Antitumor agents 247. New 4-ethoxycarbonylethyl curcumin analogs as potential antiandrogenic agents

Article abstract of DOI:10.1016/j.bmc.2005.11.034

4-Ethoxycarbonylethyl curcumin (ECECur) (3) is a current drug candidate for the treatment of prostate cancer. Due to problems inherent in the tautomerism of ECECur, 4-fluoro-4-ethoxycarbonylethyl curcumin (4) and 4- ethoxycarbonylethylenyl curcumin (5) were designed and synthesized. These two target compounds and their synthetic intermediates (4-9) were evaluated for their inhibitory activity against androgen receptor transcription in LNCaP and PC-3 prostate cancer cell lines. While the enol-keto analogs showed varying anti-androgen potencies, the di-keto analogs showed no activity. Tetrahydropyranylation of the phenoxy groups had a positive impact on the anti-AR activity of 4-ethoxycarbonylethylenyl curcumin, but a negative impact on the activity of ECECur. With potent anti-AR activity, di-tetrapyranylated 4-ethoxycarbonylethylenyl curcumin (9), which exists in only one form, is a good drug lead for further structural modification. Based on the SAR information obtained from the above study, five new compounds were designed and subsequently synthesized. Among them, compound 10 was found to be the most potent anti-AR agent and is considered to be a promising drug candidate for the treatment of prostate cancer.

Full text of DOI:10.1016/j.bmc.2005.11.034

Bioorganic & Medicinal Chemistry 14 (2006) 2527–2534
Antitumor agents 247. New 4-ethoxycarbonylethyl
curcumin analogs as potential antiandrogenic agentsq
Li Lin,a Qian Shi,b Ching-Yuan Su,b Charles C.-Y. Shihb and Kuo-Hsiung Leea,*
aNatural Products Laboratory, School of Pharmacy, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7360, USA
bAndroScience Corporation, 11175 Flintkote Ave., Suite F, San Diego, CA 92121, USA
Received 19 October 2005; revised 15 November 2005; accepted 15 November 2005
Available online 19 January 2006
Abstract—4-Ethoxycarbonylethyl curcumin (ECECur) (3) is a current drug candidate for the treatment of prostate cancer. Due
to problems inherent in the tautomerism of ECECur, 4-fluoro-4-ethoxycarbonylethyl curcumin (4) and 4-ethoxycarbonylethyl-
enyl curcumin (5) were designed and synthesized. These two target compounds and their synthetic intermediates (4–9) were
evaluated for their inhibitory activity against androgen receptor transcription in LNCaP and PC-3 prostate cancer cell lines.
While the enol–keto analogs showed varying anti-androgen potencies, the di-keto analogs showed no activity. Tetrahydropyr-
anylation of the phenoxy groups had a positive impact on the anti-AR activity of 4-ethoxycarbonylethylenyl curcumin, but a
negative impact on the activity of ECECur. With potent anti-AR activity, di-tetrapyranylated 4-ethoxycarbonylethylenyl
curcumin (9), which exists in only one form, is a good drug lead for further structural modification. Based on the SAR
information obtained from the above study, five new compounds were designed and subsequently synthesized. Among them,
compound 10 was found to be the most potent anti-AR agent and is considered to be a promising drug candidate for the
treatment of prostate cancer.
Ó 2006 Published by Elsevier Ltd.
1. Introduction
an increase in prostate-specific antigen (PSA), while
taking the nonsteroidal anti-androgens, have a decrease
in the level of PSA after withdrawal of the drugs.5 Due
to these drawbacks with the current anti-prostate cancer
drugs, we aim to develop new curcumin analogs as
potential anti-androgens for treatment of prostate cancer.
Prostate cancer is the most prevalent cancer among males
in Western countries.2 Androgen and the androgen
receptor (AR), a member of the superfamily of nuclear
receptors, have been documented to play important roles
in the growth of prostate cancer cells. The present treat-
ment for prostate cancer is surgery combined with
chemo-therapeutic agents.3 The prevailing drugs are
anti-androgens including steroidal agents and non-ste-
roidal agents. However, the steroidal agents have
drawbacks such as agonistic activity and overlapping
effects with other hormonal systems, as well as causing
some side effects.4 These drawbacks hinder their use as
ideal anti-prostate cancer drugs. The non-steroidal agents
such as hydroxyflutamide and bicalutamide were thought
to have no agonistic activity. However, after a long period
of administration, they may result in ‘‘anti-androgen
withdrawal syndrome’’ in which patients who experience
Curcumin (1) [diferuloyl methane; 1,7-bis-(4-hydroxy-3-
methoxyphenyl)-1,6-heptadiene-3,5-dione] (Fig. 1) is the
major constituent in the rhizome of Curcuma longa
(Zingiberaceae), commonly named turmeric. Over a
long period of study, curcumin has been found to
possess various biological activities, such as anti-
inflammatory,6 antioxidant,7 anti-HIV-1 integrase,8 che-
mo-preventive,9 anti-angiogenic,10 and anti-cancer.11
Therefore, various groups worldwide have used curcu-
min as a lead compound to develop numerous analogs
with different bioactivities. In our laboratory, we are
interested in the design and synthesis of new curcumin
analogs possessing anti-androgen receptor activity.12,13
In a previous paper, we reported that two curcumin ana-
logs, dimethyl curcumin (2) and 4-ethoxycarbonylethyl
curcumin (ECECur) (3) (Fig. 1), showed greater potency
in anti-AR assays than hydroxyflutamide, an anti-
androgen, currently used clinically.12 ECECur has also
Keywords: 4-Ethoxycarbonylethyl curcumin; Antiandrogenic agents;
Prostate cancer.
q
Antitumor agents 247. For paper 246, see Ref. 1.
Corresponding author. Tel.: +1 919 962 0066; fax: +1 919 966
3893; e-mail: khlee@unc.edu
*
0968-0896/$ - see front matter Ó 2006 Published by Elsevier Ltd.
doi:10.1016/j.bmc.2005.11.034

2528
L. Lin et al. / Bioorg. Med. Chem. 14 (2006) 2527–2534
OH
O
OH
O
H3CO
HO
OCH3
OH
H3CO
H3CO
OCH3
OCH3
4
4'
4'
1 Curcumin
2 Dimethyl curcumin
O
O
OH
O
H3CO
HO
OCH3
OH
H3CO
HO
OCH3
OH
O
OEt
O
OEt
3 4-Ethoxycarbonylethyl curcumin (ECECur)
Figure 1. Structures of compounds 1–3.
exhibited interesting biological activities other than po-
tent anti-AR activity (unpublished data). It is regarded
as a promising drug candidate for the treatment of pros-
tate cancer, as well as certain other diseases. However,
the tautomerism of ECECur, which causes it to exist
in both enol–keto and di-keto forms, may hinder its
potential as a clinically useful drug. In this paper, we
present a resolution to the tautomerism of ECECur. In
addition, we investigated which form of ECECur may
be responsible for its anti-AR activity. This SAR study
was used to guide the further design of five new analogs,
among which compound 10 exhibited great potential as
a drug candidate to treat prostate cancer. Extensive
SAR correlations were established based on this study.
salt (SelectFluorTM) in the presence of sodium hydride
in DMF.16 Deprotection of compound 7 with PPTS in
ethanol afforded the target compound 4.
The preparation of analog 5 (Scheme 2) started from
curcumin 1, which was obtained by recrystallization of
commercially available curcumin (Aldrich, Inc.). After
the two phenoxy groups of 1 were protected as the
THP ethers, the resulting compound 8 was alkylated
at the C-4 position using sodium hydride in anhydrous
tetrahydrofuran, followed by the addition of ethyl
propiolate to afford monosubstituted compound 9.
The target compound 5 was obtained by cleavage of
the THP ethers as described above for 4.
Compounds 10 and 11 were derived directly from
dimethyl curcumin (2), which was prepared from
commercially available 3,4-dimethoxybenzaldehyde as
described above for compound 3. Dimethyl curcumin
reacted in a Michael addition with ethyl propiolate to
give compound 10 and with methyl propiolate to give
compound 11 (Scheme 2).
2. Chemistry
Compound 4, the C-4 fluorinated analog of 3, was mod-
ified from ECECur as shown in Scheme 1. Commercially
available vanillin was condensed with ethyl 4-acetyl-5-
oxo-hexanoate using the method of Pedersen et al.14 to
obtain compound 3. Boric anhydride was used to form
a boron complex with ethyl 4-acetyl-5-oxo-hexanoate
in order to prevent Knoevenagel condensation.14 Com-
pound 3 was then protected as its tetrahydropyranyl
(THP) ether by using DHP in the presence of PPTS in
dry dichloromethane.15 The THP ether 6 was fluorinat-
ed with 1-alkyl-4-fluoro-1, 4-diazabicyclo[2,2,2]octane
As shown in Scheme 3, compound 12 was obtained by
reacting dimethyl curcumin with N-ethylpropiolamide,
which was prepared from ethylamine by the slow addi-
tion of methyl propiolate at ꢀ30 °C.17 Compound 13
was obtained by the reaction of dimethyl curcumin with
propiolic acid amide under reflux. Propiolic acid amide
O
O
O
O
O
O
B2O3,
CHO
H3CO
OCH3
OH
H3CO
THPO
OCH3
OTHP
H3CO
HO
DHP, PPTS
CH2Cl2
(CH ) COOEt
2
2
(nBuO)3B
nBuNH2, HCl
HO
vanillin
O
OEt
O
OEt
6
3
O
O
F
O
O
NaH, DMF;
SelectFluor
H3CO
THPO
OCH3
OTHP
H3CO
HO
OCH3
OH
PPTS,
EtOH
F
O
OEt
O
OEt
7
4
Scheme 1. Synthesis of compound 4.

L. Lin et al. / Bioorg. Med. Chem. 14 (2006) 2527–2534
2529
Scheme 2. Synthesis of compounds 5, 8–11, and 14.
OH
O
OH
O
H3CO
H3CO
OCH3
OCH3
H3CO
H3CO
OCH3
OCH3
NaH, THF
O
R'
NHR
2
12, R' = CONHEt
13, R' = CN
R = H or Et
RNH2
O
OMe
Scheme 3. Synthesis of compounds 12 and 13.
was prepared from concentrated aqueous ammonia and
methyl propiolate at ꢀ30 °C.18
line and in the androgen-independent PC-3 cell line
transfected with wild-type AR. Cytotoxicity was also
examined in the LNCaP cell line. The bioassay results
are shown in Figure 2 and generally were consistent over
the three models. The parent compound ECECur (3)
showed the highest activity, followed by compound 9.
Compounds 5 and 6 showed weak activity, and di-keto
compounds 4, 7, and 8 were inactive. Compounds 5 and
9 exist exclusively in the enol–keto form; thus, the
enol–keto form is likely the active form for anti-AR
activity. Because fluorination of ECECur at the C-4
position (4) abolished activity in all bioassay models,
we propose that the di-keto form may not contribute
to the anti-AR activity of ECECur. In a previous paper,
we also reported that a di-keto curcumin analog with di-
substitution at the C4 position did not exhibit anti-AR
activity in either human prostate cancer cell line.13 The
conformational difference between the di-keto and
enol–keto forms of ECECur may help to explain the
different anti-AR activity of these two analog types
(Fig. 3). In the enol–keto conformation, the two phenyl
rings and the unsaturated linker are in the same plane,
because of the high degree of conjugation as well as
the formation of a strong hydrogen bond between the
enol proton and the carbonyl oxygen. However, in the
di-keto conformation, ECECur does not stabilize in a
planar structure, because the two electron-rich carbonyl
oxygen atoms repel each other. Compound 6, the THP
ether of ECECur (3), showed decreased activity, while
compound 9, the equivalent analog of 5, showed greater
anti-AR activity. Thus, the THP-protecting group has a
dissimilar influence on the anti-AR activity of different
compounds. With ethoxycarbonylethyl substitution at
C-4, tetrahydropyranylation of both phenols had a
negative impact on the anti-AR activity. However, with
Compound 10 was reduced to the allylic alcohol 14 with
Dibal-H at ꢀ78 °C (Scheme 2).
3. Results and discussion
To circumvent the tautomerism of ECECur and deter-
mine the needed structural features for anti-androgen
receptor activity, we designed two target compounds 4
and 5. Ideally, target compounds should be stable in
only one form as well as have minimal structural change
from the parent compound ECECur in order to most
likely retain similar potency.
For the di-keto ECECur analog, we designed compound
4, in which the C-4 hydrogen has been replaced with a
fluorine atom. Hydrogen and fluorine have similar atom
sizes, and therefore, replacing hydrogen with fluorine
does not significantly alter the overall molecular size.
We proposed compound 5 as the enol–keto ECECur
analog. The only structural difference between 5 and
its parent compound 3 occurs in the C-4 side chain.
The ethoxycarbonylethyl side chain in ECECur contains
a saturated single (C–C) bond, while the ethoxycarbony-
lethylenyl side chain of 5 contains an unsaturated dou-
ble (C@C) bond. Compound 5 exists exclusively in an
enol–keto form due to the highly conjugated polyene.
The target ECECur analogs and their synthetic interme-
diates were tested for inhibitory activity against AR
transcription in the androgen-dependent LNCaP cell

2530
L. Lin et al. / Bioorg. Med. Chem. 14 (2006) 2527–2534
an ethoxycarbonylethylenyl group at C-4, it had a posi-
anti-AR activity in LNCaP
tive influence on the anti-AR activity. Therefore, both
the C-4 position and the C-40 moieties on the phenyl
rings of curcumin analogs are important pharmaco-
phores with respect to the anti-AR activity. In order
to obtain an optimal anti-AR agent, we next focused
on modification of the substituents at these positions.
200
150
100
50
Based on the preceding SAR information, we designed
and synthesized compounds 10–14, which have C-40
methoxy groups on both phenyl rings and various ethyl-
enyl side chains at C-4. As shown in Figure 4, among the
five new analogs, compound 10 exhibited the highest
potency in all three bioassay models. Compound 13
showed moderate anti-AR activity in LNCaP cells, weak
anti-AR activity in PC-3 cells transfected with wild-type
androgen receptor, and weak inhibitory activity against
the growth of LNCaP cells. Compounds 11, 12, and 14
showed either weak or no activity in the prostate cancer
cells. Compared with ECECur (3), compound 10 showed
similar anti-AR potency in the LNCaP cell line, was
slightly less potent in PC-3 cells transfected with wild-
type androgen receptor, and was more cytotoxic in the
LNCaP cell growth assay (Fig. 3). Since compound 10
showed significant activity and does not undergo or be
limited by tautomerism, it is regarded as a very promising
drug candidate for in vivo investigation. Our SAR con-
clusions were extended on the basis of the structural fea-
tures and bioactivity of these five new analogs. Although
the structures of these analogs vary only in the side chain
at C-4, the replacement of ethoxy (compound 10) with
methoxy (compound 11) or N-ethylamino (compound
12) resulted in loss of activity, which implies that a long
chain ester may be more favorable. The reduction of ester
to alcohol also led to loss of activity, emphasizing the
necessity for an ester in the side chain. From comparing
compounds 10 and 13, the nitrile functional group does
not enhance the anti-AR activity or the cytotoxicity in
LNCaP cells. Therefore, in our future study, we will
focus on optimization of the ester length.
0
DHT
EtOH
DHT+ D2HT+ D3HT+ D4HT+ D5HT+ D6HT+ D7HT+ D8HT+ 9
DHT+HF
DHT (1 nM) + tested compound (3 uM)
anti-AR activity in PC-3 + wt. AR
120
100
80
60
40
20
0
DHT (1 nM) + tested compound (3 uM)
LNCaP cell growth
120
100
80
60
40
20
0
DHT
EtOH
DHT+ D2HT+ 3DHT+ 4DHT+ D5HT+ 6DHT+ D7HT+ 8DHT+ 9
4. Conclusion
DHT (1 nM) + tested compound (3 uM)
In summary, we eliminated the tautomerism problem
of 4-ethoxycarbonylethyl curcumin 3 by synthesis of
4-fluoro-4-ethoxycarbonylethyl curcumin 4 and 4-eth-
oxycarbonylethylenyl curcumin 5. Evaluation of the
anti-androgen activity of these ECECur analogs (3–14)
in LNCaP cells and PC-3 cells indicated that di-keto
Figure 2. Anti-AR activity and cytotoxicity of compounds 2–9 (3 lM)
in prostate cancer cells.
H
O
O
O
O
OCH3
OH
H3CO
HO
H3CO
HO
OEt
.
H
.
O
O
OEt
HO
OCH3
di-keto conformation
enol-keto conformation
Figure 3. The conformations of ECECur.

L. Lin et al. / Bioorg. Med. Chem. 14 (2006) 2527–2534
2531
5. Experimental
Anti-AR activity in LNCaP
Melting points were determined with a Fisher–Johns
melting apparatus and are uncorrected. 1H NMR
spectra were recorded on a Varian Gemini-300 spec-
trometer. The chemical shifts are presented in terms
of parts per million with TMS as the internal reference.
MS spectra were recorded on an API-3000 LC/MS/MS
spectrometer. Column chromatography was carried out
on CombiFlashÒ CompanionTM (Isco, Inc.), and thin-
layer chromatography was performed on pre-coated
silica gel or aluminum plates (Aldrich, Inc.). Elemental
analyses were performed by Atlantic Microlab Inc.,
Norcross, GA, and agreed with theoretical values to
within 0.4%.
120
100
80
60
40
20
0
DHT (1 nM) + tested compound (5 uM)
Anti-AR activity in PC-3 +wt. AR
5.1. Ethoxycarbonylethyl curcumin (3)
120
100
80
60
40
20
0
The synthetic procedure was modified from the meth-
od first introduced by Pedersen et al.14 Ethyl 4-acetyl-
5-oxo-hexanoate (1.87 mL, 10 mmol) and boric anhy-
dride (0.5 g, 7 mmol) were dissolved in 10 mL EtOAc.
The solution was stirred for 30 min at 40 °C. Vanillin
(3.1 g, 20 mmol) and tributyl borate (5.4 mL,
20 mmol) were added, and the mixture was stirred
for 30 min. Butylamine (1.49 mL, 15 mmol) dissolved
in 10 mL EtOAc was added dropwise during 15 min.
The stirring was continued for 24 h at 40 °C. The mix-
ture was then hydrolyzed by adding 30 mL of 4 N
hydrochloric acid and heating for 30 min at 60 °C.
The organic layer was separated, and the aqueous
layer was extracted three times with EtOAc. The com-
bined organic layers were washed until neutral and
dried over anhydrous sodium sulfate. The solvent
was removed in vacuo. The crude product was puri-
fied by CombiFlashÒ chromatography eluting with
hexane–EtOAc.
DHT (1 nM) + tested compound (5 uM)
LNCaP cell growth
150
100
50
0
Dimethyl curcumin (2) was prepared from 2,4-pentane-
dione and 3,4-dimethoxybenzaldehyde using the same
procedure.
DHT (1 nM) + tested compound (5 uM)
Figure 4. Anti-AR activity and cytotoxicity of compounds 2, 3, and
10–14 (5 lM) in prostate cancer cells.
The structures of 1–3 were confirmed by comparison of
their physical data with those reported in the
literature.12,14
ECECur analogs did not exhibit anti-AR activity, but
enol–keto ECECur analogs showed varying anti-AR
potencies. Our SAR study revealed that: (1) the enol–keto
moiety is responsible for the anti-AR activity, and the
di-keto form probably is not an active form; (2) tetrahy-
dropyranylation or methylation of the phenoxy groups
impacts the anti-AR potency; (3) the –CH@CHCOOEt
side chain is superior to –CH@CHCOOMe, –CH@
CHCONHEt, –CH@CHCN, and –CH@CHCH2OH.
Proper substitutionsat the C-4 and C-40 positions are vital
for anti-AR activity as well as cytotoxicity in LNCaP
cells. Compound 10, which bears the optimal features
identified to date for potent anti-AR activity and exists
exclusively in the enol–keto form, has been identified as
a promising curcumin analog and is undergoing contin-
uing evaluation as a potential clinical trial candidate for
the treatment of prostate cancer.
5.2. 7-[3-Methoxy-4-(tetrahydropyran-2-yloxy)-phenyl]-
4-{3-[3-methoxy-4-(tetrahydropyran-2-yloxy)-phenyl]-
acryloyl}-5-oxo-hept-6-enoic acid ethyl ester (6)
Compound 3 (153.1 mg, 0.33 mmol) and dihydropyran
(0.73 mL, 7.32 mmol) were dissolved in 3 mL of dry
dichloromethane containing PPTS (8.3 mg, 0.033 mmol).
The solution was stirred at rt for 48 h. The solution was
then washed with water, and the solvent was removed
in vacuo. The crude product was purified by Combi-
FlashÒ chromatography eluting with hexane–EtOAc
to give 6 (134 mg), 59% yield, yellow powder, mp
60–61 °C (acetone); ESI MS m/z 635.2 (MꢀH)+; 1H
NMR (300 MHz, CDCl3): d 1.25 (3H, t), 1.57–2.17 (12H,
m), 2.96 (0.57H, t), 3.62 (4H, t), 3.91 (6H, s), 4.13 (2H,
q), 5.47 (2H, t), 6.72 (2H, d, J = 15.6 Hz), 6.90–7.18 (6H,
m), 7.44 (2H, d, J = 15.6 Hz); Anal. (C36H44O10) C, H.

2532
L. Lin et al. / Bioorg. Med. Chem. 14 (2006) 2527–2534
5.3. 4-Fluoro-7-[3-methoxy-4-(tetrahydropyran-2-yloxy)-
phenyl]-4-{3-[3-methoxy-4-(tetrahydropyran-2-yloxy)-
phenyl]-acryloyl}-5-oxo-hept-6-enoic acid ethyl ester (7)
0 °C. The solution was stirred at 0 °C for 30 min and
then at rt for 2 h. To the sodium salt solution, ethyl pro-
piolate (0.02 mL, 0.20 mmol) was added. After stirring
for 2 h, the solution was extracted with EtOAc and
washed with 5% H2SO4 (10 mL) and saturated NaHCO3
solution (10 mL). Then the solvent was evaporated in
vacuo. The crude product was purified by CombiFlashÒ
chromatography eluting with hexane–EtOAc to afford 9
(39.4 mg), 62% yield, orange powder; mp 72–73 °C; ESI
MS m/z 634.7 (M+H)+; 1H NMR (300 MHz, CDCl3): d
1.34 (3H, t), 1.5–2.2 (12H, m), 3.62 (4H, t), 3.92 (6H, s),
4.28 (2H, q), 5.49 (2H, t) 5.96 (H, d, J = 15.6 Hz),
7.00 (2H, d, J = 15.6 Hz), 7.08–7.16 (6H, m), 7.76 (2H,
d, J = 15.3 Hz), 7.83 (H, d, J = 15.9 Hz); Anal.
(C36H42O10) C, H.
A DMF solution (3 mL) of 6 (30 mg, 0.047 mmol) was
added to an oil-free suspension of NaH (10 mg of
60%, 6 mg, 0.25 mmol) in DMF (2 mL) under nitrogen
at 0 °C. The solution was stirred at 0 °C for 30 min
and then at rt for 2 h. To the sodium salt solution,
SelectFluorTM (70 mg, 0.2 mmol) in DMF (2 mL) was
added. After stirring for 1 h, the solution was extracted
with EtOAc and washed with 5% H2SO4 (10 mL)
and subsequently with saturated NaHCO3 solution
(10 mL). The solvent was evaporated in vacuo, and the
crude product was purified by CombiFlashÒ chromatog-
raphy eluting with hexane–EtOAc to afford 7 (4 mg),
13% yield, yellow powder, mp 59–60 °C; ESI MS m/z
5.7. 5-Hydroxy-7-(4-hydroxy-3-methoxyphenyl)-4-[3-(4-
hydroxy-3-methoxyphenyl)-acryloyl]-hepta-2,4,6-trienoic
acid ethyl ester (5)
1
677.3 (M+Na)+, 655.3 (M+H)+; H NMR (300 MHz,
CDCl3): d 1.25 (3H, t), 1.5–2.1 (12H, m), 2.45 (2H,
m), 2.65 (2H, m), 3.62 (4H, t), 3.90 (6H, s), 4.13 (2H,
q), 5.49 (2H, t), 7.06 (2H, dd, J = 3 Hz, J = 15.6 Hz),
7.12–7.15 (6H, m), 7.75 (2H, d, J = 15.6 Hz); Anal.
(C36H44FO10) C, H.
Compound 9 was converted to 5 using the same proce-
dure described above for 4 from 7. 93% yield, orange
powder, mp 106–106.5 °C; ESI MS m/z 465.2
(MꢀH)+; 1H NMR (300 MHz, CDCl3): d 1.34 (3H,
t), 3.95 (6H, s), 4.29 (2H, quart), 5.96 (2H, d,
J = 15.6 Hz), 6.95 (2H, d, J = 8.2 Hz), 6.96 (1H, d,
J = 15.6 Hz), 7.05 (2H, d, J = 2.1 Hz), 7.17 (2H, dd,
J = 8.2 Hz, J = 2.1 Hz), 7.75 (2H, d, J = 15.3 Hz),
7.90 (1H, d, J = 15.9 Hz); Anal. (C26H26O8Æ11/8H2O)
C, H.
5.4. 4-Fluoro-7-(4-hydroxy-3-methoxyphenyl)-4-[3-(3-
methoxy-4-methylphenyl)-acryloyl]-5-oxo-hept-6-enoic
acid ethyl ester (4)
The EtOH (3 mL) solution of compound 7 (12.5 mg,
0.019 mmol) and PPTS (5 mg, 0.02 mmol) was stirred
at rt for 3 h. The solution was evaporated in vacuo,
and compound 4 (9 mg) was obtained by CombiFlashÒ
chromatography eluting with hexane–EtOAc. 97% yield,
yellow powder, mp 63–63.5 °C; EIMS m/z 509.3
(M+Na)+, 487.3 (M+H)+; 1H NMR (300 MHz, CDCl3):
d 1.25 (3H, t), 2.45 (2H, m), 2.65 (2H, m), 3.95 (6H, s),
6.93 (2H, d, J = 7.8 Hz), 7.06 (2H, dd, J = 3 Hz,
J = 15.6 Hz), 7.09 (2H, d, J = 1.8 Hz), 7.16 (2H, dd,
J = 1.8 Hz, J = 7.8 Hz), 7.75 (2H, d, J = 15.6 Hz); Anal.
(C26H27FO8Æ3/4H2O) C, H.
5.8. 7-(3,4-Dimethoxyphenyl)-4-[3-(3,4-dimethoxy
phenyl)-acryloyl]-5-hydroxy-hepta-2,4,6-trienoic acid
ethyl ester (10)
Dimethylated curcumin (2) (360 mg, 1 mmol) was react-
ed with NaH (24 mg, 1 mmol) and ethyl propiolate
(0.2 mL, 1.97 mmol), as described above for preparation
of 9, to afford 10 (268 mg), 54% yield, red powder, mp
170–171 °C; ESI MS m/z 494.6 (M+H)+; 1H NMR
(300 MHz, CDCl3): d 1.34 (3H, t), 3.95 (12H, s), 4.29
(2H, quart), 5.98 (2H, d, J = 15.6 Hz), 6.95 (2H, d,
J = 8.4 Hz), 7.00 (1H, d, J = 15.6 Hz), 7.08 (2H, d,
J = 1.8 Hz), 7.22 (2H, dd, J = 8.4 Hz, J = 1.8 Hz), 7.77
(2H, d, J = 15.3 Hz), 7.91 (1H, d, J = 15.6 Hz). Anal.
(C28H30O8Æ1/4H2O) C, H.
5.5. 5-Hydroxy-1,7-bis-[3-methoxy-4-(tetrahydropyran-2-
yloxy)-phenyl]-hepta-1,4,6-trien-3-one (8)
Recrystallized curcumin (1, 1.08 g, 2.94 mmol) was pro-
tected as its tetrahydropyranyl ethers by reaction with
dihydropyran (2 mL, 20 mmol) as described above for
conversion of 3 to 6. CombiFlashÒ chromatography elut-
ing with hexane–EtOAc gave pure 8 (1.12 g), 66.8% yield,
yellow powder, mp 67–69 °C; ESI MS m/z 535.0 (MꢀH)+;
1H NMR (300 MHz, CDCl3): d 1.57–2.17 (12H, m),
3.62 (4H, t), 3.91 (6H, s), 5.47 (2H, t), 5.83 (1H, s), 6.50
(2H, d, J = 15.9 Hz), 7.09–7.16 (6H, m), 7.60 (2H, d,
J = 15.9 Hz); Anal. (C31H36O8Æ1/4H2O) C, H.
5.9. 7-(3,4-Dimethoxyphenyl)-4-[3-(3,4-dimethoxy
phenyl)-acryloyl]-5-hydroxy-hepta-2,4,6-trienoic acid
methyl ester (11)
Dimethylated curcumin (2) (200 mg, 0.5 mmol) was
reacted with NaH (20 mg of 60%, (4.2 mg, 0.5 mmol)
in THF (3 mL) then with methyl propiolate (0.08 mL,
1 mmol), as described above for preparation of 9, to af-
ford 11 (121 mg), 50% yield, orange powder; mp 167–
168 °C; ESI MS m/z 481.6 (M+H)+; 1H NMR
(300 MHz, CDCl3): d 3.82 (3H, s), 3.93 (12H, s), 5.98
(H, d, J = 15.6 Hz), 6.90 (2H, d, J = 8.4 Hz), 6.98
(2H, d, J = 15.6 Hz), 7.07 (2H, d, J = 1.8 Hz), 7.20
(2H, dd, J = 8.4 Hz, J = 1.8 Hz), 7.76 (2H, d,
J = 15.6 Hz), 7.90 (H, d, J = 15.9 Hz); Anal.
(C27H28O8Æ1/4H2O) C, H.
5.6. 5-Hydroxy-7-[3-methoxy-4-(tetrahydropyran-2-
yloxy)-phenyl]-4-{3-[3-methoxy-4-(tetrahydropyran-2-
yloxy)-phenyl]-acryloyl}-hepta-2,4,6-trienoic acid ethyl
ester (9)
A THF solution (3 mL) of 8 (55 mg, 0.10 mmol) was
added to an oil-free suspension of NaH (7 mg of 60%,
4.2 mg, 0.17 mmol) in THF (2 mL) under nitrogen at

L. Lin et al. / Bioorg. Med. Chem. 14 (2006) 2527–2534
2533
5.10. 7-(3,4-Dimethoxyphenyl)-4-[3-(3,4-dimethoxy
phenyl)-acryloyl]-5-hydroxy-hepta-2,4,6-trienoic acid
ethylamide (12)
J = 8.4 Hz), 6.97 (2H, d, J = 15.6 Hz), 7.06 (2H, d,
J = 1.8 Hz), 7.17 (2H, dd, J = 8.4 Hz, J = 1.8 Hz), 7.68
(2H, d, J = 15.6 Hz); Anal. (C26H28O7Æ3/4H2O) C, H.
Ethylamine (6 mL, 2.0 M in MeOH) and water (6 mL)
were cooled in dry ice-isopropyl alcohol. Methyl propi-
olate (840 mg, 10 mmol, 0.84 mL) was added gradually
to the stirring solution. After 10 min, the solvent was
evaporated in vacuo. Crude N-ethylpropionamide was
obtained. A THF solution (3 mL) of 2 (100 mg,
0.25 mmol) was added to an oil-free suspension of
NaH (10 mg of 60%, 2.4 mg, 0.25 mmol) in THF
(3 mL) under nitrogen at 0 °C. The solution was stirred
at 0 °C for 30 min and then at rt for 2 h. To the sodium
salt solution, crude N-ethylpropionamide (0.40 mL) was
added. After stirring for 22 h, the solution was extracted
with EtOAc and washed with 5% H2SO4 (10 mL) and
subsequently saturated NaHCO3 solution (10 mL).
Then the solvent was evaporated in vacuo. The crude
product was purified by CombiFlashÒ chromatography
eluting with hexane–EtOAc to afford 12 (20 mg), yellow
powder, mp 219–221 °C; ESI MS m/z 516.2 (M+Na)+;
1H NMR (300 MHz, CDCl3): d 1.32 (3H, t), 3.92 (6H,
s), 3.92 (6H, s), 4.00 (2H, quart), 5.85(H, d,
J = 15.3 Hz), 6.88 (2H, d, J = 8.4 Hz), 6.95 (2H, d,
J = 15.6 Hz), 7.06 (2H, d, J = 1.5 Hz), 7.18 (2H, dd,
J = 8.4 Hz, J = 1.5 Hz), 7.73 (2H, d, J = 15.6 Hz), 7.82
(H, d, J = 15.3 Hz); Anal. (C28H31NO7Æ3/2H2O) C, H.
5.13. Cell culture and gene transfection
Human prostate cancer LNCaP and PC-3 cells were
maintained in RPMI medium and Dulbeccoꢀs minimum
essential medium (DMEM), respectively. Both media
were strengthened with penicillin (25 units/mL), strepto-
mycin (25 lg/mL), and 10% fetal calf serum. Androgen
receptor transactivation assay, an androgen-dependent
reporter gene transcription test, was employed as the pri-
mary screening for potential anti-androgen identifica-
tion. This assay was first performed in LNCaP cells,
which express a clinically relevant mutant AR. Once
anti-androgenic activity was detected in the LNCaP
AR transactivation assay, compounds were re-examined
for their potential activity against wild-type AR. Wild-
type AR transactivation assay was performed in PC-3
host cells, which lack an endogenous, functional AR.
The method and conditions of cell and gene transfection
have been described previously.12,13 In brief, cells were
plated in 24-well tissue culture dishes 24 (PC-3 cells) or
48 (LNCaP cells) hours prior to transfection. Subse-
quently, LNCaP cells were transfected with a reporter
gene, MMTV-luciferase, which contains MMTV-LTR
promoter and androgen receptor binding element, and
PRL-SV40, which served as an internal control for trans-
fection efficiency. PC-3 cells were transfected with a wild-
type AR expression plasmid, pSG5AR, in addition to the
above-mentioned MMTV-luciferase reporter gene and
PRL-SV40 internal control. SuperFect (Qiagen, Chats-
worth, CA) was employed as the transfection reagent fol-
lowing manufacturerꢀs recommendations. At the end of a
5 h transfection, the medium was changed to DMEM or
RPMI supplemented with 10% charcoal dextran-
stripped, that is, androgen-depleted, serum. Twenty-four
hours later, the cells were treated with 1 nM DHT and/or
test compounds at the designated concentration for
another 24 h. The cells were harvested for luciferase
activity assay using Dual Luciferase Assay System (Pro-
mega, Madison, WI). The derived data were expressed as
relative luciferase activity normalized to the internal
luciferase control. Cells cultured in medium containing
DHT (androgen), as a positive control, induced a marked
reporter gene expression. Test compounds capable of sig-
nificantly suppressing this DHT-induced reporter gene
expression were identified as potential anti-androgens.
5.11. 7-(3,4-Dimethoxyphenyl)-4-[3-(3,4-dimethoxy
phenyl)-acryloyl]-5-hydroxy-hepta-2,4,6-trienenitrile (13)
Compound 2 was reacted with propionamide prepared
from concentrated aqueous ammonia (14 mL) and
methyl propiolate (4.2 g, 50 mmol) giving 2.1 g of off-
white prisms with cooling below 0 °C as described above
in the preparation of 12. CombiFlashÒ chromatography
eluting with hexane–EtOAc afforded 13 (40 mg). mp
204–205 °C; ESI MS m/z 448.3 (M+H)+; 1H NMR
(300 MHz, DMSO): d 3.70 (3H, s), 3.75 (3H, s), 3.79
(3H, s), 3.80 (3H, s), 6.32 (H, d, J = 9.6 Hz), 6.99 (2H,
dd, J = 8.4 Hz, J = 1.5 Hz), (2H, d, J = 15.6 Hz), (2H,
d, J = 1.8 Hz), (2H, dd, J = 8.4 Hz, J = 1.8 Hz), (2H, d,
J = 15.6 Hz), (H, d, J = 15.9 Hz); Anal. (C26H25NO6Æ9/
4H2O) C, H.
5.12. 1,7-Bis-(3,4-dimethoxyphenyl)-5-hydroxy-4-(3-
hydroxypropenyl)-hepta-1,4,6-trien-3-one (14)
Compound 10 (49.4 mg, 0.1 mmol) was dissolved in
anhydrous THF (10 mL). At ꢀ78 °C, Dibal-H (0.4 mL
1.5 M in toluene, 0.6 mmol) was added to the THF solu-
tion. Stirring was continued for 75 min at ꢀ78 °C.
Water was then added to quench the excess Dibal-H.
The mixture was filtered and extracted with EtOAc.
The filtrate was washed with water and saturated NaCl
solution, and then dried over anhydrous Na2SO4. The
organic solvent was removed in vacuo. PTLC offered
14 (8.4 mg), red powder, mp 178–179 °C; ESI MS m/z
453.2 (M+H)+; 1H NMR (300 MHz, CDCl3): d 3.92
(6H, s), 3.93 (6H, s), 4.40 (2H, d, J = 4.5 Hz), 5.30
(0.4H, s), 5.88 (H, triplet of doublet, J = 15.6 Hz,
J = 4.5 Hz), 6.59 (H, d, J = 15.6 Hz), 6.88 (2H, d,
5.14. LNCaP cell growth assay
As mentioned above, LNCaP cells contain a substantial
amount of mutant AR, and thus the growth of these cells
can be significantly activated upon androgen incubation.
LNCaP cell growth assay was used to further confirm
the anti-androgenic activity detected by the above-men-
tioned AR transactivation assay. An MTT assay, which
relies upon the conversion of a colorless substrate to re-
duced color tetrazolium by the mitochondrial dehydro-
genase (possessed by all viable cells), was used to
measure cell growth. The experimental conditions have
been detailed elsewhere.13 Briefly, cells were plated in

2534
L. Lin et al. / Bioorg. Med. Chem. 14 (2006) 2527–2534
2. Landis, S. H.; Murray, T.; Bolden, S.; Wingo, P. A.
Cancer J. Clin. 1998, 48, 6–29.
3. Crawford, E. D.; Eisenberger, M. A.; McLeod, D. G.;
Spaulding, J. T.; Benson, R.; Dorr, F. A.; Blumenstein, B.
A.; Davis, M. A.; Goodman, P. J. N. Eng. J. Med. 1989,
321, 419–424.
4. Neumann, F.; Jacobi, G. H. J. Clin. Oncol. 1982, 1, 41–65.
5. Yeh, S.; Miyamoto, H.; Chang, C. Lancet 1997, 349, 852–
853.
6. Rao, T. S.; Basu, N.; Siddiqui, H. H. Indian J. Med. Res.
1982, 75, 574–578.
96-well tissue culture plates and then incubated for five
consecutive days in the presence of 1 nM DHT and/or
test compounds in 10% charcoal dextran-stripped serum
containing RPMI. At five days of incubation, the cells
were given MTT (5 mg/mL in PBS) for 3 h at 37 °C.
The resultant precipitate was dissolved in a lysis buffer
and then quantitated at a wavelength of 595 nm using
a microplate reader. To ensure the accuracy of data de-
rived from the MTT analysis, cell count was performed
using duplicate samples. Test compounds that displayed
an adverse effect on the androgen-induced prostate
tumor cell growth were identified as potential anti-
androgens and anti-prostate cancer agents.
7. Sharma, O. P. Biochem. Pharmacol. 1976, 25, 1811–1812.
8. Mazumder, A.; Raghavan, K.; Weinstein, J.; Kohn, K.
W.; Pommier, Y. Biochem. Pharmacol. 1995, 49, 1165–
1170.
9. Kawamori, T.; Lubet, R.; Steele, V. E.; Kelloff, G. J.;
Kaskey, R. B.; Rao, C. V.; Reddy, B. S. Cancer Res. 1999,
59, 597–601.
Acknowledgment
10. Arbiser, J. L. PCT Int. Appl. (Emory University, USA).
Wo, 2001, 30 pp.
11. Aggarwal, B. B.; Kumar, A.; Bharti, A. C. Anticancer Res.
2003, 23, 363–398.
This work was supported by NIH Grant CA-17625 from
the National Cancer Institute, awarded to K. H. Lee.
12. Ohtsu, H.; Xiao, Z.; Ishida, J.; Nagai, M.; Wang, H. K.;
Itokawa, H.; Su, C. Y.; Shih, C.; Chiang, T.; Chang, E.;
Lee, Y.; Tsai, M. Y.; Chang, C.; Lee, K. H. J. Med. Chem.
2002, 45, 5037–5042.
13. Ohtsu, H.; Itokawa, H.; Xiao, Z.; Su, C. Y.; Shih, C. C.
Y.; Chiang, T.; Chang, E.; Lee, Y.; Chiu, S. Y.; Chang, C.;
Lee, K. H. Bioorg. Med. Chem. 2003, 11, 5083–5090.
14. Pedersen, U.; Rasmussen, P. B.; Lawesson, S. O. Liebigs
Ann. Chem. 1985, 1557–1569.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmc.2005.11.034.
References and notes
15. Miyashita, N.; Yoshikoshi, A.; Grieco, P. A. J. Org.
Chem. 1977, 42, 3772–3774.
1. Sakurai, N.; Nakagawa-Goto, K.; Ito, J.; Sakura, Y.;
Nakanishi, Y.; Bastow, K. F.; Cragg, G.; Lee, K. H.
‘‘Antitumor Agents 246. Cytotoxic Alangium Alkaloids
from Alangium longiflorum’’, Phytochemistry, submitted
for publication.
16. Lal, G. S. J. Org. Chem. 1993, 58, 2791–2796.
17. Crow, W. D.; Leonard, N. J. J. Org. Chem. 1965, 30,
2660–2665.
18. Jung, M. E.; Buszek, K. R. J. Am. Chem. Soc. 1988, 110,
3965–3969.