Development of Novel AKR1C3 Inhibitors as New Potential Treatment for Castration-Resistant Prostate Cancer

Article abstract of DOI:10.1021/acs.jmedchem.0c00939

Aldo-keto reductase (AKR) 1C3 catalyzes the synthesis of active androgens that promote the progression of prostate cancer. AKR1C3 also contributes to androgen-independent cell proliferation and survival through the metabolism of prostaglandins and reactive aldehydes. Because of its elevation in castration-resistant prostate cancer (CRPC) tissues, AKR1C3 is a promising therapeutic target for CRPC. In this study, we found a novel potent AKR1C3 inhibitor, N-(4-fluorophenyl)-8-hydroxy-2-imino-2H-chromene-3-carboxamide (2d), and synthesized its derivatives with IC50 values of 25-56 nM and >220-fold selectivity over other AKRs (1C1, 1C2, and 1C4). The structural factors for the inhibitory potency were elucidated by crystallographic study of AKR1C3 complexes with 2j and 2l. The inhibitors suppressed proliferation of prostate cancer 22Rv1 and PC3 cells through both androgen-dependent and androgen-independent mechanisms. Additionally, 2j and 2l prevented prostate tumor growth in a xenograft mouse model. Furthermore, the inhibitors significantly augmented apoptotic cell death induced by anti-CRPC drugs (abiraterone or enzalutamide).

Full text of DOI:10.1021/acs.jmedchem.0c00939

Subscriber access provided by University of Rochester | River Campus & Miner Libraries
Article
Development of novel AKR1C3 inhibitors as new potential
treatment for castration-resistant prostate cancer
Satoshi Endo, Hiroaki Oguri, Jin Segawa, Mina Kawai, Dawei Hu, Shuang Xia, Takuya
Okada, Katsumasa Irie, Shinya Fujii, Hiroaki Gouda, Kazuhiro Iguchi, Takuo Matsukawa,
Naohiro Fujimoto, Toshiyuki Nakayama, Naoki Toyooka, Toshiyuki Matsunaga, and Akira Ikari
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00939 • Publication Date (Web): 27 Aug 2020
Downloaded from pubs.acs.org on August 27, 2020
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 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Development of novel AKR1C3 inhibitors as new potential treatment for castration-resistant
prostate cancer
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
†
,
†
†
†
‡
‡
‡
Satoshi Endo *, Hiroaki Oguri , Jin Segawa , Mina Kawai , Dawei Hu , Shuang Xia , Takuya Okada ,
Katsumasa Irie§,∥, Shinya Fujii , Hiroaki Gouda , Kazuhiro Iguchi , Takuo Matsukawa , Naohiro
⊥
#
∇
○
○
◆
‡
¶
†
Fujimoto , Toshiyuki Nakayama , Naoki Toyooka , Toshiyuki Matsunaga , Akira Ikari
†
Laboratory of Biochemistry, Gifu Pharmaceutical University, Gifu 501-1196, Japan.
‡
Graduate School of Innovative Life Science, University of Toyama, Toyama 930-8555, Japan
Cellular and Structural Physiology Institute, Nagoya University, Furo-cho, Chikusa, Nagoya, 464-
601, Japan.
§
8
∥
Graduate School of Pharmaceutical Sciences, Nagoya University, Furo-cho, Chikusa, Nagoya, 464-
601, Japan.
8
⊥
Institute of Biomaterials and Bioengineering, Tokyo Medical and Dental University, Tokyo 101-
0
062, Japan
#
School of Pharmacy, Showa University, Tokyo 142-8555, Japan.
∇
Laboratory of Community Pharmacy, Department of Pharmacy, Gifu Pharmaceutical University,
Gifu 501-1196, Japan.
○
Department of Urology, University of Occupational and Environmental Health, Kitakyushu 807-
8555, Japan.
◆
Department of Pathology, University of Occupational and Environmental Health, Kitakyushu 807-
8
555, Japan.
¶
Education Center of Green Pharmaceutical Sciences, Gifu Pharmaceutical University, Gifu 502-8585,
Japan
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 50
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
ABSTRACT
Aldo-keto reductase (AKR) 1C3 catalyzes the synthesis of active androgens that promote the
progression of prostate cancer. AKR1C3 also contributes to androgen-independent cell proliferation
and survival through the metabolism of prostaglandins and reactive aldehydes. Due to its elevation in
castration-resistant prostate cancer (CRPC) tissues, AKR1C3 is a promising therapeutic target for
CRPC. In this study, we found a novel potent AKR1C3 inhibitor, N-(4-fluorophenyl)-8-hydroxy-2-
imino-2H-chromene-3-carboxamide (2d), and synthesized its derivatives with IC50 values of 25-56
nM and >220-fold selectivity over other AKRs (1C1, 1C2 and 1C4). The structural factors for the
inhibitory potency were elucidated by crystallographic study of AKR1C3 complexes with 2j and 2l.
The inhibitors suppressed proliferation of prostate cancer 22Rv1 and PC3 cells through both androgen-
dependent and -independent mechanisms. Additionally, 2j and 2l prevented prostate tumor growth in
a xenograft mouse model. Furthermore, the inhibitors significantly augmented apoptotic cell death
induced by anti-CRPC drugs (abiraterone or enzalutamide).
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
INTRODUCTION
Prostate cancer is the second most common cancer in Europe and USA with 1.3 million new cases
1
diagnosed in 2018. The causes of prostate cancer have not been fully clarified, but as androgens play
an important role in growth of prostate tumor, androgen deprivation therapy is the standard first-line
approach for metastatic hormone-sensitive prostate cancer. Despite the initial efficacy of androgen
deprivation therapy, unfortunately most patients with prostate cancer eventually develop resistance to
this therapy and tumor progression to castration-resistant prostate cancer (CRPC). The two
mechanisms, androgen-dependent and androgen-independent pathways, are suggested to be involved
2
in the development of CRPC. In the androgen-dependent pathway, it has been reported that the
development of CRPC is proceeded by upregulated androgen receptor (AR) gene expression,3, 4
a
switch from antagonist to agonist by missense mutations such as W741C and T877A in the ligand
5
binding domain of AR, and hyperproliferation by expression changes and/or mutations of AR
2
ACS Paragon Plus Environment

Page 3 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6
coupling factor. By contrast, the androgen-independent pathway includes several mechanisms:
Constitutive activation of AR signaling pathway by expression of AR splicing variants such as
ARv567es lacking AR ligand binding domain and AR-V7,7, 8 increased expression of anti-apoptotic
protein B-cell lymphoma 2 (Bcl-2), and a suppression of tumor suppressor signals due to gene
mutations in phosphatidylinositol 3,4,5-trisphosphate 3-phosphatase and dual-specificity protein
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
9
phosphatase and p53. Recent hormone-targeting drugs such as abiraterone acetate (a cytochrome
P450 17A1 inhibitor), enzalutamide and apalutamide (AR antagonists) are reported to significantly
increase survival rate of CRPC patients.10, 11 However, long-term use of these drugs is associated with
side effects, and their tolerance is common. It is suggested to increase clinical efficacy of the above
agents by combined use of two drugs acting on different targets, although several mechanisms have
been proposed for the drug resistance. In addition, it is desired to search for new molecular targets for
anti-CRPC drug development.
One of the new targets for anti-CRPC drug development is aldo-keto reductase (AKR)1C3, which
belongs to the AKR1C subfamily (https://www.med.upenn.edu/akr/). Human members of this
subfamily are four NADPH-dependent hydroxysteroid dehydrogenases (HSDs) with different
positional and stereo selectivity with respect to their steroid substrates: AKR1C1 (20α-HSD),
AKR1C2 (type-3 3α-HSD), AKR1C3 (type-2 3α-HSD/type-5 17β-HSD) and AKR1C4 (type-1 3α-
HSD).12, 13 Although the four AKR1Cs share over 84% amino acid sequence identity, AKR1C3
uniquely catalyzes the metabolism of prostaglandin (PG) H and PGD into PGF and 9α,11β-PGF ,
2
2
2α
2
respectively.14 AKR1C3 is highly expressed in human prostate and its cancerous tissues, where the
enzyme is involved both in the synthesis of androgens such as testosterone and 5α-dihydrotestosterone
1
5
12,14
16
(
DHT) and in the decreased levels of antiproliferative PGD and 15-deoxy-Δ
-PGJ . In addition,
2
2
silencing of AKR1C3 and the treatment with its inhibitor result in growth inhibition of prostate cancer
cells.17-19 In CRPC, AKR1C3 is overexpressed and mediates the intracrine formation of androgens.15,
2
0-22
Furthermore, AKR1C3 participates in the chemotherapeutic resistance to abiraterone and
enzalutamide.21, 22 Thus, AKR1C3 closely relates to cell proliferation of prostate cancer and is
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 50
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
attracting attention as a potential therapeutic target not only for the prevention and treatment of CRPC
but also for overcoming the drug resistance.
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
Many synthetic and natural inhibitors for AKR1C3 have been reported and covered in recent
reviews.23-25 Since AKR1C3 is structurally similar to other AKR1C isoforms (AKR1C1, AKR1C2,
and AKR1C4) having different physiological roles, the development of specific inhibitors for
AKR1C3 is desired. AKR1C3 inhibitors are classified into steroidal and nonsteroidal types.25, 26 The
typical steroidal inhibitors are medroxyprogesterone acetate27 and steroid lactone.28
Medroxyprogesterone acetate has poor selectivity to AKR1C3, and also inhibits AKR1C4. Steroid
lactone is an estrogen derivative having a lactone on the D ring of the steroid, and shows an nM order
of IC value, but its inhibitory activity for other AKR1C isoforms has not been investigated. On the
5
0
other hand, non-steroidal inhibitors with high selectivity to AKR1C3 over other AKR1C isoforms are
analogues of flufenamic acid,29, 30 indomethacin,31 naproxen,32 cinnamic acid18, 33-35 and long-chain
3
6
polyunsaturated fatty acids. So far, two following compounds were tested in clinical trials. ASP9521
was tested in a small Phase I-IIb clinical trial for CRPC, but unfortunately it was shown to be well
tolerated but without efficacy.37, 38 The phase II clinical trial of BAY1128688, which has been
developed to treat endometriosis by Bayer Healthcare, was recently terminated on account of the
3
9
hepatotoxicity due to the non-specific inhibition of AKR1D1. Thus, no AKR1C3 selective inhibitors
have been clinically applied, although many AKR1C3 inhibitors have been reported.
In this study, we searched for a novel AKR1C3 inhibitor having distinct molecular skeletons from
the above known inhibitors by examining the inhibitory activities of chromene derivatives that had
been synthesized for development of carbonyl reductase 1 (CBR1) and AKR1B10 inhibitors.40, 41
Among the derivatives, N-(4-fluorophenyl)-8-hydroxy-2-imino-2H-chromene-3-carboxamide (2d)
with low CBR1 inhibitory potency was found to potently inhibit AKR1C3. To obtain more potent and
selective AKR1C3 inhibitors, derivatives of 2d were synthesized and evaluated for their inhibitory
potency and selectivity towards AKR1C3. To comprehend binding modes of the inhibitors in the
active site of AKR1C3, we also solved the crystal structures of AKR1C3 with the potent inhibitors.
4
ACS Paragon Plus Environment

Page 5 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Furthermore, to verify the efficacy of the AKR1C3 inhibitors as therapeutic agents for prostate cancer,
their effects on the survival and growth of the androgen-dependent and androgen-independent prostate
cancer cells and the combined effects with CRPC agents were evaluated.
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 AND DISCUSSION
Design and Synthesis of Chromene Derivatives. Desired 8-OH chromene derivatives (2a-z) were
synthesized by the Knoevenagel condensation of 2,3-dihydroxybenzaldehyde and corresponding
cyanoamide, and subsequent ring closure to iminochromene system (Scheme 1). The derivatives with
a hydroxyl group at different positions on the chromene ring (3-5), and 8-methoxy derivative (6) were
also synthesized by the same reaction process for 2a-z by the condensation of 1d with appropriate
benzaldehyde as shown in Scheme 2. Compound 2d showing low inhibitory potency for CBR1 (IC50
4
40 nM) inhibited AKR1C3 more potently (IC 25 nM) among the chromene derivatives previously
50
synthesized.48 The IC50 value was much lower than that (4.1 μM) for the most used commercial
AKR1C3 inhibitor, indomethacin, determined under the same assay conditions. To define the
structural factors responsible for the inhibitory potency of 2d for AKR1C3, we first synthesized the
derivatives (3-5) with a hydroxy group at different positions on the chromene ring, and examined their
inhibitory activity (Table 1). These derivatives and an 8-methoxy derivative of 2d (6) showed less
inhibitory potency than 2d, indicating that the 8-hydroxy group on the chromene ring is a key structural
requisite for the potent inhibition. Next, we synthesized 2d-derivatives with different substituents on
the phenyl ring to examine the effect of the substituents on AKR1C3 inhibition. As the IC value of
50
the derivative with no substituent (2a) was similar to that of 2d, no large changes in inhibitory potency
were observed by introduction of fluoro (2b, 2c, 2r), chloro (2e-g), methyl (2h-j), ethyl (2k-m),
trifluoromethyl (2s-u, 2v) and isopropyl (2w, 2x) groups at 2, 3 and 4 positions on the phenyl ring
(Table 2). Apart from three derivatives (2q, 2n and 2p) exhibiting weak inhibitory activities, the
structure around the phenyl ring may not significantly contribute to the inhibitory potency of 2d.
Table 1. Inhibitory Effects of the Chromene Derivatives on AKR1C3
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 50
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
Entry
R1
IC (nM)
50
2
d
8-OH
5-OH
6-OH
7-OH
8-OMe
25 ± 2.0
370 ±10
> 10,000
> 10,000
> 10,000
3
4
5
6
6
ACS Paragon Plus Environment

Page 7 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2. Inhibition of AKR1C3 by Derivatives with Different Substituents on the Phenyl Ring of 2d
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
ꢀ
ꢀꢀꢀꢀ
Entry
R2
IC (nM)
Entry
R2
IC (nM)
50
50
2
a
H
36 ± 0.5
64 ± 5
2m
2n
2o
2p
2q
2r
2s
4’-Et
2’-OH
45 ± 0.5
>10,000
60 ± 0
2
b
2’-F
2
c
3’-F
29 ± 3
3’-OH
2
d
4’-F
25 ± 2
4’-OH
>10,000
220 ± 30
47 ± 0.5
20 ± 0.1
50 ± 0.3
24 ± 0.1
58 ± 0.1
30 ± 2.2
47 ± 0.3
2e
2f
2’-Cl
3’-Cl
4’-Cl
2’-Me
3’-Me
4’-Me
2’-Et
3’-Et
56 ± 7
3’, 4’-di-F
3’,4’,5’-F
2’-CF3
46 ± 0.7
52 ± 4
2g
2h
2i
52 ± 5.5
55 ± 0.9
27 ± 0.8
21 ± 0.7
34 ± 0.5
2t
3’-CF3
2u
2v
2w
2x
4’-CF3
2j
2’,4’-di-CF3
2’-iso-Pr
3’-iso-Pr
2k
2l
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 50
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
To examine the inhibitory selectivity for AKR1C3, we compared the effects of representative potent
AKR1C3-inhibitory derivatives on the activities of structurally similar AKR1C1, AKR1C2 and
AKR1C4 (Table 3). Although the lead compound 2d inhibited AKR1C1 weakly with an IC value of
5
0
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
6
.0 μM, the other derivatives did not inhibit other AKR1C enzymes at 10 μM concentrations. These
derivatives still retained CBR1-inhibitory potency, but the IC values for the derivatives other than
5
0
2
e were higher than those against AKR1C3. As the IC values of all the synthesized compounds to
50
CBR1 are shown in a supplemental Table S3, only 2e showed the lowest IC value, but the values for
5
0
other compounds were higher than 75 nM.
Table 3. Selectivity of Compounds (2c-f, 2j, 2l and 2s)
IC (nM)
5
0
Entry
AKR1C3
29 ± 3
AKR1C1
AKR1C2
AKR1C4
CBR1
>
10,000
>10,000
(>340)
>10,000
(>340)
370 ± 45 b
(13)
2
c
(
>340) a
6,000 ± 1,000
(240)
>10,000
(>400)
>10,000
(>400)
440 ± 86 b
(18)
2d
25 ± 2
56 ± 7
>
10,000
>180)
10,000
>220)
>10,000
(>180)
>10,000
(>180)
34 ± 3.5 b
(0.6)
2
e
(
120 ± 15 b
>
>10,000
(>220)
>10,000
(>220)
2
f
46 ± 0.7
27 ± 0.8
34 ± 0.5
(
(3)
>10,000
(>370)
>10,000
(>370)
>10,000
(>370)
270 ± 15
(10)
2j
>
10,000
>300)
>10,000
(>300)
>10,000
(>300)
250 ± 28
(7)
2
l
(
8
ACS Paragon Plus Environment

Page 9 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
>
10,000
>500)
>10,000
(>500)
>10,000
(>500)
75 ± 7.9
(4)
2
s
20 ± 0.1
(
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
a
b
Ratios of other enzymes/AKR1C3 are given in parentheses.
40
The IC values for CBR1 are taken from Ref.
5
0
Binding Modes of 2j and 2l Co-crystallized in Complex with AKR1C3. To understand the enzyme-
+
inhibitor interactions, we determined the crystal structures of AKR1C3 in complex with NADP and
inhibitors, 2j and 2l, which differ in the substitutes (4-methyl and 3- ethyl, respectively) on the phenyl
+
+
ring. The structures of the binary (AKR1C3/NADP ) and ternary complexes (AKR1C3/NADP /2j and
+
AKR1C3/NADP /2l) were solved and refined at 1.7 Å and 1.86 Å resolution, respectively, and their
refinement statistics for the geometry and stereochemistry of the final models are summarized in Table
S1. The final models of the ternary complexes consisted of 319 amino acid residues, missing the first
four N-terminal residues due to lack of corresponding electron density, whereas one of the binary
complex consisted of 316 amino acid residues except for the first four N-terminal residues and the last
+
three C-terminal residues. As the AKR1C3/NADP /2j structure was shown in Figure 1A, two enzyme
molecules per asymmetric unit were found for the three complexes, and each enzyme structure was
+
comprised of an eight-stranded α/β barrel with NADP bound in an extended conformation adjacent
to an active site located at the C-terminal end of the barrel. In the active site of the AKR1C3/NADP+
structure, an acetate molecule was found with its carboxyl group pointing toward the catalytic residues,
+
Tyr55 and His117 (Figure 1B). In the AKR1C3/NADP /2j structure, well-defined electron density
observed for the inhibitor molecules allowed unambiguous fitting of the inhibitor that was anchored
from its 8-hydroxyl group by forming H-bonds with Tyr55 and His117 (Figure 1C). The inhibitor-
binding pocket was lined by side-chains contributed by 14 amino acid residues (Tyr54, Tyr55, Trp86,
Ser87, His117, Ser118, Met120, Leu122, Phe139, Asn167, Trp227, Phe306, Phe311 and His314). In
addition, the side-chain OH of Ser129 formed an H-bond network with the carbonyl oxygen of the
amide-bond of 2j via a water molecule (Figure 1D), and its conformation differed from that in the
+
AKR1C3/NADP structure, suggesting that this H-bond network is resulted from an ‘induced fit’
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 50
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
mechanism due to the binding of 2j. The above two H-bond interactions were also observed in the
+
AKR1C3/NADP /2l structure, which is shown as two enzyme molecules in the asymmetric unit
(Figures 1E and 1F). Although 2l was surrounded by the same 14 amino acid residues as in 2j, the 3-
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
ethylphenyl ring of 2l was bound upside down to the two enzyme molecules, which suggests that the
interactions of the ethylphenyl moiety to the surrounding residues (Ser118, Met120, Leu122, Phe139,
Phe311 and His314) are weaker than those of other parts of the inhibitor to the enzyme residues. This
flexibility would make the binding of 2l weaker than that of 2j entropically, and may explain why the
modification of its phenyl ring moiety (Table 2) did not significantly improve the inhibitory potency
of 8-hydroxychrome inhibitor.
10
ACS Paragon Plus Environment

Page 11 of 50
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
+
Figure 1. Crystal structures of AKR1C3 binary and ternary complexes with NADP and 2j or 2l. The
+
+
+
atomic coordinates for the AKR1C3/NADP , AKR1C3/NADP /2j and AKR1C3/NADP /2l were
deposited in the Protein Data Bank (PDB codes 7C7F, 7C7G and 7C7H, respectively), and will be
released upon article publication. (A) Schematic representation of the overall structure of the ternary
+
complex with 2j (magenta) and NADP (yellow) in the asymmetric unit. The AKR1C3 chains, A and
+
B, are shown in pale cyan and pale green, respectively. (B) Active site of the AKR1C3/NADP binary
complex structure, showing a bound acetate (green). Fo-Fc electron density maps contoured at 3.0 σ
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 50
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
around the active site pocket (blue mesh). To provide Fo-Fc map, structures of the protein, NADP+
and solvent water (red) are used for the phasing. The H-bonds are shown as black dotted lines with
corresponding distances in Å. (C) 2j-bound site of the ternary complex, showing the surrounding
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
+
residues (within 4.0 Å from 2j, magenta) and NADP (yellow). (D) Close-up view of the H-bond
network between Ser129 and 2j through a water molecule (red) in the ternary complex, which are
+
superimposed with the AKR1C3/NADP binary complex. The residues of the binary and ternary
complexes are shown in white and light blue, respectively. (E, F) Binding modes of 2l in the binding
+
pockets of AKR1C3 chain A (E) and chain B (F) in the asymmetric unit. NADP (yellow) and residues
(light blue) within 4.0 Å from 2l (magenta) are depicted with H-bonds (dotted line) and their distances
are shown in Å.
Effects of Novel AKR1C3 Inhibitors on Androgen-Dependent Proliferation of Prostate Cancer
2
2Rv1 Cells. To evaluate this effect, we selected androgen-dependent 22Rv1 cells, because the cells
expressed both AKR1C3 and AR (Figure S1). The proliferation of the cells was promoted by treatment
with pregnenolone, a precursor of androgens (Figure 2A). Since this promotive effect by pregnenolone
was in a dose-dependent manner up to 20 nM, the cells were treated with 20 nM pregnenolone in
subsequent proliferation experiments. The pretreatment of the cells with 2d or 2j resulted in a
significant suppression of the pregnenolone-induced proliferation, which was comparable to those by
AR antagonists, flutamide and enzalutamide, and a CYP17A1 inhibitor, abiraterone (Figure 2B).
When the changes in cell cycle were analyzed, the treatment with pregnenolone significantly increased
S-phase cell population, and the pretreatment with 2d or 2j caused not only a decrease in the S-phase
cell population, but also an increase in the G0/G1-phase cell population (Figure 2C). These data
suggest that 2d and 2j arrest the cells in the G0/G1 phase.
Next, we investigated whether the antiproliferative effects of 2d and 2j result from suppression of
androgen signaling through inhibition of AKR1C3 in the cells. AR that is activated by binding of
androgens initially localizes in the cytoplasm in an inactive complex with heat shock proteins.
Androgen binding releases the heat shock proteins and AR translocates to the nuclei, where it binds to
12
ACS Paragon Plus Environment

Page 13 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the androgen-response element within the regulatory regions of target genes, leading to enhancement
of cell proliferation through promoting transcription of target genes such as prostate specific antigen
(
PSA) and transmembrane protease, serine 2 (TMPRSS2).42, 43 To confirm that 2d and 2j work as
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
AKR1C3 inhibitors in the cells, their effects on cellular metabolism of 4-androstene-3,17-dione (4AD)
to testosterone were first examined by quantifying testosterone concentration (Figure 2D). The
testosterone concentration was increased nearly double by the pretreatment with 2 μM finasteride,
which is an inhibitor of steroid 5α-reductase that catalyzes the conversion of testosterone to DHT.44,
4
5
Under the culture conditions using finasteride, the pretreatment of 2d and 2j significantly reduced
the testosterone production. In the absence of finasteride, no significant change in the testosterone
production was observed by the pretreatment of 2d and 2j, suggesting that the compounds do not affect
catalytic (both agonistic and antagonistic) potency of steroid 5α-reductase (Data not shown). Since 2d
and 2j did not affect expression of AKR1C3 in 22Rv1 cells (Figure S2), the data clearly indicate that
the cellular target of the two compounds is AKR1C3. This was further supported by marked lowering
of AR nuclear translocation by the pretreatment with 2d or 2j: In culture conditions without finasteride,
the addition of 4AD greatly increased the nuclear translocation, which was decreased by the
pretreatment of 2d or 2j (Figure 2E). Using pregnenolone as an androgen precursor, we finally
examined the effects of 2d and 2j on AR and AR-target gene expression. As shown in Figures 2F-2H,
pregnenolone enhanced the expression levels of mRNAs for AR, TMPRSS2 and PSA, which were
decreased not only by flutamide, abiraterone and enzalutamide, but also by 2d and 2j. These results
indicate that 2d and 2j inhibit androgen-dependent prostate cancer cell proliferation through
suppression of androgen synthesis by AKR1C3 inhibition. In addition, several synthesized 8-
hydroxychromene derivatives including 2d and 2j exhibited antagonistic activities for AR (Table 4),
although their AR-agonistic activities were not detected (Figure S3). The antagonistic activities of the
8
-hydroxychromene derivatives (IC 2.4-8.1 μM) were lower than that of a representative antagonist,
50
hydroxyflutamide, but are comparable to that (IC50 4.7 μM) of 3-[4’-(nitronaphthalen-1-
amino)]benzoic acid reported as a bifunctional inhibitor for AKR1C3 and AR, although the results
4
6
were obtained in different experiments. Thus, the present 8-hydroxychromene derivatives are potent
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 50
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
suppressors of the androgen signaling suppressors. The compounds still retained moderate CBR1
inhibitory activity, but CBR1 expression is elevated or decreased in several non-prostatic cancers, and
4
7
its role in the cancers is unknown. In addition, CBR1 do not metabolize pregnenolone and 4AD that
were used as the precursors of androgens. The CBR1 inhibition by 2d and 2j may not be involved in
their antiproliferative 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
14
ACS Paragon Plus Environment

Page 15 of 50
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
Figure 2. Suppression of androgen-dependent proliferation of prostate cancer 22Rv1 cells by 2d and
2
j. (A) Cell proliferation by pregnenolone (Preg). The cells were cultured in the absence or presence
of the indicated concentrations of Preg, and the viable cell numbers were estimated at 72 h. Data are
expressed as percentages of the cell number in the Preg-free control cells. Significant difference from
#
#
#
the control cells, p < 0.01, p < 0.05. (B) Effects of 2d, 2j and drugs (Fl, flutamide; Ab, abiraterone;
En, enzalutamide) on the Preg-induced cell proliferation. The cells were pretreated for 2 h without or
with the indicated concentrations of 2d, 2j or the drugs, and then treated for 72 h with 20 nM Preg.
Data are expressed as percentages of the cell numbers in the Preg-free control cells. Significant
#
#
difference from the control cells ( p < 0.01) and the cells treated with Preg alone (**p < 0.01, *p <
.05). (C) Effects of 2d and 2j on cell cycle. The cells were pretreated for 2 h without or with 10 μM
0
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 50
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
2
d or 2j, and then treated for 24 h with 20 nM Preg. The percentages of G0/G1, S, and G2/M phases
#
#
were analyzed using a flow cytometer. Significant difference from Preg-free control cells ( p < 0.01,
#p < 0.05) and the cells treated with Preg alone (**p < 0.01, *p < 0.05). (D) Effects of 2d and 2j on
the metabolism of 4AD into testosterone. The cells were treated for 2 h without or with 2 μM
finasteride and/or AKR1C3 inhibitor (2d or 2j), and then cultured for 24 h in the presence of 100 nM
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
4
AD. The concentration of testosterone in the cell extracts was analyzed by EIA, and is expressed as
#
#
percentage of that in the finasteride-free control cells. Significant difference from the control cells ( p
0.01) and the cells cultured with finasteride alone (**p < 0.01). (E) Effects of 2d and 2j on nuclear
translocation of AR. The cells were pretreated for 2 h without or with 10 μM AKR1C3 inhibitor 2d or
j, and then treated for 24 h with 10 nM 4AD, and the cell extracts were analyzed by Western blotting.
The expression levels of nuclear AR (normalized to that of Lamin A+C) are expressed as percentages
<
2
#
#
of that in the control cells. Significant difference from the control cells ( p < 0.01) and the cells
cultured with 4AD alone (**p < 0.01, *p < 0.05). (F-H) Effect of 2d, 2j and drugs (Fl, Ab, En) on
expression of mRNAs for AR (F), TMPRSS2 (G) and PSA (H). The cells were pretreated for 2 h with
the indicated concentrations of the agents, and then treated with 20 nM Preg for 24 h. The mRNA
expression level in the cells was analyzed by RT-qPCR analysis, and represented the ratio to that in
#
#
the Preg-free control cells. Significant difference from the control cells ( p < 0.01) and the cells
cultured with Preg alone (**p < 0.01, *p < 0.05).
16
ACS Paragon Plus Environment

Page 17 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 4. AR antagonistic activity by means of inhibition of DHT-induced SC-3 cell proliferation. IC50
value represents the concentration required for 50% of the maximum antagonistic activity.
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
Compounds
Hydroxyflutamide
2b
IC (μM)
50
0.23
4.4
3.3
6.5
3.3
2.4
8.1
5.8
3.6
2
c
2
d
2
e
2
j
2
k
2l
2m
Effects of 2d and 2j on Androgen-Independent Proliferation of Prostate Cancer Cells. AKR1C3
was also reported to be involved in proliferation of several cancer cells such as leukemia, breast cancer
and skin squamous cell carcinoma through activation of mitogen-activated protein kinase (MAPK)
pathway or phosphatidylinositol-3 kinase (PI3K) pathway via biosynthesis of PGF and 9α,11β-PGF ,
2
α
2
which are ligands of PGF receptor (Figure 3A).19, 48-50 Then, we evaluated the effects of 2d and 2j on
the above androgen-independent cell proliferation using prostate cancer PC3 cells that express
AKR1C3 but not AR (Figure S1). When the cells were cultured in the presence of 2d or 2j for 72 h,
viable cell numbers were significantly reduced compared to the inhibitor-free cells (Figures 3B and
3
C), indicating that AKR1C3 plays an important role in the above androgen-independent proliferation.
We previously established androgen-low-sensitive LNCaP-E9 cells51 isolated from LNCaP cells,
which consist of heterogeneous cell population with naturally occurring differences in androgen
sensitivity. 52 Due to the androgen-low sensitivity of LNCaP-E9 cells, their pregnenolone-induced
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 50
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
proliferation was lower than that of the parental LNCaP cells, but was reduced to levels similar to
those of LNCaP cells by the treatment with 2d or 2j (Figure 3D). The results imply that the two
inhibitors suppress proliferation of LNCaP cells via inhibition of both androgen-dependent and
androgen-independent pathways. Next, using PC3 cells and their AKR1C3-overexpressing cells, the
antiproliferative effects of 2d and 2j were compared to those of AL-8810 (a PGF receptor antagonist),
U0126 (a MAPK inhibitor) and wortmannin (a PI3K inhibitor) (Figure 3E). The overexpression of
AKR1C3 increased cell proliferation by 1.4-fold compared to the control PC3 cells. The proliferation
of PC3 cells and their AKR1C3-overexpressing cells was significantly suppressed by the treatment
with the four agents other than U0126. The antiproliferative effects of the four agents on the AKR1C3-
overexpressing cells were greater than those on the mock cells. Although future work on molecular
mechanisms underlying the inhibition of androgen-independent cell proliferation by 2d and 2j is
needed, these results suggest, at least in these prostatic cancer cells, that the AKR1C3 inhibitors
suppress the androgen-independent cell proliferation through inhibition of the PGF signaling, where
the MAPK pathway may not be involved.
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
18
ACS Paragon Plus Environment

Page 19 of 50
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
Figure 3. Inhibitory effects of 2d and 2j on androgen-independent proliferation. (A) PG biosynthetic
cascade. PGF2α and 9α,11β-PGF are formed from PGH and PGD , respectively, by AKR1C3, and
2
2
2
bind to PGF receptor, leading to cell proliferation through activation of several signaling pathways
including mitogen-activated protein kinase (MAPK), phosphatidylinositol-3-kinase (PI3K) and/or
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 50
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
nuclear factor-kappa B (NFκB). (B, C) Growth curves of androgen-independent PC3 cells in the
absence or presence of 2d (B) and 2j (C). Significant difference from the inhibitor-free control cells,
**p < 0.01. (D) Effects of 2d and 2j on pregnenolone (Preg)-induced proliferation of androgen-
sensitive LNCaP and androgen-low sensitive LNCaP-E9 cells. The cells were pretreated for 2 h
without or with the indicated concentrations of 2d or 2j, and then treated for 72 h with 20 nM Preg.
The viable cell numbers are expressed as percentages of that of the non-treated control LNCaP cells.
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
#
#
#
p < 0.01, p < 0.05 compared with the control cells. Significant difference from the Preg-free cells
(##
p < 0.01, p < 0.05) and the cells treated with Preg alone ( p < 0.01). (E) Effects of inhibitors on
#
**
growth of AKR1C3-transfected and mock PC3 cells. Upper panel: Extracts (40 μg) of the cells
transfected with the expression vector harboring the cDNA for AKR1C3 (2) and the empty vector (1)
were applied to Western blot analysis using the antibodies against the enzyme and β-actin. The cells
were cultured for 72 h in the absence or presence of 1 µM wortmannin (Wo), 10 µM AL-8810 (AL),
U0126 (U), 2d and 2j. The viable cell numbers are expressed as percentages of that of the inhibitor-
#
#
#
free control mock cells. Significant difference from the control mock cells ( p < 0.01, p < 0.05) and
*
*
between mock and AKR1C3-transfected cells treated with the same agent ( p < 0.01).
Effects of 2j and 2l on Prostate Tumor Growth in a Xenograft Model. To evaluate the anticancer
efficacy of the present AKR1C3 inhibitors in vivo, we investigated the suppressive effects of 2j and
2
l on prostate tumor growth in a xenograft mouse model using 22Rv1 cells. The tumor volume was
significantly reduced by ip injection of the inhibitors, of which 2j was more effective (Figure 4A).
Histopathological analysis revealed that the inhibitor treatment increased necrotic region in the tumor
tissues of mice, in contrast to no specific feature in non-tumor region surrounded the tumor (Figure
4
B). It should be noted that no histopathological changes were observed in other tissues of the
inhibitor-treated mice, as shown representative sections of the lung, liver and kidney of the 2j-treated
mice (Figure 4B). In addition, there was no statistical difference in the average body weights among
the inhibitor-treated and control mice. The results suggest a possibility that the present AKR1C3
inhibitors, especially 2j, are useful as candidates for novel therapeutic agents for CRPC, although
20
ACS Paragon Plus Environment

Page 21 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
future works are needed to elucidate their metabolism and pharmacokinetics.
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
Figure 4. In vivo anticancer effects of 2j and 2l on mouse xenograft models. (A) Tumor sizes of mice
which were intraperitoneally injected with vehicle (control, n=15), 2j (50 mg/kg, n=10) and 2l (100
#
#
#
mg/kg, n=10). Significant difference from the control group, p < 0.01, p < 0.05. (B) Representative
histological sections of the tumor tissue, liver, kidney and lung of the 2j-trated mice. The sections
were stained with hematoxylin and eosin (magnification ×40). An arrow indicates areas of necrosis.
Effects of 2d and 2j on Action of Anti-CRPC Drugs. Recent literature shows that AKR1C3 is
associated with anticancer-drug resistance, and that its inhibitors are effective both to overcome the
resistance and to enhance anticancer potency of anti-CRPC drugs (such as enzalutamide and
abiraterone).21, 22, 53-55 We first investigated whether the present AKR1C3 inhibitors 2d and 2j
contribute to enhancing the anticancer potency of enzalutamide using 22Rv1 cells, which were
cultured in the presence of pregnenolone. The anticancer potency of enzalutamide was evaluated in
terms of DNA fragmentation (Figure 5A), changes in apoptotic signals (Figures 5B and 5C) and cell
proliferation (Figure 5D). The treatment with 50 μM enzalutamide slightly decreased the viable cell
numbers and increased the expression of cleaved caspase-3, but no significant changes were observed
in expression levels of apoptosis-associated proteins (Bax, cleaved poly ADP-ribose polymerase
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 50
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
(
PARP), Bcl-2). By contrast, 2d and 2j slightly induced apoptotic cell death, which is evidenced by
DNA fragmentation and the expression of cleaved caspase-3 and cleaved PARP. The cotreatment of
d or 2j with enzalutamide greatly induced the apoptotic phenotypes (i.e., DNA laddering, increases
2
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 Bax, cleaved PARP and cleaved caspase-3 expression, and a decrease in Bcl-2 expression), which
may cause significant cell death. Next, similar experiments were done using another anti-CRPC drug
abiraterone (Figures 5E and S4). Like enzalutamide, abiraterone slightly provoked apoptotic cell death,
which was markedly enhanced by its cotreatment with 2d or 2j. Since it was reported that the anti-
CRPC drugs such as enzalutamide and abiraterone induce apoptosis to prostate cancer cells through
androgen depletion56, 57 and the castration treatment in rats also induces apoptosis, the enhancement
of the anti-CRPC drug-induced anticancer activity by co-treatment of 2d and 2j may result from
further suppression of intracellular androgen signaling.
58
22
ACS Paragon Plus Environment

Page 23 of 50
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
Figure 5. Enhancement of enzalutamide (En)-evoked apoptotic cell death by cotreatments of 2d and
2
j. (A-D) The 22Rv1 cells were pretreated for 2 h with or without 10 μM 2d, 2j and/or 50 μM En, and
then treated for 24 h with 20 nM pregnenolone (Preg). (A) DNA fragmentation. DNA fragmentation
was visualized as oligonucleosome-sized fragmentation in ethidium bromide after DNA agarose gel
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 50
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
electrophoresis. (B) Western blot analysis of proapoptotic and anti-apoptotic proteins. The cell extracts
were applied to the analysis using antibodies against Bax, Bcl-2, cleaved PARP and β-actin. The ratio
of Bax/Bcl-2 is expressed as a percentage of that in the control cells. Significant difference from the
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
#
control cells ( p < 0.05) and the cells co-treated with Preg and En (**p < 0.01). (C)
Immunofluorescence analysis. The cell lysate was immunoblotted with anti-cleaved caspase-3
antibody (green) and 4',6-diamidino-2-phenylindole stained DNA (DAPI, blue). Merged images were
also shown. (D) Cell viability. (E) Viability of the cells pretreated with or without 10 μM 2d, 2j and/or
1
0 μM abiraterone (Ab). Data are expressed as percentages of the cell viability in the control cells
#
#
treated with the vehicle DMSO alone. Significant difference from the cells treated with Preg alone ( p
#
<
0.01, p < 0.05) and the cells co-treated with Preg and En or Ab (**p < 0.01).
AKR1C3 is also reported to be involved in cancer cell survival via reductive detoxification of
reactive aldehydes such as 4-hydroxy-2-nonenal and 4-oxo-2-nonenal, which are produced through
lipid peroxidation during induction of oxidative stress.59 To mechanistically elucidate the
enhancement of apoptotic cell death by the cotreatment of 2d and 2j with the anti-CRPC drugs, we
furthermore examined production of reactive oxygen species (ROS) in the agents-treated cells (Figure
6
). Intracellular ROS were produced by a single treatment of any anti-CRPC drugs or inhibitors, and
their amounts were further increased by the cotreatment of the drug and inhibitor in a synergistic or
interactive manner (Figures 6A and 6B). Therefore, it was suggested that 2d and 2j increase the
intracellular ROS levels by inhibiting AKR1C3, which is involved in the elimination of ROS-mediated
reactive aldehydes generated by enzalutamide and abiraterone. CBR1 catalyzes the reduction of
anthracyclines (doxorubicin and daunorubicin) to less potent but cardiotoxic metabolites, and its
inhibition leads to the increase in chemotherapeutic efficacy and prevention of cardiotoxic side effect
as reviewed by Shi and Di.47 However, there have been no reports on the role of CBR1 in the
metabolism and anti-cancer actions of enzalutamide and abiraterone. Therefore, we consider that the
CBR1 inhibition by the chromene derivatives may not affect the sensitivity to the CRPC drugs such
as enzalutamide and abiraterone. Most recently, it has been reported that AKR1C3 is involved in
24
ACS Paragon Plus Environment

Page 25 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
acquired resistance to novel approved drugs for CRPC such as apalutamide and darolutamide as well
as enzalutamide and abiraterone, and that targeting AKR1C3 confers the drug resistance. 60, 61 We also
suggested that AKR1C3 polymorphism is a prognostic factor of limitation to androgen deprivation
therapy in patients with metastatic prostate cancer.62 Therefore, targeting AKR1C3 may be a
potentially promising approach for treatment of prostate cancer and prevention and/or overcoming the
CRPC drug resistance.
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
Figure 6. Increase in abiraterone (Ab) or enzalutamide (En)-induced ROS production by AKR1C3
inhibitors. The 22Rv1 cells were treated with 10 μM 2d, 2j and/or Ab, 50 μM En for 3h, and their
intracellular ROS levels were measured by flow cytometric analysis. The DCF-positive cells (shown
with a red horizontal line) in the cytometric data (A) are expressed as percentages of those in the
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 50
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
control cells treated with the vehicle alone in the bar graph (B). Significant difference from the control
#
#
#
cells ( p < 0.01, p < 0.05) and the cells treated with Ab or En alone (**p < 0.01).
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
CONCLUSIONS
Herein, we have described the discovery of potent AKR1C3 inhibitors with high selectivity to this
enzyme over other human AKR1C subfamily members. The crystallographic study of the enzyme-
+
NADP -inhibitor (2j or 2l) ternary complexes revealed the structural determinants for their potent
inhibitory activities. Our ex vivo and in vivo experiments demonstrated that the inhibitors exhibit
excellent antiproliferative activity against prostate cancer 22Rv1 cells through suppression of
androgen signaling by dual inhibition of AKR1C3 and AR. The inhibitors also suppressed the
proliferation of androgen-independent and androgen-low sensitive prostate cancer cells probably
through inhibition of PGF signaling. Moreover, the inhibitors enhanced apoptotic cell death induced
by CRPC agents, abiraterone and enzalutamide, which suggest the effectiveness of their combination
therapy both for increase in clinical efficacy of the known drugs and for overcoming the resistance.
Taken together, we conclude that the present inhibitors may be useful as candidates of drugs for
treatments against prostate cancer overexpressing AKR1C3.
26
ACS Paragon Plus Environment

Page 27 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
EXPERIMENTAL SECTION
. Chemistry:
General
1
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
Chemicals were purchased from Sigma-Aldrich, Merck (Darmstadt, Germany), FUJIFILM Wako
Chemicals (Osaka, JAPAN), Nakalai Tesque, Tokyo Chemical Industry (Tokyo, Japan) and Kanto
Chemical (Tokyo, Japan), and used without further purification. Column chromatography was done
on Cica silica gel 60N (spherical, neutral; particle size, 63-210 nm, Kanto Chemical), while thin-layer
chromatography (TLC) was performed using Merck silica gel 60F254 plates. Melting points were taken
on a Yanaco micromelting point apparatus and are uncorrected. The nuclear magnetic resonance
(NMR) spectra were acquired in the specified solvent, in a JEOL JNM-A400 (400 and 100 MHz for
1H and 13C, respectively) or JEOL JNM-ECX500 (500 and 125 MHz for H and 13C, respectively).
The chemical shifts (δ) are reported in ppm downfield from TMS and coupling constants (J) are
expressed in Hertz. IR spectra were measured with a JASCO FT/IR-460 Plus spectrophotometer
1
(JASCO Corp., Tokyo, Japan). The low-resolution mass spectra (MS) and high-resolution mass
spectra (HRMS) were obtained with a Shimadzu GCMS-QP 500 mass spectrometer (Shimadzu Corp.,
Kyoto, Japan), JEOL D-200, or JEOL AX505 mass spectrometer (JEOL Ltd., Tokyo, Japan) in the
electron impact mode at the ionization potential of 70 eV. The purities (> 95%) of 2a-x were analyzed
by elemental analysis and HPLC. The elemental analyses of 2a-x was performed using a Thermo
Scientific FlashEA1112 analyzer. Analytical HPLC of 2d, 2j, and 2l performed on HITACHI
Chromaster 5110 pump with HITACHI Chromaster 5410 UV detector using Mightysil RP-18 GP (5
µL, 4.6 mm × 150 mm) column and MeCN : 0.1% formic acid = 2 : 1 as mobile phase with flow rate
of 2.0 mL/min at 20 °C.
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 50
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
Scheme 1. Synthesis of 8-OH chromene derivatives 2a-x
General procedure for the synthesis of cyanoamides (1a-x).
To a stirred solution of cyanoacetic acid (100 mg, 1.18 mmol) in CH Cl (5 mL) were added
2
2
corresponding aniline (1.18 mmol), EDC•HCl (450 mg, 2.35 mmol), and DMAP (29 mg, 0.24 mmol).
The resulting mixture was stirred at room temperature for 24 h. The volatiles were evaporated, and the
residue was chromatographed on column chromatography (SiO , hexane : acetone = 5 : 1 ~ 1 : 1) to
2
40
63
64
63
65
66
40
67
68
67
69
afford the corresponding amides (1a-g, 1h, 1i, 1j-k, 1l, 1m, 1n-p, 1q, 1r, 1s, 1t, 1u,
63
67
70
1
v-w , 1x ) as a colorless solid.
General procedure for the synthesis of 8-OH chromene derivatives (2a-x).
To a stirred solution of amide (1a-x, 0.43 mmol) in EtOH (3 mL) were added 2,3-
dihydroxybenzaldehyde (59 mg, 0.43 mmol), and piperidine (3 drops). The resulting mixture was
stirred at room temperature for 24 h. The insoluble materials are corrected by filtration, washed with
EtOH, and dried to give rise to the corresponding amides (2a-x) as a pale yellow solid.
8
-Hydroxy-2-imino-N-phenyl-2H-chromene-3-carboxamide (2a)
1
Yield: 82%; mp: 235-237 °C; H-NMR (500 MHz, DMSO-d ) δ 7.08-7.11 (2H, m), 7.12 (1H, t, J =
6
28
ACS Paragon Plus Environment

Page 29 of 50
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
8
.0 Hz), 7.23 (1H, dd, J = 8.0, 3.0 Hz), 7.37 (1H, t, J = 8.0 Hz), 7.67 (2H, dd, J = 8.0, 1.0 Hz), 8.50
1
3
(1H, d, J = 1.0 Hz), 9.09 (1H, d, J = 1.0 Hz), 10.23 (1H, br); C-NMR (125 MHz, DMSO-d ) δ 119.4,
6
1
3
2
19.6, 119.7, 119.9, 120.0, 124.0, 124.1, 129.1, 138.3, 142.0, 142.1, 144.0, 156.0, 159.8; IR (KBr):
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
-1
+
298, 1674, 1598, 1473 cm ; MS (EI): m/z 280 (M ); HRMS: Calcd for C H N O 280.0848, Found
1
6
12
2
3
80.0847; Anal. Calcd for C H N O : C, 68.56; H, 4.32; N, 9.99, Found: C, 68.22; H, 4.03; N, 10.03.
1
6
12
2
3
N-(2-Fluorophenyl)-8-hydroxy-2-imino-2H-chromene-3-carboxamide (2b)
1
Yield: 82%; mp: 247-249 °C; H-NMR (500 MHz, DMSO-d ) δ 7.07-7.12 (2H, m), 7.13-7.17 (1H,
6
m), 7.21 (1H, d, J = 8.0 Hz), 7.24 (1H, dd, J = 8.0, 2.0 Hz), 7.31 (1H, ddd, J = 11.0, 8.0, 1.0 Hz), 8.46
1
3
(1H, td, J = 8.0, 2.0 Hz), 8.53 (1H, d, J = 2.0 Hz), 9.12 (1H, s), 10.23 (1H, br); C-NMR (125 MHz,
DMSO-d ) δ 115.1 (d, J = 19.4 Hz), 119.3, 119.6, 119.8, 120.2, 121.4, 124.1, 124.5 (d, J = 7.3 Hz),
6
1
24.7 (d, J = 3.6 Hz), 126.6 (d, J = 9.8 Hz), 142.0, 142.6, 144.0, 152.4 (d, J = 241.8 Hz), 155.8, 160.1;
-1
+
IR (KBr): 3310, 1607, 1569, 1473 cm ; MS (EI): m/z 298 (M ); HRMS: Calcd for C H FN O
3
1
6
11
2
2
98.0754, Found 298.0754; Anal. Calcd for C H FN O : C, 64.43; H, 3.72; N, 9.39, Found: C, 64.61;
16 11 2 3
H, 3.40; N, 9.33.
N-(3-Fluorophenyl)-8-hydroxy-2-imino-2H-chromene-3-carboxamide (2c)
1
Yield: 51%; mp: 225-227 °C; H-NMR (500 MHz, DMSO-d ) δ 6.95 (1H, td, J = 8.0, 2.0 Hz), 7.06-
6
7
.11 (2H, m), 7.23 (1H, dd, J = 8.0, 2.0 Hz), 7.28 (1H, d, J = 8.0 Hz), 7.39 (1H, q, J = 8.0 Hz), 7.74
1
3
(1H, dt, J = 11.0, 2.0 Hz), 8.50 (1H, d, J = 2.0 Hz), 9.11 (1H, s), 10.22 (1H, br); C-NMR (125 MHz,
DMSO-d ) δ 107.0 (d, J = 23.2 Hz), 110.0 (d, J = 23.2 Hz), 116.0, 119.8, 120.1, 120.3, 120.6, 124.6
6
(d, J = 1.3 Hz), 131.1 (d, J = 10.4 Hz), 140.4 (d, J = 10.4 Hz), 142.5, 142.9, 144.5, 156.4, 160.7, 162.8
-1
+
(d, J = 239.4 Hz); IR (KBr): 3305, 1609, 1579, 1473 cm ; MS (EI): m/z 298 (M ); HRMS: Calcd for
C H FN O 298.0754, Found 298.0754; Anal. Calcd for C H FN O : C, 64.43; H, 3.72; N, 9.39,
16 11
2
3
16 11
2
3
Found: C, 64.36; H, 3.35; N, 9.49.
N-(4-Fluorophenyl)-8-hydroxy-2-imino-2H-chromene-3-carboxamide (2d)
29
ACS Paragon Plus Environment