Discovery, optimization, and pharmacophore modeling of oleanolic acid and analogues as breast cancer cell migration and invasion inhibitors through targeting Brk/Paxillin/Rac1 axis

Article abstract of DOI:10.1111/cbdd.12380

Bioassay-guided fractionation of Terminalia bentzoe L. leaves methanol extract identified the known triterpene oleanolic acid (1) as its major breast cancer cell migration inhibitor. Further chemical optimization afforded five new (9-12 and 15) and seven

Full text of DOI:10.1111/cbdd.12380

Chem Biol Drug Des 2014
Research Article
Discovery, Optimization, and Pharmacophore Modeling
of Oleanolic Acid and Analogues as Breast Cancer Cell
Migration and Invasion Inhibitors Through Targeting
Brk/Paxillin/Rac1 Axis
Heba E. Elsayed1,2, Mohamed R. Akl1, Hassan Y.
as the second leading cause of cancer mortality in women
(1,2). Breast cancer metastatic cascade involves compli-
cated events, including cell invasion of the basement mem-
brane, migration through the extracellular matrix, and
Ebrahim1, Asmaa A. Sallam1, Eman G. Haggag2,
Amel M. Kamal2 and Khalid A. El Sayed1,*
1Department of Basic Pharmaceutical Sciences, School of
transport by blood vessels and/or lymphatics to distant
Pharmacy, University of Louisiana at Monroe, Monroe, LA
71201, USA
organs (3). Targeting molecular signaling pathways underly-
ing cancerous cell migration and invasion is among the
promising strategies to control metastatic malignancies.
2Department of Pharmacognosy, Faculty of Pharmacy,
Helwan University, Helwan, Cairo 11795, Egypt
*Corresponding author: Khalid A. El Sayed,
elsayed@ulm.edu
The vast majority of current anticancer drugs are origi-
nated from or inspired by natural entities (4). Natural prod-
uct secondary metabolites are characterized by their
structural novelty and diversity, qualifying them as promis-
ing scaffolds for the discovery of new anticancer entities
(5). The abundance of bioactive plant metabolites offers
the advantage for bioguided investigation of their extracts
to identify pure bioactive compounds for further optimiza-
tion. Plants of genus Terminalia (Combretaceae) are native
to Africa and are widely distributed in tropical and subtrop-
ical regions (6). They are known as a rich source of diverse
bioactive secondary metabolites, including phenolics and
triterpenes (7,8). Terminalia bentzoe L. is a medium-sized
tree, which is neither chemically nor biologically studied,
except for its essential oil content (9).
Bioassay-guided fractionation of Terminalia bentzoe L.
leaves methanol extract identified the known triterpene
oleanolic acid (1) as its major breast cancer cell migra-
tion inhibitor. Further chemical optimization afforded
five new (9–12 and 15) and seven known (4–8, 13, and
14) semisynthetic analogues. All compounds were
tested for their ability to inhibit human breast cancer
MDA-MB-231 cells migration, proliferation, and inva-
sion. The results revealed that 3-O-[N-(30-chloro-
benzenesulfonyl)-carbamoyl]-oleanolic acid (11) and
3-O-[N-(50-fluorobenzenesulfonyl)-carbamoyl]-oleanolic
acid (12) were the most active hits at low lM concen-
tration. Western blot analysis indicated the activity of
1, 11, and 12 might be related, at least in part, to the
suppression of Brk/Paxillin/Rac1 signaling pathway.
Pharmacophore modeling study was conducted to bet-
ter understand the common structural binding epitopes
important for the antimigratory activity. The sulfonyl
carbamoyl moiety with an optimal bulkiness electron-
deficient phenyl ring is associated with improved activ-
ity. This study is the first to discover the antimigratory
and anti-invasive activities of oleanolic acid and ana-
logues through targeting the Brk/Paxillin/Rac1 axis.
Oleanolic acid (1) is a pentacyclic triterpene widely distrib-
uted in food and traditional herbal remedies and exhibits
diverse therapeutic effects (10). The anticancer activity of
1 was reported to be mediated through different mecha-
nisms including the inhibition of mitogen-activated protein
kinase/extracellular signal-regulated kinase (MAPK/ERK)
pathway and inhibition of topoisomerases I and II (11,12).
Oleanolic acid has been subjected to various chemical
modifications to uncover its structure–activity relationship
and optimize its anticancer effect. Various substituents
were linked at different positions, while the pentacyclic core
skeleton was retained. Examples of prior 1 modification are
the derivatizations of C-3 hydroxyl with acids bearing differ-
ent aliphatic and heterocyclic side chains, which showed
improved antitumor activities (13). Other examples include
the 2-substituted-1-en-3-one derivatives like 2-cyano-
3,12-dioxooleana-1,9(11)-dien-28-oic acid (CDDO) as well
as its C-28 methyl ester that were synthesized by Honda
et al. (14). Expanding ring A system by incorporating an
Key words: bioassay-guided, breast cancer, Brk/Paxillin/Rac1
axis, Combretaceae, oleanolic acid, Terminalia bentzoe L
Received 8 March 2014, revised 7 May 2014 and accepted
for publication 7 June 2014
Breast cancer is the most common malignancy among the
female population worldwide (1). The high metastatic ability
of some breast cancer phenotypes, including tumor cell dis-
semination to bones, lungs, or brain, ranked breast cancer
ª 2014 John Wiley & Sons A/S. doi: 10.1111/cbdd.12380
1

Elsayed et al.
azaheterocyclic ring at C-2/3 afforded hybrids with promis-
ing antiproliferative and apoptosis-inducing effects on
hepatic carcinoma cells (BEL-7404) (15).
tional Corporation, Morrisville, NC, USA) was used for col-
umn chromatography, and analytical TLC was carried out
on precoated Si gel 60 F254 TLC plates (EMD Chemicals
Inc., Gibbstown, NJ, USA), using appropriate developing
systems as reported for each experiment. TLC spots were
visualized by spraying with a solution of freshly prepared p-
Breast tumor kinase (Brk), also called protein tyrosine
kinase 6 (PTK6), is a ~60 kDa non-receptor protein kinase,
which is highly expressed in approximately 86% of breast
cancers, but minimally expressed in normal human mam-
mary epithelial cells, suggesting a tumor cell-specific func-
tion (16). Brk activation and/or overexpression can promote
proliferation, migration, and invasion of breast cancer cells
in response to different ligands through multiple mecha-
nisms. For instance, Brk mediates epidermal growth factor
(EGF)-induced migration of breast tumor cells through a
mechanism involving Ras-related C3 botulinum toxin sub-
strate 1 (Rac1) mediated by paxillin (16). Paxillin, a multido-
main cytoskeletal focal adhesion adaptor protein, acts as a
molecular scaffold for other adaptors and structural pro-
teins (17). Moreover, it mediates signal transduction and
cross talk between adhesion and growth factor signaling
(17). Paxillin has proven to regulate diverse physiological
processes including gene expression, cell proliferation,
adhesion, and motility (17). It is phosphorylated in response
to numerous stimuli, including various growth factors such
as insulin-like growth factor 1 (IGF-1), EGF, and hepatocyte
growth factor (HGF). It is also a substrate for non-receptor
tyrosine kinases such as focal adhesion kinase (FAK) (16).
Rac1 is a member of the Rho GTPase family reported to
play an essential role in cancer progression, invasion, and
metastasis. Documented overexpression of Rac1 proteins
was reported in most invasive human cancers, including
metastatic breast cancer (16,17).
anisaldehyde–ethanol–acetic
acid–sulfuric
acid
(2:170:20:10 v/v) and heated at 105 °C until maximal spots
color developed. All chemicals were purchased from
Sigma-Aldrich (St. Louis, MO, USA) or Alfa Aesar (Ward Hill,
MA, USA).
Plant material
Terminalia bentzoe L. (Combretaceae) leaves were col-
lected at the Giza ZOO, Cairo, Egypt, in November 2012. A
ꢀ ꢁ
specimen was identified by Dr. Therese Labib, Taxonomy
Specialist at Al Orman Botanical Garden Herbarium, Giza,
Egypt. A voucher specimen (T.B. No. 1) was deposited at
the Department of Pharmacognosy, Faculty of Pharmacy,
Helwan University, Cairo, Egypt. The fresh leaves (5 kg)
were dried in shade and grounded prior to extraction.
Extraction and bioassay-guided isolation
The dried leaves of T. bentzoe L. (1.5 kg) were extracted
with methanol at 50 °C (7 L 9 3). Extracts were combined
and solvent was removed under reduced pressure at 50 °C
to yield a brown viscous mass which was then lyophilized
to give 250 g dry extract. Subsequent ultrasonication at
room temperature with n-hexane (2 L 9 3) and then ace-
tone (3 L 9 3) afforded hexane-soluble portion (5 g), ace-
tone-soluble portion (94 g), and a residue (141 g). All were
subjected to biological screening for their ability to inhibit
migration of human breast cancer MDA-MB-231 cells in
wound-healing assay (WHA) at different concentrations (25,
50, 100, 200 and 400 lg/mL). The bioactive acetone por-
tion was then subjected to vacuum liquid chromatography
on Si gel 60 eluted with n-hexane-CH2Cl2, CH2Cl2-EtOAc,
and EtOAc-Me2CO gradient systems by 50% increment/
2 L. Fractions were pooled together based on their TLC
similarity pattern affording four collective fractions. The most
active cFIII eluted by CH2Cl2-EtOAc 5:5 to 0:10 was further
purified by flash chromatography on Si gel 60 using gradient
of CH2Cl2-EtOAc each with 10% increment/500 mL. Three
triterpene acids were obtained: oleanolic 1 (CH2Cl2-EtOAc
8:2, 170 mg), masilinic 2 (CH2Cl2-EtOAc 6:4, 30 mg), and
arjunolic 3 (CH2Cl2-EtOAc 2:8, 20 mg).
This study reports the discovery of oleanolic acid through
bioassay-guided fractionation of Terminalia bentzoe L.
leaves methanol extract as its principle breast cancer cell
migration inhibitor. A series of five new (9–12 and 15) and
seven known analogues were semisynthesized. All com-
pounds were evaluated for their antiproliferative, antimigra-
tory, and anti-invasive potentials against MDA-MB-231
human breast cancer cell line. We also report the discov-
ery of Brk/Paxillin/Rac1 axis as a novel molecular target for
the common dietary supplement triterpene oleanolic acid
and analogues. Important structural pharmacophoric fea-
tures of the active analogues were also deduced.
Methods and Materials
General experimental procedures
1H and 13C NMR spectra were recorded at 400 and
100 MHz, respectively, in appropriate deuterated NMR sol-
vent, using TMS as an internal standard, on a JEOL Eclipse
ECS-400 NMR spectrometer (Boston, MA, USA). The ESI-
MS experiments were conducted using a 3200 Q-trap LC/
MS/MS system (Applied Biosystems, Foster City, CA, USA)
using ANALYST version 1.4.1 software (MDS Sciex, Toronto,
ON, Canada). Si gel 60 (230–400 mesh; Natland Interna-
Semisynthetic reactions
Oxidation of oleanolic acid with CrO3
To 2 mL solution of 1 (10 mg, 0.021 mmol) in toluene/
acetone 4:1, an equimolar amount of CrO3 was added
(18). The reaction mixture was stirred for 1 h at room tem-
perature after which 10 mL water was added to quench
the reaction. The solution was then extracted with EtOAc
2
Chem Biol Drug Des 2014

Oleanolic Acid and Analogues as Breast Cancer Inhibitors
(10 mL 9 3). The collected organic layers were dried over
anhydrous Na2SO4 and evaporated. The residue was puri-
fied on Si gel 60 using n-hexane-EtOAc 8:2, isocratic sys-
tem, to afford 4 (8 mg, yield 84%).
at 110 °C for 2 h and worked up as in carbamoylation
reaction. The solid residue (14 mg) was further purified on
Si gel 60 column using gradient elution n-hexane-EtOAc
from 8:2 to 6:4 with gradient increment of 10%, and frac-
tions of 5 mL each were collected, affording compound
15 (n-hexane-EtOAc 7:3, 6 mg, yield 44%).
Carbamoylation of oleanolic acid
To solutions of 1 (20 mg, 0.043 mmol) in toluene (3 mL)
was added an equimolar amount of different isocyanates
and 10 lL of Et3N (19). Each reaction mixture was refluxed
with stirring at 110 °C for 2 h. Water (10 mL) was then
added to quench the reaction followed by shaking with
EtOAc (10 mL 9 3). The combined organic layers were
dried over anhydrous Na2SO4 and evaporated under
vacuum. The solid residue was further purified on Si gel
60 column using gradient elution of n-hexane-EtOAc
from 9:1 to 6:4, and fractions were collected each of 5 mL.
5 was eluted by n-hexane-EtOAc 8:2 (6 mg, yield 26%), 6
was eluted by n-hexane-EtOAc 8:2 (8 mg, yield 32%), 7
was eluted by n-hexane-EtOAc 7:3 (8 mg, yield 29%), 8
was eluted by n-hexane-EtOAc 8:2 (5.5 mg, yield 21%), 9
was eluted by n-hexane-EtOAc 7:3 (5 mg, yield 18%), 10
was eluted by n-hexane-EtOAc 7:3 (9 mg, yield 33%), 11
was eluted by n-hexane-EtOAc 7:3 (6 mg, yield 20%), and
12 was eluted by n-hexane-EtOAc 7:3 (6.5 mg, yield 23%).
3-O-[N-(40,50,60-trimethoxyphenyl)carbamoyl]-oleanolic
acid (9): Reaction of 1 (20 mg, 0.043 mmol) with tri-
methoxyphenyl isocyanate (9.1 mg, 0.045 mmol), affor-
ded 9. White amorphous solid; 1H and 13C NMR
(Tables 1 and 2); ESI-MS: m/z 664.4 [M ꢀ H]ꢀ (calcd.
for C40H58NO7).
3-O-[N-(benzenesulfonyl)-carbamoyl]-oleanolic acid (10):
Reaction of 1 (20 mg, 0.043 mmol) with benzenesulfo-
nyl isocyanate (5.7 lL, 0.043 mmol) was proceeded to
give 10. White amorphous solid; 1H and 13C NMR
(Tables 1 and 2); ESI-MS: m/z 638.6 [M ꢀ H]ꢀ (calcd.
for C37H52NO6S).
3-O-[N-(30-chlorobenzenesulfonyl)-carbamoyl]-oleanolic
acid (11): Reaction of 1 (20 mg, 0.043 mmol) with 2-
chlorobenzenesulfonyl isocyanate (6 lL, 0.043 mmol)
was proceeded to give 11. White amorphous solid; 1H
and 13C NMR (Tables 1 and 2); ESI-MS: m/z 672.4
and 674.2 for [M ꢀ H]ꢀ and [M + 2 ꢀ H]ꢀ, respec-
tively, (calcd. for C37H51ClNO6S).
Reaction of oleanolic acid with lead tetraacetate
To 20 mg (0.043 mmol) of 1 in 3 mL CH2Cl2, Pb(OAc)4
(19.5 mg, 0.043 mmol) was added and stirred at room
temperature for 48 h (20). The reaction was worked up by
the addition of 10 mL water, followed by extraction with
EtOAc (10 mL 9 3). The combined organic layers were
dried over anhydrous Na2SO4 and concentrated under
vacuum. The residue was subjected to purification on Si
gel 60 using gradient elution of n-hexane-EtOAc beginning
with 9:1 to 7:3, to yield 13 (4 mg, yield 20%).
3-O-[N-(50-fluorobenzenesulfonyl)-carbamoyl]-oleanolic
acid (12): Reaction of 1 (20 mg, 0.043 mmol) with 40-
fluorobenzenesulfonyl isocyanate (8 mg, 0.043 mmol)
afforded 12. White amorphous solid; 1H and 13C NMR
(Tables 1 and 2); ESI-MS: m/z 656.6 [M ꢀ H]ꢀ (calcd.
for C37H51FNO6S).
3-O-[N-(30-chlorobenzenesulfonyl)-carbamoyl]-12a-bromo-
1
olean-28,13b-olide (15): White amorphous solid; H and
13C NMR (Tables 1 and 2); ESI-MS: m/z 750.4, 752.3,
and 754.1 for [M ꢀ H]ꢀ, [M ꢀ H + 2]ꢀ and
[M ꢀ H + 4]ꢀ, respectively, (calcd. for C37H50ClBrNO6S).
Reaction of oleanolic acid with
N-bromosuccinimide
To 3 mL solution of 1 (10 mg, 0.021 mmol) in 20% aque-
ous acetone, N-bromosuccinimide (4 mg, 0.022 mmol) in
2 mL 20% aqueous acetone was added drop wise (21).
The reaction stirred for 1 h then water (5 mL) was added,
followed by extraction using EtOAc (5 mL 9 3). Organic
layers were collected, dried over anhydrous Na2SO4, and
concentrated under vacuum. The residue was purified on
Si gel 60 column eluted with n-hexane-EtOAc 7:3 isocratic
system, to yield compound 14 (8.5 mg, yield 76%).
Cell lines and culture conditions
The MDA-MB-231 human breast cancer cells were
obtained from American Type Culture Collection (ATCC,
Manassas, VA, USA). Cells were cultured in RPMI-1640
growth medium (Gibcoꢀ-Invitrogen, Grand Island, NY, USA)
and 10% (v/v) fetal bovine serum (Lonza, Basel, Switzer-
land), supplemented with 2 mM L-glutamine, 100 IU/mL
penicillin G, 100 lg/mL streptomycin, 25 lg/mL amphoteri-
cin B, 2.0 mg/mL NaHCO3, 5.9 mg/mL HEPES, and main-
tained at 37 °C in humidified incubator under 5% CO2.
Reaction of compound 14 with
o-chlorobenzenesulfonyl isocyanate
The MCF-10A non-tumorigenic immortalized human
mammary epithelial cells were purchased from American
Type Culture Collection (ATCC). Cells were cultured in
DMEM/F12 growth medium and 5% (v/v) horse serum,
To solution of the lactone 14 (10 mg, 0.018 mmol) in
2 mL toluene stirred with one equivalent of o-chloro-
benzenesulfonyl isocyanate (1.3 lL, 0.018 mmol) and 5 lL
of Et3N (19). The reaction mixture was refluxed with stirring
Chem Biol Drug Des 2014
3

Elsayed et al.
Table 1: 1H NMR spectroscopic data of compounds 9–12 and 15a
dH, mult. (J in Hz)
Position
1
9
10
11
12
15
1.62, m
1.57, m
1.54, m
1.57, m
1.63, m
1.04, m
0.90, m
0.90, m
0.91, m
1.01, m
2
3
5
6
1.70, m
4.47, dd (11.0, 4.6)
0.86, brs
1.53, m
1.50, m
4.34, dd (11.1, 4.7)
0.72, m
1.53, m
1.54, m
4.38, dd (11.8, 5.0)
0.80, m
4.35, dd (10.0, 3.4)
0.73, brs
1.45, m
4.37, dd (14.6, 5.0)
0.75, brs
1.49, m
1.51, m
1.46, m
1.46, m
1.20, m
1.26, m
1.27, m
1.33, m
7
1.44, m
1.30, m
1.55, m
1.84, m
5.27, brs
1.65, m
1.32, m
1.58, m
1.11, m
2.81, dd (15.1, 5.0)
1.60, m
1.14, m
1.36, m
1.20, m
1.78, m
1.58, m
0.91, s
0.85, s
0.93, s
0.74, s
1.12, s
0.89, s
0.93, s
6.68, s
1.42, m
1.25, m
1.49, m
1.85, m
5.23, t (3.2)
1.64, m
1.42, m
1.43, m
1.26
1.51, m
1.84, m
5.24, t (3.6)
1.64, m
1.49, m
1.37, m
1.68, m
1.62, m
4.24, brs
1.64, m
1.16, m
9
11
12
15
1.45, m
1.84, m
5.23, brs
1.62, m
16
1.60, m
1.57, m
1.60, m
2.12, m
18
19
2.78, dd (13.7, 4.1)
1.59, m
1.15, m
1.34, m
1.19, m
1.77, m
1.56, m
0.77, s
0.75, s
0.86, s
0.70, s
1.08, s
0.87, s
2.78, dd (11.9, 3.6)
1.54, m
1.12, m
1.34, m
1.18, m
2.78, dd (14.6, 5.6)
1.61, m
1.12, m
1.34, m
1.20, m
1.78, m
1.58, m
0.79, s
0.77, s
0.87, s
0.71, s
1.08, s
0.87, s
0.90, s
8.03, m
7.21, dd (8.7, 8.2)
1.96, m
2.28, m
1.95, m
1.32, m
1.24, m
2.30, m
1.72, m
0.76, s
0.73, s
0.82, s
1.17, s
1.38, s
0.96, s
0.87, s
21
22
1.75, m
23
24
25
26
27
29
30
30
0.76, s
0.74, s
0.85, s
0.71, s
1.07, s
0.87, s
0.89, s
0.89, s
7.95, dd (7.3, 1.3)
7.52, dd (7.5, 6.8)
7.62, dd (7.3, 1.3)
7.52, dd (7.5, 6.8)
7.95, dd (7.3, 1.3)
40
7.55, m
7.53, m
7.44, dd (7.8, 6.8)
8.21, d (7.7)
7.45, dd (7.8, 1.4)
7.56, dd, 8.2, 1.8)
7.53, dd (8.2, 7.8)
8.21, dd (8.2, 1.8)
50
60
7.21, dd (8.7, 8.2)
8.03, m
70
6.68, s
3.83, s
3.78, s
3.83, s
80
90
100
aIn CDCl3, 400 MHz, coupling constants (J) in Hz.
supplemented with 20 lg/mL EGF, 0.5 lg/mL hydrocorti-
sone, 0.1 lg/mL cholera toxin, 10 lg/mL insulin, 100 IU/
mL penicillin G, 100 lg/mL streptomycin, and incubated
at 37 °C in humidified atmosphere of 5% CO2.
Migration assay
Wound-healing assay (WHA) was conducted as previously
described (22). Briefly, the confluent monolayer MDA-MB-
231 cells were harvested and seeded onto a sterile 24-well
plate then allowed to attach overnight. A scratch wound
was inflicted using a sterile 200-lL pipette tip. Wounds were
photographed on an inverted microscope connected to
vista vision camera to measure the boundary of the wound
at premigration. Subsequently, cells were treated with differ-
ent concentrations of test compounds in serum-starved
media, each in triplicate, vehicle control was prepared by
adding the maximum volume of DMSO used in preparing
test compounds to serum-free media. Incubation was car-
ried out for 20–24 h till wound was just about to close in
control wells. Media was then aspirated, and cells were fixed
by methanol prior to staining with Giemsa stain. Images were
captured for each wound (3 images/well) and processed by
Test compounds
Stock solutions (25 mM) of purified tested compounds
were prepared in sterilized DMSO and stored at 4 °C.
Working solutions at their final concentrations for each
assay were prepared in appropriate culture medium imme-
diately prior to use. The vehicle control (DMSO) was pre-
pared by adding the maximum DMSO volume used to
prepare test compounds to the convenient media such
that its final concentration did not exceed 0.2% v/v. Stan-
dard oleanolic acid (purity ≥97%; Sigma-Aldrich) was con-
sidered as a positive control in all biological assays (11,15).
4
Chem Biol Drug Des 2014

Oleanolic Acid and Analogues as Breast Cancer Inhibitors
Table 2: 13C NMR spectroscopic data of compounds 9–12, 15a
dC, mult. (J in Hz)
Position
9
10
11
37.9, CH2
23.5, CH2
85.2, CH
37.9, C
12
37.2, CH2
23.4, CH2
85.1, CH
15
37.9, CH2
23.1, CH2
84.8, CH
38.0, C
1
2
3
38.1, CH2
23.9, CH2
82.6, CH
37.9, C
37.9, CH2
23.4, CH2
85.1, CH
38.1, C
4
37.2, C
5
6
7
8
55.4, CH
18.2, CH2
32.6, CH2
39.3, C
55.2, CH
18.1, CH2
32.5, CH2
39.2, C
55.2, CH
18.1, CH2
32.5, CH2
39.3, C
55.2, CH
55.3, CH
17.5, CH2
34.5, CH2
42.6, C
18.1, CH2
32.5, CH2
39.2, C
9
47.6, CH
37.0, C
47.5, CH
36.9, C
47.5, CH
36.9, C
47.3, CH
36.9, C
45.5, CH
36.3, C
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
10
23.4, CH2
23.4, CH2
122.4, CH
143.6, C
41.6, C
27.7, CH2
22.9, CH2
46.5, C
41.0, CH
45.8, CH2
30.7, C
33.8, CH2
32.4, CH2
27.9, CH3
16.6, CH3
15.4, CH3
17.1, CH3
25.9, CH3
182.9, C
33.1, CH3
23.6, CH3
150.4, C
138.7, C
128.2, CH
129.0, CH
133.9, CH
129.0, CH
128.2, CH
23.2, CH2
122.4, CH
143.6, C
41.7, C
27.6, CH2
22.9, CH2
46.5, C
41.0, CH
45.8, CH2
30.7, C
33.8, CH2
32.4, CH2
27.8, CH3
16.6, CH3
15.3, CH3
17.0, CH3
25.9, CH3
182.0, C
33.1, CH3
23.6, CH3
150.0, C
136.0, C
131.9, C
131.8, CH
134.9, CH
127.2, CH
132.8, CH
23.4, CH2
122.4, CH
143.6, C
27.5, CH2
56.1, CH
91.6, C
122.6, CH
143.6, C
41.6, C
27.7, CH2
23.0, CH2
46.5, C
41.0, CH
45.9, CH2
30.7, C
33.8, CH2
32.4, CH2
28.1, CH3
16.8, CH3
15.4, CH3
17.2, CH3
25.9, CH3
182.7, C
33.1, CH3
23.6, CH3
154.0, C
134.1, C
96.0, CH
153.4, C
134.7, C
153.4, C
96.0, CH
56.1, CH3
61.0, CH3
56.1, CH3
41.6, C
43.3, C
27.7, CH2
22.9, CH2
46.5, C
41.0, CH
45.8, CH2
30.7, C
33.8, CH2
32.4, CH2
27.9, CH3
16.6, CH3
15.4, CH3
17.1, CH3
25.9, CH3
182.0, C
29.2, CH2
21.3, CH2
45.6, C
52.3, CH
39.9, CH2
31.9, C
33.9, CH2
30.4, CH2
27.6, CH3
16.4, CH3
17.0, CH3
19.1, CH3
21.1, CH3
179.5, C
33.3, CH3
23.6, CH3
150.0, C
136.0, C
134.9, C
131.8, CH
131.9, CH
127.1, CH
132.9, CH
33.1, CH3
23.6, CH3
150.3, C
20
134.5, C
30
131.2, CH, d (9.5)
116.3, CH, d (23)
134.9, C, d (260)
116.3, CH, d (23)
131.2, CH, d (9.5)
40
50
60
70
80
90
100
C, quaternary; CH, methine; CH2, methylene; CH3, methyl carbons.
aIn CDCl3, 100 MHz, coupling constants (J) in Hz. Carbon multiplicities were determined by PENDANT experiment.
camera program (Liss View software, Radnor, PA, USA)
where the healed wound width was measured. Percent
migration was calculated using the following equation:
Proliferation assay
MDA-MB-231 cells were plated at a density of 1 9 104 cells
per well (6 wells/group) in 96-well culture plates and allowed
to adhere overnight for growth studies (23). The next day,
cells were divided into different groups and then fed serum-
free RPMI-1640 media supplemented with HGF as a mito-
gen and experimental treatments or vehicle-treated control
media. Cells in all groups were fed fresh treatment media
every other day for a 72 h treatment period. Viable cell
count was determined using the MTT colorimetric assay.
The absorbance was measured at k 570 nm on a micro-
plate reader (BioTek, Winooski, VT, USA). The number of
cells per well was calculated against a standard curve
Percentmigration=
woundwidthat zerotime - woundwidthintreatedwell
ꢁ 100
woundwidth at zero time - wound width in DMSO
The IC50 value for each compound was calculated by non-
linear regression of log concentration versus the % migra-
tion, implemented in GRAPHPAD PRISM version 5.0 (GraphPad
Software, San Diego, CA, USA).
Chem Biol Drug Des 2014
5

Elsayed et al.
prepared at the start of the experiment. The IC50 value for
each compound was calculated by nonlinear regression of
log concentration versus the % survival, implemented in
GRAPHPAD PRISM version 5.0 (GraphPad Software).
laboratories, Hercules, CA, USA). Equivalent amounts of
protein were electrophoresed on SDS–PAGE, and the gels
were electroblotted onto PVDF membranes and visualized
according to the method described previously (23). The
visualization of b-tubulin was used to ensure equal sample
loading in each lane. All experiments were repeated at least
three times, and a representative Western blot image from
each experiment is shown in Figure 1.
Invasion assay
The Trevigen’s Cultrexꢀ BME Cell Invasion Assay kit was
used for assessment of anti-invasive activity against MDA-
MB-231 cellsa. The top invasion chamber was coated with
50 lL of basement membrane extract solution (1X BME)
and incubated for 4 h at 37 °C in a 5% CO2-humidified
atmosphere. The coating solution was aspirated off, and
50 lL of cell suspension was added to each well at top
chamber. Ten microlitre of tested compounds, prepared in
serum-free medium, was added to each well in a final non-
toxic concentration of 5 or 15 lM. About 150 lL of RPMI-
1640 medium was added to lower chamber containing
10% FBS and fibronectin (1 lL/mL) and N-formyl-met-leu-
phe (10 nM) as chemotactic agent and incubated at 37 °C
in CO2-humidified atmosphere for 24 h. Media were care-
fully aspirated from both chambers after which cell dissoci-
ation/calcein–AM solution was added to each bottom
chamber well and incubated at 37 °C in CO2 incubator for
1 h. The fluorescence of each sample was measured at
485 nm excitation and 520 nm emission by microplate
reader (BioTek). The number of cells invaded per well was
calculated against a standard curve prepared by plating
various numbers of cells.
Pharmacophore modeling
Pharmacophore modeling was carried out using PHASE
€
(version 3.5, 2013) module of the Schrodinger suite imple-
mented in Maestro (Maestro 9.3.5, 2012) molecular mod-
eling package. The structures of compounds were built
using the structure drawing tool and prepared using Lig-
Prep (version 2.5, 2012). Conformers were generated using
ConfGen by applying OPLS-2005 force-field method with
implicit distance-dependent dielectric solvent model at cut-
ꢂ
off root-mean-square deviation (RMSD) of 1 A. A maximum
of 500 conformers and 100 minimization steps were set.
Compounds were divided into a training set of 12 com-
pounds and a test set of three compounds using the
leave-n-out method. The test set was used for the valida-
tion of the generated models. An arbitrary activity threshold
value was then assigned to divide training set compounds
into eight active (IC50 <15 lM) and six inactive ones
(IC50 > 15 lM). Compounds 1, 2, 4, and 13 were excluded
as their inclusion limits the number of features in each
hypothesis to hydrogen bond acceptors and hydrophobic
groups only. Four features/sites were considered in gener-
ating pharmacophore variants: hydrogen bond acceptor
Cytotoxicity assay
The confluent monolayer MCF-10A cells were harvested
and seeded into 96-well plate at density 2 9 104 cells/well
and left to recover. Twenty-four hours postseeding, cells
were treated with test compounds at different concentra-
tions in fresh serum-free media, each in triplicate, and
incubated at 37 °C in humidified atmosphere of 5% CO2
for 24 h. At the end of treatment, media were replaced by
fresh ones containing 50 lL MTT (1 mg/mL in PBS) and
incubated for further 4 h. The insoluble formazan crystals
were solubilized in 100 lL of DMSO, and absorbance was
measured at 570 nm on microplate reader (BioTek). The
average number of cells per well was calculated from the
standard curve prepared by platting various cell concen-
trations at the start of the experiment.
Western blot analysis
MDA MD-231 cells were initially plated at a density of
1 9 106 cells/100-mm culture plate, allowed to attach over-
night in RPMI-1640 media containing 10% FBS. Cells were
then washed with 1X PBS and incubated with vehicle con-
trol or treatment in serum-free media supplemented with
20 ng/mL EGF as a mitogen for 3 days in culture (23).
At the end of treatment period, cells were lysed in RIPA
buffer (Qiagen Sciences Inc., Valencia, CA, USA). Protein
concentration was determined by the BCA assay (Bio-Rad
Figure 1: Oleanolic acid and active analogues 10–12 reduced
Brk, paxillin, and Rac1 phosphorylation at 10 lM dose level
evaluated by Western blot analysis.
6
Chem Biol Drug Des 2014

Oleanolic Acid and Analogues as Breast Cancer Inhibitors
(A), negative ionizable group (N), hydrophobic group (H),
and aromatic ring (R). Resulting pharmacophore hypothe-
ses were then scored using default weights of scoring
parameters for both active (survival score) and inactive
ones (survival-inactive scores). To further assist in ranking
hypotheses, post hoc score was generated in which an
activity reward was added for hypotheses that utilize the
most active compound as the reference and a penalty
assigned for hypotheses in which the reference ligand
shows a high relative conformational energy. Hypotheses
were clustered, and those with the highest survival score in
each cluster were selected for further assessment. Four
hypotheses were eventually selected: AAHHHR.165891,
AAAHNR.322, AAHHNR.19590, and AAAHHR.808.
The compounds were identified based on their spectro-
scopic analysis and comparing with literature (24–26).
Based on its high abundance in several plants and herbal
remedies, oleanolic acid was chosen for further optimiza-
tion. Diverse semisynthetic analogues of oleanolic acid (1)
were prepared through reaction with chromium trioxide,
different isocyanates, lead tetraacetate, and N-bromosuc-
cinimide (NBS). Oleanolic acid was oxidized by CrO3
(Scheme 1) to afford 4, the corresponding 3-oxo-oleane-
12-ene-28-oic acid (oleanonic acid) (15). Additionally,
oleanolic acid was carbamoylated using various aromatic
and olefinic isocyanates (Scheme 2) providing the
expected C-3 carbamates 5–12. Known carbamate ana-
logues were identified as 3-O-[N-(allyl)-carbamoyl]-oleanol-
ic acid (5), 3-O-[N-(benzyl)-carbamoyl]-oleanolic acid (6),
3-O-[N-(biphenyl)-p-carbamoyl]-oleanolic acid (7), and
3-O-[N-(20-naphthyl)-carbamoyl]-oleanolic acid (8) based
on their spectroscopic data and compared with literature
(12). Compound 9 is the reaction product of oleanolic acid
with 3,4,5-trimethoxyphenyl isocyanate. It exhibited in ESI-
MS a molecular ion peak [M ꢀ H]ꢀ at m/z 664.4, sug-
gesting the expected incorporation of the trimethoxyphenyl
carbamoyl moiety to 1. The 1H and 13C NMR data
(Tables 1 and 2) revealed the downfield shift of H-3 to d
4.47 and C-3 to d 82.6, compared with the parent 1,
spotlighting the carbamoylation at position C-3. Addition-
ally, the aromatic protons resonance at d 6.68 (2H, s, H-
30 and H-70) connected to its carbon at d 96.0 in HMQC
spectrum, suggested the expected trisubstituted phenyl
ring. The aromatic methoxy singlets at d 3.83 (6H, s) and
3.78 (3H, s) linked to their carbons at 61.0 and 56.1 ppm,
respectively, in HMQC spectrum were attributed to
80,90,100-trimethoxy pattern on the phenyl moiety. There-
fore, analogue 9 was interpreted to be 3-O-[N-(80,90,100-
trimethoxyphenyl)-carbamoyl]-oleanolic acid. Compound
10 was furnished by the reaction of 1 with benzenesulfo-
nyl isocyanate. Its molecular ion peak [M ꢀ H]ꢀ at m/z
638.6; with an increase of 181.9 a.m.u. than the parent
(1); indicated inclusion of a benzenesulfonyl carbamoyl
moiety. The 1H and 13C NMR data (Tables 1 and 2) along
Statistical analysis
The results are presented as means ꢂ SEM of at least
three independent experiments. Differences among various
treatment groups were determined by the analysis of vari-
ance (ANOVA) followed by Dunnett’s test. A difference of
p < 0.05 was considered statistically significant as com-
pared to the vehicle-treated control group. The IC50 values
represent concentrations that induce 50% cell growth,
migration, and invasion inhibition.
Results and Discussion
Terminalia bentzoe L. leaves methanol extract showed a
considerable antimigratory activity (IC50 = 200.0 lg/mL)
against the human metastatic breast cancer cells MDA-
MB-231 in WHA. Further fractionation by hexane and
acetone followed by biological evaluation afforded active
acetone-soluble portion with IC50 = 50.0 lg/mL. Subse-
quent bioassay-guided fractionation using vacuum liquid
chromatography afforded a bioactive triterpene acid-rich
fraction (IC50 = 7.3 lg/mL) identified from its 1D NMR
spectra. The mixture was then subjected to further purifi-
cation yielding three pure known triterpene acids: oleanolic
(1), masilinic (2), and arjunolic (3) with their corresponding
antimigratory IC50 at 14.0, 9.0, and >40.0 lM, respectively.
Br
O
O
(iii)
1
H
(ii)
(i)
14
, R = H
RO
O
O
O
H
(iv)
S
N
H
15, R =
Cl
Scheme 1: Semisynthesis of 4 and 13–15.
(i) CrO3, toluene/acetone, rt, 1 h. (ii) Pb
(OAc)4, CH2Cl2, rt, 48 h. (iii) N-
H
bromosuccinimide, 20% aqueous acetone,
rt, 2 h. (iv) R-N=C=O, Et3N, toluene, reflux
110 °C, 2 h.
HO
O
H
H
4
13
Chem Biol Drug Des 2014
7

Elsayed et al.
R
5 a
R
e
H
9
6
7
8
b
c
d
10 f
COOH
11
12
(i)
g
h
1
H
RO
H
O
O
C
O
C
O
2'
C
N
H
N
H
2'
C
N
5'
N
H
1'
3'
H
a
b
c
d
8'
H3CO
Cl
O
O
C
O
S
O
O
O
C
O
C
3'
S
C
S
2'
N
N
H
H3CO
N
H
N
H
H
O
O
O
Scheme 2: Carbamoylation of oleanolic
acid (1). (i) R-N=C=O, Et3N, toluene, reflux
110 °C, 2 h.
5'
F
H3CO
10'
e
f
g
h
with 3J HMBC cross-peak between the carbamoyl carbon
C-10 (dC 150.4) and H-3 (dH 4.35) confirmed the carba-
moylation at this position. Hence, 10 was interpreted as
3-O-[N-(benzenesulfonyl)-carbamoyl]-oleanolic acid. Com-
pound 11 was obtained by the reaction of 1 with o-chlo-
robenzenesulfonyl isocyanate. Its ESI-MS spectrum
displayed characteristic isotopic cluster signals of
[M ꢀ H]ꢀ and [M ꢀ H + 2]ꢀ at m/z 672.4 and 674.2,
respectively, with ratio 3:1, indicating a mono-chloro-
benzenesulfonyl carbamate analogue. It exhibited similar
1H and 13C chemical shifts as 10 with 1,2-disubstituted
benzene moiety of four aromatic protons [(8.21, 1H, d,
J = 7.7), (7.55, 1H, m), (7.53, 1H, m), and (7.44, 1H, dd,
J = 7.8, 6.8 Hz)] linked to their carbons in the HMQC
spectrum at dC 132.8, 131.8, 134.9, and 127.2, respec-
tively. Therefore, compound 11 was elucidated as 3-O-[N-
(30-chlorobenzenesulfonyl) carbamoyl]-oleanolic acid. Com-
pound 12 is the reaction product of 1 with p-fluoro-benze-
nesulfonyl isocyanate. Its ESI-MS spectrum displayed a
molecular ion peak [M ꢀ H]ꢀ at m/z 656.6, consistent
with possible mono-fluorobenzenesulfonyl carbamate ana-
logue of 1. The 1H and 13C NMR data (Tables 1 and 2)
showed the distinct splitting pattern for 1,4-disubstituted
aromatic moiety due to coupling with the fluorine atom.
The two aromatic protons (H-40/60) appeared as doublet
of doublet with J = 8.7, 8.2 Hz due to the ortho-coupling
with H-30/70 and fluorinated carbon C-50. Their corre-
sponding carbons at dC 116.3 displayed doublet splitting
pattern (J = 23 Hz) as a result of 2J coupling with C-5‘.
The two aromatic carbons C-30/70 at dC 131.2 appeared
as doublet (J = 9.5 Hz) were linked to their corresponding
protons at dH 8.03 (2H, m) in HMQC spectrum.
The assigned C-50 at d 165.9 splitted as doublet with
J = 260 Hz due to 1J-coupling with its ipso-fluoro atom.
Thus, 12 was proved to be 3-O-[N-(50-fluorobenzenesulfo-
diene-28-oic acid (13), and 3b-hydroxy-12a-bromo-olean-
28, 13-olide (14), respectively (27,28). Finally, for both ring
C and C-3 modification, the lactone 14 was further
reacted with o-chlorobenzenesulfonyl isocyanate to get 15
(Scheme 1). Its ESI-MS showed characteristic clusters
for
a mono-chlorinated, mono-brominated compound
[M ꢀ H]ꢀ, [M ꢀ H + 2]ꢀ, and [M ꢀ H + 4]ꢀ at m/z 750.4,
752.3, and 754.1, respectively. The 1D NMR data
(Tables 1 and 2) were comparable to 14 with few varia-
tions. Proton H-3 and carbon C-3 were downfield shifted
by +1.19 and +5.7 ppm, respectively. It also showed a
characteristic disubstituted benzene pattern [d 8.21 (1H,
dd, J = 8.2, 1.8 Hz), 7.56 (1H, dd, J = 8.2, 1.8 Hz), 7.53
(1H, dd, J = 8.2, 7.8 Hz), 7.45 (1H, dd, J = 7.8, 1.4 Hz)].
The HMBC spectrum showed cross-peak between
the carbonyl C-10 (d 150.0) and proton H-3 (d 4.38),
confirming C-3 carbamoylation. Hence, 15 proved to be
3-O-[N-(30-chlorobenzenesulfonyl)-carbamoyl]-12a-bromo-o
lean-28,13b-olide.
nyl)-carbamoyl]-oleanolic acid. Ring
C
modification
was achieved by reacting 1 with Pb(OAc)4 and NBS
(Scheme 1) to afford compounds 3b-hydroxy-olean-9, 12-
8
Chem Biol Drug Des 2014

Oleanolic Acid and Analogues as Breast Cancer Inhibitors
To study the in vitro inhibitory effect of oleanolic acid and
its semisynthetic analogues against the human metastatic
breast cancer cells MDA-MB-231, the scratch WHA has
been conducted. Initially, to determine the activity role of
the C-3 OH group, it was oxidized to a ketone 4. Ana-
IC50 = 5.8 lM (Figure 2, Table 3). Comparing the activity
of 10 and the benzyl carbamate 6, the sulfonyl group’s
electron-withdrawing effect on the benzene ring improved
the activity. This may be improving the carbamoyl NH
group ionization at the physiological pH and possible
generation of a salt bridge interaction within the binding
pocket of the molecular target (29). The active benze-
nesulfonyl moiety was then maintained, and additional
electron-withdrawing group substitution on the benzene
ring at different positions was attempted. The o-chloro-
benzenesulfonyl carbamate 11 showed nearly threefold
more activity versus the unsubstituted carbamate 10
(IC50 = 2.1 lM, Figure 2). This suggested the importance
of additional electron-withdrawing groups on the phenyl
moiety. The p-fluorobenzenesulfonyl carbamate 12
logue
4
exhibited improved activity (IC50 = 5.5 lM,
Table 3), versus its parent 1, suggesting no hydrogen
bonding donating role for C-3 OH in 1 for antimigratory
activity. C-3 oxygen is likely to play a hydrogen acceptor
role for this activity. Subsequently, a strategy was adopted
for structure extension at C-3, maintaining the C-3 oxygen
and diversifying the substitution to probe activity improve-
ment. A series of various aromatic and olefinic carbamates
were synthesized and tested. For instance, the benzyl car-
bamate analogue 6 exhibited IC50 = 17.0 lM, which was
comparable to 1 (IC50 = 14.0 lM), while the olefinic allyl
carbamate 5 was less active (IC50 > 40.0 lM), indicated
that position C-3 can afford a carbamate moiety with med-
ium bulkiness. To further assess the aromatic side chain
size on the activity, the biphenyl carbamate 7 and the
naphthyl carbamate 8 were synthesized and tested. Both
analogues were less active (IC50 > 40.0 and 25.4 lM,
respectively, Table 3) than the parent oleanolic acid, aug-
menting the preliminary observation of an optimal bulki-
ness of the phenyl moiety.
To explore the effect of different substituents on the phenyl
moiety, the trimethoxyphenyl carbamate 9 was showing
reduced antimigratory activity (IC50 > 40.0 lM, Table 3)
indicating the unfavorable electron-donating group substi-
tutions on the phenyl moiety. The effect of electron-with-
drawing group substitutions at the phenyl carbamate
moiety was also assessed. For instance, the benzenesulfo-
nyl carbamate 10 showed enhanced activity, with an
Figure 2: Effect of analogues 10–12 on MDA-MB-231 cell
migration in wound-healing assay at different concentrations
compared with DMSO as vehicle control. Oleanolic acid was used
as positive control. Error bars indicate the SEM of n = 3 per
compound. *p < 0.05.
Table 3: IC50 (lM ꢂ SEM) of natural triterpenes and semisynthetic
oleanolic acid analogues in migration, proliferation, and cytotoxicity
assays
Migration
assay
Proliferation
assay
Cytotoxicity
assay
Compound MDA-MB-231
MCF-10A
1
14.0 ꢂ 2.4
9.0 ꢂ 1.4
>40.0
11.2 ꢂ 2.1
10.1 ꢂ 0.6
56.1 ꢂ 1.2
7.2 ꢂ 1.3
18.0 ꢂ 2.3
11.5 ꢂ 0.7
26.6 ꢂ 2.8
24.0 ꢂ 2.2
18.9 ꢂ 1.5
7.8 ꢂ 0.6
6.1 ꢂ 0.8
3.4 ꢂ 0.4
8.1 ꢂ 1.4
30.0 ꢂ 3.2
9.7 ꢂ 1.2
>50.0
2
30.2 ꢂ 2.3
3
>50.0
4
5.5 ꢂ 1.1
>50.0
>50.0
5
>40.0
6
17.0 ꢂ 2.3
>40.0
15.5 ꢂ 2.1
7
>50.0
8
25.4 ꢂ 3.1
29.4 ꢂ 0.7
9.3 ꢂ 0.6
4.5 ꢂ 0.4
5.7 ꢂ 1.1
17.7 ꢂ 0.8
>50.0
9
>40.0
10
11
12
13
14
15
5.8 ꢂ 0.9
2.1 ꢂ 0.5
1.4 ꢂ 0.3
6.9 ꢂ 1.1
>40.0
4.0 ꢂ 0.7
Figure 3: Effect of some active analogues on MDA-MB-231
breast cancer cells viability compared with DMSO as vehicle
control and oleanolic acid as positive control. Error bars indicate
the SEM of n = 3 per compound. *p < 0.05
12.0 ꢂ 2.0
>50.0
Values representing the mean ꢂ SEM of three independent
experiments.
Chem Biol Drug Des 2014
9

Elsayed et al.
exhibited nearly 10-fold enhancement in antimigratory
activity, compared with 1 (Figure 2, Table 3). This clearly
indicates the preference of small electron-withdrawing iso-
stere on the phenyl moiety. Additionally, previous ring C
modifications showed anticancer activity improvement of 1
(14). Therefore, D12,13 system was targeted through the
use of NBS (Scheme 1), furnishing the unexpected bromo-
lactone product 14. This product offered the opportunity
to test the importance of D12,13 system as well as the free
COOH group for the antimigratory activity. Compound 14
was less active (IC50 > 40.0 lM) than 1, pointing out the
importance of intact D12,13 system and free COOH group.
As benzenesulfonyl carbamoylation enhanced the activity
of the parent 1, compound 14 was further carbamoylated
using o-chlorobenzenesulfonyl isocyanate to afford 15.
The latter regained the antimigratory activity, with an IC50
of 12.0 lM (Table 3), suggesting the ability of o-chloro-
benzenesulfonyl carbamoylation to make up for the lost
activity due to masking of the free C-28 COOH group,
possible by offering additional binding pharmacophore at
the molecular target. Additional unsaturation (D9,11) in ring
C represented by 13, which also reported as a natural
product (eucalytptanoic acid), showed enhanced antimi-
gratory, with an IC50 of 6.9 lM (Table 3). This may suggest
that the change in ring C geometry by additional unsatura-
tion may offer better binding affinity to the molecular tar-
get, improving the overall bioactivity.
10.1 lM, respectively. Meanwhile, arjunolic acid (3) was
nearly inactive, with an IC50 of 56.0 lM. This suggested C-
2 hydroxylation has little effect on the activity of 2, while
hydroxylation of the C-23 methyl group greatly diminished
the antiproliferative effect in 3. The carbamate analogues
displayed diverse activities, based on the size and nature
of the side chain. The benzyl carbamate 6 maintained the
antiproliferative activity (IC50 = 11.5 lM), while the allyl car-
bamate 5 was less active (IC50 = 18.0 lM). Therefore,
analogous to migration, carbamates with aromatic side
chain have more favorable antiproliferative activity than
small olefinic moiety. Carbamates with more bulky chains
as 7 and 8 showed reduced activity. In addition, carba-
mates with electron-donating substituents in the aromatic
ring were less active than the benzenesulfonyl carbamates
with o- or p-position small electron-withdrawing groups.
Carbamate 12 with p-fluoro substituent was more active
than o-chloro substituted carbamate 11 (IC50 = 3.4 and
6.1 lM, respectively), suggesting small electron-withdraw-
ing group in p-position can better improve the antiprolifera-
tive activity. Demonstrated in Figure 3, the effect of various
doses of oleanolic acid and most active analogues 10–12
on the viability of the human MDA-MB-231 mammary
tumor cells after 3-day culture period.
Recently, the anti-invasive activity of oleanolic acid against
malignant glioma cells was reported (11). Therefore, natu-
ral and semisynthetic triterpenes were further assessed for
their ability to inhibit invasion of MDA-MB-231 breast can-
cer cells using Cultrexꢀ BME cell invasion kit. Each com-
pound was tested at subtoxic doses of 5 or 15 lM.
Treatment with 5 lM oleanolic acid allowed 74% of cells to
invade (Figure 4). Meanwhile, 5 lM treatments of the halo-
genated carbamates 11 and 12 allowed only 53.5 and
27.5% cell invasion, respectively. This indicated obvious
anti-invasive activity improvement parallel to antimigratory
activity. Other carbamates 6 and 15 were less active at
To assess the growth inhibitory effect of natural and semi-
synthetic triterpenes on MDA-MB-231 human breast can-
cer cells, the MTT (3-(4, 5-dimethyl-2-thiazolyl)-2, 5-
diphenyl-tetrazolium bromide) proliferation assay was
implemented (23). Oleanolic acid and some of its ana-
logues significantly inhibited MDA-MB-231 cell growth in a
dose-responsive manner as compared to vehicle-treated
cells (Table 3). The natural products oleanolic (1) and ma-
silinic (2) acids were both active with IC50 = 11.2 and
Figure 4: Anti-invasive activity of natural
and semisynthetic triterpenes against MDA-
MB-231 breast cancer cells using Cultrexꢀ
BME cell invasion assay. Oleanolic acid was
used as positive control. Error bars indicate
the SEM of n = 3 per compound. *p < 0.05.
10
Chem Biol Drug Des 2014

Oleanolic Acid and Analogues as Breast Cancer Inhibitors
A
B
Figure 5: Alignment of active analogues.
(A) analogue 12 aligned with the top-scoring
hypothesis of variant AAHHHR. (B) analogue
11 aligned with the top-scoring hypothesis
of variant AAHHNR. Light red spheres
represent hydrogen bond acceptors (A),
green spheres represent hydrophobic
groups (H), dark red spheres represent
negative ionizable groups (N), and brown
ring represents aromatic group (R).
the same dose level, allowing 93.4 and 80.1% cell inva-
sion, respectively. Carbamate 9 at 15 lM dose was allow-
ing only 30% MDA-MB-231 cells’ invasion, although it was
inactive in migration assay.
analogues 11 and 12 significantly blocked proliferation,
migration, and invasion of the highly invasive MDA-MB-231
breast cancer cells. This effect might be related, at least in
part, to the suppression of Brk/Paxillin/Rac1 signaling path-
way. In addition, treatment with analogue 12 showed
marked inhibition of phosphorylation (activation) of Akt and
ERK1/2 which are important signaling molecules for survival
and proliferation, respectively, without affecting their total
levels.
The selective cytotoxicity of all tested natural and semisyn-
thetic triterpenes toward the non-tumorigenic human
mammary epithelial MCF-10A cells was assessed using
MTT cytotoxicity assay. The results (Table 3) demonstrated
that natural triterpenes and most semisynthetic analogues
were non-toxic at relatively higher concentrations com-
pared with their corresponding IC50 in different anticancer
assays. For example, the most active analogue 12 showed
an IC50 of 17.7 lM, while it exhibited IC50 values of 1.4
and 3.4 lM in migration and proliferation assays, respec-
tively, against the breast cancer cells MDA-MB-231. These
results clearly suggest the selectivity of 12 against the
breast cancerous cells.
Identification of the important structural pharmacophores
required for the antimigratory activity of oleanolic acid
In an attempt to unravel the molecular signaling pathway
that might justify the antimigratory and anti-invasive activities
of oleanolic acid and related analogues, Western blot analy-
sis was conducted. Treatment of human breast cancer
MDA-MB-231 cells with 10 lM dose of oleanolic acid or its
most active analogues (10–12) resulted in no effect on total
Brk, paxillin, and Rac1 levels, whereas they resulted in a
large reduction in phosphorylated (activated) Brk, paxillin,
and Rac1 compared with vehicle-treated control groups. In
addition, Western blot analysis showed that intracellular lev-
els of FAK and p-FAK are similar in all treatment groups.
Treatment with analogues 11 and 12 showed superior inhi-
bition of phosphorylation (activation) of Brk/Paxillin/Rac1
signaling pathway (Figure 1). These results suggested that
Figure 6: Poor spatial alignment of inactive analogue 14 with the
pharmacophoric features. Light red spheres represent hydrogen
bond acceptors (A), green spheres represent hydrophobic groups
(H), and brown ring represents aromatic group (R).
Chem Biol Drug Des 2014
11

Elsayed et al.
A
B
Figure 7: Antimigratory activity of analogue
12 against MDA-MB-231 breast cancer
cells in WHA. (A) DMSO vehicle control. (B)
4 lM treatment of 12.
analogues was achieved through a pharmacophore mod-
pharmacophore modeling study correlated electron-defi-
cient phenyl moiety at optimal bulkiness linked to a sulfo-
nyl group as the main pharmacophores in active
analogues. Based on the potent activity and selectivity,
carbamate 12 is defined as a potential hit appropriate for
further optimization for the use to control metastatic
breast cancer.
€
eling study. PHASE module of the Schrodinger molecular
modeling software was used, and a number of pharmaco-
phore models were generated. A successful pharmaco-
phore model identifies the important binding groups
required for activity and their relative positions in space
with respect to each other. Numerous common pharma-
cophore hypotheses (models) with different combination of
variants (features) were generated.b Features explored in
generating the hypotheses include hydrogen bond accep-
tor (A), negative ionizable group (N), hydrophobic group
(H), and aromatic ring (R). Hypotheses were evaluated on
the basis of survival, survival-inactive, and post hoc scores
(30). Two-hundred hypotheses survived the scoring func-
tion; however, only the top-scoring hypothesis from each
variant combination was selected including AAHHHR,
AAAHNR, AAHHNR, and AAAHHR. These hypotheses
showed superior alignment with active analogues (IC50
<15 lM, Figure 5), displayed lower relative conformational
energy values for actives, and discriminated active and
inactive analogues (lower fitness scores for inactive ones,
IC50 > 15 lM, Figure 6) (30). The generated hypotheses
further highlighted the most important structural features
implicated in the antimigratory activity of oleanolic acid
analogues, including C-X aromatic moiety and the sulfonyl
carbamoyl group. The absence of any of these features
significantly compromised the antimigratory activity. For
example, the benzyl carbamate 6 is about twice as active
as the allyl carbamate 5 but fourfold less active than the
benzenesulfonyl carbamate 10. In addition, carbamate 12
which verified most of the essential binding epitopes (Fig-
ure 5) showed the best antimigratory activity (Figure 7).
Acknowledgments
The Egyptian Ministry of Higher Education and Helwan
University, Cairo, Egypt, are greatly acknowledged for sup-
porting H. E. Elsayed fellowship. The Louisiana Board of
Regents is also acknowledged for supporting the new MS
facility (LEQSF(2013-14)-ENH-TR-26).
Conflict of Interest
The authors of the present publication declare no conflict
of interest.
References
1. Yi E.H., Lee C.S., Lee J.K., Lee Y.J., Shin M.K., Cho
C.H., Kang K.W., Lee J.W., Han W., Noh D.Y., Kim
Y.N., Cho I.H., Yel S.K. (2013) STAT3-RANTES
autocrine signaling is essential for tamoxifen resistance
in human breast cancer cells. Mol Cancer Res;11:31–
42.
2. Cheng X., Hung M.C. (2007) Breast cancer brain
metastases. Cancer Metastasis Rev;26:635–643.
3. Chambers A.F., Groom A.C., MacDonald I.C. (2002)
Dissemination and growth of cancer cells in metastatic
sites. Nat Rev Cancer;2:563–572.
4. Harvey A.L. (2008) Natural products in drug discovery.
Drug Discov Today;13:894–901.
5. Mishra B.B., Tiwari V.K. (2011) Natural products: an
evolving role in future drug discovery. Eur J Med
Chem;46:4769–4807.
Conclusions
The common triterpene oleanolic acid from T. bentzoe L.
was identified as the breast cancer cell migration inhibitor
principle. Semisynthetic optimization of oleanolic acid
afforded 12 analogues from which carbamate 12 showed
potent selective antimigratory, antiproliferative, and anti-
invasive activities against MDA-MB-231 cells at low lM
level. Western blot analysis discovered Brk/Paxillin/Rac1
6. Owolabi M.S., Lawal O.A., Ogunwande I.A., Hauser
R.M., Setzer W.N. (2013) Chemical composition of the
axis as
a potential molecular target. Computer-aided
12
Chem Biol Drug Des 2014

Oleanolic Acid and Analogues as Breast Cancer Inhibitors
leaf essential oil of Terminalia catappa L. growing in
and HIF-1 activation in breast tumor cells. J Nat
Prod;71:396–402.
south-western Nigeria. Am
J
Essent Oils Nat
Prod;1:51–54.
20. Kyeremeh K., Baddeley T.C., Stein B.K., Jaspars M.
(2006) A homologous series of eunicellin-based diterp-
enes from Acalycigorgia sp. characterised by tandem
mass spectrometry. Tetrahedron;62:8770–8778.
21. Sawant S.S., Sylvester P.W., Avery M.A., Desai P.,
Youssef D.T., El Sayed K.A. (2004) Bioactive rear-
ranged and halogenated semisynthetic derivatives of
7. Kinoshita S., Inoue Y., Nakama S., Ichiba T., Aniya Y.
(2007) Antioxidant and hepatoprotective actions of
medicinal herb, Terminalia catappa L. from Okinawa
Island and its tannin corilagin. Phytomedicine;14:755–
762.
8. Ramesh A.S., Christopher J.G., Radhika R., Setty
C.R., Thankamani V. (2012) Isolation, characterisation
and cytotoxicity study of arjunolic acid from Terminalia
arjuna. Nat Prod Res;26:1549–1552.
the marine natural product sarcophine.
J
Nat
Prod;67:2017–2023.
22. Rodriguez L.G., Wu X., Guan J.L. (2005) Wound-heal-
ing assay. Methods Mol Biol;294:23–29.
9. Gurib-Fakim A., Demarne F. (1994) Essential oil of Ter-
€
minalia bentzoe (L.) L. f. subsp. rodriguesensis Wic-
23. Akl M.R., Ayoub N.M., Sylvester P.W. (2012) Mecha-
nisms mediating the synergistic anticancer effects of
combined c-tocotrienol and sesamin treatment. Planta
Med;78:1731–1739.
kens. J Essent Oil Res;6:533–534.
10. Sultana N., Ata A. (2008) Oleanolic acid and related
derivatives as medicinally important compounds. J
Enzyme Inhib Med Chem;23:739–756.
11. Guo G., Yao W., Zhang Q., Bo Y. (2013) Oleanolic
acid suppresses migration and invasion of malignant
glioma cells by inactivating MAPK/ERK signalling path-
way. PLoS ONE;8:e72079.
24. Seebacher W., Simic N., Weis R., Saf R., Kunert O.
1
(2003) Complete assignments of H and 13C NMR res-
onances of oleanolic acid, 18a-oleanolic acid, ursolic
acid and their 11-oxo derivatives. Magn Reson
Chem;41:636–638.
12. Ashour A., El-sharkawy S., Amer M., Abdel Bar F., Ka-
takura Y., Miyamoto T., Toyota N., Bang T.H., Kondo
R., Shimizu K. (2014) Rational design and synthesis of
topoisomerase I and II inhibitors based on oleanolic
acid moiety for new anti-cancer drugs. Bioorg Med
Chem;22:211–220.
13. Pertino M.W., Lopez C., Theoduloz C., Schmeda-Hir-
schmann G. (2013) 1,2,3-Triazole-substituted oleanolic
acid derivatives: synthesis and antiproliferative activity.
Molecules;18:7661–7674.
14. Honda T., Rounds B.V., Gribble G.W., Suh N. (1998)
Design and synthesis of 2-cyano-3,12-dioxoolean-1,9-
dien-28-oic acid, a novel and highly active inhibitor of
nitric oxide production in mouse macrophages. Bioorg
Med Chem Lett;8:2711–2714.
15. Kang X., Hu J., Gao Z., Ju Y., Xu C. (2012) Synthesis,
anti-proliferative and proapoptotic activity of novel ole-
anolic acid azaheterocyclic derivatives. Med Chem
Com;3:1245–1249.
25. Mahato S.B., Kundu A.P. (1994) 13C NMR Spectra of
pentacyclic triterpenoids–a complication and some
salient features. Phytochemistry;37:1517–1573.
26. Bisoli E., Garcez W.S., Hamerski L., tieppo C.,
Garcez F.R. (2008) Bioactive pentacyclic triterpenes
from the stems of Combretum laxum. Mole-
cules;13:2717–2728.
27. Begum S., Sultana I., Siddiqui B.S., Shaheen F., Gilani
A. (2002) Structure and spasmolytic activity of eucalyp-
tanoic acid from Eucalyptus camaldulensis var. obtusa
and synthesis of its active derivative from oleanolic
acid. J Nat Prod;65:1939–1941.
28. Pungitore C.R., Garcia M., Gianello J.C., Sosa M.E.,
Tonn C.E. (2005) Insecticidal and antifeedant effects of
Junellia aspera (Verbenaceae) triterpenes and deriva-
tives on Sitophilus oryzae (Coleoptera: Curculionidae).
J Stored Prod Res;41:433–443.
29. Carins D. (2008) Essentials of Pharmaceutical Chemis-
try, 3rd edn. London, UK: Pharmaceutical Press.
30. Dixon S.L., Smondyrev A.M., Knoll E.H., Rao S.N.,
Shaw D.E., Friesner R.A. (2006) PHASE: a new engine
for pharmacophore perception, 3D QSAR model devel-
opment, and 3D database screening: 1. Methodology
16. Chen H.Y., Shen C.H., Tsai Y.T., Lin F.C., Huang
Y.P., Chen R.H. (2004) Brk activates Rac1 and pro-
motes cell migration and invasion by phosphorylating
paxillin. Mol Cell Biol;24:10558–10572.
17. Turner C.E. (2000) Paxillin and focal adhesion signal-
ing. Nat Cell Biol;2:E231–E236.
and preliminary results.
Des;20:647–671.
J
Comput Aided Mol
18. Kim D.S., Pezzuto J.M., Pisha E. (1998) Synthesis of
betulinic acid derivatives with activity against human
melanoma. Bioorg Med Chem Lett;8:1707–1712.
19. El Sayed K.A., Khanfar M.A., Shallal H.M., Muralidha-
ran A., Awate B., Youssef D.T., Liu Y., Zhou Y.D.,
Nagle D.G., Shah G. (2008) Latrunculin A and its C-
17-O-carbamates inhibit prostate tumor cell invasion
Notes
aCultrexꢀ BME Cell Invasion Assay Protocol. (2013) (www.
trevigen.com, accessed December; 2013).
bPhase 3.5 Quick Start Guide. (2013) Schro¨ dinger press,
LLC.
Chem Biol Drug Des 2014
13