From: "Saved by Windows Internet Explorer 8" Subject: Indole-3-Carbinol and Diindolylmethane Induce Apoptosis of Human Cervical Cancer Cells and in Murine HPV16-Transgenic Preneoplastic Cervical Epithelium Date: Wed, 8 Jun 2011 14:51:04 -0700 MIME-Version: 1.0 Content-Type: multipart/related; type="text/html"; boundary="----=_NextPart_000_0020_01CC25EB.7D9D39E0" X-MimeOLE: Produced By Microsoft MimeOLE V6.1.7600.16776 This is a multi-part message in MIME format. ------=_NextPart_000_0020_01CC25EB.7D9D39E0 Content-Type: text/html; charset="utf-8" Content-Transfer-Encoding: quoted-printable Content-Location: http://jn.nutrition.org/content/131/12/3294.full =EF=BB=BF
* ; North Shore-Long Island Jewish Research Institute, = Manhasset, NY 11030 and Department of Otolaryngology, Long Island Jewish = Medical=20 Center, The Long Island Campus of Albert Einstein College of Medicine, = New Hyde=20 Park, NY 11030; and; =E2=80=A0 ; Department of = Biological=20 Sciences, St. John=E2=80=99s University, Jamaica, NY 11439
3To whom correspondence = should be=20 addressed. E-mail: carterth{at}aol.com.
Dietary indole-3-carbinol (I3C) has clinical benefits for both = cervical=20 cancer and laryngeal papillomatosis, and causes apoptosis of breast = cancer cells=20 in vitro. We asked whether I3C and its major acid-catalyzed condensation = product=20 diindolylmethane (DIM), which is produced in the stomach after = consumption of=20 cruciferous vegetables, could induce apoptosis of cervical cancer cell = lines. We=20 also asked whether this effect could be observed in vivo. In vitro, both = I3C and=20 DIM caused accumulation of DNA strand breaks in three cervical cancer = cell=20 lines. Induction of apoptosis was confirmed by nuclear morphology, = nucleosome=20 leakage, altered cytoplasmic membrane permeability and caspase 3 = activation.=20 Neither I3C nor DIM caused apoptotic changes in normal human = keratinocytes. In=20 C33A cervical cancer cells, DIM was more potent than I3C [dose at which = the=20 number of viable cells was 50% of that in untreated cultures = (LD50) =3D=20 50=E2=80=9360 =C2=B5mol/L for DIM and 200 =C2=B5mol/L = for I3C in a mitochondrial=20 function assay] and faster acting. Furthermore, I3C reduced Bcl-2 = protein in a=20 time- and dose-dependent manner. In HPV16-transgenic mice, which develop = cervical cancer after chronic estradiol exposure, apoptotic cells were = detected=20 in cervical epithelium by TdT-mediated dUTP nick-end labeling staining = and by=20 immunohistochemical staining of active caspase 3 only in mice exposed to = 17=C3=9F-estradiol (E2) and fed I3C. Rare apoptotic cells = were also=20 observed by hematoxylin and eosin staining in the spinous layer of the = cervical=20 epithelium in both control and transgenic mice. Estradiol reduced the = percentage=20 of these late-stage apoptotic cells in the cervical epithelium of = transgenic,=20 E2-treated mice, but this reduction was prevented by I3C. = These data=20 confirm the proapoptotic action of I3C on transformed cells in vitro, = extend the=20 observations to cervical cancer cells and to DIM and show for the first = time=20 that dietary I3C results in increased apoptosis in target tissues in = vivo.=20
KEY WORDS: =E2=80=A2 cervical cancer =E2=80=A2 = indole-3-carbinol =E2=80=A2=20 diindolylmethane =E2=80=A2 apoptosis =E2=80=A2 mice
Indole-3-carbinol (I3C),4=
a common=20
phytochemical in the human diet, is present in all members of the =
cruciferous=20
vegetable family, which includes cabbage, broccoli, Brussels sprouts,=20
cauliflower and kale. Recently, it has become clear that I3C has the =
potential=20
to prevent and even to treat a number of common cancers, especially =
those that=20
are estrogen-related. Pure I3C taken as a dietary supplement (the =
equivalent of=20
one third of a head of cabbage per day) reverses precancerous changes in =
women=20
with stage II and stage III cervical dysplasia (1
). A=20
diet rich in cruciferous vegetables (2
) or=20
supplements of I3C (3
) causes=20
regression of tumors or decreases their rate of growth or recurrence in =
two=20
thirds of patients with recurrent laryngeal papillomatosis. Clinical =
trials are=20
planned to test the efficacy of I3C as a preventive treatment for breast =
cancer=20
(4
). Laboratory studies suggest that this phytochemical =
can act in=20
several different ways to prevent transformation and/or tumor =
progression, as=20
well as to kill transformed cells selectively.
I3C is rapidly converted in the stomach to a variety of condensation=20
products, chiefly diindolylmethane (DIM) (5
).=20
Plasma from humans and rats fed I3C contains no detectable I3C, but =
large=20
amounts of DIM, as well as other metabolites, some of which remain=20
uncharacterized (6
; L.=20
Bjeldanes, University of California at Berkeley, personal =
communication). Thus=20
DIM, rather than I3C, is probably the major compound initially available =
to=20
cells after ingestion of I3C. I3C is also converted slowly to DIM at =
neutral pH=20
(5
), with the result that either compound is active in =
vitro. For=20
example, both I3C and DIM induce apoptosis in MCF-7 breast carcinoma =
cells=20
growing in culture (7
,8
).
We showed previously that dietary I3C prevents the appearance of =
cervical=20
cancer after chronic estrogen exposure in transgenic mice expressing the =
type 16=20
human papillomavirus (HPV) oncogenes (9
). This=20
effect is accompanied by a shift in estrogen metabolism to favor the =
production=20
of 2-OH estrone rather than the 16=CE=B1-OH metabolite, which is =
associated with=20
prolonged estrogenic activity and carcinogenesis (our unpublished =
data)(10
,11
). Thus,=20
it is likely that one of the major pathways by which I3C and its =
derivatives=20
prevent the onset of cervical cancer involves alteration of estrogen =
metabolism=20
by inducing specific cytochrome P450 isoforms via the aryl=20
hydrocarbon receptor (9
,10
) for=20
which DIM is a weak ligand (12
,13
).
However, we observed that estradiol and I3C act in opposing ways on =
the=20
balance of proliferation of cervical cancer cells (unpublished data), =
and others=20
have reported that the inhibitory effect of I3C on MCF-7 breast cancer =
cells is=20
independent of estrogen signaling (14
).=20
Therefore, at least in vitro, there must be an additional, =
estrogen-independent=20
pathway by which I3C interferes with the establishment and progression =
of=20
malignancy. Both I3C and DIM induce apoptotic changes in breast cancer =
cells in=20
vitro (7
,8
).=20
However, the published clinical data on therapeutic use of I3C involve =
laryngeal=20
papillomas and cervical cancer, which are both HPV-associated diseases. =
It is=20
therefore important to determine whether I3C and DIM can induce =
apoptosis in a=20
variety of cervical cancer cell lines, including those with and without =
HPV=20
genes, and whether apoptotic changes occur in target tissues in vivo. We =
looked=20
for apoptotic cells in dysplastic cervical epithelium of HPV16 =
transgenic mice=20
fed I3C under conditions that would otherwise lead to the development of =
cervical neoplasia. Furthermore, we asked whether normal cells in =
culture, as=20
well as normal cervical tissue, are affected by I3C/DIM. Our results =
indicate=20
that the ability of I3C and DIM to induce apoptosis extends to a variety =
of=20
cervical cancer cell lines and to preneoplastic cervical epithelium. =
This effect=20
appears to be specific for transformed cells, raising the possibility =
that these=20
dietary phytochemicals may be generally useful as therapeutic agents. =
Reagents.
17=C3=9F-Estradiol (E2) and I3C were purchased from Sigma = (St. Louis,=20 MO). Diindolylmethane was a gift from Dr. M. Zeligs, BioResponse, = Boulder, CO.=20
Cell lines and cell culture.
The cervical cancer cell lines CaSki (containing multiple copies of=20
integrated HPV16 DNA) and C33A (HPV negative, mutant p53) were obtained =
from the=20
American Type Culture Collection (ATCC, Manassas, VA). C33AE6 cells =
(15
) were=20
stably transfected with the HPV16 E6 gene (pLXSN16E6 from D. Galloway, =
Fred=20
Hutchinson Cancer Research Center, Seattle, WA) and express low levels =
of E6=20
transcripts (15
). All=20
cells were maintained as monolayer cultures at 37=C2=B0C, 7% =
CO2. Cervical=20
cancer cells and 3T3 fibroblasts were grown in Dulbecco=E2=80=99s =
modified Eagle=E2=80=99s=20
medium (DMEM) containing 4.5 g/L glucose and bicarbonate (GIBCO-BRL,=20
Gaithersburg, MD), supplemented with 110 mg/L sodium pyruvate, 200 =
mmol/L=20
glutamine, 100 mL/L fetal bovine serum, and 1 x 105 U/L each of penicillin and =
streptomycin. Normal=20
human foreskin keratinocytes were grown in F12-DMEM on feeder layers by =
the=20
method of Rheinwald and Green (16
) as=20
described previously (16
,17
).
Mice.
K14-HPV16 transgenic mice were derived and described by Arbeit et al. =
(18
), and=20
were characterized and maintained by us as described previously (9
,17
).=20
Control and experimental groups were as described previously (9
).=20
Virgin normal and transgenic mice (4=E2=80=935 wk old) were implanted =
subcutaneously=20
with 0.25 mg/d release pellets of E2 and fed diets with or =
without=20
I3C as described below. Implants were repeated every 60 d until the end =
of the=20
study. Mice were housed in groups of 5/cage. All experiments involving =
mice were=20
done in strict adherence to IACUC-approved procedures.
Diet.
Mice consumed ad libitum the AIN76a diet or AIN76a diet (19
)=20
enriched with 0.1 g/kg I3C. Diets were prepared by Ziegler (Gardner, =
PA). The=20
AIN76a diet contains 5% corn oil and supplies a total of 18.5 MJ/kg, =
with 22% of=20
energy from protein, 11% from fat and 67% from carbohydrates.
Cell viability.
C33A Cells were trypsinized, seeded at 104 cells/well in = 96-well=20 plates containing 100 =C2=B5L medium/well and incubated over = night. The next=20 day, the medium was changed to 200 =C2=B5L containing either = DIM or dimethyl=20 sulfoxide (DMSO) as solvent control, employing a minimum of four = replicate wells=20 per condition. Viability was determined at indicated times by a = mitochondrial=20 function assay [reduction of=20 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)= -2H-tetrazolium=20 (MTS)] using the Cell Titer Aqueous One kit (Promega, Madison, WI) = according to=20 the manufacturer=E2=80=99s instructions. Absorbance at 595 nm of the = solution in=20 individual wells was determined with a multiwell plate reader. Data were = analyzed by plotting the mean and SD of=20 cell viability vs. DIM concentration. Protein concentration was measured = with=20 the MicroBCA kit (Pierce, Rockford, IL) using a bovine serum albumin = (BSA)=20 standard.
Nucleosomal leakage apoptosis assay.
Nucleosomal leakage was monitored with a Cell Death Detection=20 ELISAPLUS kit from Roche Molecular Biochemicals (Mannheim, = Germany)=20 that detects histones and DNA in cytoplasmic extracts. Cells were grown = in=20 96-well plates as described above in replicates of 6 wells/condition. = Cells in=20 duplicate wells were lysed and the postnuclear supernatant solution was = analyzed=20 for nucleosomal leakage according to manufacturer=E2=80=99s = instructions. Results were=20 determined by measuring absorbance at 405 nm with a microwell plate = reader as=20 above. The second set of 4 wells from the same plate was analyzed for = viability=20 using the MTS assay by absorbance at 595 nm, and the ELISA results were=20 normalized to this value (A405/A595) to correct = for=20 reduction in viable cell number after treatment with I3C and DIM. The = mean and=20 SD of the normalized data were = plotted=20 vs. DIM concentration.
Western blotting.
Cells treated with I3C, or vehicle controls, each with or without = estradiol,=20 were lysed at room temperature in buffer containing 10 mmol/L=20 NaH2PO4, 20 g/L triton X-100, 12g/L SDS, 10g/L = sodium=20 deoxycholate supplemented just before use with 2 =C2=B5mol/L = aprotinin, 100=20 =C2=B5mol/L phenylmethylsulfonyl fluoride and 0.1 mmol/L EDTA, = boiled for 2=20 min, and centrifuged for 10 min at 12,000 x g at 4=C2=B0C. = Supernatant=20 solutions were stored at -80=C2=B0C until use. Extract protein (100 = =C2=B5g) in=20 sample buffer (125 mmol/L Tris-HCl, pH6.8, 10 g/LSDS, 20 g/L = =C3=9F-mercaptoethanol=20 and 0.1g/L bromophenol blue) was loaded onto a 12% SDS-polyacrylamide = gel. After=20 electrophoresis at 32 V for 3 h at room temperature, protein bands were=20 transferred to an Immobilon-P membrane (Millipore, Bedford, MA) by=20 electroblotting overnight in transfer buffer (192 mmol/L glycine, 25 = mmol/L Tris=20 and 200 g/L methanol). Before incubation with antibodies, the membrane = was=20 blocked with Tris buffered saline with Tween (TBST)/milk (20mmol/L = Tris-HCl,=20 137mmol/L NaCl, 15g/L nonfat dry milk and 1g/L Tween20, pH7.6) for 1 h. = BCL-2,=20 BAX and =CE=B1-tubulin were detected with specific mouse monoclonal = antibodies (Santa=20 Cruz Biotechnology, Santa Cruz, CA) by incubating dilutions = (1:500=E2=80=931:1000) for 1=20 h. After washing in TBST/milk, the filters were incubated with = horesradish=20 peroxidase (HRP)-conjugated anti-mouse immunoglobulin G antibody (Santa = Cruz) at=20 1:2000 dilution for 1 h at room temperature. Antibody bound to protein = was=20 detected using the enhanced chemiluminescence system (Amersham Life = Science,=20 Piscataway, NJ).
TdT-mediated dUTP nick-end labeling (TUNEL)=20 assay.
The assay used the in situ cell death detection kit, peroxidase = (POD), from=20 Boehringer Mannheim (Indianapolis, IN). Cells were grown for 24 h in = 8-well=20 chamber slides seeded with 105 cells/well, treated with 200=20 =C2=B5mol/L I3C and incubated at 37=C2=B0C for 48h. The slides = were washed in PBS=20 and fixed with 40 g/L paraformaldehyde for 30 min at room temperature. = Fixed=20 cells were washed in PBS, permeabilized with sodium citrate buffer = containing 1=20 g/L Triton X-100 for 2 min on ice, and then incubated with terminal=20 deoxynucleotidyl transferase for 1 h at 37=C2=B0C. After being rinsed = with PBS,=20 slides were treated with converter-POD (conjugated with HRP) at = 37=C2=B0C for 30 min=20 and mounted with a glass cover slip. At least 200 cells/well were = evaluated for=20 staining.
Assessment of DNA fragmentation by gel=20 electrophoresis.
Cells were grown for 48 h in the presence of various concentrations = of I3C or=20 DIM. After treatment, the cells were harvested, centrifuged at 500 x g for 5 min = and washed=20 with PBS. The cell pellet was lysed in 200 =C2=B5L of lysis = buffer=20 containing 50 mmol/L Tris-HCl, 20 mmol/L EDTA and 10 g/L NP-40, and = centrifuged=20 at 1600 x = g for 5=20 min. The supernatant solution was incubated with (5 mg/L) RNase A for 2 = h at=20 56=C2=B0C and then digested with proteinase K (2.5 mg/L) at 37=C2=B0C = for 16 h. DNA was=20 precipitated with an equal volume of 10 mol/L ammonium acetate and 2.5 = volumes=20 of ethanol. The precipitates were rinsed with 700 g/L ethanol, air = dried,=20 dissolved in Tris/EDTA buffer (10 mmol/L Tris-HCl, pH 7.5, 1 mmol/L = EDTA),=20 electrophoresed through a 1.5% agarose gel, stained with ethidium = bromide and=20 photographed on a UV transilluminator.
Immunohistochemistry.
Tissues were procured, fixed and processed for immunostaining as =
described=20
(9
). For detection of activated caspase 3, both cells in =
culture and=20
tissue slices were fixed in 10 g/L paraformaldehyde for 60 min at room=20
temperature, followed by three washes in PBS. After being blocked =
overnight with=20
150 g/L BSA in PBS, the cells or tissue slices were incubated with a =
polyclonal=20
antibody specific for the activate form of caspase 3 (Promega) overnight =
at 4=C2=B0C.=20
After three washes in PBS, samples were incubated with =
peroxidase-conjugated=20
goat anti-rabbit second antibody (Santa Cruz Biochemicals); the signal =
was=20
developed according to the manufacturer=E2=80=99s instructions. For =
TUNEL staining of=20
tissue sections, paraffin-embedded tissues were sectioned at a thickness =
of 5=20
=C2=B5m and processed using the Complete ApopTag in situ =
hybridization kit=20
(Intergen, Purchase, NY). Hematoxylin and eosin (H&E) staining was=20
accomplished as described (9
).
Cytologic detection of apoptosis.
Two groups of mice (n =3D 50 transgenic and 50 =
nontransgenic) were=20
treated continuously with E2 starting at 4=E2=80=935 wk of =
age. Half of each=20
group was fed a diet containing 2 g/kg I3C and the other half a control =
diet=20
without I3C. At 6 mo, when the majority of E2-treated =
transgenic mice=20
fed the control diet were found to have cervical tumors (9
), all=20
mice were killed by CO2 inhalation and their cervical tissue =
examined=20
by H&E staining as described (9
).=20
Samples were coded and the determination of apoptotic cells was carried =
out by a=20
single individual unaware of treatment groups (M.Q.).
Fluorescence staining of nuclei.
C33A cells and normal human keratinocytes growing in monolayer were = fixed in=20 paraformaldehyde as above, followed by 10 =C2=B5g/L=20 4,6-diamidino-2-phenylindole (DAPI) in methanol (Boehringer Mannheim, = Germany)=20 for 30 min at 37=C2=B0C. Stained cells were mounted with Aqueous = Mounting Medium=20 (Biomedia, Loomis, CA) before fluorescence microscopy.
Fluorescence-activated cell sorting = analysis.
To determine altered permeability, C33A cells were treated with or = without=20 200 =C2=B5mol/L I3C for 48 h and trypsinized. After washing = with PBS, the=20 cells were incubated in PBS containing 8 mg/L of 7-amino actinomycin D=20 (7-AAD)-fluorescein isothiocyanate (Sigma) in the dark at 4=C2=B0C for = 20 min and=20 then washed in PBS + 150 g/L BSA + 0.2 g/L NaN3 containing 20 = mg/L of=20 nonfluorescent actinomycin D (AD; Sigma). The resulting cell suspension = was then=20 analyzed cytofluorometrically on a Coulter Elite flow cytometer, set for = single=20 color, ungated fluorescence intensity.
Statistical analysis of data.
Standard deviations were calculated for all quantitative data as = indicated in=20 the figures and figure legends. Significant differences (P < = 0.05)=20 were determined using Student=E2=80=99s t test.
I3C and DIM induce DNA fragmentation in cervical cancer=20 cells.
After treatment with either I3C or DIM, DNA from I3C-treated C33A =
cervical=20
cancer cells accumulated double-strand breaks that generated the =
characteristic=20
apoptotic pattern of nucleosomal "laddering" (Fig.=20
1A
, lanes=20
1=E2=80=935), and the effect was dose dependent because an equal amount =
of DNA from=20
cells treated with increasing concentrations of I3C gave progressively =
more=20
intense patterns of nucleosomal repeat-sized DNA fragments. A similar =
apoptotic=20
pattern was obtained when cells were treated with DIM (Fig. =
1B
). In=20
contrast, nontransformed cells such as 3T3 fibroblasts, did not =
accumulate=20
double-strand DNA breaks after DIM treatment (Fig. 1A
, lanes=20
6=E2=80=9310).
We next used TUNEL to determine the fraction of cells undergoing =
apoptosis.=20
I3C treatment substantially increased the fraction of apoptotic cells in =
each of=20
three cell lines (C33A, C33AE6 and CaSki), from a 100% increase with =
CaSki cells=20
to a >200% increase with C33A (Fig. 2
).=20
C33AE6, which express theHPV16 E6 oncogene (15
),=20
showed an intermediate response (Fig. 2)
. Thus,=20
it appeared that the induction of DNA strand breaks by I3C was a general =
phenomenon for cervical cancer cell lines and was independent of viral =
gene=20
expression because C33A cells do not express detectable viral gene =
products=20
(20
), whereas both CaSki and C33AE6 do (15
,21
).
Cell death induced by I3C in C33A cervical cancer cells = has the=20 characteristics of apoptosis.
If the DNA fragmentation caused by I3C were the result of late-stage=20
apoptosis, then it should be possible to detect early apoptotic changes =
in these=20
cells. We next determined whether I3C could cause the cytoplasmic =
membrane=20
changes characteristic of this process. Extroversion of =
phosphatidylserine=20
during apoptosis (22
) is=20
accompanied by altered membrane permeability such that a fluorescent =
derivative=20
of AD, namely, 7-AAD, can enter apoptotic cells and bind to DNA in the =
nucleus=20
(23
,24
).=20
Fluorescence-activated cell sorting (FACS) analysis of =
7-AAD=E2=80=93stained C33A cells=20
that had been treated with 200 =C2=B5mol/L I3C confirmed that =
the compound=20
increased the percentage of stained cells from 9.8% of the total =
(Fig.=20
3A
) to 55%=20
of the total (Fig. 3B
) after=20
48 h of exposure to I3C.
DIM is a more effective inducer of apoptosis than=20 I3C.
If the DIM, the dimeric adduct of I3C, were the active molecular =
form, or=20
more proximate to the active form, then DIM would be expected to act =
faster and=20
at a lower concentration than I3C because the latter is converted only =
slowly to=20
DIM in cell culture. Figure 4A
shows=20
the effect of increasing concentrations of DIM and I3C on C33A cells =
after a=20
48-h exposure, using a mitochondrial function assay (MTS assay) as an =
indirect=20
measure of cell viability. We observed a reduction in the number of =
viable cells=20
in cultures treated with as little as 40 =C2=B5mol/L DIM, =
whereas I3C had no=20
observable effect at concentrations <100 =C2=B5mol/L (Fig. =
4A
). The=20
dose at which the number of viable cells is 50% of that in untreated =
cultures=20
(LD50) for DIM was =E2=88=BC60 =C2=B5mol/L, compared =
with =E2=88=BC200=20
=C2=B5mol/L for I3C. We next compared the rate of cell killing =
by DIM and=20
I3C; 100 =C2=B5mol/L DIM began to have an observable effect =
between 16 and=20
20 h after addition to growing cells, whereas 300 =C2=B5mol/L =
I3C had no=20
observable effect until =E2=88=BC36 h (Fig. 4B
). To=20
confirm that DIM was in fact inducing apoptosis, the dose-response =
experiment in=20
Figure 4A
was=20
repeated, and ELISA was used to detect nucleosomal leakage into the =
cytoplasm.=20
Histones and DNA were detected in the cytoplasm of cells treated with 75 =
=C2=B5mol/L DIM, whereas release in I3C-treated cells required=20
concentrations several fold higher (Fig. 4C
).
Two additional hallmarks of apoptotic cell death are nuclear =
condensation and=20
activation of caspases. Figure 5
shows an=20
immunofluorescence photomicrograph of DAPI-stained C33A cells =
(panels A=20
and B) and normal human keratinocytes (panels C and=20
D) treated with DIM. Because DAPI stains only DNA and =
chromatin, the=20
outline of the stained region indicates the relative size and shape of =
the=20
nucleus, as well as the distribution of DNA within it. At concentrations =
as low=20
as 30 =C2=B5mol/L, DIM caused observable changes in the nuclear =
morphology=20
of virtually all of the C33A cells (data not shown); at 60 =
=C2=B5mol/L DIM,=20
condensed, fragmented late-stage apoptotic nuclei were observed in C33A =
cells=20
and virtually all of the nuclei were contracted compared with those in =
untreated=20
cells (panel B, white arrows). At 100 =C2=B5mol/L, few =
C33A cells=20
remained on the cover slip after staining (not shown). In contrast, =
normal=20
keratinocytes did not exhibit these changes even at 100 =
=C2=B5mol/L DIM=20
(compare Fig. 5C
and=20
D
). When duplicate cultures of C33A cells were stained =
for=20
activated caspase 3 by immunohistochemistry, a similar dose response was =
observed (Fig. 6
).
Bcl-2 is reduced in cells treated with I3C and=20 DIM.
One of the points at which several pathways of apoptotic induction =
converge=20
is the breakdown of mitochondrial membrane integrity. This process is =
thought to=20
be controlled in part by the relative abundance of various members of =
the Bcl-2=20
family of proteins, notably BCL-2 itself and BAX, the former acting as =
an=20
antiapoptotic agent and the latter as an apoptotic inducer [reviewed in =
(15
)]. I3C=20
caused a time- and dose-dependent reduction in the amount of Bcl-2 =
protein=20
detected by Western blot (Fig. 7A
and=20
B), whereas BAX was not affected by treatment with 300 =
=C2=B5mol/L=20
I3C for 72 h. DIM reduced the amount of BCL-2 to undetectable levels by =
72 h.=20
Indole-3-carbinol increases apoptosis in cervical = epithelium of=20 HPV16 transgenic mice.
The observation that I3C and its major condensation product DIM can =
induce=20
apoptotic changes in cervical cancer cells in vitro led us to ask =
whether I3C=20
had the same effect in vivo. Transgenic mice expressing the human HPV E6 =
and E7=20
oncogenes under control of the keratin 14 promoter all develop cervical =
cancer=20
when exposed chronically to estradiol (9
,18
), but a=20
diet supplemented with I3C protects nearly all of these mice (9
). We=20
examined sections of cervical epithelium from normal and transgenic mice =
by=20
H&E staining in a double-blind protocol, scoring for apoptotic cell =
nuclei=20
(Fig. 8A
and=20
B
). A small number of apoptotic nuclei were observed =
throughout the=20
cervical epithelium of normal mice, but this number did not change =
significantly=20
in mice treated with E2 (Fig. 8C)
,=20
although the E2-treated mice showed increased cervical =
dysplasia (9
).=20
Transgenic mice exposed to estradiol, on the other hand, had =
significantly fewer=20
apoptotic cells (Fig. 8C
). When=20
these estradiol-treated transgenic mice were also fed I3C, the number of =
apoptotic cells detected in H&E stained sections returned to normal. =
We next used TUNEL as an alternative method to detect apoptotic cells =
in=20
sections of cervical tissue. Figure 9
shows=20
clear evidence of stained cells in cervical epithelium only from =
transgenic mice=20
chronically exposed to estrogen, i.e., those expected to develop =
cervical cancer=20
but also fed I3C. We did not observe TUNEL staining cells in the =
cervical=20
epithelium of normal mice or in transgenic mice exposed to amounts of =
estrogen=20
that induce epithelial hypertrophy and dysplasia.
Finally, we asked whether activation of caspase 3 could be detected =
in the=20
cervical epithelium of the transgenic mice fed I3C. Using the=20
immunohistochemical staining method described above for Figure =
6C
,=20
individual cells stained positively for active caspase 3 in the =
suprabasal=20
epithelium of the transgenic mice fed I3C (Fig. =
10B
), but=20
not of placebo-fed controls (Fig. 10D
).
The results described in this paper support the conclusion that both =
I3C and=20
DIM induce apoptotic changes specifically in transformed cervical cells. =
Many of=20
the hallmarks of apoptosis (25
) are=20
present in I3C- and DIM-treated cells growing in culture, i.e., altered =
nuclear=20
morphology (Fig. 5)
, DNA=20
fragmentation (Figs. 1
, 2)
and=20
nucleosomal leakage (Fig. 4C)
,=20
changes in membrane permeability (Fig. 3)
and=20
activation of caspase 3 (Fig. 6)
. The=20
quantitative data from the various assays used to detect cell killing =
and=20
apoptosis were in good agreement. For example, exposure to 300 =
=C2=B5mol/L=20
I3C for 48 h caused 35% of the cells to stain positively for DNA strand =
breaks=20
in the TUNEL assay (a late apoptotic event) and induced membrane =
permeability=20
changes (an early event) in 55%. Similarly, 100 =C2=B5mol/L DIM =
for 48 h=20
killed between 80 and 90% of the cells (Fig. 4A
and=20
B
) and active caspase 3 was visible by immunostaining in =
roughly=20
half of the surviving cells (Fig. 6)
. In=20
contrast to cervical cancer cells, normal human foreskin keratinocytes, =
cells=20
that are similar (although clearly not identical) to those in the cervix =
that=20
most commonly undergo malignant transformation, did not exhibit major =
changes in=20
nuclear morphology associated with apoptosis upon exposure to DIM (Fig. =
5
,=20
panels C and D).
The dose-response curves for I3C and DIM (Fig. 4)
are=20
consistent with slow conversion of I3C to DIM in cell culture medium, =
assuming=20
that DIM is the active compound or is further metabolically converted to =
the=20
active compound(s). However, serum levels of DIM in rats fed I3C were =
only =E2=88=BC10%=20
of those required to induce apoptosis in vitro in our experiments (6
). The=20
fact that relatively high concentrations of DIM (and I3C), pharmacologic =
as=20
opposed to dietary levels, were required to obtain biologic effects in =
vitro is=20
consistent with a requirement for metabolic conversion of DIM to a =
secondary=20
active compound. However, it is now widely recognized that serum levels =
of=20
bioactive compounds are often uninformative because other factors such =
as=20
carrier proteins, intracellular accumulation and localized concentration =
in=20
specific tissues may all come into play. It is therefore not surprising =
that the=20
concentrations of DIM required to produce effects in vitro are higher =
than those=20
normally attainable in vivo. Similar observations have been made for =
other=20
bioactive natural products, for example isoflavones from soy (26
,27
). In=20
any case, the fact that dietary I3C appeared to induce apoptosis in the =
cervical=20
epithelium of HPV16 transgenic mice suggests that this compound or its =
active=20
metabolites do indeed reach effective intracellular levels in vivo.
Detection of apoptotic cells in cervical epithelia was undoubtedly=20
facilitated by our experimental model system. In the HPV transgenic =
mouse, just=20
as in human cervical cancer, nearly every cell is expressing HPV =
oncogenes and=20
is thus initiated for transformation. If, as our results suggest, =
transformed=20
cells or cells undergoing preneoplastic conversion are differentially =
sensitive=20
to I3C/DIM, then the enrichment for these cells in cervical epithelium =
as a=20
result of VP6 and VP7 gene expression would explain our ability to =
detect what=20
in other tissues or experimental systems would be a rare event. In =
cervical=20
epithelium from HPV16 mice, chronic exposure to elevated E2 =
is=20
required for expression of the transformed phenotype (9
,18
). The=20
observed reduction in late-stage apoptotic cells identifiable by routine =
histopatholgy in samples from HPV16 mice exposed to E2 (Fig.=20
8C
) is therefore consistent with the generally observed =
phenomenon=20
that apoptosis is inhibited in neoplastic and preneoplastic cells =
[reviewed in=20
(25
)]. By contrast, when E2-treated HPV =
transgenic mice=20
were also fed I3C, we did not observe the E2-related =
reduction in=20
apoptotic cells. This suggests that I3C either prevented transformation =
or=20
killed the cells as they became transformed. Our in vitro data are =
consistent=20
with the latter explanation.
We do not yet know either the mechanism by which these phytochemicals =
induce=20
apoptosis or what determines their apparent specificity for transformed =
cells.=20
Our data and published reports from other laboratories rule out both HPV =
and=20
estrogen effects as requirements for induction of apoptosis by I3C/DIM, =
although=20
both mechanisms may play ancillary roles in specific instances. Apart =
from sex,=20
two differences of potential relevance between cervical cancer cells, =
which are=20
sensitive to DIM, and foreskin keratinocytes, which are not, are the =
presence of=20
HPV oncogenes and estrogen receptors in the former. However C33A cells, =
unlike=20
C33AE6 and CaSki cells, do not express HPV oncogenes or any other viral =
genes at=20
detectable levels (15
);=20
therefore, it is unlikely that viral gene products account for the =
sensitivity=20
of cervical cancer cells to I3C and DIM. The reduced sensitivity of =
C33AE6 cells=20
to killing by I3C compared with the parental C33A cell line (Fig. 2)
may have=20
been caused by the reported antiapoptotic effect of E6 (28
=E2=80=9330
).=20
However, even the protection afforded by expression of this viral =
oncogene was=20
not complete.
Cell killing by DIM or I3C similarly does not seem to require =
estrogen; in=20
fact, I3C and E2 have opposing effects on cell survival. =
Although=20
C33A cells possess estrogen receptors (our unpublished data), and =
estrogen=20
induces proliferation of epithelial cells in the cervix, estradiol =
interferes=20
with induction of apoptosis caused by a number of agents, including I3C =
(our=20
unpublished observations). Conversely, I3C interferes with signaling =
from the=20
estrogen receptor (31
). The=20
cytotoxic specificity of I3C/DIM for transformed cells also appears to =
be=20
reflected in cervical epithelium in vivo, regardless of whether mice =
were=20
exposed chronically to elevated estrogen (Figs. 8
9
10)
. During=20
the course of multiple studies spanning >2 y, we have never seen =
apoptotic=20
cells in cervical epithelium or any evidence of cervical histopathology =
in=20
normal mice fed a diet containing DIM. We conclude that induction of =
apoptosis=20
in cervical cancer cells exposed to I3C or DIM both in vitro and in vivo =
occurs=20
by a pathway separate from, and independent of estrogen-responsive =
mechanisms,=20
and that it is specific for transformed cells.
How does I3C/DIM induce apoptosis of cancer cells? Several =
possibilities are=20
suggested by published work from our laboratory and others. One =
mechanism that=20
could account for the sensitivity of transformed cells is cell cycle =
inhibition.=20
I3C inhibits cdk6 expression in MCF-7 cells (32
) by=20
interfering with transcription (33
). Thus=20
it might be that proliferating cells are more sensitive to I3C and DIM =
due to a=20
need for cdk6 activation, which these agents prevent. Another =
possibility=20
involves the induction of proapoptotic genes, for example, via the aryl=20
hydrocarbon receptor (12
,13
) or=20
perhaps the sensitization of cells to the cytotoxic effects of cytokines =
or=20
other factors in the cellular microenvironment.
Our results provide further mechanistic underpinning for the = epidemiologic=20 data supporting the efficacy of a diet rich in cruciferous vegetables = for=20 reducing the incidence of cervical cancer. However, this is also one of = the=20 first reports of apoptotic changes induced by chemotherapeutic agents in = vivo,=20 particularly involving solid tumors. The fact that normal keratinocytes = did not=20 show evidence of apoptotic morphologic changes after prolonged exposure = to high=20 DIM concentrations, and that normal cervical epithelium never showed any = indication of apoptosis extensive enough to score positively in TUNEL = staining,=20 suggests that I3C/DIM may have a utility beyond dietary prevention, as a = relatively innocuous chemotherapeutic agent in the treatment of cervical = cancer=20 and, potentially, other malignancies.
1 Presented in part =
at the=20
scientific session of the 18th International Papillomavirus Meeting, =
June 2000,=20
Barcelona, Spain [Chen, D.-Z., Carter, T. H. & Auborn, K. (2000)=20
Indole-3-carbinol and diindolylmethane induce apoptotic changes in =
cervical=20
cancer cells in vitro and in vivo. HPV 2000: 273 (abs.)] and the =
Keystone=20
Symposium on the Molecular Basis of Cancer, January, 2001, Taos, NM =
[Carter, T.=20
H., Chen, D.-Z., Qi, M. & Auborn, K. J. (2001) Indole-3-carbinol and =
diindolylmethane sensitize cancer cells to induction of apoptosis via =
the TNF=CE=B1=20
pathway. p. 122 (abs.)]. ![]()
2 Supported by =
National=20
Institutes of Health, National Cancer Institute grant CA73385 (to =
K.J.A.). ![]()
4 Abbreviations =
used: 7-AAD,=20
7-amino actinomycin D; AD, actinomycin D; BSA, bovine serum albumin; =
DAPI,=20
4,6-diamidino-2-phenylindole; DIM, diindolylmethane; DMEM, =
Dulbecco=E2=80=99s modified=20
Eagle=E2=80=99s medium; DMSO, dimethyl sulfoxide; E2, =
17=C3=9F-estradiol; H&E,=20
hematoxylin and eosin; HPV, human papillomavirus; HRP, horseradish =
peroxidase;=20
I3C, indole-3-carbinol; LD50, dose at which the number of =
viable=20
cells is 50% of that in untreated cultures; MTS,=20
3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)=
-2H-tetrazolium;=20
POD, peroxidase; TUNEL, TdT-mediated dUTP nick-end =
labeling. ![]()
Manuscript received June 6, 2001. Initial review completed July 12, = 2001.=20 Revision accepted September 19, 2001.