
Influence on Cell Cycle and Cell Death. HCT 116 cells were seeded
2
5
in 56.7 cm dishes (6.5 ꢀ 10 cells in 10 mL of medium each) (Nunclon Δ
2
(37 °C, 5% carbon dioxide atmosphere with maximal humidity) and
56.7 cm dishes (Nunc, Denmark)) and incubated overnight. The medium
grown in McCoy’s 5A medium w GlutaMAX (Gibco, USA) (contain-
ing 10% FCS (lot 1046 FF, Biochrom AG, Germany) and 45 μg/mL
gentamicin (Gibco, USA)). Subcultivation was done every second or third
day before 80% confluence was reached. During subcultivation, cells were
treated with Versene (containing 0.05% trypsin) (from (Gibco, USA)) for
was removed and replaced with medium containing 0.1% DMSO and no
compound (controls), 100 μM or 200 μM 4MeOI3C, or 250 μM I3C.
Three replicates per concentration were used. After 48 h of incubation, the
medium and 5 mL of washing PBS were combined in centrifuge tubes in
order to collect floating cells. The remaining cells were trypsinized in 1 mL
of Versene and resuspended in 5 mL of medium before adding to the tubes,
which were spun at 300g and 4 °C for 5 min (standard) afterward. The
pellet was dissolved in PBS, and the cell number was estimated. At least
10 min. At passage 30, cells were discarded.
Growth Inhibition. To investigate the time- and concentration-
dependent effect of 4MeOI3C on proliferating cells, 24 well plates were
4
5
used (Nunclon Δ 24-well plates (Nunc, Denmark)); 3.5 ꢀ 10 cells in 1 mL
5.0 ꢀ 10 cells were transferred to a new tube, washed in PBS, and

Article
J. Agric. Food Chem., Vol. 58, No. 14, 2010 8455
resuspended in 200 μL of PBS. Under whirl mixing, 2 mL of ice cold 70%
ethanol was slowly added, and the tubes were kept at -20 °C until FACS
analysis was performed. For the analysis, cells in ethanol were spun,
washed in PBS, and resuspended in 400 μL of PBS. This was followed by
whirl mixing with 50 μL of 1 mg/mL RNase in PBS, and subsequently,
Table 1. Time- and Concentration-Dependent Effect of 4MeOI3C on Pro-
liferating Human Cancer Cells in Vitro (Mean of 3 Identical Experiments ( SD)
living cells in percent of control
concentration
b
50 μL of 0.4 mg/mL propidium iodide (PI) in PBS was added, and the cells
cell line
DLD-1
(μM)
24 h
48 h
72 h
96 h
were incubated in the dark for 30 min at room temperature. All samples
kept in the dark were analyzed on a flow cytometer within 2 h
a
a
a
a
25
89.4 ( 1.7a 88.6 ( 0.3a 92.7 ( 1.4 89.6 ( 3.6
83.1 ( 4.9 75.4 ( 6.1 80.6 ( 3.0 79.6 ( 3.7
73.8 ( 3.4a 56.5 ( 4.3a 56.5 ( 3.9 58.5 ( 1.3
45.0 ( 6.7 23.5 ( 7.6 12.1 ( 3.9 6.08 ( 1.7
a
5
1
2
0
(
FACSCalibur, Becton Dickinson used with BD Bioscience CellQuest
Pro 4.0.2 operating software and Verity Software House Inc. ModFit LT
data processing software). Living cells were categorized in the G /G
phase, S phase, and G /M phase from their content of DNA. Registrations
in the sub-G /G area in the histograms were considered as the sum of
a
a
a
00
00
a
0
1
HCT 116
25
89.0 ( 4.8 89.3 ( 1.7a 93.3 ( 2.4 92.7 ( 4.1
2
a
a
5
1
2
0
79.1 ( 11.7a 71.5 ( 5.9 76.0 ( 8.5 82.4 ( 3.6
0
1
a
a
a
a
00
00
63.5 ( 12.0 46.3 ( 5.9a 44.2 ( 5.9 59.6 ( 6.8
dead/dying cells and debris.
a
a
43.7 ( 9.4 16.9 ( 2.9 7.33 ( 1.9 4.90 ( 2.0
Data Processing and Statistics. Unless otherwise stated, all experi-
ments were performed three times, and the results are given as the mean
with standard deviation (SD) . Microsoft Excel 2007 in combination with
SYSTAT 11 for Windows was used for data processing and statistics. All
data were tested for being normally distributed according to the Shapiro-
Wilk test. When appropriate, variance analysis was performed using one-
way ANOVA with Bonferroni post hoc test at statistical significance levels
p < 0.05 and p < 0.01.
a
b
p < 0.01 significant from the control. All cells, including control cells, were
exposed to 0.1% DMSO.
Table 2. Median Effect Concentrations (IC50) of 4MeOI3C in Micromolar over
Time (Mean of 3 Identical Experiments ( SD)
cell line
24 h
48 h
72 h
96 h
DLD-1
HCT 116
189 ( 23
162 ( 62
116 ( 14
96 ( 13
119 ( 8
98 ( 15
116 ( 5
120 ( 10
RESULTS
Structure Elucidation. NMR spectral data of 3-formyl-4-meth-
oxyindole were in agreement with assignments given elsewhere (19)
deviation at 24 h blurs this time-dependent pattern. However,
the compound seems to be an equally potent inhibitor of both
cell lines.
Additionally, we calculated the effect of 4MeOI3C on cell
doubling time (T ). Forty-eight hours of treatment with 100 μM
(data not shown). The final product, 4MeOI3C, was dissolved in
1
CDCl , and spectral data were as follows: H NMR (300 MHz,
3
CDCl ) δ 3.27 ppm (t, 1, J = 6.8 Hz, OH), 3.98 (s, 3, OCH ),
3
3
4.80 (dd, 2, J = 6.8 Hz, 0.6 Hz, CH ), 6.55 (dd, 1, J = 7.8 Hz,
2
d
0.4 Hz, H-5), 6.93 (dd, 1, J = 2.4 Hz, 0.3 Hz, H-2), 6.97 (dd, 1, J =
4
MeOI3C increased Td to 27.2 h ((1.1 h) for DLD-1 cells
compared to T = 18.5 h ((0.4 h) in untreated cells. Treating
HCT 116 cells similarly, we found that T was increased to 26.2 h
8.2 Hz, 0.7 Hz, H-7), 7.11 (dd, 1, J = 8.2 Hz, 7.8 Hz, H-6), 8.30
13
d
(s, br, 1, NH). C NMR (75 MHz, CDCl ) δ 55.5 ppm (OCH ),
3 3
d
5
1
8.2 (CH OH), 99.7 (C-5), 105.2 (C-7), 116.4 (C-3), 116.8 (C-3R),
21.0 (C-2), 123.2 (C-6), 138.4 (C-7R), 153.2 (C-4). Calculations
2
((1.7 h), compared with T = 16.3 h ((0.9 h) in untreated cells.
d
In both cases, the Td at 100 μM treatment was statistically
significantly different from that of controls at the p < 0.01 level.
Increasing the concentration to 200 μM 4MeOI3C, we found that
the cell number was more or less unchanged.
could not predict which of the 4MeOI3C protons H-5 and H-7
should have chemical shifts at 6.55 ppm and 6.97 ppm, respec-
tively. For this reason, we performed a NOE-dif experiment in
order to point out the H-5 shift. The doublet at 6.55 ppm was
heavily affected when the methoxy group was irradiated. Neither
could our calculations predict the C-5 and C-7 chemical shifts.
HMQC 2D NMR showed couplings between H-5 and the 99.7
ppm carbon. Also, a coupling between H-7 and the 105.2 ppm
carbon was revealed, thus confirming the position of these critical
carbons. Furthermore, couplings between H-2 and C-2 as well as
H-6 and C-6 were confirmed. Finally, H-C couplings within the
methoxy group and the hydroxymethyl group were observed.
4
MeOI3C and I3C Exhibit Nearly Additive to Slightly Syner-
gistic Effects. In three molar 4MeOI3C/I3C ratios, i.e., 1:1, 1:2,
and 2:1, we investigated whether additive effects were achieved
during the combinatorial treatment of the compounds. The CI
was plotted for each series with respect to AF (Figure 2C,D),
which is positively related to the effect concentrations of the
compounds (Figure 2A,B). In Figure 2C,D, the line CI = 1 repre-
sents additive effects, whereas values below and above represent
synergism and antagonism, respectively. In DLD-1 cells, near-
additive to slightly synergistic effects were achieved, when using
equal or half amounts of 4MeOI3C relative to I3C, whereas
the responses achieved were closer to an additive effect when
4
MeOI3C Inhibits Cell Growth in a Time- and Concentration-
Dependent Manner. Using cell counting, we estimated the effect of
MeOI3C on proliferating DLD-1 and HCT 116 cells. Cells were
4
grown for 24, 48, 72, or 96 h under exposure to increasing
concentrations of 4MeOI3C (25-200 μM). Surviving cells at
different concentrations were compared to control cells percen-
tage-wise (Table 1). The cell number was inhibited in a concen-
tration- and time-dependent manner by 4MeOI3C. Already after
4
MeOI3C was in double amounts relative to I3C. All series in
HCT 116 cells exhibited near-additive effects.
MeOI3C Influences Cell Cycle Phase Distribution in a Con-
4
centration-Dependent Manner. The distribution of HCT 116 cells
in the phases of the cell cycle was investigated by FACS analysis.
Cells were untreated (control) or treated with a medium (100 μM)
or a high (200 μM) concentration of 4MeOI3C or with a medium
concentration of I3C (250 μM) for 48 h, and the distribution is
shown in Figure 3. Although our cell counting indicated that
100 μM 4MeOI3C is close to IC , no differences in phase
distribution were observed, when compared to that of the control
cells after 48 h. Increasing the 4MeOI3C to 200 μM, a large build
up in the G /G phase was observed (increased from 44% to
2
4
4 h of treatment, a statistically significant effect of 50 μM
MeOI3C on DLD-1 and of 100 μM on HCT 116 cells was
observed. Furthermore, these data revealed that 200 μM kept
the cell number almost unchanged throughout the experimen-
tal period. From concentration-effect curves, we calculated the
median inhibitory effect concentrations (IC ) for each cell line
50
50
at all time points (Table 2). In order to reach the median effect
after only 24 h, a relatively high dose of 189 ( 23 μM 4MeOI3C
was necessary in DLD-1 cells. Extending the time of exposure
reduced the concentrations needed for the median inhibitory
effect concentration. In HCT 116 cells, a relatively large standard
0
1
74%), and cells in S phase were reduced dramatically (from 37%
to 8%). These observations indicate that new cellular targets of
4MeOI3C are introduced along with increasing concentrations.

8456 J. Agric. Food Chem., Vol. 58, No. 14, 2010
Kronbak et al.
Figure 2. Combinatory effect of treatment of 4MeOI3C and I3C. The affected fraction (AF) of cells compared to controls after 48 h of compound treatment is
plotted with respect to total compound concentration in micromolar for DLD-1 (A) and HCT 116 (B) cells, respectively. The concentration is in this way either
pure 4MeOI3C or I3C, oritis thesum of concentrations ina seriesof molar 4MeOI3C/I3Cratios 1:1, 1:2, or 2:1. Means of four identical experiments( STD are
given, and the lines are tendency lines. The Combination Index (CI) is plotted with respect to AF, as well for DLD-1 (C) and HCT 116 (D) cells. CI relationships
of threemolar 4MeOI3C/I3C ratios aregivenforeach cellline as the meansoffouridentical experiments( STD. CI< 1, CI= 1, and CI> 1 represent synergism,
additive effect, and antagonism, respectively. See text for further explanations of the effect.
whereas 200 μM exhibited strong cytotoxicity by increasing sub-
G /G1 registrations dramatically to 88% ((5%), statistically
0
significantly different from controls at p < 0.01 level (data not
shown). However, the sub-G /G area of histograms is very poorly
0
1
defined, and further investigations are necessary in order to evalu-
ate the ability of 4MeOI3C to induce apoptosis. Thus, 100 μM
4MeOI3C can be considered as a cytostatic concentration with an
overall-slowdown effect not related to a specific cell cycle phase.
Following 48 h of treatment with 250 μM I3C, which corre-
sponds to IC , an increased accumulation in the G /G phase
50
0
1
was observed. The level of counts in the sub-G /G area was
0
1
17% ((1%).
DISCUSSION
In cruciferous vegetables, various indoles have been identified,
and they may in part explain the proposed cancer preventive
activity of these vegetables. The cancer preventive effects of
indoles, mainly I3C, have been related to the modulation of meta-
bolism of procarcinogens, but in the last 10 years, the inhibitory
effect on cell proliferation has provided a relevant additional
explanation.
In general terms, the parental compounds of I3C and NI3C are
present at the highest concentrations in, e.g., broccoli (24), but in
some cases, the level of 4MeOGB in broccoli and Brussels sprouts
is shown to exceed the GB and NeoGB levels (25). Previous
multivariate analysis of the effect of broccoli showed significant
contribution from all of the glucosinolates found in broccoli (26),
which is why a closer analysis of the cellular effects of 4MeOI3C is
relevant compared with I3C. Previously, such a comparison was
done for NI3C and I3C (17).
Figure 3. Effect of 4MeOI3C and I3C on the cell cycle of HCT 116 cells.
Living cells (100%) distributed in phases of the cell cycle after 48 h of
treatment with 100 or 200 μM 4MeOI3C or 250 μM I3C. No test compound
was introduced to control cells, and all cells were exposed to 0.1% DMSO.
Means of three identical experiments ( STD are given. * Means of only
two identical experiments ( half of the range are given. Within each phase
(
i.e., horizontally compared), only groups not sharing identical letters
(
a-c) are statistically significantly different at the p < 0.01 level.
After 48 h of culture, 9% ((2%) of all registrations were in
the sub-G /G area for control cells. Exposure to 100 μM of
0
1
4
MeOI3C increased this number to 21% ((2%), which was
not statistically significant, indicating milder cytotoxic effects,

Article
Pure 4MeOI3C was produced in two steps from 4-methoxy-
indole in an overall yield of 38% after the final recrystallization.
Likewise, pure NI3C has been produced in an overall yield of
J. Agric. Food Chem., Vol. 58, No. 14, 2010 8457
In these experiments, two different methods of estimating the
cell number were used. Comparing these two assays, we observed
55-70% higher IC50 values for 4MeOI3C at 48 h, when perform-
ing the NR assay instead of the cell counting experiments
(compare Figure 2A,B and Table 2). When counting the cells,
the limit between living and dead cells is set manually and may
therefore differ from the NR assay. In the NR assay, only living
cells take up NR, and the measure of light emission defines the
surviving fraction of cells relative to untreated cells. These two
assays are in fact two different ways of viewing the same problem,
but different results may well appear from the differences in
experimental conditions. Different plates are used (24-well versus
96-well dishes), trypsination for Coulter counting may damage
some of the living cells, and in the Coulter counting experiment,
the indole-containing medium is replacing the medium used for
seeding the cells the day before. However, both our Coulter
counter experiments and the CI analysis were quite consistent
when reproduced, indicating reliable results.
The combinatory effects of several compounds have been
described by different models: Chou and Talalay introduced the
Combination Index method and included the median effect plot
(MEP) as a mathematical tool to improve the linearity of the
concentration-effect curves (22). The concentration-effect curves
after 48 h of treatment with the indoles exhibited nearly perfect
linearity in all series. When these data were transformed to MEP,
the lines were slightly disturbed, and the use of MEP transforma-
tion did not alter the final AF-CI relationship. For these reasons,
the MEP was not included in our final data processing.
20% in a similar two-step synthesis (27). However, the synthesis
of NI3C differed from that of 4MeOI3C in that it involved two
distillations, one after each step, as both NI3C and the interme-
diate compound 3-formyl-1-methoxyindole are oils. 3-Formyl-
4-methoxyindole was purified on p-TLC plates and therefore not
further recrystallized before use in the second reaction step. That
the synthesis of 4MeOI3C is about doubly as effective as that of
NI3C might in part be a consequence of the different purification
methods and in part of the different reactivities of the reactant
indoles.
This is the first analysis of the cellular effect of 4MeOI3C, and
the potency is about 2-fold higher than I3C on the basis of IC50
values in both colon cell lines. NI3C was previously found to be
about 10-fold more potent compared with that of I3C in the same
cell lines as those used here (17). These differences in potency
indicate that the actual combinations of various indoles in the
cruciferous vegetables may be crucial for the biological effects of
the vegetable.
I3C is highly sensitive to the acidic milieu of the stomach but
also the near-neutral pH (28). As reviewed elsewhere (5), the
indolic carbinols likely react to form a range of products through-
out the digestive system, e.g., di- and oligomers and products
from reactions with other plant substances present such as
ascorbic acid. Therefore, our in vitro model does not perfectly
reflect the in vivo situation following the ingestion of cruciferous
vegetables but is useful for evaluating the combinatory effect of
the indoles.
The CI analysis was chosen over other methods for describing
combinatorial treatments, e.g., fractional product method and
isobologram analysis, in part because the analysis seems kineti-
cally independent of the two compounds, in part because further
investigations can include more compounds which are still
comparable to results achieved so far, and in part because the
degree of synergism, additive effect, and antagonism is easily
described in a curved relationship to an interval of cellular
inhibition (expressed as AF).
Even though several studies have been performed on the in
vitro effect of I3C, it has been proposed that little or no effect can
be ascribed to the compound itself. Bradlow (29) states that I3C
in cell culture medium under normal conditions might be so
unstable that the effect can only be ascribed to spontaneously
0
formed condensation products of I3C, mainly, 3,3 -diindolyl-
methane (DIM). We observed well-defined, concentration-
dependent inhibitory effects of 4MeOI3C and I3C, separately
treated to cells or combined, and the CI analysis was performed
on the basis of these observations. The calculations of CI were
based on initial indole concentrations, as the compounds dis-
solved in medium were only added once at the beginning of the
testing period. The experiments do not give information on the
conversion of the indoles included in the analyses, but we suggest
that future experiments should focus on this aspect.
As indicated above, the combined effect will strongly depend
on the actual combination of the indoles present. Therefore, we
wanted to show the combined effect of 4MeOI3C and I3C. Near-
additive to slightly synergistic effects were achieved in DLD-1
cells, whereas the combination of the two indoles only exhibited
a near-additive effect in HCT 116 cells. The results indicate that
the strongest effect is obtained, when the level of I3C matches
or, in DLD-1 only, exceeds that of 4MeOI3C; however, the effect
is only slightly synergistic. In extracts from some cruciferous
plants such as watercress (30), nozawana (31), and broccoli
sprouts (32), 4MeOGB levels higher than GB levels have been
observed, but in most plant extracts, the level of GB exceeds
that of 4MeOGB (33-35). No strong synergism is achieved
during combinatorial treatment, which may indicate that the two
compounds to a high degree have a similar mode of action.
Neave et al. (17) proposed a higher degree of combinatory effect
of NI3C and I3C, on the basis of the fact that they act differently
on the cell cycle. Both the cell cycle analysis and the combinatory
experiments indicate that the mode of action depends on the
concentration of 4MeOI3C.
Comparing the effects of the three indoles on the cell cycle
distribution in HCT 116 cells indicates that they act by different
mechanisms: I3C causes an accumulation in G /G phase, NI3C
0
1
accumulates the cells in G /M phases (17), and the present study
2
showed no change of cell phase distribution, when cells were
exposed to 100 μM 4MeOI3C, whereas a higher concentration
caused G /G phase accumulation. The major difference in the
0
1
response to I3C and NI3C was the fact that NI3C besides the p21
protein also induces the p27 protein (17). However, we performed
no analysis of the modulation of cell cycle related proteins by
4MeOI3C. The fact that 100 μM 4MeOI3C increased the doubl-
ing time by about 50%, combined with the observation that the
cell cycle phase distribution was not affected, indicates that
4MeOI3C at 100 μM has an overall-slowdown effect on the cell
cycle. Continued studies should reveal, whether cell cycle related
proteins are targets of 4MeOI3C during this slowdown.
4MeOI3C was continuously found to be about 2-fold as potent
as I3C on the cells investigated. However, only when tested at a
high concentration (200 μM) does 4MeOI3C seem to have the
same effect as I3C, in that cells are arrested in the G /G phase
0
1
or killed. Even if 4MeOI3C and I3C have different molecular
targets, the end point effects of both compounds are crucial to
the cell, as proliferation is inhibited, and the compounds are,
therefore, in agreement with the statements of Chou (23), most
likely at least partly dependent on the presense of each other. This
supports our use of the Combination Index equation for mutually
exclusive compound interaction, i.e., when the compounds are
not totally independent of each other.

8458 J. Agric. Food Chem., Vol. 58, No. 14, 2010
Kronbak et al.
In conclusion, 4MeOI3C was successfully synthesized and
in pancreatic cancer prevention: Effects of indole-3-carbinol (I3C)
and genistein. Anticancer Res. 2004, 24, 133–137.
purified. Treatment of the compound on human colon cancer
cell lines DLD-1 and HCT 116 in vitro resulted in inhibition of
their proliferation in a time- and concentration-dependent man-
ner. The compound was about 2-fold as potent as the structurally
familiar I3C in performing this inhibition. Combinatorial treat-
ment of the two compounds exhibited nearly additive to slightly
synergistic effects after 48 h. 4MeOI3C levels close to IC50
induced an overall-slowdown mechanism in HCT 116 cells.
However, when treated at the double concentration, 4MeOI3C
(
14) Weng, J. R.; Tsai, C. H.; Kulp, S. K.; Chen, C. S. Indole-3-carbinol
as a chemopreventive and anti-cancer agent. Cancer Lett. 2008, 262,
153–163.
(
15) Kim, Y. S.; Milner, J. A. Targets for indole-3-carbinol in cancer
prevention. J. Nutr. Biochem. 2005, 16, 65–73.
(
16) Aggarwal, B. B.; Ichikawa, H. Molecular targets and anticancer
potential of indole-3-carbinol and its derivatives. Cell Cycle 2005, 4,
1201–1215.
(
17) Neave, A. S.; Sarup, S. M.; Seidelin, M.; Duus, F.; Vang, O.
Characterisation of the N-methoxyindole-3-carbinol (NI3C)-
induced cell cycle arrest in human colon cancer cell lines. Toxicol.
Sci. 2005, 83, 126–135.
caused the cells to accumulate in the G /G1 phase, and the
0
cytotoxicity was increased dramatically, possibly through the
induction of apoptosis. Thus, the mode of action seems to be
concentration-dependent.
On the basis of our observations, we find 4MeOI3C to be
an interesting substance, which does not antagonize the anti-
cancer effect of I3C but contributes by being both cytostatic and
cytotoxic to cancer cells in vitro; however, we suggest future
investigations to elusidate the molecular fate in vivo and the role
of possible condensation products of the compound following
cruciferous vegetable consumption.
(18) Bonnesen, C.; Stephensen, P. U.; Andersen, O.; Sørensen, H.; Vang, O.
Modulation of cytochrome P-450 and glutathione S-transferase iso-
form expression in vivo by intact and degraded indolyl glucosinolates.
Nutr. Cancer 1999, 33, 178–187.
(
19) Pindur, U.; Witzel, H. The reactions of pindolol, mepindolol,
carazolol, and related model compounds with triethyl orthoformate:
pathways and products of a new colour reaction. Arch. Pharm.
(
Weinheim, Ger.) 1990, 323, 427–432.
(
20) Somei, M.; Yamada, F.; Kunimoto, M.; Kaneko, C. The chemistry
of indoles. 22. A practical one pot synthesis of 4-alkoxy-3-formy-
lindoles. Heterocycles 1984, 22, 797–801.
LITERATURE CITED
(
1) Higdon, J. V.; Delage, B.; Williams, D. E.; Dashwood, R. H.
Cruciferous vegetables and human cancer risk: epidemiologic evi-
dence and mechanistic basis. Pharmacol. Res. 2007, 55, 224–236.
2) Zareba, G.; Serradelf, N. Chemoprotective effects of broccoli and
other Brassica vegetables. Drugs Future 2004, 29, 1097–1104.
(21) Chou, T. C.; Talalay, P. Analysis of combined drug effects: A new
look at a very old problem. TIPS 1983, 4, 450–454.
(22) Chou, T. C.; Talalay, P. Quantitative analysis of dose-effect relation-
ships: the combined effects of multiple drugs or enzyme inhibitors.
Adv. Enzyme Regul. 1984, 22, 27–55.
(23) Chou, T. C. Theoretical basis, experimental design, and compute-
rized simulation of synergism and antagonism in drug combination
studies. Pharmacol. Rev. 2006, 58, 621–681.
(
(
3) Verhoeven, D. T. H.; Verhagen, H.; Goldbohm, R. A.; Van den Brandt,
P. A.; van Poppel, G. A review of mechanisms underlying anti-
carcinogenicity by Brassica vegetables. Chem.-Biol. Interact. 1997, 103,
7
9–129.
(24) IARC. IARC Handbooks of Cancer Prevention. Cruciferous Vege-
tables, Isothiocyanates and Indoles; IARC Press: Lyon, France,
2004; Vol. 9.
(25) Kushad, M. M.; Brown, A. F.; Kurilich, A. C.; Juvik, J. A.; Klein,
B. P.; Wallig, M. A.; Jeffery, E. H. Variation of glucosinolates in
vegetable crops of Brassica oleracea. J. Agric. Food Chem. 1999, 47,
1541–1548.
(26) Vang, O.; Mortensen, J.; Andersen, O. Biochemical effects of dietary
intake of different broccoli samples. II. Multivariate analysis of
contributions of specific glucosinolates in modulating cytochrome
P-450 and antioxidant defense enzyme activities. Metabolism 2001,
50, 1130–1135.
(27) Hanley, A. B.; Parsley, K. R.; Lewis, J. A.; Fenwick, G. R. Chemistry
of indole glucosinolates: Intermediacy of indole-3-ylmethyl isothio-
cyanates in the enzymatic hydrolysis of indole glucosinolates.
J. Chem. Soc. Perkin Trans. 1 1990, 2273–2276.
(
(
(
4) Buskov, S. H.; Sørensen, H.; Sørensen, J. C. Oligomeric compounds
formed under acidic processing of food containing hydroxybenzyl and
indol-3-ylmethylglucosinolates. Czech J. Food Sci. 2000, 18, 31–33.
5) Agerbirk, N.; De Vos, M.; Kim, J. H.; Jander, G. Indole glucosino-
late breakdown and its biological effects. Phytochem. Rev. 2009, 8,
1
01–120.
6) Nguyen, H. H.; Aronchik, I.; Brar, G. A.; Nguyen, D. H.; Bjeldanes,
L. F.; Firestone, G. L. The dietary phytochemical indole-3-carbinol
is a natural elastase enzymatic inhibitor that disrupts cyclin E protein
processing. Proc. Natl. Acad. Sci. U.S.A. 2008, 105, 19750–19755.
7) Sundar, S. N.; Kerekatte, V.; Equinozio, C. N.; Doan, V. B.;
Bjeldanes, L. F.; Firestone, G. L. Indole-3-carbinol selectively
uncouples expression and activity of estrogen receptor subtypes in
human breast cancer cells. Mol. Endocrinol. 2006, 20, 3070–3082.
8) Cover, C. M.; Hsieh, S. J.; Tran, S. H.; Hallden, G.; Kim, G. S.;
Bjeldanes, L. F.; Firestone, G. L. Indole-3-carbinol inhibits the
(
(
(28) Vang, O.; Frandsen, H.; Hansen, K. T.; Nielsen, J. B.; Andersen, O.
Modulation of drug-metabolising enzyme expression by condensa-
tion products of indole-3-ylcarbinol, an inducer in cruciferous
vegetables. Pharmacol. Toxicol. 1999, 84, 59–65.
1
expression of cyclin-dependent kinase-6 and induces a G cell cycle
arrest of human breast cancer cells independent of estrogen receptor
signaling. J. Biol. Chem. 1998, 273, 3838–3847.
(
9) Garikapaty, V. P.; Ashok, B. T.; Tadi, K.; Mittelman, A.; Tiwari,
R. K. 3,3 -Diindolylmethane downregulates pro-survival pathway
in hormone independent prostate cancer. Biochem. Biophys. Res.
Commun. 2006, 340, 718–725.
(29) Bradlow, H. L. Review. Indole-3-carbinol as a chemopro-
0
tective agent in breast and prostate cancer. In Vivo 2008, 22,
441–445.
(30) Boyd, L. A.; McCann, M. J.; Hashim, Y.; Bennett, R. N.; Gill, C. I.;
Rowland, I. R. Assessment of the anti-genotoxic, anti-proliferative,
and anti-metastatic potential of crude watercress extract in human
colon cancer cells. Nutr. Cancer 2006, 55, 232–241.
(31) Suzuki, C.; Ohnishi-Kameyama, M.; Sasaki, K.; Murata, T.;
Yoshida, M. Behavior of glucosinolates in pickling cruciferous
vegetables. J. Agric. Food Chem. 2006, 54, 9430–9436.
(32) Tian, Q.; Rosselot, R. A.; Schwartz, S. J. Quantitative determination
of intact glucosinolates in broccoli, broccoli sprouts, Brussels
sprouts, and cauliflower by high-performance liquid chromatogra-
phy-electrospray ionization-tandem mass spectrometry. Anal. Bio-
chem. 2005, 343, 93–99.
(
10) Hsu, J. C.; Dev, A.; Wing, A.; Brew, C. T.; Bjeldanes, L. F.;
Firestone, G. L. Indole-3-carbinol mediated cell cycle arrest of
LNCaP human prostate cancer cells requires the induced production
of activated p53 tumor suppressor protein. Biochem. Pharmacol.
006, 72, 1714–1723.
11) Chinni, S. R.; Li, Y. W.; Upadhyay, S.; Koppolu, P. K.; Sarkar, F. H.
Indole-3-carbinol (I3C) induced cell growth inhibition, G cell cycle arrest
and apoptosis in prostate cancer cells. Oncogene 2001, 20, 2927–2936.
12) Kim, D. S.; Jeong, Y. M.; Moon, S. I.; Kim, S. Y.; Kwon, S. B.; Park,
E. S.; Youn, S. W.; Park, K. C. Indole-3-carbinol enhances ultra-
violet B-induced apoptosis by sensitizing human melanoma cells.
Cell. Mol. Life Sci. 2006, 63, 2661–2668.
2
(
(
1
(33) Tiedink, H. G. M.; Davies, J. A.; Van Broekhoven, L. W.; van der
Kamp, H. J.; Jongen, W. M. Formation of mutagenic N-nitroso
(
13) Lian, J. P.; Word, B.; Taylor, S.; Hammons, G. J.; Lyn-Cook, B. D.
Modulation of the constitutive activated STAT3 transcription factor
compounds in vegetable extracts upon nitrite treatment:
a

Article
J. Agric. Food Chem., Vol. 58, No. 14, 2010 8459
comparison with the glucosinolate content. Food Chem. Toxicol.
988, 26, 947–954.
(35) Verkerk, R.; Dekker, M.; Jongen, W. M. F. Post-harvest increase of
indolyl glucosinolates in response to chopping and storage of
Brassica vegetables. J. Sci. Food Agric. 2001, 81, 953–958.
1
(
34) Nilsson, J.; Olsson, K.; Engqvist, G.; Ekvall, J.; Olsson, M.; Nyman,
M.; Akesson, B. Variation in the content of glucosinolates, hydro-
xycinnamic acids, carotenoids, total antioxidant capacity and low-
molecular-weight carbohydrates in Brassica vegetables. J. Sci. Food
Agric. 2006, 86, 528–538.
Received for review May 11, 2010. Revised manuscript received June 15,
2010. Accepted June 21, 2010. During this study, a scholarship was given
by The Danish Cancer Society to R.K., for which we are highly grateful.