Antitumor agents. 217. Curcumin analogues as novel androgen receptor antagonists with potential as anti-prostate cancer agents

Article abstract of DOI:10.1021/jm020200g

A number of curcumin analogues were prepared and evaluated as potential androgen receptor antagonists against two human prostate cancer cell lines, PC-3 and DU-145, in the presence of androgen receptor (AR) and androgen receptor coactivator, ARA70. Compounds 4 [5-hydroxy-1,7-bis(3,4-dimethoxyphenyl)-1,4,6-heptatrien-3-one], 20 [5-hydroxy-1,7-bis[3-methoxy-4-(methoxycarbonylmethoxy)phenyl]-1, 4,6-heptatrien-3-one], 22 [7-(4-hydroxy-3-methoxyphenyl)-4-[3(4-hydroxy-3-methoxyphenyl) acryloyl]-5-oxohepta-4,6-dienoic acid ethyl ester], 23 [7-(4-hydroxy-3-methoxyphenyl)-4-[3-(4-hydroxy-3-methoxyphenyl)acryloyl] 5-oxohepta-4,6-dienoic acid], and 39 [bis(3,4-dimethoxyphenyl)-1,3-propanedione] showed potent antiandrogenic activities and were superior to hydroxyflutamide, which is the currently available antiandrogen for the treatment of prostate cancer. Structure-activity relationship (SAR) studies indicated that the bis(3,4-dimethoxyphenyl) moieties, the conjugated β-diketone moiety, and the intramolecular symmetry of the molecules seem to be important factors related to antiandrogenic activity. The data further suggest that the coplanarity of the β-diketone moiety and the presence of a strong hydrogen bond donor group were also crucial for the antiandrogenic activity, which is consistent with previous SAR results for hydroxyflutamide analogues. When the pharmacophoric elements of dihydrotestosterone (DHT) and compound 4 are superposed, the resulting construct implies that the curcumin analogues may function as a 17α-substituted DHT. Compounds 4, 20, 22, 23, and 39 have been identified as a new class of antiandrogen agents, and these compounds or their new synthetic analogues could be developed into clinical trial candidates to control androgen receptor-mediated prostate cancer growth.

Full text of DOI:10.1021/jm020200g

J . Med. Chem. 2002, 45, 5037-5042
5037
An titu m or Agen ts. 217.† Cu r cu m in An a logu es a s Novel An d r ogen Recep tor
An ta gon ists w ith P oten tia l a s An ti-P r osta te Ca n cer Agen ts
Hironori Ohtsu,‡ Zhiyan Xiao,‡ J unko Ishida,§ Masahiro Nagai,§ Hui-Kang Wang,‡ Hideji Itokawa,‡
Ching-Yuan Su,| Charles Shih,| Tzuying Chiang, Eugene Chang, YiFen Lee, Meng-Yin Tsai,
Chawnshang Chang, and Kuo-Hsiung Lee‡,*
Natural Products Laboratory, School of Pharmacy, University of North Carolina at Chapel Hill,
Chapel Hill, North Carolina 27599-7360, Hoshi University, 2-4-41 Ebara, Shinagawa, Tokyo 158-0098, J apan,
AndroScience Corporation, 11175 Flintkote Avenue, Suite F, San Diego, California 92121, and George Whipple
Laboratory for Cancer Research, Departments of Pathology, Urology and Radiation Oncology, University of Rochester
Medical Center, 601 Elmwood Avenue, Box 626, Rochester, New York 14642
Received May 10, 2002
A number of curcumin analogues were prepared and evaluated as potential androgen receptor
antagonists against two human prostate cancer cell lines, PC-3 and DU-145, in the presence
of androgen receptor (AR) and androgen receptor coactivator, ARA70. Compounds 4 [5-hydroxy-
1,7-bis(3,4-dimethoxyphenyl)-1,4,6-heptatrien-3-one], 20 [5-hydroxy-1,7-bis[3-methoxy-4-(meth-
oxycarbonylmethoxy)phenyl]-1,4,6-heptatrien-3-one], 22 [7-(4-hydroxy-3-methoxyphenyl)-4-[3-
(4-hydroxy-3-methoxyphenyl)acryloyl]-5-oxohepta-4,6-dienoic acid ethyl ester], 23 [7-(4-hydroxy-
3-methoxyphenyl)-4-[3-(4-hydroxy-3-methoxyphenyl)acryloyl]5-oxohepta-4,6-dienoic acid], and
39 [bis(3,4-dimethoxyphenyl)-1,3-propanedione] showed potent antiandrogenic activities and
were superior to hydroxyflutamide, which is the currently available antiandrogen for the
treatment of prostate cancer. Structure-activity relationship (SAR) studies indicated that the
bis(3,4-dimethoxyphenyl) moieties, the conjugated â-diketone moiety, and the intramolecular
symmetry of the molecules seem to be important factors related to antiandrogenic activity.
The data further suggest that the coplanarity of the â-diketone moiety and the presence of a
strong hydrogen bond donor group were also crucial for the antiandrogenic activity, which is
consistent with previous SAR results for hydroxyflutamide analogues. When the pharmacoph-
oric elements of dihydrotestosterone (DHT) and compound 4 are superposed, the resulting
construct implies that the curcumin analogues may function as a 17R-substituted DHT.
Compounds 4, 20, 22, 23, and 39 have been identified as a new class of antiandrogen agents,
and these compounds or their new synthetic analogues could be developed into clinical trial
candidates to control androgen receptor-mediated prostate cancer growth.
Ch a r t 1. Structures of Cyprosterone,
Hydroxyflutamide, and Bicalutamide
In tr od u ction
The androgen receptor (AR) is a member of a large
family of ligand-dependent transcriptional factors known
as the steroid receptor superfamily.2,3 Androgens and
the AR play an important role in the growth of prostate
cancer and normal prostate. Prostate cancer represents
the most common male malignancy in the United
States.4 Recently, antiandrogens such as hydroxyfluta-
mide (HF) in combination with surgical or medical
castration have been widely used for the treatment of
prostate cancer.5 Both steroidal and nonsteroidal de-
rivatives are presently available and have shown clinical
benefit as chemotherapeutic agents for prostate cancer
(Chart 1). The synthetic steroidal antiandrogen cypros-
terone is one of the first antiandrogens used clinically
in Europe;6 however, this steroidal antiandrogen showed
agonistic activity and overlapping effects with other
hormonal systems, in addition to a range of unpleasant
* To whom correspondence should be addressed. Phone: (919)-962-
0066, Fax: (919)-966-3893. E-mail: khlee@unc.edu.
† Antitumor Agents. 217. For paper 216, see ref 1.
‡ University of North Carolina at Chapel Hill.
§ Hoshi University.
side effects.7 HF and bicalutamide are nonsteroidal
antiandrogens, both of which are thought to be pure
antiandrogens without agonistic activity. Bicalutamide
has a longer half-life (6 days) and a higher binding
affinity to the AR than HF.8,9
| AndroScience Corporation.
University of Rochester Medical Center.
10.1021/jm020200g CCC: $22.00 © 2002 American Chemical Society
Published on Web 10/12/2002

5038 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23
Ohtsu et al.
F igu r e 1. Structures of Curcumin Analogues 1-20.
The agonistic activity of the antiandrogen may result
in “antiandrogen withdrawal syndrome”.10 A currently
accepted hypothesis postulates that mutations in ARs
may account for why HF, the active metabolite of
flutamide, can activate AR target genes and stimulate
prostate cancer growth.10 The same mechanism is used
to explain the “flutamide withdrawal syndrome”, in
which patients who experience an increase in prostate-
specific antigen (PSA) while taking flutamide, have a
decrease in PSA after withdrawal of treatment. Indeed,
HF can activate AR target genes, such as PSA and
MMTV-LTR (a reporter gene which expresses lu-
ciferase), in the presence of ARA70, the first identified
AR coactivator.11 Because this syndrome often leads to
the failure of androgen-ablative therapy, it is desirable
to develop better antiandrogens without agonist activity.
we have prepared a number of curcumin analogues and
evaluated their antagonistic activity against the AR in
the presence of ARA70, using two human prostate
cancer cell lines, PC-3 and DU-145. 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 ARA70.
Ch em istr y
Figures 1 and 2 show the structures of curcumin
analogues and 1,3-diaryl-1,3-diketopropane derivatives.
Curcumin (1), demethoxycurcumin (2), and bisdemethox-
ycurcumin (3) were obtained by column chromatography
(silica gel, CHCl3-MeOH) of commercially available
curcumin (Aldrich), which contained 2 and 3 as minor
components. Treatment of 1 with diazomethane gave
dimethylated curcumin (4) and monomethylated cur-
cumin (9). Methylation of 1 with methyl iodide and K2-
CO3 furnished the trimethylated derivative 10, in which
a methyl group was also introduced at the C-4 position.
Compounds 5-8 were synthesized by heating 1-4 with
histidine hydrazide, AcOH, and p-TsOH overnight.
Hydrogenation of 1 with 10% Pd-C gave a mixture of
11-13. Similarly, compounds 14-16 and 17, 18 were
obtained by hydrogenation of 4 and 10, respectively.
Heating 1 with methyl chloroacetate, NaI, and K2CO3
in acetone furnished a mixture of mono(methoxycarbo-
The phenolic diarylheptanoid curcumin (1) is the
major pigment in turmeric. Curcumin and its analogues
show potent antioxidant activity, antiinflammatory
activity,12 cytotoxicity against tumor cells,13 and anti-
tumor-promoting activities.14,15 In a previous paper, we
reported that two cyclic diarylheptanoids, 13-oxomyri-
canol and myricanone, exhibited potent antitumor-
promoting effects on DMBA-initiated and TPA-induced
mouse skin carcinogenesis.16 Very recently, we have also
evaluated the cytotoxic effects of curcumin analogues
and 1,3-diaryl-1,3-diketopropane derivatives against a
panel of human tumor cell lines.17 In the present study,

Novel Androgen Receptor Antagonists
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23 5039
F igu r e 2. Structures of Curcumin Analogues 21-44.
nylmethyl) ether 19 and bis(methoxycarbonylmethyl)
ether 20, which were separated by preparative TLC
(PLC). Compounds 21-23 were prepared from benzene
or vanillin and ethyl 4-acetyl-5-oxohexanoate as de-
scribed previously in the literature.22 Compounds 21-
23 constitute an unseparable mixture of keto-enol
tautomeric isomers. The syntheses of 24-38 were
described in our previous paper.16,17 Compounds 39-
44 were purchased from Aldrich, Inc. (Milwaukee, WI).
4 (Figure 3A). Thus, the bis(3,4-dimethoxyphenyl) groups
of 4 are important to the activity. Compounds 14 and
15, which were obtained by hydrogenation of 4, were
as potent as HF with a 18% reduction in DHT-induced
AR activity but were considerably less active compared
to 4 (Figure 3A). Introducing a methyl group at C-4 of
4 (10) resulted in decreased activity (Figure 3B). Con-
verting the â-diketone moiety of 4 to the corresponding
pyrazole derivative 8 greatly reduced the activity (data
not shown). Furthermore, 1,3-bis(3,4-dimethoxyphenyl)-
1,3-propandione (39), which contains the bis-aryl groups
found in 4, but lacks the conjugated double bonds, was
less active than 4 (Figure 3A and 3B), indicating that
the conjugated double bonds also contribute to the
activity of 4. These observations suggested that the bis-
(3,4-dimethoxyphenyl) groups and the conjugated â-dike-
tone moiety are crucial for the activity.
Resu lts a n d Discu ssion
Forty-four curcumin derivatives (1-44) were tested
for antagonistic activity against the AR using two
different human prostate cancer cells, PC-3 and DU-
145 (Figures 3A-C). The parental compound, curcumin
(1), was inactive in all cases. However, dimethylated
curcumin (4) showed significant antagonistic activity
(reducing 70% of DHT-induced AR activity) when as-
sayed in PC-3 cells transfected with wild-type AR
(wt.AR) and was more potent than HF (which reduced
16% of DHT-induced AR activity, Figure 3A). Compound
4 also showed the highest antagonist activity when
assayed in DU-145 cells transfected with a mutant
LNCaP AR and ARA70 (showing a 45% reduction in
DHT-induced AR activity, Figure 3B), indicating that
compound 4 is an effective antagonist for both wt.AR
and mutant AR.
Data in Figure 3C shows a somewhat different cell
assay system where antiandrogen activity was assayed
in DU-145 cells transfected with wt.AR and ARA70.
Compounds 4, 20, 22, 23, and 39 showed comparable
or more potent antiandrogen activity than HF in this
assay system. Compounds 20 and 22 were almost
equipotent (54% and 53.8% reduction, respectively) and
were slightly more active than 4 (49.9%). Because
curcumin (1) itself was not active, introducing either
methoxycarbonylmethyl groups at the phenolic hy-
droxyls (20) or an ethoxycarbonylethyl group at C-4 (22)
greatly contributed to the anti-AR activity in DU-145
cells in the presence of wt.AR and ARA70.
To determine the structural requirements for AR
antagonist activity in this series of compounds, an SAR
study was conducted in PC-3 and DU-145 cell assay
systems. Compared with 4, monomethylated curcumin
(9) lacks one O-methyl group at the para position on
one benzene ring and was significantly less active than
In this study, we also examined the antiandrogen
activity of fluorodiarylheptanoids 24-29 and cyclic

5040 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23
Ohtsu et al.
F igu r e 3. Suppression of DHT-mediated MMTV (mouse memory tumor virus) transcription AR activity by hydroxyflutamide
(HF) and selected compounds. PC-3 and DU-145 cell lines were seeded and cotransfected with reporter MMTV-luciferase (A and
C) or mutant AR expression plasmid (B) and ARA70 (B and C) using SuperFect. Subsequently the transfected cells were harvested
and replated in 10% charcoal-stripped fetal bovine serum DMEM (Dulbecco’s Modified Eagle Media) medium. The cells were
then treated with dehydrotestosterone (DHT, 1 nM) and antiandrogens (1 µM) and harvested for detection of the luciferase activity.
The results were averaged from two independent experiments.
diarylheptanoids 30-38. Compounds 24-29 have fluo-
rine or trifluoromethyl substituents on both benzene
rings but showed weak activity or were inactive (data
not shown). Among the cyclic diarylheptanoids 30-38,
compound 30 was the most active and was almost as
active as HF (Figure 3A and 3C). The remaining cyclic
diarylheptanoids showed weak antagonistic activity.
By closely examining the structures and activities of
the curcumin analogues, additional structural features,
besides the bis(3,4-dimethoxyphenyl) groups and the
conjugated â-diketone moiety, seem to be related to the
antiandrogenic activity. The conjugated â-diketone moi-
ety in the most active compounds (4, 20, 22, 23, and
39) can be effectively fixed into a conformation that
ensures coplanarity. However, in compounds 11-18, the
coplanarity of the â-diketone moiety is lost and the
antiandrogenic activity decreases dramatically. This
observation is consistent with the correlation between
the antiandrogenic activity and the coplanarity of the
-NH-CO-OH- moiety in HF analogues.18 The hydro-
gen bond donor ability of the hydroxyl group in the most
active compounds is enhanced by the conjugated system
(this ability will be decreased by the formation of
intramolecular hydrogen bond with the carbonyl group;
however, the conjugated system makes the carbonyl
group itself a very weak hydrogen bond acceptor), but
certain compounds (e.g., 30-38) lacking such a hydroxyl
group are significantly less active. The presence of a
strong hydrogen bond donor group is also consistent
with the previous SAR for HF analogues.18 In addition,
intramolecular symmetry seems to be another impor-
tant factor related to antiandrogenic activity.
F igu r e 4. Superposition of DHT and compound 4. DHT and
compound 4 are represented in white and red, respectively.
Both molecules were constructed and optimized with SYBYL
6.0. Compound 4 was subjected to a genetic algorithm confor-
mational search. The lowest-energy conformer of 4 was
superposed to DHT by the six-membered rings (A-ring of DHT
and phenyl of 4) and the hydroxyl groups and was then field-
fit minimized using DHT as template.
DHT, the naturally occurring ligand of AR, with com-
pound 4 (Figure 4) by aligning the six-membered rings
(A-ring of DHT and a phenyl ring of 4) and the hydroxyl
groups. The resulting superposition suggests that the
curcumin analogues may assume a conformation similar
to that of 17R-substituted DHT in order to exert their
antiandrogenic activity. The remaining structural ele-
ments of compound 4 extend into the D-ring C17
vicinity, which coincides with a sterically favorable area
defined by a previous three-dimensional quantitative
structure-activity relationship (3D-QSAR) study using
the comparative molecular field analysis (CoMFA)
technique.20 The consistency between the SAR of cur-
cumin analogues and known antiandrogens implies that
this new class of compounds may have a ligand-
receptor recognition pattern and interaction mode simi-
lar to that of known antiandrogens. This knowledge
would facilitate future mechanic investigations of cur-
cumin analogues.
It was previously suggested that the steroid A-ring
is a pharmacophoric component necessary for molecular
recognition between steroid receptors and the ligands,
while the remaining steroid structure confers receptor
selectivity/specificity.19,20 Additionally, the “17â-hydroxyl-
3-one” substructures have been considered essential for
effective binding of steroids to AR.21 Using these pre-
defined pharmacophoric elements, we have superposed

Novel Androgen Receptor Antagonists
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23 5041
Tetr a h yd r ocu r cu m in (11). White powder, mp 92-93 °C
(lit.23 95-96 °C), 1H NMR (300 MHz, CDCl3): δ 2.53-2.58 (3H,
m), 2.78-2.88 (5H, m), 3.87 (6H, s, OCH3 × 2), 5.43 (1H, s,
1-H), 5.50 (2H, s, ArOH), 6.65 (2H, d, J ) 8 Hz), 6.69 (2H, s),
6.83 (2H, d, J ) 8 Hz); 13C NMR (300 MHz, CDCl3): δ 31.3,
40.4, 55.8, 99.8, 111.0, 114.3, 120.8, 132.6, 144.0, 146.4, 193.2.
Hexa h yd r ocu r cu m in (12). White powder, mp 87-88 °C
(lit.23 78-80 °C), 1H NMR (300 MHz, CDCl3): δ 1.60-1.81 (2H,
m), 2.53-2.97 (8H, m), 3.85 (6H, s, OCH3 × 2), 4.06 (1H, m,
2-H), 6.70 (4H, m), 6.80 (2H, d, J ) 8 Hz); 13C NMR (300 MHz,
CDCl3): δ 29.7, 31.7, 38.8, 45.8, 49.8, 56.3, 67.4, 111.5, 111.6,
114.8, 114.9, 121.2, 121.4, 133.0, 134.2, 144.2, 144.5, 146.9,
147.9, 211.9.
Octa h yd r ocu r cu m in (13). Colorless oil, 1H NMR (300
MHz, CDCl3): δ 1.61 (2H, m), 1.75 (4H, m), 2.53-2.70 (4H,
m), 3.80 (6H, s, OCH3 × 2), 3.91 (2H, brs), 6.13 (2H, s, ArOH),
6.65 (2H, d, J ) 8 Hz), 6.69 (2H, bs) 6.82 (2H, bd, J ) 8 Hz),
13C NMR (300 MHz, acetone-d6): δ 31.1, 39.8, 42.6, 35.6, 72.0,
111.0, 114.3, 120.6, 133.6, 143.6, 146.4.
Con clu sion
In conclusion, we have prepared a number of cur-
cumin analogues and evaluated their potential antian-
drogen activity in three different assay conditions using
human prostate cancer cell lines. Compounds 4 showed
promising antiandrogen activities in all assays. Com-
pounds 4, 20, 22, 23, and 39 have been identified as a
new class of antiandrogen agents. SAR studies revealed
that bis(3,4-dimethoxyphenyl) moieties, a conjugated
â-diketone, and an ethoxycarbonylethyl group at the C-4
position play important roles in the antagonistic activ-
ity. Further mechanistic studies and the development
of new antiandrogens are actively underway in our
laboratory.
Exp er im en ta l Section
Com p ou n d 14. White powder (yield 26.0%), mp 60-61 °C,
1H NMR (300 MHz, CDCl3): δ 2.56 (3H, m), 2.86 (5H, m), 3.85
(12H, s, OCH3 × 4), 5.44 (1H, s, 1-H), 6.71 (4H, m), 6.78 (2H,
bd); Anal. Calcd (theoretical) for C23H28O6‚1/4H2O: C, 68.21;
H, 7.09. Found: C, 68.25; H, 7.06.
Melting points were determined on a Fisher-J ohns melting
apparatus and are uncorrected. Optical rotations were deter-
mined with a DIP-1000 polarimeter. 1H NMR spectra were
recorded on a Bruker AC-300 and J MN GX-500 spectrometer.
The chemical shifts are presented in terms of ppm with TMS
as the internal reference. MS spectra were recorded on
HP5989A and J MS D-300 instruments. Elemental analyses
were performed by Atlantic Microlab Inc., Norcross, GA, and
agreed with theoretical values to within (0.4%. New com-
pounds 9, 16, and 18 were homogeneous by HPLC analyses
in three different solvent systems. Compounds 1-3 were
obtained by column chromatography (silica gel, CHCl3-MeOH)
of commercially available curcumin (Aldrich), which contained
2 and 3 as minor components. Compounds 39-44 were
purchased from Aldrich, Inc (Milwaukee, WI).
Dim eth ylcu r cu m in (4). Curcumin (1) in Et2O and MeOH
was treated with excess of diazomethane in ether for 24 h.
The solvents were removed in vacuo, and the residue was
purified by silica gel column chromatography and PLC to yield
yellow needles of 4 (yield 19.8%); mp 129-130 °C (MeOH) (lit.23
128-130 °C); 1H NMR (300 MHz, CDCl3): δ 3.93 (12H, s,
OCH3 × 4), 5.82 (1H, s, 1-H), 6.48 (2H, d, 16 Hz), 6.88 (2H, d,
J ) 8 Hz), 7.08 (2H, bs), 7.15 (2H, bd), 7.61 (2H, J ) 16 Hz);
13CNMR (300 MHz, CDCl3): δ 55.9, 56.0, 101.3, 109.8, 111.1,
122.0, 122.6, 128.1, 140.4, 149.2, 151.0, 183.2.
Com p ou n d 15. White powder (yield 20.0%), mp 94-95 °C,
1H NMR (300 MHz, CDCl3): δ 1.65-1.80 (2H, m), 2.53-2.84
(8H, m), 3.85 (12H, s, OCH3 × 4), 4.05(1H, bs, 2′-H), 6.68-
7.23 (4H, m), 6.79 (2H, bd), Anal. Calcd (theoretical) for
C
23H30O6‚1/4H2O: C, 67.88; H, 7.55. Found: C, 67.73; H, 7.49.
Com p ou n d 16. Colorless oil (yield 4.2%), mp 60-61 °C, δ
1H NMR (300 MHz, CDCl3): δ 1.55-1.65 (4H, m), 1.73-1.82
(3H, m), 2.60-2.72 (3H, m), 3.86 (6H, s, OCH3 × 2), 3.87 (8H,
bs, OCH3 × 2, 2,2′-H), 6.72-6.78 (4H, m), 6.79 (2H, bd), 7.27-
(2H, s, OH × 2), EIMS m/z: 404 (M+), HRFAB-MS m/z
404.219070 (M + H)+ (calcd for C23H32O6: 404.2198891).
Com p ou n d 17. Colorless oil (yield 5.9%),1H NMR (300
MHz, CDCl3): δ 1.10 (3H, d), 1.80 (1H, m), 2.43-2.82 (8H,
m), 3.86 (6H, s, OCH3 × 2), 3.87 (6H, s, OCH3 × 2), 3.94 (1H,
bs, 2′-H),6.70-6.78 (6H, m), EIMS m/z 416 (M+).
Com p ou n d 18. Colorless oil (yield 6.95%), 1H NMR (300
MHz, CDCl3): δ 0.95 (3H, d, 1-CH3), 1.52 (1H, m), 1.84 (2H,
m), 2.67 (6H, m), 3.83 (14H, bs, OCH3 × 4, 2, 2′-H), 6.78 (6H,
m); EIMS m/z: 418 (M+), HRFAB-MS m/z 418.236618 (M +
H)+ (calcd for C24H34O6: 418.2355392).
P r ep a r a tion of P yr a zole Der iva tive 8. To a solution of
1-4 in butanol and ethanol were added histidine hydrazide
(1 equiv), acetic acid, and p-TsOH. The solution was refluxed
for 24 h, and then the solvent was removed in vacuo. The
residue was purified by silica gel column chromatography and
PLC.
Com p ou n d 8. Yellow powder (yield 17.5%), mp 166-168
°C (MeOH); 1H NMR (300 MHz, CDCl3): δ 3.92 (6H, s, OCH3
× 2), 3.94 (6H, s, OCH3 × 2), 6.62 (1H, s, 1-H), 6.86 (2H, d, J
) 8 Hz), 6.93 (2H, d, J ) 16 Hz), 7.04 (2H, dd, J ) 8, 2 Hz),
7.06 (2H, bs), 7.05 (2H, d, J ) 16 Hz); 13CNMR (300 MHz,
CDCl3): δ 55.8, 55.9, 99.6, 108.6, 111.2, 115.8, 120.1, 129.7,
130.6, 149.1, 149.3; Anal. Calcd (theoretical) for C23H24N2O4‚
5/4H2O: C, 66.57; H, 6.44; N, 6.75. Found: C, 66.44; H, 6.19;
N, 6.27.
Mon om eth ylcu r cu m in (9). Curcumin (1) in MeOH was
treated with excess diazomethane in Et2O for 24 h. After
removal of solvents, the residue was purified by silica gel
column chromatography and PLC to yield a yellow amorphous
solid (yield 20%); mp 89-91°C, [R]D -3.6° (c ) 1.14, CHCl3);
1H NMR (300 MHz, CDCl3): δ 3.93 (9H, s, OCH3 × 3) , 5.81
(1H, s, 1-H), 5.94 (1H, bs, OH), 6.49 (2H, bd, J ) 15 Hz), 6.93
(1H, d, J ) 8 Hz), 6.97 (1H, d, J ) 8 Hz), 7.10 (4H, m), 7.60
(2H, bd, J ) 15 Hz); EIMS m/z 382 (M+), HRFABMS 382.1396
(M + H+) (calcd for C22H22O6: 382.1416).
Hyd r ogen a tion of 1, 4, a n d 10 (11-18). A solution of
starting material in EtOAc was shaken with 10% Pd-C under
H2 (45 psi) overnight using a Parr’s apparatus. The solution
was filtered and concentrated in vacuo to give a residue, which
was purified by silica gel column chromatography and PLC.
P r ep a r a tion of 19 a n d 20. A mixture of curcumin (1, 100
mg, 0.81 mmol) in acetone (20 mL) with methyl chloroacetate
(2 mL) and NaI (20 mg) was refluxed with anhydrous potas-
sium carbonate (176 mg) for 24 h with stirring. After filtration
and removal of solvent, the residue was purified by silica gel
column chromatography to yield the corresponding methyl
acetates 19 and 20.
Com p ou n d 19: Yellow powder (yield 20.0%), mp 60-61 °C,
mp 66-67 °C, [R]D -2.4° (c ) 2.08, CHCl3); 1H NMR (300 MHz,
acetone-d6): δ 3.73 (3H, s, COOCH3), 3.86 (6H, s, OCH3 × 2),
4.79 (2H, s, OCH2COO), 5.99 (1H, s, 1-H), 6.70 and 6.73 (both
1H, d, J ) 15.3 Hz), 6.88 (1H, d, J ) 8 Hz), 6.94 (1H, d, J )
8 Hz), 7.17 (2H, m), 7.33 (2H, m), 7.59 and 7.61 (both 1H, d,
J ) 15.3 Hz), 13C NMR (300 MHz, CDCl3): d 51.8, 55.9, 55.9,
65.9, 101.4, 111.2, 111.6, 114.3, 115.9, 121.8, 122.6, 123.0,
123.5, 127.6, 128.7, 129.8, 140.3, 141.3, 148.4, 149.8, 150.0,
150.4, 169.4, 183.4, 184.6; Anal. Calcd (theoretical) for C24H24O8‚
3/4H2O: C, 63.50; H, 5.66. Found: C, 63.53; H, 5.65.
Com p ou n d 20: Yellow powder (yield 20.0%), mp 141-142
°C (MeOH), [R]D -0.29° (c ) 5.86, CHCl3); 1H NMR (300 MHz,
CDCl3): δ 3.80 (6H, s), 3.93 (6H, s), 4.73 (4H, s, OCH2COO ×
2), 5.82 (1H, s, 1-H), 6.50 (2H, d, J ) 16 Hz), 6.79 (2H, d, J )
8 Hz), 7.09 (4H, bs), 7.58 (2H, d, J ) 16 Hz), 13C NMR (300
MHz, CDCl3): δ 52.3, 56.0, 66.0, 101.4, 110.7, 113.6, 122.0,
122.7, 129.5, 140.1, 149.0, 149.7, 169.0, 183.1; Anal. Calcd
(theoretical) for C27H28O10‚1/2H2O: C, 62.18; H, 5.60. Found:
C, 62.31; H, 5.57.
Com p ou n d 21: Yellow amorphous solid (yield 3.0%), 1H
NMR (300 MHz, CDCl3): δ 2.58 (2H, m), 2.95 (2H, m), 7.12
(2H, d, J ) 15 Hz), 7.40 (6H, m), 7.60 (4H, m), 7.81(2H, d, J

5042 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 23
15 Hz), 12.65(1H, bs); Anal. Calcd (theoretical) for
Ohtsu et al.
(6) McLeod, D. G. Antiandrogenic Drugs. Cancer 1993, 71, 1046-
1049.
(7) Neumann, F.; J acobi, G. H. Antiandrogens in Tumor Therapy.
J . Clin. Oncol. 1982, 1, 41-65.
)
C
22H20O4: C, 75.84, H, 5.79. Found: C, 75.56, H, 5.74.
Com p ou n d 22: Yellow amorphous solid (yield 25.0%),
Anal. Calcd (theoretical) for C26H28O8: C, 66.66, H, 6.02.
Found: C, 66.38, H, 6.16.
Com p ou n d 23: Yellow powder (yield 45.0%), mp 144-146
°C (MeOH) [lit.22 71-73 °C (CH2Cl2)]; Anal. Calcd (theoretical)
for C24H26O8‚5/2H2O: C, 59.87; H, 5.23. Found C, 59.94; H, 5.11.
The structures of 1-4, 10-13, 22, and 23 were confirmed
by comparison of their physical spectral data with those
reported in the literature.22,23
(8) Verhelst, J .; Denis, L.; Vliet, V.; Poppel, H. V.; Braeckman, J .;
Cangh, P. V.; Mattelaer, J .; D′Hulster, D.; Mahler, C. Endocrine
Profile during Administration of the New Nonsteroidal Anti-
androgen Casodex in Prostate Cancer. Clin. Endocrinol. 1994,
41, 525-530.
(9) Kelly, W. K.; Scher, H. I. Prostate Specific Antigen Decline after
Antiandrogen Withdrawal: the Flutamide Withdrawal Syn-
drome. J . Urol. 1993, 149, 607-609. (b) Scher, H. I.; Kelly, W.
K. Flutamide Withdrawal Syndrome: Its Impact on Clinical
Trials in Hormone-refractory Prostate Cancer. J . Clin. Oncol.
1993, 11, 1566-1572.
(10) Miyamoto, H.; Yeh, S.; Wilding, G.; Chang, C. Promotion of
Agonist Activity of Antiandrogens by the Androgen Receptor
Coactivator, ARA70, in Human Prostate Cancer DU145 Cells.
Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 7379-7384.
(11) Yeh, S.; Miyamoto, H.; Chang, C. Hydroxyflutamide May Not
Always Be a Pure Antiandrogen. The Lancet 1997, 349, 852-
853.
(12) Nurfina, A. N.; Reksohadiprodjo, M. S.; Timmerman, H.; J enie,
U. A.; Sugiyanto, D.; van der Goot, H. Synthesis of Some
Symmetrical Curcumin Derivarives and their Antiinflammatory
Activity. Eur. J . Med. Chem. 1997, 32, 321-328.
(13) Syu, W., J r.; Shen, C. C.; Don, M. J .; Ou, J . C.; Lee, G. H.; Sun,
C. M. Cytotoxicity of Curcuminoids and Some Novel Compounds
from Curcuma zedoaria. J . Nat. Prod. 1998, 61, 1531-1534.
(14) Sugiyama, Y.; Kawakishi, S.; Osawa, T. Involvement of the
â-Diketone Moiety in the Antioxidative Mechanism of Tetrahy-
drocurcumin. Biochem. Pharmacol. 1996, 52, 519-525.
(15) Ruby, A. J .; Kuttan, G.; Babu, K. D.; Rajasekharan, K. N.;
Kettan, R. Anti-tumor and Antioxidant activity of Natural
Curcuminoids. Cancer Lett. 1995, 94, 79-83.
(16) Ishida, J .; Kozuka, M.; Wang, H.-K.; Konoshima, T.; Tokuda,
H.; Okuda. M.; Mou, X.-Y.; Nishino, H.; Sakurai, N.; Lee, K.-H.;
Nagai, M. Antitumor-promoting Effects of Cyclic Diarylhep-
tanoids on Epstein-Barr Virus Activation and Two-Stage Mouse
Skin Carcinogenesis. Cancer Lett., 2000, 159. 135-140.
(17) Ishida, J .; Ohtsu, H.; Tachibana, Y.; Nakanishi, Y.; Bastow, K.
F.; Nagai, M.; Wang, H. K.; Itokawa, H.; Lee, K. H. Synthesis
and Evaluation of Curcumin Analogues as Cytotoxic Agents.
Unpublished data.
(18) Morris, J . J .; Hughes, L. R.; Glen, A. T. and Taylor, P. J . Non-
Steroidal Antiandrogens. Design of Novel Compounds Based on
an Infrared Study of the Dominant Conformation and Hydrogen-
Bonding Properties of a Series of Anilide Antiandrogens. J . Med.
Chem. 1991, 34, 447-455.
(19) McKinney, J . D. and Waller, C. L. Polychlorinated Biphenyls
as Hormonally Active Structural Analogues. Environ. Health.
Perspect. 1994, 102, 290-297.
Su p p r ession of DHT-Med ia ted Tr a n scr ip tion Activity.
Cell Cu ltu r e a n d Tr a n sfection s. Human prostate cancer
DU145 and PC-3 cells were maintained in Dulbecco’s mini-
mum essential medium (DMEM) containing penicillin (25
units/mL), streptomycin (25 µg/mL), and 10% fetal calf serum
(FCS). For AR transactivation assay, PC-3 cells were trans-
fected with an AR expression plasmid and reporter gene.
Because of a low content of endogenous AR coactivators, DU-
145 cells were transfected with expression plasmids for AR
and ARA70, and reporter gene. The previously described
conditions were followed with minor modifications.10 Trans-
fections were performed using the SuperFect kit according to
manufacturer’s procedures (Qiagen, Chatsworth, CA). Briefly,
1 × 105 cells were plated on 35-mm dishes 24 h before
transfection, and then a reporter plasmid, MMTV-Luciferase,
which contains MMTV-LTR promoter and AR-binding ele-
ment, was cotransfected with AR expression plasmid (wild type
or mutant), or pSG5ARA70. PRL-TK was used as an internal
control for transfection efficiency. The total amount of DNA
was adjusted to 3.0 µg with pSG5 in all transcriptional
activation assays. After a 2 h transfection, the medium was
changed to DMEM-10% charcoal stripped serum medium, and
14-16 h later, the cells were treated with DHT, antiandrogen,
or test compounds. After another 14-16 h, the cells were
harvested and tested for luciferase activity in luciferase assays
(Promega, Dual Luciferase Assay System, Madison, WI). Data
were expressed as relative luciferase activity as compared to
an internal luciferase positive control.
Ack n ow led gm en t. This work was supported by
National Cancer Institute Grant CA-17625 awarded to
K. H. Lee and DK-60905 to C. Chang.
Refer en ces
(1) For the previous paper in this series, see Ohtsu, H.; Nakanishi,
Y.; Bastow, K. F.; Lee, K. H. Antitumor Agents 216. Synthesis
and Evaluation of Paclitaxel-Camptothecin Conjugates as Novel
Cytotoxic Agents. Bioorg. Med. Chem., in press.
(2) Chang, C.; Kokontis, J .; Liao, S. Structural Analysis of Comple-
mentary DNA and Amino acid Sequences of Human and Rat
Androgen Receptors. Proc. Natl. Acad. Sci. U.S.A. 1988, 85,
7211-7215.
(3) Beato, M. Gene Regulation by Steroid Hormones. Cell 1989, 56,
335-344.
(4) Landis, S. H.; Murray, T.; Bolden, S.; Wingo, P. A. Cancer
Statistics 1998. Cancer J . Clin. 1998, 48, 6-29.
(5) 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 . A Controlled Trial of Leuprolide with and
without Flutamide in Prostate Carcinoma. New Engl. J . Med.
1989, 321, 419-424.
(20) Waller, C. L.; J uma, B. W.; Gray. L. E., J r.; Kelce, W. R. Three-
Dimensional Quantitative Structure-Activity Relationships for
Androgen Receptor Ligands. Toxicol. Appl. Pharmacol. 1996,
137, 219-227.
(21) Singh, S. M.; Gauthier S.; Fernand L. Androgen Receptor
Antagonists (Antiandrogens): Structure-Activity Relationships.
Curr. Med. Chem. 2000, 7, 211-247.
(22) Pedersen, U.; Rasmussen, P. B.; Lawesson, S. O. Synthesis of
Naturally Occurring Curcuminoids and Related Compounds.
Liebigs Ann. Chem. 1985, 1557-1569.
(23) Roughley, P. J .; Whiting, D. A. Experiments in the Biosynthesis
of Curcumin. J . Chem. Soc., Perkin Trans. 1 1973, 2379-2388.
J M020200G