Discovery and biological evaluation of darolutamide derivatives as inhibitors and down-regulators of wild-type AR and the mutants

Article abstract of DOI:10.1016/j.ejmech.2019.111608

Androgen receptor (AR) has been a target of prostate cancer (PC) for nearly six decades. Recently, downregulating or degrading AR and the mutants especially the splice variant 7 (AR-V7) lacking ligand binding domain (LBD) emerged as an advantageous therapeutic approach to overcome drug resistance. Here, the structural modification of darolutamide resulted in the discovery of dual-action AR inhibitors and down-regulators. Unlike other traditional AR antagonists targeting the AR-LBD, compounds 4k and 4b not only inhibit the activities of wt-AR and AR-F876L mutant but also downregulate the protein expression of full-length (AR-full) and AR variant 7 (AR-V7) at mRNA level. In cell proliferation assays, compounds 4k and 4b exhibited better antiproliferative activities than darolutamide and enzalutamide against AR-V7-positive 22Rv1 cells and VCaP cells. In addition, 4k demonstrated better antitumor activity than clinically used enzalutamide in castration-resistant VCaP xenograft model. Collectively, combining the activities of AR inhibition and downregulation, compound 4k is proposed as an advantageous lead compound to disrupt AR signaling and overcome resistance.

Full text of DOI:10.1016/j.ejmech.2019.111608

Journal Pre-proof
Discovery and biological evaluation of darolutamide derivatives as inhibitors and
down-regulators of wild-type AR and the mutants
Jiang Yu, Peiting Zhou, Mingxing Hu, Liuqing Yang, Guoyi Yan, Ruixue Xu, Yufang
Deng, Xinghai Li, Yuanwei Chen
PII:
S0223-5234(19)30742-1
DOI:
https://doi.org/10.1016/j.ejmech.2019.111608
EJMECH 111608
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 25 June 2019
Revised Date: 26 July 2019
Accepted Date: 7 August 2019
Please cite this article as: J. Yu, P. Zhou, M. Hu, L. Yang, G. Yan, R. Xu, Y. Deng, X. Li, Y. Chen,
Discovery and biological evaluation of darolutamide derivatives as inhibitors and down-regulators
of wild-type AR and the mutants, European Journal of Medicinal Chemistry (2019), doi: https://
doi.org/10.1016/j.ejmech.2019.111608.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©
2019 Published by Elsevier Masson SAS.

Discovery and biological evaluation of Darolutamide derivatives as
inhibitors and down-regulators of wild-type AR and the mutants
†
†
†
∥
#
†
Jiang Yu , Peiting Zhou , Mingxing Hu , Liuqing Yang , Guoyi Yan , Ruixue Xu ,
†
‡
†‡*
Yufang Deng , Xinghai Li and Yuanwei Chen
†
State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, Sichuan University
and Collaborative Innovation Center for Biotherapy, Chengdu, 610041, China
‡
Hinova Pharmaceuticals Inc., 4th Floor, Building RongYao A, No. 5, Keyuan South Road,
Chengdu, 610041, China
∥
Department of Pharmacy, Shanxi Medical University, Taiyuan, Xinjian Road 56, 030001, China
#
Department of Hepatobiliary Pancreatic Surgery, Henan Province People's Hospital, Zhengzhou
4
*
50003, China.
Corresponding Author, E-mail: ywchen@scu.edu.cn.
¢
Graphical Abstract
Enza Daro
4b
Enza Daro
4k
(
µM)
AR inhibitors and down-regulators
AR-full
Tublin
Growth inhibition of VCaP cells in vivo
1
1
500
000
Control
Enza 30mg/kg
4
4
k 30mg/kg
k 60mg/kg
NS
5
00
0
**
*
**
0
10
20
30
40
Days after treatment

¢
Abbreviations
AR, androgen receptor; LBD, ligand binding domain; AR-full, full-length AR; AR-V7,
AR variant 7; PC, prostate cancer; CRPC, castration resistant prostate cancer; SAR,
structure-activity relationship; PSA, prostate-specific antigen; FKBP5, FK506 Binding
Protein 5; Enza, Enzalutamide; Abi, Abiraterone; IC , 50 % inhibitory concentration;
5
0
CCK-8, cell counting kit-8; wt-AR, wild-type AR; Bort, bortezomib; DCM,
dichloromethane; EA, ethyl acetate; DMF, N,N-Dimethyl formamide; DMA,
N,N-dimethylacetamide; PEG, poly(ethylene glycol); NBS, N-Bromosuccinimide; NMP,
1
-Methyl-2-pyrrolidinone; THF, Tetrahydrofuran; DIAD, Diisopropyl azodicarboxylate;
EDCI, 1-Ethyl-3- (3-dimethylaminopropyl)carbodiimide; HOBt, 1-Hydroxybenzotriazole;
DIPEA, N,N-Diisopropylethylamine; BINAP, 2,2′-bis(diphenylphosphino)-1,1′-bina
phthyl; DMAP, 4-Dimethylaminopyridine; RT, room temperature.
¢
Abstract
Androgen receptor (AR) has been a target of prostate cancer (PC) for nearly six
decades. Recently, downregulating or degrading AR and the mutants especially the splice
variant 7 (AR-V7) lacking ligand binding domain (LBD) emerged as an advantageous
therapeutic approach to overcome drug resistance. Here, the structural modification of
darolutamide resulted in the discovery of dual-action AR inhibitors and down-regulators.
Unlike other traditional AR antagonists targeting the AR-LBD, compounds 4k and 4b not
only inhibit the activities of wt-AR and AR-F876L mutant but also downregulate the

protein expression of full-length (AR-full) and AR variant 7 (AR-V7) at mRNA level. In
cell proliferation assays, compounds 4k and 4b exhibited better antiproliferative activities
than darolutamide and enzalutamide against AR-V7-positive 22Rv1 cells and VCaP cells.
In addition, 4k demonstrated better antitumor activity than clinically used enzalutamide
in castration-resistant VCaP xenograft model. Collectively, combining the activities of
AR inhibition and downregulation, compound 4k is proposed as an advantageous lead
compound to disrupt AR signaling and overcome CRPC resistance.
Keywords: Androgen receptor inhibitors; Androgen receptor down-regulators; F876L
mutant; Darolutamide derivatives; Castration-resistant prostate cancer.
1
. Introduction
AR is a ligand-dependent transcription factor. Numerous studies demonstrated that
prostate cancer was driven by aberrant activation of AR. After being activated by ligand
binding, AR is translocated into nucleus, dimerizes and interacts with androgen-response
1
-3
elements, which leads to biological responses including cell growth and survival. AR
antagonists are classified into steroidal and nonsteroidal compounds. Despite the
effectiveness of steroidal antiandrogens on patients with advanced PC, some undesirable
side effects including impotence and loss of libido limited their clinical application and
4
-6
led to the development of nonsteroidal AR inhibitors. As shown in Figure 1, flutamide,
bicalutamide and nilutamide are first-generation nonsteroidal AR antagonists with better
7
selectivity for AR. However, after a period of treatment, most PC patients eventually

progressed into the castration-resistant stage that is associated with AR amplifications,
8
-14
AR overexpression, point mutations in LBD and active AR splice variants.
For
example, T877A and H874Y mutants mediated resistance to nilutamide and
1
5,16
hydroxyflutamide treatment.
And bicalutamide worked as an agonist for both W741C
1
7
and W741L mutant.
The improved understanding of mechanisms to resistance led to the development of
second-generation AR antagonists. Since 2011, enzalutamide and apalutamide (formerly
known as ARN509) with higher binding affinity than first-generation AR antagonists
18-20
have been approved by FDA for the treatment of CRPC.
Although some success has
been achieved, the secondary resistance to both enzalutamide and apalutamide inevitably
occurred, which is considered to be partially mediated by AR-F876L and AR-V7
2
1,22
mutants.
Other second-generation AR antagonists such as darolutamide (Figure 1)
2
3,24
and HC-1119 have entered clinical studies.
HC-1119 is a deuterated AR antagonist
2
5
with higher drug exposure and potentially better safety profile than enzalutamide.
NO2
F
NC
NO2
F3C
O
O
O
O
S
F3C
N
F3C
N
NH
N
H
H
OH
O
O
Flutamide
Bicalutamide
Nilutamide
F
F
OH
O
O
O
N
CN
S
CN
F3C
S
N
Cl
(S) N
N
N
N
N
N
H
N
NH
F3C
O
CN
O
Enzalutamide
Apalutamide
Darolutamide (1a)
Figure 1. Molecular structures of AR antagonists.

Darolutamide (formerly known as ODM-201) is a structurally unique AR antagonist
targeting LBD. It inhibits AR variants including W741L, T877A, H874Y and F876L
mutants, competitively binds to AR-LBD with high binding affinity and significantly
inhibits the growth of enzalutamide-resistant prostate cancer cells in vivo. Unlike
enzalutamide and apalutamide, darolutamide showed negligible brain penetrance which
decreased the likelihood of seizures possibly due to its distinctly different chemical
26,27
structure.
In phase III trial, darolutamide was found to significantly extend
28
metastasis-free survival compared to placebo in patients with non-metastatic CRPC.
However, resistance to these traditional antiandrogens targeting AR-LBD due to various
mutations still remains a major challenge in CRPC treatment. Therefore, there still exists
unmet needs to develop new compounds targeting AR and the mutants. Recently,
downregulating or degrading ARs especially the AR-V7 emerged as an advantageous
2
9-43
therapeutic approach to overcome drug resistance.
For example, galeterone, a
multi-target oral small molecule inhibitor, blocked the AR signaling through ARs
37
downregulation, AR-LBD antagonism and inhibition of androgen synthesis. Here, we
described the discovery and biological evaluation of dual-action AR inhibitors and
down-regulators deriving from darolutamide.
2
. Chemistry
As outlined in Scheme 1, starting from the commercially available 3-chloro-5-
fluoroaniline (5) or 4-bromo-2-chlorobenzonitrile (9), intermediates 14-15 and products

2
6,38
1
a and 1c were successfully prepared by following literature procedures.
For the synthesis of intermediate (R)-16, (R)-tert-butyl-1-hydroxypropan-2-yl
2
carbamate was used. To investigate the effect of substitutes at R position of compounds
bearing amide group, EDCI, HOBt, DIPEA and corresponding carboxylic acids were
used to afford compounds 1c-1g, 1k-1o and intermediates 13-14. Subsequently, the
reduction of ketone substituents in compounds 1c, 1e, 1k, 17 by NaBH and hydrolysis of
4
1
8 by hydrochloric acid yielded compounds 1a, 1b, 1i, 1h and 1j, respectively. Direct
sulfonylation of intermediate 15 with different sulfonyl chlorides gave the desired
sulfonamides (2d-2f) in good to moderate yields. The reaction of intermediate 14 with
triphosgene
formed
an
isocyanate
intermediate,
which
reacted
with
5
-(tert-butyl)-1H-pyrazol-3-amine to yield the carbamide derivative 2g. For the synthesis
4
of compounds 3a-3p and 4a-4m with various substitutes at R position, several
approaches including nucleophilic substitution reaction, Cu-catalyzed and Pd-catalyzed
C-N
-chloro-4-(1H-pyrazol-3-yl) benzonitrile (13), a three-step procedure was carried out to
obtain compounds 2a-2c.
coupling
reactions
were
used
(Scheme
1).
Starting
from
2

O
N
O
N
Cl
NH2
Cl
Br
NH2
Cl
NH2
Cl
X
B
Cl
N
O
Cl
NH
O
N
N
a
b
NC
c
NC
d
e
NC
NC
R1
R1
R1
R1
F
R1
8
9
, R1 = F; X = I
, R1 = H; X = Br
10, R1 = F
11, R1 = H
12, R1 = F
13, R1 = H
5
6, R1 = F
7, R1 = F
OBu-t
O
HO
O
O
N
Cl
N
Cl
N
H
R2
R2
O
N
HO
N
N
(
R or S)
NH2
H
NC
g
NC
f
R1
R1
1c, (S)-Me; R1 = H
d, (S)-Me; R1 = F
e, (R)-Me; R1 = H
17, (S)-Me; R1 = H; R2 =
O
1
1
1
1
1
1
1
4, R1 = F; (S)-Me
5, R1 = H; (S)-Me
6, R1 = H; (R)-Me
O
O
f-1g, (S)-Me; R1 = H 18, (S)-Me; R1 = H; R2
k-1o, (S)-Me; R1 = H
=
p
o
j/ k/ l/ m/ n
h or i
O
C
O
O
O
R4
Cl
NC
N
(S)
R2
Cl
N
Cl
N
Cl
N
S
R2'
N
N
N
(S)
N
N
N
N
N
(S)
N
R2
H
H
H
H
H
NC
NC
NC
R1
F
1
1
1
a, (S)-Me; R1 = H
b, (R)-Me; R1 = H
h-1j, (S)-Me; R1 = H
3
a-3p, 4a-4i, 4k, 4m, (S)-Me
2g
2d-2f
4
j, 4l, (R)-Me
Scheme 1. Synthesis of compounds 1a-1o, 2d-2g, 3a-3p and 4a-4m. Reagents and
conditions: (a) NBS, CH
NaNO , KI, <10 ℃ , overnight; (d) Suzuki coupling, boronic ester, Na
THF, 40 ℃ , overnight; (e) HCl, EtOH, reflux, overnight; NaOH, H O, rt; (f) PPh
DIAD, 0 ℃ - rt; HCl, H O, rt, overnight; (g) EDCI, HOBt, DIPEA, DMF, rt, overnight.
3
CN, <10 ℃ , 4h; (b) CuCN, NMP, 160 ℃ , 6h; (c) H
2
SO
4
,
2
2
CO , PdCl (PPh
3
2
3
)
2
,
2
3
, THF,
2
(
h) NaBH , MeOH-THF, rt, 2 h; (i) 1M HCl, THF, rt, overnight; (j) Et N, EtOH, 90 ℃ , 24
4
3
h; (k) t-BuONa, X-phos, Pd (dba) , PhMe, 110 ℃ , overnight; (l) Metformin-HCl, CuI,
2
3
Cs CO , EtOH, reflux, 24 - 48 h; (m) t-BuONa, BINAP, Pd(OAc) or Pd (dba) , PhMe,
2
3
2
2
3
1
10 ℃ , overnight; (n) t-BuONa, Dave-Phos, Pd (dba) , dioxane, MW 150 ℃ , 30 min; (o)
2 3
DIPEA, DCM, rt, overnight; (p) Et N, triphosgene, DCM, 0 ℃ -rt, 6 h; Pyridine, rt, 6 h.
3

As shown in Scheme 2, the electrophilic addition of methyl methacrylate with
intermediate 13 in refluxing acetonitrile gave the racemic mixture 19. Then, subsequent
hydrolysis reaction of 19 followed by a coupling reaction with arylamine yielded the
desired compounds 2a-2c.
Scheme 2. Synthesis of compounds 2a-2c. Reagents and conditions: (a) K
2
3 3
CO , CH CN,
reflux, overnight; (b) LiOH-H
DCM, 30 ℃ , overnight.
2
O, THF, MeOH, 70 ℃ , overnight; (c) EDCI, DMAP,
3
. Results and Discussion
3
.1 Discovery of compounds 4b and 4k.
Firstly, we evaluated the activities of darolutamide (1a), its active metabolite
ORM-15341(1c) and compounds 1b, 1e bearing (R)-Me group in AR-positive human PC
2
0,39
cells (LNCap/AR, VCaP, 22Rv1)
via CCK-8 cell proliferation assay (SAB, CP002).
3
9
The AR-negative human PC cells (PC-3) and non-cancer originated human hepatocyte
cells (L-02) were also used to rule out the non-AR-mediated toxicity. As shown in Table
1
, analogues 1a-1c and 1e exhibited equal antiproliferative activity against LNCaP/AR
cells with IC values from 1.08 µM to 1.65 µM suggesting that the inhibitory effect of
5
0
darolutamide did not rely on the (S)-configuration of isopropylamine chain. Unexpectedly,

2
all compounds bearing 1H-pyrazole group at R position (1a-1f) inhibited the
proliferation of VCaP cells in a concentration-independent manner, whose inhibition
curves can’t be normally converged by GraphPad Prism 5 (Figure 2A). Interestingly,
when we replaced the 1H-pyrazole moiety with phenyl (1g-1l) or alkyl groups (1m-1o)
the concentration-independent activity against VCaP cells almost disappeared (Table 1),
2
suggesting the substituent at R position was of great influence. Like other traditional
antagonists targeting AR-LBD, all the analogues in Table 1 exhibited poor effects on the
3
9
activity of AR-V7-positive 22Rv1 cells and low toxicity toward AR-negative PC-3 and
non-cancer originated L-02 cells. Among them, derivatives 1h-1j were more potent than
darolutamide and enzalutamide against VCaP cells with IC50 values in submicromolar
range.
a
Table 1. In Vitro Antiproliferative Activities of 1a-1o.
2
IC (µM)
Compd
R
50
LNCaP/AR
0.19
VCaP
22Rv1
>30
PC-3
>30
L-02
Enzalutamide
Darolutamide
_
50%@30
NC
17.10
1
.65
>30
>30
>30
>30
>30
NT
(
1a)
c
1
b
1.65
NC

1
c
1.13
1.10
1.08
1.00
1.83
NC
NC
NC
NC
NC
>30
>30
>30
>30
NT
27.69
NT
b
1
d
c
1
e
15.54
>30
NT
1
f
18.73
>30
>30
NT
1g
NT
1
h
2.28
3.69
0.74
0.90
0.91
0.59
20.08
9.05
>30
>30
>30
>30
NT
>30
>30
1
i
1
j
1
k
3.60
3.41
1.81
2.46
3.04
1.40
1.54
2.19
5.23
22.47
NT
NT
NT
NT
NT
>30
>30
>30
>30
>30
15.33
NT
1
l
1
m
NT
1n
NT
1
o
NT
a
IC , 50% inhibitory concentration. IC values are the mean value of at least two
5
0
50
experiments with duplicate measurements. The deviations were less than 35%. 50%@30,
0％ inhibition at 30 µM. NT, not tested. NC, not converged by GraphPad Prism 5.
5
1
b
1
c
Unless otherwise noted, R = H. R = F. (R)-Me.
(A)

VCaP
VCaP
1
50
1
1
50
00
4
4
k
b
1
1
1
1
1
f
e
d
c
b
100
Enza
5
0
0
5
0
0
1a
0
1
2
3
4
5
0
1
2
3
4
5
Log (nM)
Log (nM)
(B)
Antiproliferative activities of 4b
Antiproliferative activities of 4k
LNCaP/AR
22Rv1
PC-3
1
00
100
50
0
LNCaP/AR
22Rv1
PC-3
L-02
L-02
5
0
0
0
1
2
3
4
5
0
1
2
3
4
5
Log (nM)
Log (nM)
Figure 2. (A) Growth inhibition curves in VCaP cells. (B) Growth inhibition curves of
compounds 4b and 4k toward different cells. Enza, enzalutamide.
3
As exhibited in Table 2, further modification was focused on the amide group at R
position. The antiproliferative activities of compounds 1f and 1h were decreased or
completely abolished by reversing the position of amine and carbonyl groups (2a, 2b),
confirming the crucial role of the amide group. Compared with derivative 2c, compound
2
b with hydroxyl side chain showed decreased potency against LNCaP/AR and VCaP
3
cells. Next, the replacement of the carbonyl group (1k) with sulfonyl group (2d) at R
position caused a slight decrease in activities against LNCaP/AR and VCaP. This result
suggests that the carbonyl group of amide moiety is tolerated.
a
Table 2. In Vitro Antiproliferative Activities of 2a-2g, 1g, 1f and 1j.

3
2
IC (
µM)
Compd
R
R
50
LNCaP/AR
0.19
VCaP
PC-3
>30
L-02
Enzalutamide
Darolutamide
_
_
_
_
50%@30
NC
17.10
1.65
1.00
2.28
3.60
>30
4.15
1.54
5.23
6.14
>30
1.63
>30
18.73
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
NT
(
1a)
1f
NC
1
1
h
k
0.90
1.40
>30
15.33
NT
b
2
a
H
b
N
2
b
>30
NT
O
O
H
N
b
2
c
2.98
3.95
2.45
13.59
5.12
NT
2
d
NT
2
e
NT
2f
NT
c
2
g
15.00
a
IC , 50% inhibitory concentration. IC values are the mean value of at least two
5
0
50
experiments with duplicate measurements. The deviations were less than 35%. 50%@30,
0% inhibition at 30 µM. NT, not tested. NC, not converged by GraphPad Prism 5.
5
1
b
c 1
Unless otherwise noted, R = H. Racemate. R = F.

Subsequently, we removed the carbonyl group in amide moiety and introduced
various aromatic rings including phenyl, pyridyl, pyrimidinyl, naphthyl, quinolyl,
4
isoquinolyl and indolyl groups at R position (Table 3). Fortunately, such modification
decreased the IC50 values of compounds to low- or sub-micromolar range for 22Rv1 (3g,
3
i, 3k), VCaP (3f, 3g, 3o) and LNCaP/AR (3a, 3b, 3k, 3p) cells without
non-AR-mediated toxicity. These data suggested that the 4-(1-(2-aminopropyl)-1H-
pyrazol-3-yl)-2-chlorobenzonitrile scaffold was worthy of further investigation. Next, we
4
introduced more different aromatic bicyclic substituents at R position.
a
Table 3. In Vitro Antiproliferative Activities of 3a-3p.
4
IC (µM)
Compd
R
50
LNCaP/AR VCaP
22Rv1
>30
PC-3
>30
L-02
Enzalutamide
Darolutamide
_
_
0.19
1.65
0.71
0.43
1.11
2.56
2.07
50%@30
NC
17.10
>30
>30
>30
3.21
>30
NT
>30
29.39
>30
NT
(
1a)
3
a
>30
3
b
>30
21.62
>30
3
c
NC
3
d
9.90
>30
11.50
>30
NT
NT
3
e
>30
>30

3
f
1.18
1.33
1.40
7.70
13.04
0.51
2.66
1.91
1.87
2.30
0.49
0.48
0.85
>30
4.61
>30
2.26
3.22
1.58
1.94
0.87
2.13
12.54
4.76
>30
>30
>30
>30
NT
>30
>30
NT
3
g
3
h
3
i
4.61
NT
26.81
NT
3
j
NT
3
k
2.98
5.82
11.98
6.04
7.95
11.98
NT
NT
3
l
11.20
NT
NT
3m
NT
3
n
NT
NT
3
o
>30
NT
NT
N
3
p
NT
a
IC , 50% inhibitory concentration. IC values are the mean value of at least two
5
0
50
experiments with duplicate measurements. The deviations were less than 35%. 50%@30,
5
0% inhibition at 30 µM. NT, not tested. NC, not converged by GraphPad Prism 5.
As displayed in Table 4, more compounds with aromatic bicyclic substituents at R
4
position were synthesized and evaluated. Among them, compounds 4b and 4k featuring
quinoline (4b) and purine (4k) groups exhibited better antiproliferative activities than
3
9
darolutamide and enzalutamide against AR-overexpressing VCaP cells
and
AR-V7-positive 22Rv1 cells without concentration-independent antiproliferative
activities (Figure 2B). Interestingly, the compound 4k bearing (S)-Me and purine groups

was more potent than the enantiomer (4l) against VCaP and 22Rv1 cells. The addition of
Cl on purine ring (4m) also decreased the inhibitory activities toward LNCaP/AR and
VCaP cells. These results suggest that the position of nitrogen atom on aromatic ring was
of great importance for the inhibitory activities in vitro and the quinoline and purine
4
groups were the preferred substituents at R position.
a
Table 4. In Vitro Antiproliferative Activities of 4a-4n.
4
IC (µM)
Compd
R
50
LNCaP/AR VCaP
22Rv1
>30
PC-3
>30
L-02
Enzalutamide
Darolutamide
_
_
0.19
1.65
0.87
1.08
2.10
4.81
2.28
0.87
3.09
0.79
0.51
50%@30
NC
17.10
>30
>30
>30
NT
>30
NT
(
1a)
4
a
28.10
0.73
4b
3.92
3.57
4.29
NT
7.96
>30
NT
16.60
NT
4
c
1.47
4d
2.41
NT
4
e
12.40
>30
NT
NT
4
f
21.20
4.72
>30
NT
NT
4
g
4.02
5.13
NT
NT
4h
>30
NT
4
i
1.31
4.36
>30
12.97

b
4
4
4
j
0.97
0.96
0.85
>30
0.64
2.00
9.71
4.24
NT
14.08
1.56
NT
27.90
NT
k
b
l
10.61
4
m
2.66
2.70
4.38
>30
14.53
a
IC , 50% inhibitory concentration. IC values are the mean value of at least two
5
0
50
experiments with duplicate measurements. The deviations were less than 35%. 50%@30,
b
5
0% inhibition at 30 µM. NT, not tested. (R)-Me.
Subsequently, the effects of some analogues on AR transactivation and protein
expression were evaluated through luciferase reporter gene assay (CCS-1019L; QIAGEN)
and western blot, respectively. To measure the inhibition of AR, HEK 293 cells were
transiently co-transfected with expression vector encoding wt-AR and an
androgen-responsive luciferase reporter gene construct. As shown in Figure 3A, these
compounds strongly suppressed the transcriptional activities of wt-AR, among which
compounds 4k, 4b, 4i, 3b, 1h and 2b were the most potent. In addition, these data were
consistent with the antiproliferative activities of compounds against AR-positive prostate
cancer cells. However, only compounds 4k, 4b and 4i obviously decreased the protein
expression of AR in LNCaP cells (Figure 3B). This result suggests that these
darolutamide derivatives maintain inhibitory activity on AR and some of them can
down-regulate the protein expression of AR. Therefore, compounds 4b and 4k were
selected for further in vitro and in vivo characterization and mechanism studies.

(A)
wt-AR Transcriptional Activity
1
00
5
0
0
(B)
LNCaP
AR-full
Tublin
(C)
Grayscale analysis
Grayscale analysis
1
1
0
0
.5
.0
.5
.0
1.5
1.0
0.5
0.0
Figure 3. (A) The inhibition of wt-AR by compounds studied in luciferase reporter gene
assay. The final concentration of testosterone and test compound were 2 nM and 10 µM,

respectively. Data were expressed as the mean ± SD (n = 3). Enza, enzalutamide. Daro,
darolutamide. (B) Western blot analysis of AR expression in LNCaP cells treated with 20
µM compounds or DMSO for 24 h. (C) Grayscale analysis of western blot by Image J.
3
.2 compounds 4b and 4k downregulate ARs expression at mRNA level
First, the effects of compounds 4b and 4k on the protein expression of ARs in
LNCaP, VCaP and 22Rv1 cells were further confirmed. As exhibited in Figure 4A, unlike
darolutamide or enzalutamide, 4b and 4k decreased the protein expression of not only
AR-full but also the truncated variant AR-V7 with higher potency than galeterone under
the same condition. To elucidate the possible mechanism of, we next measured the
mRNA expression of ARs in prostate cancer cells. As exhibited in Figure 4B, the mRNA
of AR-full and AR-V7 in LNCaP, VCaP and 22Rv1 cells were decreased by the treatment
of 4b or 4k, which was in accord with the downregulation of ARs protein. This result
suggests that our compounds suppress the protein level of AR-full and AR-V7 through
decreasing mRNA expression.
(A)
LNCaP
Enza Daro
4b
Enza Daro
4k
Gal (µM)
(µM)
(µM)
AR-full
AR-full
AR-full
Tublin
Tublin
Tublin

VCaP
Enza Daro
4b
4k
Daro
Enza
(µM)
(µM)
Gal (µM)
AR-full
AR-Vs
AR-full
AR-Vs
AR-full
AR-Vs
Tublin
Tublin
Tublin
2
2Rv1
Enza Daro
4b
4k
EnzaDaro
Gal
Enza
(µM)
(µM)
(µM)
AR-full
AR-V7
AR-full
AR-V7
AR-full
AR-V7
Tublin
Tublin
Tublin
(B)
VCaP
1
1
0
0
.5
.0
.5
.0
NS
AR-full
AR-V7
NS
NS
NS
*
*
*
**
** ***
**
***
*
** ***
*
**
(µM)
(µM)
Enza Daro
4b
4k
4b
4k
4b
4k
(C)
LNCaP
DMSO
4k
Gal
Enza
4k
DMSO
Bort:
-
+
-
+
-
+
-
+
Bort/MG132:
B
M
-
B
M
-
AR-full
AR-V7
AR-full
AR-V7
Tublin
Tublin
Figure 4. (A) Compounds 4b and 4k reduced the protein expression of AR-full and
splice variant AR-Vs (AR-Vs) in LNCaP, VCaP and 22Rv1 cells. Cells cultured in
medium supplemented with 10% (v/v) FBS were treated with compounds or DMSO for
2
4h. Daro, darolutamide. Enza, enzalutamide. (B) qRT-PCR analysis for AR-full and

AR-V7 mRNA expression in LNCaP, VCaP and 22Rv1 cells treated with DMSO or test
compounds for 24 h. The relative mRNA expression was normalized to GAPDH. Data
are expressed as the mean ± SD (n = 3). *indicates significance as p< 0.05, ∗∗p < 0.01,
∗
∗∗p < 0.001 compared to samples treated with DMSO by two-tailed t test. NS, no
significant difference. (C) The effect of proteasome inhibitors on the decrease in ARs
protein caused by 4k. LNCaP cells cultured in medium supplemented with 10% (v/v)
FBS were treated with compounds (30
of 20 M proteasome inhibitors for 26h (left) or 10h (right). Gal, galeterone. Bort,
bortezomib. B, Bort. M, MG132.
µM) or DMSO for 24 h in the presence or absence
µ
Next, we examined the effect of 4k on ARs protein degradation utilizing proteasome
inhibitor bortezomib and MG132. Noteworthily, increase in AR-V7 protein level was
observed in both vehicle and test groups after bortezomib treatment for 26 hours (Figure
4
4
C
), which was considered related to the effect of bortezomib itself but not the compound
k. Meanwhile, bortezomib didn’t relieve the decrease of AR-full protein caused by 4k
and similar results were observed by the treatment of MG132 and bortezomib for 10
hours in LNCaP cells, which suggests compound 4k didn’t work through the degradation
of ARs protein. On the other hand, considering the inhibition of bromodomain-containing
protein 4(BRD4) may led to decrease in ARs expression, we also examined the effects of
compounds 4k and 4b on BRD4. The results demonstrated that all of the test compounds
including 4k, 4b, darolutamide and enzalutamide didn’t inhibit the activity of BRD4 even

at a concentration of 30 µM (Figure S1).
These data suggest that our compounds downregulate the protein of AR-full and
AR-V7 through decreasing mRNA expression but not degrading ARs protein by
proteasome or inhibiting BRD4.
3
.3 Compounds 4b and 4k inhibit wild-type AR (wt-AR) and the F876L mutant
To determine whether the selected compounds affect AR-regulated genes
transcription, qRT-PCR analysis was performed in PC cells. As shown in Figure 5A, the
mRNA expression of PSA and FKBP5 in both LNCaP and VCaP cells were suppressed
by compounds 4b and 4k in the presence of 0.2 nM R1881. And the 50% inhibition of
transcription was observed at ~1µM, which was equal potent to that of darolutamide.
Next, the activities of compounds 4b and 4k on the AR-F876L mutant that exhibited
2
1,22
resistance to Enzalutamide
To avoid interference from endogenous AR expression, the AR-negative PC-3 cells were
used. As represented in Figure 5B 4b and 4k significantly decreased the
were measured via the dual luciferase reporter gene assay.
,
androgen-induced luciferase expression in cells transfected with wt-AR or AR-F876L
expression vector. As expected, high-concentration enzalutamide exhibited weak efficacy
compared to those of low concentration (0.1 and 1
µM) for AR-F876L mutant, which was
2
6,27
in accord with Moilanen’s and Borgmann’s reports.
To further demonstrate the inhibitory activities of compounds 4b and 4k on ARs, the

4
0
nuclear translocation assay was performed. WPMY-1 cells were transfected with
enhanced green fluorescence protein (EGFP) tagged wt-AR or AR-F876L mutant, which
were treated with vehicle or compounds in the presence or absence of 0.5 nM R1881 for
2
4 h. As represented in Figure 5C-D, 4b and 4k prevented not only wt-AR but also
AR-F876L mutant from translocating into nucleus. However, AR-F876L mostly located
in nucleus after the treatment of enzalutamide under the same conditions. These results
taken together demonstrate that compounds 4b and 4k can effectively suppress the
activities of wt-AR and the AR-F876L mutant.
(A)
LNCaP
VCaP
1
1
0
0
.5
.0
.5
.0
1.5
1.0
0.5
0.0
Daro
Daro
4k
4
4
k
b
4b
NS NS
NS
*
*
**
**
**
**
**
**
**
*
** ***
**
**
**
***
Concentration (µM)
Concentration (µM)
VCaP
LNCaP
1
.5
1.0
0.5
0.0
1
1
0
0
.5
Daro
Daro
NS
NS
4
k
4
4
k
b
4b
.0
.5
.0
NS
NS
NS
*
**
*
*
**
**
**
*
***
***
** **
**
Concentration (µM)
Concentration (µM)
(B)

AR-F876L Transcriptional Activity
wt-AR Transcriptional Activity
Daro
Daro
4k
4
k
1
00
100
50
0
NS NS
NS
4
b
4b
Enza
*
*
**
Enza
*
*
*
*
*
**
5
0
0
***
*
*
***
**
***
**
**
**
***
*
** *** ***
**
**
*
** *** ***
**
*
**
** ***
**
*
** ***
*** ***
***
*
**
Concentration (µM)
Concentration (µM)
(C)
wt-AR-EGFP Nucleus
Merge
(D)
AR-F876L-EGFP Nucleus
Merge
DMSO
DMSO
Enza+
R1881
DMSO +
R1881
4
k +
4k +
R1881
R1881
4
b +
4b +
R1881
R1881
Figure 5. (A) qRT-PCR analysis of 0.2 nM R1881-induced PSA and FKBP5 mRNA
expression in different cells treated with vehicle (DMSO) or compounds for 24 h. The
relative mRNA expression was normalized to GAPDH. Data are expressed as the mean ±
SD (n = 3). *indicates significance as p<0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, compared to
samples treated with DMSO by two-tailed
t test. NS, no significant difference. Daro,
Darolutamide. (B) inhibition of androgen-induced transactivation of wt-AR and
AR-F876L mutant. Data were expressed as the mean ± SD (n = 3). Enza, enzalutamide.
(C, D) Nuclear translocation of (C) wt-AR and (D) AR-F876L mutant in WPMY-1 cells

treated with vehicle or compounds (20
µM) in the presence or absence of 0.5 nM R1881
for 24 h. Hoechst 33342 (blue) was used to label the nuclei.
3
.4 Physicochemical Characterization of compounds 4k and 4b.
The solubility of compounds 4k and 4b was measured in phosphate buffered saline
4
1
(
PBS) at 25
quantification were exhibited in Figure S3. As reported in Table S1, although compound
showed around 10 times lower solubility than darolutamide, it’s sufficient for the in
℃ through saturation shake-flask method. The calibration curves used for
4
k
vitro tests. Compound 4b can’t be detected even the injection volume was increased to 50
ul. However, 4b was efficient and concentration-dependent in vitro possibly due to the
increased solubility in nutrient medium with DMSO at 37℃ . To determine whether the
compounds are stable under the experimental conditions, stability assay was performed in
nutrient medium with fetal bovine serum and antibiotic. As shown in Figure S4, they
were all stable in 6 days at 37℃ without obvious decrease in concentration.
3
.5 Molecular modeling
To gain atomic insights into the possible binding modes of compounds 4b and 4k in
4
2,43
the setting of AR-LBD (PDB ID: 3V49),
structure-based computer modeling was
4
4
performed using Autodock 4.0. Considering darolutamide is a mixture of two
pharmacologically active diastereomers, we analyzed the two diastereomers respectively.
As shown in Figure 6B-C, both of 4k and 4b established two key hydrogen bonds with
Gln711 and Thr877 residues through the nitrogen atom of cyano group and the hydrogen

atom of isopropylamine chain, which was more similar to that of darolutamide with
S)-OH (Figure 6A In addition, the nitrogen atom on purine ring of compound 4k
(
).
formed additional hydrogen bond interaction with the side chain of His874. Differently,
compound 4b established a sigma-pi interaction with Thr877 residue through the benzene
ring of quinolyl group. Noteworthily, the loss of carbonyl group in the amide moiety of
4
b and 4k didn’t affect the hydrogen bond interaction of hydrogen atom in
isopropylamine with Thr877 residue, which was in accordance with the tolerance of
carbonyl group observed in SAR study.
(A)
(S)-OH
(R)-OH
(B)
(C)

Figure 6. The predicted binding poses and 2D diagram of (A) darolutamide (B)
compound 4k (C) compound 4b in the androgen binding site of AR (PDB ID: 3V49). In
the 3D diagram of darolutamide, the diastereomer with (S)-OH was drawn in gray and
diastereomer with (R)-OH was in yellow.
3
.6 Compound 4k inhibits CRPC tumor growth in vivo.
To assess the potential of our compounds in vivo, pharmacokinetic (PK) analysis was
performed in Sprague-Dawley (SD) rats. The plasma levels of 4b and 4k were measured
after an oral dose of 30 mg/kg (Figure 7A). Unlike 4b, compound 4k can’t be detected in
plasma at 24 h after administration and the quick clearance of compound 4k in plasma
resulted in a short half-life (T1/2 = 1.79 h). However, compound 4k displayed higher
maximum plasma concentration (2648.88 ng/mL) and more desirable systemic exposure
with an AUC value of 7280.76 ng*h/ml than 4b (Figure 7B), which may be associated
0
-t
with its better solubility. Therefore, compound 4k was selected for further in vivo
pharmacological studies.

(A)
(B)
Compds
4k
1.79
4b
T1/2 (h)
Tmax (h)
3.14
0.65
0.33
Cmax (ng/ml)
2648.88
521.00
783.55
796.54
AUC0-t (ng*h/ml) 7280.76
AUC0-∞(ng*h/ml) 7379.80
(C)
(D)
*
Growth inhibition of VCaP cells in vivo
*
2
2
1
.5
.0
.5
*
*
NS
1
1
500
000
Control
Enza 30mg/kg
4
4
k 30mg/kg
k 60mg/kg
1.0
NS
0
0
.5
.0
5
00
0
**
*
**
0
10
20
30
40
Days after treatment
Figure 7. (A) The concentration-time curves in rats. Compounds were formulated in a
solvent of 10% DMA and 90% PEG200 for an oral dose of 30 mg/kg. (B)
Pharmacokinetic parameters of compounds 4b and 4k
enzalutamide and compound 4k on CRPC xenograft model. NS, no significant difference.
D) Individual tumor weight of each group. The tumor volume was expressed as the mean
. (C) Growth inhibitory effects of
(
±
SEM (N = 8-9). Enza, enzalutamide. NS, no significant difference.
As shown in Figure 7C, the castration-resistant VCaP xenografts generated in male
BALB/c nude mice were treated with 4k (30 mg/kg and 60 mg/kg), enzalutamide (30

mg/kg) and vehicle once a day for 40 days. Satisfactorily, the growth of VCaP tumors
was significantly inhibited by the treatment of compound 4k with 77.8% and 64.7%
tumor growth inhibition (TGI) for 60mg/kg and 30mg/kg, respectively. However,
enzalutamide was less potent than 4k at the dose of 30 mg/kg with a TGI value of 43.0%.
During the period of treatment, no significant weight loss was observed for any groups
(
Figure S2) and the tumor weight of each group was recorded at the end of treatment
Figure 7D).
(
4
. Conclusion
In summary, we synthesized and evaluated a series of darolutamide derivatives,
which led to the identification of dual-action AR inhibitors and down-regulators, 4b and
. Preliminary SAR study demonstrates that the inhibitory activity of darolutamide does
4
k
not rely on the (S)-configuration of isopropylamine chain. Meanwhile, the
concentration-independent antiproliferative activities observed by darolutamide and its
2
4
analogues in VCaP cells are mainly caused by the substituents at R and R positions.
Considering the tolerance of carbonyl group in amide moiety, numerous compounds
without carbonyl group were synthesized resulting in the discovery of 4b and 4k
.
Compounds 4b and 4k not only inhibit wt-AR and AR-F876L mutant that confers
resistance to enzalutamide but also downregulate the protein expression of AR-full and
AR-V7 at mRNA level, which might contribute to their better antiproliferative activities
toward AR-V7-positive 22Rv1 and AR-overexpressing VCaP cells than darolutamide and

enzalutamide. Furthermore, compound 4k effectively inhibits the growth of
castration-resistant VCaP xenografts in vivo without severe toxicity. Combining the dual
functions of AR inhibition and downregulation, compound 4k holds promise to overcome
secondary resistant caused by various AR mutants. These evidences suggest compound
4k as a potent and orally available lead compound to overcome resistance in CRPC.
5
. Chemical experiments
5
.1 General Chemistry. Reagents and solvents were purchased and used without further
purification. Biotage Isolera One apparatus and Agela flash column silica-CS Flash were
used for column chromatography. NMR spectra were recorded on a Bruker AMX 400
1
13
spectrometer ( H = 400 MHz, C = 101 MHz). Chemical shifts were expressed in parts
per million (ppm) and referenced to trimethylsilane (TMS) or residual deuterated solvent
(
CDCl or DMSO-d ). The purity of compounds was determined by high-performance
3 6
liquid chromatography (HPLC), performed on Dionex ultimate 3000 HPLC instrument,
Waters e2695 Series chromatographs and Waters xBridge column (5 m, 4.6mm × 150
µ
mm). HRMS spectra were recorded on a Q-TOF Premier mass spectrometer (Micromass,
Manchester, UK). Melting points were measured by WRR-Y melting point apparatus.
5
.2 General procedure A for the synthesis of intermediates 14-16.
Step 1. 4-Bromo-3-chloro-5-fluoroaniline (6). 3-Chloro-5-fluoroaniline (5.0 mmol,
7
3
30.0 mg) was dissolved in CH CN (12.0 ml) and the solution was cooled to 0 °C. NBS
(5.0 mmol, 890.0 mg) was added to the reaction mixture in small portions keeping the