Scholarly article on 1,1-bis(4-hydroxyphenyl)-2-methylprop-1-ene 93434-07-2 from European Journal of Medicinal Chemistry

European Journal of Medicinal Chemistry 224 (2021) 1 13733

Contents lists available at ScienceDirect

European Journal of Medicinal Chemistry

journal homepage: h ttp://www.elsevier.com/locate/ejmech

Balanced dual acting compounds targeting aromatase and estrogen

receptor

a as an emerging therapeutic opportunity to counteract

estrogen responsive breast cancer

,

Jessica Caciolla ^{a 1}, Silvia Martini ^{b 1}, Angelo Spinello ^c, Matic Pavlin ^{c d}, Eleonora Turrini ^e,

Federica Simonelli ^c, Federica Belluti ^a, Angela Rampa ^a, Alessandra Bisi ^a,

Carmela Fimognari ^e, Nadia Zaffaroni ^b, Silvia Gobbi ^a^*, Alessandra Magistrato ^c

,

, **

^aDepartment of Pharmacy and Biotechnology, Alma Mater Studiorum-University of Bologna, Via Belmeloro 6, 40126, Bologna, Italy

^bFondazione IRCSS Istituto Nazionale Dei Tumori, Via Amadeo 42, 20113, Milano, Italy

^cNational Research Council of Italy Institute of Materials (CNR-IOM) C/o SISSA, Via Bonomea 265, 34136, Trieste, Italy

d

ꢀ ꢀ

Laboratory of Microsensor Structures and Electronics, Faculty of Electrical Engineering, University of Ljubljana, Trzaska Cesta 25, SI-1000 Ljubljana,

Slovenia

^eDepartment for Life Quality Studies, Alma Mater Studiorum-University of Bologna, Corso D'Augusto 237, 47921, Rimini, Italy

a r t i c l e i n f o

a b s t r a c t

Article history:

Breast Cancer (BC) is a leading cause of death in women, currently affecting 13% of female population

worldwide. First-line clinical treatments against Estrogen Receptor positive (ERþ) BC rely on suppressing

estrogen production, by inhibiting the aromatase (AR) enzyme, or on blocking estrogen-dependent pro-

Received 1 July 2021

Received in revised form

27 July 2021

Accepted 28 July 2021

Available online 31 July 2021

oncogenic signaling, by targeting Estrogen Receptor (ER)

SERDs). The development of dual acting molecules targeting AR and ER

a

with selective Modulators/Degraders (SERMs/

represents a tantalizing alter-

a

native strategy to ﬁght ER þ BC, reducing the incidence of adverse effects and resistance onset that limit

the effectiveness of these gold-standard therapies.

Keywords:

Here, in silico design, synthesis, biological evaluation and an atomic-level characterization of the

binding and inhibition mechanism of twelve structurally related drug-candidates enable the discovery of

breast cancer

Multitarget

Aromatase inhibitors

SERM

SERD

Molecular dynamics

QM/MM

multiple compounds active on both AR and ER

played a balanced low-nanomolar IC₅₀towards the two targets, SERM activity and moderate selectivity

towards a BC cell line. Moreover, most of the studied compounds reduced ER levels, suggesting a po-

a in the sub-mM range. The best drug-candidate 3a dis-

a

tential SERD activity. This study dissects the key structural traits needed to obtain optimal dual acting

drug-candidates, showing that multitarget compounds may be a viable therapeutic option to counteract

ER þ BC.

1. Introduction

estimated that 13% of women will develop an invasive BC in their

lifetime [2]. A high percentage of BC (around 80%) is estrogen

Breast cancer (BC) is the most common malignant tumor and

the second leading cause of cancer-related death in women [1].

According to Brest Cancer Fact and Figures 2019e2020, it is

dependent and needs hormones to grow and proliferate [3].

Key enzyme in the biosynthesis of estrogens is aromatase (AR), a

member of the cytochrome P450 superfamily [4], expressed in

Abbreviations: AI, aromatase inhibitor; AR, aromatase; BC, breast cancer; E2, estradiol; END, endoxifen; ER, estrogen receptor; ERþ, estrogen receptor positive; ICI

(182,780), fulvestrant; LBC, ligand binding cavity; LBD, ligand binding domain; MD, molecular dynamics; PGR, progesterone receptor; QM/MM, quantum mechanics/mo-

lecular mechanics; SERD, selective estrogen receptor degraders; SERM, selective estrogen receptor modulators; TAM, tamoxifen;

DG_b, binding free energy.

* Corresponding author.

** Corresponding author.

E-mail addresses: silvia.gobbi@unibo.it (S. Gobbi), alessandra.magistrato@sissa.it (A. Magistrato).

These authors equally contributed to this manuscript.

1

https://doi.org/10.1016/j.ejmech.2021.113733

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

ovaries and in many other tissues, responsible for peripheral es-

trogens production in postmenopausal women [5]. Upon binding to

Alternatively, SERDs have been developed, which bind to ERa

triggering its degradation. Unlike SERMs, the interaction of SERDs

with H12 of ER leads to alterations in its conformation and

disruption of its structure, exposing different hydrophobic amino

acid residues, thus causing the ubiquitination and subsequent

proteasomal degradation of the receptor. The prototype SERD, ful-

vestrant (ICI 182,780), although being a pure antagonist, exhibits

poor pharmacokinetic properties, requiring intravenous adminis-

tration and resulting in limited efﬁcacy. Intense research efforts

have been recently devoted to the discovery of SERDs that could

counteract the resistance to prolonged therapeutic regiments due

the Estrogen Receptor (ER)

gens trigger a conformational change of its helix (H)12, inducing

the formation of an ER homodimer, which then binds DNA and

a ligand binding domain (LBD), estro-

a

control gene expression [6,7].

Two main strategies are currently employed in the clinics to

treat hormone-dependent BC patients: (i) preventing estrogens

biosynthesis by inhibiting AR or (ii) modulating ERa activity with

Selective Estrogen Receptor Modulators or Degraders (SERMs/

SERDs). The third-generation Aromatase Inhibitors (AIs) exemes-

tane, letrozole and anastrozole are highly speciﬁc and characterized

by fewer side effects when compared to previous reported agents.

AIs are also classiﬁed as steroidal (exemestane) and non-steroidal

derivatives (anastrozole and letrozole), exhibiting a different inhi-

bition mechanism: exemestane binds irreversibly to AR, acting as a

suicide inhibitor, while anastrozole and letrozole are reversible and

competitive inhibitors [8,9], supposedly able to form a coordination

bond between their azole moiety and the heme iron atom of AR

[10]. Nevertheless, no experimental structural data are currently

available to ensure their binding mode, which has been recently

questioned [1 1,12]. The inhibition of AR leads to estrogens depletion

in the whole body, thus inducing signiﬁcant adverse side effects,

such as musculoskeletal disorders, arthralgia, increase of osteopo-

rosis and fractures, and cardiovascular events [13e15].

to the onset of somatic mutations in the LBD of ER

a [20e22].

Although fulvestrant and AZD-9496, currently in phase I clinical

trials [23], retain efﬁcacy against the most diffused and aggressive

ER

needed.

In principle, AI and SERM/SERD may act synergistically to inhibit

a isoforms, alternative therapeutic strategies are urgently

the growth and proliferation of BC cells, leading to an increased

efﬁcacy when used in a combined treatment. Nevertheless, con-

ﬂicting evidence emerged in distinct clinical studies. While Brodie

and coworkers observed that the combination of letrozole and

fulvestrant was more effective than the two drugs administered

singularly [24], the ATAC (anastrozole, tamoxifen, alone or in

combination) trial showed, instead, that anastrozole alone was

more effective in terms of disease-free survival and time to recur-

rence when compared to TAM alone or to anastrozole and TAM in

combination [25,26]. As a result, an efﬁcacy boost is not warrantied

for any SERM and AI combination.

SERMs bind to ERa inducing opposite activity in different tis-

sues. They act as agonists, mimicking estrogens activity in bones

and heart tissues, while acting as antagonists in breast tissue [16]

and, depending on the proﬁle of each drug, they are widely used for

the treatment/prevention of BC (and other cancers) and osteopo-

rosis. The most representative and currently used SERM for BC

treatment is tamoxifen (Z-1-{4-[2-(dimethylamino)ethoxy]

phenyl}-1,2-diphenyl-1-butene, TAM, Fig. 1). Despite its clinical

usefulness, TAM exerts signiﬁcant side effects, the most important

being the increased risk of endometrial cancer, due to its agonist

activity in uterine cells, and thromboembolic events [1 7]. Addi-

tionally, the prolonged use of TAM leads to the possible develop-

ment of intrinsic or acquired drug resistance [18]. Besides, its wide

biological proﬁle also includes antifungal, antioxidant, and antiviral

activities, antiangiogenesis and apoptosis induction properties.

Pharmacological studies have shown that TAM is a prodrug, and its

anticancer activity is due to its active metabolites, the most abun-

dant and active being endoxifen (END, Fig. 1) [19].

Among the novel strategies recently perused to counteract

ER þ BC is the development of compounds with dual AI/SERM ac-

tivity, able to simultaneously block estrogen production and ERa-

mediated cell proliferation in BC tissues. Additionally, these dual

acting agents may be endowed with fewer side effects, since the

agonistic activity of SERM-like compounds in bone-tissue would

alleviate the typical disturbances caused by AI-induced estrogen

depletion, while the reduction of estrogen biosynthesis may

counteract the agonist effects of SERM-like molecules on uterine

tissue.

In this context, in 2012 Flockhart and coworkers reported that

some TAM metabolites, namely END, N-desmethyltamoxifen and

norendoxifen, act as AIs (K_ivalues of 4.0

respectively), while still binding with good afﬁnity to ER

Strikingly, these molecules, simultaneously blocking the activity of

mM, 15.9

mM and 35 nM,

a

[27,28].

Fig. 1. The most representative Selective Estrogen Receptor Modulator (SERM) tamoxifen (TAM) and its main metabolites.

2

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

two distinct targets involved in ER þ BC, laid the foundations for

developing dual AI/SERM lead compounds. Recently, the structures

of the aforementioned TAM metabolites were subjected to different

chemical modiﬁcations, among which the removal of the amino

side chain of SERMs and its substitution by a hydroxyl group,

2. Results

2.1. Design and synthesis

Building on the successful realization of other dual acting

compounds [29], we conceived 1-6a,b drug-candidates (Fig. 2) on

the basis of their potential ability to coordinate the heme moiety of

AR, their structural similarity to TAM metabolites, and their syn-

thetic feasibility.

In particular, the ethyl moiety, a distinctive feature of triphe-

nylethylene SERMs, was maintained, and the key nitrogen atom,

believed to establish a coordination bond with the heme iron of AR,

was positioned in the existing aromatic ring, by substituting ben-

zene with a pyridine or imidazole. A small spacer (1e2 carbon

atoms) was also introduced between the nitrogen-containing

moiety and the double bond, as in the literature molecule A, to

evaluate the effect of its length on the activity, and the shortening

of the ethyl chain to a methyl one was also explored. For these

compounds docking simulations were performed to assess their

resulting into

a symmetrical bisphenol structure [29]. TAM

bisphenol, another known metabolite of this drug (Fig. 2) had also

been previously shown to possess signiﬁcant afﬁnity for ERs [30],

supporting the hypothesis that the removal of the side chain did

not affect the antiestrogenic activity of this class of compounds.

Moreover, different nitrogen-containing moieties were also inser-

ted on the ethyl group to coordinate the heme iron of the AR

enzyme, mimicking the letrozole/anastrozole metal coordinating

moiety. As an example, the introduction of an imidazole ring led to

a potent and effective compound (A, Fig. 2). Nevertheless, no dual-

acting drug-candidate has so far been object of preclinical studies,

thus calling attention to this burgeoning research ﬁeld.

Inspired by these promising results, and taking advantage of our

long-lasting experience in developing AIs and TAM-related com-

pounds [31e34], here we designed different series of potential

multitarget agents (Fig. 2), to gain new insight on TAM-derived

molecules.

ability to bind to the active site of AR and to the LBC of ERa.

Due to the well-known ﬂaws of the docking simulations to

reproduce the coordination bonds to the heme iron of AR, we

imposed a constraint on the distance between the imidazole/ni-

trogen and the iron [35]. As a result, all ligands snugly ﬁt within the

AR active site with good docking scores (Table S1), but it is unclear

if/unlikely that some of the compounds (1a, 1b, 2a and 2b) would

establish a coordination bond to the iron atom (Fig. S1) owing to the

lack of a spacer between the nitrogen-containing moiety and the

double bond. As well, the twelve ligands were docked on a repre-

After validating the aptitude of the designed ligands to ﬁt into

AR active site and ER

tested in vitro in order to evaluate their ability to inhibit AR and

bind to ER , their growth inhibition potency in BC and healthy

a LBC, the compounds were synthesized and

a

breast cell lines, and their potential SERM and/or SERD effects. All-

atom molecular dynamics (MD) simulations ﬁnally dissected the

key structural traits underlying the striking inhibitory potency of

the newly synthesized drugs-candidates.

sentative structure of the ER

a LBD as obtained from previous

classical MD simulations [6]. All ligands revealed good docking

scores (ranging from ꢀ8 to ꢀ11 kcal/mol, Table S1), similarly to

Fig. 2. Design of the studied compounds.

3

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

Scheme 1. Synthesis of compounds 1e3, 5, 6. Reagents and conditions: i) Zn, TiCl₄, THF, N₂, reﬂux, 3e5 h.

those predicted for END. On the basis of these results, we per-

formed the synthesis of these ligands.

hydroxyl groups, by acid hydrolysis afforded compounds 4a and 4b.

While the ketone intermediates 7a,b and 8b, for the synthesis of

compounds 1a,b and 3b respectively, were commercially available,

the other intermediates (8a, 9-11a,b) were synthesized following

different strategies (Schemes S1eS3).

For the synthesis of compounds 1-3a,b and 5-6a,b (Scheme 1),

4,4⁰-dihydroxybenzophenone and the appropriate ketone (7-11a,b)

were reacted under the McMurry reaction conditions, namely in

the presence of titanium chloride and zinc powder to form a

carbon-carbon double bond.

2.2. In vitro assays of the synthesized compounds

For the synthesis of compounds 4a,b (Scheme 2), 4,4⁰-dihy-

droxybenzophenone and acetone or methylethylketone were

reacted in the McMurry reaction to give 12a,b. The hydroxyl groups

were protected with chloromethylmethylether (MOM-Cl) and the

obtained intermediates 13a,b were ﬁrst brominated with N-bro-

mosuccinimide (NBS) and then reacted with imidazole to give

14a,b. Cleavage of MOM ethereal functions, to obtain the free

The compounds were tested at different concentrations (from

0.0001 to 10 mM) for their ability to inhibit estrogen production by

AR, by monitoring the conversion of a ﬂuorogenic substrate of the

enzyme into a highly ﬂuorescent metabolite (Aromatase Inhibitor

Screening Kit, BioVision) and results were reported in Table 1. For

comparison letrozole was also tested as reference inhibitor. All

Scheme 2. Synthesis of compounds 4a,b. Reagents and conditions: i) Zn, TiCl₄, THF, N₂, reﬂux, 3 h; ii) NaH 60%, THF, MOM-Cl, N₂, r.t., 3 h; iii) NBS, CCl₄, BPO, h

imidazole, CH₃CN, N₂, reﬂux, 5 h; v) methanol, HCl, r.t., 12 h.

n, reﬂux 4 h; iv)

4

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

Table 1

Biological proﬁles of 1-6a, b: compound concentrations inhibiting by 50% AR activity, ER

A comparison with letrozole (LET) and endoxifen (END) as reference inhibitors for AR and ER

a

binding and the proliferation of MCF7 (ERþ) BC cells and MCF10A healthy breast cells.

a

, respectively.

Compound

Code

AR inhibition^aIC₅₀

mM

ER

a

binding^aIC₅₀

mM

MCF-7 cells^aIC₅₀m

M

MCF-10a cells^aIC₅₀

17.0 0.5

mM

1a

1.253

0.129

>10

10.0 2.1

1b

>10

0.076

0.036

5.3 1.1

41.7 0.3

2a

2b

0.952

0.078

1.073

1.010

60.0 0.9

>100

0.596

0.045

0.223

0.071

58.0 8.5

3a

0.023

0.018

0.019

0.001

27.6 0.4

>100

3b

0.861

0.062

1.770

0.368

41.2 1.6

58.7 0.9

4a

0.035

0.001

1.803

0.221

24.3 0.9

34.7 0.5

4b

1.002

0.171

4.950

0.613

37.0 1.0

52.5 0.5

5a

0.900

0.339

0.033

0.003

42.1 1.1

20.0 1.6

(continued on next page)

5

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

Table 1 (continued)

Compound

Code

AR inhibition^aIC₅₀

mM

ER

a

binding^aIC₅₀

mM

MCF-7 cells^aIC₅₀m

>100

M

MCF-10a cells^aIC₅₀

mM

5b

3.240

0.438

0.313

0.00^a

>100

6a

6b

0.052

0.004

0.197

0.031

37.1 1.1

>100

0.063

0.003

1.737

0.318

39.8 4.7

50.4 0.7

Endoxifen

Letrozole

>10

0.004 0.001

0.030 0.023

>10

0.1 0.1

4.1 1.1

>10

a

Data represent the mean SD of at least three independent experiments performed in triplicate.

compounds, except for 1b, proved to be effective AIs and, remark-

ably, several ligands exhibited IC₅₀in the nM range. Overall, the

obtained results suggested that the presence of the TAM-like ethyl

moiety seemed to be preferred to the shorter methyl. As well, the

introduction of either imidazole or pyridine led to potent de-

rivatives, providing appropriate interactions of the designed mol-

ecules with AR.

In detail, the role and contribution of the heterocycle seemed to

be dependent on the length of the spacer between the azole ring

and the double bond. Indeed, apart from the weak AIs 1-2a,b,

bearing no linker (n ¼ 0) and giving dissimilar results, both het-

erocycles led to equipotent compounds when separated from the

vinyl moiety by a single-methylene spacer (n ¼ 1, 3-4a,b). Instead,

imidazole seemed to be preferred to pyridine in the ethyl-linker

derivatives (n ¼ 2, 5-6a,b).

the activity of compounds bearing the ethyl group, in the pyridine

series only the insertion of a methylene spacer on compound 1a

allowed

a

signiﬁcant afﬁnity gain for the receptor (3a,

IC₅₀¼ 19 nM), while the elongation of the linker (5a, IC₅₀¼ 33 nM)

did not further inﬂuence it. Concerning the corresponding imida-

zolyl derivatives, going from 2a to 4a a low activity was maintained,

which was, instead, boosted by the insertion of another methylene

unit (6a). Thus, the linker length had a lower impact on the

measured IC₅₀in the pyridine series than in the corresponding

imidazole group. Remarkably, only among compounds with no

linker (1-2a,b), the presence of the methyl group on the double

bond, rather than the ethyl, led to sub-mM afﬁnity towards ERa,

regardless of the choice of the heterocycle, being 1b the most active

derivative (IC₅₀¼ 76 nM). Compounds 3a and 5a, both carrying a

pyridine ring and the ethyl group on the double bond, and a one- or

two-methylene spacer, respectively, resulted to be the most inter-

esting compounds, exhibiting IC₅₀comparable (5a) or lower (3a)

Taking into account the data for the pyridine series, the intro-

duction of a methylene between the azole ring and the double bond

signiﬁcantly boosted the activity of the compounds (3a,b), while its

extension to two methylene units (5a,b) induced a decrease. As

regards imidazole derivatives, the low micromolar activity of 2a,

carrying an ethyl moiety on the double bond, increased with the

insertion of the methylene spacer (n ¼ 1, 4a, IC₅₀¼ 35 nM), and was

not affected by its lengthening (n ¼ 2, 6a). In contrast, the activity of

the corresponding methyl derivatives was slightly reduced going

from 2b to 4b (n ¼ 1), but was remarkably improved by elongating

the linker to two methylene units (n ¼ 2, 6b, IC₅₀¼ 63 nM). Alto-

gether, among the synthesized compounds, 3a, 4a, 6a and 6b

resulted to be the most potent AIs (IC₅₀s of 23, 35, 52 and 63 nM,

respectively), exhibiting a slightly lower potency with respect to

the reference inhibitor letrozole. Conversely, END had an

than the reference END. Conversely, letrozole had an IC₅₀>10 mM.

Moreover, we evaluated the ability of these compounds to

inhibit the growth of the MCF-7 ER þ BC cell line and e for

comparative purpose e that of a healthy breast cell line (MCF10A)

(Table 1). Several molecules showed a moderate capability to

inhibit cell growth in MCF-7 cells with IC₅₀values>10 mM. Inter-

estingly, almost all compounds required slightly higher concen-

trations to inhibit MCF10A healthy breast cells proliferation,

suggesting a mild non-speciﬁc antiproliferative effect on cancer

cells. Notably 2a, 2b, 3a and 6a had no effect on healthy cells at any

concentration tested. Remarkably, the antiproliferative activity of

previously reported dual acting AR/ER

measured [29].

a drug candidates was not

IC₅₀>10

The relative afﬁnity of the new compounds and of the most

active TAM metabolite END towards ER (Table 1) was then eval-

uated using the PolarScreen ER Competitor Assay Kit Green

(Invitrogen) by measuring their ability to prevent estrogen binding

to ER . Apart from 1a, all tested compounds were able to bind and

displace the estrogen from ER , showing IC₅₀values in the M- or

sub- M range. From these results it clearly appeared that, except

for 1-2a,b, the introduction of a pyridine ring led in general to

higher afﬁnity for ER as compared to imidazole, and that the TAM-

like ethyl was preferred to the methyl moiety. Taking into account

mM.

2.3. Analysis of progesterone receptor (PGR) mRNA expression in

MCF-7 cells

a

A functional assay was then performed to assess the ability of

two of the most promising compounds based on the above re-

ported IC₅₀s, 3a and 6a, to antagonize the mRNA estrogen-

stimulated expression of Progesterone Receptor (PGR) in MCF-

7 cells [29]. Estradiol (E2) signiﬁcantly increased the expression of

PGR compared to untreated cells and E2-stimulated mRNA

expression of PGR was considered 100%. The antiestrogenic activity

a

m

a

6

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

activity on ER

a

degradation. Similar to fulvestrant, compound 5b

almost completely degraded ER

(313 nM vs 1.0 M). Interestingly, compound 3a, which showed the

highest binding afﬁnity towards ER in the binding assay

(IC₅₀¼ 19 nM), also demonstrated an outstanding degradation

activity against ER in MCF-7 cells, as shown in Fig. 4.

a, but at a lower concentration

m

a

2.5. Molecular mechanism of inhibition

Aiming to exhaustively dissect the structural traits underlying

the experimentally measured efﬁcacy of the synthesized com-

pounds, we performed extensive force ﬁeld-based MD and mixed

quantum-classical (QM/MM) MD simulations. These allowed to

predict the inhibitors binding mode to AR and to the LBD of ERa and

to identify the key interactions they established with both target

biomolecules.

Fig. 3. PGR mRNA expression level after treatment of MCF-7 cells with E2 (10 nM)

alone (considered as 100% PGR mRNA expression) or E2 in association with com-

pounds 3a or 6a (10e20 mM) or endoxifen (1 mM), used as positive control for its ability

Consistently with previous studies [35], the docking scores of

the inhibitors on AR do not correlate with the measured IC₅₀, owing

to the inability of the docking simulations to describe coordination

bonds. Thus, in order to understand which drug may owe its po-

tency to a coordination bond to the iron atom, we performed QM/

MM MD simulations to accurately account for the metal-ligand

interaction, while explicitly considering the biomolecular envi-

ronment [38]. Due to the high computational cost of QM/MM MD

simulations, we selected a subset of compounds (3a, 3b, 4a, 4b, 5a,

6a, 6b) to ﬁnd a rationale between their structural motifs and

measured IC₅₀s.

Each AR/drug adduct was initially relaxed by 100 ns of explicitly

solvated classical MD simulations and its equilibrated structure

underwent 10 ps of QM/MM MD simulations. In line with our

previous study [35], a structural analysis of the resulting adducts

clearly disclosed that the coordination geometry (i.e. bond lengths

to antagonize stimulated PGR expression. ***p < 0.001 versus E2-stimulated mRNA

expression of PGR.

of compounds 3a and 6a was then monitored as reduction of the

aforementioned expression (Fig. 3). END, used as positive control,

antagonized the E2-stimulated mRNA expression of PGR to 5.6%

(Fig. 3), consistent with previously published results [36]. As well,

3a and 6a were able to dose-dependently antagonize the E2-

stimulated expression of PGR, although to a weaker extent

compared to the positive control END. Namely, they antagonized it

up to 29 and 28%, respectively, at the highest tested concentration

(20

mM) (Fig. 3). These values are in line with those observed for

previously reported dual-acting AR/ER

a

compounds [29].

2.4. Assessment of ER

a expression in MCF-7 cells

Besides the capacity to prevent the binding of estrogen to ER

a

,

Table 2

crucial property of SERDs is the ability to degrade the receptor. To

evaluate the efﬁciency of the studied molecules as potential mul-

tipotent agents, as a preliminary assay the ability of the compounds

Structural characteristics for the binding of compounds 3a, 3b, 4a, 4b, 5a, 6a, 6b to

the active site of AR. The table reports the coordination bonds (Angstrom) and angles

(deg) of the azole/pyridine moiety of the ligand to the heme iron atom. For com-

parison, the clinically used inhibitor letrozole (LTZ), having an IC₅₀of 4 nM, is also

reported [1 1].

to modulate the expression of ER

analysis, together with fulvestrant (ICI) as positive control. Based on

the results of ER binding assay, all compounds, except for 1a

(IC₅₀>10 M), were used at their speciﬁc IC₅₀calculated with the

cell-free assay. Fulvestrant was used at a ﬁxed concentration of

a was evaluated by Western Blot

a

Distance (Å)

Angle (deg)

m

LTZ

3a

3b

4a

4b

5a

6a

6b

2.33 0.15

2.27 0.11

e

91.8 2.7

92.2 6.7

e

1

m

M.

2.16 0.07

2.20 0.11

e

84.4 7.8

89.6 5.1

e

As shown in Fig. 4, a reduction of protein levels, although to a

different extent, was obtained with the tested compounds, with the

exception of 2a, 3b, 4b and 6b. Compounds 2b, 4a and 5a, with IC₅₀s

of 223 nM, 1800 nM and 33 nM, respectively, showed excellent

2.19 0.08

2.20 0.11

95.0 6.1

87.2 4.7

Fig. 4. Representative western blotting showing the expression of ER

and fulvestrant (ICI, 1 M).

control and referred to untreated control.

a

in MCF-7 cells exposed to vehicle (DMSO) or to the studied compounds (IC₅₀calculated by ER

a

binding assay)

m

b-actin was used to conﬁrm equal protein loading on the gel. For each protein, band density was quantiﬁed using ImageJ [37] normalized to loading

7

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

and angles) of the inhibitors was not strictly entwined with the

observed potency. Indeed, the most potent compounds 3a, 4a, 6a

and 6b presented a binding geometry similar to 4b, which is,

instead, characterized by lower inhibitory potency. As well, the

clinically used letrozole exhibited even longer coordination bonds

distances [35]. Strikingly, 3b and 5a in QM/MM MD trajectories do

not establish any coordination bond to the heme iron (Table 2),

conﬁrming common intuition that the possibility of binding the

metal moiety boost the inhibitory potency. The AR active site is

small and rigid and adapts with difﬁculty to ligands of shape

different from its endogenous substrate. For this reason, even small

differences among distinct inhibitors can strongly affect the bind-

ing properties. As an example, in the case of compound 5a, the

simultaneous presence of a two-methylene unit spacer, connecting

the vinyl moiety to the pyridine, and of an ethyl substituent made

the ligand very bulky, preventing the formation of a coordination

bond (Fig. 5). Conversely, from the docking poses of 2a and 2b

(Fig. S1), it appeared that the absence of a methylene linker be-

tween the double bond and the heterocycle made the latter highly

affected its ability to rotate and establish the coordination bond to

the metal center.

We also monitored the role of the hydroxyl substituents in

establishing hydrogen (H)-bonds to the active site residues. In spite

of the similar coordination geometry, among the ligands predicted

to coordinate the heme, the hydroxyl groups of 4a, 6a and 6b

established two persistent H-bonds to the backbones of Leu372 and

Asp309, while 3a H-bonded only to the backbone of Leu372 (Fig. S2

and Table S2). Notably, also the clinically used inhibitors letrozole

and exemestane can establish persistent H-bonds to the backbone

of Met374 [35]. The presence of these H-bonds appeared therefore

as another critical element to boost AI potency to the sub mM range.

In general, even the ligands unable to coordinate the iron atom,

owing to geometric constraints, as 4b, can engage a persistent H-

bond with the backbone of Leu372, providing a rationale for their

activity in the low

of inhibitors with the AR catalytic site, we also calculated their

binding free energy ( G_b) with the Molecular Mechanics General-

mM range. In order to dissect the key interactions

D

ized Born Surface Area (MM-GBSA) method [39]. Next, by per-

forming a decomposition analysis, we picked the residues and the

rigid, due to the delocalization of the

pꢀorbitals. This strikingly

Table 3

Total and Decomposition of the binding free energy (

enzyme (a) and to Estrogen receptor (ER) (b) as obtained from this QM/MM and classical molecular dynamics trajectories analyses, respectively. The residues which

G_bby more than 0.5 kcal/mol are listed.

DG_b, kcal/mol) obtained from Molecular Mechanics Generalized Surface Area for the binding of 3a to the aromatase (AR)

a

contribute to

D

(a)

3a/AR

DG_btot ¼ ꢀ15.95 8.93

Residue

DG_{b per residue}

Arg115

Ile133

ꢀ0.86 0.66

ꢀ1.21 0.41

ꢀ0.80 0.20

ꢀ1.52 0.31

ꢀ0.64 0.45

ꢀ0.71 0.40

ꢀ0.93 0.52

ꢀ1.92 0.32

ꢀ1.54 0.57

ꢀ0.89 0.28

ꢀ1.15 0.35

ꢀ1.18 0.38

ꢀ7.31 4.36

Phe134

Trp224

Ala306

Asp309

Thr310

Val370

Leu372

Val373

Met374

Leu477

Ligand

(b)

3a/ER

a

Residue

Monomer A

Monomer B

DG_btot ¼ ꢀ37.50 3.70

DG_btot ¼ ꢀ34.28 2.54

DG_{b per residue}

Met343

Leu346

Thr347

Leu349

Ala350

Trp383

Leu384

Leu387

Met388

Leu391

Phe404

Val418

Glu419

Met421

Ile424

ꢀ0.79 0.29

ꢀ1.19 0.50

ꢀ0.69 0.24

ꢀ0.92 0.26

ꢀ1.79 0.30

ꢀ1.57 0.44

ꢀ0.74 0.27

ꢀ1.93 0.32

ꢀ0.78 0.31

ꢀ0.92 0.40

ꢀ0.86 0.29

ꢀ0.75 0.24

ꢀ1.18 0.43

ꢀ0.74 1.09

ꢀ1.99 0.52

ꢀ0.56 0.25

ꢀ1.11 0.41

ꢀ1.46 0.63

ꢀ1.01 0.32

ꢀ1.16 0.34

ꢀ0.91 0.27

ꢀ0.56 0.24

ꢀ1.12 0.27

ꢀ0.63 0.22

ꢀ0.62 0.21

Leu428

Met522

Hie524

Leu525

Met528

Ligand

ꢀ0.51 0.14

ꢀ0.72 0.51

ꢀ1.43 0.46

ꢀ1.49 0.55

ꢀ17.84 1.85

ꢀ0.78 0.37

ꢀ3.47 0.42

ꢀ14.90 1.31

8

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

Fig. 5. Binding pose of letrozole (LTZ) and the selected newly synthesized inhibitors 3a, 3b, 4a, 4b, 5a, 6a and 6b to the aromatase (AR) active site as obtained from representative

frames of QM/MM MD trajectories. AR structure is displayed in green new cartons, the heme moiety, the cysteine and the inhibitors are shown in balls and sticks, the residues

forming hydrogen bonds to the ligands are shown in licorice. All atoms are coloured by atom name (H white, O red, S yellow and N blue, C cyan).

9

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

nature of the interactions that more strongly contribute to

D

G_bof

synthesized, subjected to biological evaluation and structurally

characterized, in order to generate drug-candidates that may offer

improved efﬁcacy and lower side effects compared to a single drug

or a combination of AI and SERM/SERD.

each inhibitor. We remark that this type of analysis, performed at

force ﬁeld level on the QM/MM geometries, completely neglects

the contribution of the inhibitor-Fe coordination bond and only

provides information on classical non bonded interactions (hy-

drophobic or electrostatic interactions) between the ligand and the

protein. Interestingly, the most potent inhibitors 3a, 4a, 6a and 6b

established hydrophobic interactions between their ethyl moiety

and Ile133 and Trp244, and between their phenol rings and Val370

and Leu372 (Table S3).

The analysis of predicted binding poses hence provided a

structural explanation of previous SAR considerations, disclosing

that an optimal spacer length, which enables the coordination to

the heme coordinating moiety without making the substituent too

bulky, and the ethyl group, which, instead, ﬁlls the hydrophobic

cavity, allow for a ﬂexible inhibitor architecture adaptable to the

tight structural constraints of the AR active site.

Considering AR inhibition activity, the presence of the TAM-like

ethyl group and a spacer between the double bond and the het-

erocycle appeared as critical hallmarks for high potency. In detail,

for compounds carrying a pyridine ring, only the methylene spacer

proved to have the optimal length for appropriately accommoda-

ting into AR active site. Conversely, for imidazole-based derivatives,

the extension of the linker to two methylene units did not reduce

the afﬁnity for the target, leading to an equipotent activity of

elongated compounds carrying a methyl (6b) and ethyl (6a) moiety

on the double bond. Even though most of the compounds exhib-

iting the best IC₅₀s as AIs carried imidazole as heterocycle (4a, 6a

and 6b), the most potent compound resulted to be 3a, which,

instead, contains a pyridine group. This evidence suggests that both

heterocycles allowed for the appropriate interactions with AR,

when a suitable and AR-adaptable structure of the whole molecule

was provided, leading to the development of potent AIs.

Next, we performed extensive classical MD simulations on the

docking pose of each inhibitor/ER

the activity of the compounds measured on ER

a

adduct to provide a rationale for

. For each of the 12

a

adducts we performed 500 ns of MD simulations.

Considering the afﬁnity of the compounds towards ERa, the

Compounds 3a and 5a proved to be the most potent of the se-

presence of a spacer, together with the TAM-like ethyl moiety,

again seemed to be critical structural traits for optimal binding.

Indeed, the most potent compounds (3a and 5a, IC₅₀s of 19 nM and

33 nM, respectively) carried the ethyl moiety on the double bond,

beside a pyridine ring and a spacer. As an exception, 1b with no

linker and a methyl substituent on the double bond, also proved to

have high afﬁnity. Nonetheless, no other clear SAR information

ries, inhibiting estrogen binding to ER

33 nM, respectively. Strikingly, these molecules presented among

the highest G_bto ER (Table S5), consistently with docking scores

a with IC₅₀values of 19 and

D

a

(Table S1). A decomposition analysis clearly revealed that their

binding is stabilized by hydrophobic interactions (Fig. S5 and

Tables 3 and S5-S7). In particular, 3a establishes hydrophobic in-

teractions with Ala350, Leu384 and Met421 (Table S7) and H-bonds

between the hydroxyl of one phenol ring to the backbone of Glu419

and His424 (Fig. 6 and Table S8). As well, 5a is mostly stabilized by

hydrophobic interactions, being Leu346, Ala350, Leu384, Leu387,

could be gained, since similar sub-mM potency was seen for both

rigid no-spacer molecules and ﬂexible one- or two-methylene

linker derivatives. This outcome may be due to the ﬂexible nature

of the receptor LBC, which adapts more easily than AR to the

binding of the drug-candidates. Stunningly, the best performing

compound presents a better IC₅₀than the reference END.

and Leu525 the residues mostly contributing to its

ligand shares with 3a the same persistent H-bond to the backbone

of Glu419. Ostensibly, most compounds, irrespectively of their IC₅₀

DG_b. This latter

,

In view of developing novel multitargeting strategies, the most

interesting result was the identiﬁcation of compounds able to

often H-bond to Glu419 and/or Glu353 (Table S2) and are stabilized

by a similar set of hydrophobic interactions of the most active in-

modulate both AR and ER

balanced potency. In particular, both 2b and 6a proved to be sub-

AIs and ER ligands, although their activities were not perfectly

balanced. Most importantly, 3a was endowed with a striking ac-

tivity in the low nanomolar range (IC₅₀s of 23 (AR) and 19 (ERa)

a, in some cases, with signiﬁcant and

hibitors, pinpointing that the

their measured IC₅₀

In order to assess the impact of the inhibitors on the structural

and dynamical properties of ER , we also inspected how their

D

G

_bper se does not fully account for

mM

.

a

binding affected the H-bond network of H12, which is known to

trigger the conformational switch between the active and the

nM) on both targets, showing a moderate selective antiproliferative

effects on a BC cell line. Notably, this compound showed a more

inactive state of ER

pointed by us to a signature of ER

our previous study, where END was observed to stabilize the

antagonist conformation of ER by preventing the formation of H-

bonds between Glu380 and Tyr537 or the backbone of Leu536, 3a

and 5a, characterized by the best IC₅₀, exhibited a reduced H-bond

network around Glu380 with respect to the inhibitors endowed

with lower activity.

a

. This H-bond network was previously pin-

balanced activity when compared to previously reported AR/ER

dual acting compounds.

The most interesting compounds identiﬁed, 3a and 6a, also

proved to be endowed with SERM activity, being able to antagonize

the expression of progesterone receptor. Notably, most of the

a

activation [6]. Consistently with

a

studied compounds, displaying sub-mM IC₅₀towards ERa, reduced

ER levels to an extent comparable to that observed with fulves-

a

trant. 5b showed the most remarkable effect at a lower dose than

the reference compound, and could be worth further evaluation to

access its potential as a promising SERD candidate.

Given these results, 3a proved to be the most promising and

balanced dual acting AI-SERM candidate of the series, displaying a

The presence/absence of this H-bond network must translate in

the presence/lack of speciﬁc cross correlation contacts (i.e.

dynamically correlated motion) between H12 and H3/H5, under-

lying the activation of ER

version of the cross-correlation matrix to monitor the dynamical

coupling of the speciﬁc ER structural elements (Fig. 7 and S5, S6).

a [6]. Hence, we computed a coarse

signiﬁcant reduction of ER

a protein amount and no cytotoxicity

a

towards healthy mammary cells.

Remarkably, all ligands exhibited small correlation coefﬁcients

between H12 and H3/H5, similarly to END, suggesting their binding

Remarkably, our detailed characterization provides a practical

guideline of the key structural traits typifying a potent and

stabilize the antagonist conformation of ER

consistently with their activity as SERM/SERD compounds.

a

(Figs. 6 and S5),

balanced AR/ER

a

multitarget anti ER þ BC compound, such as the

most potent and balanced compound 3a: (i) the presence of the

hydroxyl substituents enables the formation of a persistent H-bond

to the backbone of Leu372 of AR, and to the backbone of Glu419 of

3. Discussion

ER

of ERa

a

, (ii) the ethyl moiety ﬁlls both the active site of AR and the LBC

, allowing in AR the formation of a coordination bond to the

In this study, multitarget AI/ER

a

ligands were designed,

10

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

11

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

iron atom and of an optimal set of hydrophobic interactions in both

interactions with the target cavity. In detail, the intramolecular p-

targets. Remarkably the AR active site and the LBC of ER

highly similar content of hydrophobic residues (the residues pre-

sent within 3 Å of 3a are 3 Leu, 2 Val, 2 Phe,1 Thr,1 Met, 2 Trp,1 Arg,

a

share a

stacking of the pyridine and the phenol extends to Trp224, which

closes the AR active site, while Val370 is packed between the two

inhibitor's phenol rings. Conversely in ERa, the pyridine/phenol are

for AR and 5 Leu, 1 Val,1 Thr, 3 Met, 1 Phe for ER

of a methylene linker, connecting the pyridine moiety to the vinyl

core, enables an internal -stacking, between the pyridine and the

phenol ring in cis position on the double bond core, while hydro-

phobic interactions occur between the remaining ethyl and phenol

moieties. This inhibitor's scaffold enables the formation of optimal

a

); (iii) the presence

lined by Ala350 and Leu384 and two methionine residues (Met388

and Met421) insert between the phenols, while Leu525 and

Met528 line the ethyl moiety of the ligand. Hence, a dual acting

drug-candidate must be able to mainly establish hydrophobic in-

teractions with the functional sites of the two targets, as well as

intra-ligand hydrophobic interactions, which allow it to pack and

p

Fig. 6. Binding pose of the synthesized inhibitors to the ligand-binding cavity of ER

mode of endoxifen (END) and AZD-9496 (AZD), are also reported for comparison. ER

a

as obtained from a representative cluster (Table S4) of classical MD trajectory. The binding

is shown as white new cartoons, the drug is displayed with back carbon atoms, the residues

forming hydrophobic interactions are shown in green, those forming hydrogen-bonds are shown in blue, residues disfavouring the binding (in the decomposition energy analysis

reported in Table S7) are displayed in orange. The rest of the atoms are coloured by atom name (H white, O red, S yellow and N blue).

Fig. 6. (continued).

12

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

grade. Reaction progress was followed by thin layer chromatog-

raphy (TLC) on precoated silica gel plates (Merck Silica Gel 60 F254)

and then visualized with a UV254 lamplight. Chromatographic

separations were performed on silica gel columns by ﬂash method

(Kieselgel 40, 0.040e0.063 mm, Merck). Melting points were

determined in open glass capillaries, using a Büchi apparatus and

are uncorrected. ¹H NMR and ¹³C NMR spectra were recorded on a

Varian Gemini spectrometer 400 MHz and 101 MHz, respectively,

in CDCl₃solutions unless otherwise indicated, and chemical shifts

(d) were reported as parts per million (ppm) values relative to

tetramethylsilane (TMS) as internal standard; coupling constants

(J) are reported in Hertz (Hz). Standard abbreviations indicating

spin multiplicities are given as follow: s (singlet), d (doublet), t

(triplet), br (broad), q (quartet) or m (multiplet). Mass spectra were

recorded on Waters Xevo G2-XS QTof apparatus operating in

electrospray mode (ES). All tested compounds were found to have

>95% purity, as determined by HPLC analysis, performed on a

chromatograph PU-1587 UV model equipped with a 20

valve (Jasco Europe, Italy) by using a Phenomenex Luna 5

m

L loop

m C18

column (150 ꢁ 4.60 mm) as stationary phase and a mixture of H₂O/

MeCN (60:40, v/v) as mobile phase for compounds 1a,b, 3a,b and

5a,b, or a Waters Spherisorb 5

stationary phase and a mixture of H₂O/MeCN (70:30, v/v) as mobile

phase for compounds 2a,b, 4a,b and 6a,b; for all compounds,

m

m C18 column (250 ꢁ 4.60 mm) as

detection at

l

¼ 254 nm, ﬂow rate of 1.0 mL/min. Compounds were

named relying on the naming algorithm developed by Cam-

bridgeSoft Corporation and used in ChemDraw Professional 19.0.

4.1.2. General procedure of McMurry reaction

Method A. Zinc powder (10 eq) was suspended in anhydrous

THF, the mixture was cooled to 0 ^ꢂC and TiCl₄(5 eq) was added

dropwise under nitrogen atmosphere. At the end of the addition,

the mixture was brought to room temperature and then warmed to

reﬂux for 2 h. After cooling, a solution of benzophenone (1 eq) and

the corresponding ketone (3 eq) in dry THF was added and the

mixture was heated to reﬂux in the dark for 3 h. After cooling, the

mixture was poured into a 10% K₂CO₃solution and extracted with

diethylether (3 ꢁ 25 mL). The combined organic layers were

washed with brine, dried over anhydrous Na₂SO₄and concentrated

to dryness. The crude compound was then puriﬁed by ﬂash column

chromatography with a suitable eluent to provide the desired

McMurry product.

Method B. Zinc powder (6 eq) was suspended in anhydrous THF,

the mixture was cooled to 0 ^ꢂC and TiCl₄(3 eq) was added dropwise

under nitrogen atmosphere. At the end of the addition, the mixture

was brought to room temperature and then warmed to reﬂux for

1 h. After cooling, a solution of benzophenone (1 eq) and the cor-

responding ketone (1 eq) in dry THF was added and the mixture

was heated to reﬂux for 2 h. After cooling, the mixture was poured

into a 10% K₂CO₃solution and extracted with diethyl ether

(3 ꢁ 25 mL). The combined organic layers were washed with brine,

dried over anhydrous Na₂SO₄and concentrated to dryness. The

crude compound was then puriﬁed by ﬂash column chromatog-

raphy with a suitable eluent to provide the desired McMurry

product.

Fig. 7. Sum of per-residue cross-correlation coefﬁcients of H12 residues with the rest

of the motifs composing Ligand Binding Domain (LBD). Left and right columns refer to

monomers A and B, respectively. Top report the correlation coefﬁcient of 1-6a,b bound

to ERa, bottom reports the internal correlation of 1b and 5a to Y537S ERa. ERa LBD

monomer structural motifs are labeled as, H1: residues 303e311; H2: residues

312e322; coil2e3 (C2e3): residues 323e338 (orange); H3: residues 339e363; H4:

residues 364e371; H5: residues 372e382; H6: residues 383e396; S1, S2: residues

397e411; H7: residues 412e419; H8: residues 420e439; H9: residues 440e463; H10:

residues 464e494; H11: residues 495e531; H12: residues 532e552. See Fig. S6 for a

visual map of the decomposition.

snugly ﬁt in the binding cavity.

In conclusion, this study provides novel potent, selective and

balanced dual-targeting AR/ERa modulators as antitumor drug-

candidates, unprecedentedly unveiling the potential of a pyridine

substituent as a structural moiety to be employed in the develop-

ment of novel multipotent anticancer agents. This study strongly

remarks that the development of multipotent compounds, able to

simultaneously modulate these utmost important BC targets, could

represent an appealing and viable therapeutic strategy to expand

the current arsenal of pharmacological strategies in the ever-

harsher battle against BC.

4.1.2.1. 4,4’-(2-(pyridine-3-yl)but-1-ene-1,1-diyl)diphenol

(1a).

Bis(4-hydroxyphenyl)methanone (0.50 g, 2.30 mmol) and 1-(pyr-

idin-3-yl)propan-1-one (7a) (0.29 mL, 2.30 mmol) in dry THF

(11.80 mL) were reacted according to Method B. The crude was

puriﬁed by ﬂash column chromatography (petroleum ether/ethyl

acetate 1:1) to give 1a (0.18 g, yield 24%), mp 240e241 ^ꢂC. ¹H NMR

4. Experimental section

4.1. Chemistry

4.1.1. General Methods

Starting materials, unless otherwise speciﬁed, were used as high

grade purity commercial products. Solvents were of analytical

(methanol-d₄):

d

0.95 (t, J ¼ 7.4 Hz, 3H, CH₃), 2.55 (q, J ¼ 7.4 Hz, 2H,

CH₂), 6.46 (d, J ¼ 8.8 Hz, 2H, arom), 6.67 (d, J ¼ 8.4 Hz, 2H, arom),

6.78 (d, J ¼ 8.8 Hz, 2H, arom), 7.05 (d, J ¼ 8.4 Hz, 2H, arom), 7.29 (dd,

13

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

J ¼ 4.8 and 7.6 Hz, 1H, arom), 7.67 (dd, J ¼ 2.0 and 7.6 Hz, 1H, arom),

8.16 (d, J ¼ 1.6 Hz, 1H, arom), 8.22 (dd, J ¼ 1.6 and 4.8 Hz, 1H, arom).

(8 mL) were reacted according to Method B. The crude was puriﬁed

by ﬂash column chromatography (petroleum ether/ethyl acetate

1:1) to give 3b (0.13 g, yield 28%), mp 205e207 ^ꢂC. ¹H NMR

¹³C NMR (methanol-d₄):

d 13.8, 29.2, 115.5 (2C), 115.9 (2C), 124.7,

131.5 (2C), 133.2 (2C), 135.3, 135.6, 137.4, 139.1, 140.9, 142.9, 146.9,

151.5, 156.9, 157.7. HRMS C₂₁H₁₉NO₂[MþH] calcd: 318.14886,

found: 318.14897.

(acetone-d₆): d 1.67 (s, 3H, CH₃), 3.52 (s, 2H, CH₂Py), 6.77e6.80 (m,

4H, arom), 7.02 (d, J ¼ 8.8 Hz, 2H, arom), 7.06 (d, J ¼ 8.4 Hz, 2H,

arom), 7.28 (dd, J ¼ 4.8 and 7.7 Hz, 1H, arom), 7.58 (d, J ¼ 8.0 Hz, 1H,

arom), 8.23 (d, J ¼ 8.8 Hz, 1H, arom), 8.39e8.41 (m, 1H, arom). ¹³

C

4.1.2.2. 4,4’-(2-(pyridine-3-yl)prop-1-ene-1,1-diyl)diphenol

(1b).

NMR (acetone-d₆): d 19.7, 39.0, 115.4 (2C), 115.6 (2C), 124.1, 130.9,

Bis(4-hydroxyphenyl)methanone (0.50 g, 2.30 mmol) and 1-(pyr-

idin-3-yl)ethan-1-one (7b) (0.78 mL, 7.00 mmol) in dry THF (19 mL)

were reacted according to Method A. The crude was puriﬁed by

ﬂash column chromatography (petroleum ether/ethyl acetate 3:2,

then 1:1, then 2:3) and crystallized from ethyl acetate to give 1b

131.1 (2C), 131.3 (2C), 135.1, 135.2, 136.5, 136.9, 140.5, 147.9, 150.6,

156.6, 156.7. HRMS C₂₁H₁₉NO₂[MþH] calcd: 318.14886, found:

318.14866.

4.1.2.7. 4,4'-(2-ethyl-4-(pyridin-3-yl)but-1-ene-1,1-diyl)diphenol

(5a). Bis(4-hydroxyphenyl)methanone (0.07 g, 0.30 mmol) and 9a

(0.05 g, 0.30 mmol) in dry THF (1.50 mL) were reacted according to

Method B. The crude was puriﬁed by ﬂash column chromatography

(dichloromethane/acetone 9:1) to give 5a (0.02 g, yield 19%), mp

(0.17 g, yield 24%), mp ˃ 230 ^ꢂC. ¹H NMR (methanol-d₄):

d 2.18 (s, 3H,

CH₃), 6.50 (d, J ¼ 8.4 Hz, 2H, arom), 6.71 (d, J ¼ 8.4 Hz, 2H, arom),

6.80 (d, J ¼ 8.8 Hz, 2H, arom), 7.08 (d, J ¼ 8.0 Hz, 2H, arom), 7.29 (dd,

J ¼ 5.0 and 7.8 Hz, 1H, arom), 7.71 (d, J ¼ 8.0 Hz, 1H, arom),

8.19e8.23 (m, 2H, arom). ¹³C NMR (methanol-d₄):

d

23.8, 116.4 (2C),

210e212 ^ꢂC. ¹H NMR (methanol-d₄):

d

1.02 (t, J ¼ 7.4 Hz, 3H, CH₃),

116.7 (2C), 125.5, 131.5, 132.9 (2C), 134.2 (2C), 136.2, 136.4, 139.6,

143.5, 144.0, 147.6, 151.9, 157.9, 158.4. HRMS C₂₀H₁₇NO₂[MþH]

calcd: 304.13321, found: 304.13319.

2.21 (q, J ¼ 7.6 Hz, 2H, CH₂), 2.46 (t, J ¼ 7.6 Hz, 2H, CH₂), 2.73 (t,

J ¼ 7.8 Hz, 2H, CH₂), 6.64e6.69 (m, 4H, arom), 6.72e6.75 (m, 2H,

arom), 6.84e6.86 (m, 2H, arom), 7.29 (dd, J ¼ 4.9 and 7.8 Hz, 1H,

arom), 7.46e7.48 (m, 1H, arom), 8.21 (d, J ¼ 1.2 Hz, 1H, arom), 8.32

4.1.2.3. 4,4'-(2-(1H-imidazole-4-yl)but-1-ene-1,1-diyl)diphenol (2a).

Bis(4-hydroxyphenyl)methanone (0.50 g, 2.30 mmol) and 10a

(0.29 g, 2.30 mmol) in dry THF (12 mL) were reacted according to

Method B. The crude was puriﬁed by ﬂash column chromatography

(petroleum ether/ethyl acetate 1:1, then 3:7, then 2:8) to give 2a

(dd, J ¼ 5.2 and 1.6 Hz, 1H, arom). ¹³C NMR (methanol-d₄):

d 13.6,

26.0, 32.6, 34.0, 115.7 (2C), 115.8 (2C), 125.0, 131.1 (2C), 131.2 (2C),

136.6, 136.2, 138.4, 138.6, 139.8, 140.7, 147.3, 150.0, 156.8, 156.8.

HRMS C₂₃H₂₃NO₂[MþH] calcd: 346.18016, found: 346.17999.

(0.06 g, yield 9%), mp 147e150 ^ꢂC. ¹H NMR (methanol-d₄):

d

0.99 (t,

4.1.2.8. 4,4'-(2-methyl-4-(pyridin-3-yl)but-1-ene-1,1-diyl)diphenol

(5b). Bis(4-hydroxyphenyl)methanone (0.26 g, 1.20 mmol) and 9b

(0.18 g, 1.20 mmol) in dry THF (6 mL) were reacted according to

Method B. The crude was puriﬁed by ﬂash column chromatography

(petroleum ether/ethyl acetate 2:3) to give 5b (0.18 g, yield 45%),

J ¼ 7.4 Hz, 3H, CH₃), 2.47 (q, J ¼ 7.4 Hz, 2H, CH₂), 6.42 (s, 1H, arom),

6.57 (d, J ¼ 8.8 Hz, 2H, arom), 6.74 (d, J ¼ 8.4 Hz, 2H, arom), 6.80 (d,

J ¼ 8.4 Hz, 2H, arom), 6.99 (d, J ¼ 8.8 Hz, 2H, arom), 7.48 (s, 1H,

arom). ¹³C NMR (methanol-d₄):

d 14.4, 28.0, 115.7 (2C), 115.9 (2C),

123.6, 131.3 (3C), 132.2 (2C), 135.3, 136.0, 136.4 (2C), 141.1, 157.1,

mp 206e207 ^ꢂC. ¹H NMR (methanol-d₄):

d 1.81 (s, 3H, CH₃), 2.45 (t,

157.4. HRMS

307.14422.

C

₁₉H₁₈N₂O₂[MþH] calcd: 307.14410, found:

J ¼ 7.6 Hz, 2H, CH₂), 2.79 (t, J ¼ 7.4 Hz, 2H, CH₂), 6.63e6.70 (m, 6H,

arom), 6.81e6.84 (m, 2H, arom), 7.29 (dd, J ¼ 4.9 and 7.8 Hz, 1H,

arom), 7.50 (d, J ¼ 7.6 Hz, 1H, arom), 8.24 (d, J ¼ 1.6 Hz, 1H, arom),

8.33 (dd, J ¼ 4.8 and 1.6 Hz, 1H, arom). ¹³C NMR (methanol-d₄):

4.1.2.4. 4,4'-(2-(1H-imidazole-4-yl)prop-1-ene-1,1-diyl)diphenol

(2b). Bis(4-hydroxyphenyl)methanone (0.34 g, 1.60 mmol) and 10b

(0.18 g, 1.60 mmol) in dry THF (8 mL) were reacted according to

Method B. The crude was puriﬁed by ﬂash column chromatography

(petroleum ether/ethyl acetate 2:3, then 0.5:9.5) to give 2b (0.07 g,

d

19.9, 32.4, 37.8, 115.6 (2C), 115.7 (2C), 125.0, 131.4 (2C), 131.5 (2C),

132.3, 136.0, 136.1, 138.4, 139.6, 140.6, 147.3, 150.0, 156.7, 156.8.

HRMS C₂₂H₂₁NO₂[MþH] calcd: 332.16451, found: 332.16434.

yield 15%) mp 161e162 ^ꢂC. ¹H NMR (methanol-d₄):

d

2.10 (s, 3H,

4.1.2.9. 4,4'-(2-ethyl-4-(1H-imidazole-4-yl)but-1-ene-1,1-diyl)diphe-

nol (6a). Bis(4-hydroxyphenyl)methanone (0.11 g, 0.50 mmol) and

11a (0.08 g, 0.50 mmol) in dry THF (2.50 mL) were reacted ac-

cording to Method B. The crude was puriﬁed by ﬂash column

chromatography (petroleum ether/ethyl acetate 0.5:9.5; then ethyl

acetate; then ethyl acetate/methanol 8:2) to give 6a (0.02 g, yield

CH₃), 6.37 (s, 1H, arom), 6.63 (d, J ¼ 8.0 Hz, 2H, arom), 6.74 (d,

J ¼ 8.8 Hz, 2H, arom), 6.84 (d, J ¼ 8.8 Hz, 2H, arom), 6.99 (d,

J ¼ 8.4 Hz, 2H, arom), 7.45 (s, 1H, arom). ¹³C NMR (methanol-d₄):

d

21.0, 115.8 (2C), 115.9 (2C), 125.0, 131.9 (3C), 132.2 (2C), 135.2,

135.9,136.5 (2C),140.8,157.3, 157.3 HRMS C₁₈H₁₆N₂O₂[MþH] calcd:

293.12845, found: 293.12837.

11%), mp 139e141 ^ꢂC. ¹H NMR (methanol-d₄):

d

1.01 (t, J ¼ 7.4 Hz,

3H, CH₃), 2.18 (q, J ¼ 7.6 Hz, 2H, CH₂), 2.45 (t, J ¼ 7.8 Hz, 2H, CH₂),

2.68 (t, J ¼ 7.8 Hz, 2H, CH₂), 6.64e6.70 (m, 5H, arom), 6.82 (d,

J ¼ 8.8 Hz, 2H, arom), 6.89 (d, J ¼ 8.4 Hz, 2H, arom), 7.51 (s, 1H,

4.1.2.5. 4,4’-(2-(pyridine-3-ylmethyl)but-1-ene-1,1-diyl)diphenol

(3a). Bis(4-hydroxyphenyl)methanone (0.20 g, 0.90 mmol) and 8a

(0.14 g, 0.90 mmol) in dry THF (4 mL) were reacted according to

Method B. The crude was puriﬁed by ﬂash column chromatography

(ethyl acetate) to give 3a (0.03 g, yield 10%), mp 202e204 ^ꢂC. ¹H

arom). ¹³C NMR (methanol-d₄):

d 13.9, 26.0, 26.2, 32.4, 115.9 (2C),

116.0 (3C), 131.5 (2C), 131.5 (2C), 133.8, 136.0, 136.5, 136.6, 139.4,

140.4, 156.9, 156.9. HRMS C₂₁H₂₂N₂O₂[MþH] calcd: 335.17540,

found: 335.17552.

NMR (methanol-d₄):

d

0.97 (t, J ¼ 7.2 Hz, 3H, CH₃), 2.05 (q, J ¼ 7.2 Hz,

2H, CH₂), 3.57 (s, 2H, CH₂Py), 6.69e6.72 (m, 4H, arom), 6.97e6.99

(m, 4H, arom), 7.35 (dd, J ¼ 4.7 and 7.8 Hz, 1H, arom), 7.66 (d,

J ¼ 7.6 Hz, 1H, arom), 8.32e8.34 (m, 2H, arom). ¹³C NMR (methanol-

4.1.2.10. 4,4'-(4-(1H-imidazole-4-yl)-2-methylbut-1-ene-1,1-diyl)

diphenol

(6b). Bis(4-hydroxyphenyl)methanone

(0.13

g,

d₄):

d

13.5, 26.0, 35.5, 115.8 (2C), 115.9 (2C), 125.1, 131.2 (2C), 131.3

0.60 mmol) and 11b (0.08 g, 0.60 mmol) in dry THF (3 mL) were

reacted according to Method B. The crude was puriﬁed by ﬂash

column chromatography (ethyl acetate; then ethyl acetate/meth-

anol 8:2) to give 6b (0.03 g, yield 16%), mp 95e99 ^ꢂC. ¹H NMR

(methanol-d₄): 1.78 (s, 3H, CH₃), 2.44 (t, J ¼ 7.6 Hz, 2H, CH₂), 2.73 (t,

J ¼ 7.6 Hz, 2H, CH₂), 6.65e6.68 (m, 5H, arom), 6.78 (d, J ¼ 8.8 Hz, 2H,

arom), 6.87 (d, J ¼ 8.8 Hz, 2H, arom), 7.54 (s, 1H, arom). ¹³C NMR

(2C), 135.8, 135.9, 137.4, 138.3, 138.7, 141.9, 147.5, 150.1, 157.0, 157.1.

HRMS C₂₂H₂₁NO₂[MþH] calcd: 332.16451, found: 332.16485.

4.1.2.6. 4,4'-(2-methyl-3-(pyridin-3-yl)prop-1-ene-1,1-diyl)diphenol

(3b). Bis(4-hydroxyphenyl)methanone (0.32 g, 1.50 mmol) and 1-

(pyridin-3-yl)propan-2-one (8b) (0.20 g, 1.50 mmol) in dry THF

14

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

(methanol-d₄):

d

19.8, 26.1, 36.4, 115.5 (2C), 115.7 (2C), 118.7, 131.5

chromatography (toluene/ethyl acetate 7:3, 1:1, ethyl acetate) to

(2C), 131.7 (2C), 133.1, 135.2, 136.2, 136.4, 137.6, 139.9, 156.7, 156.7.

give 14a (0.10 g, yield 19% two steps) as brown oil. ¹H NMR:

d 0.95 (t,

HRMS C₂₀H₂₀N₂O₂[MþH] calcd: 321.15975, found: 321.15967.

J ¼ 7.6 Hz, 3H, CH₃), 2.03 (q, J ¼ 7.4 Hz, 2H, CH₂), 3.46 (s, 6H, 2 OCH₃),

4.63 (s, 2H, CH₂-imi), 5.14 (s, 4H, 2 OCH₂), 6.88 (s, 1H, arom),

6.95e6.99 (m, 4H, arom), 7.04e7.07 (m, 5H, arom), 7.45 (s, 1H,

arom).

4.1.2.11. 4,4'-(2-methylbut-1-ene-1,1-diyl)diphenol

(12a).

Bis(4-hydroxyphenyl)methanone (0.60 g, 2.80 mmol) and methyl-

ethylketone (0.25 mL, 0.20 g, 2.80 mmol) in dry THF (13 mL) were

reacted according to Method B. The crude was puriﬁed by ﬂash

column chromatography (petroleum ether/ethyl acetate 3,5:1,5) to

give 12a (0.53 g, yield 73%) mp 175e176 ^ꢂC (lit [40]. mp

4.1.4.2. 1-(3,3-bis(4-(methoxymethoxy)phenyl)-2-methylallyl)-1H-

imidazole (14b). Starting from 13b (0.35 g, 1.10 mmol), a crude

compound was obtained that was puriﬁed by ﬂash column chro-

matography (toluene/ethyl acetate 9:1, 7:3, 1:1, ethyl acetate) to

give 14b (0.08 g, yield 19% two steps) mp 231e232 ^ꢂC. ¹H NMR:

179e180 ^ꢂC). ¹H NMR:

d

1.03 (t, J ¼ 7.4 Hz, 3H, CH₃),1.76 (s, 3H, CH₃),

2.11 (q, J ¼ 7.4 Hz, 2H, CH₂), 4.68 (br, 2H, 2OH), 6.72e6.74 (m, 4H,

arom), 6.97e7.00 (m, 4H, arom).

d

1.71 (s, 3H, 2 CH₃), 3.28 (s, 6H, 2 OCH₃), 4.60 (s, 2H, CH₂-imi), 5.24

(s, 4H, 2 OCH₃), 6.89 (s, 1H, arom), 6.95e6.97 (m, 4H, arom),

7.03e7.05 (m, 5H, arom), 7.47 (s, 1H, arom).

4.1.2.12. 4,4'-(2-methylprop-1-ene-1,1-diyl)diphenol

(12b).

Bis(4-hydroxyphenyl)methanone (0.60 g, 2.80 mmol) and acetone

(0.20 mL, 0.16 g, 2.80 mmol) in dry THF (13 mL) were reacted ac-

cording to Method B. The crude was puriﬁed by ﬂash column

chromatography (petroleum ether/ethyl acetate 4:1) to give 12b

(0.56 g, yield 83%) mp 180e182 ^ꢂC (lit [40]. mp 188e189 ^ꢂC). ¹H

4.1.5. General procedure for the synthesis of compound 4a,b

To a solution of 14a or 14b in methanol (5 mL), 37% HCl (0.05 mL)

was added. The reaction mixture was stirred at room temperature

for 12 h and then the solvent was evaporated to dryness. The res-

idue was dissolved in H₂O and neutralized with saturated solution

of NaHCO₃. The formed precipitate was collected by vacuum

ﬁltration to afford compound 4a or 4b.

NMR:

d 1.79 (s, 6H, 2CH₃), 4.60 (br, 2H, 2OH), 6.71e6.75 (m, 4H,

arom), 6.97e7.00 (m, 4H, arom).

4.1.3. General procedure for the synthesis of compounds 13a,b

A suspension of 12a or 12b (1 eq) and NaH (60%, 4.1 eq) in dry

THF (15e20 mL) was stirred at room temperature for 30 min under

an inert N₂atmosphere. Then, chloromethylmethylether (4 eq) was

added and the resulting mixture was stirred at room temperature

for 3 h. The reaction was quenched with aq NaHCO₃, the organic

solvent was removed under reduced pressure and the aqueous

layer was extracted with ethyl acetate. The organic phase was dried

and concentrated in vacuum and the crude compound was then

puriﬁed by ﬂash column chromatography.

4.1.5.1. 4,4'-(2-((1H-imidazole-1-yl)methyl)but-1-ene-1,1-diyl)

diphenol (4a). Starting from 14a (0.10 g, 0.20 mmol), 4a was ob-

tained (0.06 g, yield 83%), mp>250 ^ꢂC. ¹H NMR (acetone-d₆):

d 0.95

(t, J ¼ 7.4 Hz, 3H, CH₃), 2.00 (q, J ¼ 7.4 Hz, 2H, CH₂), 4.76 (s, 2H, CH₂-

imi), 6.79e6.84 (m, 4H, arom), 6.98 (s, 1H, arom), 7.03e7.09 (m, 5H,

arom), 7.65 (s, 1H, arom). ¹³C NMR (acetone-d₆):

d 13.4, 24.0, 47.9,

115.8 (2C), 115.9 (2C),129.2, 130.8 (2C), 131.0 (2C), 131.4, 132.5, 134.2,

134.4, 134.5, 143.6, 157.3, 157.4. HRMS C₂₀H₂₀N₂O₂[MþH] calcd:

321.15975, found: 321.15993.

4.1.3.1. 4,4'-(2-methylbut-1-ene-1,1-diyl)bis((methoxymethoxy)ben-

zene) (13a). Starting from 12a (0.53 g, 2.10 mmol), a crude com-

pound was obtained that was puriﬁed by ﬂash column

chromatography (petroleum ether/ethyl acetate 9.5:0.5) to give 13a

4.1.5.2. 4,4'-(3-(1H-imidazole-1-yl)-2-methylprop-1-ene-1,1-diyl)

diphenol (4b). Starting from 14b, (0.08 g, 0.20 mmol), 4b was ob-

tained (0.03 g, yield 30%), mp>250 ^ꢂC. ¹H NMR (acetone-d₆):

d 1.64

(s, 3H, CH₃), 4.69 (s, 2H, CH₂-imi), 6.79 (d, J ¼ 8.6 Hz, 2H, arom), 6.83

(d, J ¼ 8.5 Hz, 2H, arom), 6.93 (s, 1H, arom), 6.99e7.02 (m, 3H,

arom), 7.06 (d, J ¼ 8.5 Hz, 2H, arom), 7.51 (s, 1H, arom), 8.39 (br, 1H,

(0.44 g, yield 62%) as orange oil. ¹H NMR:

d

1.03 (t, J ¼ 7.4 Hz, 3H,

CH₃), 1.76 (s, 3H, CH₃), 2.11 (q, J ¼ 7.4 Hz, 2H, CH₂), 3.47 (s, 6H, 2

OCH₃), 5.14 (s, 4H, 2 OCH₂), 6.91e6.95 (m, 4H, arom), 7.01e7.04 (m,

4H, arom).

OH), 8.45 (br, 1H, OH). ¹³C NMR (acetone-d₆):

d 17.9, 51.2, 115.6 (2C),

116.0 (2C), 119.6, 128.6, 129.8, 131.4 (2C), 131.4 (2C), 134.2, 134.5,

138.2, 142.8, 157.3, 157.4. HRMS

307.14410, found: 307.14464.

C

₁₉H₁₈N₂O₂[MþH] calcd:

4.1.3.2. 4,4'-(2-methylprop-1-ene-1,1-diyl)bis((methoxymethoxy)

benzene) (13b). Starting from 12b (0.41 g, 1.70 mmol), a crude

compound was obtained that was puriﬁed by ﬂash column chro-

matography (petroleum ether/ethyl acetate 8:2) to give 13b (0.35 g,

d 1.79 (s, 6H, 2 CH₃), 3.48 (s, 6H, 2

OCH₃), 5.16 (s, 4H, 2 OCH₂), 6.92e6.95 (m, 4H, arom), 7.02e7.04 (m,

4.1.6. 1-(Pyridine-3-yl)butan-2-one (8a). A solution of 2-(pyridine-

3-yl)acetic acid (0.58 g, 4.20 mmol), pyridine (2 mL) and propionic

anhydride (2 mL) was reﬂuxed under N₂for 6 h. Solvents were

evaporated and the residue was dissolved in dichlorometane and

washed with 2 M NaOH solution. The organic layer was dried and

concentrated in vacuum and the crude compound was puriﬁed by

ﬂash chromatography (ethyl acetate) to give 8a (0.14 g, yield 22%)

yield 62%) as an oil. ¹H NMR:

4H, arom).

4.1.4. General procedure for the synthesis of compound 14a,b

A mixture of 13a or 13b (1 eq), N-bromosuccunimide (NBS, 1 eq)

and a catalytic amount of benzoyl peroxide (BPO) in CCl₄(15 mL)

was reﬂuxed for 4 h. The mixture was hot ﬁltered and evaporated to

dryness. The resulting residue, without further puriﬁcation, was

dissolved in acetonitrile and imidazole (3 eq) was added. The re-

action mixture was reﬂuxed for 6 h under N₂atmosphere, the

solvent was evaporated, and the crude compound was puriﬁed by

ﬂash column chromatography with a suitable eluent.

like a yellow oil. ¹H NMR:

d

1.07 (t, J ¼ 6.8 Hz, 3H, CH₃), 2.53 (q,

J ¼ 6.8 Hz, 2H, CH₂), 3.71 (s, 2H, CH₂Py), 7.22e7.30 (m, 1H, arom),

7.55 (dd, J ¼ 1.7 and 7.9 Hz, 1H, arom), 8.45 (d, J ¼ 1.8 Hz, 1H, arom),

8.52 (d, J ¼ 7.6 Hz, 1H, arom).

4.1.7. General procedure for the synthesis of compounds 15-16a,b

A solution of the selected aldehyde (1 eq), piperidine (1.8 eq)

and acetic acid (1.6 eq) in acetone or methylethylketone was stirred

at room temperature until TLC indicated complete conversion of

the starting material. The solvent was evaporated to dryness and

the obtained crude compound was puriﬁed by ﬂash column chro-

matography with a suitable eluent.

4.1.4.1. 1-(2-(bis(4-(methoxymethoxy)phenyl)methylene)butyl)-1H-

imidazole (14a). Starting from 13a (0.44 g, 1.30 mmol), a crude

compound was obtained that was puriﬁed by ﬂash column

15

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

4.1.7.1. (E)-1-(pyridin-3-yl)pent-1-en-3-one (15a). Starting from 3-

pyridinecarboxyaldehyde (1,00 g, 0.88 mL, 9.30 mmol) in methyl-

ethylketone (20 mL), a crude was obtained that was puriﬁed by

ﬂash column chromatography (toluene/acetone 6:4) to give 15a

CH₃), 2.45 (q, J ¼ 7.4 Hz, 2H, CH₂), 2.82 (t, J ¼ 6.0 Hz, 2H, CH₂), 2.89

(t, J ¼ 6.0 Hz, 2H, CH₂), 4.17 (br, 1H, NH), 6.82 (s, 1H, arom), 7.66 (s,

1H, arom).

(0.48 g, yield 32%) as brown oil. ¹H NMR:

d

1.18 (t, J ¼ 7.2 Hz, 3H,

4.1.8.4. 4-(1H-imidazole-4-yl)butan-2-one (11b). Starting from 16b

(0.18 g, 1.30 mmol), a crude compound was obtained that was used

without further puriﬁcation (0.12 g, 66%) as brown oil. ¹H NMR:

CH₃), 2.72 (q, J ¼ 7.2 Hz, 2H, CH₂), 6.81 (d, J ¼ 16.4 Hz, 1H, CH]C),

7.34e7.37 (m, 1H, arom), 7.55 (d, J ¼ 16.4 Hz, 1H, CH]C), 7.88 (d,

J ¼ 8.0 Hz, 1H, arom), 8.62e8.64 (m, 1H, arom), 8.78 (s, 1H, arom).

d

2.16 (s, 3H, CH₃), 2.78e2.86 (m, 4H, 2 CH₂), 6.80 (s, 1H, arom), 7.65

(s, 1H, arom).

4.1.7.2. (E)-4-(pyridin-3-yl)but-3-en-2-one (15b). Starting from 3-

pyridinecarboxyaldehyde (0.50 g, 0.44 mL, 4.70 mmol) in acetone

(8 mL), a crude was obtained that was puriﬁed by ﬂash column

chromatography (petroleum ether/ethyl acetate 1:9, 0.5:9.5, ethyl

4.1.9. 1-(1H-imidazole-4-yl)ethan-1-one (10b).

4(5)-cyanoimidazole [41] (0.48 g, 5.20 mmol) in THF (5 mL) was

added to a 3 M solution of methylmagnesium bromide (8.6 mL,

26 mmol) and the mixture was stirred at rt for 3 h. Then, H₂O and

10% H₂SO₄solution were added dropwise and the mixture was

stirred for 30 min and then brought to pH 8 with 30% NaOH solu-

tion. After the organic layer was separated, the aqueous layer was

extracted with ethyl acetate. The organic layers were combined,

washed with NaHCO₃saturated solution and brine, and concen-

trated under reduced pressure. The crystals formed were collected

by ﬁltration to give 10b (0.30 g, yield 53%) mp 179e180 ^ꢂC (lit [42].

acetate) to give 15b (0.38 g, yield 55%) as brown oil. ¹H NMR:

d 2.38

(s, 3H, CH₃), 6.76 (d, J ¼ 16.4 Hz, 1H, CH]C), 7.30e7.46 (m, 1H,

arom), 7.48 (d, J ¼ 16.4 Hz, 1H, CH]C), 7.84 (d, J ¼ 8.0 Hz, 1H, arom),

8.59 (d, J ¼ 4.4 Hz, 1H, arom), 8.74 (s, 1H, arom).

4.1.7.3. (E)-1-(1H-imidazole-4-yl)pent-1-en-3-one

Starting from 1H-imidazole-4-carboxyaldehyde

(16a).

(0.50 g,

5.20 mmol) in methylethylketone (10 mL), a crude compound was

obtained that was puriﬁed by ﬂash column chromatography

(dichlorometane/methanol 8:2) to give 16a (0.70 g, yield 89%) as

mp 172 ^ꢂC). ¹H NMR:

1H, arom).

d 2.52 (s, 3H, CH₃), 7.76 (s, 1H, arom), 7.80 (s,

brown oil. ¹H NMR (methanol-d₄):

(q, J ¼ 7.2 Hz, 2H, CH₂), 6.88 (d, J ¼ 16.8 Hz, 1H, CH]C), 7.53 (d,

J ¼ 16.0 Hz, 1H, CH]C), 7.92 (s, 1H, arom), 9.04 (s, 1H, arom).

d

1.14 (t, J ¼ 7.2 Hz, 3H, CH₃), 2.75

4.2. Biological assays

4.1.7.4. (E)-4-(1H-imidazole-4-yl)but-3-en-2-one

Starting from 1H-imidazole-4-carboxyaldehyde

5.20 mmol) in acetone (10 mL), a crude compound was obtained

(16b).

(0.50 g,

4.2.1. Aromatase inhibition assays

Inhibition of HA was quantiﬁed by the Aromatase Inhibitor

Screening Kit (BioVision Inc., San Francisco, USA), monitoring the

conversion of a ﬂuorogenic substrate into a highly ﬂuorescent

metabolite as catalyzed by the AR enzyme. After the reconstitution

of the reagents, a standard curve was generated by diluting the

ﬂuorescent standard. Test compounds were dissolved in DMSO at a

ﬁnal concentration of ꢃ0.25% (v/v), after having veriﬁed that such

concentration of solvent does not signiﬁcantly affect the activity of

the enzyme. Each mother solution was diluted in the assay buffer to

generate a range of concentrations instrumental to build a multi-

point dose-response curve. The reaction was prepared by adding

Aromatase mix (containing Recombinant Human Aromatase (2X),

the aromatase assay buffer as well as NADPH-generating system) to

test compounds, inhibitor control, background control and positive

that was puriﬁed by ﬂash column chromatography (acetone þ 0.5%

NH₃) to give 16b (0.24 g, yield 34%) as brown oil. ¹H NMR:

d 2.31 (s,

3H, CH₃), 6.75 (d, J ¼ 16.0 Hz, 1H, CH]C), 7.34 (s, 1H, arom), 7.47 (d,

J ¼ 16.0 Hz, 1H, CH]C), 7.72 (s, 1H, arom).

4.1.8. General procedure for the synthesis of compounds 9a,b and

11a,b

A mixture of 15a,b or 16a,b (1 eq) and Zn (5 eq) in acetic acid

(10e15 mL) was heated to reﬂux for 3 h. The reaction mixture was

poured in water, basiﬁed with 2 M NaOH solution and extracted

with dichlorometane. The organic phase was washed with NaHCO₃

saturated solution, dried over anhydrous Na₂SO₄and concentrated

in vacuum. The crude compound, when necessary, was puriﬁed by

ﬂash column chromatography with a suitable eluent.

control (1 m

M LTZ). The reaction mixture was preincubated at 37 ^ꢂC

for 10 min in order to allow the tested inhibitors to interact with the

enzyme. Subsequently, the reaction initiated after the addition of

4.1.8.1. 1-(Pyridin-3-yl)pentan-3-one (9a). Starting from 15a

(0.48 g, 2.80 mmol), a crude compound was obtained and was

puriﬁed by column chromatography (toluene/acetone 7:3) to give

30

aromatase substrate and

done in 96-well microtiter plates (Corning Incorporated, Corning,

ME, USA) in a ﬁnal reaction volume of 100 L/well. The ﬂuorescence

of the sample was measured using a TECAN Ultra microplate reader

(Tecan Trading AG, Switzerland) at dual wavelengths of 488/

527 nm for 60 min. Results were expressed as relative ﬂuorescence

units (RFUs). Experiments were carried out in triplicate and the

average values were used to construct the dose-response curves.

The percentage of inhibition was calculated as the ratio between

the RFUs of control and test compound, according to the manu-

facturer's instructions. Tests were done at different concentrations

m

L of aromatase substrate/NADP þ mixture (containing buffer,

b

-NADP^þ100X stock). The assays were

9a (0.29 g, yield 62%) as brown oil. ¹H NMR:

d

1.04 (t, J ¼ 7.4 Hz, 3H,

m

CH₃), 2.41 (q, J ¼ 7.4 Hz, 2H, CH₂), 2.74 (t, J ¼ 7.4 Hz, 2H, CH₂), 2.90 (t,

J ¼ 7.4 Hz, 2H, CH₂), 7.18e7.22 (m, 1H, arom), 7.51 (d, J ¼ 8.0 Hz, 1H,

arom), 8.44e8.46 (m, 2H, arom).

4.1.8.2. 4-(Pyridin-3-yl)butan-2-one (9b). Starting from 15b (0.41 g,

2.80 mmol), a crude compound was obtained that was used

without further puriﬁcation (0.18 g, yield 43%) as brown oil. ¹H

NMR:

d

2.15 (s, 3H, CH₃), 2.79 (t, J ¼ 7.2 Hz, 2H, CH₂), 2.92 (t,

J ¼ 7.4 Hz, 2H, CH₂), 7.27e7.29 (m, 1H, arom), 7.61 (d, J ¼ 7.6 Hz, 1H,

arom), 8.46e8.48 (m, 2H, arom).

(10, 1, 0.1, 0.01 0.001 and 0.0001

mM). For each compound, the

concentration able to inhibit enzymatic activity by 50% (IC₅₀) was

calculated by nonlinear regression of the experimental data to a

tetraparametric logistic curve (SigmaPlot 13.0 - Systat Software

Inc.). In order to test the sensitivity of the kit and assess the quality

of the measurements the IC₅₀of letrozole and END were also tested

with the same procedure.

4.1.8.3. 1-(1H-imidazole-4-yl)pentan-3-one (11a). Starting from 16a

(0.35 g, 2.30 mmol), a crude compound was obtained and was pu-

riﬁed by column chromatograph. y (dichlorometane/methanol 9:1)

to give 11a (0.15 g, 43%) as brown oil. ¹H NMR: 1.06 (t, J ¼ 7.4 Hz, 3H,

16

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

4.2.2. Estrogen receptor binding assays

4.2.5. Progesteron receptor (PGR) mRNA expression analysis

The PolarScreen ER

a

Competitor Assay Kit Green (Cat.No.15883;

MCF-7 cells were seeded at a density of 10⁵cells/well in a six

well plate and preconditioned with minimum essential media

(MEM) supplemented with 10% charcoal-stripped FBS 72 h before

treatment, to remove the estrogens from the culture medium

containing 10% FBS. Cells were treated with 10 nM estradiol (E2)

Invitrogen) was used to determine the relative afﬁnity of tested

compounds, of the reference SERM END and of letrozole (as a

negative control) towards ER

protocol. Brieﬂy, serial dilutions of each test compound

(10e0.001 M) were prepared at 100X in 100% DMSO in a DMSO

a, according to the manufacturer's

m

alone, or in combination with i) 1

mM endoxifen, used as positive

tolerant plate. The compounds were further diluted 50-fold in

control, ii) 3a 10 or 20 M, iii) 6a 10 or 20 m

m

M. After 24 h from

screening buffer to get a 2X ﬁnal concentration and 2% DMSO. Ten

treatment, cells treated with E2 in combination with the tested

compounds or endoxifen were harvested for RNA extraction using

PureLink™ RNA Mini Kit (Thermo Fisher Scientiﬁc, Invitrogen,

Waltham, MA, USA), according to manufacturer's instructions.

m

L of the 2X compound concentration was added to a 10

containing 2X full length ER and Fluormone™ ES2 Green

(75 nM ER ; 4.5 nM Fluormone™ ES2 Green) in a black low volume

384-well assay plate (Corning NY, US) to obtain 20 L ﬁnal volume.

The negative control contained 10 L of screening buffer and 10

of ER /Fluormone™ ES2 Green Complex. This control was used to

determine the polarization value when no competitor was present

(theoretical maximum polarization). Ten L of ER /Fluormone™

ES2 Green Complex and 10 L of 20 -estradiol (endogenous

ligand) (ﬁnal concentration in well of 10 M) were used as a posi-

mL mixture

a

m

Brieﬂy, cells were lysed using Lysis Buffer supplemented with 1% b-

m

mL

mercaptoethanol. Lysed samples were added to an equal volume of

ethanol and mixed thoroughly. The solution was passed through a

ﬁlter cartridge containing a silica-based membrane to which RNA

binds. The membranes were then washed with Wash Buffer I and

two times with Wash Buffer II. Total RNA was ﬁnally eluted with

RNase-free water and stored at ꢀ20 ^ꢂC. RNA was quantiﬁed using

a

m

a

m

mM b

m

€

tive control (minimum polarization value). The assay plate was

then incubated in the dark at room temperature for 2 h. Polariza-

tion values were measured on a Tecan Spark Plate Reader with

485 nm excitation and 530 nm emission interference ﬁlters with

bandwidths of 25/20 nm. All measurements were done in triplicate.

The polarization values were converted to percent inhibition using

the equation:

the Nanoquant plate (Tecan, Mannedorf, Switzerland) using the

multiplate reader Inﬁnite 200 PRO (Tecan). Complementary DNA

(cDNA) was synthetized using the High-Capacity cDNA Reverse

Transcription kit (Life Technologies, Carlsbald, CA, USA). Brieﬂy,

200 ng of RNA in a total volume of 10

mL in RNase free water were

added to 10 L of master mix prepared according to manufacturer's

m

instructions with RNase inhibitor and subjected to appropriate

thermocycling conditions using MiniAmp thermal cycler (Thermo

Fisher Scientiﬁc). Finally, relative quantiﬁcation was performed by

real time PCR (BIORAD CFX Connect) using Universal Master Mix

(Themo Fisher Scientiﬁc) and Taqman gene expression assay

(Thermo Fisher Scientiﬁc) for PGR gene (FAM, Hs01556702_m1)

and glyceraldehyde 3-phosphate dehydrogenase gene (GADPH,

FAM, Hs99999905_m1), used as endogenous control to normalize

each sample. Each measurement was performed in duplicate for a

total of 40 ampliﬁcation cycles. Quantitative values of ampliﬁcation

were obtained from the threshold cycle (Ct), deﬁned as the number

of cycles required for the ﬂuorescent signal to exceed the back-

ground level. The relative mRNA expression was analyzed using the

I% ¼ (P₀eP)/(P₀eP₁₀₀) ꢁ 100

where P₀is the polarization value at 0% inhibition, P₁₀₀is the po-

larization value at 100% inhibition, and P is the observed FP at each

concentration point. We used free

b-estradiol (100% inhibition) as a

positive control and ER /Fluormone™ ES2 Green Complex (0% in-

a

hibition) as a negative control. We transformed polarization values

into percent inhibition to normalize the differences at 0% inhibition

for each run.

We then analyzed the percent inhibition versus competitor

concentration curves by nonlinear least-squares curve ﬁtting and

determined the concentration of competitor needed to displace

half of the bound ligand (IC₅₀). IC₅₀values were calculated using

GraphPad Prism software.

DD

2^ꢀ^Ctmethod (Livak KJ, Schmittgen TD (2001). Analysis of relative

gene expression data using real-time quantitative PCR and the 2(-

Delta Delta C(T)) Method. Methods, 25(4): 402e8) and untreated

MCF-7 cells were used as calibrator of the experiment. The results

were expressed as percentage of antiestrogenic effects compared to

E2-stimulated mRNA expression of PGR, considered as 100%.

4.2.3. Cell lines and culture conditions

MCF-7 and MCF10A cell lines were obtained from American

Type Culture Collection (ATCC). MCF-7 cells were grown in

RPMI þ10% FBS, while MCFA10 in MEBM medium (Lonza) supple-

mented with 10 ng/mL EGF, 10 ng/mL FGF, 600 U/L eparin and B-27

Supplement (Thermo Fisher Scientiﬁc). Cells were tested for the

absence of Mycoplasma fortnightly and maintained in logarithmic

growth phase as a monolayer in a humidiﬁed 5% CO2 atmosphere at

37 ^ꢂC.

4.2.6. Protein extraction and Western blot analysis

MCF-7 cells were incubated with vehicle (DMSO), test com-

pounds (IC₅₀calculated by ERa binding assay) or fulvestrant (1 mM)

for 72 h. Then, cells were harvested using a scraper and lysed with

RIPA buffer (Thermo Scientiﬁc Pierce, #89900). Protein samples

were sonicated for 20 s and quantiﬁed through the BCA assay

method (Thermo Fisher Scientiﬁc, Waltham, MA, USA). Total

cellular lysates were separated on a 4e12% NuPAGE bis-tris gel

(Thermo Fisher Scientiﬁc, Waltham, MA, USA) and transferred to

nitrocellulose using standard protocols. Filters were blocked in

PBS-Tween 20 with 5% skim milk and incubated overnight with

4.2.4. Cell proliferation assays

Cells were seeded in triplicate in 12-well plates (50.000 cells/

well, 70.000 cells/well, 90.000 cells/well for MDA-MB-231 cells and

MCF-7 cells, and MCF10A, respectively). After 24 h cells were

exposed to increasing concentrations of compounds 1a-6b and of

the two reference compounds letrozole and END for 72 h. Culture

medium was then removed and adherent cells were harvested

using trypsin and counted with a cell counter (Beckman Coulter,

S.p.A., Milan, Italy). All experiments were performed in triplicate.

The IC₅₀value was deﬁned as the concentration of a drug inhibiting

cell proliferation by 50%.

primary antibodies speciﬁc for ER

a (ab16660; Abcam, Cambridge,

UK). -actin (ab8226; Abcam, Cambridge, UK) antibody was used to

b

conﬁrm equal protein loading on the gel. Band density was quan-

tiﬁed by scanning ﬁlms and processing image intensities with the

ImageJ 1.47v Software.

17

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

4.3. Computational details

set (GPW) [65]. In particular, we employed a double z (MOLOPT)

basis set, an auxiliary PW basis set with a density cutoff of 400 Ry

and Goedecker-Teter-Hutter (GTH) pseudopotentials [66,67]. This

protocol has been used successfully in many QM/MM MD simula-

tions of biomolecules, including aromatase [45,68,69]. The dangling

bonds between the quantum and classical regions were saturated

using capping hydrogens atoms. An integration step of 0.5 fs was

used in all the QM/MM MD simulations in the NVT ensemble. All

systems were initially optimized and equilibrated at 300 K without

4.3.1. Docking calculations

Docking simulations of compounds 1a-6b on aromatase enzyme

were performed with Glide [43] using the single precision (SP)

protocol on the AR crystal structure (PDB code 3S79). A van der

Waals (vdW) radius scaling factor of 0.80 Å for protein and ligands

atoms having a partial charge less than 0.15 was used to mimic

protein ﬂexibility. A metal constraint was incorporated in order to

mimic a binding pose in which the nitrogen atom was at coordi-

nation distance from the heme iron atom, as in previous study [35].

Possible tautomerism of the imidazole ring of 2a-b and 6a-b have

been explicitly considered.

ꢁ

constraints for 2 ps, using a Nose-Hoover thermostat [70], followed

by other 8 ps after the stable formation of the coordination bond

has occurred. The heme iron atom was simulated with a doublet

spin state, consistently with other studies [35]. The MM region was

described with the same force ﬁeld of the classical MD simulations.

In order to predict the binding mode of the newly synthesized

compound to ER

END obtained from MD simulations of our previous study [6]. The

1a-6b compounds were docked in the LBD cavity of ER with the

Glide software using the same protocol detailed above for AR.

Validation of the protocol has been performed reproducing the

binding poses of endoxifen and exemestane, as extracted from their

a we used the structure of ERa in complex with

4.3.4. Classical MD simulations of ER

After docking the inhibitor to the LBD of ER

antagonist adducts in its dimeric form was built on the crystal

structures of ER the LDB dimer in the antagonist state by super-

imposing this monomeric antagonist ER s structures, on the both

monomers of the antagonist ER crystal structure (PDB id: 1qku),

a

monomer. The

a

crystallographic structures bound to ER

respectively (See Table S1).

a [44] and AR [10],

a

following an established computational protocol [6]. Nonetheless,

we veriﬁed the stability of the built dimers for each complex 500 ns

of MD simulations. Each model of the drug/ERa was solvated with

4.3.2. Classical MD simulations of AR

We performed classical MD simulations on selected models of

AR in complex with the docked inhibitors, using a protocol adopted

in previous studies [45]. Namely, the protonation states under

physiological conditions were calculated with the webserver Hþþ

[46], while Asp309 was considered its neutral form, as reported in

literature [47,48]. We have used Parm99SB AMBER force ﬁeld for

the protein [49,50], the Shahrokh et al. parameters for the heme

moiety and Cys437 [51] and the general Amber FF (GAFF) for the

inhibitors [52]. The ESP charges [53] were derived by performing

geometry optimization of the inhibitors at Hartree-Fock level of

theory with a 6-31G* basis set using the Gaussian 09 software and

were subsequently transformed in RESP charges by Antechamber

module of Ambertools 18 [54]. Our models were solvated along the

x and y direction with the TIP3P waters model [55] leading to a total

of 66342 atoms. The systems topology was built with Ambertools

18, and later converted to a GROMACS format with the software

acpype.

The GROMACS 2018.2 [56] code was ﬁnally used to perform 100

ns MD simulations using an integration time step of 2 fs and all

covalent bonds involving hydrogen atoms constrained with the

LINCS algorithm. MD simulations were done in the NPT ensemble,

at a temperature of 300 K, using a velocity-rescaling thermostat

[57]. An initial energy minimization step was done with the

steepest descend algorithm. A preliminary equilibration of the

model was performed for 20 ns with the protein and ligand atoms

TIP3P water molecules in the rectangular shaped box with the di-

mensions of at least 12 Å in each direction from the solute, and

charge neutrality was achieved by the addition of Na^þions [71],

leading to a total of 60000 atoms. In all simulations we used the

same protocol detailed above.

4.3.5. Analysis

Cluster analysis and root mean square deviation (RMSD) of the

MD trajectories were carried out using g_cluster tool, based on the

Daura et al. algorithm [72] and g_rmsd gmx_rmsf modules of the

GROMACS 2018.2 package. The cpptraj module of Amber program

was used for hydrogen bond analysis (cut-off parameters for H-

bond was 3.3 Å and 35^ꢂ), cross-correlation matrices. As in previous

studies we have introduced a simpliﬁed form of the cross correla-

tion matrix in which the cross correlation coefﬁcients are summed

by structural elements [6,73].

The Amber 18 tool MM_PBSA.py [74] was used to perform Mo-

lecular Mechanics Generalized Born Surface Area (MM-GBSA) free

energy calculation, taking 500 frames from an equilibrated part of

the MD trajectories. An improved generalized born solvation model

was employed (igb ¼ 8) [75], and a salt concentration of 0.1 M. The

conformational entropic contribution of the free energy was not

considered, as it was suggested that this term does not improve the

quality of the results using the MM-G(P)BSA [76,77]. Visualization

of the MD trajectories and images were done using the VMD pro-

gram [78] and UCSF Chimera [79].

harmonically

restrained

using

a

force

constant

of

1000 kJ mol^ꢀ1nm^ꢀ2. Next, constraints were released, and the

model was thermalized to the target temperature of 300 K within

10 ns. In order to relax the AR structure to the presence of the ligand

and monitor the stability of the docking poses each system un-

derwent by 100 ns-long MD simulation, while keeping restrained

the distance of the ligand to the Fe of the heme moiety.

5. Supplementary data

Figs. S1eS6, Tables S1eS9, ligand parameters, synthetic

Schemes S1eS3 and representative NMR spectra of studied com-

pounds (3a).

4.3.3. QM/MM molecular dynamics simulations of AR

Due to the difﬁculties in describing coordination bonds with

classical FF [58,59], QM/MM Born Oppenheimer MD simulations

were performed using CP2K 6.1 program [60e62]. The QM layer,

including the heme group and the inhibitors (90 atoms), was

simulated in a cubic box with sides of 24 Å and it was described at

Density Functional Theory (DFT) using the BLYP [63,64] exchanges

correlation functional with a dual Gaussian-type/Plane Waves basis

Declaration of competing interest

The authors declare that they have no known competing

ﬁnancial interests or personal relationships that could have

appeared to inﬂuence the work reported in this paper.

18

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

ꢂ

[22] M. Pavlin, L. Gelsomino, I. Barone, A. Spinello, S. Catalano, S. Ando,

Acknowledgments

A. Magistrato, Structural, thermodynamic, and kinetic traits of antiestrogen-

compounds selectively targeting the Y537S mutant estrogen receptor

a

AM thanks the Italian Association for Cancer Research (AIRC,

MFAG Grant No 17134). AS was supported by a FIRC-AIRC ‘Mario e

Valeria Rindi’ fellowship for Italy.

transcriptional activity in breast cancer cell lines, Front. Chem. 7 (2019) 602.

[23] E.P. Hamilton, M.R. Patel, A.C. Armstrong, R.D. Baird, K. Jhaveri, M. Hoch,

T. Klinowska, J.P.O. Lindemann, S.R. Morgan, G. Schiavon, H.M. Weir, S.A. Im,

A ﬁ rst-in-human study of the new oral selective estrogen receptor degrader

AZD9496 for ER, Clin. Canc. Res.. 24 (15) (2018) 3510e3518.

[24] D. Jelovac, L. Macedo, O.G. Goloubeva, V. Handratta, A.M. Brodie, Additive

antitumor effect of aromatase inhibitor letrozole and antiestrogen fulvestrant

Appendix A. Supplementary data

in

a postmenopausal breast cancer model, Canc. Res.. 65 (12) (2005)

Supplementary data to this article can be found online at

https://doi.org/10.1016/j.ejmech.2021.1 13733.

5439e5444.

[25] M. Baum, A. Buzdar, J. Cuzick, J. Forbes, J. Houghton, A. Howell, T. Sahmoud,

ATAC (Arimidex, Tamoxifen Alone or in Combination) Trialists ’ Group, Anas-

trozole alone or in combination with tamoxifen versus tamoxifen alone for

adjuvant treatment of postmenopausal women with early-stage breast can-

cer: results of the ATAC (Arimidex, Tamoxifen Alone or in Combination) trial

ef ﬁ cacy and safety update analyses, Cancer 98 (9) (2003) 1802e1810.

[26] A. Howell, J. Cuzick, M. Baum, A. Buzdar, M. Dowsett, J.F. Forbes, G. Hoctin-

Boes, J. Houghton, G.Y. Locker, J.S. Tobias, A.T. Group, Results of the ATAC

(Arimidex, Tamoxifen, Alone or in Combination) trial after completion of 5

years' adjuvant treatment for breast cancer, Lancet 365 (9453) (2005) 60e62.

[27] W.J. Lu, Z. Desta, D.A. Flockhart, Tamoxifen metabolites as active inhibitors of

aromatase in the treatment of breast cancer, Breast Canc. Res. Treat. 131 (2)

(2012) 473e481.

References

[1] American Cancer Society, Cancer Facts & Figures 2020, 2020. Atlanta.

[2] American Cancer Society, Breast Cancer Facts & Figures 2019-2020, 2019.

Atlanta.

[3] Y. Feng, M. Spezia, S. Huang, C. Yuan, Z. Zeng, L. Zhang, X. Ji, W. Liu, B. Huang,

W. Luo, B. Liu, Y. Lei, S. Du, A. Vuppalapati, H.H. Luu, R.C. Haydon, T.C. He,

G. Ren, Breast cancer development and progression: risk factors, cancer stem

cells, signaling pathways, genomics, and molecular pathogenesis, Genes Dis. 5

(2) (2018) 77e106.

[28] W.J. Lu, C. Xu, Z. Pei, A.S. Mayhoub, M. Cushman, D.A. Flockhart, The tamoxifen

metabolite norendoxifen is a potent and selective inhibitor of aromatase

(CYP19) and a potential lead compound for novel therapeutic agents, Breast

Canc. Res. Treat. 133 (1) (2012) 99e109.

[4] A. Spinello, I. Ritacco, A. Magistrato, The catalytic mechanism of steroidogenic

cytochromes P450 from all-atom simulations: entwinement with membrane

environment, redox partners, and post-transcriptional regulation, Catalysts 9

(1) (2019).

[29] W. Lv, J. Liu, T.C. Skaar, E. O'Neill, G. Yu, D.A. Flockhart, M. Cushman, Synthesis

of triphenylethylene bisphenols as aromatase inhibitors that also modulate

estrogen receptors, J. Med. Chem. 59 (1) (2016) 157e170.

[30] V. Lubczyk, H. Bachmann, R. Gust, Investigations on estrogen receptor binding.

The estrogenic, antiestrogenic, and cytotoxic properties of C2-alkyl-

substituted 1,1-bis(4-hydroxyphenyl)-2-phenylethenes, J. Med. Chem. 45 (24)

(2002) 5358e5364.

[31] M. Recanatini, A. Bisi, A. Cavalli, F. Belluti, S. Gobbi, A. Rampa, P. Valenti,

M. Palzer, A. Palusczak, R.W. Hartmann, A new class of nonsteroidal aromatase

inhibitors: design and synthesis of chromone and xanthone derivatives and

inhibition of the P450 enzymes aromatase and 17 alpha-hydroxylase/C17,20-

lyase, J. Med. Chem. 44 (5) (2001) 672e680.

[32] S. Gobbi, A. Cavalli, M. Negri, K.E. Schewe, F. Belluti, L. Piazzi, R.W. Hartmann,

M. Recanatini, A. Bisi, Imidazolylmethylbenzophenones as highly potent

aromatase inhibitors, J. Med. Chem. 50 (15) (2007) 3420e3422.

[33] S. Gobbi, C. Zimmer, F. Belluti, A. Rampa, R.W. Hartmann, M. Recanatini,

A. Bisi, Novel highly potent and selective nonsteroidal aromatase inhibitors:

synthesis, biological evaluation and structure-activity relationships investi-

gation, J. Med. Chem. 53 (14) (2010) 5347e5351.

[34] E. Catanzaro, F. Seghetti, C. Calcabrini, A. Rampa, S. Gobbi, P. Sestili, E. Turrini,

F. Maffei, P. Hrelia, A. Bisi, F. Belluti, C. Fimognari, Identi ﬁ cation of a new

tamoxifen-xanthene hybrid as pro-apoptotic anticancer agent, Bioorg. Chem.

86 (2019) 538e549.

[35] J. Caciolla, A. Spinello, S. Martini, A. Bisi, N. Zaffaroni, S. Gobbi, A. Magistrato,

Targeting orthosteric and allosteric pockets of aromatase via dual-mode novel

azole inhibitors, ACS Med. Chem. Lett. 11 (5) (2020) 732e739.

[36] Y.C. Lim, Z. Desta, D.A. Flockhart, T.C. Skaar, Endoxifen (4-hydroxy-N-des-

methyl-tamoxifen) has anti-estrogenic effects in breast cancer cells with

(2005) 471e478.

[5] N. Heldring, A. Pike, S. Andersson, J. Matthews, G. Cheng, J. Hartman,

€

M. Tujague, A. Strom, E. Treuter, M. Warner, J.A. Gustafsson, Estrogen re-

ceptors: how do they signal and what are their targets, Physiol. Rev. 87 (3)

(2007) 905e931.

[6] M. Pavlin, A. Spinello, M. Pennati, N. Zaffaroni, S. Gobbi, A. Bisi, G. Colombo,

A. Magistrato, A computational assay of estrogen receptor

a antagonists re-

veals the key common structural traits of drugs effectively ﬁ ghting refractory

breast cancers, Sci. Rep. 8 (1) (2018) 649.

[7] H.K. Patel, T. Bihani, Selective estrogen receptor modulators (SERMs) and

selective estrogen receptor degraders (SERDs) in cancer treatment, Pharmacol.

Ther. 186 (2018) 1e24.

[8] M.D.C. Barros-Oliveira, D.R. Costa-Silva, D.B. Andrade, U.S. Borges, C.B. Tavares,

R.S. Borges, J.M. Silva, B.B.D. Silva, Use of anastrozole in the chemoprevention

and treatment of breast cancer: a literature review, Rev. Assoc. Med. Bras. 63

(4) (1992) 371e378.

[9] K. Van Asten, P. Neven, A. Lintermans, H. Wildiers, R. Paridaens, Aromatase

inhibitors in the breast cancer clinic: focus on exemestane, Endocr. Relat.

Canc. 21 (1) (2014) R31eR49.

[10] D. Ghosh, J. Lo, D. Morton, D. Valette, J. Xi, J. Griswold, S. Hubbell, C. Egbuta,

W. Jiang, J. An, H.M. Davies, Novel aromatase inhibitors by structure-guided

design, J. Med. Chem. 55 (19) (2012) 8464e8476.

[11] C. Egbuta, J. Lo, D. Ghosh, Mechanism of inhibition of estrogen biosynthesis by

azole fungicides, Endocrinology 155 (12) (2014) 4622e4628.

[12] A. Spinello, S. Martini, F. Berti, M. Pennati, M. Pavlin, J. Sgrignani, G. Grazioso,

G. Colombo, N. Zaffaroni, A. Magistrato, Rational design of allosteric modu-

lators of the aromatase enzyme: an unprecedented therapeutic strategy to

ﬁ ght breast cancer, Eur. J. Med. Chem. 168 (2019) 253e262.

[13] P. Hadji, M.S. Aapro, J.J. Body, M. Gnant, M.L. Brandi, J.Y. Reginster,

M.C. Zillikens, C.C. Glüer, T. de Villiers, R. Baber, G.D. Roodman, C. Cooper,

B. Langdahl, S. Palacios, J. Kanis, N. Al-Daghri, X. Nogues, E.F. Eriksen, A. Kurth,

R. Rizzoli, R.E. Coleman, Management of Aromatase Inhibitor-Associated Bone

Loss (AIBL) in postmenopausal women with hormone sensitive breast cancer:

joint position statement of the IOF, CABS, ECTS, IEG, ESCEO IMS, and SIOG,

J. Bone Oncol. 7 (2017) 1e12.

[14] J. Foglietta, A. Inno, F. de Iuliis, V. Sini, S. Duranti, M. Turazza, L. Tarantini,

S. Gori, Cardiotoxicity of aromatase inhibitors in breast cancer patients, Clin.

Breast Canc. 17 (1) (2017) 11e17.

[15] M.S. Ewer, S. Glück, A woman's heart: the impact of adjuvant endocrine

therapy on cardiovascular health, Cancer 115 (9) (2009) 1813e1826.

[16] J.S. Lewis-Wambi, V.C. Jordan, Treatment of postmenopausal breast cancer

with selective estrogen receptor modulators (SERMs), Breast Dis. 24 (2005)

93e105.

[37] M.D. Abramoff, P.J. Magelhaes, S.J. Ram, Image processing with ImageJ, Bio-

phot. Int. 11 (7) (2004).

ꢀ

[38] P. Janos, A. Spinello, A. Magistrato, All-atom simulations to studying metal-

lodrugs/target interactions, Curr. Opin. Chem. Biol. 61 (2020) 1e8.

[39] I. Massova, P. Kollman, Combined molecular mechanical and continuum sol-

vent approach (MM-PBSA/GBSA) to predict ligand binding, Perspect. Drug

Discov. Des. 18 (2000) 113e135.

[40] J.F. Miquel, K. Olsson, B. Sundbeck, Synthesis of unsymmetrical dipheny-

lalkenes, J. Med. Chem. 6 (1963) 774e780.

[41] J. Kawakami, K. Kimura, M. Yamaoka, A convenient synthesis of 4(5)-alky-

lacyl-1H-imidazoles from 4(5)-imidazolecarboxaldehyde, Synthesis-Stuttgart

(5) (2003) 677e680.

[42] S. Iwasaki, Photochemistry of imidazolides. I. The photo-Fries-type rear-

rangement of N-substituted imidazoles, Helv. Chim. Acta 59 (8) (1976)

2738e2752.

[17] S. Martinkovich, D. Shah, S.L. Planey, J.A. Arnott, Selective estrogen receptor

modulators: tissue speci ﬁ city and clinical utility, Clin. Interv. Aging 9 (2014)

1437e1452.

[43] R.A. Friesner, J.L. Banks, R.B. Murphy, T.A. Halgren, J.J. Klicic, D.T. Mainz,

M.P. Repasky, E.H. Knoll, M. Shelley, J.K. Perry, D.E. Shaw, P. Francis,

P.S. Shenkin, Glide: a new approach for rapid, accurate docking and scoring. 1.

Method and assessment of docking accuracy, J. Med. Chem. 47 (7) (2004)

1739e1749.

[18] A. Ring, M. Dowsett, Mechanisms of tamoxifen resistance, Endocr. Relat. Canc.

11 (4) (2004) 643e658.

[19] I. Shagufta Ahmad, Tamoxifen a pioneering drug: an update on the thera-

peutic potential of tamoxifen derivatives, Eur. J. Med. Chem. 143 (2018)

515e531.

[44] P.Y. Maximov, B. Abderrahman, S.W. Fanning, S. Sengupta, P. Fan, R.F. Curpan,

D.M.Q. Rincon, J.A. Greenland, S.S. Rajan, G.L. Greene, V.C. Jordan, Endoxifen,

4-hydroxytamoxifen and an estrogenic derivative modulate estrogen receptor

complex mediated apoptosis in breast cancer, Mol. Pharmacol. 94 (2) (2018)

812e822.

[20] W. Toy, H. Weir, P. Razavi, M. Lawson, A.U. Goeppert, A.M. Mazzola, A. Smith,

J. Wilson, C. Morrow, W.L. Wong, E. De Stanchina, K.E. Carlson, T.S. Martin,

S. Uddin, Z. Li, S. Fanning, J.A. Katzenellenbogen, G. Greene, J. Baselga,

S. Chandarlapaty, Activating, Canc. Discov. 7 (3) (2017) 277e287.

[21] J.T. Lei, X. Gou, S. Seker, M.J. Ellis, Alterations and metastasis in estrogen re-

ceptor positive breast cancer, J. Canc. Metastasis Treat 5 (2019).

[45] A. Spinello, M. Pavlin, L. Casalino, A. Magistrato,

A dehydrogenase dual

19

J. Caciolla, S. Martini, A. Spinello et al.

European Journal of Medicinal Chemistry 224 (2021) 113733

hydrogen abstraction mechanism promotes estrogen biosynthesis: can we

expand the functional annotation of the aromatase enzyme? Chemistry 24

(42) (2018) 10840e10849.

[62] T. Laino, F. Mohamed, A. Laio, M. Parrinello, An ef ﬁ cient linear-scaling elec-

trostatic coupling for treating periodic boundary conditions in QM/MM sim-

ulations, J. Chem. Theor. Comput. 2 (5) (2006) 1370e1378.

[63] A. Becke, Density-functional exchange-energy approximation with correct

asymptotic-behavior, Phys. Rev. 38 (6) (1988) 3098e3100.

[64] C. Lee, W. Yang, R. Parr, Development of the colle-salvetti correlation-energy

formula into A functional of the electron-density, Phys. Rev. B 37 (2) (1988)

785e789.

[65] J. VandeVondele, J. Hutter, Gaussian basis sets for accurate calculations on

molecular systems in gas and condensed phases, J. Chem. Phys. 127 (11)

(2007).

[66] S. Goedecker, M. Teter, J. Hutter, Separable dual-space Gaussian pseudopo-

tentials, Phys. Rev. B 54 (3) (1996) 1703e1710.

[46] R. Anandakrishnan, B. Aguilar, A.V. Onufriev, H þþ 3.0: automating pK pre-

diction and the preparation of biomolecular structures for atomistic molecular

modeling and simulations, Nucleic Acids Res. 40 (2012) W537eW541 (Web

Server issue).

[47] G. Di Nardo, M. Breitner, A. Bandino, D. Ghosh, G.K. Jennings, J.C. Hackett,

G. Gilardi, Evidence for an elevated aspartate pK(a) in the active site of human

aromatase, J. Biol. Chem. 290 (2) (2015) 1186e1196.

[48] J. Sgrignani, A. Magistrato, In ﬂ uence of the membrane lipophilic environment

on the structure and on the substrate access/egress routes of the human

aromatase enzyme. A computational study, J. Chem. Inf. Model. 52 (6) (2012)

1595e1606.

[67] C. Hartwigsen, S. Goedecker, J. Hutter, Relativistic separable dual-space

[49] L. Wickstrom, A. Okur, C. Simmerling, Evaluating the performance of the

ff99SB force ﬁ eld based on NMR scalar coupling data, Biophys. J. 97 (3) (2009)

853e856.

[50] K. Lindorff-Larsen, S. Piana, K. Palmo, P. Maragakis, J.L. Klepeis, R.O. Dror,

D.E. Shaw, Improved side-chain torsion potentials for the Amber ff99SB

protein force ﬁ eld, Proteins 78 (8) (2010) 1950e1958.

Gaussian pseudopotentials from H to Rn, Phys. Rev. B 58 (7) (1998)

3641e3662.

[68] J. Sgrignani, M. Iannuzzi, A. Magistrato, Role of water in the puzzling mech-

anism of the ﬁ nal aromatization step promoted by the human aromatase

enzyme. Insights from QM/MM MD simulations, J. Chem. Inf. Model. 55 (10)

(2015) 2218e2226.

[51] K. Shahrokh, A. Orendt, G.S. Yost, T.E. Cheatham, Quantum mechanically

derived AMBER-compatible heme parameters for various states of the cyto-

chrome P450 catalytic cycle, J. Comput. Chem. 33 (2) (2012) 119e133.

[52] J. Wang, R.M. Wolf, J.W. Caldwell, P.A. Kollman, D.A. Case, Development and

[69] I. Ritacco, A. Spinello, E. Ippoliti, A. Magistrato, Post-translational regulation of

CYP450s metabolism as revealed by all-atoms simulations of the aromatase

enzyme, J. Chem. Inf. Model. 59 (6) (2019) 2930e2940.

[70] S.A. Nose, Uni ﬁ ed formulation of the constant temperature molecular-

dynamics methods, J. Chem. Phys. 81 (1984) 511e519.

testing of

a general amber force ﬁ eld, J. Comput. Chem. 25 (9) (2004)

1157e1174.

[71] I. Joung, T. Cheatham, Determination of alkali and halide monovalent ion

parameters for use in explicitly solvated biomolecular simulations, J. Phys.

Chem. B 112 (30) (2008) 9020e9041.

[72] X. Daura, K. Gademann, B. Jaun, D. Seebach, W. van Gunsteren, A. Mark,

Peptide folding: when simulation meets experiment, Angew. Chem. Int. Ed. 38

(1e2) (1999) 236e240.

[73] L. Casalino, G. Palermo, A. Spinello, U. Rothlisberger, A. Magistrato, All-atom

simulations disentangle the functional dynamics underlying gene maturation

in the intron lariat spliceosome, Proc. Natl. Acad. Sci. U. S. A 115 (26) (2018)

6584e6589.

[53] C. Bayly, P. Cieplak, W. Cornell, P. Kollman, A well-behaved electrostatic po-

tential based method using charge restraints for deriving atomic charges - the

resp model, J. Phys. Chem. 97 (40) (1993) 10269e10280.

[54] J. Wang, W. Wang, P.A. Kollman, D.A. Case, Automatic atom type and bond

type perception in molecular mechanical calculations, J. Mol. Graph. Model. 25

(2) (2006) 247e260.

[55] W.L. Jorgensen, J. Chandrasekhar, J.D. Madura, R.W. Impey, M.L. Klein, Com-

parison of simple potential functions for simulating liquid water, J. Chem.

Phys. 79 (1983) 926e935.

[56] D. Van Der Spoel, E. Lindahl, B. Hess, G. Groenhof, A.E. Mark, H.J. Berendsen,

GROMACS: fast, ﬂ exible, and free, J. Comput. Chem. 26 (16) (2005)

1701e1718.

[74] B. Miller, T. McGee, J. Swails, N. Homeyer, H. Gohlke, A. Roitberg, MMPBSA.py:

an ef ﬁ cient program for end-state free energy calculations, J. Chem. Theor.

Comput. 8 (9) (2012) 3314e3321.

[57] G. Bussi, D. Donadio, M. Parrinello, Canonical sampling through velocity

rescaling, J. Chem. Phys. 126 (1) (2007).

[75] H. Nguyen, D. Roe, C. Simmerling, Improved generalized born solvent model

parameters for protein simulations, J. Chem. Theor. Comput. 9 (4) (2013)

2020e2034.

[76] S. Genheden, U. Ryde, The MM/PBSA and MM/GBSA methods to estimate

ligand-binding af ﬁ nities, Expet Opin. Drug Discov. 10 (5) (2015) 449e461.

[77] J. Borisek, A. Saltalamacchia, A. Spinello, A. Magistrato, Exploiting cryo-EM

structural information and all-atom simulations to decrypt the molecular

mechanism of splicing modulators, J. Chem. Inf. Model. 60 (5) (2020)

2510e2521.

[78] W. Humphrey, A. Dalke, K. Schulten, VMD: visual molecular dynamics, J. Mol.

Graph. Model. 14 (1) (1996) 33e38.

[79] E. Pettersen, T. Goddard, C. Huang, G. Couch, D. Greenblatt, E. Meng, T. Ferrin,

UCSF chimera - a visualization system for exploratory research and analysis,

J. Comput. Chem. 25 (13) (2004) 1605e1612.

[58] J. Sgrignani, L. Casalino, F. Doro, A. Spinello, A. Magistrato, Can multiscale

simulations unravel the function of metallo-enzymes to improve knowledge-

based drug discovery? Future Med. Chem. 11 (7) (2019) 771e791.

[59] A. Spinello, A. Magistrato, An omics perspective to the molecular mechanisms

of anticancer metallo-drugs in the computational microscope era, Expet Opin.

Drug Discov. 12 (8) (2017) 813e825.

[60] J. VandeVondele, M. Krack, F. Mohamed, M. Parrinello, T. Chassaing, J. Hutter,

QUICKSTEP: fast and accurate density functional calculations using a mixed

Gaussian and plane waves approach, Comput. Phys. Commun. 167 (2) (2005)

103e128.

[61] T. Laino, F. Mohamed, A. Laio, M. Parrinello, An ef ﬁ cient real space multigrid

OM/MM electrostatic coupling, J. Chem. Theor. Comput.

1 (6) (2005)

1176e1184.

20

Article Doi

Balanced dual acting compounds targeting aromatase and estrogen receptor α as an emerging therapeutic opportunity to counteract estrogen responsive breast cancer

DOI: 10.1016/j.ejmech.2021.113733

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.ejmech.2021.113733

Full text of DOI:10.1016/j.ejmech.2021.113733

Products guided by the article

R&D Labs maybe for 93434-07-2

Relevant to this article

Enantioselective cyanosilylation of α,α-dialkoxy ketones catalyzed by proline-derived in-situ-prepared N-oxide as bifunctional organocatalyst

Unusual α-hydroxyaldehyde with a cyclopentane framework from verbenol epoxide

Carbon Elongations with 1,5-Diazapentadienium Salts: Reaction with Aralkyl Ketones

Mass Spectra of N'-Isopropylidenealkanohydrazides

1D chain Cd(II) and Co(II) coordination polymers: Synthesis, crystal structures and luminescence properties