Novel non-steroidal/non-anilide type androgen antagonists with an isoxazolone moiety

Article abstract of DOI:10.1016/S0968-0896(01)00421-7

3-Substituted (Z)-4-(4-N,N-dialkylaminophenylmethylene)-5(4H)-isoxazolones and related compounds were designed and prepared as candidates for structurally novel androgen antagonists. Several compounds showed potent anti-androgenic activity as assessed by nuclear androgen receptor binding assay and growth inhibition assay using androgen-dependent Shionogi carcinoma cells SC-3. They were approximately 10-220 times more potent than flutamide in these assay systems. They also showed anti-androgenic activity toward prostate tumor cell line LNCaP, which has an aberrant nuclear androgen receptor.

Full text of DOI:10.1016/S0968-0896(01)00421-7

Bioorganic & Medicinal Chemistry 10 (2002) 1555–1566
Novel Non-Steroidal/Non-Anilide Type Androgen
Antagonists with an Isoxazolone Moiety
Toshiyasu Ishioka,a Asako Kubo,b Yukiko Koiso,a Kazuo Nagasawa,a
Akiko Itaib and Yuichi Hashimotoa,*
aInstitute of Molecular and Cellular Biosciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo, 113-0032, Japan
bInstitute of Medicinal Molecular Design, Key Molecular Inc., 4-24-5 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
Received 30 October 2001; accepted 26 November 2001
Abstract—3-Substituted (Z)-4-(4-N,N-dialkylaminophenylmethylene)-5(4H)-isoxazolones and related compounds were designed
and prepared as candidates for structurally novel androgen antagonists. Several compounds showed potent anti-androgenic activity
as assessed by nuclear androgen receptor binding assay and growth inhibition assay using androgen-dependent Shionogi carcinoma
cells SC-3. They were approximately 10–220 times more potent than flutamide in these assay systems. They also showed anti-
androgenic activity toward prostate tumor cell line LNCaP, which has an aberrant nuclear androgen receptor. # 2002 Elsevier
Science Ltd. All rights reserved.
Introduction
ligand binding pocket, being in the closed conformation
when an agonist occupies the binding pocket and in the
Androgens, typically testosterone, are endogenous
ligands for the nuclear androgen receptor (AR) and are
essential for the development and maintenance of the
male reproductive system and secondary male sex char-
acteristics.1 Androgens play diverse physiological and
pathophysiological roles in both males and females.1,2
Among the pathophysiological effects elicited by
androgens, a role as endogenous tumor promoters,
especially for prostate tumor, is well known. The effect
is considered to be mediated by androgen-binding to
AR. The AR is a member of the steroid/thyroid/reti-
noid/vitamin D3 nuclear receptor superfamily and acts
as an androgen-dependent specific transcription fac-
tor.3,4 The nuclear receptor consists of three main func-
tional domains including the ligand-binding domain
(LBD or E-region), DNA-binding domain (DBD or C-
region) and amino-terminal domain (A/B region).4ꢀ6
The general structure of the LBD has been elucidated to
be composed of 12 a helices and a small b sheet by X-
ray crystallography of a number of different nuclear
receptors. Among the substructures, the helix H12 plays
a critical role in the ligand-dependent activation of the
receptor, that is, H12 functions as a lid covering the
open conformation without the agonist or when an
antagonist occupies the pocket.7,8 Recently, the crystal
structure of LBD of human AR complexed with an
agonist, which coincides with the general structure of
LBD (vide supra), was reported.9
Androgen antagonists are compounds which antagonize
the biological responses induced by endogenous or exo-
genous androgens, by inhibiting competitively their
binding to AR. Because of the known action of andro-
gens as tumor promoters, androgen antagonists are
expected to be effective for treatment of these androgen-
dependent tumors, especially for prostate tumor.2 There
are two structural types of androgen antagonists which
have been used in therapeutics, that is, steroidal and
anilide types (Fig. 1). A typical steroidal androgen
antagonist is cyproterone acetate (CPA: 1), which is the
first drug to have been used orally, but CPA also inter-
acts with progestin and glucocorticoid receptors.10,11
Anilide analogues were the first group of non-steroidal
androgen antagonists to be discovered, and three of
them have been used therapeutically for the treatment
of prostate tumor so far.2,12 Flutamide (2) was the first
non-steroidal androgen antagonist to enter clinical use,
but its active metabolite, 2-hydroxyflutamide (3) is
cleared so rapidly that it requires frequent dosing (three
times a day).2,13 A cyclic derivative of 2, nilutamide (4),
*Corresponding author. Tel.: +81-3-5841-7847; fax: +81-3-5841-
8495; e-mail: hashimot@iam.u-tokyo.ac.jp
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00421-7

1556
T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
can be administered once a day, but it elicits side effects
including hot flushes, transient night blindness and
alcohol intolerance.2,14 Bicalutamide (5), which is mar-
keted as a recemic mixture with the R-isomer being the
eutomer, is superior to both flutamide (2) and niluta-
mide (4) in terms of its pharmacokinetics and side effect
profile.2,15 Discovery of these non-steroidal androgen
antagonists stimulated researchers to develop other
structural types of androgen antagonists.
esis, and structure–activity relationship study of another
structurally novel type of androgen antagonists, 3-sub-
stituted (Z)-4-(4-N,N-dialkylaminophenylmethylene)-
5(4H)-isoxazolones.
Results and Discussion
Molecular design
Recently, we reported phthalimide-type androgen
antagonists, including compounds 6–10, (Fig. 2).16ꢀ19
These androgen antagonists were derived from DIMP
(11)20 and thalidomide, the former of which is a well
known androgen antagonist, though not used clin-
ically.16ꢀ19 An anilide analogue 12, which was derived
from tetrafluorophthalimide analogues 6, 7, 9 and 10,
also showed anti-androgenic activity.19 We also repor-
ted azo/azoxy-benzene types of androgen antagonists,
including compounds 13 and 14.19,20 Hamann et al.
reported linear tricyclic quinolines, with the generic
structure 15 as a novel androgen antagonist scaf-
fold.2,21,22 In this article, we describe the design, synth-
Androgen antagonists which we wish to develop are
ligands which bind the LBD of AR in a competitive
manner with androgens. The X-ray crystallographic
structure of the LBD of AR,9 as well as our previous
studies on the structural development of retinoids whose
target molecule is also a member of the nuclear receptor
superfamily,5,6 suggested that compounds with two
aromatic groups connected with two to four carbon unit
mimics as a spacer are candidate ligands for AR.19
Roughly speaking, our androgen antagonists 6–9 and
12–14, and a classical androgen antagonist, DIMP (11),
fall into this structural category. Because all members of
the nuclear receptor superfamily are assumed to have
Figure 1. Structures of androgen antagonists used in therapeutics (1–5). [Values in parentheses are the RBA values (see text). Values in italics are
reported by other researchers.]
Figure 2. Structures of non-steroidal and non-anilide types of androgen antagonists (6–15). [A value in parentheses is the RBA value (see text).]

T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
1557
evolved from a single gene and to possess a similarly
shaped ligand-binding pocket,7,23 it might be expected
that a ligand superfamily also exists.8 From such a
standpoint, ligands for nuclear receptors should be
roughly structurally superimposable, and there should
exist a general antagonistic substituent group for
nuclear receptor ligands.8,19 Such a general antagonistic
group, if it exists, is expected to hinder conformational
change (folding of the helix H12) of the receptor to the
active form without reducing the binding affinity of the
ligand to which it is attached. Our previous studies on
the development of antagonists of nuclear receptor
ligands suggested that a dialkylamino or a dialkylam-
inophenyl group might be such a general antagonistic
group, as found in the well known estrogen antagonist,
tamoxifen, and other steroid antagonists.8,19 On this
basis, we designed a generic structure 16 as a possible
androgen antagonist skeleton (Fig. 3). The structure 16
was constructed through the following steps: (i) extrac-
tion of an oxazine substructure from DIMP (11), (ii)
usage of the oxazine substructure as a spacer linking
two aromatic groups, (iii) modification of the oxazine
moiety to a more rigid isoxazolone or a pyrazolone
substructure, and (iv) introduction of a dialkylamino
group as a possible antagonistic group.
ponent condensation reaction with benzaldehyde (pre-
pared by Vilsmeier reaction)24 or acetophenone
derivatives, b-keto esters, and hydroxylamine or hydra-
zines in the presence of phosphoric acid according to the
procedures reported by Donleavy and Gilbert.25 The
isoxazolone 24 was synthesized with hydroxylamine and
b-keto ester, which was prepared from the condensation
reaction of N-phenylpyrrolidine and ethyl acetoacetate
with formaldehyde. All the compounds were character-
ized by 1H NMR, mass spectroscopy and elemental
analysis or high-resolution mass spectroscopy. The ste-
reochemistries of olefins were determined by means of
NOESY experiments (NOE was observed between the
methyl at C3 and the vinyl proton at C5 positions).
Biological assay
Androgenic and anti-androgenic activities of the com-
pounds were assessed by using a combination of two
assay systems, that is receptor-binding assay and growth
promotion/inhibition assay using androgen-dependent
Shionogi carcinoma cells SC-3.15,26,27 The ability of the
compounds to be ligands for AR can be assessed by the
former assay. The latter assay is to determine whether
the compounds are androgen agonists or antagonists.
None of the compounds prepared showed growth-pro-
moting activity toward SC-3 cells, suggesting that none
was an androgen agonist. Therefore, only growth-inhi-
bitory activities of the compounds (IC50 values) toward
testosterone-promoted growth of SC-3 cells are pre-
sented in this paper (Table 1). Briefly, SC-3 cells were
incubated under usual conditions with 10 nM testoster-
one in the presence of various concentrations of a test
compound. The increase of cell number in the absence
of test compound was defined as 100%, and relative/
expedient anti-androgenic activity of each test com-
pound was presented as IC50 value, that is, the con-
centration at which the testosterone-promoted increase
of cell number was reduced to 50%.
Data supporting the idea that a structure related to 16
can be a ligand for AR were obtained by computer-
assisted molecular design. A number of compounds
were identified as possible ligands for AR by computer-
assisted docking search of a database of commercially
available compounds (unpublished results). Among the
candidate compounds presented by the computer, there
was
(Z)-4-(4-N,N-diethylaminophenylmethylene)-3-
methyl-5(4H)-isoxazolone (17). Assays for androgenic
and anti-androgenic activity of 17 indicated that 17 is an
androgen antagonist with moderate activity (vide infra).
Therefore, we sought to develop potent androgen
antagonists based on the structures of 16 and 17.
For receptor-binding assay, we adopted competitive
binding assay using [3H]testosterone and recombinant
human AR. The recombinant AR was prepared from
cytosol of Escherichia coli transformed with a human
AR LBD expression vector (GST-hARLBD) which
Synthesis
Compounds 18–50 were prepared by the method shown
in Figure 4. The isoxazolones (18–23, 25 and 30–50) and
the pyrazolones (26–29) were synthesized by three-com-
Figure 3. Molecular design of generic structure of androgen antagonists (16) and structure of a candidate compound presented by the computer
(17).

1558
T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
codes amino acids 627–919 of hAR fused with GST
protein under the lac promoter (see Experi-
mental).19,28,29 The binding activity of the GST-
hARLBD thus prepared toward [3H]testosterone, that
is, the Kd (dissociation constant) value, was determined
to be 1.4 nM by Scatchard analysis, and this value is
close to that of AR prepared from intact SC-3 cells. The
binding activity of a test compound was determined by
incubation of GST-hARLBD with [3H]testosterone in
the presence of various concentrations of the test
compound.30ꢀ32 The concentration of a test compound
that inhibits [3H]testosterone-binding by 50% (IC50
value) was determined after log-logit transformation,
and the Ki value was calculated based on the Kd of
[3H]testosterone. The Ki value of flutamide (2) thus
obtained was 1.3 mM. The receptor-binding activity of
each test compound is presented as relative binding
affinity (RBA), which was defined as (Ki value of fluta-
mide)/(Ki value of test compound) (Figs 5–8).
Table 1. Growth-inhibitory activities toward SC-3 and LNCaP cells
For compounds which showed potent anti-androgenic
activity, growth inhibition assay using human prostate
tumor cells LNCaP was also performed. LNCaP cells
are known to possess an aberrant AR with a mutation
of 877Thr to Ala, and some androgen antagonists
including flutamide (2) show androgen-agonistic activity
toward LNCaP cells.29 The data are presented as IC50
values measured by the same method used in the case of
SC-3 assay (vide supra) (Table 1).
Compd
IC50 value (mM)
LNCaP cells
SC-3 cells
2: flutamide
3: hydroxyflutamide
2.7
0.8
0.6
3.0
1.6
1.3
0.4
3.5
5.8
1.9
6.2
1.6
2.9
1.4
1.0
0.7
1.1
0.3
0.2
0.4
0.3
0.4
0.3
0.5
2.9
6.1
4.0
1.7
3.8
>10a
>10a
9.7
>10 (47%)b
9.0
>10 (32%)b
>10 (38%)b
>10 (24%)b
9.5
n.t.
>10 (7%)b
n.t.
17
19
20
21
22
23
26
28
29
30
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
The values (IC50 values in SC-3 and LNCaP assays, and
RBA values in receptor-binding assay) thus measured
showed some deviation from experiment to experiment.
However, the results were basically reproducible. Each
experiment was performed in triplicate and repeated at
least three times, and typical sets of data (mean value of
the triplicate) are presented in this paper.
>10 (34%)
>(37%)b
5.6
>10 (28%)b
>10 (15%)b
7.9
>10 (18%)b
>10 (38%)b
9.8
Structure–activity relationship studies (1): (Z)-4-(4-di-
alkylaminophenylmethylene)-3-methyl-5(4H)-isoxazolones
and their related compounds (17–33) (Fig. 5). As already
mentioned, computer-assisted docking search of a com-
mercially available compound database yielded 17 as a
candidate ligand. Biological assays indicated that 17 is
an androgen antagonist with moderate activity. There-
fore, we started our study by modification of the struc-
ture of 17. Structural development was mainly based on
receptor (AR)-binding activity, because this reflects
direct (intrinsic) activity of the compound, excluding
>10 (22%)b
>10 (28%)b
>10 (2%)b
>10 (24%)b
2.5
2.2
1.9
2.1
aKnown to act as an androgen agonist toward LNCaP cells.
bPercent inhibition of testosterone (10 nM)-induced increase of cell
growth at 10 mM. n.t., not tested.
Figure 4. Synthetic methods for isoxazolones (18–25, and 30–50) and pyrazolones (26–29).

T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
1559
Figure 5. Structures of (Z)-4-(4-N,N-dialkylaminophenylmethylene)-3-methyl-5(4H)-isoxazolones and related compounds (17–33). [Values in
parentheses are the RBA values (see text).]
Figure 6. Structures of 3-substituted (Z)-4-(4-pyrrolidinophenylmethylene)-5(4H)-isoxazolones and related compounds (34–39). [Values in
parentheses are the RBA values (see text).]
involvement of cell permeability, metabolic inactivation,
and so on. (Of course, cell level assay is mandatory to
determine whether the ligand acts as an agonist or an
antagonist.) Therefore, structure–activity relationships
are discussed on the basis of RBA values in the follow-
ing four sections.
The non-alkylated derivative 18, was inactive. The cor-
responding hydroxyl analogue 33, was also inactive.
Introduction of two alkyl groups into the amino group
(19–23) resulted in appearance of the activity. The
activity of normal alkyl chain-introduced compounds
increased in the following order: H: 18 (RBA=0)
<CH3: 19 (1.3)<C2H5: 17 (20.3)<C3H7: 20 (57.0)
<C4H9: 22 (134.3). This order is not merely based on
the increase of hydrophobicity alone, because the diiso-
propyl analogue 21 has the weakest activity among
compounds 17–22. Cyclization of the diethyl group of
17 resulted in an increase of the activity, that is, 23 with
Figure 7. Structures of (Z)-4-(4-pyrrolidinophenylmethylene)-3-
phenyl-5(4H)-isoxazolones and related compounds (40–45). [Values in
parentheses are the RBA values (see text).]

1560
T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
an RBA value of 33.3. However, the ring expansion of
23 to give a piperidino analogue 30 resulted in a
decrease of the activity, and introduction of an oxygen
atom into the piperidino moiety to give the morpholino
analogue 32 abolished the activity. The acetamido analo-
gue 31 corresponding to 18 was also completely inactive.
Structure–activity relationship studies (3): effect of a sub-
stituent introduced at the 3-phenyl group (40–45) (Fig.
7). The effect of a substituent introduced into the 3-
phenyl group of 35 was investigated. Introduction of a
halogen atom (fluorine or chlorine) dramatically
increased the activity (40–43). The introduction of
chlorine at the para-position of 35 to give 41 increased
the activity almost 2-fold. Compound 41 showed the
most potent activity among the prepared compounds.
The para-position seems to be a superior position for
substitution compared to the ortho-position in the case
of chloro derivatives; the 40-chloro derivative 41
(RBA=220.0) is more potent than the 20-chloro deriva-
tive 40 (158.0). However, the order was reversed in the
case of fluorine derivatives; the 20-fluoro derivative 42
(140.6) is slightly more active than the corresponding 40-
fluoro derivative 43 (125.0). Introduction of another
electron-withdrawing group, a nitro group, into 35 to
give 44 and 45 moderately decreased the activity. The
decreasing effect was more apparent when a nitro group
was introduced at the para-position (45) compared to
introduction at the meta-position.
Further structural modification was performed based on
23, because we anticipated that the very potent activity
of 22, 134.4 times more potent than flutamide (2:
RBA=1.00; Fig. 1), might be attributed to its high
lipophilicity, not its molecular shape, at least in part.
Saturation of the 4-methylene moiety to give 24 resulted
in complete disappearance of the activity. The (E)-iso-
mer of 3-methylated 23, that is, 25, showed moderate
activity which was far weaker than that of 23. An aza
analogue 26 corresponding to 23 showed almost the
same activity as 23. Introduction of a substituent onto
the nitrogen atom of 26 resulted in a decrease of the
activity.
Structure–activity relationship studies (2): 3-substituted
(Z)-4-(4-pyrrolidinophenylmethylene)-5(4H)-isoxazolones
and related compounds (34–39) (Fig. 6). Next, based on
the ligand superfamily concept (vide supra, Fig. 3), we
introduced a second aromatic group into the 3-position
of 23. As expected, introduction of a phenyl group and
a benzyl group to give 35 and 36, respectively, resulted
in dramatic increases of the activity, with RBA values of
116.0 and 214.0, respectively. These compounds (35 and
36) are 1.6–2.9 times more potent than hydroxy-
flutamide (3: RBA=73.7; Fig. 1), an active metabolite
of flutamide (2). They appear to be more active than the
(R)-form of bicalutamide [eutomer of bicalutamide (5:
reported RBA=115.9; Fig. 1)], based on the reported Ki
values.33 Substitution of the 3-methyl group of 23 with
an ethyl group to give 34 slightly increased the activity.
Structure–activity relationship studies (4): (Z)-4-(4-dibu-
tylaminophenylmethylene)-3-benzyl-5(4H)-isoxazolone
analogues (46–50) (Fig. 8). The features of the struc-
ture–activity relationship of (Z)-4-(4-N,N-dialkylami-
nophenylmethylene)-5(4H)-isoxazolone
derivatives
established so far can be summarized as follows. Firstly,
in the series of (Z)-4-(4-N,N-dialkylaminophenylmethyl-
ene)-3-methyl-5(4H)-isoxazolones (17–23), the N,N-
dibutylamino derivative (22) shows the most potent
receptor-binding affinity (Fig. 5). Secondly, introduction
of a phenyl or a benzyl group at the 3-position con-
tributes to potent receptor affinity (Fig. 6). In the series
of (Z)-4-(pyrrolidinophenylmethylene)-3-phenyl-5(4H)-
isoxazolones (35, 40–45), the 40-chlorophenyl analogue
(41) showed the most potent receptor affinity (Fig. 7).
Overall these results led us to design hybrid molecules
possessing a N,N-dibutylaminophenyl group and a 40-
chlorophenyl (46) or a chlorinated benzyl group (48–
50). However, these compounds did not show as potent
activity as expected. Although a benzyl analogue 47
showed rather potent activity with a RBA value of
A pyridyl group is also effective as the 3-substituent
(37–39), but less so than a phenyl or a benzyl group.
However, pyridyl derivatives are expected to possess
higher polarity and hydrophilicity, and might possess
superior characteristics as lead compounds for andro-
gen-antagonistic therapeutics (vide infra).
Figure 8. Structures of (Z)-4-(4-N,N-dibutylaminophenlymethylene)-3-phenyl-5(4H)-isoxazolone analogues (40–50). [Values in parentheses are the
RBA values (see text).]

T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
1561
113.2, it was less potent than the corresponding pyrro-
lidino analogue 36. Introduction of chlorine (48–50)
dramatically lowered the activity. Surprisingly, the 40-
chlorophenyl analogue (46) showed only moderate
activity, which is much weaker than those of the corre-
sponding N-pyrrolidino analogue 41 and its non-chlor-
inated analogue 35. Though we cannot interpret the
findings at this stage, one possible explanation might be
different conformations of the ligand–receptor complex
between the N,N-dibutylaminophenyl series (22, 46–50)
and pyrrolidinophenyl series (23–29, 34–45).
testosterone-promoted growth of LNCaP cells with IC50
values of 1.9–2.5 mM. The potency of growth-inhibiting
activity toward LNCaP cells does not seem to correlate
to that toward SC-3 cells. Though we cannot analyze
this phenomenon further at this stage, it might be
attributed to the difference between AR expressed in
SC-3 cells (normal form) and AR expressed in LNCaP
cells (mutated aberrant form).29 These results suggest
that our compounds might be superior lead compounds
for development of therapeutically useful androgen
antagonists with different features from those of fluta-
mide (2).
Cell level activity
None of our compounds (17–50) showed growth-pro-
moting activity toward SC-3 or LNCaP cells cultured
in the absence of testosterone, indicating that our
compounds do not possess androgen-agonistic activity
(data not shown). In addition, under the experimental
conditions used, our compounds did not show
apparent cytotoxicity. On the other hand, some of
our compounds inhibited testosterone-promoted cell
growth of SC-3 to the baseline (the cell number was
almost the same as that when the cells were incubated in
the absence of testosterone), indicating these com-
pounds with AR-affinity are androgen antagonists. The
results of growth inhibition assay of testosterone-
induced cell growth of SC-3 and LNCaP cells are shown
in Table 1.
Conclusion
We designed novel non-steroidal/non-anilide type of
androgen antagonists based on the ligand superfamily
concept and a computer-assisted docking search of a
compound database. The molecular design and the
structural development studies reached potent andro-
gen antagonists with a (Z)-4-(4-pyrrolidinophenyl-
methylene)-3-phenyl/benzyl-5(4H)-isoxazolone
struc-
ture. Among them, 36 and 41 possess very potent AR-
binding affinity, more than 200 times greater than that
of flutamide (2). They showed potent SC-3 cell growth-
inhibitory activity, and they also acted as androgen
antagonists toward generally androgen antagonist-
resistant LNCaP cells. In the cell level assays, 22, 37,
and 39–45 showed potent SC-3 cell growth-inhibitory
activity. Among them, 40–43 were revealed to act as
androgen antagonists toward LNCaP cells. These com-
pounds are expected to be useful as lead compounds for
the development of therapeutically useful androgen
antagonists with different types of structures from
known androgen antagonists. Further biological eva-
luation of these compounds is in progress.
Concerning SC-3 assay, there seems to be only a rough
correlation between the growth-inhibitory activity and
AR-affinity (RBA values). However, compounds with
potent AR-affinity generally showed potent growth-
inhibiting activity. For example, compounds with RBA
values more than 120, that is, 22, 36, and 40–43, showed
potent growth-inhibitory activity on testosterone-
induced SC-3 growth with IC50 values of less than 1.0
mM, being more potent than flutamide (2: IC50=2.7
mM). The IC50 values of 22 and 40–43 are 0.2–0.4 mM,
indicating they are more potent than hydroxyflutamide
(3: IC50=0.8 mM), a metabolically activated form of
flutamide (2). All of the potent SC-3 cell growth inhibi-
tors with IC50 values of less than 1.0 mM possess RBA
values of more than 20. In this sense, 47 showed much
weaker SC-3 growth-inhibitory activity (IC50=6.1 mM)
than expected from its rather high RBA value (113.3).
This discrepancy might be attributed to the highly lipo-
philic nature of the compound.
Experimental
Chemicals
[1,2,6,7-3H]Testosterone ([3H]testosterone, 101 Ci/
mmol), isopropyl b-d-thiogalactoside and Sephadex G-
25 were purchased from Amersham Pharmacia Biotech
(England). Tris–base, EDTA, glycerol, DTT, sodium
chloride, sodium molybdate and testosterone were pur-
chased from Wako Pure Chemical, Ltd. (Japan).
Bacto1 Trypton and Bacto1 Yeast Extract were pur-
chased from Difco Laboratories (USA). Coomassie
Protein Assay Kit was purchased from Pierce (USA).
Antibody against hAR-LBD was purchased from Santa
Cruz Biotechnology, Inc. (USA). Atomlight scintillation
cocktail was purchased from Packard Bioscience
(Netherlands). MEM (modified Eagle’s medium) and
flutamide (2) were purchased from Sigma-Aldrich Japan
K.K. (Japan). FBS (fetal bovine serum) was purchased
from Gibco (USA). Cell Counting Kit was purchased
from Dojindo (Japan). All other chemicals were pur-
chased from Tokyo Kasei Kogyo Co., Ltd. (Japan),
Kanto Chemical Co., Inc. (Japan), Sigma-Aldrich
Japan K.K. (Japan) or Lancaster Synthesis (UK).
Some of our typical compounds were also assessed by
LNCaP assay. Generally, LNCaP cells, cloned from
human prostate tumor, are resistant to known androgen
antagonists, possibly because of their mutated aber-
rant AR.33 Flutamide (2) is reported to act as an
androgen agonist on LNCaP cells. As shown in
Table 1, LNCaP cells were rather resistant to our
compounds, though some of them (36, 39, and 47–50)
acted as androgen antagonists on the cells with IC50
values of less than 8 mM. No androgen-agonistic
activity was found, as far as checked. The N,N-dibu-
tylaminophenyl methylene-3-benzylisoxazolone analo-
gues 47–50 showed potent inhibitory activity on

1562
T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
1
General synthetic methods. Commercially available
reagents were used as supplied. Routine thin layer
chromatography (TLC) was performed on silica gel 60
F254 plates (Merck, Germany). Column chromato-
graphy was performed using silica gel 60 spherical
(Kanto Chemical Co., Inc., Japan). Melting points were
determined on a MP-J3 melting point apparatus
(Yanaco, Japan) and are uncorrected. Fast atom bom-
bardment mass spectra (FAB-MS) were measured by a
MS-JEOL JMS-HX110 mass spectrometer using a
nitrobenzyl alcohol matrix as appropriate. Proton
nuclear magnetic resonance (NMR) spectra data were
(MH+); H NMR (500 MHz, CDCl3) d 8.36 (d, J=9.0
Hz, 2H), 7.19 (s, 1H), 6.87 (d, J=9.0 Hz, 2H), 4.08 (qq,
J=7.0, 7.0 Hz, 2H), 2.25 (s, 3H), 1.38 (d, J=7.0 Hz,
12H). Anal. calcd for C17H22N2O2: C, 71.30; H, 7.74; N,
9.78. Found: C, 71.28; H, 7.83; N, 9.62.
(Z)-4-(4-N,N-Dibutylaminophenylmethylene) - 3 - methyl-
5(4H)-isoxazolone (22). Yield: 75%; red oil; MS(FAB)
m/z 315 (MH+); 1H NMR (500 MHz, CDCl3) d 8.4
(brs, 2H), 7.14 (s, 1H), 6.68 (d, J=9.5 Hz, 2H), 3.40 (t,
J=8.0 Hz, 4H), 2.24 (s, 3H), 1.63 (tt, J=8.0, 8.0 Hz,
4H), 1.39 (tq, J=8.0, 8.0 Hz, 4H), 0.98 (t, J=8.0 Hz,
6H). Anal. calcd for C19H26N2O2: C, 72.58; H, 8.33; N,
8.91. Found: C, 72.39; H, 8.22; N, 9.05.
obtained on
a
JEOL ALPHA500 spectrometer
(500 MHz) and were consistent with assigned structures.
Chemical shifts are given in parts per million (ppm)
downfield from internal reference TMS in d units, and
coupling constants (J values) are given in hertz (Hz).
Elemental analysis was performed on a MT-6 elemental
analyzer (Yanagimoto, Tokyo).
(Z)-4-(4-pyrrolidinophenylmethylene) - 3 - methyl - 5(4H)-
isoxazolone (23). Yield: 69%; red needles (benzene); mp
192 ꢁC; MS(FAB) m/z 257 (MH+), 256 (M+) ;1H NMR
(500 MHz, CDCl3) d 8.42 (d, J=9.0 Hz, 2H); 7.21 (s,
1H), 6.61 (d, J=9.0 Hz, 2H), 3.47 (tt, J=3.5, 3.5 Hz,
4H), 2.25 (s, 3H), 2.08 (tt, J=3.5, 3.5 Hz, 4H). Anal.
calcd for C15H16N2O2: C, 70.29; H, 6.29; N, 10.93.
Found: C, 70.39; H, 6.36; N, 10.86.
Preparation of 4 - N,N - dipropylaminobenzaldehyde.
Under an argon atmosphere, POCl3 (0.770 mL, 8.40
mmol) was slowly added to DMF (5 mL) on an ice
bath. To this reaction mixture was added N,N-dipropyl-
aniline (1.35 mL, 7.00 mmol), and the resulting mixture
was carefully heated at 80 ꢁC for 2 h. The reaction mix-
ture was poured onto cracked ice, neutralized with
aqueous sodium hydroxide and extracted with ethyl
acetate. The extracts were washed with brine, dried over
magnesium sulfate, filtered and concentrated in vacuo.
The residue was purified by column chromatography
(hexane/ethyl acetate, 10:1) to give 4-N,N-dipropylami-
nobenzaldehyde (1.06 g, 74%) as a white solid.
Ethyl-2-(4-pyrrolidinophenylmethyl)acetoacetate. A mix-
ture of N-phenylpyrrolidine (1.03 g, 7.00 mmol), ethyl
acetoacetate (1.82 g, 14.0 mmol) and formalin (3.00 mL
of 35%) in acetic acid (7 mL) was stirred at room
temperature for 10–12 h. The reaction mixture was
neutralized with ice-cold saturated Na2CO3 solution
and extracted with ethyl acetate. The extracts were
washed with brine, dried over magnesium sulfate,
filtered and concentrated in vacuo. The residue was
purified by column chromatography (hexane/ethyl ace-
tate, 10:1) to give ethyl 2-(4-pyrrolidinophenylmethyl)-
acetoacetate (230 mg, 11%). MS(FAB) m/z 179 (MH+),
178 (M+); 1H NMR (500 MHz, CDCl3) d 7.02 (dt,
J=8.5, 2.0 Hz, 2H), 6.47 (dt, J=8.5, 2.0 Hz, 2H), 4.16
(q, J=7.0 Hz, 2H), 3.73 (t, J=7.5 Hz, 1H), 3.24 (tt,
J=3.5, 3.5 Hz, 4H), 3.06 (t, J=7.5 Hz, 2H), 2.17 (s,
3H), 1.98 (tt, J=3.5, 3.5 Hz, 4H), 1.28 (t, J=7.0 Hz,
3H).
(Z)-4-(4-Aminophenylmethylene)-3-methyl-5(4H)-isoxa-
zolone (18). Yield: 23%; orange crystals (CH2Cl2); mp
210–211 ꢁC; MS(FAB) m/z 203 (MH+); 1H NMR
(500 MHz, DMSO-d6) d 8.35 (d, J=8.5 Hz, 2H), 7.54 (s,
1H), 7.09 (s, 2H), 6.66 (d, J=8.5 Hz, 2H), 2.19 (s, 3H).
Anal. calcd for C11H10N2O2: C, 65.34; H, 4.98; N,
13.85. Found: C, 65.37; H, 5.16; N, 13.67.
(Z)-4-(4-N,N-Dimethylaminophenylmethylene)-3-methyl-
5(4H)-isoxazolone (19). Yield: 47%; red needles
(EtOAc); mp 207–208 ꢁC; MS(FAB) m/z 231 (MH+),
230 (M+); 1H NMR (500 MHz, CDCl3) d 8.41 (d,
J=10.0 Hz, 2H), 7.23 (s, 1H), 6.73 (d, J=10.0 Hz, 2H),
3.16 (s, 6H), 2.25 (s, 3H). Anal. calcd for C15H18N2O2:
C, 67.81; H, 6.13; N, 12.17. Found: C, 67.91; H, 6.40; N,
11.92.
4-(4-Pyrrolidinophenylmethyl)-3-methyl-5(4H)-isoxazo-
lone (24). A mixture of ethyl 2-(4-pyrrolidinophenyl-
methyl)acetoacetate (434 mg, 1.50 mmol), hydroxylamine
hydrochloride (206 mg, 3.00 mmol) and sodium acetate
(369 mg, 4.50 mmol) in ethanol (30 mL) was heated
under reflux for 3 h. The reaction mixture was diluted
with ethyl acetate, and the organic layer was washed
with brine, dried over magnesium sulfate, filtered and
concentrated in vacuo. The residue was purified by col-
umn chromatography (hexane/ethyl acetate, 2:1) to give
4-(4-pyrrolidinophenylmethyl) - 3 - methyl - 5(4H)-iso-
xazolone (24) (245 mg, 63%) as a brown solid, which
was recrystallized from hexane/ethyl acetate; mp 139–
142 ꢁC; MS(FAB) m/z 259 (MH+), 258 (M+); 1H NMR
(500 MHz, CDCl3) d 7.10 (dd, J=8.0, 1.0 Hz, 2H), 6.48
(dd, J=8.0, 1.0 Hz, 2H), 3.51 (t, J=5.5 Hz, 1H), 3.25
(tt, J=3.5, 3.5 Hz, 4H), 3.16 (dd, J=14.5, 5.5 Hz, 1H),
3.09 (dd, J=14.5, 5.5 Hz, 1H), 1.99 (tt, J=3.5, 3.5 Hz,
4H), 1.99 (s, 3H). HRMS (FAB, MH+) C15H19N2O2,
calcd for 259.1447, found 259.1484.
(Z)-4-(4-N,N-Dipropylaminophenylmethylene)-3-methyl-
5(4H)-isoxazolone (20). Yield: 42%; red crystals (hex-
ane/EtOAc); mp 117 ꢁC; MS(FAB) m/z 287 (MH+); 1H
NMR (500 MHz, CDCl3) d 8.4 (brs, 2H), 7.19 (s, 1H),
6.69 (d, J=9.0 Hz, 2H), 3.38 (t, J=8.0 Hz, 4H), 2.25 (s,
3H), 1.68 (tq, J=8.0, 8.0 Hz, 4H), 0.98 (t, J=8.0 Hz,
6H). Anal. calcd for C17H22N2O2: C, 71.30; H, 7.74; N,
9.78. Found: C, 71.32; H, 7.97; N, 9.65.
(Z) - 4 - (4 - N,N - Diisopropylaminophenylmethylene) - 3-
methyl-5(4H)-isoxazolone (21). Yield: 11%; red crystals
(hexane/EtOAc); mp 134–135 ꢁC; MS(FAB) m/z 287

T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
1563
(E)-4-(4-Pyrrolidinophenyl-1-ethylene)-3-methyl-5(4H)-
isoxazolone (25). Yield: 33%; orange flakes (EtOAc);
mp 147–150 ꢁC; MS(FAB) m/z 271 (MH+), 270 (M+);
1H NMR (500 MHz, CDCl3) d 7.20 (dt, J=9.0, 2.0 Hz,
2H), 6.57 (dt, J=9.0, 2.0 Hz, 2H), 3.39 (tt, J=3.5, 3.5
Hz, 4H), 2.80 (s, 3H), 2.07 (tt, J=3.5, 3.5 Hz, 4H), 1.82
(s, 3H). Anal. calcd for C15H16N2O2: C, 71.09; H, 6.71;
N, 10.36. Found: C, 71.09; H, 6.78; N, 10.17. Forma-
tion of the corresponding (Z)-form was not observed
(the stereochemistry of the olefin was determined by
means of NOESY experiments). NOE was observed
between the methyl at C3 and the aromatic protons at
20- and 60-positions.
(Z)-4-(4-Acetoaminophenylmethylene)-3-methyl-5(4H)-
isoxazolone (31). Yield: 41%; yellow solid (MeOH/
EtOAc); mp 253–257 ꢁC (decomposition); MS(FAB) m/
1
z 245 (MH+); H NMR (500 MHz, DMSO-d6) d 10.48
(s, 1H), 8.46 (d, J=8.5 Hz, 2H), 7.83 (s, 1H), 7.77 (d,
JH=8.5 Hz, 2H), 2.26 (s, 3H), 2.11 (s, 3H). Anal. calcd
for C13H12N2O2: C, 63.93; H, 4.95; N, 11.47. Found: C,
63.82; H, 5.06; N, 11.35.
(Z)-4-(4-Morpholinophenylmethylene)-3-methyl-5(4H)-
isoxazolone (32). Yield: 46%; red needles (CH2Cl2); mp
202 ꢁC; MS(FAB) m/z 273 (MH+); 1H NMR
(500 MHz, CDCl3) d 8.42 (d, J=9.0 Hz, 2H), 7.25 (s,
1H), 6.90 (d, J=9.0 Hz, 2H), 3.86 (t, J=4.5 Hz, 4H),
3.86 (t, J=4.5 Hz, 4H), 2.26 (s, 3H). Anal. calcd for
C15H16N2O3: C, 66.16; H, 5.92; N, 10.29. Found: C,
65.93; H, 5.98; N, 10.27.
(Z)-4-(4-Pyrrolidinophenylmethylene)-3-methyl-5(4H)-
pyrazolone (26). Yield: 40%; red crystals (MeOH/
EtOAc); mp 213–214 ꢁC; MS(FAB) m/z 256 (MH+);
1H NMR (500 MHz, CDCl3) d 8.53 (d, J=8.5 Hz,
2H), 8.41 (s, 1H), 7.22 (s, 1H), 6.60 (d, J=8.5 Hz,
2H), 3.44 (tt, J=3.5, 3.5 Hz, 4H), 2.13 (s, 3H), 2.07
(tt, J=3.5, 3.5 Hz, 4H). Anal. calcd for C15H17N3O : C,
70.56; H, 6.71; N, 16.46. Found: C, 70.28; H, 6.86; N,
16.23.
(Z)-4-(4-hydroxyphenylmethylene)-3-methyl-5(4H)-isoxa-
zolone (33). Yield: 40%; yellow flakes (EtOAc); mp 207–
217 ꢁC; MS(FAB) m/z 204 (MH+); 1H NMR
(500 MHz, CDCl3) d 8.45 (d, J=9.0 Hz, 2H), 7.66 (s,
1H), 6.90 (td, J=9.0, 1.5 Hz, 2H), 3.30 (tt, J=1.5, 1.5
Hz, 1H), 2.26 (s, 3H). Anal. calcd for C11H9NO3: C,
65.02; H, 4.46; N, 6.89. Found: C, 64.95; H, 4.52; N,
6.73.
(Z)-4 - (4 - Pyrrolidinophenylmethylene) - 1,3 - dimethyl -
5(4H) - pyrazolone (27). Yield: 61%; red crystals
(EtOAc); mp 170–172 ꢁC; MS(FAB) m/z 270 (MH+),
269 (M+); 1H NMR (500 MHz, CDCl3) d 8.58 (d,
J=8.5 Hz, 2H), 7.19 (s, 1H), 6.50 (d, J=8.5 Hz, 2H),
3.45 (tt, J=3.5, 3.5 Hz, 4H), 3.40 (s, 3H), 2.23 (s, 3H),
2.06 (tt, JH=3.5, 3.5 Hz, 4H). Anal. calcd for
C16H19N3O: C, 71.35; H, 7.11; N, 15.60. Found: C,
71.36; H, 7.19; N, 15.51.
Synthesis of (Z) - 4 - (4 - pyrrolidinophenylmethylene) - 3 -
ethyl-5(4H)-isoxazolone (34). Phosphoric acid (2.94 g)
was added to a mixture of 4-pyrrolidinobenzaldehyde
(526 mg, 3.00 mmol) and 3-keto-n-valerate (1.95 g, 15.0
mmol), and the whole mixture was stirred at room
temperature for 10–12 h. The reaction mixture was
poured into ice-cold sodium hydroxide solution and
extracted with dichloromethane. The extracts were
washed with brine, dried over magnesium sulfate, fil-
tered and concentrated in vacuo. The residue was pur-
ified by column chromatography (hexane/ethyl acetate,
3:2) to give (Z)-4-(4-pyrrolidinophenylmethylene)-3-
ethyl-5(4H)-isoxazolone (34) (171 mg, 21%) as red
crystals, which were further purified by recrystallization
from hexane/ethyl acetate to give orange crystals; mp
147 ꢁC; MS(FAB) m/z 271 (MH+), 271 (M+); 1H NMR
(500 MHz, CDCl3) d 8.41 (d, J=8.5 Hz, 2H), 7.23 (s,
1H), 6.60 (d, J=8.5 Hz, 2H), 3.47 (tt, J=3.5, 3.5 Hz,
4H), 2.64 (q, J=7.5 Hz, 2H), 2.09 (tt, J=3.5, 3.5 Hz,
4H), 1.34 (t, J=7.5 Hz, 3H). Anal. calcd for
C16H18N2O2: C, 71.09; H, 6.71; N, 10.36. Found: C,
71.00; H, 6.82; N, 10.21.
(Z)-4-(4-Pyrrolidinophenylmethylene)-1-phenyl-3-methyl-
5(4H) - pyrazolone (28). Yield: 47%; red crystals
(EtOAc); mp 205 ꢁC; MS(FAB) m/z 332 (MH+), 332
1
(M+); H NMR (500 MHz, CDCl3) d 8.59 (d, J=8.5
Hz, 2H), 8.01 (dd, J=8.0, 1.0 Hz, 2H), 7.41 (td, J=8.0,
1.0 Hz, 2H), 7.15 (td, J=8.0, 1.0 Hz, 2H), 6.60 (d,
J=8.5 Hz, 2H), 3.45 (tt, J=3.5, 3.5 Hz, 4H), 2.33 (s,
3H), 2.07 (tt, J=3.5, 3.5 Hz, 4H). Anal. calcd for
C21H21N3O: C, 76.11; H, 6.39; N, 12.68. Found: C,
76.25; H, 6.55; N, 12.58
(Z)-4-(4-Pyrrolidinophenylmethylene)-1-(4-cyanophenyl)-
3-methyl-5(4H)-pyrazolone (29). Yield: 67%; red nee-
dles (CH2Cl2/EtOAc); mp 235–236 ꢁC; MS(FAB) m/z
1
357 (MH+), 356 (M+); H NMR (500 MHz, CDCl3) d
8.55 (d,J=8.5 Hz, 2H), 8.25 (dt,J=8.0, 2.0 Hz, 2H),
7.67 (dt,J=9.0, 2.0 Hz, 2H), 7.28 (s, 1H), 6.64 (d,J=8.5
Hz, 2H), 3.47 (tt, J=3.5, 3.5 Hz, 4H), 2.09 (s, 3H), 2.07
(tt, J=3.5, 3.5 Hz, 4H). Anal. calcd for C22H20N4O : C,
74.14; H, 5.66; N, 15.72. Found: C, 74.20; H, 5.95; N,
15.75.
(Z)-4-(4-Pyrrolidinophenylmethylene)-3-phenyl-5(4H)-
isoxazolone (35). Yield: 81%; red crystals (CH2Cl2/
EtOAc); mp 172–173 ꢁC; MS(FAB) m/z 319 (MH+),
318 (M+); 1H NMR (500 MHz, CDCl3) d 8.4 (brs, 2H),
7.5–7.6 (m, 5H), 7.370 (s, 1H), 6.59 (d, J=9.0 Hz, 2H),
3.47 (tt, J=3.0, 3.0 Hz, 4H), 2.08 (tt, J=3.0, 3.0 Hz,
4H). Anal. calcd for C20H18N2O2: C, 75.45; H, 5.70; N,
8.80. Found: C, 75.35; H, 5.82; N, 8.69.
(Z)-4-(4-Piperazinophenylmethylene) - 3 - methyl - 5(4H) -
isoxazolone (30). Yield: 34%; orange needles (hexane/
EtOAc); mp 163 ꢁC; MS(FAB) m/z 271 (MH+); 1H
NMR (500 MHz, CDCl3) d 8.39 (d, J=9.0 Hz, 2H),
7.20 (s, 1H), 6.87 (d, J=9.0 Hz, 2H), 3.5 (brs, 4H), 1.7
(brs, 6H). Anal. calcd for C16H18N2O2; C, 71.09; H,
6.71; N, 10.36. Found: C, 70.97; H, 6.88; N, 10.45.
(Z)-4-(4-Pyrrolidinophenylmethylene)-3-phenylmethyl-
5(4H) - isoxazolone (36). Yield: 82%; orange flakes
(CH2Cl2/EtOAc); mp 196–197 ꢁC; MS(FAB) m/z 333
1
(MH+), 332 (M+); H NMR (500 MHz, CDCl3) d 8.2

1564
T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
(brs, 2H), 7.3–7.4 (m, 5H), 7.16 (s, 1H), 6.54 (d, J=8.5
Hz, 2H), 3.98 (s, 2H), 3.44 (tt, J=3.0, 3.0 Hz, 4H), 2.06
(tt, JH=3.0, 3.0 Hz, 4H). Anal. calcd for C21H20N2O2:
C, 75.88; H, 6.06; N, 8.43. Found: C, 75.79; H, 6.11; N,
8.34.
7.55 (d,J=7.5 Hz, 1H), 7.53 (d, J=9.0 Hz, 1H), 7.31 (d,
J=7.5 Hz, 1H), 7.23 (d, J=9.0 Hz, 1H), 7.14 (d, J=2.0
Hz, 1H), 6.65 (d, J=9.0 Hz, 2H), 3.47 (tt, J=3.5, 3.5
Hz, 4H). 2.08 (tt, J=3.5, 3.5 Hz, 4H). Anal. calcd for
C20H17N2O2F; C, 71.42: H, 5.09; N, 8.33. Found: C,
71.46; H, 5.26; N, 8.12.
(Z)-4-(4-Pyrrolidinophenylmethylene) - 3 - (2- pyridyl) -
5(4H)-isoxazolone (37). Yield: 16%; red crystals (hex-
ane/EtOAc); mp 180–181 ꢁC; MS(FAB) m/z 320
(Z)-4-(4-Pyrrolidinophenylmethylene)-3-(4-fluorophenyl)-
5(4H)-isoxazolone (43). Yield: 53%; purple crystals
(hexane/EtOAc); mp 217 ꢁC; MS(FAB) m/z 337
1
(MH+); H NMR (500 MHz, CDCl3) d 8.77 (s, 1H),
1
8.75 (dd, J=5.0, 1.5 Hz, 1H), 8.50 (d, J=9.0 Hz, 2H),
8.03 (d, J=8.0 Hz, 1H), 7.84 (td, J=8.0, 1.5 Hz, 1H),
7.40 (ddd, J=8.0, 5.0, 1.0 Hz, 1H), 6.61 (d, J=9.0 Hz,
2H), 3.48 (tt, J=3.5, 3.5 Hz, 4H), 2.09 (tt, J=3.5, 3.5
Hz, 4H). Anal. calcd for C19H17N3O2: C, 71.46; H, 5.37;
N, 13.16. Found: C, 71.61; H, 5.58; N, 12.92.
(MH+); H NMR (500 MHz, CDCl3) d 8.4 (brs, 2H),
7.58 (ddd, J=8.5, 5.5, 2.0 Hz, 2H), 7.31 (s, 1H), 7.23
(td, J=8.5, 2.0 Hz, 2H), 6.60 (d, J=9.0 Hz, 2H), 3.48
(tt, J=3.5, 3.5 Hz, 4H), 2.09 (tt, J=3.5, 3.5 Hz, 4H).
Anal. calcd for C20H17N2O2F; C, 71.42: H, 5.09; N,
8.33. Found: C, 71.51; H, 5.29; N, 8.21.
(Z)-4-(4-Pyrrolidinophenylmethylene) - 3 - (3 - pyridyl) -
5(4H) - isoxazolone (38). Yield: 49%; red needles
(CH2Cl2/EtOAc); mp 145–146 ꢁC; MS(FAB) m/z 320
(Z)-4-(4-Pyrrolidinophenylmethylene)-3-(3-nitrophenyl)-
5(4H) - isoxazolone (44). Yield: 79%; red crystals
(CH2Cl2/MeOH); mp 215 ꢁC; MS(FAB) m/z 364
1
1
(MH+); H NMR (500 MHz, CDCl3) d 8.85 (d, J=1.5
(MH+); H NMR (500 MHz, CDCl3) d 8.47 (t, J=2.0
Hz, 1H), 8.79 (dd, J=5.0, 1.5 Hz, 1H), 8.4 (brs, 2H),
7.95 (dt, J=8.0, 1.5 Hz, 1H), 7.40 (ddd, J=8.0, 5.0, 0.5
Hz, 1H), 7.31 (s, 1H), 6.61 (d, J=9.0 Hz, 2H), 3.49 (tt,
J=3.5, 3.5 Hz, 4H), 2.09 (tt, J=3.5, 3.5 Hz, 4H). Anal.
calcd for C19H17N3O2: C, 71.46; H, 5.37; N, 13.16.
Found: C, 71.53; H, 5.55; N, 12.91.
Hz, 1H), 8.41 (ddd, J=9.0, 2.0, 1.0 Hz, 2H), 7.96 (dt,
J=8.0, 1.0 Hz, 1H), 7.74 (d, J=8.0 Hz, 1H), 7.29 (s,
1H), 6.61 (d, J=9.0 Hz, 2H), 3.50 (tt, J=3.5, 3.5 Hz,
4H), 2.10 (tt, J=3.5, 3.5 Hz, 4H). Anal. calcd for
C20H17N3O4: C, 66.11; H, 4.92; N, 11.56. Found: C,
66.00; H, 4.92; N, 11.84.
(Z)-4-(4-Pyrrolidinophenylmethylene) - 3 - (4 - pyridyl) -
5(4H) - isoxazolone (39). Yield: 51%; red needles
(CH2Cl2/EtOAc); mp 210–211 ꢁC; MS(FAB) m/z 320
(MH+); 1H NMR (500 MHz, CDCl3) d 8.81 (dd,
JH=6.0, 0.5 Hz, 2H), 8.4 (brs, 2H), 7.52 (dd, J=6.0, 0.5
Hz, 2H), 7.32 (s, 1H), 6.61 (d, J=9.0 Hz, 2H), 3.50 (tt,
J=3.5, 3.5 Hz, 4H), 2.10 (tt, J=3.5, 3.5 Hz, 4H). Anal.
calcd for C19H17N3O2: C, 71.46; H, 5.37; N, 13.16.
Found: C, 71.26; H, 5.44; N, 13.46.
(Z)-4-(4-Pyrrolidinophenylmethylene)-3-(4-nitrophenyl)-
5(4H) - isoxazolone (45). Yield: 73%; red crystals
(CH2Cl2); mp 178–180 ꢁC; MS(FAB) m/z 364 (MH+);
1H NMR (500 MHz, DMSO-d6) d 8.5 (brs, 2H), 8.40
(dt, J=9.0, 1.0 Hz, 2H), 7.92 (dt, J=9.0, 1.0 Hz, 2H),
7.48 (s, 1H), 6.74 (d, J=9.0 Hz, 2H), 3.49 (tt, J=3.5, 3.5
Hz, 4H), 1.99 (tt, J=3.5, 3.5 Hz, 4H). Anal. calcd for
C20H17N3O4: C, 66.11; H, 4.92; N, 11.56. Found: C,
65.81; H, 4.90; N, 11.37.
(Z)-4-(4-Pyrrolidinophenylmethylene)-3-(2-chlorophenyl)
- 5(4H) - isoxazolone (40). Yield: 35%; red crystals
(EtOAc); mp 220–221 ꢁC; MS(FAB) m/z 353 (MH+);
1H NMR (500 MHz, CDCl3) d 8.4 (brs, 2H), 7.54 (dd,
J=7.5, 1.0 Hz, 1H), 7.49 (dd, J=7.5, 1.0 Hz, 1H), 7.48
(td, J=7.5, 1.0 Hz, 1H), 7.42 (td, J=7.5, 1.0 Hz, 1H),
6.97 (s, 1H), 6.58 (d,J=9.0 Hz, 2H), 3.47 (tt, J=3.5, 3.5
Hz, 4H), 2.08 (tt, J=3.5, 3.5 Hz, 4H). Anal. calcd for
C20H17N2O2Cl; C, 68.09; H, 4.86; N, 7.94. Found: C,
68.36; H, 5.08; N, 7.90.
(Z)-4-(4-N,N-Dibutylaminophenylmethylene)-3-(4-chloro-
phenyl) - 5(4H) - isoxazolone (46). Yield: 62%; red oil;
MS(FAB) m/z 411 (MH+); 1H NMR (500 MHz,
CDCl3) d 8.4 (brs, 2H), 7.53 (dd, J=7.0, 2.5 Hz, 2H),
7.51 (dd, J=7.0, 2.5 Hz, 2H), 7.28 (s, 1H), 6.67 (d,
J=9.0 Hz, 2H), 3.41 (t, J=7.5 Hz, 4H), 1.63 (tt, J=7.5
Hz, 4H), 1.39 (tq, J=7.5, 7.5 Hz, 4H), 0.98 (t, J=7.5
Hz, 6H). HRMS (FAB, MH+) C24H28N2O235Cl, calcd
for 411.1839, found 411.1849.
(Z)-4-(4-N,N - Dibutylaminophenylmethylene) - 3 - phenyl-
methyl-5(4H)-isoxazolone (47). Yield: 48%; red needles
(hexane/EtOAc); mp 118 ꢁC; MS(FAB) m/z 391
(Z)-4-(4-Pyrrolidinophenylmethylene)-3-(4-chlorophenyl)
- 5(4H) - isoxazolone (41). Yield: 62%; red needles
(CH2Cl2/EtOAc); mp 170–174 ꢁC; MS(FAB) m/z 353
1
(MH+); H NMR (500 MHz, CDCl3) d 8.25 (brs, 2H),
1
(MH+); H NMR (500 MHz, CDCl3) d 8.4 (brs, 2H),
7.33–7.25 (m, 4H), 7.13 (s, 1H), 6.62 (d,J=9.4 Hz, 2H),
3.97 (s, 2H), 3.37 (m, 4H), 1.59 (m, 4H), 1.36 (m, 4H),
0.96 (t, J=7.4 Hz, 6H). Anal. calcd for C25H30N2O2: C,
76.89; H, 7.74; N, 7.17. Found: C, 76.80; H, 7.82; N, 7.17.
7.54 (dd, J=7.0, 2.5 Hz, 2H), 7.51 (dd, J=7.0, 2.5 Hz,
2H), 7.31 (s, 1H), 6.60 (d, J=9.0 Hz, 2H), 3.48 (tt,
J=3.5, 3.5 Hz, 4H), 2.09 (tt, J=3.5, 3.5 Hz, 4H). Anal.
calcd for C20H17N2O2Cl: C, 68.09; H, 4.86; N, 7.94.
Found: C, 68.08; H, 5.11; N, 8.06.
(Z)-4-(4-N,N-Dibutylaminophenylmethylene)-3-(2-chloro-
phenylmethyl) - 5(4H)-isoxazolone (48). Yield: 56%;
orange solid (hexane/EtOAc); mp 101–117 ꢁC;
MS(FAB) m/z 425 (MH+); 1H NMR (500 MHz,
CDCl3) d 8.31 (brs, 2H), 7.41 (d, J=7.3, 6.9 Hz, 2H),
7.36–7.2 (m, 2H), 7.22 (s, 1H), 6.64 (d, J=9.4 Hz, 2H),
(Z)-4-(4-Pyrrolidinophenylmethylene)-3-(2-fluorophenyl)-
5(4H)-isoxazolone (42). Yield: 48%; red needles (hex-
ane/EtOAc); mp 200–201 ꢁC; MS(FAB) m/z 337
1
(MH+); H NMR (500 MHz, CDCl3) d 8.4 (brs, 2H),

T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
1565
4.10 (s, 2H), 3.38 (m, 4H), 1.60 (m, 4H), 1.35 (m, 4H),
0.96 (t, J=7.3 Hz, 6H). Anal. calcd for C25H29N2O2Cl:
C, 70.66; H, 6.88; N, 6.59. Found; C, 70.49; H, 6.93; N,
6.48.
1000-fold excess of nonradioactive testosterone.
Separation of ligand-bound protein fraction and free
radioactive ligand was achieved by Sephadex G-25 gel
filtration column chromatography. The protein fraction
was collected and the radioactivity was determined with
a liquid scintillation counter. Under the experimental
conditions, 10 nM [3H]testosterone was enough to
saturate the binding sites in the GST-hARLBD frac-
tion. All experiments were performed in triplicate or
more. To determine the Kd of [3H]testosterone, data
were analyzed using a modified form of the Scatchard
equation: B/F=(ꢀ1/Kd) ꢂ B+Bmax/Kd, where B is the
concentration of specifically bound [3H]testosterone, F
is the concentration of free [3H]testosterone, and Bmax is
the maximum concentration of available binding sites.
Binding competition studies for the test compounds
were then performed under identical conditions by
incubating increasing concentrations (100 nM to 10
mM) of test compound (dissolved in DMSO) with the
GST-hARLBD fraction in the presence of a saturat-
ing concentration of [3H]testosterone (10 nM) at 4 ꢁC
for 12–18 h. The concentration of test compounds that
inhibited [3H]testosterone-binding to the extent of 50%
was quantified (IC50) after log–logit transformation.
The Ki values were determined by application of
the Cheng–Prussof equation to the IC50 values,
where Ki=IC50/{1+([3H]testosterone)/Kd}. Relative
binding affinity (RBA) was defined as: RBA=KiFlu/Ki,
where KiFlu is Ki of flutamide (2), the reference
compound.
(Z)-4-(4-N,N-Dibutylaminophenylmethylene)-3-(3-chloro-
phenylmethyl) - 5(4H) - isoxazolone (49). Yield: 52%;
orange flakes (hexane/EtOAc); mp 101 ꢁC; MS(FAB) m/
1
z 425 (MH+); H NMR (500 MHz, CDCl3) d 8.27 (brs,
2H), 7.4–7.2 (m, 4H), 7.09 (s, 1H), 6.63 (d, J=9.4 Hz,
2H), 3.94 (s, 2H), 3.38 (m, 4H), 1.60 (m, 4H), 1.37 (m,
4H), 0.96 (t, J=7.2 Hz, 6H). Anal. calcd for
C25H29N2O2Cl: C, 70.66; H, 6.88; N, 6.59. Found: C,
70.54; H, 6.96; N, 6.54.
(Z)-4-(4-N,N-Dibutylaminophenylmethylene)-3-(4-chloro-
phenylmethyl) - 5(4H) - isoxazolone (50). Yield: 43%;
orange flakes (hexane/EtOAc); mp 123–124 ꢁC;
MS(FAB) m/z 425 (MH+); 1H NMR (500 MHz,
CDCl3) d 8.26 (brs, 2H), 7.3–7.2 (m, 4H), 7.07 (s, 1H),
6.63 (d, J=9.4 Hz, 2H), 3.94 (s, 2H), 3.38 (m, 4H), 1.60
(m, 4H), 1.37 (m, 4H), 0.96 (t, J=7.2 Hz, 6H). Anal.
calcd for C25H29N2O2Cl: C, 70.66; H, 6.88; N, 6.59.
Found: C, 70.64; H, 6.94; N, 6.57.
Receptor binding assay
Binding affinities of test compounds for hAR (human
androgen receptor) were measured in competition
experiments using [3H]testosterone and cytosolic frac-
tion of hAR-LBD (hAR ligand-binding domain)-trans-
formed E. coli as described previously.20,28,30,31 A hAR-
LBD expression plasmid vector which codes GST-
hARLBD (627-919 aa, EF domain) fusion protein
under the lac promoter (provided by Prof. S. Kato,
Univ. of Tokyo) was transfected into E. coli strain HB-
101. An overnight culture (10 mL) of the bacteria was
added to 1 L of LB medium and incubated at 27 ꢁC
until its optical density reached 0.6–0.7 at 600 nm. Fol-
lowing the addition of IPTG to a concentration of 1
mM, incubation was continued for an additional 4.5 h.
Cells were harvested by centrifugation at 4000g at 4 ꢁC
for 15 min and stored at ꢀ80 ꢁC until use. All sub-
sequent operations were performed at 4 ꢁC. The bacter-
ial pellet obtained from 40 mL of culture was
resuspended in 1 mL of ice-cold TEGDM buffer (10
mM Tris, 1 mM EDTA, 10% glycerol, 10 mM DTT, 10
mM sodium molybdate). This suspension was subjected
to sonication using 12ꢂ7 s bursts on ice (USP-600A
sonicator, Shimadzu, Japan) and crude GST-hARLBD
fraction was prepared by centrifugation of the suspen-
sion at 12,000g for 30 min at 4 ꢁC. Western blot analysis
showed that GST-hARLBD existed in the supernatant
cytosolic fraction. This crude receptor fraction was
diluted to a protein concentration of 0.3–0.5 mg/mL
and used in binding assays as GST-hARLBD frac-
tion. Total protein was determined using a Coomassie
Protein Assay Kit. In preliminary experiments, the
equilibrium dissociation constant (Kd) of testosterone
was determined by incubating the GST-hARLBD frac-
tion with increasing concentrations of [3H]testosterone
(from 0.1 to 100 nM final concentration) at 4 ꢁC for 12–
18 h. Non-specific binding was assessed by addition of
SC-3/LNCaP growth inhibition assay
Shionogi Carcinoma-3 (SC-3) cells were cloned from
Shionogi Carcinoma 115 cells, which were established
from a mouse breast cancer. SC-3 shows androgen-
dependent growth.26,27 In this assay, androgenic and
anti-androgenic activities of test compounds were
determined in terms of SC-3 growth promotion and
inhibition, respectively. SC-3 cells were cultured in the
presence of MEM supplemented with 10% FBS and 10
nM testosterone at 37 ꢁC 5% CO2. All experiments were
performed in triplicate or more. For SC-3 cell growth-
inhibition assay, the cells were trypsinized, and diluted
to 3.0 ꢂ 104 cell/mL with MEM supplemented with
10% charcoal resin-stripped fetal bovine serum. This
cell suspension was transferred to 96-well microtiter
plates, and various concentrations of test compound
(from 100 nM to 10 mM DMSOsolution) and/or
testosterone ethanol solution (final concentration 10
nM) were added. Then the plates were incubated at
37 ꢁC 5% CO2 for 3 days, and the cell number was
determined using the WST-1 method with a Cell
Counting Kit and an MPR-A4i2 micro plate reader
(TOSOH, Japan). The number of cells on wells with
testosterone alone was defined as 100%. The con-
centration of test compounds that inhibited the
increase of the cell number induced by 10 nM testoster-
one by 50% was quantified (IC50) after log–logit trans-
formation. LNCaP growth promotion and inhibition
assays were performed by the same method as described
for SC-3 assay, except that RPMI1640 medium was
used.

1566
T. Ishioka et al. / Bioorg. Med. Chem. 10 (2002) 1555–1566
14. Raynaud, J. P.; Bonne, C.; Moguilewsky, M.; Lefebvre,
Acknowledgements
F. A.; Belanger, A.; Labrie, F. The Prostate 1984, 5, 299.
15. Tyrrell, C. J.; Denis, L.; Newling, D.; Soloway, M.;
Channer, K.; Cockshott, I. D. Eur. Urol. 1998, 9, 39.
16. Miyachi, H.; Azuma, A.; Kitamoto, T.; Hayashi, K.;
Kato, S.; Koga, M.; Sato, B.; Hashimoto, Y. Bioorg. Med.
Chem. Lett. 1997, 7, 1483.
17. Hashimoto, Y. Curr. Med. Chem. 1998, 5, 163.
18. Hashimoto, Y. Recent Res. Develop. Org. Chem. 2000, 4,
87.
19. Hashimoto, Y. Bioorg. Med. Chem. 2002, 10, 461.
20. Takahashi, H.; Ishioka, T.; Koiso, Y.; Sodeoka, M.;
Hashimoto, Y. Biol. Pharm. Bull. 2000, 23, 1387.
21. Hamann, L. G.; Higuchi, R. I.; Zhi, L.; Edwards, J. P.;
Wang, X. N.; Marschke, K. B.; Kong, J. W.; Farmer, L. J.;
Jones, T. K. J. Med. Chem. 1998, 41, 623.
The authors are grateful to Prof. S. Kato for his gener-
ous supply of human androgen receptor gene. The work
described in this paper were partially supported by
Grants-in-Aid for Scientific Research from The Minis-
try of Education, Science, Sports and Culture, Japan,
and by funds from the Program for Promotion of Fun-
damental Studies in Health Sciences of the Organization
for Drug ADR Relief, R&D Promotion and Product
Review, Japan.
References and Notes
22. Hamann, L. G.; Mani, N. S.; Davis, R. L.; Wang, X. N.;
Marschke, K. B.; Jones, T. K. J. Med. Chem. 1999, 42, 210.
23. Laudet, V.; Hanni, C.; Coll, J.; Catzeflis, F.; Stehelin, D.
EMBO J. 1992, 11, 1103.
24. Campaigne, E.; Archer, W. L. J. Amer. Chem. Soc. 1953,
77, 989.
1. Mooradian, A. D.; Morley, J. E.; Korenman, S. G. Endocr.
Rev. 1987, 8, 1.
2. Zhi, L.; Martinborough, E. Ann. Rep. Med. Chem. 2001,
36, 169.
3. McEwan, I. J. Biochem. Soc. Trans. 2000, 28, 369.
4. MacLean, I. J.; Warne, G. L.; Zajac, J. D. J. Steroid Bio-
chem. Molec. Biol. 1997, 62, 233.
25. Donleavy, J. J.; Gilbert, E. E. J. Amer. Chem. Soc. 1937,
59, 1072.
5. Hashimoto, Y.; Shudo, K. Cell Biol. Rev. 1991, 25, 209.
6. Hashimoto, Y. Cell Struct. Funct. 1991, 16, 113.
7. Wurtz, J.; Bourguet, W.; Renaud, J.; Vivat, V.; Chambon,
P.; Moras, D.; Gronemeyer, H. Nature Struct. Biol. 1996, 3, 87.
8. Eyrolles, L.; Kawachi, E.; Matsushima, Y.; Nakajima, O.;
Kagechika, H.; Hashimoto, Y.; Shudo, K. Med. Chem. Res.
1992, 2, 361.
9. Matias, P. M.; Donner, P.; Coelho, R.; Thomaz, M.; Peix-
oto, C.; Macedo, S.; Otto, N.; Joschko, S.; Scholz, P.; Wegg,
A.; Basler, S.; Schfer, M.; Egner, U.; Carrondo, M. A. J. Biol.
Chem. 2000, 275, 26164.
10. Neumann, F. Horm. Metab. Res. 1977, 9, 1.
11. Moguilewsky, M.; Bouton, M. M. J. Steroid Biochem.
1988, 31, 699.
12. Singh, S. M.; Gauthier, S.; Labrie, F. Curr. Med. Chem.
2000, 7, 211.
26. Kasayama, S.; Saito, H.; Kouhara, H.; Sumitani, S.; Sato,
B. J. Cell. Physiol. 1993, 154, 254.
27. Hiraoka, D.; Nakamura, N.; Nishizawa, Y.; Uchida, N.;
Noguchi, S.; Matsumoto, K.; Sato, B. Cancer Res. 1987, 47,
6560.
28. Roehborn, C. G.; Zoppi, S.; Gruber, J. A.; Wilson, C. M.;
McPhaul, M. J. Mol. Cell. Endocrinol. 1992, 84, 1.
29. Beer, H.-D.; Florence, C.; Dammeier, J.; McGuire, L.;
Werner, S.; Duan, D. R. Oncogene 1997, 15, 2211.
30. Sawada, T.; Hashimoto, Y.; Li, Y.; Kobayashi, H.; Iwa-
saki, S. Biochem. Biophys. Res. Commun. 1991, 178, 558.
31. Sawada, T.; Kobayashi, H.; Hashimoto, Y.; Iwasaki, S.
Biochem. Pharmacol. 1993, 45, 1387.
32. Teutsch, G.; Goubet, F.; Battmann, T.; Bonfils, A.; Bou-
choux, F.; Cerede, E.; Gofflo, D.; Gaillard-Kelly, M.; Phili-
bert, D. J. Steroid Biochem. Molec. Biol. 1994, 48, 111.
33. Mukherjee, A.; Kirkovsky, L.; Yao, X. T.; Yates, R. C.;
Miller, D. O.; Dalton, J. T. Xenobiotica 1996, 26, 117.
13. Neri, R.; Peets, E.; Watnick, A. Biochem. Soc. Trans.
1979, 7, 565.