Discovery and preliminary SARs of keto-indoles as novel indoleamine 2,3-dioxygenase (IDO) inhibitors

Article abstract of DOI:10.1016/j.ejmech.2011.02.049

Indoleamine 2,3-dioxygenase (IDO) is an important new therapeutic target for the treatment of cancer. With the aim of discovering novel IDO inhibitors, a virtual screen was undertaken and led to the discovery of the keto-indole derivative 1a endowed with an inhibitory potency in the micromolar range. Detailed kinetics were performed and revealed an uncompetitive inhibition profile. Preliminary SARs were drawn in this series and corroborated the putative binding orientation as suggested by docking.

Full text of DOI:10.1016/j.ejmech.2011.02.049

European Journal of Medicinal Chemistry 46 (2011) 3058e3065
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Short communication
Discovery and preliminary SARs of keto-indoles as novel
indoleamine 2,3-dioxygenase (IDO) inhibitors
b
c
d
a
ꢀ ꢁ a
Eduard Dolusic , Pierre Larrieu , Sébastien Blanc , Frédéric Sapunaric , Jenny Pouyez ,
Laurence Moineaux a, Delphine Colette c, Vincent Stroobant b, Luc Pilotte b, Didier Colau b, Thierry Ferain c,
Graeme Fraser c, Moreno Galleni d, Jean-Marie Frère d, Bernard Masereel a, Benoît Van den Eynde b,
Johan Wouters a, Raphaël Frédérick a
,
*
a Drug Design and Discovery Center, University of Namur, 61 Rue de Bruxelles, 5000 Namur, Belgium
b Ludwig Institute for Cancer Research, Université Catholique de Louvain, 74 Avenue Hippocrate, 1200 Brussels, Belgium
c Euroscreen SA, 47 Rue Adrienne Boland, 6041 Gosselies, Belgium
d Centre d’Ingénierie des Protéines, Université de Liège, Allée du 6 août, 4000 Liège, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Indoleamine 2,3-dioxygenase (IDO) is an important new therapeutic target for the treatment of cancer.
With the aim of discovering novel IDO inhibitors, a virtual screen was undertaken and led to the
discovery of the keto-indole derivative 1a endowed with an inhibitory potency in the micromolar range.
Detailed kinetics were performed and revealed an uncompetitive inhibition profile. Preliminary SARs
were drawn in this series and corroborated the putative binding orientation as suggested by docking.
Ó 2011 Elsevier Masson SAS. All rights reserved.
Received 1 December 2010
Received in revised form
17 February 2011
Accepted 21 February 2011
Available online 26 February 2011
Keywords:
Indoleamine 2,3-dioxygenase
Enzyme inhibitors
Keto-indoles
IDO
Structure-based drug discovery
Structureeactivity relationships
1. Introduction
described that, except for some dendritic cells, IDO is not expressed
in normal tissue [5,6]. However, it is highly produced in the
Immunotherapy is a promising new strategy for cancer therapy.
It consists of the therapeutic vaccination of cancer patients to
stimulate their (natural) immune system against cancer cells. This
approach, however, showed limited efficacy in vivo. Cancer cells are
actually able to develop enzymatic mechanisms allowing tumors to
resist or escape immune rejection. Among the enzymes involved,
indoleamine 2,3-dioxygenase (IDO) plays a key role [1].
IDO is an intracellular monomeric heme-containing enzyme
that catalyzes the rapid degradation of tryptophan (Trp) in the
initial step of the kynurenine pathway, the de novo biosynthetic
route for nicotinamide adenine dinucleotide (NAD) production
[2,3]. This results in a local Trp depletion that severely affects the
proliferation of T lymphocytes and is thereby profoundly immu-
nosuppressive [4]. Pioneering studies by Munn and Mellor
placenta and is required for the fetus immune tolerance by the
organism of the mother during pregnancy [6,7]. Recently, Van den
Eynde showed that many types of human tumors express IDO in
a constitutive manner [1]. The expression of IDO in human ovarian
and colon carcinomas is also associated with poor prognosis and
reduced survival [8,9]. Finally, because IDO expression is strongly
induced by interferon-gamma [10], even tumors that do not
constitutively express IDO may do so when exposed to inflamma-
tory conditions, e.g. resulting from an ongoing immune response.
IDO thus noticeably protects foreign cells against immune rejec-
tion. Based on its implication in immunosuppression, and partic-
ularly in cancer tumors, IDO clearly represents an attractive target
for the development of inhibitors [11e14].
Until recently, the best known IDO inhibitors [15] displayed
affinities in the micromolar range and comprised mainly Trp
derivatives such as 1-methyl-
and -carbolines [17]. In 2006, potent nanomolar inhibitors were
isolated from marine invertebrate extracts [18,19] At the same time,
L-tryptophan (1MT) [16] (Ki w 34 mM)
b
* Corresponding author. Tel.: þ32 81 72 42 90; fax: þ32 81 72 42 38.
E-mail address: raphael.frederick@fundp.ac.be (R. Frédérick).
0223-5234/$ e see front matter Ó 2011 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2011.02.049

ꢀ ꢁ
E. Dolusic et al. / European Journal of Medicinal Chemistry 46 (2011) 3058e3065
3059
new brassinin-based IDO inhibitors were published [20,21], the
best having a Ki of 12 M. In 2008, two new classes of compounds
were reported based on the earlier discovery of the marine inver-
tebrate inhibitors. Carr et al. identified tryptamine quinone as the
core pharmacophore of exiguamine A and described a series of
preferential binding of some inhibitors to the inactive ferric form of
IDO, and the redox activity of others. Since many of the IDO
inhibitors that have been discovered recently might fail in in vivo
testing, there is still considerable interest in discovering novel IDO
inhibitor families [29].
m
derivatives, the best showing a Ki of 0.2
m
M [22]. Kumar et al.
The two three-dimensional structures of IDO, in complex with
PIM and the cyanide ion (CNꢀ), provide important data for the
structure-based drug design of novel IDO inhibitors [25]. From
a structural point of view, IDO is characterized by a small and highly
lipophilic active site (Fig. 1a).
reported optimization of the naphthoquinone core of annulin B and
designed derivatives, partly based on structural modeling, with IC50
values reaching 60 nM [23]. In 1989, 4-phenylimidazole (PIM) was
identified as a modestly potent IDO inhibitor [24]. Despite the
uncompetitive inhibition kinetics, the authors showed through
spectroscopic studies that PIM binds into the active site of IDO.
Moreover, they discovered a preferred binding of PIM to the inac-
tive ferric (Fe3þ) form of the enzyme, which may explain this
apparent discrepancy. The first crystal structure of human IDO
complexed with the inhibitor reported later [25] confirmed the
binding of PIM in the active site. A recently undertaken structure-
based study [26] of PIM-derived molecules prompted by Sugimo-
to’s work confirmed the antecedent findings while providing access
to novel inhibitors with low micromolar potencies.
In the crystal structure, PIM interacts with the heme iron
through its N-1 atom, projecting its phenyl ring toward the lipo-
philic cavity (Pocket A, Fig. 1b). The PIM binding site consists of
residues Tyr126, Cys129, Val130, Phe163, Phe164, Ser167, Leu234,
Gly262, Ser263, Ala264, and the heme ring (Fig. 1c). Putative
hydrogen bonding sites are the thiol group of Cys129, the hydroxyl
group of Ser167, the CO group of Gly262, the NH group of Ala264,
and the heme 7-propionate group. Ligands larger than PIM may
also interact with Phe226, Arg231, Ser235, Phe291, Ile354, and
Leu384, which are located at the binding site entrance. Here,
additional hydrogen bonds are possible with the side chain of
Arg231. A hydrophobic pocket in this region is provided by Phe163,
Phe226, Leu234, Ile354, and the heme ring (Pocket B, Fig. 1b). The
cyanide-bound structure (PDB entry 2D0U) differs from the PIM-
bound structure mainly in the access to pocket A, which is hindered
by an inward movement of the backbone of the loop residues
262e266 and the side chain of Phe163, suggesting some flexibility
in the active site. In both structures, two buffer molecules (2-[N-
cyclohexylamino]ethanesulfonic acid) are bound at the entrance of
Very recently, novel natural product inhibitors (IC50 z 2 mM)
[27] and potent competitive inhibitors including a hydroxyamidine
chelating motif (IC50 values up to 60 nM) have also been reported
[28].
Many IDO inhibitors show either noncompetitive or uncom-
petitive inhibition kinetics similar to those of PIM and norharman,
which have been shown to bind directly to the heme iron and to
occupy the presumed Trp binding site [26]. The interpretation of
IDO inhibition kinetics may, however, be complicated due to the
Fig. 1. (a) Lipophilicity mapped onto the Connoly surface of IDO (Color code: brown ¼ high lipophilicity, green ¼ medium lipophilicity and blue ¼ high hydrophilicity; picture made
with MOLCAD) [30] and (b and c) active site of IDO with PIM as observed in the X-ray crystal structure (PDB code 2D0T). A red sphere was used for the heme iron. (Picture made
using PYMOL) [31]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

ꢀ ꢁ
E. Dolusic et al. / European Journal of Medicinal Chemistry 46 (2011) 3058e3065
3060
Directory) and the ZINC database [33] (over 2.5 million commer-
cially-available compounds) can be used to perform VS. We chose
the latter as it is free, web-accessible, offers the ligands in a ready-
to-dock 3D format, and hits are marketed for confirmatory bio-
logical evaluation.
Since careful docking of the entire ZINC library would take
a prohibitively long time, filters were used to limit the number of
compounds to be considered before the docking experiment. The
VS flowchart we employed is depicted in Fig. 2.
(i) Lipinski-style rules were first applied [34]. As our primary aim
at this very first step was to select only compounds that could
serve as leads for future medicinal chemistry development, we
used simple molecular descriptors such as the molecular
weight (MW ꢁ 250), the hydrophobicity (LogP ꢁ 2.5), and the
number of rotatable bonds (RB ꢁ 5) as the first filter. Then, all
of the qualifying structures were downloaded from the ZINC
website and transformed into a UNITY hit list file (Sybyl
version 7.3) [30,35] for compatibility with our modeling
system. This resulted in a library of roughly 62,000 structures
(Fig. 2, (i)).
Fig. 2. IDO virtual screening flowchart. (i) Lipinski-style rules, (ii) goldscore ꢂ50, (iii)
Cscore ꢂ4, (iv) Visual analysis and commercial-availability.
(ii) All fragments were then docked within the IDO active site
(PDB code 2D0T) by means of the automated GOLD program
[36]. For each ligand, 3 conformations were generated and
ranked according to the GOLDSCORE. Only the conformations
possessing the best score were subsequently retained and the
top-ranked ligands (GOLDSCORE ꢂ50) were isolated. This led
to a set of 18 550 structures.
(iii) This set was further refined using Cscore calculations [30].
Cscore is a combination of five different scoring functions,
namely G-, D-, PMF-, F- and Chemscore. The Cscore value is an
integer comprised between 5 (all of the five scoring function
have identified the ligand as a good potential binder) and
0 (none of the 5 scoring functions has identified the ligand as
a binder) and helps distinguishing the best potential ligands
from the less probable ones. Concomitantly to Cscore evalua-
tion, the individual contributions (van der Waals, H-bonding,
etc.) of Chemscore were also analyzed and particularly the
the active site, forming a hydrogen bond to the heme propionate
and interacting with pocket B.
As the enzyme active site is now well known, it represents a very
good tool to undertake a direct approach of receptor-based drug
discovery and design. Among these, virtual screening (VS) [32]
involves the computational screening of very large libraries of
commercially-available chemicals that complement targets of
known structure, followed by the experimental testing of those with
the best predicted binding energies. We have applied this strategy in
the present work for the discovery of new IDO inhibitors.
2. Results
Various chemical libraries, for example the NCI database
(National Cancer Institute), the ACD library (Available Chemical
Fig. 3. IDO inhibition of the selected compounds at [I] ¼ 100
mM.

ꢀ ꢁ
E. Dolusic et al. / European Journal of Medicinal Chemistry 46 (2011) 3058e3065
3061
Table 1
uncompetitive inhibition mode (Fig. 4). It is characterized by an
inhibition constant (Ki) of w190 M (Table 1), in a range similar to
Inhibitory IDO activity of 1a identified using VS.
m
the IC50. These data indicate that, as an uncompetitive inhibitor, 1a
would only bind to the substrate-bound enzyme. This might be
surprising taking into account the way it was discovered i.e. virtual
screening/docking to the heme substrate IDO binding site.
However, recently Kumar et al. observed similar inhibition kinetic
profiles for phenylimidazoles [26], despite their binding to the
heme iron at the active site as experimentally established through
X-ray studies [25]. One hypothesis that could, at least in part,
explain this uncompetitive profile would be a high affinity of these
molecules to the inactive ferric vs. active ferrous form of the
Cmpd
IC50
(
m
M)
Ki (
m
M)
Type of inhibition
1MT
100 ꢄ 20
65 ꢄ 7
7
190
Competitive
Uncompetitive
Zinc00387478 ¼ 1a
chemscore-metal value which reflects the formation of an
interaction between the ligand and the metal, namely, the
heme iron. Thus, keeping compounds characterized by
a Cscore value ꢂ4 and a chemscore-metal value of 1, respec-
tively, led to a final set of 183 derivatives.
(iv) These 183 structures were visually analyzed within the IDO
active site and 39 of them were selected based on structural
diversity, synthetic accessibility, appropriate conformation and
real commercial availability. This step introduces some subjec-
tivity in the selection and relies on the expertise of the medicinal
chemist(s) who perform it. The 39 structures were obtained
fromthe vendors (Chembridge, Sigma, Specs, Maybridge, Acros).
enzyme. Similar hypotheses were also reported earlier for b-car-
boline derivatives and could probably explain the kinetic profile
observed for 1a [17]. To corroborate our hypothesis, we performed
UV analysis of IDO with or without 1a and analyzed the effect on
the Soret band, characteristic of the heme group. The optical
absorption spectrum of the substrate-free ferric IDO exhibits
a Soret maximum at 404 nm and the binding of L-Trp causes, as
reported previously, a red-shift of the Soret maximum to 410 nm
together with a hypochromic effect (Fig. 5a) [37]. In the case of 1a
addition (Fig. 5b), a dose-dependent hypochromic effect without
red-shift is observed thus confirming a close interaction of 1a with
the heme binding site.
The biological activity of the 39 compounds was evaluated using
a colorimetric in vitro IDO inhibition assay [1]. Briefly, the inhibitors
are incubated (6 min, 37 ꢃC) with IDO and
L-Trp, the natural
substrate. Trichloroacetic acid (TCA) is then added to quench the
reaction and the N-formylkynurenine generated is hydrolyzed to
kynurenine during the next 30 min. After that time, p-dimethyla-
minobenzaldehyde (pDMAB) is added to the reaction mixture to
form a Schiff base with kynurenine. The IDO residual activity is
measured at 490 nM and corresponds to the Schiff base formed.
Six derivatives (Zinc code 276331, 3081571, 2492429, 1052355,
448979, and 387478) out of the 39 compounds analyzed displayed
an inhibitory potency >30% at a concentration of 100 mM, two of
them (Zinc 276331 and Zinc 387478) possessing more than 50% IDO
inhibition at this concentration (Fig. 3).
Unfortunately, the low solubility of the benzotriazole
(Zinc00276331) did not allow determination of a proper IC50 for this
compound. In contrast, the indole derivative (Zinc00387478 ¼ 1a)
proved to be at least as potent as 1MT, being characterized by an IC50
Next, preliminary SARs around 1a were investigated. Indeed,
one attracting feature of virtual screening is that it allows, once a hit
is identified, to directly analyze the interactions stabilizing the
molecule within the active site cavity of the target, and subse-
quently to suggest pharmacomodulations on the basis of the
putative binding orientation. Thus, the interactions stabilizing 1a
inside the IDO binding cleft were analyzed in detail and revealed
that the oxygen atom of the ketone function, being situated w2 Å
above the plane of the heme, coordinates the heme iron. The indole
ring is found to be stabilized in the so-called lipophilic A-pocket of
IDO. Indeed, this stabilization of the indole in pocket A was already
suggested by other groups to explain substrate binding as well as
interaction of 1MT. In this position, the 3-pyridyl group is projected
toward the entrance of the active site and is stabilized in the
aromatic pocket B through T-shape interactions with Phe163 and
Phe226 (Fig. 6). In this orientation, excellent complementarities are
found between the indole nucleus and pocket A. Indeed, apart from
the 4-, and 5-position as well as the 1-position (indole NH), only
little space is left available to accommodate substituents (Fig. 6).
A few analogs of 1a were then synthesized in order to assess
whether or not the binding mode suggested by docking could
explain preliminary experimental SARs and thus profitably offer up
hypotheses for further structure-based design effort. According to
of 65 mM for IDO inhibition (Table 1). This compound is characterized
by an indole nucleus substituted in the 2-position by a 3-pyridinyl-
ethanone function. Interestingly, no biological activity was reported
so far for this compound, making it an interesting tool for chemical
optimization to further enhance IDO inhibition. Compared to 1MT,
1a also has the advantage of containing no chiral center.
Detailed kinetic experiments were performed on 1a to elucidate
its mode of inhibition. In contrast to 1MT which shows a competi-
tive IDO inhibition profile (Table 1), this compound exhibited an
90
a
[I] = 0 µM
b
180
160
140
120
100
80
[I] = 0 µM
[I] = 20 µM
[I] = 40 µM
[I] = 200 µM
[I] = 20 µM
[I] = 40 µM
[I] = 60 µM
80
70
60
50
40
30
20
60
40
10
0
20
0
120
80
100
0
20
40
60
80
100
120
0
20
40
60
[S], µM
[S], µM
Fig. 4. Hanes plot of IDO activity against L-tryptophan in presence of increasing concentration of 1-methyl-L-tryptophan (1MT) (a) and 1a (b).

ꢀ ꢁ
E. Dolusic et al. / European Journal of Medicinal Chemistry 46 (2011) 3058e3065
3062
Fig. 5. (a) Absorption spectra of ferric IDO alone or with L-Trp 20 mM. (b) Absorption spectra of ferric IDO alone or with increasing concentration of 1a. All the data were obtained
with 12.6 M IDO in 50 mM Tris buffer (pH 8.0) at room temperature.
m
the structural analysis reported above, we hypothesized that
substitution of the indole in the 1-, 4- and 5-position through
introduction of small groups such as halogens, methyl, methoxy or
nitro moieties could improve the IDO inhibitory potency in this
series. We also investigated the importance of the ketone function
of 1a suggested to be directly involved in the interaction with the
heme iron, through replacement of the ethanone linker by an
ethylene, a hydroxyethylene or an amide.
A synthetic pathway allowing to access these structures was
therefore elaborated (Scheme 1). It involves the deprotonation of
3-methylpyridine in presence of lithium diisopropylamide (LDA).
The carbanion generated in situ then reacts with a reactive indole
carbonyl compound at 0 ꢃC to provide the desired keto-indole
compounds of the general formula 1 The WolffeKishner reduction
and the catalytic hydrogenation of 1a (R ¼ R0 ¼ H), respectively
afforded the ethylene linked compound 2 and secondary alcohol 3.
Finally, amide 4 was made from the corresponding 1H-indole-2-
carboxylic acid by reaction with thionyl chloride followed by
nucleophilic substitution with 3-aminopyridine (Scheme 1).
The IC50’s of the novel indole derivatives were determined on
isolated enzyme and are summarized in Table 2.
derivative 1g is deprived of any IDO inhibitory potency. This
preliminary SARs suggests, in agreement with the docking study
reported above, that only rather small and lipophilic groups are
tolerated in this position.
Regarding the importance of the ketone function on 1a,
replacement of the ethanone linker with an ethylene (2) or even
a hydroxyl-ethylene spacer (3) totally suppresses the IDO inhibitory
potency thus confirming the importance of the ketone pharmaco-
phore for IDO inhibition in this series. Only its replacement with an
amide (4) afforded a compound that is still active against IDO.
We also appraised the ability of our derivatives to inhibit
tryptophan degradation and kynurenine production in cells
expressing murine IDO. The compounds were also screened for
their inhibitory activity in cells expressing murine tryptophan 2,3-
dioxygenase (TDO), an enzyme functionally related to IDO, with
a view to investigate their selectivity. For all cell lines, cell viability
was evaluated at the end of the assay. This cellular assay is
informative for drug development as it evaluates not only the IDO/
TDO inhibitory effect of the compounds but also their capacity to
permeate the cell, their potential cytotoxicity, inhibition of tryp-
tophan transporters, and the effects of their metabolites. Con-
cerning the ability of these derivatives to inhibit tryptophan
degradation in cells expressing murine IDO, 1c, 1d and 1e are the
most potent derivatives with 20e25% inhibition of tryptophan
As expected from our docking analysis, a two-fold improvement
in the IDO inhibitory potency is observed when fluorine is intro-
duced in the 5-position (1c) around the indole scaffold. In contrast,
its introduction in the 4-position (1b) appears to be detrimental. A
bulkier halogen in this 5-position such as a chlorine (1d) also led to
a potent IDO inhibitor. The 5-methyl (1e) and the 5-methoxy (1f)
analogs are still active but their IDO inhibition is not improved
compared to the unsubstituted compound 1a. Finally, the 5-nitro
degradation at 20
mM, most of the other compounds being
deprived of any cellular activity. The relative discrepancy in IDO
inhibition between the isolated enzyme and cellular assays
probably reflects a poor permeation in this series. Interestingly,
the more potent IDO inhibitors in cell-based assay are deprived of
Fig. 6. (a and b) Predicted binding mode for 1a as docked in GOLD (Pictures made with MOLCAD).

ꢀ ꢁ
E. Dolusic et al. / European Journal of Medicinal Chemistry 46 (2011) 3058e3065
3063
3.2. General synthetic procedure for compounds 1aei [39]
An oven-dried flask was purged with argon while hot, then
allowed to cool down to room temperature under argon and
charged with dry THF (2 mL/mmol of starting indole) and diiso-
propylamine (4.0 equivalents). The solution was cooled to ꢀ78 ꢃC in
a dry ice/isopropanol bath and n-butyl lithium, 1.6 M solution in
hexanes (4.0 equivalents), was added to give a bright yellow solu-
tion. The mixture was stirred for 30 min below ꢀ60 ꢃC whereupon
a solution of picoline (4.0 equivalents) in THF (2 mL/mmol of
starting indole) was added to give a yellow colored mixture. During
a further 30 min, the bath temperature was allowed to rise to
approx. 0 ꢃC during which time the reaction mixture became an
orange suspension. The temperature was then kept at 0 ꢃC (using an
ice/water bath) for an additional 30 min. At this point, a solution of
the indole-2-carbonyl compound (1.0 equivalent) in THF (4 mL/
mmol) was added dropwise over the flask wall. The reaction
mixture darkened (dark purple to dark brown color in most cases)
and was then allowed to attain room temperature overnight. At
18 h after the addition of the indole component, the reaction was
quenched with saturated aqueous ammonium chloride solution
(10 mL/mmol of starting compound) and stirred for several
minutes. Upon partition between ethyl acetate and water, the
aqueous phase was extracted twice with ethyl acetate and the
combined organic layers were dried over MgSO4. The title products
were purified by chromatography eluting with a gradient of ethyl
acetate in cyclohexane (the products typically eluted with 70e80%
of ethyl acetate). The products were then precipitated by concen-
tration of the pooled column fractions combined with the addition
of cyclohexane. The precipitates formed were filtered off, washed
twice with cyclohexane and dried at 40 ꢃC in vacuo to yield
analytically pure samples.
Scheme 1. (i), LDA, THF/hexane, ꢀ78 to 0 ꢃC, 1 h; (ii) THF/hexane, 0 ꢃC to rt, 16e24 h;
(iii), H2NNH2, KOH, (CH2OH)2, mW, 1 h then H2O/NH4Cl; (iv) HCO2NH4, Pd black, MeOH,
rt, 3 days; (v) SOCl2,
D
, 15 min; (vi) 3-aminopyridine, DIPEA, THF, 0 ꢃC / rt, 30 min.
any cellular activity on TDO (Table 2), suggesting some selectivity
in this series.
In conclusion, a virtual screening strategy combining various
filters including high-throughput docking was used to search for
new IDO inhibitors. From the 39 final compounds identified and
assayed, six derivatives displayed an inhibitory potency >30% at
a concentration of 100
mM, two of them possessing more than
50% IDO inhibition at this concentration. Detailed kinetics
revealed an uncompetitive inhibition profile for the best inhib-
itor 1a. Its binding mode inside the IDO binding cleft was eval-
uated by means of docking and revealed essential features
responsible for the IDO inhibition potency in this series.
Preliminary SARs around 1a corroborated this putative binding
orientation and support the interest in this series for further
drug design effort.
3.3. Biology
3.3.1. Protein expression and purification of recombinant human
IDO
The coding region for human IDO (Ala2-Gly403) was cloned into
a derivative of plasmid pET9 (Novagen). The recombinant plasmid,
pETIDO, encodes a histidine tag at the N-terminus of IDO. Bacterial
strain BL21 AI (Invitrogen) was used for overexpression of IDO and
transformed with the pETIDO plasmid. The transformed cells were
grown on a rotary shaker at 37 ꢃC and 220 rpm to an OD600 of 1.2,
3. Experimental section
3.1. Chemistry
Details of synthetic procedures used to prepare compounds
1aei, 2, 3 and 4 as well as structural and analytical characteriza-
in LB medium supplemented with 25
mg/mL kanamycin, 50
mg/mL
ꢀ ꢁ
tions can be found in Dolusic et al. [38].
L-tryptophan and 10 mM bovine hemin (Sigma). The culture was
Table 2
Biological activity of the indole derivatives 1aei, 2, 3, 4.
Compound
R
R0
IDO IC50
M)
IDO cell assay
inhibition %
TDO cell assay
inhibition %
Cell viability
@ 20 M (%)
(
m
m
@ 20
m
Ma
@ 20 m
Ma
1MT
1a
1b
1c
1d
1e
1f
1g
1h
1i
100 ꢄ 20
65 ꢄ 7
153 ꢄ 23
36.0 ꢄ 0.1
24.6 ꢄ 0.1
87 ꢄ 11
49 ꢄ 1
NI
34 ꢄ 2
92 ꢄ 15
NI
NI
94 ꢄ 2
20
13
12
24
24
21
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
NI
No toxicity
No toxicity
80
No toxicity
90
No toxicity
No toxicity
H
4-F
5-F
5-Cl
5-Me
5-MeO
5-NO2
H
H
H
H
H
H
H
H
Me
Me
90
No toxicity
5-F
95
90
90
75
2
3
4
NI ¼ no inhibition.
a
The indicated values are the result of two independent determinations; variation between experiments is no more than ꢄ5%.

ꢀ ꢁ
E. Dolusic et al. / European Journal of Medicinal Chemistry 46 (2011) 3058e3065
3064
cooled in a water/ice bath and supplemented again with 50
m
g/mL
of tryptophan was degraded in the absence of inhibitor, and an
equimolar amount of kynurenine was produced. The percentages of
inhibition of tryptophan degradation and kynurenine production
by the compounds were calculated in reference to this maximal
activity. The initial wells containing the cells in the remaining
L
-tryptophan and 10 M bovine hemin. The expression of Histagged
m
IDO was induced by the addition of 1% (w/v) arabinose. Induced
cells were grown at 20 ꢃC and 60 rpm for 20 h. Cells (1 L culture)
were collected by centrifugation, resuspended in 40 mL of 25 mM
Mes, 150 mM KCl, 10 mM imidazole, and protease inhibitors
(complete EDTA free, Roche Applied Science) (pH 6.5), and dis-
rupted with a French press. The extract was clarified by centrifu-
volume of 50
MTT assay. To that end, 50
with 10% FCS and amino acids) were added to the wells together
with 50 l of SDS/
l of MTT. After 3e4 h of incubation at 37 ꢃC, 100
ml were used to estimate cell viability in a classical
m
l of culture medium (Iscove medium
gation and filtration on a 0.22
mm filter. The enzyme was purified by
m
m
IMAC using Ni2þ as a ligand and an IMAC HITRAP column (5 mL, GE
Healthcare). Briefly, the extract was loaded on the column with
25 mM Mes, 150 mM KCl, and 10 mM imidazole (pH 6.5). The
column was washed with 50 mL of the same buffer with the
imidazole concentration adjusted to 100 mM. Finally, the protein
was eluted with 25 mM Mes, 150 mM KCl, and 50 mM EDTA (pH
6.5). The buffer was then exchanged into 25 mM Mes and 150 mM
KCl (pH 6.5) using a HITRAP desalting column (GE Healthcare). The
purity of the enzyme was estimated to be >95% based on the SDS-
PAGE gel and Coomassie blue staining. The ratio of absorbance at
404 nm to that at 280 nm of the protein was around 1.9.
DMF were added to dissolve the crystals of formazan blue and the
absorbance at 570 nm/650 nm was measured after overnight
incubation at 37 ꢃC.
3.3.4. Steady states kinetics and inhibition studies
Human IDO with an N-terminal His tag was expressed in
Escherichia coli and purified to homogeneity as described by Oda
[41]. The enzymatic reactions were conducted at 37 ꢃC for 12 min in
50 mM potassium phosphate pH 6.5, neutralized ascorbic acid
10 mM, methylene blue 5
-tryptophan and 8.5 nM IDO. The reaction was initiated by addition
of IDO. At various intervals (2 min), aliquots (200 L) were removed,
mixed with 40 l of trichloroacetic acid (30% w/v), and incubated at
65 ꢃC for 10 min to achieve the conversion of N-formylkynurenine to
kynurenine. The final -tryptophan concentrations varied between
12.5 M and 100
mM, DMSO 5%, catalase 100 units per ml,
L
m
3.3.2. Enzymatic assay
The enzymatic inhibition assays were performed as described
by Takikawa et al. [40] with some modifications. Briefly, the reac-
m
L
tion mixture (200
(50 mM, pH 6.5) ascorbic acid (10 mM), methylene blue (5
purified recombinant IDO (43 M), -Trp (100 M), and DMSO
L). The inhibitors were serially diluted 10-fold from 1000 to
M in pure DMSO or, if not soluble at 1000 M, by four orders of
m
L) contained potassium phosphate buffer
m
mM. The concentration of L-tryptophan was deter-
mM),
mined using the absorption coefficient e280 ¼ 5500 Mꢀ1 cmꢀ1 and
m
L
m
the substrate was filtered through a membrane filter prior to each
(10
0.1
m
m
experiment (pore size, 0.22 mm). The spectrophotometer method to
m
measure kynurenine with the Ehrlich reagent was carried out
according to Takikawa [40] and the initial rates were calculated from
the absorbance increase at 480 nm (e480 ¼ 15820 Mꢀ1 cmꢀ1) [42].
Each experiment was performed at least three times. The initial rates
were fit to the MichaeliseMenten equation by nonlinear regression
analysis using Graphpad prism software. Apparent Km and Ki values
were determined by varying the concentrations of the substrate and
the inhibitors. The mode of inhibition was determined by plotting
[S]/v against [S] (Hanes plot) and 1/v against [I] (Dixon plot).
Spectroscopic measurements. Optical absorption spectra were
recorded on a Specord 50 photometer (Analytik Jena). All experi-
ments were performed in 50 mM TriseHCl buffer (pH 8.0) at 25 ꢃC.
magnitude from their highest soluble concentration. The reaction
was conducted at 37 ꢃC for 6 min and stopped by addition of 30%
(w/v) trichloroacetic acid (40
to kynurenine, the tubes were incubated at 37 ꢃC for 30 min, fol-
lowed by centrifugation at 20 000 g for 20 min. Finally, 150 L of
supernatant is added to 150 L of p-dimethylaminobenzaldehyde
mL). To convert N-formylkynurenine
m
m
(pDMAB) (2%, v/v) in acetic acid to generate a Schiff base with
kynurenine that was detected at a wavelength of 480 nm.
3.3.3. Cellular tests
Cell Line. A plasmid construct encoding murine IDO was trans-
fected into mouse mastocytoma line P815B. Clone P185B-mIDO
clone 6 [1], which overexpresses IDO, was selected and used for the
cellular assay. Mouse IDO shares a 62% sequence identity with
human IDO, and the active site residues are 100% conserved. For the
assays evaluating inhibition of mouse TDO, a plasmid construct
encoding murine TDO was transfected into mouse mastocytoma
cell line P815B [1]. Clone P815B-mTDO clone 12, which over-
expresses TDO, was selected and used for the cellular assay. Mouse
TDO shares 88% sequence identity with human TDO, and the active
site residues are 100% conserved.
3.4. Molecular modeling
Molecular modeling studies were carried out on a Linux work-
station. The compounds were downloaded from the ZINC website
(http://zinc.docking.org) and transformed into a UNITY hit list by
mean of SYBYL (version 8.0) [30]. Docking was performed using the
3D-coordinates of IDO (pdb code 2DOT) with the help of the
automated GOLD program [36] (active site definition: residues
within 7 Å around phenylimidazole) using parameters especially
designed for virtual screening experiments (7e8 fold acceleration).
The hits that were obtained after ranking using GOLDSCORE and
CSCORE were re-docked for confirmatory evaluation and their
geometry subsequently optimized using the MINIMIZE module. The
minimization process uses the Powell method with the Tripos force
field (dielectric constant 1r) to reach a final convergence of
Assay. The assay was performed in 96-well flat bottom plates
seeded with 2 ꢅ 105 cells in a final volume of 200
ml. To determine
whether compounds were significant IDO or TDO inhibitors, the
cells were incubated overnight at 37 ꢃC in HBSS (Hanks Balanced
Salt Solution, Invitrogen) supplemented with 80
and 20 M of the compound. The plates were then centrifuged
10 min at 300 g, and 150 l of the supernatant were collected. The
mM L-Tryptophan
m
m
0.01 kcal molꢀ1
.
supernatant was analyzed by HPLC to measure the concentration of
tryptophan and kynurenine, based on the retention time and the
UV absorption (280 nm for tryptophan, 360 nm for kynurenine). For
Acknowledgments
the HPLC analysis, 50 ml of supernatant were mixed with 500 ml
acetonitrile to precipitate the proteins. After centrifugation, the
supernatant was collected, concentrated on a speedvac, resus-
This work is supported by the FNRS (Belgium) and the Walloon
region (CANTOL grant no 5678). R.F. is greatly indebted to the
Belgian “Fonds de la Recherche ScientifiqueeFNRS” for the award of
a postdoctoral research grant.
pended in a final volume of 100
ml water and injected in the HPLC
(C18 column). In those conditions, about 50% of the initial amount

ꢀ ꢁ
E. Dolusic et al. / European Journal of Medicinal Chemistry 46 (2011) 3058e3065
3065
[22] G. Carr, M.K. Chung, A.G. Mauk, R.J. Andersen, Synthesis of indoleamine 2,3-
dioxygenase inhibitory analogues of the sponge alkaloid exiguamine A, J. Med.
Chem. 51 (2008) 2634e2637.
[23] S. Kumar, W.P. Malachowski, J.B. DuHadaway, J.M. LaLonde, P.J. Carroll,
D. Jaller, R. Metz, G.C. Prendergast, A.J. Muller, Indoleamine 2,3-dioxygenase is
the anticancer target for a novel series of potent naphthoquinone-based
inhibitors, J. Med. Chem. 51 (2008) 1706e1718.
[24] M. Sono, S.G. Cady, Enzyme kinetic and spectroscopic studies of inhibitor and
effector interactions with indoleamine 2,3-dioxygenase. 1. Norharman and 4-
phenylimidazole binding to the enzyme as inhibitors and heme ligands,
Biochemistry 28 (1989) 5392e5399.
[25] H. Sugimoto, S. Oda, T. Otsuki, T. Hino, T. Yoshida, Y. Shiro, Crystal structure of
human indoleamine 2,3-dioxygenase: catalytic mechanism of O2 incorpora-
tion by a heme-containing dioxygenase, Proc. Natl. Acad. Sci. U S A 103 (2006)
2611e2616.
[26] S. Kumar, D. Jaller, B. Patel, J.M. LaLonde, J.B. DuHadaway, W.P. Malachowski,
G.C. Prendergast, A.J. Muller, Structure based development of phenyl-
imidazole-derived inhibitors of indoleamine 2,3-dioxygenase, J. Med. Chem.
51 (2008) 4968e4977.
[27] G. Carr, W. Tay, H. Bottriell, S.K. Andersen, A.G. Mauk, R.J. Andersen, Plec-
tosphaeroic acids A, B, and C, indoleamine 2,3-dioxygenase inhibitors
produced in culture by a marine isolate of the fungus Plectosphaerella cucu-
merina, Org. Lett. 11 (2009) 2996e2999.
[28] E.W. Yue, B. Douty, B. Wayland, M. Bower, X. Liu, L. Leffet, Q. Wang,
K.J. Bowman, M.J. Hansbury, C. Liu, M. Wei, Y. Li, R. Wynn, T.C. Burn,
H.K. Koblish, J.S. Fridman, B. Metcalf, P.A. Scherle, A.P. Combs, Discovery of
potent competitive inhibitors of indoleamine 2,3-dioxygenase with in vivo
pharmacodynamic activity and efficacy in a mouse melanoma model, J. Med.
Chem. 52 (2009) 7364e7367.
[29] U.F. Rohrig, L. Awad, A. Grosdidier, P. Larrieu, V. Stroobant, D. Colau,
V. Cerundolo, A.J. Simpson, P. Vogel, B.J. Van den Eynde, V. Zoete, O. Michielin,
Rational design of indoleamine 2,3-dioxygenase inhibitors, J. Med. Chem. 53
(2010) 1172e1189.
References
[1] C. Uyttenhove, L. Pilotte, I. Theate, V. Stroobant, D. Colau, N. Parmentier,
T. Boon, B.J. Van den Eynde, Evidence for a tumoral immune resistance
mechanism based on tryptophan degradation by indoleamine 2,3-dioxyge-
nase, Nat. Med. 9 (2003) 1269e1274.
[2] S.A. Rafice, N. Chauhan, I. Efimov, J. Basran, E.L. Raven, Oxidation of L-tryp-
tophan in biology: a comparison between tryptophan 2,3-dioxygenase and
indoleamine 2,3-dioxygenase, Biochem. Soc. Trans. 37 (2009) 408e412.
[3] M. Sono, M.P. Roach, E.D. Coulter, J.H. Dawson, Heme-containing oxygenases,
Chem. Rev. 96 (1996) 2841e2888.
[4] P. Hwu, M.X. Du, R. Lapointe, M. Do, M.W. Taylor, H.A. Young, Indoleamine
2,3-dioxygenase production by human dendritic cells results in the inhibition
of T cell proliferation, J. Immunol. 164 (2000) 3596e3599.
[5] A.L. Mellor, D.H. Munn, IDO expression by dendritic cells: tolerance and
tryptophan catabolism, Nat. Rev. Immunol. 4 (2004) 762e774.
[6] D.H. Munn, M. Zhou, J.T. Attwood, I. Bondarev, S.J. Conway, B. Marshall,
C. Brown, A.L. Mellor, Prevention of allogeneic fetal rejection by tryptophan
catabolism, Science 281 (1998) 1191e1193.
[7] Y. Kudo, C.A. Boyd, Human placental indoleamine 2,3-dioxygenase: cellular
localization and characterization of an enzyme preventing fetal rejection,
Biochim. Biophys. Acta 1500 (2000) 119e124.
[8] G. Brandacher, A. Perathoner, R. Ladurner, S. Schneeberger, P. Obrist,
C. Winkler, E.R. Werner, G. Werner-Felmayer, H.G. Weiss, G. Gobel,
R. Margreiter, A. Konigsrainer, D. Fuchs, A. Amberger, Prognostic value of
indoleamine 2,3-dioxygenase expression in colorectal cancer: effect on
tumor-infiltrating T cells, Clin. Cancer Res. 12 (2006) 1144e1151.
[9] A. Okamoto, T. Nikaido, K. Ochiai, S. Takakura, M. Saito, Y. Aoki, N. Ishii,
N. Yanaihara, K. Yamada, O. Takikawa, R. Kawaguchi, S. Isonishi, T. Tanaka,
M. Urashima, Indoleamine 2,3-dioxygenase serves as a marker of poor prog-
nosis in gene expression profiles of serous ovarian cancer cells, Clin. Cancer
Res. 11 (2005) 6030e6039.
[10] B.S. Leung, L.E. Stout, E.G. Shaskan, R.M. Thompson, Differential induction of
indoleamine-2,3-dioxygenase (IDO) by interferon-gamma in human gyneco-
logic cancer cells, Cancer Lett. 66 (1992) 77e81.
[11] A.J. Muller, J.B. DuHadaway, P.S. Donover, E. Sutanto-Ward, G.C. Prendergast,
Inhibition of indoleamine 2,3-dioxygenase, an immunoregulatory target of
the cancer suppression gene Bin1, potentiates cancer chemotherapy, Nat.
Med. 11 (2005) 312e319.
[30] Sybyl 8.0, Tripos Inc., in, 1699 South Hanley Rd., St. Louis, Missouri, 63144,
USA.
[31] W.L. Delano, The PyMOL Molecular Graphics System on World Wide
Web.http://pymol.sourceforge.net/ (accessed 08.02.11).
[32] B.K. Shoichet, Virtual screening of chemical libraries, Nature 432 (2004)
862e865.
[33] J.J. Irwin, B.K. Shoichet, ZINCea free database of commercially available
compounds for virtual screening, J. Chem. Inf. Model. 45 (2005) 177e182.
[34] S.J. Teague, A.M. Davis, P.D. Leeson, T. Oprea, The design of leadlike combi-
natorial libraries, Angew. Chem. Int. Ed. Engl. 38 (1999) 3743e3748.
[35] R.D. Clark, A. Strizhev, J.M. Leonard, J.F. Blake, J.B. Matthew, Consensus scoring
for ligand/protein interactions, J. Mol. Graph Model. 20 (2002) 281e295.
[36] G. Jones, P. Willett, R.C. Glen, A.R. Leach, R. Taylor, Development and valida-
tion of a genetic algorithm for flexible docking, J. Mol. Biol. 267 (1997)
727e748.
[12] A.J. Muller, W.P. Malachowski, G.C. Prendergast, Indoleamine 2,3-dioxygenase
in cancer: targeting pathological immune tolerance with small-molecule
inhibitors, Expert Opin. Ther. Targets 9 (2005) 831e849.
[13] A.J. Muller, G.C. Prendergast, Marrying immunotherapy with chemotherapy:
why say IDO? Cancer Res. 65 (2005) 8065e8068.
[14] A.J. Muller, G.C. Prendergast, Indoleamine 2,3-dioxygenase in immune
suppression and cancer, Curr. Cancer Drug Targets 7 (2007) 31e40.
[15] A. Macchiarulo, E. Camaioni, R. Nuti, R. Pellicciari, Highlights at the gate of
tryptophan catabolism: a review on the mechanisms of activation and regu-
lation of indoleamine 2,3-dioxygenase (IDO), a novel target in cancer disease,
Amino Acids 37 (2009) 219e229.
[16] S.G. Cady, M. Sono, 1-Methyl-dl-tryptophan, beta-(3-benzofuranyl)-dl-alanine
(the oxygen analog of tryptophan), and beta-[3-benzo(b)thienyl]-dl-alanine
(the sulfur analog of tryptophan) are competitive inhibitors for indoleamine
2,3-dioxygenase, Arch. Biochem. Biophys. 291 (1991) 326e333.
[17] N. Eguchi, Y. Watanabe, K. Kawanishi, Y. Hashimoto, O. Hayaishi, Inhibition of
indoleamine 2,3-dioxygenase and tryptophan 2,3-dioxygenase by beta-car-
boline and indole derivatives, Arch. Biochem. Biophys. 232 (1984) 602e609.
[18] H.C. Brastianos, E. Vottero, B.O. Patrick, R. Van Soest, T. Matainaho, A.G. Mauk,
R.J. Andersen, Exiguamine A, an indoleamine-2,3-dioxygenase (IDO) inhibitor
isolated from the marine sponge Neopetrosia exigua, J. Am. Chem. Soc. 128
(2006) 16046e16047.
[19] A. Pereira, E. Vottero, M. Roberge, A.G. Mauk, R.J. Andersen, Indoleamine 2,3-
dioxygenase inhibitors from the Northeastern Pacific Marine Hydroid Garveia
annulata, J. Nat. Prod. 69 (2006) 1496e1499.
[20] T. Banerjee, J.B. Duhadaway, P. Gaspari, E. Sutanto-Ward, D.H. Munn,
A.L. Mellor, W.P. Malachowski, G.C. Prendergast, A.J. Muller, A key in vivo
antitumor mechanism of action of natural product-based brassinins is inhi-
bition of indoleamine 2,3-dioxygenase, Oncogene 27 (2008) 2851e2857.
[21] P. Gaspari, T. Banerjee, W.P. Malachowski, A.J. Muller, G.C. Prendergast, J. DuHadaway,
S. Bennett, A.M. Donovan, Structure-activity study of brassinin derivatives as indole-
amine 2,3-dioxygenase inhibitors, J. Med. Chem. 49 (2006) 684e692.
[37] C. Lu, Y. Lin, S.R. Yeh, Spectroscopic studies of ligand and substrate binding to
human indoleamine 2,3-dioxygenase, Biochemistry 49 (2010) 5028e5034.
ꢀ ꢁ
[38] E. Dolusic, P. Larrieu, S. Blanc, F. Sapunaric, B. Norberg, L. Moineaux, D. Colette,
V. Stroobant, L. Pilotte, D. Colau, T. Ferain, G. Fraser, M. Galeni, J.M. Frere, B.
Masereel, B. Van den Eynde, J. Wouters, R. Frederick, Indol-2-yl ethanones as
novel indoleamine 2,3-dioxygenase (IDO) inhibitors, Bioorg. Med. Chem. 19
(2011) 1550e1561.
[39] R.G. Sundberg, J.G. Luis, R.L. Parton, S. Schreiber, P.C. Srinivasan, P. Lamb,
P. Forcier, R.B. Bryan, Chloroacetamide photocyclization of indole derivatives.
synthesis, stereochemistry, and crystal structure of 3,7-methano-3-azacy-
cloundecino5[,4-blindole (deethylquebrachamine) derivatives, J. Org. Chem.
23 (1978) 4859e4865.
[40] O. Takikawa, T. Kuroiwa, F. Yamazaki, R. Kido, Mechanism of interferon-
gamma action. Characterization of indoleamine 2,3-dioxygenase in cultured
human cells induced by interferon-gamma and evaluation of the enzyme-
mediated tryptophan degradation in its anticellular activity, J. Biol. Chem. 263
(1988) 2041e2048.
[41] S. Oda, H. Sugimoto, T. Yoshida, Y. Shiro, Crystallization and preliminary
crystallographic studies of human indoleamine 2,3-dioxygenase, Acta Crys-
tallogr. Sect. F Struct. Biol. Cryst Commun. 62 (2006) 221e223.
[42] E. Alegre, A.S. Lopez, A. Gonzalez, Tryptophan metabolites interfere with the
Ehrlich reaction used for the measurement of kynurenine, Anal. Biochem. 339
(2005) 188e189.