Antitumor agents 222. Synthesis and anti-androgen activity of new diarylheptanoids

Article abstract of DOI:10.1016/j.bmc.2003.08.029

Fifteen new diarylheptanoids were synthesized and evaluated for antagonistic activity against androgen receptor (AR)-mediated reporter gene transcription using DU145, PC-3, and LNCaP prostate cancer cell lines. Most compounds showed activity in a 5α-dihydrotestosterone (DHT)-induced reporter gene expression assay in DU145 cells transfected with wild-type AR. Ten compounds (5, 8, 10, 14-15, and 18-22) were equipotent with hydroxyflutamide (HF), the anti-androgen currently available for the treatment of prostate cancer. However, except for compounds 5 and 10, none of the tested compounds was significantly effective in attenuating DHT-induced reporter gene expression in LNCaP cells, which contain endogenous mutant AR.

Full text of DOI:10.1016/j.bmc.2003.08.029

Bioorganic& Medicinal Chemistry 11 (2003) 5083–5090
Antitumor Agents 222.y Synthesis and Anti-androgen Activity of
New Diarylheptanoids
Hironori Ohtsu,a Hideji Itokawa,a Zhiyan Xiao,a Ching-Yuan Su,b Charles C.-Y. Shih,b
Tzuying Chiang,c Eugene Chang,c YiFen Lee,c Shang-Yi Chiu,c
Chawnshang Changc 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 Avenue, Suite F, San Diego, CA 92121, USA
cGeorge Whipple Laboratory for Cancer Research, Department of Pathology, Urology and Biochemistry, University of Rochester Medical
Center, 601 Elmwood Avenue, Box 626, Rochester, NY 14642, USA
Received 22 May 2003; accepted 22 August 2003
Abstract—Fifteen new diarylheptanoids were synthesized and evaluated for antagonistic activity against androgen receptor
(AR)-mediated reporter gene transcription using DU145, PC-3, and LNCaP prostate cancer cell lines. Most compounds showed
activity in a 5a-dihydrotestosterone (DHT)-induced reporter gene expression assay in DU145 cells transfected with wild-type AR.
Ten compounds (5, 8, 10, 14–15, and 18–22) were equipotent with hydroxyflutamide (HF), the anti-androgen currently available for
the treatment of prostate cancer. However, except for compounds 5 and 10, none of the tested compounds was significantly effective
in attenuating DHT-induced reporter gene expression in LNCaP cells, which contain endogenous mutant AR.
# 2003 Elsevier Ltd. All rights reserved.
Introduction
effects with other hormonal systems, causes a range of
unpleasant side effects including thrombosis, fluid
retention, and loss of libido, which hinder their use as
anti-prostate cancer drugs.8 The agonisticactivity of the
anti-androgen may result in ‘anti-androgen withdrawal
syndrome’ and failure of androgen-ablative therapy
after several months’ treatment. The predominant
mechanism of action of the available anti-androgens is
to compete with the natural androgens at their binding
site on AR. Due to the high sensitivity of AR to
androgens, it is desirable to develop anti-androgens
with high specificity and affinity to AR as potential anti-
prostate cancer drugs.
Prostate cancer is the second most common cancer
diagnosed in US males and the androgen receptor (AR)
has been well documented to play an essential role in its
progression.2ꢀ4 The AR functions as a ligand-dependent
transcription factor and belongs to the super family of
nuclear receptors.5,6 High levels of androgens may con-
stitutively activate the AR signaling pathway so to be
responsible for the progression of prostate cancer.
Logical approaches to treat prostate cancer include
ablation of androgens by either anti-androgens (AR
antagonists) neutralization or bilateral orchiectomy.
Currently, anti-androgens, in combination with surgical
or medical castration, are widely used for the treatment
of prostate cancer.7
The phenolic diarylheptanoid, curcumin (1), is the
major pigment in turmeric(the rhizome of Curcuma
longa). Several curcumin analogues showed cytotoxicity
against multiple tumor cell lines. Notably, some com-
pounds were moderately active (with ED50 ranging from
2.8 to 8.0 mg/mL) against two prostate cancer cell lines
(PC-3 and LNCaP clone FGC), which are usually not
responsive to chemotherapeutic agents.9 In a previous
paper, we reported that several diarylheptanoids
showed potent antagonisticactivity against the andro-
gen receptor (AR) (compounds 2–7, Fig. 1).10 Among
Both steroidal and non-steroidal anti-androgens are
presently available and have shown clinical benefit as
chemotherapeutic agents for prostate cancer. However,
their agonisticatcivity, as well as their overlapping
*Corresponding author. Tel.: +1-919-962-0066; fax: +1-919-966-
3893; e-mail: khlee@unc.edu
yFor Part 221, see ref 1.
0968-0896/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2003.08.029

5084
H. Ohtsu et al. / Bioorg. Med. Chem. 11 (2003) 5083–5090
Figure 1. Structures of diarylheptanoids.
Chemistry
these prior compounds, 3 [5-hydroxy-1,7-bis(3,4-dimeth-
oxyphenyl)-1,4,6-heptatrien-3-one] and 5 [7-(4-hydroxy-
3-methoxyphenyl)-4-[3-(4-hydroxy-3-methoxyphenyl)-
acryloyl]-5-oxo-hept-6-enoic acid ethyl ester] showed
the most promising anti-AR activity and were as potent
as hydroxyflutamide (HF), which is the currently avail-
able anti-androgen for the treatment of prostate cancer.
The previous structure–activity relationship (SAR)
study indicated that the bis(3,4-dimethoxyphenyl) and
conjugated b-diketone moieties are important to the
activity but modification of the b-diketone moiety led to
decreased potency. However, chemical modification on
the phenyl rings was well tolerated and, thus, could
probably result in significantly optimized activity pro-
files. In an effort to obtain compounds with more potent
antagonistic activity against the AR, we next focused on
chemical modification of the bis-aromatic ring system.
Phenyl rings bearing fluorine, nitro, and dimethyl-
amino, as well as methoxy groups were incorporated into
new diarylheptanoid derivatives as described herein.
The general procedure for synthesizing diarylheptanoids
8–20 is summarized in Scheme 1. 2,4-Pentanedione or
ethyl 4-acetyl-5-oxo-hexanoate was condensed with the
appropriate benzaldehyde in EtOAcat 40 ꢁC using the
method of Pedersen et al.11 Protection with a boron
complex was necessary in order to avoid Knoevenagel
condensation at C-3 of 2,4-pentanedione. Although
such protection should not be needed with ethyl 4-ace-
tyl-5-oxo-hexanoate, reactions without the complexa-
tion did not succeed.11 Two commercially unavailable
benzaldehydes were synthesized as shown in Scheme 2.
To prepare 4-dimethylamino-3-methoxybenzaldehyde,
4-amino-3-methoxybenzoicacid was first esterified with
diazomethane then N-dimethylated in good yield with
formaldehyde and sodium borohydride in aqueous
THF.12 Attempted direct reduction of the carbomethoxy
group to the aldehyde with 1.1 equiv of DIBAL-H
resulted in a mixture of alcohol and aldehyde. Reduction
Scheme 1. Synthesis of diarylheptanoids 7–19.

H. Ohtsu et al. / Bioorg. Med. Chem. 11 (2003) 5083–5090
5085
Scheme 2. Synthesis of aldehydes.
of the methyl benzoate with excess DIBAL-H (2 equiv)
followed by Swern oxidation gave 4-dimethylamino-3-
methoxybenzaldehyde in an acceptable yield (60% in
two steps). 5,6-Dimethoxynicotinaldehyde was prepared
from 2-chloro-3-pyridinol.13 Compound 13 was synthe-
sized by treating compound 12 with trimethylsilyldiazo-
methane and N,N-diisopropylethylamine in methanol
(41% yield). Alkylation of curcumin with iodopropane
and K2CO3 in refluxing acetone for 48 h gave a mixture
of tetrapropyl (compound 21, m/z 559 [M+Na]+) and
tripropyl (compound 22, m/z 517 [M+Na]+) deriva-
tives (Scheme 3). The structures of all new compounds
induced reporter gene expression to various degrees as
shown in Table 1A (column 1). Ten compounds (5, 8,
10, 14–15, and 18–22) were equipotent with hydroxy-
flutamide (HF), the anti-androgen currently available
for the treatment of prostate cancer. Surprisingly, com-
pounds 16 and 17 did not show any suppression. These
results were confirmed using PC-3 cells transfected with
wild-type AR (Table 1A, column 2). In the latter assay
system, compounds 5–13 displayed detectable activity
and were able to block DHT-induced wild-type AR
transactivation; however, they were less active than HF
in this assay system. The greater potency of these com-
pounds in the DU145 cells, in which AR/ARA70-medi-
ated reporter gene expression is involved, suggests that
ARA70 plays a role in enhancing the antagonist activity
of these diarylheptanoid compounds.
1
were confirmed by H NMR and MS data.
Results and Discussion
The curcumin analogues 2–22 were first evaluated for
their antagonisticaticvity against the AR in an
AR-mediated transactivation assay following the prior
conditions.14 Three different human prostate cancer cell
lines, PC-3, DU-145, and LNCaP, were chosen. PC-3
cells are androgen-independent tumor cells that do not
express functional AR; DU-145 cells are androgen-
independent tumor cells that express neither functional
AR nor endogenous ARA (AR co-activator); and
LNCaP are prostate cancer cells expressing mutated AR
(threonine 877 to alanine). The DU145 cells allow a
more sensitive detection of AR agonist or antagonist
activity than PC3 or LNCaP cells.
We further tested if these compounds could block
mutated AR-mediated transactivation and cell growth
in LNCaP cells (Table 1A and B, respectively). At a test
dose of 1 mM (Table 1A, column 3), only compounds 5
and 10 were significantly effective (equipotent with HF)
in attenuating DHT-induced reporter gene expression in
LNCaP cells. When the test dose was raised to 2.5 mM
(Table 1A, column 4), compounds 19 and 20 showed
detectable activity in blocking AR-mediated transacti-
vation. However, at 2.5 mM, compounds 19 and 20 also
negatively affected the growth of fetal rat heart-derived
H9c2 myocytes (a 10–20% reduction, compared to
vehicle control). Therefore, the above-observed AR
transactivation blockage may reflect a nonspecific cyto-
toxicity of compounds 19 and 20 at an elevated dose.
These data suggest that most of these compounds may
have no or only marginal activity in blocking mutant
AR activation by DHT.
Using DU145 cells transfected with wild-type AR and
ARA70, compounds 2 and 4–22 were tested for their
ability to block AR function in the presence of DHT. At
a dose of 1 mM, most compounds inhibited DHT-
Scheme 3. Synthesis of diarylheptanoids 21 and 22.

5086
H. Ohtsu et al. / Bioorg. Med. Chem. 11 (2003) 5083–5090
Table 1. Suppression of DHT-induced AR activation and cell growth by compounds 2–22
Compd
A: Suppression of DHT-induced AR activationa
B: Suppression of cell growthb
DU145/wtARc
PC3/wtARc
LNCaPc
LNCaPd
PC3/mtARc
LNCaP+DHTc
LNCaP-DHTc
HF
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
+++
++
++++
++
++
+
++
++
++
0
+
++
0
++
0
+
0
+
+
0
+
+
+
++
+++
++
++
+++
+
+++
++
++
+++
++
++
+
+
+
++
++
++
++
+
+
+
0
++
+
+
+
0
0
0
0
0
0
0
0
0
++
+
+
0
+
+
0
0
0
0
++
+++
+++
0
0
0
0
0
0
+++
+++
+++
+++
+++
0
++
++
0
0
0: not effective (ꢀ5 to +5%) of DHT value; +: 10–20% reduction; ++: 20–30% reduction; +++: 30–40% reduction; ++++: 40–50%
reduction.
aDU-145, PC-3, and LNCaP cell lines were seeded and cotransfected with reporter MMTV-luciferase (all cell lines), wild-type (wt) or mutant (mt)
AR expression plasmid (DU-145, PC-3) and ARA70 (DU-145) using SuperFect. Subsequently, the transfected cells were harvested and re-plated in
10% charcoal-stripped fetal bovine serum DMEM medium. The cells were then treated with dehydrotestosterone (DHT, 1 nM), and antiandrogens
(1 mM) and harvested for detection of the luciferase activity. (cf. Experimental).
bSuppression of cell growth was accessed with the MTT assay (cf. Experimental).
c1.0 mM of test compound.
d2.5 mM of test compound.
It should be noted that, although compound 10 was
able to block DHT-induced mutant AR transactivation
in LNCaP cells, it did not effectively reduce DHT-
induced reporter gene expression in PC-3 cells trans-
fected with the same mutant AR (compare the third and
fifth columns, Table 1A). Cellular factors in different
cell lines may affect the antagonistic activity of com-
pound 10 toward mutant AR; thus, compound 10
would not have consistent antagonistic activity among
various prostate cancer cell types. It is also noteworthy
that when compounds 11–13 were further studied using
a LNCaP cell growth assay (Table 1B), they caused only
minimal suppression of prostate cancer cell growth in
the presence of DHT.
H9c2 muscle cells (American Type Cell Collection).
Absence of the AR was demonstrated by Western blot-
ting in primary cultures of both cell lines. Their growth
after a 5-day incubation with compound 3 at 1–5 mM
was comparable to that with vehicle control (0.2%
DMSO). Therefore, we concluded that compound 3,
and probably other diarylheptanoid compounds, might
selectively block wild-type AR function and, thus, their
inhibitory activity against prostate cancer cell growth
might be mediated by the AR.
Furthermore, the new compounds were evaluated for
AR agonist activity in several different systems (Table
2A). In this study, HF, an anti-androgen with residual
agonist activity, was included as a positive control to
validate the conditions for detection of weak or moder-
ate AR agonist activity. In PC-3 cells transfected with
wild-type or mutant AR, compounds 8–10 at a dose of 1
mM, did not stimulate reporter gene expression,
although HF at the same test dose caused an ꢂ200%
increase in wild-type AR-mediated reporter gene
expression. Using DU145 cells transfected with wild-
type AR and ARA70 (which enhances AR’s response to
agonists), compounds 11–22 at a concentration of 1 mM
also did not increase reporter gene expression, although
HF at the same dose evoked AR/ARA70-mediated
transactivation by ꢂ50%. Compounds 11–13 were also
evaluated in an LNCaP cell growth assay (Table 2B) and,
at a dose of 1 mM, did not stimulate cell growth. There-
fore, we concluded that these compounds have little AR
agonist activity and do not stimulate the biological
Dorai et al.15 reported an inhibitory effect of curcumin
(ꢂ10 mM) on the growth of LNCaP (AR-positive) and
PC-3 (AR-negative) prostate cancer cell lines, which
implied that general non-specific cytotoxicity might be
responsible for the above observed effects. To eliminate
this possibility, compound 3 was tested in parallel with
HF against PC-3 and DU145 cells, which are AR nega-
tive and, hence, androgen-insensitive. The results indi-
cated that, at a dose up to 5 mM, compound 3 displayed
no adverse effect on the growth of either PC-3 or
DU145 cells. The reference compound, HF, had no
inhibitory effect on the growth of PC-3 cells, but slightly
increased the growth of DU145 cells. Compound 3 was
further tested for growth inhibition of two AR-negative
normal non-prostate cells, human coronary artery
endothelial cells (Clonetics) and fetal rat heart-derived

H. Ohtsu et al. / Bioorg. Med. Chem. 11 (2003) 5083–5090
5087
Table 2. AR agonist activity of compounds 2–22
on aluminum plates (Aldrich, Inc.). Elemental analyses
were performed by Atlantic Microlab Inc., Norcross,
GA, USA.
Compd
A: AR activation
B: LNCaP cell
growth inhibition
DU145/
wtAR
PC3/ PC3/
wtAR mtAR
LNCaP
Preparation of 3-dimethylamino-4-methoxybenzaldehyde
HF
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
+
++
++
+
0
0
0
0
3-Amino-4-methoxybenzoicaicd (5 g, 30 mmol) was
treated with excess diazomethane-etherate to give
methyl 3-amino-4-methoxybenzoate (5.16 g, 95% yield).
A slurry of methyl 3-amino-4-methoxybenzoate (3.7 g,
20.4 mmol) and finely crushed sodium borohydride
(4.42 g, 117 mmol) in dry THF (40 mL) was added to a
stirred solution of 3 M H2SO4 (6 mL) and 37% aqueous
formaldehyde (6.1 mL, 81.6 mmol) in THF (40 mL) at
0ꢂ20 ꢁC (ice bath to room temperature). After the first
half of the addition, the mixture was acidified with 3 M
H2SO4 (6 mL) and the addition was continued. To the
resultant mixture, water (50 mL) was added with stir-
ring, and then the solution was made strongly basicby
addition of solid KOH. The organicphase was sepa-
rated and the aqueous phase was extracted with diethyl
ether (50 mLꢃ2). The combined organic layers were
washed with brine and dried over Na2SO4. The solution
was concentrated in vacuo to give methyl 3-dimethyla-
mino-4-methoxybenzoate as a brown oil (4.18 g, 98%
yield). IR nmax (CHCl3) 2800, 1710, 1598, 1509; 1H
NMR (CDCl3) d 2.19 (6H, s), 3.89 (3H, s), 3.95 (3H, s),
6.87 (1H, d, J=8.4 Hz), 7.63 (1H, d, J=1.8 Hz), 7.72 (1
h, dd, J=8.4, 1.8 Hz); positive ESI-MS m/z 210
[M+1]+.
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0: not effective (ꢀ5 to +5%); +: 50–60% increase above vehicle con-
trol; ++: 175–200% increase. The same protocols described in Table
1’s legend were employed.
function of wild-type or mutant AR, which represents a
significant therapeutic advantage of this compound
class over the currently available anti-androgens.
To a stirred solution of 3-dimethylamino-4-methoxy-
benzoate (3.06 g, 14.6 mmol) in dry CH2Cl2 (20 mL) at
0 ꢁC, DIBAL-H (1.5 M toluene solution, 20 mL, 30
mmol) was added in one portion. The reaction was
stirred for 30 min at 0 ꢁC, and then quenched by adding
MeOH (5 mL). After stirring at room temperature for
an additional 30 min, the resulting slurry was filtered
and washed thoroughly with CH2Cl2. The solvent was
removed in vacuo to give 3-dimethylamino-4-methoxy-
benzylalcohol (2.16 g, 82% yield) as a pale-yellow oil.
IR nmax (CHCl3) 3432 (OH), 2789, 1602, 1509; 1H
NMR (CDCl3) d 2.77 (6H, s), 3.87 (3H, s), 4.58 (2H, s),
6.81 (1H, brd, J=8.2 Hz), 6.94 (1H, dd, J=8.2, 2.4 Hz),
6.95 (1H, brs); positive ESI-MS m/z 182 [M+1]+.
At the present time, accurate SAR conclusions are dif-
ficult to formulate from semi-quantitative biological
data. Compounds with various substitution patterns
and both electron-attracting and -releasing substituents
showed anti-androgen activity. The sole compound (16)
with a nitrogen heteroatom incorporated into the
phenyl ring was inactive.
In conclusion, driven by previous SAR, we synthesized
a number of new curcumin analogues by diversely sub-
stituting the phenyl rings. Biological evaluation of these
compounds confirmed our previous observations that
curcumin analogues are a novel class of anti-androgens.
Further chemical modification is ongoing to optimize
their anti-androgen activity.
Dry DMSO (3.35 mL, 47.2 mmol) in CH2Cl2 (5 mL)
was added to oxalyl chloride (2.06 mL, 23.6 mmol) in
CH2Cl2 (5 mL) at ꢀ78 ꢁC, and the mixture was stirred
for 10 min at ꢀ78 ꢁC. 3-Dimethylamino-4-methoxy-
benzylalcohol (2.14 g, 11.8 mmol) in dry CH2Cl2 (10
mL) was added dropwise to the mixture over 10 min
and stirring was continued at ꢀ78 ꢁC for 15 min then at
ꢀ40 to ꢀ45 ꢁC for 2 h. Triethylamine (8.2 mL, 59.0
mmol) was added to the mixture, which was stirred for
20 min at 0 ꢁC. The reaction was quenched with sat.
NH4Cl solution, and the organiclayer was separated.
The aqueous layer was extracted with CH2Cl2, and the
combined CH2Cl2 solution was washed successively
with water and brine, dried (Na2SO4), and concentrated
to give 3-dimethylamino-4-methoxybenzaldehyde (1.55
g, 74% yield) as a brown oil. IR nmax (CHCl3) 2800,
Experimental
Melting points were determined on a Fisher-Johns
melting apparatus and are uncorrected. All reagents and
solvents were reagent grade or were purified by standard
methods before use. 1H NMR spectra were recorded on
a Varian Gemini-300 spectrometer. The chemical shifts
are presented in terms of ppm with TMS as the internal
reference. IR spectra were recorded on a MIDAC M
series spectrometer. MS spectra were recorded on
HP5989A and JMS D-300 instruments. Column chro-
matography was carried out on silica gel 60 (Merck
70–230 and 230–400 mesh), and thin-layer chromato-
graphy (TLC) was performed using pre-coated silica gel
1
1686 (Ar–CHO), 1591, 1506; H NMR (CDCl3) d 2.83

5088
H. Ohtsu et al. / Bioorg. Med. Chem. 11 (2003) 5083–5090
(6H, s), 3.98 (3H, s), 6.97 (1H, d, J=8.1 Hz), 7.48 (1H,
dd, J=8.1, 1.8 Hz), 7.52 (1H, d, J=1.8 Hz), 9.85 (1H,
s); positive ESI-MS m/z 180 [M+1]+.
der, mp 280 ꢁC (decomp, n-hexane–EtOAc). H NMR
1
(DMSO-d6) d 3.94 (6H, s), 6.12 (1H, s), 6.97 (2H, d,
J=15.3 Hz), 7.61 (2H, d, J=15.3 Hz), 7.63 (2H, s), 7.82
(2H, s). ESI-MS m/z 457 [Mꢀ1]+. Anal. calcd for
C21H18N2O10: C, 55.03; H, 3.96. Found: C, 55.07; H, 3.98.
General procedure for synthesis of diarylheptanoids
(8–20)
Compound 13. To a solution of 12 (116 mg, 0.253 mmol)
in dioxane–MeOH (10 mL, 1:1), N,N-diisopropylethyl-
amine (122 mL, 0.70 mmol) and 2 M hexane solution of
trimethylsiliyldiazomethane (350 mL, 0.70 mmol) were
added and the mixture was stirred at room temperature
overnight under N2. The solvent was removed in vacuo
and the resulting residue was recrystallized from hex-
ane–EtOActo give 40.8 mg of pure 13. The mother
liquor was subjected to silica gel column chromato-
graphy (CHCl3) to give additional 13 (10 mg). Forty-
one percent yield, yellow powder, mp 209–210 ꢁC
The procedure reported by Pedersen et al.11 was used to
prepare compounds 8–20. In general, 2,4-pentanedione
(1 equiv, for compounds 4–17) or ethyl 4-acetyl-5-oxo-
hexanoate (1 equiv, for compounds 18–20) and boric
anhydride (0.7 equiv) dissolved in 10 mL of EtOAcwere
stirred for 30 min at 40 ꢁC. The appropriate benzalde-
hyde (2 equiv) and tributylborate (4 equiv) were added,
and the mixture was stirred for 30 min. n-Butylamine
(1.5 equiv) dissolved in 10 mL of EtOAcwas added
dropwise over 20 min. Stirring was continued for 18 h at
40 ꢁC. The mixture was then hydrolyzed by adding 10
mL of 1 N HCl and heating at 60 ꢁC for 1 h. The
organiclayer was separated, and the aqueous layer was
extracted three times with EtOAc. The combined
organiclayers were washed with H 2O until neutral, then
dried over Na2SO4, and the solvent was removed in
vacuo. The crude products were purified by silica gel
column chromatography eluting with n-hexane–EtOAc.
1
(n-hexane–EtOAc). H NMR (acetone-d6) d 3.99 (6H,
s), 4.06 (6H, s), 6.12 (1H, s), 7.01 (2H, d, J=16.0 Hz),
7.68 (2H, d, J=16.0 Hz), 7.70 (4H, s). ESI-MS m/z 485
[Mꢀ1]+. Anal. calcd for C23H22N2O10: C, 56.79; H,
4.56. Found: C, 56.50; H, 4.55.
Compound 14. 20% yield (started with 5 mmol of
3-dimethylamino-4-methoxybenzaldehyde), orange nee-
dles, mp 126–127 ꢁC (n-hexane-EtOAc). 1H NMR
(CDCl3) d 2.83 (12H, s), 3.93 (6H, s), 5.83 (1H, s), 6.50
(2H, d, J=16.0 Hz), 6.86 (2H, d, J=16.0 Hz), 7.14 (2H,
d, J=1.8 Hz), 7.20 (2H, d, J=8.4, 1.8 Hz), 7.61 (2H, d,
J=16.0 Hz). ESI-MS m/z 421 [Mꢀ1]+. Anal. calcd for
C25H30N2O4: C, 71.07; H, 7.16. Found: C, 71.17; H, 7.35.
Compound 8. 3% yield (started with 5 mmol of
2,4-difluorobenzaldehyde), yellow powder, mp 155.5–
156 ꢁC (n-hexane–EtOAc). 1H NMR (CDCl3) d 5.86
(1H, s), 6.68 (2H, d, J=16.0 Hz), 6.84–6.96 (4H, m),
7.56 (2H, m), 7.71 (2H, d, J=16.0 Hz). ESI-MS m/z 347
[Mꢀ1]+. Anal. calcd for C19H12F4O2: C, 65.52; H, 3.47.
Found: C, 65.66; H, 3.57.
Compound 15. 28% yield (started with 5 mmol of
3,4,5-trimethoxybenzaldehyde), yellow needles, mp
189–190 ꢁC (n-hexane–EtOAc, lit:16 mp 108–109 ꢁC).
1H NMR (CDCl3) d 3.90 (6H, s), 3.91 (12H, s), 5.87
(1H, s), 6.53 (2H, d, J=15.6 Hz), 6.79 (4H, s), 7.59
(2H, d, J=15.6 Hz). ESI-MS m/z 455 [Mꢀ1]+. Anal.
calcd for C25H28O8: C, 65.78; H, 6.18. Found: C,
65.66; H, 6.16.
Compound 9. 10% yield (started with 5 mmol of
2-fluoro-4-methoxybenzaldehyde), yellow powder, mp
163–165 ꢁC (n-hexane–EtOAc). 1H NMR (CDCl3) d
3.84 (6H, s), 5.82 (1H, s), 6.62 (2H, d, J=16.0 Hz), 6.70
(4H, m), 7.48 (2H, t, J=8.7 Hz), 7.71 (2H, d, J=16.0
Hz). ESI-MS m/z 371 [Mꢀ1]+. Anal. calcd for
C21H18F2O4: C, 67.74; H, 4.87. Found: C, 67.58; H,
4.92.
Compound 16. 27% yield (started with 5 mmol of
3,4-diethoxybenzaldehyde), yellow powder, mp 138–139 ꢁC
(n-hexane–EtOAc). 1H NMR (CDCl3) d 1.47 (6H, t,
J=6.9 Hz), 1.48 (6H, t, J=6.9 Hz), 4.14 (8H, q, J=6.9
Hz), 5.80 (1H, s), 6.47 (2H, d, J=15.6 Hz), 6.87 (2H, d,
J=8.4 Hz), 7.09 (2H, brs), 7.11 (2H, brd, J=15.6 Hz),
7.56 (2H, d, J=15.6 Hz). ESI-MS m/z 451 [Mꢀ1]+.
Anal. calcd for C27H32O6: C, 71.66; H, 7.13. Found: C,
71.46; H, 7.06.
Compound 10. 9% yield (started with 5 mmol of
2-fluoro-6-methoxybenzaldehyde), yellow needles, mp
144–145 ꢁC (n-hexane–EtOAc). 1H NMR (CDCl3) d
3.82 (6H, s), 5.90 (1H, s), 6.72 (2H, d, J=16.0 Hz), 6.88
(2H, m), 7.56 (2H, m), 7.00-7.07 (4H, m), 7.74 (2H, d,
J=16.0 Hz). ESI-MS m/z 371 [Mꢀ1]+. Anal. calcd for
C21H18F2O4: C, 67.74; H, 4.87. Found: C, 67.60; H,
4.95.
Compound 17. 17% yield (started with 5 mmol of 5,6-
Compound 11. 29% yield (started with 5 mmol of
6-nitroveratraldehyde), yellow needles, mp 240–241 ꢁC
(n-hexane-EtOAc). H NMR (CDCl3) d 3.99 (6H, s),
4.04 (6H, s), 6.05 (1H, brs), 6.50 (2H, brd, J=15.3 Hz),
7.03 (2H, brs), 7.64 (2H, brs), 8.18 (2H, d, J=15.3 Hz).
EI–MS m/z 440 [MꢀNO2]+. Anal. calcd for
C23H22N2O10: C, 56.79; H, 4.56. Found: C, 56.58; H,
4.65.
dimethoxynicotinaldehyde),
yellow
needles,
mp
168–170 ꢁC (n-hexane–EtOAc). 1H NMR (CDCl3) d
3.94 (6H, s), 4.06 (6H, s), 5.84 (1H, s), 6.50 (2H, d,
J=16.0 Hz), 7.21 (2H, brs), 7.61 (2H, dd, J=16.0 Hz),
7.89 (2H, brs). ESI-MS m/z 397 [Mꢀ1]+. Anal. calcd
for C21H22N2O6: C, 63.31; H, 5.57; N, 7.03. Found: C,
63.84; H, 5.40; N, 6.99.
1
Compound 18. N,N-Diisopropylethylamine (185 mL,
1.06 mmol) and trimethylsilyldiazomethane (0.53 mL,
2 M hexane solution, 1.06 mmol) were added to a solution
Compound 12. 64% yield (started with 5 mmol of 4 -
hydroxy-3-methoxy-5-nitrobenzaldehyde), yellow pow-

H. Ohtsu et al. / Bioorg. Med. Chem. 11 (2003) 5083–5090
5089
of compound 5 (116 mg, 0.354 mmol) in MeOH–benzene
(10 mL, 1:1), and the mixture was stirred at room
temperature overnight under N2. The solvent was
removed in vacuo and the resulting residue was subjected
to silica gel column chromatography using n-hexane–
EtOActo give 78 mg of pure 18 as a yellow oil. 44%
yield, yellow oil. 1H NMR (CDCl3) d 1.25 (3H, t, J=7.2
Hz), 2.33–2.58 (2H, m), 2.95 (1H, m), 3.91 (6H, s), 3.93
(3H, s), 3.96 (3H, s), 4.14 (2H, q, J=7.2 Hz), 6.72 (2H,
d, J=16.0 Hz), 6.85–7.22 (6H, m), 7.73 (2H, d, J=16.0
Hz). ESI-MS m/z 495 [Mꢀ1]+.
517 [M+Na]+. Anal. calcd for C30H38O6: C, 72.85; H,
7.74. Found: C, 72.89; H, 7.77.
Biological assays
Cell culture and gene transfection.14 Human prostate
cancer DU145, PC-3, and LNCaP cells were maintained
in Dulbecco’s minimum essential medium (DMEM) or
RPMI medium containing penicillin (25 units/mL),
streptomycin (25 g/mL), and 10% fetal calf serum. For
gene transfection, the previously described conditions
were followed with minor modifications. LNCaP cells,
containing an endogenous mutant AR, were transfected
with reporter gene. PC-3, lacking functional AR, were
transfected with an AR expression plasmid and reporter
gene. DU145 cells with a low content of endogenous
AR coactivators, were transfected with expression plas-
mids for AR and ARA70 coactivator and a reporter gene.
Transfections were performed using the SuperFect kit
(Qiagen, Chatsworth, CA, USA). Briefly, 1–2ꢃ105 cells
were plated on each 35-mm dish 24 h before transfection,
and then a reporter gene, MMTV-luciferase, which con-
tains MMTV-LTR promoter and AR-binding element,
was co-transfected with AR expression plasmid (wild-
type or mutant), or pSG5ARA70. pRL-TK was used as
an internal control for transfection efficiency. After a 3-h
transfection, the medium was changed to DMEM or
RPMI supplemented with 10% charcoal stripped serum,
and 20 h later, the cells were treated with DHT (1 nM)
and/or test compounds for another 20 h. The cells were
harvested and assayed for luciferase activity using Dual
Luciferase Assay System (Promega, Madison, WI, USA).
Data were expressed as relative luciferase activity nor-
malized to the internal luciferase positive control.
Compound 19. 25% yield (started with 5 mmol of 3,4,5-
trimethoxybenzaldehyde), yellow oil. 1H NMR (CDCl3)
d 1.24 (3H, t, J=7.2 Hz), 2.38 (1H, m), 2.56 (1H, t,
J=7.2 Hz), 2.97 (1H, t, J=7.2 Hz), 3.90 (6H, s), 3.93
(12H, s), 4.13 (2H, q, J=7.2 Hz), 6.84 (4H, s), 7.04 (2H,
d, J=15.3 Hz), 7.70 (2H, d, J=15.3 Hz); ESI-MS m/z
555 [Mꢀ1]+. Anal. calcd for C30H36O10: C, 64.70; H,
6.52. Found: C, 64.86; H, 6.48.
Compound 20. 28% percent yield (started with 5 mmol
of 3-dimethylamino-4-methoxybenzaldehyde), yellow
1
oil. H NMR (CDCl3) d 1.26 (3H, t, J=6.9 Hz), 2.37
(1H, m), 2.55 (1H, s), 2.80 (6H, s), 2.84 (6H, s), 2.96
(1H, m), 3.92 (3H, s), 3.94 (3H, s), 4.09–4.19 (2H, m),
6.69–7.25 (8H, m), 7.73 (2H, d, J=15.3 Hz). ESI-MS
m/z 521 [Mꢀ1]+.
7-(3-Methoxy-4-propyloxyphenyl)-4-propyl-4-[3-(3-meth-
oxy-4-propyloxyphenyl)-acryloyl]-hept-6-en-5-one (21)
and 7-(3-methoxy-4-propyloxyphenyl)-4-[3-(3-methoxy-
4-propyloxyphenyl)-acryloyl]-hept-6-en-5-one (22). Iodo-
propane (3.34 mL, 34.2 mmol) was added to a stirred
solution of curcumin (628 mg, 1.71 mmol) and K2CO3
(473 mg, 34.2 mmol) in dry acetone (35 mL), and the
mixture was refluxed for 48 h. After cooling the mixture
to room temperature and filtering, the solvent was
removed in vacuo. The resulting yellow solid was sub-
jected to silica gel column chromatography (n-hexane–
EtOAc) to give compounds 21 (283 mg, 31% yield) and
22 (112 mg, 13% yield).
LNCaP and H9c2 cell growth assay.17 MTT assay, a
colorimetric assay that measures the conversion of a
colorless substrate to reduce into a colored tetrazolium
by mitochondrial dehydrogenase, was employed to
assess cell proliferation. Briefly, LNCaP and H9c2 cells
were plated at a density of 1ꢃ103/well and 3ꢃ103/well,
respectively, in 96-well tissue culture plates. Subse-
quently, the cells were incubated with 10% charcoal
stripped serum-containing RPMI supplemented with
DHT (1 nM) and/or test compounds for 5 consecutive
days. At the end of incubation, the cells were added
with an MTT-containing solution (5 mg/mL in PBS) in
1/10 of volume for 3 h at 37 ꢁC. The resultant pre-
cipitate was dissolved in lysis buffer containing 50%
dimethyl formamide, 5% sodium dodecyl sulphate,
0.35 M acetic acid and 50 mM HCl. The lysate was read
at a wavelength of 595 nm using a microplate reader. To
ensure the accuracy of the data derived from the MTT
assay, cell counts from duplicate wells was also per-
formed in some growth assays.
Compound 21. Pale-yellow needles, mp 118–119 ꢁC
(n-hexane–EtOAc). 1H NMR (CDCl3) d 0.93 (6H, t,
J=6.9 Hz), 1.03 (6H, t, J=7.5 Hz), 1.26 (3H, m),
1.80–1.91 (4H, m), 1.98–2.05 (4H, m), 3.88 (6H, m), 3.99
(4H, t, J=6.9 Hz), 6.64 (2H, d, J=15.3 Hz), 6.82 (2H,
d, J=8.1 Hz), 6.99 (2H, d, J=1.8 Hz), 7.11 (2H, dd,
J=8.1, 1.8 Hz), 7.68 (2H, d, J=15.3 Hz); ESI-MS m/z
559 [M+Na]+. Anal. calcd for C33H44O6: C, 73.85; H,
8.26. Found: C, 73.97; H, 8.53.
Compound 22. Pale-yellow needles, mp 132–133 ꢁC
(n-hexane–EtOAc). 1H NMR (CDCl3) d 0.97 (3H, t,
J=6.9 Hz), 1.05 (3H, t, J=7.2 Hz), 1.06 (3H, t, J=7.2
Hz), 1.05 (2H, m), 1.80-1.95 (4H, m), 2.44 (2H, t, J=8.4
Hz), 2.69 (1H, dd, J=16.5, 3.3 Hz), 2.94 (1H, dd,
J=16.5, 13.8 Hz), 3.90 (3H, s), 3.91 (3H, s), 4.01 (2H, t,
J=6.9 Hz), 4.02 (2H, t, J=6.9 Hz), 5.31 (1H, dd,
J=14.0, 3.0 Hz), 6.84 (2H, d, J=8.1 Hz), 6.93 (1H, d,
J=9.0 Hz), 7.01 (2H, dd, J=6.0, 1.8 Hz), 7.09 (1H, dd,
J=8.1, 1.8 Hz), 7.32 (2H, d, J=16.0 Hz). ESI-MS m/z
Acknowledgements
This work was supported by grants from the National
Cancer Institute, NIH: Grant CA-17625 awarded to
K. H. Lee, and Grant 1 R43 CA-96189-01 awarded to
C. C-Y. Shih.

5090
H. Ohtsu et al. / Bioorg. Med. Chem. 11 (2003) 5083–5090
9. Ishida, J.; Ohtsu, H.; Tachibana, Y.; Nakanishi, Y.; Bas-
References and Notes
tow, K. F.; Nagai, M.; Wang, H. K.; Itokawa, H.; Lee, K. H.
Bioorg. Med. Chem. 2002, 10, 3481.
1. For previous paper in series, Hayashi, K.; Nakanishi, Y.;
Bastow, K. F.; Cragg, G.; Nozaki, H.; Lee, K. H. A. J. Nat.
Prod. 2003, 66, 125.
2. Visakorpi, T.; Hyytinen, E.; Koivisto, P.; Tanner, M.; Kei-
nanen, R.; Palmberg, C.; Palotie, A.; Tammela, T.; Isola, J.;
Kallioniemi, O. P. Nat. Genet. 1995, 9, 401.
10. Ohtsu, H.; Xiao, Z.; Ishida, J.; Nagai, M.; Wang, H. K.; Ito-
kawa, H.; Su, C. Y.; Shih, C.; Chiang, T.; Chang, E.; Lee, Y. F.;
Chiu, S. Y.; Chang, C.; Lee, K. H. J. Med. Chem. 2002, 45, 5037.
11. Pedersen, U.; Rasmussen, P. B.; Lawesson, S. O. Liebigs
Ann. Chem. 1985, 1557.
3. Brinkmann, A. O.; Trapman, J. Nat. Med. 2000, 6, 628.
4. Taplin, M. E.; Bubley, G. J.; Shuster, T. D.; Frantz, M. E.;
Spooner, A. E.; Ogata, G. K.; Keer, H. N.; Balk, S. P. New
Engl. J. Med. 1995, 332, 1393.
5. Chang, C.; Kokontis, J.; Liao, S. Proc. Natl. Acad. Sci.
U.S.A. 1988, 85, 7211.
6. Beato, M. Cell 1989, 56, 335.
12. Giumanini, A. G.; Chiavari, Giuseppe; Musiani, M. M.;
RossiP. Synthesis 1980, 743.
13. Stogryn, E. L. J. Heterocycl. Chem. 1974, 11, 251.
14. Myiamoto, H.; Yeh, S.; Wilding, G.; Chang, C. Proc.
Natl. Acad. Sci. U.S.A. 1998, 95, 7379.
15. Dorai, T.; Gehani, N.; Katz, A. Prostate Cancer Prostatic
Dis. 2000, 3, 84.
7. 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. New Engl. J. Med. 1989, 321,
419.
16. Nurfina, A. N.; Reksohadiprodjo, M. S.; Timmerman, H.;
Jenie, U. A.; Sugiyanto, D.; van der Goot, H. Eur. J. Med.
Chem. 1997, 32, 321.
17. Su, C. Y.; Chong, K. Y.; Chang, J. X.; Ryter, S.; Khar-
dori, R.; Lai, C. C. J. Mol. Cell. Cardiol. 1999, 31, 845.
8. Furr, B. J.; Tucker, H. Urology 1996, 47, 13.