3-(3,4-dihydroisoquinolin-2(1 H)-ylsulfonyl)benzoic acids: Highly potent and selective inhibitors of the type 5 17-β-hydroxysteroid dehydrogenase AKR1C3

Article abstract of DOI:10.1021/jm3007867

A high-throughput screen identified 3-(3,4-dihydroisoquinolin-2(1H)- ylsulfonyl)benzoic acid as a novel, highly potent (low nM), and isoform-selective (1500-fold) inhibitor of aldo-keto reductase AKR1C3: a target of interest in both breast and prostate cancer. Crystal structure studies showed that the carboxylate group occupies the oxyanion hole in the enzyme, while the sulfonamide provides the correct twist to allow the dihydroisoquinoline to bind in an adjacent hydrophobic pocket. SAR studies around this lead showed that the positioning of the carboxylate was critical, although it could be substituted by acid isosteres and amides. Small substituents on the dihydroisoquinoline gave improvements in potency. A set of reverse sulfonamides showed a 12-fold preference for the R stereoisomer. The compounds showed good cellular potency, as measured by inhibition of AKR1C3 metabolism of a known dinitrobenzamide substrate, with a broad rank order between enzymic and cellular activity, but amide analogues were more effective than predicted by the cellular assay.

Full text of DOI:10.1021/jm3007867

Article
pubs.acs.org/jmc
3‑(3,4-Dihydroisoquinolin-2(1H)‑ylsulfonyl)benzoic Acids: Highly
Potent and Selective Inhibitors of the Type 5 17-β-Hydroxysteroid
Dehydrogenase AKR1C3
Stephen M. F. Jamieson,† Darby G. Brooke,† Daniel Heinrich,† Graham J. Atwell,† Shevan Silva,†
Emma J. Hamilton,† Andrew P. Turnbull,§ Laurent J. M. Rigoreau,‡ Elisabeth Trivier,‡ Christelle Soudy,‡
Sharon S. Samlal,‡ Paul J. Owen,‡ Ewald Schroeder,‡ Tony Raynham,‡ Jack U. Flanagan,†
and William A. Denny*,†
†Auckland Cancer Society Research Centre, The University of Auckland, Private Bag 92019, Auckland 1142, New Zealand
§Cancer Research Technology Ltd, Birkbeck College, University of London, London, U.K.
‡Cancer Research Technology Ltd, Wolfson Institute for Biomedical Research, The Cruciform Building, Gower Street, London
WC1E 6BT, U.K.
S
* Supporting Information
ABSTRACT: A high-throughput screen identified 3-(3,4-dihydroisoquinolin-
2(1H)-ylsulfonyl)benzoic acid as a novel, highly potent (low nM), and
isoform-selective (1500-fold) inhibitor of aldo-keto reductase AKR1C3: a
target of interest in both breast and prostate cancer. Crystal structure studies
showed that the carboxylate group occupies the oxyanion hole in the enzyme,
while the sulfonamide provides the correct twist to allow the dihydroisoquinoline to bind in an adjacent hydrophobic pocket.
SAR studies around this lead showed that the positioning of the carboxylate was critical, although it could be substituted by acid
isosteres and amides. Small substituents on the dihydroisoquinoline gave improvements in potency. A set of “reverse
sulfonamides” showed a 12-fold preference for the R stereoisomer. The compounds showed good cellular potency, as measured
by inhibition of AKR1C3 metabolism of a known dinitrobenzamide substrate, with a broad rank order between enzymic and
cellular activity, but amide analogues were more effective than predicted by the cellular assay.
dihydrotestosterone reductase activity, but with 100-fold less
catalytic efficiency than the reduction of PGD2.8
he type 5 17-β-hydroxysteroid dehydrogenase is a
Tmember of the aldo-keto reductase (AKR) superfamily
of enzymes,1 where it is also known as AKR1C3. It is expressed
in the human prostate and mammary gland, where it is
responsible for reducing androst-4-ene-3,17-dione, estrone, and
progesterone to, respectively, testosterone, 17β-estradiol, and
20α-hydroxprogesterone.2 This production of growth-stimula-
tory steroid hormones by AKR1C3 makes it a target of interest
in both breast and prostate cancer. AKR1C3 is also known as
prostaglandin F2 synthase,3 as it transforms PGH2 to PGF2α
and PGD2 to 9α,11β-PGF2α, thus diverting PGD2 (1) from the
PGJ2 pathway that governs cellular differentiation,4 making it
also a potential drug target for some leukemias. AKR1C3 has
also been reported to metabolize some drugs; notably,
inactivating the topoisomerase poison doxorubicin5 and
catalyzing the aerobic activation of the hypoxia prodrug PR-
104.6
There are three closely related isoforms of AKR1C3
(AKR1C1, AKR1C2, and AKR1C4), which are also involved
in steroid hormone metabolism. These enzymes, in particular
AKR1C1 and AKR1C2, are not ideal drug targets, since they
can prevent androgen signaling and cell proliferation by
catalyzing the reduction of the potent androgen 5α-
dihydrotestosterone to 5α-androstane-3β,17β-diol and 5α-
androstane-3α,17β-diol, respectively.7 AKR1C3 has weak 5α-
Among the classes of compounds studied9,10 as potential
competitive inhibitors of the AKR1C family of enzymes are
structural mimics of prostaglandin D2. In this class are
compounds such as the plant stress hormone jasmonic acid
(2), which shows Ki values for the inhibition of AKR1C1-4 of
106, 18, 162, and 37 μM, respectively.11 Simpler analogues such
as the benzylidenecyclopentanone 3 also showed modest
inhibitory activity (IC50 values approximately 35 μM for
AKR1C-1 and -3).12
Based on kinetic data, it was suggested that the reductive
reactions promoted by AKR1C3 proceed through an oxyanion
transition state and that steroid carboxylates might be suitable
competitive inhibitors.2 A later study13 showed this to be the
case, with (for example) the cholanic acid analogue (4)
showing IC50 values of 14.0, 0.04, and 1.02 μM, respectively, for
inhibition of AKR1C1, -2, and -3, with high selectivity with
respect to inhibition of COX1 and COX2 (IC50 values > 250
μM). The crystal structure of AKR1C3 complexed with the
flavanoid rutin (5)14 (PBD 1RY8) showed H-bond formation
with the conserved Y55 and H117 residues that constitute the
Received: June 6, 2012
Published: August 9, 2012
© 2012 American Chemical Society
7746
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
Table 1. Selected Known Inhibitors of AKR1C3
b
AKR isoform IC50 (μM)
1C2
a
no.
compd
1C1
1C3
1C4
11
12
13
14
15
16
flufenamic acid
indomethacin
naproxen
2.64 0.68
3.14 0.50
0.41 0.15
0.73 0.16
1.1 0.47
>100
>100
>100
>100
>100
>100
>100
>100
3.16 1.18
>100
31.3 16.3
8.74 4.29
42.4 21.0
32.5 19.6
meclofenamic acid
S(+)-ibuprofen
0.54 0.14
32.7 12.7
1.56 0.83
>100
>100
flurbiprofen
a
b
See Figure 1 for structures. IC50 values were determined using a competitive fluorescence assay; see Experimental Section.
“oxyanion hole”. The related compound 2′-hydroxyflavanone
(6) showed good in vitro potency for AKR1C3 (IC50 0.10 μM,
with 62-fold selectivity over AKR1C2).15 Cinnamic acids,
which are flavonoid precursors, also show some inhibitory
activity; the bis(cinnamic acid) baccharin (7) has an IC50 of
0.11 μM and high selectivity for AKR1C3.16
A series of substituted N-phenylanthranilic acids also showed
good AKR1C3 potencies, with high selectivity over COX
enzymes.13 For example, 8 had IC50 values of 4.0, 0.96, 0.39, 33,
and 225 μM, respectively, for the inhibition of AKR1C-1, -2,
and -3 and COX1 and COX2. Studies13,17 on a series of
substituted 3-(phenylamino)benzoic acids showed that
AKR1C3 inhibitory activity correlated positively with the
electron-withdrawing capability of 4-substituents (which affects
the carboxylate pKa), e.g, compounds 9 and 10.
The widespread observation that nonsteroidal anti-inflam-
matory drugs (NSAIDs) appear to protect against a variety of
cancers, including prostate carcinoma, gastrointestinal tumors,
and leukemia,18 prompted the suggestion19 that inhibition of
AKR1C3 might be a contributing mechanism of action. Crystal
structures of the NSAIDs flufenamic acid (11) and
indomethacin (12) with AKR1C3 show that the carboxylate
of the former occupies the oxyanion hole, but in the latter, the
ketone oxygen is in close proximity to this binding site.19
Indomethacin was an effective inhibitor of AKR1C3 (IC50 2.3
μM), with a 20−40-fold selectivity over AKR1C1 and -C2.
Several other known NSAIDS, including naproxen, (13),
meclofenamic acid (14), S(+)-ibuprofen (15), and flurbiprofen
(16) are also known20 to have AKR1C isoform activity (Table
1).
In a project seeking novel, more potent, and highly selective
inhibitors of AKR1C3, we conducted a high-throughput screen.
AKR1C3 was cloned in an E. coli expression vector with a N-
terminal His tag for assay development. The assay used a
fluorescent readout (probe 5; see Supporting Information for
structure). A screen of 99,000 compounds from an in-house
library gave 8000 primary hits, which were trimmed to 960 by
removal of kinase library compounds and by inspection.
Generation of 5-point IC50 values gave 187 compounds with
estimated activities of <1 μM, from which 3-(3,4-dihydroiso-
quinolin-2(1H)-ylsulfonyl)benzoic acid (17) (Table 2) was
selected as a lead. In this paper we discuss the development of
this lead, including the development of structure−activity
relationships (SAR) and a crystal structure-derived model of
the binding of the class to AKR1C3.
Figure 1.
literature method.21 Compound 17 of Table 4 was previously
reported as an intermediate in the synthesis of a series of
PPARδ partial agonists.22
Energy minimization was performed with SZYBKI using the
MM-PBSA implicit solvent model. For the free ligand,
minimization was performed with a dielectric constant of 80
and a microcavitation term of 0.25. RMSD values are reported
in angstroms, and RMSD[protein] is the RMSD after ligand
minimization within the AKR1C3 active site. E[Bound] is the
ligand energy in the conformation minimized in the active site,
while E[Free] is the energy after minimization of that
conformation in the absence of protein. ΔE[Bound−Free] is the
difference between the two energies. Energy is reported as
kilocalories per mole. RMSD[Bound−Free] is the difference in
angstroms between the two conformations. LPE is the Ligand−
Protein energy reported by SZYBKI.
The majority of the compounds of Table 3 were prepared by
coupling of sulfonyl chloride 96 with commercially available or
previously reported substituted 1,2,3,4-tetrahydroisoquinolines
(96a−x) (Scheme 1A). Reaction of the 7-nitrile 51 with NaN3
gave the tetrazole 45. Suzuki coupling of iodide 50 with
CHEMISTRY
■
Compounds 17−24 of Table 2 were prepared by direct
coupling of sulfonyl chloride 95 and the required amines
(Scheme 1A). Compound 24 of Table 2 was prepared by a
7747
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
a
Table 2. Variation of the Dihydroisoquinoline Unit of 17
Scheme 1. Compounds of Tables 2 and 3
a
IC50 values were determined using a competitive fluorescence assay;
b
see Experimental Section. Concentration of drug to inhibit the
conversion of nitroaromatic prodrug PR-104A to the hydroxylamine in
HCT-116 cells engineered to overexpress AKR1C3, as determined by
c
LC-MS/MS; see Experimental Section. Reference 21.
a
Reagents and conditions: (i) K2CO3 or Et3N, THF/water or Me2CO,
N2 atmosphere, 0−20 °C over 3−72 h; (ii) NaN3, NH4Cl, DMF, 120
°C, 170 h; (iii) Pd(PPh3)2Cl2, piperidine, 20 °C, 12 h; (iv) Cs2CO3,
MeOH, 20 °C, 100 min; (v) Pd(PPh3)4, CuI, TMS-CCH, DIPE, 20
°C, 48 h; (vi) 2,4-dichloro-5,5-dimethylhydantoin, MeCN/AcOH/
water, 0−20 °C, 110 min; (vii) LiOH, MeOH, 20 °C, 80 min.
(trimethylsilyl)acetylene (97) gave the adduct 98, which was
deprotected with Cs2CO3 to give 43 (Scheme 1B). Similar
reaction of 50 with 1-ethynyl-4-(trifluoromethoxy)benzene
(99) gave 44 directly. Methyl 5-(chlorosulfonyl)nicotinate
(101) for the synthesis of 55 was prepared from methyl 5-
mercaptonicotinate (100) by the method of Isabel et al.,23,24
except the last oxychlorination was carried out under conditions
developed by Pu et al.25 Coupling of 99b with tetrahydroiso-
quinoline 96a and hydrolysis of the product ester then gave 55
(Scheme 1C). Compounds 34, 37, 42, and 48 of Table 3 were
prepared conveniently (albeit in moderate to low yields) from
the open-chain analogues 58, 63, 64, and 59, respectively
(Table 6), on cyclization with methanesulfonic acid in s-trioxan
(Scheme 2).
Finally, the “reverse sulfonamides” 84−94 of Table 7
(Scheme 4) were constructed from known sulfonyl chlorides
106, 107a−f and known amines 108, 109a−e.
RESULTS AND DISCUSSION
■
High-Throughput Chemical Screening. Library screen-
ing against purified recombinant human AKR1C3 identified a
number of compounds with IC50 values of <200 nM, and these
were tested for activity against the AKR isoforms, 1C1, 1C2,
and 1C4. The frequency of hits for AKR1C3 was kept low by
using a high probe 5 substrate concentration (10× Km) at a
saturating NADPH concentration. The accuracy of the
biochemical assay was confirmed by generating IC50 values
for the known inhibitors (11−16), comparable with those
previously published.26
The open-chain analogues 56−64 of Table 5 were prepared
(Scheme 2) by coupling the sulfonyl chloride 95 and
substituted phenylethanamines 102a−h under mild basic
conditions (K2CO3, 35 °C) in a two-phase system in modest
yields (10−50%, unoptimized).
The bulk of the compounds of Table 6 were similarly
prepared (Scheme 3) by coupling of known 3-substituted
benzenesulfonyl chlorides 103a−e with 1,2,3,4-tetrahydroiso-
quinoline (100a). The 3-nitro analogue 66 prepared in this way
was reduced to the amine 67, and this was further reacted with
ethyl 2-chloro-2-oxoacetate, and the resulting ester 104 was
hydrolyzed with Cs2CO3 to give the oxoacetic acid 72. The 3-
nitrile 65 was treated with NaN3 to give tetrazole 74 (Scheme
3A). The 3-amides 75−82 were prepared from the acid
chloride of 17 and appropriate amines (Scheme 3B).
Compound 83 was prepared by Pd-assisted coupling of 2-(4-
iodophenylsulfonyl)-1,2,3,4-tetrahydroisoquinoline (105) and
2-pyrrolidinone (Scheme 3C).
Of several potent and selective compounds identified, the 3-
(3,4-dihydroisoquinolin-2(1H)-ylsulfonyl)benzoic acid (17)
(Table 2) represented a new class of highly potent AKR1C3
inhibitors. In repeat assays it showed an IC50 of 0.013 0.003
μM for AKR1C3, compared with 20.3
3.8 μM against
AKR1C1 and >30 μM against AKR1C2 and AKR1C4. This
compares favorably for both potency and specificity with other
known AKR1C3 inhibitors (Table 1). Representative com-
pounds 17, 79, and 85 were screened for inhibition of COX-1
and COX-2 but were found to be inactive at 10 μM.
X-ray Crystallography. To understand how this potency is
achieved, we determined the structure of 17 bound to AKR1C3
in the presence of oxidized NADP+, and the crystallographic
7748
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
Table 3. SAR of Dihydroisoquinoline-Substituted for Inhibition of AKR Isoforms
a
IC50 (μM)
b
no.
R
1C1
1C2
1C3
1C4
HCT-116
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
9-Me
2-Me
>30
>30
>30
>30
>30
>30
>30
0.027 0.022
0.0086 0.0066
0.053 0.047
0.0089 0.0030
0.011 0.0021
0.011 0.0001
0.014 0.008
0.016 0.005
0.016 0.008
0.017 0.003
0.022 0.016
0.029 0.001
0.0087 0.0035
0.0061 0.0013
0.039 0.03
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
0.024
0.176
0.193
0.029
0.016
0.014
0.020
0.453
0.039
0.052
5-NH2
5-NO2
5-Cl
24.6 5.7
>30
9.32 3.20
>30
5-Br
>30
>30
5-I
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
5-OH
5-OMe
6-Me
6-NO2
6-CN
6-Cl
25.6 9.3
>30
>30
>30
>1
17.6 6.7
7.48 0.46
>30
13.9 3.0
>30
0.039
0.012
6-Br
6-I
6-OH
6-OMe
7-Me
7-CCH
0.027 0.005
0.038 0.016
0.013 0.001
0.042 0.034
1.31 0.46
>1
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
0.39 0.02
0.029
c
7-CCX
7-C-tetrazole
7-NO2
7-F
0.068 0.038
0.017 0.015
0.021 0.009
0.020 0.019
0.012 0.0028
0.014 0.004
0.034 0.016
0.029 0.006
0.019 0.014
0.16 0.009
9.81 0.55
>30
0.072
0.024
>30
>30
>30
>30
>30
>30
>30
>30
>30
7-Cl
>1
7-Br
0.02 0.01
0.045
7-I
7-CN
0.266
7-OMe
8-Cl
0.74 0.13
12.1 3.5
0.028
6,7-diOMe
>30
>30
>30
>30
>1
́
H (5-aza)
0.042 0.007
a,b
c
As for Table 2. X = 4-trifluoromethoxyphenyl.
a
cavity, allowing the pendant tetrahydroquinoline to occupy a
lipophilic cavity in the active site pocket defined by the side
chains of residues Met120, Asn167, Tyr216, Phe306, Ser308,
Phe111, Tyr 317, and Tyr319, and an ordered water (Figure
2A). The water molecule is located in a pocket defined by the
side chain phenol OH groups of Tyr216 and Ty319 and by the
carboxylate of Glu192. There are a number of ethylene glycol
molecules in the structure, carried over from cryoprotecting the
crystal. One is bound in the active site opposite the sulfonamide
group of 17 and forms contacts to the side chain hydroxyl
groups of Ser87 and Ser118 and the backbone carbonyl of
Met120, clearly identifying new solvent exposed target sites on
the protein.
Structures for the related compounds 79 and 85 (Figure 2B,
C) were also determined, and the binding modes were in good
agreement with that of 17; the structure data are presented in
Table S2. The carboxylate-containing 85 showed good overlap
with the benzoic acid, sulfonyl, and 6,6 cyclic systems of 17,
although the napthyl and tetrahydroquinoline groups do have
alternative orientations. This structure also showed an acetate
bound to the ethylene glycol binding site described above,
forming a hydrogen bond with the Ser82 side chain hydroxyl
Scheme 2. Compounds of Tables 3 and 5
a
Reagents and conditions: (i) Et3N or K2CO3, Me2CO, THF, 0−20
°C, 3−90 h.
data is presented in Table S2 (see Supporting Information).
The structure illustrates that the inhibitor binds in the active
site cavity with its carboxylate group occupying the oxyanion
hole, where it interacts with Tyr55 and His117, while the
benzylic unit is located in a pocket defined by residues Tyr24,
Trp227, and Phe306. The sulfonyl extends out into the central
7749
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
a
transitions through the same space as the sulfonamide group in
17, positioning, the CF3 group in close proximity to the
piperidine ring of the tetrahydroquinoline in 17. While it is
likely that the increased potency of 17 is related to the better
occupation of an active site pocket by the tetrahydroquinoline,
its complementarity with the active site may also be aided by
the rigid “butterfly” shape formed by the sulfonamide.
Moreover, the small difference in activity between 17 and 79
indicates that occupation of the oxyanion hole by a carboxylate
group has a minor but detectable role in inhibition.
Scheme 3. Compounds of Table 6
To gain some insight into the complementarity between the
rigid “butterfly” shape of the embedded sulfonamide linker and
the AKR1C3 active site, we compared energy and conforma-
tional differences for ligands minimized in and out of the
AKR1C3 active site (Table 4). Ligands were initially minimized
within the active site and for the sulfonamide compounds 17
and 79, as well as the N-phenylanthranilic acid flufenamic acid
(11); the rmsd of the respective minimized conformation was
within 0.26, 0.43, and 0.33 Å of the initial structure respectively,
while the minimized conformation of 85 showed greater
deviation at 0.67 Å. These ligand conformations were then
minimized without the active site constraint, and the lower
energy difference between the bound and free minimized
conformations implied that the sulphonamides 17 and 79 are
less strained in the active site than flufenamic acid (11). This
was also to some extent supported by the lower rmsd values,
with 17 and 79 having the lowest deviation from the minimized
active site geometry while flufenamic acid (11) had the highest.
The low differences in strain between the bound and unbound
forms indicate that 17 and 79 are likely more rigid and better
adapted to the AKR1C3 active site than flufenamic acid (11);
this is also evident in the calculated ligand−protein interaction
energy (LPE) and may in part account for the difference in
potency.
a
Reagents: (i) K2CO3 or Et3N, pyridine, dioxan or Me2CO, 20 °C, 2−
90 h; (ii) Fe, AcOH, DMF, EtOH, H2O, reflux, 1 h; (iii) NaN3,
NH4Cl, DMF, 120 °C, 69 h; (iv) ClCOCO2Et, pyridine, THF, 10−20
°C, 1 h; (v) Cs2CO3, THF, H2O; (vi) 17, (COCl)2, CH2Cl2, DMF
(cat.) or CDI, THF, then RNH2, 20 °C, 1−4 h; (vii) 2-pyrrolidinone,
Pd2dba3, Xantphos, Cs2CO3, dioxane, 100 °C, 18 h.
a
Scheme 4. Compounds of Table 7
STRUCTURE ACTIVITY RELATIONSHIPS
■
The AKR1C3 preference for the tetrahydroquinoline system of
17 was investigated with compounds 18−23 in Table 2 and
with 56−64 in Table 5. The data in Table 2 shows that altering
the size or orientation of the tetrahydroquinoline piperidine
group, while retaining AKR1C3 inhibitory activity, decreases
the potency from 2- to 46-fold compared to 17. Comparison of
the data for 17 and the tetrahydroquinoline isostere 19, when
compared to that of 20−22, clearly shows that the potency of
17 depends on the orientation of tetrahydroquinoline group.
The 46-fold reduction in IC50 of 23 compared to 17 also
indicates the importance of the ring-embedded sulfonamide
linker.
a
Reagents and conditions: (i) Et3N or K2CO3, Me2CO or THF, N2,
0−20 °C, 3−90 h.
A series of open chain analogues presented in Table 5
provided a more divergent set of sulfonamide probes that
explore the active site’s capacity to bind different structures.
These substantially more flexible compounds were weaker
inhibitors than 17. Thus, compound 56, the direct open-chain
analogue of 17, was 17-fold less potent, and the set of
compounds were on average 28-fold less potent than 17 against
AKR1C3. Overall, these data clearly support the preference of
the AKR1C3 active site for a rigid bicyclic system.
Analogue sets around the tetrahydroquinoline core of 17
were developed to probe the interaction with the buried part of
the active site, and the results are presented in Table 3. Based
on the binding mode of 17, substitutions around the 5-position
may better fill part of the active site defined by the side chains
of Met120, Pro318, Asn167, Tyr319, and Tyr216 and three
group (Figure 2C). The structure of 79 bound showed that the
amide carbonyl group of the methylbenzamide analogue 79 was
within hydrogen bond distance of His117 and Tyr55 (Figure
2B). To accommodate the carboxamide, the carbonyl group has
rotated away from the oxyanion hole, displacing the benzyl
group relative to that of 17, causing the sulfonamide to move
further into the active site pocket and the position occupied by
an ethylene glycol, and acetate molecule in 17 and 85,
respectively, is now occupied by a water molecule.
When compared to other carboxylate-containing NSAIDs
with known binding modes, including flufenamic acid (11), the
binding mode of 17 most closely resembled that of 11 (Figure
2D, E). There is good agreement between the benzoic acid
moieties of both compounds, while the aniline unit of 11
7750
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
Figure 2. X-ray structures for AKR1C3 in complex with compounds 17 (2A), 79 (2B), and 85 (2C). Carbon atoms for the protein, compound, and
NADP+ are colored gray, green, and cyan, respectively. Key hydrogen bonds are represented as dotted lines. The surface of the active site cavity is
shown in gray with residues defining the pocket labeled. The water molecule located in a pocket formed by the side chains of Y216, Y319, and E192
is represented as a red sphere. An ethylene glycol molecule carried over from crystal freezing is bound in the active site of the structure with
compound 17, whereas an acetate molecule carried over from the crystallization condition occupies an equivalent position in the structure with
compound 85 (2D and 2E). Comparison between the structures of AKR1C3 in complex with compound 17 (green carbon atoms) and compound
11 (flufenamic acid; PDB code = 1S2C; purple carbon atoms) viewed from the side (2D) and above the plane of the nicotinamide ring of the
NADP+ (2E). Differences in binding modes cause a shift in position of the side chains of F311 and W227. Figure prepared using PyMOL (The
PyMOL Molecular Graphics System, Version 1.3, Schrodinger, LLC).
̈
Table 4. Comparison of Strain Energies for the Bound and
Unbound Ligands
Table 5. SAR for Analogous “Open-Chain” 3-(N-
Phenethylsulfamoyl)benzoic Acids
compd
E[Bound]
E[Free]
ΔE
rmsd
LPE
11
17
79
85
−7.46
−75.60
−32.98
−99.61
−14.75
−77.29
−35.91
−106.25
7.29
1.69
2.93
6.64
0.62
0.16
0.23
0.32
−9.87
−20.90
−16.87
−4.15
a,b
IC50 (μM)
b
no.
R
1C1
1C2
1C3
1C4
HCT-116
>1
56
57
58
59
60
61
62
63
64
H
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
0.22 0.16
0.33 0.13
0.21 0.05
0.14 0.08
0.74 0.29
1.44 0.11
0.53 0.21
0.22 0.05
0.64 0.07
>30
>30
>30
>30
>30
>30
>30
>30
>30
ordered water molecules not accessed by the tetrahydroquino-
line core. Consistent with this, substitutions that introduced
hydrogen bond donor and acceptor groups or lipophilic
moieties were well tolerated. Similarly, substitutions at the 6-
position were also well tolerated. By contrast, informed by the
experimentally derived binding mode, substitutions at the 7-
position were expected to be less tolerated, as this position is
4.04 Å of the backbone amide unit of Asn302. As seen in Table
3, a range of simple substituents at the 6-position appear to be
very well tolerated, with many compounds having AKR1C3
IC50 values comparable to (and up to 2-fold superior; 38) that
of the parent 17. Several of these also demonstrate improved
selectivity relative to AKR1C1. However, there are limits to the
size of this binding pocket; while the acetylene 43 is only 3-fold
2-F
3-Cl
>1
>1
4-Cl
3-OMe
4-OMe
3-OPh
3-Me
4-Me
a,b
As for Table 2.
less potent that 17, the much larger 4-trifluoromethoxypheny-
lacetylene 44 is 100-fold less potent. Overall, the most potent
compound was the 6-bromo analogue 38, with about 5000-fold
7751
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
selectivity for AKR1C3 over the other three isoforms.
Superposition of all human AKR1C3 entries in the PDB
indicated that the loop structure adjacent to position 7 can
undergo some conformational change and may be enough to
accommodate some substitutions.
The function of the carboxylate in ligand binding was
examined with the compounds presented in Table 6. This
compounds exhibited slightly reduced potency compared to
acid 17 but still had IC50 values in the submicromolar range.
Both simple secondary (79) and tertiary (80) amides were
acceptable, as was a range of analogues bearing solubilizing
units. The least effective of these was the strong base 78, which
was 70-fold less effective than 17. Strikingly, the 4-
pyrrolidinone 83 was only 3-fold less potent than the parent
compound, 17, with an IC50 of 0.042 μM. It is also noteworthy
that 3-(4-pyridinyl)ethyl amide 82 is also reasonably active,
being only 4-fold less potent than 17. This tolerance of the
extended aromatic substitution of 82 suggests some flexibility
inherent to the cavity about the A-ring.
Table 6. SAR for Ring A Substituents for Inhibition of AKR
Isoforms
Finally, Table 7 shows a set of “reverse sulfonamides”.
Compounds 84−88 have the carboxylic acid attached to a set
Table 7. SAR for “Reverse” Sulfonamides
a,b
As for Table 2.
showed that both the nature and positioning of the acid group
were critical. A variety of other 3-substituents, including H-
bond donor groups such as amino, were completely ineffective
(compounds 65−70), and even positioning the acid group at
the neighboring 4-position gave a compound (71) nearly 100-
fold less potent than 17. However, compounds 72 and 74
showed that acid isosteres were acceptable as H-bond acceptors
at the 3-position.
Inspection of the structure of 17 indicated that substituents
around the phenyl ring may occupy an extended cavity distal to
the oxyanion hole that has one wall defined by the cofactor
NADPH, and a small series of 3-amides of varying size, pKa,
and rigidity (75−82) were made to explore this. These
a,b
c
As for Table 2. 84 showed <10% inhibition at 3 μM.
of alicyclic and aromatic heterocycles that position it differently.
Compound 85 is of particular interest as a potent inhibitor of
AKR1C3. Determination of the crystal structure of its complex
with the enzyme showed that alternative ring systems could
appropriately present the carboxylate group to the oxyanion
hole. The IC50 data for compounds 84 and 85 showed the S
stereoisomer was about 12-fold less active (IC50s 0.40 and
0.032 μM, respectively). The pyrrolidone 86 and pyrazole 88
analogs had similar potency to 84, while the azetidine 87 was
7752
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
inactive. Compounds 89−94 retain the (racemic) piperidine
carboxylic acid and explore an alternative for the bicyclic unit.
Compounds 90, 91, and 93 retain some activity, but those (89,
92, 94) with more bulky units were inactive, demonstrating
again the limits of the lipophilic binding pocket.
selectivity over the other three isoforms for compound 38).
The broad requirement for an acid, isostere, or amide at the 3-
position is consistent with the crystal structure data (interaction
with Tyr55 and His117 at the oxyanion hole site). The 30-fold
loss in potency in going from 17 to 18 (loss of the benzene
ring) shows that the dihydroisoquinoline makes important
hydrophobic interactions in a cavity defined by residues
Asn167, Tyr216, Phe306, Tyr319, and Tyr317 (Figure 2B). A
wide variety of small substituents on the benzene were
tolerated, including the 7-alkyne 43, but the corresponding 2-
(4-methoxyphenyl)alkyne 44 was 100-fold less potent,
suggesting limits to this cavity. The compounds also showed
good potency in a cellular assay, blocking the ability of
AKR1C3 to metabolize a proven substrate. The high potency
and selectivity of the class for AKR1C3 over both other AKR1C
isoforms and over COX1 and COX2 make them of
considerable interest as both therapeutic inhibitors and
biological tools.
While the vast majority of the compounds showed no
discernible activity (IC50 > 30 μM) against the other three
AKR1 isoforms, there were a few exceptions, from which some
SAR can be discerned. Compound 24, one of three compounds
with substantial activity (IC50 < 1 μM) against AKR1C1, is also
the only compound in the set with an additional fused aromatic
ring on the acid-bearing unit. Of the only other compounds
with AKR1C1 IC50s < 1 μM, 53 is the only 8-substituted
analogue and 94 has the most extended amine-bearing unit.
Compound 94 also has 33-fold selectivity for AKR1C3 over the
other three isoforms, which may reflect differences in the
shape/size of the hydrophobic pocket. The only compound
among the NSAIDs of Table 1 or the {dihydroisoquinolin-
2[1H]-yl}sulfonyl)benzoic acids studied here with an IC50 < 10
μM against AKR1C4 was the 7-tetrazole 45. This, taken
together with the relative loss of AKR1C3 activity for
compound 44, which bears a very bulky 7-substituent, suggests
that exploration of further (hetero)aromatic systems at this
position might yield selective AKR1C4 analogues. Finally, the
5-chloro (29) and 8-chloro (53) analogues were the only
compounds to show any activity against AKR1C2; while this
was only weak (IC50 9.3 and 12.1 μM, respectively), it was
unexpected.
EXPERIMENTAL SECTION
■
Combustion analyses were performed by the Campbell Micro-
analytical Laboratory, University of Otago, Dunedin, New Zealand.
Melting points were determined on an Electrothermal IA9100 melting
point apparatus and are as read. NMR spectra were measured on a
Bruker Avance 400 spectrometer at 400 MHz for 1H and are
referenced to Me4Si. Chemical shifts and coupling constants are
recorded in units of parts per million and hertz, respectively. High-
resolution electron impact (HREIMS) and fast atom bombardment
(HRFABMS) mass spectra were determined on a VG-70SE mass
spectrometer at nominal 5000 resolution. High-resolution electrospray
ionization (HRESIMS) and atmospheric pressure chemical ionization
(HRAPCIMS) mass spectra were determined on a Bruker micrOTOF-
Q II mass spectrometer. Low-resolution atmospheric pressure
chemical ionization (APCI) mass spectra were measured for organic
solutions on a ThermoFinnigan Surveyor MSQ mass spectrometer,
connected to a Gilson autosampler. Thin-layer chromatography was
carried out on aluminum-backed silica gel plates (Merck 60 F254), with
exposure to I2, or by staining in permanganate and phosphomolybdic
acid dips and by UV light (254 and 365 nm). Column chromatography
was carried out on silica gel (Merck 230−400 mesh). Compounds of
Tables 1 and 2 were isolated following trituration in Et2O, unless
otherwise indicated. Tested compounds were ≥95% pure, as
determined by combustion analysis, or by HPLC conducted on an
Agilent 1100 system, using a reversed phase C8 column with diode
array detection.
CELLULAR ACTIVITY
■
Selected compounds displaying the greatest potency against the
isolated AKR1C3 isoenzyme were also evaluated for their
effectiveness in human HCT-116 colon cancer cells engineered
to overexpress AKR1C3. The assay utilized an exogenous
dinitrobenzamide substrate, PR-104A (see Supporting Infor-
mation for structure), which is exclusively metabolized by
AKR1C3 under aerobic conditions to its cytotoxic 4-hydroxyl-
amine (PR-104H) and 4-amine (PR-104M) metabolites.6 The
effectiveness of compounds at inhibiting AKR1C3 activity in
these cells was measured by their ability to inhibit this
metabolism. The parent compound 17 was able to inhibit
AKR1C3 activity (inhibiting PR-104H formation) with an IC50
of 0.027 0.002 μM. For the majority of the 45 compounds
evaluated in the HCT-116 assay, there was a broad rank order
between enzymatic and cellular activity, with some exceptions.
Compounds showing significantly less activity in the cellular
assay than expected by the above ranking were the more polar
ones (the OH- and NO2-substituted acids 32 and 35 and the
tetrazole isostere 74), and the aminooxoacetic acid 72, which
may be due to poorer uptake. This is reinforced by the fact that
most of the neutral amides in Table 6 (e.g., 76, 79, 80, 83)
showed better than expected cellular activity. Three other
compounds whose poorer cellular activity could not be
explained on pharmacological grounds did have unique
structures: the only 2-substituted dihydroisoquinoline 26 and
the two angular-substituted analogues 20 and 22.
Chemistry: General Procedures. Method A.22 A solution/
suspension of the amine or amine hydrochloride salt (1.00 or greater
equiv) in acetone or THF (∼0.03−0.30 M, AR grade) under N2 was
treated with Et3N or K2CO3 (1.0 equiv, or greater, especially if amine
was an HCl salt) and DMF (∼0.50−3.50 M, dry), and the resulting
mixture was cooled to 0 °C in an ice−water bath. A solution of the
sulfonyl chloride (1 equiv) in acetone (∼0.03−0.30 M, AR grade) was
then slowly added dropwise, and the resulting mixture was allowed to
warm to 20 °C over 3−90 h. After analysis (TLC, APCI) showed the
reaction was complete, the reaction mixture was filtered through a
Celite pad, and the solvent was removed under reduced pressure. The
residue was extracted into 1 M NaOH, and the resulting suspension
was filtered through a Celite pad. The filtrate was acidified with dilute
HCl (1 M) to afford a suspension, filtration of which afforded solid
crude product. (Alternatively, the residue was resuspended in EtOAc
and extracted with 1 M NaOH. The aqueous layer was filtered and
then acidified with 1 M HCl, and the resulting suspension was filtered
to give a solid.) Reprecipitation of this solid from an appropriate
solvent system, and/or purification by column chromatography on
silica gel, furnished the desired sulfonamide.
CONCLUSIONS
■
The results show that the 3-(3,4-dihydroisoquinolin-2(1H)-
ylsulfonyl)benzoic acids are novel, very potent, and highly
isoform-selective inhibitors of the AKR1C3 enzyme (up to an
IC50 of 6.1 nM for AKR1C3 with approximately 5,000-fold
Method B.27 Methanesulfonic acid (0.4 mL per mmol of
sulfonamide) and s-trioxan (0.3 equiv) were added to a solution/
7753
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
suspension of sulfonamide (1.00 equiv) in 1,2-dichloroethane (∼0.2
M) under N2, and the resulting solution/suspension was stirred at 35
°C overnight. After analysis (TLC, APCI) showed the reaction was
complete, the reaction mixture was cooled to 20 °C and then filtered.
The collected solid was purified by reprecipitation from EtOAc/
hexanes, with further purification by column chromatography on silica
gel, if necessary, to afford the product sulfonamide.
and extracted with EtOAc (×4).28,30 The combined organic extracts
were washed with brine, dried (MgSO4), and evaporated under
reduced pressure to afford crude product. This was combined with a
second run from 161 mg of 98 and by chromatography on silica gel,
eluting with 0−50% EtOAc/hexanes + 0.5% v/v AcOH followed by
reprecipitation from EtOAc/hexanes, to yield 43 (21 mg, 11%,
1
unoptimized) as an amorphous pale pink solid: mp 224−227 °C; H
NMR [(CD3)2SO]: δ 13.54 (v br s, 1H), 8.27 (t, J = 1.6 Hz, 1H), 8.21
(dt, J = 7.8, 1.3 Hz, 1H), 8.06 (d, J = 8.0 Hz, 1H), 7.76 (t, J = 7.8 Hz,
1H), 7.32 (s, 1H), 7.23 (dd, J = 7.9, 1.5 Hz, 1H), 7.11 (d, J = 8.0 Hz,
1H), 4.25 (s, 2H), 4.11 (s, 1H), 3.35 (t, J = 6.0 Hz, 2H), 2.84 (t, J =
5.9 Hz, 2H); HRESIMS calcd for C18H16NO4S m/z [M + H]+
342.0795, found 342.0792; HPLC purity 96.7%.
Compounds of Table 2 (Scheme 1A). 3-({5,6-Dihydropyridin-
1[2H]-yl}sulfonyl)benzoic Acid (18). Reaction of 3-(chlorosulfonyl)-
benzoic acid (95) and 1,2,3,6-tetrahydropyridine by method A,
followed by recrystallization of the crude product from EtOAc/
hexanes, afforded 18 (58%, unoptimized) as cream flaky crystals; mp
1
166−168 °C; H NMR [(CD3)2SO]: δ 13.54 (br s, 1H), 8.23 (m,
2H), 8.01 (dt, J = 7.9, 1.5 Hz, 1H), 7.78 (m, 1H), 5.73 (m, 1H), 5.65
(m, 1H), 3.53 (m, 2H), 3.14 (t, J = 5.7 Hz, 2H), 2.12 (m, 2H);
HRESIMS calcd for C12H14NO4S m/z [M + H]+ 268.0638, found
268.0641; HPLC purity 98.9%.
See Supporting Information for details of the syntheses of related
compounds 19−23 of Table 2 from sulfonic acid 95.
3-((7-((4-(Trifluoromethoxy)phenyl)ethynyl)-3,4-dihydroisoquino-
lin-2(1H)-yl)sulfonyl)benzoic Acid (44). Based on the procedures
described by Gu et al.,31 1-ethynyl-4-(trifluoromethoxy)benzene (99)
(0.035 mL, 0.23 mmol, 1.00 equiv) was added to a mixture of iodide
50 (100 mg, 0.23 mmol, 1.00 equiv) and Pd(PPh3)2Cl2 (6 mg, 0.009
mmol, 0.04 equiv) in piperidine (1.00 mL, 2.26 mmol, 45 equiv) at 20
°C. The resulting semisolid yellow-white mixture was heated to 85 °C
to form an orange solution and was then stirred at this temperature for
a further 12 h. The reaction mixture was cooled to 20 °C, diluted with
H2O and 15% HCl, and then extracted with Et2O (×4). The combined
organic layers were washed with water until neutral and then dried
(MgSO4) and concentrated to dryness under reduced pressure. The
residue was reprecipitated from EtOAc/hexanes to furnish 44 (21 mg,
20%, unoptimized) as an amorphous off-white solid; mp 235−238 °C;
1H NMR [(CD3)2SO]: δ 13.62 (v br s, 1H), 8.28 (t, J = 1.6 Hz, 1H),
8.21 (dt, J = 7.8, 1.4 Hz, 1H), 8.05 (d, J = 8.0 Hz, 1H), 7.75 (t, J = 7.8
Hz, 1H), 7.66 (ddd, J = 9.4, 4.8, 2.7 Hz, 2H), 7.42 (m, 3H), 7.33 (dd, J
= 7.9, 1.5 Hz, 1H), 7.16 (d, J = 8.0 Hz, 1H), 4.29 (s, 2H), 3.38 (t, J =
6.0 Hz, 2H), 2.87 (t, J = 5.8 Hz, 2H); HRESIMS calcd for
C25H18F3NO5S m/z [M + H]+ 502.0931, found 502.0940; HPLC
purity 98.7%.
Compounds of Table 3 (Scheme 1B, C). 3-({3,4-Dihydroiso-
quinolin-2[1H]-yl}sulfonyl)benzoic Acid (17).21 Reaction of 3-
(chlorosulfonyl)benzoic acid (95) and 1,2,3,4-tetrahydroisoquinoline
(96a) by method A, followed by reprecipitation of the crude product
from EtOAc/hexanes, afforded 17 (yield 50%, unoptimized) as an
1
amorphous white solid; mp 276−280 °C; H NMR [(CD3)2SO]: δ
13.53 (br s, 1H), 8.27 (t, J = 1.6 Hz, 1H), 8.21 (d, J = 7.8 Hz, 1H),
8.07 (dt, J = 7.8, 1.5 Hz, 1H), 7.76 (t, J = 7.8 Hz, 1H), 7.13 (m, 4H),
4.24 (s, 2H), 3.35 (toverlap water peak, J = 6.0 Hz, 2H), 2.84 (t, J = 6.0 Hz,
2H); HRESIMS calcd for C16H15NO4S m/z [M + H]+ 318.0795,
found 318.0794; HPLC purity 99.8%.
See Supporting Information for details of the syntheses of related
compounds 25−33, 35, 36, 38−41, 46, 47, and 49−50 of Table 3
from sulfonic acid 95.
3-({6-Chloro-3,4-dihydroisoquinolin-2[1H]-yl}sulfonyl)benzoic
Acid (37). Reaction of sulfonamide 63 with methanesulfonic acid and
s-trioxan by method B, followed by reprecipitation of the crude
product from EtOAc/hexanes, and then column chromatography on
silica gel (eluting with 0−40% EtOAc/hexanes + 0.5% v/v
CH3CO2H), afforded 37 (35%, unoptimized) as an amorphous
3-((7-(2H-Tetrazol-5-yl)-3,4-dihydroisoquinolin-2(1H)-yl)-
sulfonyl)benzoic Acid (45). Following the method of Nakamura et
al.,32 sodium azide (102 mg, 1.57 mmol, 1.25 equiv) was added to a
mixture of nitrile 51 (430 mg, 1.27 mmol, 1.00 equiv) and NH4Cl (84
mg, 1.57 mmol, 1.25 equiv) in DMF, and the resulting mixture was
stirred at 120 °C for 171 h. After this time, the reaction mixture was
cooled to 20 °C and then concentrated to dryness under reduced
pressure. Trituration of the resulting brown gum in an ultrasonic bath
afforded a glutinous tan solid, which was reprecipitated from MeOH/
EtOAc/hexanes to furnish 45 (327 mg, 68%) as an amorphous white
1
white solid; mp 202−203 °C; H NMR [(CD3)2SO]: δ 13.51 (br s,
1H), 8.27 (t, J = 1.6 Hz, 1H), 8.21 (dt, J = 7.8, 1.3 Hz, 1H), 8.06 (d, J
= 8.2 Hz, 1H), 7.76 (t, J = 7.8 Hz, 1H), 7.21 (m, 3H), 4.25 (s, 2H),
3.35 (toverlap water peak, J = 6.0 Hz, 2H), 2.84 (t, J = 6.0 Hz, 2H);
HRESIMS calcd for C16H15Cl NO4S m/z [M + H]+ 354.0376,
352.0405, found 354.0385, 352.0407; HPLC purity 97.1%.
1
solid; mp: 246 °C (dec.); H NMR [(CD3)2SO]: δ 14.77 (br s, 1H),
See Supporting Information for details of the syntheses of related
compounds 34, 42, and 48 of Table 3 from compounds 58, 64, and
59, respectively, of Table 5.
8.28 (t, J = 1.6 Hz, 1H), 8.21 (dt, J = 7.9, 1.3 Hz, 1H), 8.09 (dd, J =
7.9, 1.8, 1.2 Hz, 1H), 7.86 (s, 1H), 7.77 (m, 2H), 7.31 (d, J = 8.0 Hz,
1H), 4.37 (s, 2H), 3.41 (t, J = 6.0 Hz, 2H), 2.91 (t, J = 6.0 Hz, 2H)
{tetrazole -NH- not visible}; HRESIMS calcd for C17H15N5NaSO4Cl
m/z [M + Na]+ 408.0737, found 408.0748; HPLC purity 96.9%.
5-((3,4-Dihydroisoquinolin-2(1H)-yl)sulfonyl)nicotinic Acid (55).
2,4-Dichloro-5,5-dimethylhydantoin (2.42 g, 12.27 mmol, 1.61 equiv)
was added portionwise to a solution of methyl 5-mercaptonicotinate23
(100) in MeCN/AcOH/water (40:1.5:1.0, 128 mL) cooled at 0 °C
such that the reaction temperature remained <10 °C. Once the
addition was complete, the reaction mixture was allowed to warm to
and stirred at 20 °C for 110 min; then almost all solvent was removed
under reduced pressure (<30 °C), and the resulting concentrated
solution was diluted with CH2Cl2 (8 mL) and cooled to 0 °C. This
solution was treated with 5% NaHCO3 (53 mL), at such a rate that the
temperature remained <10 °C. The resulting mixture was stirred at 0−
5 °C for 15 min and then separated. The organic layer was washed
with a cooled (<10 °C) solution of 10% brine and then dried
(MgSO4) and filtered. Solvent was removed under reduced pressure
(bath temperature 30 °C) to afforded methyl 5-(chlorosulfonyl)-
nicotinate (101) as a yellow viscous oil (1.18 g), which was used
without further purification, as soon as possible.
3-((7-Ethynyl-3,4-dihydroisoquinolin-2(1H)-yl)sulfonyl)benzoic
Acid (43). A solution of iodide 50 (0.43 g, 0.96 mmol, 1.00 equiv) in
CH3CN (10 mL) and DMF (15 mL) was treated sequentially with
Pd(PPh3)4 (60 mg, 0.05 mmol, 0.05 equiv), CuI (11 mg, 0.06 mmol,
0.06 equiv), (trimethylsilyl)acetylene (97) (0.23 mL, 1.63 mmol, 1.70
equiv) and N,N-diisopropyl ethylamine (0.28 mL, 1.63 g, 1.70
equiv).28,29 The resulting mixture was deoxygenated by five N2-flush−
evacuate cycles and then stirred at 20 °C for 48 h. The reaction
mixture was then diluted with water and saturated aqueous NH4Cl and
then extracted with CH2Cl2 (×4). The combined organic layers were
washed with brine, dried (MgSO4), and evaporated under reduced
pressure. The resulting crude products from two equal runs were
combined and chromatographed on silica gel, eluting with 0−50%
EtOAc/hexanes + 0.5% v/v AcOH to afford 3-((7-((trimethylsilyl)-
ethynyl)-3,4-dihydroisoquinolin-2(1H)-yl)sulfonyl)benzoic acid (98)
(0.27 g, 37% combined yield) as an amorphous off-white solid
(APCI−): m/z 413.1; TLC Rf 0.47 (50% EtOAc/hexanes + 3 drops
AcOH)), which was used without further purification.
A solution of the TMS ethyne 98 (78 mg, 0.19 mmol, 1.00 equiv) in
MeOH (4 mL) was treated with anhydrous Cs2CO3 (0.19 g, 0.59
mmol, 3.12 equiv), and the resulting dark brown solution was stirred at
20 °C for 100 min and then diluted with saturated aqueous NH4Cl
101 (0.066 g, 0.28 mmol, 1.00 equiv) was reacted with 1,2,3,4-
tetrahydroisoquinoline (96a) (0.42 mL, 0.34 mmol, 1.20 equiv) by
method A, and the resulting crude ester (39 mg, 0.12 mmol, 1.00
7754
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
equiv) was dissolved in MeOH (5.8 mL) and water (3.9 mL) and
treated with anhydrous LiOH (9 mg, 0.40 mmol, 3.38 equiv) with
stirring at 20 °C for 80 min, when all starting material was consumed
(TLC). Solvent was removed under reduced pressure, the residue was
redissolved in H2O (50 mL) and acidified to pH 1 by addition of 1 M
HCl (∼1 mL), and the product was collected by filtration, washed with
hexanes, and dried to afford crude 55 as an amorphous pale yellow
solid. The product from two reactions was combined and
reprecipitated from EtOAc/hexanes to give pure 55 (87 mg, 30%)
as an amorphous tan solid; mp 233−237 °C; 1H NMR [(CD3)2SO]: δ
14.06 (br s, 1H), 9.26 (d, J = 1.9 Hz, 1H), 9.17 (d, J = 2.2 Hz, 1H),
8.48 (d, J = 2.1 Hz, 1H), 7.22−7.05 (m, 4H), 4.33 (s, 2H), 3.44
(toverlap water peak, J = 6.0 Hz, 2H), 2.84 (t, J = 6.0 Hz, 2H); HRESIMS
calcd for C15H15N2O4S m/z (M + H)+ 319.0747, found 319.0747;
HPLC purity 98.6%.
Compounds of Table 5 (Scheme 2). 3-(N-Phenethylsulfamoyl)-
benzoic Acid (56). Reaction of 3-(chlorosulfonyl)benzoic acid (95)
and 2-phenylethanamine (102a) (5.00 equiv), under the conditions
used to prepare acid 49, gave 56 (53%) as an amorphous white solid;
mp 188−190 °C; 1H NMR [(CD3)2SO]: δ = 13.44 (br s, 1H), 8.31 (t,
J = 1.6 Hz, 1H), 8.16 (d, J = 7.8 Hz, 1H), 7.99 (d, J = 8.1 Hz, 1H),
7.85 (t, J = 5.7 Hz, 1H), 7.71 (dd, J = 7.8, 7.7 Hz, 1H), 7.25 (m, 2H),
7.18 (m, 1H), 7.14 (m, 2H), 2.99 (m, 2H), 2.81 (dd, J = 7.6, 7.2 Hz,
2H); HRESIMS calcd for C15H15NNaO4S m/z [M + Na]+ 328.0614,
found 328.0610; TLC; HPLC purity 98.8%.
4.33 (q, J = 6.8 Hz, 2H), 4.21 (s, 2H), 3.29 (after D2O exchange, s,
2H), 2.87 (s, 2H), 1.32 (t, J = 6.8 Hz, 3H).
A stirred suspension of 104 (0.25 g, 0.64 mmol) in THF (6 mL)
was treated dropwise with a solution of Cs2CO3 (0.23 g, 0.71 mmol)
in water (3 mL). The mixture was stirred at 20 °C for 3 h and then
diluted with 0.1 H HCl (80 mL). The resulting solid was extracted
with dilute NH4OH, and the filtrate was acidified to give a solid.
Recrystallization of this from MeOH/water gave 72 (176 mg, 76%);
1
mp 194 °C; H NMR [(CD3)2SO]: δ 14.33 (v br s, 1H), 11.06 (s,
1H), 8.33 (t, J = 7.0 Hz, 1H), 8.04 (dt, J = 7.8, 1.7 Hz, 1H), 7.61 (t, J =
7.8 Hz, 1H), 7.57 (dt, J = 7.9, 1.5 Hz, 1H), 7.17−7.08 (m, 4H), 4.20
(2, 2H), 3.28 (after D2O exchange, t, J = 6.0 Hz, 2H), 2.87 (t, J = 5.9
Hz, 2H); HRESIMS calcd for C17H15N2O5 m/z [M − 1]− 359.0707,
found 359.0715; HPLC purity 99.7%.
2-((3-(2H-Tetrazol-5-yl)phenyl)sulfonyl)-1,2,3,4-tetrahydroisoqui-
noline (74). A mixture of nitrile 65 (269 mg, 0.90 mmol, 1.00 equiv)
and NH4Cl (60 mg, 1.13 mmol, 1.25 equiv) in DMF (14 mL) was
treated with solid sodium azide (73 mg, 1.13 mmol, 1.25 equiv), and
the resulting mixture was stirred at 120 °C for 69 h (method of
Nakamura et al.32). The reaction mixture was cooled to 20 °C and
concentrated to dryness under reduced pressure. Trituration of the
resulting orange-brown residue in an ultrasonic bath afforded a pale
orange-brown solid, which was reprecipitated from EtOAc/hexanes to
give tetrazole 74 (107 mg, 35%) as an amorphous off-white solid; mp
1
250−251 °C; H NMR [(CD3)2SO]: δ 17.07 (br s, 1H), 8.45 (t, J =
See Supporting Information for details of the syntheses of related
compounds 57−64 of Table 5 from sulfonyl chloride 95 and amines
102b−h.
1.6 Hz, 1H), 8.33 (dt, J = 7.9, 1.4 Hz, 1H), 8.00 (dt, J = 7.9, 1.4 Hz,
1H), 7.84 (t, J = 7.8 Hz, 1H), 7.12 (m, 4H), 4.29 (s, 2H), 3.39 (t, J =
6.0 Hz, 2H), 2.86 (t, J = 6.0 Hz, 2H); HRESIMS calcd for
C16H16N5O2S m/z [M + H]+ 342.1019, found 342.1027; HPLC purity
99.5%.
Compounds of Table 6 (Scheme 3A). 3-((3,4-Dihydroisoqui-
nolin-2(1H)-yl)sulfonyl)benzonitrile (65). A suspension of 1,2,3,4-
tetrahydroisoquinoline (96a) (0.76 mL, 6.05 mmol, 1.00 equiv) and 3-
cyanobenzene-1-sulfonyl chloride (103a) (1.22 g, 6.05 mmol, 1.00
equiv) in acetone (17 mL) was treated with Et3N (0.84 mL, 6.05
mmol, 1.00 equiv), and the resulting viscous suspension was stirred
vigorously for 3 days at 20 °C. The mixture was then filtered, and the
crude product was collected and reprecipitated to give 65 (247 mg,
14%, unoptimized) as an amorphous white solid; mp 192−194 °C; 1H
NMR [(CD3)2SO]: δ 8.25 (t, J = 1.6 Hz, 1H), 8.14 (m, 2H), 7.81 (t, J
= 7.9 Hz, 1H), 7.21−7.12 (m, 3H), 7.12−7.06 (m, 1H), 4.29 (s, 2H),
3.39 (t, J = 6.0 Hz, 2H), 2.83 (t, J = 6.0 Hz, 2H); HRESIMS calcd for
C16H14N2NaO2S m/z [M + Na]+ 321.0668, found 321.0661; calcd for
C16H15N2O2S m/z [M + H]+ 299.0849, found 299.0842; HPLC purity
99.5%.
Compounds of Table 6 (Scheme 3B). 3-((3,4-Dihydroisoqui-
nolin-2(1H)-yl)sulfonyl)-N-sulfamoylbenzamide (73). A solution of
acid 17 (1.80 g, 5.67 mmol, 1.00 equiv) in THF (15 mL) was treated
with CDI (2.00 g, 12.33 mmol, 2.17 equiv), and the resulting mixture
was stirred for 70 min at 60 °C to give the intermediate imidazolide.
The mixture was cooled to 20 °C and treated with sulfamide (1.105 g,
11.502 mmol, 2.028 equiv) for 30 min at 20 °C, and then a solution of
DBU (1.72 mL, 11.50 mmol, 2.03 equiv) was added dropwise. After
the exotherm, the mixture was stirred at 20 °C for 2.5 h and then
diluted with EtOAc (56 mL) and CH2Cl2 (68 mL) and washed with 1
M HCl (68 mL × 2). The combined aqueous layers were extracted
with CH2Cl2 (38 mL x 2), and the remaining aqueous suspension was
filtered through a pad of Celite to afford an amorphous yellow solid.
Reprecipitation of this from EtOAc/hexanes/MeOH furnished 73 (82
See Supporting Information for details of the syntheses of related
compounds 66, 68−71, and 72 of Table 6 from tetrahydroisoquinoline
96a and sulfonyl chlorides 103b−e.
1
mg, unoptimized) as an amorphous white solid; mp 188−189 °C; H
NMR [(CD3)2SO]: δ 12.25 (s, 1H), 8.36 (t, J = 1.6 Hz, 1H), 8.20 (dt,
J = 8.0, 1.3 Hz, 1H), 8.02 (d, J = 7.8 Hz, 1H), 7.74 (t, J = 7.8 Hz, 1H),
7.37 (br s, 2H), 7.18−7.09 (m, 4H), 4.25 (s, 2H), 3.35 (t, J = 6.0 Hz,
2H), 2.85 (t, J = 5.9 Hz, 2H); HRESIMS calcd for C16H17N3NaO5S2
m/z [M + Na]+ 418.0502, found 418.0503; HPLC purity 98.7%.
See Supporting Information for details of the syntheses of related
compounds 75−82 of Table 6 from the acid chloride of 17 and a
variety of known amines.
3-(3,4-Dihydroisoquinolin-2(1H)-ylsulfonyl)aniline (67). A solu-
tion of the 3-nitro analogue 66 (1.00 g, 3.14 mmol) in a mixture of
AcOH (1 mL), DMF (15 mL), EtOH (20 mL), and water (2 mL) was
treated at reflux with Fe powder (2.0 g, excess). The mixture was
heated for a further hour and then treated with concentrated NH4OH
(2 mL), filtered through Celite, and concentrated to a small volume.
Dilution with water gave a solid that was recrystallized from EtOAc/
petroleum ether to give 67 (0.81 g, 89%); mp 208−209 °C; 1H NMR
[(CD3)2SO]: δ 7.23 (t, J = 7.9 Hz, 1H), 7.19−7.07 (m, 4 H), 7.00 (t, J
= 1.9 Hz, 1H), 6.89 (d, J = 7.7 Hz, 1H), 6.81 (dd, J = 1.5, 1.0 Hz, 1H),
5.60 (s, 2H), 4.15 (s, 2H), 3.25 (t, J = 5.9 Hz, 2H), 2.86 (t, J = 5.8 Hz,
2H); HRESIMS calcd for C15H16NaN2O2S m/z [M + Na]+ 311.0825,
found 311.0845; HPLC purity 99.9%.
Compounds of Table 6 (Scheme 3C). 1-(4-(3,4-Dihydroisoqui-
nolin-2(1H)-ylsulfonyl)phenyl)pyrrolidin-2-one (83). 2-((4-
Iodophenyl)sulfonyl)-1,2,3,4-tetrahydroisoquinoline (107) (42 mg,
105 μmol), Pd2dba3 (25 mg, 27 μmol, 0.2 equiv), Xantphos (30 mg,
52 μmol, 0.4 equiv), 2-pyrrolidinone (11 mg, 126 μmol, 1.2 equiv),
and Cs2CO3 (52 mg, 160 μmol, 1.5 equiv) were dissolved in dry
dioxane (1 mL, 0.1 M) and heated at 100 °C for 18 h. The reaction
mixture was cooled to 20 °C, and solvent was removed under reduced
pressure. The residue was purified by column flash-chromatography on
silica to give 83 (71%); mp 180−181 °C; 1H NMR (CDCl3): δ 7.85−
7.80 (m, 4 H), 7.16−7.14 (m, 2 H), 7.17−7.14 (m, 2), 7.14−7.06 (m,
1 H), 7.03−7.01 (m, 1 H), 4.26 (s, 2 H), 3.89 (t, J = 7.20 Hz, 2 H),
3.37 (t, J = 6.00 Hz, 2 H), 2.92 (t, J = 6.00 Hz, 2 H), 2.65 (t, J = 8.00
Hz, 2 H), 2.22 (hept, J = 3.20 Hz, 2 H). HRESIMS m/z [M + H]+
357.1267; found: 357.1258; [M + Na]+ calculated: 379.1087; found:
379.1088; HPLC purity 99.8%. Anal. Calcd for C19H20N2O3S.0.5H2O:
C, H, N.
2-(3-(3,4-Dihydroisoquinolin-2(1H)-ylsulfonyl)phenylamino)-2-
oxoacetic Acid (72). A stirred suspension of 67 (400 mg, 1.39 mmol)
in a mixture of THF (30 mL) and pyridine (5 mL) was treated
dropwise with ethyl 2-chloro-2-oxoacetate (0.20 mL, 179 mmol) at 10
°C. The mixture was stirred at 20 °C for 1 h and then diluted with
water (200 mL) and stirred for a further 30 min. The resulting solid
was collected and washed with hexane to give ethyl 2-(3-(3,4-
dihydroisoquinolin-2(1H)-ylsulfonyl)phenylamino)-2-oxoacetate
(104) (513 mg, 95%) [mp (EtOAc/petroleum ether) 156−158 °C]
which was used directly; 1H NMR [(CD3)2SO]: δ 11.09 (s, 1H), 8.29
(s, 1H), 8.04 (d, J = 7.1 Hz, 1H), 7.68−7.53 (m, 2H), 7.14 (br s, 4H),
7755
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
Compounds of Table 7 (Scheme 4A). (S)-1-(Naphthalen-2-
ylsulfonyl)piperidine-3-carboxylic Acid (84). Reaction of naphtha-
lene-2-sulfonyl chloride (106) and (S)-piperidine-3-carboxylic acid
(107a) by method A gave acid 86 (yield 53%); mp 103−105 °C; H
NMR [(CD3)2SO]: δ 12.49 (sbr, 1 H), 8.45 (d, J = 1.44 Hz, 1 H), 8.23
(d, J = 7.96 Hz, 1 H), 8.17 (d, J = 8.72 Hz, 1 H), 8.08 (d, J = 14.04 Hz,
1 H), 7.78−7.66 (m, 3 H), 3.60 (dd, J = 8.20 Hz, J = 2.84 Hz, 1 H),
3.44−3.36 (m, 1 H), 2.60 (t, J = 9.88 Hz, 1 H), 2.55−2.45 (m, 2 H),
1.8−1.65 (m, 2 H), 1.57−1.43 (m, 1 H), 1.40−1.28 (m, 1 H) ppm.
HRMS (ESI): m/z = [M + H]+ calculated: 320.0951; found:
320.0955; [M + Na]+ calculated: 342.0770; found: 342.0777; [M +
K]+ calculated: 358.0510; found: 358.0508. HPLC purity 99.6%. Anal.
Calcd for C16H17NO4S: C,H,N.
See Supporting Information for details of the syntheses of related
compounds 85−88 of Table 7 from sulfonyl chloride 106 and
aminoacids 107b−e.
Compounds of Table 7 (Scheme 4B). 1-(4′-Chlorobiphenyl-4-
ylsulfonyl)piperidine-3-carboxylic acid (89). Reaction of piperidine-3-
carboxylic acid (108) and 4′-chlorobiphenyl-4-sulfonyl chloride
(109a) by method A gave acid 89 (yield 56%); mp 257−260 °C;
1H NMR [(CD3)2SO]: δ 12.51 (sbr, 1 H), 7.95 (dd, J = 6.73 Hz, J =
1.81 Hz, 2 H), 7.84−7.76 (m, 4 H), 7.58 (mc, 2 H), 3.49 (d, J = 8.93
Hz, 1 H), 3.29 (d, J = 11.24 Hz, 1 H), 2.60−2.40 (m, 3 H), 1.85−1.78
(m, 1 H), 1.78−1.68 (m, 1 H), 1.58−1.43 (m, 1 H), 1.42−1.30 (m, 1
H) ppm. HRMS (ESI): m/z = [M + H]+ calculated: 380.0718; found:
380.0707; [M + Na]+ calculated: 402.0537; found: 402.0529; [M +
K]+ calculated: 418.0277; found: 418.0260; HPLC purity 96.5%. Anal.
Calcd for C18H18ClNO4S: C,H,N.
See Supporting Information for details of the syntheses of related
compounds 90−94 of Table 7 from 108 and sulfonyl chlorides 109b−
f.
SYBYL8.0.3 (Tripos, St. Louis, MO). Preparation included stripping
water and addition of hydrogens followed by visual inspection.
Minimization of inhibitors within the AKR1C3 active site was
performed with SZYBKI v1.5.1 (OpenEye Scientific Software, Santa
Fe, NM) using the conjugate gradient method until convergence at
0.05 kcal/(mol·A). Only the protein polar hydrogens within 8 Å were
allowed to optimize along with the ligand. The MMFF94s potential
function was used with the exact analytical VdW potential, protein
dielectric set at 2, the Poisson−Boltzmann solvent model, and the
VdW protein−ligand interaction sphere at 18.0 Å; all frozen terms
were used in the single-point calculation. All other settings were used
at default values. Where the ligand was minimized in the absence of
protein, the MMFF94s potential function was used with the exact
analytical VdW potential, and the Poisson−Boltzmann solvent model
with a solvent dielectric of 1 and the cavity salvation term at 0.025.
The conjugate gradient method was used until convergence at 0.05
kcal/(mol·A); all other settings were used at default values.
COX Assays. (carried out by GVK Biosciences Ltd., Biology, 28A,
IDA Nacharam, Hyderabad 500 076 India: www.gvkbio.com)
COX 1: Human whole blood collected by venipuncture into
heparin (60 U/mL) tubes was half diluted into RPMI medium
(without FBS) and treated with test compound (10 μM) for 1 h at 37
°C in 5% CO2 in a 96-well plate. Calcium ionophore A23187 (50 μM)
was added, and the plates were incubated for another 30 min under
the same conditions and then centrifuged for 10 min at 250G/4 °C,
and supernatant was collected and frozen at −80 °C for ELISA.
COX 2: Heparinized human whole blood was treated with aspirin
(12 μg/mL) to inactivate COX 1 and then incubated for 6 h, diluted 2-
fold, and treated with test compound (10 μM) as above. LPS was then
added (final concentration 10 μg/mL), and the plates were incubated
for another 18 h and then centrifuged, and supernatant was collected
as above.
1
Crystallization and Structure Determination. C-terminal His6-
tagged human AKR1C3 was purified following a published
procedure,19 with the incorporation of a final Blue Sepharose 6B
step to produce stable, crystallization-grade protein.33 Two conditions
from the PACT premier screen (Molecular Dimensions Ltd.) were
found to give highly reproducible, single crystals for the binary
complex of AKR1C3 with NADP+: 25% (w/v) PEG1500, 0.1 M
PCTP buffer, pH 8.0 (condition C5) and 20% (w/v) PEG3350, 0.2 M
sodium acetate (condition E7). Sitting drop experiments were set up
on a Mosquito crystallization robot (TTP Labtech) using drops
comprising 800 nL of protein (20 mg·mL−1) + 800 nL of reservoir
solution at 16 °C. Crystals with a rod shaped morphology appeared
within 3 days. Crystals were soaked with 5 mM compound 79 or
compound 85 for 3 days and were subsequently flash frozen in liquid
nitrogen using 20% ethylene glycol + 5 mM compound in reservoir
solution as cryoprotectant. Crystals for AKR1C3 in complex with
compound 17 were obtained by incubating protein at 20 mg·mL−1
overnight in the presence of 2.4 mM compound prior to setting up
coarse screen cocrystallization experiments. Drops comprising 250 nL
of protein + 250 nL of reservoir solution were set up at 16C. Crystals
with a rodlike morphology grew from the PACT premier screen
condition, E4 (0.2 M potassium thiocyanate, 20% (w/v) PEG3350).
These crystals were cryopRotected with 20% ethylene glycol + 5 mM
compound in reservoir solution prior to flash freezing in liquid
nitrogen.
Data were collected at 100 K on a Rigaku Micromax-007 HF X-ray
generator with Varimax HF Optics and a Rigaku Saturn 944 CCD
detector. These data were processed using the program Structur-
eStudio (Rigaku). The structures were solved by molecular
replacement using the program PHASER34 and the structure of
AKR1C3+NADP+ (PDB code = 2FGB) as the search model. The
models were rebuilt using COOT35 and refined using REFMAC5.36
The refinement statistics are summarized in S2 (see Supporting
Information). The coordinates and diffraction amplitudes have been
deposited in the Protein Data Bank with accession codes 4FAM (17),
4FAL (80), and 4FA3 (86).
The supernatants were then assayed for thromboxane B2 (TBX2)
using an Assay Designs (Enzo Life Sciences Ltd.) immunoassay kit.
Production of Enzyme Expression Vectors. The AKR1C3 gene
was amplified from an IMAGE cDNA clone (3682448; Source
Bioscience, Nottingham, U.K.) by PCR using forward (AACCTTCA-
TATGGATTCCAAACACCAG) and reverse (AACCTTCTCGAGT-
TAATATTCATCTGAAT) primers that introduced an NdeI re-
striction site overlapping the start codon and a downstream XhoI site
before the stop codon. The digested fragment was ligated into a
pET28a plasmid vector (Merck Chemicals, Nottingham, U.K.), and
the nucleotide sequence of the cloned AKR1C3 gene was confirmed
by DNA sequencing. The expression plasmid incorporated an N-
terminal His-tag to aid purification. AKR1C1, AKR1C2, and AKR1C4
expression vectors were a gift from Dr Chris Bunce, University of
Birmingham.
Expression and Purification of Recombinant His-Tagged
Enzyme. The AKR1C enzyme constructs were transformed in
Escherichia coli BL21 (DE3) cells (Invitrogen, Paisley, Scotland,
U.K.) and plated on L-agar supplemented with 30 μg/mL kanamycin.
Colonies were cultured in 100 mL L-broth with 30 μg/mL kanamycin
for 6 h at 37 °C with shaking at 220 rpm. Cultures were expanded by
diluting 1:500 in Terrific broth supplemented with 30 μg/mL
kanamycin and were grown at 37 °C for 8 h with shaking at 160
rpm. Cells were harvested by centrifugation at 10000g for 30 min at 4
°C.
The cell pellet was resuspended in 50 mM potassium phosphate
buffer (pH = 7.2) with protease inhibitor cocktail (Roche Diagnostics
Ltd., Burgess Hill, West Sussex, U.K.), lysed by sonication on ice and
centrifuged at 20000g for 45 min at 4 °C. Protein was purified using an
AKTA FPLC (GE Healthcare, Chalfont St Giles, Buckinghamshire,
U.K.), by loading the cell supernatant onto an IMAC column charged
with Ni2+ using a P960 pump at 4 mL/min. Unbound protein was
washed with 50 mM potassium phosphate buffer (pH = 7.2) and 5
mM imidazole buffer at 1 mL/min. The recombinant protein was
eluted using a linear 5−500 mM imidazole gradient in 50 mM
potassium phosphate buffer (pH = 7.2) at 1 mL/min. Pooled fractions
containing overexpressed enzyme, as determined by SDS-PAGE, were
dialyzed against 50 mM potassium phosphate buffer (pH = 7.2) and 1
Energy Minimization. The structure for flufenamic acid bound to
AKR1C3 was obtained from the Protein Data Bank, code 1S2C, and
was prepared along with the structures determined here using
7756
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
mM DTT at 4 °C for a minimum of 4 h to remove imidazole. The
protein was concentrated on a Vivaspin concentrator (GE Healthcare)
with a 10000 Mw cutoff for size exclusion chromatography on a 26/60
Superdex 200 column (GE Healthcare) at 0.5 mL/min in 50 mM
potassium phosphate buffer (pH = 7.2), 1 mM DTT, and 150 mM
NaCl. The expression of purified N-terminal His-tagged recombinant
enzyme was confirmed by SDS-PAGE.
ABBREVIATIONS
■
AKR, aldo-keto reductase; COX, cyclooxygenase; APCI, low-
resolution atmospheric pressure chemical ionization mass
spectrometry; DMF, dimethylformamide; HREIMS,
HRFABMS, HRESIMS and HRAPCIMS, high-resolution
electron impact, fast atom bombardment, electrospray ioniza-
tion, and atmospheric pressure chemical ionization mass
spectrometry; NADP, nicotinamide adenine dinucleotide
phosphate; NSAID, nonsteroidal anti-inflammatory; PBD,
protein databank; THF, tetrahydrofuran
Measurement of AKR1C Enzyme Activity. A competitive
fluorescence assay was used to measure AKR1C enzyme activity,
where a nonfluorescent ketone probe (probe 5)37 selective for the
AKR1C enzyme isoforms is reduced to a fluorescent alcohol in the
presence of AKR1C enzyme and NADPH. Briefly, purified protein (2
μg/mL AKR1C1, 1 μg/mL AKR1C2, 2 μg/mL AKR1C3, and 5 μg/
mL AKR1C4) were incubated with 40 μM probe 5, test compounds,
and 50 μM NADPH in an assay buffer of 10 mM MOPS (pH = 7.2),
130 mM NaCl, 1 mM DTT, and 0.01% Triton-X-100 for 1 h at 37 °C.
The reaction was stopped by addition of 35 mM NaOH, and
fluorescence was read in a SpectraMax M2 microplate reader
(Molecular Devices, Sunnyvale, CA, USA) at excitation/emission
wavelengths of 420/510 nM. The compounds and known AKR1C3
inhibitors (flufenamic acid, indomethacin, naproxen, meclofenamic
acid, S(+)-ibuprofen and flurbiprofen; Sigma-Aldrich, Auckland, New
Zealand) were tested at multiple concentrations between 0.1 nM and
100 μM in 2% DMSO to generate AKR1C enzyme inhibition data.
Compound IC50 values were calculated by fitting the inhibition data to
a four-parameter logistic sigmoidal dose−response curve using Prism
5.02 (GraphPad, La Jolla, CA, USA).
REFERENCES
■
(1) Barski, O. A.; Tipparaju, S. M.; Bhatnagar, A. The aldo-keto
reductase superfamily and its role in drug metabolism and
detoxification. Drug Met. Rev. 2008, 40, 553−624.
(2) Penning, T. M.; Burczynski, M. E.; Jez, J. M.; Hung, C.-F.; Lin,
H.-K.; Ma, H.; Moore, M.; Palackal, N.; Ratnam, K. Human 3α-
hydroxysteroid dehydrogenase isoforms (AKR1C1−AKR1C4) of the
aldo-keto reductase superfamily: functional plasticity and tissue
distribution reveals roles in the inactivation and formation of male
and female sex hormones. Biochem. J. 2000, 351, 67−77.
(3) Komoto, J.; Yamada, T.; Watanabe, K.; Takusagawa, F. Crystal
structure of human prostaglandin F synthase (AKR1C3). Biochemistry
2004, 43, 2188−2198.
(4) Desmond, J. C.; Mountford, J. C.; Drayson, M. T.; Walker, E. A.;
Hewison, M.; Ride, J. P.; Luong, Q. T.; Hayden, R. E.; Vanin, E. F.;
Bunce, C. M. The aldo-keto reductase AKR1C3 is a novel suppressor
of cell differentiation that provides a plausible target for the non-
cyclooxygenase-dependent antineoplastic actions of nonsteroidal anti-
inflammatory drugs. Cancer Res. 2003, 63, 505−512.
(5) Novotna, R.; Wsol, V.; Xiong, G.; Maser, E.. Inactivation of the
anticancer drugs doxorubicin and oracin by aldo-keto reductase
(AKR)1C3. Toxicol. Lett. 2008, 181, 1−6.
(6) Guise, C. P.; Abbattista, M. R.; Singleton, R. S.; Holford, S. D.;
Connolly, J.; Dachs, G. U.; Fox, S. B.; Pollock, R.; Harvey, J.; Guilford,
P.; Donate, F.; Wilson, W. R.; Patterson, A. V. The bioreductive
prodrug PR-104A is activated under aerobic conditions by human
aldo-keto reductase 1C3. Cancer Res. 2010, 70, 1573−1584.
(7) Steckelbroeck, S.; Jin, Y.; Gopishetty, S.; Oyesanmi, B.; Penning,
T. M. Human cytosolic 3α-hydroxysteroid dehydrogenases of the
aldo−keto reductase superfamily display significant 3β-hydroxysteroid
dehydrogenase activity: implications for steroid hormone metabolism
and action. J. Biol. Chem. 2003, 279, 10784−10795.
(8) Penning, T. M.; Steckelbroeck, S.; Bauman, D. R.; Miller, M. W.;
Jin, Y.; Peehl, D. M.; Fung, K. M.; Lin, H. K. Aldo-keto reductase
AKR1C3: role in prostate disease and the development of specific
inhibitors. Mol. Cell. Endocrinol. 2006, 248 (1−2), 182−191.
(9) Brozic, P.; Turk, S.; Rizner, T. L.; Gobec, S. Inhibitors of aldo-
keto reductases AKR1C1-AKR1C4. Curr. Med. Chem. 2011, 18, 2554−
2565.
Inhibition of Cellular AKR1C3 Activity. AKR1C3 activity was
determined in HCT-116 cells engineered to overexpress AKR1C3 by
measuring the major metabolite (PR-104H) of PR-104A 2-electron
reduction under aerobic conditions, a reaction catalyzed selectively by
AKR1C3.6 The synthesis of PR-104A, PR-104H, and tetradeuterated
PR-104H,38,39 the transfection of the HCT-116/AKR1C3 cell line,6
and the measurement of PR-104H by LC-MS/MS have been
described previously.40 The compounds were administered 2 h prior
to 100 μM PR-104A at multiple concentrations between 1 nM and 3
μM. PR-104H formation was quantitated against a PR-104H standard
curve ranging from 1 to 1000 nM. Compound IC50 values were
calculated from four-parameter logistic sigmoidal dose−response
curves that were fitted to the inhibition data using Prism 5.02.
Representative compounds were tested over repeat assays to ensure
assay reproducibility.
ASSOCIATED CONTENT
■
S
* Supporting Information
Additional experimental procedures and characterizations for
the compounds in Tables 2, 3, and 5−7, as well as combustion
analytical data and the structures of probe 5 and PR-104A. This
material is available free of charge via the Internet at http://
pubs.acs.org.
AUTHOR INFORMATION
■
(10) Byrns, M. C.; Jin, Y.; Penning, T. M. Inhibitors of type 5 17β-
hydroxysteroid dehydrogenase (AKR1C3): Overview and structural
insight. J. Steroid Biochem. Mol. Biol. 2011, 125, 95−104.
(11) Davies, N. J.; Hayden, R. E.; Simpson, P. J.; Birtwistle, J.; Mayer,
K.; Ride, J. P.; Bunce, C. M. AKR1C isoforms represent a novel
cellular target for jasmonates alongside their mitochondrial-mediated
effects. Cancer Res. 2009, 69, 4769−4775.
Corresponding Author
*Tel.: +64 9 923 6144. Fax: +64 9 3737502. E-mail address: b.
denny@auckland.ac.nz.
Notes
The authors declare no competing financial interest.
(12) Stefane, B.; Brozic, P.; Vehovc, M.; Rizner, T. L.; Gobec, S. New
cyclopentane derivatives as inhibitors of steroid metabolizing enzymes
AKR1C1 and AKR1C3. Eur. J. Med. Chem. 2009, 44, 2563−2571.
(13) Adeniji, A. O.; Twenter, B. M.; Byrns, M. C.; Jin, Y.; Chen, M.;
Winkler, J. D.; Penning, T. M. Development of potent and selective
inhibitors of aldo-keto reductase 1C3 (type 5 17β-hydroxysteroid
dehydrogenase) based on N-phenyl-aminobenzoates and their
structure-activity relationships. J. Med. Chem. 2012, 55, 2311−2323.
ACKNOWLEDGMENTS
■
This work was partially funded by the Deutsche Forschungsge-
meinschaft under Grant HE6009/1-1 (to D.H.). The authors
thank Dr. Chris Guise and Dr. Adam Patterson for the HCT-
116/AKR1C3 cell line, and Dr. Chris Bunce for the AKR1C1,
AKR1C2, and AKR1C4 expression vectors.
7757
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758

Journal of Medicinal Chemistry
Article
(14) Komoto, J.; Yamada, T.; Watanabe, K.; Takusagawa, F. Crystal
structure of human prostaglandin F synthase (AKR1C3). Biochemistry
2004, 43, 2188−2198.
(15) Skarydova, L.; Zivna, L.; Xiong, G.; Maser, E.; Wsol, V.
AKR1C3 as a potential target for the inhibitory effect of dietary
flavonoids. Chem.-Biol. Interact. 2009, 178, 138−144.
(16) Endo, S.; Matsunaga, T.; Kanamori, A.; Otsuji, Y.; Nagai, H.;
Sundaram, K.; El-Kabbani, O.; Toyooka, N.; Ohta, S.; Hara, A.
Selective inhibition of human type-5 17β-hydroxysteroid dehydrogen-
ase (AKR1C3) by baccharin, a component of Brazilian propolis. J. Nat.
Prod. 2012, 75, 716−721.
(17) Adeniji, A. O.; Twenter, B. M.; Byrns, M. C.; Jin, Y.; Winkler, J.
D.; Penning, T. M. Discovery of substituted 3-(phenylamino)benzoic
acids as potent and selective inhibitors of type 5 17β-hydroxysteroid
dehydrogenase (AKR1C3). Bioorg. Med. Chem. Lett. 2011, 21, 1464−
1468.
(18) Xu, X. C. Cox-2 inhibitors in cancer treatment and prevention, a
recent development. Anti-Cancer Drugs 2002, 13, 127−137.
(19) Lovering, A. L.; Ride, J. P.; Bunce, C. M.; Desmond, J. C.;
Cummings, S. M.; White, S. A. Crystal structures of prostaglandin D2
11-ketoreductase (AKR1C3) in complex with the nonsteroidal anti-
inflammatory drugs flufenamic acid and indomethacin. Cancer Res.
2004, 64, 1802−1810.
(20) Byrns, M. C.; Penning, T. M. Type 5 17β-hydroxysteroid
dehydrogenase/prostaglandin F synthase (AKR1C3): Role in breast
cancer and inhibition by non-steroidal anti-inflammatory drug analogs.
Chem.-Biol. Interact. 2009, 178, 221−227.
(21) Kakefuda, A.; Kondoh, Y.; Hirano, M.; Kamikawa, A.; Enjo, K.;
Furutani, T. Benzimidazole derivatives. WO 2009/014150 A1,
published June 29th, 2009.
as new, potent, and selective 20-hydroxy-5,8.11,14-eicosatetraenoic
acid synthase inhibitors. J. Med. Chem. 2003, 46, 5416−5427.
(33) Zhou, M.; Wei Qiu, W.; Chang, H.-J.; Gangloff, A.; Lin, S.-X.
Purification, crystallization and preliminary X-ray diffraction results of
human 17β-hydroxysteroid dehydrogenase type 5. Acta Crystallogr.
2002, D58, 1048−1050.
(34) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M.
D.; Storoni, L. C.; Read, R. J. Phaser crystallographic software. J. Appl.
Crystallogr. 2007, 40, 658−674.
(35) Emsley, P.; Cowtan, K. Coot: model-building tools for
molecular graphics. Acta Crystallogr. 2004, D60, 2126−2132.
(36) Murshudov, G. N.; Lebedev, A.; Vagin, A. A.; Wilson, K. S.;
Dodson, E. J. Efficient anisotropic refinement of Macromolecular
structures using FFT. Acta Crystallogr. 1999, D55, 247−255.
(37) Yee, D. J.; Balsanek, V.; Sames, D. New tools for molecular
imaging of redox metabolism: development of a fluorogenic probe for
3α-hydroxysteroid dehydrogenases. J. Am. Chem. Soc. 2004, 126,
2282−2283.
(38) Yang, S.; Atwell, G. J.; Denny, W. A. Synthesis of asymmetric
halomesylate mustards with aziridineethanol/alkali metal halides:
application to an improved synthesis of the hypoxia prodrug PR-
104. Tetrahedron 2007, 63, 5470−5476.
3
2
(39) Atwell, G. J.; Denny, W. A. Synthesis of H- and H4-labelled
versions of the hypoxia-activated pre-prodrug 2-[(2-bromoethyl)-2,4-
dinitro-6-[[[2-(phosphonooxy)ethyl]amino] carbonyl]anilino]ethyl
methanesulfonate (PR-104). J. Labelled Compd. Radiopharm. 2007,
50, 7−12.
(40) Singleton, R. S.; Guise, C. P.; Ferry, D. M.; Pullen, S. M.; Dorie,
M. J.; Brown, J. M.; Patterson, A. V.; Wilson, W. R. DNA cross-links in
human tumor cells exposed to the prodrug PR-104A: relationships to
hypoxia, bioreductive metabolism, and cytotoxicity. Cancer Res. 2009,
69, 3884−3891.
(22) Shearer, B. G.; Patel, H. S.; Billin, A. N.; Way, J. M.; Winegar, D.
A.; Lambert, M. H.; Xu, R. X.; Leesnitzer, L. M.; Merrihew, R. V.;
Huet, S.; Willson, T. M. Discovery of a novel class of PPARδ partial
agonists. Bioorg. Med. Chem. Lett. 2008, 18, 5018−5022.
(23) Isabel, E.; Black, W. C.; Bayly, C. I.; Grimm, E. L.; Janes, M. K.;
McKay, D. J.; Nicholson, D. W.; Rasper, D. M.; Renaud, J.; Roy, S.;
Tam, J.; Thornberry, N. A.; Vaillancourt, J. P.; Xanthoudakis, S.;
Zamboni, R. Nicotinyl aspartyl ketones as inhibitors of caspase-3.
Bioorg. Med. Chem. Lett. 2003, 13, 2137−2140.
(24) Black, C.; Grimm, E. L.; Isabel, E.; Renaud, J. Nicotinyl aspartyl
ketones as inhibitors of caspase-3. Patent WO 01/27085 A1, 2001.
(25) Pu, Y.-M.; Christesen, A.; Ku, Y.-Y. A simple and highly effective
oxidative chlorination protocol for the preparation of arenesulfonyl
chlorides. Tetrahedron Lett. 2010, 51, 418.
(26) Byrns, M. C.; Steckelbroeck, S.; Penning, T. M. An
indomethacin analogue, N-(4-chlorobenzoyl)-melatonin, is a selective
inhibitor of aldo-keto reductase 1C3 (type 2 3α-HSD, type 5 17β-
HSD, and prostaglandin F synthase), a potential target for the
treatment of hormone dependent and hormone independent
malignancies. Biochem. Pharmacol. 2008, 75, 484−493.
(27) Orazi, O. O.; Corral, R. A.; Giaccio, H. Synthesis of fused
heterocycles: 1,2,3,4-tetrahydroisoquinolines and ring homologues via
sulphonamidomethylation. J. Chem. Soc., Perkin Trans. 1 1986, 1977.
(28) Feng, D.-J.; Wang, P.; Li, X.-Q.; Li, Z.-T. Self-assembly of
hydrazide-based heterodimers driven by hydrogen bonding and donor-
acceptor interaction. Chin. J. Chem. 2006, 24, 1200−1208.
(29) Myers, A. G.; Dragovich, P. S. Synthesis of functionalises enynes
by palladium/copper-catalyzed coupling reactions of acetylenes with
(Z)-2,3-dibromopropenoic acid ethyl ester: (Z)-2-bromo-5-(trime-
thylsilyl)-2-penten-4-ynoic acid ethyl ester. Org. Synth. 1995, 72, 104−
111.
(30) Nicolaou, K. C.; Tang, Y.; Wang, J. Total synthesis of sporolide
B. Angew. Chem., Int. Ed. 2009, 48, 3449−3453.
(31) Gu, Z.; Li, Z.; Liu, Z.; Wang, Y.; Liu, C.; Xiang, J. Simple,
efficient copper-free Sonogashira coupling of haloaryl carboxylic acids
or unactivated aryl bromides with terminal alkynes. Catal. Commun.
2008, 9, 2154−2157.
(32) Nakamura, T.; Sato, M.; Kakinuma, H.; Miyata, N.; Taniguchi,
K.; Bando, K.; Koda, A.; Kameo, K. Pyrazole and isoxazole derivatives
7758
dx.doi.org/10.1021/jm3007867 | J. Med. Chem. 2012, 55, 7746−7758