Scaffold oriented synthesis. Part 2: Design, synthesis and biological evaluation of pyrimido-diazepines as receptor tyrosine kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2008.03.021

We report the discovery of the pyrimido-diazepine scaffolds as novel adenine mimics. Structure-based design led to the discovery of analogs with potent inhibitory activity against receptor tyrosine kinases, such as KDR, Flt3 and c-Kit. Compound 14 exhibited low nanomolar KDR enzymatic and cellular potencies (IC₅₀ = 9 and 52 nM, respectively).

Full text of DOI:10.1016/j.bmcl.2008.03.021

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 2691–2695
Scaffold oriented synthesis. Part 2: Design, synthesis
and biological evaluation of pyrimido-diazepines
as receptor tyrosine kinase inhibitors
Vijaya Gracias, Zhiqin Ji, Irini Akritopoulou-Zanze,* Cele Abad-Zapatero,
Jeffrey R. Huth, Danying Song, Philip J. Hajduk, Eric F. Johnson, Keith B. Glaser,
Patrick A. Marcotte, Lori Pease, Nirupama B. Soni, Kent D. Stewart,
Steven K. Davidsen, Michael R. Michaelides and Stevan W. Djuric
Global Pharmaceutical Research and Development, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL 60044, USA
Received 14 January 2008; revised 5 March 2008; accepted 6 March 2008
Available online 10 March 2008
Abstract—We report the discovery of the pyrimido-diazepine scaffolds as novel adenine mimics. Structure-based design led to the
discovery of analogs with potent inhibitory activity against receptor tyrosine kinases, such as KDR, Flt3 and c-Kit. Compound 14
exhibited low nanomolar KDR enzymatic and cellular potencies (IC50 = 9 and 52 nM, respectively).
Ó 2008 Elsevier Ltd. All rights reserved.
Protein kinases have proven to be attractive targets for a
variety of therapeutic indications1 and several kinase
inhibitors have progressed to become marketed drugs,
particularly for cancer.2 All kinases possess a structur-
ally conserved catalytic domain that binds ATP3 and
most kinase research programs target this ATP binding
site with structurally related adenine mimics. Conse-
quently, many patents have been filed around similar
or in many cases identical chemotypes resulting in a con-
gested intellectual property situation.
tors; an effort intended to help address the above patent
challenge.4 We have placed an emphasis on the novelty
of the designed chemotypes and attempted to create
structures that have never been described in patents or
in published literature as kinase inhibitors. Specifically,
we were interested in identifying novel adenine replace-
ments that upon further functionalization would furnish
proprietary kinase inhibitors.
NH2
NH2
N
We have recently reported on our efforts to enhance the
Abbott compound collection with novel kinase inhibi-
X
N
NH2
O
O
5 X = H
6 X = NH2
a
H2N
H2N
H2N
R
R
N
N
N
N
b
N
N
N
3
4
R1
R1
R1
N
H
N
H2N
N
N
H
N
H
N
1
2
Proposed New Scaffolds
Adenine
NH2
N
Figure 1. De-novo design of pyrimido-diazepines.
N
1a-f and 1o,p X = H
2a-d X = NH2
X
N
N
H
Keywords: KDR inhibitors; VEGFR; PDGFR; cKit inhibitors; Flt3
inhibitors; Adenine mimics; Structure-based drug design.
*
Scheme 1. Reagents and conditions: (a) Eschenmosher’s salt, MeCN,
80 °C, >20%; (b) EtOH, reflux, 1–14%.
Corresponding author. Tel.: +1 847 935 5006; fax: +1 847 935
0310; e-mail: irini.zanze@abbott.com
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.03.021

2692
V. Gracias et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2691–2695
In this context we have explored the idea of replacing
the imidazole ring of adenine with a seven-membered
ring diazepine as shown in Figure 1. We hypothesized
that the new pyrimido-diazepine scaffold 1 would bind
to the hinge region of the target kinase in a manner sim-
ilar to that exhibited by adenine with the formation of
two key hydrogen bond donor/acceptor interactions.
We anticipated that additional functionalization of the
molecule based on molecular modeling considerations
would allow us to explore various sites with the expecta-
tion of gains in potency and selectivity depending on the
kinase targeted. To exploit the potential of a third
hydrogen bond in the hinge region, we designed scaffold
2 which possess an additional amino group in the 2-po-
sition of the pyrimidine ring.
O
a
b
Cl
NH
NH
7
8
O
O
O
O
O
c
The molecules were initially synthesized following previ-
ously described procedures.5 Reactions of ketones 3
(Scheme 1) with Eschenmosher’s salt yielded intermedi-
ates 4 which were isolated and subsequently subjected to
heating with commercially available pyrimidine-4,5,6-
triamines 5 or 6 to provide the pyrimidoazepines 1 or
2, respectively, in low overall yields.
N
H2N
2 HCl
10
NH
O
NH2
O
9
NH2
NH2
N
NO2
Cl
NH2
N
Subsequently, for the synthesis of more complex ana-
logs, a new synthetic protocol was devised as shown in
Scheme 2. Friedel–Crafts acylation of protected aniline
7 with 3-chloropropanoyl chloride provided precursor
8, which was subjected to nucleophilic substitution with
potassium phthalimide to yield compound 9 in good
yield. Acid deprotection of both amino groups of 9 pro-
vided 10 as the diacid salt in good yield. Nucleophilic
aromatic substitution of 6-chloro-5-nitropyrimidin-4-
amine with 10 in the presence of triethylamine furnished
compound 11. At this point various isocyanates were re-
acted with aniline 11 to yield exclusively ureas 12. Final
compounds 1 were obtained upon reduction of the nitro
group and spontaneous imine formation of the resulting
amine with the carbonyl group of 12.
NO2
R2 NCO
e
N
d
N
N
H
O
11
R2
HN
H
H
N
N
R2
HN
O
O
NH2
N
NH2
N
N
f
NO2
N
N
N
H
O
N
H
12
1g-n
Scheme 2. Reagents and conditions: (a) 3-Chloropropanoyl chloride,
AlCl3, CS2, 64%; (b) potassium phthalimide, DMF, 125 °C, 59%; (c)
HCl/HOAc, reflux, 72%; (d) Et3N, 0–75 °C, 83%; (e) DMF, 56%; (f)
SnCl2, EtOH, 40–70%.
Initially, simple analogs of 1 and 2 were tested against a
preliminary panel of five kinases of interest4,6 at low
concentration of ATP (10 lM) in order to detect even
Table 1. Kinase inhibitory activity of pyrimido-diazepine analogs 1 and 2 against an exploratory kinase panel
R1
NH2
N
N
X
N
N
H
a
a
a
a
a
Compound
X
R1
KDR IC50 (lM)
Plk1 IC50 (lM)
Pak4 IC50 (lM)
CK2 IC50 (lM)
Akt1 IC50 (lM)
1a
1b
1c
1d
1e
2a
2b
2c
H
H
3b
3
>100
50
>100
>100
>100
>100
>100
>100
>100
>100
4b
12
92
H
3-Me
4-Me
3-Cl
4-Cl
H
>100
>100
>100
>100
>100
>100
>100
H
4
>100
>100
>100
>100
>100
>100
>100
64%c
>100
>100
21
H
0.6
3
H
NH2
NH2
NH2
35
21
9
3-Cl
4-Cl
>100
a IC50 values are based on an eleven point curve at 10 lM ATP concentration, >100 indicates less than 50% inhibition at that concentration or an
IC50 > 100 lM.
b Average of two values.
c Indicates % inhibition at 100 lM.

V. Gracias et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2691–2695
2693
weak inhibitory activity trends. Most of the analogs
exhibited modest activity against KDR in the single digit
micromolar and high submicromolar range and some
were active against CK2 (Table 1). The presence of the
extra amino group in analogs 2 had a negative effect
on the KDR inhibitory activity of the compounds (1a
vs 2a, 1d vs 2b, and 1e vs 2c).
Encouraged by these initial results we decided to further
investigate the KDR inhibitory potency of the series by
employing molecular modeling. Superimposition of
plain compound 1a with previously reported KDR
inhibitor 137 (Fig. 2) provided us with insights as to
how to best take advantage of the hydrophobic ‘back
pocket’ of KDR. Introduction of urea functionality such
as in 13 has been shown to have beneficial effects on
KDR inhibitory activity (KDR IC50 = 52 nM) and such
compounds have been part of a program at Abbott aim-
ing to develop PDGFR and VEGFR multitargeted ki-
nase inhibitors.7–10
Based on modeling considerations, both para- and meta-
positions of the phenyl ring provided reasonable vectors
to extend the urea functionality. We selected to prepare
a small group of key urea analogs to validate our mod-
eling hypothesis. The urea analogs were synthesized
according to Schemes 1 and 2 from the appropriate
starting materials and tested against KDR, Flt3, and
c-Kit in an HTRF assay format at 1 mM ATP.
We immediately observed an improvement in KDR po-
tency with the simple para-phenyl urea analog 1f (Table
2) in comparison to 1a. However, its 2-amino analog
equivalent 2d was not active under the assay conditions
and these analogs were not further pursued. Introduc-
tion of a trifluoromethyl group in the meta-position of
the urea phenyl ring (compound 1g) dramatically in-
creased the in vitro KDR potency. 1g was also very po-
tent against Flt3 and c-Kit and in the KDR cellular
assay. The presence of the same group in the para-posi-
Figure 2. Model of compound 1a, green carbons, bound to KDR
kinase in ‘inactive’ conformation with Phe 1047 in the DFG-out
position [model created as in Refs. 7 and 10]. Hinge hydrogen bonds to
the backbone Glu 917 C@O and Cys 917 N–H are shown with black
dotted lines. A model of thienopyrimidine KDR inhibitor previously
published7 is shown with pink carbons and with its urea H-bond to
Glu 885.
Table 2. Kinase inhibitory activity of para-urea analogs 1 and 2
H
N
H
N
R2
O
NH2
N
N
X
N
N
H
R2
KDR IC50 (lM)
KDR (cell)b IC50 (lM)
Flt3 IC50 (lM)
c-Kit IC50 (lM)
a
c
a
a
Compound
X
1f
H
Ph
Ph
0.951
>13
NDd
ND
ND
ND
>13
2d
1g
1h
1i
NH2
H
2.858
0.002
0.009
0.006
0.003
0.007
0.016
0.013
0.002
3-CF3-Ph
4-CF3-Ph
3-Cl-Ph
0.003
0.110
0.019
0.049
0.005
0.006
0.065
0.004
0.013
2.440
0.050
1.520
0.345
0.049
0.311
0.121
0.006
0.035
0.018
0.028
0.022
0.050
0.037
0.011
H
H
1j
H
2-F-3-CF3-Ph
4-F-3-CF3-Ph
4-Cl-3-CF3-Ph
2-F-5-Me-Ph
2-F-5-CF3-Ph
1k
1l
H
H
1m
1n
H
H
a IC50 values are based on an eleven point curve at 1 mM ATP concentration.
b Activity against VEGF-induced KDR phosphorylation in 3T3-murine fibroblast cells.
c Average of two experiments.
d ND stands for not determined.

2694
V. Gracias et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2691–2695
Table 3. Kinase inhibitory activity of meta-urea analogs
R3
O
R4
N
H
N
NH2
H
N
N
N
N
H
a
a
a
Compound
R3
R4
KDR IC50 (lM)
KDR (cell)b IC50 (lM)
Flt3 IC50 (lM)
c-Kit IC50 (lM)
1o
1p
14
H
Ph
3-CF3-Ph
3-CF3-Ph
0.888
0.126
0.006
NDc
2.696
0.145
0.135
3.022
0.091
0.012
H
Me
0.790
0.052
a IC50 values are based on an eleven point curve at 1 mM ATP concentration.
b Activity against VEGF-induced KDR phosphorylation in 3T3-murine fibroblast cells.
c ND stands for not determined.
tion (1h) resulted in losses in KDR potency. Di-substi-
tuted analogs containing the meta-trifluoromethyl group
and a fluoro group in various positions resulted in com-
pounds with similar in vitro KDR potency, (1k and 1l)
however, their cellular potency was inferior to that of
compound 1g.
and 14 is potentially due to the different gatekeeper res-
idues for the two enzymes, Val 916 for KDR and Phe
691 for Flt3, which is in the direct vicinity of the aryl
Me group of 14.
In conclusion, we have identified pyrimido-diazepines as
novel kinase hinge binders and were able to rapidly
transform them to potent VEGFR and PDGFR inhibi-
tors using modeling and our expertise in kinase inhibi-
tion. We are in the process of finding other
applications for this unique core and our findings will
be reported in due course.
The meta-substituted phenyl urea 1o (Table 3) exhibited
a similar drop in potency when compared to 1a. How-
ever, the meta-trifluoromethyl phenyl analog of the
meta-urea 1p was not as potent as 1g. Modeling of 1p
in the active site of KDR (Fig. 3) suggested the presence
of a small pocket that could be filled with additional
functionality. Thus compound 14 was prepared follow-
ing procedures similar to those in Scheme 1. To our de-
light compound 14 exhibited improved in vitro and
cellular KDR potency and an apparent selectivity for
KDR versus Flt3 (Table 3). This is in stark contrast to
the para-urea series (Table 2) where most of the com-
pounds were equipotent for Flt3 and to meta-urea ana-
logs such as 1p lacking the methyl group. The altered
KDR/Flt3 enzyme potency ratio for compounds 1p
Acknowledgments
The authors thank the Abbott High Throughput Purifi-
cation and Structural Chemistry groups for their sup-
port with purification and analytical data, respectively.
References and notes
1. Akritopoulou-Zanze, I. Idrugs 2006, 9, 481.
´
´
2. (a) de Carcer, G.; Perez de Castro, I.; Malumbres, M.
Curr. Med. Chem. 2007, 14, 969; (b) Schenone, S.;
Bondavalli, F.; Botta, M. Curr. Med. Chem. 2007, 14,
2495; (c) Steeghs, N.; Nortier, J. W. R.; Gelderblom, H.
Ann. Surg. Onc. 2007, 14, 942.
3. Liao, J. J.-L. J. Med. Chem. 2007, 50, 409.
4. Akritopoulou-Zanze, I.; Darczak, D.; Sarris, K.; Phe-
lan, K. M.; Huth, J. R.; Song, D.; Johnson, E. F.; Jia,
J.; Djuric, S. W. Bioorg. Med. Chem. Lett. 2005, 16,
96.
5. (a) Saitz, C.; Jullian, C. J. Heterocyclic Chem. 2000, 37,
401; (b) Insuasty, B.; Perez, A.; Valencia, J.; Quiroga, J. J.
Heterocyclic Chem. 1997, 34, 1555.
6. Luo, Y.; Smith, R. A.; Guan, R.; Liu, X.; Klinghofer, V.; Shen,
J.; Hutchins, C.; Richardson, P.; Holzman, T.; Rosenberg, S.
H.; Giranda, V. L. Biochemistry 2004, 43, 1254.
7. Dai, Y.; Guo, Y.; Frey, R. R.; Ji, Z.; Curtin, M. L.;
Ahmed, A. A.; Albert, D. H.; Arnold, L.; Arries, S. S.;
Barlozzari, T.; Bauch, J. L.; Bouska, J. J.; Bousquet, P. F.;
Cunha, G. A.; Glaser, K. B.; Guo, J.; Li, J.; Marcotte, P.
A.; Marsh, K. C.; Moskey, M. D.; Pease, L. J.; Stewart, K.
Figure 3. Model of compound 1p, green carbons, bound to KDR with
the urea of 1p hydrogen-bonded to Glu 885. The arrow shows a vector
that projects into a small hydrophobic volume comprising the
sidechains of Val 848, Val 916, and the methylenes of Lys 868.

V. Gracias et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2691–2695
2695
D.; Stoll, V. S.; Tapang, P.; Wishart, N.; Davidsen, S. K.;
Michaelides, M. R. J. Med. Chem. 2005, 48, 6066.
K. D.; Yates, M.; Davidsen, S. K.; Michaelides, M. R.
Bioorg. Med. Chem. Lett. 2006, 16, 4326.
8. Dinges, J.; Ashworth, K. L.; Akritopoulou-Zanze, I.; Arnold,
L. D.; Baumeister, S. A.; Bousquet, P. F.; Cunha, G. A.;
Davidsen, S. K.; Djuric, S. W.; Gracias, V. J.; Michaelides, M.
R.; Rafferty, P.; Sowin, T. J.; Stewart, K. D.; Xia, Z.; Zhang,
H. Q. Bioorg. Med. Chem. Lett. 2006, 16, 4266.
9. Ji, Z.; Ahmed, A. A.; Albert, D. H.; Bouska, J. J.;
Bousquet, P. F.; Cunha, G. A.; Glaser, K. B.; Guo, J.; Li,
J.; Marcotte, P. A.; Moskey, M. D.; Pease, L. J.; Stewart,
10. Dai, Y.; Hartandi, K.; Ji, Z.; Ahmed, A. A.; Albert, D. H.;
Bauch, J. L.; Bouska, J. J.; Bousquet, P. F.; Cunha, G. A.;
Glaser, K. B.; Harris, C. M.; Hickman, D.; Guo, J.; Li, J.;
Marcotte, P. A.; Marsh, K. C.; Moskey, M. D.; Martin,
R. L.; Olson, A. M.; Osterling, D. J.; Pease, L. J.; Soni, N.
B.; Stewart, K. D.; Stoll, V. S.; Tapang, P.; Reuter, D. R.;
Davidsen, S. K.; Michaelides, M. R. J. Med. Chem. 2007,
50, 1584.