biochemical pharmacology 73 (2007) 56–67

available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/biochempharm

Growth inhibitory activity of cucurbitacin glucosides

isolated from Citrullus colocynthis on human breast
cancer cells

Tehila Tannin-Spitz a, Shlomo Grossman a,*, Sara Dovrat a, Hugo E. Gottlieb b,

Margalit Bergman a
a
The Mina & Everard Goodman Faculty of Life Sciences, Bar-Ilan University, Ramat-Gan 52900, Israel
b
Department of Chemistry, Bar-Ilan University, Ramat-Gan 52900, Israel

article info abstract

Article history: Our aim was to study the effects of cucurbitacin glucosides extracted from Citrullus
Received 11 July 2006 colocynthis leaves on human breast cancer cell growth. Leaves were extracted, resulting
Accepted 11 September 2006 in the identification of cucurbitacin B/E glucosides. The cucurbitacin glucoside combination
(1:1) inhibited growth of ER+ MCF-7 and ER MDA-MB-231 human breast cancer cell lines.
Cell-cycle analysis showed that treatment with isolated cucurbitacin glucoside combination
Keywords: resulted in accumulation of cells at the G2/M phase of the cell cycle. Treated cells showed
Apoptosis rapid reduction in the level of the key protein complex necessary to the regulation of G2 exit
Breast cancer and initiation of mitosis, namely the p34CDC2/cyclin B1 complex. cucurbitacin glucoside
Cucurbitacin glucoside treatment also caused changes in the overall cell morphology from an elongated form to a
F-actin round-shaped cell, which indicates that cucurbitacin treatment caused impairment of actin
G2/M filament organization. This profound morphological change might also influence intracel-
p34CDC2 lular signaling by molecules such as PKB, resulting in inhibition in the transmission of
survival signals. Reduction in PKB phosphorylation and inhibition of survivin, an anti-
Abbreviations: apoptosis family member, was observed. The treatment caused elevation in p-STAT3 and in
ER, estrogen receptor p21WAF, proven to be a STAT3 positive target in absence of survival signals. Cucurbitacin
FACS, fluorescence activated cell glucoside treatment also induced apoptosis, as measured by Annexin V/propidium iodide
sorting staining and by changes in mitochondrial membrane potential (DC) using a fluorescent dye,
FCS, foetal calf serum JC-1. We suggest that cucurbitacin glucosides exhibit pleiotropic effects on cells, causing
JC-1, 5,50 ,6,60 -tetrachloro-1,10 ,3, both cell cycle arrest and apoptosis. These results suggest that cucurbitacin glucosides
30 -tetraethyl- might have therapeutic value against breast cancer cells.
benzamidazolocarbocyanin iodide # 2006 Elsevier Inc. All rights reserved.
MTT, 3-(4,5-dimethylthiazol-2-yl)-2,
5-diphenyl tetrazolium bromide
NMR, nuclear magnetic resonance
PI, propidium iodide
TLC, thin layer chromatography

* Corresponding author. Tel.: +972 3 5318050; fax: +972 3 6356869/5351824.

E-mail address: [email protected] (S. Grossman).
0006-2952/$ – see front matter # 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2006.09.012
 biochemical pharmacology 73 (2007) 56–67 57

1. Introduction explored the mechanism by which they reduce the prolifera-

tion of human breast cancer cells and found that cucurbitacin
Breast cancer affects one in every 10 women in Western B and E glucosides can cause accumulation of cells in the G2/M
Europe and the USA [1], and it is the second leading cause of phase of the cell cycle accompanied with induction of
cancer-related deaths [2]. Treatments include surgery, radia- apoptosis. In addition, we looked into the signaling mechan-
tion, and in some cases, drugs that are specifically targeted ism by which the cucurbitacin glucosides exert their beneficial
such as herceptin, in the case of tumors that overexpress EGF effects in the chemoprevention of breast cancer cells.
receptor, or tamoxifen, in the case of estrogen-dependent
tumors [3]. However, the majority of cases, especially those
that result in metastasis, are still treated with conventional 2. Material and methods
chemotherapy. The problem of drug resistance is a major
obstacle in chemotherapeutic treatment, therefore, there is a 2.1. Extraction and isolation procedures
great need for the development of new therapeutic drugs that
will be more efficient or will synergize with the existing ones. Leaves of C. colocynthis (L.) Shrad (collected from an open field
There has been a growing interest in the use of herbs as a at Bar-Ilan University) were mixed 1:4 (w/v) with distilled
potent source of new therapeutic anticancer drugs. Plants water and homogenized. The homogenate was filtered,
contain a wide variety of chemicals that have potent biological centrifuged and the supernatant was dried in a lyophilizer
effects, including anticancer activity [4]. Recently, we pre- (0.07 mbar, 48 8C). The powder was further extracted in 70%
sented the mechanism of NAO (natural antioxidants) from chloroform/methanol, filtered, and evaporated under reduced
spinach and derived purified components [5,6] acting through pressure. The resulting extract was tested by thin-layer
G1 arrest [7]. Moreover, we reviewed the efficacy of these chromatography (TLC) (silica gel 60 F254 plates, Merck Eurolab
natural compounds in various biological systems [8–10]. In this SA, France) using the solvent system: chloroform/methanol
research, we focused on the antiproliferative effect of the (9:1). The fraction with Rf = 0.56 was a mixture of two
cucurbitacin glucosides isolated from Citrullus colocynthis. cucurbitacin B and E glucosides (1:1). By further TLC using
C. colocynthis (L.) Shrad (Cucurbitaceae), locally known as benzene–ethanol (8:2), we succeeded in separating between
Sherry or Handal, is used in folk medicine by people in rural the cucurbitacin glucosides.
areas as a purgative, antirheumatic, and as a remedy for skin
infections. The plant contains cucurbitacins A, B, C, and D, a- 2.2. Identification of molecular structure by NMR
elaterin, and probably other constituents [11].
The cucurbitacins (highly oxygenated tetracyclic triter- NMR spectra of the cucurbitacin glucosides were obtained on a
pens) are of great interest because of the wide range of Bruker DMX-600 spectrometer, at 600.1 MHz (1H) and
biological activity they exhibit in plants and animals. They are 150.9 MHz (13C), respectively, at room temperature. The
predominantly found in the cucurbitaceae family but are also solvent used was a mixture of 10% CD3OD and 90% CDCl3,
present in several other families of the plant kingdom. Species containing 0.1% TMS as internal reference. NMR analysis was
of the plants in which they are found have been used for facilitated by the use of 2D experiments such as COSY (1H  1H
centuries in various pharmacopoeia. A number of compounds correlation), HMQC (one-bond 13C  1H correlation), and HMBC
of this group have been investigated for their cytotoxic, (long-range 13C  1H correlation).
hepatoprotective, anti-inflammatory, cardiovascular effects
and anti-diabetic effects [12]. Several studies were published 2.3. Cell culture
pointing out that different cucurbitacin species exhibit various
biological activities against tumor expansion. For instance, MCF-7 (ER positive) and MDA-MB-231 (ER negative) were
cucurbitacin I and Q were shown to specifically inhibit STAT3 obtained from the American type culture collection (Rockville,
phosphorylation, which contributes to the proliferation of MD, USA). Cells were grown in RPMI-1640 medium (Biological
many cancerous cells [13,14]. Another study reported that Industries Inc., Kibbutz Beit Haemek, Israel) supplemented
cucurbitacin E inhibits the proliferation of prostate cancer with 10% (v/v) foetal calf serum (FCS), 1% penicillin–strepto-
cells and causes disruption of the cytoskeleton structure of mycin–nystatin. Cells were maintained at 37 8C in a humidi-
actin and vimentin [15,16]. Cucurbitacin B, D, E, and I inhibited fied atmosphere of 5% CO2 in air.
the growth of several cancer cell lines and inhibited COX-2
enzyme rather than COX-1 [12]. Cucurbitacin B was shown to 2.4. Cell proliferation and viability assay
exhibit anti-inflammatory activity [17,18].
The purpose of the present study was to examine the The effect of the cucurbitacin glucoside treatment on
effects of the cucurbitacin glucosides (B and E) extracted from proliferation of MCF-7 and MDA-MB-231 cells was measured
C. colocynthis on human breast cancer cells, including the by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bro-
effects on cellular growth, cell-cycle distribution, apoptosis, mide (MTT) (Sigma, MO, USA) assay based on the ability of live
and the expression of proteins involved in cell-cycle regula- cells to cleave the tetrazolium ring to a molecule that absorb at
tion, utilizing both estrogen-dependent (MCF-7) and estrogen- 570 nm [19]. 7  103 cells were grown on a 96-well microliter
independent (MDA-MB-231) human breast cancer cell lines. plate and incubated with cucurbitacin B/E glucoside or their
To this end, we succeeded in isolating and describing by combination (4, 8, and 20 mM) in RPMI-10% FCS medium. After
nuclear magnetic resonance (NMR) analysis the molecular 24, 48, and 72 h, the medium was changed and 130 ml/well of
structure of both cucurbitacin B and E glucosides. We also fresh RPMI-1640 media was added. Next, 20 ml MTT reagent
58 biochemical pharmacology 73 (2007) 56–67

(5 mg/1 ml PBS) was added to each well, and the cells were media for 4 or 24 h. Cells were fixed by 4% (w/v) PFA in PBS
further incubated at 37 8C for 2 h. To determine lysis of the for 15 min at room temperature, washed in PBS, and
cells, 100 ml N,N-dimethyl formamid solution (50% final permeabilized with 0.5% (v/v) Triton X-100 for 30 min at RT.
concentration of N,N-dimethyl formamid and 20% of sodium Cells were washed three times in PBS and incubated with
dodecyl sulphate, pH 4.7) was added to each well for an TRITC-labeled-phalloidin (P1951, Sigma) for 40 min in the
additional 7 h, followed by reading on a scanning multiwell dark. Preparations were washed and stained with Hoechst for
spectrophotometer. 5 min, washed again in PBS and finally mounted in mounting
medium 80% glycerol in PBS, 1% n-propyl gallate (Sigma).
2.5. Flow cytometry Microscopic observations were performed and photographed
using a fluorescence microscope (Axiovert 200M) equipped
MCF-7 and MDA-MB-231 cells (5  105) were seeded in 100-mm with a HBO 100 W mercury lamp.
culture dishes, and allowed to attach overnight. The medium
was replaced with fresh complete medium containing the 2.8. Annexin V/PI flow cytometric staining technique
desired concentration of cucurbitacin B/E glucoside (B: 20 and
33 mM; E: 15 and 25 mM) or their combination (8 and 20 mM). Apoptosis was determined based on morphological change.
Cells were incubated for 24 h at 37 8C. The cells were washed Apoptotic cells were quantified by Annexin V-FITC and PI
with PBS and then spun at 300  g. The pellet was resuspended double-staining by using a staining kit purchased from MBL
in 250 ml PBS and 250 ml propidium iodide (PI) (50 mg/ml final Co. Ltd. (Watertown, MA, USA).
concentration) for 15 min, and analyzed using a flow cyt-
ometer. 2.9. JC-1 mitochondrial membrane potential detection
assay
2.6. Western blot analysis
We used JC-1 (Sigma, MO, USA) for in-situ detection of
MCF-7 and MDA-MB-231 cells (1  106) were seeded in 100-mm mitochondrial membrane transition events in live cells, which
culture dishes. The cells were treated with the cucurbitacin provides an early indication of the initiation of cellular
glucoside combination (20 mM) in RPMI-1640 media for 1, 4 or apoptosis. The collapse in the electrochemical gradient across
24 h. The media was then aspirated, and cells were washed the mitochondrial membrane (DC) was measured using a
with cold PBS. The cells were scraped and washed twice by fluorescent cationic dye 5,50 ,6,60 -tetrachloro-1,10 ,3,30 -tetra-
centrifugation at 500  g for 5 min at 4 8C. The pellet was ethyl-benzamidazolocarbocyanin iodide, known as JC-1. In
resuspended in lysis buffer supplemented with proteases and non-apoptotic cells, JC-1 exists as a monomer in the cytosol
phosphotase inhibitors and incubated for 1 h at 4 8C. The (green) and also accumulates as aggregates in the mitochon-
lysate was collected by centrifugation at 14,000  g for 40 min dria (red). In apoptotic and necrotic cells, JC-1 exists in
at 4 8C, and the supernatant (total cell lysate) was stored at monomeric form and stains the cytosol green. For this assay,
20 8C. For Western blot analysis, 40 mg protein was resolved cells were treated with the cucurbitacin glucoside combina-
over 12% polyacrylamide gels and transferred to a nitrocellu- tion (0, 8, and 20 mM) and maintained at 37 8C in a humidified
lose membrane. The blot was blocked in blocking buffer (1% 5% CO2 atmosphere for 24 h. Cells were washed with PBS and
non-fat dry milk/1% Tween 20 in PBS) for 1 h at room spun at 300  g. The pellet was resuspended in 500 ml PBS, and
temperature, incubated with appropriate monoclonal primary 2 ml of 1 mg/ml JC-1 reagent was added for 20 min in 37 8C in
antibodies (human reactive anti-p34CDC2, anti-Survivin, anti- the dark. Cells were washed with PBS and analyzed using a
STAT3 from Santa Cruz Biotechnology, anti-Cyclin B1 from flow cytometer. Analysis was performed using Cell Quest
Biosource, and anti-p21WAF from BD Biosciences) or polyclonal software (BD Biosciences, San Jose, CA, USA) for apoptosis.
primary antibodies (human reactive anti-phospho-p34CDC2
(Thr14 and Tyr15) from Santa Cruz, anti-phospho-STAT3 (Tyr 2.10. Statistical analysis
705), anti-AKT, and anti-phospho-AKT (Ser 473) from cell
signaling) in blocking buffer overnight at 4 8C. The blot was All experiments were performed at least three times. Where
then incubated with appropriate secondary antibody horse- appropriate, the data are expressed as the mean  standard
radish peroxidase conjugate and detected by chemilumines- error of the mean (S.E.M.). Probability (P) was calculated using
cence and autoradiography using X-ray film. b-actin was a Student t-test. P-values lower than 0.05 were considered
detected on the same membrane and used as a loading significant.
control. Protein bands (p34CDC2 and p-p34CDC2 after 1, 4 or 24 h
treatment) were quantified with the image analysis software
Image J. Results were presented as normalized fold changes in 3. Results
relation to control. Normalization has been done using b-actin
as a loading control. 3.1. Spectroscopic identification of isolated fraction

2.7. Fluorescent microscopy Fractions were isolated from the water extract of C. colocynthis
leaves using TLC as described in Section 2, and their structure
F-actin detection was performed on cells grown on coverslips was determined by NMR spectroscopy (see Supplementary
and placed into a six-well plate. Cells were treated with the data). The fractions isolated are glycosylated cucurbitacin B
cucurbitacin glucoside combination (20 mM) in RPMI-1640 and E (Fig. 1).
 biochemical pharmacology 73 (2007) 56–67 59

Fig. 1 – Chemical structure of cucurbitacin B glucoside Fig. 2 – Dose-dependent effect of cucurbitacin glucoside
(MW = 739). Structure of cucurbitacin E glucoside combination (B and E in 1:1 ratio) on MDA-MB-231
(MW = 737) is the same as cucurbitacin B glucoside except proliferation. MDA-MB-231 cells at 7  103 cells/well were
for a double bond between carbons 1 and 2. cultured with cucurbitacin glucoside combination (4, 8,
and 20 mM) in RPMI-10% FCS. Daily cell proliferation was
determined by MTT assay. The data represent the
mean W S.E. of three independent experiments. Effect of
3.2. Effect of cucurbitacin B and E glucoside combination
cucurbitacin glucoside combination is statistically
on MCF-7 and MDA-MB-231 cell proliferation
significant in comparison with control (P < 0.05).

To evaluate the effect of cucurbitacin glucosides on cell

proliferation of human breast cancer cell lines MCF-7 and 3.3. Effect of cucurbitacin B/E glucosides and their
MDA-MB-231, we initially treated both cell lines with escalat- combination on cell cycle
ing doses of the cucurbitacin glucoside combination (1:1). Cell
proliferation was estimated by the MTT assay (as described in To gain insights into the mechanism by which cell reduction is
Section 2). The effect at various concentrations was studied achieved, we investigated the effect on cell cycle distribution
after 24, 48, and 72 h of cell growth. Fig. 2 demonstrates that by fluorescence-activated cell sorting (FACS) analysis, and the
the cucurbitacin glucoside combination inhibited cellular results are summarized in Figs. 3 and 4. A 24-h exposure of
proliferation in a dose-dependent manner. The proliferation MDA-MB-231 and MCF-7 cells to 20 mM cucurbitacin B gluco-
of MDA-MB-231 cells was reduced by 50% upon a 48-h side or 25 mM cucurbitacin E glucoside caused 3-fold
exposure to 8 mM cucurbitacin glucoside combination. Treat- enrichment of cells in G2/M phase (Fig. 3). Exposure of MCF-
ing MCF-7 cells with cucurbitacin glucoside combination also 7 cells to 20 mM of the cucurbitacin glucoside combination (1:1)
inhibited cellular proliferation (data not shown). Antiproli- also resulted in the accumulation of cells in the G2/M phase.
ferative activity of the cucurbitacin glucoside combination G2/M enrichment was accompanied by a reduction in S phase
was further confirmed by counting live cells (using trypan blue cells, from 55% cells in the control to 3% cells following
exclusion dye). Proliferation of MDA-MB-231 cells was inhib- treatment (Fig. 4A). A histogram representing this G2/M arrest
ited upon treatment in dose- and time-dependent manner is shown in Fig. 4B. In the following experiments, we used the
(data not shown). combination of B and E cucurbitacin glucosides (1:1) since their

Fig. 3 – Effect of cucurbitacin B/E glucoside on cell-cycle distribution in MDA-MB-231 and MCF7 cells: (A) MDA-MB-231 cells
treated with cucurbitacin B glucoside; (B) MDA-MB-231 cells treated with cucurbitacin E glucoside; (C) MCF-7 cells treated
with cucurbitacin B glucoside; (D) MCF-7 cells treated with cucurbitacin E glucoside. The cells (5  105) were treated with
cucurbitacin glucoside and analyzed at 24 h by DNA flow cytometry. Values indicate the percentage of cells in the indicated
phases of the cell cycle. The data shown are representative of three independent experiments with similar findings.
60 biochemical pharmacology 73 (2007) 56–67

Fig. 4 – Effect of cucurbitacin glucoside combination (E + B) on cell-cycle distribution. (A) MCF7 cells treated with cucurbitacin
glucoside combination (8 and 20 mM). The values indicate the percentage of cells in the indicated phases of the cell cycle.
The data shown are representative of three independent experiments with similar findings. (B) Histograms show number
of cells per channel (vertical axis) vs. DNA content (horizontal axis). MCF7 cells (5  105) were treated with 8 and 20 mM of
cucurbitacin glucoside combination and analyzed at 24 h by DNA flow cytometry.

effect on cell cycle distribution was similar to the effect of each investigation. These results suggest that changes in expres-
individual cucurbitacin glucoside (B/E). sion of G2/M regulating proteins may contribute to cucurbi-
tacin glucoside-mediated cell-cycle arrest in MCF-7 and MDA-
3.4. Effect of cucurbitacin glucoside combination on p34CDC MB-231 cells.
and cyclin B1 expression
3.5. Effect of cucurbitacin glucoside combination on cell
In an attempt to understand the effect of the molecular events morphology and actin organization
involved in cucurbitacin glucoside activity on cell cycle
progression, we next investigated the effect of the cucurbi- As previously described in the literature, cucurbitacin E was
tacin glucoside combination on the expression of proteins shown to interfere with cell cytoskeleton by causing marked
that are pivotal for G2/M transition, including p34CDC2 and disruption of the actin cytoskeleton [15,16]. In order to verify
cyclin B1. G2/M transition requires activity of cyclin-depen- if the cucurbitacin glucoside combination had an effect on
dent kinase, p34CDC2. p34CDC2 is positively regulated by cell morphology, especially on cytoskeleton elements, we
association with cyclin B1 and negatively regulated by used light microscopy and performed staining analysis.
phosphorylation of amino acids Thr14 and Tyr15. Fig. 5A Viewing cells after treatment (20 mM) compared to control,
represents a typical Western blot from treated (20 mM using light microscopy, clearly showed a deep change in the
cucurbitacin glucoside combination) and untreated cells. A overall morphology of both MCF-7 and MDA-MB-231 cells
decrease in the level of cyclin B1 and in the level of p34CDC2 from their typical morphology to round-shaped cells (Fig. 6).
protein was observed at 24 h post-treatment. In many cases, Interestingly, this effect on cell morphology appeared 1 h
especially those elicited by DNA damage agents, reduction in post-treatment (data not shown). Changes in cell shape are
p34CDC2 expression is preceded by p34CDC2 phosphorylation at associated with reorganization of actin filaments [20],
the inhibitory positions Thr14 and Tyr15. In order to find out if therefore, we examined the effect of cucurbitacin glucosides
cucurbitacin glucoside combination treatment induces inhi- on actin distribution. Cells were stained with TRITC-labeled-
bitory phosphorylation of p34CDC2, we followed the Thr14 and phalloidin, which binds selectively to F-actin. As seen in
Tyr15 phosphorylation status of p34CDC2 at early time points. Fig. 7, shortly after treatment (4 h), and 24 h later on, we
As can be seen in Fig. 5B, a reduction in phospho-p34CDC2 was observed changes in F-actin distribution from organized
observed as early as 1 h post-treatment. It seems that the actin filament network underneath cell membrane and at
cucurbitacin glucoside combination causes rapid elimination the edges of control cells to accumulation of F-actin in the
of p34CDC2 protein, which is responsible for blocking G2 to M cytoplasm next to the nucleus, in cucurbitacin glucoside-
transition. The mechanism by which this phenomenon treated cells. This effect was probably not a result of an
occurs is not yet known but it is the focus of our present apoptotic process, which is also characterized by round-cell
 biochemical pharmacology 73 (2007) 56–67 61

3.6. Effect of cucurbitacin glucosides on STAT3 activity

and signaling

Recently, a study was published which demonstrated that

cucurbitacin I and Q can interfere with STAT3 signaling by
specifically inhibiting STAT3 phosphorylation [13]. Therefore,
we were interested to find out the effect of the cucurbitacin
glucoside combination on STAT3 phosphorylation in our cell
system. Unexpectedly, we found that the cucurbitacin gluco-
side combination enhances rather than decreases STAT3
phosphorylation. This effect was detected early after treat-
ment (4 h), and remained after 24 h treatment as well (Fig. 8).
Constitutive activation of STAT3 has been shown in many
different tumors. This activation usually results in an anti-
apoptotic effect and promotes cell proliferation [23]. The effect
of cucurbitacin glucosides on STAT3 phosphorylation is,
therefore, incompatible with our previous results showing
decrease in cell proliferation. We found a partial explanation
for this enigma in the literature by combining the data of two
different recently published papers concerning both STAT3
signaling and the consequences of cytoskeleton disruption.
First, STAT3 was found to exhibit inhibitory activity on cell
proliferation by enhancing p21WAF expression in a specific
situation in which both STAT3 activation and inhibition of PKB
signaling occur simultaneously [24]. Another published study
Fig. 5 – Western blot analysis of extracts obtained from showed that disruption of actin cytoskeleton network reduces
MCF7 and MDA-MB-231 cells treated with the cucurbitacin
PKB signaling, culminating in reduction in the expression level
glucoside combination. (A) Cells were treated with 0 (S)
of survivin (an IAP family member) which leads to both G2/M
and 20 mM (+) cucurbitacin glucoside combination for 24 h.
arrest and apoptosis [22]. Since we observed both STAT3
Extracts were prepared and analyzed by Western blotting
activation and cell transition to round-cell morphology, which
with an antibody to p34CDC2 and cyclin B1. (B) MDA-MB-
is a hallmark of cytoskeleton disruption, we suspected that
231 cells were treated with 0 (S) and 20 mM (+) cucurbitacin
cucurbitacin treatment brings about the specific condition in
glucoside combination for 1, 4, and 24 h. Extracts were
which STAT3 activity is exploited in favor of preventing cell
prepared and analyzed by Western blotting with an proliferation rather than enhancing it. To verify this hypoth-
antibody to p34CDC2 and phospho-p34CDC2. An antibody to esis, we looked at the key elements in this signaling scheme,
b-actin was used as a loading control. Western blots are namely p21WAF, PKB, and survivin. Indeed, we can demon-
representative of three independent experiments. strate that cucurbitacin glucoside treatment elevated p21WAF
Quantification of the bands was performed and expression, caused reduction in PKB phosphorylation, and,
normalized change folds in relation to non-treated cells
consequently, inhibited survivin expression, exactly as
are shown in parenthesis. Normalization has been done
expected (Fig. 9A–C).
using b-actin as a loading control.
3.7. Cucurbitacin glucoside combination induces apoptosis

morphology, since it appeared shortly after treatment when The apoptosis inducing effect of the cucurbitacin glucoside
there was no evidence of cells undergoing apoptosis. Using a combination was evaluated by Annexin V/PI binding. One of
transmission electron microscope, we confirmed this actin the earliest events of apoptosis is the loss of plasma-
reorganization. Control cells showed typical morphology membrane polarity, accompanied by translocation of phos-
while, in treated cells, the actin accumulated next to the phatidylserine (PS) from the inner to outer membrane leaflets,
nucleus (data not shown). thereby exposing PS to the external environment [25]. The
Lately, accumulating evidence in the literature indicates phospholipid-binding protein Annexin V has a high affinity for
that cytoskeletal components are involved in downstream PS and can bind to cells with externally exposed PS. Positive
signal transduction pathways regulating cell survival [21]. For staining with fluorescently labeled Annexin V correlates with
instance, it is reported that cytochalasin B, which causes loss of membrane polarity but precedes the complete loss of
cytoskeletal dearrangement [22], exhibits inhibition of pro- membrane integrity that accompanies the later stages of cell
liferation, G2/M arrest, and apoptosis. These effects closely death resulting from either apoptosis or necrosis [25]. In
resemble the effects of cucurbitacin glucosides, therefore, we contrast, PI can only enter cells after loss of membrane
hypothesize that cucurbitacin glucosides mediate their integrity. Thus, dual staining with Annexin V and PI allows
effects, at least partially, by altering cytoskeleton network, clear discrimination between unaffected cells (Annexin V
causing changes in cell morphology, which lead to both G2/M negative, PI negative), early apoptotic cells (Annexin V positive,
arrest and apoptosis. PI negative), and late apoptotic cells (Annexin V positive, PI
62 biochemical pharmacology 73 (2007) 56–67

Fig. 6 – Effect of cucurbitacin glucosides combination treatment on (A) MCF-7 cell morphology and (B) MDA-MB-231 cell
morphology. Cells were treated with 20 mM of the cucurbitacin glucoside combination (24 h) as indicated, and
photographed (light microscopy, 20 magnification).

positive). As can be seen in Fig. 10, the cucurbitacin glucoside In these studies, the leaves of C. colocynthis were extracted
combination treatment (20 mM) increased the percentage of late with water, followed by a few more extractions (described in
apoptotic cells (presented in figure as UR) in MDA-MB-231 cells Section 2) to yield cucurbitacin B glucoside and cucurbitacin E
from 18.43% in control untreated cells to 61.35% in treated cells glucoside purified fractions. Cucurbitacin B and E glucoside
(Fig. 10A) (similar data were obtained for MCF-7 cells, data not combination (1:1) exhibited growth inhibitory activity of human
shown). When the cucurbitacin glucoside combination was breast cancer cell lines in a dose- and time-dependent manner
tested in two escalating concentrations (8 and 20 mM), induction as can be seen by MTT assay (Fig. 2). Cucurbitacin B and E
of apoptosis showed dose dependency (Fig. 10B), which combination treatment inhibited cell growth with IC50 value of
indicates that the apoptotic effect specifically resulted from 8 mM for MDA-MB-231 cell line after 48 h. In order to elucidate
the cucurbitacin glucoside treatment. the mechanism by which these extracts inhibited cell growth,
To further confirm our data, we used additional in-situ we performed cell-cycle analysis. The cucurbitacin glucoside
detection of apoptosis by measuring the effects of treatment combination and each of the cucurbitacin glucosides alone
on mitochondrial membrane potential (DC) using a fluor- induced cell-cycle arrest in the G2/M phase of the cell cycle.
escent cationic dye 5,50 ,6,60 -tetrachloro-1,10 ,3,30 -tetraethyl- Because of the similar effect on cell cycle distribution observed
benzamidazolocarbocyanin iodide, known as JC-1. In healthy for each cucurbitacin glucoside alone or their combination, we
cells, JC-1 exists as a monomer in the cytosol (FL1 positive; decided to use a combination of the cucurbitacin glucosides in a
green) and also accumulates as aggregates in the mitochon- 1:1 ratio for our subsequent experiments.
dria (FL2 positive; red). In apoptotic and necrotic cells, JC-1 In order to get a better understanding of the mechanism by
exists exclusively in monomer form and produces a green which the cucurbitacin combination leads to G2/M arrest, we
cytosolic signal. As shown in Fig. 11, cucurbitacin glucoside followed the status of key proteins known to regulate G2/M
combination treatment in MDA-MB-231 cells elevated the transition, such as p34CDC2 and cyclin B1. The p34CDC2 protein
fraction of apoptotic cells from 11.17% in control cells to kinase is generally acknowledged to be the key mediator of G2/
48.58% in cells treated with 20 mM cucurbitacin glucoside M phase transition in all eukaryotic cells [26]. The active
combination. Herein we also observed dose response mitotic kinase (MPF, or mitosis-promoting factor) is a dimer
(Fig. 11). comprised of a catalytic subunit, a B-type cyclin, and a cyclin-
dependent kinase termed p34CDC2 or CDK1. The activity of the
p34CDC2 kinase not only depends on its association with cyclin
4. Discussion B1, but also on its phosphorylation state. Phosphorylation of
either Thr14 or Tyr15 inhibits p34CDC2 kinase activity, while
In the present study, we studied the inhibitory effect of phosphorylation on Thr161 by CDK7 kinase is required for
cucurbitacin glucosides (B and E) isolated from C. colocynthis kinase activity. In addition, the dephosphorylation of Thr14 or
leaves, on breast cancer cells. Tyr15 by CDC25C phosphatase is a final step for p34CDC2 kinase
 biochemical pharmacology 73 (2007) 56–67 63

Fig. 8 – Effect of cucurbitacin glucoside combination on

phospho-STAT3 expression. phospho-STAT3 expression
was detected by Western blot analysis of extracts obtained
from MCF7 and MDA-MB-231 cells treated with the
cucurbitacin glucoside combination. Cells were treated
with 0 (S) and 20 mM (+) cucurbitacin glucoside
combination for 4 h (A) and 24 h (B). Extracts were
prepared and analyzed by Western blotting with an
antibody to STAT3 and phospho-STAT3 (Tyr 705).
Antibody to b-actin was used as a loading control.
Fig. 7 – Effect of cucurbitacin glucoside combination (20 mM) Western blots are representative of three independent
on F-actin distribution in MDA-MB-231 cells, 4 h post- experiments.
treatment (A) in MCF-7 cells, 24 h post-treatment (B). Cells
were treated with cucurbitacin glucoside combination
(20 mM), fixed and stained with TRITC-labeled-phalloidin activities against tumor expansion. For instance, cucurbita-
and Hoechst and photographed (fluorescence microscope, cin I and Q were shown to specifically inhibit STAT3
100 magnification). Upper photographs present DIC phosphorylation which contributes to the proliferation of
images, lower photographs present double-staining many cancerous cells [13,14]. In our study, we demonstrated a
images (TRITC-labeled-phalloidin and Hoechst). rapid and sustained elevation in the phosphorylated form of
STAT3 (Fig. 8). This result was not compatible with the
published data and at first seemed to be in contradiction to
activity [27,28]. We found that cucurbitacin glucoside treat- our own results concerning the antiproliferative effect of
ment caused a reduction at the protein level of both p34CDC2 cucurbitacin glucosides treatment. This conflict was partially
and cyclin B1. The reduction of p34CDC2 was very rapid and resolved when we considered the effects of cucurbitacin
could be detected as early as 1 h post-treatment. By 24 h, only glucosides on cell morphology. We noticed a marked change
remnants of p34CDC2 and cyclin B1 protein were observed in cell morphology, as was mentioned before in a study
(Fig. 5A). This phenomenon is even more impressive if we take reporting that cucurbitacin E inhibits the proliferation of
into consideration that by 24 h of treatment, most of the cells prostate cancer cells and causes disruption of the cytoske-
were arrested in the G2/M stage in which p34CDC2 and cyclin B1 leton structure of actin and vimentin [15,16]. cucurbitacin
are mostly expressed. Moreover, when we followed phos- glucoside treatment caused disruption of the elongated shape
phorylated p34CDC2, we found a sharp decrease in phosphor- of the cells toward a round-shaped appearance as was
ylation status that followed a similar kinetics to the inhibition demonstrated by both light microscopy visualization
at the protein level (Fig. 5B). Thus, it is reasonable to (Fig. 6) and F-actin staining (Fig. 7). The F-actin network
hypothesize that cucurbitacin glucoside treatment causes was shown to shift toward an accumulated form next to the
cell cycle arrest at G2/M by reducing the amount and, hence, nucleus instead of supporting cell shape by organization
the activity of p34CDC2/cyclin B1 complex, which is necessary under the cell membrane. Two recently published articles
for G2 to M transition, and not by inhibition of p34CDC2 activity revealed data that might connect between the multiple
by its phosphorylation. effects of cucurbitacin glucoside treatment. On one hand,
Recently, several studies were published pointing out that disruption of actin cytoskeleton was shown to inhibit PKB
different cucurbitacin species exhibit various biological phosphorylation and signaling, resulting in inhibition of
64 biochemical pharmacology 73 (2007) 56–67

connections [20,30,31]. Substances that inhibit cytoskeleton

formation, such as okadaic acid, or cause actin depolymeriza-
tion, such as cytochalasin B, were shown to induce apoptotic
cell death, which involves changes in mitochondrial mem-
brane potential [21,22,32]. Inhibition of survivin, which is a
member of a family of proteins that inhibits apoptosis (IAP),
can also contribute to induction of apoptosis, since survivin
was shown to inhibit both activated caspases 9 and 3 [22]. The
fact that cucurbitacin glucosides greatly inhibited survivin
expression and disrupted actin polymerization implies that
cucurbitacin glucosides might induce apoptosis. Therefore,
we examined whether cucurbitacin treatment can induce
apoptotic cell death by using two methods: cell-surface
Annexin V binding, which measures the appearance of
phosphatidylserine on the external plasma membrane, and
JC-1 binding, which indicates changes in mitochondrial
membrane potential [33,34]. Collapse of mitochondrial mem-
brane potential is an accessory in the execution of apoptosis,
which is directly dependent on cytoskeletal changes [21]. The
two methods tested clearly showed (Figs. 10 and 11) an
induction up to 3.5-fold in apoptotic cell fraction following
cucurbitacin glucoside treatment. The differences in the
percentage of apoptotic control cells between the methods
used for apoptosis detection (Figs. 10 and 11), might partly
reflect the difference in the sensitivity of the methods used.
Fig. 9 – Western blot analysis of extracts obtained from As mentioned above, cucurbitacin glucoside treatment
MCF7 and MDA-MB-231 cells treated with the cucurbitacin induced cell-cycle arrest in the G2/M phase (68% of the cells)
glucoside combination. Cells were treated with 0 (S) and (Fig. 4) and, at the same time, displayed 60% cells in apoptosis
20 mM (+) cucurbitacin glucoside combination for 24 h. (Fig. 10). These results might seem contradictory, but can be
Extracts were prepared and analyzed by Western blotting explained based on the fact that apoptotic cells are often
with an antibody to p21WAF (A); antibody to PKB and accompanied by apoptotic body formation, which is reflected
phospho-PKB (Ser 473) (B); antibody to survivin (C). as sub-G1 peak in FACS analysis. Since cucurbitacin glucosides
Antibody to b-actin was used as a loading control. interfere with actin cytoskeleton dynamics, which is a
Western blots are representative of three independent prerequisite for apoptotic body formation, no apoptotic bodies
experiments. appear and therefore the apoptotic fraction is invisible by
FACS analysis. The phenomenon of absence of apoptotic
bodies was mentioned in the literature [32] regarding
survivin expression [22], and on the other hand, another cytochalasin D, which is known to interfere with actin
published paper reported that STAT3 activation can lead to dynamics. Therefore, for apoptosis detection we used meth-
p21WAF induction and, hence, inhibition of cell proliferation ods directed toward biochemical changes typical for cells in
when STAT3 activation occurs concomitantly with PKB apoptosis, such as translocation of phosphatidyl serine from
inhibition [24]. We speculate that these two occurrences the inner to outer membrane leaflets (Fig. 10) and changes in
explain the mechanism activated by cucurbitacin glucoside mitochondrial membrane potential (Fig. 11). These methods
treatment, meaning that treatment causes both STAT3 reveal the apoptotic fraction that was unseen by FACS
activation (Fig. 8) and, at the same time, inhibits PKB analysis. In other words, we suggest that at least part of the
phosphorylation through disruption of cytoskeleton fila- cells diagnosed as apoptotic cells are presented as cells in G2/M
ments. As can be seen in Fig. 9, cucurbitacin glucoside phase using FACS analysis, since they still display 4N DNA
treatment induced p21WAF expression as expected in a content.
situation where STAT3 was enhanced in the presence of In conclusion, the results of the present study indicate that
inhibited PKB phosphorylation. In addition, as expected from cucurbitacin glucosides effectively inhibited proliferation of
PKB dephosphorylation, we also detected a sharp reduction in both estrogen-dependent and estrogen-independent human
survivin expression level (Fig. 9). breast cancer cells by causing G2/M phase arrest and
These results suggest that cucurbitacin glucoside treat- apoptosis. The mechanism by which cucurbitacin glucosides
ment might have an even greater impact on tumor expansion mediated this biological effect was partially revealed (Fig. 12).
than just G2/M arrest and inhibition of cell proliferation since It is reasonable to hypothesize that cucurbitacin glucosides
both survivin inhibition and discarding of actin filament can can be useful for chemotherapy of both estrogen-dependent
cause apoptotic cell death. Epithelial cells especially depend and -independent breast cancers. A major issue concerning
on their interaction with the extracellular matrix for survival cucurbitacin therapeutic value is the overall toxicity of these
[29], and the organization of actin filaments plays an compounds in vivo. The toxicity of C. colocynhtis fruits was
important role in the formation of extracellular matrix tested on both rats and sheep by oral administration and
 biochemical pharmacology 73 (2007) 56–67 65

Fig. 10 – Annexin V binding and propidium iodide uptake induced by the cucurbitacin glucoside combination treatment: (A)
MDA-MB-231 cells were treated with the cucurbitacin glucoside combination (20 mM) for 24 h, stained with Annexin V and
propidium iodide, and analyzed by flow cytometry. The horizontal (FL1-H) and vertical (FL2-H) axes represent labeling with
Annexin V and PI, respectively; (B) graphic presentation of data obtained by Annexin/PI staining after 24 h treatment using
8 and 20 mM cucurbitacin glucoside combination. UR represents late apoptotic cells (positive for both Annexin and PI), LL
represents live cells. The data shown are representative of three independent experiments with similar findings.

although toxic effects were recorded, they were not fatal we hope that therapeutic doses will not cause a fatal toxic
[11,35]. In a recently published study, a comparison between effect and will be tolerable. Our future studies will focus on
the cytotoxic effects of different cucurbitacin molecules was that issue to gain more information and deepened insight into
conducted and the final conclusion was that glycosylated the molecular targets of these molecules inside the cells in
molecules were less toxic than those that did not contain a order to get a better comprehension of the way they induce
glycoside moiety [36]. Since our cucurbitacin is glycosylated, cell-cycle arrest and apoptosis.

Fig. 11 – Effect of cucurbitacin glucoside combination on mitochondrial membrane potential. Cells were treated with
cucurbitacin glucoside combination (8 and 20 mM) for 24 h to follow the extent of apoptosis by determination of
mitochondrial membrane potential using JC-1 reagent. Cells were analyzed on a FACScan cytometer. Dot plots of red (FL2)
vs. green fluorescence (FL1) show live cells with intact mitochondrial membrane potential and dead cells with lost
mitochondrial potential, respectively. The data shown are representative of three independent experiments with similar
findings.
66 biochemical pharmacology 73 (2007) 56–67

[6] Bergman M, Perelman A, Dubinsky Z, Grossman S.

Scavenging of reactive oxygen species by a novel
glucurinated flavonoid antioxidant isolated and purified
from spinach. Phytochemistry 2003;62:753–62.
[7] Bakshi S, Bergman M, Dovrat S, Grossman S. Unique
natural antioxidants (NAOs) and derived purified
components inhibit cell cycle progression by
downregulation of ppRb and E2F in human PC3 prostate
cancer cells. FEBS Lett 2004;573:31–7.
[8] Lomnitski L, Bergman M, Nyska A, Ben-Shaul V, Grossman
S. Composition, efficacy, and safety of spinach extracts.
Nutr Cancer 2003;46:222–31.
[9] Tam NN, Nyska A, Maronpot RR, Kissling G, Lomnitski L,
Suttie A, et al. Differential attenuation of oxidative/
nitrosative injuries in early prostatic neoplastic lesions in
TRAMP mice by dietary antioxidants. Prostate 2006;66:57–69.
[10] Breitbart E, Lomnitski L, Nyska A, Malik Z, Bergman M,
Sofer Y, et al. Effects of water-soluble antioxidant from
Fig. 12 – Proposed model for cucurbitacin glucoside- spinach, NAO, on doxorubicin-induced heart injury. Hum
mediated G2/M arrest and apoptosis. Following Exp Toxicol 2001;20:337–45.
cucurbitacin glucoside treatment, dearrangement of actin [11] Adam SE, Al Yahya MA, Al Farhan AH. Response of Najdi
filaments rapidly occurs, culminating in round-cell sheep to oral administration of Citrullus colocynthis fruits,
morphology. Dephosphorylation of PKB and activation of Nerium oleander leaves or their mixture. Small Rumin Res
STAT3 were observed as well. This unique situation of 2001;40:239–44.
[12] Jayaprakasam B, Seeram NP, Nair MG. Anticancer and
simultaneous inhibition of PKB activity and STAT3
antiinflammatory activities of cucurbitacins from Cucurbita
activation is possibly responsible for the observed andreana. Cancer Lett 2003;189:11–6.
induction in p21WAF expression. PKB inhibition might also [13] Blaskovich MA, Sun J, Cantor A, Turkson J, Jove R, Sebti SM.
lead to reduction in survivin, which promotes Discovery of JSI-124 (cucurbitacin I), a selective Janus
programmed cell death. Reduction in the protein level of kinase/signal transducer and activator of transcription 3
key cell-cycle regulatory proteins such as p34CDC2 and signaling pathway inhibitor with potent antitumor activity
against human and murine cancer cells in mice. Cancer Res
cyclin B1, was also detected. Taken together, these overall
2003;63:1270–9.
changes lead to both G2/M arrest and apoptosis.
[14] Sun J, Blaskovich MA, Jove R, Livingston SK, Coppola D,
Sebti SM, et al. Cucurbitacin Q: a selective STAT3 activation
inhibitor with potent antitumor activity. Oncogene
2005;24:3236–45.
Acknowledgements [15] Duncan KL, Duncan MD, Alley MC, Sausville EA.
Cucurbitacin E-induced disruption of the actin and
The authors would like to thank Prof. Zvi Malik and Ms. Judy vimentin cytoskeleton in prostate carcinoma cells.
Biochem Pharmacol 1996;52:1553–60.
Hanania for assistance with the use of the Transmission
[16] Duncan MD, Duncan KL. Cucurbitacin E targets
Electron Microscopy, and Mr. Uri Karo for assistance with the
proliferating endothelia. J Surg Res 1997;69:55–60.
FACS. This research was partially supported by the Vinograd [17] Peters RR, Saleh TF, Lora M, Patry C, de Brum-Fernandes AJ,
Foundation. Farias MR, et al. Anti-inflammatory effects of the products
from Wilbrandia ebracteata on carrageenan-induced pleurisy
in mice. Life Sci 1999;64:2429–37.
Appendix A. Supplementary data [18] Yesilada E, Tanaka S, Sezik E, Tabata M. Isolation of an anti-
inflammatory principle from the fruit juice of Ecballium
elaterium. J Nat Prod 1988;51:504–8.
Supplementary data associated with this article can be [19] Mosmann T. Rapid colorimetric assay for cellular growth
found, in the online version, at doi:10.1016/j.bcp.2006.09.012. and survival: application to proliferation and cytotoxicity
assays. J Immunol Meth 1983;65:55–63.
[20] Burridge K, Chrzanowska-Wodnicka M. Focal adhesions,
references
contractility, and signaling. Annu Rev Cell Dev Biol
1996;12:463–518.
[21] Cabado AG, Leira F, Vieytes MR, Vieites JM, Botana LM.
[1] Chang JC. A review of breast cancer chemoprevention. Cytoskeletal disruption is the key factor that triggers
Biomed Pharmacother 1998;52:133–6. apoptosis in okadaic acid-treated neuroblastoma cells.
[2] Parker SL, Tong T, Bolden S, Wingo PA. Cancer statistics. CA Arch Toxicol 2004;78:74–85.
Cancer J Clin 1997;47:5–27. [22] Liang YL, Wang LY, Wu H, Ma DZ, Xu Z, Zha XL. PKB
[3] Carolin KA, Pass HA. Prevention of breast cancer. Crit Rev phosphorylation and survivin expression are cooperatively
Oncol Hematol 2000;33:221–38. regulated by disruption of microfilament cytoskeleton. Mol
[4] Sporn MB, Suh N. Chemoprevention of cancer. Cell Biochem 2003;254:257–63.
Carcinogenesis 2000;21:525–30. [23] Yu H, Jove R. The STATs of cancer: new molecular targets
[5] Bergman M, Varshavsky L, Gottlieb HE, Grossman S. The come of age. Nat Rev Cancer 2004;4:97–105.
antioxidant activity of aqueous spinach extract: chemical [24] Barré B, Avril S, Coqueret O. Opposite regulation of Myc and
identification of active fractions. Phytochemistry p21WAF1 transcription by STAT3 proteins. J Biol Chem
2001;58:143–52. 2003;278:2990–6.
 biochemical pharmacology 73 (2007) 56–67 67

[25] Vermes I, Haanen C, Steffens-Nakken H, Reutelingsperger [32] Yamazaki Y, Tsuruga M, Zhou D, Fujita Y, Shang X, Dang Y,
C. A novel assay for apoptosis. Flow cytometric detection of et al. Cytoskeletal disruption accelerates caspase-3
phosphatidylserine expression on early apoptotic cells activation and alters the intracellular membrane
using fluorescein labelled Annexin V. J Immunol Meth reorganization in DNA damage-induced apoptosis. Exp Cell
1995;184:39–51. Res 2000;259:64–78.
[26] Liang YC, Tsai SH, Chen L, Lin-Shiau SY, Lin JK. Resveratrol- [33] Bijl M, Horst G, Bijzet J, Bootsma H, Limburg PC,
G2 arrest through the inhibition of CDK7 and p34CDC2 Kallenberg CG. Serum amyloid P component binds to late
kinases in colon carcinoma HT29 cells. Biochem Pharmacol apoptotic cells and mediates their uptake by monocyte-
2003;65:1053–60. derived macrophages. Arthritis Rheum 2003;48:
[27] Molinari M. Cell cycle checkpoints and their inactivation in 248–54.
human cancer. Cell Prolif 2000;33:261–74. [34] Nihal M, Ahmad N, Mukhtar H, Wood GS. Anti-proliferative
[28] Hartwell LH, Kastan MB. Cell cycle control and cancer. and proapoptotic effects of ()-epigallocatechin-3-gallate
Science 1994;266:1821–8. on human melanoma: possible implications for the
[29] Martin SS, Leder P. Human MCF10A mammary epithelial chemoprevention of melanoma. Int J Cancer 2005;114:
cells undergo apoptosis following actin depolymerization 513–21.
that is independent of attachment and rescued by Bcl-2. [35] Al Yahya MA, Al Farhan AH, Adam SE. Preliminary toxicity
Mol Cell Biol 2001;21:6529–36. study on the individual and combined effects of Citrullus
[30] Maniotis AJ, Chen CS, Ingber DE. Demonstration of colocynthis and Nerium oleander in rats. Fitoterapia
mechanical connections between integrins, cytoskeletal 2000;71:385–91.
filaments, and nucleoplasm that stabilize nuclear [36] Bartalis J, Halaweish FT. Relationship between
structure. Proc Natl Acad Sci USA 1997;94:849–54. cucurbitacins reversed-phase high-performance liquid
[31] Parsons JT, Martin KH, Slack JK, Taylor JM, Weed SA. Focal chromatography hydrophobicity index and basal
adhesion kinase: a regulator of focal adhesion dynamics cytotoxicity on HepG2 cells. J Chromatogr B Anal Technol
and cell movement. Oncogene 2000;19:5606–13. Biomed Life Sci 2005;818:159–66.