Design and Synthesis of Basic Selective Estrogen Receptor Degraders for Endocrine Therapy Resistant Breast Cancer

Article abstract of DOI:10.1021/acs.jmedchem.9b01580

The clinical steroidal selective estrogen receptor (ER) degrader (SERD), fulvestrant, is effective in metastatic breast cancer, but limited by poor pharmacokinetics, prompting the development of orally bioavailable, nonsteroidal SERDs, currently in clinical trials. These trials address local breast cancer as well as peripheral metastases, but patients with brain metastases are generally excluded because of the lack of blood-brain barrier penetration. A novel family of benzothiophene SERDs with a basic amino side arm (B-SERDs) was synthesized. Proteasomal degradation of ERα was induced by B-SERDs that achieved the objectives of oral and brain bioavailability, while maintaining high affinity binding to ERα and both potency and efficacy comparable to fulvestrant in cell lines resistant to endocrine therapy or bearing ESR1 mutations. A novel 3-oxyazetidine side chain was designed, leading to 37d, a B-SERD that caused endocrine-resistant ER+ tumors to regress in a mouse orthotopic xenograft model.

Full text of DOI:10.1021/acs.jmedchem.9b01580

Subscriber access provided by Karolinska Institutet, University Library
Article
Design and Synthesis of Basic Selective Estrogen Receptor
Degraders (B-SERDs) for Endocrine Therapy Resistant Breast Cancer
Yunlong Lu, Lauren Gutgesell, Rui Xiong, Jiong Zhao, Yangfeng Li, Carlo Rosales, Michael Hollas, Zhengnan
Shen, Jesse Gordon-Blake, Katherine Dye, Yue-Ting Wang, Sue Lee, Hu Chen, Donghong He, Oleksii
Dubrovyskyii, Huiping Zhao, Fei Huang, Amy W. Lasek, Debra A. Tonetti, and Gregory RJ Thatcher
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b01580 • Publication Date (Web): 20 Nov 2019
Downloaded from pubs.acs.org on November 20, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 72
Journal of Medicinal Chemistry
1
2
3
4
Design and Synthesis of Basic Selective Estrogen
Receptor Degraders (B-SERDs) for Endocrine
Therapy Resistant Breast Cancer
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Yunlong Lu,b,§ Lauren M. Gutgesell,b,§, Rui Xiong,a,b Jiong Zhao, Yangfeng Li, Carlo Rosales, Michael
a
a
b
a
a
b
a
a
b
c
Hollas, Zhengnan Shen, Jesse Gordon-Blake, Katherine Dye, Yueting Wang, Sue Lee, Hu Chen,
c
a
b
a
c
b
Donghong He, Oleksii Dubrovyskyii, Huiping Zhao, Fei Huang, Amy W. Lasek, Debra A. Tonetti,
and Gregory R. J. Thatcher a,c,
*
a
UICentre (drug discovery @ UIC), University of Illinois at Chicago, 833 S Wood St, Chicago, Illinois
0612, USA
6
b
Department of Pharmaceutical Sciences, University of Illinois College of Pharmacy, University of Illinois
at Chicago, 833 S Wood St, Chicago, Illinois 60612, USA
c
Department of Psychiatry, University of Illinois at Chicago, 1601 W Taylor St, Chicago, Illinois 60612,
USA
*Corresponding author: Dr. Gregory RJ Thatcher; thatcher@uic.edu; Phone: 312-355-5282; University
of Illinois at Chicago, College of Pharmacy, 833 S Wood St, Room 539, Chicago, IL 60612, USA
Email addresses: Y Lu (ylu65@uic.edu), L Gutgesell (gutgese1@uic.edu), , R Xiong (rxiong3@uic.edu),
J
Zhao (jiongz@uic.edu),
Y
Li (yfli3@uic.edu),
C
Rosales (crosal7@uic.edu),
M
Hollas
(
(
(
(
(
mhollas@uic.edu), Z Shen (zhnshen@uic.edu), J Gordon-Blake (jgordo27@uic.edu), K Dye
kedye@uic.edu), Y Wang (wangyt@uic.edu), S Lee (suelee1@uic.edu), H Chen (hch@uic.edu), D He
dhe@uic.edu),
O
Dubrovskyi (aledub@uic.edu),
H
Zhao (hpzhao58@uic.edu),
F
Huang
fhuangch@uic.edu), A Lasek (alasek@uic.edu), D Tonetti (dtonetti@uic.edu), G Thatcher
thatcher@uic.edu)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7
5
50
2
5
****
****
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
0
-25
-
-
50
75
NT
37d
Fulvestrant
Cmax ratio (brain/plasma) = 1.54
F% = 22%
T
1/2 = 4.5 h
Abstract
The clinical steroidal selective estrogen receptor (ER) degrader (SERD), fulvestrant, is effective
in metastatic breast cancer, but limited by poor pharmacokinetics, prompting the development of orally
bioavailable, non-steroidal SERDs, currently in clinical trials. These trials address local breast cancer as
well as peripheral metastases, but patients with brain metastases are generally excluded due to the lack
of blood-brain barrier penetration. A novel family of benzothiophene SERDs with a basic amino side arm
(
B-SERDs) was synthesized. Proteasomal degradation of ERα was induced by B-SERDs that achieved
the objectives of oral and brain bioavailability, while maintaining high affinity binding to ERα and both
potency and efficacy comparable to fulvestrant in cell lines resistant to endocrine therapy or bearing
ESR1 mutations. A novel 3-oxyazetidine side chain was designed, leading to 37d, a B-SERD that caused
endocrine-resistant ER+ tumors to regress in a mouse ectopic xenograft model.
Keywords: Breast cancer, tamoxifen resistance, endocrine therapy, estrogen receptor, Selective
Estrogen Receptor Degraders
ACS Paragon Plus Environment

Page 3 of 72
Journal of Medicinal Chemistry
1
2
3
4
Introduction
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
in 8 women will develop invasive breast cancer during their lifetime and, in Europe and the
1
United States, approximately 190,000 women are expected to die from breast cancer in 2019. The
majority of breast cancers express estrogen receptor α (ERα), which drives proliferation and survival of
these tumors. 2, 3 Endocrine therapy of ER positive (ER+) breast cancer has had a remarkable effect on
long-term survival. In premenopausal and postmenopausal women, treatment with tamoxifen (TAM, 1a,
Figure 1) and aromatase inhibitors (AI), respectively, provides effective first line and adjuvant therapy. 4,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
Aromatase inhibitors (AIs) prevent the synthesis of estrogens through inhibiting the aromatase enzyme;
whereas, the selective ER modulator (SERM), TAM, binds to ER, stabilizing an antagonist conformation
in breast cancer cells, resulting in antiproliferative signaling.6, 7 Notwithstanding these successes, more
women die annually of ER+ breast cancer than triple-negative breast cancer (TNBC), despite TNBC
8
having no safe, targeted therapy. This is explained by the high prevalence of ER+ breast cancer and the
9
high rate of resistance to endocrine therapy. The battlefield has therefore shifted to endocrine-resistant,
metastatic breast cancer (MBC), in which TAM and AIs have lost efficacy.
Figure 1. Structures of the SERD fulvestrant (2), oral SERDs (3,4,8,9), together with “SERM/SERDs” (5, 7), and
SERMs (1, 6). Several of these SERDs are currently being studied in clinical trials: NCT01823835, NCT02316509,
NCT03332797, NCT03236974, NCT03616587, NCT02734615, NCT03284957, NCT02338349, NCT03455270.
In the resistance setting, up to 90% of tumors remain ER+ (i.e. express ERα), wherein ER
provides prosurvival signaling even in the absence of estrogens and/or the presence of TAM.10 ER,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
therefore, remains a target for therapeutics in endocrine-resistant breast cancer, amply demonstrated by
the clinical efficacy of the selective ER degrader (SERD), fulvestrant, 2, which has moved to first line
therapy for metastatic ER+ breast cancer.11 Importantly, resistance to one endocrine therapy does not
12
lead to “cross-resistance” to all ER-targeted therapeutics. For example, patients who have progressed
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
on SERM treatment will commonly be treated with AIs or fulvestrant.5, 13 Similarly, patients with breast
14
cancer tumors resistant to AIs will generally respond to fulvestrant. Resistance to endocrine therapy is
multifactorial, including upregulation of growth factor signaling, modifications in prosurvival pathways, ER
functioning as a ligand-independent transcription factor, and mutations in ER.15-18 Hence, Cdk4/6
inhibitors that block an alternate growth pathway in ER-expressing breast cancer, used in combination
with AI’s and fulvestrant, have rapidly changed standard of care.19-22
While effective in ER+ breast cancer, fulvestrant (ICI182,780, 2) has significant pharmaceutical
liabilities including poor solubility and pharmacokinetics (PK), which require intramuscular injection. The
combination of these issues leads to significant clinical problems in establishing stable and efficacious
drug levels.23, 24 This has been a strong impetus for the recent entry into clinical trials of orally bioavailable
SERDs, including GDC-0810(3a)/GDC-0927(3b)/GDC-9545 (Genentech), AZD9496(4)/AZD9833
(AstraZeneca), LSZ-102 (Novartis), SAR439859 (Sanofi), G1T48 (G1, 9b), and RAD1901 (Radius, 5)
(Figure 1). The classification of a drug as a SERD requires demonstration of enhanced proteosomal
25
degradation of ERα, in addition to antagonist actions on binding to ER. In the clinical setting, SERMs
are the standard-of-care for premenopausal women because SERMs do not act as antiestrogens in all
tissues, in contrast to AIs and fulvestrant; and indeed, the SERM, raloxifene (6), is used clinically to treat
post-menopausal osteoporosis.26 The SERM, bazedoxifene (7), approved for post-menopausal
27
osteoporosis in Europe, has been described as a SERM/SERD hybrid; and RAD1901 (5), marketed as
28
28, 29
a SERD, preserves bone mass, thereby reclassifying it as a SERM/SERD hybrid (Figure 1).
A majority of SERDs in clinical trials are non-steroidal acrylates. The acrylate side chain engages
in a hydrogen bonding network with helix 12 (H12), as observed in the co-crystal structure of GW5638
(8a) with ERα (pdb 1R5K; Figure 1), causing displacement and destabilization of H12, a conformational
ACS Paragon Plus Environment

Page 5 of 72
Journal of Medicinal Chemistry
1
2
3
trigger to expose a hydrophobic surface, leading to proteasomal degradation of ER. The acrylate side
chain can limit brain bioavailablity, and in some cases clinical trials of oral SERDs exclude patients with
brain metastases. Patients with brain metastases have extremely poor prognosis; therefore, our
motivation in developing a SERD with a basic amino side arm (B-SERD) was to ensure good brain
bioavailability to allow treatment of this population.30-32 The replacement of an acrylate anion with a basic
amino group would be expected to improve blood-brain barrier penetration. We recently developed and
optimized a unique benzothiophene chemical scaffold as the basis for a family of potent acrylate
benzothiophene SERDs (e.g. 9) with oral bioavailability and in vivo efficacy. 33 We therefore used this
scaffold to explore a variety of basic side arms, with the objective of maintaining the excellent potency
and efficacy of 2, whilst gaining the oral and brain bioavailability lacking in 2.
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Structure Design
We have made numerous modifications to benzothiophene scaffolds to diversify the biological
activity of ER ligands.34-41 To successfully generate the potent, oral SERD, 9, we designed a unique
scaffold substituted with an acrylate containing side chain. 33 Co-crystal structures of SERMs, containing
the archetypical SERM 2-phenoxyethylamino side chain, bound to ERα reveal the key salt-bridge
interaction between the SERM side chain amino group and Asp-351, and we hypothesized that retaining
this interaction and extending the aliphatic side chain would displace H12, expose its hydrophobic surface
and result in ERα degradation.42-45 Molecular modeling of putative B-SERDs using the co-crystal structure
of 4 bound to ERα (pdb 5ACC) supported this hypothesis; and suggested that favorable interactions
targeted for 9 with the two hydrophobic cavities formed by Leu 384 and Leu 428 (pdb 1R5K) could be
maintained in a B-SERD (Figure 2).
The choice of constrained basic side arm for a B-SERD ranges from the pyrrolidine, piperidine,
and azepane rings found in SERMs and SERDs, to the azetidine ring found in a SERD reported in 2019,
after completion of our lead optimization campaign (3b).46 Effective side arms would presumably need
to maintain a salt bridge or H-bond with Asp-351, with the strength of this interaction influenced by amine
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
basicity. In addition, SERMs are well known to undergo oxidative metabolism leading to metabolites
formed from N-desalkylation. Since both characteristics are expected to be strongly influenced by ring
size (hybridization and ring strain) and electron-withdrawing substituents, calculations were performed
on candidate side arms using DFT molecular orbital calculations at RI-MP2/6-311++G**//B3LYP/6-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
31+G** (Figure 2 and Supplemental Information). The influence of ring size is to decrease nitrogen
basicity as ring size contracts, and to further decrease basicity with electron-withdrawing group
substitution. In general, a similar trend is seen for ionization energy, indicating that azetidine side arms
are less susceptible to oxidation, although a cyclopropyl substituent on the azetidine ring stabilizes the
radical cation formed on oxidation. The relative energy for H atom abstraction (dHA) α to N is an indicator
of susceptibility to Phase 1 oxidation, potentially leading to N-desalkylation (Figure 2D). Again, the
azetidine ring carbons are predicted to be less readily oxidized, especially in the fluoro-substituted
derivatives. Although calculated electronic contributions can be overwhelmed by ligand binding site
interactions and microenvironment, the calculated trends encouraged exploration of azetidine side arms.
Figure 2. Structure Design. H-bonding of the amine side chain to Asp-351 in the ERα ligand-binding pocket should
allow engagement of ring substituents with hydrophobic pockets formed in the region of Leu-384 and Leu-428
(increasing affinity), whilst displacing H12 (causing ERα degradation). The design of the amine side chain using a
ACS Paragon Plus Environment

Page 7 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
conformationally restricted heterocycle considered: the ability to interact with H12 (A); the amine basicity (B); and
susceptibility to oxidation (C, D). DFT molecular orbital calculations of proton affinity (dPA) (B), ionization energy
(
dIE) (C), and H-atom abstraction (dHA) (D) were normalized relative to the calculated free energy for the piperidine
side chain: dHAr corresponds to heterocyclic ring-C oxidation; dHA refers to oxidation of the alternate carbon. The
R group in A is modeled by H in calculations.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Chemistry
Scheme 1. Synthetic routes for precursor synthons
To construct the 2-keto-benozothiophene core required by our structure design, commercially
available 4-methoxycinnamic acid was used, followed by cyclization to afford 3-chloro-6-
methoxybenzo[b]thiophene-2-carbonyl chloride (10) (Scheme 1). The acyl chloride was converted to a
Weinreb amide, which was subsequently reduced to the corresponding aldehyde (12) with DIBAL-H at -
o
7
8 C. A variety of Grignard reagents were shown to react with the aldehyde to afford diverse secondary
alcohols 13a-d that were oxidized to generate phenyl ketones 14b-e. Meanwhile, ketones 14a, 14f, 14h,
o
1
4i were obtained from the reaction of Weinreb amide with Grignard reagents at 0 C. Compound 14g
was obtained from Grignard reagent reacting with acyl chloride 10. The deprotection of the methoxy
33
group of compounds 14 was performed with Lewis acid BF
3
·Me S in an ice bath. The resulting phenol
2
was protected with tetrahydropyran (THP) to afford the primary precursor synthons 16a-i. To obtain the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
azetidine synthon required by our structure design, we started with t-butyl 3-(hydroxymethyl)azetidine-1-
carboxylate 17, installing the fluoro substitution using tetrabutylammonium fluoride under reflux to
generate compound 19. The azetidine synthon 20 was obtained by deprotection using 4M HCl; and the
pyrrolidine synthon 24 was obtained in a similar way from 21 (Scheme 1). For construction of the side
chain, 2-(4-(2-bromoethoxy)phenoxy)tetrahydro-2H-pyran 26 was obtained from 4-((tetrahydro-2H-
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
pyran-2-yl)oxy)phenol 25. Coupling of 26 to the appropriate amine under basic conditions (K
2
CO
3
) gave
the THP-protected side chain: 27a-f (Scheme 2). Deprotection of the THP protecting group afforded
compounds 28a-f that underwent S
N
Ar reaction with synthons 16a-i to produce 29a-o, which gave final
products 30a-o after deprotection.
Scheme 2. Synthetic routes for candidate piperidine, azetidine, and pyrrolidine B-SERDs
Scheme 3. Synthetic routes for candidate reverse-azetidine B-SERDs
ACS Paragon Plus Environment

Page 9 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
The 3-oxy-azetidine, or “reverse azetidine”, side arm was derived from reaction of t-butyl 3-
iodoazetidine-1-carboxylate
with
4-(benzyloxy)phenol
31
to
afford
t-butyl
3-(4-
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
(
benzyloxy)phenoxy)azetidine-1-carboxylate 32 (Scheme 3). Deprotection in trifluoroacetic acid (TFA) to
give 33 was followed by alkylation with 1-bromo-3-fluoropropane or (1-ethoxycyclopropoxy)
trimethylsilane under basic conditions (NaH), to give 34a, and 34b, respectively. Deprotection was
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
performed under Pd/C and H to afford compounds 35a-b, which were coupled with the appropriate
2
precursor synthon (16a,b,d,f,g) to yield 36a-f using the procedure shown in Scheme 2. Deprotection of
THP under mild acidic conditions gave 37a-f.
Biological Testing
Our objective in structure optimization was to develop novel, potent, orally bioavailable B-SERDs
with efficacy comparable to fulvestrant (2). Since the ultimate goal was an oral SERD suitable for
treatment of brain metastases in endocrine-resistant ER+ breast cancer, lead optimization would need
to be driven by efficacy in endocrine-resistant, ER+ breast cancer cell lines. A development lead would
need to demonstrate efficacy in breast cancer cells bearing mutations in ESR1, brain penetration, and
cause regression of endocrine-resistant tumors in a mouse xenograft study.
Optimization of the B-SERD scaffold was driven by in vitro assays in 2D and 3D breast cancer
cell cultures. The primary objective, to develop new oral SERDs for treatment of MBC resistant to
endocrine therapy, required use of breast cancer cell lines modeling TAM and AI resistance. The
MCF7:TAM1 cell line models resistance to TAM and AIs, having been developed by long-term exposure
of parental, endocrine-dependent MCF7:WS8 cell cultures to the active metabolite of TAM, 4OH-TAM
(1b, Figure 1) with concurrent long-term estrogen deprivation (LTED); whereas the MCF7:5C cell line
was developed from LTED of MCF7:WS8 cell cultures.34, 47 Both cell lines undergo estrogen-independent
growth and are insensitive to treatment with 1b, in contrast to the parent MCF7:WS8 cell line, growth of
which is dependent on estrogens and inhibited by 1b. In addition to these cell lines, we evaluated the
T47D:Y537S and T47D:D538G cell lines obtained from CRISPR-Cas9 manipulation of T47D-WT cells
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
as described previously.48 ESR1 mutations are prevalent in ER+ MBC after AI therapy and the Y537S
49
and D538G mutations are associated with AI resistance and more aggressive disease. Not only do the
T47D ESR1 mutant cell lines provide an additional model of resistance, the T47D cell line itself provides
a different ER+ tumor background.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In our previous development of oral SERDs, exemplified by 9, we used oral SERD 3a as a
benchmark; however, this oral SERD failed in clinical trials; thus, we now use fulvestrant, 2, as a
benchmark, since it is currently the only clinical SERD. It should be noted that: 2 demonstrates high
efficacy and potency in cell cultures and cell-derived xenografts (CDX); and the severe pharmacokinetic
limitations experienced in the clinic, which limit efficacy, are not recapitulated in preclinical models.
Therefore, a desired B-SERD will match the high potency and efficacy of 2 in cell lines and CDX, but in
contrast to 2, will demonstrate oral bioavailability and brain exposure. All cell lines used for B-SERD
optimization are sensitive to growth inhibition by 2. In both cell cultures and in vivo measurements of
target engagement and side effects, we have also compared against the SERDs: 3b and 9; and the
SERM/SERDs, 5 and 7.
Cell viability assays were performed in 2D monolayer cell culture and 3D spheroidal cell cultures
to measure the ability of B-SERDs to inhibit growth of both endocrine-resistant and parent cell lines. Full
concentration-response curves were obtained for B-SERDs and for 2 to derive potency for inhibition of
cell growth in 2D monolayer cultures of MCF7:WS8 and MCF7:5C cell lines, measuring cell DNA content.
The maximum efficacy for inhibition is reported relative to vehicle (0%) and cell medium only (100%),
such that the maximum efficacy of 2 was measured as 66% and 48% inhibition of cell growth in parental
and endocrine-resistant breast cancer cell lines, respectively. Inhibition of estrogenic activity, required for
SERD activity, was measured in the endocrine-dependent parental cell line in competition with E , using
2
a transient ERE-luciferase transcriptional reporter and compared to relative binding affinity (RBA) to ERα
using a radioligand binding assay. The effect of treatments on ER protein level was measured using in-
cell westerns (ICW) in MCF7:WS8 cells and confirmed by western blots in presence and absence of a
proteasome inhibitor to show proteasomal degradation. While monolayer cell culture is higher throughput
ACS Paragon Plus Environment

Page 11 of 72
Journal of Medicinal Chemistry
1
2
3
and allows multi-plate measurement of ER content using ICW, 3D spheroids provide a more
physiologically relevant cell model, more closely mimicking the increased cell-cell signaling and hypoxic
core, observed in solid tumors.50 Cell viability for 3D spheroid cultures was measured in parent
MCF7:WS8 and endocrine-resistant MCF7:TAM1 cell cultures as well as in three T47D cell lines,
including those with Y537S and D538G mutations in ESR1. The novel B-SERD showing a superior PK
profile was; 1) compared with 2 in an MCF7:TAM1 orthotopic CDX mouse model; 2) compared with 2
and 7 in the assessment of unwanted uterotrophic effects in juvenile rats; and 3) compared with 9 in
assessment of target engagement (ERα immunoassay) in uterus and ovaries of intact female mice.
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Results
The benchmark SERD, 2, inhibits growth of endocrine-resistant and parental cell lines in 2D
cultures with high potency (pIC50 = 8.8 - 9.2) and with observed reduced maximal efficacy in the
endocrine-resistant MCF7:5C cell line (Table 1). Of the B-SERDs with a pyrrolidine side arm, 30l showed
superior potency and efficacy to 2 in endocrine-dependent and -independent cell cultures (Table 1). The
enantiomer, 30m was marginally inferior to 30n in both cell lines. As also observed for 2, all pyrrolidines
studied (30l-n) lost significant efficacy in endocrine-resistant MCF7:5C cells. In contrast to 2 and to the
pyrrolidines, the piperidine 30o did not inhibit growth of MCF7:5C cells. MCF7:5C cells are tamoxifen
resistant and are cross-resistant to the SERM raloxifene 6 (Table 1). The very similar cellular phenotype
induced by 6 and 30o strongly suggests that 30o is a SERM and that cross-resistance in MCF7:5C cells
also extends to this SERM.
Moving from a pyrrolidine to an azetidine side arm led to no loss of efficacy and potency in growth
inhibition of MCF7:WS8 cells (9.0 < pIC50 < 10.4; Table 2). Comparison of the 2-Me-phenyl substituted
series of azetidine ligands bearing different substitutions on the azetidine ring, 30a, 30b, 30d, showed
identical efficacy, with 30b having the higher growth inhibition potency of the parental cell line. As we saw
for 30o (and 6) high potency in the parental cell line can translate to total loss of efficacy in the MCF7:5C
cell line for a SERM (Table 1). The unadorned 4-membered azetidine ring of 30a confers potent SERD
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
activity in inhibiting growth of both cell lines, but significantly reduced efficacy in MCF7:5C cells.
Structurally, 30a resembles a SERM; however, the observed activity shows that tamoxifen-resistant
MCF7:5C cells that are cross-resistant to the piperidine SERMs, 6 and 30o, are not cross-resistant to
30a, leading to speculation that 30a may fall in the SERM/SERD classification. Regardless, the lack of
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
cross-resistance supports the selection of the azetidine side arm for further exploration.
ACS Paragon Plus Environment

Page 13 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Comparison of the nine 3-fluormethyl azetidine derivatives, 30c-k, allows comparison of the effect
of different substitution patterns of the benzoyl ring; however, all derivatives potently inhibited growth in
MCF7:WS8 cells (8.9 < pIC50 < 9.8). In endocrine-resistant cells, many derivatives showed a small loss
of potency, and a loss of efficacy; with the exception of 30g that lost efficacy in MCF7:5C cells. We
speculate that the provision of a second phenolic group in 30g and 30h, in addition to the benzothiophene
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
phenol, facilitates an alternative binding mode that is more akin to a SERM, and MC7:5C cells are, in
general, resistant to SERMs. The meta substitution in 30j is also not a preferred substitution as indicated
by efficacy in MCF7:5C cells. These observations support the importance of the o-methyl group in
exploiting the small hydrophobic pockets created by Leu-428 and Leu-438, explored in optimization of
the SERD 9, in maintaining potency in the endocrine-resistant cell line.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
We next explored derivatives with a “reverse azetidine”, or 3-oxy-azetidine, side arm. This
modification moves the conformational lock of the azo-ring closer to the more rigid benzothiophene core.
To induce the displacement of H12, which underlies the high potency and efficacy observed with
azetidine ligands (Table 2), a longer and flexible fluoropropyl chain was conjugated to the azetidine
nitrogen. Given the potential for the N-cyclopropyl substituent to increase stability to Phase 1 metabolism
(
Figure 2), compound 37f was prepared; although this ligand was potent and effective in MCF7:WS8
cells, disappointingly, it showed a loss of efficacy in tamoxifen-resistant cells (Table 3). All reverse
azetidines, 37a-f, were potent (9.3 < pIC50 < 10.2) and efficacious inhibitors of MCF7:WS8 cell growth.
As with the 3-fluoromethyl azetidine series, the predicted reduced basicity of the N-fluoropropyl
azetidine nitrogen did not influence potency or efficacy. Having established the efficacy of the putative
B-SERDs in inhibiting growth of endocrine-resistant and parental cell lines, it was essential to measure
ERα levels to confirm the functional identity of these compounds as SERDs. Potency towards
degradation of ERα was studied in MCF7:WS8 cells measured by ICW. All B-SERDs that induced growth
inhibition also caused loss of ERα with potency comparable to 2 (Table 4). The SERM, 30o, did not
induce ER degradation (Figure S1). The cyclopropylidene derivatized reverse azetidine, 37f, showed low
efficacy in the ICW assay and therefore despite predicted stability against metabolic N-desalkylation (see
Figure 2) was not considered as a development candidate. Interestingly, the unadorned azetidine (30a)
ACS Paragon Plus Environment

Page 15 of 72
Journal of Medicinal Chemistry
1
2
3
and fluoromethyl pyrrolidine (30n) derivatives also manifested reduced efficacy, as did the bisphenolic
derivative, 30h. These compounds also showed lower efficacy in growth inhibition of MCF7:5C cells.
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
SERDs and SERMS, by definition, antagonize the actions of E
binding of a repressed ER transcriptional complex to the ER response element (ERE) causing inhibition
of selective gene transcription. All B-SERDs tested antagonized ERE-luciferase induction by E with low
nanomolar potency, with 30h, 30m, and 37f excluded from further consideration, based on lower potency
2
in breast cancer cells, by inducing
2
(
Table 4). For selected compounds, RBA was measured to confirm sub-nanomolar affinity for ERα (Table
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
4
). The selectivity for ERα/ERβ given by RBA measurements was less than fourfold for all B-SERDs
tested (Table S1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Based upon the foregoing data collected in 2D cell cultures, several B-SERDs were excluded
from further study; however, many examples remained with high potency and efficacy, and clear evidence
ACS Paragon Plus Environment

Page 17 of 72
Journal of Medicinal Chemistry
1
2
3
of target engagement. The success of the 2,6-dimethyl,4-halo acrylate SERDs (9) in our previous
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
33
preclinical studies biased us towards these derivatives; therefore, 30f, 30m, and 37d, were prioritized
for a preliminary study of oral brain bioavailability. The three B-SERDs were tested delivered orally at
1
00 mg/kg to C57/BL6 mice with plasma and brain concentrations measured by LC-MS/MS at 30 and
20 min and compared to fulvestrant delivered by the standard s.c. route (Table 5). B-SERDs showed
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
the desired brain bioavailability; for example, the concentration of 30f in brain at 0.5 and 2 h was 600-
750 nM and that of 37d exceeded 1 µM, whereas the concentration of 2 in the brain was below LOQ at
30 min. The 3-oxyazetidine side-armed B-SERD, 37d, gave substantially higher plasma and brain
concentrations compared to the analogues bearing either azetidine (30f) or pyrrolidine side-arms (30m).
The pyrrolidine, 30m, was not studied further.
B-SERDs 30f and 37d were compared with congeners, 30i and 37b, respectively, and to SERD
2 in 3D cultures following treatment for 14 days (Figure 3A). In the parental, MCF7:WS8 cell line, all
SERDs (10 nM) were equally effective in inhibiting growth of spheroids as shown by spheroid size and
viability measured by ATP content (Figure 3A,B). However, in the tamoxifen-resistant MCF7:TAM1
spheroids, significant differences were observed between treatments, with 2, 30f and 37d having greater
efficacy. Growth of MCF7:TAM1 cells is endocrine-independent and tamoxifen-resistant, and since this
is an LTED cell line, it models resistance to AI; however, it does not harbor mutations in ESR1, which are
known to be associated with acquired AI resistance. The efficacy of B-SERD 37d was compared to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
SERDs (2, 3b) and SERM (1b) in T47D cell lines expressing only mutant ERα (TYS = Y537S;TDG =
D538G) or WT ERα.48 B-SERD 37d was equally as effective as the SERDs fulvestrant (2), GDC-0927
(
3b), and as the SERM 1b (Figure 3C).
Western blots supported the observations on B-SERD-induced ER degradation in cell cultures
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
made using ICW, and using co-treatment with the proteasomal inhibitor, MG-132, western blots also
confirmed the role of the proteasome in degradation (Figure 4A). Representative western blots
demonstrating ERα degradation from treatment of the parent cell line are shown in Figure 4B (and Figure
S2), providing a comparison of 37d with the structurally related acrylate, 9a, clinical SERD, 2, and
SERM/SERD, 5. We extended these in vitro observations to examine the effects of oral administration of
the B-SERD, 37d, compared to the congenic acrylate SERD, 9a, in intact female mice (Figure 4C). In
these experiments, animals were sacrificed, and tissues flash frozen, at either 5 or 7 hr after oral drug
administration. In the acute treatment paradigm, mice received only one dose before sacrifice; whereas,
in chronic treatment, mice were treated with drug for 3 days. Treatment with 37d or 9a (50 mg/kg) under
all conditions led to significant (p < 0.01 control vs treated groups) loss of ERα, as quantified by western
blot of tissue homogenates (Figure 4C). After single dose administration, degradation of ERα in ovaries
was significantly greater with 37d versus acrylate SERD 9a. After multiple doses, degradation of ERα in
the uterus was significantly greater with 37d versus acrylate SERD 9a.
The juvenile rat model is widely used for assessment of uterotrophic activity of ER ligands, since
the young female rat is highly sensitive to ER ligands, but without a background of high levels of
circulating endogenous estrogens. Ethinylestradiol (EE2) was administered as a positive control, yielding
a significant increase in uterine weight (Figure 4D). There was no significant uterotrophic effect of the
SERM/SERD, 7, nor the SERDs, 2 and 37d, versus the vehicle control.
ACS Paragon Plus Environment

Page 19 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Full PK analysis for 37d was conducted in CD-1 mice to determine rigorous oral bioavailability
parameters, demonstrating absolute bioavailability of 22%, a half-life of 4.5 h, and a brain/plasma ratio
for Cmax and AUClast of 1.54 and 1.26, respectively (Table 6 and Figure S3).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 6. Pharmacokinetics of 37d
Dose
T
1/2
T
max
C
max
AUClast
AUCInf
AUC%Extrap
MRTInfobs
(hrs)
AUClast/D
(*ng/mL)
33.6
F
(mg/kg) (h)
(h)
(ng/mL) (h*ng/mL)
(h*ng/mL) obs (%)
(%)
21.9
Plasma 50 p.o.
4.48
1.00
1.00
413
1678
2106
766
2292
3450
778
26.8
6.05
Brain
50 p.o.
635
Plasma 5 i.v.
1.68
1592b
1.52
1.07
153
a) PK parameters from measurements at 0.25, 0.5, 1, 2, 4, 8 and 24 h (N=5). b) C .
0
Efficacy in MCF7:TAM1 spheroids is predictive of efficacy in xenografts and therefore 37d was
tested in breast cancer tumors that had been established by injecting MCF7:TAM1 cells into the
mammary fat pads of nude mice. Tumors in mice treated with vehicle or 1 continued to grow over the
treatment period, whereas in mice treated with 2 (s.c.) or 37d (p.o.) tumor regression was observed
(
Figure 5A). The observations with 1a reinforced that these breast cancer tumors are tamoxifen resistant.
Analysis of individual tumors is shown for SERD treatment compared to the no treatment group, to show
statistical significance in tumor regression (Figure 5B).
Figure 5. Effect on tumor growth. (A) Tumor growth or regression of MCF7:TAM1 cell-derived xenografts grown
2
to 0.3 cm before initiating daily treatment with TAM (1a), 37d (100 mg/kg oral gavage) or 2 (5 mg s.c.). (B) Individual
tumor % area change after 4 weeks treatment: **** p<0.0001 versus no treatment (NT) group, by one-way ANOVA
with Dunnett’s test.
ACS Paragon Plus Environment

Page 21 of 72
Journal of Medicinal Chemistry
1
2
3
4
Discussion
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The search for an orally, brain bioavailable B-SERD, started with our unique benzothiophene
scaffold developed for the acrylate SERD (9b), currently in clinical trials. Modification of this scaffold on
the benzoyl ring, and with a basic side chain yielded a number of interesting ERα ligands. The piperidine
derivative, 30o, demonstrated SERM-like characteristics; whereas most derivatives that substituted the
piperidine side arm with a pyrrolidine, azetidine, or reverse azetidine, demonstrated SERD activity. The
azetidine 30a, a direct analogue of 30o, inhibited growth of a tamoxifen-resistant cell line and lowered
ERα levels; however, the observed reduced efficacy indicates a mixed SERM/SERD activity. Several
other derivatives, such as 30g, and the cyclopropyl azetidine 37f, showed activity profiles worthy of further
study; however, our objective was to select a SERD with comparable potency and efficacy to the clinical
SERD 2, but with oral and brain bioavailability. The compound selected, 37d, bears a novel reverse
azetidine side arm.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The success of 2 in treating advanced, metastatic ER+ breast cancer provides the rationale for
oral SERD development; therefore, efficacy in endocrine-resistant breast cancer cells is crucial. In parent
cell lines, pyrrolidine, azetidine, and reverse azetidine derivatives with potent and high efficacy SERD
activity demonstrated high potency and high efficacy inhibition of cell growth. However, almost all
compounds demonstrated reduced efficacy in growth inhibition of the endocrine-resistant cell line,
including 2. In this cell line, derivatives containing fluoromethyl pyrrolidine, fluoromethyl azetidine, or N-
fluoropropyl azetidine, bearing benzoyl rings substituted o- or p- with fluoro or methyl groups,
demonstrated higher potency (pIC50 = 9.3 ± 0.4) and identical efficacy to 2 (Emax = 48 ± 9%). With no clear
bias for benzoyl substitution, analogues with the 2,5-dimethyl-4-fluoro substitution, contained in clinical
candidate 9b, were selected for progression. Of these, 37d demonstrated markedly superior oral
bioavailability and excellent brain bioavailability.
Compound 37d, fulfilled the requirements for a brain bioavailable B-SERD with efficacy in
multiple ER+ breast cancer cell lines, including two endocrine-resistant cell lines, and two ESR1 mutant
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
cell lines, in 2D and 3D cultures. This B-SERD showed equivalent potency and efficacy to the clinical
SERD 2 both in vitro and in vivo (Figure 3-6), but importantly was both orally bioavailable and brain
penetrant, in contrast to 2. As measured by western blotting for ERα, co-administration of a proteasome
inhibitor blocked the actions of both 2 and 37d, demonstrating that 37d is indeed a SERD, inducing
proteasomal degradation of ERα. Almost complete degradation of ERα in gynecological tissues was
observed on oral administration of 37d to female mice and 37d was devoid of uterotrophic effects.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A
B
C
1
.0
1.0
1
0
0
.0
.8
.6
37d
2
0.8
0.8
IC50 = 0.55 nM
IC50 = 0.49 nM
0
.6
IC50 = 1.0 nM
0
0
0
0
.6
.4
.2
.0
0.4
0
0
.2
.0
0.4
-11
-10
-9
-8
-7
-6
-11
-10
-9
-8
-7
-6
-11
-10
-9
-8
-7
-6
log[SERD] (M)
log[SERD] (M)
log[SERD] (M)
Figure 6. Summary of B-SERD in vitro activity. 2D growth inhibition in MCF7:WS8 (A) and MCF7:5C (B) cells
as a result of treatment with 37d (green) or 2 (orange). Data normalized to vehicle, DMSO (1.0) and no cells (0.0)
shown as mean ± SEM, from three biological and analytical replicates. (C) ER level following treatment for 24 hr
with 37d or 2 evaluated by ICW. Data corrected to vehicle (1) and 1 µM 2 shown as mean ± SEM, from three
biological and analytical replicates.
During completion of this work, a paper was published reporting the substitution of the piperidine
ring of the SERM acolbifene (EM-652) with a pyrrolidine ring (38),45 which can be seen as a similar
substitution to that of the piperidine ring in 30o with pyrrolidine in 30l-n. Interestingly, R-38 demonstrated
SERD-like properties, whereas S-38 did not lower ERα and showed SERM-like properties, which was
explained by stereospecific interactions in 38/ERα co-crystal structures causing H12 destabilization only
for the R-isomer. In our benzothiophene series, reported herein, 30o showed SERM-like properties;
however both 3-fluoromethyl-pyrrolidine isomers (30m,n) showed SERM/SERD or SERD-like properties.
The R-38/ERα structure (pdb 5UFX) was used to model the potential interactions of 30f and 37d with
ACS Paragon Plus Environment

Page 23 of 72
Journal of Medicinal Chemistry
1
2
3
ERα: the H-bond between Asp-351 and the azetidine N forces the fluoropropyl arm to interact with and
destabilize H12 (Figure 7).
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In various assays, we compared the B-SERD, 37d, with: the analogous acrylate SERD 9a; SERD
2
; the SERM/SERDs 5 and 7; and the fluoromethyl azetidine SERD 3b. The search for an oral SERD
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
was driven initially by 8b a derivative of the SERM, TAM, which appends an acrylate side arm in place of
the SERM amine side arm. This led to the acrylate SERD 3a that includes a modification of the tamoxifen
scaffold. Similarly, the amine side arm of the SERM acolbifene was replaced with a methyl pyrrolidine in
R-38; whereas SERD 3b contains a modification of the acolbifene scaffold and a fluoromethyl azetidine
side arm. Recent reports on R-38 and 3b posit that ER degradation is subservient to silencing of the ER
transcriptional complex in growth inhibition of breast cancer cells and xenografts by SERDS.45, 51 In
particular, detailed ChIP-seq, ATAC-seq, and RNA-seq analysis of the acrylate SERD 3a was shown
inferior to 3b in this respect. Similar sequencing assays to compare the B-SERD 37d with the directly
analogous acrylate SERD 9b could provide further mechanistic insight and direction for drug
development.
Figure 7. B-SERD docking to ER. B-SERDs 30f and 37d have a common basic side-arm motif: 3-fluoro-N-(2-
oxyethyl)propan-1-amine. The structure design envisaged occupation of the hydrophobic pockets formed by
leucines 348, 428, 525, and isoleucine 424 by the substituted benzoyl ring, allowing a salt bridge interaction
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
between the amine side arm amine and Asp-351 in the ERα ligand binding site. B-SERD 37d (green) was docked
to the ligand binding domain of ERα obtained from the co-crystal structure with the SERD R-38 (magenta) (pdb
5UFX) confirming the proposed binding site interactions leading to destabilization of H12.
Conclusion
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
B-SERDs were identified with very high potency and efficacy in models of treatment-resistant
breast cancer, building out from the unique benzothiophene scaffold optimized in our previous pursuit of
acrylate SERDs.33 The novel N-fluoropropyl 3-oxyazetidine side armed B-SERD, 37d, was selected as
a development candidate, based on the combination of sub-nanomolar potency and superior PK
characteristics. The preclinical data support the development of B-SERDs for treatment of patients with
metastatic ER+ breast cancer, including those with brain metastases.
Experimental Section
Cell Lines and Culture Conditions. The parental, ET-sensitive cell line, MCF7:WS8, was
derived from a single-cell clone by the Jordan group. This cell line was maintained in phenol red-
containing RPMI-1640 medium supplemented with 10% FBS, 1% antibiotic-antimycotic, 1% glutamax,
and 10 ng/mL insulin at 37 °C and 5% CO . The MCF7:5C cell line was also derived from a single cell
2
clone by the Jordan group.47 These cells were maintained in phenol red-free RPMI 1640 medium
supplemented with 10% charcoal-dextran treated FBS, 1% antibiotic-antimycotic, and 1% glutamax at
37 °C and 5% CO2. The MCF7:TAM1 cell line was generated through long-term exposure to increasing
concentrations of 1b until resistance was established (>25 passages).34 MCF7:TAM1 cells are
maintained in 1b (1 µM) and phenol red-free, stripped RPMI-1640. The T47D:TYS, T47D:TDG and parent
WT T-47D cell lines were a kind gift from David Shapiro (UIUC) and were cultured and maintained as
previously described.48
DNA Content Assay. MCF7:WS8 cells were stripped of estrogens for 2 days prior to plating each
experiment by changing media to phenol red-free RPMI1640 and 10% FBS. Cells were seeded in a 96-
well, clear, flat bottom microplate at a density of 5000 cells/well and treated with either 0.01% (v/v) DMSO,
2
1 nM E , or compound of interest. All compounds were stored dissolved in DMSO and diluted to the
ACS Paragon Plus Environment

Page 25 of 72
Journal of Medicinal Chemistry
1
2
3
specific treatment concentration through serial dilution. On Day 5 (MCF7:WS8) or Day 7 (MCF7:5C),
media was removed and cells were lysed through hydrolysis at -80°C overnight. DNA content was
determined by Hoechst 33258 dye in TNE buffer (1 mg/mL Hoechst in TE buffer + 2M NaCl).
Fluorescence signal was measured using a Synergy Neo (BioTek).
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ER ICW Assay. MCF7:WS8 cells were stripped using phenol red-free media and stripped FBS
4
for 2 days prior to plating at 2.5 x 10 cells/well in black, clear bottom 96-well microplate. Cells were
incubated for 48 hrs prior to treatment for 24 hrs. Fixation, detection of ER (H10, Santa Cruz
Biotechnologies) and imaging were performed per LI-COR manufacturer’s protocol using the In-Cell
Western™ Assay Kits and LI-COR Odyssey SA imaging system. IRDye 800CW (anti-rabbit) signal was
normalized to CellTag 700 stain.
ER Western Assay. Cells were stripped of estrogens for 3 days prior to plating. Cells were grown
to 80% confluency in a 6-well plate and treated with vehicle or MG-132 (1 µM) for 30 minutes followed
by 2 hours of SERD treatment (10 nM). For comparison of SERDs and B-SERDs, similar experiments
were performed without prior treatment with MG-132. ER protein content was quantified by western blot
where ERα (antibody Cell Signaling 8644) was normalized to actin or GAPDH. Blots were quantified
using LI-COR Odyssey SA imaging system.
Spheroid Growth Assay. The cells were plated at a density of 1000-1500 cells/well in Corning®
96-well black, clear round bottom, ultra-low attachment spheroid microplates and grown in the absence
of treatment for 1-3 days. Spheroids were then treated with 2X treatment media following the removal of
half of the media (100 µL) from each well. Treatment at a 1X concentration was repeated every 2-3 days
for 14 days. CellTiter-Glo® 3D Cell Viability Assay protocol was used to determine growth inhibition of
the spheroids, as per manufacturer’s instructions. Luminescence signal was read using a Synergy Neo
(
Biotek). Data was normalized to blank (media with CellTiter Glo 3D reagent).
Binding Affinity Studies. Binding affinities were also determined by a competitive radiometric
3
binding assay using 2 nM [ H]estradiol as tracer (PerkinElmer, Waltham, MA) and full-length purified
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
human ERα (Pan Vera/Invitrogen, Carlsbad, CA), as reported previously.52, 53 The RBA values were
calculated using the following equation: IC50 estradiol/IC50 compound × 100.
Estrogenicity Assay. MCF7:WS8 cells were grown in phenol red-free RPMI1640 and stripped
5
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
FBS for 3 days prior to plating. Cells were then plated at a density of 1X10 cells/well in clear, flat bottom,
48-well plates and incubated for 24 hours. The cells were co-transfected with 5 μg of the pERE-luciferase
plasmid per plate, which contains three copies of the Xenopus laevis vitellogenin A2 ERE upstream of
firefly luciferase and 0.5 μg of pRL-TK plasmid (Promega, Madison, WI) containing a cDNA encoding
Renilla luciferase. Transfection was performed for 6 hrs using 2 µL/well Lipofectamine 2000 transfection
reagent (Invitrogen) in Opti-MEM media. Cells were then treated with test compounds. Luciferase activity
was measured after 18 hrs of treatment using the Dual Luciferase assay system (Promega) with Synergy
Neo (BioTek).
Animal Experiments. The Animal Care and Use Committee of the University of Illinois at Chicago
approved all animal procedures. Animal care adhered to the National Institutes of Health Guide for the
Care and Use of Laboratory Animals.
Juvenile Rat Study. 12 days old female Sprague Dawley rats were purchased from Envigo, USA.
Sucklings were housed with foster nursing dams (6 per dam). After one week of acclimation period, at
age 19 days, treatment was initiated by administering: 37d or 7 (Sigma, USA) (p.o. 10 mg/kg suspended
in 2% Tween 80, 0.5% methylcellulose); EE2 (Sigma, USA) (p.o. 0.1 mg/kg in peanut oil); 2 (s.c. 2 mg/kg
in peanut oil). 24 hours after last of 3 daily doses, animals were euthanized with CO followed by
2
exsanguination from the abdominal aorta. Uteri, including uterine horn were carefully excised and
weighed after removing fat and mesentery. Uteri were cut in the sagittal plane into two equal halves: one
half was fixed in PBS (pH 7.4) buffered 4% PFA for 24 h, rinsed with water, washed twice with 70%
ethanol, paraffin embedded and processed for hematoxylin eosin staining and imaging.
Mouse Gynecological Tissue Immunoassay. Female C57BL/6J mice (8 W) were purchased
from the Jackson Laboratory (Bar Harbor, ME) and treated at 9-12 W. Mice were housed in groups of 3-
ACS Paragon Plus Environment

Page 27 of 72
Journal of Medicinal Chemistry
1
2
3
4
5
5
with food and water available ad libitum. Dosing: 9a (p.o. 50 mg/kg in PEG 400, 0.5% carboxymethyl
cellulose, polyvinylpyrrolidone and Tween-80; 9:90:0.5:0.5 v/v); 37d (p.o. 50 mg/kg in PEG 400, 10% (2-
hydroxypropyl)-beta-cyclodextrin(HPCD); 1:9 v/v); the vehicle control group received PEG/HPCD vehicle.
In the acute experiment, mice (n=5) were given a single administration of drug and euthanized 5 h later.
In the chronic experiments, mice were administered drug once daily for 3 days and euthanized 5 h (n=5)
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
or 7 h (n=6) after the last drug administration. Mice were euthanized by CO
2
inhalation followed by
o
decapitation. Ovaries and uterus were immediately removed and frozen on dry ice and stored at -80 C
until being processed for western blots. Tissue samples were homogenized in 200 μl of lysis buffer (Cell
Signaling Technology, Danvers, MA; 20 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1%
Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na
3
VO , 1 µg/ml
4
leupeptin, and 1 µg/ml aprotinin) containing the Complete protease inhibitor cocktail (Roche, Indianapolis,
o
IN) and clarified by centrifugation for 15 min at 10,000 rpm at 4 C. Protein concentrations were
determined using the BCA Protein Assay Kit (Thermo Fisher Scientific). Equal amounts of protein (20 μg)
were subjected to SDS-PAGE on Novex 10% Tris-glycine gels (Thermo Fisher Scientific) and transferred
to PVDF membranes. Membranes were blocked with 5% bovine serum albumin (BSA) in TBST (25 mM
Tris-HCl, 137 mM NaCl and 0.1% Tween-20) and then incubated with primary antibodies overnight at
4°C (anti-ERα, 1:1000, Millipore Sigma, #06-935; anti-β-actin, 1:10,000; Millipore Sigma, #A5441).
Membranes were incubated with IRDye donkey anti-rabbit and donkey anti-mouse secondary antibodies
(
LI-COR Biosciences, Lincoln, NE, #925-32213 and #925-68072) at room temperature. Blots were
imaged using the Odyssey Fc Imaging System (LI-COR Biosciences). Band intensities were determined
using with LI-COR Image Studio software and ERα normalized to β-actin.
Mouse Xenograft Experiments. MCF7:TAM1 tumors were established in 4–6 week old
ovariectomized athymic female nude mice (Harlan Laboratories) and E was administered via silastic
2
capsules (1.0 cm) implanted subcutaneously between the scapulae as previously described.54, 55 SERD
37d and SERM 1a were administered (p.o. 100 mg/kg/day) in 10% HPCD:PEG-400 (9:1 v/v) solution,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 72
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
whereas 2 was administered (s.c. 5 mg/week) in sesame oil. Tumor cross-sectional area was determined
weekly using Vernier calipers and calculated using the formula (length/2) × (width/2) × π as described
previously.54, 55 Mean tumor area was plotted against time (in weeks) to monitor tumor growth.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Pharmacokinetic Experiments. PK screening was conducted in female 3M C57BL/6 mice at
two time points (N=3), collecting plasma and brain tissues from perfused mice post-mortem, after gavage
administration of 30f, 30m, or 37d in HPβCD:PEG-400, or s.c. administration of 2. Standard curves were
established in the corresponding biological matrix using a common internal standard and optimization of
analyte measurement by LC-MS/MS in the MRM mode. Further PK measurements were performed by
Pharmaron Inc. at ten time points, administering 37d in solution (p.o. 50 mg/kg in water and 10% SBE-
β-CD:PEG-400 9:1 v/v). Working solutions were made by serial dilution of analyte in 50% acetonitrile in
water. Plasma samples were diluted in 50% acetonitrile to achieve a range of dilutions for analysis and
quantitation by LC-MS/MS.
General Synthetic Procedures. All chemicals and solvents were purchased from Sigma-Aldrich,
Fisher Scientific, Matrix Scientific,or Oakwood Chemical and were used without further purification.
Synthetic intermediates were purified using Biotage flash chromatography system on 230−400 mesh
1
13
silica gel. H and C NMR spectra were obtained using Bruker DPX-400 or Avance-400 spectrometer at
400 and 100 MHz, respectively. NMR chemical shifts are described in δ (ppm) using residual solvent
1
13
peaks as standard (CDCl3, 7.26 ppm (1H), 77.16 ppm (13C); CD3OD, 3.31 ppm ( H), 49.00 ppm ( C);
1
13
1
13
DMSO-d , 2.50 ppm ( H), 39.52 ppm ( C); acetoned6, 2.05 ppm ( H), 29.84 ppm 206.26 ppm ( C)).
6
Data are reported in the following format: chemical shift, multiplicity (s = singlet, d = doublet, dd = doublet
of doublet, t = triplet, q = quartet, br = broad, m = multiplet), coupling constants, and number of protons.
High-resolution mass spectral data were measured using a Shimadzu IT-TOF LC/MS for all final
compounds. All compounds submitted for biological testing were confirmed to be ≥96% pure by analytical
1
HPLC, supported by H analysis, unless otherwise stated. The purity of final compounds were determined
by HPLC using Agilent Eclipse XDB-C18 column (4.6x250 mm, 5 µm) with UV absorbance detection at
ACS Paragon Plus Environment

Page 29 of 72
Journal of Medicinal Chemistry
1
2
3
2
54 nm, eluting with a linear gradient from 10% aqueous MeCN to 90% MeCN over 18 mins, holding at
0% MeCN for a further 5 min. For the synthesis of Grignard reagents, the following procedure was used:
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
9
To a dried round-bottomed flask were added aryl bromide (1 eq.) in anhydrous tetrahydrofuran and
magnesium turnings (1.1 eq.) under an argon atmosphere. One granule of iodine was added to initiate
the reaction along with hot fan. The solution turned pale white and then brownish color along with strong
heat release. The Grignard reagent was ready for use without further purification when the magnesium
was consumed. For full experimental details of all compounds, see Supporting Information.
Representative synthetic methods, spectral data, and HRMS for novel compounds are described in detail
below. Full spectra and chromatograms are supplied in Supplemental Information.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
-Chloro-6-methoxybenzo[b]thiophene-2-carbonyl chloride (10). To
chlorobenzene was added (10g, 56.18 mmol) (E)-3-(4-methoxyphenyl)acrylic acid, (40 mL, 561.8 mmol)
SOCl , (0.45 mL, 5.618 mmol) pyridine and molecular sieves. The reaction was heated at reflux for 3
a
solution of
2
days. Excess thionyl chloride was removed under reduced pressure and the remaining material was
suspended in hot hexane then filtered. The filtrate collected and evaporated under reduced pressure to
give compound 10 as a pale solid (yield: 7g, 40%).
3-Chloro-N,6-dimethoxy-N-methylbenzo[b]thiophene-2-carboxamide (11). To an oven-dried
round-bottom flask was dissolved compound 10 (2g, 7.7 mmol) in (15 mL) of anhydrous dichloromethane
under argon atmosphere. N,O-Dimethylhydroxylamine hydrochloride (0.83g, 8.5 mmol) was added in one
portion followed by Et
temperature and monitored by TLC. Upon completion, the reaction was quenched by water and extracted
with ethyl acetate, washed by water, brine and dried over Na SO . The organic extracts were evaporated
3
N (5.4 mL, 38.8 mmol ) dropwise to the mixture. The reaction was stirred at room
2
4
under reduced pressure and purified by flash chromatography (10-30% ethyl acetate in hexanes) to give
1
a light yellow solid (yield: 1.7 g, 76%). H NMR (400 MHz, CDCl
3
) δ 7.80 (d, J = 8.9 Hz, 1H), 7.22 (d, J =
2
.2 Hz, 1H), 7.08 (dd, J = 8.9, 2.3 Hz, 1H), 3.88 (s, 3H), 3.72 (s, 3H), 3.38 (s, 3H). 13C NMR (101 MHz,
3
CDCl ) δ 161.45, 159.31, 139.78, 129.61, 124.47, 123.58, 122.66, 115.53, 103.71, 61.47, 55.29, 33.16.
ACS Paragon Plus Environment