Dimethoxyaurones: Potent inhibitors of ABCG2 (breast cancer resistance protein)

Article abstract of DOI:10.1016/j.ejps.2008.07.008

A series of 4,6-dimethoxyaurones were synthesized by reacting 4,6-dimethoxybenzofuran-3(2H)-one with various benzaldehydes in a base-catalyzed aldol reaction. A Z configuration was assigned to the aurones based on spectroscopic and crystallographic data.

Full text of DOI:10.1016/j.ejps.2008.07.008

european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/ejps
Dimethoxyaurones: Potent inhibitors of ABCG2
(breast cancer resistance protein)
Hong-May Sim, Chong-Yew Lee, Pui Lai Rachel Ee, Mei-Lin Go∗
Department of Pharmacy, National University of Singapore, 18 Science Drive 4, Singapore 117543, Singapore
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 4,6-dimethoxyaurones were synthesized by reacting 4,6-dimethoxybenzofuran-
3(2H)-one with various benzaldehydes in a base-catalyzed aldol reaction. A Z configuration
was assigned to the aurones based on spectroscopic and crystallographic data. The aurones
were tested for their ability to modulate ABCG2 (breast cancer resistance protein)-mediated
multidrug resistance in vitro. Several members (0.5 ␮M) increased the accumulation of
mitoxantrone (MX) in human breast cancer cells (MDA-MB-231) transfected with ABCG2
and re-sensitized these cells to the cytotoxic effects of MX. In the re-sensitization assay,
aurones at 0.5 ␮M reduced the resistance of the transfected cells to MX to just twice that
of the parental cells, exceeding fumitremorgin C (FTC) tested at the same concentration.
The aurones (10 ␮M) also increased calcein-AM accumulation in MDCKII/MDR1 cells that
were transfected with ABCB1 (P-glycoprotein), at levels comparable to verapamil tested at
the same concentration. Structure–activity analysis showed that substitution of the ben-
zylidene ring B of the aurone template was less important for ABCG2 inhibition, with little
variation in activity noted for compounds with an unsubstituted ring B or one that was
substituted. In contrast, substitution of ring B gave rise to better inhibitors of ABCB1. A pref-
erence for the 3ꢀ position of ring B was noted. There was also some indication from the
data that aurones with good ABCG2 inhibitory activity were poor ABCB1 inhibitors and vice
versa, but further confirmation would be required. Limited antiproliferative activity (>70%
cell survival) was observed for many aurones on four different cell lines. Thus, functional-
ized 4,6-dimethoxyaurones are promising ABCG2 inhibitors that combine good activity at
submicromolar concentrations with limited antiproliferative activity.
Received 28 May 2008
Received in revised form
10 July 2008
Accepted 24 July 2008
Published on line 3 August 2008
Keywords:
4,6-Dimethoxyaurones
Modulators
ABCG2
ABCB1
Antiproliferative activity
Structure–activity relationship
© 2008 Elsevier B.V. All rights reserved.
1.
Introduction
resistance protein, BCRP) are the most relevant in the clini-
cal setting. These transporters actively extrude a broad panel
One of the major obstacles to effective cancer chemotherapy
is the acquisition of multidrug resistance (MDR) by cancer
cells (Szakács et al., 2006; Shukla et al., 2008). This phe-
nomenon is commonly associated with the overexpression of
ATP-binding cassette (ABC) transporters, of which ABCB1 (P-
glycoprotein, Pgp), ABCC1 (MRP1), and ABCG2 (breast cancer
of therapeutically useful cytotoxic drugs, suppressing their
intracellular levels to below cell-killing thresholds, thus result-
ing in treatment failures and poor patient outcomes. Various
strategies have been proposed to circumvent this problem and
these may be broadly classified as those involved in the co-
administration of transporter inhibitors and cytotoxic agents
∗
Corresponding author at: Department of Pharmacy, Faculty of Science, National University of Singapore, 18 Science Drive 4, Singapore
117543, Singapore. Tel.: +65 6516 2654; fax: +65 6779 1554.
E-mail address: phagoml@nus.edu.sg (M.-L. Go).
0928-0987/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.ejps.2008.07.008

294
european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
the compounds for their abilities to increase mitoxantrone
(MX) accumulation in human breast cancer cells (MDA-MB-
231) transfected with the human wild type (482R) BCRP cDNA,
and to re-sensitize these cells to the cytotoxic effects of
MX. To assess the selectivity of these aurones for ABCG2,
their effects on the ABCB1 transporter were investigated by
the calcein-AM accumulation assay with MDCKII cells trans-
fected with human MDR1 cDNA. Finally, antiproliferative
activities were evaluated by the MTT assay on several cell
lines.
(“engage”), discovering and developing cytotoxic drugs that
bypass transporter-mediated efflux (“evade”), and “exploit-
ing” the paradoxical sensitivity of MDR cells (Szakács et al.,
2006). The first approach of developing clinical inhibitors of
ABC transporters has been vigorously pursued but has not
resulted in a suitable clinical candidate (Robert and Jarry, 2003;
Breedveld et al., 2006).
Aurones (2-benzylidenebenzofuran-3(2H)-ones) are
a
group of lesser known flavonoids that are structural isomers
of flavones. They impart a golden yellow color to flowers
and may function as phytoalexins in plants (Pare et al.,
1991). In contrast to flavonoids, there are relatively few
reports on the biological properties of aurones. A review
cited their potential in cancer chemotherapy, as agents for
the treatment of parasitic and microbial infections, and as
inhibitors of an enzyme involved in the metabolism of thyroid
hormones (Boumendjel, 2003). They have also been reported
to be antiproliferative agents interfering with G2/M phase
of the cell cycle (Lawrence et al., 2003), tyrosinase inhibitors
(Okombi et al., 2006) and as potentially useful imaging agents
for detecting ␤-amyloid plaques in Alzheimer’s disease (Ono
et al., 2007). Several reports have focused on aurones as mod-
ulators of Pgp-mediated multidrug resistance (Boumendjel et
al., 2002a; Boumendjel, 2003; Hadjeri et al., 2003; Václavíková
et al., 2006). The inhibitory activity was ascribed to the
structural mimicry between the benzofuranone ring of the
aurone and the adenine of ATP (Boumendjel, 2003). Support
for this hypothesis came from a study that demonstrated
the binding of aurones to the cytosolic nucleotide binding
domain of mouse Pgp (Boumendjel et al., 2002a). The same
mimicry has been proposed for the chromenone ring of
flavones, but when compared with other flavonoids, aurones
showed greater binding affinity and more potent activity
as Pgp modulators (Boumendjel, 2003; Hadjeri et al., 2003;
Václavíková et al., 2006). This led to the proposal that the
conformationally restrained aurone template may have a bet-
ter fit in the ATP-binding site of the transporter (Boumendjel,
2003).
Some of the shortcomings of flavonoids as modulators of
ABC transporters are their moderate potencies and their broad
spectrum of biological activities (Boumendjel et al., 2002b).
In this respect, the greater potencies reported for aurones
may be an advantage. On the other hand, a functionalized
4,6-dihydroxyaurone was reported to have a good docking
fit in the ATP-binding site of CDK2 (Schoepfer et al., 2002).
This finding implied that aurones may have affinities for the
same site in other kinases and related proteins. If so, they
would have a broader spectrum of activities than presently
reported. Although there are several reports on the Pgp modu-
latory activity of aurones, the structural variations made to the
template has been quite limited (Boumendjel et al., 2002a,b;
Boumendjel, 2003; Hadjeri et al., 2003; Václavíková et al., 2006).
Moreover, little is known of how aurones affect ABCG2, a
clinically relevant ABC transporter that is known to interact
with flavonoids (S. Zhang et al., 2005). With this background
in mind, we prepared a series of 15 4,6-dimethoxyaurones
with different substituents on ring B. Our objectives are
to evaluate their potential as inhibitors of ABCG2, and to
assess the structure–activity relationships associated with
these modifications. Our investigations involved evaluating
2.
Materials and methods
2.1.
Synthesis of aurone library: general details
Reagents (synthetic grade or better) for the synthesis of the
aurones were obtained from Sigma–Aldrich Chemical Com-
pany Inc. (St. Louis, MO, USA) and used without further
purification. Melting points were measured on a Gallenkamp
melting point apparatus and reported as uncorrected values.
Mass spectra were captured on an LCQ Finnigan MAT fit-
ted with a chemical ionization (APCI probe) and m/z ratios
for the molecular ion were reported. 1H and 13C NMR spec-
tra were determined on a Bruker-Spectrospin 300 Ultrashield
spectrometer and referenced to TMS. Infrared spectra were
collected on solid KBr disks using a Perkin Elmer Precisely
Spectrum 100 FT-IR Spectrometer. Thin layer chromatogra-
phy was carried out on Merck Silica 60 F254 plates. Flash
chromatography was carried out using Merck silica gel,
0.040–0.063 mm. The purity of final compounds were verified
by combustion analysis (C,H) on a Perkin-Elmer PE 2400 Ele-
mental Analyzer and by high pressure liquid chromatography
(HPLC) for compound 14.
2.1.1. 3,5-Dimethoxyphenoxyacetic acid (16)
The method of Lawrence et al. (2003) was followed. To a solu-
tion of 3,5-dimethoxyphenol (0.83 g, 5.44 mmol) in anhydrous
DMF (10 ml) was added sodium hydride (0.33 g, 13.6 mmol).
Chloroacetic acid (0.51 g, 5.44 mmol) in DMF (10 ml) was added
dropwise and the mixture stirred under cover of nitrogen
gas at room temperature for 12 h. After this time, the mix-
ture was quenched with 4 M HCl (100 ml) and extracted with
dichloromethane (3× 50 ml). The organic phase was washed
with brine, dried with anhydrous Na2SO4 and evaporated in
vacuo to give brownish oil which solidified to a white solid on
cooling in an ice-bath. Recrystallisation in chloroform–hexane
gave 16 in 76% yield. White needle-like crystalline solid; mp
148–149 ◦C; 1H NMR (CDCl3, 300 MHz): ı 6.14 (t, 1H, J = 2.0 Hz),
6.11 (d, 2H, J = 2.1 Hz), 4.65 (s, 2H), 3.74 (s, 6H); MS (APCI) m/z
[M+1]+ 212.5.
2.1.2. 4,6-Dimethoxybenzofuran-3(2H)-one (17)
The method of Lawrence et al. (2003) was followed. 20 g of
polyphosphoric acid was heated to 80 ◦C after which was
added 0.5 g of compound 16. The mixture was stirred at
90 ◦C for 8 h and then poured into ice-water (100 ml) and sub-
sequently extracted with dichloromethane (4× 50 ml). The
combined organic phase was washed with brine, dried with
anhydrous Na2SO4 and evaporated in vacuo. The residue

european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
295
was purified by column chromatography, with hexane:ethyl
acetate 2:1 as eluting solvent. 17 was obtained as a pale yel-
low solid, yield: 56%; mp 137–139 ◦C; 1H NMR (CDCl3, 300 MHz):
ı 6.16 (d, 1H, J = 1.5 Hz), 6.02 (d, 1H, J = 1.1 Hz), 4.60 (s, 2H), 3.91
(s, 3H), 3.87 (s, 3H); MS (APCI) m/z [M+1]+ 195.1.
grown in DMEM medium containing 10% fetal bovine serum
and 0.1 mg/ml streptomycin sulfate and penicillin G. Human
colon carcinoma HCT116 and normal human diploid embry-
onic lung fibroblasts CCL-186 were purchased from American
Type Culture Collection (Rockville, MD). HCT116 cells were cul-
tured in McCoy’s 5A modified medium supplemented with 10%
FBS, 0.1% penicillin G and 0.1% streptomycin. CCL-186 cells
were grown in Eagle’s minimal essential medium (EMEM) sup-
plemented with 10% FBS, 0.3% l-glutamine, 0.1% penicillin G
and 0.1% streptomycin. All the cells were sub-cultured when
they reached 80–90% confluency and used within 10 passages
for the assays.
2.1.3. General procedure for the synthesis of aurones
The method of Lawrence et al. (2003) was followed. To a
solution of 17 (100 mg, 0.52 mmol) in methanol (10 ml) was
added the substituted benzaldehyde (0.76 mmol), followed by
a solution of KOH (500 mg, 8.92 mmol) in distilled water (1 ml).
The solution was stirred at room temperature for 1–3 h. The
desired compound was obtained as a precipitate, removed
by suction filtration, washed with cold methanol and crys-
tallized in ethanol or methanol to give the purified aurone
(1–15). In the case of aurones 6 and 7, additional protec-
tion and deprotection of the phenolic hydroxyl group on the
benzaldehyde were required. Spectroscopic and other ana-
lytical details of aurones 1–15 are given in Supplementary
Information.
2.3.
Western blot analysis
Fifteen microgram of cell lysates from MDA-MB-231 (R and
V cells), and 40 ␮g of cell lysates from MDCKII (wild type)
and MDCKII/MDR1 were subjected to electrophoresis on 7.5%
sodium dodecyl sulphate-polyacrylamide gel (SDS-PAGE) and
transferred to nitrocellulose membranes (Bio-Rad Labora-
tories Pte Ltd., Singapore) for overnight blocking at room
temperature. The blots were probed with primary antibodies
followed by horseradish peroxidase-conjugated anti-mouse
secondary antibody. The primary antibodies used were anti-
BCRP (BXP-21), anti-Pgp (C219), anti-MRP1 (MRPr-1) and
anti-MRP2 (M2III-6). The blots were washed and incubated
with West Femto luminal/enhancer solution (Pierce Chemi-
cals, Rockford, IL) and stable peroxide solution for 5 min. The
membranes were analyzed with FluorChemTM 9900 (Alpha
Innotech Corporation, San Leandro, CA).
2.1.4. Protection and deprotection of phenolic hydroxyl
groups on benzaldehydes for syntheses of aurones 6 and 7
3-Hydroxybenzaldehyde
or
4-hydroxylbenzaldehyde
(3 mmol), pyridinium p-toluenesulphonate (50 mg, 0.2 mmol)
and 3,4-dihydro-2H-pyran (673 mg, 8 mmol) were dissolved
in dichloromethane (10 ml) and stirred for 4 h at room tem-
perature. The reaction mixture was then washed with 1 M
Na2CO3 (3× 20 ml). The organic layer was dried over anhy-
drous Na2SO4 and concentrated in vacuo to give the crude
tetrahydropyranyl ether (18) as a yellow oil which was used
without purification for the condensation reaction with 17
as described in Section 2.1.3. After the period of stirring,
the reaction mixture was acidified with 4 M HCl, stirred
for 4 h at room temperature, and filtered under vacuum to
give the crude product (6, 7) which was recrystallized from
methanol.
2.4.
Mitoxantrone accumulation studies
The accumulation of MX was determined by flow cytometry
on MDA-MB-231/R and MDA-MB-231/V cells by a previously
described method (Minderman et al., 2002). Stock solutions
of aurones, MX and FTC were prepared in DMSO and diluted
with media to give the desired concentrations. The final con-
centration of DMSO was kept at no more than 0.1% (v/v).
Briefly, cells were grown in 75-cm3 flasks to about 90% conflu-
ency, trypsinized, washed with ice-cold PBS and resuspended
in serum-free RPMI medium at a cell density of 106 cells/ml.
MDA-MB-231/R cells were incubated with various concentra-
tions of aurones (0.05 ␮M, 0.5 ␮M, 5 ␮M) or vehicle (0.1% DMSO
in RPMI) at 37 ◦C for 15 min, followed by the addition of MX
(3 ␮M) for 30 min. FTC, a specific inhibitor of BCRP, was tested
at 10 ␮M as a positive control. After the incubation period,
ice-cold PBS (10 ml) was added to stop the process and the
cells were removed by centrifugation (2000 rpm, 5 min, 4 ◦C).
The cell pellet was washed with ice-cold PBS (3×) and resus-
pended in cold PBS for the determination of the intracellular
concentration of MX on a FACScan flow cytometer (CyAn®
Research Flow Cytometer, Dako, Glostrup, Denmark). Cells
were excited at 488 nm and the emission recorded via a 680 nm
long-pass filter for the detection of MX fluorescence. The data
were analyzed with the instrument’s software (Summit ver-
sion 4.3, Dako, Glostrup, Denmark). MX levels in the cells
were normalized to the vehicle control (0.1% DMSO) which
was taken as 100%. At least three independent determinations
were carried out for each test compound. MX accumulation in
2.2.
Materials for biological assays
Mitoxantrone (MX), fumitremorgin
C
(FTC), calcein ace-
toxymethyl ester (calcein-AM), verapamil, cyclosporin A,
3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide
(MTT) were purchased from Sigma–Aldrich, St. Louis, MO,
USA. The breast cancer cell line, MDA-MB-231, stably trans-
fected with expression vectors for wild type 482R BCRP (R
cells) and pcDNA3.1 (parental V cells) were kindly provided
by Dr. Douglas D. Ross (Greenebaum Cancer Center, Univer-
sity of Maryland, Baltimore, USA). Both MDA-MB-231/V and
MDA-MB-231/R cells were cultured in 75-cm3 flasks with RPMI
1640 (Invitrogen Corporation, CA, USA) culture media supple-
mented with 10% fetal bovine serum (Hyclone, UT, USA) at
37 ◦C in a 5% CO2 humidified atmosphere. The culture media
contained 0.1 mg/ml streptomycin sulfate and 0.1 mg/ml peni-
cillin G (Sigma Chemical Co., St. Louis, MO, USA) and 1.0 mg/ml
geneticin (Invitrogen Corporation, CA, USA). Mardin-Darby
canine kidney (MDCK) cells transfected with the expression
vector for human mdr1 cDNA was a gift from Dr. Anton
Berns (Netherlands Cancer Institute, Antoni van Leeuwen-
hoek Hospital, Amsterdam, Netherlands). The cells were

296
european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
the MDA-MB-231 cells was expressed by the following equa-
tion:
2.7.
Uptake of calcein-AM in MDCK/MDR1 cells
A previously described method was followed with some
modifications (Liu et al., 2008). Briefly, MDCKII/MDR1
and MDCKII/WT cells were grown to 80–90% confluency,
trypsinized and seeded in 96-well plates at a cell den-
sity of 2.5 × 104 cells/well. After incubation at 37 ◦C for
MX accumulation (%)
area(cells+test compound+MX) − area(cells)
=
× 100
area(cells+MX) − area(cells)
24 h,
a monolayer of cells was obtained at the bottom
of the well. The medium was decanted and the mono-
layer carefully washed with PBS. Stock solutions (10 mM)
of test compounds were prepared in DMSO and diluted
with Hank’s buffered saline solution (HBSS) such that addi-
tion of an aliquot (100 ␮l) gave a final concentration of
10 ␮M test compound in the well. The final concentration of
DMSO in each well did not exceed 1% (v/v). After incubation for
30 min (37 ◦C, 5% CO2 atmosphere), calcein-AM in HBSS–DMSO
(100 ␮l) was added to each well to give a final concentration
of 2 ␮M. Incubation was continued for another 10 min, after
which fluorescence was measured at 10 min intervals for
50 min on a microplate reader with ꢀexcitation of 485 nm and
ꢀemission of 535 nm. Concurrent determinations were made
for the positive controls verapamil and cyclosporin A, both at
10 ␮M. The accumulation of calcein-AM was calculated at the
50th minute using the following equation:
2.5.
Mitoxantrone sensitization
The ability of aurones to re-sensitize MDA-MB-231/R cells to
MX was assessed by determining the growth inhibitory IC50
of MX in the presence of various concentrations of the test
aurone. Stock solutions of MX, FTC and test aurones were
prepared in DMSO and diluted with media so that final con-
centrations in the well did not exceed 1% (v/v) DMSO. Cells
(104) were seeded in 96-well plates and incubated for 24 h,
after which, the culture medium in each well was replaced
with fresh medium containing various concentrations of MX
and test aurone (0.05 ␮M or 0.5 ␮M). FTC (10 ␮M) was tested as
positive control. After 72 h of incubation, the drug-containing
medium was removed and cells were washed once with PBS.
Hundred microlitres of 0.5 mg/ml MTT was then added to each
well, incubated for 3 h at 37 ◦C, after which the MTT solu-
tion was decanted and DMSO (100 ␮l) was added to dissolve
the purple formazan crystals. Absorbance was read at 590 nm
on a microplate reader with at least three separate determi-
nations made for each concentration of test compound. The
absorbance values obtained at a known concentration of MX
were averaged and adjusted by subtraction of absorbance of
empty wells (“blank”). This value was expressed as a percent-
age of the average absorbance obtained from control wells
(wells with cells and media containing 1% DMSO) which were
also adjusted by subtraction of the blank to give % surviv-
ing cells. The IC50 value of MX was determined from the
sigmoidal curve obtained by plotting % surviving cells at vari-
ous concentrations of MX using GraphPad Prism (Version 4.00,
GraphPad Software, San Diego, CA). The software used the
four-parameter logistic equation given by the expression for
calcein-AM accumulation (%)
FMDCKII/MDR1+test compound
=
× 100
FMDCKII/MDR1
2.8.
Molecular modeling
Structures of aurones were drawn and geometry minimized
using the Hamiltonian forcefield MMFF94x in MOE® (Chemical
Computing Group, Montreal, Canada). The following molec-
ular descriptors were collected for the aurones using the
QuaSAR module of MOE®: constitutive (number of rotatable
bonds, hydrogen bond donors, hydrogen bond acceptors),
lipophilic (calculated log P in oil/water), geometric (van der
Waals surface area of hydrogen bond acceptors and hydrogen
bond donors, total hydrophobic van der Waals surface area,
topological polar surface area) and electronic (sum of atom
polarizabilities). These descriptors are given in Table 1 of Sup-
plementary Information The software was also used to deter-
mine distances and angles between rings and hydrogen
bond donor groups in selected energy-minimized compounds.
Principal component analysis (PCA) and projection to latent
structures by partial least squares analysis (PLS), available
on SIMCA-P 11 (Umetrics AB, Umea, Sweden), were used to
develop structure–activity relationships.
determination of IC50
:
bottom + (top − bottom)
Y =
,
−X)×HillSlope)
(1 + 10(logIC
)
50
where X is the logarithm of concentration and Y is the
response.
2.6.
Cytotoxicity studies
The cytotoxicity profiles of the aurones on different cell lines
were determined by the MTT assay as described in Section 2.5.
The cell lines tested were the human breast cancer cell line
MDA-MB-231 (V and R), human colon carcinoma cell line HCT
116 and the normal human fibroblast CCL 186. Stock solutions
of aurones were prepared in DMSO and diluted with media to
give concentrations of 10 ␮M for investigations on MDA-MB-
231 (V and R) cells, and 5 ␮M on the other two cell lines. The
final concentration of DMSO in the well did not exceed 1%
(v/v). Cell survival was determined as a % of control cells and
at least two determinations were made for each concentration
of aurone.
2.9.
X-ray crystallography
Crystals of 5-hydroxy-2-(4ꢀ-methoxybenzylidene)-benzo-
furan-3(2H)-one and aurone 8 were grown in methanol and
dichloromethane solutions, respectively and mounted on
glass fibres. X-ray data were collected with a Bruker AXS
SMART APEX diffractometer, using Mo K␣ radiation at 223 K,
with the SMART suite of Programs (SMART version 5.628
(200), Bruker AXS Inc., Madison, WI). Data were processed

european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
297
and corrected for Lorentz and polarization effects with SAINT
Table 1 – Structures of 4,6-dimethoxyaurones (1–15)
(SAINT+ version 6.22a (2001) Bruker AXS Inc., Madison, WI),
and for absorption effect with SADABS (SADABS, version
2.10 (2001), University of Göttingen). Structural solution
and refinement were carried out with the SHELXTL, suite
of programs (SHELXTL, Version 6.14 (2000), Bruker AXS Inc.,
Madison, WI). The structure was solved by direct methods
to locate the heavy atoms, followed by difference maps for
the light, non-hydrogen atoms. All non-hydrogen atoms
were generally given anisotropic displacement parameters in
the final model whereas H-atoms were placed at calculated
positions.
Compound
R
1
2
3
H
2ꢀ-Cl
3ꢀ-Cl
4
4ꢀ-Cl
5
6
7
8
2ꢀ-OH
3ꢀ-OH
4ꢀ-OH
2ꢀ-OCH3
3ꢀ-OCH3
4ꢀ-OCH3
2ꢀ-CH3
3ꢀ-CH3
4ꢀ-CH3
3ꢀ-CN
4ꢀ-CN
2.10. Statistical analysis
Data were analyzed for statistically significant differences
using one-way ANOVA followed by a Dunnett’s post hoc test
(SPSS 15.0 for Windows, Chicago, IL). p-Values <0.05 were con-
sidered statistically significant.
9
10
11
12
13
14
15
3.
Results
3.1.
Synthesis of aurones (1–15)
The substituents (chloro, methoxy, hydroxy, methyl, cyano)
were selected based on their different locations on the Craig
Plot (Craig, 1971). Hence, substituents with a range of lipophilic
and electron withdrawing/donating effects were represented
in this series. Each substituent was attached to three different
positions (2ꢀ, 3ꢀ or 4ꢀ) on ring B, with the exception of the cyano
group where only two regioisomers (3ꢀ, 4ꢀ) were prepared.
The 4,6-dimethoxyaurones were synthesized following
previous reports (Beney et al., 2001; Lawrence et al., 2003) and
the key steps are illustrated in Scheme 1. The first step (Step
a) involved the synthesis of 3,5-dimethoxyphenoxyacetic
The structures of the synthesized aurones (1–15) are given
in Table 1. These aurones had methoxy groups at positions
4 and 6 of ring A and differed only in the type of substituent
on ring B. The choice of methoxy groups on the ring A was
prompted by the docking results of Schoepfer et al. who
showed that a functionalized 4,6-dihydroxaurone interacted
with the ATP-binding site of CDK2 by hydrogen bonding via the
4-hydroxy and carbonyl groups (Schoepfer et al., 2002). Intro-
ducing methoxy in place of hydroxyl groups would minimize
interactions with the ATP-binding sites of CDK2 and related
proteins, and possibly reduce unwanted interactions. More-
over, Hadjeri and co-workers identified 4,6-dimethoxyaurone
as an outstanding modulator of Pgp-mediated multidrug resis-
tance, which showed that methoxy groups did not adversely
affect, and may even increase affinity for the transport protein
(Hadjeri et al., 2003).
acid (16) by
a hydride-catalyzed condensation of 3,5-
dimethoxyphenol with chloroacetic acid. The product was
subjected to intramolecular cyclization via Friedal–Craft acy-
lation (Step b) to form the benzofuran-3-one ring which was
then reacted with various benzaldehydes in a base-catalyzed
aldol reaction to give the target aurones in moderate yields.
Scheme 1 – General procedure for the synthesis of 4,6-dimethoxyaurones. Reagents and conditions: (a) chloroacetic acid,
NaH, DMF, rt, 12 h. (b) Polyphosphoric acid, 80 ◦C, 8 h. (c) Substituted benzaldehyde, 50% KOH in MeOH/H2O, rt, 3 h.

298
european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
Scheme 2 – Protection and deprotection of phenolic hydroxyl group on 3ꢀ-OH benzaldehyde for synthesis of aurone 6.
Reagents and conditions: (a) pyridinium p-toluenesulphonate (0.2 mmol), CH2Cl2, rt, 4 h. (b) 50% KOH in MeOH/H2O, rt, 3 h.(c)
4 M HCl, rt, 4 h.
which implied the presence of an E configuration. Thus, the
configuration of the 2ꢀ-substutituted aurones could not be
unequivocally assigned based on chemical shifts alone. In
order to resolve the problem, the X-ray structure of aurone 8
(2ꢀ-methoxy on ring B) in the solid state was examined (Fig. 1a).
It was found to have the Z configuration and this led us to
assign the same configuration to the remaining 2ꢀ-substituted
aurones. The X-ray structure of another aurone, 5-hydroxy-2-
(4ꢀ-methoxybenzylidene)-benzofuran-3(2H)-one that was not
included in the present series, was also available (Lee, unpub-
lished results). This aurone had a 4ꢀ-methoxy substituent on
ring B and a chemical shift of ı 6.87 ppm for its ␤ hydro-
gen. As anticipated, the X-ray analysis showed that it had
a Z configuration. Thus, spectroscopic and crystallographic
data supported the assignment of a Z configuration for the
aurones. The down field shift of the ␤ hydrogen observed
for the 2ꢀ-substituted aurones may be due to the anomalous
“ortho effect” often associated with such substituents.
For aurones 6 and 7, the reacting benzaldehyes (3 and
4-hydroxybenzaldehydes) required protection of their pheno-
lic hydroxyl groups by conversion to the tetrahydropyranyl
ether prior to condensation with 17. Thereafter, the protect-
ing moiety was removed by acid hydrolysis to give the desired
hydroxylated aurones 6 and 7 (Scheme 2). Protection was not
necessary for 2-hydroxybenzaldehyde.
3.2.
Assignment of configuration of aurones
The 4,6-methoxyaurones may exist as either E (trans) or Z (cis)
isomers, with the Z isomer generally regarded as the thermo-
dynamically more stable form (Ur-Rahman et al., 2001) Many
authors have made the Z/E assignment of aurones based on
the chemical shift of the olefinic (␤) proton (Beney et al., 2001;
Thakkar and Cushman, 1995; Okombi et al., 2006). The olefinic
proton in the E isomer is deshielded to a greater extent than
the same proton in the Z isomer and hence has a larger chem-
ical shift. In the literature, a chemical shift of ı 6.70 ppm was
cited for the Z isomer and ı 7.01 ppm for the E isomer (Beney
et al., 2001; Thakkar and Cushman, 1995). Z/E isomers were
also distinguished by the 13C chemical shift of the exocyclic
carbon ( CH) (Pelter et al., 1979). In the Z isomer, the chemical
shift was reported at ı 111 ppm while in the E isomer, it was
observed at a higher frequency of ı 120–130 ppm.
Both 13C and 1H chemical shifts were considered for the
assignment of the Z/E configuration of the aurones. The 13C
chemical shifts of the exocyclic carbon in all compounds
appeared at ı 105–111 ppm, in keeping with a Z configu-
ration of the exocyclic double bond. However, there was
some ambiguity when the 1H chemical shifts of the ␤ hydro-
gen were considered. Most of the aurones had values in
the range of ı 6.68–6.77 ppm, consistent with reported val-
ues for known Z aurones. On the other hand, aurones with
2ꢀ-substituents (2, 5, 8, 11) had values of ı 7.00–7.38 ppm,
3.3.
Effect of aurones on cell survival of MDA-MB-231,
HCT116 and CCL-186 cells
The aurones were evaluated for antiproliferative activity on
human breast cancer cells that were transfected with the plas-
mid vector carrying human wild type ABCG2 (MDA-MB-231/R)
or the empty vector plasmid (MDA-MB-231/V). The presence
of ABCG2 in MDA-MB-231/R cells was verified by Western blot
analysis. ABCG2 was not detected in the MDA-MB-231/V cells
(Fig. 2). The cell model was also investigated for the expres-
sion of other transport proteins (ABCB1, ABCC1, and ABCC2)
and these were not detected (data not shown).
When tested at 10 ␮M against MDA-MB-231/V and R cells,
some aurones were found to be slightly more potent against V
cells than R cells. These were 2, 4, 6, 8 which at 10 ␮M caused
about 30–85% survival of V cells but 100% survival of R cells.

european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
299
Fig. 1 – Molecular structures of (a) 4,6-dimethoxy-2-(2ꢀ-methoxybenzylidene)-benzofuran-3(2H)-one (8) and (b)
5-hydroxy-2-(4ꢀ-methoxybenzylidene)-benzofuran-3(2H)-one, showing the Z configuration of the exocyclic double bond.
Aurone 2 had the greatest cytotoxic effect on V cells, with 37%
cell survival at 10 ␮M and 62% survival at 5 ␮M. However, only
two determinations were made for these compounds against
the V cells and would require further confirmation.
To determine if the effects were tissue- or tumour-specific,
the aurones (except 11, 12, 14) were tested at 5 ␮M against a
colon cancer cell line (HCT116) and a normal cell line (CCL186).
All the aurones had limited activity (>80% cell survival) against
the normal cell line but affected the survival of HCT116 cells to
a greater extent, with aurones 2 and 3 associated with partic-
ularly low levels of cell survival (37%, 50%). Aside from these
compounds, the other aurones showed limited antiprolifera-
tive activity against the panel of four cell lines (Table 2).
3.4.
Effect of aurones on the efflux of mitoxantrone
Fig. 2 – Western blot analyses of ABCG2 in MDA-MB-231 (R
and T) cells. aAnalyses of ABCG2 expression in
MDA-MB-231/R and MDA-MB-231/V cells. The protein
loading for each sample was 15 ␮g as determined by
protein assay. b␤-Actin was used as a positive control for
the Western blot analyses.
from ABCG2-expressing MDA-MB-231 cells
Mitoxantrone is a substrate of ABCG2 and accumulates to a
lesser extent in the ABCG2 expressing MDA-MB-231/R cells
than non-ABCG2 expressing MDA-MB-231/V cells. In our
hands, V cells accumulated 2.4-fold more MX than R cells.

300
european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
Table 2 – Effects of aurones on percentage cell survival in various cell linesa
Compound % Cell survival
MDA-MB-231/V
R
MDA-MB-231/R
HCT116
CCL 186
10 ␮M
10 ␮M
5 ␮M
5 ␮M
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
H
122 (9)
115
101
37
102
87
101
83
156
86
133
122
134
90
120
158
179
76
27
50
60
56
93
94
67
>100
93
>100
>100
>100
97
>100
88
>100
92
NDc
NDc
>100
NDc
>100
2ꢀ-Clb
3ꢀ-Cl
4ꢀ-Cl
103 (16)
126 (21)
115 (16)
105 (11)
114 (8)
126 (31)
112 (12)
136 (16)
116 (32)
83 (32)
116 (40)
113 (8)
108 (7)
2ꢀ-OH
3ꢀ-OH
4ꢀ-OH
2ꢀ-OCH3
3ꢀ-OCH3
4ꢀ-OCH3
2ꢀ-CH3
3ꢀ-CH3
4ꢀ-CH3
3ꢀ-CN
4ꢀ-CN
74
80
NDc
NDc
81
NDc
94
a
Cell survival (%) = (average absorbance of wells with test compounds at 590 nm/average absorbance of wells with medium and 1% DMSO) × 100.
Values are mean of two or more independent determinations. S.D. in parenthesis for n = 3 determinations.
Percentage cell survival in MDA-MB-231/V cells for compound 2 at 5 ␮M = 62%, mean of two independent determinations.
ND, not determined.
b
c
The accumulation of MX in R cells was determined in the
presence of three different concentrations of test aurone
(5 ␮M, 0.5 ␮M, 0.05 ␮M) (Table 3). The results showed that the
extent of MX accumulation varied according to the concen-
tration used. Aurones tested at 0.05 ␮M had no significant
effect on MX accumulation. When the concentration was
increased to 0.5 ␮M, 10 aurones increased MX accumulation
to a significant extent (p < 0.05) compared to untreated cells
and at a higher concentration of 5 ␮M, 11 members caused
significant MX accumulation. Thus, a large number of the
aurones were effective at 0.5 ␮M and they caused accumula-
tion of MX to levels comparable to that of the known ABCG2
inhibitor FTC which was tested at a higher concentration of
10 ␮M.
Table 3 – Accumulation of MX in MDA-MB-231/R cells in presence of aurones
Compound
R
MDA-MB-231/R MX accumulation (% of control)a
Aurone concentration
5 ␮M
0.5 ␮M
100
0.05 ␮M
Control (0.1% DMSO)
FTC (10 ␮M)
H
237 (60.9)*b
209 (32.7)*
226 (22.6)**
231 (30.9)**
195 (44.0)
214 (29.9)*
235 (29.1)**
219 (26.7)*
217 (25.1)*
222 (31.9)*
215 (30.1)*
163 (53.6)
203 (60.2)*
130 (27.1)
172 (40.5)
161 (50.9)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
283 (86.6)*
275 (62.9)*
275 (55.3)*
233 (23.8)
235 (50.5)
273 (65.5)*
263 (54.3)*
251 (46.8)
294 (70.9)**
278 (57.1)*
321 (146)**
291 (46.2)**
268 (48.7)*
358 (64.7)***
187 (39.6)
150 (5.1)
2ꢀ-Cl
148 (30.3)
160 (20.9)
125 (8.7)
3ꢀ-Cl
4ꢀ-Cl
2ꢀ-OH
131 (16.3)
139 (11.1)
138 (23.5)
148 (18.2)
169 (18.0)
146 (21.8)
72.6 (24.0)
117 (14.1)
88.8 (13.8)
117 (29.4)
81.0 (31.4)
3ꢀ-OH
4ꢀ-OH
2ꢀ-OCH3
3ꢀ-OCH3
4ꢀ-OCH3
2ꢀ-CH3
3ꢀ-CH3
4ꢀ-CH3
3ꢀ-CN
4ꢀ-CN
a
MX
accumulation
(%
of
control)
in
MDA-MB-231/R
cells
calculated
as:
MX
accumulation
(%
of
control)
(%) = [area(cells+test compound+MX) − area(cells)/area(cells+MX) − area(cells)] × 100. Values are expressed as mean of two or more independent
determinations. S.D. in parenthesis for n = 3 determinations.
Significant difference (*p < 0.05, **p < 0.01, ***p < 0.001) between MX accumulation in the presence of test aurone and control (0.1% DMSO)
(one-way ANOVA followed by Dunnett post hoc test).
b

european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
301
Western blots showed that MDA-MB-231/R cells had no
detectable levels of ABCB1, ABCC1 or ABCC2 expression.
Therefore the increase in MX accumulation in MDA-MB-231/R
cells was ascribed to the inhibition of ABCG2 by the aurones.
To further confirm this point, the effect of aurones at 5 ␮M
was investigated in MDA-MB-231/V cells that were not trans-
fected with ABCG2. It was found that levels of accumulation
were not different from those observed in untreated control
V cells (Supplementary Information, Table 2). Therefore, the
higher levels of MX accumulation in the R cells resulted from
the inhibition of ABCG2 activity by the aurones.
Fig. 3 – IC50 determinations of mitoxantrone (MX) in
3.5.
Effect of selected aurones (3,6,9,14) on the IC50 of
parental MDA-MB-231/V cells (ꢀ), MDA-MB-231/R cells (ꢁ),
MDA-MB-231/R cells in presence of 3 (0.5 ␮M) (ꢂ).
mitoxantrone in MDA-MB-231/R cells
In order to confirm the functional relevance of the aurone-
induced increase in MX accumulation in MDA-MV-231/R cells,
selected aurones (3, 6, 9, 14) were tested for their ability to
restore MX sensitivity to the ABCG2 expressing MDA-MV-231/R
cells. 3, 6 and 9 were chosen because of their good outcomes in
the MX accumulation assays, and 14 was selected as a negative
control as it had a lesser effect on MX accumulation (Table 3).
Re-sensitization was assessed by determining the effect of a
fixed concentration of aurone on the IC50 of MX on MDA-MB-
231/R cells. In the absence of aurones, the IC50 values of MX
were 0.09 ␮M (V cells) and 3.45 ␮M (R cells) (Table 4). The 38-
fold difference in IC50 values was ascribed to ABCG2 which
extruded MX from R cells but not V cells. The IC50 of MX was
reassessed in the presence of either 0.05 ␮M or 0.5 ␮M aurone.
At these concentrations, the aurones did not adversely affect
cell survival, as described in Section 3.2. FTC at 0.05 and 0.5 ␮M
were included as positive controls. A representative curve for
the IC50 determination of MX in the presence of aurone 3 is
given in Fig. 3.
A statistical comparison of the IC50 values of MX in
untreated R cells (no aurone) and treated R cells (with aurone)
showed a significant lowering of the IC50 of MX at both aurone
concentrations (0.5 ␮M, 0.05 ␮M) (p < 0.05). At the higher con-
centration of 0.5 ␮M, the aurones caused more than 10-fold
increase in MX potency, with aurones 3 and 9 associated with
a 23–25-fold increase. At this concentration, the aurones were
also more effective than FTC in restoring sensitivity of the cells
to MX. It was noted that aurone 14 which was meant to be the
negative control also re-sensitized the R cells to MX.
Table 4 – Effect of selected aurones on IC50 of
mitoxantrone in MDA-MB-231/R cells
The effect of the aurones on the fold resistance index of MX
on the R and V cell lines was also considered. In the absence of
aurones, the fold resistance index of MX on the two cell lines
was 38(IC50 MX(R cells)/IC50 MX(V cells), which meant that MX was
38 times less sensitive on the transfected R cells compared
to the parental V cells. With FTC (0.5 ␮M), the difference in
sensitivity was reduced to sevenfold, but with aurones 3 and
9, the index was further reduced to 1.6–1.7. It is likely that
when tested at a higher concentration, these aurones would
restore MX to its original sensitivity against V cells.
Compound
R
IC50 of mitoxantrone (␮M)a,b
MDA-MB-231/R
Fold resistance
indexc
Control
3.45 (1.34)
(3.45/0.09 = 38.3)
0.05 ␮M
3
6
9
3ꢀ-Cl
1.67 (0.52)d*
1.77 (0.67)*
1.11 (0.19)*
0.95 (0.27)**
0.87 (0.22)**
18.5
19.7
12.3
10.6
9.7
3ꢀ-OH
3ꢀ-OCH3
3ꢀ-CN
FTC
14
0.5 ␮M
3
6
9
3.6.
Effect of aurones on the accumulation of
3ꢀ-Cl
0.14 (0.07)***
0.26 (0.10)***
0.15 (0.04)***
0.30 (0.17)***
0.67
1.6
2.9
1.7
3.3
7.4
calcein-AM in MDCKII/MDR1 cells
3ꢀ-OH
3ꢀ-OCH3
3ꢀ-CN
FTCe
The aurones were investigated at 10 ␮M for their effects on the
accumulation of calcein-AM, a ABCB1 substrate, by MDCKII
cells transfected with ABCB1 (MDCKII/MDRI). Western blot
analysis confirmed the presence of ABCB1 in these cells (Fig. 4).
The parental cells (MDCKII/WT) showed constitutive levels of
ABCB1 but at lower levels compared to the transfected cell line.
Calcein-AM is non-fluorescent until it is taken up by cells
where it is hydrolyzed by enzymes to give calcein which flu-
oresces. Thus, the accumulation of calcein-AM is assessed
from the intensity of fluorescence and inhibitors of ABCB1 will
cause an increase in the fluorescence intensity. In this experi-
ment, known ABCB1 inhibitors—cyclosporin A and verapamil
(both at 10 ␮M) were used as controls and they increased
14
a
IC50 values are presented as mean of two or more independent
determinations. S.D. in parenthesis for n = 3 determinations.
IC50 value of MX in MDA-MB-231/V cells = 0.09 (0.07) ␮M.
b
c
Fold
resistance
index
to
MX = IC50
of
MX
in
R
cells(absence or presence of aurone)/IC50 of MX in V cells.
d
Significant difference (*p < 0.05, **p < 0.01, ***p < 0.001) between IC50
of MX in the presence of test aurone and the absence of test
aurone (control) in MDA-MB-231/R cells (one-way ANOVA fol-
lowed by Dunnett post hoc test).
IC50 value of MX in MDA-MB-231/R cells in the presence of FTC at
10 ␮M = 0.33 ␮M determined from duplicate determinations.
e

302
european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
3.7.
Structure–activity relationships
A comparison of inhibitory trends among aurones based on
the levels of MX and calcein-AM accumulation (at 0.5 ␮M and
10 ␮M test aurone, respectively) showed some interesting dif-
ferences between these two activities. First, aurone 1 with
no substituent on ring B significantly increased MX but not
calcein-AM accumulation, which would suggest that a sub-
stituted ring B was less important for inhibition of ABCG2
as compared to ABCB1. Second, several aurones with 3ꢀ-
substitutents (3, 6, 9, 12) were associated with better ABCB1
inhibitory activities than their respective 2ꢀ and 4ꢀ regioi-
somers. Such a trend was not seen with ABCG2 inhibition.
Third, a larger number of aurones significantly increased MX
accumulation at 0.5 ␮M (Table 3) as compared to calcein-AM
accumulation (Table 5), even though a higher concentration of
aurone was used for the latter. Lastly, there was some indica-
tion from the data that aurones with good ABCG2 inhibitory
activity were poor ABCB1 inhibitors and vice versa. Aurones
1, 2, 5, 7, 8 were significantly better ABCG2 inhibitors (0.5 ␮M,
compared to control, p < 0.05) but these same aurones were no
better than the control when evaluated as ABCB1 inhibitors
(Tables 3 and 5). In the case of aurones 13 and 14, they were
better inhibitors of ABCB1 than ABCG2, but when 14 was sub-
sequently tested for re-sensitization of MBA-MB-231/R cells
to MX, it fared just as well as the other active aurones (3,
6, 9). This may imply that there are limitations in using a
non-concentration related dependent variable (like MX accu-
mulation at a fixed concentration) for analysis.
Fig. 4 – Western blot analyses of ABCB1 in MDCK II/MDR1
and MDCK II/WT cells. aAnalyses of ABCB1 expression in
MDCKII/MDR1 and MDCKII/WT cells. The protein loading
for each sample was 40 ␮g as determined by protein assay.
b␤-Actin was used as a positive control for the Western blot
analyses.
fluorescence intensity as anticipated. The accumulation of
calcein-AM was also monitored in parental MDCK II cells
and a ninefold difference was detected between the parental
and transfected cell lines, with more calcein-AM accumulated
in the parental MDCK II cells. Unlike the MX accumulation
assay, the aurones were tested at a relatively high concen-
tration of 10 ␮M on the calcein-AM assay. The choice of this
concentration was to allow comparisons to be made with ver-
apamil. In our hands, verapamil caused a significant increase
in calcein-AM accumulation (300%) at 10 ␮M. Lower concen-
trations would cause smaller increases which would hamper
comparisons with the test aurones. As it turned out, none
of the aurones were significantly better than verapamil even
at 10 ␮M (Table 5) although several (3, 6, 9, 10, 12, 13, 14)
increased calcein-AM to the same extent as verapamil.
In order to have
a better understanding of these
structure–activity trends, a quantitative structure–activity
analysis was attempted. Descriptors that encompassed con-
stitutive, geometric, lipophilic and electronic characteristics
were collated and used in the analysis. ABCG2 inhibitory activ-
ity was expressed in terms of MX accumulation (normalized to
100%) at 0.5 ␮M test compound, and ABCB1 inhibitory activity
(normalized to 100%) was expressed in terms of calcein-AM
accumulation at 10 ␮M test compound. PCA and projection to
latent structures by PLS were attempted but both approaches
did not provide meaningful models for interpretation. Possi-
ble reasons may be the limited number of compounds in this
series, their restricted structural diversity, and the absence of a
concentration term (like EC50 for increasing MX or calcein-AM
accumulation to 50% of the maximum) as dependent variables
for the analysis.
Matsson and co-workers (2007) proposed a three point
pharmacophore for inhibitors of ABCG2 which comprised of
two hydrophobic features and one hydrogen (H) bond accep-
tor function. This model was derived from 28 structurally
heterogenous ABCG2 inhibitors and characterized by spec-
ified distances and angles. We were interested to evaluate
the fit of the aurones on this model in view of their ABCG2
inhibitory activity. Aurone 3 was chosen as a representa-
tive member and two other chalcones (20 and 21) that were
tested for ABCG2 inhibitory activity in our laboratory were
also assessed. Chalcone 20 at 0.5 ␮M re-sensitized MDA-MB-
231/R cells to MX with a fold resistance index of 8.5 (Han et
al., 2008) while chalcone 21 had no effect on MX accumula-
tion in ABCG2-overexpressing MCF-7 cells (Liu et al., 2008).
Although the three compounds were tested in our laboratories
Table 5 – Accumulation of calcein-AM in MDCKII/MDR1
cells in the presence of aurones
Compounda
R
Calcein-AM
accumulation
(% of control)b
MDCKII/MDR1 (control)
MDCKII/MDR1 (verapamil)
MDCKII/MDR1 (cyclosporin A)
100
301 (30.2)c ***
1000 (76.0)***
182 (7.87)
1
2
3
H
2ꢀ-Cl
108 (1.48)
242 (21.3)***
137 (6.26)
3ꢀ-Cl
4
5
6
7
8
9
10
11
12
13
14
15
4ꢀ-Cl
2ꢀ-OH
3ꢀ-OH
4ꢀ-OH
2ꢀ-OCH3
3ꢀ-OCH3
4ꢀ-OCH3
2ꢀ-CH3
3ꢀ-CH3
4ꢀ-CH3
3ꢀ-CN
4ꢀ-CN
107 (5.27)
209 (30.1)**
121 (9.89)
122 (11.1)
303 (55.6)***
294 (15.7)***
111 (4.26)
362 (61.0)***
255 (65.7)***
245 (47.6)***
126 (9.57)
a
Determined at 10 ␮M test compound at the 50th minute. Values
are expressed as mean of n = 3 determinations. S.D. given in paren-
thesis.
b
c
Calcein-AM
[FMDCKII/MDR1] × 100.
accumulation
(%) = [FMDCKII/MDR1 + test compound]/
Significant difference (**p < 0.01, ***p < 0.001) between calcein-AM
accumulation in the presence of test aurone and control (0.1%
DMSO) (one-way ANOVA followed by Dunnett post hoc test).

european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
303
Table 6 – Distances and angles of the pharmacophoric triangle for aurone 3 and chalcones 20, 21
Mattson et al.
Aurone 3b
Chalcone 20b
Chalcone 21b
Distances (Å) between pharmacophore pointsa
Ring A–Ring B
Ring A–HBA
Ring B–HBA
6.74
3.47
9.84
6.76
2.76
8.62
7.53
2.75
10.04
7.53
2.75
10.04
Angles (◦) at each pharmacophore pointa
Ring A
Ring B
HBA
146.9
11.1
22.0
124.1
15.4
40.5
151.7
7.5
20.9
151.7
7.5
20.9
The dimensions reported by Matsson et al. (2007) are included for comparison.
a
Measured from energy minimized structures drawn on MOE® (CCG, Montreal, Canada). The centroids of rings A and B were used for mea-
surements involving the rings.
b
Rings A, B and hydrogen bond acceptor (in brackets) in aurone 3 and chalcones 20 and 21. The 6-OCH3 and 4-OH groups in aurone 3 and the
chalcones gave the best fit distances to the model proposed by Matsson et al. (2007)
.
at different times and using assays that were not completely
It is seen that the distance between the hydrophobic rings
(A, B) in aurone 3 complied with the optimal distance in the
pharmacophore model (6.76 Å vs. 6.74 Å), but the other two
distances (ring B-HBA, ring A-HBA) deviated from the pro-
posed values. The distances in the two chalcones (20, 21)
were almost identical and quite different from those found in
aurone 3. In particular, the distances between the hydropho-
bic rings in the chalcones (7.53 Å) were greater than the same
distance in the aurone (6.76 Å) and the pharmacophore model
(6.74 Å). On the other hand, the distances between ring B and
similar, it is reasonable to rank their ABCG2 inhibitory prop-
erties in the order of aurone 3 > chalcone 20 > chalcone 21.
The hydrophobic rings and H bond acceptor groups in each
compound are indicated in Table 6, alongside the relevant
distances and angles. Since more than one H bond accep-
tor group was present in each compound, the group that
complied best with the reported distances was chosen. A
depiction of the pharmacophore points in aurone 3 is given in
Fig. 5.
Fig. 5 – Depiction of hydrophobic rings (A, B) and H bond acceptor (6-OCH3) group in aurone 3. Distances are in Å units and
measured by MOE.

304
european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
2008), dipyridamole (Y. Zhang et al., 2005), the thiosemicar-
bazone NSC73306 (Wu et al., 2007) were reported to have dual
substrate and inhibitor profiles. The possibility of aurones
functioning as dual ABCG2 substrates and inhibitors warrants
further investigation as it would have important implications
on their modulatory properties. One approach would be to
investigate how these compounds affected ABCG2-mediated
ATP hydrolysis. The ability of ABCG2 and ABCB1 to extrude
agents out of cells is driven by ATPase activity and ATP pro-
duction. Substrates would be expected to increase ATPase
activity and enhance ATP production while inhibitors would
have the opposite effect. The reality however may be less
straightforward. For example, pheophorbide A, a substrate of
ABCG2 had a biphasic effect on ABCG2 ATPase activity, stim-
ulating ATP hydrolysis at lower concentrations but inhibiting
ATP hydrolysis at higher concentrations (Wu et al., 2007). Cur-
cumin stimulated ABCG2 ATPase activity even though it was
cited as an inhibitor of ABCG2 (Chearwae et al., 2006).
ABC transport modulators can either bind and compete
with substrate binding at the substrate binding site(s) or
interfere with ATP binding or hydrolysis by binding to the
ATP-binding site. Photoaffinity labeling of ABCG2 with differ-
ent photoaffinity analogues would provide a useful means of
distinguishing between these interactions. [125I] Iodoarylazi-
doparzosin, a photoaffinity analogue of prazosin, has been
used to characterize the substrate binding sites of ABCG2
(Shukla et al., 2006). If aurones are substrates, they would
inhibit the photolabeling of ABCG2 with this reagent. On the
other hand, the photoaffinity ATP analogue [␣-32P] 8-azidoATP
binds specifically to the nucleotide binding domains of numer-
ous ABC transporters and it may be used, alongside the ATPase
assay, to test if aurones interacted with the ATP-binding site
of ABCG2 (Wu et al., 2007).
A meaningful QSAR could not be derived from the avail-
able data but qualitative deductions on structural trends
were evident. The most important observation related to the
different structural requirements for inhibition of the two
transport proteins, even for a relatively simple template like
the aurones. The benzylidene ring B was less critical for ABCG2
inhibition. An unsubstituted ring B had as much inhibitory
effect as one that was substituted with groups of different
lipophilicities and electron donating/withdrawing properties.
This was not the case for ABCB1 inhibition where substitu-
tion of ring B was important, with greater inhibition associated
with substitution at the 3ꢀ position. The nature of the group
(lipophilic or hydrophilic, electron donating or withdrawing)
at 3ꢀ was however less critical. These different requirements
would imply that aurones were either good inhibitors of
ABCG2 or ABCB1, but not both. This was indeed noted for sev-
eral aurones but would require further validation with a larger
number of compounds.
The 4,6-dimethoxyaurone template was also compared to
a reported pharmacophore model for ABCG2 inhibitors. Of
the three distances in the reported model, only the distance
between the hydrophobic rings (A and B) of the aurone showed
a good fit. On the other hand, a chalcone with no ABCG2
inhibitory activity also showed compliance but to a different
distance in the model. The inference was that some distances
in the model may be more critical than others and based on the
three test compounds and their relative potencies as ABCG2
the H bond acceptor in the chalcones (10.04 Å) were closer
to the distance (9.84 Å) in the model. Similarly, the angles in
the pharmacophore model and the test compounds showed
varying degrees of compliance. Neither the aurones nor the
chalcones showed a convincing fit to the reported pharma-
cophore model. Comparison of distances and angles showed
that the two chalcones were identical in spite of their differing
activities as ABCG2 inhibitors.
4.
Discussion
Several 4,6-dimethoxyaurones increased the accumulation
of the MX in MDA-MB-231/R cells which were transfected
with ABCG2. These cells did not show detectable levels of
other transporters (ABCB1, ABCC1, ABCC2) in immunoblotting
experiments. The aurones did not increase MX accumula-
tion in MDA-MD-231/V cells that were not transfected with
ABCG2. These findings were consistent with the inhibition
of the ABCG2 efflux transporter. The aurones increased MX
accumulation at 0.5 ␮M and they were more potent than FTC
which brought about the same level of accumulation but at a
higher concentration (10 ␮M). Several aurones were found to
re-sensitize MDA-MB-231/R cells to MX, with some members
(3, 9) reducing the difference in sensitivities between trans-
fected (R) and parental (V) cell lines to a mere twofold at 0.5 ␮M.
The aurones were again more effective than FTC (0.5 ␮M) in
restoring MX sensitivity to MDA-MB-231/R cells.
The good ABCG2 inhibitory activity of the aurones was
complemented by their low antiproliferative activity on cul-
tured cell lines, as shown from the high levels of survival of
cancerous (MBA-MB-231/R, MBA-MB-231/V, HCT 116) and non-
cancerous (CCL 186) cell lines exposed to fixed concentrations
(5 ␮M or 10 ␮M) of most aurones. Although growth inhibitory
IC50 values were not determined in this investigation, values
exceeding 10 ␮M can be expected. The weak antiprolifera-
tive activities of the aurones were in sharp contrast to those
reported by Lawrence et al. (2003) which had submicromo-
lar IC50 values against a human myelogenous leukemia cell
line (K562). Those aurones were heavily methoxylated on both
rings A and B and the most promising compound had four
methoxy groups, with three on ring A. In comparison, the
present series of aurones had only two methoxy groups on
ring A. It may be that antiproliferative activity is influenced
by the number of methoxy groups attached to the benzofu-
ranone ring of the aurone. Unfortunately, the aurones were
not more cytotoxic on the MDA-MB-231/R cells that overex-
pressed ABCG2 than the MDA-MB-231/V cells. This would have
been a very desirable feature because the aurones could then
overcome drug resistance by eliminating the ABCG2 overex-
pressing cells. A thiosemicarbazone NSC73306 was reported
to selectively kill Pgp expressing multidrug resistant can-
cer cells besides inhibiting ABCG2 efflux activity (Wu et al.,
2007).
It was observed that some aurones (2, 4, 6, 8) were more
cytotoxic against the parental MDA-MB-231/V cells than the
transfected R cells, raising the possibility that these com-
pounds may be ABCG2 substrates. These compounds would
then be actively extruded by ABCG2 in the transfected R
cells, resulting in lower cytotoxicity. Imatinib (Lemos et al.,

european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
305
multidrug resistance by flavonoid derivatives and analogues.
J. Med. Chem. 46, 2125–2131.
Han, Y., Riwanto, M., Go, M.L., Ee, P.L.R., 2008. Modulation of
inhibitors, this may be the distance between the hydrophobic
rings.
In summary, this study has demonstrated the potential of
breast cancer resistance protein (BCRP/ABCG2) by non-basic
chalcone analogues. Eur. J. Pharm. Sci. 35, 30–41.
Lawrence, N.J., Rennison, D., Mcgown, A.T., Hadfield, J.A., 2003.
The total synthesis of an aurone isolated from Uvaria
hamiltonii: aurones and flavones as anticancer agents. Bioorg.
Med. Chem. Lett. 13, 3759–3763.
Lemos, C., Jansen, G., Peters, G.J., 2008. Drug transporters: recent
advances concerning BCRP and tyrosine kinase inhibitors. Br.
J. Cancer 98, 857–862.
Liu, X.L., Tee, H.W., Go, M.L., 2008. Functionalized
chalcones as selective inhibitors of P-glycoprotein
and breast cancer resistance protein. Bioorg. Med. Chem. 16,
171–180.
the 4,6-dimethoxy aurones as inhibitors of ABCG2. Aurones
were found to re-sensitize ABCG2 expressing cancer cells
to MX at submicromolar concentrations and to exhibit low
antiproliferative activity against several cell lines. Different
structural requirements were observed for aurones that were
good inhibitors of ABCG2 or ABCB1, implying that appropriate
changes in structure may give rise to aurones with selectivity
for one protein. If subsequent investigations provide further
support for the latter, aurones may have considerable utility
in advancing the understanding of ABCG2 function.
Matsson, P., Englund, G., Ahlin, G., Bergström, C., Norinder, U.,
Artursson, P., 2007. A global drug inhibition pattern for the
human ATP-binding cassette transporter breast cancer
resistance protein (ABCG2). J. Pharmacol. Exp. Ther. 323,
19–30.
Minderman, H., Suvannasankha, A., O’Loughlin, K., Scheffer, G.,
Scheper, R., Robey, R., Baer, M., 2002. Flow cytometric analysis
of breast cancer resistance protein expression and function.
Cytometry 48, 59–65.
Acknowledgements
We wish to thank Dr. Douglas D. Ross (Greenebaum Cancer
Center, University of Maryland, Baltimore, USA) for the gener-
ous gift of MDA-MB-231 (R and V) cells, and Dr. Anton Berns
(Netherlands Cancer Institute, Antoni van Leeuwenhoek Hos-
pital, Amsterdam, Netherlands) for his kind provision of MDCK
II/MDR1 cells. This work was made possible by funding RP
148 000 084 112 to MLG, RP 148 050 078 101/133 to PLRE,
and National University of Singapore graduate scholarships
to HMS, CYL.
Okombi, S., Rival, D., Bonnet, S., Mariotte, A., Perrier, E.,
Boumendjel, A., 2006. Discovery of
benzylidenebenzofuran-3(2H)-one (aurones) as inhibitors of
tyrosinase derived from human melanocytes. J. Med. Chem.
49, 329–333.
Ono, M., Maya, Y., Haratake, M., Ito, K., Mori, H., Nakayama, M.,
2007. Aurones serve as probes of beta-amyloid plaques in
Alzheimer’s disease. Biochem. Biophys. Res. Commun. 361,
116–121.
Pare, P.W., Dmitrieve, N., Mabry, T.J., 1991. Phytoalexin aurone
induced in cephalocereus senilis liquid suspension culture.
Phytochemistry 30, 1133–1135.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.ejps.2008.07.008.
Pelter, A., Ward, R.S., Heller, H.G., 1979. Carbon-13 nuclear
magnetic resonance spectra of (Z)- and (E)-aurones. J. Chem.
Soc., Perkin Trans. 1, 328–329.
r e f e r e n c e s
Robert, J., Jarry, C., 2003. Multidrug resistance reversal agents. J.
Med. Chem. 46, 4805–4817.
Schoepfer, J., Fretz, H., Chaudhuri, B., Muller, I., Seeber, E., Meijer,
I., Lozach, O., Vangrevelinghe, E., Furet, P., 2002.
Structure-based design and synthesis of
Beney, C., Mariotte, A.-M., Boumendjel, A., 2001. An efficient
synthesis of 4,6 dimethoxyaurones. Heterocycles 55, 967–972.
Boumendjel, A., 2003. Aurones: a subclass of flavones with
promising biological potential. Curr. Med. Chem. 10,
2621–2630.
Boumendjel, A., Beney, C., Deka, N., Mariotte, A., Lawson, M.,
Trompier, D., Baubichon-Cortay, H., Di Pietro, A., 2002a.
4-Hydroxy-6-methoxyaurones with high-affinity binding to
cytosolic domain of P-glycoprotein. Chem. Pharm. Bull. 50,
854–856.
Boumendjel, A., Di Pietro, A., Dumontet, C., Barron, D., 2002b.
Recent advances in the discovery of flavonoids and analogs
with high-affinity binding to P-glycoprotein responsible for
cancer cell multidrug resistance. Med. Res. Rev. 22, 512–529.
Breedveld, P., Beijnen, J., Schellens, J., 2006. Use of P-glycoprotein
and BCRP inhibitors to improve oral bioavailability and CNS
penetration of anticancer drugs. Trends Pharmacol. Sci. 27,
17–24.
Chearwae, W., Shukla, S., Limtrakul, P., Ambudkar, S.V., 2006.
Modulation of the function of the multidrug resistance-linked
ATP-binding cassette transporter ABCG2 by the cancer
chemopreventive agent curcumin. Mol. Cancer Ther. 5,
1995–2006.
2-benzylidene-benzofuran-3-ones as flavopiridol mimics. J.
Med. Chem. 45, 1741–1747.
Shukla, S., Robey, R.W., Bates, S.E., Ambudkar, S.V., 2006. The
calcium channel blockers, 1,4-dihydropyridines, are
substrates of the multidrug resistance-linked ABC drug
transporter, ABCG2. Biochemistry 45, 8940–8951.
Shukla, S., Wu, C., Ambudkar, S., 2008. Development of inhibitors
of ATP-binding cassette drug transporters: present status and
challenges. Expert Opin. Drug Metab. Toxicol. 4, 205–223.
Szakács, G., Paterson, J., Ludwig, J., Booth-Genthe, C., Gottesman,
M., 2006. Targeting multidrug resistance in cancer. Nat. Rev.
Drug Discov. 5, 219–234.
Thakkar, K., Cushman, M., 1995. A novel oxidative cyclization of
2ꢀ-hydroxychalcones to 4,5-dialkoxyaurones by thallium(III)
nitrate. J. Org. Chem. 60, 6499–6510.
Ur-Rahman, A., Choudhary, M.I., Hayat, S., Kahn, A.M., Ahmed,
A., 2001. Two new aurones from marine brown alga
spatoglossum variabile. Chem. Pharm. Bull. 49, 105–107.
Václavíková, R., Boumendjel, A., Ehrlichová, M., Kovár, J., Gut, I.,
2006. Modulation of paclitaxel transport by flavonoid
derivatives in human breast cancer cells. Is there a correlation
between binding affinity to NBD of P-gp and modulation of
transport? Bioorg. Med. Chem. 14, 4519–4525.
Craig, N.P., 1971. Interdependence between physical parameters
and selection of substituent groups for correlation studies. J.
Med. Chem. 14, 680–684.
Hadjeri, M., Barbier, M., Ronot, X., Mariotte, A., Boumendjel, A.,
Boutonnat, J., 2003. Modulation of P-glycoprotein-mediated

306
european journal of pharmaceutical sciences 3 5 ( 2 0 0 8 ) 293–306
Wu, C., Shukla, S., Calcagno, A., Hall, M., Gottesman, M.,
Ambudkar, S., 2007. Evidence for dual mode of action of a
thiosemicarbazone, NSC 73306: a potent substrate of the
multidrug resistance linked ABCG2 transporter. Mol. Cancer
Ther. 6, 3287–3296.
relationships for the flavonoid-mediated inhibition of
breast cancer resistance protein. Biochem. Pharmacol. 70,
627–639.
Zhang, Y., Gupta, A., Wang, H., Zhou, L., Vethanayagam, R.R.,
Unadkat, J.D., Mao, Q., 2005. BCRP transports dipyridamole
and is inhibited by calcium channel blockers. Pharm. Res. 22,
2023–2034.
Zhang, S., Yang, X., Coburn, R.A., Morris, M.E., 2005. Structure
activity relationships and quantitative structure activity