Phenethyl pyridines with non-polar internal substitutents as selective ligands for estrogen receptor beta

Article abstract of DOI:10.1016/j.ejmech.2009.03.013

To create estrogen receptor beta (ERβ)-selective ligands with improved biological characteristics, we have extended our investigations of structurally simple bibenzyl-core ligands by preparing a series of compounds in which one phenol is replaced by a 3-h

Full text of DOI:10.1016/j.ejmech.2009.03.013

European Journal of Medicinal Chemistry 44 (2009) 3560–3570
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Phenethyl pyridines with non-polar internal substitutents
as selective ligands for estrogen receptor beta
Michael Waibel a, Karen J. Kieser a, Kathryn E. Carlson a, Fabio Stossi b,
,
Benita S. Katzenellenbogen b, John A. Katzenellenbogen a
*
a Department of Chemistry, University of Illinois, 600 South Mathews Avenue, Urbana, IL 61801, United States
b Department of Molecular and Integrative Physiology, University of Illinois, 600 South Matthews Avenue, Urbana, IL 61801, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
To create estrogen receptor beta (ERb)-selective ligands with improved biological characteristics, we
Received 25 September 2008
Received in revised form
6 March 2009
Accepted 12 March 2009
Available online 24 March 2009
have extended our investigations of structurally simple bibenzyl-core ligands by preparing a series of
compounds in which one phenol is replaced by a 3-hydroxypyridine ring. These phenethyl pyridines
were obtained by picoline anion alkylation, and compounds with different patterns of alkyl substitution
on the central two carbon units were prepared. Binding affinities for ER
a and ERb were determined, and
ligands with promising affinities and selectivities for ER were further tested for their gene transcrip-
b
tional activity. Several compounds had high affinity selectivity and good potency selectivity in tran-
scription assays. This study advances our understanding of compounds having ER-subtype selectivity and
will help to direct efforts in developing novel ER ligands.
Keywords:
Estrogen receptor
Estrogen receptor beta
Stilbestrol
Ó 2009 Elsevier Masson SAS. All rights reserved.
Hexestrol
Pyridine estrogen
Selective ligand
1. Introduction
used in menopausal hormone replacement therapy to maintain
bone mineral density [2] and for the prevention of breast cancer in
Estrogens, which act through the estrogen receptors (ERs),
regulate physiological processes not only in female reproductive
tissues [1], but also in the skeletal [2], cardiovascular [3] and central
nervous systems [4] of both males and females. While estrogen
pharmaceuticals can provide health benefits in some tissues, they
also promote cancer in other sites, such as the breast and the uterus
[5,6]. It has proved possible, nevertheless, to create selective
estrogen receptor modulators (SERMs), which are compounds that
activate the ER selectively only in those tissues where estrogen
action is desired, thus supporting estrogen health benefits, while
minimizing the risk of cancer [7]. Tamoxifen [8] and raloxifene [9]
are well known compounds that show such tissue-dependent
estrogen agonist or antagonist effects, and they are currently being
women at high risk [10]. The tissue selectivity of these SERMs is
thought to result from conformational characteristics of the ER-
ligand complexes that exploit differences in their interactions with
cell- and gene-specific coregulatory factors [7,11].
The ER system consists of two subtypes, ER
different patterns of tissue distribution and transcriptional regulation
[12–14], and ER has proven to have antiproliferative effects in breast
cancer cells when it is present together with ER [15,16]. Although the
amino acid sequence identity of ER and ER ligand binding domains
a and ERb, which have
b
a
a
b
is only 58%, the residues lining the ligand binding pockets are almost
identical. Of the ca. 25 residues lining the pocket, only two are
different (Met421 and Leu384 in ER
Met336 in ER ); also, the internal volume of the ligand pocket in ER
about 100 Å3 smaller than that of ER
[17]. This high sequence identity
a
are replaced by Ile373 and
b
b
is
a
in the binding pockets of the two ERs creates a challenge in designing
highly beta-selective ligands.
Abbreviations: DPN, diarylpropionitrile; ER, estrogen receptor; LDA, lithium
diisopropylamide; RBA, relative binding affinity; RTP, relative transcriptional
potency; SERM, selective estrogen receptor modulator; TBAF, tetrabutylammonium
fluoride; TBS, t-butyldimethylsilyl.
The phytoestrogen genistein 2 (Fig. 1) has a 20-fold binding
selectivity relative to estradiol 1 [12], and we have described
a number of ER
selective diarylpropionitrile 3 (DPN) [22]. Others have reported
ER -selective benzothiazoles [23], benzimidazoles [24] and
b-selective ligands [18–22], for example, the 70-fold
* Corresponding author. Tel.: þ1 217 333 6310; fax: þ1 217 333 7325.
b
E-mail address: jkatzene@uiuc.edu (J.A. Katzenellenbogen).
0223-5234/$ – see front matter Ó 2009 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2009.03.013

M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
3561
2. Chemistry
OH
OH
OH
O
O
2.1. Synthesis of a-substituted phenethyl pyridines
Scheme 1 shows the syntheses of phenethyl pyridines bearing
different substituents on the linker carbon next to the pyridine ring.
The reaction sequences started with the protection of 5-hydroxy-2-
methyl pyridine 11, as previously described [32]. The bulky t-
butyldimethylsilyl (TBS)-group was chosen to prevent lithiation of
the rather acidic C-6 position of the pyridine ring in the subsequent
treatments with LDA. The methyl group of pyridine 12 was
smoothly lithiated with LDA at low temperatures, and the anion
was reacted with 4-methoxybenzyl chloride [20], yielding phe-
nethyl pyridine 13. Lithiation at the same position was repeated in
the next step, followed by reaction with alkyl halides, producing
compounds 14–16. We attempted to improve the low yield of
compound 14 by employing different alkylating agents: Use of
methyl triflate gave only a very polar compound, possibly the N-
methylated product; by contrast, use of methyl bromide or methyl
iodide gave the desired compound in low yield, but a large amount
of starting material could be recovered. Yields of 14 could be
improved simply by repeating the reaction on recovered starting
material.
Deprotection of compounds 13–16 was first attempted in two
steps, by reacting them first with tetrabutylammonium fluoride
(TBAF) to remove the TBS-group, followed by treatment with BBr3
in CH2Cl2 to cleave the phenol ether. The mono deprotected
compounds resulting from the first reaction, however, were highly
polar and insoluble in CH2Cl2. They could be dissolved in DMSO or
acetic acid, but no deprotection took place when they were refluxed
with NaCN in DMSO for several days. On the other hand, refluxing
with hydriodic acid in acetic [29] gave good yields of the target
compounds 17–20 in short reaction times. In addition, we found
that the TBS ether function of compounds 13–16 was also cleaved
under these conditions. Thus, compounds 17–20 could be obtained
from molecules 13–16 in one single step.
HO
HO
1, Estradiol (E2)
2, Genistein
OH
OH
N
O
CN
HO
HO
3, Diarylpropionitrile (DPN)
4
F
Fig. 1. Important ER ligands.
molecules having other types of heterocyclic cores [25–27], in
particular, the benzoxazole 4, with a 200-fold binding preference
for ERb [25].
We recently observed that bibenzyl-core compounds having
small, non-polar internal substituents showed unexpectedly high
selectivity for ERb in terms of binding and transcriptional potency
[28]. The most selective ligand of this series was the gem-dimethy-
lated compound 5 (Fig. 2), but monoalkylated analogs that are
homologs of isobutestrol 7 [20,29], such as ligand 6, also showed
good ER
polar internal substituents were ER
because most ER -selective non-steroidal ligands described previ-
b
selectivity. This finding, that simple bisphenols with non-
b
selective, was surprising,
b
ously have distinctly more polar interior substituents [20–27,30], as is
illustrated by the ligands 2–4 in Fig. 1. We, therefore, thought that
replacing one of the phenols of the bibenzyl core by a 3-hydrox-
ypyridine ring, which would reintroduce an interior polar element
into these compounds, might lead to a further increase in their ER
b
We also attempted to introduce two alkyl chains at the carbon
next to the pyridine ring by alkylating twice. However, when
compound 14 was treated with LDA, the deep purple color indic-
ative of the lithiated pyridine, which was observed in all of the
alkylations of compounds 12 and 13, did not appear, and upon
addition of the alkyl halide, no reaction took place. Employing the
LIKOR and LIDAKOR superbases [33], which are reported to remove
benzylic rather than aromatic protons, gave the same result. Our
assumption that the pyridine methylene position in compound 14
is too sterically hindered to be deprotonated by these bases is
supported by the fact that literature examples of such anion
formations are only reported with less hindered systems [33–35].
Our attempts to use a stronger base, n-BuLi, gave a mixture of
decomposition compounds, probably due to lithiation of the
aromatic rings or addition to the pyridine ring.
selectivity. In addition, we recently reported related pyridine
compounds 8 and 9, analogs of hexestrol 10 [20], with both linker
carbons bearing one ethyl chain. These compounds showed some-
what promising ERb selectivity, but they had quite low affinity for
both ERs, presumably because they were missing the second
hydroxyl function, which is considered important for high affinity ER
binding [31]. Therefore, we focused on making the pyridinol analogs
having two hydroxyl groups.
OH
OH
R
HO
CH3
H3C
HO
2.2. Synthesis of b-substituted phenethyl pyridines
6 R=Me
5
7 R=Et, isobutestrol
Scheme 2 illustrates the syntheses of phenethyl pyridines with
either one alkyl chain at the benzylic position or one substituent on
both linker carbons. In the first step, benzylic alcohols 21 and 22
were chlorinated using SOCl2 to produce compounds 23 and 24
[36]. These reactions proceeded rapidly and in high yield at ꢀ78 ꢁC
to give the benzylic chlorides. They were then reacted with lithi-
ated 12, as described above in the preparation of 13, to give phe-
nethyl pyridines 25 and 26. As before, deprotection using hydriodic
acid gave the target compounds 27 and 28. Further lithiation and
ethylation of compounds 25 and 26 yielded the mixtures of dia-
stereomers 29 and 30. The target compounds 31a, 31b, 32a and 32b
H3C
N
H3C
H
OH
H
H
H
HO
CH3
CH3
HO
8, R*,R*-isomer
9,R*,S*-isomer
10, Hexestrol
Fig. 2. ER ligands having either a bibenzyl- or a phenethyl pyridyl core.

3562
M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
CH3
R
R
CH3
a
a
OH
Cl
N
N
HO
TBSO
MeO
MeO
12 (92 %)
OCH3
11
21 R = Me
22 R = Et
23 R = Me (75 %)
24 R = Et (83 %)
b
OCH3
CH3
b
N
TBSO
N
13 (61 %)
TBSO
N
R
TBSO
OCH3
R
25 R = Me (44 %)
26 R = Et (39 %)
12
c
N
14 R = Me (16 %)
15 R = Et (53 %)
16 R = Pr (60 %)
TBSO
OH
c
N
R
27 R = Me (71 %)
28 R = Et (57 %)
HO
OH
R
d
OCH3
H3C
N
HO
d (from 25,26)
17 R = H (from 13) (64 %)
18 R = Me (84 %)
19 R = Et (71 %)
20 R = Pr (75 %)
N
R
29 R = Me (31 % cis, 12 % trans)
30 R = Et (12 % cis, 12 % trans)
TBSO
Scheme 1. (a) Imidazole, TBS–Cl, DMF; (b) LDA, 4-(CH3O)-PhCH2Cl, THF; (c) LDA, RX
(14: X ¼ I; 15, 16: X ¼ Br), THF; (d) HI, HOAc.
H3C
OH
c
could be obtained very efficiently, since it was possible to separate
the diastereomers by chromatography after the final deprotection
of 29 and 30.
N
R
HO
31a R = Me,cis (66 %)
31b R = Me, trans (62 %)
32a R = Et, cis (72 %)
32b R = Et, trans (75 %)
3. Pharmacology
Scheme 2. (a) SOCl2, CH2Cl2; (b) LDA, 23 (for 25)/24 (for 26), THF; (c) HI, HOAc; (d)
LDA, EtBr, THF.
3.1. Estrogen receptor binding assays
The phenethyl pyridines were tested in competitive radiometric
binding assays, using purified full-length human ERa and ERb, to
determine their binding affinities for the two ERs [37,38]. The
binding affinities are expressed relative to that of estradiol, which is
set to 100%, to give relative binding affinity (RBA) values. Table 1
shows the RBA values of the synthesized pyridines and additionally
the bibenzyls 6 and 7, as well as hexestrol 10 and phenethyl pyri-
dines 8 and 9. C Log P values are also listed in Table 1.
The most ERb-selective phenethyl pyridines were among the
ones bearing one alkyl chain on the linker carbon next to the
pyridine ring, namely compounds 18 (54-fold) and 19 (80-fold).
from the more general finding that increased substitution results in
higher affinities but lower selectivity, a behavior illustrated by
isobutestrol 7 compared to hexestrol 10 [20], as well as by their
stilbene analogs [28]. Placing the alkyl substituent on the benzylic
carbon also resulted in compounds having quite good ERb selec-
tivities, but these were lower than those of the first compound
series (27 vs. 18, 28 vs. 19). Curiously, at the benzylic position,
methyl substitution gave better affinity and selectivity than ethyl
substitution (18, 19 vs. 27, 28).
Both 18 and 19 were more ER
bibenzyl analogs, compound 6 (29-fold) and isobutestrol 7 (18-
fold) [28]. This result confirmed our expectations and is consistent
with the general finding that ERb selectivity can be achieved with
ligands, such as the phenethyl pyridines, that are more polar on the
interior than the bibenzyls. However, although the phenethyl
pyridines were more ERb selective in their binding, their overall
affinity was almost 100-fold less than that of the corresponding
bibenzyls.
The general trend in this first compound series is that both
affinity and selectivity for ERb increase with the chain length of the
substituents, reaching a maximum at two carbon atoms (17 vs. 18
vs. 19); further extension of the chain lowered both affinity and
selectivity (20). Interestingly, the trend shown by 17–19 differs
b
selective than their corresponding
The fact that compound 19 showed such good selectivity,
together with the observed trend that higher substitution resulted
in both increased affinity and selectivity, led us to examine
whether the affinity and selectivity of 19 might be improved by
introducing a second alkyl chain, on the benzylic position. The anti
(R*,S*) isomers of both doubly substituted compounds (31b, 32b)
showed markedly increased affinity as desired, but selectivity was
reduced in comparison to their monosubstituted analogs (31b vs.
19, 27; 32b vs. 19, 28). On the other hand, the syn (R*,R*) isomers
(31a, 32a) had much lower affinity than the corresponding anti-
isomers (31a vs. 31b, 32a vs. 32b). This trend is also known for the
meso (R*,S*) and dl (R*,R*) diastereomers of hexestrol [29], as well
as for the monohydroxy reference compounds 8 and 9, studied
earlier [20].

M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
3563
Table 1
ER and ERb relative binding affinity (RBA), relative transcriptional potency (RTP), transcriptional activity (EC50 values), and affinity and potency selectivities for phenethyl
a
pyridines and related ligands.
Ligand
Ligand binding
Transcription potency
RBAa (%)
ER
b
/
a
C Log Pc
RTPd (%)
ER
b
/
a
EC50 (nM)
b/a
potency
ratiob
e
affinity
ratiob
potency
ratiob
ER
a
b
ER
a
b
ER
a
ERb
OH
<0.004
0.025 ꢂ 0.006
0.27 ꢂ 0.07
0.72 ꢂ 0.05
0.101 ꢂ 0.03
0.17 ꢂ 0.02
0.15 ꢂ 0.02
0.119 ꢂ 0.002
5.3 ꢂ 1.3
6
54
80
5
2.71
NDf
0.03
0.15
ND
ND
0.83
2.94
ND
0.29
ND
ND
45
ND
NTf
330
65
NT
60
17
NT
5.5
3.8
NT
2.6
NT
NT
0.1
0.6
N
HO
HO
HO
HO
HO
HO
17
OH
OH
OH
OH
OH
CH3
<0.005
3.11
3.64
4.17
3.11
3.64
4.04
4.04
4.57
28
N
H3C
N
18
19
20
0.009 ꢂ 0.003
20
H3C
0.021
ND
NT
NT
170
NT
NT
1.1
26
N
N
N
<0.006
28
0.023
13
440
CH3
27
28
0.018 ꢂ 0.001
0.016 ꢂ 0.004
0.7 ꢂ 0.2
8
ND
ND
ND
0.5
3.0
NT
CH3
H3C
N
OH
OH
OH
7
ND
NT
CH3
31a
HO
HO
HO
H3C
N
7
91
0.11
CH3
31b
32a
H3C
N
0.035 ꢂ 0.006
0.67 ꢂ 0.14
19
0.63
1.9
16
CH3
(continued on next page)

3564
M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
Table 1 (continued )
Ligand
Ligand binding
Transcription potency
RBAa (%)
ER
b
/
a
C Log Pc
RTPd (%)
ER
b
/
a
EC50 (nM)
b/a
potency
ratiob
e
affinity
ratiob
potency
ratiob
ER
a
b
ER
91
a
b
ER
a
ERb
H3C
OH
3.3 ꢂ 0.7
7.5 ꢂ 1.6
2
4.57
82
42
0.9
0.11
0.61
0.2
N
HO
CH3
32b
OH
OH
OH
CH3
0.71 ꢂ 0.08
2.2 ꢂ 0.4
277 ꢂ 57
20.5 ꢂ 5.0
39.5 ꢂ 8.4
697 ꢂ 194
29
18
3.65
4.18
5.11
1.5
ND
ND
27
6.5
NT
NT
1.2
NT
NT
5
HO
6
H3C
ND
ND
ND
ND
NT
NT
HO
Isobutestrol, 7
H3C
2.5
HO
CH3
Hexestrol, 10
H3C
0.010
0.081
0.023
0.684
2
8
5.23
5.23
ND
ND
ND
ND
ND
ND
NT
NT
NT
NT
NT
NT
N
HO
CH3
CH3
8g
H3C
N
HO
9g
a
[estradiol]
[compound]
Relative binding affinity (RBA) values are determined by competitive radiometric binding assays and are expressed as EC
/EC
ꢃ 100 (RBA, estradiol ¼ 100).
50
50
In these assays, the Kd for estradiol is 0.2 nM for ER
a
and 0.5 nM for ERb. For details, see Experimental section.
b
For each set of affinity or potency values, the
b/a
ratio is calculated such that the ratio is >1 for compounds having higher affinity or potency on ER
b
than on ER
a
.
c
C Log P values are calculated within ChemDraw Ultra 7.0.3.
Relative transcriptional potency (RTP) values are expressed as EC
d
(estradiol)
(ligand)
/EC
ꢃ 100 (RTP, estradiol ¼ 100). In these assays, the EC50 for estradiol is 0.1 nM for ER
a
and
50
50
0.5 nM for ERb.
e
Transcriptional activity was measured using a cotransfection assay in human endometrial cancer (HEC-1) cells (see Experimental section and Fig. 3). Transcriptional
potency ¼ EC50
.
f
ND ¼ not determined; NT ¼ not tested.
g
Data are from De Angelis et al. [20].
The lower binding affinity of the monohydroxy phenethyl pyri-
Also of note is the modest affinity of all pyridine derivatives
compared to the binding of their bibenzyl analogs. Because the
pyridine derivatives are more polar than the bibenzyls (compare
estimated C Log P values, Table 1: 18, 27 vs. 6; 19, 28 vs. 7; 32a, 32b
vs. 10), it seems reasonable that the pyridines would experience
a greater desolvation energy penalty upon moving from water to
the ligand binding pocket of the ERs, thus disfavoring the binding
process energetically. This desolvation energy penalty experienced
by polar ER ligands is evident in other ligand series: The moderately
dines 8 and 9 (prepared previously [20]) compared to the corre-
sponding dihydroxy analogs 32a and 32b is notable. It also suggests
that these two ligand sets might be binding to ER in opposite
directions: The pyridinol ring of 8 and 9 most probably mimics the A-
ring of estradiol, and binding is rather poor because the more polar
ring has a less favorable interaction with Phe356 which is near the
ligand A-ring in ERb; with the isomers 32, on the other hand,
a phenol could be acting as an A-ring mimic, with consequently
higher affinity due to a more favorable interaction of this less polar
polar alkyl-triaryl-pyrazoles bind much better to ER
a than their
ring with Phe356 in ER
estrogens, a second hydroxyl group seems important for high affinity
binding [31].
b
. Interestingly, with most non-steroidal
isostructural but more polar alkyl-triaryl-imidazoles [39], but less
well than their isostructural but less polar furans [40]. Similarly, the
pyridine derivatives of deoxyhexestrol bind better than their more

M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
3565
100
50
ERβ
18
19
ERβ
ERα
ERα
0
100
31b
27
ERβ
50
ERα
ERβ
ERα
0
100
32b
32a
ERα
ERβ
50
ERα
ERβ
0
-12 -11 -10
-9
-8
-7
-6
100
E2
ERβ
50
0
ERα
-12 -11 -10
-9
-8
-7 -6
Concentration (Log M)
Fig. 3. Transcriptional efficacy of ligands 18, 19, 27, 31b, 32a and 32b on ER
include as well the values corresponding to a ligand concentration of 0 M (V ¼ negative control). Human endometrial cancer (HEC-1) cells were transfected with expression vectors
for ER and ER and the estrogen-responsive gene 2ERE-pS2-Luc, and were incubated with the indicated concentrations of ligand for 24 h. Luciferase activity and -galactosidase
a and ERb at 0.1, 10, and 1000 nM, with estradiol (E2) for comparison (lowest left panel). The diagrams
a
b
b
activity were assayed as described [41]. Estradiol activity at 10 nM is set at 100%.
polar pyrimidine analogs, consistent with the higher C Log P values
of the former system [20].
[41]. The dose–response curves for these compounds are given in
Fig. 3, and the EC50 values are listed in Table 1. This table also
includes, for comparison, the EC50 values for other related
compounds determined in a recent study [28].
To facilitate comparisons of the ER subtype transcriptional potencies
of our compounds with their ER subtype binding affinities, the EC50
values from the transcription assays were converted to relative tran-
scriptional potency (RTP) values (See Table 1, footnote d). The RTP
values provide a measure of transcriptional potency relative to that of
estradiol and thus can be better compared to their binding affinities,
which are also measured relative to estradiol by the competitive
radiometric binding assays. Estradiol has a 2.5-fold preference in favor
While it is notable that both anti-isomers, 31b and 32b, are less
polar (based on their chromatographic behavior) than the corre-
sponding syn-isomers, 31a and 32a, which might account for their
higher affinity, the markedly higher ER binding affinity of anti-
bibenzyl diols is thought to arise from their preference to adopt
a biaryl anti conformation, preferred by the ER ligand binding
pocket, rather than a syn conformation [29].
3.2. Activity on gene transcription
The transcriptional activities of six selected compounds (18, 19,
of ER
5-fold preference in terms of transcriptional potency (EC50
[ER ] ¼ 0.5 nM).
] ¼ 0.1 nM vs. [ER
As shown in Fig. 3 and listed in Table 1, the three compounds
having the highest ER relative binding selectivity (18,19, and 27) also
showed good ER selectivity in terms of transcriptional potency. In
a
in terms of binding (Kd [ER
a
] ¼ 0.2 nM vs. [ER
b
] ¼ 0.5 nM) and
27, 31b, 32a, 32b) that showed good ER
b
binding selectivity were
a
tested for their agonistic character as regulators of gene tran-
scription, together with estradiol. They were assayed in human
endometrial cancer (HEC-1) cells, transfected with expression
a
b
b
plasmids for ER
a
or ER
b
and an estrogen-responsive reporter gene
b

3566
M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
fact, the
b
/
a
ratio of their RBA values and their RTP values is quite
were recorded on Varian UNITY 400 or 500 MHz spectrometers.
Chemical shifts are reported in ppm and referenced from solvent
references. NMR coupling constants are reported in hertz. Mass
spectra were obtained on a Micromass 70-VSE spectrometer (EI,
70 eV; CI, methane). Melting point (mp) determinations were
carried out on Thomas Hoover Unimelt capillary apparatus and are
uncorrected. All target compounds shown in Scheme 1 could be re-
crystallized, while all compounds in Scheme 2 could not. They were
obtained as oils that spontaneously foamed upon drying to form
amorphous solids, which might explain their less distinct melting
points. Elemental analyses were performed by the Microanalysis
Service Laboratory of the University of Illinois. Analyses indicated
by the symbols of the elements or functions were within ꢂ0.4% of
the theoretical values.
comparable. The other three compounds (31b, 32a, and 32b), which
have generally lower ER relative binding selectivity, are also less ER
selective in terms of their RTP values. Compound 31b, in fact, has an
ER -potency preference that is greater than that of estradiol; 32b is
similar to that of estradiol, but 32a is more ER selective than estradiol.
b
b
a
b
Irrespective of their ER-subtype selectivities, the higher absolute
affinities of the anti disubstituted compounds (31b and 32b)
compared to the others (18, 19, 27, 31a, and 32a) are reflected in their
overall much higher transcriptional potencies. Compounds 31b and
32b, in particular, were found to be extremely potent, thus being the
most potent ligands of the whole series, although they are not at all
ERb selective.
4. Conclusion
This study extends our investigation of simple, non-steroidal ERb-
5.2. Chemical synthesis
selective bibenzyl diol ligands having non-polar internal substituents
to their corresponding phenethyl pyridinols. The nitrogen-for-carbon
substitution in one ring, which transforms a phenol into a pyridinol
and makes the interior of the ligand more polar, gives ligands that
conform more closely to the ‘‘narrow but somewhat polar interior’’
5.2.1. 5-(tert-Butyl-dimethyl-silanyloxy)-2-[2-(4-methoxy-phenyl)-
ethyl]-pyridine (13)
Lithium diisopropylamide (LDA) (10.5 mL of a 2 M solution in
heptane/THF/ethylbenzene) was added to 3.5 mL THF at ꢀ78 ꢁC,
and then compound 12 (4.00 g, 17.94 mmol, dissolved in 3.5 mL
THF) was added dropwise over a time period of 15 min. After
45 min, 1-chloromethyl-4-methoxy-benzene (2.91 mL, 21.5 mmol,
dissolved in 3.5 mL THF) was added in the same manner as 12 and
the reaction was stirred for 1 h. The mixture was allowed to go to rt
during a 1.5 h time period and stirred for 1 h at rt. The reaction was
quenched by adding NH4Cl (sat. aqueous solution) and extracted
three times with EtOAc. The organic extracts were dried over
Na2SO4 and the solvent was removed under vacuum. After purifi-
cation by flash chromatography (10% EtOAc/hexanes), 13 was
obtained as a colorless oil (3.73 g, 61%). 1H NMR (500 MHz, CDCl3)
that characterizes the pharmacophore of most ER
reported to date. Some of the monosubstituted phenethyl pyridinols
did exhibit high ER -selective relative binding affinity and transcrip-
b-selective ligands
b
tional potency that in some cases was greater than that of the corre-
sponding bibenzyl diols, but their overall binding affinities were much
lower. This reduction in affinity probably comes from the greater
desolvation penalty experienced by the more polar pyridine
compounds, which tracks with their lower lipophilic character as
measured by their lower C Log P values. In reporter gene transcription
assays, the ERb-selective transcriptional potencies of the phenethyl
pyridinols mirrored quite closely their ERb-selective binding affinities.
The vicinally disubstituted phenethyl pyridinols, like the disubstituted
bibenzyl diols, showed enhanced ER binding affinity and transcrip-
d
8.16 (d, J ¼ 2.8, 1H), 7.08 (d, J ¼ 8.7, 2H), 7.03 (dd, J ¼ 8.4, 2.8, 1H),
6.92 (d, J ¼ 8.4, 1H), 6.80 (d, J ¼ 8.7, 2H), 3.78 (s, 3H), 3.01–2.92 (m,
4H), 0.98 (s, 9H), 0.20 (s, 6H); 13C NMR (500 MHz, CDCl3)
157.77,
tional potencies, but reduced ERb selectivities.
d
Unlike other ER -selective ligands thus far described, polar
b
153.91, 150.21, 141.55, 133.78, 129.38, 127.28, 123.17, 113.68, 55.19,
39.55, 35.29, 25.58, 18.17, ꢀ4.51; MS (EI) m/z 343 (Mþ, 23).
elements in the interior region of these structurally simple bibenzyl
diol and phenethyl pyridinol ligands are not required for their ER-
subtype selectivity; in fact, while introducing the polar nitrogen into
5.2.2. General alkylation procedure
a benzene to make a pyridine maintained ER
improve it, and it uniformly reduced binding affinity. Thus, polar
interior elements are not a prerequisite for non-steroidal ER -selec-
tive ligands. Our findings provide further understanding of the design
elements that should be considered in the continuing efforts to create
b selectivity, it did not
LDA was added to THF at ꢀ78 ꢁC, and then the starting material 13
(1 eq, dissolved in THF) was added dropwise over a time period of
15 min. After 45 min, the alkyl halide (dissolved in THF) was added, and
after further 45 min pure alkyl halide was added, both times the
addition was done over a 15 min time period. The mixture was allowed
to go to rt during a 1.5 h time period and stirred for 45 min at rt. The
reaction was quenched by adding NH4Cl (sat. aqueous solution) and
extracted three times with EtOAc. The organic extracts were dried over
Na2SO4 and the solvent was removed under vacuum, and the crude
product was purified by flash chromatography.
b
ERb-selective ligands with improved biological characteristics.
5. Experimental section
5.1. Materials and methods
Reagents and solvents were purchased from Aldrich and Fisher
Scientific. THF and dichloromethane were dried using a solvent-
dispensing system (SDS) (neutral alumina columns) built by J.C.
Meyer on the basis of a design developed by Pangborn et al. [42].
Glassware was oven- or flame-dried, assembled while hot, and
cooled under nitrogen atmosphere. All reactions were performed in
anhydrous solvents and under nitrogen atmosphere, unless stated
otherwise. Reaction progress was monitored by thin-layer chro-
matography (TLC) using 0.25 mm Merck silica gel 60 glass plates
containing F254 UV-Indicator. The plates were visualized by either
UV light (254 nm), or dipping in a solution of potassium perman-
ganate followed by heating. Column chromatography was per-
5.2.3. 5-(tert-Butyl-dimethyl-silanyloxy)-2-[2-(4-methoxy-
phenyl)-1-methyl-ethyl]-pyridine (14)
According to the general alkylation procedure, lithiation of 13
was performed by adding 13 (890 mg, 2.6 mmol, in 0.5 mL THF) to
a solution of LDA (1.5 mL of a 2 M solution in heptane/THF/ethyl-
benzene) in 1 mL THF. Adding methyl iodide (1st addition: 487
mL,
7.8 mmol, in 0.5 mL THF, 2nd addition: 487 L pure) to a stirred
m
solution of lithiated 13 gave 14 as a colorless oil (149 mg, 16%) after
reaction work up and purification (10% EtOAc/hexanes). 1H NMR
(500 MHz, CDCl3)
d
8.18 (d, J ¼ 2.9, 1H), 7.00 (dd, J ¼ 8.4, 2.9, 1H),
6.95 (d, J ¼ 8.7, 2H), 6.86 (d, J ¼ 8.4, 1H), 6.74 (d, J ¼ 8.7, 2H), 3.75 (s,
3H), 3.09–3.03 (m, 1H), 2.99–2.95 (m, 1H), 2.77–2.73 (m, 1H), 1.25
(d, J ¼ 6.9, 3H), 0.98 (s, 9H), 0.20 (s, 6H); 13C NMR (500 MHz, CDCl3)
formed using Woelm 32–63 m
m silica gel packing. 1H and 13C NMR

M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
3567
d
158.09, 157.68, 150.16, 141.33, 132.81, 130.01, 127.21, 122.16, 113.42,
hydriodic acid (57 wt.% in water) in 3.8 mL acetic acid and refluxed
for 2.5 h. After reaction work up, purification by flash chromatog-
raphy (pure EtOAc) and re-crystallization from EtOAc/acetone, 18
was obtained as a white solid (137 mg, 84%, mp 219 ꢁC). 1H NMR
(500 MHz, acetone-d6) 8.14 (d, J ¼ 2.9, 1H), 7.05 (dd, J ¼ 8.4, 2.9, 1H),
6.93 (d, J ¼ 8.4, 1H), 6.89 (d, J ¼ 8.6, 2H), 6.66 (d, J ¼ 8.6, 2H), 3.06–
2.99 (m, 1H), 2.94–2.90 (m, 1H), 2.69–2.65 (m, 1H), 1.17 (d, J ¼ 6.9,
55.12, 43.05, 42.68, 25.58, 19.93, 18.17, ꢀ4.50; MS (CI) m/z 358
(Mþ þ 1, 72).
5.2.4. 5-(tert-Butyl-dimethyl-silanyloxy)-2-[1-(4-methoxy-
benzyl)-propyl]-pyridine (15)
According to the general alkylation procedure, lithiation of
890 mg of 13 was performed as described in the case of compound
3H); 13C NMR (500 MHz, acetone-d6)
d 157.33, 156.27, 152.55,
14. Adding ethyl bromide (1st addition: 233
THF, 2nd addition: 233 L pure) to a stirred solution of lithiated 13
gave 15 as a colorless oil (507 mg, 53%) after reaction work up and
purification (10% EtOAc/hexanes). 1H NMR (500 MHz, CDCl3)
8.20
m
L, 3.1 mmol, in 0.5 mL
137.86, 132.63, 130.77, 123.14, 122.95, 115.63, 43.76, 43.25, 20.65;
MS (CI) m/z 230 (Mþ þ 1, 100). HRMS (CI) calcd for C14H16NO2:
230.1181, found 230.1181. Anal. C14H15NO2$0.1H2O: Calcd. C, 72.77;
H, 6.63; N, 6.06. Found C, 72.73; H, 6.50; N, 6.12.
m
d
(d, J ¼ 2.8, 1H), 6.97 (dd, J ¼ 8.5, 2.8, 1H), 6.90 (d, J ¼ 8.8, 2H), 6.77 (d,
J ¼ 8.5, 1H), 6.71 (d, J ¼ 8.8, 2H), 3.74 (s, 3H), 2.95–2.76 (m, 3H),
1.78–1.66 (m, 2H), 0.98 (s, 9H), 0.76 (t, J ¼ 7.4, 3H), 0.20 (s, 6H); 13C
5.2.9. 6-[1-(4-Hydroxy-benzyl)-propyl]-pyridin-3-ol (19)
According to the general deprotection procedure, compound 15
(360 mg, 0.97 mmol) in 5.5 mL acetic acid was added to 1.4 mL
hydriodic acid (57 wt.% in water) in 5.5 mL acetic acid and refluxed for
1.5 h. After reaction work up, purification by flash chromatography
(75% EtOAc/hexanes) and re-crystallization from EtOAc/hexanes, 19
was obtained as a white solid (167 mg, 71%, mp 196 ꢁC). 1H NMR
(500 MHz, acetone-d6) 8.16 (d, J ¼ 2.9, 1H), 7.03 (dd, J ¼ 8.4, 2.9, 1H),
6.86–6.82 (m, 3H), 6.63 (d, J ¼ 8.6, 2H), 2.90–2.84 (m, 1H), 2.81–2.75
(m, 2H), 1.75–1.60 (m, 2H), 0.70 (t, J ¼ 7.5, 3H); 13C NMR (500 MHz,
NMR (500 MHz, CDCl3)
d 157.57, 156.64, 150.10, 141.53, 132.92,
129.96, 126.91, 123.52, 113.35, 55.09, 50.86, 41.17, 27.51, 25.58, 18.18,
12.06, ꢀ4.50; MS (CI) m/z 372 (Mþ þ 1, 100).
5.2.5. 5-(tert-Butyl-dimethyl-silanyloxy)-2-[1-(4-methoxy-
benzyl)-butyl]-pyridine (16)
According to the general alkylation procedure, lithiation of
890 mg of 13 was performed as described in the case of compound
14. Adding propyl bromide (1st addition: 283
mL, 3.1 mmol, in
acetone-d6) d 156.17, 155.74, 152.52, 138.07, 132.68, 130.73, 124.38,
0.5 mL THF, 2nd addition: 283 L pure) to a stirred solution of
m
122.82, 115.58, 51.50, 41.86, 28.42, 12.32; MS (EI) m/z 243 (Mþ, 43).
HRMS (EI) calcd for C15H17NO2: 243.1259, found 243.1260. Anal.
(C15H17NO2)$0.2H2O: Calcd. C, 73.08; H, 8.20; N, 5.33. Found C, 73.10; H,
7.46; N, 5.73.
lithiated 13 gave 16 as a colorless oil (605 mg, 60%) after reaction
work up and purification (10% EtOAc/hexanes). 1H NMR (500 MHz,
CDCl3)
d
8.19 (d, J ¼ 2.8, 1H), 6.96 (dd, J ¼ 8.4, 2.8, 1H), 6.88 (d,
J ¼ 8.8, 2H), 6.76 (d, J ¼ 8.4, 1H), 6.70 (d, J ¼ 8.8, 2H), 3.74 (s, 3H),
2.93–2.81 (m, 3H), 1.77–1.69 (m, 1H), 1.65–1.58 (m, 1H), 1.19–1.10
(m, 2H), 0.98 (s, 9H), 0.83 (t, J ¼ 7.3, 3H), 0.20 (s, 6H); 13C NMR
5.2.10. 6-[1-(4-Hydroxy-benzyl)-butyl]-pyridin-3-ol (20)
According to the general deprotection procedure, compound 16
(216 mg, 0.56 mmol) in 5.7 mL acetic acid was added to 1.4 mL hydri-
odic acid (57 wt.% in water) in 5.7 mL acetic acid and refluxed for 2 h.
After reaction work up, purification by flash chromatography (80%
EtOAc/hexanes) and re-crystallization from EtOAc/hexanes, 20 was
obtained as a white solid (108 mg, 75%, mp 157 ꢁC). 1H NMR (500 MHz,
acetone-d6) 8.15 (d, J ¼ 2.9, 1H), 7.01 (dd, J ¼ 8.4, 2.9, 1H), 6.85–6.82 (m,
3H), 6.62 (d, J ¼ 8.6, 2H), 2.92–2.84 (m, 2H), 2.79–2.73 (m, 1H), 1.75–
1.68 (m, 1H), 1.59–1.52 (m, 1H), 1.16–1.03 (m, 2H), 0.79 (t, J ¼ 7.4, 3H);
(500 MHz, CDCl3)
d 157.57, 156.89, 150.08, 141.48, 132.91, 129.95,
126.95, 123.42, 113.34, 55.10, 48.90, 41.45, 36.89, 25.58, 20.66, 18.18,
14.08, ꢀ4.50; MS (CI) m/z 386 (Mþ þ 1, 100).
5.2.6. General deprotection procedure
A solution of starting material in acetic acid was added dropwise
to a solution of hydriodic acid (57 wt.% in water) in acetic acid. The
mixture was refluxed, cooled to rt as soon as deprotection was
complete and diluted with water. By adding NaOH (10% aqueous
solution) and NaHCO3 (sat. aqueous solution), a slightly basic pH
value was adjusted (buffer area of NaHCO3). The mixture was
extracted three times with EtOAc and the combined organic extracts
were washed with NaHSO3 (10% aqueous solution). The organic
phase was dried over Na2SO4, the solvent was removed under
vacuum, and purification was performed by flash chromatography.
13C NMR (500 MHz, acetone-d6)
d 156.17, 156.00,152.47,138.09,132.69,
130.73, 124.23, 122.76, 115.57, 49.49, 42.18, 37.96, 21.25, 14.36; MS (EI)
m/z 257 (Mþ, 32). HRMS (EI) calcd for C16H19NO2: 257.1416, found
257.1421.
5.2.11. 1-(1-Chloro-ethyl)-4-methoxy-benzene (23)
SOCl2 (2.47 mL, 34.0 mmol) in 6 mL CH2Cl2 was added to a stir-
red solution of alcohol 21 (4.32 g, 28.4 mmol) in 20 mL CH2Cl2 at
ꢀ78 ꢁC over a time period of 15 min. The reaction was quenched
with water at low temperature and extracted three times with
EtOAc. The organic extracts were dried over Na2SO4 and the solvent
was removed under vacuum. After distillation using a ‘‘Kugelrohr’’,
23 was obtained as a colorless oil (3.62 g, 75%, boiling at 90 ꢁC/
5.2.7. 6-[2-(4-Hydroxy-phenyl)-ethyl]-pyridin-3-ol (17)
According to the general deprotection procedure, compound 13
(765 mg, 2.23 mmol) in 12 mL acetic acid was added to 3 mL
hydriodic acid (57 wt.% in water) in 12 mL acetic acid and refluxed
for 4 h. After reaction work up, purification by flash chromatog-
raphy (pure EtOAc) and re-crystallization from EtOAc/hexane, 17
was obtained as a white solid (308 mg, 64%, mp 136 ꢁC). 1H NMR
2.5 Torr). 1H NMR (500 MHz, CDCl3)
J ¼ 8.8, 2H), 5.10 (quart, J ¼ 6.8, 1H), 3.81 (s, 3H), 1.85 (d, J ¼ 6.8, 3H);
d
7.35 (d, J ¼ 8.8, 2H), 6.88 (d,
(500 MHz, acetone-d6)
d
8.14 (d, J ¼ 2.9, 1H), 7.12 (dd, J ¼ 8.4, 2.9,
13C NMR (500 MHz, CDCl3)
d 159.34, 134.94, 127.74, 113.86, 58.76,
1H), 7.02–6.99 (m, 3H) 6.71 (d, J ¼ 8.6, 2H), 2.94–2.84 (m, 4H); 13C
55.24, 26.37; MS (CI) m/z 135 (Mþ ꢀ 35, 100).
NMR (500 MHz, acetone-d6) d 156.34, 153.11, 152.61, 137.90, 133.52,
130.14, 123.87, 123.28, 115.85, 40.09, 35.97; MS (EI) m/z 215 (Mþ,
72). HRMS (EI) calcd for C13H13NO2: 215.0946, found 215.0943.
Anal. C13H13NO2$0.2H2O: Calcd. C, 71.34; H, 6.17; N, 6.30. Found C,
71.11; H, 6.09; N, 6.38.
5.2.12. 1-(1-Chloro-propyl)-4-methoxy-benzene (24)
The reaction was performed as described above for 23. Usage
of SOCl2 (3.94 mL, 54.24 mmol) in 10 mL CH2Cl2, and alcohol 22
(7.50 g, 45.2 mmol) in 30 mL CH2Cl2 gave 24 as a colorless oil
(6.91 g, 83%, boiling at 95 ꢁC/1.5 Torr) after purification by distil-
5.2.8. 6-[2-(4-Hydroxy-phenyl)-1-methyl-ethyl]-pyridin-3-ol (18)
According to the general deprotection procedure, compound 14
lation. 1H NMR (500 MHz, CDCl3)
d
7.32 (d, J ¼ 8.8, 2H), 6.89
(d, J ¼ 8.8, 2H), 4.79 (t, J ¼ 7.3, 1H), 3.81 (s, 3H), 2.21–2.03 (m, 2H),
(253 mg, 0.71 mmol) in 3.8 mL acetic acid was added to 955
mL
1.00 (d, J ¼ 7.3, 3H); 13C NMR (500 MHz, CDCl3)
d 159.38, 133.92,

3568
M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
128.19, 113.86, 65.44, 55.22, 33.10, 11.75; MS (CI) m/z 149
(Mþ ꢀ 35, 100).
5.2.17. Mixture of cis- and trans-5-(tert-butyl-dimethyl-
silanyloxy)-2-[1-ethyl-2-(4-methoxy-phenyl)-propyl]-pyridine (29)
According to the general alkylation procedure, 25 was lithi-
ated by adding 25 (3.56 g, 9.97 mmol, in 3.8 mL THF) to a solution
of LDA (5.00 mL of a 2 M solution in heptane/THF/ethylbenzene)
5.2.13. 5-(tert-Butyl-dimethyl-silanyloxy)-2-[2-(4-methoxy-
phenyl)-propyl]-pyridine (25)
The reaction was carried out as described above for the
synthesis of compound 13. Lithiation of 12 was performed by
adding 12 (5.06 g, 22.70 mmol, in 4.2 mL THF) to a solution of LDA
(13.62 mL of a 2 M solution in heptane/THF/ethylbenzene) in
4.2 mL THF. Adding 23 (3.86 g, 22.70 mmol, in 4.2 mL THF) to
a stirred solution of lithiated 12 gave 25 as a yellow oil (3.57 g, 44%)
after reaction work up and purification (10% EtOAc/hexanes). 1H
in 3.8 mL THF. Adding ethyl bromide (1st addition: 750
mL,
10.05 mmol, in 3.8 mL THF, 2nd addition: 750 L pure) to a stirred
m
solution of lithiated 25 gave the mixture of diastereomers 29 as
a yellow oil (cis-isomer: 1.19 g, 31%, trans-isomer: 470 mg, 12%)
after reaction work up and purification (10% EtOAc/hexanes). 1H
NMR (500 MHz, acetone-d6)
d
8.24 (d, J ¼ 2.9, 0.4H), 8.11 (d,
J ¼ 2.9, 1H), 7.14 (d, J ¼ 8.8, 0.8H), 7.08 (dd, J ¼ 2.9, J ¼ 8.4, 0.4H),
6.97 (d, J ¼ 8.4, 0.4H), 6.87–6.82 (m, 3.8H) 6.64 (d, J ¼ 8.8, 2H),
6.54 (d, J ¼ 8.4, 1H), 3.80 (s, 1.2H), 3.71 (s, 3H), 3.05–2.93 (m,
1.4H), 2.71–2.60 (m, 1.4H), 1.93–1.85 (m, 1H), 1.80–1.71 (m, 1H),
1.53–1.47 (m, 0.4H), 1.44–1.38 (m, 0.4H), 1.31 (d, J ¼ 7.1, 3H), 1.00
(s, 3.6H), 0.95–0.92 (m, 10.2H), 0.70 (t, J ¼ 7.4, 3H), 0.54 (t, J ¼ 7.4,
1.2H), 0.23 (s, 2.4H), 0.15 (d, J ¼ 1.7, 6H). 13C NMR (500 MHz,
NMR (500 MHz, CDCl3)
d
8.14 (d, J ¼ 2.8, 1H), 7.08 (d, J ¼ 8.8, 2H),
6.96 (dd, J ¼ 2.8, 8.4, 1H), 6.79 (d, J ¼ 8.8, 2H), 6.77 (d, J ¼ 8.4, 1H),
3.77 (s, 3H), 3.21–3.14 (m,1H), 2.96–2.86 (m, 2H),1.23 (t, J ¼ 7.1, 3H),
0.97 (s, 9H), 0.19 (s, 6H); 13C NMR (500 MHz, CDCl3)
d 157.75,153.35,
150.10, 141.47, 138.86, 127.87, 127.00, 123.80, 113.59, 55.15, 46.49,
39.69, 25.57, 21.40, 18.16, ꢀ4.53; MS (CI) m/z 358 (Mþ þ 1, 100).
acetone-d6)
d 157.79, 157.34, 156.46, 156.12, 150.12, 149.68, 141.69,
5.2.14. 5-(tert-Butyl-dimethyl-silanyloxy)-2-[2-(4-methoxy-
phenyl)-butyl]-pyridine (26)
141.19, 138.77, 138.42, 128.44, 126.82, 126.41, 124.11, 124.05,
113.64, 113.05, 56.02, 55.89, 55.15, 55.01, 44.52, 44.03, 26.52,
25.58, 24.67, 21.12, 19.20, 18.17, 12.92, 12.15, ꢀ4.49, ꢀ4.45; MS (CI)
m/z 386 (Mþ þ 1, 100).
The reaction was carried out as described above for the
synthesis of compound 13. Lithiation of 12 was performed by
adding 12 (5.22 g, 23.41 mmol, in 4.3 mL THF) to a solution of LDA
(14.00 mL of a 2 M solution in heptane/THF/ethylbenzene) in
4.3 mL THF. Adding 24 (4.30 g, 23.37 mmol, in 4.3 mL THF) to
a stirred solution of lithiated 12 gave 26 as a yellow oil (3.38 g, 39%)
after reaction work up and purification (10% EtOAc/hexanes). 1H
5.2.18. Mixture of cis- and trans-5-(tert-butyl-dimethyl-
silanyloxy)-2-[1-ethyl-2-(4-methoxy-phenyl)-butyl]-pyridine (30)
According to the general alkylation procedure, 26 was lithi-
ated by adding 26 (2.90 g, 7.82 mmol, in 3 mL THF) to a solution
of LDA (4.69 mL of a 2 M solution in heptane/THF/ethylbenzene)
NMR (500 MHz, CDCl3)
d
8.11 (d, J ¼ 2.8, 1H), 6.99 (d, J ¼ 8.7, 2H),
6.90 (dd, J ¼ 8.4, 2.8, 1H), 6.76 (d, J ¼ 8.7, 2H), 6.68 (d, J ¼ 8.4, 1H),
3.76 (s, 3H), 3.04–2.98 (m, 1H), 2.91–2.84 (m, 2H),1.74–1.66 (m,1H),
1.64–1.55 (m, 1H), 0.96 (s, 9H), 0.77 (t, J ¼ 7.4, 3H), 0.17 (s, 6H); 13C
in 3 mL THF. Adding ethyl bromide (1st addition: 700
mL,
9.37 mmol, in 3 mL THF, 2nd addition: 700 L pure) to a stirred
m
solution of lithiated 26 gave a 1:1 mixture of diastereomers 30 as
a yellow oil (cis-isomer: 374 mg, 12%, trans-isomer: 374 mg, 12%)
after reaction work up and purification (10% EtOAc/hexanes). 1H
NMR (500 MHz, CDCl3)
d 157.71, 153.43, 149.98, 141.39, 136.80,
128.60, 126.90, 123.82, 113.47, 55.09, 47.52, 45.00, 28.69, 25.57,
18.14, 12.05, ꢀ4.53; MS (CI) m/z 372 (Mþ þ 1, 100).
NMR (500 MHz, CDCl3)
d
8.24 (d, J ¼ 2.8, 1H), 8.08 (d, J ¼ 2.8, 1H),
7.11–7.06 (m, 3H), 6.96 (d, J ¼ 8.4, 1H), 6.85 (d, J ¼ 8.6, 2H), 6.81–
6.76 (m, 3H), 6.62 (d, J ¼ 8.8, 2H), 6.50 (d, J ¼ 8.4, 1H), 3.80 (s, 3H),
3.70 (s, 3H), 2.79–2.66 (m, 4H), 2.01–1.88 (m, 2H), 1.76–1.68 (m,
1H), 1.60–1.51 (m, 1H), 1.48–1.21 (m, 4H), 0.99 (s, 9H), 0.94 (s, 9H),
0.73–0.69 (m, 6H), 0.54–0.49 (m, 6H), 0.23 (s, 6H), 0.14 (d, J ¼ 1.7,
5.2.15. 6-[2-(4-Hydroxy-phenyl)-propyl]-pyridin-3-ol (27)
According to the general deprotection procedure, compound 25
(300 mg, 0.84 mmol) in 4.3 mL acetic acid was added to 1.14 mL
hydriodic acid (57 wt.% in water) in 4.3 mL acetic acid and refluxed
for 2 h. After reaction work up, purification by flash chromatog-
raphy (80% ether/CH2Cl2) and additional purification by preparative
TLC (80% ether/CH2Cl2), 27 was obtained as a white solid (136 mg,
71%, mp 152 ꢁC). 1H NMR (500 MHz, acetone-d6) 8.10 (d, J ¼ 2.9,1H),
7.09 (dd, J ¼ 2.9, 8.4, 1H), 7.03 (d, J ¼ 8.5, 2H), 6.88 (d, J ¼ 8.4, 1H),
6.71 (d, J ¼ 8.5, 2H), 3.18–3.10 (m, 1H), 2.91–2.70 (m, 2H), 1.15 (d,
6H); 13C NMR (500 MHz, CDCl3)
d 157.78, 157.32, 156.73, 156.16,
150.08, 149.61, 141.67, 141.09, 136.30, 135.75, 129.37, 129.23,
126.79, 126.34, 124.17, 123.98, 113.53, 112.90, 55.14, 55.04, 54.98,
54.74, 52.31, 51.97, 27.42, 26.54, 25.77, 25.59, 25.26, 18.18, 12.27,
12.10, ꢀ4.47, ꢀ4.56; MS (CI) m/z 400 (Mþ þ 1, 100).
J ¼ 6.9, 3H); 13C NMR (500 MHz, acetone-d6)
d 156.36, 152.61,
5.2.19. Separated cis- and trans-6-[1-ethyl-2-(4-hydroxy-phenyl)-
propyl]-pyridin-3-ol (31a/31b)
152.48, 138.66, 137.77, 128.64, 124.60, 123.20, 115.84, 46.67, 40.31,
22.02; MS (CI) m/z 230 (Mþ þ 1, 26). HRMS (CI) calcd for C14H16NO2:
230.1181, found 230.1182.
According to the general deprotection procedure, the mixture of
diastereomers 29 (780 mg; cis-isomer: 562 mg, 1.46 mmol, trans:
218 mg, 0.57 mmol) in 22 mL acetic acid was added to 5.5 mL
hydriodic acid (57 wt.% in water) in 22 mL acetic acid and refluxed
for 1.5 h. After reaction work up, the two diastereomers were
purified by flash chromatography (75% ether/CH2Cl2) to give two
fractions, each one enriched in one of the two isomers. Subse-
quently, preparative TLCs (50% ether/CH2Cl2) were performed with
both fractions to give the pure cis-isomer 31a (white solid, 371 mg,
66%, mp 90–95 ꢁC) and trans-isomer 31b (white solid, 135 mg, 62%,
mp 80–85 ꢁC). 31a: 1H NMR (500 MHz, acetone-d6) 8.07 (d, J ¼ 2.9,
1H), 6.89 (dd, J ¼ 2.8, 8.5, 1H), 6.82 (d, J ¼ 8.5, 2H), 6.66 (d, J ¼ 8.5,
1H), 6.56 (d, J ¼ 8.5, 2H), 3.03–2.97 (m, 1H), 2.73–2.69 (m, 1H), 1.86–
1.71 (m, 2H), 1.24 (d, J ¼ 7.1, 3H), 0.63 (t, J ¼ 7.4, 3H); 13C NMR
5.2.16. 6-[2-(4-Hydroxy-phenyl)-butyl]-pyridin-3-ol (28)
According to the general deprotection procedure, compound
26 (200 mg, 0.53 mmol) in 5.6 mL acetic acid was added to
1.46 mL hydriodic acid (57 wt.% in water) in 5.6 mL acetic acid and
refluxed for 2 h. After reaction work up, purification by flash
chromatography (80% ether/CH2Cl2) and additional purification
by preparative TLC (80% ether/CH2Cl2), 28 was obtained as
a white solid (74 mg, 57%, mp 83–88 ꢁC). 1H NMR (500 MHz,
acetone-d6) 8.08 (d, J ¼ 2.9, 1H), 6.99–6.95 (m, 3H), 6.81 (d, J ¼ 8.4,
1H), 6.70 (d, J ¼ 8.6, 2H), 2.95–2.81 (m, 3H), 1.68–1.60 (m, 1H),
1.57–1.49 (m, 1H), 0.72 (t, J ¼ 7.4, 3H); 13C NMR (400 MHz,
acetone-d6) d 156.30, 152.50, 152.41, 137.70, 136.49, 129.40, 124.54,
122.96, 115.71, 48.08, 45.14, 29.45, 12.35; MS (CI) m/z 244 (Mþ þ 1,
(500 MHz, acetone-d6)
d 155.93, 155.25, 152.19, 138.30, 137.56,
129.25, 125.00, 122.50, 115.35, 56.16, 44.61, 25.17, 19.70, 12.53; MS
13). HRMS (CI) calcd for C15H18NO2: 244.1338, found 244.1341.

M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
3569
(CI) m/z 258 (Mþ þ 1, 55). HRMS (CI) calcd for C16H20NO2: 258.1494,
found 258.1497. 31b: 1H NMR (500 MHz, acetone-d6) 8.23 (d,
J ¼ 2.8, 1H), 7.17 (dd, J ¼ 2.8, 8.4, 1H), 7.09–7.06 (m, 3H), 6.78 (d,
J ¼ 8.4, 2H), 2.96–2.90 (m, 1H), 2.67–2.62 (m, 1H), 1.56–1.46 (m, 1H),
1.40–1.31 (m, 1H), 0.86 (d, J ¼ 7.1, 3H), 0.49 (t, J ¼ 7.4, 3H); 13C NMR
incubated at 37 ꢁC in a 5% CO2-containing incubator for 6 h. The
medium was then replaced with fresh Improved MEM supple-
mented with 5% CDCS plus the desired concentrations of ligands.
Cells were harvested 24 h later. Luciferase activity and
sidase activity were assayed as described [41].
b-galacto-
(500 MHz, acetone-d6)
d 156.41, 155.49, 152.64, 138.52, 138.21,
129.19, 125.11, 122.92, 115.90, 56.35, 45.20, 27.23, 21.71, 12.43; MS
(CI) m/z 258 (Mþ þ 1, 100). HRMS (CI) calcd for C16H20NO2:
258.1494, found 258.1486.
Acknowledgments
This work was supported through grants from the National
Institutes of Health (R37DK15556 and P01AG024387 to J.A.K., and
R01CA18119 to B.S.K.). We are grateful to Dr. Meri De Angelis for
chemical synthesis advice and contributions to the start of this
project.
5.2.20. Separated cis- and trans-6-[1-ethyl-2-(4-hydroxy-phenyl)-
butyl]-pyridin-3-ol (32a/32b)
According to the general deprotection procedure, the mixture of
diastereomers 30 (374 mg (187 mg of each isomer), 0.94 mmol) in
20 mL acetic acid was added to 2.6 mL hydriodic acid (57 wt.% in
water) in 10 mL acetic acid and was refluxed for 1.5 h. After reaction
work up, the two diastereomers were purified by flash chroma-
tography (75% ether/CH2Cl2) to give two fractions, each one
enriched in one of the two isomers. Subsequently, preparative TLCs
(50% ether/CH2Cl2) were performed with both fractions to give the
pure cis-isomer 32a (white solid, 91 mg, 72%, mp 83–88 ꢁC) and
trans-isomer 32b (white solid, 95 mg, 75%, mp 83–88 ꢁC). 32a: 1H
NMR (500 MHz, acetone-d6) 8.05 (d, J ¼ 2.8, 1H), 6.87 (dd, J ¼ 2.8,
8.4, 1H), 6.76 (d, J ¼ 8.6, 2H), 6.62 (d, J ¼ 8.4,1H), 6.57 (d, J ¼ 8.6, 2H),
2.81–2.73 (m, 2H), 1.97–1.82 (m, 2H), 1.78–1.69 (m, 1H), 1.58–1.49
(m, 1H), 0.69–0.63 (m, 6H); 13C NMR (500 MHz, acetone-d6)
References
[1] J.A. Katzenellenbogen, B.S. Katzenellenbogen, Chem. Biol. 3 (1996) 529–536.
[2] L. Gennari, D. Merlotti, F. Valleggi, G. Martini, R. Nuti, Drugs Aging 24 (2007)
361–379.
[3] M.E. Mendelsohn, Am. J. Cardiol. 89 (2002) 12E–17E discussion 17E–18E.
[4] C. Behl, Nat. Rev. Neurosci. 3 (2002) 433–442.
[5] S.A. Beresford, N.S. Weiss, L.F. Voigt, B. McKnight, Lancet 349 (1997) 458–461.
[6] B. Zumoff, Proc. Soc. Exp. Biol. Med. 217 (1998) 30–37.
[7] D.P. McDonnell, Trends Endocrinol. Metab. 10 (1999) 301–311.
[8] B. Fisher, J.P. Costantino, D.L. Wickerham, C.K. Redmond, M. Kavanah,
W.M. Cronin, V. Vogel, A. Robidoux, N. Dimitrov, J. Atkins, M. Daly, S. Wieand,
E. Tan-Chiu, L. Ford, N. Wolmark, J. Natl. Cancer Inst. 90 (1998) 1371–1388.
[9] S.R. Cummings, S. Eckert, K.A. Krueger, D. Grady, T.J. Powles, J.A. Cauley,
L. Norton, T. Nickelsen, N.H. Bjarnason, M. Morrow, M.E. Lippman, D. Black,
J.E. Glusman, A. Costa, V.C. Jordan, JAMA J. Am. Med. Assoc. 281 (1999) 2189–
2197.
[10] S. Palacios, Menopause Int. 13 (2007) 27–34.
[11] J.A. Katzenellenbogen, B.W. O’Malley, B.S. Katzenellenbogen, Mol. Endocrinol.
10 (1996) 119–131.
[12] G.G. Kuiper, B. Carlsson, K. Grandien, E. Enmark, J. Haggblad, S. Nilsson,
J.A. Gustafsson, Endocrinology 138 (1997) 863–870.
[13] G.G. Kuiper, E. Enmark, M. Pelto-Huikko, S. Nilsson, J.A. Gustafsson, Proc. Natl.
Acad. Sci. U.S.A. 93 (1996) 5925–5930.
d
155.94, 155.27, 152.06, 137.54, 135.54, 130.24, 125.02, 122.29,
115.24, 54.92, 52.52, 26.30, 25.70, 12.55; MS (CI) m/z 272 (Mþ þ 1,
66). HRMS (CI) calcd for C17H22NO2: 272.1651, found 272.1650. 32b:
1H NMR (500 MHz, acetone-d6) 8.21 (d, J ¼ 2.9, 1H), 7.14 (dd, J ¼ 2.9,
8.4, 1H), 7.06–7.04 (m, 3H), 6.79 (d, J ¼ 8.4, 2H), 2.71–2.67 (m, 2H),
1.50–1.41 (m, 1H), 1.35–1.16 (m, 3H), 0.50–0.45 (m, 6H); 13C NMR
(500 MHz, acetone-d6)
d 156.48, 155.77, 152.64, 138.23, 135.97,
[14] S. Mosselman, J. Polman, R. Dijkema, FEBS Lett. 392 (1996) 49–53.
[15] S. Paruthiyil, H. Parmar, V. Kerekatte, G.R. Cunha, G.L. Firestone, D.C. Leitman,
Cancer Res. 64 (2004) 423–428.
[16] E.C. Chang, J. Frasor, B. Komm, B.S. Katzenellenbogen, Endocrinology 147
(2006) 4831–4842.
130.02, 125.07, 122.98, 115.89, 54.45, 53.01, 28.14, 27.30, 12.44,
12.36; MS (CI) m/z 272 (Mþ þ 1, 100). HRMS (CI) calcd for
C17H22NO2: 272.1651, found 272.1659.
[17] A.C. Pike, A.M. Brzozowski, R.E. Hubbard, T. Bonn, A.G. Thorsell, O. Engstrom,
J. Ljunggren, J.A. Gustafsson, M. Carlquist, EMBO J. 18 (1999) 4608–4618.
[18] D.R. Compton, K.E. Carlson, J.A. Katzenellenbogen, Bioorg. Med. Chem. Lett. 14
(2004) 5681–5684.
[19] D.R. Compton, S. Sheng, K.E. Carlson, N.A. Rebacz, I.Y. Lee, B.S. Katzenellenbogen,
J.A. Katzenellenbogen, J. Med. Chem. 47 (2004) 5872–5893.
[20] M. De Angelis, J.A. Katzenellenbogen, Bioorg. Med. Chem. Lett. 14 (2004)
5835–5839.
5.3. Estrogen receptor binding affinity assays
Relative binding affinities were determined by a competitive
radiometric binding assay as previously described [37,38], using
10 nM [3H]estradiol as tracer (GE Healthcare, Piscataway, NJ), and
purified full-length human ER
a and ERb (PanVera/InVitrogen,
[21] M. De Angelis, F. Stossi, M. Waibel, B.S. Katzenellenbogen, J.A. Katzenellenbogen,
Bioorg. Med. Chem. 13 (2005) 6529–6542.
Carlsbad, CA). Incubations were for 18–24 h at 0 ꢁC, then the
receptor–ligand complexes were adsorbed onto hydroxyapatite
(BioRad, Hercules, CA) and unbound ligand was washed away. The
binding affinities are expressed as relative binding affinity (RBA)
values, with the RBA of estradiol set to 100. The values given are the
average ꢂ range or SD of two or more independent determinations.
[22] M.J. Meyers, J. Sun, K.E. Carlson, G.A. Marriner, B.S. Katzenellenbogen,
J.A. Katzenellenbogen, J. Med. Chem. 44 (2001) 4230–4251.
[23] B. Barlaam, P. Bernstein, C. Dantzman, P. Warwick, Preparation of benzox-
azoles and benzothiazoles as selective ligands for human b-estrogen receptor,
in PCT Int. Appl. (Astrazeneca AB, Swed.). An 2002:504768, WO02/051821,
2002, pp. 71.
[24] B. Barlaam, S. Dock, J. Folmer, Preparation of benzimidazoles as selective
estrogen receptor-b ligand, in PCT Int. Appl. (Astrazeneca AB, Swed.). An
2002:449659, WO02/046168, 2002, pp. 46.
Estradiol binds to ER
0.5 nM.
a with a Kd of 0.2 nM and to ERb with a Kd of
[25] H.A. Harris, L.M. Albert, Y. Leathurby, M.S. Malamas, R.E. Mewshaw, C.P. Miller,
Y.P. Kharode, J. Marzolf, B.S. Komm, R.C. Winneker, D.E. Frail, R.A. Henderson,
Y. Zhu, J.C. Keith Jr., Endocrinology 144 (2003) 4241–4249.
[26] M.S. Malamas, E.S. Manas, R.E. McDevitt, I. Gunawan, Z.B. Xu, M.D. Collini,
C.P. Miller, T. Dinh, R.A. Henderson, J.C. Keith Jr., H.A. Harris, J. Med. Chem. 47
(2004) 5021–5040.
[27] U. Schopfer, P. Schoeffter, S.F. Bischoff, J. Nozulak, D. Feuerbach, P. Floersheim,
J. Med. Chem. 45 (2002) 1399–1401.
[28] M. Waibel, M. De Angelis, F. Stossi, K.J. Kieser, K.E. Carlson,
B.S. Katzenellenbogen, J.A. Katzenellenbogen, Eur. J. Med. Chem. (2009).
[29] M.R. Kilbourn, A.J. Arduengo, J.T. Park, J.A. Katzenellenbogen, Mol. Pharmacol.
19 (1981) 388–398.
[30] M. De Angelis, F. Stossi, K.A. Carlson, B.S. Katzenellenbogen,
J.A. Katzenellenbogen, J. Med. Chem. 48 (2005) 1132–1144.
[31] H. Gao, J.A. Katzenellenbogen, R. Garg, C. Hansch, Chem. Rev. 99 (1999) 723–744.
[32] R.W. Chesnut, R.R. Cesati, C.S. Cutler, S.L. Pluth, J.A. Katzenellenbogen,
Organometallics 17 (1998) 4889–4896.
5.4. Cell culture and transient transfections
Human endometrial cancer (HEC-1) cells were maintained in
Minimum Essential Medium (MEM) plus phenol-red supplemented
with 5% calf serum and were changed to phenol-red-free Improved
MEM and 5% charcoal dextran-treated calf serum (CDCS) for 3–4
days before use in experiments. Transfection assays were per-
formed in 24-well plates using a mixture of 0.35 mL of serum-free
improved MEM medium and 0.15 mL of Hank’s balanced salt
solution containing 5
thersburg, MD), 20 g of transferrin (Sigma, St. Louis, MO), 0.2
pCMV -galactosidase as internal control, 0.5 g of 2ERE-pS2-Luc,
and 50 ng of ER expression vector per well. The cells were
mg of lipofectin (Life Technologies, Inc., Gai-
m
mg of
b
m
[33] R. Tedesco, R. Fiaschi, E. Napolitano, Synthesis (1995) 1493–1495.

3570
M. Waibel et al. / European Journal of Medicinal Chemistry 44 (2009) 3560–3570
[34] R.B. Bates, T.J. Siahaan, J. Org. Chem. 51 (1986) 1432–1434.
[35] S. Takagishi, M. Schlosser, Synlett (1991) 119–121.
[36] P. Wyatt, H. Eley, J. Charmant, B.J. Daniel, A. Kantacha, Eur. J. Org. Chem. (2003)
4216–4226.
[37] K.E. Carlson, I. Choi, A. Gee, B.S. Katzenellenbogen, J.A. Katzenellenbogen,
Biochemistry 36 (1997) 14897–14905.
[39] B.E. Fink, D.S. Mortensen, S.R. Stauffer, Z.D. Aron, J.A. Katzenellenbogen, Chem.
Biol. 6 (1999) 205–219.
[40] D.S. Mortensen, A.L. Rodriguez, K.E. Carlson, J. Sun, B.S. Katzenellenbogen,
J.A. Katzenellenbogen, J. Med. Chem. 44 (2001) 3838–3848.
[41] E.M. McInerney, M.J. Tsai, B.W. O’Malley, B.S. Katzenellenbogen, Proc. Natl.
Acad. Sci. U.S.A. 93 (1996) 10069–10073.
[38] J.A. Katzenellenbogen, H.J. Johnson Jr., H.N. Myers, Biochemistry 12 (1973)
4085–4092.
[42] A.B. Pangborn, M.A. Giardello, R.H. Grubbs, R.K. Rosen, F.J. Timmers, Organo-
metallics 15 (1996) 1518–1520.