4-Aryl-5-cyano-2-aminopyrimidines as VEGF-R2 inhibitors: Synthesis and biological evaluation

Article abstract of DOI:10.1016/j.bmcl.2007.04.021

A novel series of 4-aryl-5-cyano-2-aminopyrimidines were synthesized and found to have potent VEGF-R2 kinase inhibitory activity. Structure-activity relationships were investigated and compound 14a was shown to be efficacious in a mouse model of corneal neovascularization.

Full text of DOI:10.1016/j.bmcl.2007.04.021

Bioorganic & Medicinal Chemistry Letters 17 (2007) 3266–3270
4-Aryl-5-cyano-2-aminopyrimidines as VEGF-R2
inhibitors: Synthesis and biological evaluation
Terry V. Hughes,a,* Stuart L. Emanuel,a Amanda K. Beck,a Steven K. Wetter,a
Peter J. Connolly,a Prabha Karnachi,a Michael Reuman,a Jabed Seraj,a
Angel R. Fuentes-Pesquera,a Robert H. Gruninger,a Steven A. Middleton,a
Ronghui Lin,a Jeremy M. Davisb and David F. C. Moffatb
aJohnson & Johnson Pharmaceutical Research and Development, L.L.C, PO Box 300, 1000 Route 202, Raritan, NJ 08869, USA
bUCB Pharma, 216 Bath Road, Slough SL14EN, UK
Received 3 February 2007; revised 1 April 2007; accepted 4 April 2007
Available online 10 April 2007
Abstract—A novel series of 4-aryl-5-cyano-2-aminopyrimidines were synthesized and found to have potent VEGF-R2 kinase inhib-
itory activity. Structure–activity relationships were investigated and compound 14a was shown to be efficacious in a mouse model of
corneal neovascularization.
Ó 2007 Elsevier Ltd. All rights reserved.
Angiogenesis is the formation of new blood vessels, and
is an essential process for tumor growth. The selective
inhibition of tumor angiogenesis might be a useful strat-
egy for inhibiting tumor growth and metastasis. Patho-
logical angiogenesis in the retina is a major cause of
blindness and is responsible for the deterioration of vi-
sion that occurs in diabetic retinopathy, age-related
macular degeneration, and retinopathy of prematurity.
Several growth factors are elevated in the eye of patients
with active neovascularization that contribute to the for-
mation of new vessels.1,2 Vascular endothelial growth
factor (VEGF) and basic fibroblast growth factor
(bFGF) are critical components in the development
of pathological angiogenesis and are involved in the
etiology of these diseases. VEGF stimulates ocular
neovascularization, increases vascular permeability, and
correlates with disease progression.3 This makes the
VEGF signal transduction pathway an attractive target
for therapeutic intervention in ischemic ocular disease.
We have discovered a series of potent VEGF-R2 kinase
inhibitors that are potentially useful as anti-angiogenesis
agents for the treatment of cancer and diabetic retinop-
athy. We recently reported the in vitro anti-angiogenic,
in vivo anti-tumor activity of 14a (JNJ-17029259) as well
as its activity against VEGF-R2 and other biologically
relevant kinases.4 Herein, we report the SAR develop-
ment, synthesis, and biological activity of 4-aryl-5-cya-
no-2-aminopyrimidines that display potent inhibition
of VEGF-R2 kinase.5 We also describe the proposed
binding mode of the pharmacophore and the rationale
for structure–activity relationships.
Routine screening for selective VEGF-R2 inhibitors
provided compound 1 (Scheme 1) as a hit with VEGF-
R2 IC50 = 1.0 lM and cyclin-dependent kinase-1
(CDK1) IC50 > 100 lM.
Compound 1 was docked into a homology model of
VEGF-R2 using the software GLIDE.6,7 The proposed
binding mode (Fig. 2) shows the top reranked pose of
the ligand as overlapped with ATP along with the bind-
ing pockets of the ATP site.8 The amino pyrimidine
forms hydrogen bonds with backbone C@O and NH
of CYS 917 in the hinge region in the ATP site. The
cyano group is positioned at the opening of the hydro-
phobic pocket.
Synthesis of 4-substituted-5-cyano-2-aminopyrimidines
(2) is outlined in Scheme 2. The key step involved reac-
tion of a vinylogous amide with a guanidinium salt to
form the pyrimidine ring. Specifically, conversion of an
Keywords: VEGF-R2; Kinase; KDR; Cancer; Angiogenesis; Kinase
inhibitor; 2-Aminopyrimidines.
*
Corresponding author. E-mail: hughe007@hotmail.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.04.021

T. V. Hughes et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3266–3270
3267
H
electron rich (e.g., indolyl, pyrrolyl) were synthesized
via Scheme 3 as they were not stable to the Sandmeyer
conditions outlined in Scheme 2.
N
N
H
N
N
N
R2
NC
N
Examination of the R2 substituent data revealed that
electron-donating substituents such as alkyl or methoxy
(2g, i, j) are tolerated at the para and meta positions, but
VEGF-R2 activity is considerably reduced with an elec-
tron-withdrawing chloride (2o). Substitution at the ortho
position also results in loss of kinase activity, possibly
due to unfavorable steric interactions (2n,p). Analogues
having heteroaromatic rings such as pyrimidine (2m) or
benzimidazole (2l) directly attached to the 2-amino
group were poor VEGF-R2 inhibitors. However, the in-
dol-5-yl derivative 2h, which is attached to the 2-amino
group through the benzene ring, was about twice as po-
tent as the parent compound 1. Interestingly, 2h was
4-fold more potent than the all-carbon naphthalen-2-yl
analogue 2k, consistent with the trend toward enhanced
kinase potency with electron-rich ring systems. Expand-
ing on the favorable para-alkyl substitution pattern, the
attachment of polar groups to the alkyl chain led to ana-
logues with improved kinase potency. Thus, hydroxyl
substitution on the para-alkyl group (2f) resulted in an
increase in VEGF-R2 potency compared to 2j, and
amine substitution increased potency even more (2a–
c). Attachment of the amino group via a 1-carbon linker
(2a) displayed similar activity when compared to the
same group attached by a 2-carbon linker (2c). The meta
aminoalkyl analogue 2e and the morpholine amide 2d
were somewhat weaker VEGF-R2 inhibitors than the
preferred para aminoalkyl compounds.
NC
R1
2
1
Scheme 1. Screening hit compound
pharmacophore 2.
1
and general structure of
N
NH2
MeO
O
O
b, c
a
NC
N
R1
R1
NC
R1
3
4
5
d
H
N
N
N
Cl
R2
e
N
N
NC
NC
R1
2
R1
6
Scheme 2. Reagents and conditions: (a) n-BuLi, CH3CN, THF, À78 to
À45 °C (80–100%); (b) DMF-DEA, DMF, 25 °C; (c) guanidine
nitrate, DMF, NaHCO3 (92%, over 2 steps); (d) SbCl3, tert-butyl
nitrite, 1,2-dichloroethane, 25 °C (47%); (e) R2NH2, THF, 25 °C to
reflux, NaHCO3 (73–95%).
aryl methyl ester 3 (R1 = aryl) to the corresponding a-
cyanoketone 4 was achieved via formation of the lithium
salt of acetonitrile by treatment with n-BuLi at À78 °C
followed by reaction with the ester 3 at À45 °C. Subse-
quent treatment of the a-cyanoketone 4 with N,N-
dimethylformamide diethyl acetal (DMF-DEA) formed
a vinylogous amide in situ that was reacted with guani-
dine nitrate in DMF at 100 °C to form the 2-amino-4-
aryl-5-cyanopyrimidine 5. The Sandmeyer reaction of
the aminopyrimidine 5 was accomplished by treatment
with antimony trichloride and tert-butylnitrite in
1,2-dichloroethane at 25 °C to smoothly afford the 2-chlo-
ropyrimidine 6.9 The displacement of the Cl of 6 with
aliphatic amines proceeded at 25 °C and with aromatic
amines in refluxing THF to afford the pharmacophore
2. The synthesis of analogues via Scheme 2 was
divergent and allowed for the late stage modification of
R2. Initial exploration of the SAR of the R2 substituent
was determined by reacting 2-chloro-5-cyano-4-phenyl-
pyrimidine 6 (R1 = phenyl) with more than 65 different
amines. The VEGF-R2 and CDK1 enzyme inhibitory
activities of representative analogues are summarized in
Table 1.
With SAR for R2 groups in hand, we fixed R2 and
varied R1. Several analogues with various R1 groups
were made via the synthesis of specific arylguanidines
8 as detailed in Scheme 3. Table 2 details a representa-
tive selection of analogues 9 made with R2 fixed as
4-(2-hydroxyethyl)phenyl. A variety of aromatic and
heteroaromatic rings was tolerated at R1 but the tert-
butyl (9j) and indol-2-yl (9i) groups led to significant
loss of kinase potency. Some of the better substituents
for VEGF-R2 kinase inhibition such as 4-(2-amino-2-
propyl)phenyl (9a) and indol-5-yl (9c) contained an
NH group capable of acting as an H-bond donor.
Molecular modeling indicated that the NH2 of 9a
and the NH of 9c participate in a favorable H-bond
with the backbone C@O of Arg 177 in the phosphate
binding site.
Next, the hydroxyl groups of 9 were transformed into
pendant amino groups via mesylation and displacement
with an amine or via oxidation and reductive amination
to afford the 1 and 2-carbon linked analogues 10 as de-
tailed in Scheme 4. Numerous analogues 14, where R1 is
fixed as 4-(2-amino-2-propyl)phenyl and R2 is varied,
were synthesized to optimize inhibition of VEGF-R2 ki-
nase (Scheme 5).10 Some examples of the VEGF-R2
activity of analogues 14 are shown in Table 3. All ana-
logues were selective for VEGF-R2 and had greatly re-
duced potency against CDK1. Notably, compound 14a
was a very potent VEGF-R2 kinase inhibitor with an
IC50 value of 27 nM. Para-substitution (14a) was shown
A second synthetic route, outlined in Scheme 3, is con-
vergent and relies on the synthesis of different guanidine
salts 8 to vary the R2 substituent. The guanidine salts 8
were synthesized by reacting amines 7 with cyanamide
and 12 N HCl in refluxing EtOH.5 The a-cyanoketone
4 was treated with DMF-DEA in DMF at 25 °C and
the resulting vinylogous amide was reacted with the
guanidine salt 8 and NaHCO3 at 100 °C in a one-pot
reaction to yield analogues 2. Analogues where R1 was

3268
T. V. Hughes et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3266–3270
Table 1. SAR for R2 substituent of compounds 1 and 2a–p (R1 = Ph)
R2
VEGF-R2 IC50 (lM)
CDK1 IC50 (lM)
a
a
Compound
1
Phenyl
1.0
>10
6.2
2a
2b
2c
2d
2e
2f
4-(Pyrrolidin-1-yl-CH2)phenyl
3-Methoxy-4-(pyrrolidin-1-yl-CH2)phenyl
4-(Pyrrolidin-1-yl-CH2CH2)phenyl
4-(3-Morpholin-4-yl-3-oxopropyl)phenyl
3-(Morpholin-4-yl-CH2)phenyl
4-(2-Hydroxyethyl)phenyl
3-Methoxyphenyl
0.061
0.091
0.11
0.18
0.36
0.20
0.38
0.41
0.49
1
9.0
6.5
2.1
2.0
2.0
4.2
2g
2h
2i
2-Methylindol-5-yl
4-Methoxyphenyl
3.3
3.3
2j
4-Methylphenyl
Naphthalen-2-yl
3.5
2k
2l
2.1
10
>10
>10
>10
>10
>10
>10
Benzimidazol-2-yl
Pyrimidin-2-yl
2m
2n
2o
2p
>10
>10
>10
>10
2-Chlorophenyl
4-Chlorophenyl
2-Methoxyphenyl
a IC50 data are the average of at least two separate experiments. IC50 values listed as >10 indicate no observed 50% inhibition at 10 lM, nor was an
inhibition maximum observed. Assay performed at 5 lM ATP concentration.
H
H
HN
N
N
N
N
a
NH2
R2 NH2
a or b
R2 NH
OH
n
NR2R3
n
N
N
NC
NC
7
8
R1
R1
H
N
N
10 (n = 1,2)
9 (n = 1,2)
R2
O
b, c
NC
N
Scheme 4. Reagents and conditions: (a) CH3SO2Cl, Et3N, 0–25 °C;
NHR1R2, DMF, 70 °C (73%); (b) Dess-Martin periodinane, THF,
25 °C; NHR1R2, NaHB(OAc)3, THF, 25 °C.
NC
R1
R1
4
2
Scheme 3. Reagents and conditions: (a) cyanamide, 12 N HCl, EtOH,
reflux (78%); (b) DMF-DEA, DMF, 25 °C; (c) 8, DMF, NaHCO3,
100 °C (60%).
H
N
N
CN
R
N
N
O
O
O
NC
Table 2. SAR for R1 substituent for analogues
HO(CH2)nPh)
9
(R2 = 4-
c, d, e
f, g, h, i
a, b
Compound
n
R1
VEGF-R2 CDK1
a
a
CN
12
CN
11
IC50 (lM) IC50 (lM)
NHBoc
NH2
13
14
9a
2
4-(2-Amino-2-propyl)
phenyl
2-Thienyl
0.052
1.6
Scheme 5. Reagents and conditions: (a) DMF-DEA, DMF, 25 °C; (b)
NH2OHÆHCl, MeOH, 25 °C (75%); (c) powdered CeCl3, MeLi, 25 °C,
sonication; (d) BOC2O, toluene, reflux (82%); (e) NaOH, EtOH, 25 °C
(92%); (f) DMF-DEA, DMF, 25 °C; (g) N-[4-(2-hydroxyethyl)phenyl]-
guanidine nitrate, DMF, NaHCO3, 100 °C (83%); (h) CH3SO2Cl,
Et3N, 0–25 °C; (i) NH2R, DMF, 70 °C (65%); (j) TFA, DCM, then
K2CO3 (95%).
9b
9c
9d
9e
9f
2
2
2
2
2
2
2
0.082
0.090
0.15
0.18
0.20
0.34
0.47
0.41
1.5
5-Indolyl
4-Methoxyphenyl
4-Bromophenyl
4-Sulfamoylphenyl
Benzothiazol-2-yl
2.0
1.9
9g
9h
0.56
2.7
4-(2-Hydroxy-2-propyl) 0.41
phenyl
9i
9j
2
2
Indol-2-yl
tert-butyl
1.4
1.6
1.7
1.1
inhibited neovascularization in a corneal pocket mod-
el.14 To investigate if the antiangiogenic effects of these
analogues would translate to efficacy in animal models
of retinopathy, compound 14a was evaluated in mice
in a corneal micropocket model for inhibition of growth
factor-induced neovascularization in the eye as de-
scribed (Fig. 1). A marked reduction in vessel growth
was observed at all doses following oral administration
of compound 14a. Although the 100 mg/kg dose reduced
vessel growth, the effects at this dose only approached
the cutoff for significance (p = 0.059). Compound 14a
a IC50 data are the average of at least two separate experiments. Assay
performed at 5 lM ATP concentration.
to confer better potency than meta-substitution (14b).
Compound 14e, which has an oxygen inserted in the
linker, was shown to be 20-fold less potent against
VEGF-R2 than 14a.11
Corneal micropocket model.12,13 The VEGFR-2 inhibitor
(SU-5416) has been reported in the literature to have

T. V. Hughes et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3266–3270
Table 3. SAR for R substituent for analogues 14
3269
Hydrophobic
pocket
Compound
R
VEGF-R2 CDK1
a
a
IC50 (lM) IC50 (lM)
O
N
H
H
14a
14b
14c
14d
14e
4-(Morpholin-4-yl-
CH2CH2)phenyl
0.027
0.055
0.094
4.1
5.7
4.8
5.6
N
N
H
CYS
917
N
N
3-(Morpholin-4-yl-
CH2CH2)phenyl
N
O
Phosphate
binding site
N
N
N
O
H
N
4-(2-Ethylimidazol-1-yl-
CH2CH2)phenyl
N
O
P
O
4-(4-Methylpiperazin-1-yl- 0.021
CH2CH2)phenyl
sugar pocket
O
4-(Morpholin-4-yl-
CH2CH2O)phenyl
0.45
21
Figure 2. Proposed binding mode of compound 1 in the active site of
VEGF-R2 as overlapped with ATP.
a IC50 data are the average of at least two separate experiments. Assay
performed at 5 lM ATP concentration.
above resulted in statistically significant reductions in
neovascularization relative to the animals dosed with
vehicle alone. Compound 14a was efficacious at inhibit-
ing corneal neovascularization at several doses and daily
oral therapy was well tolerated with no decrease in body
weight and no treatment-related deaths.
In summary, we have described the SAR development
and synthesis of a novel series of 4-aryl-5-cyano-2-
aminopyrimidines that are potent and selective inhibi-
tors of VEGF-R2. The SAR development for R2 of
the pharmacophore 1 relied upon a divergent synthesis
(Scheme 2), while the SAR optimization for R1 focused
on a more convergent approach (Scheme 3). Compound
14a displayed potent inhibition of VEGF-R2 in vitro
and was shown to be orally efficacious at inhibiting cor-
neal neovascularization in vivo. Interfering with the
VEGF signal transduction pathway with pharmacologi-
cal agents represents a potential method for reducing or
preventing vision loss associated with uncontrolled reti-
nal angiogenesis.
References and notes
1. Aiello, L. P.; Avery, R. L.; Arrigg, P. G.; Keyt, B. A.;
Jampel, H. D.; Shah, S. T.; Pasquale, L. R.; Thieme, H.;
Iwamoto, M. A.; Park, J. E.; Nguyen, H. V.; Aiello, L. M.;
Ferrara, N.; Park, G. L. New Engl. J. Med. 1994, 1480.
2. Adamis, A. P.; Miller, J. W.; Bernal, M. T.; D’Amico, D.
J.; Folkman, J.; Yeo, T. K.; Yeo, K. T. Am. J. Opthamol.
1994, 445.
3. Ozaki, H.; Seo, M.-S.; Ozaki, K.; Yamada, H.; Yamada,
E.; Okamoto, N.; Hofmann, F.; Wood, J. M.; Campo-
chiaro, P. A. Am. J. Pathol. 2000, 697.
4. Emanuel, S.; Gruninger, R. H.; Fuentes-Pesquera, A.;
Connolly, P. J.; Seamon, J. A.; Hazel, S.; Tominovich, R.;
Hollister, B.; Napier, C.; D’Andrea, M. R.; Reuman, M.;
Bignan, G.; Tuman, R.; Johnson, D.; Moffatt, D.;
Batchelor, M.; Foley, A.; O’Connell, J.; Allen, R.; Perry,
M.; Jolliffe, L.; Middleton, S. A. Mol. Pharmacol. 2004,
66, 635.
Figure 1. Corneal micropocket assay. Pellets containing bFGF were
implanted in one eye of C57BL6 mice to induce a vascular response.
(a) Representative photos of neovascularization induced by pellets
containing bFGF and dosed with 0.5% methylcellulose (vehicle) or
125 mg/kg compound 14a. (b) Inhibition of bFGF-induced retinal
neovascularization by compound 14a was assessed in mice following
daily oral administration for 6 days at the indicated dose levels. Results
are expressed as total vessel length in mm. Bars, SE; *p < 0.05,
**p< 0.01, compared to the vehicle + bFGF positive control group by
one-way ANOVA with Dunnett’s post hoc test.
5. Batchelor, M. J.; Moffat, D. F. C.; Davis, J. M.;
Hutchings, M. C. U.S. Patent 6,579,983, 2003.
6. Glide 3.5, Schrodinger, New York, 2005.
7. (a) Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren,
T. A.; Klicic, J. J.; Mainz, D. T.; Repasky, M. P.; Knoll,
E. H.; Shaw, D. E.; Shelley, M.; Perry, J. K.; Francis, P.;
Shenkin, P. S. J. Med. Chem. 2004, 1739; (b) Doe, J. S.;
Smith, J. J.; Roe, R. P. J. Am. Chem. Soc. 1968, 8234.
was orally available in nude mice (44 %F,
AUC = 15.6 lM-h, dose = 3 mg/Kg). The plasma levels
of 14a reach 12 lM with a t1/2 of 7 h with an oral dose
of 100 mg/kg in nude mice. Doses of 125 mg/kg and

3270
T. V. Hughes et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3266–3270
8. No scaling factors were applied to the van der Waals radii.
model in nude mice. While most compounds displayed
good oral efficacy, compound 14a was the most efficacious
analogue at suppressing tumor growth.4
Default settings were used for all the remaining parame-
ters. The top 20 poses based on Glidescore were energy-
minimized with Macromodel using both the OPLS-AA,
and the refined poses reranked using Glidescore.
9. Liu, M.; Luo, M.; Mozdziesz, D. E.; Sartorelli, A. C.
Nucleosides Nucleotides 2005, 45.
12. Kenyon, B. M.; Voest, E. E.; Chen, C. C.; Folkman, J.;
D’Amato, R. J. Investig. Ophtalmol. Vis. Sci. 1996, 37,
1625.
13. Angiogenesis was induced with bFGF in these models
because it produced a more robust vascular response than
VEGF (data not shown).
10. Details of the synthetic transformation of 11–14 will be
reported elsewhere.
11. More than 40 potent and selective VEGF-R2 kinase
inhibitors from the 4-aryl-5-cyano-2-aminopyrimidine
series were tested in an A375 in vivo xenograft tumor
14. Hirata, A.; Ogawa, S.-I.; Komentani, T.; Kuwano, T.;
Naito, S.; Kuwano, M.; Ono, M. Cancer Res. 2002,
2554.