Discovery of potent competitive inhibitors of indoleamine 2,3-dioxygenase with in vivo pharmacodynamic activity and efficacy in a mouse melanoma model

Article abstract of DOI:10.1021/jm900518f

A hydroxyamidine chemotype has been discovered as a key pharmacophore in novel inhibitors of indoleamine 2,3-dioxygenase (IDO). Optimization led to the identification of 5l, which is a potent (HeLa IC50=19 nM) competitive inhibitor of IDO. Testing of 5l i

Full text of DOI:10.1021/jm900518f

7364 J. Med. Chem. 2009, 52, 7364–7367
DOI: 10.1021/jm900518f
IDO in tumor cells, as well as in the dendritic cells that localize
to the tumor draining lymph nodes, has been shown to be an
independent prognostic variable for reduced overall survival
in patients with in a wide variety of tumors, including ovarian,
colorectal, and pancreatic cancers, melanoma, and hematolo-
gical malignancies.7 Further, it was shown that 1-methyltryp-
tophan (1-MT), a weak competitive inhibitor (Ki =34 μM)
of the enzyme, increases the efficacy without increased toxicity
of chemotherapeutic agents, such as paclitaxel, gemcitabine,
and cyclophosphamide, in several mouse tumor models.9 As
a single agent, 1-MT impairs the growth of granulocyte-
macrophage colony-stimulating factor (GM-CSF) expressing
B16 melanomas.8 These effects were not observed in T cell
deficient mice, suggesting that the results were a consequence
of pharmacological inhibition of IDO mediated immunosup-
pression within the tumor microenvironment.
Discovery of Potent Competitive Inhibitors of
Indoleamine 2,3-Dioxygenase with in Vivo
Pharmacodynamic Activity and Efficacy in a
Mouse Melanoma Model
Eddy W. Yue, Brent Douty, Brian Wayland, Michael Bower,
Xiangdong Liu, Lynn Leffet, Qian Wang, Kevin J. Bowman,
Michael J. Hansbury, Changnian Liu, Min Wei, Yanlong Li,
Richard Wynn, Timothy C. Burn, Holly K. Koblish,
Jordan S. Fridman, Brian Metcalf, Peggy A. Scherle, and
Andrew P. Combs*
Incyte Corporation, Experimental Station, Route 141 and
Henry Clay Road, Wilmington, Delaware 19880
Received April 23, 2009
To date, only a few structural classes are known to be IDO
inhibitors. Most of these compounds have been shown to be
noncompetitive inhibitors and likely are involved in redox
reactions with the iron bound to the heme of IDO.8-12 Nearly
all of the reported competitive IDO inhibitors are structurally
related tryptophan analogues, such as1-MT.13-16 The modest
potencies and poor physical properties of these indole-based
compounds led us to search for a new chemotype. A screening
effort was initiated with the objective of identifying a novel
lead compound that could be optimized to afford potent and
selective inhibitors of IDO with suitable permeability and
metabolic stability to allow evaluation of the functional
consequences of IDO inhibition in vivo. Herein, we report
the discovery and optimization of a novel structural series of
potent and competitive inhibitors of IDO that strongly sup-
port the role of IDO in tumor progression and its pharmaco-
logical inhibition as a means to treat IDO overexpressing
malignancies.
Abstract: A hydroxyamidine chemotype has been discovered as a
key pharmacophore in novel inhibitors of indoleamine 2,3-dioxy-
genase (IDO). Optimization led to the identification of 5l, which is a
potent (HeLa IC50 = 19 nM) competitive inhibitor of IDO. Testing
of 5l in mice demonstrated pharmacodynamic inhibition of IDO, as
measured by decreased kynurenine levels (>50%) in plasma and
dose dependent efficacy in mice bearing GM-CSF-secreting B16
melanoma tumors.
Indoleamine 2,3-dioxygenase (IDOa) and tryptophan 2,3-
dioxygenase (TDO) are the two key heme-containing dioxy-
genases that catalyze the rate limiting step in the cata-
bolism of the essential amino acid tryptophan to N-formyl-
kynurenine by oxidative cleavage of the indole 2,3 double
bond.1 This reaction is the initial step in the de novo biosyn-
thetic route, known as the kynurenine pathway, which leads to
a series of biologically active metabolites, including neurotran-
smitters serotonin and melatonin, excitoxin quinolinic acid,
N-methyl-D-aspartate (NMDA) receptor antagonist kynure-
nic acid, and ultimately the production of nicotinamide ade-
nine dinucleotide (NAD). IDO is expressed in various tissues
throughout the body but predominately in cells within the
immune system where it is specifically induced in dendritic cells
and macrophages at sites of inflammation by cytokines, such
as interferon γ (IFN-γ).1 The overexpression of IDO has been
implicated in a variety of diseases, including cancer, neuro-
degenerative disorders (Alzheimer’s disease), age-related cat-
aract, and HIV encephalitis.2-6 In contrast, TDO is almost
entirely located in the liver where it maintains proper trypto-
phan balance in response to dietary intake.
Figure 1. IDO lead 1.
High throughput screening (HTS) of Incyte’s corporate
collection identified 4-amino-1,2,5-oxadiazole-3-carboximida-
mide (1) as a micromolar IDO enzyme inhibitor (Figure 1).
Enzyme kinetic assays measuring the conversion of tryptophan
to N-formylkynurenine demonstrated that 1 is a competitive
inhibitor (Ki=1 μM). Direct binding to the heme active site was
Recently, IDO has been shown to play an important role in
the process of immune evasion by tumors.1 IDO mediated
depletion of local tryptophan levels, and production of toxic
tryptophan metabolites results in suppression of T cell activa-
tion and induction of T cell apoptosis. The overexpression of
confirmed by absorption spectroscopy of the ferrous (Fe2+
)
form of IDO. Changes in the strength and maximum wave-
length for the Soret peak suggest direct binding to the heme
moiety. Testing in a human HeLa cell line stimulated with IFN-
γ to overexpress IDO provided further evidence that 1 is a
potent inhibitor (IC50 = 1 μM) of IDO. Counterscreening
against TDO (IC50 = 10 μM) demonstrated this class of
inhibitor was selective for IDO. The combination of low
molecular weight (<250), good ligand efficiency (LE =
0.49), low micromolar cellular activity, TDO selectivity, and
high caco-2 permeability (39 Â 10-6 cm/s) of lead 1 provided an
excellent starting point for our discovery program.
*To whom correspondences should be address. Phone: 302-498-6832.
Fax: 302-425-2750. E-mail acombs@incyte.com.
a Abbreviations: IDO, indoleamine 2,3-dioxygenase; TDO, trypto-
phan 2,3-dioxygenase; NAD, nicotinamide adenine dinucleotide; IFN-
γ, interferon γ; 1-MT, 1-methyltryptophan; NMDA, N-methyl-D-aspar-
tate; GM-CSF, granulocyte-macrophage colony-stimulating factor; Kyn,
kynurenine; HTS, high throughput screening; SAR, structure-activity
relationships; LE, ligand efficiency; TGC, tumor growth control.
r
pubs.acs.org/jmc
Published on Web 06/09/2009
2009 American Chemical Society

Letter
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7365
Table 2. SAR of the Hydroxyamidine
Scheme 1. Synthesis of 4-Amino-1,2,5-oxadiazole-3-carboxi-
midamides (5)a
IC50 (nM)a
compd
R
IDO
90
HeLa
20
5c
6
N-OH
N-OMe
NH
>20000
>20000
1400
>2000
>2000
>2000
>2000
>2000
7
8
N-NH2
O
9
10
>20000
>20000
a Reagents: (a) NaNO2, 2 N HCl, NH2OH HCl; then 10 N NaOH, Δ,
3
S
2 h (see ref 18); (b) NaNO2, HCl, H2O, 0 °C, 1.5 h, 30-40% (see ref 19);
(c) aniline, Et3N, EtOH, 0-25 °C, 30-60%.
a IC50 values are the mean of at least two independent assays.
is observed and enabled optimization by iterative cycles of
synthesis and testing.
Table 1. SAR of Phenyl Ring of 4-Amino-1,2,5-oxadiazole-3-carbox-
imidamides (5)
Early structure-activity relationships (SAR) demonstrated
that phenyl derivatives 5 were significantly more potent than
the original benzyl lead 1 and other alkylamine derivatives.
Selected phenyl analogues and their SAR are shown in
Table 1. A strong preference for meta-substitution was de-
monstrated by the 50-fold improvement in cellular potency
when m-chloro 5c was compared to o-chloro 5b and p-chloro
5d. A variety of other ortho- and para-substituents larger than
fluoro were also significantly less active (data not shown).
Further optimization revealed that meta-halogens and small
meta-alkyl substituents provided the most potent IDO inhi-
bitors in this series (Br, Cl >Me, Et>isopropyl, F >H, OMe
tert-butyl), including m-chloro 5c and m-bromo 5g (HeLa
IC50 = 19 and 17 nM, respectively). Addition of a p-fluoro
substituent to these derivatives was tolerated and gave 5l and
5m, which showed improved in vitro human intrinsic clear-
ances (5c =1.1, 5l = 0.57 L/h/kg) without loss in IDO cel-
lular potency. 5l was chosen for further in vitro and in vivo
studies because of its improved physical chemical properties
compared to 5m. Counterscreening against TDO showed 5l
was inactive (IC50>10 μM) in our TDO enzyme and cellular
assays. Membrane permeability remained excellent for all com-
pounds in the series, for example, 5l caco-2 permeability= 36 Â
10-6 cm/s. We speculate the reason compounds within this
chemotype have high permeability is due to a combination of
their low molecular weight and low energy conformation
containing two internal hydrogen bonds.
SAR of the unique hydroxylamidine functional group demon-
strated its important role in binding IDO. All modifications of
the hydroxyamidine pharmacophore gave significantly less
potent derivatives (Table 2). Amide 9 and thioamide 10 were
devoid of activity. Elimination of only the hydroxyl group in
amidine 7 or methylation of the hydroxyl group in methoxy-
amidine 6 revealed the requirement for a free hydroxyl group
to bind IDO. Only the aminoamidine 8 showed measurable
activity, albeit 300-fold less active in the HeLa cell assay. In
accordance with the SAR established here, the known binding
of compounds containing hydroxyamidines to iron hemes,
and our internal enzyme kinetics and spectroscopic results, we
hypothesize thattheseIDO inhibitors form a dative bond with
the heme iron in the ferrous state through the oxygen of the
hydroxyamidine.
IC50 (nM)a
compd
R
IDO
HeLa
490
5a
5b
5c
5d
5e
5f
H
3200
6500
86
2-Cl
3-Cl
4-Cl
3-F
3-Br
1500
19
6000
500
73
2400
270
17
5g
5h
5i
3-Me
3-Et
550
430
2100
25000
4400
67
120
54
3-isopropyl
3-tert-butyl
3-OMe
225
6800
970
19 (46)b
12
5j
5k
5l
3-Cl,4-F
3-Br,4-F
5m
59
a IC50 values are the mean of at least three independent assays.
b Murine cellular B16 assay IC50 (nM).
4-Amino-1,2,5-oxadiazole-3-carboximidamides
5
were
synthesized in three steps (Scheme 1). Treatment of malono-
nitrile 2 with hydroxylamine, sodium nitrite, and hydrochloric
acid gave hydroxyamidine 3,17 which was diazotized under
acidic conditions18 to provide the hydroximoyl chloride 4.19
Coupling of 4 to a variety of benzylamines and anilines gave
moderate yields of analogues 1 and 5, respectively.
A diverse library of carboximidamides were synthesized
and tested for their ability to inhibit human IDO in the
standard enzymatic assay and a HeLa cellular assay measur-
ing kynurenine formation spectrophotometrically (Table 1).
We did observe that most compounds are slightly more potent
in the more physiologically relevant HeLa cell-based assays
than in the enzyme assays. This may be due to the complexity
of the enzyme assay which utilizes a methylene blue ascorbate
regeneration system to maintain IDO in the active reduced
form or unidentified differences between the recombinant
IDO used in the enzyme assay and native IDO in HeLa cells.
Nevertheless, a very good correlation between the two assays
A computer model of 5lbound toIDO wasgenerated on the
basis of the published crystal structure of IDO and established

7366 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23
Yue et al.
Table 3. 5l Suppresses Kynurenine Generation in Vivoa
group
plasma Kyn (μM)
Kyn (% inh)
5l (μM)
control
2.5 h
1.1 ( 0.17
0.45 ( 0.2
60
2.5 ( 0.6
a Data are averaged from three individual mice per treatment per time
point.
Table 4. 5l Inhibits B16-GMCSF Melanoma Growth in C57BL/6
Micea
5l dose (mg/kg b.i.d.)
TGC (day 20) (%)
5l at 1 h (μM)
25
50
75
13
8.9 ( 1.2
12 ( 1.2
29 ( 13
36b
50c
a Data provided are from one of two experiments with similar results.
b p < 0.05 vs vehicle control. c p < 0.01 vs vehicle control.
growth was correlated with increasing exposures of 5l in
plasma (Table 4).
Figure 2. Computer generated model of 5l bound to hIDO using
the Molecular Operating Environment (MOE) from Chemical
Computing Group. The coordination of the oxygen of the hydro-
xyamidine to the iron (red ball) of the heme (ball and stick) is shown
as a dotted yellow line.
These levels exceed those necessary to reduce kynurenine in
plasma by >50%. A maximal effect of 50% tumor growth
control (TGC) was seen when 5l was administered subcuta-
neously at 75 mg/kg b.i.d. (p < 0.01). For comparison, 1-MT
demonstrated ∼45% TGC when administered as subcuta-
neous pellets (5 mg/day).23 Taken together, these data suggest
that this novel hydroxyamidine structural class, represented
by 5l, can inhibit IDO in vivo, as measured by pharmaco-
dynamic reduction in kynurenine levels and tumor growth
suppression.
A competitive hydroxyamidine containing IDO inhibitor
1 was identified by high throughput screening. Modification
of the scaffold afforded a novel and potent IDO inhibitor
5l with >100-fold improvement in enzyme and cellular
potency. A computer model of 5l bound to IDO through
the oxygen of the hydroxyamidine to the iron of the heme in
the ferrous state was consistent with enzyme kinetics and
SAR for the series. Testing in mice demonstrated that 5l
decreased kynurenine levels by >50% in plasma and in-
hibited B16-GM-CSF tumor growth in a dose dependent
fashion. These data provide convincing support for pharma-
cological inhibition of IDO as a means to disable tumor-
induced immune tolerance and to treat a variety of cancers
where IDO is overexpressed.
SAR for the series (Figure 2).20,21 The initial modeling of 5l to
IDO assumed that the oxygen of the hydroxyamidine coordi-
nates with the heme iron in the catalytically active, ferrous
(Fe2+) state. Conformational analysis of 5l in solution re-
vealed that the compound exists in a low energy cis-conforma-
tion which is consistent with published crystal structures of
similar hydroxyamidines that also form an intramolecular
hydrogenbondbetweenthe aniline nitrogenand theoxygen of
the hydroxyamidine.22
The phenyl ring was then placed deep into the hydrophobic
pocket where it is perfectly situated to extend a meta-sub-
stituent into a small hydrophobic channel. In this conforma-
tion, the phenyl ring fits tightly into the active site, providing
strong rationale for the lack of activity of compounds bearing
substituents larger than hydrogen or fluoro in the ortho- and
para-positions. The furazan moiety extends toward solvent,
making loose contacts with the enzyme. The amino substitu-
ent is situated in the vicinity of a propanoic acid group on the
heme ring where it may form a hydrogen bond.
IDO inhibitor 5l was selected for further evaluation in vivo.
The reduction of kynurenine levels in plasma was used as a
pharmacodynamic marker for inhibition of IDO activity.
Initial oral pharmacokinetic studies showed that 5l was
rapidly cleared (t1/2 < 0.5 h) and that oral administration
would not be a suitable dosing method for in vivo studies.
Therefore, a single subcutaneous 100 mg/kg dose of 5l was
administered to naive C57BL/6 mice. Blood was harvested
from individual mice over 8 h. Kynurenine and inhibitor 5l
concentrations were measured by LCMS. Reductions of
kynurenine levels by ∼50-60% were observed between 2
and 4 h, with maximum inhibition seen at 2.5 h (Table 3).
The measured plasma exposure (2.5 μM) of 5l during this
period exceeded our calculated mouse protein binding ad-
justed B16 cellular IC50 (PBadjIC50 = 1.0 μM, murine cellular
B16 IC50 = 46 nM). Notably, kynurenine levels increased
back to baseline after 4 h as 5l exposure levels decreased below
the mouse PBadjIC50 from 1.0 to 0.1 μM.
Acknowledgment. We thank Laurine Galya, Mei Li, James
Doughty, and Karl Blom for their analytical assistance. We
thank Mary Becker-Pasha and Mark Rupar for production of
the enzyme and development of the IDO and TDO enzyme
assays, respectively. We also thank Dr. Hyam Levitsky at
Johns Hopkins University School of Medicine for providing
us with the B16-GM-CSF cell line.
Supporting Information Available: Synthesis procedures with
supporting analytical data and assay protocols. This mate-
rial is available free of charge via the Internet at http://pubs.
acs.org.
References
(1) Munn, D. H.; Mellor, A. L. Indoleamine 2,3-dioxygenase and
tumor-induced tolerance. J. Clin. Invest. 2007, 117, 1147–1154.
(2) Takikawa, O. Biochemical and medical aspects of the indoleamine
2,3-dioxygenase-initiated L-tryptophan metabolism. Biochem. Bio-
phys. Res. Commun. 2005, 338, 12–19.
(3) Boasso, A.; Herbeuval, J.-P.; Hardy, A. W.; Anderson, S. A.;
Dolan, M. J.; Fuchs, D.; Shearer, G. M. HIV inhibits CD4+
T-cell proliferation by inducing indoleamine 2,3-dioxygenase in
plasmacytoid dendritic cells. Blood 2007, 109, 3351–3359.
The efficacy of 5l was tested in C57BL/6 mice bearing GM-
CSF-secretingB16 tumors, where1-MT has previously shown
activity.23 Tumors were allowed to grow until day 7 when
14 days of subcutaneous dosing of 5l at 25, 50, and 75 mg/kg
b.i.d. was initiated. Dose dependent inhibition of tumor

Letter
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 23 7367
(4) Potula, R.; Poluektova, L.; Knipe, B.; Chrastil, J.; Heilman, D.
Dou, H.; Takikawa, O.; Munn, D. H.; Gendelman, H. E.; Persidsky,
Y. Inhibition of indoleamine 2,3-dioxygenase (IDO) enhances elimi-
nation of virus-infected macrophages in an animal model of HIV-1
encephalitis. Blood 2005, 106, 2382–2390.
(5) Guillemin, G. J.; Brew, B. J.; Noonan, C. E.; Takikawa, O.;
Cullen, K. M. Indoleamine 2,3 dioxygenase and quinolinic acid
immunoreactivity in Alzheimer’s disease hippocampus. Neuropath.
Appl. Neurobiol. 2005, 31, 395–404.
(6) Fujigaki, H.; Saito, K. Inhibition of increased indoleamine 2,3-
dioxygenase activity exacerbates neuronal cell death in various
CNS disorders. Int. Congr. Ser. 2007, 1304, 314–323.
(7) Curti, A.; Trabanelli, S.; Salvestrini, V.; Baccarani, M.; Lemoli, R.
M. The role of indoleamine 2,3-dioxygenase in the induction of
immune tolerance: focus on hematology. Blood 2009, 113,
2394–2401.
(8) Peterson, A. C.; La Loggia, A. J.; Hamaker, L. K.; Arend, R. A.;
Fisette, P. L.; Ozaki, Y.; Will, J. A.; Brown, R. R.; Cook, J. M.
Evaluation of substituted beta-carbolines as noncompetitive in-
doleamine 2,3-dioxygenase inhibitors. Med. Chem. Res. 1993,
473–482.
(9) Carr, G.; Chung, M. K. W.; Mauk, A. G.; Andersen, R. J. Syn-
thesis of indoleamine 2,3-dioxygenase inhibitory analogues
of the sponge alkaloid exiguamine A. J. Med. Chem. 2008, 51,
2634–2637.
(10) Pereira, A.; Vottero, E.; Roberge, M.; Mauk, A. G.; Andersen, R.
J. Indoleamine 2,3-dioxygenase inhibitors from the Northeastern
Pacific marine hydroid Garveia annulata. J. Nat. Prod. 2006, 69,
1496–1499.
(11) Brastianos, H. C.; Vottero, E.; Patrick, B. O.; Van Soest, R.;
Matainaho, T.; Mauk, A. G.; Andersen, R. J. Exiguamine A, an
indoleamine-2,3-dioxygenase (IDO) Inhibitor isolated from the
marine sponge Neopetrosia exigua. J. Am. Chem. Soc. 2006, 128,
16046–16047.
(13) Li, X.; Yin, W.; Sarma, P. V. V. S.; Zhou, H.; Ma, J.; Cook, J. M.
Synthesis of optically active ring-A substituted tryptophans as IDO
inhibitors. Tetrahedron Lett. 2004, 45, 8569–8573.
(14) Gaspari, P.; Banerjee, T.; Malachowski, W. P.; Muller, A. J.;
Prendergast, G. C.; DuHadaway, J.; Bennett, S.; Donovan, A.
M. Structure-activity study of brassinin derivatives as indole-
amine 2,3-dioxygenase inhibitors. J. Med. Chem. 2006, 49,
684–692.
(15) Muller, A. J.; DuHadaway, J. B.; Donover, P. S.; Sutanto-Ward,
E.; Prendergast, G. C. Inhibition of indoleamine 2,3-dioxygenase,
an immunoregulatory target of the cancer suppression gene Bin1,
potentiates cancer chemotherapy. Nat. Med. (N.Y., NY, U.S.)
2004, 11, 312–319.
(16) Muller, A. J.; Malachowski, W. P.; Prendergast, G. C. Indoleamine
2,3-dioxygenase in cancer: targeting pathological immune toler-
ance with small-molecule inhibitors. Expert Opin. Ther. Targets
2005, 9, 831–849.
(17) Ichikawa, T.; Kato, T.; Takenishi, T. New synthesis of adenine and
4-aminoimidazole-5-carboxamide. J. Heterocycl. Chem. 1965, 2,
253–255.
(18) Kocevar, M.; Polanc, S.; Sollner, M.; Tisler, M.; Vercek, B. Simple
procedure for the synthesis of pyridinecarbohydroximoyl chlorides
and bromides. Synth. Commun. 1988, 18, 1427–1432.
(19) Andrianov, V.; Eremeev, A. 4-Aminofurazan-3-carbohydroximic
acid halides. Khim. Geterotsikl. Soedin. 1994, 420–425.
(20) Gajewski, T. F.; Meng, Y.; Harlin, H. Immune suppression in the
tumor microenvironment. J. Immunother. 2006, 29, 233–240.
(21) Oda, S. I.; Sugimoto, H.; Yoshida, T.; Shiro, Y. Crystallization and
preliminary crystallographic studies of human indoleamine 2,3-
dioxygenase. Acta Crystallogr., Sect. F: Struct. Biol. Cryst. Com-
mun. 2006, F62, 221–223.
(22) Buzykin, B. I.; Dokuchaev, A. S.; Kharitonova, O. A. Crystal and
molecular structure of N-phenyl-4-nitrobenzamidoxime. Izv.
Akad. Nauk, Ser. Khim. 1995, 1516–1519.
(12) Kumar, S.; Malachowski, W. P.; DuHadaway, J. B.; LaLonde, J.
M.; Carroll, P. J.; Jaller, D.; Metz, R.; Prendergast, G. C.; Muller,
A. J. Indoleamine 2,3-dioxygenase is the anticancer target for a
novel series of potent naphthoquinone-based inhibitors. J. Med.
Chem. 2008, 51, 1706–1718.
(23) Hou, D.-Y.; Muller, A. J.; Sharma, M. D.; DuHadaway, J.;
Banerjee, T.; Johnson, M.; Mellor, A. L.; Prendergast, G. C.;
Munn, D. H. Inhibition of indoleamine 2,3-dioxygenase in den-
dritic cells by stereoisomers of 1-methyl-tryptophan correlates with
antitumor responses. Cancer Res. 2007, 67, 792–801.