A new class of nonsteroidal aromatase inhibitors: Design and synthesis of chromone and xanthone derivatives and inhibition of the P450 enzymes aromatase and 17α-hydroxylase/C17,20-lyase

Article abstract of DOI:10.1021/jm000955s

Aromatase (P450arom) is a target of pharmacological interest for the treatment of breast cancer. In this paper, we report the design, synthesis, and in vitro biological evaluation of a series of new (di)benzopyranone-based inhibitors of this enzyme. The d

Full text of DOI:10.1021/jm000955s

672
J . Med. Chem. 2001, 44, 672-680
A New Cla ss of Non ster oid a l Ar om a ta se In h ibitor s: Design a n d Syn th esis of
Ch r om on e a n d Xa n th on e Der iva tives a n d In h ibition of th e P 450 En zym es
Ar om a ta se a n d 17r-Hyd r oxyla se/C17,20-Lya se
Maurizio Recanatini,* Alessandra Bisi, Andrea Cavalli, Federica Belluti, Silvia Gobbi, Angela Rampa,
Piero Valenti, Martina Palzer,† Anja Palusczak,† and Rolf W. Hartmann†
Department of Pharmaceutical Sciences, University of Bologna, Via Belmeloro 6, I-40126 Bologna, Italy, and Pharmaceutical
and Medicinal Chemistry, Universita¨t des Saarlandes, P.O. Box 15 11 50, D-66041 Saarbru¨cken, Germany
Received March 28, 2000
Aromatase (P450arom) is a target of pharmacological interest for the treatment of breast cancer.
In this paper, we report the design, synthesis, and in vitro biological evaluation of a series of
new (di)benzopyranone-based inhibitors of this enzyme. The design of the new compounds was
guided by a CoMFA model previously developed for a series of nonsteroidal aromatase inhibitors.
Both the chromone and the xanthone nuclei were taken as molecular skeletons, and the
functions supposed to be critical for binding to the aromatase active site - a heterocyclic ring
(imidazole or 1,3,4-triazole) linked to the aromatic moiety by a methylene unit and an H-bond
accepting function (CN, NO2, Br) located on the aromatic ring at a suitable distance from the
heterocyclic nitrogen carrying the lone pair - were attached to them. The chromone, xanthone,
and flavone derivatives were prepared by conventional synthetic methods from the appropriate
methyl analogues. Aromatase inhibitory activities were determined by the method of Thompson
and Siiteri, using human placental microsomes and [1â,2â-3H]testosterone as the labeled
substrate. All the compounds were also tested on 17R-hydroxylase/C17,20-lyase (P450 17), an
enzyme of therapeutic interest for the treatment of prostatic diseases. The goal to find new
potent inhibitors of aromatase was reached with the xanthone derivatives 22d ,e (IC50 values
43 and 40 nM, respectively), which exceeded the potency of the known reference drug fadrozole
and also showed high selectivity with respect to P450 17. Moreover, compounds 22g-i based
on the same xanthonic nucleus showed fairly high potency as P450 17 inhibitors (IC50 values
220, 130, and 42 nM, respectively). Thus, they might be new leads for the development of drug
candidates for androgen-dependent diseases.
Ch a r t 1
In tr od u ction
The role of endogenous estrogens in the development
of hormone-dependent breast cancer has long been
recognized,1 and two main approaches have been de-
vised to antagonize the action of these hormones. It is
possible either to act directly at the estrogen receptor
by means of antagonists, among which tamoxifen is the
most popular one,2 or to block the biosynthesis of the
hormones, by inhibiting a key enzyme of the process.
In the latter strategy, since the early days, the research
efforts have been focused on aromatase,3 a cytochrome
P450 enzyme that catalyzes the conversion of androgens
into estrogens by aromatization of the steroid A ring.
Recently, however, reports showing the therapeutic
interest of other enzymes of the estrogen biosynthetic
pathway, like estrone sulfatase and estrone sulfotrans-
ferase, have appeared.4
via a competitive mechanism that involves the coordi-
nation of the heme iron. These compounds cause a
bathochromic shift in the UV absorption spectrum Soret
band of the heme deriving from the coordination of the
iron contained in the porphyrin ring by a heteroatom
(usually N, but also S or O) of the inhibitor.9
With the aim to design a new class of aromatase
inhibitors, we relied on (S)-fadrozole10 as the reference
compound and on some recently described comparative
molecular field analysis (CoMFA) models accounting for
the three-dimensional structure-activity relationships
(SAR) of a rather large series of nonsteroidal aromatase
inhibitors.11 As the molecular scaffolds on which to
insert the pharmacophoric functions typical of the azole-
Aromatase inhibitors can be both steroidal and non-
steroidal compounds.5 The latter class evolved from the
first marketed drug aminoglutethimide to the potent
and selective azoles, some of which have already been
marketed (fadrozole,6 anastrozole,6 letrozole7) or are in
phase III clinical trial (vorozole8) (Chart 1).
Most of the nonsteroidal aromatase inhibitors of
therapeutic importance act by binding to the enzyme
* To whom correspondence should be addressed. Tel: 051 2099700.
Fax: +39 051 2099734. E-mail: mreca@alma.unibo.it.
† Universita¨t des Saarlandes.
10.1021/jm000955s CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/24/2001

New Class of Nonsteroidal Aromatase Inhibitors
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 5 673
Ch a r t 2
static properties of both favorable and unfavorable zones
in the inhibitor space, after the compounds have been
aligned onto a template molecule on the basis of some
pharmacophoric hypothesis.22 As regards the nonste-
roidal aromatase inhibitors, the alignment was sug-
gested by a hypothesis of Furet et al.,10 who recognized
the possibility for (S)-fadrozole to bind to the enzyme
in a way similar to that of some steroidal competitive
inhibitors. According to this hypothesis, the imidazolic
N coordinates the heme iron and the CN function acts
as an H-bond acceptor in the same way as the 17-
carbonyl group of steroids does. Given that the clearest
(and statistically most significant) indication coming
from the CoMFA models for the nonsteroidal aromatase
inhibitors was the existence of a wide favorable steric
zone located around the phenyl ring of fadrozole-type
inhibitors,11 the consequent operative conclusion was to
design molecules able to fill this area.
type compounds (see below), we chose both the ben-
zopyran-4-one (chromone) molecule and its benzo ana-
logue, xanthone (Chart 2). Actually, it is known that
other benzopyranone derivatives, namely flavones, are
able to inhibit human aromatase.12 Moreover, in a
number of previous cases, we showed the versatility of
the benzopyranone moiety as a carrier of pharmacoph-
oric groups in as many different therapeutic fields as
those of analeptics,13 adrenergic â-blocking agents,14
clofibrate-like hypolipemic drugs,15 bradycardic16 and
antitumor17 agents, and acetylcholinesterase inhibi-
tors.18
An important criterion for assessing the possible
clinical usefulness of aromatase inhibitors in the treat-
ment of breast cancer is their selectivity with respect
to other P450 enzymes of physiological importance. On
the other hand, recently, it has been shown that some
aromatase inhibitors based on the benzocycloalkene
structure are able to inhibit 17R-hydroxylase/C17,20-
lyase (P450 17) as well.19 A finding like this is critical
and must be taken into consideration when designing
new analogues in view of SAR studies and also when
planning the development of the lead compound(s). In
fact, inhibitors of P450 17 have great therapeutic
interest, as they might be an alternative to antiandro-
gens in the treatment of androgen-dependent diseases
such as prostate cancer.20
In Table 1, the structures of the molecules studied as
aromatase inhibitors in the present paper are reported.
They were designed by taking the chromone and the
xanthone nuclei as molecular skeletons and by inserting
on them the functions supposed to be critical for the
binding to the aromatase active site. These functions
are (a) a heterocyclic ring (imidazole or 1,3,4-triazole)
linked to the aromatic moiety by a methylene unit and
(b) an H-bond accepting function (CN, NO2, Br) located
on the aromatic ring at a suitable distance from the
heterocyclic nitrogen atom carrying the lone pair.23
Compounds 22a -f are characterized by the presence
of an imidazole moiety linked through a methylene
bridge to position 8 of the chromone nucleus (22a -c)
or, correspondingly, to position 4 of the xanthone
nucleus (22d -f) (substituents R, Chart 2). These com-
pounds carry also an H-bond acceptor group or atom
(R′′) in a position para to the former substituent. In
compounds 22g-i, the positions of the two functions on
the aromatic nucleus of xanthone were inverted, to
investigate whether the presence of the ethereal oxygen
or, in turn, of the carbonyl group close to the electron-
attracting function could modulate the affinity for the
target enzyme.
In this paper, we report the design, synthesis, and in
vitro biological evaluation of a series of new (di)-
benzopyranonic aromatase inhibitors, together with a
discussion of their SARs. Based on the above consider-
ations, all the newly designed aromatase inhibitors were
also tested in a P450 17 inhibition assay, to obtain a
preliminary biological profile of the series indicating
which compounds might be worthy of further investiga-
tion and in which direction the optimization efforts must
be addressed.
Conformational analyses were carried out on all the
compounds with the aim to find an energetically acces-
sible conformation of each molecule that could allow an
overlapping of the two pharmacophoric functions on the
corresponding ones of (S)-fadrozole. It resulted that, for
the molecules in their putative active conformation, all
or part of both the chromone and xanthone moieties can
occupy the favorable CoMFA steric region (green);
moreover, no part of the molecules touches the unfavor-
able region (yellow), as shown in Figure 1a for com-
pound 22d .
In h ibitor Design
In previous studies, we attempted the rationalization
of the SARs of a rather large series of nonsteroidal
aromatase inhibitors in terms of their three-dimensional
properties.11 CoMFA analyses were carried out on series
of inhibitors belonging to two different structural
classes: one related to (S)-fadrozole and one related to
(E)-2-(4-pyridylmethylene)-1-tetralone.21 A comprehen-
sive CoMFA model was obtained accounting for the
three-dimensional SAR of the whole series and leading
to the individuation of regions in the space around the
molecules where their steric or electrostatic character-
istics cause variations of the inhibitory activity. The
design of the series presented here was guided by these
theoretical models, which, however, were not considered
as definitive and unmodifiable, such that some choice
was determined by classical SAR considerations.
In compounds 22j-m , the position of the five-
membered heterocycle on the (di)benzopyranonic ring
was again varied, and the possibility that the nuclear
carbonyl function could act as an H-bond acceptor was
explored. We verified that the introduction of the
coordinating heterocycle both in position 7 of the
chromone nucleus (R′ in 22j,k ) and in the corresponding
position 3 of the xanthone ring (R′ in 22l,m ) allowed
for the alignment of the carbonyl group of these mol-
ecules onto the CN function of (S)-fadrozole. In these
cases too, part of each molecule was engulfed in the
It is known that the CoMFA analysis can provide a
quantitative description in terms of steric and electro-

674 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 5
Recanatini et al.
Sch em e 1a
a
Reagents: (i) HNO3/H2SO4, 0-5 °C; (ii) N-bromosuccinimide,
CCl4, reflux 5 h; (iii) imidazole, CH3CN, reflux 6 h under N2; (iv)
H2, Pd/CaCO3; (v) NaNO2, HCl, 0-5 °C; (vi) NaCN/CuCN, 100 °C,
20 min; (vii) CuBr, rt, 12 h.
Sch em e 2a
F igu r e 1. Orthogonal views of the fit of (a) compound 22d
(cyan) and (b) compound 22j (magenta) into the CoMFA model
used for inhibitor design; the reference compound (S)-fadrozole
(white) is also shown superimposed to the new derivatives.
The occupancy of the large sterically favorable green volume
by part of the xanthone ring of 22d (a, left) and by the
chromone ring of 22j (b, left) is evident; the yellow sterically
forbidden volumes are not touched by any part of the inhibitors
(a,b, right).
favorable CoMFA volumes, thus satisfying the steric
requirement (Figure 1b, compound 22j).
Compounds 22n ,o were prepared in order to probe
the possibility that flavones fit the CoMFA model for
the inhibition of aromatase. In our extended CoMFA
model,11b a series of pyridyltetralones lacking a suitably
located H-bond acceptor function was aligned in an
alternative way with respect to (S)-fadrozole. In the
design hypothesis, compounds 22n ,o mimic the relative
orientation of the tetralones in the CoMFA model,
possessing the coordinating heterocycle but lacking the
suitably located H-bonding group.
The triazole analogues 22k ,m ,o were prepared with
the aim of exploring the effects of the variation of the
aza heterocycle. It has been reported that the replace-
ment of the imidazolic moiety with a triazolic one can
lead to better in vivo activity of the compound.23
a
Reagents: (i) Cu, nitrobenzene, K2CO3, 170 °C, 6 h; (ii) PPA;
(iii) N-bromosuccinimide, CCl4, reflux, 5 h; (iv) imidazole, CH3CN,
reflux 6 h under N2; (v) H2, Pd/CaCO3; (vi) NaNO2, HCl, 0-5 °C;
(vii) NaCN/CuCN, 100 °C, 20 min; (viii) CuBr, rt, 12 h.
cinimide and condensed with imidazole to give com-
pound 22b (Scheme 1). Reduction of compound 2 with
H2/Pd/CaCO3 afforded 5-amino-8-methyl-4H-benzopy-
ran-4-one (3) that was treated with an aqueous solution
of NaNO2, keeping the temperature below 5 °C: the
obtained diazonium salt was then cautiously added to
a boiling solution of NaCN/CuCN to give 5-cyano-8-
methyl-4H-benzopyran-4-one (4) or to a solution of CuBr
to yield 5-bromo-8-methyl-4H-benzopyran-4-one (5).
Compounds 22a ,c were obtained by bromination of
intermediates 4 and 5 followed by condensation with
imidazole.
Since several attempts to obtain 4-methyl-1-nitro-9H-
9-xanthenone (6) by nitration of 4-methyl-9H-9-xan-
thenone failed, this compound was synthesized from
2-methyl-5-nitrophenol, through reaction with O-chlo-
robenzoic acid in nitrobenzene affording the correspond-
Ch em istr y
The chromone, xanthone, and flavone derivatives of
Table 1 were prepared according to Schemes 1-4 from
the appropriate methyl analogues that were obtained
by conventional synthetic methods. To prepare the
xanthone derivatives 22d -f, an alternative route was
followed with respect to the corresponding chromones.
8-Methyl-4H-benzopyran-4-one (1) was nitrated with
HNO3 in H2SO4 to give 8-methyl-5-nitro-4H-benzopy-
ran-4-one (2), which was brominated with N-bromosuc-

New Class of Nonsteroidal Aromatase Inhibitors
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 5 675
Ta ble 1. Structure and Analytical Data of the Chromone (22a -c,j,k ,n ,o) and Xanthone (22d -i,l,m ) Derivatives
no.
scaffold
R
R′
R′′
CN
NO2
Br
CN
NO2
Br
CH2Im
CH2Im
CH2Im
H
R′′′
R′′′′
formulaa
mp (°C)
yield (%)
22a
22b
22c
22d
22e
22f
22g
22h
22i
22j
22k
22l
22m
22n
22o
chromone
chromone
chromone
xanthone
xanthone
xanthone
xanthone
xanthone
xanthone
chromone
chromone
xanthone
xanthone
chromone
chromone
CH2Imb
CH2Im
CH2Im
CH2Im
CH2Im
CH2Im
CN
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Ph
Ph
C14H9N3O2
C13H9N3O4
192-5c
207-8c
141-3c
255-7c
234-6c
218-20c
172-4d
152-4d
211-4d
216-8d
198-200d
158-62d
235-6d
188-9d
191-3d
50
55
60
40
55
40
40
45
40
50
40
40
40
45
40
C13H9BrN2O2
C18H11N3O2
C17H11N3O4
C17H11BrN2O2
C18H11N3O2
C17H11N3O4
C17H11BrN2O2
C13H10N2O2
C12H9N3O2
C17H12N2O2
C16H11N3O2
C20H16N2O3
C19H15N3O3
NO2
Br
H
H
H
H
H
H
CH2Im
CH2Trie
CH2Im
CH2Tri
OMe
H
H
H
H
CH2Im
CH2Tri
OMe
H
a
b
C, H, and N analyses were within (0.4% of the theoretical values. CH2Im: imidazol-1-ylmethyl. c Crystallizing solvent: petroleum
ether. Crystallizing solvent: ethanol. e CH2Tri: 1,3,4-triazol-1-ylmethyl.
d
Sch em e 3a
Sch em e 4a
a
Reagents: (i) N-bromosuccinimide, CCl4, reflux
5 h; (ii)
a
Reagents: (i) HNO3/H2SO4, 0-5 °C; (ii) N-bromosuccinimide,
imidazole/1,3,4-triazole, CH3CN, reflux 6 h under N2.
CCl4, reflux 5 h; (iii) imidazole, CH3CN, reflux 6 h under N2; (iv)
H2, Pd/CaCO3; (v) NaNO2, HCl, 0-5 °C; (vi) NaCN/CuCN, 100 °C,
20 min; (vii) CuBr, rt, 12 h.
19, or 21, respectively, which was then condensed with
imidazole or 1,3,4-triazole to give the desired com-
pounds.
ing diphenyl ether, and a subsequent cyclization by
means of polyphosphoric acid, as described in Scheme
2. Bromination with N-bromosuccinimide, followed by
reaction with imidazole, afforded compound 22e.
The nitro derivative 6 was hydrogenated to 1-amino-
4-methyl-9H-9-xanthenone (7) and treated, as previ-
ously described, with NaNO2 in the presence of NaCN/
CuCN or CuBr to give 4-methyl-9-oxo-9H-1-xanthene-
carbonitrile (8) or 1-bromo-4-methyl-9H-9-xanthenone
(9), respectively. Derivatives 8 and 9 were brominated,
and the resulting bromomethyl derivatives 15d ,f were
condensed with imidazole to afford 22d ,f, respectively
(Scheme 2).
The synthesis of compounds 22g-i was accomplished
as illustrated in Scheme 3, using the same procedure
described for compounds 22a -c (Scheme 1) and starting
from 1-methyl-9H-9-xanthenone (10).
Compounds 22j-o were prepared as shown in Scheme
4: the selected methyl derivative (16, 17, or 20) was
brominated with N-bromosuccinimide in refluxing CCl4
to yield the corresponding bromomethyl derivative 18,
Biologica l P r op er ties
In h ibition of P 450a r om . The method of Thompson
and Siiteri, i.e., human placental microsomes and
[1â,2â-3H]testosterone, was used for the determination
of inhibitory activities.24 In Table 2, IC50 values and
inhibitory potencies of the compounds relative to ami-
noglutethimide are shown; the IC50 value of fadrozole
tested in our test system25 is also reported for compari-
son.
Considering the inhibitors carrying the N-imidazolyl-
methyl substituent (R, Chart 2) both in position 8 of the
chromone ring (compounds 22a -c) and in position 4 of
the xanthone nucleus (compounds 22d -f), it is evident
that the xanthone-based molecules are much better
aromatase inhibitors than the chromone-based ones. As
R′′ substituents, CN and, in case of the xanthones, NO2
turned out to be superior to Br. Thus, compounds 22d ,e
are extremely potent aromatase inhibitors exceeding the

676 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 5
Recanatini et al.
Ta ble 2. Inhibition of the Steroidogenic P450 Enzymes
Aromatase and P450 17 by Chromone (22a -c,j,k ,n ,o) and
Xanthone (22d -i,l,m ) Derivatives
human aromatasec
human 17R-hydroxylased
a
a
% inhib IC50
rel
% inhib
(2.5 µM)
IC50
no.
22a
22b
22c
22d
22e
22f
22g
22h
22i
22j
22k
22l
(25 µM) (µM) potb
(µM)
98
95
94
99
99
99
93
95
95
98
85
97
27
94
21
99
0.24
1.1
2.1
0.043 430
0.040 463
0.089 208
76
17
9
3
5
26
14
4
38
92
94
98
9
5
33
8
44
14
2.3
8
19
20
185
5
0.22
0.13
0.042
0.97
0.94
0.10
3.7
F igu r e 2. Lineweaver-Burk plot for the inhibition of aro-
matase activity by compound 22e. Initial rates (V) are ex-
pressed as pmol/h/mg of 3H2O formed in the aromatization of
the tritiated substrate and were obtained as described in
ref 26.
0.39
47
22m
22n
22o
fadrozole
0.55
33
0.052 359
a
show only marginal inhibitory activity. The xanthone-
derived compounds 22g-i with the “inverted” position
of the substituents are rather strong inhibitors of P450
17, being 3-18 times (compounds 22g,i, respectively)
more potent than the reference compound ketoconazole
(IC50 ) 0.74 µM27).
The given values are mean values of at least three experi-
b
ments. The deviations were within (5%. Relative potency cal-
culated with respect to aminoglutethimide (IC50 ) 18.5 µM).
c Human placental microsomes, concentration of testosterone
(substrate): 2.5 µM; [1â,2â-3H]testosterone: 0.225 µCi. Human
d
testicular microsomes, concentration of progesterone (substrate):
2.5 µM.
Discu ssion
enzyme inhibitory activity of the well-known nonste-
roidal inhibitor fadrozole: their relative potencies with
respect to fadrozole are 1.2 and 1.3, respectively. The
exchange of the N-imidazolylmethyl substituent for the
electron-withdrawing group (R against R′′) dramatically
decreases the enzyme inhibitory activity (compounds
22g-i). Derivatives bearing the N-imidazolylmethyl
group as R′ substituent show an increase in aromatase
inhibition in the case of the chromone scaffold (22j) and
a decrease in the xanthone series (22l). The 3-imida-
zolylmethyl-substituted (R′′′, Chart 2) chromone 22n
shows mediocre aromatase inhibition. The replacement
of imidazole with 1,3,4-triazole in compounds 22j,l,n
strongly decreases the inhibitory activity (compounds
22k ,m ,o, respectively).
The mode of aromatase inhibition was investigated
more in depth for the most potent compound 22e. By
performing difference spectroscopy experiments, we
found that this derivative shows a characteristic type
II UV-vis difference spectrum (minimum at 390 nm,
maximum at 420 nm) indicating the formation of a
coordination bond between the imidazole nitrogen and
the heme iron of P450arom. A study of the steady-state
kinetics of 22e was then performed, using a microsomal
aromatase preparation derived from human placenta
and the tritiated water method.26 The xanthone deriva-
tive exhibited an inhibition constant (Ki) of 1.05 nM
obtained from a Lineweaver-Burk plot (Figure 2) that
revealed also a typical competitive type of inhibition.
The corresponding Km and Vmax values were 65.1 nM
and 353.2 pmol/h/mg, respectively.
The series of compounds whose structure and biologi-
cal activity are presented in Tables 1 and 2, respectively,
shows that it is possible to obtain potent inhibitors of
the P450arom and P450 17 by suitably modifying the
(di)benzopyran-4-one nucleus.
Our first goal was to find new potent inhibitors of
aromatase, and it was reached with compounds 22d ,e
that exceeded the potency of the marketed drug fadro-
zole (Afema). However, depending on the substitution
pattern, the aromatase inhibitory activity varies sub-
stantially, and the reasons thereof might be tentatively
looked for in some physicochemical properties of the
compounds. The most striking evidence is the dramatic
loss of activity that occurs upon replacement of the
imidazole ring of compounds 22j,l,n with the 1,3,4-
triazole in compounds 22k ,m ,o. This observation was
also made with other inhibitors of P450arom19c and
P450 17,28 and an explanation advanced for this finding
involved the different atomic charges of the heterocyclic
nitrogens responsible for the interaction with the heme
iron.28 The HOMO for compounds 22l,m calculated with
the SYBYL program using the AM1 method29 is shown
in Figure 3a,b, respectively. It appears evident that the
molecular orbital involved in the eventual coordination
bond with the iron atom is located on the imidazole ring
of compound 22l, but it is spread over the xanthone
nucleus in the case of 22m . The relative unavailability
of the nitrogen electron pair of the latter compound
explains reasonably well the low inhibitory activity of
the 1,3,4-triazol-1-yl-substituted derivatives. Regarding
the effects of different patterns of structural modifica-
tion, it appears from the data of Table 2 that the
combination xanthone, imidazol-1-ylmethyl, and strong
H-bond acceptor (CN or NO2) is the best one for an
efficient aromatase inhibition (compounds 22d -f). This
is true, considering both the xanthone/chromone re-
placement (compounds 22a -c) and the inversion of the
In h ibition of P 450 17. The assay was performed as
recently described by us, using human testicular mi-
crosomes as source of the enzyme and nonlabeled
progesterone as substrate.27 Percent inhibition values
at an inhibitor concentration of 2.5 µM (Table 2) indicate
that, with the exception of compounds 22g-i, all other
compounds do not significantly inhibit the enzyme or

New Class of Nonsteroidal Aromatase Inhibitors
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 5 677
molecules could serve as an H-bond acceptor either in
the same orientation of the fadrozole CN or in an
alternative one. This hypothesis seems to be confirmed.
However, it has to be noted that any conjecture about
the actual inhibitor binding mode remains speculative
until the structure of some ligand-enzyme complex is
resolved by an experimental method (X-ray crystal-
lography, NMR). Interestingly, for these three com-
pounds, the order of potency is reversed with respect to
lipophilicity; in fact, the calculated log P values for
compounds 22j,l,n are 0.11, 1.71, and 2.13, respectively.
As stated in the Introduction, to assess the selectivity
of the new aromatase inhibitors, we extended the
biological investigation of the compounds to another
P450 enzyme, namely P450 17. The data reported in
Table 2 show that indeed the most potent aromatase
inhibitors 22d -f are selective, as they do not inhibit
P450 17. On the other hand, it is quite remarkable that
the “inverted” xanthone derivatives 22g-i are good
inhibitors of 17R-hydroxylase, even so they are less
selective than compounds 22d -f. Interestingly, slight
structural modifications decrease P450arom inhibition
strongly and concurrently increase P450 17 inhibition
markedly. Similar evidences have been obtained by us
with 3- and 4-pyridyl-substituted tetrahydrocyclopropa-
naphthalenes,19b imidazolyl(methyl)-substituted tetra-
hydronaphthalenes,19c and imidazolylmethyl-substitut-
ed biphenyls.27 Considering the relevance of the target
and the high P450 17 inhibitory activity, and taking into
account also that compound 22i is rather selective with
respect to aromatase, it might be a lead for the develop-
ment of a drug candidate for prostatic diseases.
Finally, a comment about the use of the CoMFA
models as an aid in the design of the new aromatase
inhibitors is deserved. As stated above, we used theo-
retical models and classical SAR concepts in conjunction
to design the structural modification of the chromone
and xanthone nuclei. Perhaps, it would have been too
ambitious to expect accurate quantitative predictions
from the CoMFA models applied to molecules structur-
ally different enough from those of the training set.
Actually, the predicted activity data for aromatase
inhibition differ from about 0.2 up to more than 2.5 log
units from the experimentally observed values. The
reasons may be numerous, but they probably reside in
the inaccurate description of the molecular properties
by the steric and electrostatic CoMFA fields and in the
arbitrariness of the molecules’ alignment. Nonetheless,
the statistical model proved to be of help in supporting
some decision about structural modifications (substitu-
ent positions) and in verifying the consistency of some
molecules with the design hypothesis. The conclusion
might be that an appropriate balance of quantitative
and qualitative thinking can still be considered the best
approach for the molecular design applied to pharma-
cological targets.
F igu r e 3. HOMO of compounds 22l (a) and 22m (b), showing
the localization of the orbital over the imidazole ring and over
the xanthone nucleus, respectively.
positions of the imidazolylmethyl and electron-attract-
ing groups on the xanthone scaffold (compounds 22g-
i). A reason for the higher activity of the xanthone-based
compounds with respect to the chromone-based ones
might be the higher lipophilicity of the former deriva-
tives, which could facilitate the access and/or the
binding to the aromatase active site. We calculated the
log P values30 of compounds 22a -f: -0.45, -0.14, 0.98,
1.14, 1.45, and 2.57, respectively. From molecular
biology studies,31 it is known that the aromatase active
site contains several hydrophobic residues, and conse-
quently, the hydrophobic effect could play a critical role
in the stabilization of the enzyme-inhibitor complex.
The importance of hydrophobicity in the molecular
recognition was earlier pointed out by Hansch32 and
recently brought to the attention of drug designers by
Davis and Teague.33
Of course, lipophilicity cannot be the only driving force
leading to a strong binding to aromatase, as shown by
the good activity of the chromone derivative 22j (despite
its log P ) 0.11) and by the relatively lower potency of
the “inverted” xanthone derivatives 22g-i. For the
latter, which show hydrophobic and electronic properties
similar to those of compounds 22d -f, the only reason-
able hypothesis to explain the lower activity is the
possibility of some conformational restraint that might
prevent the molecules from assuming a conformation
that locates the pharmacophoric functions in optimal
spatial orientation for binding to the enzyme.
Con clu sion s
In conclusion, we have shown that by appropriately
modifying the (di)benzopyran-4-one nucleus, it is pos-
sible to obtain potent and selective aromatase inhibitors.
Compounds endowed with such an activity can be
considered for the development as anticancer drugs,
provided that they are selective in their action against
The fairly good inhibitory activity of compounds
22j,l,n shows that the presence of a strong H-bonding
function (like CN or NO2) is not strictly necessary to
obtain aromatase inhibition. Our idea (see Inhibitor
Design) was that the carbonyl group present in these

678 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 5
Recanatini et al.
saturated solution of NaHCO3. The suspension obtained was
filtered and acidified. The solid was collected by filtration,
washed with water and dried, obtaining 17 g (60%) of 2-(2-
methyl-5-nitrophenoxy)benzoic acid: mp 145-147 °C (ligroin);
1H NMR δ 2.45 (s, 3H), 6.9-8.2 (m, 7H, arom). 16.4 g (0.06
mol) of the previous compound was added portionwise to 240
mL of H2SO4 concentrated, keeping the temperature below 50
°C. The reaction mixture was then heated at 110 °C for 40
min, cooled and poured into ice. After filtration, the product
was dried and crystallized from toluene, affording 12 g (80%)
of 6: mp 255-257 °C; 1H NMR δ 2.55 (s, 3H), 7.4-8.1 (m, 6H,
arom).
the target. The most interesting derivatives presented
here (compounds 22d ,e) show a human aromatase
inhibitory activity superior to that of the known drug
fadrozole and are selective with respect to another
human P450 enzyme, namely P450 17. On the other
hand, based on the same xanthone nucleus, we obtained
compounds 22g-i that show fairly high potency as P450
17 inhibitors, the latter enzyme being a target for the
development of drugs for the treatment of prostatic
diseases.
Further experiments on the selectivity toward other
P450 enzymes and on the in vivo activity must be
performed to answer the question whether the xantho-
nes described in this paper are novel drug candidates
for the treatment of steroidal hormone-dependent dis-
eases.
1-Am in o-4-m eth yl-9H-9-xa n th en on e (7). Using the same
procedure described for compound 3, compound 7 (6.1 g, 90%)
was obtained from compound 6 (7.65 g, 0.03 mol): mp 146-
1
148 °C (ligroin); H NMR δ 2.35 (s, 3H), 6.4 (d, J ) 7.3 Hz,-
1H), 6.55 (broad, 2H, NH2), 7.2-8.25 (m, 5H, arom).
4-Meth yl-9-oxo-9H-1-xa n th en eca r bon itr ile (8). Using
the same procedure described for compound 4, compound 8
(1.2 g, 50%) was obtained from 7 (2.25 g, 0.01 mol): mp 220-
Exp er im en ta l Section
1
222 °C dec (toluene); H NMR δ 2.6 (s, 3H), 7.3-8.3 (m, 6H,
Ch em istr y. Gen er a l Meth od s. All melting points were
determined in open glass capillaries using a Bu¨chi apparatus
and are uncorrected. 1H NMR were recorded on a Varian
Gemini 300 spectrometer in CDCl3 solutions, with Me4Si as
the internal standard. Mass spectra were recorded on a VG
7070 E spectrometer. Elemental analyses were within 0.4%
of theoretical value unless otherwise indicated.
8-Meth yl-5-n itr o-4H-ben zop yr a n -4-on e (2). A solution of
8-methyl-4H-benzopyran-4-one (1; 3.2 g, 0.02 mol) in H2SO4
(500 mL) was treated with fuming HNO3 (1.26 g, 0.02 mol)
keeping the temperature between 0 and 5 °C. The reaction
mixture was stirred 1 h at room temperature, then poured into
ice. The separated solid was collected by filtration, and
compound 2 (2.8 g, 70%) was obtained: mp 172-174 °C
(toluene); 1H NMR δ 2.55 (s, 3H), 6.4 (d, J ) 6.1 Hz, 1H, H-2),
7.3 (d, J ) 6.3 Hz, 1H), 7.6 (d, J ) 6.3 Hz, 1H), 7.95 (d, J )
6.1 Hz, 1H, H-3).
5-Am in o-8-m eth yl-4H-ben zop yr a n -4-on e (3). A solution
of 2 (9 g, 0.043 mol) in THF was hydrogenated at room
temperature using Pd/CaCO3 as catalyst. After filtration, the
solvent was removed under reduced pressure and the residue
was crystallized from toluene to yield compound 3 (6.15 g,
80%): mp 141-143 °C; 1H NMR δ 2.3 (s, 3H), 6.15 (d, J )
4.64 Hz, 1H, H-2), 6.3 (broad, 2H, NH2), 6.35 (d, J ) 7.8 Hz,
1H), 7.15 (d, J ) 7.8 Hz, 1H), 7.7 (d, J ) 4.64 Hz, 1H, H-3).
8-Meth yl-5-cya n o-4H-ben zop yr a n -4-on e (4). To a solu-
tion of 3 (1.75 g, 0.01 mol) in H2SO4, a solution of NaNO2 (0.69
g, 0.01 mol) in water was added dropwise, keeping the
temperature below 5 °C. The reaction mixture was stirred 1 h
at room temperature, and then it was added to a solution of
1.5 g of CuCN and 2 g of NaCN in H2O at 50-60 °C. The
mixture was heated at 100 °C for 20 min, stirred overnight at
room temperature and filtered. The residue, on crystallizing
from toluene, gave 0.94 g (50%) of 4: mp 193-195 °C; 1H NMR
δ 2.55 (s, 3H), 6.45 (d, J ) 6.0 Hz,1H, H-2), 7.55-7.7 (m, 2H),
7.9 (d, J ) 6.0 Hz, 1H, H-3).
5-Br om o-8-m eth yl-4H-ben zop yr a n -4-on e (5). Compound
3 (1.75 g, 0.01 mol) was treated in HCl 1:1 with a solution of
NaNO2 (0.69 g, 0.01 mol) in water as previously described. The
solution obtained was added to a boiling solution of CuBr (1.7
g) in water. The mixture was stirred overnight at room
temperature and then extracted with CH2Cl2, washed with
water, dried (Na2SO4) and evaporated to dryness. The residue
was purified by flash chromatography (eluent: petroleum
ether/ethyl acetate 7/3) to give 1.2 g of 5 (50%): mp 85-87
°C; 1H NMR δ 2.45 (s, 3H), 6.35 (d, J ) 6.0 Hz, 1H, H-2), 7.25-
7.4 (m, 2H), 7.85 (d, J ) 6.0 Hz, 1H, H-3).
arom).
1-Br om o-4-m eth yl-9H-9-xa n th en on e (9). Using the same
procedure described for compound 5, compound 9 (1.73 g, 60%)
was obtained from 7 (2.25 g, 0.01 mol): mp 147-148 °C
1
(ligroin); H NMR δ 2.5 (s, 3H), 7.2-8.3 (m, 6H, arom).
1-Meth yl-4-n itr o-9H-9-xa n th en on e (11). A solution of
1-methyl-9H-9-xanthenone (10; 6.0 g, 0.028 mol) in H2SO4 (700
mL) was treated with fuming HNO3 (1.78 g, 0.028 mol),
keeping the temperature between 0 and 5 °C. The reaction
mixture was stirred 1 h at room temperature and poured into
ice. The separated solid was collected by filtration and purified
by flash chromatography (eluent: hexane/ethyl acetate, 9/1).
5.0 g (70%) of 11 was obtained: mp 199-200 °C (toluene); 1H
NMR δ 3.05 (s, 3H), 7.4-8.3 (m, 6H, arom).
4-Am in o-1-m eth yl-9H-9-xan th en on e (12). Using the same
procedure described for compound 4, compound 12 (3.0 g, 90%)
was obtained from 11 (3.8 g, 0.015 mol): mp 137-140 °C
(ligroin); 1H NMR δ 2.8 (s, 3H), 3.6 (broad, 2H, NH2), 7.2-
8.25 (m, 5H, arom).
1-Meth yl-9-oxo-9H-4-xa n th en eca r bon itr ile (13). Using
the same procedure described for compound 4, compound 13
(1.2 g, 50%) was obtained from 12 (2.25 g, 0.01 mol): mp 158-
160 °C (toluene); 1H NMR δ 3.1 (s, 3H), 7.3-8.3 (m, 6H, arom).
4-Br om o-1-m eth yl-9H-9-xan th en on e (14). Using the same
procedure described for compound 5, compound 14 (1.73 g,
60%) was obtained from 12 (2.25 g, 0.01 mol): mp 159-161
1
°C (ligroin); H NMR δ 3.0 (s, 3H), 7.2-8.2 (m, 6H, arom).
8-(Br om om eth yl)-5-n itr o-4H-ben zop yr a n -4-on e (15b).
A mixture of 8-methyl-5-nitro-4H-benzopyran-4-one (2; 2.8 g,
0.0137 mol), N-bromosuccinimide (2.44 g, 0.0137 mol) in the
presence of a catalytic amount of benzoyl peroxide in 100 mL
of CCl4 was refluxed 4 h and then hot filtered. The solvent
was evaporated to dryness and the residue crystallized from
ligroin, to give 2.7 g of the desired compound (70%): mp 151-
1
153 °C; H NMR δ 4.65 (s, 2H), 6.4 (d, J ) 4.8 Hz,1H, H-2),
7.35 (d, J ) 7.2 Hz, 1H), 7.8 (d, J ) 7.2 Hz, 1H), 7.95 (d, J )
4.8 Hz, 1H, H-3).
8-(Br om om eth yl)-5-cya n o-4H-ben zop yr a n -4-on e (15a ).
Using the previous procedure, compound 15b (0.92 g, 70%)
was obtained from 4 (0.94 g, 0.005 mol): mp 202-204 °C
(ligroin); 1H NMR δ 4.8 (s, 2H), 6.45 (d, J ) 7.6 Hz, 1H, H-2),
7.7-7.85 (m, 2H), 7.95 (d, J ) 7.6 Hz, 1H, H-3).
5-Br om o-8-(br om om eth yl)-4H-ben zop yr a n -4-on e (15c).
Using the previous procedure, compound 15c (1.1 g, 70%) was
obtained from 5 (1.2 g, 0.005 mol): mp 115-117 °C (ligroin);
1H NMR δ 4.65 (s, 2H), 6.35 (d, J ) 6.0 Hz,1H, H-2), 7.35 (d,
J ) 8.0 Hz, 1H), 7.6 (d, J ) 8.0 Hz, 1H), 7.85 (d, J ) 6.0 Hz,
1H, H-3).
4-Meth yl-1-n itr o-9H-9-xa n th en on e (6). A mixture of 15.3
g (0.1 mol) of 2-methyl-5-nitrophenol, 12.3 g (0.1 mol) of
o-chlorobenzoic acid, 20 g of K2CO3, 0.5 g of Cu and 0.5 g of
CuI in 300 mL of nitrobenzene was heated in an oil bath at
170-180 °C for 8 h. The solvent was steam distilled, and the
residue was filtered and acidified with HCl. The separated
solid was filtered, washed with water and suspended in a
4-(Br om om eth yl)-1-n itr o-9H-9-xa n th en on e (15e). Using
the previous procedure, compound 15e (1.7 g, 65%) was
obtained from 6 (2.0 g, 0.008 mol): mp 83-85 °C (toluene);
1H NMR δ 5.5 (s, 2H), 7.4-8.2 (m, 6H, arom).

New Class of Nonsteroidal Aromatase Inhibitors
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 5 679
4-(Br om om e t h yl)-9-oxo-9H -1-xa n t h e n e ca r b on it r ile
(15d ). Using the previous procedure, 0.96 g (60%) of oily
compound was obtained from 8 (1.2 g, 0.005 mol), and it was
used in the subsequent step without further purification.
1-Br om o-4-(br om om eth yl)-9H-9-xa n th en on e (15f). Us-
ing the previous procedure, 0.73 g (60%) of oily compound was
obtained from 9 (0.96 g, 0.003 mol), and it was used in the
subsequent step without further purification.
1-(Br om om eth yl)-4-n itr o-9H-9-xa n th en on e (15h ). Using
the previous procedure, 0.4 g (60%) of oily compound was
obtained from 11 (0.5 g, 0.002 mol), and it was used in the
subsequent step without further purification.
1-(Br om om e t h yl)-9-oxo-9H -4-xa n t h e n e ca r b on it r ile
(15g). Using the previous procedure, 0.28 g (60%) of oily
compound was obtained from 13 (0.47 g, 0.002 mol), and it
was used in the subsequent step without further purification.
4-Br om o-1-(br om om eth yl)-9H-9-xa n th en on e (15i). Us-
ing the previous procedure, 0.44 g (60%) of oily compound was
obtained from 14 (0.57 g, 0.002 mol), and it was used in the
subsequent step without further purification.
(relative abundance) 276 (M+, 10), 261 (80), 152 (66), 76 (100).
22n : 1H NMR δ 3.9 (s, 3H), 5.05 (s, 2H), 6.8-8.2 (m, 11H,
arom + C-H imidazole); MS m/z (relative abundance) 332 (M+,
62.7), 266 (17.5), 265 (100), 115 (69.9).
Gen er a l Meth od for P r ep a r a tion of 1,3,4-Tr ia zol-1-yl
Der iva tives 22k ,m ,o. A mixture of the selected bromomethyl
derivative (0.005 mol) and 1,3,4-triazole (0.015 mol) in 50 mL
of acetonitrile was refluxed for 7 h under nitrogen. The solvent
was removed under reduced pressure and the residue was
purified by flash chromatography (eluent: toluene/acetone 4/1).
1H NMR a n d MS for Com p ou n d s 22k ,m ,o (CDCl3). 2k :
1H NMR δ 5.5 (s, 2H), 6.35 (d, J ) 5.33 Hz, 1H, H-2), 7.3-
8.25 (m, 6H, arom + CH triazole + H-3); MS m/z (relative
abundance) 227 (M+, 100), 200 (48), 159 (47), 131 (44). 22m :
1H NMR δ 5.45 (s, 2H), 7.0-8.3 (m, 9H, arom + C-H triazole);
MS m/z (relative abundance) 277 (M+, 100), 250 (14.6), 209
(42), 181 (36). 22o: 1H NMR δ 3.9 (s, 3H), 5.25 (s, 2H), 6.8-
8.5 (m, 10H, arom + C-H triazole); MS m/z (relative abun-
dance) 333 (M+, 17.2), 278 (51.1), 264 (100), 115 (31.7).
Biologica l Meth od s. En zym e P r ep a r a tion s. The en-
zymes were prepared according to described methods: human
placental aromatase,34 human testicular 17R-hydroxylase as
described for the rat enzyme.19a
En zym e Assa ys. The enzyme assays were performed as
described: aromatase,34 17R-hydroxylase,27 kinetic study of
22e.26
Molecu la r Mod elin g. The construction of molecular mod-
els and the interpolation with the previously developed CoMFA
models were performed by means of the SYBYL software.35
Details about conformational analysis, structure optimization,
and CoMFA alignment are reported in ref 19b.
7-(Br om om eth yl)-4H-ben zop yr a n -4-on e (18). Using the
previous procedure and starting from 1.6 g (0.01 mol) of
7-methyl-4H-benzopyran-4-one (16), 1.43 g (60%) of the desired
1
compound was obtained: mp 89-93 °C (ligroin); H NMR δ
4.5 (s, 2H), 6.35 (d, J ) 5.2 Hz,1H, H-2), 7.3-7.5 (m, 2H), 7.8
(d, J ) 5.2 Hz, 1H, H-3), 8.2 (d, J ) 6.3 Hz, 1H).
3-(Br om om eth yl)-9H-9-xa n th en on e (19). Using the pre-
vious procedure and starting from 2.1 g (0.01 mol) of 3-methyl-
9H-9-xanthenone (17), 1.73 g (60%) of the desired compound
was obtained: mp 104-107 °C (ligroin); 1H NMR δ 5.5 (s, 2H),
6.85-8.35 (m, 7H, arom).
3-(Br om om eth yl)-7-m eth oxy-2-p h en yl-4H-ben zop yr a n -
4-on e (21). Using the previous procedure and starting from
2.6 g (0.01 mol) of 3-methyl-7-methoxy-2-phenyl-4H-benzopy-
ran-4-one (20), 1.89 g (55%) of the desired compound was
obtained: mp 115-117 °C (ligroin); 1H NMR δ 3.9 (s, 3H), 5.5
(s, 2H), 6.85-8.5 (m, 8H, arom).
Ack n ow led gm en t. The Saarbru¨cken group thanks
the Fonds der Chemischen Industrie for financial sup-
port.
Refer en ces
(1) Cuzick, J .; Wang, D. Y.; Bulbrook, R. D. The Prevention of Breast
Cancer. Lancet 1986, 8472, 83-86.
Gen er a l Meth od for P r ep a r a tion of Im id a zol-1-yl
Der iva tives 22a -j,l,n . A mixture of the selected bromomethyl
derivative (0.005 mol) and imidazole (0.015 mol) in 50 mL of
acetonitrile was refluxed for 7 h under nitrogen. The solvent
was removed under reduced pressure and the residue was
purified by flash chromatography (eluent: toluene/acetone 4/1).
1H NMR a n d MS for Com p ou n d s 22a -j,l,n (CDCl3).
22a : 1H NMR δ 5.45 (s, 2H), 6.5 (d, J ) 6.52 Hz, 1H, H-2),
6.95-7.75 (m, 5H, arom + CH imidazole), 7.9 (d, J ) 6.52 Hz,
1H, H-3); MS m/z (relative abundance) 251 (M+, 46.7), 184
(100), 158 (29.3), 32 (16.8). 22b: 1H NMR δ 5.45 (s, 2H), 6.45
(d, J ) 6.25 Hz, 1H, H-2), 6.95-7.65 (m, 5H, arom + CH
imidazole), 7.95 (d, J ) 6.25 Hz, 1H, H-3); MS m/z (relative
abundance) 271 (M+, 10.1), 174 (27.4), 68 (100), 41 (48.2).
22c: 1H NMR δ 5.35 (s, 2H), 6.4 (d, J ) 5.55 Hz, 1H, H-2),
6.95-7.65 (m, 5H, arom + CH imidazole), 7.8 (d, J ) 5.55 Hz,
1H, H-3); MS m/z (relative abundance) 305 (M+, 10.2), 260
(36.5), 193 (100), 167 (28.9). 22d : 1H NMR δ 5.4 (s, 2H), 7.0-
8.3 (m, 9H, arom + C-H imidazole); MS m/z (relative
abundance) 301 (M+, 9.2), 235 (100), 68 (97.9), 41 (61.5). 22e:
1H NMR δ 5.45 (s, 2H), 7.0-8.3 (m, 9H, arom + C-H
imidazole); MS m/z (relative abundance) 321 (M+, 44.8), 237
(40.9), 224 (100), 68 (92.9). 22f: 1H NMR δ 5.55 (s, 2H), 7.1-
8.3 (m, 9H, arom + C-H imidazole); MS m/z (relative
abundance) 354 (M+, 7.8), 276 (31.8), 243 (100), 209 (96.5).
22g: 1H NMR δ (DMSO-d6) 6.0 (s, 2H), 6.8-8.3 (m, 9H, arom
+ C-H imidazole); MS m/z (relative abundance) 301 (M+, 7.9),
145 (100), 67 (16.6), 40 (35.6). 22h : 1H NMR δ 6.1 (s, 2H),
7.0-8.3 (m, 9H, arom + C-H imidazole); MS m/z (relative
abundance) 321 (M+, 7.2), 68 (100), 44 (95.5), 41 (53). 22i: 1H
NMR δ 6.2 (s, 2H), 7.0-8.3 (m, 9H, arom + C-H imidazole);
MS m/z (relative abundance) 354 (M+, 6.2), 275 (35.6), 242
(100), 179 (28.9). 22j: 1H NMR δ 5.25 (s, 2H), 6.3 (d, J ) 4.89
Hz, 1H, H-2), 6.9-8.05 (m, 6H, arom + CH imidazole), 8.2 (d,
J ) 4.89 Hz, 1H, H-3); MS m/z (relative abundance) 226 (M+,
96), 159 (100), 131 (42.4), 51 (24.1). 22l: 1H NMR δ 5.35 (s,
2H), 6.9-8.3 (m, 10H, arom + C-H imidazole); MS m/z
(2) de Cupis, A.; Schettini, G.; Favoni, R. E. New vs Old Fashioned
Oestradiol Antagonists in Mammary Carcinoma: ‘In Vitro’ and
‘In Vivo’ Pharmacological Approaches. Pharmacol. Res. 1999, 39,
335-344.
(3) Brodie, A. M.; Schwarzel, W. C.; Shaikh, A. A.; Brodie, H. J .
The Effect of an Aromatase Inhibitor, 4-Hydroxy-4-androstene-
3,17-dione, on Estrogen-Dependent Processes in Reproduction
and Breast Cancer. Endocrinology 1977, 100, 1684-1695.
(4) Pasqualini, J . R.; Chetrite, G. S. Estrone Sulfatase versus
Estrone Sulfotransferase in Human Breast Cancer: Potential
Clinical Applications. J . Steroid Biochem. Mol. Biol. 1999, 69,
287-292.
(5) (a) Banting, L.; Nicholls, P. J .; Shaw, M. A.; Smith, H. J . Recent
Developments in Aromatase Inhibition as a Potential Treatment
for Oestrogen-Dependent Breast Cancer. Prog. Med. Chem. 1989,
26, 253-298. (b) Banting, L. Inhibition of Aromatase. Prog. Med.
Chem. 1996, 33, 147-184. (c) O′Reilly, J . M.; Brueggemeier, R.
W. Development of Steroidal and Nonsteroidal Inhibitors of
Aromatase for the Treatment of Hormone-Dependent Breast
Cancer. Curr. Med. Chem. 1996, 3, 11-22.
(6) Cheng, X.-M. To Market, To Market - 1995. Annu. Rep. Med.
Chem. 1996, 31, 337-355.
(7) Galatsis, P. To Market, To Market - 1996. Annu. Rep. Med.
Chem. 1997, 32, 305-326.
(8) Goss, P. E.; Winer, E. P.; Tannock, I. F.; Schwartz, L. H.
Randomized Phase III Trial Comparing the New Potent and
Selective Third-Generation Aromatase Inhibitor Vorozole with
Megestrol Acetate in Postmenopausal Advanced Breast Cancer
Patients. North American Vorozole Study Group. J . Clin. Oncol.
1999, 17, 52-63.
(9) Cole, P. A.; Robinson, C. H. Mechanism and Inhibition of
Cytochrome P-450 Aromatase. J . Med. Chem. 1990, 33, 2933-
2942.
(10) Furet, P.; Batzl, C.; Bhatnagar, A.; Francotte, E.; Rihs, G.; Lang,
M. Aromatase Inhibitors: Synthesis, Biological Activity, and
Binding Mode of Azole-Type Compounds. J . Med. Chem. 1993,
36, 1393-1400.
(11) (a) Recanatini, M. Comparative Molecular Field Analysis of Non-
Steroidal Aromatase Inhibitors Related to Fadrozole. J . Comput.-
Aided Mol. Des. 1996, 10, 74-82. (b) Recanatini, M.; Cavalli, A.
Comparative Molecular Field Analysis of Non-Steroidal Aro-
matase Inhibitors: An Extended Model for Two Different
Structural Classes. Bioorg. Med. Chem. 1998, 6, 377-388.

680 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 5
Recanatini et al.
(12) Kellis, J . T., J r.; Vickery, L. E. Inhibition of Human Estrogen
Synthetase (Aromatase) by Flavones. Science 1984, 225, 1032-
1034.
(13) Da Re, P.; Sagramora, L.; Mancini, V.; Valenti, P.; Cima, L.
Central Nervous System Stimulants of the Xanthone Group. J .
Med. Chem. 1970, 13, 527-531.
(14) Da Re, P.; Valenti, P.; Borraccini, A.; Primofiore, G. P. â-Adr-
energic Blocking Agents of the Chromone and Xanthone Groups.
J . Med. Chem. 1972, 15, 198-199.
(15) Gaion, R. M.; Valenti, P.; Montanari, P.; Da Re, P. Xanthone
Analogues of Clofibrate. Synthesis and Biological Evaluation as
Antagonists of Lipolysis in Vitro. Arzneim. Forsch./Drug Res.
1982, 32, 499-502.
(16) Valenti, P.; Chiarini, A.; Gasperi, F.; Budriesi, R. Xanthone 1,4-
Dihydropyridine Derivatives with a Potent Selective Bradycardic
Effect. Arzneim. Forsch./ Drug Res. 1990, 40, 122-125.
(17) (a) Valenti, P.; Recanatini, M.; Da Re, P.; Galimbeni, W.; Avanzi,
N.; Filippeschi, S. Xanthone Analogues of Geiparvarin. Arch.
Pharm. 1985, 318, 923-926. (b) Valenti, P.; Da Re, P.; Rampa,
A.; Montanari, P.; Carrara, M.; Cima, L. Benzo-γ-pyrone Ana-
logues of Geiparvarin: Synthesis and Biological Evaluation
Against B16 Melanoma Cells. Anti-Cancer Drug Des. 1993, 8,
349-360. (c) Valenti, P.; Fabbri, G.; Rampa, A.; Bisi, A.; Gobbi,
S.; Da Re, P.; Carrara, M.; Sgevano, A.; Cima, L. Synthesis and
Antitumor Activity of Some Analogues of Flavone Acetic Acid.
Anti-Cancer Drug Des. 1996, 11, 243-252.
5959-5967. (b) Kim, K. H.; Greco, G.; Novellino, E. A Critical
Review of Recent CoMFA Applications. In 3D QSAR in Drug
Design: Recent Advances. Kubinyi, H., Folkers, G., Martin, Y.
C., Eds.; KLUVER/ESCOM: The Netherlands, 1998; pp 257-
315.
(23) Lang, M.; Batzl, C.; Furet, P.; Bowman, R.; Ha¨usler, A.;
Bhatnagar, A. S. Structure-Activity Relationships and Binding
Model of Novel Aromatase Inhibitors. J . Steroid Biochem. Mol.
Biol. 1993, 44, 421-428.
(24) Thompson, E. A., J r.; Siiteri, P. K. Utilization of Oxygen and
Reduced Nicotinamide Adenine Dinucleotide Phosphate by
Human Placental Microsomes During Aromatization of Andros-
tenedione. J . Biol. Chem. 1974, 249, 5364-5372.
(25) Hartmann, R. W.; Bayer, H.; Gru¨n, G.; Sergejew, T.; Bartz, U.;
Mitrenga, M. Pyridyl-Substituted Tetrahydrocyclopropa[a]naph-
thalenes: Highly Active and Selective Inhibitors of P450 arom.
J . Med. Chem. 1995, 38, 2103-2111.
(26) Purba, H. S.; Bhatnagar, A. S. A Comparison of Methods
Measuring Aromatase Activity in Human Placenta and Rat
Ovary. J . Enzyme Inhib. 1990, 4, 169-178.
(27) Wachall, B. G.; Hector, M.; Zhuang, Y.; Hartmann, R. W.
Imidazole Substituted Biphenyls: A New Class of Highly Potent
and in vivo Active Inhibitors of P450 17 as Potential Therapeu-
tics for Treatment of Prostate Cancer. Bioorg. Med. Chem. 1999,
7, 1913-1924.
(18) Rampa, A.; Bisi, A.; Valenti, P.; Recanatini, M.; Cavalli, A.;
Andrisano, V.; Cavrini, V.; Fin, L.; Buriani, A.; Giusti, P.
Acetylcholinesterase Inhibitors: Synthesis and Structure-Activ-
ity Relationships of ω-[N-Methyl-N-(3-alkylcarbamoyloxyphe-
nyl)methyl]aminoalkoxyheteroaryl Derivatives. J . Med. Chem.
1998, 41, 3976-3986.
(19) (a) Sergejew, T.; Hartmann, R. W. Pyridyl Substituted Benzo-
cycloalkenes: New Inhibitors of 17R-Hydroxylase/C17,20-lyase
(P450 17R). J . Enzyme Inhib. 1994, 8, 113-122. (b) Wa¨chter, G.
A.; Hartmann, R. W.; Sergejew, T.; Gru¨n, G. L.; Ledergerber,
D. Tetrahydronaphthalenes: Influence of Heterocyclic Substit-
uents on Inhibition of Steroidogenic Enzymes P450 arom and
P450 17. J . Med. Chem. 1996, 39, 834-841. (c) Hartmann, R.
W.; Frotscher, M.; Ledergerber, D., Wa¨chter, G. A.; Gru¨n, G. L.;
Sergejew, T. Synthesis and Evaluation of Azole-Substituted
Tetrahydronaphthalenes as Inhibitors of P450 arom, P450 17,
and P450 TxA2. Arch. Pharm. 1996, 329, 251-261.
(28) Zhuang, Y.; Wachall, B. G.; Hartmann, R. W. Novel Imidazolyl
and Triazolyl Substituted Biphenyl Compounds: Synthesis and
Evaluation as Nonsteroidal Inhibitors of Human 17R-Hydroxy-
lase-C17,20-Lyase (P450 17). Bioorg. Med. Chem. 2000, in press.
(29) Dewar, M. J .; Zoebisch, E. G.; Healy, E. F.; Stewart, J . J . P.
AM1: a New General Purpose Quantum Mechanical Molecular
Model. J . Am. Chem. Soc. 1985, 107, 3902-3909.
(30) C-QSAR, version 1.87; BioByte: Claremont, CA, 1994.
(31) Kao, Y. C.; Zhou, C.; Sherman, M.; Laughton, C. A.; Chen, S.
Molecular Basis of the Inhibition of Aromatase (Estrogen
Synthetase) by Flavone and Isoflavone Phytoestrogens; a Site-
Directed Mutagenesis Study. Environ. Health Persp. 1998, 106,
85-92.
(32) Hansch, C.; Muir, R. M.; Fujita, T.; Maloney, P. P.; Geiger, F.;
Streich, M. The Correlation of Biological Activity of Plant Growth
Regulators and Chloromycetin Derivatives with Hammett Con-
stants and Partition Coefficients. J . Med. Chem. 1963, 6, 2817-
2824.
(33) Davis, A.; Teague, S. J . Hydrogen Bonding, Hydrophobic Inter-
actions, and Failure of the Rigid Receptor Hypothesis. Angew.
Chem., Int. Ed. Engl. 1999, 38, 736-749.
(20) Van Wauwe, J . P.; J anssen, P. A. Is There a Case for P-450
Inhibitors in Cancer Treatment? J . Med. Chem. 1989, 32, 2231-
2239. (b) Hartmann, R. W. Selective Inhibition of Steroidogenic
P450 Enzymes: Current Status and Future Perspectives. Eur.
J . Pharm. Sci. 1994, 2, 15-16.
(34) Hartmann, R. W.; Batzl, C. Aromatase Inhibitors. Synthesis and
Evaluation of Mammary Tumor Inhibiting Activity of 3-Alky-
lated 3-(4-Aminophenyl)piperidine-2,6-diones. J . Med. Chem.
1986, 29, 1363-1369.
(21) Hartmann, R. W.; Bayer, H.; Gru¨n, G. Aromatase Inhibitors.
Syntheses and Structure-Activity Studies of Novel Pyridyl-
Substituted Indanones, Indans, and Tetralins. J . Med. Chem.
1994, 37, 1275-1281.
(22) (a) Cramer, R. D., III; Patterson, D. E.; Bunce, J . D. Comparative
Molecular Field Analysis (CoMFA). 1. Effect of Shape on Binding
of Steroids to Carrier Proteins. J . Am. Chem. Soc. 1988, 110,
(35) SYBYL, version 6.5; Tripos Assoc.: St. Louis, MO, 1998.
J M000955S