Carcinogenesis vol.33 no.2 pp.413–419, 2012
Advance Access publication December 9, 2011
a-Mangostin, a xanthone from mangosteen fruit, promotes cell cycle arrest in prostate
cancer and decreases xenograft tumor growth
Jeremy J.Johnson?, Sakina M.Petiwala, Deeba N.Syed1,
John T.Rasmussen, Vaqar M.Adhami1, Imtiaz A.Siddiqui1,
Amanda M.Kohl1and Hasan Mukhtar1
Department of Pharmacy Practice, University of Illinois at Chicago College of
Pharmacy, 833 South Wood Street, Chicago, IL 60612-7230, USA and
1Department of Dermatology, University of Wisconsin School of Medicine
and Public Health, Madison, WI 53706, USA
?To whom correspondence should be addressed. Tel: þ1 312 996 4368;
Fax: þ1 312 996 0379;
There is a need to characterize promising dietary agents for
chemoprevention and therapy of prostate cancer (PCa). We exam-
ined the anticancer effect of a-mangostin, derived from the
mangosteen fruit, in human PCa cells and its role in targeting cell
cycle-related proteins involved in prostate carcinogenesis. Using an
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide as-
say, we found that a-mangostin significantly decreases PCa cell
viability in a dose-dependent manner. Further analysis using flow
cytometry identified cell cycle arrest along with apoptosis. To
establish a more precise mechanism of action, we performed a cell
free biochemical kinase assay against multiple cyclins/cyclin-
dependent kinases (CDKs) involved in cell cycle progression; the
mostsignificantinhibition in the cellfree-based assayswas CDK4,a
critical component of the G1 phase. Through molecular modeling,
we evaluated a-mangostin against the adenosine triphosphate-
binding pocket of CDK4 and propose three possible orientations
that may result in CDK4 inhibition. We then performed an in vivo
animal study to evaluate the ability of a-mangostin to suppress
tumor growth. Athymic nude micewere implantedwith22Rv1 cells
and treated with vehicle or a-mangostin (100 mg/kg) by oral
gavage. At the conclusion of the study, mice in the control cohort
had a tumor volume of 1190 mm3, while the treatment group had
a tumor volume of 410 mm3(P < 0.01). The ability of a-mangostin
to inhibit PCa in vitro and in vivo suggests a-mangostin may be
a novel agent for the management of PCa.
Of all cancers, prostate cancer (PCa) is ideal for evaluating chemo-
preventive agents for four main reasons: (i) it is typically diagnosed
in men .50 years old, (ii) its high latency period, (iii) upon initial
diagnosis patients often undergo ‘watchful waiting’ and (iv) it could
be targeted at various stages of disease development. For these reasons,
even a slight delay in the pathogenesis of PCa with chemoprevention
has the potential to result in a substantial reduction in the incidence of
PCa as well as significantly increase the quality of life in these patients.
Over the last several decades, epidemiological, human migratory
studies, preclinical and even early phase clinical trials have suggested
that selected dietary constituents may offer a protective effect in
reducing the incidence of multiple cancers, including cancer of the
prostate (1–3). Given the potential that some of these phytochemicals
have shown it is essential to further identify and develop promising new
agents in the hope of creating a broad spectrum of cancer chemopre-
ventive agents that could be used alone or in combination.
The purple mangosteen (Garcinia mangostana) is a slow growing
tropical tree native to India, Myanmar, Malaysia, Philippines, Sri
Lanka and Thailand and is reported to reach heights of 6–25 m
(20–80 ft) (4,5). The purple mangosteen (here after referred to as
mangosteen) is related to several other fruits including button
mangosteen (G. prainiana) and the lemon drop mangosteen
(G. madruno). For clarification, the mangosteen is unrelated to the
mango (Mangifera spp.). A class of compounds known as xanthones
have been isolated from the mangosteen with well over 60 different
xanthones isolated from the fruit, leaves, bark and roots (4,5). Avariety
of health-promoting attributes have been associated with the
mangosteen, which include antiinflammatory (6–8), antibacterial
activity (9), cardioprotective (10,11) and antioxidant activity (12–16).
Of all the xanthones, a-mangostin (1,3,6-Trihydroxy-7-methoxy-2,
8-bis(3-methyl-2-butenyl)-9H-xanthen-9-one) has been identified as
the most abundant xanthone and as a result has received the most
attention for its health-promoting properties.
In preclinical settings, crude mangosteen extract containing a-
mangostin and c-mangostin was shown to decrease preneoplastic lesions
in the colon of rat exposed to dimethylhydrazine (17). A significant
suppression in the development of aberrant crypt foci was evident in rats
administered a-mangostin as a part of their daily diet. A significant
reduction in dysplastic foci and b-catenin accumulated crypts was also
observed (P , 0.05). Recently, a-mangostin has been shown to inhibit
matrix metalloproteinase-2/9 through the c-jun N-terminal kinase (JNK)
signaling pathway in PC3 cells (18).
We hypothesized that a highlypurified form of a-mangostin (.95%)
may offer cancer chemopreventive and/or chemotherapeutic effects
against PCa. For our study, we evaluated a-mangostin for its ability
toinhibit PCa cell growth, induceapoptosis,inhibit deregulated kinases
in PCa and suppress tumor growth in 22Rv1 xenograft mice.
Materials and methods
a-Mangostin (Xanomax 95TM, .95%) was obtained from Avesthagen
(Chatsworth, CA), 3,3#-Diindolylmethane (.98%), Genistein (.98%), Epigallo-
catechin 3-gallate (.97%)was obtained from Sigma (St Louis, MO), 5-(N-(4-
?95%) and trans-4-((6-(ethylamino)-2-((1-(phenylmethyl)-1H-indol-5-yl)ami-
from EMD Chemicals (Gibbstown, NJ). All antibodies for western blot analysis
were obtained from Cell Signaling Technology (Danvers, MA). Protein assay kit
was obtained from Pierce (Rockford, IL). APO-DIRECTTMkit was obtained from
Phoenix Flow Systems (San Diego CA). Cleaved caspase-3 kit was obtained from
Cell Signaling Technology. General caspase inhibitor (z-vad-fmk) was obtained
from R&D Systems (Minneapolis, MN).
Cell culture and treatment
LNCaP, PC3, DU145 and 22Rv1 cells were obtained from American Type
Culture Collection (Manassas, VA). These cells were cultured in RPMI
(LNCaP, PC3 and 22Rv1) or DMEM (DU145) supplemented with 10% fetal
bovine serum and 1% penicillin/streptomycin. Human normal prostate epithe-
lial cells (PrEC) were obtained from Lonza (Basel, Switzerland) and grown
according to the manufacturer’s instructions. All cells were maintained under
standard cell culture conditions as described previously (19).
Cell viability was determined after 48 h by 3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyltetrazolium bromide (MTT) assay as described previously (20).
The APO-DIRECTTMkit (Phoenix Flow Systems) was used for measuring apo-
cells were 50–60% confluent and were serum depleted (i.e. 0.1%) for 36 h to
promote cell synchronization. Cells were then treated with complete media con-
taining a-mangostin for 24 and 48 h. The kit was followed per protocol directions.
22Rv1and PC3 cells were seeded in six well plates at a concentration of 1000
cells per well and incubated for 48 h. Cells were then treated with increasing
Abbreviations: ATP, adenosine triphosphate; CDK, cyclin-dependent kinase;
MTT, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide; PCa,
prostate cancer; PrEC, prostate epithelial cells.
? The Author 2011. Published by Oxford University Press. All rights reserved. For Permissions, please email: firstname.lastname@example.org
concentrations of a-mangostin (2.5–20 lM). Fresh media containing a-
mangostinwasreplacedevery 3–4 daysuntilthe conclusionof thestudy, which
was 15 days after initiating the study. Media was removed and wells were
briefly rinsed with 1X phosphate-buffered saline and removed. Colonies were
then stained with crystal violet in triplicate and colonies were counted visually.
After 24 h of cell growth, the complete media was removed and replaced with
media containing a-mangostin with the appropriate assay concentrations for
the whole-cell lysates as depicted in figures. Cell lysates were prepared as
previously described (20).
a-Mangostin was diluted into assay mixture where concentrations of a-
mangostin ranged from 1 to 10 lM of a-mangostin to determine inhibition
of JNK1, JNK2, JNK3, CyclinA1/CDK2, CyclinA2/CDK2, CyclinD1/
CDK4, CyclinD3/CDK6 (Kinexus, Vancouver, British Columbia, Canada).
The IC50was determined for a-mangostin against CyclinD1/CDK4 with six
dilutions ranging from 2.5 to 15 lM. The assay was initiated by the addition
of [33P]-ATP and the reaction mixture incubated at room temperature for 20–
40 min. After the incubation period, the assay was terminated by spotting 10
ll of the reaction mixture onto multiscreen phosphocellulose P81 plate. The
multiscreen phosphocellulose P81 plate was washed three times for ?15 min
each in a 1% phosphoric acid solution. The radioactivity on the P81 plate was
counted in the presence of scintillation fluid in a Trilux scintillation counter.
Protein kinase-specific activity of [33P]-ATP incorporated per minute per
sample was determined. A total counts per minute for each reaction sample
was determined for blanks (without substrate), control (without a-mangos-
tin) and samples (with a-mangostin). The corrected activity for control sam-
ples (i.e. without a-mangostin) represented 100% kinase activity and was
used to determine the percent of kinase activity.
In vivo 22Rv1tumor xenograft model
Athymic (nu/nu) malenudemice(Harlan Laboratory, Madison, WI)7–8 weeks
old were housed under pathogen-free conditions with a 12 h light/12 h dark
schedule and fed with an autoclaved AIN-76A diet ad libitumas described
previously (21). 22Rv1 cells were used for determining the in vivo effects of
a-mangostin based on the fact that these cells form rapid and reproducible
tumors in nude micewith tumor xenografts established as described previously
(21). Fourteen animals were randomly divided into two groups, with
seven animals in each group. The animals in group 1 received vehicle (100
ll) by oral gavage and served as control. The animals in group 2 received
a-mangostin (100 mg/kg)by oral gavage five timesweekly. Body weights were
recorded once weekly throughout the study. All procedures conducted were in
accordance with the guidelines for the use and care of laboratory animals.
All statistical analysis was performed by using VassarStats software. Data are
expressed as mean with standard deviation for all groups. Statistical signifi-
cance of differences in all measurements between control and treated groups
was determinedby one-way analysis of variancefollowed by Tukey’s HSD test
for multiple comparisons. Student’s paired t test was used for pair wise group
comparisons, as needed. All statistical tests were two-sided, and P , 0.05 was
considered statistically significant.
a-Mangostin decreases PCa cell viability
PCa cells (LNCaP, 22Rv1, DU145 and PC3) were treated for 48 h
with a-mangostin, Epigallocatechin 3-gallate, 3,3#-Diindolylme-
thane, and genistein and evaluated for decreasing PCa cell viability
using the MTT assay. These cells were selected based on their
Fig. 1. (A) Chemical structure of a-mangostin and nomenclature of a xanthone. (B) Cell viability assay was performed treating cells with a-mangostin,
Epigallocatechin 3-gallate, Genistein and 3,3#-Diindolylmethane up to 70 lM against four PCa cell lines (LNCaP, 22Rv1, DU145 and PC3) for 48 h. Data points
are represented by the average of three values with standard deviation and representative of two different experiments.
J.J.Johnson et al.
response to androgens with LNCaP being androgen dependent, 22Rv1
being androgen independent but androgen sensitive and both DU145
and PC3 being androgen independent. The IC50of a-mangostin was
calculated to be 5.9, 6.9, 22.5 and 12.7 lM in all four cell lines
(LNCaP, 22Rv1, DU145 and PC3, respectively) (Figure 1B).
a-Mangostin inhibits colony formation in PCa cells
Next, we evaluated a-mangostin for decreasing the clonogenic
potential of 22Rv1 and PC3 cells. 22Rv1 and PC3 cells were seeded
at a concentration of 1000 cells per well and treated with increasing
concentrations of a-mangostin every 3–4 days. We observed a
statistically significant (P , 0.01) decrease in colony formation at
doses of 2.5 lM in both cell lines and a complete inhibition of colony
formation at 10 lM after 12–14 days (Figure 2A). a-Mangostin was
not found to inhibit normal human PrEC using a-mangostin at 7.5 and
15 lM after 48 h as shown in Figure 2B. These cells were also
evaluated for viability using MTT at 5, 10, 15, 20, 25, 30 and 35
lM a-mangostin that had 104.7% (±5.2), 104.6% (±1.6), 103.7%
(±3.0), 103.3% (±9.7), 58.3% (±7.2), 51.5% (±5.2), 51.1% (±5.2) cell
viability compared with control PrEC.
a-Mangostin induces G1cell cycle arrest in PCa cells
22Rv1 and PC3 cells treated with a-mangostin for 24 h were evaluated
by flow cytometry for cell cycle arrest. a-Mangostin resulted in a
statistically significant (P , 0.01) higher number of cells in G1. As
summarized in Figure 2C, after 24 h, the number of 22Rv1 cells in G1
Fig. 2. (A) Colony formation was performed using 22Rv1 and PC3 cells. For
colony formation, cells were plated at ?1000 cells per well and incubated for
48 h. After 48 h, mediawas replaced with fresh media containing a-mangostin
andwasrepeatedevery 3–4daysuntil completionof the experiment.Statistical
analysis performed using VassarStats software by one way analysis of variance
and the Tukey test. ‘a’ versus control, ‘b’ versus 2.5 lM, ‘c’ versus 5 lM.
(B) Phase contrast microscopic images (10x and 40x) of PrECs, PC3 and
22Rv1 cells with control and a-mangostin-treated cells (7.5 and 15 lM) are
shown. Cells were treated for 48 h. (C) Cell cycle analysis was performed per
manufacturer’s protocol using synchronized PC3 cells. Cells were
synchronized for 36 h and then treated with a-mangostin for 24 h. These
experiments were performed in triplicate and are represented by the mean
along with standard deviation. These results were consistent with 22Rv1 cells
treated with a-mangostin. Statistical analysis was performed using VassarStats
software by one way analysis of variance and statistical significance was
performed by the Tukey test with?P , 0.01 and??P , 0.001.
Fig. 3. (A)PC3and22Rv1cellsweretreatedwitha-mangostinfor24and48h
and evaluated for apoptosis per manufacturer’s protocol. Cells were
synchronized for 36 h and then treated with a-mangostin for 24 h. These
with standard deviation. Statistical analysis was performed using VassarStats
software by one way analysis of variance and statistical significance was
performed by the Tukey test with?P , 0.01.‘a’ versus control, ‘b’ versus 7.5
lM. (B) 22Rv1 cells were grown to 60–70% confluence and treated with
a-mangostin up to 24 h and whole cell lysates were prepared as described in
Materials and methods. Seventy micrograms of protein was subjected to
sodium dodecyl sulfate–polyacrylamide gel electrophoresis followed by
western blot analysis and chemiluminescence detection. Equal loading of
(C) 22Rv1 cells were treated with a-mangostin for 24 h and whole cell lysates
were collected to evaluate for active caspase-3 protein by enzyme-linked
immunosorbent assay. These experiments were performed in triplicate and are
represented by the mean along with standard deviation. Statistical analysis was
performed using VassarStats software by one way analysis of variance and
statistical significance was performed by the Tukey test with?P , 0.01. ‘a’
versus control, ‘b’ versus 7.5 lM.D, 22Rv1 cells were treated with general
caspase inhibitor z-vad-fmk (100 lM) for 2 h and then treated with
a-mangostin for 18 h. Whole cell lysates were collected to evaluate for active
caspase-3 protein by enzyme-linked immunosorbent assay. These experiments
were performed in triplicate and are represented by the mean along with
a-Mangostin targets CDK4
phase increased from the control (42.3% ± 1.19)comparedwith7.5 lM
a-mangostin (59.6%± 3.09; P , 0.01) and 15 lM a-mangostin (62.4%
± 2.4; P , 0.01). In addition, the number of PC3 cells in G1increased
from the control (37.35% ± 0.77) compared with 7.5 lM a-mangostin
(55.24% ± 2.51; P , 0.01) and 15 lM a-mangostin (53.06% ± 1.09;
P , 0.01). These results suggest that an inhibition in cell viability may
be associated with the induction of cell cycle arrest.
a-Mangostin induces apoptosis in PCa cells
Using flow cytometry, we evaluated PC3 and 22Rv1 cells treated with
a-mangostin for apoptosis by end labeling DNA fragments with
fluorescent tagged deoxyuridine triphosphate nucleotides (F-dUTP).
After 48 h, treatment with a-mangostin (15 lM) resulted in 38.56%
± 12.93 (P , 0.01) of the cells undergoing apoptosis (Figure 3A). We
also observed similar results with 22Rv1 cells treated with a-mangostin
57.0% ± 26.9 cells undergoing apoptosis at 15 lM, respectively (P ,
0.01). By western blot, we observed a decrease in pro-caspase-3 and an
increase in the active form of caspase-3 in 22Rv1 cells (Figure 3B)
allowing for a qualitative confirmation that apoptosis was occurring. To
accurately quantify the extent of caspase-3 activation, we performed an
enzyme-linked immunosorbent assay to detect activated caspase-3 in
in the cleaved form of caspase-3 (Figure 3C) at 7.5 and 15 lM by 345
and 894%, respectively, compared with control cells. When PCa cells
treatment of a-mangostin, the cleavage of caspase-3 to its active form
was inhibited (Figure 3D).
a-Mangostin inhibits kinases in a cell free biochemical based-assay
Using a cell free-based kinase assay, we evaluated a-mangostin (1 and
10 lM) for its ability to inhibit several kinases that included the JNK
and cyclin/cyclin-dependent kinase (CDK) proteins (Figure 4A).
A significant inhibition (P , 0.01) was observed at the following
cyclins/CDKs: cyclinA1/CDK2 (34.2% ± 0.1) and cyclinA2/CDK2
(13.2% ± 0.4), cyclinD1/CDK4 (59.2 ± 1.4) and cyclinD3/CDK6
(5.5% ± 0.3) with 10 lM a-mangostin. No direct inhibition of kinase
activity was observed with treatment of JNK1 and JNK2 at either 1 or
10 lM. A slight but significant decrease (P , 0.01) in kinase activity
at 10 lM was observed with JNK3 (7% ± 2.8). The greatest decrease
in kinase activity by a-mangostin was observed with cyclinD1/CDK4.
As a result, cyclinD1/CDK4 was evaluated in further studies. The
EC50 of a-mangostin and cyclinD1/CDK4 was determined to be
7.4 lM (Figure 4B).
a-Mangostin modulates proteins involved in cell cycle in PCa cells
Next, we assessed the effect of a-mangostin to inhibit proteins
upstream of CDK4 that included the INK proteins (p15 INK4B,
p16 INK4A), p27 Kip1 and CDK4 (Figure 5A) in 22Rv1 cells. An
increase in p27 Kip1 was observed in cells treated with a-mangostin,
however no effect was observed on p15 INK4B, p16 INK4A or
CDK4. A decrease in protein expression was observed with both
cyclins D1 and D3. Next, we evaluated the downstream targets of
CDK4, specifically the phosphorylation sites of retinoblastoma
protein. A dose-dependent decrease in the levels of phosphorylated
retinoblastoma protein at residues Ser780, Ser795 and Ser807/811
was observed (Figure 5B). The retinoblastoma protein has been
shown to regulate the expression of cyclin E and we were able to
observe a decrease in protein expression of cyclin E consistent with
the notion that a-mangostin decreases the phosphorylated retinoblas-
CDK4 inhibitors decrease cell viability
Using an MTT cell viability assay, 22Rv1 cells were treated with
a CDK4 inhibitor and a CDK4/6 inhibitor for 48 h (Figure 5C).
Interestingly, we found that the CDK4 inhibitor increased the cell
viability at a lower dose of 2.5 lM to 134.9% ± 0.9 followed by
a decrease to 41.2% ± 4.5 when treated with 5 lM with an estimated
IC50of 4.4 lM. Similarly, we found that a CDK4/6 inhibitor increased
the cellular viability at 2.5–10 lM followed by a decrease to 23.7% ±
1.3 at 20 lM with an estimated IC50of 16.7 lM. These stimulatory
effects were not observed when cells were treated with a-mangostin.
In addition, we performed in silico docking using the CDK4 and
CDK4/6 inhibitors with the ATP binding pocket of CDK4. Using
AutoDock 4.2, the CDK4 inhibitor was found to be positioned with
the methoxy groups going into the binding pocket first, possibly
owing to their hydrophobic properties. The cyclohexyl-OH functional
group on the CDK4/6 inhibitor was found to position itself deep into
the binding pocket with the phenyl group positioned out into solution.
a-Mangostin/CDK 4-binding hypothesis by molecular modeling
Next, through in silico docking using AutoDock 4.2,we evaluated 30
possible binding configurations of a-mangostin with the binding
pocket of CDK4 to generate a binding hypothesis. The structure of
CDK4 in the present study is from the crystal structure of cyclinD1/
CDK4 and was used for molecular modeling studies (22). From the 30
different arrangements, several clusters were generated, and we have
presented the three most common clusters designating them cluster
1 (yellow), cluster 2 (green) and cluster 3 (gold) (Figure 5D). In
cluster 1, the isoprenyl group at the second carbon of a-mangostin
was found to position itself deep within the hydrophobic ATP-binding
pocket of CDK4 positioning itself near the phenylalanine-93. In
clusters 2 and 3, the opposite end of a-mangostin (i.e. carbons 5–8)
Fig. 4. (A) Cell free biochemical-based kinase assay was evaluated
(a-mangostin 1 and 10 lM) against JNK 1/2/3, CylinA1/CDK2, cyclinA2/
CDK2, CyclinD1/CDK4 and CyclinD3/CDK6. Statistical analysis was
performed using VassarStats software by one way analysis of variance and
statistical significance was performed by the Tukey test with?P , 0.01.
(B) Cell free biochemical kinase assay established the IC50of a-mangostin
against CyclinD1/CDK4. Data points are represented by the average of three
values with standard deviation. As a positive control staurosporine, a potent
microbial-derived general kinase inhibitor was used (data not shown).
Statistical analysis was performed using VassarStats software by one way
analysis of variance and statistical significance was performed by the Tukey
test with?P , 0.01.‘a’versus control, ‘b’versus 2.5 lM, ‘c’versus 5 lM, ‘d’
versus 7.5 lM, ‘e’ versus 10 lM.
J.J.Johnson et al.
went into the binding pocket first. In the second and third cluster, there
appears to be limitations in that they do not fill the binding pocket of
CDK4 as well as cluster 1 making them less favored. In each of the
different cluster arrangements, the hydrogen bonding could occur
with the three hydroxyl groups on a-mangostin at aspartate-99,
arginine-101 and threonine-102. The carbonyl group did not interact
with any amino acids but rather seemed to stabilize the hydroxyl
group at the first carbon of a-mangostin.
a-Mangostin inhibits the growth of human prostate carcinoma 22Rv1
cells in athymic nude mice
Athymic nude micewere implanted with 22Rv1 cells and divided into
two cohorts receiving vehicle or a-mangostin (100 mg/kg). We found
a-mangostin was as tolerable as the vehicle throughout the experi-
ment, this was evidenced by body weight measurements (Figure 6A).
In the control cohort, tumors were measurable on day 14 while the
cohort treatment with a-mangostin at 100 mg/kg did not have
Fig. 5. (A and B) 22Rv1 cells were grown to 60–70% confluence and treated with a-mangostin up to 24 h and whole cell lysates were prepared as described in
Materials and methods. Forty micrograms of protein was subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis followed by western blot analysis
and chemiluminescence detection. Equal loading of protein was confirmed by stripping the immunoblot and reprobing it for b-actin. (C) Cell viability assay was
performed treating cells with 5-(N-(4-Methylphenyl)amino)-2-methyl-4,7-dioxobenzothiazole (CDK4 Inhibitor) and trans-4-((6-(ethylamino)-2-((1-(phenylmethyl)-
1H-indol-5-yl)amino)-4-pyrimidinyl)amino)-cyclohexanol(CDK6 Inhibitor) up to 20 lM against 22Rv1 cells for 48 h. Data points are represented by the average of
three values with barsrepresenting standard deviation. Molecular modeling of CDK4 interacting with CDK4 and CDK4/6 inhibitor shown at the ATP binding site and
wasanalyzed usingAutodock4.2withsurfacestructure imagescreatedusingPyMOL.(D) MolecularmodelingofCDK4interacting witha-mangostinisshownatthe
ATPbindingsiteandwasanalyzedusing Autodock4.2 with surfacestructure imagescreatedusing PyMOL (protein colors: gray 5 carbon,blue5 nitrogenandred 5
oxygen). Thirty different binding configurations were created with the three highest scoring configurations presented. a-Mangostin is presented as a different color in
each configuration: cluster 1 (yellow), cluster 2 (green) and cluster 3 (gold).
a-Mangostin targets CDK4
measurable tumors until day 18 (Figure 6C). The average computed
tumor volume was significantly inhibited in mice receiving a-
mangostin and can be visualized in Figure 6B. In the control group,
the average tumor volume of 1190 mm3was reached on day 34 after
tumor cell inoculation, whereas mice receiving a-mangostin had an
average tumor volume of 410 mm3(70 mg/kg). This represents a
significant suppression in tumor growth by 65% (P , 0.01), respec-
tively, compared with control.
PCa is most often defined by a slow progression through well-defined
stages beginning with hyperplasia through adenocarcinoma. This
slow progression represents an important opportunity to either stop
or delay the progression of this disease. One strategy to manage this
progression is through chemoprevention using either dietary or
synthetic agents that can delay, prevent or reverse the process of
carcinogenesis. In our study, we observed a-mangostin to promote
cell cycle arrest by targeting CDK4 using a cell free kinase assay and
confirmed the promotion of cell cycle arrest in PC3 and 22Rv1 cells.
Furthermore, a-mangostin promoted apoptosis through the activation
of caspase-3 in 22Rv1 cells. Using athymic nude mice, 22Rv1 xeno-
grafts, we observed that the oral administration of a-mangostin
resulted in a 65% smaller tumor compared with vehicle-treated mice.
Ideally, new chemical entities that are being evaluated as an
investigational drug should have little to no toxicity. Xanthones
isolated from the pericarp of the mangosteen are not genotoxic in
mutagenesis studies (23). In addition, a-mangostin (10 lM) was not
found to be toxic to human peripheral blood lymphocytes even though
identical treatment in leukemia cell lines lead to apoptosis (24). In our
results, by MTT assay, we did not observe a-mangostin to decrease
cell viability of PrEC compared with PCa cells suggestive of selec-
tivity. In two different cardioprotective studies, a-mangostin was
administered 200 mg/kg by oral gavage to rats for up to 8 days with
no observable adverse effects in solid organ systems (10,11). In
another study, a-mangostin was tolerable with no observable adverse
effects during a 5 weeks intervention with custom-blended food
pellets that contained 0.02 or 0.05% a-mangostin in F344 rats (17).
In addition, mangosteens that contain a-mangostin are consumed by
humans in the forms of juices (e.g. Xango?), dietary supplements
standardized to a-mangostin (e.g. Nature’s Way? 80 mg a-mangos-
tin/capsule) and as a fruit. Investigators have shown that when rats are
administered 40 mg/kg body wt of a-mangostin in corn oil the max-
imum plasma concentration was 11.7 lM, suggesting that under those
conditions a-mangostin is well absorbed (25).
Multiple molecular pathways have been shown to be deregulated in
PCa including the CDKs. These enzymes are responsible for promot-
ing cell cycle progression and transcription of genes involved in cell
replication and survival. It has been well established that early on in
prostate carcinogenesis, a deregulation of CDKs/cyclins occurs lead-
ing to the phosphorylation of retinoblastoma proteins leading to an
inability to regulate the cell cycle. One of the earlier events in carci-
nogenesis is the loss of cell cyclecontrol(26) and would seem tobe an
ideal target for cancer chemoprevention. Another potential benefit of
targeting these cell cycle regulatory targets is that cell cycle deregu-
lation occurs over the full spectrum of cancers. Recently, investigators
have begun to evaluate CDK inhibitors as chemopreventive agents
and have been able to successfully show that CDK inhibitors are
effective in preventing colon tumorigenesis in a mouse model (27).
The results presented in this study suggest that a-mangostin directly
inhibits CDK4 as shown in cell-free biochemical kinase assays. Fur-
thermore, using molecular modeling, we have generated hypothetical
binding arrangements between CDK4 and a-mangostin. It is tempting
to speculate that a-mangostin due to its flat/planar structure, hydro-
phobic isoprenyl groups, along with neighboring hydrogen-bonding
donors at positions 1 and 3 are able to fit within the deep and narrow
ATP binding pocket. Based on our analysis, it would seem that the
isoprenyl groups may offer an advantage to other nonisoprenylated
chemicals by encouraging interactions with amino acids deep within
the hydrophobic molecular binding pocket. Furthermore, the isopren-
yl groups may allow for flexible docking and subsequent better filling
of the ATP binding pocket. It is important to note, based on our
modeling, there still appears to be room to fill the pocket possibly
affording an opportunity to further modify a-mangostin in order to
develop more effective CDK4 inhibitors. Alternatively, there may be
other xanthones available as natural products from the mangosteen
fruit that are better inhibitors of CDK4. Cocrystallization studies of
a-mangostin with CDK4 are needed to determine how a-mangostin
interacts with CDK4 and determine if it is a competitive inhibitor or
an allosteric inhibitor of CDK4 or even a combination of the two. The
retinoblastoma protein has up to 14 different phosphorylation sites
and further analysis by mass spectrometry is needed to determine
which phosphorylation sites are specifically inhibited. Once this in-
hibition profile is established a side-by-side comparison with other
CDK4 inhibitors (i.e. xanthones and nonxanthones) is needed to iden-
tify any unique properties of a-mangostin. Alternatively, there may be
other xanthones that target pathways other than CDK4/cyclin D1/
retinoblastoma encouraging a multitargeted strategy.
Previously, a-mangostin has been shown to suppress JNK protein
expression leading to an inhibition of matrix metalloproteinases-2/9
in PCa cells (18). We did not observe a strong direct effect of a-
mangostin directly inhibiting JNK kinase activity which taken together
suggests that a-mangostin may target proteins upstream of JNK. Sev-
eral studies have reported the proapoptotic effects of a-mangostin in
colon and leukemia cell lines (24,28–30). This was in agreement with
our analysis of PCa cells treated with a-mangostin where we observed
the activation of the caspase cascade followed by apoptosis in both PC3
and 22Rv1 cells after 24 and 48 h. Another interesting observation is
that a-mangostin appeared to be more effective at decreasing cell
viability in LNCaP cells and 22Rv1 cells which have mutated receptors
at T877A and H874Y, respectively. Based on these results, we are
evaluating the targeting of androgen receptor with a-mangostin.
In summary, the major finding of the present study is the demon-
stration that a-mangostin, derived from the purple mangosteen fruit,
inhibits CDK4 leading to a decrease in the phosphorylation of the
retinoblastoma proteins leading to G1cell cycle arrest in 22Rv1PCa
Fig. 6. (A) Body weights of athymic nude mice treated with a-mangostin
were measured one time weekly. (B) Fourteen animals were subcutaneously
injected in each flank of the mouse with ?1.5 ? 10622Rv1 cells to initiate
tumor growth. Twenty-four hours after cell implantation, the animals in each
cohort received by oral gavage vehicle or a-mangostin (100 mg/kg). The
average tumor volume of control and a-mangostin treated mice were plotted
over days after tumor cell inoculation. Data points represent the mean of 14
tumors from seven mice; bars represent standard deviation of the mean,?P ,
0.01. (C) Four representative tumors from each cohort at the end of the study
are shown comparing control mice to mice treated with a-mangostin.
J.J.Johnson et al.
cells at doses as low as 7.5 lM after 24 h of treatment. Apoptosis was
observed at 15 lM after 48 h of treatment. This data suggests that
lower sustained doses will promote cell cycle arrest, whereas higher
doses have the potential to promote apoptosis, however further work
is needed to characterize this. We observed that a-mangostin signif-
icantly suppressed tumor formation in nude mice implanted with
22Rv1 cells. This is an important observation in that many patients
diagnosed with PCa that are undergoing watchful waiting may be
excellent candidates for chemopreventive agents that delay and slow
the process of carcinogenesis. Together, these findings provide a more
thorough molecular understanding to previous reports (24,28,29) of
a-mangostin decreasing cell viability and promoting G1cell cycle
arrest. Given the evidence of a-mangostin’s anticancer activity,
and/or chemopreventive agent.
Clinical and Translational Science Award (CTSA) KL2 program NIH
1KL2RR025012-01 and 1R03CA138953-01A1 (to J.J.J); US PHS
Grant T32AR055893 (to I.A.S); NIEHS Grant T32ES007015 (to
Conflict of Interest Statement: None declared.
1.Johnson,J.J. (2011) Carnosol: a promising anti-cancer and anti-inflammatory
agent. Cancer Lett., 305, 1–7.
2.Johnson,J.J. et al. (2010) Green tea polyphenols for prostate cancer chemo-
prevention: a translational perspective. Phytomedicine, 17, 3–13.
3.Johnson,J.J. et al. (2007) Curcumin for chemoprevention of colon cancer.
Cancer Lett., 255, 170–181.
4.Obolskiy,D. et al. (2009) Garcinia mangostana L.: a phytochemical and
pharmacological review. Phytother. Res., 23, 1047–1065.
5.Pedraza-Chaverri,J.etal.(2008)Medicinal propertiesof mangosteen(Garcinia
mangostana). Food Chem. Toxicol., 46, 3227–3239.
6.Chen,L.G. et al. (2008) Anti-inflammatory activity of mangostins from
Garcinia mangostana. Food Chem. Toxicol., 46, 688–693.
7.Chomnawang,M.T. et al. (2007) Effect of Garcinia mangostana on inflam-
mation caused by Propionibacterium acnes. Fitoterapia, 78, 401–408.
8.Gopalakrishnan,C. et al. (1980) Effect of mangostin, a xanthone from Garci-
nia mangostana Linn. in immunopathological & inflammatory reactions.
Indian J. Exp. Biol., 18, 843–846.
9.Sundaram,B.M. et al. (1983) Antimicrobial activities of Garcinia mangos-
tana. Planta Med., 48, 59–60.
10.Devi Sampath,P. et al. (2007) Cardioprotective effect of alpha-mangostin,
a xanthone derivative from mangosteen on tissue defense system against iso-
11.Sampath,P.D. et al. (2008) Ameliorative prospective of alpha-mangostin,
a xanthone derivative from Garcinia mangostana against beta-adrenergic
cathecolamine-induced myocardial toxicity and anomalous cardiac TNF-
alpha and COX-2 expressions in rats. Exp. Toxicol. Pathol., 60, 357–364.
12.Jung,H.A. et al. (2006) Antioxidant xanthones from the pericarp of
Garcinia mangostana (Mangosteen). J. Agric. Food Chem., 54, 2077–2082.
13.Marquez-Valadez,B. et al. (2009) The natural xanthone alpha-mangostin
reduces oxidative damage in rat brain tissue. Nutr. Neurosci., 12, 35–42.
14.Pedraza-Chaverri,J. et al. (2009) ROS scavenging capacity and neuropro-
tective effect of alpha-mangostin against 3-nitropropionic acid in cerebellar
granule neurons. Exp. Toxicol. Pathol., 61, 491–501.
15.Weecharangsan,W. et al. (2006) Antioxidative and neuroprotective activi-
ties of extracts from the fruit hull of mangosteen (Garcinia mangostana
Linn.). Med. Princ. Pract., 15, 281–287.
16.Williams,P. et al. (1995) Mangostin inhibits the oxidative modification of
human low density lipoprotein. Free Radic. Res., 23, 175–184.
17.Nabandith,V. et al. (2004) Inhibitory effects of crude alpha-mangostin,
a xanthone derivative, on two different categories of colon preneoplastic
lesions induced by 1, 2-dimethylhydrazine in the rat. Asian Pac. J. Cancer
Prev., 5, 433–438.
18.Hung,S.H. et al. (2009) Alpha-mangostin suppresses PC-3 human prostate
carcinoma cell metastasis by inhibiting matrix metalloproteinase-2/9 and
urokinase-plasminogen expression through the JNK signaling pathway.
J. Agric. Food Chem., 57, 1291–1298.
19.Suh,Y. et al. (2009) A plant flavonoid fisetin induces apoptosis in colon
cancer cells by inhibition of COX2 and Wnt/EGFR/NF-kappaB-signaling
pathways. Carcinogenesis, 30, 300–307.
20.Johnson,J.J. et al. (2008) Carnosol, a dietary diterpene, displays growth
inhibitory effects in human prostate cancer PC3 cells leading to G2-phase
cell cycle arrest and targets the 5#-AMP-activated protein kinase (AMPK)
pathway. Pharm. Res., 25, 2125–2134.
21.Johnson,J.J. et al. (2010) Disruption of androgen and estrogen receptor
activity in prostate cancer by a novel dietary diterpene carnosol: implica-
tions for chemoprevention. Cancer Prev. Res. (Phila.), 3, 1112–1123.
22.Day,P.J. et al. (2009) Crystal structure of human CDK4 in complex with
a D-type cyclin. Proc. Natl Acad. Sci. USA, 106, 4166–4170.
23.Settheetham,W. et al. (1995) Study of genotoxic effects of antidiarrheal
medicinal herbs on human cells in vitro. Southeast Asian J. Trop. Med.
Public Health, 26 (suppl. 1), 306–310.
24.Matsumoto,K. et al. (2003) Induction of apoptosis by xanthones from
mangosteen in human leukemia cell lines. J. Nat. Prod., 66, 1124–1127.
25.Syamsudin,F.aizatun et al. (2009) HPLC analysis and pharmacokinetic
study of mangostin after oral administration in rats. Pharm. Res., 2, 43–49.
26.Syljuasen,R.G. et al. (1999) Loss of normal G1 checkpoint control is an
early step in carcinogenesis, independent of p53 status. Cancer Res., 59,
27.Boquoi,A. et al. (2009) Chemoprevention of mouse intestinal tumorigene-
sis by the cyclin-dependent kinase inhibitor SNS-032. Cancer Prev. Res.
(Phila.), 2, 800–806.
28.Matsumoto,K. et al. (2005) Xanthones induce cell-cycle arrest and apoptosis
in human colon cancer DLD-1 cells. Bioorg. Med. Chem., 13, 6064–6069.
29.Matsumoto,K. et al. (2004) Preferential target is mitochondria in alpha-
mangostin-induced apoptosis in human leukemia HL60 cells. Bioorg. Med.
Chem., 12, 5799–5806.
30.Menasria,F. et al. (2008) Apoptotic effects on B-cell chronic lymphocytic
leukemia (B-CLL) cells of heterocyclic compounds isolated from Guttiferaes.
Leuk. Res., 32, 1914–1926.
Received December 9, 2010; revised November 23, 2011;
accepted December 3, 2011
a-Mangostin targets CDK4