Bioisosteres of Indomethacin as Inhibitors of Aldo-Keto Reductase 1C3

Article abstract of DOI:10.1021/acsmedchemlett.8b00484

Aldo-keto reductase 1C3 (AKR1C3) is an attractive target in drug design for its role in resistance to anticancer therapy. Several nonsteroidal anti-inflammatory drugs such as indomethacin are known to inhibit AKR1C3 in a nonselective manner because of COX-off target effects. Here we designed two indomethacin analogues by proposing a bioisosteric connection between the indomethacin carboxylic acid function and either hydroxyfurazan or hydroxy triazole rings. Both compounds were found to target AKR1C3 in a selective manner. In particular, hydroxyfurazan derivative is highly selective for AKR1C3 over the 1C2 isoform (up to 90-times more) and inactive on COX enzymes. High-resolution crystal structure of its complex with AKR1C3 shed light onto the binding mode of the new inhibitors. In cell-based assays (on colorectal and prostate cancer cells), the two indomethacin analogues showed higher potency than indomethacin. Therefore, these two AKR1C3 inhibitors can be used to provide further insight into the role of AKR1C3 in cancer.

Full text of DOI:10.1021/acsmedchemlett.8b00484

Subscriber access provided by AUT Library
Letter
Bioisosteres of Indomethacin as Inhibitors
of Aldo-Keto Reductase 1C3 (AKR1C3)
Marco L Lolli, Irene Maria Carnovale, Agnese Chiara Pippione, Weixiao Yuan Wahlgren, Davide Bonanni,
Elisabetta Marini, Daniele Zonari, Margherita Gallicchio, Valentina Boscaro, Parveen Goyal, Rosmarie
Friemann, Barbara Rolando, Renzo Bagnati, Salvatore Adinolfi, Simonetta Oliaro-Bosso, and Donatella Boschi
ACS Med. Chem. Lett., Just Accepted Manuscript • DOI: 10.1021/acsmedchemlett.8b00484 • Publication Date (Web): 28 Jan 2019
Downloaded from http://pubs.acs.org on January 30, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 8
ACS Medicinal Chemistry Letters
1
2
3
4
5
6
7
8
9
Bioisosteres of Indomethacin as Inhibitors of Aldo-Keto Reductase
1
C3 (AKR1C3)
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
†
†
†
‡ ∥
Marco L. Lolli, Irene M. Carnovale, Agnese C. Pippione, Weixiao Y. Wahlgren, Davide
†
†
†
†
†
Bonanni, Elisabetta Marini, Daniele Zonari, Margherita Gallicchio, Valentina Boscaro,
‡
‡
†
§
†
Parveen Goyal, Rosmarie Friemann, Barbara Rolando, Renzo Bagnati, Salvatore Adinolfi,
*†
*†
Simonetta Oliaro-Bosso and Donatella Boschi .
†
Department of Science and Drug Technology, University of Turin, via Pietro Giuria 9, 10125 Turin (Italy).
‡
Department of Chemistry and Molecular Biology, University of Gothenburg, Box 462, S-40530 Gothenburg
(
Sweden).
§
Istituto di Ricerche Farmacologiche "Mario Negri" IRCCS, Via La Masa 19, 20156 Milan (Italy).
KEYWORDS AKR1C3, Prostate cancer; bioisosterism; scaffold hopping; indomethacin; X-ray crystallography.
ABSTRACT: Aldo-keto reductase 1C3 (AKR1C3) is an attractive target in drug design for its role in resistance to anticancer
therapy. Several non-steroidal anti-inflammatory drugs as indomethacin are known to inhibit AKR1C3 in a non-selective
manner because of COX-off target effects. Here we designed two indomethacin analogues by proposing a bioisosteric
connection between the indomethacin carboxylic acid function and either hydroxyfurazan or hydroxy-1,2,5-N1-
methyltriazole rings. Both compounds were found to target AKR1C3 in a selective manner. In particular, hydroxyfurazan
derivative is highly selective for AKR1C3 over the 1C2 isoform (up to 90-times more), and inactive on COX enzymes. High-
resolution crystal structure of its complex with AKR1C3 shed light onto the binding mode of the new inhibitors. In cell-
based assays (on colorectal and prostate cancer cells), the two indomethacin analogues showed higher potency than
indomethacin. Therefore, these two AKR1C3 inhibitors can be used to provide further insight into the role of AKR1C3 in
cancer.
AKR1C3 is a soluble enzyme member of the aldo-
ketoreductase family involved in numerous enzymatic
reactions. It catalyses the NADPH-dependent reduction of
weak steroidal precursors, such as androstenedione, to the
AKR1C3 is inhibited by several classes of compounds,1, 12
although no drugs have yet been approved for clinical use.
Many AKR1C3 inhibitors were discovered by starting
investigations from non-steroidal anti-inflammatory drug
(NSAID) structures, such as flufenamic acid (FLU) and
indomethacin (INDO) (Fig. 1). Such compounds are
usually carboxylic acids and transport into cells is often
1
potent steroidal receptor ligand testosterone. Elevated
AKR1C3 expression has been identified in the sex hormone-
dependent tissues of prostate and breast cancer, as well as
in sex hormone-independent tissues, such as
adenocarcinoma, squamous cell carcinoma, columnar
13
carrier-mediated. The development of non-carboxylate
inhibitors of AKR1C3 is therefore essential. The cross
inhibition of AKR1C subfamily members is a critical
concern when designing AKR1C3 inhibitors, given their
high amino acid sequence identity and structural
similarity. AKR1C1 and AKR1C2 isoforms are closely related
to AKR1C3 and are needed to inactivate DHT in the
2
epithelium, and colorectal cancer. Although a few recent
studies have made controversial observations about the in
3
, 4
vivo effectiveness of AKR1C3-based therapies,
other
studies have supported the use of AKR1C3 as a therapeutic
target in castration-resistant prostate cancer (CRPC) and
5
, 6
1
in Acute Myeloid Leukemia. AKR1C3’s role in resistance
to radiation, and pharmacological, therapies is also an
prostate. Furthermore, COX inhibition has often been
retained by NSAID-analogue AKR1C3 inhibitors. NSAID
analogues that can inhibit AKR1C3 need to be designed
without the structural features that are required for
inhibition of COX, so to be devoid of gastrointestinal
7
emerging theme. AKR1C3 induction appears correlated
with changes in the drug efficacy of important
8
chemotherapeutic agents, such as abiraterone,
9
10
12
enzalutamide, and anthracycline. Up-regulated AKR1C3
is also involved in colon cancer cisplatin resistance and the
treatment of colon-cancer cells with specific AKR1C3
toxicity. Furthermore, selective AKR1C3 inhibitors are an
important tool to better understand AKR1C3’s role in the
development of cancer and other diseases.
1
1
inhibitors improved sensitivity to cisplatin toxicity.
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
The removal of the carboxylic acid moiety in fenamate
Page 2 of 8
The present paper reports synthetic strategies and
biochemical studies for AKR1C3 inhibitors 1 and 2 (Figure
1), also tested for in vitro cytotoxicity against two panels of
human tumour cells: colorectal (HCT 116, HT-29) and
prostate cancer (LNCaP and VCaP) cell lines. To study the
inhibitor interactions within the AKR1C3 binding site, we
determined the high-resolution crystal structure of
AKR1C3 in complex with compound 1. In addition,
physicochemical characterization and preliminary drug-
like properties are also presented.
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
derivatives coincides with a reduced COX1 activity, as
demonstrated in our recent reports on the design of new
1
4, 15
AKR1C3 inhibitors.
In those works, a scaffold hopping
strategy based on the replacement of the FLU benzoate
moiety with three different hydroxyazoles was used. The
design of those ligands enabled us to demonstrate that
hydroxyazoles are valid bioisosters of the carboxylic acid
function versus AKR1C3 inhibitory activity, which can be
deprotonated to various degrees at physiological pH,
1
6-18
depending on their acidic properties.
As part of this
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
Several creative methods for the assembly of N1-
substituted indole compounds have been reported. These
include variations of the Fischer Indole synthesis, as
reported in the refined standard procedure to produce
series of reports, we demonstrated how replacing the
benzoic acid of the known NSAID and AKR1C3 inhibitor
FLU (Figure 1), with
a
4-hydroxy-N1-substituted
triazolecarbonylic moiety (compounds A and B, Figure 1)
led to a bioisosteric replacement for AKR1C3 activity, with
20
INDO. In this procedure, p-methoxyphenylhydrazone 3
(Scheme 1) is benzoylated with 4 at the N1-position to give
the N1,N1-disubstituted hydrazone 5, and then hydrolyzed
with hydrogen chloride to give hydrazine hydrochloride
salt 6, which is cyclized, in the final reaction step, with
levulinic acid to afford INDO. This procedure was modified
by replacing levulinic acid with the appropriate 3-
oxobuthylazoles, 12 and 13, giving the two desired
hydroxyazole derivatives 1 and 2 in a convergent scheme
(Scheme 1).
no
associated
COX1/2
effect.
Moreover,
the
regiosubstitution of the nitrogen atoms on the
hydroxytriazole allowed us to perform a structural
1
9
refinement, leading to improved AKR1C3 binding and to
reduced AKR1C2 and COX1/2 off-target binding.
F3C
NH
OH
F3C
NH OH
N
O
O
N
N
Scheme 1. Synthesis of target compounds 1 and 2a
F3C
NH
O
N
N
N
OH
O
BnO
Br
N
N
N
NH
X
O
O
F3C
3
7 (X=O)
Flufenamic acid
FLU)
A
B
8 (X=N-CH3)
Cl
(
O
a
Cl
O
O
4
c
HO
HO
HO
O
OH
O
9
N
N
O
N
X
O
N
O
N
N
O
N
N
O
O
O
N
N
N
N
N
O
BnO
O
Cl
O
O
O
OtBu
N
N
O
1 (X=O)
Cl
X
5
2 (X=N-CH3)
10 (X=O)
Indomethacin
INDO)
1
2
1
1 (X=N-CH3)
(
f
b
Figure 1. Chemical structures of AKR1C3 inhibitors: FLU,
hydroxytriazoles bioisosteres of FLU (structures A and B),
INDO and the INDO bioisosteres studied in this work
(compounds 1 and 2).
d
BnO
O
O
O
N
X
NH2 . HCl
e
N
BnO
N
N
+
O
N
N
INDO (Figure 1), a potent but unselective AKR1C3
inhibitor with COX1/2 inhibitory activity, overcomes
X
O
Cl
Cl
1
2 (X=O)
3 (X=N-CH3)
15 (X=O)
16 (X=N-CH3)
8
9
6
resistance to new agents (abiraterone, and enzalutamide )
that in turn have been shown to prolong survival in
patients with metastatic CRPC. Considering the potential
clinical use of INDO and the successful modulation of FLU,
we report a bioisosteric strategy based on the replacement
of the INDO carboxylic function with two hydroxyazoles
with different pKa values. Hydroxyfurazan (pKa  3) and
N1-methyl-3-hydroxy-1,2,5-triazole (pKa  6) were chosen
to improve potency and selectivity towards AKR1C3, and
remove COX1/2 off-target effects.
1
aReagents and conditions: (a) Dry pyridine, dry toluene, rt; (b)
HCl (gas), dry toluene, dry MeOH, 10°C; (c) NaH (60% w/w),
dry THF; (d) PTSA, dry toluene, 90°C; (e) CH
3
COOH, 80°C; (f)
H
2
, Pd/C, dry THF.
The benzyl-protected bromomethylhydroxyazoles 7,21
and 8, were the key starting materials for the synthesis of
the 3-oxobuthylazoles 12 and 13 and were prepared as
2
2, 23
previously reported.
Their reaction with tert-butyl
acetoacetate in the presence of NaH in dry THF furnished
the tert-butyl esters 10 and 11, which were hydrolysed and
ACS Paragon Plus Environment

Page 3 of 8
ACS Medicinal Chemistry Letters
decarboxylated via treatment with p-toluenesulfonic acid
PTSA) in dry toluene at 90° C. It is important to note that
hydroxytriazole indomethacin analogue 2 was more active
1
2
3
4
5
6
7
8
9
(
than INDO, and gave an IC50 value (0.94 µM) that is 8-
times lower than that of INDO (IC50 7.35 µM).
Furthermore, it was more selective towards AKR1C3, with
an AKR1C2 IC50 value that is 15-times higher (IC50 14.25 µM).
The new hydroxyfurazan indomethacin analogue 1 showed
a better biochemical profile, presenting a IC50 value (0.30
µM) that is 25-times lower than that of INDO (IC50 7.35
µM), and was also much more selective towards AKR1C3,
with a AKR1C2 IC50 value that is 90-times higher (IC50 26.75
µM). All these data support the bioisosterism between the
carboxylic acid function of INDO and the 3-
the presence of the tert-butyl ester group in 10 and 11 is
mandatory for decarboxylation, as the corresponding ethyl
esters were observed to undergo deacethylation under
these conditions. The key reagent for the underlying
Fischer indolisation step, 4-chloro-N-(4-methoxyphenyl)-
benzohydrazide hydrochloride (6), was prepared
according to more recently reported literature
24
procedures, and allowed to react with ketones 12 and 13
in glacial acetic acid at 80° C for 1 h, giving indole
derivatives 15 and 16. Finally, the catalytic hydrogenation
of 15 and 16 gave the desired hydroxyazoles 1 and 2,
respectively, in quantitative yields.
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
hydroxyfurazan and 3-hydroxyl-N -methyl-1,2,5-triazole
moieties, as AKR1C activity was retained. Moreover, the
bioisosteric replacement let to these new scaffolds having
higher affinity for the AKR1C3 enzyme than for AKR1C2.
The free carboxylic acid, which exists predominantly as the
negatively charged carboxylate at physiological pH, is
thought to contribute to the general inhibition of AKR1C
The selective targeting of AKR1C3 over 1C2 is considered
critical for the development of effective prostate cancer
2
therapy. Not only do the two isoforms share over 86%
sequence identity, but AKR1C2 is also involved in DHT
inactivation
Accordingly, the inhibitory potencies of compounds 1 and
were determined for both AKR1C2 and AKR1C3, and were
making
its
inhibition
undesirable.
24
enzymes by NSAIDs. Accordingly, the replacement of the
carboxylic acid function of INDO with the more highly
acidic hydroxyfurazan led to new analogue 1 having higher
AKR1C3 activity than the less acidic 3-hydroxyl-N(1)-
methyl-1,2,5-triazole derivative 2.
2
compared with INDO, which was used as a control. In
order to investigate their selectivity, the IC50 ratio of
AKR1C2 and AKR1C3 inhibition was used as an indicator
(
Table 1).
The compounds were also evaluated against COX1 and
COX2 enzymes to prove that they are not active as NSAIDs
and to ensure no COX related off-target effects occurred.
INDO, compounds 1 and 2 were assayed for their inhibitor
effect against COX1 and COX2 enzymes using ovine COX1
(oCOX1) and human COX2 (hCOX2). The activity of the
tested compounds is shown in Table 1. INDO, a non-
Table 1. Inhibitory effect of compounds 1, 2 and INDO
against AKR1C3, AKR1C2, COX1 and COX2
recombinant purified enzymes.
AKR
AKR1C3 AKR1C2
Cpd IC50±SE IC50±SE
oCOX1
hCOX2
1
C2/
AKR
C3a
COX1/
AKR1C
3a
IC50±SE IC50±SE
(µM)
27
selective, non-covalent slowly reversible inhibitor,
significantly inhibited COX1 and COX2 with IC50 values of
.10 and 0.61 M, respectively. These values are consistent
(µM)
(µM)
(µM)
1
0
INDO 7.35±0,56 8.97±0,22 1.2 0.10±0.01 0.61±0.09 0.01
with the reported inhibitory potencies for INDO against
>100
0)b
>100
(12±4)b
24
COX. Notably, neither of the two new hydroxyazole
1
0.30±0.02 26.75±2.89 89
>333
>106
(
indomethacin analogues displayed significant inhibitory
activity against either of the two COX isoforms at the
highest concentration evaluated (100 µM) (Table 1). They
showed high selectivity for AKR1C3 versus COX1/2. These
data support the absence of bioisosterism between the
carboxylic acid function and the 3-hydroxyfurazan and 3-
0
.94±0.05
7
>100
>100
2
14.25±3.39 15
b
b
(37±1)
(26±2)
a)
Ratio of IC50 value on the indicated enzymes b) % of
inhibition ± SE at 100 µM.
1
hydroxyl-N -methyl-1,2,5-triazole moieties in the INDO
The inhibitory activity assay was performed using
recombinant purified enzymes with the oxidation of S-
scaffold for COX1/2 activity and demonstrate that these
scaffolds display higher affinity for the AKR1C3 enzyme
than for COX1/2. These data are in accordance with the
known SAR of INDO, which highlights the importance of
the distance between the carboxylate and the indole ring
+
tetralol in the presence of NADP , and compared to INDO.
In our assays, INDO was shown to inhibit AKR1C3 with an
IC50 value of 7.35 M and to be weakly selective against
AKR1C2 (ratio AKR1C2/AKR1C3 IC50 value 1.2). In order to
evaluate the bioisosterism between its carboxylic acid
function and new hydroxyazole moieties, the carboxylic
acid function of INDO was replaced with the
hydroxyfurazan and hydroxy-N1-methyl-1,2,5-triazole
moieties, which modified both electronic properties and
lipophilicity. Hydroxyfurazan is the more acidic
heterocycle of the hydroxyazoles studied, as its
24
in obtaining COX activity. In compounds 1 and 2, the
distance between the ionizable hydroxyazoles and indole
is longer than the distance between the carboxylate and
indole in INDO.
Its higher activity on the isolated enzyme led to
compound 1 being selected for co-crystallization to
rationalise the interactions of the new bioisostere INDO
analogues within the AKR1C3 binding site. The crystal
structure of AKR1C3 in complex with compound 1 (PDB ID:
25
experimental pKa value was assessed to be around 3. It is
26
therefore more acidic than INDO itself (pKa 4.42), while
6GXK) was determined at a resolution of 1.7 Å. The AKR1C3
hydroxy-N1-methyl-1,2,5-triazole is much less acidic, with
1
6, 23
active site can be divided into five sub-sites: SP1, SP2, SP3,
an experimental pKa value of around 6.
The new
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
the oxyanion site (OS) and a steroid channel (SC). SP1 is
Page 4 of 8
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
lined by the amino acid residues Ser118, Asn167, Phe306,
Phe311 and Tyr319. In all available structures co-crystallized
28
with small molecules, SP1 is occupied with ligands bound.
SP2 is formed by residues Trp86, Leu122 and Ser129. SP3 is
constituted of Tyr24, Glu192, Ser221 and Tyr305, and INDO
is one of the few small molecules that interacts with this
sub-pocket. The OS site is responsible for aldehyde or
+
ketone reduction and is formed by the cofactor NADP and
the catalytic residues Tyr55 and His117. The SC-site is an
open channel that is gated by residues Tyr24, Leu54, Ser129
and Trp227. The omit electron density map of 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
28
complex in the active site corresponds to compound 1
(
Figure S2), and displays a unique binding conformation.
It is well known that INDO presents two different binding
modes; INDO carboxylate group establishes a hydrogen
bond with the phosphate group of the cofactor at
crystallization conditions at pH 6.0, while it is instead
2
8
24
bound in the OS sub-site at pH 7.5. Liedtke et al. have
suggested that the binding mode of INDO is linked to the
+
protonation state of the NADP phosphate group, which
shifts ionization state according to pH conditions. The
binding mode of compound 1, obtained from crystals
grown at pH 6.0, clearly indicates that it interacts with the
OS-site using the hydrofurazan portion (Figure 2); the
interactions with the protonated phosphate groups, which
are present in the INDO complex crystallised at the same
pH, are missing (PDB ID: 1S2A). The different steric
encumbrances of the hydroxyfurazan and carboxyl groups,
which hinder 1 from establishing hydrogen-bond with the
cofactor inside SP3, are a probable explanation. The
overlay of compound 1 and INDO co-crystallized with
AKR1C3 (Figure 2) shows two similar binding modes. The
key binding interaction of 1 is the hydrogen bonding
between the OS residues Tyr 55 and His117 and oxygen
atom of hydroxyfurazan motif, which overlays and mimic
carboxylic oxygen of INDO. Interestingly, in many NSAIDs
complexes with AKR1C3, an additional hydrogen bond
interaction with a water molecule adjacent to carboxylic
Figure 2. AKR1C3 co-crystallized with compound 1 (orange
sticks) (PDB ID: 6GXK), superimposed with the AKR1C3
+
structure in complex with INDO (green sticks). NADP is
represented in purple sticks. Molecular graphics and analyses
29
were performed using the UCSF Chimera package.
A molecular docking study was performed on COX-2 and
AKR1C2 in order to provide further insight into compound
28
acid is present, actually that water is missing in the
binding site of compound 1-AKR1C3 complex. The p-
chlorobenzoyl ring is observed roughly perpendicular
respect the plane of indole ring and occupies the SP1,
making few interactions with the protein. The indole ring
extends toward SP3 sub-site and is involved in a hydrogen
bond between the methoxy oxygen and a water molecule
present in SP3.
1’s selectivity toward AKR1C3. The hydroxyfurazan motif is
unable to overlay the INDO carboxylic group in the
predicted conformation of compound 1 versus COX-2 and
is not able to interact with its Arg89 and Tyr324 via
hydrogen bond (Figure S3). Moreover, the (bio)isteric
modulation of INDO changes the link length between the
anionic portion and the indole ring; this difference makes
compound
1 unable to satisfy the pharmacophore
3
0
requirements, for activity toward COX, making it
selective against AKR1C3. Moving now to C2 isoform, it is
well known that AKR1C2 is characterized by a smaller SP1
2
than the C3 isoform. This difference allows the design of
compounds selective for the latter isoform. Docking
analyses of compound 1 with AKR1C2 (Figure S4) indicate
that the molecule, which places the fluorophenyl motif
toward SC and the hydroxyfurazan into SP1, does not
maintain the key interaction with the OS. This may explain
its reduced activity towards the C2 isoform.
ACS Paragon Plus Environment

Page 5 of 8
ACS Medicinal Chemistry Letters
In the last phase of our study, we evaluated the
not ionized form (pKa value = 6.10, see Table 3, %
ionization ca 3%). Consequently compound 2 is much
more lipophilic than 1 and INDO at physiological pH (see
LogD in Table 3) therefore it may cross membranes by
passive diffusion more easily than 1 and INDO. Further
membrane permeability assays would be necessary to
confirm this hypothesis.
1
2
3
4
5
6
7
8
9
antiproliferative effects of the compounds in two different
panels of human cancer cells; colorectal and prostate
cancer cells. Two cell lines that expressed (LNCaP and
HCT116 cells) or overexpressed (VCaP and HT-29 cells)
AKR1C3 were selected for each tissue. AKR1C3 protein
expression was confirmed by western blot results prior to
7.4
chemosensitivity
screening
(Figure
S5).
The
In view of testing in the near future the efficacy of these
compounds in vivo, i.e. in a rodent model of prostate
cancer by oral administration, we determined the
solubility and the stability of compounds 1, 2 and INDO in
Simulated Gastric Fluid (SGF) and Simulated Intestinal
Fluid (SIF) (Table 3 and Figure S1). As expected, the
solubility of the tested compounds is higher at pH 7 respect
antiproliferative activity of the compounds was then
investigated by assessing cell viability and dosing ATP
content in a luminescent cell viability assay. Results are
reported in Figure 3. The IC50 values of the new compounds
were determined and are reported in Table 2. INDO
inhibited cell proliferation with IC50 values that ranged
from about 313 μM to 137 μM, showing the lowest activity
against HT-29 colon cancer cells. Both hydroxyazole
indomethacin analogues 1 and 2 showed similar or greater
activity than INDO in all the tested cell lines. Compound 2
showed higher potency than compound 1 against all the
cell lines, except in HCT 116 (data were not statistically
significant). The best effect was observed in both AKR1C3
overexpressing prostate cancer VCaP and colorectal cancer
HT-29 cells. These results also confirmed the
bioisosterism, in these cellular models, between the
carboxylic acid function and the hydroxyfurazan and
hydroxy-N1-methyl-1,2,5-triazole moieties. In LNCaP,
VCaP and HT-29 cells, the behaviour of the azole
indomethacin analogues was opposite to the trend
observed in the isolated enzyme experiments.
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
7.4
to pH 2, according to their measured log D . Serum
stability and Plasma Protein Binding (PPB) were also
evaluated (Figure S1 and Table S2). These data on the
whole prove that behavior of 1 and 2 is similar to the INDO
one, being stable in gastrointestinal fluids and in serum
with a very high PPB (> 99%).
Table 3. Ionization, lipophilicity and solubility of
compounds 1, 2 and INDO.
Solubility
(
mg/l)
clogP
Cpd
pKaa
log D7.4
b
pH 2
SGF
pH 7
SIF
1.01
±0.03
INDO 4.41 ±0.04c
4.18
7.19
1819.5
Figure 3. The effect of 1, 2 and INDO on selected cell lines.
Compounds were utilized at 10-200 μM. Cell viability was
estimated by determining ATP content in three replicate
wells. Results are normalized to the growth of cells that were
treated with dimethyl sulfoxide (DMSO) and are represented
as meanSEM of at least three independent experiments.
1
.80
1
4.09 ±0.03
6.10 ±0.01
5.21
0.48
0.12
276.3
2.40
±
0.03
2
4.54
3.1 ±0.2
a)
potentiometric titration; b) clogP calculated using Bio-
Loom for Windows, vers.1.5.
Table 2. Inhibitory effect of compounds 1, 2 and INDO
on cell proliferation.
In summary, we have presented a new generation of
AKR1C3 inhibitors that have been designed using a
bioisosteric approach and by replacing the carboxylic acid
function of INDO with hydroxylated azoles. The two newly
reported indomethacin analogues, compounds 1 and 2,
were found to selectively inhibit AKR1C3 without any
significant AKR1C2 and COX1/2 off-target effects and with
greater potency than INDO. The crystal structure of
AKR1C3 in complex with compound 1 allowed to rationalise
the interactions of the new inhibitors within the AKR1C3
binding site. Compounds 1 and 2 were also able to inhibit
the cell proliferation of AKR1C3-expressing prostate and
colorectal cancer cells in a dose-dependent manner. They
demonstrated higher potency than INDO against most cell
lines. Compound 2 showed higher potency than compound
1 especially in AKR1C3 overexpressing cancer cell lines. We
are aware that the activities of 1, 2, and INDO itself are
found only at relatively high concentration (IC50 ≥ 70 μM).
Anyway, this is the first report of indomethacin analogues,
optimised for AKR1C3 inhibition, ever tested on cellular
models of prostate and colorectal cancers. The novel
chemical scaffolds reported herein provide a promising
starting point for the design of more potent and selective
LNCaP
Cpd IC50±SEM
(µM)
VCaP
IC50±SEM IC50±SEM IC50±SEM
(µM) (µM) (µM)
136.9±1.03 136.8±1.04 312.6±1.07
HCT 116
HT-29
INDO 238.4±1.11
1
141.2±1.04**
138.1±1.17
78.86±1.20 134.7±1.24*
2
116.6±1.05***69.11±1.04****,§ 110.7±1.21 81.96±1.20**
*p<0.05, **p<0.01, ***p<0.001, **** p<0.0001 Compound 1 or 2
§
vs INDO; p<0.0001 Compound 2 vs 1.
A cell-based system is a more complex environment than
a purified enzyme and multiple parallel outcomes that are
connected to other pathways are possible. Moreover, the
different lipophilicity and ionization grades of 1 and 2 may
have an important role to play in their cellular
internalization and consequently in their cellular activity.
While compound 1 and INDO are largely ionized at
physiological pH (pKa value respectively 4.09 and 4.41, see
Table 3, % ionization > 99%), compound 2 is prevalently in
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
Page 6 of 8
AKR1C3 inhibitors, and for the study of AKR1C3’s role in
cancer development and other disorders.
(5) Seisen, T.; Roupret, M.; Gomez, F.; Malouf, G. G.; Shariat, S.
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
F.; Peyronnet, B.; Spano, J. P.; Cancel-Tassin, G.; Cussenot, O. A
comprehensive review of genomic landscape, biomarkers and
treatment sequencing in castration-resistant prostate cancer.
Cancer Treat Rev 2016, 48, 25-33.
(6) Khanim, F. L.; Hayden, R. E.; Birtwistle, J.; Lodi, A.; Tiziani,
S.; Davies, N. J.; Ride, J. P.; Viant, M. R.; Gunther, U. L.;
Mountford, J. C.; Schrewe, H.; Green, R. M.; Murray, J. A.;
Drayson, M. T.; Bunce, C. M. Combined bezafibrate and
medroxyprogesterone acetate: potential novel therapy for acute
myeloid leukaemia. PLoS One 2009, 4, e8147.
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI : xxxxxxxxxxxxxxxxxxxx.
Experimental procedures for the biological experiments,
solubility and stability measures, synthetic procedures
and characterization data for the compounds, X-ray
diffraction details, docking pose of compound 1 in COX2
and AKR1C2 binding site (PDF)
(
7) Sun, S. Q.; Gu, X.; Gao, X. S.; Li, Y.; Yu, H.; Xiong, W.; Yu,
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
H.; Wang, W.; Li, Y.; Teng, Y.; Zhou, D. Overexpression of AKR1C3
significantly enhances human prostate cancer cells resistance to
radiation. Oncotarget 2016, 7, 48050-48058.
(8) Liu, C.; Armstrong, C. M.; Lou, W.; Lombard, A.; Evans, C.
P.; Gao, A. C. Inhibition of AKR1C3 activation overcomes
resistance to abiraterone in advanced prostate cancer. Mol Cancer
Ther 2017, 16, 35-44.
AUTHOR INFORMATION
Corresponding Author
*
*
Tel: +39-0116707195. E-mail: donatella.boschi@unito.it
Tel: +39-0116706864. E-mail: simona.oliaro@unito.it.
(
9) Liu, C.; Lou, W.; Zhu, Y.; Yang, J. C.; Nadiminty, N.;
Gaikwad, N. W.; Evans, C. P.; Gao, A. C. Intracrine androgens and
AKR1C3 activation confer resistance to enzalutamide in prostate
cancer. Cancer Res 2015, 75, 1413-1422.
(10) Zhong, T.; Xu, F.; Xu, J.; Liu, L.; Chen, Y. Aldo-keto
reductase 1C3 (AKR1C3) is associated with the doxorubicin
resistance in human breast cancer via PTEN loss. Biomed
Pharmacother 2015, 69, 317-325.
Author Contributions
The manuscript was written through contributions of all
authors.
Funding Sources
This research was financially supported by the University of
Turin (Ricerca Locale grants 2015-2018) and Fondazione Cassa
di Risparmio di Torino (grant BOSD_CRT_17_2). PG is
supported by the Wellcome Trust/DBT Alliance Fellowship
(IA/E/16/1/502999).
(
11) Matsunaga, T.; Hojo, A.; Yamane, Y.; Endo, S.; El-Kabbani,
O.; Hara, A. Pathophysiological roles of aldo-keto reductases
AKR1C1 and AKR1C3) in development of cisplatin resistance in
(
human colon cancers. Chem Biol Interact 2013, 202, 234-242.
(12) Penning, T. M. Aldo-Keto Reductase (AKR) 1C3 inhibitors:
a patent review. Expert Opin Ther Pat 2017, 27, 1329-1340.
Notes
The authors declare no competing financial interest.
(
13) Choi, J. S.; Jin, M. J.; Han, H. K. Role of monocarboxylic acid
transporters in the cellular uptake of NSAIDs. J Pharm Pharmacol
005, 57, 1185-1189.
14) Pippione, A. C.; Giraudo, A.; Bonanni, D.; Carnovale, I. M.;
ACKNOWLEDGMENT
2
The authors wish to thank Dr Dale Lawson for proofreading
the final manuscript and Dr Antonio Foti for the help in
synthesis of compound 1 during his thesis training period.
(
Marini, E.; Cena, C.; Costale, A.; Zonari, D.; Pors, K.; Sadiq, M.;
Boschi, D.; Oliaro-Bosso, S.; Lolli, M. L. Hydroxytriazole
derivatives as potent and selective aldo-keto reductase 1C3
ABBREVIATIONS
(
AKR1C3) inhibitors discovered by bioisosteric scaffold hopping
Aldo-keto reductase 1C3 isoform (AKR1C3), castration-
resistant prostate cancer (CRPC), flufenamic acid (FLU),
indomethacin (INDO), cyclooxygenase (COX), aldo-keto
reductase 1C2 isoform (AKR1C2), p-toluenesulfonic acid
approach. Eur J Med Chem 2017, 139, 936-946.
(15) Pippione, A. C.; Carnovale, I. M.; Bonanni, D.; Sini, M.;
Goyal, P.; Marini, E.; Pors, K.; Adinolfi, S.; Zonari, D.; Festuccia,
C.; Wahlgren, W. Y.; Friemann, R.; Bagnati, R.; Boschi, D.; Oliaro-
Bosso, S.; Lolli, M. L. Potent and selective aldo-keto reductase 1C3
(AKR1C3) inhibitors based on the benzoisoxazole moiety:
application of a bioisosteric scaffold hopping approach to
flufenamic acid. Eur J Med Chem 2018, 150, 930-945.
(
PTSA), non-steroidal anti-inflammatory drug (NSAID).
REFERENCES
(1) Pippione, A. C.; Boschi, D.; Pors, K.; Oliaro-Bosso, S.; Lolli,
M. L. Androgen-AR axis in primary and metastatic prostate
cancer: chasing steroidogenic enzymes for therapeutic
intervention. J Cancer Metastasis Treat 2017, 3, 328-361.
(
16) Pippione, A. C.; Dosio, F.; Ducime, A.; Federico, A.;
Martina, K.; Sainas, S.; Frolund, B.; Gooyit, M.; Janda, K. D.;
Boschi, D.; Lolli, M. L. Substituted 4-hydroxy-1,2,3-triazoles:
synthesis, characterization and first drug design applications
through bioisosteric modulation and scaffold hopping
approaches. Medchemcomm 2015, 6, 1285-1292.
(
2) Byrns, M. C.; Jin, Y.; Penning, T. M. Inhibitors of type 5
1
7beta-hydroxysteroid dehydrogenase (AKR1C3): overview and
structural insights. J Steroid Biochem Mol Biol 2011, 125, 95-104.
(3) Zhou, W.; Limonta, P. AKR1C3 inhibition therapy in
castration-resistant prostate cancer and breast cancer: lessons
from responses to SN33638. Front Oncol 2014, 4, 162.
(
17) Sainas, S.; Pippione, A. C.; Giorgis, M.; Lupino, E.; Goyal, P.;
Ramondetti, C.; Buccinna, B.; Piccinini, M.; Braga, R. C.; Andrade,
C. H.; Andersson, M.; Moritzer, A. C.; Friemann, R.; Mensa, S.; Al-
Kadaraghi, S.; Boschi, D.; Lolli, M. L. Design, synthesis, biological
evaluation and X-ray structural studies of potent human
dihydroorotate dehydrogenase inhibitors based on hydroxylated
azole scaffolds. Eur J Med Chem 2017, 129, 287-302.
(
4) Loriot, Y.; Fizazi, K.; Jones, R. J.; Van den Brande, J.; Molife,
R. L.; Omlin, A.; James, N. D.; Baskin-Bey, E.; Heeringa, M.; Baron,
B.; Holtkamp, G. M.; Ouatas, T.; De Bono, J. S. Safety, tolerability
and anti-tumour activity of the androgen biosynthesis inhibitor
ASP9521 in patients with metastatic castration-resistant prostate
cancer: multi-centre phase I/II study. Invest New Drugs 2014, 32,
(
18) Sainas, S.; Pippione, A. C.; Giraudo, A.; Martina, K.; Bosca,
F.; Rolando, B.; Barge, A.; Ducime, A.; Federico, A.; Grossert, S. J.;
White, R. L.; Boschi, D.; Lolli, M. L. Regioselective N‐Alkylation of
Ethyl 4‐Benzyloxy‐1,2,3‐triazolecarboxylate: A Useful Tool for the
995-1004.
ACS Paragon Plus Environment

Page 7 of 8
ACS Medicinal Chemistry Letters
Synthesis of Carboxylic Acid Bioisosteres. Journal of Heterocyclic
Chemistry 2018, DOI: 10.1002/jhet.3426.
19) Lolli, M.; Narramore, S.; Fishwick, C. W.; Pors, K. Refining
the chemical toolbox to be fit for educational and practical
purpose for drug discovery in the 21st Century. Drug Discov Today
Dehydrogenase/Prostaglandin F Synthase) in Castrate-Resistant
Prostate Cancer. J Med Chem 2013, 56, 2429-2446.
1
2
3
4
5
6
7
8
9
(
(25) Lolli, M. L.; Hansen, S. L.; Rolando, B.; Nielsen, B.;
Wellendorph, P.; Madsen, K.; Larsen, O. M.; Kristiansen, U.;
Fruttero, R.; Gasco, A.; Johansen, T. N. Hydroxy-1,2,5-oxadiazolyl
Moiety as Bioisoster of the Carboxy Function. Synthesis,
Ionization Constants, and Pharmacological Characterization of γ-
Aminobutyric Acid (GABA) Related Compounds. J Med Chem
2006, 49, 4442-4446.
(26) Bouchard, G.; Carrupt, P. A.; Testa, B.; Gobry, V.; Girault,
H. H. Lipophilicity and solvation of anionic drugs. Chemistry
2002, 8, 3478-3484.
2
015, 20, 1018-1026.
20) Liedtke, A. J.; Kim, K.; Stec, D. F.; Sulikowski, G. A.;
(
Marnett, L. J. Straightforward protocol for the efficient synthesis
of varied N(1)-acylated (aza)indole 2-/3-alkanoic acids and esters:
optimization and scale-up. Tetrahedron 2012, 68, 10049-10058.
(
21) Bonomo, S.; Tosco, P.; Giorgis, M.; Lolli, M.; Fruttero, R.
The role of fluorine in stabilizing the bioactive conformation of
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
dihydroorotate dehydrogenase inhibitors. J Mol Model 2013, 19,
(27) Cena, C.; Tosco, P.; Marini, E.; Lazzarato, L.; Piccinini, M.;
Ramondetti, C.; Lupino, E.; Fruttero, R.; Gasco, A. Nitrooxyacyl
derivatives of salicylic acid: aspirin-like molecules that covalently
inactivate cyclooxygenase-1. ChemMedChem 2011, 6, 523-530.
(28) Flanagan, J. U.; Yosaatmadja, Y.; Teague, R. M.; Chai, M. Z.
L.; Turnbull, A. P.; Squire, C. J. Crystal structures of three classes
of non-steroidal anti-inflammatory drugs in complex with aldo-
keto reductase 1C3. PLoS One 2012, 7, e43965.
(29) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.;
Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. UCSF Chimera--a
visualization system for exploratory research and analysis. J
Comput Chem 2004, 25, 1605-1612.
(30) Kurumbail, R. G.; Stevens, A. M.; Gierse, J. K.; McDonald,
J. J.; Stegeman, R. A.; Pak, J. Y.; Gildehaus, D.; Miyashiro, J. M.;
Penning, T. D.; Seibert, K.; Isakson, P. C.; Stallings, W. C.
Structural basis for selective inhibition of cyclooxygenase-2 by
anti-inflammatory agents. Nature 1996, 384, 644-648.
1
099-1107.
(
22) Lolli, M. L.; Giordano, C.; Pickering, D. S.; Rolando, B.;
Hansen, K. B.; Foti, A.; Contreras-Sanz, A.; Amir, A.; Fruttero, R.;
Gasco, A.; Nielsen, B.; Johansen, T. N. 4-Hydroxy-1,2,5-oxadiazol-
3
-yl Moiety as Bioisoster of the Carboxy Function. Synthesis,
Ionization Constants, and Molecular Pharmacological
Characterization at Ionotropic Glutamate Receptors of
Compounds Related to Glutamate and Its Homologues. J Med
Chem 2010, 53, 4110-4118.
(
23) Giraudo, A.; Krall, J.; Nielsen, B.; Sorensen, T. E.; Kongstad,
K. T.; Rolando, B.; Boschi, D.; Frolund, B.; Lolli, M. L. 4-Hydroxy-
,2,3-triazole moiety as bioisostere of the carboxylic acid function:
1
a novel scaffold to probe the orthosteric gamma-aminobutyric
acid receptor binding site. Eur J Med Chem 2018, 158, 311-321.
(
24) Liedtke, A. J.; Adeniji, A. O.; Chen, M.; Byrns, M. C.; Jin, Y.;
Christianson, D. W.; Marnett, L. J.; Penning, T. M. Development
of Potent and Selective Indomethacin Analogues for the
Inhibition
of
AKR1C3
(Type
5
17β-Hydroxysteroid
ACS Paragon Plus Environment

ACS Medicinal Chemistry Letters
Page 8 of 8
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
8
8x35mm (600 x 600 DPI)
ACS Paragon Plus Environment