Inhibitors of the tyrosine kinase EphB4. Part 1: Structure-based design and optimization of a series of 2,4-bis-anilinopyrimidines

Article abstract of DOI:10.1016/j.bmcl.2008.04.015

A series of bis-anilinopyrimidines have been identified as potent inhibitors of the tyrosine kinase EphB4. Structural information from two alternative series identified from screening efforts was combined to identify the initial leads.

Full text of DOI:10.1016/j.bmcl.2008.04.015

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 2776–2780
Inhibitors of the tyrosine kinase EphB4. Part 1:
Structure-based design and optimization of a series
of 2,4-bis-anilinopyrimidines
Catherine Bardelle, Darren Cross, Sara Davenport, Jason G. Kettle,* Eun Jung Ko,
Andrew G. Leach, Andrew Mortlock, Jon Read, Nicola J. Roberts,
Peter Robins and Emma J. Williams
AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK
Received 25 March 2008; revised 4 April 2008; accepted 6 April 2008
Available online 10 April 2008
Abstract—A series of bis-anilinopyrimidines have been identified as potent inhibitors of the tyrosine kinase EphB4. Structural infor-
mation from two alternative series identified from screening efforts was combined to identify the initial leads.
Ó 2008 Elsevier Ltd. All rights reserved.
The erythropoietin-producing hepatoma amplified se-
quence (Eph) family is the largest known family of
receptor tyrosine kinases, with 14 receptors and 8 cog-
nate ephrin ligands identified.1 There is growing evi-
dence that Eph receptor signalling may contribute to
tumorigenesis in a wide variety of human cancers, either
on tumour cells directly, or indirectly via modulation of
vascularisation. Many Eph receptors are over-expressed
in various tumour types2,3 and expression of EphB4 is
up-regulated in tumours such as neuroblastomas, leu-
kaemia’s, breast, liver, lung and colon. Moreover, vari-
ous in vitro and in vivo studies particularly regarding
EphB4 have indicated that over-expression of Eph
receptors on cancer cells is able to confer tumourigenic
phenotypes such as proliferation and invasion, consis-
tent with the speculated role in oncogenesis.
have shown that disruption of EphB47 or its ligand eph-
rinB28 causes embryonic lethality associated with vascu-
lar modelling defects consistent with a critical role in
vessel development. EphB4 activation also stimulates
endothelial cell proliferation and migration in vitro9
and inhibition of EphB4 signalling using soluble extra-
cellular domains of EphB4 has been shown to inhibit tu-
mour growth and angiogenesis in in vivo xenograft
studies.10,11
O
O
HN
O
O
Cl
N
Inhibition of EphB4 expression using interfering-RNA
or antisense oligodeoxynucleotides inhibited prolifera-
tion, survival and invasion of PC3 prostate cancer cells
in vitro and in vivo.4 In addition, there is good evidence
that EphB4 may contribute to tumour vascularisa-
tion.5,6 Members of the Eph family including EphB4
are expressed on endothelial cells and transgenic studies
N
N
1
O
O
HN
N
Cl
O
N
N
O
N
Keywords: Kinase; EphB4; Structure-based design; Lead identification.
*
H
Corresponding author. Tel.: +44 1625517920; fax: +44 1625586707;
e-mail: jason.kettle@astrazeneca.com
2
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.04.015

C. Bardelle et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2776–2780
2777
In our search for inhibitors of EphB4 tyrosine kinase,
preliminary screening12 of libraries of kinase inhibitors
highlighted compound 1, a known anilinoquinazoline
inhibitor of the c-Src tyrosine kinase from in-house re-
search13 as a potent inhibitor, and also 2, a modestly po-
tent bis-anilinotriazine, analogues of which have
previously been reported as kinase inhibitors.14 Deter-
mination of the binding modes of these two inhibitors
was achieved by soaking into the catalytic domain of
EphB4 (D598-G892 with a Y774E mutation)15 and re-
vealed as expected that both occupied the ATP-binding
site. Compound 1 forms a key hydrogen bond between
the quinazoline N-1 and Met696 (Fig. 1(a)), while com-
pound 2 presents a donor–acceptor pair to Met696 (Fig.
1(b)). Intriguingly, protein overlay of these two struc-
tures (Fig. 1(c)) highlighted exact convergence of the
heterocyclic hinge-binding groups, with the different
C-4 anilino groups opposed. The dioxole aniline of the
more potent 1 is buried in the selectivity pocket, the
hydrophobic pocket beyond the Thr693 gatekeeper res-
idue, whereas the aniline of 2 points towards the ribose
binding site,16 suggesting that a hybrid of the aniline of 1
and the core of 2 may also result in potent inhibition of
EphB4.
In the event, we elected to examine both triazine and
pyrimidine core hybrids. Synthesis of these analogues
is outlined in Scheme 1, and is representative of the syn-
thesis of related analogues prepared in the course of this
study. Accordingly, both 2,4- and 4,6-dichloropyrimi-
dine, and 2,4-dichlorotriazine 3 were reacted with 5-
chlorobenzo[1,3]dioxol-4-amine to give intermediates
4, which in turn were reacted with 3,4,5-trimethoxyani-
line at elevated temperatures to give the target regioiso-
meric pyrimidines 6 and 7, and triazine 5.
The compounds listed in Table 1 were evaluated in an
EphB4 kinase assay measuring the inhibition of phos-
phorylation of a synthetic peptide substrate at Km
ATP concentration.17 The dioxole-containing triazine
5 and 4,6-pyrimidine isomer 6 both showed a modest
improvement in activity compared with hit 2, although
we did not have structural confirmation that this was
due to any shift in the aniline orientation. Gratifyingly,
2,4-pyrimidine analogue 7 showed potent activity at a
level comparable to quinazoline hit 1, and Figure 1(d)
shows the structure of this compound bound to the cat-
alytic domain of EphB4 confirming indeed that the C-4
aniline has re-orientated towards the selectivity pocket
Figure 1. (a) Structure of anilinoquinazoline 1 bound to the catalytic domain of EphB4. (b) Structure of bis-anilinotriazine 2 bound to the catalytic
domain of EphB4. (c) Overlay of the structures of 1 and 2 obtained through alignment of protein structure. (d) Structure of bis-anilinopyrimidine 7,
prepared on the basis of this alignment, bound to the catalytic domain of EphB4.

2778
C. Bardelle et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2776–2780
When the ligands are constrained to the conformation
O
required for binding, the hydrogen bonding acceptor
ability (logKb values)19 of the nitrogen that forms the
primary interaction with the hinge backbone can be esti-
mated by calculation.20 These calculations indicate that
quinazoline 1 and 2,4-pyrimidine 7 both have conforma-
tional energies which are closer to the global minimum,
and demonstrate the most effective hydrogen bonding
capability when compared to triazine 5 and 4,6-pyrimi-
dine 6, and this would appear to be reflected in the
experimentally observed enzyme data. It is noteworthy
that 4,6-pyrimidine 6 has a conformational energy
above the global minimum that is approximately the
same as that calculated for the more active 2,4-pyrimi-
dine 7 (both 0.8 kcal/mol), that is until the anilines at
the 6-position (for 6) and 2-position (in the case of 7)
are constrained to the observed dihedral angles. This
produces steric repulsion in the 4,6-pyrimidine that is
not present in the 2,4-pyrimidine due to the adjacent
nitrogen in the pyrimidine core. The hydrogen bonding
ability of N-1 is also dependent on this conformation
and before the constraint is imposed, the 2,4- and 4,6-
isomers have rather similar logKb values (2.05 and
2.14, respectively).
O
Cl
N
HN
Cl
(a)
X
Y
X
Y
Cl
Cl
N
3 X, Y = N or CH
4 X, Y = N or CH
O
(b)
O
O
HN
5 X, Y = N
O
O
Cl
6 X = CH, Y = N
7 X = N, Y = CH
X
Y
N
H
N
Scheme 1. Synthesis of bis-anilinotriazines and pyrimidines. Reagents
and conditions: (a) 5-chlorobenzo[1,3]dioxol-4-amine,13 NaH, DMA, 0–
25 °C; (b) 3,4,5-trimethoxyaniline, HCl in Et2O,n-BuOH, DMA, 120 °C.
Since we considered that the trimethoxyaniline repre-
sented a potential DMPK liability, we next elected to
examine alternative anilines at the C-2 position, and in
particular the electronic, steric and positional effects of
various substituents. Table 2 highlights the systematic
examination of these factors, whilst keeping the C-4 ani-
line fixed. As expected, the three methoxy substituents in
7 showed a key contribution to the observed potency,
with the simple aniline analogue 8 showing a 6-fold drop
in activity. All of the substituents examined when placed
in the 2-position resulted in loss of activity relative to 8.
This effect was particularly dramatic for electron-with-
drawing substituents such as cyano, sulfonamide and
sulfone (compounds 21, 24 and 27) but was observed
for simple substituents such as F and methoxy (com-
pounds 9 and 18). It is clear from the crystal structures
in Figure 1((b) and (d)) that preference exists for the C-2
aniline to be co-planar with the central heterocyclic ring,
allowing the key NH-donor to be presented to the hinge
region in the correct orientation whilst allowing optimal
electron delocalisation in the ligand itself, and any
ortho-substitution might be expected to adversely im-
pact this. When comparing triads of the same substitu-
ent across 2-, 3- and 4-positions it is evident that
potency is significantly enhanced at the 3-position, with
the activity order 3 < 4 ꢀ 2 maintained for all substitu-
ents (compare 9–11, 12–14, etc.). All the 3-substituents
examined showed enhancements in activity relative to
unsubstituted parent 8, although the most potent activ-
ity was observed for the bulky, strongly electron-with-
drawing sulfonamide (25) and sulfone (28) groups.
Importantly, these demonstrated potency improvements
over initial hybrid 7, indicating that a strongly electron-
rich C-2 aniline is not a requirement for activity. In
addition, a range of both large and small, electron-
donating and -withdrawing substituents furnished sub-
micromolar activity at this position including methyl
(16), methoxy (19), cyano (22) and morpholinyl (31).
Table 1. EphB4 inhibition data for compounds 1, 2 and 5–7 and
conformational energies and computed hydrogen bond acceptor
abilities of the different cores
Entry Core
EphB4 IC50 Relative energy Calculated
(lM)a
of bound
logKb
or N-1
conformation
and global
minimum
(kcal/mol)
1
2
5
6
7
Quinazoline
Triazine
Triazine
0.35 0.13 0.7
16 —
2.96 0.71 1.5
2.65
—
1.42
1.72
2.12
4,6-Pyrimidine 2.54 0.35 2.8
2,4-Pyrimidine 0.24 0.16 0.9
a For determinations where n P 2, standard deviation is given.
as hypothesised. It is interesting to note that this most
active of the hybrids has a CH group at C-5 which in
both 1 and 2 is N.
The combined effect of the hydrogen bonding ability of
the different cores and their innate conformational pref-
erences can rationalize these observations. The crystal
structures of both 1 and 7 reveal that the dioxole ring
sits in the selectivity pocket and consequently demands
a conformation of the aniline ring whose energy is rea-
sonably close to the global minimum energy of the li-
gand. A further conformational restriction applies to
the pyrimidine and triazine cores because the region
adjacent to the hinge in which they form their principal
interactions requires a relatively flat ligand, such as that
found for quinazoline 1 or indeed the adenine of ATP.
The ligand-bound structures reveal the dihedral angles
that are preferred in the protein complex, and when
these are applied as constraints in quantum mechanical
calculations, the energies relative to presumed global
minima are given in Table 1.18

C. Bardelle et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2776–2780
Table 2. EphB4 inhibition data for C-2 aniline variants
2779
O
O
HN
3
Cl
4
2
N
N
H
N
a
Substituent
EphB4 IC50 (lM)
Entry
2-
Entry
3-
Entry
4-
H
8
10
13
16
19
22
25
28
31
1.55 0.07
1.15 0.07
1.33 0.79
0.46 0.04
0.64 0.07
0.55 0.15
0.04 0.02
0.09 0.05
0.58 0.12
F
Cl
9
12
15
18
21
24
27
30
5.00 0.28
11.32 0.58
13.96 1.41
3.88 0.57
20.49 0.71
35.50 0.71
41.00 0.00
7.85 3.47
11
14
17
20
23
26
29
32
2.30 0.10
6.00 1.00
1.60 0.00
1.94 0.21
1.25 0.07
0.30 0.12
0.63 0.01
1.00 0.00
Me
MeO
CN
SO2NH2
SO2Me
1-Morpholinyl
a For determinations where n P 2, standard deviation is given.
Examination of substituent effects at the 4-position indi-
cated these inhibitors to be less effective than those with
the same substituents at the 3-position, generally at least
2-fold but greater for some examples. It is interesting to
note that the most potent substituents in the 4-position
mirror those in the 3-position, namely the sulfonamide
(26) and sulfone (29) both show sub-micromolar activity
comparable with lead 7. Halogen substituents (F,Cl)
were seen to be either poorly tolerated or offer no benefit
over the unsubstituted parent in all positions (com-
pounds 9–11 and 12–14).
On the basis of these data we chose to examine the 3-sul-
fone 28 in greater detail. This compound was examined
in a cellular assay of EphB4 inhibition,21 and was found
to inhibit phosphorylation with an IC50 of
0.19 0.07 lM, in good agreement with its enzyme inhi-
bition value. In a panel of kinase assays, 28 was shown
to be remarkably selective. Table 3 highlights the data
from this selectivity panel. Compound 28 is inactive or
weakly active against most of the kinases tested. Only
the EphB2 and Src enzymes are potently inhibited,
and this is expected both on the basis of high sequence
homology to EphB4 (85% and 44%, respectively, in
the kinase domains) and the presence of a C-4 aniline
previously optimised for potent Src inhibition in the qui-
nazoline series.13 Table 4 summarises selected physico-
chemical and pharmacokinetic parameters in the rat
for compound 28. logD is acceptable, and solubility is
modest with a reasonably high degree of protein bind-
ing, although not prohibitively so, as a lead for further
optimisation. The in vivo profile in rat is also encourag-
ing, showing a modest but acceptable bioavailability of
30%, and low clearance, potentially indicating solubility
limited absorption.
Table 3. Selectivity data for compound 28
Kinase
IC50 (lM)
EphB2
Src
0.05
<0.10
10.7
17.6
19.7
20.5
27.1
>30
>30
>30
>30
>30
>30
>30
>30
>30
CSK
CDK2
FAK
JAK3
p38
IGF
ZAP70
EGF
KDR
FGF
PLK
CHK1
JNK
PKA
In summary, we have used structural information from
two distinct series bound in the active site of EphB4 to
Table 4. Selected physicochemical and rat pharmacokinetic dataa for compound 28
logD
(pH 7.4)
Solubilityb
Rat PPB%
free drugc
Bioavailability
(%)
Cl
(ml/min/kg)
Vdss
t1/2
AUC (po)d
(lM)
(L/kg)
(po) (h)
(lM h)
3.13
30
0.74
30
6.1
0.5
5.3
4.0
a From an oral dose of 5 lmol/kg, iv dose of 0.5 lmol/kg in Han Wistar rats.
b Aqueous solubility measured in pH 7.4 buffer on pure but generally amorphous material.
c Measured at 37 °C.
d Total AUC at the oral dose administered.

2780
C. Bardelle et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2776–2780
9. Steinle, J. J.; Meininger, C. J.; Forough, R.; Wu, G.; Wu,
M. H.; Granger, H. J. J. Biol. Chem. 2002, 277, 43830.
10. Martiny-Baron, G.; Korff, T.; Schaffner, F.; Esser, N.;
generate a novel series of 2,4-bis-anilinopyrimidine
inhibitors. SAR studies around the C-2 hinge-binding
aniline indicate a strong preference for meta-substitu-
tion, particularly with electron-withdrawing groups.
One example from this work, 28, has shown an encour-
aging kinase selectivity and pharmacokinetic profile and
serves as the basis for further hit to lead optimisation.
Further optimisation of, in particular the C-2 aniline
of this class of inhibitors is described in the subsequent
communication.
´
Eggstein, S.; Marme, D.; Augustin, H. G. Neoplasia 2004,
6, 248.
11. Kertesz, N.; Krasnoperov, V.; Reddy, R.; Leshanski, L.;
Kumar, S. R.; Zozulya, S.; Gill, P. S. Blood 2006, 107,
2330.
12. The assay detects inhibitors of recombinant EphB4-
mediated phosphorylation of a polypeptide substrate in
presence of magnesium-ATP using Alphascreen (Packard
Bioscience) luminescence detection technology. For full
details see Ref. 17.
´
13. Ple, P. A.; Green, T. P.; Hennequin, L. F.; Curwen, J.;
Acknowledgment
Fennell, M.; Allen, J.; Lambert-van der Brempt, C.;
Costello, G. J. Med. Chem. 2004, 47, 871.
We thank Jonathan Wingfield for technical assistance.
14. Armistead, D. M.; Bemis, J. E.; Buchanan, J. L.; Dipietro,
L. V.; Elbaum, D.; Habgood, G. J.; Kim, J. L.; Marshall,
T. L.; Geuns-Meyer, S. D.; Novak, P. M.; Nunes, J. J.;
Patel, V. F.; Toledo-Sherman, L. M.; Zhu, X. PCT Int.
Appl. WO2001025220.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2008.04.015.
15. By analogy with Wybenga-Groot, L. E.; Baskin, B.;
Ong, S. H.; Tong, J.; Pawson, T.; Sicheri, F. Cell 2001,
106, 745.
16. There is structural precedent for aniline oriented
towards the ribose site in a related series of 4,6-bis-
anilinopyrimidines binding to CDK2: Beattie, J. F.;
Breault, G. A.; Ellston, R. P. A.; Green, S.; Jewsbury,
P. J.; Midgley, C. J.; Naven, R. T.; Minshull, C. A.;
Pauptit, R. A.; Tucker, J. A.; Pease, J. E. Bioorg. Med.
Chem. Lett. 2003, 13, 2955.
References and notes
1. Kullander, K.; Klein, R. Nat. Rev. Mol. Cell Biol. 2002, 3,
475.
2. Surawska, H.; Ma, P. C.; Salgia, R. Cytokine Growth
Factor Rev. 2004, 15, 419.
3. Nakamoto, M.; Bergemann, A. D. Microsc. Res. Tech.
2002, 59, 58.
4. Xia, G.; Kumar, S. R.; Masood, R.; Zhu, S.; Reddy, R.;
Krasnoperov, V.; Quinn, D. I.; Henshall, S. M.; Suther-
land, R. L.; Pinski, J. K.; Daneshmand, S.; Buscarini, M.;
Stein, J. P.; Zhong, C.; Broek, D.; Roy-Burman, P.; Gill,
P. S. Cancer Res. 2005, 65, 4623.
5. Brantley-Sieders, D.; Schmidt, S.; Parker, M.; Chen, J.
Curr. Pharm. Des. 2004, 10, 3431.
6. Cheng, N.; Brantley, D. M.; Chen, J. Cytokine Growth
Factor Rev. 2002, 13, 75.
7. Gerety, S. S.; Wang, H. U.; Chen, Z.-F.; Anderson, D. J.
Mol. Cell 1999, 4, 403.
8. Wang, H. U.; Chen, Z.-F.; Anderson, D. J. Cell 1998, 93,
741.
17. Kettle, J. G.; Read, J.; Leach, A.; Barlaam, B. C.; Ducray,
R.; Lambert-Van Der Brempt, C. M. P.; PCT Int. Appl.
WO2007085833.
18. The methodology, geometry and energies are provided as
supporting information.
19. Abraham, M. H.; Duce, P. P.; Prior, D. V.; Barratt, D. G.;
Morris, J. J.; Taylor, P. J. J. Chem. Soc., Perkin Trans. 2
1989, 1355.
20. Kenny, P. W. J. Chem. Soc., Perkin Trans. 2 1994, 2, 199.
21. CHO–K1 cells were engineered to stably express an
EphB4-Myc-His construct. The endpoint assay used a
sandwich ELISA to detect EphB4 phosphorylation status.
Myc-tagged EphB4 from treated cell lysate were captured
via an anti-c-Myc antibody and the phosphorylation
status of captured EphB4 was then measured using a
generic phosphotyrosine antibody. For further details see
Ref. 17.