Diphenyl quinolines and isoquinolines: Synthesis and primary biological evaluation

Article abstract of DOI:10.1016/S0968-0896(00)00194-2

The synthesis of a series of 35 substituted 3,4-diphenyl quinolines and isoquinolines is described. The majority of these molecules differ from all other triphenylethylene based antiestrogens by a different spatial location of the aminoalkyl side chain. T

Full text of DOI:10.1016/S0968-0896(00)00194-2

Bioorganic & Medicinal Chemistry 8 (2000) 2629±2641
Diphenyl Quinolines and Isoquinolines: Synthesis and Primary
Biological Evaluation
Martine Croisy-Delcey,a,* Alain Croisy,b Daniele Carrez,b Christiane Huel,b
Angele Chiaroni,c Pierre Ducrot,a Emile Bisagni,a Lu Jind
and Guy Leclercqd
aUMR 176 CNRS Institut Curie-Recherche, Laboratoire Raymond Latarjet, BaÃt. 110-112,
Centre Universitaire, 91405 Orsay Cedex, France
b
Â
Unite 350 INSERM, Institut Curie-Recherche, Laboratoire Raymond Latarjet, Bat. 110-112,
Ã
Centre Universitaire, 91405 Orsay Cedex, France
cInstitut de Chimie des Substances Naturelles, CNRS, 91198 Gif-sur-Yvette, France
d
Â
Â
Laboratoire J.-C. Heuson de Cancerologie Mammaire, Service de Medecine Interne,
Â
Institut Jules Bordet, Rue Heger-Bordet, 1, 1000 Brussels, Belgium
Received 17 May 2000; accepted 11 July 2000
AbstractÐThe synthesis of a series of 35 substituted 3,4-diphenyl quinolines and isoquinolines is described. The majority of these
molecules di€er from all other triphenylethylene based antiestrogens by a di€erent spatial location of the aminoalkyl side chain.
The binding anity of the most representative molecules (8, 9, 19, 20, 21, 23 and 25), including analogues 8 and 21 without the side
chain, for the estrogen receptor a (ER)y was determined. The ability of these molecules to induce the progesterone receptor was also
studied. Antiproliferative activity was evaluated on MCF-7 human breast cancer cells, while intrinsic cytotoxic/cytostatic properties
resulting from interaction with other targets than ER were assayed on L1210 murine leukemia cells. Introduction of an aminoalk-
ylamino side chain at carbon 2 confers strong cytotoxic properties to diphenylquinolines 9 and 10 as well as pure antiestrogenic
activities. However, cytotoxicity is so high with respect to antiestrogenicity that the latter was clearly observable only in one case
(9b). The structure of compound 9b was determined by X-ray crystallography. Molecular modeling of its docking within the hor-
mone-binding domain of the receptor was subsequently undertaken. According to our results, the design of molecules with the side
chain bound to the ethylene part of the triphenyl ethylene skeleton might generate compounds of potential pharmacological inter-
est. # 2000 Elsevier Science Ltd. All rights reserved.
Introduction
clinical trials are in progress to evaluate the use of
tamoxifen in chemoprevention for women at high risk of
breast cancer.5 7
Tamoxifen 1 is one of the most widely prescribed anti-
cancer drugs (for reviews, see refs 1, 2). This triar-
ylethylene compound, which elicits partial antiestrogenic
activity, is today the drug of choice for palliative ther-
apy of advanced breast cancer.3,4 It is also an ecient
adjuvant therapy for early stage disease.1 Furthermore,
Unfortunately tamoxifen resistance can develop which
limits its eciency.8 Side e€ects, including endometrial
carcinoma, are also sometimes observed as a major
adverse consequence of drug treatment.9 11 In response
to this situation, the search for antiestrogens without
residual estrogenic activity has developed considerably
over the last 10 years. In this context, we recently
described the synthesis of a series of 3,4-diphenyl iso-
quinolines,12 while Kihara and his group investigated a
series of 3,4-diphenyl 1,2-dihydroisoquinolines 2.13,14
Abbreviations: ER, estrogen receptor a; HBD, hormone binding
domain; E2, estradiol; DES, diethylstilbestrol; PgR, progesterone
receptor; OH-Tam, 4-hydroxytamoxifen; NMR, nuclear magnetic
resonance.
*Corresponding author. Tel.: +33-(0)1-69-86-30-89; e-mail: martine.
croisy@curie.u-psud.fr
yTwo forms of ER have been identi®ed (a and b). The present paper
refers to ERa since experiments were conducted with MCF-7 cells
which are known to solely express high amounts of this receptor form.
Several studies have clearly shown that tamoxifen inhi-
bits mammary tumor cell growth mainly through bind-
ing to their estrogen receptor a (ER). However, intrinsic
0968-0896/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(00)00194-2

2630
M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
toxic properties, detected in both ER-positive and ER-
negative tumors, may also contribute to the antitumor
potency of the drug. Indeed, interaction with targets of
lower binding anity may lead to cell death (cytotoxic
e€ect).15 Estrogen secretion and/or administration
abrogate ER-mediated e€ects of tamoxifen while main-
taining fully its cytotoxicity.
distinguished from other triphenylethylenic antiestro-
gens by a di€erent spatial location of their aminoalkyl
side chain in respect to the aromatic part of the mole-
cule (i.e., this chain is situated on the nitrogen hetero-
cycle bearing the ethylene group, rather than on one of
the phenyl groups). As a consequence, it is expected that
upon binding to ER they would confer an alternate
structural conformation to the HBD. The potential
in¯uence of such a change on the endocrinological/
antitumoral activity was assessed. The binding anity
for ER of compounds 9a, 9b, 19a, 19c, 19d, 19g, 19h,
20c, 20e, 23 and 25b, as well as compounds 8b, 8c and
21a±c without the crucial aminoalkyl side chain, was
determined. The capacity to activate ER (i.e. induction
of progesterone receptor), as well as antiproliferative
activity of these compounds on MCF-7 breast cancer
cells, was also analyzed. Potential cytotoxic activity
resulting from interaction with targets other than ER
was evaluated on L1210 cells.
The basic aminoalkyl side chain of tamoxifen (i.e.,
-OCH2CH2N(Me)2) plays a major role in its anti-
estrogenic/antitumoral activity. Indeed, analogues in
which this side chain is absent do not inhibit the growth
of mammary tumor cells in monolayer culture. Recent
structural studies concerning the interaction between
ER and 4-hydroxytamoxifen (OH-Tam), a metabolite
showing a ꢀ100-fold higher binding anity for ER,
provide the ®rst explanation for this phenomenon.16
Transcriptional activation of the receptor is mediated
by separate regions (`activation functions' AF-1 and
AF-2), of which AF-2 is solely responsive to ligand
binding.17 Estrogenic ligands trigger AF-2 expression,
while OH-Tam does not, although as recently revealed
by crystal structure analysis and mutants selection, they
bind at the same site within the core of the hormone
binding domain (HBD).16,18 20 Data relative to both the
unliganded and the estradiol (E2)/diethylstilbestrol (DES)
saturated HBD indicate that estrogenic hormones
induce a dramatic repositioning of the HBD helix 12,
leading to an AF-2 surface appropriate for speci®c
interactions with co-activators.19 OH-Tam16 (as well as
the antiestrogen raloxifene 318) impedes such a dis-
placement of helix 12, maintaining the receptor in an
inactive status.
Chemistry
3,4-Diphenyl quinoline. As shown in Scheme 1, N-acyla-
tion of 2-amino-benzophenone 4 through reaction with
p-methoxyphenylacetyl chloride 5 provided compound
6 in 80% yield. Cyclization of this intermediate under
alkaline conditions gave diaryl quinolone 7,21 which was
converted to the chloro derivative 8a by reaction with
phosphorus oxychloride. By re¯uxing 8a with an excess
of the appropriate diamine, the substituted diaryl qui-
nolines 9a and 9b were obtained. O-Demethylation to
the phenol derivatives 10a,b was achieved by re¯uxing
9a,b in HBr for 6 h under a nitrogen atmosphere.
Reductive dechlorination of 8a (Zn, acetic acid, H2O,
re¯ux) a€orded 8b, which was subsequently demethy-
lated to 8c.
The key residues Glu-353, Arg-394 and Hist-524 in the
HBD are essential for the binding of E2, DES and the
estrogenic stilbene and stilbene-like core of OH-Tam
and raloxifene.18,19 Speci®c interactions of the amino-
alkyl side chain of these antiestrogens with residues in a
channel located at the entrance of the HBD (Asp-351,
Thr-347, Ala-350 and Trp-383) have also been estab-
lished.16 Certain of these contacts, including the interac-
tion of Asp-351 with the nitrogen atom of the terminal
dimethylamino group, are thought to be responsible for
the antiestrogen properties. Since the orientation of this
side chain within this channel is governed mainly by its
presence at a speci®c position on the triphenylethylene
core, it may be anticipated that its attachment at any
other position may induce a dramatic change in the
endocrinological pro®le. This hypothesis is analyzed in
the present study.
3,4-Diaryl isoquinolines. The synthesis of the mono- and
dimethoxy1-1-chloro-3,4-diarylisoquinolines 18a±c was
previously described.12 In essentially the same manner, the
preparation of trimethoxy derivative 18d was achieved
Herein, we describe the preparation and evaluation of a
series of 35 substituted 3,4-diphenyl quinolines and
isoquinolines. The majority of these molecules can be
Scheme 1. Reagents and conditions: (i) CH2Cl2, 0 ꢁC, 3 h; (ii) C2H5-
OH, KOH, re¯ux; (iii) POCl3; (iv) diamine, re¯ux; (v) AcOH, Zn,
75 ꢁC, 1 h; (vi) HBr, re¯ux.

M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
2631
(Scheme 2). Thus, the metallation of 2-bromo-5-meth-
oxybenzoic acid 11 using two molar equivalents of n-
butyllithium at 100 ꢁC followed by condensation with
4-methoxymethyl benzoate 12 a€orded 2-(4-methoxy-
benzoyl)-5-methoxybenzoic acid 13.22 Low temperature
was an important factor in this reaction since at more
elevated temperatures self-condensation occured,23
leading after acidic work up to a mixture of 13 and
phthalate 14, which are dicult to separate.
directly dehydrated (re¯uxing formic acid, 0.5 h) to iso-
quinolin-1-one 17. Compound 17 was converted to the
chloro derivative 18d.
Conversion of chloro compounds 18a±d to products
19a±i was achieved by re¯uxing with an excess of the
required diamine. Subsequent reductive dechlorination
of chloroisoquinolines 18a±d using Zn in warm acetic
acid gave the diaryl isoquinolines 21a±d. Demethylation
of 19c±i and 21b±d was accomplished as for 9a,b
a€ording 20c±i and 22b±d. (Scheme 3).
Under classical amide forming conditions, the acid chlo-
ride obtained from 13 (SOCl2 at room temperature) was
reacted with p-methoxybenzylamine to a€ord a mixture
of the expected compound N-(4-methoxybenzyl)-1-
hydroxy-1-(4-methoxyphenyl)-isoindol-3-one 15 and 6-
methoxy-2-(4-methoxybenzyl)-3-(4-methoxyphenyl)-3-(4-
methoxybenzylamino)-2,3-dihydro-isoindol-1-one 16 in
moderate yield (21 and 22%, respectively).
Compound 23, bearing the tamoxifen side chain, was
obtained by O-alkylation of 22b with 2-dimethylamino-
ethyl chloride.25
Condensation of diphenyl isoquinolin-1-one 24 with
dimethylaminoethylamine in DMF, in the presence of
K2CO3, gave 25a, which was demethylated under classical
conditions to 25b.
A more synthetically useful yield of 15 (89%) was
obtained when the transformation was carried out using
diphenyl phosphoryl azide (DPPA) in DMF at 0 ꢁC.24
As described earlier,12 the phthalimide salt generated by
the reaction of 15 with 4 equiv of lithium di-isopropyl-
amide in THF at room temperature underwent pyrrole
ring opening followed by subsequent closure to a mix-
ture of cis/trans 3,4-diaryl-tetrahydro-4-hydroxy-2H-
isoquinolin-1-ones. These isomeric products were
Biological evaluation
Binding anity for ER. At high concentration (10 mM),
the methoxy substituted compounds 8b, 9b, 19h and
21b,c as well as the chloroisoquinoline 21a produced a
signi®cant (30±70%) inhibition of [3H]E2 (0.005 mM)
binding to cytosolic ER from MCF-7 cells. At lower
concentrations, they were without e€ect, suggesting a
Scheme 2. Reagents and conditions: (i) n-BuLi,
2 equiv THF,
100 ꢁC; (ii) CH2Cl2, SOCl2, 4-methoxybenzylamine; (iii) DMF, 4-
methoxybenzylamine, DPPA, (C2H5)3N, 20 ꢁC, 6 h; (iv) THF, LDA
4 equiv, 78 ꢁC; (v) HCO2H, re¯ux, 0.5 h; (vi) POCl3, re¯ux, 3 h.
Scheme 3. Reagents and conditions: (i) diamine, re¯ux; (ii) HBr,
re¯ux; (iii) Zn, AcOH/H2O, re¯ux.

2632
M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
weak binding anity. All the other compounds investi-
gated (see Table 1), and in particular those with the cor-
responding free OH group, failed to display any binding
ability. The positioning of the oxygen substituent on the
phenyl group attached at position 3 of the quinoline (or
isoquinoline) nucleus, which is essentially opposite to
that for the hydroxyl group in typical antiestrogens, is
most probably at the origin of this di€erence in activity.
Indeed, the 4-OH group in OH-Tam occupies e€ectively
the same space as the 3-OH group in E2, whereas the
OCH3 groups in our molecules correlate best with the
17b-OH in E2. In E2, it is well established that transfer of
the acidic 3-OH proton to the protein plays an impor-
tant role in binding.
The weak estrogenic activity of several of these com-
pounds was re¯ected by their ability to stimulate MCF-
7 cell growth at 1 mM (% of 0.01 mM E2: 8b=102;
8c=85; 21a=73; 21b=92; 21c=82). At 10 mM this
stimulatory e€ect was either largely diminished or
totally absent. With the exception of 9b, which inhibited
growth by 40%, none of the other compounds displayed
any signi®cant activity at 1 mM, while only a few cells
were still adhering to the culture dishes at 10 mM, due to
a high cytotoxicity (see below).
The tamoxifen analogue 23 increased PgR concentra-
tion without a€ecting ER, which distinguishes its beha-
vior from that of the parent compound (see above). It
stimulated the growth of MCF-7 cells at 1 mM (40% of
optimal stimulatory e€ect of E2), but displayed an IC50
of 10 mM. It should therefore be considered as a weak
antiestrogen with residual estrogenic properties. Eval-
uated in parallel, OH-Tam (0.01 mM ) displayed a con-
ventional inhibitory potency (50% growth inhibition).
Estrogenic/antiestrogenic activity
The in¯uence of these representative compounds on ER
and PgR levels from MCF-7 cells was assessed. The
weak binding anity of compounds 8b and 21a±c for
ER was re¯ected by their ability, at 1 mM, to decrease its
concentration (down-regulation) and to induce PgR.26
Compound 8c, devoid of signi®cant binding anity for
ER (15% inhibition of [3H]E2 binding at 10 mM),
behaved similarly. At 0.1 mM, 8b, 21a and 21c still
induced PgR (36, 41 and 18% of optimal induction
produced by E2, respectively), but at 0.01 mM they were
ine€ective.
Antiproliferative activity
The cytostatic and/or cytotoxic properties of these
molecules were further evaluated on two cell lines: the
L1210 murine leukemia, which can be considered as a
good model for exploring the intrinsic toxicity of the
drugs, and the MCF-7 human mammary carcinoma,
which might give some speci®c response to molecules
interfering with the estrogen dependent growth process.
The IC50 (drug concentration inhibiting cell growth by
50%) data are presented in Table 2.
Compounds 9b and 19h, which also weakly bind to ER,
decreased its level without any induction of PgR. Such
behavior suggested an antiestrogenic activity. Interest-
ingly, compound 19g behaved similarly, while it failed
to signi®cantly bind to ER (22% inhibition of [3H]E2
binding at 10 mM). These observations are also sum-
marized in Table 1. For additional comparison, note
that at 0.1 mM OH-Tam invariably eliminated ER and
induced PgR by ꢀ50% of the optimal potency of E2.
In the quinoline series (compounds 8±10), the majority
of the tested substances had an IC50 in the micromolar
range (4 to 10 mM). Introduction of an aminoalkyl side
chain as well as replacing the OCH3 group by a free OH
residue did not signi®cantly improve either the growth
inhibition or the cytotoxic properties. Even 9b, which
appeared to have a slight binding anity for ER and
some antiestrogenic properties, elicited a moderate
growth inhibiting e€ect on MCF-7 cells. No signi®cant
di€erence was observed for the e€ect on L1210 and
MCF-7 lines excepted for compounds 8b and 8c, which
have, respectively, a 2 and 3 times higher IC50 on L1210
than on MCF-7.
Table 1. In¯uence of 1 mM of compounds on ER and PgR con-
centration in MCF-7 cells
Compound
ER
% of
controls
PgRa
% of optimal induction
produced by E2
E2 (0.0 1mM)
OH-Tam (0.1 mM)
8b
8c
9a
9b
19a
19c
19d
19g
19h
20c
20e
21a
21b
23
0
0
18
6
47
48
90
134
107
40
100
50
561
611
12
In the isoquinoline series (compounds 19±22), those
molecules without the aminoalkylamino side chain
(21a±c and 22b,c) were weakly cytostatic in the L1210
assay, with IC50s ranging from about 25 mM to more
than 100 mM, and relatively highly cytotoxic toward
MCF-7 cells. However, in this study, at low doses (0.1
to 1 mM), compounds 21b and 21c, and to a lesser extent
21a and 22b, increased signi®cantly the growth of this
cell line (data not shown), thus acting as e€ective estro-
genic molecules.
02
02
23
13
23
45
03
114
123
17
27
116
104
13
43
933
604
65
05
Introduction of a dimethylaminoalkylamino side chain
at the C1 position of the isoquinoline ring generally led
to an increase in the cytotoxic properties in both cell lines
with IC50 ranging from 2 to 10 mM. The trihydoxy deri-
vatives 20h,i were exceptions to this trend, having IC50s
25b
a5 independent experiments. PgR induction by E2 (% of controls) : 1
ꢂ 301 ; 2 ꢂ 792 ; 3 ꢂ 540 ; 4 ꢂ 520 ; 5 ꢂ 205.

M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
2633
Table 2. Cytotoxicity on L1210 murine leukemia and on MCF-7
human mammary carcinoma cells in vitro. Results are expressed as
IC50 in mmol
inhibiting pro®le comparable to that observed with
MCF-7 cells. On the other hand, tamoxifen appeared
less e€ective, con®rming the higher toxicity of 9b. A low
binding anity of 9b may explain this behavior.
Ref.
L1210
MCF-7
8b
8c
10.9
17
5.6
4.3
Structure of compound 9b
9a
9b
10a
10b
19a
19b
19c
19d
19e
19f
19g
19h
19i
20c
20d
20e
20f
20g
20h
20i
21a
21b
21c
22b
22c
22d
23
25b
Mitomycin
OH-Tam
Tamoxifen
10
16.6
7.1
8
6.5
4.3
3.1
3
3.1
2.4
2.4
2.4
6.4
5.9
3.4
3.2
6
7.5
7.4
37
33
9
8
6
9.7
The X-ray crystal structure for 9b revealed that the two
phenyl rings at C3 and C4 are respectively tilted by 110
and 105 degrees with respect to the quinoline system.
The chain substituted at C2 is disordered, existing pre-
dominantly in two possible six-membered ring `chair
forms', generated by a folding of the C24, C25, C26
atoms to permit H-bonding between the H atom of N23
with N27. This chair type six-membered ring structure is
nearly co-planar to the quinoline ring (dihedral angle of
20ꢁ and 7.8ꢁ for the two forms of the chain respectively;
Figure 1). This con®guration appears to be a general
characteristic of the three-carbon chain since the 1H
NMR signal for their NH is shifted slightly down®eld
(0.5 to 0.8 ppm) with respect to the two-carbon chain.
4.1
nta
5
3.6
2.7
5.1
2.4
2
2.3
2.5
nt
nt
7
6.2
4.2
9.1
3.5
110
140
>100
26
33
35
>100
nt
Starting from the recently published X-ray structures of
ER co-crystallized with either OH-Tam16 or ralox-
ifene,18 we carried out docking experiments of 9b within
the HBD of the receptor. Comparison of both ralox-
ifene and OH-Tam/ER structures shows how the HBD
pocket might accommodate the ligand. The raloxifene/
ER complex structure appeared more appropriate for
docking 9b because of its larger binding cavity. On this
basis, molecular modeling of the docking of 9b within
the receptor pocket was carried out using the protein
core of the raloxifene/ER complex.
31.5
82
9.5
>100
nt
nt
0.04
nt
4.5
nt
0.08
7.5 (4 in ref 31)
ant: not tested.
Whereas superposition of the side chains of 9b and
raloxifene (tail±tail orientation; Figure 2(a)) led to a
high anity complex, superposition of the triphenyl
ethylene skeleton (head±tail orientation; Figure 2(b))
resulted in far higher binding energies (ÁEꢂ21 kcal/
mol). Indeed, the more stable complexes in the head±tail
orientation have the amino side chain in a hydrophobic
pocket. In contrast, the tail±tail orientation places the
polar side chain toward the outside of the cavity. It is
therefore unambiguous that the tail±tail orientation is
far more accurate. From this result, it is suggested that a
biological e€ect might be observed for a three±carbon
side chain. The lower growth inhibition potency of the
two±carbon analogue 9a supports this view.
between 30 and 40 mM for MCF-7 cells. Surprisingly, in
this series, the hydroxy derivatives appeared to be less
toxic than the corresponding methoxylated compounds.
Antiestrogenic pro®le of 9b
The antiestrogen activity observed for compound 9b led
us to characterize it more fully. As shown in Table 3, at
1 mM, it produced a weak growth inhibition of MCF-7
cells. This e€ect was partly suppressed by 0.01 mM E2.
At the same concentration tamoxifen induced a sig-
ni®cant cytostatic activity which was inhibited by the
hormone. At 10 mM, 9b appeared more cytotoxic than
tamoxifen, since its growth inhibition could not be
antagonized by E2. In RTx6 cells (a tamoxifen resistant
clone from MCF-7 cells27), compound 9b had a growth
From these data we can conclude that 9b ®ts the receptor
pocket of ER. However, the OH group position of OH-
Table 3. Comparison of growth inhibitory potency of 9b and tamoxifen (% of control growth in the absence or presence of 0.01 mM estradiola,b
MCF-7 cells RTx6 cells
Tamoxifen Tamoxifen
)
9b
+ E2
+ E2
9b
+ E2
+ E2
1 mM
10 mM
73
4
77b
5
36
9
73
29
82
27
75
9
66
29
92
28
aAssessment of DNA amounts after 120 h of culture,32 culture performed in quadruplicate (mg DNA in controls, MCF-7=35.9Æ1.7,
RTx6=21.5Æ2.4).
bValues after corrections for stimulatory e€ect of E2.

2634
M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
resistance to tamoxifen (RTx6 cells), indicating that this
phenomenon is relevant to (a) target(s) non-related to
ER dependent systemic toxicity. On the other hand,
compound 23, bearing an aminoalkyl side chain at the
same position as tamoxifen, displayed an antiestrogenic
activity (growth inhibition of MCF-7 cells) without total
loss of estrogenic activity (PgR induction). This behavior
is reminiscent of that of conventional antiestrogens
(tamoxifen, nafoxidine, raloxifene). Thus, positioning of
the side chain at the ethylene part of the triphenyl ethyl-
ene skeleton might be a convenient way to attain anti-
estrogenic molecules, provided that the potency issue
can by addressed.
Modeling of the binding of OH-Tam and raloxifene to
ER revealed a speci®c orientation of their side chain
with respect to a channel located at the entrance of the
HBD.16,18 According to our data, this channel would
also accept the side chain of 9b, although the latter is
attached on the opposite side of the triphenyl ethylene-
like core of the molecule. Hence, interaction of the
amine function of this side chain with (a) residue(s) of
the channel (i.e., Asp 35118) would compensate for the
absence of the phenyl ring usually bearing this side
chain. Association with ER and thereby estrogen
antagonism would, however, be dramatically a€ected.
Hydrophobic substituents linked at position 11b of E2,
which more or less mimic this benzene ring, are indeed
known to strongly increase binding anity to ER.28,29
Although the observed decrease of binding anity may
hamper the design of `reversed' antiestrogens such as 9b,
one may assume that the introduction of small hydro-
phobic substituent in their side chain may overcome the
present de®ciency. E2 derivatives in which the 11b or 7a
hydrogen (corresponding position for 9b) is substituted
by a hydrophobic group or the side chain of tamoxifen
are, indeed, almost equivalent in terms of binding to
ER.30 Such reversed antiestrogens may di€er from those
presently available in terms of cytotoxicity, metabolism
and clearance. Therefore, one may suggest that linkage
of a side chain at the non-aromatic part of the ethyl-
triphenyl-ethylene skeleton may generate compounds of
potential pharmacological interest. E€orts in this direc-
tion are currently being made.
Figure 1. Computer-drawn ellipsoid representation of compound 9b
from X-ray data. The benzene solvent molecule linking two molecules
in the cell was omitted.
Tam is opposite to that of the OCH3 of 9b. This fact
may explain the low anity of 9b for ER as well as its
antiestrogenic property, free of residual estrogen activity.
This hypothesis is partly con®rmed by the lack of dif-
ference in the binding anity of 9b and its correspond-
ing OH analogue 10b (data not shown).
Discussion
Although they bind to ER with a very low binding a-
nity, diphenyl quinolines and isoquinolines which
incorporate the triphenyl ethylene structural motif can
interfere with estrogen dependent physiological processes.
In the absence of aminoalkylamine substitution, biolo-
gically active compounds of this study display weak
estrogenic properties. Introduction of an aminoalkyl-
amino side chain appears to confer strong cytotoxicity,
as well as pure antiestrogenic properties. However, cyto-
toxicity is so high that antiestrogenicity was detectable in
only one case (compound 9b). This compound appears
more toxic than tamoxifen, possibly due to its lower
binding anity to ER. Interestingly, it failed to overcome
Figure 2. Superposition of raloxifene (yellow) and 9b (gray) within the HBD domain of ER: (a) best docking hit in the tail±tail orientation and (b)
best docking hit in the head±tail orientation. Panel c shows the superposition of raloxifene (yellow), OH-Tam (green) and 9b (purple, tail±tail
orientation), under their lowest energy of interaction.

M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
2635
Experimental
H20 H60); 7.55 (d, 1H, J=2, 4 Hz, H6); 7.32 (d, 1H,
J=8.6 Hz, H3); 7.13 (dd, 1H, J=2.4, 8.4 Hz, H4); 6.89
(d, 2H, J=8.6 Hz, H30 H50); 3.90 (s, 3H, OCH3); 3.86 (s,
3H, OCH3). Anal. (C16H14O5) C, H, O.
Chemistry
1H and 13C NMR spectra were performed on Bruker
200 or 400 AC spectrometers operating at respectively
200 or 400 MHz (1H) and 50 or 100 MHz (13C ). Che-
mical shifts are reported in d (ppm) with the following
abbreviations: singlet (s), doublet (d), doublet doublet
(dd), triplet (t), multiplet (m), br s (broad singlet); Ph
refers to phenyl rings at positions 3 and 4 of the het-
erocyclic rings. Melting points were determined on
Reichert hot stage microscope and are uncorrected.
Microanalytical results (combustion analysis) were
obtained from CNRS, Institut des substances naturelles,
Gif sur Yvette; values for the mentioned elements are
within Æ0.4% of the theoretical values.
Where the temperature was allowed to rise above 90ꢁ
self-condensation led to another product, namely 5-
methoxy-2-[5-methoxy-1-(4-methoxyphenyl)-3-oxo-1,3-
dihydro-isobenzofuran-1-yl]-benzoic acid (14) as color-
ꢁ
1
less powder, mp 188 C (toluene); H NMR (CDCl3) d
7.59 (d, 1H, J=8.5 Hz, H7); 7.35 (d, 1H, J=2.2 Hz,
H30); 7.27±7.18 (m, 3H, H4 H40 H50); 7.11 (s, 1H, H1);
7.13 (d, 2H, J=8.5 Hz, H200 H600); 6.90 (q, 1H, J=2.6,
8.8 Hz, H6); 6.78 (d, 2H, H300 H500); 3.86 (s, 3H, OCH3
5); 3.82 (s, 3H, OCH3 50); 3.73 (s, 3H, OCH3 400). Anal.
(C24 H20O7) C, H.
N-(2-Benzoylphenyl)-2-(4-methoxyphenyl)-acetamide (6).
To a solution of 2-aminobenzophenone (4.0 g, 20 mmol)
in CH2Cl2 (100 mL), 4-methoxyphenyl acetyl chloride
(4.4 g, 24 mmol) was added dropwise at 0 ꢁC with stir-
ring. The reaction mixture was allowed to warm to rt,
stirred for 3 h and quenched with a saturated NaHCO3
solution. After partitioning the organic layer was dried
over MgSO4 and evaporated to a crude residue (5.6 g,
80%) which was used directly in the next step. An ana-
lytical sample was puri®ed by column chromatography
on silica gel eluting with CH2Cl2 to give pure 6 as yellow
3-Hydroxy-6-methoxy-2-(4-methoxybenzyl)-3-(4-methox-
yphenyl)-2,3-dihydro-isoindol-1-one (15). To a solution
of acid 13 (5.7 g, 20 mmol) and 4-methoxybenzylamine
(3.28 g, 24 mmol) in dry DMF (200 mL) at 0 ꢁC under
nitrogen, freshly distilled DPPA (12.65 g, 46 mmol,
10 mL) was added. The reaction mixture was cooled to
20ꢁ under stirring over 10 min, when thiethylamine
(10 mL, 7.2 g, 70 mmol) was added dropwise. After 6 h
the reaction was warmed to rt and DMF removed in
vacuo. Water was added and the mixture extracted with
CH2Cl2, which was then washed with water, dried and
evaporated. Crystallization from toluene gave 15 as
colorless microcrystals (6.9 g, 89%); mp 174 ꢁC. 1H
NMR (CDCl3) d 7.16 (m, 6H, H4 H7, H20 H60, H2 H6
benzyl); 6.95 (dd, 1H, J=2.4, 8 Hz, H5); 6.76 (d, 2H,
J=8.9 Hz, H30 H50); 6.66 (d, 2H, J=8.6 Hz, H3 H5 ben-
zyl); 4.65 (d, 1H, CH2); 4.02 (d, 1H, J=14.8 Hz, CH2);
3.80 (s, 3H, OCH3); 3.76 (s, 3H, OCH3); 3.69 (s, 3H,
OCH3); 3.07 (s, 1H, OH). Anal. (C24H23NO5) C, H, N.
1
oil. H NMR (DMSO-d6) d 10.26 (s, 1H, NH); 7.72±
7.47 (m, 8H); 7.42 (dd, 1H, J=1. 5, 7.61 Hz, H3); 7.34
(m, 1H, H4); 7.07 (d, 2H, J=8.7 Hz, H20 H60); 6.84 (d,
2H, H30 H50); 3.74 (s, 3H, OCH3); 3.41 (s, 2H, CH2).
3-(4-Methoxyphenyl)-4-phenyl-1H-quinolin-2-one (7). A
mixture of 6 (5.1 g, 15 mmol) and KOH (2.5 g, 44 mmol)
in absolute ethanol (150 mL) was re¯uxed for 1 h. On
cooling, the resultant precipitate was collected by ®ltra-
tion, washed with further ethanol and dried. Recrys-
tallization from ethanol a€orded pale yellow needles
(4.3 g, 89%), mp>320 ꢁC. 1H NMR (DMSO-d6) d 12.07
(s, 1H, NH); 7.57±7.44 (m, 1H, H8 or H5); 7.44±7.13
(m, 4H); 7.2±7.08 (m, 4H) ; 7.04 (d, 2H, J=8.6 Hz, H2
H6 Ph3); 6.73 (d, 2H, H3 H5 Ph3); 3.70 (s, 3H, OCH3).
Anal. (C22H17NO2+1/4 C2H5OH) C, H, N.
When the method previously described12 was employed,
15 was obtained (21%) and a second compound which
was isolated by chromatography on silica gel, eluting
with CH2Cl2 and identi®ed as 6-methoxy-2-(4-methoxy-
benzyl) - 3 - (4 - methoxybenzylamino) - 3 - (4 - methoxy-
phenyl)-2,3-dihydroisoindol-1-one (16) (22%). Crystal-
lization from toluene gave colorless microcrystals of mp
128 ꢁC. 1H NMR (CDCl3) d 7.38 (d, 1H, J=2.2 Hz, H7);
7.07 (d, 1H, J=8.5Hz, H4); 6.96 (dd, 1H, J=2.2, 8.5 Hz,
H5); 7.32 (d, 2H, J=8.5 Hz, H2 H6 phenyl); 7.27 (d, 2H,
J=8.3Hz, H2 H6 benzyl); 6.73 (d, 2H, J=8.3Hz, H2 H6
benzylamine); 6.81 (d, 2H, J=8.5Hz, H3 H5 phenyl); 6.75
(d, 2H, J=8.5 Hz, H3 H5 benzyl); 6.69 (d, 2H, J=8.5 Hz,
H3 H5 benzylamine); 5.07 (d, 1H, J=15 Hz, CH2 phenyl)
3.80 (d, 1H, J=15 Hz, CH2 benzyl); 3.84 (s, 3H, OCH3);
3.77 (s, 3H, OCH3); 3.74 (s, 6H, 2OCH3); 2.84 (t, 1H,
J=12Hz and 9 Hz, CH2 benzylamine); 2.70 (d, 1H,
J=12 Hz, CH2 benzylamine); 1.92 (d,1H, J=9Hz, NH).
13C NMR (CDCl3) d 168 (CˆO), 45 (CH2 benzylamine),
42 (CH2 benzyl). Anal. (C24H23NO5) C, H, N.
5 - Methoxy - 2 - (4 - methoxybenzoyl) - benzoic acid (13).
Lithium O-lithiobenzoate was prepared from 2-bromo-
5-methoxybenzoic acid 11 (5.78 g, 25 mmol) in dry THF
(100 mL) by addition of n-BuLi (25 mL of 2 M solution
in cyclohexane, 50 mmol). The mixture was protected
under nitrogen and the temperature was controlled by a
liquid nitrogen±diethyl ether bath maintained below
95 ꢁC over 1 h. 4-Methoxymethylbenzoate (4.15 g,
25 mmol) in THF (50 mL) was then added dropwise, the
mixture stirred for 2 h whereupon it was warmed to rt.
The reaction was quenched on pouring into 5% aqueous
HCl (100 mL) and extracted with CH2Cl2. The organic
layer was washed with H2O and extracted with 10%
NaOH solution. The aqueous layer was then acidi®ed,
extracted with CH2Cl2 and the organic layer dried
(MgSO4) before evaporation in vacuo. Crystallization
from toluene gave 13 as a colorless powder (3 g, 43%);
7-Methoxy-3,4-bis(4-methoxyphenyl)-2H-isoquinolin-1-
one (17). A solution of LDA mono tetrahydrofuran
(6.6 mL, 1.5 M in cyclohexane, 9.8 mmol) was added
dropwise to a stirred solution of 15 (1.0 g, 2.47 mmol) in
dry THF (30 mL) at 78 ꢁC under an argon atmosphere,
mp 179 ꢁC; H NMR (CDCl3) d 7.71 (d, 2H, J=8.6 Hz,
1

2636
M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
a highly blue color appearing immediately. The mixture
was allowed to warm to rt and left stirring overnight,
whereupon it was quenched with saturated NH4Cl solu-
tion and extracted with CH2Cl2. The organic layer was
washed with brine and dried over MgSO4 before solvent
removal in vacuo. The resultant residue was dehydrated
on re¯uxing in formic acid over 30 min. After cooling,
the mixture was evaporated under reduced pressure.
After partitioning between H2O and CH2Cl2, the
organic layer was washed with H2O, dried and evapo-
rated. The resultant crude quinolone 17 was crystallized
from toluene as colorless needles (480 mg, 50%), mp
233 C. H NMR (CDCl3) d 9.14 (br s, 1H, NH); 7.83
(d, 1H, J=2.4 Hz, H8); 7.30±7.95 (m, 6H); 6.82 (d, 2H,
J=8.4 Hz, H3 H5 Ph3); 6.73 (d, 2H, J=8.7 Hz, H30 H50
Ph4); 3.91 (s, 3H, OCH3); 3.78 (s, 3H, OCH3); 3.74 (s,
3H, OCH3). Anal. (C24H21NO4) C, H, N.
(40% w/v) This solution was then extracted with
CH2Cl2 and the organic layer washed with H2O, dried
(MgSO4), evaporated and the residue puri®ed by chro-
matography. Physical data are reported in Table 5.
General procedure for demethylation. The appropriate
ethers 8b, 9a,b, 19c±i and 21b,c (1.5 mmol) were re¯uxed
in HBr (10 mL) under nitrogen for 6 h (24 h for the tri-
methoxy derivatives) and, after cooling, the mixture was
basi®ed with NH4OH and the crude products puri®ed
by recrystallization from ethanol or methanol. Physical
data are reported in Table 6.
ꢁ
1
4-[4-(2-Dimethylaminoethoxy)phenyl]-3-phenylisoquino-
line (23). Compound 22b (0.8 g, 2.7 mmol) was gradu-
ally added with stirring to a suspension of NaH, freshly
washed with hexane (0.105 mg as 60% w/w mineral oil
dispersion, 3.8 mmol) in DMF (5 mL) at rt and allowed
to stand for 1.5 h. 2-Chloroethyldimethylamine (0.32 g,
3 mmol) (obtained by treatment of the corresponding
hydrochloride with 10% NaOH solution) was then
added and, following 2 h stirring at 100 ꢁC, the reaction
mixture was cooled and evaporated under reduced
pressure. The resultant residue was extracted with
CH2Cl2 and the organic layer washed with 2N HCl. The
aqueous layer was adjusted to pH 10 on addition of a
10% NaOH solution and extracted with CH2Cl2, dried,
evaporated and puri®ed by chromatography (Al2O3)
eluting with CH2Cl2. Compound 23 was obtained as
white needles on recrystallization from ethanol (200 mg,
20%), mp 115 ꢁC. 1H NMR (CDCl3) d 9.36 (s, 1H, H1);
8.08 (m, 1H, H8); 7.72 (m, 1H, H5); 7 62 (m, 2H, H6
H7); 7.39 (m, 2H, H2 H6 Ph3); 7.22 (m, 3H, H3 H5
Ph3); 7.15 (d, 2H, J=8.7 Hz, H2 H6 Ph4); 6.93 (d, 2H,
H3 H5 Ph4); 4.11 (t, 2H, O-CH2); 2.77 (t, 2H, N-CH2);
2.37 (s, 6H, N(CH3)2). Anal. (C25H24N2O) C, H, N.
1-Chloro-5-methoxy-3,4-bis(4-methoxyphenyl)-isoquino-
line (18d). A mixture of 17 (2.38 g, 6.2 mmol) in POCl3
(20 mL) was re¯uxed for 3 h. After cooling, excess POCl3
was evaporated in vacuo, and the residue poured into ice-
water to which was added a saturated aqueous K2CO3
solution. The mixture was extracted with CH2Cl2, the
organic layer washed twice with H2O, dried and evapo-
rated. The crude chloro compound was puri®ed by
chromatography on silica gel (eluting with CH2Cl2) and
crystallized froꢁm cyclohexane as colorless needles (2.1 g,
84%), mp 138 C. 1H NMR (CDCl3) d 7.57 (m, 2H, H5
H8); 7.29 (d, 2H, J=8.9 Hz, H2 H6 Ph3); 7.26 (m, 1H,
H6); 7.12 (d, 2H, J=8.8 Hz, H2 H6 Ph4) 6.91 (d, 2H,
J=8.8 Hz, H30 H50 Ph4); 6.73 (d, 2H, J=8.9 Hz, H3 H5
Ph3); 3.99 (s, 3H, OCH3); 3.84 (s, 3H, OCH3); 3.76 (s,
3H, OCH3). Anal. (C24H20 Cl NO3) C, H, N.
2-Chloro-3-(4-methoxyphenyl)-4-phenylquinoline (8a).
Using the procedure described above, 8a was obtained
from quinolone 7 as pale yellow needles (74%), mp
2-(2-Dimethylaminoethyl)-4-(4-hydroxyphenyl)-3-phenyl-
isoquinolin-1-one (25b). 4-(4-Methoxyphenyl)-3-phenyl-
isoquinolin-1-one 24 (1.0 g, 3 mmol) was dissolved in
dry DMF (36 mL) at 60 ꢁC. After cooling to rt the mix-
ture was sequentially treated with potassium carbonate
(5.0 g, 36 mmol) and diethylaminoethylamine dropwise
(0.45 g, 35 mmol), obtained from alkali treatment of
diethylaminoethylamine hydrochloride. The reaction
mixture was stirred at 40 ꢁC over 24 h, whereupon it was
cooled, ®ltered and DMF removed under reduced pres-
sure. The residue was extracted with CH2Cl2 and
washed with water, dried and evaporated. The solid was
recrystallized from ethanol to yield 25a as colorless nee-
dles (0,6g, 50%), mp 179 ꢁC. 1H NMR (CDCl3) d 8.55 (m,
1H, H8); 7.51 (m, 2H, H6 H7); 7.30±7.20 (m, 3H, H6, H2
H6 Ph3); 7.21±7.14 (m, 3H, H5, H3 H5 Ph3); 6.97 (d, 2H,
J=8.4 Hz, H2 H6 Ph4); 6.73 (d, 2H, H3 H5 Ph4); 4.04 (m,
2H, CH2a); 3.75 (s, 3H, OCH3); 2.52 (m, 2H, CH2b); 2.07
(s, 6H, N-(CH3)2). Anal. (C26H26N2O2) C, H, N.
ꢁ
1
199 C. H NMR (CDCl3) d 8.10 (1H, J=8 Hz, H8);
7.72 (m, 1H, H7); 7.49 (m, 2H, H6 H5); 7.26±7.08 (m,
5H Ph4); 7.02 (d, 2H, J=8.8 Hz, H2 H6 Ph3); 6.75 (d,
2H, H3 H5 Ph3); 3.96 (s, 3H, OCH3). Anal.
(C22H16ClNO) C, H, N.
General procedure for amino substituted quinolines and
isoquinolines (9a,b, 19a±i). A mixture of the relevant
chloro derivative (1.5 mmol) and the appropriate amine
(15 mL, large excess) was re¯uxed under nitrogen for 24h.
Evaporation of the diamine under reduced pressure pro-
vided a residue which was recrystallized from either etha-
nol or methanol. Trimethoxy derivatives were prepared in
an autoclave at 160 ꢁC over 3 days and the residues
obtained after the evaporation of the diamine were pur-
i®ed by chromatography on alumina and recrystallized
from cyclohexane. Physical data are reported in Table 4.
General procedure for diaryl quinoline and isoquinolines
(8b and 21a±d). The relevant chloro compounds, 8a
and 18a±d (1.5 mmol), were dissolved in a mixture of
acetic acid (20 mL) and water (1.7 mL) and heated to
75ꢁ, when powdered zinc (6 mmol) was added After 1 h
the reaction was quenched on addition of water and the
mixture made basic on addition of aqueous NaOH
Demethylation was performed as described above and
yielded the hydrobromide of hydroxy compound 25b
(yield 30%) as a colorless powder from methanol, mp
261 ꢁC. H NMR (DMSO-d6) d 9.39 (s, 1H, HBr); 9.32
1
(br s, 1H, OH); 8.41 (m, 1H, H8); 7.85±7.55 (m, 1H, H6
H7); 7.38 (s, 5H, Ph3); 7.15 (m, 1H, H5); 6.92 (d, 1H,

M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
Table 4. Physical data for methoxy derivatives 9a,b, 19a±i
2637
Compound no.
R1
H
R2
H
R3
n
Formulaa
Yield
(%)
Mp (ꢁC)
Solvent
NMR dH (CDCl3)
9a
OCH3
2
C26H27N3O
87
73
72
139
Methanol
7.78 (m, 1H, H8); 7.50 (m, 1H, H6); 7.31±7.17 (m, 4H, H5, H7,
H2H6 Ph4); 7.12±7.05 (m, 3H, H3H4H5 Ph4); 7.04 (d, 2H,
J=8.6 Hz, H2H6 Ph3); 6.76 (d, 2H, H3H5 Ph3); 5.06 (t, 1H,
NH); 3.74 (s, 3H, OCH3); 3.64 (q, 2H, J=6.3 and 11.6 Hz,
CH2a); 2.47 (t, 2H, J=6.3 Hz, CH2b); 2.19 (s, 6H, N(CH3)2).
9b
H
H
H
H
OCH3
3
2
C27H29N3O
137
Ethanol
7.76 (m, 1H, J=8.2 Hz, H8); 7.48 (m, 1H, J=1.5 Hz, 6.8 Hz, H6);
7.25±7.17 (m, 4H, H5H7, H2H6 Ph4); 7.10 (m, 5H, H3H4H5
Ph4, H2H6 Ph3); 6.76 (d, 2H, J=8.75 Hz, H3H5 Ph3); 5.83 (t,
1H, NH); 3.74 (s, 3H, OCH3), 3.63 (m, 2H, CH2a); 2.29 (t, 2H,
J=6.3 Hz, CH2g); 1.97 (s, 6H, N(CH3)2); 1.72 (m, 2H, CH2b).
19a
H
C25H25N3
119
Methanol
7.90 (m, 1H, H8); 7.59±7.12 (m, 13H); 6.10 (t, 1H, NH); 3.77 (q,
2H, J=10.5 and 6.3 Hz, CH2a); 2.67 (t, 2H, J=6.3 Hz, CH2b);
2.33 (s, 6H, N(CH3)2).
19b
19c
H
H
H
H
H
3
2
C26H27N3
82
91
160
7.77 (m, 1H, H8); 7.51±7.12 (m, 14H, HAr, NH); 3.79 (m, 2H,
CH2a); 2.56 (t, 2H, J=6 Hz, CH2g); 2.36 (s, 6H, N(CH3)2); 1.91
(m, 2H, CH2b).
7.89 (m, 1H, H8); 7.60±7.40 (m, 5H, H5H6H7, H2H6 Ph3);
7.20±7.08 (m, 5H, H3H4H5 Ph3, H2H6 Ph4); 6.88 (d, 2H,
J=8.7 Hz, H2H6 Ph4); 6.05 (t, 1H, NH); 3.82 (s, 3H, OCH3);
3.76 (m, 2H, CH2a); 2.66 (t, 2H, J=5.9 Hz, CH2b); 2.32 (s,
6H, N(CH3)2).
OCH3
C26H27N3O
164
Ethanol
19d
H
OCH3
H
3
C27H29N3O
78
158
Ethanol
7.75 (m, 1H, H8); 7.54±7.37 (m, 6H, H5H6H7, H2H6 Ph3, NH);
7.21±7.08 (m, 5H, H3H4H5 Ph3, H2H6 Ph4); 6.86 (d, 2H,
J=8.7 Hz, H3H5 Ph4); 3.82 (s, 3H, OCH3); 3.80 (m, 2H, CH2a);
2.55 (t, 2H, J=6 Hz, CH2g); 2.35 (s,6H, N(CH3)2); 1.90
(m, 2H, CH2b).
19e
19f
H
H
H
OCH3
3
2
C27H29N3O
87
78
154
Methanol
7.74 (m, 1H, H8); 7.52±7.15 (m, 11H, 10 HAr, NH); 6.68 (m,
2H, J=8.8 Hz, H3H5 Ph3); 3.79 (m, 2H, CH2a) ,3.74 (s, 3H,
OCH3); 2.55 (t, 2H, CH2g); 2.35 (s, 6H, N(CH3)2); 1.90 (m,
2H, CH2b).
OCH3 OCH3
C27H29N3O2
162
Methanol
7.89 (m, 1H, H8); 7.50±7.40 (m, 5H, H7H6H5, H2H6 Ph3);
7.13 (d, 2H, J=8.7 Hz, H2H6 Ph4); 6.90 (d, 2H, H3H5 Ph4);
6.72 (d, 2H, J=8.8 Hz, H3H5 Ph3); 6.09 (m, 1H, NH); 3.84
(s, 3H, OCH3); 3.79 (m, 2H, CH2a); 3.76 (s, 3H, OCH3); 2.78
(t, 2H, J=6 Hz, CH2b); 2.33 (s, 6H, N(CH3)2).
19g
19h
19i
H
OCH3 OCH3
3
2
3
C28H31N3O2
C28H31N3O3
C29H33N3O3
77
65
64
142
Methanol
7.73 (m, 1H, H8); 7.50±7.34 (m, 6H, H5H6H7, H2H6 Ph3, NH);
7.12 (d, 2H, J=8.6 Hz, H2H6 Ph4); 6.88 (d, 2H, H3H5 Ph4); 6.70
(d, 2H, J=8.7 Hz, H3H5 Ph3); 3.83 (s, 3H, OCH3); 3.79 (m, 2H,
CH2a); 3.75 (s, 3H, OCH3); 2.55 (t, 2H, J=6 Hz, CH2g); 2.35
(s, 6H, N(CH3)2); 1.90 (m, 2H, CH2b).
OCH3 OCH3 OCH3
168
7.57 (d, 1H, J=9 Hz, H5); 7.47 (d, 2H, J=8.9 Hz, H2H6 Ph3);
Cyclohexane 7.27 (m, 4H, H8H6, H2H6 Ph4); 6.98 (d, 2H, J=8.7 Hz, H3H5
Ph4); 6.80 (d, 2H, H3H5 Ph3); 5.91 (t, 1H, NH); 4.04 (s, 3H,
OCH3); 3.93 (s, 3H, OCH3); 3.89 (m, 2H, CH2a); 3.85 (s, 3H,
OCH3); 2.77 (t, 2H, J=5.8 Hz, CH2b); 2.42 (s, 6H, N(CH3)2).
OCH3 OCH3 OCH3
138
7.65 (m, 1H, H8); 7.48 (t, 1H, NH); 7.35±7.19 (m, 4H, H5H6,
Cyclohexane H2H6 Ph3); 7.10 (d, 2H, J=8.2 Hz, H2H6 Ph4); 6.97 (d, 2H,
H3H5 Ph4); 6.75 (d, 2H, J=8.8 Hz, H3H5 Ph3); 3.94 (s, 3H,
OCH3); 3.81 (s, 3H, OCH3); 3.73 (s, 3H, OCH3); 3.55 (m, 2H,
CH2a); 2.39 (m, 2H, CH2g); 2.20 (s, 6H, N(CH3)2); 1.90
(m, 2H, CH2b).
aElemental analyses (C, H, N) of all compounds agreed to within Æ0.4% of theoretical values.
J=8.4 Hz, H2 H6 Ph4); 6.65 (d, 1H, H3 H5 Ph4); 4.10
(t, 2H, CH2a); 3.22 (t, 2H, CH2b); 2.75 (s, 6H,
N(CH3)2). Anal. (C25H24N2O2; HBr) Br.
MI). They were maintained in monolayer culture at 37 ꢁC
in Earle's base MEM containing 10% heat inactivated
(56 ꢁC, 1 h) fetal calf serum (FCS).
Tamoxifen resistant RTx6 cells (from Dr. J. C. Faye
INSERM U 168, Toulouse, France) were maintained under
the same conditions in RPMI 1640 in the presence of 1 mM
tamoxifen, which was removed before each experiment.
Biological studies
MCF-7 cells used for receptor assays and growth stimula-
tion were from the Michigan Cancer Foundation (Detroit,

2638
M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
Table 5. Physical data for derivatives 8b, 21a±c
Compound no.
R1
H
R2
H
R3
Formulaa
Yield Mp (ꢁC)
NMR dH (CDCl3)
(%)
Solvent
8b
OCH3 C22H17NO
62
162
8.98 (s,1H, H2); 8.17 (m, 1H, H8); 7.74±7.65 (m, 2H, H5H6);
7.46 (m, 1H, H7); 7.41±7.15 (m, 5H, Ph4); 7.10 (d, 2H, J=8.9
Hz, H2H6 Ph3); 6.77 (d, 2H, H3H5 Ph3); 3.77 (s, 3H, OCH3).
G. Berti33
9.34 (s, 1H, H1); 8.03 (m, 1H, H8); 7.75±7.52 (m, 3H, H5H6 H7);
7.42±7.32 (m, 2H, H2H6 Ph3); 7.28±7.18 (m, 3H, H3H4H5 Ph3)
7.15 (d, 2H, J=8.8 Hz, H2H6 Ph4); 6.90 (d, 2H, H3H5Ph4); 3.83
(s, 3H, OCH3).
21a
21b
H
H
H
OCH3
H
H
C22H17N
C22H17NO
45
40
155
123
21c
21d
H
OCH3 OCH3 C23H19NO
39
53
167
158
9.31 (s, 1H, H1); 8.00 (m, 1H, H8); 7.72±7.53 (m, 3H, H5H6H7);
7.32 (d, 2H, J=8.9 Hz, H2H6 Ph3); 7.16 (d, 2H, J=8.7 Hz, H2H6
Ph4); 6.92 (d, 2H, H3H5 Ph4); 6.75 (d, 2H, H3H5 Ph3); 3.85 (s,
3H, OCH3 Ph4); 3.77 (s, 3H, OCH3 Ph3).
9.02 (s, 1H, H1); 7.58 (d, 1H, J=8 Hz, H8); 7.35±7.20 (m, 4H,
H5H6, H2H6 Ph3); 7.14 (d, 2H, J=8.7 Hz, H2H6 Ph4); 6.91 (d,
2H, H3H5 Ph4); 6.74 (d, 2H, J=8.9 Hz, H3H5 Ph3); 3.96 (s, 3H,
OCH3 7); 2.34 (s, 3H, OCH3 Ph4); 2.26 (s, 3H, OCH3 Ph3).
OCH3 OCH3 OCH3 C23H19NO
aElemental analyses (C, H, N) of all compounds agreed to within Æ0.4% of theoretical values.
Cells were cultured at 37 ꢁC in a water-jacketed CO2
incubator (5% CO2) with l-glutamine, penicillin, strep-
tomycin and gentamicin at the usual concentrations (all
materials from Gibco, Life Technologies, Cergy-Pon-
toise, France).
medium was replaced. Cells were harvested 72 h later
and their growth evaluated by measuring their DNA
content by the diphenyl amine method. Each culture
was performed in quadruplicate.
Cytotoxicity (measurement of IC50 values) was deter-
mined in L1210 and MCF-7 cells by an investigator
(DC) ignorant of the growth data established on MCF-
7 cells by another investigator (GL). Murine leukemia
L1210 cells (ATCC-CCL 219) were cultivated in Dul-
becco's MEM supplemented with 10% FCS. Cells were
seeded at 105 cells/mL in 1-mL multiwell plates. After
24 h (usually 3 to 4Â105 cells/mL), tested compounds
were added in duplicate at various concentrations and
the plates incubated for 24 h before counting with a
Coulter-Counter ZM (Coultronics Inc.). The dose inhi-
biting the growth by 50% (IC50) was extrapolated from
regression curves obtained with experimental points free
of signi®cant toxicity. MCF-7 cells (ATCC HTB 22)
were cultivated in the same medium. Cells were seeded
in 1-mL multiwell plates (2Â105 cells/well). After 24 h,
tested compounds were added in triplicate at various
concentrations and cells were incubated for 7 days. Cell
density was then determined using the MTT assay.36
Absorbance at 540 nm was measured with a Bio-Rad
microplate reader (Model 450). Survival was expressed as
% of untreated controls and IC50 was extrapolated from
regression curves obtained from experimental points.
Measurement of binding anity of compounds for ER,
modulation of ER and PgR levels as well as growth sti-
mulation were performed according to experimental
protocols previously described.26,32,34
Measurement of binding anity for ER
Cytosol of MCF-7 cells was overnight incubated at 0±
4 ꢁC with 5 nM [3H]E2 in the absence or presence of
increasing amounts of either unlabeled E2 (control) or
the investigated compound (1 nM to 10 mM). Unbound
ligands were then removed by a dextran-coated charcoal
(DCC) treatment and the residual radioactivity mea-
sured by scintillation counting.
E€ect on ER and PgR levels26
Measurements were carried out on cytosols from MCF-
7 cells incubated for 3 days in the absence (control) or
presence of increasing amounts (range 0.01 to 1 mM) of
the investigated compound.
Cytosolic ER levels were measured by Scatchard plot
analysis according to the procedure established by the
EORTC Receptor Group35 using [3H]E2 as the labeling
agent. Progesterone receptors (PgR) were measured
according to this procedure using [3H]ORG-2058 as
labeled ligand.
X-ray crystal structure determination of 9b. Colorless
crystal (0.13Â0.16Â0.53mm) (benzene): C27H29N3O, 0.5
C6H6, Mw=450.59, triclinic system, space group P-1,
Z=2, a=9.971 (6), b=10.213 (8), c=13.643 (12) A,
a=106.94 (6), b=89.04 (5), g=103.42 (4)ꢁ, V=1290.8 A3,
dc=1.16 g cm 3, F(000)=482, l (Cu Ka)=1.5418 A,
m=0.55 mm 1; 4913 intensities measured on a Nonius
CAD-4 di€ractometer ( 114h411, 124k411, l: 0±
16) of which 4694 unique (Rint=0.022) and 3681
observed with Fo=4s(Fo). The structure was solved by
direct methods using SHELXS-8637 and re®ned by full-
matrix least squares based upon a unique Fo2 with
SHELXL-93.38 The N,N-dimethyldiamine chain sub-
stituent at C2 was found to be disordered, existing in
E€ect on cell growth
E€ect of the investigated compounds on breast cancer
cell lines was assessed after 120 h of culture.32 Cells were
plated in 25 mm Petri dishes (MCF-7: 2Â104 cells/mL;
RTx6: 4Â104 cells/mL) in 10% DCC treated serum
(DCC-FCS in MEM or RPMI). After 24 h, tested com-
pound was added to the medium and 48 h later the

M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
Table 6. Physical data for hydroxy compounds
2639
Compound no. R1 R2 R3
n
Formulaa
Yield Mp (ꢁC)
(%) Solvent
NMR dH (DMSO)
8c
H
H
H
H
OH
OH
Ð
C
21H15NO
63
239
9.53 (br s, 1H, OH); 8.96 (s, 1H, H2); 8.14 (m, 1H, H8); 7.80
(m, 1H, H6); 7.65±7.50 (m, 2H, H5H7); 7.51±7.20 (m, 5H, Ph4);
7.05 (d, 2H, J=8.6 Hz, H2H6 Ph3); 6.68 (d, 2H, H3H5 Ph3).
9.46 (s, 1H, OH); 7.7 (m, 1H, H8); 7.58 (m, 1H, H6); 7.35±7.22
(m, 3H, H5, H2H6 Ph4); 7.17±7.07 (m, 4H, H7, H3H4H5 Ph4);
6.95 (d, 2H, J=8.4 Hz, H2H6 Ph3); 6.68 (d, 2H, H3H5 Ph3);
6.00 (t, 1H, NH); 3.77 (m, 2H, CH2a); 2.95 (s, 6H, N(CH3)2);
2.53 (m, 2H, CH2b).
9.41 (s, 1H, OH); 7.63 (m, 1H, J=8.2 Hz, H8); 7.50 (m, 1H, H6);
7.38±7.20 (m, 3H, H5, H2H6 Ph4); 7.20-7.00 (m, 4H, H7, H3H4H5
Ph4); 6.93 (d, 2H, J=8.6 Hz, H2H6 Ph3); 6.67 (d, 2H, H3H5 Ph3);
6.27 (t, 1H, NH); 3.52 (m, 2H, CH2a); 2.53 (s, 6H, N(CH3)2); 2.30
(t, 2H, CH2g); 1.66 (m, 2H, CH2b).
10a
2
3
2
3
C
C
25H25N3O HBr
55
235
10b
20c
20d
H
H
H
H
OH
OH
H
OH
H
26H27N3O HBr
64
65
70
203
263
261
C26H25N3O HBr
C26H27N3O HBr
10.09 (br s, 1H, HBr); 9.48 (s, 1H, OH); 8.39 (m, 1H, H8); 7.97
(t, 1H, NH); 7.75±7.55 (m, 2H, H5H6); 7.52±7.35 (m, 3H, H7,
H2H6 Ph3); 7.32±7.08 (m, 3H, H3H4H5 Ph3); 6.97 (d, 2H, J=8.4
Hz, H2H6 Ph4); 6.79 (d, 2H, H3H5 Ph4); 3.93 (m, 2H, CH2a);
3.46 (t, 2H, CH2b); 2.83 (s, 6H, N(CH3)2).
H
9.46 (s, 1H, OH); 8.32 (m, 1H, H8); 7.78 (t, 1H, NH); 7.70±7.55
(m, 2H, H5H6); 7.52-7.37 (m, 3H, H7, H2H6 Ph3); 7.36±7.20 (m,
3H, H3H4H5 Ph3); 6.97 (d, 2H, J=8.4 Hz, H2H6 Ph4); 6.77 (d, 2H,
H3H5 Ph4); 3.66 (m, 2H, CH2a); 3.09 (m, 2H, CH2g); 2.58 (s, 6H,
N(CH3)2); 2.06 (m, 2H, CH2b).
9.96 (br s, 1H, HBr); 9.48 (s, 1H, OH); 8.33 (m, 1H, J=7.7 Hz, H8);
7.85 (t, 1H, NH); 7.70±7.15 (m, 10H); 6.61 (d, 2H, J=8.6 Hz, H3H5
Ph3); 3.70 (m, 2H, CH2a); 3.16 (t, 2H, CH2g); 2.61 (s, 6H, N(CH3)2);
2.10 (m, 2H, CH2b).
9.40 (br s, 2H, OHPh3, Ph4); 8.24 (m, 1H, J=7.6 Hz, H8); 7.60±7.32
(m, 4H, H5H6H7, NH); 7.23 (d, 2H, J=8.6 Hz, H2H6 Ph3); 6.96 (d,
2H, J=8.4 Hz, H2H6 Ph4); 6.80 (d, 2H, H3H5 Ph4); 6.57 (d, 2H,
H3H5 Ph3); 3.69 (m, 2H, CH2a); 2.62 (m, 2H, CH2b); 2.26 (s, 6H,
N(CH3)2).
10.07 (br s, 1H, HBr); 9.48 (s, 1H, OH Ph4); 9.46 (s, 1H, OH Ph3);
8.33 (m, 1H, J=7.8 Hz, H8); 7.82 (t, 1H, NH); 7.70±7.40 (m, 3H,
H5H6H7); 7.23 (d, 2H, J=8.6 Hz, H2H6 Ph3); 6.96 (d, 2H, J=8.2 Hz,
H2H6 Ph4); 6.80 (d, 2H, H3H5 Ph4); 6.64 (d, 2H, H3H5 Ph3); 3.70
(m, 2H, CH2a); 3.15 (t, 2H, CH2g); 2.59 (s, 6H, N(CH3)2); 2.08 (m,
2H, CH2b).
20e
20f
H
H
OH
3
2
C
26H27N3O HBr
57
75
272
257
OH OH
OH OH
C
25H25N3O2 HBr
20g
H
3
C26H27N3O2 HBr
70
296
20h
20i
OH OH OH
OH OH OH
2
3
C25H25N3O3
50
39
274
248
10.72 (br s, 1H, HBr); 9.93 (s, 1H, OH7); 9.43 (s, 1H, OH Ph4); 9.39
(s, 1H, OH Ph3); 7.60 (t, 1H, NH); 7.52 (m, 1H, H8); 7.38-7.12 (m,
4H, H5H6, H2H6 Ph3); 6.93 (d, 2H, J=8.2 Hz, H2H6 Ph4); 6.77 (d,
2H, H3H5 Ph4); 6.62 (d, 2H, J=8.5 Hz, H3H5 Ph3); 3.84 (m, H,
CH2a); 3.37 (m 2H, CH2b); 2.74 (s, 6H, N(CH3)2).
12.04 (br s, 1H, HBr); 10.59 (br s, 1H, OH7); 9.64 (br s, 2H, OH Ph4,
Ph3); 9.35 (br s, 1H, NH); 7.97 (m, 1H, H8); 7.55±7.31 (m, 2H, H5H6);
7.15 (d, 2H, J=8.6 Hz, H2H6 Ph3); 6.94 (d, 2H, J=8.5 Hz, H2H6
Ph4); 6.77 (d, 2H, H3H5 Ph4); 6.72 (d, 2H, H3H5 Ph3); 3.80 (m, 2H,
CH2a); 3.22 (m, 2H, CH2g); 2.84 (s, 6H, N (CH3)2); 2.15
(m, 2H, CH2b).
C
26H27N3O3
22b
22c
H
H
OH
H
Ð
Ð
C
21H15NO
63
65
259
9.61 (s, 1H, OH); 9.44 (s, 1H, H1); 8.24 (m, 1H, H8); 7.83±7.61 (m, 3H,
H5H6H7); 7.43±7.22 (m, 5H, Ph3); 7.06 (d, 2H, J=8.4 Hz, H2H6 Ph4);
6.82 (d, 2H, H3H5 Ph4).
OH OH
C21H15NO2
> 260 9.39 (s, 1H, H1); 8.19 (m, 1H, H8); 7.84±7.53 (m, 3H, H5H6H7); 7.22
(d, 2H, J=8.4 Hz, H2H6 Ph3); 7.05 (d, 2H, J=8.2 Hz, H2H6 Ph 4);
6.85 (d, 2H, H3H5 Ph4); 6.64 (d, 2H, H3H5 Ph3); 6.17 (br s for
exangeable protons HBr and OH).
22d
OH OH OH
Ð
C21H15NO3
59
> 260 10.00 (br s, 1H, OH7); 9.50 (br s, 2H, OH); 9.16 (s, 1H, H1); 7.46 (d,
1H, J=9 Hz, H5); 7.34 (d, 1H, J=1.9 Hz, H8); 7.28 (dd, 1H, J=9 and
1.9 Hz, H6); 7.16 (d, 2H, J=8.5 Hz, H2H6 Ph3); 7.02 (d, 2H, J=8.4 Hz,
H2H6 Ph4); 6.82 (d, 2H, H3H5 Ph4); 6.61 (d, 2H, H3H5 Ph3).
aElemental analyses (C, H, N) of all compounds agreed to within Æ0.4% of theoretical values.
two alternate chair-like structures (2/3:1/3 occupancy
for atoms C24 to C29). Atoms of weight 1/3 (C240±
C290) were re®ned only isotropically. Indeed, the split-
ting results in a `¯ip-¯ap' motion of atoms C25 and C26
around the pivot atoms C24 and N27, as shown by
deviations of these atoms (C25. . .C250=0.68, C26. . .
C260=0.91 A) larger than those of the other atoms
of the chain (C24. . .C240=0.18, N27. . .N270=0.36,
C28. . .C280=0.37, C29. . .C290=0.57 A). Each folding of
the chain allows formation of a strong intramolecular
hydrogen bond between nitrogen atoms N23 and N27
(mean distances N23. . .N27=2.884(4), HN23. . .N27=
2.20 A, angle N-H. . .N=136ꢁ). This H-bonding renders
the chain rigid. In addition, a benzene solvent molecule
was found around the symmetry center at (0, 1/2, 1/2),
linking the two molecules in the cell. All the H atoms

2640
M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
were located and ®tted at idealized positions, except
those of the minor chain (calculated). All hydrogens
were assigned an isotropic displacement parameter
equivalent to that of the bonded atom, plus 20% (or
30% for those of the methyl groups). Thus, re®nement
converged to R1(F)=0.0574 for the 3681 observed Fo,
and wR2(F2)=0.1861 for all the 4693 data with `good-
ness-of-®t' S=1.070. The residual electron density was
References
1. Furr, B. J. A.; Jordan, V. C. Pharmacol. Ther. 1985, 25,
127.
2. Roberts, N. J. Oncology (USA) 1997, 11 (2, suppl 1), 15.
3. Fornier, M.; Munster, P.; Seidman, A. D. Oncology (USA)
1999, 13, 647.
4. Kimmick, G. G.; Muss, H. B. Cancer Treat. Res. 1998, 94,
231.
3
found between 0.20 and 0.45 e A in the ®nal di€er-
ence map, near the benzene molecule. In the crystal,
only van der Waals contacts are observed. Full crystal-
lographic results have been deposited at the Cambridge
Crystallographic Data Centre (CCDC), UK, as Supple-
mentary Material (CIF ®le).
5. Fisher, B., Costantino, J. P.; Wickerham, D. L.; Redmond,
C. K.; Kavanah, M.; Cronin, W.M.; Vogel, V.; Robidoux, A.;
Dimitrov, N.; Atkins, J.; Daly, M.; Wieand, S.; Tan-Chiu, E.;
Ford, L.; Wolmark, N. and other National Surgical Adjuvant
Breast and Bowel Project Investigators; J. Natl. Cancer Inst.
1998, 90, 1371. See also the report of the Progress Review
Group of NCI related to the Phase III chemopreventive study
of breast cancer (charting the course: priorities for breast can-
cer research, August 1998, at http://wwwosp.nci.nih.gov/plan-
ning/prg/bprgprevention.htm).
6. Powles, T. J. Oncology (USA) 1997, 12 (3, suppl 5), 28.
7. McKean, V. A. J. Obstet. Gynecol. Neonatal Nurs. 1997, 26,
79.
8. Johnston, S. R. Anticancer Drugs 1997, 10, 911.
9. Carcangiu, M. L. Anat. Pathol. 1997, 2, 53.
10. Neveu, P.; Vergote, I. Curr. Opin. Obstet. Gynecol. 1998,
10, 9.
Molecular modeling calculations
Molecular modeling calculations were carried out using
SYBYL (SYBYL 6.5.2, Tripos Associates, Inc., St.
Louis, MO) on a Silicon Graphics workstation. Com-
pound 9b geometry issued from X-ray crystal structure
data was built from a standard fragment library, and its
geometry was optimized using Tripos Force Field39
without electrostatic terms. Minimization was carried
out using Powell's method until the gradient value was
11. Leo, L.; Tessarolo, M.; Echo, G.; Farina, C.; Nuzzo, L.;
Arduino, S.; Wierdis, T.; Lanza, A. Rev. Eur. J. Gynaecol.
Oncol. 1997, 18, 429.
12. Croisy-Delcey, M.; Huel, C.; Bisagni, E. Heterocycles
1995, 41, 1721.
.
smaller than 0.01 kcal/mol A. Conformational search
was then performed to identify lowest energy con-
formations using random search technique and the fol-
lowing
options:
maximum
hits=10,
RMS
threshold=0.2 A, and energy cut-o€=10 kcal/mol.
Each of the 149 conformers was then superposed onto
X-ray data extracted from the raloxifene study, using
the FIT option of SYBYL with centroids of cycles A
and C, and the side chain's nitrogen as template (tail±
tail orientation). After Gasteiger±Huckel charges were
added, the 37 conformations having a superposition
RMS less than 1.2 A2 were then docked within the
estrogen receptor a raloxifene binding site. The best 11
hits (binding energy within 10 kcal/mol) were subse-
quently minimized treating the receptor as ¯exible
within an 8 A sphere around the ligand.
13. Kihara, M.; Ikendri, M.; Nagao, Y. Drug Des. Discov.
1995, 12, 259.
14. Kihara, M.; Ikeuchi, M.; Yamanchi, A.; Nukatsuka,
M.; Matsumoto, H.; Toko, T. Chem. Pharm. Bull. 1997, 45,
939.
15. Wol€, D. M.; Fuqua, S. A. W. Cancer Treat. Rev. 1998,
21, 247.
16. Shiau, A. K.; Barstad, D.; Loria, P. M.; Cheng, L.;
Kushner, P. J.; Agard, D. A.; Greene, G. L. Cell 1998, 95,
927.
17. Kumar, V.; Green, S.; Stack, G.; Berry, M.; Jin, J. R.;
Chambon, P. Cell 1987, 51, 941.
18. Brzozowski, A.; Pike, A.; Dauter, Z.; Hubbard, R.; Bonn,
T.; Engstrom, O.; Ohman, L.; Greene, G.; Gustafsson, J.;
Cariquist, M. Nature 1997, 389, 753.
19. Tanenbaum, D. M.; Wang, Y.; Williams, S. P.; Sigler, P.
B. Proc. Natl. Acad. Sci. USA 1998, 95, 5998.
20. Wurtz, J. M.; Egner, U.; Heinrich, N.; Moras, D.; Muel-
ler-Fahrnow, A. J. Med. Chem. 1998, 41, 1803.
21. Marsili, A. Tetrahedron Letters 1963, 18, 1143.
22. Parham, W. E.; Bradsher, C. K.; Edgar, K. J. J. Org.
Chem. 1981, 46, 1057.
In parallel, the 149 conformations were superposed to
raloxifene using A, C and D aromatic centroõds (head±
tail). All conformations were docked manually within
the raloxifene-binding site, 85 of the conformations
were rejected and the 64 conformations left were mini-
mized keeping the receptor as rigid body. The 7 best hits
(binding energy within 20 kcal/mol) were subsequently
minimized treating the receptor as ¯exible within an 8 A
sphere around the ligand.
23. Parham, W. E.; Sayed, Y. A. J. Org. Chem. 1974, 39,
2051.
24. Shioiri, T.; Ninomiya, K.; Yamada, K. J. Am. Chem. Soc.
1972, 94, 6203.
25. Ogawa, K.; Mabushita, Y. I.; Yamawaki, I.; Kasseda, M.;
Shibata, J.; Toko, T.; Asao, T. Chem. Pharm. Bull. 1991, 39,
911.
Acknowledgements
26. Legros, N.; Leclercq, G.; McCague, R. Biochem. Pharma-
col. 1991, 42, 1837.
27. (a) Faye, J. C.; Jozan, S.; Redeuilh, G.; Baulieu, E. E.;
Bayard, F. Proc. Natl. Acad. Sci. USA 1983, 80, 3158. (b)
Borras, M.; Jin, L.; Bouhoute, A.; Legros, N.; Leclercq, G.
Biochem. Pharmacol. 1994, 48, 2015.
We are indebted to David Grierson for critical reading
of the manuscript. Biological studies conducted at the
I.J.B. were supported by funds MEDIC and SERVIER
Co. (Courbevoie, F.). L. Jin was a recipient of the
Fond J. C. Heuson de Cancerologie Mammaire. Mrs
Nathalie Marie is greatly acknowledged for editorial
assistance.
28. Napolitano, E.; Fiaschi, R.; Carlson, K. E.; Katzen-
ellenbogen, J. A. J. Med. Chem. 1995, 38, 429.

M. Croisy-Delcey et al. / Bioorg. Med. Chem. 8 (2000) 2629±2641
2641
29. Anstead, G. M.; Carlson, K. E.; Katzenellenbogen, J. A.
Steroids 1997, 62, 268.
30. Poupaert, J. H.; Lambert, D. M.; Vamecq, J.; Abul-Hajj,
Y. J. Bioorg. Med. Chem. Lett. 1995, 8, 839.
31. Stevens, M. F.; McCall, C. J.; Lelieveld, P.; Alexander, P.;
Richer, A.; Davis, D. E. J. Med. Chem. 1994, 37, 1689.
32. Leclercq, G.; Devleeschouwer, N.; Henson, J. C. J. Steroid
Biochem. 1983, 19, 75.
33. Berti, G.; Corti, P. Gazz. Chim. Ital. 1958, 88, 704.
34. Jin, L.; Borras, M.; Lacroix, M.; Legros, N.; Leclercq, G.
Steroids 1995, 60, 512.
36. Mosmann, T. J. Immunol. Meth. 1983, 65, 55.
37. Sheldrick, G. M. (1985) SHELXS-86. Program for the
Solution of Crystal Structures. Univ. of Gottingen, Germany.
38. Sheldrick, G. M. (1993) SHELXL-93. Program for the
Re®nement of Crystal Structures. Univ. of Gottingen, Ger-
many.
39. Clark, M.; Cramer, R. D. III; Van Opdenbosch, N. J.
Comput. Chem. 1989, 10, 982. Clark, M., Cramer, R. D. III;
Jones, D. M.; Patterson, D. E.; Simeroth, P. E. Tetrahedron
Comput. Methods 1990, 3, 47. 3D-QSAR in Drug Design. The-
ory, Methods and Applications; Kubinyi, H., Ed.; ESCOM:
Leiden, 1993. This includes many applications and cross-
references of the CoMFA methodology in medicinal chem-
istry.
35. EORTC Receptor Group, Revision of the standard for
the assessment of hormone receptors in human breast cancer.
Eur. J. Cancer 1980, 16, 1513.