Design, synthesis, and biological evaluation of new raloxifene analogues of improved antagonist activity and endometrial safety

Article abstract of DOI:10.1016/j.bioorg.2020.104482

Raloxifene agonism of estrogen receptor (ER) in post-menopausal endometrium is not negligible. Based on a rational drug design workflow, we synthesized 14 analogues of raloxifene bearing a polar group in the aromatic ring of the basic side chain (BSC) and/or changes in the bulkiness of the BSC amino group. Analogues with a polar BSC aromatic ring and amino group substituents of increasing volume displayed increasing ER antagonism in Ishikawa cells. Analogues with cyclohexylaminoethoxy (13a) or adamantylaminoethoxy BSC (13b) lacking a polar aromatic ring displayed high ER-binding affinity and ER antagonism in Ishikawa cells higher than raloxifene and similar to fulvestrant (ICI182,780). The endometrial surface epithelium of immature female CD1 mice injected with 13b was comparable to that of vehicle-treated mice, while that of mice treated with estradiol, raloxifene or 13b in combination with estradiol was hyperplastic. These findings indicate that raloxifene analogues with a bulky BSC amino group could provide for higher endometrial safety treatment of the menopausal syndrome.

Full text of DOI:10.1016/j.bioorg.2020.104482

Journal Pre-proofs
Design, Synthesis, and Biological Evaluation of New Raloxifene Analogues
of Improved Antagonist Activity and Endometrial Safety
George Lambrinidis, Cedric Gouedard, Sotiria Stasinopoulou, Angeliki
Angelopoulou, Vassiliki Ganou, Aggeliki K. Meligova, Dimitra J. Mitsiou,
Panagiotis Marakos, Nicole Pouli, Emmanuel Mikros, Michael N. Alexis
PII:
S0045-2068(20)31780-6
DOI:
https://doi.org/10.1016/j.bioorg.2020.104482
YBIOO 104482
Reference:
To appear in:
Bioorganic Chemistry
Received Date:
Revised Date:
Accepted Date:
2 August 2020
21 October 2020
13 November 2020
Please cite this article as: G. Lambrinidis, C. Gouedard, S. Stasinopoulou, A. Angelopoulou, V. Ganou, A.K.
Meligova, D.J. Mitsiou, P. Marakos, N. Pouli, E. Mikros, M.N. Alexis, Design, Synthesis, and Biological
Evaluation of New Raloxifene Analogues of Improved Antagonist Activity and Endometrial Safety, Bioorganic
Chemistry (2020), doi: https://doi.org/10.1016/j.bioorg.2020.104482
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier Inc.

1
Design, Synthesis, and Biological Evaluation of New
Raloxifene Analogues of Improved Antagonist Activity
and Endometrial Safety
George Lambrinidis,*[a] Cedric Gouedard,[b] Sotiria Stasinopoulou,[b] Angeliki
Angelopoulou,[b] Vassiliki Ganou,[b] Aggeliki K. Meligova,[b] Dimitra J. Mitsiou,[b]
Panagiotis Marakos,[a] Nicole Pouli,[a] Emmanuel Mikros,*[a] and Michael N. Alexis*[b]
[a] Department of Pharmacy, National and Kapodistrian University of Athens,
Panepistimioupoli Zografou, 15771, Athens, Greece
[b] Institute of Chemical Biology, National Hellenic Research Foundation, 48,
Vassileos Constantinou Avenue, 11635, Athens, Greece
* Corresponding authors’ E-mails: lambrinidis@pharm.uoa.gr; mikros@pharm.uoa.gr;
mnalexis@eie.gr;
Highlights
• Rational design, synthesis and evaluation of 14 basic side chain (BSC) analogues
of raloxifene
• The analogues’ BSC bore a polar group in the aromatic ring and/or changes in
amino group bulkiness
• Analogues with amino group substituents of increasing volume displayed increasing
ER antagonism
• Two analogues w/o a polar aromatic ring had ER-antagonism in Ishikawa cells
higher than raloxifene
• The adamantylaminoethoxy analogue did not stimulate the endometrial epithelium
of immature mice
Abstract:
Raloxifene agonism of estrogen receptor (ER) in post-menopausal endometrium is
not negligible. Based on a rational drug design workflow, we synthesized 14
analogues of raloxifene bearing a polar group in the aromatic ring of the basic side
chain (BSC) and/or changes in the bulkiness of the BSC amino group. Analogues
with a polar BSC aromatic ring and amino group substituents of increasing volume
displayed increasing ER antagonism in Ishikawa cells. Analogues with
cyclohexylaminoethoxy (13a) or adamantylaminoethoxy BSC (13b) lacking a polar
aromatic ring displayed high ER-binding affinity and ER antagonism in Ishikawa cells
higher than raloxifene and similar to fulvestrant (ICI182,780). The endometrial
surface epithelium of immature female CD1 mice injected with 13b was comparable

2
to that of vehicle-treated mice, while that of mice treated with estradiol, raloxifene or
13b in combination with estradiol was hyperplastic. These findings indicate that
raloxifene analogues with a bulky BSC amino group could provide for higher
endometrial safety treatment of the menopausal syndrome.
Graphical Abstract
Keywords: drug design, SERMs, endometrial safety, estrogen receptors, hormones,
raloxifene
1. Introduction
Menopausal disorders can be treated with hormone therapy and more safely
with selective ER modulators (SERMs) [1–3]. FDA-approved SERMs include among
others, tamoxifen (for breast cancer prevention and treatment), raloxifene (for
osteoporosis prevention and treatment and for invasive breast cancer prevention)
and bazedoxifene in combination with equine estrogen (for prevention of
osteoporosis and treatment of vasomotor symptoms) [2]. The health benefits of
estrogen and SERMs are predominantly mediated by ERα and ERβ1, the estrogen-
binding isoform of ERβ. These are ligand-dependent transcription factors capable of
regulating tissue and organ physiology by binding a large structurally diverse group
of extraneous natural and synthetic chemicals besides estrogen [4,5]. ERα and ERβ1
modulate each other’s transcriptional activity by acting as heterodimers as well as
homodimers, with the heterodimers potentially targeting nearly half of the chromatin-
binding sites that are accessible to the homodimers [6]. The pharmacology of SERMs
is determined by the potential of SERM-bound ER to recruit functionally distinct co-
regulators (co-activators and co-repressors) of gene transcription in a cell-dependent
manner, thus displaying different ER-agonist activities in different estrogen target
cells [2,3]. Knockdown of Steroid Receptor Coactivator 1 expression by siRNA
abolished ER-agonism of tamoxifen in uterine cells, suggesting that ER-agonism is
dependent on the relative levels of expression of co-activators versus co-repressors

3
[7]. ER-agonism in the uterus is the trait that predominantly determines the safety of
SERMs [2,3]. Clinical studies have demonstrated that raloxifene maintains most of
the breast cancer chemopreventive potential of tamoxifen and compared to the latter,
it is associated with lower endometrial cancer risk [8]. On the whole, clinical studies
support a safety profile for raloxifene [9]. Nevertheless, preclinical and clinical studies
have shown that raloxifene can stimulate rat and human endometrium [10,11].
Bazedoxifene displayed lower uterotrophic activity compared to raloxifene and when
tested in combination with conjugated estrogen (CE) failed to stimulate the
endomedrium [12]. In contrast, the combination of raloxifene with 17β-estradiol (E2)
or CE caused endometrial hypertrophy and hyperplasia [13].
ER-dependent recruitment of co-regulators to estrogen target gene
enhancers is cell-, promoter- and ligand-specific [3]. Crystallographic studies have
shown that the agonist and antagonist activities of ER ligands are associated with
distinct positions of the carboxyl-terminal alpha-helix 12 (Η12) of the Ligand Binding
Domain (LBD) of ERα and ERβ1 [14–17]; binding of E2 to ERα allows H12 to form
together with LBD H3 and H5 a binding site for co-activators, while binding of
raloxifene allows its BSC to extend out of the ligand-binding pocket (LBP), interact
with Asp351 and relocate H12 to the coactivator-docking site, thus stabilizing the
antagonist conformation [14,17]. Likewise, positioning of H12 in-between the agonist
and antagonist positions is associated with partial ER agonism, while failure of H12
to properly fold over the LBD exposes a sizeable hydrophobic domain, leading to
proteasomal degradation of ER [15,16,18]. Recently, antiestrogens with an
adamantyl core structure and BSC of varying length and/or bulkiness were reported,
including two analogues with adamantylcarbonylamino or 3-hydroxylpropylamido
BSC end, of which the latter displayed ERα-antagonist efficacies comparable to
fulvestrant (ICI182,780), a selective ER degrader (SERD) [19,20].
The aim of this study was to explore the impact of increasing bulkiness of the
BSC amino group of raloxifene on ER agonism as well as the possibility to increase
ER-binding affinity by introducing a forth hydrogen bond. We initially synthesized two
consecutive series of analogues with a BSC amino group of increasing bulkiness and
a hydroxyl or an acetamide group at position 3’ of the BSC aromatic ring for which
preliminary Molecular Docking Simulations (MDS) predicted to form hydrogen bond
with Thr347 of ERα or Thr299 of ERβ. Since the analogues of both these series
failed to display appreciable ER-binding affinity and/or low ER agonism, we next
synthesized cyclohexylamino (13a) and adamantylamino (13b) BSC analogues
lacking a 3’-derivatization and finally their respective amides for which preliminary
MDS predicted similar ERα-binding affinity to raloxifene. Compound 13a has been
patented by Eli Lilly but hasn’t been biologically evaluated. The ER-binding affinity
and inhibitory potency of the analogues and their efficacy of ER agonism and
antagonism in breast and endometrial adenocarcinoma cells were assessed. We
show that derivative 13b is deprived of agonist activity in breast and endometrial cells
and in the immature mouse uterus. These findings indicate that the adamantylamino
BSC may help to develop SERMs of high endometrial safety.
2. Results and Discussion

4
2.1 Rational design of modifications of the BSC of raloxifene:
Crystal structures of raloxifene in complexes with the LBD of ERα and ERβ1
(hereafter referred to as ERβ) revealed that the BSC extends outwards from the
center of the LBP through a predominantly hydrophobic channel to form a salt bridge
between the piperidinium nitrogen and Asp351 of ERα (or Asp303 of ERβ), thereby
displacing H12 [15,16]. We followed the solvent mapping strategy in order to identify
modifications that could increase chemical affinity. Using the SZMAP algorithm, as
implemented on OpenEye Suite Software solvent mapping calculations, we identified
significantly favorable or unfavorable regions of solvent thermodynamics in the LBP
of ERα and ERβ in the absence of ligand (apo form) as shown in (Figure 1A). Water
molecules interacting with the protein and stabilized through H bonds are shown with
yellow spheres, while water molecules forcibly trapped within the protein are shown
in purple. Superimposition of raloxifene shows that both core structure hydroxyl
groups and BSC nitrogen coincide with yellow waters while all aromatic and
hydrophobic surfaces overlap perfectly with purple waters. Drug design-wise
however there is a specific yellow water molecule that does not match to a
counterpart on raloxifene. This specific water molecule interacts with the hydroxyl
group of Thr347 of ERα (Thr299 of ERβ) located in the hydrophobic channel and
oriented towards the aromatic ring of raloxifene BSC at a distance of 3.98Å. This
observation was utilized to design analogues initially with a hydroxyl group and
subsequently with an acetamide group at position 3’ of the BSC aromatic group that
potentially could form one and two H-bonds, respectively, with Thr347 (Thr299) and
thus increase ER-binding affinity. The 3’-derivatized analogues were endowed with
increasingly voluminous amino group substituents to examine whether increasing the
basicity and bulkiness of the BSC amino group could perturb the conformational
equilibrium of H12 and impact ER agonism by volume-induced perturbations (Figure
1B, 5a-e and 9a-e). Since the experimental ER-binding affinity for these analogues
was found to be considerably lower than that of raloxifene, though not as lower for
those with a bulky BSC amino group as for the rest (Table 1 and Figure 2), we next
synthesized cyclohexylamino and adamantylamino BSC analogues lacking 3’-
derivatization (13a, 13b) and finally their amides (20a, 20b) to examine how the
basicity and bulkiness of the BSC affect ER-binding affinity and ER agonism in the
absence of interference from a 3’-derivatization. Notably, MDS indicated that the
mode of binding of 13b to ERα is driving the adamantylamino BSC to the outer space
of the protein, which could stabilize H12 in the antagonist position and thus increase
ER antagonist efficacy. Figure 2 conveys the rationales and the consequential
experimental findings that guided the consecutive rounds of modifications of the BSC
of raloxifene.
2.2 Theoretical Molecular Simulations:
Preliminary docking simulations were run using the Induced Fit Docking (IFD)
algorithm as implemented on Schrödinger Suite 2017. The first 10 analogues fitted
well inside the binding pocket of ER, forming crucial hydrogen bonds. Like in
raloxifene, the secondary amino groups of the BSC of 5d (Figure 1C) forms a salt
bridge with Asp351, while a second inter-molecular hydrogen bond between the
acidic phenolic hydroxyl group of its BSC and Thr347 stabilizes the conformation of
the ligand. These 10 analogues were initially predicted to have similar binding affinity

5
to raloxifene. The average variation of the free energy of binding (GlideScore) was
relatively small, ~1.5 kcal/mol for ERα and ~2 kcal/mol for ERβ, which is close to the
standard error of prediction for molecular mechanic calculations. We then proceeded
to more accurate Free Energy Perturbation (FEP) calculations for both raloxifene and
analogue binding to ERα-LBD. These calculations could attribute differences in
binding affinity to differences in solvation energy. For each pair of perturbations, we
calculated the difference in the free energy of binding (ΔΔG Theoretical) (Table 1).
As not expected, these analogues, with the exception of 5d, displayed lower ER-
binding affinity compared to raloxifene. Nevertheless, we proceeded to their
synthesis in order to evaluate how increasing the bulkiness of the BSC amino group
would affect ER agonism. Then we carried out IFD simulations for the two bulky BSC
analogues lacking a 3’-derivatization (13a, 13b) and their respective amides (7a, 7b)
and FEP calculations for the former (Figure 1D). While IFD simulations predicted for
13a and 13b affinities comparable to or lower than that of raloxifene, FEP
calculations predicted affinities higher than raloxifene (Table 1). In the light of these
findings, we proceeded to the synthesis of all 4 analogues lacking a 3’-derivatization.
Table 1: Predicted logP and experimental
and theoretical ΔΔG of binding to ERα (1 column Table)
Test
Cmp
RBA LogP
ΔG
ΔΔG
ΔΔG
Expa
Pred
Expb
Expc
Predd
Ral
5a
62.9
7.01
6.48
4.69
5.35
5.43
4.59
5.20
4.09
4.80
5.06
4.41
5.23
5.59
6.39
5.63
5.60
-13.37
-12.02
-12.43
-11.98
-12.77
-12.10
-11.96
-11.99
-11.88
-11.82
-12.38
-13.19
-12.68
-12.89
-12.43
0.00
1.35
0.94
1.40
0.61
1.27
1.42
1.39
1.50
1.55
1.00
0.18
0.69
0.49
0.95
0.00
3.54
1.82
2.05
-0.12
1.05
4.79
2.89
4.89
3.85
0.90
-0.95
-2.34
nd
5b
13.68
6.51
5c
5d
23.45
8.01
5e
9a
6.33
9b
6.64
9c
5.53
9d
5.08
9e
12.45
46.61
20.51
28.56
13.51
13a
13b
20a
20b
nd
a Experimental relative binding affinity to ERα (RBAα) is expressed as % of that of raloxifene
(cf Supplementary Table S1); b Experimental ΔG (kcal/mol) of binding was calculated using
c
Cheng Prusoff equation; ΔΔG (kcal/mol): difference between the experimental ΔG of
d
analogues from that of raloxifene;
ΔΔG (kcal/mol) of analogues from raloxifene as
calculated using the FEP algorithm. Cmp: compound; Exp: experimental; Pred: predicted;
Ral: raloxifene; nd: not determined

6
Figure 1 (2 column Figure): (A) SZMAP analysis of ERα-LBD in complex with raloxifene.
Hydrophobic and hydrophilic interfaces are colored purple and yellow, respectively. (B)
Raloxifene analogues designed and synthesized based on SZMAP analysis. (C, D)
Superimposition of the crystal structure of raloxifene (Ral) and the global minimum structure
of 5d (C) and 13b (D) in complex with ERα-LBD; Ad: 1-adamantyl; Cy: cyclohexyl.
Figure 2 (2 column Figure): Design and synthesis of basic side chain (BSC) analogues of
raloxifene: Introduction of a hydroxyl and subsequently of an acetamide group at position 3’ of
the BSC aromatic ring and of increasingly voluminous amino group substituents to the 3’-
derivatives gave rise to analogues of appreciably lower RBAα compared to raloxifene. Since
RBAα decrease was comparatively lower for analogues with bulkier BSC amino groups (5d,
9e), cyclohexylamino and adamantylamino BSC analogues lacking 3’-derivatization (13a,
13b) were synthesized next followed by their amides (20a, 20b) in order to examine how the
basicity and bulkiness of the BSC amino group affected RBAα in the absence of 3’-
derivatization.

7
2.3 Synthesis:
Derivatives 4a-e (Scheme 1A) were prepared from the benzoylchloride 1,
which was easily obtained from the corresponding carboxylic acid [21,22]. The
chloride
1
was then used for the acylation of 6-methoxy-2-(4-
methoxyphenyl)benzo[b]thiophene (2) [23] and the resulting benzothiophene 3 was
treated with the appropriate primary or secondary amines to give compounds 4a-e,
which upon deprotection yielded the target compounds 5a-e. Analogues 9a-e
(Scheme 1B) were prepared from the mixed anhydride 6 by an analogous procedure.
Compound 6 was prepared from ethyl 3-amino-4-hydroxybenzoate [24] by
successive N-acetylation, etherification, ester hydrolysis and treatment of the
substituted benzoic acid with ethyl chloroformate (for compound synthesis, isolation
and characterization see Supplementary Material).
Scheme 1 (2 column Scheme): (A), a) 6-methoxy-2-(4-methoxyphenyl)benzothiophene (2),
AlCl3, 1,2 DCE, rt, 1h; b) Amine, EtOH, reflux, 24h; c) AlCl3, EtSH, CH2Cl2, rt, 2h. (B), a) 2,
AlCl3, 1,2 DCE, rt, 1h; b) Amine, EtOH, rt, 24h; c) AlCl3, EtSH, CH2Cl2, rt, 2h
The secondary amines 13a,b (Scheme 2A) were prepared by the procedure
proposed by Bradley et al [25] using the benzothiophene 2 [23] as starting material.
Compound 2 was acylated with 4-bromobenzylchloride and the bromine atom of the
resulting ketone 10 [26,27] was displaced by 2-hydroxyethoxide to give compound

8
11a [25,27], which was converted to the corresponding mesylate 11b. Treatment of
the mesylate with the suitable amines resulted in the benzothiophenes 12a,b which
upon demethylation produced the target amines 13a and 13b [27]. Finally, the
amides 20a,b (Scheme 2B) were prepared from ethyl 4-hydroxybenzoic acid (15)
which reacted with N-cyclohexyl- or N-(1-adamantyl)-2-chloracetamide to give the
carboxamides 16a and 16b, respectively [both compounds were synthesized,
although they could be purchased from chemical vendors]. The carboxamides were
then converted first to the corresponding carboxylic acids 17a,b and then to the
mixed anhydrides 18a,b, which were used for the Friedel-Crafts acylation of 6-
acetyloxy-2-(4-acetyloxyphenyl)benzo[b]thiophene [23] that provided the benzo-
thiophenes 19a,b. These were deprotected to give the target compounds 20a,b.
Scheme 2 (2 column Scheme): (A), a) EtOH, HCl (gas), reflux, 2h; b) K2CO3, acetone,
ClCH2CONHR1, reflux, 12h; c) ΝaΟΗ 40%, EtOH, rt, 1h; d) toluene, ClCO2Et, Et3N, 0 οC, 1h;
e) 6-acetyloxy-2-(4-acetyloxyphenyl)benzo[b]thiophene, AlCl3, 1,2-DCE, rt, 1h; f) MeOH/ΝΗ3,
rt, 1h. (B), a) 4-bromobenzylchloride, AlCl3, 1,2 DCE, rt, 1h; b) NaH, HOCH2CH2OH, DMF 90
οC, 1h; c) MsCl Et3N, CH2Cl2, rt, 1h; d) cyclohexylamine or 1-adamantanamine, ΚΙ, toluene,
reflux, 12h; e) AlCl3, EtSH, CH2Cl2, rt, 2h.

9
2.4 ERα- and ERβ-binding affinities:
The affinity and selectivity of analogue binding to ERα and ERβ were
assessed relative to the binding affinity (set to 100) and selectivity (set to 1) of E2
using competitor assay kits, as already described [28]. The relative binding affinities
of the analogues for ERα (RBAα) were from 1.3-fold (13a) to 12.3-fold (9d) lower
compared to raloxifene, while for ERβ (RBAβ) were from 1.6-fold (13a) to 42-fold
(9b) lower; as well, the RBAα:RBAβ ratio of the analogues ranged between 1.2 (13b)
and 7.3 (9b), while that of raloxifene was 1.7, suggesting very low to average
selectivity of the analogues for ERα. Interestingly, 13b displayed a RBAα:RBAβ of
1.2, indicating that the adamantylamino group can be accommodated nearly equally
well in the LBD of ERα and ERβ (for RBA values and ratios see Supplementary
Table S1, columns 2-4). The introduction of a hydroxyl or an acetamide group at
position 3’ of the BSC phenyl group was expected to increase RBA as already
explained in section 2.1. However, derivatization of raloxifene at position 3’ to
generate 5c and 9c decreased RBAα 9.7- and 11.4-fold, respectively, and RBAβ
10.8 and 25.3-fold, respectively, indicating that these derivatizations are incompatible
with BSC channel constrains of either ER. In fact, the RBAα and RBAβ of the 3’
derivatives were 4.6- to 12.3-fold and 5.8- to 42.1-fold lower, respectively, than those
of raloxifene, implying that, irrespective of the volume of the amino group substituent,
derivatization at position 3’ is more incompatible with the hydrophobicity constrains of
the BSC channel of ERβ than with those of ERα. The role of hydrophobicity in
inducing affinity variations among the analogues 5a-e and 9a-e is discussed in the
below subsection. Given that tertiary amines are less basic than secondary ones, it
was not expected that replacement of a tertiary BSC amino group (raloxifene) by a
more basic secondary one (13a, 13b) (a replacement that would increase the pKa
value from 8.46 to 10.07, as calculated from Marvin Sketch software), would lower
RBAα and RBAβ. It appears that the lower RBA of 13a and 13b compared to
raloxifene may reflect steric hindrance owing to the increased length and/or bulkiness
of the BSC and/or entropy-enthalpy compensation as discussed in the following
subsection. Delocalization of the lone electron pair of amide nitrogen likely accounts
for the lower RBAα and RBAβ of 20a and 20b compared to 13a and 13b. Notably,
stabilization by a salt bridge is known to be considerably stronger than stabilization
by weakly charged partners. Finally, the approx. 2.2-fold lower RBAα of 20b
compared to 20a and of 13b compared to 13a could reflect the increased bulkiness
of the adamantylamino group compared to the cyclohexylamino one.
The preliminary FEP calculations of ΔΔG evidently overestimated the binding
affinity of the cyclohexylamino and adamantylamino analogues (Table 1). We
therefore explored other factors influencing the binding interactions. Calculated logP
values (Marvin Sketch 19.20) were in very good correlation with ΔGbinding (Figure 3A),
depicting the importance of hydrophobic interactions and the entropic term of the free
energy of these interactions for the ER-binding affinity. Using Ligand and Structure-
Based Descriptors analysis (LSBD, Schrödinger Inc), 180 descriptors were
calculated and partial least squares (PLS) analysis resulted to an improved
correlation between experimental and calculated data (Figure 3B). LSBD analysis
takes into account topological and lipophilicity descriptors of each ligand, as well as
interaction energy terms from all different algorithms (Glide, Liaison, Embrace, etc),

10
combining them to a final prediction model. The descriptors influencing most the final
model were associated with lipophilicity and solute-accessible surface area.
Since secondary amines are less lipophilic than tertiary ones (logP13a =5.59
while logPRAL =6.48), it is possible that the logP value of raloxifene is a crucial cutoff
favoring binding vs solubility. Although docking calculations predicted that an OH or a
NHCOCH3 group at the 3’-position of BSC phenyl group could improve affinity, all the
3’-derivatives displayed lower RBA than raloxifene. A rationalization based on the
decrease of lipophilicity of all the 3’-derivatives compared to raloxifene would be most
suitable, given the trends observed in Figure 3A. It is clear, however, that binding to
ER is a multivariate process and that several factors (partial charge of BSC nitrogen,
logP, pKa) affect the binding affinity. In the FEP calculations we obtained a very good
overall correlation (r=0.8) between predicted and experimental ΔΔG (Figure 3B).
However, while FEP calculations showed that 13a, 13b and possibly 5d should have
displayed better binding affinity compared to raloxifene (Table 1, column 6), for most
of the remainder analogues the solvation energy of perturbation was predicted to be
lower compared to the experimental one. ERβ FEP calculations predicted even
better binding optimization (data not shown). Overall, however, the introduction of a
polar group on raloxifene BSC in order to improve the enthalpy of interaction through
formation of H bond(s) with Thr347 (Thr299) appears to be detrimental for affinity,
probably because is counterbalanced by unfavorable entropic contributions
commonly referred to as enthalpy-entropy compensation [29].

11
Figure 3 (1 column Figure): Correlation between the experimental ΔG of binding of
raloxifene and its analogues to ERα and ERβ and, A) calculated logP and, B) ΔG of binding
as calculated using LSBD analysis
2.5 Potency of antagonism of ER-mediated effects:
We investigated whether and how RBA correlated with the potency of
antagonism of E2-induced ER-dependent, (i) gene transcription in MCF7:D5L cells
(a clone of MCF7 cells stably transfected with an ERE-endowed luciferase reporter
gene [30]), (ii) proliferation of wild-type MCF7 cells and, iii) alkaline phosphatase
(AlkP) expression in Ishikawa cells; while (i) and (ii) depend on ERα, (iii) depends on
both ERα and ERβ [31]. The potency of antagonism (IC50) was assessed using cells
growing in medium supplemented with charcoal-treated heat-inactivated (i.e. steroid-
free) fetal bovine serum (chFBS) and 0.1 nM E2 i.e. post-menopausal level of
estrogen. The E2-repleted cells were exposed to vehicle (0.1% DMSO) or to
increasing concentrations of test compounds for 24 h (MCF7:D5L cells) or 72 h
(MCF7 and Ishikawa cells) (Figures 4A,B,C and 5A,C). Only 2 analogues (13a,13b)
displayed IC50 of antagonism of E2-induced transcription lower than 20 nM (Table 2,
column 4); their IC50 of antagonism of E2-induced MCF7 cell proliferation were 1.51
nM (13a) and 5.80 nM (13b) and of E2-induced AlkP expression 10.6 nM (13a) and
8.27 nM (13b) (for IC50 comparisons see Supplementary Table S1, columns 5-7).
The structure-activity relationships (SARs) of the analogues based on the IC50 of ER-
dependent gene transcription are presented in Table 2. Interestingly, the IC50 of gene
transcription and those of cell proliferation correlated with RBAα [Pearson’s R=-0.419
(p=0.021) and -0.791 (p=0.001), respectively], as expected from cells (MCF7:D5L
and MCF7) known to express only ERα, while the IC50 of AlkP expression correlated
with RBAβ (R=-0.547; p=0.043), in line with the involvement of ERβ in the E2-
dependent expression of AlkP [31].
2.6 Efficacy of agonism and antagonism of ER-mediated effects:
The agonist and antagonist efficacies of analogues compared to raloxifene
were assessed using cells growing in chFBS supplemented with vehicle (agonist
mode) or 0.1 nM E2 (antagonist mode); the full antagonist ICI182,780 (ICI) was used
as control. In the antagonist mode (Figures 4A,B,C and 5A,C), with the ER-
antagonist efficacy of ICI (1 μM) and E2 (0.1 nM) set equal to 100% and 0%,
respectively, 13a and 13b displayed lower antagonist efficacy compared to ICI in
MCF7:D5L cells and similar antagonist efficacy to ICI in MCF7 cells, while raloxifene
displayed similar antagonist efficacy to ICI in either cell. In Ishikawa cells, however,
the antagonist efficacy of 13b was higher compared to ICI, while that of 13a was
comparable to ICI and that of raloxifene and all the other analogues was lower
compared to ICI (for antagonist efficacy comparisons see Supplementary Table S2,
columns 2-4). Notably, the rank order of antagonist efficacies of the key analogues in
Ishikawa cells was, 13b>ICI≈13a>raloxifene (p<0,05; ANOVA), indicating that ER
antagonism increased as the bulkiness of the BSC amino group increased.
The higher antagonist efficacy of 13b compared to 13a in Ishikawa cells may
reflect a higher potential of 13b to favor formation of ERα/ERβ heterodimers over

12
homodimers, resulting in improved corepressor recruitment through ERβ. It is
anticipated that formation of ERα/ERβ heterodimers is favored when RBAα and
RBAβ are comparable, which is the case with 13b more than with any other
analogue. It has been reported that in ERα/ERβ-expressing cells the coregulator
RIP140 can undertake corepressor functions upon recruitment by ERβ [6].
Interestingly, using glutamate-challenged HT22 cells to assess the antioxidant
activity of 13b vs that of raloxifene, as previously reported [32], revealed that 13b
was moderately active while raloxifene was inactive in this assay (Supplementary
Figure S1), presumably reflecting adamantane’s electron donor potential due to its
hyperconjugation properties [33]. Whether this potential is somehow involved in
stabilizing the antagonist conformation of ERα and/or ERβ is unknown.
Table 2: SARs of raloxifene analogues (1 column Table)
a
IC50
Agonismb
Cmp
R1
R2
(nM) (% of E2)
5a CH2CH2N(CH3)2
5b CH2CH2N(CH2CH3)2
5c CH2CH2N(CH2)5
OH
OH
OH
OH
OH
~1000
61.5
109
40.9
86.9
734
222
263
124
56.4
43.5
46.4
25.1
7.4
41.6
38.9
34.3
29.2
25.7
1.1
5d
5e
CH2CH2NH-Cy
CH2CH2NH-Ad
9a CH2CH2N(CH3)2 NHCOCH3
9b CH2CH2N(CH2CH3)2 NHCOCH3
9c CH2CH2N(CH2)5 NHCOCH3
9d
9e
CH2CH2NH-Cy NHCOCH3
CH2CH2NH-Ad NHCOCH3 77.7
13a CH2CH2NH-Cy
13b CH2CH2NH-Ad
20a CH2CONH-Cy
20b CH2CONH-Ad
Ral CH2CH2N(CH2)5
H
H
H
H
H
4.59
19.4
35.2
60.4
1.11
-9.8
41.5
15.1
10.0
a Potency of antagonism of ER-dependent gene transcription in MCF7:D5L cells; b Efficacy of
agonism of AlkP expression in Ishikawa cells; Cmp: compound; E2: estradiol; Ad: 1-
adamantyl; Cy: cyclohexyl.
In the agonist mode (Figures 4D and 5B,D), with the agonist efficacy of E2
(0.1 nM) and ICI (1 μΜ) set equal to 100% and 0%, respectively, 13a displayed
higher agonist efficacy than ICI in MCF7:D5L cells and similar agonist efficacy to ICI
in MCF7 cells, while 13b and raloxifene displayed similar agonist efficacy to ICI in
both cells. In Ishikawa cells, however, the agonist efficacy of 13b was lower
compared to ICI, while that of 13a was comparable to ICI and that of raloxifene and
all the other analogues was higher compared to ICI (Supplementary Table S2,
columns 5-7). Notably again, the rank order of agonist efficacies of the key

13
analogues at 1 μΜ in Ishikawa cells was, raloxifene>ICI≈13a>13b, implying that ER
agonism decreased as the bulkiness of the BSC amino group increased. In line with
this notion, the rank order of relative agonist efficacies in the 5a:b:c:d:e series was
1.0:0.8:0.8:0.4:0.1 and in the 9a:b:c:d:e series 1.0:0.9:0.8:0.7:0.6. Again, these data
argue in favor of a suppressive effect of bulkiness on ER agonism in Ishikawa cells.
The SARs of the analogues as ER agonists are summarized in Table 2. The
inference from the SARs data of Table 2 is that finding a means (e.g. a core structure
modification) for increasing the antagonist potency (IC50) of 13b (lowest agonist
efficacy hit) might generate SERMs of therapeutic potential as well as endometrial
safety.
A
B
140
120
100
80
140
120
100
80
Ral
13a
13b
20a
20b
ICI
60
60
5a
5b
5c
5d
5e
ICI
40
40
20
20
0
0
0.1
1
10
100
1000
0.1
1
10
100
1000
Concentration (nM)
Concentration (nM)
C
D
140
120
100
80
100
80
60
40
20
0
60
9a
9b
9c
9d
9e
ICI
#
*
40
#
*
20
*
*
*
*
*
*
*
*
*
0
a
b
c
d
e
a
b
c
d
e
a
b
a
b
0.1
1
10
100
1000
5
9
13
20
Concentration (nM)
Figure 4 (2 column Figure): Estrogen receptor agonism and antagonism of raloxifene and
its analogues in ERE-dependent gene transcription. MCF7:D5L cells growing in culture
medium supplemented with 5% chFBS and either 0.1 nM estradiol (E2) (antagonist mode, A-
C) or vehicle (agonist mode, D), were treated for 16 h with increasing concentrations of
raloxifene (Ral) or Ral analogue or 1 μM ICI182,780 (ICI) (A-C); or with 1 μM Ral or Ral
analogue or ICI or 0.1 nM E2 (D). Luciferase expression was expressed relative to that in the
presence of 0.1 nM E2, set equal to 100. Data are mean±SEM of at least three independent
experiments carried out in triplicate. *,# p<0.05 vs incubation with ICI or raloxifene,
respectively (ANOVA); shFBS, charcoal-treated heat-inactivated fetal bovine serum

14
A
B
80
60
40
20
0
120
100
80
Ral
5b
5d
13a
13b
20a
20b
ICI
#
*
#
*
#
*
#
*
60
*
*
40
-20
20
a
b
c
5
d
e
a
b
c
9
d
e
a
13
b
a
20
b
0,1
1
10
100
1000
Concentration (nM)
C
D
80
60
40
20
0
120
100
80
60
40
20
0
#
*
#
*
#
*
#
*
#
*
#
*
#
*
#
*
#
*
#
*
Ral
5b
5d
13a
13b
20a
20b
ICI
#
*
*
*
#
#
*
#
b
-20
a
b
c
d
e
a
b
c
d
e
a
a
b
0.1
1
10
100
1000
Concentration (nM)
5
9
13
20
Figure 5 (2 column Figure): Estrogen receptor agonism and antagonism of raloxifene and
its analogues in cell proliferation and alkaline phosphatase expression. MCF7 cells (A, B) and
Ishikawa cells (C, D) growing in culture medium supplemented with 5% chFBS plus either 0.1
nM estradiol (E2) (antagonist mode, A, C) or vehicle (agonist mode, B, D) were treated for 72
h with increasing concentrations of raloxifene (Ral) or Ral analogue or with 1 μM ICI182,780
(ICI) (A, C); or with 0.1 nM E2, or 1 μM of Ral or Ral analogue or ICI (B, D). Relative cell
numbers were assessed using MTT (A, B), AlkP expression was assessed by measuring
pNPP hydrolysis at 405 nm (C, D), and both were expressed relative to the respective values
in the presence of 0.1 nM E2 set equal to 100. Data are mean±SEM of at least three
independent experiments carried out in triplicate. *,# p<0.05 vs incubation with ICI or
raloxifene, respectively (ANOVA); pNPP, para-Nitrophenylphosphate; shFBS, charcoal-
treated heat-inactivated fetal bovine serum
The ability of 13a and 13b to antagonize MCF7 cell proliferation as effectively
as raloxifene and AlkP expression in Ishikawa cells more effectively than raloxifene
prompted a comparison of their effects on the proliferation of these cells in the
presence of FBS and chFBS. In the presence of FBS and 0.1 nM E2, raloxifene, 13a,
13b and ICI (all at 1 μM) suppressed MCF7 cell proliferation fully and Ishikawa cell
proliferation marginally (raloxifene) or partially (13a, 13b and ICI) (Supplementary
Figures S2A,B). Interestingly, in Ishikawa cells, the partial suppressive effects of 13a
and 13b were similar to the effect of ICI. Likewise, in the presence of chFBS and 0.1
nM E2, raloxifene, 13a, 13b and ICI (all at 1 μM) suppressed MCF7 and Ishikawa cell
proliferation fully and partially, respectively (Supplementary Figures S2C,D). Again, in
Ishikawa cells, the suppressive effects of 13a and 13b were similar to the effect of

15
ICI. Since the ERα of MCF7 cells was degraded in the presence of ICI but not in the
presence of raloxifene, 13a or 13b (inset to Supplementary Figure S2C and data not
shown), the mechanism of action of raloxifene and its analogues is likely different
from that of ICI. These findings classify 13a and 13b as SERMs rather than SERDs.
Given the many adverse effects of ICI, including detrimental effect on bone mineral
density and femoral strength [3], it is tempting to speculate that 13a and 13b might
be devoid of such effects. Recently, antiestrogens with an adamantyl ligand core
structure and an adamantylamido group in the BSC end facing H11 and H12 were
shown to display considerable ERα-antagonist and ERα-degrading activities, albeit
lower compared to those of ICI [20].
2.7 Effects on the immature mouse uterus:
Analogue 13b was singled out for in vivo assessment. Immature female CD1
mice weighing 13.1±0.5 g received daily for three consecutive days subcutaneous
injections of vehicle (20 μl of DMSO), E2 (55 μg/kg), raloxifene (9.5 mg/kg), 13b (10
mg/kg) or 13b followed 1 h later by E2 (55 μg/kg) before subjecting them to
euthanasia on the fourth day. Histological examination of hematoxylin/eosin-stained
uterine sections revealed that in the vehicle-treated group of animals, the surface
endometrium had a mean height of 51.76 μm. The architecture and distribution of
glands in the stroma appeared to be normal, without features of hyperplasia. Rare
mitotic figures were recognized in the surface and glandular endometrial epithelium
(1 Mitotic Figure/10 High Power Fields) (Figure 6A). The endometrial surface
epithelium of E2-treated mice was prominently hyperplastic compared to that of
vehicle-treated animals, with a mean height of 69.55 μm. Hyperplasia of the
glandular epithelium and endometrial glands (proliferation of glands, with an increase
in the gland to stroma ratio compared to normal endometrium, and back-to-back
glands) was also observed. Furthermore, there was an increase in the number of
apoptotic bodies in the glandular epithelium and in the mitotic activity of the stroma
(7-8 MF/10 HPF) compared to vehicle-treated mice (Figure 6B). In the group of
raloxifene-treated animals, prominent hyperplasia of the surface endometrium was
observed, with a mean height of 61.24 μm. Mitotic figures were numerous in the
surface epithelium (7 MF/10 HPF). The endometrial glands were hyperplastic with
fewer apoptotic bodies compared to the E2 group. The endometrial stroma appeared
to be edematous (Figure 6C). The mean height of surface epithelium of 13b-treated
animals was 55.08 μm, showing extremely mild hyperplasia compared to vehicle-
treated mice. Hyperplasia of endometrial glands was present. Mitotic activity was
increased in the surface and in the glandular epithelium compared to vehicle-treated
animals, though not as excessively as in the raloxifene-treated ones (4 MF/10 HPF)
(Figure 6D). In the group of animals treated with 13b followed by E2, the surface
endometrium appeared highly hyperplastic with a mean height of 71.63 μm.
Hyperplasia and increased mitotic activity (5 MF/10 HPF) were observed in the
endometrial glands compared to vehicle-treated mice. The endometrial stroma
appeared mildly edematous in some areas, with a slight increase in mitotic figures (2
MF/10 HPF) compared to vehicle-treated mice though not so pronounced as in E2-
treated animals (Figure 6E). Analysis of microscope images using ImageJ
(http://imagej.net) revealed that, compared to the surface epithelium of vehicle-

16
treated mice, that of the other animals was significantly thicker, with the exception of
the surface epithelium of 13b-treated ones (Figure 6F).
It has been reported that raloxifene exerts a trophic effect in the immature
rodent uterus [10,12]. The above findings confirm these reports and further show that
while raloxifene caused prominent hyperplasia of surface endometrium and
endometrial glands, 13b was only mildly active in either respect; and that 13b was
able to suppress hormonal induction of mitotic activity in the stroma but not in the
surface endometrium and endometrial glands.
10 μm
10 μm
10 μm
F
80
*
*
*
60
40
20
0
10 μm
10 μm
Veh Ral E2 13b 13b+E2
Figure 6 (2 column Figure): A (vehicle): Normal endometrium. Glands with normal
architecture and distribution [inset: mitosis in surface endometrium (arrow)]; B (estradiol):
Hyperplasia of surface endometrial epithelium [inset: Increased stromal mitotic activity
(arrows) and crowded endometrial glands with apoptotic bodies (arrowhead)]; C (raloxifene):
Hyperplasia and increased mitotic activity in surface endometrial epithelium (arrows) [inset:
edematous stroma and glandular hyperplasia]; D (13b):Mild hyperplasia and mitotic activity in
surface endometrial epithelium (arrows) [inset: Endometrial glandular hyperplasia]; E
(13b+estradiol): Hyperplasia of surface endometrial epithelium and endometrial glands, and
edematous stroma [inset: Increased mitotic activity in glands and stroma (arrows)]; F:
Average luminal epithelial height: values are mean±SEM of ten independent assessments (*
p<0.05 compared to vehicle; t-test. A-E, mag x400
3. Conclusion
Preliminary molecular mechanic calculations suggested that introducing a
hydroxyl or an acetamido group at position 3’ of the phenyl group of the BSC of
raloxifene in order to H-bond Thr347 (ERα) or Thr299 (ERβ) as well as increasing the
bulkiness of its BSC amino group would generate analogues of similar ER-binding
affinity to raloxifene. Since the ER-binding affinity of the 3’-derivatives was found to
be considerably lower than that of raloxifene, the inference is that the 3’-
derivatizations clash with the BSC channel constraints. The two analogues lacking a
3’-derivatization and bearing cyclohexylamino (13a) and adamantylamino BSC (13b)
subsequently synthesized displayed high affinity for ERα and ERβ and antagonist

17
efficacy in Ishikawa cells higher than raloxifene and similar to (13a) or higher than
(13b) ICI. Unlike the latter, however, these two analogues preserved ERα integrity,
implying that they could be classified as SERMs rather than SERDs. Importantly,
while the endometrial surface epithelium of immature female mice injected with 13b
was comparable to that of vehicle-treated mice, the surface epithelium of mice
treated with 13b in combination with estradiol was highly hyperplastic. The above
findings suggest that inventing a means for increasing the ER-binding affinity and
consequently the antagonist potency of the adamantylaminoethoxy analogue may
give rise to SERMs of higher endometrial safety for the treatment of menopausal
syndrome.
4. Experimental Section
The Experimental Section is presented in the Supplementary Material
Declaration of Competing Interest
None of the authors has a conflict of interest to declare
Acknowledgments
We thank Nikos Youroukos, Animal Facility Unit, National Hellenic Research
Foundation, for his assistance with animal experiments. This study was carried out
with the financial support of Thales 2011 program “SERMENCO MIS 375617”, which
was co-funded by the European Social Fund and National Resources.
Supplementary material:
Supplementary Material to this article can be found online at ………..
Supplementary Material includes Tables S1-2; Figures S1-2; Theoretical Molecular
1
Methods; Chemical Synthesis Methods; Biological Evaluation Methods; H and 13C
spectra of synthesized compounds; References.
References
[1] J. Marjoribanks, C. Farquhar, H. Roberts, A. Lethaby, J. Lee, Long-term hormone
therapy for perimenopausal and postmenopausal women, Cochrane Database of
Systematic Reviews. (2017). https://doi.org/10.1002/14651858.CD004143.pub5
[2] J.V. Pinkerton, S. Thomas, Use of SERMs for treatment in postmenopausal women, The
Journal of Steroid Biochemistry and Molecular Biology. 142 (2014) 142–154.
https://doi.org/10.1016/j.jsbmb.2013.12.011

18
[3] S.E. Wardell, E.R. Nelson, D.P. McDonnell, From empirical to mechanism-based
discovery of clinically useful Selective Estrogen Receptor Modulators (SERMs), Steroids.
90 (2014) 30–38. https://doi.org/10.1016/j.steroids.2014.07.013
[4] G.G.J.M. Kuiper, J.G. Lemmen, B. Carlsson, J.C. Corton, S.H. Safe, P.T. van der Saag, B.
van der Burg, J.-Å. Gustafsson, Interaction of Estrogenic Chemicals and Phytoestrogens
with
Estrogen
Receptor
β,
Endocrinology.
139
(1998)
4252–4263.
https://doi.org/10.1210/endo.139.10.6216
[5] G. Lambrinidis, M. Halabalaki, E.S. Katsanou, A.-L. Skaltsounis, M.N. Alexis, E. Mikros,
The estrogen receptor and polyphenols: molecular simulation studies of their
interactions, a review, Environmental Chemistry Letters. 4 (2006) 159–174.
https://doi.org/10.1007/s10311-006-0065-y
[6] Z. Madak-Erdogan, T. Charn, Y. Jiang, E.T. Liu, J.A. Katzenellenbogen, B.S.
Katzenellenbogen, Integrative genomics of gene and metabolic regulation by estrogen
receptors α and β, and their coregulators, Mol Syst Biol. 9 (2013) 676.
https://doi.org/10.1038/msb.2013.28
[7] Y. Shang, M. Brown, Molecular Determinants for the Tissue Specificity of SERMs,
Science. 295 (2002) 2465–2468. https://doi.org/10.1126/science.1068537
[8] C.D. Runowicz, J.P. Costantino, D.L. Wickerham, R.S. Cecchini, W.M. Cronin, L.G. Ford,
V.G. Vogel, N. Wolmark, Gynecologic conditions in participants in the NSABP breast
cancer prevention study of tamoxifen and raloxifene (STAR), American Journal of
Obstetrics
and
Gynecology.
205
(2011)
535.e1-5.
https://doi.org/10.1016/j.ajog.2011.06.067
[9] S. Martino, D. Disch, S.A. Dowsett, C.A. Keech, J.L. Mershon, Safety assessment of
raloxifene over eight years in a clinical trial setting, Current Medical Research and
Opinion. 21 (2005) 1441–1452. https://doi.org/10.1185/030079905X61839
[10] J. Ashby, J. Odum, J.R. Foster, Activity of Raloxifene in Immature and Ovariectomized
Rat Uterotrophic Assays, Regulatory Toxicology and Pharmacology. 25 (1997) 226–231.
https://doi.org/10.1006/rtph.1997.1108
[11] S. Mirkin, D.F. Archer, H.S. Taylor, J.H. Pickar, B.S. Komm, Differential effects of
menopausal therapies on the endometrium:, Menopause. 21 (2014) 899–908.
https://doi.org/10.1097/GME.0000000000000186
[12] B.S. Komm, Y.P. Kharode, P.V.N. Bodine, H.A. Harris, C.P. Miller, C.R. Lyttle,
Bazedoxifene Acetate: A Selective Estrogen Receptor Modulator with Improved
Selectivity, Endocrinology. 146 (2005) 3999–4008. https://doi.org/10.1210/en.2005-
0030
[13] D.W. Stovall, W.H. Utian, M.L.S. Gass, Y. Qu, D. Muram, M. Wong, L. Plouffe, The
effects of combined raloxifene and oral estrogen on vasomotor symptoms and
endometrial
safety:,
Menopause.
14
(2007)
510–517.
https://doi.org/10.1097/GME.0b013e318031a83d
[14] A.M. Brzozowski, A.C. Pike, Z. Dauter, R.E. Hubbard, T. Bonn, O. Engström, L. Öhman,
G.L. Greene, J.-A. Gustafsson, M. Carlquist, Molecular basis of agonism and antagonism
in the oestrogen receptor, Nature. 389 (1997) 753-758. https://doi.org/10.1038/39645
[15] A.C. Pike, A.M. Brzozowski, R.E. Hubbard, T. Bonn, A.-G. Thorsell, O. Engström, J.
Ljunggren, J.-A. Gustafsson, M. Carlquist, Structure of the ligand-binding domain of
oestrogen receptor beta in the presence of a partial agonist and a full antagonist, The
EMBO Journal. 18 (1999) 4608–4618. https://dx.doi.org/10.1093/emboj/18.17.4608
[16] A.C. Pike, A.M. Brzozowski, J. Walton, R.E. Hubbard, A.-G. Thorsell, Y.-L. Li, J.-A.
Gustafsson, M. Carlquist, Structural insights into the mode of action of a pure
antiestrogen, Structure.
2126(01)00568-8
9
(2001) 145–153. https://doi.org/10.1016/S0969-

19
[17] A.K. Shiau, D. Barstad, P.M. Loria, L. Cheng, P.J. Kushner, D.A. Agard, G.L. Greene, The
Structural Basis of Estrogen Receptor/Coactivator Recognition and the Antagonism of
This
Interaction
by
Tamoxifen,
Cell.
95
(1998)
927–937.
https://doi.org/10.1016/S0092-8674(00)81717-1
[18] Y.-L. Wu, X. Yang, Z. Ren, D.P. McDonnell, J.D. Norris, T.M. Willson, G.L. Greene,
Structural Basis for an Unexpected Mode of SERM-Mediated ER Antagonism,
Molecular Cell. 18 (2005) 413–424. https://doi.org/10.1016/j.molcel.2005.04.014
[19] J. Min, V.S. Guillen, A. Sharma, Y. Zhao, Y. Ziegler, P. Gong, C.G. Mayne, S. Srinivasan,
S.H. Kim, K.E. Carlson, K.W. Nettles, B.S. Katzenellenbogen, J.A. Katzenellenbogen,
Adamantyl Antiestrogens with Novel Side Chains Reveal a Spectrum of Activities in
Suppressing Estrogen Receptor Mediated Activities in Breast Cancer Cells, J. Med.
Chem. 60 (2017) 6321–6336. https://doi.org/10.1021/acs.jmedchem.7b00585
[20] L. Wang, V.S. Guillen, N. Sharma, K. Flessa, J. Min, K.E. Carlson, W. Toy, S. Braqi, B.S.
Katzenellenbogen, J.A. Katzenellenbogen, S. Chandarlapaty, A. Sharma, New Class of
Selective Estrogen Receptor Degraders (SERDs): Expanding the Toolbox of PROTAC
Degrons,
ACS
Med.
Chem.
Lett.
9
(2018)
803–808.
https://doi.org/10.1021/acsmedchemlett.8b00106
[21] C.D. Jones, M.G. Jevnikar, A.J. Pike, M.K. Peters, L.J. Black, A.R. Thompson, J.F. Falcone,
J.A. Clemens, Antiestrogens. 2. Structure-activity studies in a series of 3-aroyl-2-
arylbenzo [b] thiophene derivatives leading to [6-hydroxy-2-(4-hydroxyphenyl) benzo
[b] thien-3-yl]-[4-[2-(1-piperidinyl) ethoxy] phenyl] methanone hydrochloride (LY
156758), a remarkably effective estrogen antagonist with only minimal intrinsic
estrogenicity, Journal of Medicinal Chemistry. 27 (1984) 1057–1066.
https://doi.org/10.1021/jm00374a021
[22] D.A. Bradley, A.G. Godfrey, C.R. Schmid, Synergistic methodologies for the synthesis of
3-aroyl-2-arylbenzo[b]thiophene-based selective estrogen receptor modulators. Two
concise syntheses of raloxifene, Tetrahedron Letters. 40 (1999) 5155–5159.
https://doi.org/10.1016/S0040-4039(99)00955-7
[23] J.A. Dodje, C.D. Jones, T.M.L. Bourgeois, Preparation of 3-[4-(2-
heterocyclylethoxy)benzoyl-2-phenylbenzothiophenes for use in alleviating the
symptoms of post-menopausal syndrome., Eur. Pat. Appl. (1996), 67pp. CODEN.
EPXXDW EP 738725 A2 19961023 CAN 126: 7985 AN 1996: 740256 77, n.d.
[24] P. Shinde, S. Shinde, A. Renge, G. Patil, A. Rode, R. Pawar, An Efficient Synthesis of
Raloxifene in Ionic Liquid:
A
Green Approach, LOC.
6
(2009) 8–10.
https://doi.org/10.2174/157017809787003133
[25] A. Pandey, D.L. Volkots, J.M. Seroogy, J.W. Rose, J.-C. Yu, J.L. Lambing, A.
Hutchaleelaha, S.J. Hollenbach, K. Abe, N.A. Giese, R.M. Scarborough, Identification of
Orally Active, Potent, and Selective 4-Piperazinylquinazolines as Antagonists of the
Platelet-Derived Growth Factor Receptor Tyrosine Kinase Family, J. Med. Chem. 45
(2002) 3772–3793. https://doi.org/10.1021/jm020143r
[26] N. Shiyeyoshi, H. Kenji, S. Hidetaka, H. Takashi, O. Hiroyuki, T. Takeshi, S. Shinobu,
Process for producing 3,4-dihydroquinazoline-4-one derivatives., PCT Int. Appl. (2003),
101pp PIXXD2 WO 2003064399, CAN 139:164805., n.d.
[27] F. Haviv, J.D. Ratajczyk, R.W. DeNet, F.A. Kerdesky, R.L. Walters, S.P. Schmidt, J.H.
Holms, P.R. Young, G.W. Carter, 3-[1-(2-Benzoxazolyl)hydrazino]propanenitrile
derivatives: inhibitors of immune complex induced inflammation, J. Med. Chem. 31
(1988) 1719–1728. https://doi.org/10.1021/jm00117a010
[28] S. Djiogue, M. Halabalaki, X. Alexi, D. Njamen, Z.T. Fomum, M.N. Alexis, A.-L.
Skaltsounis, Isoflavonoids from Erythrina poeppigianaꢀ: Evaluation of Their Binding
Affinity for the Estrogen Receptor, J. Nat. Prod. 72 (2009) 1603–1607.
https://doi.org/10.1021/np900271m

20
[29] J.D. Chodera, D.L. Mobley, Entropy-Enthalpy Compensation: Role and Ramifications in
Biomolecular Ligand Recognition and Design, Annu. Rev. Biophys. 42 (2013) 121–142.
https://doi.org/10.1146/annurev-biophys-083012-130318
[30] N. Fokialakis, G. Lambrinidis, D.J. Mitsiou, N. Aligiannis, S. Mitakou, A.-L. Skaltsounis, H.
Pratsinis, E. Mikros, M.N. Alexis, A new class of phytoestrogens: evaluation of the
estrogenic activity of deoxybenzoins, Chemistry & Biology. 11 (2004) 397–406.
https://doi.org/10.1016/j.bmc.2012.03.012
[31] N. Fokialakis, X. Alexi, N. Aligiannis, D. Siriani, A.K. Meligova, H. Pratsinis, S. Mitakou,
M.N. Alexis, Ester and carbamate ester derivatives of Biochanin A: Synthesis and in
vitro evaluation of estrogenic and antiproliferative activities, Bioorganic & Medicinal
Chemistry. 20 (2012) 2962–2970. https://doi.org/10.1016/j.bmc.2012.03.012
[32] M. Koufaki, E. Theodorou, X. Alexi, F. Nikoloudaki, M.N. Alexis, Synthesis of tropolone
derivatives and evaluation of their in vitro neuroprotective activity, European Journal
of
Medicinal
Chemistry.
45
(2010)
1107–1112.
https://doi.org/10.1016/j.ejmech.2009.12.006
[33] T. Laube, First Crystal Structure Analysis of an Aliphatic Carbocation—Stabilization of
the 3,5,7-Trimethyl-1-adamantyl Cation by C-C Hyperconjugation, Angew. Chem. Int.
Ed. Engl. 25 (1986) 349–350. https://doi.org/10.1002/anie.198603491

21