Potent aromatase inhibitors and molecular mechanism of inhibitory action

Article abstract of DOI:10.1016/j.ejmech.2017.11.057

Estrogen is a significant factor in the maintenance and progression of hormone-dependent breast cancer. As well known, aromatase mediates the production of estrogen. Thus, inhibition of aromatase with chemical molecules has been considered to be an effective treatment for estrogen receptor-positive (ER+) breast cancer. In this work, we designed and synthesized a series of novel non-steroidal molecules containing 2-phenylindole scaffold and moiety of either imidazole or 1,2,4-triazole to enhance their binding capacity with the aromatase. Among these molecules, a compound named as 8o was confirmed experimentally to have the highest inhibitory activity to aromatase. Further cell activity assay proved that compound 8o has low cytotoxicity and is a promising lead for developing novel aromatase inhibitors. Molecular modeling and simulation techniques were performed to identify the binding modes of letrozole and 8o with the aromatase. Analysis of energy of the two compound-aromatase complexes revealed that the 8o has low binding energy (strong binding affinity) to the aromatase as compared to letrozole, which was in accordance with the experimental results. As concluded, a combination of experimental and computational approaches facilitates us to understand the molecular mechanism of inhibitory action and discover more potent non-steroidal AIs against aromatase, thereby opening up a novel therapeutic strategy for hormone-dependent breast cancer.

Full text of DOI:10.1016/j.ejmech.2017.11.057

Accepted Manuscript
Potent aromatase inhibitors and molecular mechanism of inhibitory action
Hongjun Kang, Xingqing Xiao, Chao Huang, Yan Yuan, Dongyan Tang, Xiaochang
Dai, Xianghui Zeng
PII:
S0223-5234(17)30960-1
10.1016/j.ejmech.2017.11.057
EJMECH 9932
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 9 August 2017
Revised Date: 28 October 2017
Accepted Date: 20 November 2017
Please cite this article as: H. Kang, X. Xiao, C. Huang, Y. Yuan, D. Tang, X. Dai, X. Zeng, Potent
aromatase inhibitors and molecular mechanism of inhibitory action, European Journal of Medicinal
Chemistry (2017), doi: 10.1016/j.ejmech.2017.11.057.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Potent aromatase inhibitors and molecular mechanism of inhibitory action
Hongjun Kanga,1, Xingqing Xiaob,1, Chao Huangc, Yan Yuand, Dongyan Tanga, Xiaochang Daia,*,
Xianghui Zenga, e,*
a Key Laboratory of Medicinal Chemistry for Natural Resource, Yunnan University, Ministry of Education, School of Chemical
Science and Technology, Yunnan University, Kunming 650091, P.R. China
b
Department of Chemical and Biomolecular Engineering, North Carolina State University, Raleigh, North Carolina 27695-7905,
USA
c
Engineering Research Center of Biopolymer Functional Materials of Yunnan, Yunnan Minzu University, Kunming, 650500, P.R.
China
d School of Ethnic Medicine, Yunnan Minzu University, Kunming, 650504, P.R. China
e
Current address: Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen. DK-2100
Copenhagen, Denmark
1 Co-first author
Correspondence: xchdai@ynu.edu.cn (X.D.), zeng_89@hotmail.com (X.Z.)
Highlights
►
►
►
►
Discovery of novel non-steroidal inhibitors against aromatase
Novel compounds demonstrate higher potencies in aromatase inhibitory than letrozole
Computational models of aromatase with different substrates are built
Binding mode of substrates to aromatase has significant impact on pharmacological effect
Abstract
Estrogen is a significant factor in the maintenance and progression of hormone-dependent breast
cancer. As well known, aromatase mediates the production of estrogen. Thus, reduction of
aromatase with chemical molecules has been considered to be an effective treatment for estrogen
receptor-positive (ER+) breast cancer. In this work, we designed and synthesized a series of novel
non-steroidal molecules containing 2-phenylindole scaffold and moiety of either imidazole or 1,2,4-
triazole to enhance their binding capacity with the aromatase. Among these molecules, a compound
named as 8o was confirmed experimentally to have the highest inhibitory activity to aromatase.
Further cell activity assay proved that compound 8o has low cytotoxicity and is a promising lead for
1

ACCEPTED MANUSCRIPT
developing novel aromatase inhibitors. Molecular modelling and simulation techniques were
performed to identify the binding modes of letrozole and 8o with the aromatase. Analysis of energy
of the two compound-aromatase complexes revealed that the 8o has a low binding energy (strong
binding affinity) to the aromatase as compared to letrozole, which was in accordance with the
experimental results. As concluded, a combination of experimental and computational approaches
facilitates us to understand the molecular mechanism of inhibitory action and discover more potent
non-steroidal AIs against aromatase, thereby opening up a novel therapeutic strategy for hormone-
dependent breast cancer.
Graphical Abstract
Keywords
Estrogen; breast cancer; aromatase inhibitors (AIs); letrozole; molecular mechanism of inhibitory
action
1. Introduction
Breast cancer, which is well known as a common disease in women, remains a high rate of
death for female patients worldwide [1, 2]. It is widely accepted that the level of estrogens and the
profile of estrogen receptors are two measurable indicators towards a risk assessment for breast
tumors [3]. Since nearly 70% of the patients with breast cancer are estrogen receptor-positive (ER+)
and may be estrogen dependent, estrogen deprivation has been considered an attractive therapeutic
strategy for ER-positive breast cancer [4]. Aromatase, otherwise called CYP19，is a rate-limiting
2

ACCEPTED MANUSCRIPT
enzyme in the biosynthesis of estrogens that is responsible for the conversion of androgens
including androstenedione and testosterone into estrone and estradiol [5]. For the postmenopausal
women with ER-positive breast cancer, aromatase is usually overexpressed in the breast tissues and
is the main source of local estrogen production in neoplastic tissues [5]. Interfering with aromatase
activity and reducing the level of estrogens in tumor tissues may slow down the growth of breast
cancer cells, eventually extending the lifespan of patients [6].
According to the diversity in molecular skeletons, aromatase inhibitors can be categorized into
two types: steroidal and non-steroidal blockers [6]. Steroidal AI (e.g., exemestane in Fig.1) derived
from the substrate androstenedione interacts with aromatase through chemical actions, resulting in
an irreversible binding process of the species; while non-steroidal AI (e.g., anatrozole and letrozole
in Fig.1) binds to enzyme through non-covalent interactions, resulting in a reversible binding
process.
Fig.1. Chemical structures for steroidal (exemestane) and nonsteroidal (anastrozole, letrozole) aromatase inhibitors
(AIs).
Nowadays, aromatase inhibitors (AIs) constitute the first-line drugs for ER-positive breast
cancer in postmenopausal women. The third generation of AIs (exemestane, anatrozole, and
letrozole) was approved by Food and Drug Administration as first-line therapy for hormonally-
responsive breast cancer in postmenopausal women, since they had been proven to be superior to
tamoxifene known as a representative of selective estrogen receptor modulators (SERMs) [7].
However, with the broad applications of AIs in clinical practices, some unexpected problems are
gradually shown up, such as non-responses to some of patients, resistance to AI treatment and
inhibition of some CYP450 enzymes [8]. Hence, there are urgent needs to discover and develop
new generation of AIs to overcome the defects.
3

ACCEPTED MANUSCRIPT
Fig.2. Chemical structures for selective estrogen receptor modulators (SERMs) based on 2-phenyl indole scaffold.
2-Phenyl indole, an estrogen-mimicking chemical structure, has been used widely in drug
designs for treatment of estrogen-related diseases [9, 10, 11]. Some SERMs (e.g., bazedoxifene and
pipendoxifene, Fig.2) are based on 2-phenyl indole skeletons. D-15414 (Fig.2), a 2-phenyl indole
analogue, is a non-steroidal estrogen with fairly high binding affinity for the estrogen receptor [12].
Thus, we are reasonable to postulate that 2-phenyl indole derivatives may enter into the substrate-
binding pocket of aromatase, because they are mimics of estrone and estradiol, the products of
aromatization reactions, with high binding affinity to aromatase enzyme. On the other hand,
imidazole and 1,2,4-triazole are nitrogen-containing heterocyclic rings that can coordinate with the
heme of aromatase [13]. Based on the knowledge above, a series of novel non-steroidal aromatase
inhibitors were designed by introducing the azole group at the 3 position in the 2-phenyl indole
framework (Fig.3A). The indole moiety accounts for fitting the binding site of aromatase, whereas
the azole moiety chelates the iron atom in the heme existing in the active site of aromatase. Since
our designed molecules are structurally similar to letrozole (Fig.3B), it is possible for the azole
substituted 2-phenyl indole to have certain bioactivity in inhibiting aromatase.
To study the impact of the modified 2-phenylindoles by either imidazole or 1,2,4-triazole on
aromatase activity, we synthesized twenty imidazole or 1,2,4-triazole substituted 2-phenyl indoles
(namely, compounds 8a-t (Table 1)), and tested their inhibitory activities against aromatase in our
lab. To explore the molecular mechanism of inhibitory action, an integrated computational strategy
was then applied to investigate the binding of letrozole and compound 8o to aromatase.
4

ACCEPTED MANUSCRIPT
Fig.3. (A) Rational design of new AIs based on 2-phenyl indole scaffold; (B) The overlapping of one designed
molecule (red) and letrozole (yellow).
2. Results and Discussion
2.1. Chemistry
The synthesis route of the twenty imidazole or 1,2,4-triazole substituted 2-phenyl indole
aromatase inhibitors 8a-t is shown in Scheme 1. The precursors 5a-j were synthesized by Wittig
reaction of substituted 2-nitrobenzaldehydes 2a-d and quaternary phosphonium salts 4a-c in the
presence of organic base DBU. Reductive cyclization of nitrostyrene derivatives 5a-j to the 2-
phenyl indole intermediates 6a-j was readily accomplished by triphenylphosphine using a
dichlorodioxomolybdenum (VI) complex (MoO2Cl2(dmf)2) as catalyst. Compounds 7a-j were
prepared from the precursors 6a–j by Mannich reaction. Finally, the target compounds 8a-t were
prepared by treatment of 3-dimethylaminomethyl-2-phenyl indoles 7a–j with imidazole or 1,2,4-
triazole in hot xylene. The structures of all target compounds 8a-t were confirmed by IR, LC-MS,
1H NMR and 13C NMR spectroscopy.
5

ACCEPTED MANUSCRIPT
Scheme 1. Synthesis route of twenty imidazole or 1,2,4-triazole substituted 2-phenyl indole AIs 8a-t. Reagents
and conditions: (a) i. DMF-DMA, pyrrolidine, DMF, 125℃, 6h, 95-100%; ii. NaIO4, THF/H2O, rt, overnight, 63-
67%; (b) PPh3, toluene, reflux, 6h, 72-98%; (c) DBU, THF, reflux, overnight, 91-96%; (d) MoO2Cl2(dmf)2, PPh3,
toluene, reflux, 16h, 50-67%; (e) CH2=N+(CH3)2Cl-, CH2Cl2, rt, 8h, 88-95%; (f) azole, xylene, 130℃, 1h, 25-56%.
2.2. Biological activity
2.2.1. Evaluate the inhibitory potencies of AIs against aromatase
6

ACCEPTED MANUSCRIPT
Aromatase inhibitory activities of the twenty 2-phenyl indole compounds 8a-t were determined
using enzyme-linked immunosorbent assay (ELISA) as described in the experimental section, and
the values of IC50/aromatase were given in Table 1. The most potent compounds in this series are 8o,
8c and 8e (IC50/aromatase: 14.1 nM, 32.3 nM, and 36.1 nM, respectively), which are more powerful
in the inhibition of aromatase, as compared to letrozole (IC50/aromatase: 49.5 nM). Additionally,
compounds 8a, 8k and 8s are also found to have good inhibitory activity to aromatase
(IC50/aromatase: 51.2 nM, 54.4 nM and 58.3 nM, respectively).
As summarized in Table 1, we learnt that the R1-substituted compounds with a large methoxy
group in benzene ring of indole exhibit a weak inhibitory activity to aromatase, indicating that the
R1 position is sensitive to steric hindrance. Thus, small group is required at the R1 position. The
atom F with larger electronegativity is a better choice rather than the atom Cl, attributing to the
steric flexibility for the polar amino acids on aromatase in vicinity of the R1 position. However,
when the methoxy group is replaced in the 2-phenyl group for the R3-substituted compounds, they
become potent AIs. Comparing the imidazole- and triazole-substituted compounds, it seems that
triazole has a stronger AI activity. This can be explained by the different polarities in the active
pocket of the enzyme as well as the spatial orientations.
Table 1
Aromatase inhibition activity and cytotoxicity of the twenty 2-phenyl indole AIs 8a-t and letrozole.
IC50/aromatase
(nM)
IC50/MCF-7
Compound No.
R1
R2
R3
Y
(µM)
83.2
45.0
8a
8b
8c
8d
8e
8f
F
F
H
H
H
H
H
H
H
H
F
Cl
CH
CH
CH
CH
CH
CH
CH
CH
51.2
82.9
41.5
Cl
F
32.3
>1000
Cl
Cl
260.8
36.1
F
OCH3
OCH3
F
>1000
>1000
>1000
>1000
Cl
276.8
108.7
1058
8g
8h
OCH3
OCH3
Cl
7

ACCEPTED MANUSCRIPT
>1000
>1000
8i
5, 6-methylenedioxy
5, 6-methylenedioxy
F
Cl
CH
CH
N
210.6
2635
54.4
8j
8k
F
F
H
H
H
H
H
H
H
H
F
>1000
8l
Cl
N
217.4
96.9
>1000
625
8m
Cl
F
N
382
325
14.4
12.2
612
334
354
4.73
8n
Cl
Cl
N
207.5
14.1
8o
F
OCH3
OCH3
F
N
8p
Cl
N
196.5
237.6
7772
58.3
8q
OCH3
OCH3
N
8r
8s
Cl
N
5, 6-methylenedioxy
5, 6-methylenedioxy
F
N
8t
Cl
N
677.4
49.5
Letrozole
2.2.2. Cytotoxicity assay
The cytotoxicity of the twenty 2-phenyl indole AIs 8a-t and letrozole was further tested in
MCF-7 cells, as listed in Table 1. Results of MCF-7 cell activity assay revealed that all the azole
substituted 2-phenyl indole AIs had lower cytotoxicity than letrozole. The most toxic compounds in
this series are compound 8p, 8q (IC50/MCF-7: 14.4 µM and 12.2 µM, respectively), which are less
toxic than letrozole (IC50/MCF-7: 4.73 µM). Compound 8o, the most potent AI in this series,
exhibits low cytotoxicity (IC50/MCF-7: 325 µM). This signified that compound 8o has a good safety
profile.
Anticancer effect of AIs is resultant of their impacts on human body by reduction of estrogen
levels in the circulation and tumor tissues, making it hard for ER-positive breast cancer cells to
grow and spread [14]. Hence, the AIs are not required to have high cytotoxic activity to directly kill
cancer cells. Novel potent, more selective, less toxic AIs are needed because of many side effects
resulted from the long time use of them. Compound 8o is an excellent lead for the development of
the new generation of AIs, as it possesses highly potent aromatase inhibitory activity and low
toxicity.
2.3. Molecular modeling studies
2.3.1. Computationally Probe binding mechanism of drug molecules with aromatase at a molecular
level
8

ACCEPTED MANUSCRIPT
To explore molecular mechanism of inhibitory action, we applied an integrated computational
strategy combining quantum mechanics (QM) calculation, molecular docking and atomistic
molecular dynamics (MD) simulation in this work to investigate the binding modes of aromatase
with letrozole and 8o (the most powerful compound, IC50/aromatase: 14.1 nM) [15]. QM
calculations were done for letrozole, 8o and iron porphyrin to optimize geometry and derive partial
atomic charges. Molecular docking technique was then used to determine binding sites of aromatase
and to predict possible binding poses of letrozole and 8o. Finally, atomistic MD simulations were
employed to examine the stability of the complexes letrozole-aromatase and 8o-aromatase. Binding
free energy was calculated on the last 5 ns simulation trajectory using molecular
mechanics/generalized born surface area (MM/GBSA) approach with variable dielectric constant
model [16, 17]. Details of our computational protocol are given in the experimental section and the
supporting information file.
2.3.2. Determine binding structures of aromatase with AIs
Structure of aromatase containing an iron porphyrin is available in Protein Data Bank (PDB
entry: 3EQM) [18]. Binding modes of letrozole and compound 8o on aromatase were studied using
AutoDocK Vina program based on the minimum-energy conformers of letrozole (Fig.S1-C) and 8o
(Fig.S2-C) obtained from geometrical optimizations [19]. Nine putative poses were yielded from
each individual docking attempt for letrozole and 8o. Six out of the nine poses showed that both the
letrozole and 8o reside in the central pocket of aromatase. (It is mentioned that the iron porphyrin is
centered in the pocket of aromatase.) The other three poses of letrozole and 8o are located in varied
surficial regions of aromatase. Thus, it is reasonable to believe that the binding site of aromatase is
in its central pocket. After removing some similar poses from the six poses of letrozole and 8o in
the aromatase pocket, three potential poses of letrozole and 8o were reserved (Fig.4). For
convenience, we named the three docking poses of letrozole as Poses I, II and III (Fig.4A), and
called the three poses of 8o as Poses 1, 2 and 3 (Fig.4B). Both the letrozole and 8o are buried in
interior of aromatase and interact with iron porphyrin.
The molecular docking results show only the static orientation and interactions of AIs with
aromatase. Therefore, it is necessary to employ explicit-solvent atomistic MD simulations to study
their dynamics properties of AIs and aromatase. We examined the binding stability of all the six
poses of letrozole and compound 8o in six independent 50-ns simulations. Binding free energy of
the AIs and aromatase over the last 5 ns of the simulation trajectory was calculated using the
9

ACCEPTED MANUSCRIPT
MM/GBSA approach with the variable dielectric constant model. Fig.4C lists our computed binding
free energies of the six docking poses associated with the letrozole-aromatase and 8o-aromatase.
For the letrozole-aromatase complex, Pose I has a lower binding free energy, -19.96±0.16 kcal/mol,
than Poses II and III. Thus, we can reasonably infer that Pose I is the most likely binding mode of
letrozole and aromatase. For the 8o-aromatase complex, Pose 1 is considered to be the most likely
binding mode due to the low binding free energy -23.43±0.21 kcal/mol. Additionally, by comparing
Pose I of the letrozole-aromatase and Pose 1 of the 8o-aromatase, we found that Pose 1 has a lower
binding free energy (-23.43±0.21 kcal/mol) than Pose I (-19.96±0.16 kcal/mol). It signifies that
aromatase has a stronger preference to compound 8o rather than letrozole, which is in agreement
with our experimental measurements in Table 1: the IC50/aromatase value (14.1 nM) of 8o is lower
than that (49.5 nM) of letrozole.
10

ACCEPTED MANUSCRIPT
Fig.4. (A) Three poses of letrozole in aromatase, viz. Poses I, II and III, are obtained via the AutoDock Vina program,
where the letrozole is buried in interior of aromatase and interacts with iron porphyrin. (B) Three poses of compound 8o
in aromatase, viz. Poses 1, 2 and 3, are obtained. Aromatase is represented by a silver ribbon, iron porphyrin is specified
11

ACCEPTED MANUSCRIPT
in yellow, and letrozole and 8o are exhibited in multi-colored bead-and-bond model. (C) Comparison of computed
binding free energies between letrozole and aromatase and between compound 8o and aromatase. (unit: kcal/mol)
2.3.3. Analysis of energy decomposition of AIs bound to aromatase
Clustering analysis was performed on the last 5 ns simulation trajectories of Pose I (letrozole-
aromatase) and Pose 1 (8o-aromatase) to obtain representative structures of the complexes at
equilibrium [20]. Figs.5A and 6A show the representative structures of the complexes letrozole-
aromatase and 8o-aromatase, respectively. Both letrozole and 8o with their triazole pointing at iron
porphyrin were buried in a limited pocket of aromatase, leading to a restricted motion of molecules.
To explore which parts of the AIs contribute to the binding affinity and which parts contribute to
binding specificity, we calculated the inter-chain van der Waals (VDW) and charge-charge
(ELE+EGB) interactions between the AIs and aromatase. The first two terms in Equation (S-3)
were calculated and plotted in Figs.5B and 6B. Detailed calculations of energy decomposition were
given in the Supporting information. To study partial energy contributions of AIs to the whole
binding with aromatases, we purposely divided letrozole into two components, viz. triazole and
other (Fig.S1-A), and purposely divided 8o into four components, viz. F, OCH3, triazole and other
(Fig.S2-A). Meanwhile, iron porphyrin was also separated into two components: HEM and Fe
(Fig.S3). The component Fe contains only one Fe atom centered in iron porphyrin, and the HEM
part involves all the non-Fe atoms in iron porphyrin.
Firstly, let us discuss the situation of letrozole bound to aromatase (Fig.5B). Triazole of
letrozole has strong inter-chain VDW interactions with the HEM of iron porphyrin, and slight inter-
chain ELE+EGB interactions with the HEM and Fe. The other part of letrozole interacts with the
HEM relying on not only mutual VDW attraction, but also mutual charge-charge (ELE+EGB)
repulsion. Besides, strong VDW interactions were observed between the other part of letrozole and
the residue Trp at site 224 (TRP 224) on aromatase. The residues VAL at site 373 (VAL 373) and
Met at site 374 (MET 373) prefer to interact with letrozole via both inter-chain VDW and
ELE+EGB energies. The interaction domain of letrozole and aromatase was enlarged in Fig.5C to
highlight the relative locations of letrozole, iron porphyrin and the key residues on aromatase. As
evidenced by our previous work [21, 22], the inter-chain VDW energy contributes to the binding
specificity for AIs to iron porphyrin-contained aromatase, and the inter-chain ELE+EGB energy
between AIs and iron porphyrin-contained aromatase is responsible for the binding affinity.
12

ACCEPTED MANUSCRIPT
Fig.5. (A) The most likely binding structure of letrozole and aromatase. The picture enlarged on the right side
demonstrates a detailed binding position and orientation of letrozole in the aromatase, where the amino acids that are
found in the aromatase to be near letrozole are shown as multi-colored isosurfaces. (B) Energy panels show the inter-
chain van der Waals (VDW) and charge-charge (ELE+EGB) interactions between letrozole and aromatase. The x-axis
represents the two primary components of letrozole: triazole and other, the y-axis represents the iron porphyrin (HEM
and Fe) and the amino acids that are near letrozole, and the color bar on the right scales the value of the energies. (C)
The snapshot shows that letrozole has strong inter-chain interactions with Phe at site 221, Trp at site 224, Val at site 373,
Met at site 374, and HEM and Fe.
Next, our focus is transferred on the situation of compound 8o bound to aromatase (Fig.6B).
Triazole of 8o attracts the HEM part of iron porphyrin through the inter-chain VDW energy, but
repels the HEM part through the inter-chain ELE+EGB energy. A medium inter-chain ELE+EGB
13

ACCEPTED MANUSCRIPT
interaction was also found in between the triazole of 8o and the component Fe of iron porphyrin.
Although the atom F of 8o does not provide strong energy contributions in the binding of aromatase,
a substitution of the atom F by a larger atom Cl easily cause an atomic overlap with aromatase due
to a limited pocket space (Fig.6C), thereby weakening the affinity of AIs. This explains why the
compound 8p exhibits a weak inhibitory activity against aromatase (Table 1). The group OCH3 of
8o, which sticks out of the central pocket of aromatase, has certain conformational flexibility so as
to contact well with the residues Phe at site 221 (PHE 221), Asp at site 309 (ASP 309) and His at
site 480 (HIS 480) via the inter-chain VDW energies, strengthening the thermal stability of the 8o-
aromatase. If the group OCH3 was substituted by stiff spherical atoms, such as F and Cl, a loss of
conformational flexibility might trigger either an inadequate contact for the small atom F or an
atomic overlap for the large atom Cl with aromatase. For this reason, the capabilities of compounds
8k and 8l in the inhibition of aromatase are relatively weaker than compound 8o (Table 1).
Based on the above discussions about letrozole and 8o, we reasonably believed that introducing
the triazole group favors molecular interactions of AIs and iron porphyrin, eventually arriving at our
initial goal of suppressing bioactivity of aromatase. Meanwhile, steric effect is also an important
factor in the design process of potent AIs. Proper substitutions chosen in AIs facilitate the binding
of AIs with aromatase by either increasing molecular contacts or avoiding inappropriate overlaps.
14

ACCEPTED MANUSCRIPT
Fig.6. (A) The most likely binding structure of compound 8o and aromatase. The picture enlarged on the right side
demonstrates a detailed binding position and orientation of 8o in the aromatase, where the amino acids that are found in
the aromatase to be near 8o are shown as multi-colored isosurfaces. (B) Energy panels show the inter-chain VDW and
ELE+EGB interactions between 8o and aromatase. The x-axis represents the four primary components of 8o: fluorin (F),
methy (OCH3), triazole and other, the y-axis represents the iron porphyrin (HEM and Fe) and the amino acids that are
near 8o. (C) The snapshot shows that 8o has strong inter-chain interactions with Phe at site 221, Val at site 370, Leu at
site 477, His at site 480, and HEM and Fe.
3. Conclusion
15

ACCEPTED MANUSCRIPT
The use of aromatase inhibitor or ER modulator as adjuvant therapy has been the mainstay of
treatment for postmenopausal women with ER-positive early-stage breast cancer. Persistent risk of
tumor recurrence remains a clinical and scientific challenge. Therefore, it is worthwhile to prevent
tumor recurrence by developing an alternative strategy with better efficacy. In this work, we strived
to develop novel non-steroidal AIs based on a 2-phenylindole scaffold to suppress the bioactivity of
aromatase, and three compounds 8o, 8c and 8e were proved experimentally to be more powerful
than letrozole does. Compound 8o is considered to be the most potent AI due to its lowest value
(14.1 nM) of IC50/aromatase. Additionally, results of MCF-7 cell activity assay revealed that
compound 8o exhibits much lower cytotoxicity (IC50/MCF-7: 325 µM) as compared to letrozole
(IC50/MCF-7: 4.73 µM), indicating a better safety profile for compound 8o. Furthermore, we
computationally studied the binding structures of letrozole and compound 8o in aromatase,
providing significant impact on the AIs’ efficacy. The discovery of new highly potent AIs and
understanding their molecular mechanism of inhibitory action is central to further improving
therapeutic options for ER-positive breast cancers.
4. Experimental and simulation section
4.1. Chemistry
All compounds were fully characterized by spectroscopic techniques. The NMR spectra were
recorded on a Bruker-Avance 300 MHz spectrometer (1H: 300 MHz, 13C: 75 MHz) with
tetramethylsilane (TMS) as the internal standard (δ 0.0 ppm), chemical shifts (δ) are expressed in
ppm, and J values are given in Hz. Deuterated DMSO was used as a solvent. IR spectra were
recorded on a FT-IR Thermo Nicolet Avatar 360 using a KBr pellet. The reactions were monitored
by thin layer chromatography (TLC) using silica gel GF254. The melting points were determined on
an XT-4A melting point apparatus and are uncorrected. HRMS was performed on an Agilent LC-
MSD TOF instrument.
All chemicals and solvents were used as received without further purification unless otherwise
stated. Column chromatography was performed on silica gel (200–300 mesh).
4.1.1. Method for the synthesis of compound 8a
Substituted 2-nitrobenzaldehydes 2a reacted with quaternary phosphonium salts 4a to produce
the precursors 5a in the presence of organic base DBU. Reductive cyclization of the nitrostyrene 5a
16

ACCEPTED MANUSCRIPT
to the 2-phenyl indole derivative 6a was readily accomplished by triphenylphosphine using a
catalyst MoO2Cl2(dmf)2. Compound 7a was prepared from the intermediate 6a by Mannich reaction.
Finally, the target compounds 8a was prepared by treatment of 7a with imidazole in hot xylene.
4.1.1.1. 4-Fluoro-2-nitrobenzaldehyde (2a)
To a stirred solution containing 4-fluro-2-nitrotoluene (1a) (1.0 g, 6.45 mmol) in 15 mL of
anhydrous N,N-dimethylformamide, was added pyrrolidine (0.69 g, 9.70 mmol) and N,N-
dimethylformamide dimethyl acetal (1.54 g, 12.92 mmol). The resulting reaction mixture was
allowed to stir at 125 for 6 hours and then poured into 20 mL of water. The product was extracted
with ethyl acetate three times (20 mL×3). The combined organic layers were washed successively
with 30 mL of water and 30 mL of brine, then dried by anhydrous sodium sulfate and concentrated
under reduced pressure to provide dark red oil. The oil was dissolved in 20 mL of THF/water (1:1,
v/v) and the resultant solution was treated with sodium periodate (4.13 g, 19.3 mmol) at 0 . The
resulting reaction mixture was allowed to stir at room temperature overnight and was filtered
through Celite. The filtrate was diluted with 10 mL of water and extracted by ethyl acetate three
times (15 mL×3). The combined organic layers were washed successively with 25 mL of water and
25 mL of brine, then dried by anhydrous sodium sulfate and concentrated under reduced pressure to
provide dark oil that was further purified by column chromatography on silica gel (6:1 petroleum
ether/ethyl acetate) to yield 4-fluoro-2-nitrobenzaldehyde (2a) (0.69 g, 4.08 mmol, 63%) as a
yellow solid [23].
4.1.1.2. (4-Flurobenzyl)triphenylphosphonium bromide (4a)
To a stirred solution containing triphenylphosphine (1.66 g, 6.33 mmol) in 15 mL of toluene,
4-fluorobenzyl bromide (3a) (1.0 g, 5.29 mmol) was added and the resulting mixture was allowed to
stir at 120 for 6 hours, then cooled to room temperature. The precipitate was filtered off, washed
with 10 ml of organic solvent (4:1 petroleum ether/ethyl acetate), and dried by vaccum to yield (4-
flurobenzyl) triphenylphosphonium bromide (4a) (2.34 g, 5.18 mmol, 98 %) as a white solid [24].
4.1.1.3. 4-Fluoro-2-nitro-4’-fluoro stilbene (5a)
4-Fluoro-2-nitrobenzaldehyde (2a) (1.0 g, 5.91 mmol) and (4-Flurobenzyl)
triphenylphosphonium bromide (4a) (2.94 g, 6.51 mmol) were dissolved in 20 mL of THF. DBU
(1.08 g, 7.10 mmol) was added to the solution. The resulting reaction mixture was refluxed for 4
hours. Then the solvent was evaporated and the residue was dissolved in 30 mL of ethyl acetate.
17

ACCEPTED MANUSCRIPT
The organic layer was washed by brine, then dried by anhydrous sodium sulfate and concentrated
under reduced pressure to provide crude 4-fluoro-2-nitro-4'-fluoro stilbene (5a) (1.45 g, 5.55 mmol,
94%) as a brown oil [24].
4.1.1.4. 6-Fluoro-2-(4-fluorophenyl)-1H-indole (6a)
To a stirring solution containing 4-fluoro-2-nitro-4’-fluoro stilbene (5a) (1.45 g, 5.55 mmol) in
25 mL of toluene, triphenylphosphine (3.88 g, 14.8 mmol) and MoO2Cl2(dmf)2 (0.20 g, 0.58 mmol)
were added successively. The resulted solution was refluxed for 16 hours under nitrogen
atmosphere. The solvent was evaporated and the residue was purified by column chromatography
on silica gel (10:1 petroleum ether/ethyl acetate) to yield 6-fluoro-2-(4-fluorophenyl)-1H-indole (6a)
(0.84 g, 3.66 mmol, 66%) as a white solid [25].
4.1.1.5. 3-Dimethylaminomethyl-6-Fluoro-2-(4-fluorophenyl)-1H-indole (7a)
6-Fluoro-2-(4-fluorophenyl)-1H-indole (6a) (1.0 g, 4.36 mmol) and (methylene)- dimethyl
ammonium chloride (0.82 g, 8.76 mmol) were dissolved in 15 mL of dry dichloromethane. The
resulting reaction mixture was stirred for 8 hours at room temperature, and then treated with 10 ml
of saturated sodium carbonate solution. The organic layer was separated and washed with water two
times (10×2), then dried by anhydrous sodium sulfate and concentrated under reduced pressure to
provide 3-dimethylaminomethyl-6-Fluoro-2-(4-fluorophenyl)-1H-indole (7a) (1.15 g, 4.01 mmol,
92 %) as a white solid [26].
4.1.1.6. 3-((1H-imidazol-1-yl)methyl)-6-fluoro-2-(4-fluorophenyl)-1H-indole (8a)
To a solution containing 3-dimethylaminomethyl-6-Fluoro-2-(4-fluorophenyl)-1H-indole (7a)
(1.0 g, 3.49 mmol) in 20 ml of xylene was added imidazole (1.19 g, 17.45 mmol). The reaction
mixture was allowed to stir at 130℃ for 1 hour. Then xylene was removed under reduced pressure
and the residue was dissolved by 30 ml of ethyl acetate. The organic layer was washed with water
three times (10 ×3), then dried by anhydrous sodium sulfate and concentrated under vaccum. The
residue was recrystalized from ethyl acetate to afford 3-((1H-imidazol-1-yl)methyl)-6-fluoro-2-(4-
methoxyphenyl)-1H-indole (8a) (0.50 g, 1.60 mmol) [27]: Yield 46 %; White solid; Mp 211-220 oC;
1
IR (KBr) (νmax, cm−1): 3144, 1502, 1226, 806, 744; H NMR (300 MHz, DMSO-d6): δ 11.75 (s,
1H), 7.66-7.62 (d, J = 10.8 Hz, 3H), 7.48-7.35 (m, 3H), 7.19-7.16 (m, 1H), 6.95-6.84 (m, 3H), 5.40
(s, 2H); 13C NMR (75 MHz, DMSO-d6): δ 163.6-160.8 (d, J = 210.0 Hz), 160.4-157.7 (d, J = 202.5
18

ACCEPTED MANUSCRIPT
Hz), 136.8, 136.5, 135.9, 130.3, 128.5, 128.1, 124.7, 119.6, 119.0, 116.1, 115.8, 108.3, 106.6, 97.6,
+
97.3, 40.7; HRMS (ESI) m/z calcd for C18H14N3F2 [M+H]+, 310.1150; found, 310.1150.
The following compounds (8b-t) were prepared using the general method described for the
synthesis of compound 8a.
4.1.2. 3-((1H-imidazol-1-yl)methyl)-2-(4-chlorophenyl)-6-fluoro-1H-indole (8b): Yield 42 %;
o
1
White solid; Mp 215-223 C; IR (KBr) (νmax, cm−1): 3327, 1625, 1221, 1080, 804, 731; H NMR
(300 MHz, DMSO-d6): δ 11.79 (s, 1H), 7.63-7.61 (m, 5H), 7.49-7.44 (m, 1H), 7.20-7.17 (d, J = 9.0
13
Hz, 1H), 6.96-6.85 (m, 3H), 5.42 (s, 2H); C NMR (75 MHz, DMSO-d6): δ 160.9, 157.8, 136.8,
136.0, 133.0, 130.4, 129.8, 129.0, 128.5, 124.7, 119.7, 119.0, 108.4, 108.1, 107.1, 97.7, 97.3, 40.7;
HRMS (ESI) m/z calcd for C18H14N3FCl+ [M+H]+, 326.0854; found, 326.0858.
4.1.3. 3-((1H-imidazol-1-yl)methyl)-6-chloro-2-(4-fluorophenyl)-1H-indole (8c): Yield 42 %;
o
1
White solid; Mp 233-238 C; IR (KBr) (νmax, cm−1): 3005, 1672, 1499, 1227, 836, 749; H NMR
(300 MHz, DMSO-d6): δ 11.81 (s, 1H), 7.67-7.61 (m, 3H), 7.48-7.36 (m, 4H), 7.07-7.04 (m, 1H),
13
6.94 (s, 1H), 6.84 (s, 1H), 5.40 (s, 2H); C NMR (75 MHz, DMSO-d6): δ 163.7, 160.5, 136.8,
136.2, 130.5, 130.5,130.4, 128.5, 127.9, 126.7, 120.0, 119.7, 119.0, 116.2, 115.9, 111.0, 106.7, 40.6;
HRMS (ESI) m/z calcd for C18H14N3FCl+ [M+H]+, 326.0854; found, 326.0853.
4.1.4. 3-((1H-imidazol-1-yl)methyl)-6-chloro-2-(4-chlorophenyl)-1H-indole (8d): Yield 44 %;
o
1
White solid; Mp 232-244 C; IR (KBr) (νmax, cm−1): 3024, 1509, 1229, 830; H NMR (300 MHz,
DMSO-d6): δ 11.82 (s, 1H), 7.62 (m, 5H), 7.48-7.43 (m, 2H), 7.07-7.04 (m, 1H), 6.94 (s, 1H), 6.83
13
(s, 1H), 5.42 (s, 2H); C NMR (75 MHz, DMSO-d6): δ 136.8, 136.4, 136.3, 133.2, 130.2, 130.2,
129.9, 129.0, 129.0, 128.5, 126.9, 126.7, 120.0, 119.9, 119.0, 111.0, 107.2, 40.5; HRMS (ESI) m/z
calcd for C18H12N2O2FCl+ [M+H]+, 342.0565; found, 342.0563.
4.1.5. 3-((1H-imidazol-1-yl)methyl)-6-fluoro-2-(4-methoxyphenyl)-1H-indole (8e): Yield 53 %;
o
1
yellow solid; Mp 210-213 C; IR (KBr) (νmax, cm−1): 3123, 1617, 1410, 1031, 829, 590; H NMR
(300 MHz, DMSO-d6): δ 11.65 (s, 1H), 7.62 (s, 1H), 7.55-7.52 (d, 2H), 7.46-7.42 (dd, 1H), 7.18-
7.15 (m, 1H), 7.11-7.09 (d, 2H), 6.92 (s, 1H), 6.91-6.89 (d, 1H), 6.85 (s, 1H), 5.39 (s, 2H), 3.81 (s,
13
3H); C NMR (75 MHz, DMSO-d6): δ 160.7-157.6 (d, J = 232.5 Hz), 159.3, 137.6, 136.8, 135.8-
19

ACCEPTED MANUSCRIPT
135.6 (d, J = 15 Hz), 129.4, 128.5, 124.9, 124.0, 119.2-119.1 (d, J = 7.5 Hz), 119.0, 114.5, 108.1-
107.8 (d, J = 22.5 Hz), 105.6, 97.6-97.2 (d, J = 30 Hz), 55.3, 40.9; HRMS (ESI) m/z calcd for
C19H17FN3O+ [M+H]+, 322.1350; found, 322.1352.
4.1.6. 3-((1H-imidazol-1-yl)methyl)-6-chloro-2-(4-methoxyphenyl)-1H-indole (8f): Yield 48 %;
o
1
yellow solid; Mp 213-221 C; IR (KBr) (νmax, cm−1): 3115, 1616, 1504, 1459, 1252, 837, 632; H
NMR (300 MHz, DMSO-d6): δ 11.70 (s, 1H), 7.61 (s, 1H), 7.55-7.47 (d, 2H), 7.44-7.41 (dd, 1H),
7.12-7.09 (m, 1H), 7.05-7.05 (d, 2H), 7.02-6.94 (d, 1H), 6.84 (s, 1H), 5.39 (s, 1H), 3.82 (s, 3H); 13C
NMR (75 MHz, DMSO-d6): δ 159.4, 137.9, 136.7, 136.1-136.0 (d, J = 7.5 Hz), 129.5, 128.5, 126.9,
126.3, 123.7, 119,8, 119,4, 118,9, 114,5, 110.9, 105.7, 55.3, 40.7; HRMS (ESI) m/z calcd for
C19H17N3OCl+ [M+H]+, 338.1054; found, 338.1052.
4.1.7. 3-((1H-imidazol-1-yl)methyl)-2-(4-fluorophenyl)-6-methoxy-1H-indole (8g): Yield 32 %;
o
1
yellow solid; Mp 208-220 C; IR (KBr) (νmax, cm−1): 3119, 1632, 1499, 1155, 1068, 857, 663; H
NMR (300 MHz, DMSO-d6): δ 11.45 (s, 1H), 7.63-7.59 (m, 3H), 7.40-7.32 (m, 3H), 6.96 (s, 1H),
6.91-6.91 (dd, 1H), 6.85 (s, 1H), 6.72-6.69 (m, 1H), 5.37 (s, 2H), 3.78 (s, 3H); 13C NMR (75 MHz,
DMSO-d6): δ 163.5-162.3 (d, J = 90 Hz), 156.3, 136.8, 134.5, 130.1, 129.9, 128.4, 122.3, 119.0,
116.1, 115.8, 109.9, 106.4, 94.5, 55.2, 40.9; HRMS (ESI) m/z calcd for C19H17N3OF+ [M+H]+,
322.1350; found, 322.1354.
4.1.8. 3-((1H-imidazol-1-yl)methyl)-2-(4-chlorophenyl)-6-methoxy-1H-indole (8h): Yield 55 %;
White solid; Mp 224-230 oC; IR (KBr) (νmax, cm−1): 3453, 3055, 1631, 1459, 1158, 1068, 837, 570;
1H NMR (300 MHz, DMSO-d6): δ 11.49 (s, 1H), 7.62 (s, 1H), 7.58 (s, 1H), 7.35-7.33 (d, 3H), 6.96,
13
6.91-6.90, 6.85, 6.72-6.69 (m, 1H), 5.39 (s, 1H), 3.78 (s, 3H); C NMR (75 MHz, DMSO-d6): δ
156.4, 136.8, 136.8, 134.0, 132.4, 130.9, 129.5, 129.0, 128.4, 122.3, 119.1, 119.0, 110.0, 106.9,
94.4, 55.2, 40.9; HRMS (ESI) m/z calcd for C19H17N3OCl+ [M+H]+, 338.1054; found, 338.1058.
4.1.9. 7-((1H-imidazol-1-yl)methyl)-6-(4-fluorophenyl)-5H-[1,3]dioxolo[4,5-f]indole (8i): Yield
56 %; White solid; Mp 232-238 oC; IR (KBr) (νmax, cm−1): 3152, 1551, 1474, 1349, 1210, 840, 660;
1H NMR (300 MHz, DMSO-d6): δ 11.45 (s, 1H), 7.60-7.56 (t, 3H), 7.37-7.31 (m, 2H), 6.96-6.92 (d,
13
3H), 6.84 (s, 1H),5.95 (1H), 5.54 (s, 2H); C NMR (75 MHz, DMSO-d6): δ 163.3, 160.0, 144.6,
142.6, 136.7, 134.3, 130.8, 129.8, 129.7, 128.6, 128.4, 121.9, 118.9, 116.0, 115.7, 106.8, 100.3,
96.9, 92.1, 40.9; HRMS (ESI) m/z calcd for C19H15N3O2F+ [M+H]+, 336.1142; found, 336.1144.
20

ACCEPTED MANUSCRIPT
4.1.10. 7-((1H-imidazol-1-yl)methyl)-6-(4-chlorophenyl)-5H-[1,3]dioxolo[4,5-f]indole (8j): Yield
o
1
51 %; White solid; Mp 244-254 C; IR (KBr) (νmax, cm−1): 3207, 1466, 1306, 1194, 1074, 825; H
NMR (300 MHz, DMSO-d6): δ 11.49 (s, 1H), 7.61-7.56 (d, 5H), 6.96-6.92 (t, 3H), 6.84 (s, 1H),
13
5.95 (s, 2H), 5.36 (s, 2H); C NMR (75 MHz, DMSO-d6): δ 144.8, 142.7, 136.7, 133.8, 132.2,
131.0, 130.9, 129.3, 128.9, 128.4, 122.0, 118.9, 107.3, 100.4, 97.0, 92.1, 40.8; HRMS (ESI) m/z
calcd for C19H15N3O2Cl+ [M+H]+, 352.0847; found, 352.0847.
4.1.11. 3-((1H-1,2,4-triazol-1-yl)methyl)-6-fluoro-2-(4-fluorophenyl)-1H-indole (8k): Yield 40 %;
o
1
White solid; Mp 218-224 C; IR (KBr) (νmax, cm−1): 3251, 1504, 1225, 1016, 838, 699; H NMR
(300 MHz, DMSO-d6): δ 11.72 (s, 1H), 8.73 (s, 1H), 7.92-7.89 (m, 1H), 7.64-7.59 (t, 2H), 7.44-
7.38 (m, 1H), 7.19-7.15 (m, 1H), 6.95-6.89 (m, 2H), 5.58 (s, 2H); 13C NMR (75 MHz, DMSO-d6):
δ 163.8, 160.9, 160.5, 157.8, 151.5, 144.0, 136.9, 135.9, 135.7, 135.6, 130.6, 130.5, 128.9, 124.6,
119.7, 119.6, 116.1, 115.8, 108.3, 108.0, 105.7, 97.7, 97.3, 43.8; HRMS (ESI) m/z calcd for
C17H12N4F2Na+ [M+Na]+, 333.0922; found, 333.0919.
4.1.12. 3-((1H-1,2,4-triazol-1-yl)methyl)-2-(4-chlorophenyl)-6-fluoro-1H-indole (8l): Yield 40 %;
o
1
White solid; Mp 247-254 C; IR (KBr) (νmax, cm−1): 3215, 1627, 1131, 830, 679; H NMR (300
MHz, DMSO-d6): δ 11.76 (s, 1H), 8.74 (s, 1H), 7.97-7.88 (m, 3H), 7.64-7.61 (d, 2H), 6.96-6.90 (t,
13
1H), 5.60 (s, 2H); C NMR (75 MHz, DMSO-d6): δ 160.9, 157.8, 151.5, 144.0, 136.5, 136.0,
135.8, 133.1, 130.3, 130.0, 128.9, 124.6, 119.8, 119.7, 108.4, 108.0, 106.1, 97.7, 97.3, 43.7; HRMS
(ESI) m/z calcd for C17H12N4FNaCl+ [M+Na]+, 349.0626; found, 349.0627.
4.1.13. 3-((1H-1,2,4-triazol-1-yl)methyl)-6-chloro-2-(4-fluorophenyl)-1H-indole (8m): Yield 49 %;
o
1
White solid; Mp 198-205 C; IR (KBr) (νmax, cm−1): 3230, 1504, 1135, 840, 512; H NMR (300
MHz, DMSO-d6): δ 11.77 (s, 1H), 8.72 (s, 1H), 7.97 (s, 1H), 7.94-7.90 (m, 3H), 7.65-7.62 (d, 1H),
7.45-7.39 (m, 3H), 7.09-7.06 (d, 1H), 5.58 (s, 2H); 13C NMR (75 MHz, DMSO-d6): δ 163.8, 160.6,
151.5, 144.0, 137.3, 136.2, 130.7, 130.6, 127.8, 126.7, 126.6, 119.9, 116.1, 115.8, 111.0, 105.7,
43.7; HRMS (ESI) m/z calcd for C17H12N4FNaCl+ [M+Na]+, 349.0626; found, 349.0626.
4.1.14. 3-((1H-1,2,4-triazol-1-yl)methyl)-6-chloro-2-(4-chlorophenyl)-1H-indole (8n): Yield 40 %;
o
1
White solid; Mp 229-234 C; IR (KBr) (νmax, cm−1): 3230, 1431, 1156, 1011, 856; H NMR (300
21

ACCEPTED MANUSCRIPT
MHz, DMSO-d6): δ 11.81 (s, 1H), 8.73 (s, 1H), 7.97 (s, 1H), 7.91-7.88 (d, 2H), 7.65-7.62 (d, 3H),
7.44-7.43 (d, 1H), 7.09-7.06 (m, 1H), 5.60 (s, 2H); 13C NMR (75 MHz, DMSO-d6): δ 151.5, 144.0,
136.9, 136.3, 133.3, 130.1, 129.0, 126.9, 126.6, 120.0, 111.0, 106.2, 43.6; HRMS (ESI) m/z calcd
+
for C17H12N4NaCl2 [M+Na]+, 365.0331; found, 365.0333.
4.1.15. 3-((1H-1,2,4-triazol-1-yl)methyl)-6-fluoro-2-(4-methoxyphenyl)-1H-indole (8o): Yield
o
1
45 %; yellow solid; Mp 206-210 C; IR (KBr) (νmax, cm−1): 3150, 1615, 1460, 1139, 842, 676; H
NMR (300 MHz, DMSO-d6): δ 11.60 (s, 1H), 8.69 (s, 1H), 7.98 (s, 1H), 7.81-7.78 (d, 2H), 7.61-
7.57 (dd, 1H), 7.17-7.12 (m, 1H), 6.94-6.88 (d, 2H), 5.57 (s, 1H), 3.82 (s, 3H); 13C NMR (75 MHz,
DMSO-d6): δ 160.7-157.5 (d, J = 240 Hz), 159.4, 151.4, 143.9, 137.9, 135.8-135.6 (d, J = 15 Hz),
129.7, 124.8, 123.9, 119.4-119.2 (d, J = 15 Hz), 114.4, 108.0-107.7 (d, J = 22.5 Hz), 104.6, 97.5-
97.2 (d, J = 22.5 Hz), 55.2, 44.0; HRMS (ESI) m/z calcd for C18H15N4OFNa+ [M+Na]+, 345.1127;
found, 345.1126.
4.1.16. 3-((1H-1,2,4-triazol-1-yl)methyl)-6-chloro-2-(4-methoxyphenyl)-1H-indole (8p): Yield
45 %; yellow solid; Mp 211-216 oC; IR (KBr) (νmax, cm−1): 3126, 1615, 1505, 1136, 1030, 841, 677;
1H NMR (300 MHz, DMSO-d6): δ 11.66 (s, 1H), 8.68 (s, 1H), 7.97 (s, 1H), 7.81-7.78 (d, J = 9 Hz,
2H), 7.61-7.58 (d, J = 9 Hz, 1H), 7.41-7.41 (d, 1H), 7.15-7.12 (d, J = 9 Hz, 2H), 7.07-7.03 (d, J =
12 Hz, 1H), 5.57 (s, 2H), 3.83 (s, 3H); 13C NMR (75 MHz, DMSO-d6): δ 159.5, 151.5, 143.9, 138.3,
136.1, 129.8, 126.8, 126.4, 123.6, 119.7-119.6 (d, J = 7.5 Hz), 114.5, 110.8, 104.8, 55.3, 43.9;
HRMS (ESI) m/z calcd for C18H15N4ONaCl+ [M+Na]+, 361.0826; found, 361.0824.
4.1.17. 3-((1H-1,2,4-triazol-1-yl)methyl)-2-(4-fluorophenyl)-6-methoxy-1H-indole (8q): Yield
25 %; yellow solid; Mp 178-180 oC; IR (KBr) (νmax, cm−1): 3126, 1632, 1504, 1160, 1028, 843, 677;
1H NMR (300 MHz, DMSO-d6): δ 11.42 (s, 1H), 8.69 (s, 1H), 7.97 (s, 1H), 7.91-7.86 (d, 5H), 7.50-
7.47 (d, 1H), 7.43 (s, 1H), 7.40-7.38 (t, 3H), 6.91-6.90 (d, H), 6.73-6.70 (m, 1H), 5.55 (s, 2H), 3.78
13
(s, 2H); C NMR (75 MHz, DMSO-d6): δ 163.5, 160.3, 156.3, 151.4, 143.9, 136.7, 135.0, 130.3,
130.2, 128.5, 122.1, 119.2, 116.0, 115.7, 109.8, 105.4, 94.5, 55.3, 44.0; HRMS (ESI) m/z calcd for
C18H15N4OFNa+ [M+Na]+, 345.1127; found, 345.1122.
4.1.18. 3-((1H-1,2,4-triazol-1-yl)methyl)-2-(4-chlorophenyl)-6-methoxy-1H-indole (8r): Yield
38 %; White solid; Mp 203-207 oC; IR (KBr) (νmax, cm−1): 3425, 3128, 1633, 1457, 1161, 1029, 837,
22

ACCEPTED MANUSCRIPT
731; 1H NMR (300 MHz, DMSO-d6): δ 11.46 (s, 1H), 8.70 (s, 1H), 7.97 (s, 1H), 7.88-7.85 (d, 2H),
13
7.62-7.48 (q, 1H), 6.90-6.89 (d, 1H), 6.73-6.70 (m, 2H), 5.56 (s, 2H), 3.78 (s, 3H); C NMR (75
MHz, DMSO-d6): δ 156.4, 151.4, 143.9, 136.8, 134.5, 132.6, 130.8, 129.8, 128.9, 122.1, 119.3,
109.9, 105.9, 94.4, 55.2, 43.9; HRMS (ESI) m/z calcd for C18H15N4ONaCl+ [M+Na]+, 361.0832;
found, 361.0829.
4.1.19. 7-((1H-1,2,4-triazol-1-yl)methyl)-6-(4-fluorophenyl)-5H-[1,3]dioxolo[4,5-f]indole (8s):
Yield 46 %; White solid; Mp 245-248 oC; IR (KBr) (νmax, cm−1): 3174, 1502, 1471, 1293, 1135, 843,
672; 1H NMR (300 MHz, DMSO-d6): δ 11.41 (s, 1H), 8.71 (s, 1H), 7.97 (s, 1H), 7.89-7.84 (t, 2H),
7.40-7.34 (d, 2H), 7.13 (s, 1H), 6.91 (s, 1H), 5.95 (s, 1H), 5.51 (s, 2H); 13C NMR (75 MHz, DMSO-
d6): δ 163.4, 160.2, 151.4, 144.6, 143.9, 142.6, 134.8, 134.6, 130.8, 130.7, 130.1, 130.0, 128.5,
121.8, 115.9, 115.6, 105.9, 100.4, 97.2, 92.2, 44.0; HRMS (ESI) m/z calcd for C18H13N4O2FNa+
[M+Na]+, 359.0914; found, 359.0913.
4.1.20. 7-((1H-1,2,4-triazol-1-yl)methyl)-6-(4-chlorophenyl)-5H-[1,3]dioxolo[4,5-f]indole (8t):
o
1
Yield 49 %; White solid; Mp 257-261 C; IR (KBr) (νmax, cm−1): 3168, 1634, 1472, 1041, 676; H
NMR (300 MHz, DMSO-d6): δ 11.46 (s, 1H), 8.73 (s, 1H), 7.97 (s, 1H), 7.86-7.83 (t, 3H), 7.60-
7.57 (d, 1H), 7.14 (s, 1H), 6.92 (s, 1H) 5.96 (s, 1H), 5.52 (s, 3H); 13C NMR (75 MHz, DMSO-d6): δ
151.4, 144.8, 144.0, 142.7, 134.3, 132.4, 131.0, 130.8, 129.6, 128.9, 121.9, 106,4, 100.4, 97.2, 92.2,
43.9; HRMS (ESI) m/z calcd for C18H13N4O2NaCl+ [M+Na]+, 375.0619; found, 375.0623.
4.2. Biological evaluation
4.2.1 Placental sample collection and microsome preparation
Studies were performed in accordance with the code of Ethics of the World Medical
Association and the ethical standards of the relevant national and institutional committees on human
investigation. Subjects were free of diseases and other substance abuse. Placenta samples were
collected and dissected immediately following delivery. Specimens were quickly washed with cold
PBS to eliminate contaminating blood. Aliquots of 103 g of placental samples were pulverized and
homogenized in TES solution (20mM TES, 1mM EDTA, 0.15M KCl, 0.1mM PMSF, pH 7.4) on
ice, and samples were centrifuged at 10,000 g for 25 min. The supernatant was centrifuged at
100,000 g for 75 min at 4 °C, and the aliquot was re-suspended in TES buffer containing 20 mM
23

ACCEPTED MANUSCRIPT
TES, 1 mM EDTA, 0.15 M KC and 0.25 M sucrose (pH 7.4). A Bradford assay (Bio-Rad) was used
for sample protein quantification, and samples was stored at −80 °C.
4.2.2 Aromatase inhibition assay
Aromatase inhibition assays were performed using the Estrone ELISA kit (BioVendor, Brno,
Czech Republic) androstenedione, NADPH and human aromatase with letrozole as positive control
[28]. The 96-well black plate was used to determine the activity of the compounds. Solutions of the
substrate and inhibitors were prepared and the ELISA procedures were performed as described in
the kits manuals. Quantification of samples was performed by applying the linear regression
equation of the standard curve to the absorbance response. Each compound was tested in triplicate
measurements and the average IC50 value can be seen in Table 1.
4.2.3. Cytotoxicity assay
MCF-7 cells (obtained from Kunming Institute of Zoology. CAS) were cultured at 37°C in a
5 % CO2 humidified atmosphere in DMEM medium supplemented with 10 % FBS, 1 % penicillin-
streptomycin. Cells were collected using 0.25% trypsin and seeded in 96 well plates at a density of
4,000 cells/well, then incubated at 37 °C overnight. After 48 h incubation with various
concentrations of tested compounds, the medium containing compounds was removed and fresh
medium containing 10 µL of MTT (5 mg/ml) was added to each well, then the resultant solution
was incubated for another 4 h. The MTT medium was discarded before adding 150 µL of DMSO to
each well. The optical densities at 490 nm were measured by an ELISA microplate reader
(Multiskan Go, Thermo Fisher Scientific, Gemany). The IC50/MCF-7 values were calculated using
GraphPad Prism 6 software.
4.3. Molecular modeling studies
Quantum mechanics (QM) calculations were performed on Gaussion09 package to scan the
potential energy profile of letrozole (Fig.S1-B) and the potential energy surface of 8o (Fig.S2-B) to
determine their minimum-energy structures in solution. The electrostatic surface potential (ESP) of
letrozole and 8o was calculated by using the DFT B3LYP method with the 6-31G(d) basis set, and
the partial atomic charges of letrozole and 8o was fit by performing two-stage restrained
24

ACCEPTED MANUSCRIPT
electrostatic potential (RESP) charge-fitting protocol [29]. The general amber force field (GAFF)
was employed to describe the atoms on letrozole and 8o [30].
The crystal structure of aromatase in complex with androstenedione is freely available in
Protein Data Bank (PDB entry: 3EQM). We removed the androstenedione and phosphate ions from
the original PDB file, and reserved the remaining 452 amino acids and an iron porphyrin for
subsequent computational study. AMBER 14SB force field is used to describe the standard amino
acids [31]. Iron porphyrin, an organometallic compound, is consisting of an Fe ion contained in the
centre of porphyrin (Fig.S3). To build the force field for the simulations of iron porphyrin with
aromatase inhibitors, we followed the Metal Center Parameter Builder (MCPB) parametrization
scheme to derive force constants and charge parameters for iron porphyrin [32].
Docking studies of letrozole and compound 8o on aromatase were carried out via AutoDock
Vina program. The energy-minimum structures of letrozole and 8o obtained from the QM
calculations (Figs.S1-C and S2-C) were set as a ligand, and the aromatase was treated as a rigid
receptor. The Gasteiger charges were used in our molecular docking studies. The dimensionality of
docking grid box was fixed at size_x=40 Å, size_y=40 Å and size_z=40 Å. By changing the
positions of box center, we conducted multiple independent docking attempts to search for possible
binding sites and determine all binding modes for the two letrozole-aromatase and 8o-aromatase
systems.
All predicted docking poses of letrozole and compound 8o on the aromatase obtained from the
docking studies need to undergo further evaluations in atomistic molecular dynamics (MD)
simulations. We solvated the aromatase complexed with letrozole and 8o into a periodic truncated
octahedral box filled with TIP3P water molecules. Chloride counterions were added to neutralize
the system. Details of our simulation procedures are given below. (1) A 10000-step energy
minimization of solvent was performed using steepest descent method with the letrozole-aromatase
and 8o-aromatase constrained by a force of 200 kcal/mol. (2) The system was heated over 10 ps at
298 K to relax the water molecules. A short 40-ps isothermal-isobaric (NPT) ensemble MD
simulation was carried out with the letrozole-aromatase and 8o-aromatase constrained by a force of
200 kcal/mol. (3) Another 10000-step energy minimization was performed with a restraint of 25
kcal/mol to the letrozole-aromatase and 8o-aromatase, after which a 20-ps NPT MD simulation was
run with a restraint of 25 kcal/mol to the letrozole-aromatase and 8o-aromatase. (4) A 10000-step
unconstrained energy minimization was conducted to all the atoms in the system, after which the
system was reheated over 40 ps at a constant volume to 298 K. (5) Production simulations were
25

ACCEPTED MANUSCRIPT
performed in 50 ns in a canonical (NVT) ensemble on the AMBER 12 package. Berendsen
thermostat was used to maintain the simulation temperature at 298 K. We used partial mesh Ewarld
summation to calculate long-ranged electrostatic interactions with a 12 Å cut-off distance and 1E-5
tolerance. The SHAKE algorithm was used to constrain the motions of all bonds involving
hydrogen atom. The calculation of the root mean square deviation (RMSD) was restrained to fit the
heavy atoms on backbone chains of the aromatase and the non-hydrogen atoms on iron porphyrin,
letrozole and 8o. The RMSD vs. time was examined over the entire simulation to ascertain whether
the binding process reached a stable state or not.
Acknowledgments
Financial support for this work was provided by the National Natural Science Foundation of
China (No. 81101688). We would like to thank NC State University High Performance Computing
Center for providing us computing resources, and Dr. Lihui Peng for assistance on placental sample
collection.
Appendix A. Supporting information
1
13
Detailed descriptions of computational protocol, and H and C NMR Spectra of compounds 8a-t
are given in the Supporting information.
References
[1] E.O. Fourkala, O. Blyuss, H. Field, R. Gunu, A. Ryan, J. Barth, I. Jacobs, A. Zaikin, A. Dawnay, U. Menon, Sex
hormone measurements using mass spectrometry and sensitive extraction radioimmunoassay and risk of estrogen
receptor negative and positive breast cancer: Case control study in UK Collaborative Cancer Trial of Ovarian
Cancer Screening (UKCTOCS), Steroids 110 (2016) 62-69.
[2] Y. Omoto, H. Iwase, Clinical significance of estrogen receptor β in breast and prostate cancer from biological
aspects, Cancer Sci. 106 (2015) 337-343.
[3] E.O. Fourkala, A. Zaikin, M. Burnell, A. Gentry-Maharaj, J. Ford, R. Gunu, Association of serum sex steroid
receptor bioactivity and sex steroid hormones, Endocr. Relat. Cancer 19 (2012) 137-147.
[4] J. Zhang, G. Li, Z. Li, X. Yu, Y. Zheng, K. Jin, H. Wang, Y. Gong, X. Sun, X. Teng, J. Cao, L. Teng, Estrogen-
independent effects of ER-α36 in ER-negative breast cancer, Steroids 77 (2012) 666-673.
[5] R.W. Brueggemeier, J.C. Hackett, E.S. Diaz-Cruz, Aromatase inhibitors in the treatment of breast cancer, Endocr.
Rev. 26 (2005) 331-345.
[6] N. Adhikari, S.A. Amin, A. Saha, T. Jha, Combating breast cancer with non-steroidal aromatase inhibitors (NSAIs):
Understanding the chemico-biological interactions through comparative SAR/QSAR study, Eur. J. Med. Chem. 137
(2017) 365-438.
26

ACCEPTED MANUSCRIPT
[7] R. Riemsma, C.A. Forbes, A. Kessels, K. Lykopoulos, M.M. Amonkar, D.W. Rea, J. Kleijnen, Systematic review of
aromatase inhibitors in the first-line treatment for hormone sensitive advanced or metastatic breast cancer, Breast
Cancer Res. Treat. 123 (2010) 9-24.
[8] G.W. Sledge, E.P. Mamounas, G.N. Hortobagyi, H.J. Burstein, P.J. Goodwin, A.C. Wolff, Past, present, and future
challenges in breast cancer treatment, J. Clin. Oncol. 32 (2014) 1979-1986.
[9] C.P. Miller, M.D. Collini, B.D. Tran, H.A. Harris, Y.P. Kharode, J.T. Marzolf, R.A. Moran, R.A. Henderson,
R.H.W. Bender, R.J. Unwalla, L.M. Greenberger, J.P. Yardley, M.A. Abou-Gharbia, C.R. Lyttle, B.S. Komm,
Design, synthesis, and preclinical characterization of novel, highly selective indole estrogens, J. Med. Chem. 44
(2001) 1654-1657.
[10] E.V. Angerer, J. Prekajac, J. Strohmeier, 2-Phenylindoles. Relationship between structure, estrogen receptor
affinity, and mammary tumor inhibiting activity in the rat, J. Med. Chem. 27 (1984) 1439-1447.
[11] W. Lv, J. Liu, D. Lu, D.A. Flockhart, M. Cushman, Synthesis of mixed (E,Z)-, (E)-,and (Z)-Norendoxifen with
dual aromatase inhibitory and estrogen receptor modulatory activities, J. Med. Chem. 56 (2013) 4611-4618.
[12] S.P. Robinson, R. Koch, V.C. Jordan, In vitro estrogenic actions in rat and human cells of hydroxylated derivatives
of D16726 (zindoxifene), an agent with known antimammary cancer activity in vivo, Cancer Res. 48 (1988) 784-
787.
[13] M.A. Neves, T.C. Dinis, G. Colombo, M.L. Sá e Melo, Fast Three Dimensional Pharmacophore Virtual Screening
of New Potent Non-Steroid Aromatase Inhibitors, J. Med. Chem. 52 (2009) 143-150.
[14] M.M. Ghorab, M.S. Alsaid, N. Samir, G.A. Abdel-Latif, A.M. Soliman, F.A. Ragab, D.A. Abou El Ella DA.
Aromatase inhibitors and apoptotic inducers: Design, synthesis, anticancer activity and molecular modeling studies
of novel phenothiazine derivatives carrying sulfonamide moiety as hybrid molecules, Eur. J. Med. Chem. 134 (2017)
304-315.
[15] X. Xiao, B. Zhao, L. Yang, X. Liang, Y. Ren, Probe the binding mode of aristololactam-β-D-glucoside to
phenylalanine transfer RNA in silico, ChemistrySelect 1 (2016) 5430-5439.
[16] H. Gohlke, C. Kiel, D.A. Case, Insights into protein-protein binding by binding free energy calculation and free
energy decomposition for the Ras-Raf and Ras-RalGDS complexes, J. Mol. Biol. 330 (2003) 891-913.
[17] K. Zhu, M.R. Shirts, R.A. Friesner, Improved methods for side chain and loop predictions via the protein local
optimization program: variable dielectric model for implicitly improving the treatment of polarization effects, J.
Chem. Theory Comput. 3 (2007) 2108-2119.
[18] D. Ghosh, J. Griswold, M. Erman, W. Pangborn, Structural basis for androgen specificity and estrogen synthesis in
human aromatase, Nature 457 (2009) 219-223.
[19] O. Trott, A.J. Olson, AutoDock Vina: improving the speed and accuracy of docking with a new scoring function,
efficient optimization, and multithreading, J. Comput. Chem. 31 (2010) 455-461.
[20] J. Shao, S.W. Tanner, N. Thompson, T.E. Cheatham, Clustering molecular dynamics trajectories: 1. characterizing
the performance of different clustering algorithms, J. Chem. Theory Comput. 3 (2007) 2312-2334.
[21] X. Xiao, P.F. Agris, C.K. Hall, Molecular recognition mechanism of peptide chain bound to the tRNALys3
anticodon loop in silico, J. Biomol. Struct. Dyn. 33 (2015) 14-27.
[22] X. Xiao, B. Zhao, P.F. Agris, C.K. Hall, Simulation study of the ability of a computationally-designed peptide to
recognize target tRNALys3 and other decoy tRNAs, Protein Sci. 25 (2016) 2243-2255.
27

ACCEPTED MANUSCRIPT
[23] D.A. Patrick, S.A. Bakunov, S.M. Bakunova, E.V Kumar, R.J. Lombardy, S.K. Jones, A.S. Bridges, O. Zhirnov,
J.E. Hall, T. Wenzler, R. Brun, R.R. Tidwell, Synthesis and in Vitro Antiprotozoal Activities of Dicationic 3,5-
Diphenylisoxazoles, J. Med. Chem., 50 (2007) 2468-2485.
[24] W. Porcal, P. Hernández, G. Aguirre, L. Boiani, M. Boiani, A. Merlino, A. Ferreira, R.D. Maio, A. Castro, M.
González, H. Cerecetto, Second generation of 5-ethenylbenzofuroxan derivatives as inhibitors of Trypanosoma
cruzi growth: Synthesis, biological evaluation, and structure–activity relationships, Bioorg. & Med. Chem., 15
(2007) 2768-2781.
[25] R. Sanz, J. Escribano, M.R. Pedrosa, R. Aguado, F.J. Arnáiz, Dioxomolybdenum(VI)-Catalyzed Reductive
Cyclization of Nitroaromatics. Synthesis of Carbazoles and Indoles, Adv. Synth. & Catal., 349 (2007) 713-718.
[26] A.P. Kozikowski, D. Ma, J. Brewer, S. Sun, E. Costa, E. Romeo, A. Guidotti, Chemistry, Binding Affinities, and
Behavioral Properties of a New Class of “Antineophobic” Mitochondrial DBI Receptor Complex (mDRC) Ligands,
J. Med. Chem, 36 (1993) 2908-2920.
[27] S. Kamiya, H. Matsui, H. Shirahase, S. Nakamura, K. Wada, M. Kanda, H. Shimaji, N. Kakeya, Thromboxane A2
Synthetase Inhibitors with Histamine H1-Blocking Activity: Synthesis and Evaluation of a New Series of Indole
Derivatives, Chem. Pharm. Bull. (Tokyo), 43 (1995) 1692-1695.
[28] K. Ohno, N. Araki, T. Yanase, H. Nawata, M. Iida, A novel nonradioactive method for measuring aromatase
activity using a human ovarian granulosa-like tumor cell line and an estrone ELISA, Toxicol Sci. 82 (2004) 443-
450.
[29] P. Cieplak, W.D. Cornell, C. Bayly, P.A. Kollman, Application of the multimolecule and multiconformational
RESP methodology to biopolymers: charge derivation for DNA, RNA, and proteins, J. Comput. Chem. 16 (1995)
1357-1377.
[30] J. Wang, R.M. Wolf, J.W. Caldwell, P.A. Kollman, D.A. Case, Development and testing of a general amber force
field, J. Comput. Chem. 25 (2004) 1157-1174.
[31] J.A. Maier, C. Martinez, K. Kasavajhala, L. Wickstrom, K.E. Hauser, C. Simmerling, ff14SB: improving the
accuracy of protein side chain and backbone parameters from ff99SB, J. Chem. Theory Comput. 11 (2015) 3696-
3713.
[32] P. Li, K.M. Merz Jr., MCPB.py: A python based metal center parameter builder, J. Chem. Inf. Model 56 (2016)
599-604.
28

ACCEPTED MANUSCRIPT
Highlights
►
Discovery of novel non-steroidal inhibitors against aromatase
►
Novel compounds demonstrate higher potencies in aromatase inhibitory than
letrozole
►
Computational models of aromatase with different substrates are built
►
Binding mode of substrates to aromatase has significant impact on pharmacological
effect