Optimization of inhibitors of the tyrosine kinase EphB4. 2. Cellular potency improvement and binding mode validation by X-ray crystallography

Article abstract of DOI:10.1021/jm301187e

Inhibition of the tyrosine kinase erythropoietin-producing human hepatocellular carcinoma receptor B4 (EphB4) is an effective strategy for the treatment of solid tumors. We have previously reported a low nanomolar ATP-competitive inhibitor of EphB4 discovered in silico by fragment-based high-throughput docking combined with explicit solvent molecular dynamics simulations. Here we present a second generation of EphB4 inhibitors that show high inhibitory potency in both enzymatic and cell-based assays while preserving the appealing selectivity profile exhibited by the parent compound. In addition, respectable levels of antiproliferative activity for these compounds have been obtained. Finally, the binding mode predicted by docking and molecular dynamics simulations is validated by solving the crystal structures of three members of this chemical class in complex with the EphA3 tyrosine kinase whose ATP-binding site is essentially identical to that of EphB4.

Full text of DOI:10.1021/jm301187e

Article
pubs.acs.org/jmc
Optimization of Inhibitors of the Tyrosine Kinase EphB4. 2. Cellular
Potency Improvement and Binding Mode Validation by X‑ray
Crystallography
Karine Lafleur,†,‡,# Jing Dong,‡,# Danzhi Huang,‡ Amedeo Caflisch,*,‡ and Cristina Nevado*,†
‡
†Department of Organic Chemistry and Department of Biochemistry, University of Zurich, Winterthurerstrasse 190, CH-8057
Zurich, Switzerland
S
* Supporting Information
ABSTRACT: Inhibition of the tyrosine kinase erythropoietin-
producing human hepatocellular carcinoma receptor B4
(EphB4) is an effective strategy for the treatment of solid
tumors. We have previously reported a low nanomolar ATP-
competitive inhibitor of EphB4 discovered in silico by
fragment-based high-throughput docking combined with
explicit solvent molecular dynamics simulations. Here we
present a second generation of EphB4 inhibitors that show
high inhibitory potency in both enzymatic and cell-based
assays while preserving the appealing selectivity profile
exhibited by the parent compound. In addition, respectable
levels of antiproliferative activity for these compounds have
been obtained. Finally, the binding mode predicted by docking and molecular dynamics simulations is validated by solving the
crystal structures of three members of this chemical class in complex with the EphA3 tyrosine kinase whose ATP-binding site is
essentially identical to that of EphB4.
recognized as an effective strategy for the therapy of solid
INTRODUCTION
■
tumors. ATP-competitive EphB4 inhibitors of type I,4,5 type
Protein kinases are involved in the phosphorylation of signaling
proteins and regulate key processes such as cell differentiation,
cell proliferation, and cell motility. Depending on their target
residue, protein kinases can be divided into serine-threonine
and tyrosine kinases. Receptor tyrosine kinases (RTKs) are
transmembrane proteins which, upon interactions with their
cognate ligands, undergo dimerization and autophosphoryla-
tion. Subsequent phosphorylation of cytoplasmic proteins by
the activated receptor triggers a signaling cascade which, in
turns, induces the transcription of specific genes. Thus,
disruption of the protein function or aberrant protein
expression has been linked to several types of diseases,
including cancer.
Eph (erythropoietin-producing human hepatocellular carci-
noma) kinases represent the largest family of RTKs. They are
divided into two families, EphA and EphB, which interact with
ephrinA and ephrinB ligands, respectively.1 Interactions
between Eph receptors and ephrin ligands induce a forward
signaling and a ligand reverse signaling involved in cell
adhesion, repulsion, and migration. Depending on the type of
tumor, Eph−ephrin interactions have been shown to either
promote or inhibit tumor growth. Interestingly, several studies
have implicated EphB4−ephrinB2 signaling in sprouting
angiogenesis and blood vessel maturation,2 and the inhibition
of vascular endothelial growth factor (VEGFR)-driven angio-
genesis by a selective EphB4 inhibitor has recently been
described.3 Therefore, inhibition of EphB4 activity has been
6,7
I1/2
,
and type II3,8 have been discovered and recently
reviewed.9
We have previously reported the identification of two series
of low micromolar (μM) type I inhibitors of EphB4 by
fragment-based high-throughput docking.10 Starting from a
library of about 700 000 compounds, our computational
strategy was designed to generate a focused subset of molecules
with promising anchor fragments for the ATP binding site of
EphB4. Automatic docking of the 21 418 compounds of the
focused library and subsequent in vitro evaluations yielded
compounds 1 and 2, which inhibit EphB4 enzymatic activity in
the low micromolar range (Table 1).10 Their predicted binding
mode involves two hydrogen bonds between the pyrimidine
ring and the Met696 backbone polar groups in the so-called
hinge region of the ATP binding pocket (Figure 1, left).
Guided by the binding mode predicted by docking and further
validated by explicit solvent molecular dynamics simulations, a
tailored library of derivatives of the parent scaffold was
synthesized to improve the binding affinity.6 In fact, addition
of a methyl and a hydroxyl group at positions 2 and 5 of the
phenyl ring, respectively, provided compound 3 (Figure 1,
right), which shows single-digit nanomolar potency in two
different enzymatic assays with the kinase domain of EphB4 in
Received: August 12, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. EphB4 Inhibition Data of Xanthine Derivatives
enzymatic assay IC50 (nM)
a
b
c
compound
R1
R2
R3
R4
R5
R6
FRET
radiometric
cellular assay IC50 (nM)
1
2
3
H
H
H
H
H
H
H
H
H
H
H
H
H
o-methoxyphenyl
butyl
3300
1900
5
4350
538
1.6
14% at 20 μM
<10% at 20 μM
130
Me
OH
o-methoxyphenyl
a
FRET-based enzymatic assay carried out using the Z′-LYTE Kinase Assay Kit−Tyr 1 Peptide (Invitrogen, Grand Island, NY) following the vendor
b
c
instructions. Measured at Reaction Biology Corporation using radiolabeled ATP. Cell IC50 values for compounds 1 and 2 were measured in a
cellular phosphorylation assay using CHO cells overexpressing EphB4. For compound 3, measurements were performed at Proqinase using MEF
cells overexpressing EphB4.
Figure 1. (Left) Predicted binding mode of compounds 1 and 2.10 (Right) Binding mode of compound 3 as predicted in silico and validated by X-
ray crystallography. The EphB4 residues involved in binding are shown. Thr693 is the so-called gatekeeper residue while Asp758 is the first residue
in the DFG motif.
solution (Table 1). In addition, compound 3 showed a
promising selectivity profile (Figure 2, right). Unfortunately, a
significant reduction of inhibitory activity on EphB4 was
observed for this compound in a cellular assay compared to the
enzymatic assays. This discrepancy could be due to reduced
permeability or increased efflux of the molecule by active
transporters. In addition, phenols can undergo glucuronidation
during phase II metabolism, thus having poor pharmacokinetic
properties.11,12
With these results in hand, we embarked on a multi-
disciplinary optimization campaign aimed at the development
of a second generation of EphB4 inhibitors that combine high
inhibitory potency in both enzymatic and cell-based assays
while retaining the appealing selectivity profile exhibited by the
parent compound. Here, we present the validation of this
approach, which has led to the discovery of a lead compound
incorporating not only high potency and selectivity but also a
promising pharmacological profile as well as respectable levels
of antiproliferative activity. In addition, we report herein the
confirmation of the binding mode of this chemical class to the
Eph receptor tyrosine kinase by means of X-ray crystallography.
The high potency displayed by compound 3 in the enzymatic
assay prompted us to explore its activity on a cellular setting.
Thus, cellular phosphorylation assays were carried out on
murine embryonal fibroblast cells (MEF) transfected with myc-
tagged human EphB4 (Proqinase). After incubation with the
corresponding inhibitor and stimulation with ephrinB2-Fc,
autophosphorylation of EphB4 was quantified via sandwich
ELISA. While compound 3 showed single-digit nanomolar
activity in enzymatic assay, its inhibition activity in cells was
reduced by almost 2 orders of magnitude, with IC50 values
higher than 100 nM (Table 1). There are multiple reasons for a
discrepancy between enzymatic and cellular assays, including
the passage through the cell membrane, the natural environ-
ment, and relevant ATP concentration present only in the
latter.13 To further evaluate the pharmacological potential of
compound 3, its cell permeability was investigated first. While
passive transport was assessed on Caco-2 monolayers, active
efflux was tested on P-glycoprotein and BCRP transporters
(Absorption Systems). Caco-2 cells express many transport
proteins present on epithelial cells of the small intestine and
form tight junctions between each other. Thus, measurements
of permeability through Caco-2 monolayers may also predict
absorption across intestinal tissues in vivo. Interestingly,
compounds 1 and 2 bearing hydrogens at position R1 and R4
showed high cell permeability, the latter showing the highest
membrane permeation, while no significant efflux was observed
(Table 2). In contrast, compound 3 with a methyl group at R1
LEAD OPTIMIZATION STRATEGY
■
In line with the results obtained in the enzymatic assay, the
inhibition activity of compounds 1 and 2 on EphB4-transfected
Chinese hamster ovary (CHO) cells has been shown to be
relatively low with IC50 values in the high micromolar range.10
B
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 2. Selectivity profile of compound 40 (left) and 3 (right) tested on a panel of 124 protein kinases. Measurements were performed at the
University of Dundee using a [γ-33P]ATP-based enzymatic assay at a concentration of 3 μM for compound 40 and 1 μM for compound 3. The
affinity is defined as strong, medium, and low (corresponding to <10%, 10−30%, and 30−70%, respectively, of remaining activity with respect to a
DMSO control). The dendrogram was obtained from KinomeScan using the KinomeTree software.
Table 2. Cell Permeability Measurements for Compounds 1−3 Using the Caco-2 Model
a
% recovery
Papp (×10−6 cm/s)
b
c
d
compound
A−B
B−A
A−B
B−A
efflux ratio
permeability classification
significant efflux
1
2
3
71
69
51
76
78
74
32.3
42.5
3.13
35.8
42.5
23.0
1.1
1.0
7.3
high
high
high
no
no
yes
a
Apparent permeability (Papp) is measured from the apical side to the basolateral side (A−B) and from the basolateral side to the apical side (B−A).
b
c
Efflux ratio = (Papp B → A)/(Papp A → B). Permeability classification: low = (Papp A → B) < 1.0 × 10−6cm/s; high = (Papp A → B) > 1.0 × 10−6
d
cm/s. Significant efflux: efflux ratio > 3.0 and (Papp B → A) > 1.0 × 10−6 cm/s.
and a hydroxyl group at R4 showed high cell permeability
although it was significantly effluxed by P-glycoprotein
transporters, which could explain the activity discrepancy
between enzymatic and cellular assays.
As phenols are known to be detrimental for the
pharmacokinetic properties of drugs,11,12 we decided to remove
the hydroxyl group at position R4. In addition, replacement of
the anisidine side chain by an alkyl chain appeared to be
promising, according to the permeability measurements
(compare compounds 1 with Papp = 32.3 × 10−6 cm/s and 2
with Papp = 42.5 × 10−6 cm/s in Table 2). Thus, we investigated
the synthesis of molecules bearing a methyl group at position
R1 and a butyl, pentyl, or hexyl chain at position R6 (a previous
shortening of the alkyl chain did not improve the affinity to
EphB4; see ref 10). In addition, we also decided to introduce a
fluorine at position R3, To explore the possibility of the
potential formation of an additional halogen bond between the
inhibitor and the Glu664 residue,14,15 we also examined the
addition of a chlorine at position R4. To improve the van der
Waals interactions between the ligand and the side chains of
Val629, Ala645, and Thr693, we also aimed at introducing a
chlorine atom or a trifluoromethyl group at position R1, with R6
being either an alkyl chain or an aromatic ring.
Rigidification of biologically active molecules has been shown
to be beneficial for selectivity properties,11 as the number of
accessible conformations able to interact with secondary targets
C
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 1
a
Reagents and conditions: (a) Primary amine, EtOH, reflux, 15 h.
a
Scheme 2
a
Reagents and conditions: (a) MeLi, THF, 25 °C, 1 h, 55%. (b) PCC, CH2Cl2, 25 °C, 1 h, 73%. (c) BuLi, acetaldehyde, THF, −78 °C, 1.5 h. (d)
PCC, CH2Cl2, 25 °C, 1 h, 56% yield over two steps. (e) KNO3, H2SO4, 0 °C, 1 h. (f) Fe, AcOH, EtOH, 90 °C, 2 h, 47% yield over two steps. (g)
NaOAc, Ac2O, CHCl3, 25 °C, 30 min, then iAmONO, 60 °C, 7 h, 58%.
is restricted. In addition, ortho-substitution of bicyclic
molecules with methyl or methoxy groups has been shown to
decrease crystal packing and improve aqueous solubility.16 We
therefore investigated the synthesis of a bis-ortho-methyl-
substituted phenyl ring on the xanthine scaffold. Finally, poor
oral bioavailability of kinase inhibitors entitling an ortho-
methyl/meta-hydroxyphenyl pattern11,17,18 led to the replace-
ment of the phenol ring by an indazole moiety.11 This
transformation aimed to generate a molecule with a similar
hydrogen bonding pattern as compound 3 with EphB4.
With these designs in mind, we set out to explore the
synthesis of a small set of customized analogues to later
evaluate their biochemical and pharmacological profile.
summarized in Scheme 2. Treatment of commercially available
4-fluoro-2-methylbenzaldehyde 5 with methyllithium followed
by oxidation with PCC provided methyl ketone 7 in 40% yield
over two steps. The synthesis of 1-(2,6-dimethylphenyl)-
ethanone started with 2-bromo-1,3-dimethylbenzene 8, which
undergoes bromo−lithium exchange in the presence of nBuli,
followed by reaction with acetaldehyde to afford secondary
alcohol 9. Subsequent oxidation using PCC gave the expected
acetophenone 10,22 which was used as starting material in the
synthesis of 1,1′-(5-methyl-1H-indazole-1,4-diyl)diethanone
13. Indeed, reaction with KNO3 selectively nitrated compound
10 at the meta position relative to the carbonyl group.23
Reduction in the presence of iron and cyclization using isoamyl
nitrite afforded the protected indazole derivative 13 in good to
moderate yields.24
α-Bromination of the acetophenones reported in Scheme 2
as well as the commercially available 2′-chloroacetophenone
(14), 3′-chloroacetophenone (15), and 2′-(trifluoromethyl)-
acetophenone (16), was achieved in the presence of copper(II)
bromide in chloroform as shown in Scheme 3.25
Alkylation at the N7 position of bromoxanthine 4 with meta-
substituted or mono-ortho-substituted α-haloketones was
accomplished using N,N-diisopropylethylamine in DMF to
give intermediates 24−27 (Scheme 4). To our surprise, in
contrast to our previous experience with anilines, the
subsequent cyclization with alkylamines under the standard
conditions shown in Scheme 1 was never observed. Instead,
only intermediates 31−36 could be isolated from the reaction
mixtures probably due to the more electron-rich nature of
alkylamines, which are able to accomplish not only the
SYNTHESIS
■
Recently, a growing interest in the synthesis of imidazox-
anthines has emerged, as a few derivatives proved to be potent
serotonin,19 A3 adenosine,20,21 or kinase receptor antagonists.6
Following our previously developed route, the synthesis of the
key intermediate 3-methyl-8-bromoxanthine 4 was achieved in
five steps starting from the corresponding alkylurea (Scheme
1). After alkylation with α-brominated acetophenones,
cyclization usually occurs by refluxing the resulting intermediate
in ethanol in the presence of a primary amine.6,19,20 However,
to the best of our knowledge, the synthesis of bis-ortho-
substituted imidazoxanthines has not been described yet. In
addition, none of the mono-ortho-substituted molecules
reported so far has an alkyl chain at position R6.
The synthesis of the noncommercially available acetophe-
nones used in the preparation of this new set of inhibitors is
D
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 3
ENZYMATIC, CELL-BASED ASSAY AND
EVALUATION OF CELL PERMEABILITY
■
The inhibitory activity of our second generation of EphB4
inhibitors 40−50 was first evaluated on an enzymatic assay
based on fluorescence resonance energy transfer (FRET).
Additionally, compounds 40, 43, and 50 were also tested in a
cell-free enzymatic assay based on radiolabeled ATP (Table 3).
Removal of the hydroxyl group at R4 and replacement of the
anisidine by a butyl group yielded 40 with an IC50 around 50
nM, thus reflecting the beneficial effect of an alkyl lateral chain
(see also Table 1, entries 1 and 2). Elongation of the butyl
chain (compounds 41 and 42) did not improve the affinity,
which is congruent with the orientation toward solvent of the
alkyl chain. The introduction of a fluorine atom at position R3
yielded compound 43 with binding affinity similar to that of
compound 40 as expected from the binding mode. However,
the addition of a chlorine at position R4 (compounds 44 and
47) as well as the replacement of the methyl group by a
chlorine atom or a trifluoromethyl moiety (compounds 45, 46,
and 48) was detrimental for binding affinity probably due
mainly to steric effects, i.e., less favorable van der Waals
interactions. Finally, compound 49 bearing methyl groups at
positions R1 and R5 showed micromolar activity, reflecting that
the introduction of an extra steric bulk on the molecule also
decreases the binding affinity. As suggested by the similar
hydrogen bonding pattern, compound 50 bearing an indazole
ring displayed almost the same potency, at least in the FRET
assay, as that of the original compound 3.
a
Reagents and conditions: (a) CuBr2, CHCl3, EtOAc, reflux, 15 h
(21−64% yield).
substitution of the bromine at C8 but also the condensation
with the carbonyl group. Fortunately, upon imine activation in
the presence of a Lewis acid such as aluminum trichloride, the
expected cyclization products 40−45 could be obtained in good
yields. Interestingly, when position R1 was substituted by a
trifluoromethyl group, the reaction with butylamine gave the
substitution product 37, which could be then converted into 46
upon treatment with the amine in the presence of AlCl3.
Cyclization of the derivatives bearing chlorine atoms at
positions R1 and R4 in the presence of anisidine was performed
as previously reported,6 giving products 47 and 48 (Scheme 5).
The synthesis of derivatives bearing a bis-ortho-phenyl-
substituted ring also proved to require further adjustments of
the synthetic sequence as summarized in Scheme 6. The
alkylation of 4 with bis-ortho-substituted α-haloketones 18 and
19 required heating, but now the subsequent cyclizations in the
presence of anisidine failed. The reactions with butylamine
provided the substitution products 38 and 39, which could
finally be cyclized by heating in a concentrated solution of
boron trifluoride diethyl etherate to give compounds 49 and 50.
Note that the reaction with butylamine also deprotected the
acetylated intermediate 30.
In addition, the inhibition activity of our most potent
compounds 3, 40, 43, and 50 was assessed on Abl, Src, Lck, and
Yes1 (Reaction Biology Corporation, Table 4). Compound 3
showed single digit nanomolar activity on Abl, Src, and Lck,
while compounds 40, 43, and 50 retained high inhibition only
on Src.
The six derivatives with nanomolar activity in the enzymatic
assays (40, 41, 43, 45, 48, and 50) were further evaluated in a
a b
,
Scheme 4.
a
Reagents and conditions: (a) α-haloketone, DIPEA, DMF, 25 °C, 17 h. (b) Primary amine, sealed tube, 180 °C, 30 min. (c) AlCl3, EtOH, sealed
b
tube, 180 °C, 2−15 h. (d) Butylamine, EtOH, sealed tube, 180 °C, 1 h. Unless otherwise stated, Rn = H.
E
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a b
,
Scheme 5.
a
b
Reagents and conditions: (a) α-haloketone, DIPEA, DMF, 25 °C, 17 h. (b) Anisidine, EtOH, sealed tube, reflux, 15 h. Unless otherwise stated, Rn
= H.
a
Scheme 6
a
Reagents and conditions: (a) 2-bromo-1-(2,6-dimethylphenyl)ethanone 18, DIPEA, DMF, 110 °C, 2 h. (b) Butylamine, EtOH, sealed tube, 180
°C, 0.5−2 h. (c) BF3·OEt2, DCM, sealed tube, 180 °C, 0.5−6 h. (d) KOH, EtOH, reflux, 2 h, then 1-(1-acetyl-5-methyl-1H-indazol-4-yl)-2-
bromoethanone 19, DMF, 110 °C, 2 h.
Table 3. EphB4 Inhibition Data of Xanthine Derivatives
enzymatic assay IC50 (nM)
a
b
compound
R1
R2
R3
R4
R5
R6
FRET
radiometric
82
cellular IC50 (nM)
40
41
42
43
44
45
46
47
48
49
50
Me
Me
Me
Me
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
F
H
H
H
H
H
H
H
H
H
H
Me
butyl
pentyl
hexyl
butyl
butyl
butyl
butyl
56
177
1300
91
50
230
n.d.
68
c
H
n.d.
H
n.d.
139
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
133
H
H
H
H
H
H
H
H
Cl
22% @ 10 μM
182
n.d.
270
n.d.
n.d.
2200
n.d.
150
Cl
H
CF3
H
H
15% @ 10 μM
12% @ 10 μM
300
Cl
o-methoxyphenyl
o-methoxyphenyl
butyl
Cl
H
Me
Me
H
49% @ 10 μM
14
indazole
butyl
a
FRET-based enzymatic assay carried out using the Z′-LYTE Kinase Assay Kit−Tyr 1 Peptide (Invitrogen) following the vendor instructions.
b
c
Measured at Reaction Biology Corporation using radiolabeled ATP. n.d.: not determined.
cellular phosphorylation assay on MEF cells transfected with
myc-tagged human EphB4 (Proqinase, Table 3, right column).
Except for compound 48 which showed micromolar affinity,
the other five compounds displayed levels of inhibitory activity
F
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 4. Abl, Lck, Src, and Yes1 Inhibition Data of Xanthine
Derivatives
Table 6. Selectivity of Compound 40 Tested on a Panel of
124 Protein Kinases at the University of Dundee
a
a
compound
Abl
0.92
Src
Lck
1.3
Yes1
n.d.
kinase
% activity
b
3
1.14
71
MKK1
RIPK2
Src
56
12
2
40
43
50
578
1282
203
493
617
264
284
110
46
496
160.4
Lck
5
a
CSK
57
14
43
58
6
Enzymatic assays were carried out at Reaction Biology Corporation
Yes1
using radiolabeled ATP. IC50 values are given in nanomolar
concentrations. n.d.: not determined.
b
Abl
BTK
Eph-A2
Eph-A4
Eph-B1
Eph-B2
Eph-B3
Eph-B4
5
in the nanomolar range, which correlates with the potencies
measured in the enzymatic assays. Of particular interest is the
affinity of compounds 40 and 43, with IC50 values of 50 and 68
nM, respectively.
10
8
13
18
To further characterize this second generation of EphB4
inhibitors, the cell permeation of compounds 40 and 50 was
evaluated on Caco-2 monolayers (Absorption Systems). Both
compounds showed a high cell permeability; however, only
compound 40 was not effluxed by transport proteins (Table 5),
and thus it was selected for further evaluation in terms of
selectivity (vide infra).
a
The percentage of kinase activity is measured in the presence of 3 μM
of 40 compared to a 100% DMSO control (only kinases with a
percentage of activity below 70% are indicated). For full data set, see
Table S1 in the Supporting Information.
In addition, the antiproliferative activity of compounds 3, 40,
41, 43, 45, and 50 was assessed on patient-derived tumor cell
lines using a propidium iodide-based proliferation assay and
dasatinib as a reference (Oncotest, Table 7). Cell lines included
colon, lung, kidney, pancreas, prostate, and stomach cancer
cells. Overall, dasatinib exhibited the highest potency, with
double-digit nanomolar activity against RXF 393NL, LXFA
983L, and PRXF DU145. Compounds 3, 40, 41, 43, 45, and 50
inhibited cell proliferation in the low micromolar range, with
compound 3 exhibiting good levels of antiproliferative activity
against RXF 393NL, PAXF 1657L, and PRXF DU145.
Compounds 45 and 50 displayed a 1−2 μM activity against
the colon cancer cell line CXF 1103L. These results confirm
the correlation between the kinase inhibitory activity of these
molecules and their antiproliferative activity, although also
reflect the need for further modifications to parallel the
performance of “gold standards” in this field, such as dasatinib.
SELECTIVITY PROFILE
■
To assess the selectivity profile of compound 40, enzymatic
assays were performed on a panel of 124 kinases (National
Centre for Protein Kinase Profiling at the University of
Dundee, Table 6, Figure 2, left, and Table S1 in the Supporting
Information). In the presence of 3 μM concentration of
compound 40, only 5 of the 124 kinases tested (Src, Lck,
EphA2, EphA4, and EphB2) had a remaining activity smaller
than 10% with respect to a DMSO control. Furthermore, 5
other kinases (RIPK2, Yes1, EphB1, EphB3, and EphB4)
showed a remaining activity between 10% and 20%. It is
important to note that these 10 kinases have a threonine as a
gatekeeper. Overall, compound 40 has a selectivity profile
similar to that of 3, with a strong inhibition of a relatively small
fraction of the human kinome (Figure 2).
CONFIRMATION OF THE BINDING MODE BY
PROTEIN X-RAY CRYSTALLOGRAPHY
ANTIPROLIFERATIVE ACTIVITY
■
■
Overexpression of the receptor tyrosine kinase EphB4 has been
linked to several types of cancer, including breast,26 colon,27
and ovary.28 To assess the potential of our inhibitors on a
cancer model, compound 3 was submitted to the NCI-60
cancer cell line panel. Compound 3 displayed remarkable levels
of cell growth inhibition against central nervous system (SNB-
75, 128 nM), leukemia (K-562, 309 nM), and breast (HS 578T,
562 nM) cancer cells (data shown in the Supporting
Information).
The binding mode of compounds 3, 40, and 50 was
investigated by X-ray crystallography using the catalytic domain
of EphA3 expressed in E. coli.29 This tyrosine kinase was
preferred to EphB4 for which only expression in insect cells has
been reported. Note that 32 of the 36 residues in the ATP
binding site of EphA3 are identical to those in EphB4.30
Moreover, the side chains involved in binding compounds 3,
40, and 50 are identical in EphA3 and EphB4. These
Table 5. Cell Permeability Measurements for Compounds 3, 40, and 50 Using the Caco-2 Model
a
% recovery
Papp (×10−6 cm/s)
b
c
d
compound
A−B
B−A
A−B
B−A
efflux ratio
permeability classification
significant efflux
3
51
85
84
74
89
96
3.13
42.7
8.02
23.0
28.1
49.0
7.3
0.7
6.1
high
high
high
yes
no
40
50
yes
a
Apparent permeability (Papp) is measured from the apical side to the basolateral side (A−B) and from the basolateral side to the apical side (B−A).
b
c
Efflux ratio = (Papp B → A)/(Papp A → B). Permeability classification: low = (Papp A → B) < 1.0 × 10−6cm/s; high = (Papp A → B) > 1.0 × 10−6
d
cm/s. Significant efflux: efflux ratio > 3.0 and (Papp B → A) > 1.0 × 10−6 cm/s.
G
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Table 7. Antiproliferative Activity against Tumor Cell Lines
a
IC50 values were determined at Oncotest using a modified propidium iodide assay. Measurements were performed after 4 days of incubation with
the corresponding compound. IC50 values are given in micromolar concentrations (μM).
observations explain the very similar potency on both targets of
our inhibitors and other Eph kinase inhibitors reported
previously.6
called back pocket with its hydroxyl group accepting a
hydrogen bond from the backbone NH of Asp758 (Asp764
for EphA3) of the DFG loop and donating a hydrogen bond to
the side chain of Glu664 (Glu670 for EphA3) (Figure 1, right).
Superposition of the crystal structures obtained for
compounds 3, 40 and 50 shows that the key features of the
binding mode are conserved, such as the hydrogen bond
interactions between the pyrimidine ring and Met702, as well as
the orientation of the phenyl ring in the back pocket (Figure 4).
Interestingly, the indazole ring of compound 50 is involved in
two hydrogen bond interactions: the NH group acts as donor
to the Glu670 side chain, and the imine nitrogen acts as
acceptor for the backbone NH of Asp764. The methoxy group
of compound 3 matches the butyl side chains of compounds 40
and 50, as both of these substituents point in the opposite
direction of the hinge region. An in-depth comparative analysis
of the binding modes of compounds 3, 40, and 50, with the X-
ray structures of two other series of EphB4/A3 inhibitors
discovered in our groups as well as with the inhibitor/Eph
complexes available in the PDB database, will be published
elsewhere (Jing Dong et al., in preparation).
Crystals were obtained using the hanging drop vapor
diffusion method (see Experimental Section). The final
resolution is 2.2 Å, 1.9 Å, and 2.1 Å for the complexes of
EphA3 with compounds 3, 40, and 50, respectively; further
details of refinement statistics are shown in Table 8.
Strikingly, the binding mode of compound 3 observed in the
crystal structure is essentially identical to that predicted by
docking and explicit solvent molecular dynamics6 (Figure 3).
The pyrimidine ring is involved in two hydrogen bonds with
the hinge region, and the phenol moiety is nestled into the so-
Table 8. X-ray Data Collection
3
40
50
space group
unit cell
a (Å)
P 1 21 1
P 1 21 1
P 1 21 1
52.6
53.29
53.06
b (Å)
38.2
38.25
38.18
Finally, the binding mode of compound 3, predicted and
6,7
c (Å)
75.7
76.04
75.84
validated as belonging to type I1/2
,
lends itself to a
modification into type II3,8 via replacement of the hydroxyl
group by an amide linker and addition of an aromatic group,
e.g., CF3-substituted benzol, to reach the so-called allosteric site
proximal to the ATP binding site. We are currently planning
the synthesis of these type II derivatives.
resolution
range (Å)
38.2−2.2
34.0−1.9
38.2−2.1
unique
reflections
15076 (2090)
23747 (3297)
17532 (2402)
⟨I/σ(I)⟩
R merge
8.8 (3.0)
8.9 (2.8)
10.6 (3.9)
0.103 (0.457)
98.4 (94.6)
0.084 (0.405)
98.5 (94.8)
0.086 (0.334)
98.8 (93.4)
completeness
(%)
CONCLUSIONS
■
multiplicity
3.7 (3.5)
3.0 (2.9)
3.5 (3.4)
Recently, a medicinal chemistry campaign inspired by the pose
of the original hit obtained by docking resulted in the discovery
of compound 3, a single-digit nanomolar inhibitor of the EphB4
tyrosine kinase. However, further evaluation of compound 3
revealed a major discrepancy between enzymatic and cellular
activity (5 nM in enzymatic assay vs 130 nM in cellular
phosphorylation assay). Furthermore, efficient efflux by P-
glycoprotein transporters was observed on caco-2 monolayers.
Aided by the docked pose of compound 3, a second
generation of EphB4 inhibitors was designed and synthesized to
refinement
resolution
range (Å)
30.68−2.20
29.79−1.90
29.70−2.10
R factor/R free 19.15/23.25
18.45/22.34
24.3
17.37/22.53
28.8
mean B factors 30.8
(Å2)
RMS bonds
(Å)
0.0077
1.505
0.0071
1.377
0.0071
1.365
RMS angles
(deg)
H
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. Comparison of docking results with the X-ray structure. (a) Superposition of the X-ray structure of the complex of EphA3 with compound
3 (magenta C atoms) and the pose of its anchor fragment (blue C atoms) as predicted by docking into EphB4. (b) Superposition of the X-ray
structure of the complex of EphA3 with compound 3 (magenta C atoms) and the binding mode into EphB4 obtained by explicit solvent molecular
dynamics (blue C atoms). The coloring scheme of the protein surface (shown only for EphA3) emphasizes donor/acceptor (blue/red) and
hydrophobic (white). The residues in the ATP binding site of EphA3 and EphB4 are identical.
Figure 4. (a) X-ray structure of compound 3 (carbon atoms in cyan) in complex with EphA3. (b) Superposition of compounds 3 (carbon atoms in
cyan), 40 (magenta C atoms), and 50 (green C atoms).
address this dichotomy. Among this small library of carefully
tailored molecules, compound 40 showed a 50 nM inhibition
activity both in enzymatic and cell-based assays. Further
biological analysis suggested that compound 40 was not
actively effluxed, thus improving the pharmacological properties
compared to the previous series. In addition, strong inhibition
activity was observed on only 5% of a 124 kinase panel which
shows that compound 40 has a selectivity profile as good as that
of compound 3.
Remarkably, the binding mode of compounds 3, 40, and 50
suggested by docking followed by MD simulations with explicit
solvent has been confirmed by X-ray crystallography. In
particular, there is a perfect overlap with the crystal structure
for the important features of the predicted binding mode of
compounds 3, 40, and 50, such as the hydrogen bonding
pattern with the hinge region and the orientation of the phenyl
or indazole ring in the back pocket.
EXPERIMENTAL SECTION
■
Chemistry. All reactions, unless otherwise stated, were carried out
under a nitrogen atmosphere using standard Schlenk techniques. All
reagents were used as received unless otherwise noted. Solvents were
purchased in the best quality available, degassed by purging thoroughly
with nitrogen, and dried over activated molecular sieves of appropriate
size. Alternatively, they were purged with argon and passed through
alumina columns in a solvent purification system (Innovative
Technology). Reactions were monitored by thin layer chromatography
(TLC) using Merck TLC silica gel 60 F254. Flash column
chromatography was performed over silica gel (230−400 mesh).
NMR spectra were recorded on AV2 400 or AV2 500 MHz Bruker
spectrometers. Chemical shifts are given in ppm. The spectra are
calibrated to the residual 1H and 13C signals of the solvents.
Multiplicities are abbreviated as follows: singlet (s), doublet (d),
triplet (t), quartet (q), doublet−doublet (dd), quintet (quint), septet
(sept), multiplet (m), and broad (br). Melting points were determined
on a Buchi Melting Point B-540 instrument. 4-Fluoro-2-methyl-
̈
benzaldehyde (5), 2-bromo-1,3-dimethylbenzene (8), 2′-chloroaceto-
phenone (14), 3′-chloroacetophenone (15), and 2′-(trifluoromethyl)-
acetophenone (16) were purchased from Fluka. 2-Bromo-1-(o-
tolyl)ethanone (23) was purchased from Synchem. Compounds 1−
4 have been previously described.6
In summary, further synthetic efforts and additional
biochemical profiling provided new EphB4 tyrosine kinase
inhibitors with improved activity in cellular assays. Thus,
compound 40 holds promise for further evaluation in in vivo
models.
High-resolution electrospray ionization mass spectrometry was
performed on a Finnigan MAT 900 (Thermo Finnigan, San Jose, CA)
double-focusing magnetic sector mass spectrometer. Ten spectra were
acquired. A mass accuracy ≤2 ppm was obtained in the peak-matching
I
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
acquisition mode by using a solution containing 2 μL of PEG200, 2 μL
of PPG450, and 1.5 mg of NaOAc (all obtained from Sigma-Aldrich,
Buchs, Switzerland) dissolved in 100 mL of MeOH (HPLC Supra
grade, Scharlau, E-Barcelona) as internal standard. The purity of all
tested compounds was determined by HPLC on a Waters Acquity
UPLC (Waters, Milford, MA) Top spectrometer using an Acquity
BEH C18 HPLC column (1.7 μm, 1 × 50 mm, Waters) with a mixture
of H2O + 0.1% HCOOH (A) and CH3CN + 0.1% HCOOH (B)
solvent (0.1 mL flow rate, linear gradient from 5% to 98% B within 4
min followed by flushing with 98% B for 1 min). UV detection was set
to 200−260 nm. Unless otherwise stated, all the compounds showed
≥95% purity.
General Procedure for the α-Bromination of Acetophe-
nones. A solution of the acetophenone (1 equiv) in CHCl3 (0.18 M)
was added to a refluxing solution of copper(II) bromide (1.99 equiv)
in EtOAc (0.45 M). The mixture was then refluxed for 10 h. The
solution was filtered through Celite and concentrated under reduced
pressure to afford a green solid. The solid was purified by flash
chromatography on silica gel to afford the corresponding products in
pure form. This method was used to obtain compounds 17−22. 2-
Bromo-1-(o-tolyl)ethanone (23) was purchased from Synchem.
General Procedure for the Preparation of Alkylated
Xanthines. 3-Methyl-8-bromoxanthine (4, 1 equiv) was dissolved in
DMF (0.4 M), and N,N-diisopropylethylamine (1.5 equiv) was added.
After the mixture was stirred for 5 min at 25 °C, α-
bromoacetophenone (1 equiv) was added. The reaction was stirred
at 25 °C for 15 h. The mixture was then concentrated under reduced
pressure, methanol was added, and the formed precipitate was filtered
off and washed with water to afford the corresponding product in pure
form. This method was used to obtain compounds 24−28.
General Procedure A for the Cyclization of Alkylated
Xanthines. A mixture of 3-alkyl-8-bromo-3,7-dihydro-7-(2-oxo-2-
phenylethyl)-1H-purine-2,6-dione (1.0 equiv) and alkylamine (0.1 M)
was heated in a sealed tube at 180 °C for 30 min. The reaction was
cooled to room temperature and evaporated to dryness. The resulting
residue was triturated with a mixture acetone/water, and the
precipitate was filtered off. The crude intermediate was heated in a
sealed tube at 180 °C in EtOH (0.1 M) in the presence of AlCl3 (2
equiv) for 2−15 h. The reaction was poured into water and extracted
with CH2Cl2, and the organic layer was dried over MgSO4, filtered,
and concentrated under reduced pressure. Purification by column
chromatography on silica gel (CH2Cl2/EtOAc 1:1) afforded the
corresponding product in pure form. This method was used to obtain
intermediates 31−36 and final products 40−45.
General Procedure B for the Cyclization of Alkylated
Xanthines. A mixture of 3-alkyl-8-bromo-3,7-dihydro-7-(2-oxo-2-
phenyl-ethyl)-1H-purine-2,6-dione (1.0 equiv) and the anisidine (4.0
equiv) in EtOH (concentration, 0.1 M) was heated in a sealed tube at
180 °C for 12 h. The reaction was cooled to room temperature, and
the formed solid was filtered off and washed with water to afford the
corresponding product in pure form. This method was used to obtain
compounds 47 and 48.
7.46−7.33 (m, 4H), 3.82 (t, J = 7.3 Hz, 2H), 3.40 (s, 3H), 2.24 (s,
3H), 1.54 (quint, J = 7.3 Hz, 2H), 1.06 (sext, J = 7.3 Hz, 2H), 0.68 (t, J
= 7.3 Hz, 3H); HRMS (ESI): m/z: calcd for C19H21N5O2Na+:
374.1588, found: 374.1585.
Synthesis of Compound 46. 8-Bromo-3-methyl-7-(2-oxo-2-(2-
(trifluoromethyl)phenyl)ethyl)-1H-purine-2,6(3H,7H)-dione (28).
White solid; yield: 54%; 1H NMR (500 MHz, DMSO-d6): δ =
11.41 (s, 1H), 8.09 (d, J = 7.5 Hz, 1H), 7.96−7.92 (m, 2H), 7.87−7.84
(m, 1H), 5.78 (s, 2H), 3.36 (s, 3H); HRMS (ESI): m/z: calcd for
C15H10BrF3N4O3H+: 430.9961, found: 430.9963.
8-(Butylamino)-3-methyl-7-(2-oxo-2-(2-(trifluoromethyl)phenyl)-
ethyl)-1H-purine-2,6(3H,7H)-dione (37). A mixture of 8-bromo-3-
methyl-7-(2-oxo-2-(2-(trifluoromethyl)phenyl)ethyl)-1H-purine-2,6-
(3H,7H)-dione (28, 490 mg, 1.13 mmol) and butylamine (0.448 mL,
4.55 mmol) in EtOH (11 mL) was heated in a sealed tube at 180 °C
for 1 h. The reaction was cooled to room temperature, and the solid
was filtered off to afford the crude product as a light yellow solid (345
1
mg, 74% yield). H NMR (500 MHz, DMSO-d6): δ = 10.62 (s, 1H),
8.15 (d, J = 7.6 Hz, 1H), 7.94−7.89 (m, 2H), 7.85−7.82 (m, 1H), 7.03
(t, J = 5.4 Hz, 1H), 5.55 (s, 2H), 3.32 (s, 3H), 1.53 (quint, J = 7.3 Hz,
2H), 1.33 (sext, J = 7.3 Hz, 2H), 0.90 (t, J = 7.3 Hz, 3H), 3.31 (s, 2H);
HRMS (ESI): m/z: calcd for C19H20F3N5O3H+: 424.1591, found:
424.1593.
8-Butyl-1-methyl-7-(2-(trifluoromethyl)phenyl)-1H-imidazo[2,1-
f ]purine-2,4(3H,8H)-dione (46). 8-(Butylamino)-3-methyl-7-(2-oxo-
2-(2-(trifluoromethyl)phenyl)ethyl)-1H-purine-2,6(3H,7H)-dione
(37, 345 mg, 0.81 mmol) was heated in a sealed tube at 180 °C in
EtOH (8 mL) in the presence of AlCl3 (543 mg, 4.07 mmol) for 15 h.
The reaction was poured in water and extracted with CH2Cl2, and the
organic layer was dried over MgSO4, filtered, and concentrated under
reduced pressure. Purification by column chromatography on silica gel
(CH2Cl2/EtOAc 1:1) afforded the desired compound as a white solid
1
(50 mg, 15% yield). mp 267−269 °C; H NMR (500 MHz, DMSO-
d6): δ = 10.98 (s, 1H), 7.98 (d, J = 7.4 Hz, 1H), 7.87 (t, J = 7.4 Hz,
1H), 7.81 (t, J = 7.4 Hz, 1H), 7.77 (d, J = 7.4 Hz, 1H), 7.63 (s, 1H),
3.81 (t, J = 7.3 Hz, 2H), 3.40 (s, 3H), 1.56 (quint, J = 7.3 Hz, 2H),
1.13 (sext, J = 7.3 Hz, 2H), 0.72 (t, J = 7.3 Hz, 3H); HRMS (ESI): m/
z: calcd for C19H18F3N5O2H+: 406.1485, found: 406.1483.
Synthesis of Compound 50. 7-(2-(1-Acetyl-5-methyl-1H-
indazol-4-yl)-2-oxoethyl)-8-bromo-3-methyl-1H-purine-2,6(3H,7H)-
dione (30). 8-Bromo-3,9-dihydro-3-methyl-1H-purine-2,6-dione (4,
250 mg, 1.02 mmol) was added to a solution of KOH (57 mg, 1.02
mmol) in EtOH (3 mL). The resulting mixture was then heated to
reflux for 2 h. EtOH was then removed under reduced pressure, and
the resulting solid was washed with cold EtOH, and filtered off to
afford a light yellow solid. A mixture of this solid (144 mg, 0.51 mmol)
was dissolved in DMF (1.3 mL), and 1-(1-acetyl-5-methyl-1H-indazol-
4-yl)-2-bromoethanone (19, 150 mg, 0.51 mmol) was added to the
mixture. The reaction was stirred at 110 °C for 2 h. The mixture was
then concentrated under reduced pressure, methanol was added, and
the formed precipitate was filtered off and washed with water to afford
the product as a light yellow solid (184 mg, 39% yield over two steps).
1H NMR (500 MHz, DMSO-d6): δ = 11.48 (s, 1H), 8.64 (d, J = 0.7
Synthesis of Compound 40. 8-Bromo-3-methyl-7-(2-oxo-2-(o-
tolyl)ethyl)-1H-purine-2,6(3H,7H)-dione (24). White solid; yield:
1
Hz, 1H), 8.42 (d, J = 8.7 Hz, 1H), 7.65 (d, J = 8.7 Hz, 1H), 5.84 (s,
2H), 3.39 (s, 3H), 2.75 (s, 3H), 2.53 (s, 3H); HRMS (ESI): m/z:
calcd for C18H15BrN6O4H+: 459.0411, found: 459.0407.
58%; H NMR (500 MHz, DMSO-d6): δ = 11.32 (s, 1H), 8.04 (dd,
J = 7.6 Hz, J = 1.1 Hz, 1H), 7.56 (dt, J = 7.6 Hz, J = 1.1 Hz, 1H), 7.44
(t, J = 7.6 Hz, 1H), 7.40 (d, J = 7.6 Hz, 1H), 5.79 (s, 2H), 3.37 (s,
3H), 2.42 (s, 3H); HRMS (ESI): m/z: calcd for C15H13BrN4O3Na+:
399.0069, found: 399.0068.
8-(Butylamino)-3-methyl-7-(2-(5-methyl-1H-indazol-4-yl)-2-ox-
oethyl)-1H-purine-2,6(3H,7H)-dione (39). A mixture of 7-(2-(1-
acetyl-5-methyl-1H-indazol-4-yl)-2-oxoethyl)-8-bromo-3-methyl-1H-
purine-2,6(3H,7H)-dione (30, 143 mg, 0.31 mmol) and butylamine
(0.123 mL, 1.24 mmol) in EtOH (3 mL) was heated in a sealed tube
at 175 °C for 30 min. The reaction was cooled to room temperature,
and the solid was filtered off to afford the crude product as a light
yellow solid (120 mg, 93% yield). 1H NMR (500 MHz, DMSO-d6): δ
= 13.35 (s, 1H), 10.70 (s, 1H), 8.46 (s, 1H), 7.68 (d, J = 8.5 Hz, 1H),
7.34 (d, J = 8.5 Hz, 1H), 7.09 (t, J = 5.1 Hz, 1H), 5.57 (s, 2H), 3.34 (s,
5H), 2.47 (s, 3H), 1.55 (quint, J = 7.3 Hz, 2H), 1.36 (sext, J = 7.3 Hz,
2H), 0.91 (t, J = 7.3 Hz, 3H); HRMS (ESI): m/z: calcd for
C20H23N7O3H+: 410.1935, found: 410.1937.
8-(Butylamino)-7-(2-(butylimino)-2-(o-tolyl)ethyl)-3-methyl-1H-
purine-2,6(3H,7H)-dione (31). White solid; 1H NMR (500 MHz,
DMSO-d6): δ = 10.53 (s, 1H), 7.31−7.25 (m, 4H), 6.90 (t, J = 5.6 Hz,
1H), 4.98 (d, J = 18.2 Hz, 1H), 4.85 (d, J = 18.2 Hz, 1H), 3.33 (s, 2H),
3.29 (s, 3H), 3.03−2.98 (m, 1H), 2.83−2.78 (m, 1H), 2.17 (s, 3H),
1.57−1.51 (m, 2H), 1.39−1.32 (m, 2H), 1.28−1.23 (m, 2H), 1.16−
1.00 (m, 2H), 0.91 (t, J = 7.3 Hz, 3H), 0.69 (t, J = 7.3 Hz, 3H); HRMS
(ESI): m/z: calcd for C23H32N6O2H+: 425.2660, found: 425.2661.
8-Butyl-1-methyl-7-(o-tolyl)-1H-imidazo[2,1-f]purine-2,4(3H,8H)-
dione (40). White solid; yield: 32% over two steps; mp 258−260 °C;
1H NMR (500 MHz, DMSO-d6): δ = 10.94 (s, 1H), 7.62 (s, 1H),
J
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
8-Butyl-1-methyl-7-(5-methyl-1H-indazol-4-yl)-1H-imidazo[2,1-
f ]purine-2,4(3H,8H)-dione (50). 8-(Butylamino)-3-methyl-7-(2-(5-
methyl-1H-indazol-4-yl)-2-oxoethyl)-1H-purine-2,6(3H,7H)-dione
(39, 40 mg, 0.09 mmol) was heated in a sealed tube at 175 °C in
CH2Cl2 (4 mL) in the presence of BF3·OEt2 (0.4 mL, 3.24 mmol) for
30 min. The reaction was poured in water and extracted with CH2Cl2,
and the organic layer was dried over MgSO4, filtered, and concentrated
under reduced pressure. Purification by column chromatography on
silica gel (gradient EtOAc:MeOH 99:1 to 98:2) afforded the desired
P
= (dCr/dt)×V/(A × CA)
app
r
(1)
percent recovery = 100 × ((V ×Crfinal) + (V ×Cdfinal))/
r
d
(V ×C0)
(2)
d
where, dCr /dt is the slope of the cumulative concentration in the
receiver compartment versus time in μM·s−1, Vr is the volume of the
receiver compartment in cm3, Vd is the volume of the donor
compartment in cm3, A is the area of the cell monolayer (1.13 cm2 for
12-well Transwell), C0 is the nominal concentration of the dosing
solution in micromolar, CA is the average of the nominal concentration
of the dosing solution and the measured donor concentration at 120
1
compound as a white solid (1.1 mg, 3% yield). H NMR (400 MHz,
DMSO-d6): δ = 13.21 (s, 1H), 10.93 (s, 1H), 7.83 (s, 1H), 7.69 (s,
1H), 7.63 (d, J = 8.5 Hz, 1H), 7.39 (d, J = 8.5 Hz, 1H), 3.77 (t, J = 7.3
Hz, 2H), 3.43 (s, 3H), 2.31 (s, 3H), 1.48 (quint, J = 7.3 Hz, 2H), 0.98
(sext, J = 7.3 Hz, 2H), 0.55 (t, J = 7.3 Hz, 3H); HRMS (ESI): m/z:
calcd for C20H21N7O2Na+: 414.1649, found: 414.1646.
min in μM, Crfinal is the cumulative receiver concentration in
final
micromolar at the end of the incubation period, and Cd
is the
FRET-Based Enzymatic Assay. Compounds were tested in the
Z′-LYTE Kinase Assay Kit−Tyr 1 Peptide (Invitrogen, USA) in a
Corning 384 well microtiter plate. Fluorescence progress curves were
measured upon excitation at 400 nm and emission at 445 and 520 nm.
The assay contained a final concentration of EphB4 and ATP of 25
ng/μL and 125 μM (which is near its Km), respectively, and was run at
room temperature for 2 h. IC50 values (inhibitor concentration at
which enzyme activity is reduced by 50%) are determined after
carrying out assays at 10 different concentrations between 20 μM and
10 pM.
[γ-33P]ATP-Based Enzymatic Assay. The enzymatic assays for
the selectivity profile were performed at the National Centre for
Protein Kinase Profiling at the University of Dundee. All assays (25.5
μL volume) were carried out robotically at room temperature and
were linear with respect to time and enzyme concentration under the
conditions used. Assays were performed for 30 min using Multidrop
Micro reagent dispensers (Thermo Electron Corporation, Waltham,
MA) in a 96-well format. The concentration of magnesium acetate in
the assays was 10 mM and [γ-33P]ATP (800 cpm/pmol) was used at 5,
20, or 50 μM to be at or below the Km for ATP for each kinase. The
assays were initiated with MgATP, stopped by the addition of 5 μL of
0.5 M orthophosphoric acid, and spotted onto P81 filter plates using a
unifilter harvester (PerkinElmer, Boston, MA). The data is presented
as mean percentage activity of duplicate assays at single concentration
compared to DMSO controls. A similar protocol was used at Reaction
Biology Corporation to measure the inhibitory activity of compounds
3, 40, 43, and 50 against Abl, Src, Lck, and Yes1.
concentration of the donor in micromolar at the end of the incubation
period.
Monolayer Assay. Cell lines were routinely passaged once or
twice weekly and maintained in culture for up to 20 passages. All cells
were cultured in RPMI 1640 medium supplemented with 10% (v/v)
fetal calf serum and 0.1 mg/mL gentamicin (medium and all
components from PAA, Colbe, Germany) at 37 °C in a humidified
̈
atmosphere with 5% CO2.
A modified propidium iodide (PI) assay was used to assess the
anticancer activity of the compounds. Briefly, cells were harvested from
exponential phase cultures, counted, and plated in 96-well flat-bottom
microtiter plates at a cell density of 4000−20 000 cells per well. After a
24 h recovery period, allowing the cells to resume exponential growth,
10 μL of culture medium (four control wells/plate) and culture
medium with the test compound were added by the liquid handling
robotic system and treatment was continued for four days. The
compounds were applied in half log increments at 10 concentrations in
duplicate. Next, cells were washed with 200 μL of PBS to remove dead
cells. Subsequently, 200 μL of a solution containing 7 μg/mL
propidium iodide (PI) and 0.1% (v/v) Triton X-100 was added. After
an incubation period of 1−2 h at room temperature, fluorescence
(FU) was measured using the Cytofluor 4000 microplate reader
(excitation λ = 530 nm, emission λ = 620 nm) to quantify the amount
of attached viable cells. For calculations, the mean value of duplicate/
quadruplicate (untreated control) data was used. Quality criteria for a
successful assay included fluorescence intensity signal of >500 units
from the untreated control wells.
Cellular Phosphorylation Assays. The following experiments
were performed at ProQinase GmbH. Mouse embryonal fibrablast
cells, stably transfected to overexpress full-length human EphB4, were
plated at 4 × 104 cells/well in DMEM supplemented with 10% FCS in
48-well culture dishes. Medium was replaced by DMEM without FCS
before test compounds prediluted in 100% DMSO were added (final
DMSO concentration of 1%). After incubation for 90 min at 37 °C,
cells were stimulated for 2 h at 4 °C using murine ephrinB2-Fc at a
final concentration of 2 μg/mL. Quantification of EphB4 phosphor-
ylation was assessed in a 96-well plate via sandwich ELISA using a myc
capture antibody and an antiphosphotyrosine detection antibody.
Cellular Permeability Assays. The following experiments were
performed at Absorption Systems. Caco-2 monolayers were grown to
confluence on collagen-coated, microporous, polycarbonate mem-
branes in 12-well Costar Transwell plates. The permeability assay
buffer for the donor chambers was Hanks Balanced Salt Solution
containing 10 mM HEPES and 15 mM glucose at a pH of 7.4. The
buffer in the receiver chambers also contained 1% bovine serum
albumin. Cells were dosed on the apical side (A-to-B) or basolateral
side (B-to-A) and incubated at 37 °C with 5% CO2 in a humidified
incubator. After 2 h, aliquots were taken from the donor and receiver
chambers. Each determination was performed in duplicate. The flux of
lucifer yellow was also measured for each monolayer after being
subjected to the test compounds to ensure that no damage was
inflicted to the cell monolayers during the flux period. All samples
were assayed by LC-MS/MS using electrospray ionization. The
apparent permeability, Papp, and percent recovery were calculated as
follows:
X-ray Crystallography. The atomic coordinates and structure
factors of EphA3 in complex with the inhibitors 3, 40, and 50 have
been deposited with the Protein Data Bank as entries 4GK2, 4GK3,
and 4GK4, respectively.
Protein Expression and Purification. A clone of the EphA3 kinase
domain (residues: 606−947) was obtained from Prof. Sirano Dhe-
Paganon’s group29 and expressed in E. coli strain BL21 (DE3). Cells
expressing EphA3 were induced with a 1 mM solution of isopropyl β-
D-thiogalactopyranoside (IPTG) for 12 h at 15 °C. Cell pellets were
resuspended in buffer A (50 mM Tris, pH 8.0, and 100 mM NaCl,
supplemented with protease inhibitors) and lysed by sonication. After
centrifugation at 15 000 rpm for 1 h, the soluble fraction of EphA3 was
purified using HisTrap FF crude and HiTrap Q HP columns (GE
Healthcare), followed by gel filtration chromatography (Superdex75;
GE Healthcare). The appropriate fractions were combined and
concentrated to ∼10 mg/mL using Amicon filter devices (10 kDa as
cutoff) in a storage solution (100 mM sodium chloride and 10 mM
Tris-HCl pH 8.0, 5% glycerol). The resulting solution was aliquoted
and stored at −80 °C for further usage.
Crystallization, Data Collection, and Structure Determination.
Crystals of the EphA3 kinase domain were grown at 20 °C using the
hanging drop vapor diffusion method. Equal volumes of protein and
reservoir solutions (0.1 M sodium cacodylate pH 6.5, 0.15 M
ammonium sulfate, 22.5% PEG 3350) were mixed, and crystals
appeared after 1 to 2 days. A 5 mM solution of inhibitor (in 100%
DMSO) was added to the hanging drop to reach a final DMSO
concentration of 10% (v/v). The crystals were soaked for 1 to 24 h
and flash-frozen in liquid nitrogen without extra cryoprotectant.
K
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Data sets were collected on a MarCCD detector and indexed,
integrated, and scaled with the XDS31 and CCP4 programs.32 The
structures were solved by molecular replacement with PHASER33
using the apo EphA3 kinase domain structure (PDB entry 2GSF) as a
search model and refined with PHENIX.34
The Simulated Annealing Composite OMIT map (Supporting
Information), in which compound 3 was omitted from structure factor
calculation, was generated in a region within 1.6 Å of compound 3
using PHENIX and Pymol (http://www.pymol.org).
(4) Bardelle, C.; Cross, D.; Davenport, S.; Kettle, J. G.; Ko, E. J.;
Leach, A. G.; Mortlock, A.; Read, J.; Roberts, N. J.; Robins, P.;
Williams, E. J. Inhibitors of the tyrosine kinase EphB4. Part 1:
Structure-based design and optimization of a series of 2,4-bis-
anilinopyrimidines. Bioorg. Med. Chem. Lett. 2008, 18, 2776−2780.
(5) Miyazaki, Y.; Nakano, M.; Sato, H.; Truesdale, A. T.; Stuart, J. D.;
Nartey, E. N.; Hightower, K. E.; Kane-Carson, L. Design and effective
synthesis of novel templates, 3,7-diphenyl-4-amino-thieno and furo-
[3,2-c]pyridines as protein kinase inhibitors and in vitro evaluation
targeting angiogenetic kinases. Bioorg. Med. Chem. Lett. 2007, 17, 250−
254.
(6) Lafleur, K.; Huang, D.; Zhou, T.; Caflisch, A.; Nevado, C.
Structure-based optimization of potent and selective inhibitors of the
tyrosine kinase erythropoietin producing human hepatocellular
carcinoma receptor B4 (EphB4). J. Med. Chem. 2009, 52, 6433−6446.
(7) Zuccotto, F.; Ardini, E.; Casale, E.; Angiolini, M. Through the
“gatekeeper door”: exploiting the active kinase conformation. J. Med.
Chem. 2010, 53, 2681−2694.
ASSOCIATED CONTENT
■
S
* Supporting Information
Selectivity data of compound 40, antiproliferative activity of
1
compound 3, H NMR and 13C NMR of selected compounds,
and HPLC traces (for purity) of tested compounds and
simulated annealing composite OMIT map. This material is
available free of charge via the Internet at http://pubs.acs.org.
(8) Mitchell, S. A.; Danca, M. D.; Blomgren, P. A.; Darrow, J. W.;
Currie, K. S.; Kropf, J. E.; Lee, S. H.; Gallion, S. L.; Xiong, J. M.;
Pippin, D. A.; DeSimone, R. W.; Brittelli, D. R.; Eustice, D. C.;
Bourret, A.; Hill-Drzewi, M.; Maciejewski, P. M.; Elkin, L. L.
Imidazo[1,2-a]pyrazine diaryl ureas: inhibitors of the receptor tyrosine
kinase EphB4. Bioorg. Med. Chem. Lett. 2009, 19, 6991−6995.
(9) Noberini, R.; Lamberto, I.; Pasquale, E. B. Targeting Eph
receptors with peptides and small molecules: progress and challenges.
Semin. Cell. Dev. Biol. 2012, 23, 51−57.
AUTHOR INFORMATION
Corresponding Author
*Phone: (41) 446353945. Fax: (41) 446353948. E-mail:
nevado@oci.uzh.ch, caflisch@bioc.uzh.ch.
■
Author Contributions
#These two authors contributed equally to this work.
Notes
(10) Kolb, P.; Kipouros, C. B.; Huang, D.; Caflisch, A. Structure-
based tailoring of compound libraries for high-throughput screening:
discovery of novel EphB4 kinase inhibitors. Proteins 2008, 73, 11−18.
(11) Bamborough, P.; Angell, R. M.; Bhamra, I.; Brown, D.; Bull, J.;
Christopher, J. A.; Cooper, A. W.; Fazal, L. H.; Giordano, I.; Hind, L.;
Patel, V. K.; Ranshaw, L. E.; Sims, M. J.; Skone, P. A.; Smith, K. J.;
Vickerstaff, E.; Washington, M. N-4-Pyrimidinyl-1H-indazol-4-amine
inhibitors of Lck: indazoles as phenol isosteres with improved
pharmacokinetics. Bioorg. Med. Chem. Lett. 2007, 17, 4363−4368.
(12) Jeong, E. J.; Liu, X.; Jia, X.; Chen, J.; Hu, M. Coupling of
conjugating enzymes and efflux transporters: impact on bioavailability
and drug interactions. Curr. Drug Metab. 2005, 6, 455−468.
(13) Manley, P. W.; Drueckes, P.; Fendrich, G.; Furet, P.; Liebetanz,
J.; Martiny-Baron, G.; Mestan, J.; Trappe, J.; Wartmann, M.; Fabbro,
D. Extended kinase profile and properties of the protein kinase
inhibitor nilotinib. Biochim. Biophys. 2010, 1804, 445−453.
(14) Auffinger, P.; Hays, F. A.; Westhof, E.; Ho, P. S. Halogen bonds
in biological molecules. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 16789−
16794.
(15) Huber, K.; Brault, L.; Fedorov, O.; Gasser, C.; Filippakopoulos,
P.; Bullock, A. N.; Fabbro, D.; Trappe, J.; Schwaller, J.; Knapp, S.;
Bracher, F. 7,8-dichloro-1-oxo-β-carbolines as a versatile scaffold for
the development of potent and selective kinase inhibitors with unusual
binding modes. J. Med. Chem. 2012, 55, 403−413.
(16) Ishikawa, M.; Hashimoto, Y. Improvement in aqueous solubility
in small molecule drug discovery programs by disruption of molecular
planarity and symmetry. J. Med. Chem. 2011, 54, 1539−1554.
(17) Maier, J. A.; Brugel, T. A.; Sabat, M.; Golebiowski, A.;
Laufersweiler, M. J.; VanRens, J. C.; Hopkins, C. R.; De, B.; Hsieh, L.
C.; Brown, K. K.; Easwaran, V.; Janusz, M. J. Development of N-4,6-
pyrimidine-N-alkyl-N′-phenyl ureas as orally active inhibitors of
lymphocyte specific tyrosine kinase. Bioorg. Med. Chem. Lett. 2006,
16, 3646−3650.
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are grateful to Maja Bollhalder, the late Sara Savaresi,
Sabrina Sonda, and Changchuan Xie for technical help in a
preliminary phase of this work. We also thank Xiao-Dan Li and
Chitra Rajendran for interesting discussions and help with data
collection. Prof. Sirano Dhe-Paganon is kindly acknowledged
for providing the plasmid for EphA3 expression. This work was
supported financially in part by a grant of the Sino-Swiss
program to A.C. and by the Forschungskredit of the University
of Zurich to K.L.
̈
ABBREVIATIONS: USED
■
Abl, abelson murine leukemia viral oncogene homologue; ATP,
adenosine triphosphate; DCM, dichloromethane; DFG,
aspartate-phenylalanine-glycine; DIPEA, diisopropylethyl-
amine; DMEM, Dulbecco’s Modified Eagle Medium; DMF,
dimethylformamide; DMSO, dimethyl sulfoxide; E. coli,
Escherichia coli; ELISA, enzyme-linked immunosorbent assay;
Eph, erythropoietin-producing human hepatocellular carcinoma
receptor; FCS, fetal calf serum; FRET, fluorescence-resonance
energy transfer; Lck, lymphocyte-specific kinase; MD, molec-
ular dynamics; Papp, apparent permeability; PBS, phosphate-
buffered saline; PEG, polyethylene glycol; PCC, pyridinium
chlorochromate; SAR, structure−activity relationship; RIPK2,
receptor interacting serine-threonine protein kinase 2; THF,
tetrahydrofuran
REFERENCES
(18) Sabat, M.; VanRens, J. C.; Laufersweiler, M. J.; Brugel, T. A.;
Maier, J.; Golebiowski, A.; De, B.; Easwaran, V.; Hsieh, L. C.; Walter,
R. L.; Mekel, M. J.; Evdokimov, A.; Janusz, M. J. The development of
2-benzimidazole substituted pyrimidine based inhibitors of lymphocyte
specific kinase (Lck). Bioorg. Med. Chem. Lett. 2006, 16, 5973−5977.
(19) Zagorska, A.; Jurczyk, S.; Pawlowski, M.; Dybala, M.; Nowak,
G.; Tatarczynska, E.; Nikiforuk, A.; Chojnacka-Wojcik, E. Synthesis
and preliminary pharmacological evaluation of imidazo[2,1-f]purine-
2,4-dione derivatives. Eur. J. Med. Chem. 2009, 44, 4288−4296.
■
(1) Pasquale, E. B. Eph receptor signalling casts a wide net on cell
behaviour. Nat. Rev. Mol. Cell. Biol. 2005, 6, 462−475.
(2) Adams, R. H. Vascular patterning by Eph receptor tyrosine
kinases and ephrins. Semin. Cell. Dev. Biol. 2002, 13, 55−60.
(3) Martiny-Baron, G.; Holzer, P.; Billy, E.; Schnell, C.; Brueggen, J.;
Ferretti, M.; Schmiedeberg, N.; Wood, J. M.; Furet, P.; Imbach, P. The
small molecule specific EphB4 kinase inhibitor NVP-BHG712 inhibits
VEGF driven angiogenesis. Angiogenesis 2010, 13, 259−267.
L
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(20) Baraldi, P. G.; Preti, D.; Tabrizi, M. A.; Fruttarolo, F.;
Romagnoli, R.; Zaid, N. A.; Moorman, A. R.; Merighi, S.; Varani, K.;
Borea, P. A. New pyrrolo[2,1-f]purine-2,4-dione and imidazo[2,1-
f]purine-2,4-dione derivatives as potent and selective human A3
adenosine receptor antagonists. J. Med. Chem. 2005, 48, 4697−4701.
(21) Baraldi, P. G.; Preti, D.; Tabrizi, M. A.; Romagnoli, R.;
Saponaro, G.; Baraldi, S.; Botta, M.; Bernardini, C.; Tafi, A.;
Tuccinardi, T.; Martinelli, A.; Varani, K.; Borea, P. A. Structure-
activity relationship studies of a new series of imidazo[2,1-f]purinones
as potent and selective A(3) adenosine receptor antagonists. Bioorg.
Med. Chem. 2008, 16, 10281−10294.
(22) Messaoudi, S.; Treguier, B.; Hamze, A.; Provot, O.; Peyrat, J. F.;
De Losada, J. R.; Liu, J. M.; Bignon, J.; Wdzieczak-Bakala, J.; Thoret,
S.; Dubois, J.; Brion, J. D.; Alami, M. Isocombretastatins A versus
combretastatins A: The forgotten isoCA-4 isomer as a highly
promising cytotoxic and antitubulin agent. J. Med. Chem. 2009, 52,
4538−4542.
(23) Alcalde, E.; Mesquida, N.; Frigola, J.; Lopez-Perez, S.; Merce, R.
Indene-based scaffolds. Design and synthesis of novel serotonin 5-
HT6 receptor ligands. Org. Biomol. Chem. 2008, 6, 3795−3810.
(24) Takami, A.; Iwakubo, M.; Okada, Y.; Kawata, T.; Odai, H.;
Takahashi, N.; Shindo, K.; Kimura, K.; Tagami, Y.; Miyake, M.;
Fukushima, K.; Inagaki, M.; Amano, M.; Kaibuchi, K.; Iijima, H.
Design and synthesis of Rho kinase inhibitors (I). Bioorg. Med. Chem.
2004, 12, 2115−2137.
(25) Bakke, B. A.; McIntosh, M. C.; Turnbull, K. D. Improved
alkylation and product stability in phosphotriester formation through
quinone methide reactions with dialkyl phosphates. J. Org. Chem.
2005, 70, 4338−4345.
(26) Kumar, S. R.; Singh, J.; Xia, G.; Krasnoperov, V.; Hassanieh, L.;
Ley, E. J.; Scehnet, J.; Kumar, N. G.; Hawes, D.; Press, M. F.; Weaver,
F. A.; Gill, P. S. Receptor tyrosine kinase EphB4 is a survival factor in
breast cancer. Am. J. Pathol. 2006, 169, 279−293.
(27) Stephenson, S. A.; Slomka, S.; Douglas, E. L.; Hewett, P. J.;
Hardingham, J. E. Receptor protein tyrosine kinase EphB4 is up-
regulated in colon cancer. BMC Mol. Biol. 2001, 2, 15.
(28) Castellano, G.; Reid, J. F.; Alberti, P.; Carcangiu, M. L.;
Tomassetti, A.; Canevari, S. New potential ligand-receptor signaling
loops in ovarian cancer identified in multiple gene expression studies.
Cancer Res. 2006, 66, 10709−10719.
(29) Choi, Y.; Syeda, F.; Walker, J. R.; Finerty, P. J., Jr.; Cuerrier, D.;
Wojciechowski, A.; Liu, Q.; Dhe-Paganon, S.; Gray, N. S. Discovery
and structural analysis of Eph receptor tyrosine kinase inhibitors.
Bioorg. Med. Chem. Lett. 2009, 19, 4467−4470.
(30) Huang, D.; Zhou, T.; Lafleur, K.; Nevado, C.; Caflisch, A.
Kinases selectivity potential for inhibitors targeting the ATP binding
site: a network analysis. Bioinformatics 2010, 26, 198−204.
(31) Kabsch, W. Automatic processing of rotation diffraction data
from crystals of initially unknown symmetry and cell constants. J. Appl.
Crystallogr. 1993, 26, 795−800.
(32) The CCP4 suite: programs for protein crystallography. Acta
Crystallogr., Sect. D: Biol. Crystallogr. 1994, 50, 760-763.
(33) 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.
(34) Adams, P. D.; Grosse-Kunstleve, R. W.; Hung, L. W.; Ioerger, T.
R.; McCoy, A. J.; Moriarty, N. W.; Read, R. J.; Sacchettini, J. C.;
Sauter, N. K.; Terwilliger, T. C. PHENIX: building new software for
automated crystallographic structure determination. Acta Crystallogr.,
Sect. D: Biol. Crystallogr. 2002, 58, 1948−1954.
M
dx.doi.org/10.1021/jm301187e | J. Med. Chem. XXXX, XXX, XXX−XXX