A Novel Potent Oral Series of VEGFR2 Inhibitors Abrogate Tumor Growth by Inhibiting Angiogenesis

Article abstract of DOI:10.1021/acs.jmedchem.5b01582

This paper describes the identification of 6-(pyrimidin-4-yloxy)-naphthalene-1-carboxamides as a new class of potent and selective human vascular endothelial growth factor receptor 2 (VEGFR2) tyrosine kinase inhibitors. In biochemical and cellular assays, the compounds exhibit single-digit nanomolar potency toward VEGFR2. Compounds of this series show good exposure in rodents when dosed orally. They potently inhibit VEGF-driven angiogenesis in a chamber model and rodent tumor models at daily doses of less than 3 mg/kg by targeting the tumor vasculature as demonstrated by ELISA for TIE-2 in lysates or by immunohistochemical analysis. This novel series of compounds shows a potential for the treatment of solid tumors and other diseases where angiogenesis plays an important role.

Full text of DOI:10.1021/acs.jmedchem.5b01582

Subscriber access provided by NORTHERN KENTUCKY UNIV STEELY
Article
A novel potent oral series of VEGFR2 inhibitors
abrogate tumor growth by inhibiting angiogenesis.
Guido Bold, Christian Schnell, Pascal Furet, Paul McSheehy, Josef Brüggen, Mestan Jürgen,
Paul William Manley, Peter Drueckes, Marion Burglin, Ursula Dürler, Jacqueline Loretan,
Robert Reuter, Markus Wartmann, Andreas Theuer, Beatrice Bauer-Probst, Georg Martiny-
Baron, Peter Allegrini, Arnaud Goepfert, Jeanette Wood, and Amanda Littlewood-Evans
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01582 • Publication Date (Web): 02 Dec 2015
Downloaded from http://pubs.acs.org on December 4, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 47
Journal of Medicinal Chemistry
1
1
2
3
A novel potent oral series of VEGFR2 inhibitors abrogate tumor growth by inhibiting angiogenesis.
4
5
6
7
8
Guido Bold*, Christian Schnell, Pascal Furet, Paul McSheehy, Josef Brüggen, Jürgen Mestan, Paul W. Manley, Peter Drückes,
Marion Burglin, Ursula Dürler, Jacqueline Loretan, Robert Reuter, Markus Wartmann, Andreas Theuer, Beatrice Bauer-
Probst, Georg Martiny-Baron, Peter Allegrini, Arnaud Goepfert, Jeanette Wood and Amanda Littlewood-Evans*
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Novartis Institutes for BioMedical Research, Oncology Research, 4002 Basel, Switzerland.
Abstract
Keywords VEGFR2, Inhibitor, tumor angiogenesis, BAW2881, BFH772
This paper describes the identification of 6ꢀ(pyrimidinꢀ4ꢀyloxy)ꢀnaphthaleneꢀ1ꢀcarboxamides as a new class of
potent and selective human vascular endothelial growth factor receptor 2 (VEGFR2) tyrosine kinase inhibitors.
In biochemical and cellular assays the compounds exhibit single digit nanomolar potency towards VEGFR2.
Compounds of this series show good exposure in rodents when dosed orally. They potently inhibit VEGFꢀ
driven angiogenesis in a chamber model and rodent tumor models at daily doses of less than 3 mg/kg by
targeting the tumor vasculature as demonstrated by ELISA for TIEꢀ2 in lysates, or by immunohistochemical
analysis. This novel series of compounds shows a potential for the treatment of solid tumors and other diseases
where angiogenesis plays an important role.
Introduction
Angiogenesis is the formation of new vessels from existing ones, whereas vasculogenesis is the formation of
new vessels from hematopoietic precursors. Both processes are reliant upon vascular endothelial growth factor
(VEGF) and its family of receptors (VEGFR).1 The VEGF family, which includes VEGFꢀA, VEGFꢀB, VEGFꢀC,
VEGFꢀD and placental growth factor (PlGF), signals via VEGFR2, VEGFR1 and VEGFR3 cell surface tyrosine
kinase receptors located on the host vascular endothelium, lymphatic and hematopoietic system.2 These
receptors become autophosphorylated at the tyrosine kinase domain upon engagement of VEGF which
subsequently leads to downstream events resulting in proliferation and migration of the vasculature.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 47
2
1
2
3
4
5
6
7
8
Although these are normal processes in wound healing and during the female reproductive cycle, aberrant
angiogenesis and vasculogenesis are observed in a number of pathological conditions such as rheumatoid
arthritis, psoriasis, retinal complications and tumor growth.3
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Growth of most solid tumors is highly dependent on the presence of newly formed vessels which provide
nutrients and oxygen to the tumor microenvironment.4 VEGF has been shown to be secreted by tumor cells
thereby encouraging new growth and maintenance of a vascular network and culminating in promotion of
metastatic spread.5
Since the main proꢀangiogenic signal for vasculature stems from VEGFR2 (also known as KDR), specific
inhibition of signal transduction via the VEGF / VEGFR system is a promising approach to starve the tumor
cells of nutrients and thus impede tumor growth6 and metastasis.7
The respective biology and progress in modulating this system have been summarized in recent review articles.8
In a collaborative effort between CIBA/Novartis and Schering AG, the phthalazine Vatalanib (PTK787 /
ZK222584)9, and the anthranilamide AAL99310 have been discovered as potent and selective VEGFR2
inhibitors (see Fig. 1). In a new screen for follow up compounds, the observation that the FLT3 inhibitor
AST487 also inhibits VEGFR2 autophosphorylation (see Fig. 1) resulted in the discovery of the 6ꢀ(pyrimidinꢀ4ꢀ
yloxy)ꢀnaphthaleneꢀ1ꢀcarboxamides, as a new class of potent and selective VEGFR inhibitors. The activity but
not the structure of two of the compounds within this series, BFH772 and BAW2881, have been described
previously.11 Herein we describe their structure, synthesis and in vitro potency, which translates to excellent in
vivo activities. We compare their profile to our forerunner compounds PTK787 and AAL993 as well as
evaluating them against established kinase inhibitors, in particular Sorafenib (Bayer/Onyx pharmaceuticals) and
Sunitinib (Pfizer).
ACS Paragon Plus Environment

Page 3 of 47
Journal of Medicinal Chemistry
3
1
2
Cl
3
4
5
6
7
8
9
10
F
F
NH
NH
F
O
N
N
HN
O
OH
N
HO
N
O
PTK787
AAL993
VEGFR2: CHO cells: IC50 = 0.034 ꢀM
VEGFR2: CHO cells: IC50 = 0.0012 ꢀM
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
N
N
N
N
NH2
NH
O
N
O
O
N
F
F
F
F
N
H
N
H
NH
F
F
BAW2881
AST487
Flt3-ITD-Ba/F3 cells: IC50 > 10 ꢀM
VEGFR2: CHO cells: IC50 = 0.004 ꢀM
Flt3-ITD-Ba/F3 cells: IC50 < 0.005 ꢀM
VEGFR2: CHO cells: IC50 = 0.23 ꢀM
Figure 1: Forerunner compounds PTK787 and AAL993 are
shown in comparison to AST487 and later generation
BAW2881. Comparative inhibitory activity towards VEGFR2
versus FLT3 in cellular assays shown for the latter two
compounds.
Chemistry
The kinase inhibitor AST487 was originally designed as an FLT3 inhibitor. It was synthesized as described by
Shieh et al.12 In the FLT3ꢀITDꢀBa/F3 cell system AST487 blocked FLT3 dependent cell proliferation with an
IC50 value <5 nM.13 However, it exhibited only moderate selectivity for FLT3.14 Besides its effects on RET
kinase, AST487 also inhibited VEGFR2, VEGFR3, KIT, and ABL. Of special interest was its inhibitory activity
against VEGFR2; with the compound demonstrating an IC50 value of 0.23 µM in a VEGFꢀdriven cellular
receptor autophosphorylation assay in CHO cells transfected with the VEGFR2 receptor15 (see Fig. 1). To
understand its binding interactions with VEGFR2, we used the crystal structure of AAL993 in complex with the
kinase domain of VEGFR2.16 AAL993 binds to the inactive “DFG out” conformation of the VEGFR2 kinase as
represented in Fig. 2. In addition to hydrogen bonds formed with residues C919, E885 and D1046, the inhibitor
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 47
4
1
2
3
makes a key hydrophobic interaction with residue V916, the “gate keeper” residue. Analogous docking of
4
5
6
7
8
AST487 into the binding pocket of AAL993 placed the aminopyrimidine residue as the hinge binding motif of
the inhibitor. Molecular modeling experiments indicated that replacing the anthranilic acid core of AAL993 by
a naphthyl moiety maintains all the favorable interactions of the inhibitor with the binding pocket. This design
concept illustrated in Fig. 2 led to the synthesis of BAW288111 in which the pyridine hinge binding moiety of
AAL993 was changed to an aminopyrimidine, as is the case for AST487. In contrast to AST487, BAW2881
showed higher potency and selectivity for VEGFR2 versus FLT3. The VEGFꢀdriven cellular receptor
autophosphorylation in CHO cells of BAW2881 was inhibited with an IC50 of 4 nM, in the same range as for
AAL993 and significantly more potent than for PKT787 (see Fig. 1).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2: AAL993 (green) bound in the ATP pocket of the
VEGFR2 kinase (PDB code 5EW3).16 Only the residues
proximal to the inhibitor are represented. Hydrogen bonds
appear as dashed lines. BAW2881 docked in the pocket
according to the design concept (yellow).
Docking of BAW2881 in a homology model of the kinase domain of FLT3 in the “DFG out” conformation17
provided a hypothesis to explain the lack of FLT3 inhibitory activity of this compound. As shown in Fig. 3,
ACS Paragon Plus Environment

Page 5 of 47
Journal of Medicinal Chemistry
5
1
2
3
4
5
6
7
8
9
BAW2881 can form interactions with FLT3 similar to those formed with VEGFR2. However, the “gate keeper”
residue in FLT3 is a phenylalanine (F691). It is bulkier than in VEGFR2 which has a valine (V916) instead. As
a consequence, there is not enough space in the FLT3 ATP binding site to accommodate the upper phenyl ring
of the naphthyl core of BAW2881. Thus, the docking model suggested a steric clash between this phenyl ring
and F691 explaining the lack of FLT3 inhibitory activity of BAW2881.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3: BAW2881 docked in the ATP binding site of a
homology model of the kinase domain of FLT3 in the “DFG
out” conformation (constructed based on PDB structure
1UWH). The steric clash between the “gate keeper” residue
F691 and the naphthyl core of the compound is indicated by a
cross. The smaller valine “gate keeper” residue of VEGFR2 is
shown in green.
The synthesis of the 6ꢀ(2ꢀaminoꢀpyrimidinꢀ4ꢀyloxy)ꢀnaphthaleneꢀ1ꢀcarboxamides is described in Scheme 1:
Diarylether formation from 2ꢀaminoꢀ4,6ꢀdichloroꢀpyrimidine and 6ꢀhydroxyꢀnaphthyl carboxylate resulted in 1.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 47
6
1
2
3
Amide formation with 3ꢀtrifluoromethylaniline with DEPC as the coupling reagent (→ 2), followed by
4
5
6
7
8
hydrogenative dechlorination led to BAW2881. Alternatively, dechlorination of 1 gave the acid 3, which then
could be converted by T3P® (Nꢀpropylphosphonic acid anhydride, cyclic trimer) to the amides 4 and 5.
Acylation of 4 with methyl chloroformate gave 6.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
H2; Pd / C
O
Cl
O
THF / NEt3
N
N
N
N
75 %
NH2
NH2
HN
O
F
HN
O
F
F
F
F
F
2
BAW2881
F
H2N
DEPC
F
F
NMM / DMF
61 %
NaOH
H2O / acetone
Cl
O
HO
Cl
Cl
62 °C
N
N
N
N
+
91 %
NH2
NH2
1
OH
O
OH
O
H2; Pd / C
THF / NEt3
46 %
R
F
O
O
H2N
F
F
N
N
N
N
T3P® / DMAP
NEt3 / DMF
NH2
NH2
O
OH
HN
R
O
3
F
MeOCO-Cl
CH2Cl2 / pyridine
F
F
O
4: R = F (59 %)
N
N
71 %
N
5: R =
(36 %)
N
O
NH
H2C
HN
F
O
F
O
F
F
6
Scheme 1: Synthesis of the 6ꢀ(2ꢀaminoꢀpyrimidinꢀ4ꢀyloxy)ꢀ
naphthaleneꢀ1ꢀcarboxylic acid amides BAW2881, 4, 5 and 6.
Etherification of 4,6ꢀdichloroꢀpyrimidine resulted in 7 (see Scheme 2). Substitution of the chlorine by azide and
reduction led to 6ꢀ(6ꢀaminoꢀpyrimidinꢀ4ꢀyloxy)ꢀnaphthalene carboxylate (8), which then was transformed into
the amides 9-11. Amide formation on 7 yielded the 6ꢀ(6ꢀchloroꢀpyrimidinꢀ4ꢀyloxy)ꢀnaphthalene carboxamide
ACS Paragon Plus Environment

Page 7 of 47
Journal of Medicinal Chemistry
7
1
2
3
4
5
6
7
8
12. Carbonylation in ethanol employing Pdꢀcatalysis then provided the ethylester 13, which was reduced by
NaBH4 in tꢀbutanol to the hydroxymethyl derivative BFH772. Replacement of the chlorine in 12 by cyanide
gave the nitrile 14.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R2
N
O
N
O
H2N
R1
N
N
T3P® / DMAP
NEt3 / DMF
NH2
NH2
O
OH
HN
O
8
R1
R2
1. NaN3 / DMF
9: R1 = CF3, R2 = H (64 %)
10: R1 = CF3, R2 = F (70 %)
11: R1 = Et, R2 = H (56 %)
57 %
65 °C
2. H2; Pd/C
N
O
HO
NaOH
N
Cl
H2O / acetone
N
+
N
87 %
Cl
Cl
O
OH
O
OH
7
F
T3P® / DMAP
NEt3 / CH3CN
H2N
F
F
92 %
O
CO
EtOH / Et3N
N
N
O
PdCl2(PPh3)2
N
N
78 %
Cl
HN
O
F
HN
O
O
O
F
F
F
F
F
12
13
KCN / DABCO
DMSO / H2O
NaBH4
tBuOH 65 °C
39 %
61 %
55 °C
N
N
O
N
O
N
N
HN
O
F
HN
O
F
HO
F
F
F
F
14
BFH772
Scheme 2: Synthesis of the 6ꢀ(pyrimidinꢀ4ꢀyloxy)ꢀ
naphthaleneꢀ1ꢀcarboxylic acid amides 9-11, 13, BFH772 and
14.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 47
8
1
2
3
Biology
4
5
6
SAR around the 6-(pyrimidin-4-yloxy)-naphthalene-1-carboxylic acid phenylamide structure type series
7
8
9
Table 1 summarizes the SAR around the 6ꢀ(pyrimidinꢀ4ꢀyloxy)ꢀnaphthaleneꢀ1ꢀcarboxylic acid phenylamide
structure type series in respect of its inhibitory potency against VEGFR2. Potency was assessed using a
biochemical assay for VEGFR inhibition and a cellular test system addressing CHO autophosphorylation
described in detail by Wood et al.15 An additional substituent at position 6 of the pyrimidine (2ꢀaminoꢀ6ꢀchloroꢀ
pyrimidine derivative 2) was not tolerated consistent with the binding model (Fig. 2) indicating a steric clash of
the chlorine atom with the kinase hinge region. Moving the amino substituent to the 6ꢀposition of the
pyrimidine led to 9, which exhibited similar activity compared to BAW2881. This was expected due to the
ability of aminopyrimidine groups to form bidentate hydrogen bonds with the hinge segment centered around
C919. A 6ꢀcyano derivative (14) was a slightly weaker inhibitor. Even less active was the 6ꢀethoxyꢀcarbonyl
derivative 13 which, according to the model, was also clashing in the hinge region. In contrast a 6ꢀ
hydroxymethyl substituent on the pyrimidine (BFH772) led to a potent derivative. In the aniline part, a blocking
fluorine at the 4ꢀposition of the phenyl ring was well tolerated (4, 10). Replacing the 3ꢀtrifluoromethyl group by
ethyl led to the highly active derivative 11. Expanding the 4ꢀposition of the anilide with a piperazinylꢀmethyl
group (5), as is the case for AST487, retained kinase activity but decreased the cellular potency for VEGFR2.
Compound 5 was less selective (data not shown) targeting other kinases more potently than BAW2881. Based
on the binding model, the piperazinyl group was expected to establish bifurcated Hꢀbonds with the backbone
carbonyl groups of VEGFR2 residues Ile1025 and His1026. These interactions have been observed in several
crystal structures of inhibitors binding to the “DFG out” conformation of kinases. The loss of selectivity
observed with this modification is consistent with the conserved nature of backbone interactions. The decrease
in VEGFR2 cellular potency despite the formation of new favorable interactions may thus be ascribed to a
detrimental effect of the piperazinyl group on compound cell permeability. Acylation of the 2ꢀaminoꢀpyrimidine
group was tolerated with regard to VEGFR2 inhibition (6).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 9 of 47
Journal of Medicinal Chemistry
9
1
2
3
R2
2
N
O
4
4
5
6
7
R4
R3
N
6
R1
O
N
H
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Inhibition of VEGFR2 Inhibition of VEGFR2
transphosphorylation
(biochemical)
autophosphorylation
(CHO cells)
cpd
R1
R2
R3
R4
(ꢀM)
(ꢀM)
AST487
0.003 ± 0.001
0.009 ± 0.003
1.18 ± 0.21
0.23 ± 0.06
0.0042 ± 0.0002
0.4 a
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
BAW2881
2ꢁNH2
2ꢁNH2
6ꢁNH2
CF3
CF3
CF3
2
9
6ꢁCl
0.004 ± 0.0002
0.002 ± 0.001
14
13
6ꢁCN
CF3
CF3
0.037 ± 0.003
0.34 ± 0.14
0.021 ± 0.005
2.0 a
6ꢁCOOEt
BFH772
6ꢁCH2OH
CF3
0.0027 ± 0.0009
0.0046 ± 0.0006
4
10
11
5
2ꢁNH2
6ꢁNH2
CF3
CF3
Et
F
F
0.007 ± 0.001
0.006 ± 0.002
0.004 ± 0.001
0.003 ± 0.001
6ꢁNH2
0.001 ± 0.0003
0.006 ± 0.003
0.001 ± 0.0002
0.05 a
2ꢁNH2
CF3
CF3
6
2ꢁNHCOOMe
F
0.025 ± 0.010
0.004 ± 0.001
Table 1: SAR of 6-(pyrimidin-4-yloxy)-naphthalene-1-carboxamides in respect to VEGFR2 inhibition ꢀ Mean IC50 values in µM
± SEM are shown; a single experiment
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 47
10
1
2
3
4
5
6
Biochemical IC50 data on VEGFR family kinase activity
7
8
9
In our biochemical assay, the naphthaleneꢀ1ꢀcarboxamide series generated potent IC50s for VEGFR2 which
were slightly more potent than the earlier generation compounds PTK787 and AAL993. The multiꢀkinase
inhibitor Sunitinib from Pfizer and Bayer and Onyx’s VEGFR, PDGFR and RAF compound Sorafenib were of
equal potency in the hVEGFR2 kinase assay. BAW2881, 4, 9 and 10 all lost potency (17ꢀ, 46ꢀ, 9ꢀ and 21ꢀfold
respectively) against murine enzyme (FLKꢀ1) compared to hVEGFR2. In addition, they also lost potency
towards human VEGFR1 (FLTꢀ1) and human VEGFR3 (FLTꢀ4) versus hVEGFR2. BFH772 was highly
effective at targeting VEGFR2 kinase with an IC50 value of 3 nM, however lost 500ꢀfold potency on FLKꢀ1,
FLTꢀ1 and FLTꢀ4 (Table 2).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Kinase
BAW2881
BFH772
4
9
10
PTK787
AAL993
Sorafenib Sunitinib
IC50 (ꢀM)
hVEGFR2 (KDR)
0.009 ±
0.0037
0.0027 ±
0.0009
0.007 ±
0.001
0.004 ±
0.0002
0.006 ±
0.002
0.055 ±
0.001
0.023 ±
0.006
0.0078 ±
0.0009
0.0102 ±
0.0019
mVEGFR2 (FLKꢁ1)
hVEGFR1 (FLTꢁ1)
hVEGFR3 (FLTꢁ4)
0.165 ±
0.064
1.5 ± 0.53
1.7 ± 0.36
1.1 ± 0.29
0.320 ±
0.097
0.037 ±
0.003
0.125 ±
0.026
0.27 ± 0.04
0.055 ±
0.013
nd
0.073 ±
0.008
0.82 ±
0.19
0.223 ±
0.077
0.081 ±
0.002
0.333 ±
0.109
0.077 ±
0.012
0.13 ±
0.081
0.402 ±
0.146
0.015 ±
0.009
0.42 ±
0.015
0.177 ±
0.052
0.050 ±
0.006
0.235 ±
0.043
0.064 ± 0.02
0.018 ±
0.001
0.245 ±
0.08
0.024 ±
0.0038
Table 2: Inhibition of VEGFR family kinases ꢀ Mean IC50values in µM ± SEM are shown. Kinase activity was tested by measuring
the phosphorylation of a synthetic substrate (poly[Glu, Tyr]), by purified GSTꢀfusion kinase domains of receptor tyrosine kinases, in
the presence of radiolabelled ATP; ATPꢀconcentrations used were optimized within the Km range for the individual kinases. nd is not
determined.
ACS Paragon Plus Environment

Page 11 of 47
Journal of Medicinal Chemistry
11
1
2
3
Selectivity – biochemical IC50 data on other kinases
4
5
6
7
8
Overall the naphthaleneꢀ1ꢀcarboxamide series displayed similar selectivity to the forerunner compounds
PTK787 and AAL993. BAW2881 inhibited a limited number of kinases including cꢀRAF, BꢀRAF, RET, ABL
and TIEꢀ2 at subꢀµM IC50s. BFH772 was highly selective; apart from inhibiting VEGFR2 at 3 nM IC50, it also
targeted BꢀRAF, RET and TIEꢀ2, albeit with at least 40ꢀfold lower potency. These two compounds were
inactive (IC50 > 10 µM; > 2 µM for cKIT) against all other tyrosine specificꢀ and serine/threonineꢀspecific
protein kinases tested. Compounds 4, 9 and 10 were also relatively specific but in addition to the
aforementioned kinases, also showed activity against a number of other protein kinases including cKIT, and
weakly against cꢀFMS. Sorafenib and Sunitinib displayed a broader spectrum of activity; with additional IC50
values generated for FLT3, cꢀSRC, cMET, EphB4, CDK1 and FGFR3 (Supplementary Fig. 1).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cellular VEGFR2 assays
In cellular VEGFꢀinduced VEGFR2 autophosphorylation assays, PTK787 was previously reported to have an
IC50 of 17 ± 2 nM and 34 ± 2 nM in HUVEC and CHO cells respectively9,15 whereas AAL993 was more potent
at 1.2 ± 0.19 nM in the latter assay.10 Sunitinib and Sorafenib, gave IC50 values of 8.75 ± 0.75 nM and 8 ± 0.5
nM respectively when profiled in our VEGFR2 autophosphorylation assay in CHO cells. The VEGFR2
inhibitors of the 6ꢀ(pyrimidinꢀ4ꢀyloxy)ꢀnaphthaleneꢀ1ꢀcarboxamide structure class as exemplified by
BAW2881, BFH772, 4, 9 and 10 revealed 5ꢀ to 15ꢀfold enhanced potency when compared to PTK787. They
were in the same range as AAL993, Sorafenib and Sunitinib, exhibiting single digit nanomolar potency in the
CHO cellular assay (Table 3).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 47
12
1
2
3
IC50 (ꢀM)
BAW2881
BFH772
4
9
10
PTK787
AAL993
4
5
6
7
VEGFR2
(CHO)
0.0042 ± 0.0002 0.0046 ± 0.0006
0.004 ± 0.001
0.002 ± 0.001
0.003 ± 0.001
0.034 ± 0.002
0.0012 ± 0.0002
8
9
VEGFR2
(HUVEC)
0.0029 ± 0.0003
0.003 ± 0
0.004
0.003
0.0023
0.017 ± 0.002
nd
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3: Inhibition of CHO and HUVEC VEGFR2 phosphorylation assays – Mean IC50 values (µM) for inhibition of KDR
tyrosine kinase in CHOꢀVEGFR2 and HUVEC cells. nd is not determined.
Cellular selectivity assays
RTK phosphorylation assays
BFH772 and BAW2881 were tested in an ELISA based assay measuring phosphorylation of the receptor
tyrosine kinase after cells were stimulated with the respective ligand.
BFH772 and BAW2881 both inhibited the ligand induced autophosphorylation of RET, PDGFR and KIT
kinases with IC50 values ranging between 30 and 160 nM. The compounds were selective (IC50 values > 0.5
µM) against the kinases of EGFR, ERBB2, INSꢀR and IGFꢀ1R and against the cytoplasmic BCRꢀABL kinase
(Supplementary Fig. 2).
BaF3 cellular assays
The potential for compounds to inhibit various kinases in a cellular context was also determined in the Ba/F3
assay system consisting of wildꢀtype ILꢀ3ꢀdependent hematopoietic Ba/F3 cell models rendered ILꢀ3
independent by transduction with various constitutively active tyrosine kinases.18 All compounds of the
naphthaleneꢀ1ꢀcarboxamide series were 20ꢀ to 100ꢀfold more potent than PTK787 and approximately the same
order of magnitude as AAL993 and Sorafenib in the cellular Ba/F3 VEGFR2 assay (Table 4). In a series of
additional Ba/F3 cell lines, all naphthaleneꢀ1ꢀcarboxamides targeted RET and PDGFRβ kinases in addition to
VEGFR2, but displayed weak (> 1 µM) activity against other kinases such as FGFR3 and cꢀKIT (Table 4).
Sorafenib and Sunitinib were less selective, generating IC50 values for many of the kinases.
ACS Paragon Plus Environment

Page 13 of 47
Journal of Medicinal Chemistry
13
1
2
3
4
5
6
7
8
Ba/F3 line
IC50 (ꢀM)
BAW2881
BFH772
4
9
10
PTK787
AAL993
Sorafenib
Sunitinib
9
TelꢁKDRꢁ
myc1.3
0.004 ±
0.001
0.002*
7.5**
>10
0.009 ±
0.002
0.002 ±
0.001
0.005 ±
0.002
0.215 ±
0.043
0.013 ±
0.004
0.014 ±
0.0023
0.238 ±
0.018
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
BcrꢁABL
4.748 ±
0.844
>10
>10
>10
9.5**
>10
>10
>10
>10
>10
>10
>10
>10
>5.5
>5.5
5.6*
5.82 ±
0.237
7.177 ±
2.534
NPMꢁALK
>10
1.015 ±
0.12
1.154 ±
0.1302
ERBB2
>10
>10
4.208 ±
0.173
3.846 ±
0.572
(V659E)
TelꢁFGFR3
8.271 ±
0.988
6.540 ±
0.290
8.257 ±
0.091
2.279 ±
0.041
2.658 ±
0.054
>10
>10
>10
>10
>10
>10
>5.5
>5.5
8.4*
>5.5
>5.5
2.9*
2.458 ±
0.156
1.561 ±
0.295
FLT3ꢁITD
>10
>10
>10
>10
>10
7.2**
>10
8.2*
5.646 ±
1.974
4.778 ±
0.152
0.012 ±
0.0021
0.035 ±
0.0028
TelꢁIGFꢁ1R
KIT(D816V)
TrpꢁMET
8.9*
>10
8.0**
>10
>10
7.8**
>10
3.681 ±
0.881
1.289 ±
0.412
8.506 ±
0.236
2.804 ±
0.33
1.469 ±
0.247
>10
>10
7.40 ±
0.786
5.124 ±
0.101
PTC3ꢁRET
0.152 ±
0.036
0.104 ±
0.026
0.226 ±
0.005
0.009 ±
0.002
0.040 ±
0.012
0.078 ±
0.012
0.297 ±
0.0078
Telꢁ
0.009 ±
0.002
0.027 ±
0.003
0.016 ±
0.001
0.004 ±
0.001
0.010 ±
0.004
0.076 ±
0.026
0.019 ±
0.0055
0.013 ±
0.0025
0.0023 ±
0.0007
PDGFRbeta
WT (untransꢁ
duced)
>10
>10
>10
>10
>10
>10
>10
>10
>10
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 47
14
1
2
3
4
5
6
7
8
9
Table 4: Cellular selectivity in Ba/F3 assay ꢀ The concentrationꢀdependent effect of test compound on proliferation and viability of
Ba/F3 cell models was determined by the resazurin sodium reduction assay. IC50 values (µM) from at least two independent
experiments, unless indicated otherwise, are presented as mean ± SEM. *Only one determination. **At least one replicate value > 10
µM. IC50 is defined as the drug concentration inhibiting Ba/F3 cell viability to 50 % of that measured for vehicleꢀtreated controls. All
cells except parental Ba/F3ꢀwt were grown in the absence of ILꢀ3.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Taking these data together, overall this new series of compounds displayed selectivity against a panel of kinases
comparable to PTK787 and AAL993, but more selective than Sorafenib and Sunitinib. The naphthaleneꢀ1ꢀ
carboxamides were several fold more potent than PTK787 against VEGFR2. Interestingly, the naphthaleneꢀ1ꢀ
carboxamides largely showed equal potency in the enzyme tyrosine kinase profiling and in cellular assays. This
suggested that the compounds efficiently penetrated the cytoplasmic membrane of cells in the latter assay, a
phenomenon that we have previously observed.19
Activity in VEGF induced proliferation (Fig. 4)
The naphthaleneꢀ1ꢀcarboxamides were subsequently tested in a functional assay before in vivo studies were
initiated. In HUVEC endothelial cell proliferation assays, PTK787 was previously demonstrated to exhibit an
IC50 of 7.1 ± 5.1 nM against VEGFꢀinduced proliferation.9, 15 AAL993 and Sunitinib exhibited an IC50 of 1 ±
0.16 nM and 7 ± 6 nM respectively, whereas the naphthaleneꢀ1ꢀcarboxamides were substantially more potent.
IC50 values of BAW2881 (0.12 ± 0.06 nM, n=4), compound 10 (0.003 ± 0.002 nM, n=2), BFH772 (<0.01 nM,
n=2 see Fig. 4) and compound 4 (0.077 nM see Fig. 4) demonstrated that they abrogated VEGF induced
proliferation at remarkably low nM concentrations. Compound 9 at IC50 values of 1.7 ± 1.6 nM, n=2 was more
similar to the anthranilic acid amide AAL993. The compounds were selective against VEGF induced
proliferation. Only 10 revealed an IC50 of 206 ± 44 nM (n=2) against FGF induced proliferation of HUVEC
cells, which nevertheless was several orders of magnitude higher than for VEGFꢀinduced proliferation. The
other compounds were >3 µM against bFGF. We speculate that the reason for the extremely low IC50 values
ACS Paragon Plus Environment

Page 15 of 47
Journal of Medicinal Chemistry
15
1
2
3
4
5
6
7
8
9
obtained in the VEGF induced HUVEC proliferation assay are due to the fact that these compounds are able to
penetrate the cytoplasmic membrane. Due to the relatively long duration (48 h) of the experiment, we
hypothesized that the inhibitors accumulated in the endothelial cells and that this translated into the sub
nanomolar efficacy. We were able to rule out potential cytotoxic effects since compounds started to affect
viability of HUVEC cells as assessed by the YoPro assay only at 10 µM and above (data not shown).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4: HUVEC proliferation assay in the presence of
VEGF (10 ng/ml) (2 left hand graphs) or bFGF (0.5 ng/ml)
(right hand graph). Basal proliferation shown with the white
bar, and growth factor induced proliferation with the black
bar. Compounds were added at the doses shown below the
bars – eg V + 0.01 is VEGF plus 0.01 nM compound; b + 0.1
is bFGF plus 0.1 nM compound. BFH772 in the left hand
graph shows full inhibitory activity at the lowest dose of 0.01
nM against VEGF induced proliferation. Cpd 4 is shown for
comparison in the middle graph. Cpd 4 showed no activity
against bFGF induced proliferation (right hand graph).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 47
16
1
2
3
Pharmacokinetics
4
5
6
7
8
To determine whether the most potent cellular and biochemical VEGFR2 compounds of the 6ꢀ(pyrimidinꢀ4ꢀ
yloxy)ꢀnaphthaleneꢀ1ꢀcarboxylic acid phenylamides series were orally bioavailable, plasma concentrations were
measured after administration to mice in a preliminary screen (not shown) and subsequently in rats as
described.9, 15 Even though manifesting low aqueous solubility only, most compounds showed exposure after
oral administration (Fig. 5A). In the preliminary mouse PK screen, the potent cellular and biochemical
VEGFR2 inhibitor compound 11, showed a shorter halfꢀlife than compounds 4, 10, 9, BFH772 and BAW2881.
We hypothesized that 11 could be metabolically labile since it contains an ethyl side chain in its structure. As
we had many other equipotent alternatives, we chose to deprioritize this inhibitor. Rat PK demonstrated that
halfꢀlives were variable with 10 having the longest duration of 18.3 h when applied intravenously. Graphs of 3
of the compounds, BAW2881, BFH772 and 4 are shown (Fig. 5B). The pharmacokinetic profile of the
compounds suggested that this series would be suitable for testing in in vivo models.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 17 of 47
Journal of Medicinal Chemistry
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5: (A) – Table showing dose normalized (dn) PK parameters over 24 h in a set of rat PK experiments (female OFA rats for
BAW2881 and BFH772; female Sprague Dawley rats for cpds 4, 10, and 9). At 5, 15, 30 min, 1, 2, 4, 6, 8, and 24 h time points, drug
concentrations in plasma was determined by HPLC/UV. Data represent SEM (n=2ꢀ4 animals per group). CL is clearance, Vss is
volume of distribution at steady state. The value Cmax. represents the highest observed concentration (dose normalized) at the indicated
time point (tmax.). T is the halfꢀlife of the compounds exposure in plasma. AUC is the area under the curve – total exposure over 0 h
½
→ infinity, dose normalized. BA is percent bioavailability. nm is not measured. (B) – Example PK, time dependent exposure profiles
for BAW2881 (n=4), BFH772 (n=2 for iv and n=3 for po) and cpd 4 (n=2 for iv and n=3 for po) shown. Grey triangles are iv
administration, and black diamonds are p.o. dosing.
Activity in vivo
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 47
18
VEGF skin implant miles assay in mouse
1
2
3
4
5
6
7
8
Two compounds of the naphthaleneꢀ1ꢀcarboxamide series were evaluated (4 and BFH772) in a VEGFꢀinduced
vessel leakage miles assay20 to confirm that the compounds would be suitable for dosing in longer term in vivo
models. Doses were chosen based on rodent PK calculations and potency of the compounds in the various in
vitro assays. For example, at 24 h, dose normalized exposure of compound 4 was 15 nM, for BAW2881 was 31
nM and for BFH772 was 83 nM, all of which were still more than the IC50 value for VEGFR2 in vitro. We
therefore opted to test inhibitors at 3 mg/kg initially since we expected that this would be sufficient to fully
inhibit the VEGFR over the time period. Evans blue was injected intravenously into the mouse tail vein and 30
min later, 2 µL of dog VEGF164 (10 ng/µL) was injected intradermally into the dorsal pinna of the ear of the
mouse. Vascular permeability was visualized by Evans blue extravasation measured with a computer assisted
system. Compound 4 was applied 48, 24, and 8 h before sacrifice and BFH772 at 48 and 24 h before sacrifice.
Compared to time 0, both VEGFR2 inhibitors, but especially 4 due to its long halfꢀlife substantially inhibited
extravasation, even at 48 h prior to application of VEGF and Evans blue (Supplementary Fig. 3).
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
VEGF-induced angiogenesis chamber model
The miles assay gave us confidence to assess the compound activity in more complex long term in vivo models.
The inhibition of a VEGFꢀmediated angiogenesis response after oral administration of the naphthaleneꢀ1ꢀ
carboxamide inhibitors was measured in a growth factor chamber implant model in mice.9, 15 Over a 4 day
period, a new blood vesselꢀrich tissue is formed around the VEGF agar containing implanted chamber. This
tissue can be removed, weighed and TIEꢀ2 protein as a measure of endothelial cell amount and therefore
vascularity, assessed in the lysate.
As BFH772 exhibited a good PK profile in rats with good Cmax and halfꢀlife characteristics as well as its
potency and relative selectivity towards VEGFR2 in vitro, we considered this compound to be one of the best in
the series. We also studied 4 in more detail as this compound exhibited a different profile to BFH772 in the
pharmacokinetic analysis (mice), having an extended halfꢀlife of 10 h compared to 4 h for BFH772 but with a
ACS Paragon Plus Environment

Page 19 of 47
Journal of Medicinal Chemistry
19
1
2
3
4
5
6
7
8
9
similar selectivity and potency in the in vitro profiling. Dose response curves of BFH772 at 0.3, 1 and 3 mg/kg
demonstrated that even at the lowest concentrations, this naphthaleneꢀ1ꢀcarboxamide inhibited VEGF induced
tissue weight and TIEꢀ2 levels, but only reach statistical significance at 1 mg/kg and above (Figure 6 A and B).
Compound 4 on the other hand, already significantly reduced the tissue weight and TIEꢀ2 levels back to basal at
a dose of 0.3mg/kg (Figure 6 C and D). Other compounds of this series, 9, 10 and BAW2881 were also
substantially more potent than the two forerunners PTK787 (IC50 of 30 mg/kg) and AAL993 (IC50 7 mg/kg)10 on
tissue chamber weight (Supplementary Fig. 4 and data not shown). The mice tolerated the compounds well and
no signs of toxicity were observed. Vandetanib, (an EGFR, VEGFR and RET tyrosine kinase inhibitor
developed by Astra Zeneca) and SU6668 (a VEGFR, PDGFR, cKIT and bFGF inhibitor from Sugen) at 50
mg/kg inhibited tissue weight with similar efficacy to AAL993 and PTK787 (supplementary figure 4).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 47
20
1
2
3
4
5
6
7
8
9
Figure 6: Dose response curve of BFH772 on tissue weight (A) and TIEꢀ2 levels (B), and compound 4 on tissue weight (C) and TIEꢀ
2 levels (D). PEG200 vehicle treated animals were either implanted with agar chambers as a negative control (black bar) or with
VEGF containing chambers as a positive control (white bar). BFH772 (0.3, 1 and 3 mg/kg) or compound 4 (0.3 and 3 mg/kg) treated
mice were implanted with VEGF containing chambers (grey bars). Porous chambers containing VEGF (2 µg/ mL) in 0.5 mL of 0.8 %
w/v agar (containing heparin, 20 U/mL) were implanted subcutaneously in the flank of female FVB mice. Compounds were
administered p.o. once daily in a vehicle of 100 % PEG200 (dose volume 5 mL/kg). Treatment was started 4ꢀ6 h before implantation
of the chambers and continued for 4 days. The chambers were removed 24 h after the last dose. Each experiment was performed on
groups of 6 mice per dose. Values are mean ± SEM.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Tumor models
B16 melanoma
The antiꢀtumor activity of these naphthaleneꢀ1ꢀcarboxamides was evaluated in an orthotopic B16 melanoma
model in B16/BL6 mice as described by LaMontagne et al.20 A primary tumor forms in the ear where the cells
are implanted and spontaneously formed cranial lymph node metastases occur in 100 % of the cases making this
a very reproducible tumor model. Briefly, the highly metastatic subline B16/BL6 derived from B16 melanoma
cells was injected intradermally under the skin of the ear of an anaesthetized mouse and primary tumor growth
measured over the following 21 days. After this time, the mice were sacrificed and cranial lymph node
metastases were excised and weighed.
ACS Paragon Plus Environment

Page 21 of 47
Journal of Medicinal Chemistry
21
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7: (A) – Table showing compound, dose, number of experiments performed and T/C (treatment/control) for B16 primary
tumor growth and metastatic weights. Treatment was started at day 7 when the primary B16 tumor was already established after
assignment of female C57Bl6 to groups of 5ꢀ6 animals with similar mean and range of tumor size. Primary tumor was measured with
computer assisted imaging software. After two weeks of daily qd treatment (day 21 after cell injection) the animals were sacrificed
and the cervical lymph nodes weighed. Vehicle was 100 % PEG 200. n=6 animals per group. (B) and (C) – TIEꢀ2 levels in the lysates
from metastases from the B16/BL6 melanoma model. Results from analysis of BAW2881 (B) and cpd 4 (C) shown. n= 6ꢀ12 animals
per group. Values are mean ± SEM. *p < 0.05, **p<0.01 statistical significance of inhibition. (DꢀF) – DCEꢀMRI analysis of
lymphnode metastases. The parameters determined by DCEꢀMRI of Ktrans (D), iAUEC (E) and Vp (F) show the changes in
individual mice and the associated mean ± SEM (n=7ꢀ8 animals), with significance determined by a 1ꢀway ANOVA, where *p<0.05,
**p<0.01 compared to vehicleꢀtreated mice.
BAW2881, 4, 9, 10 and BFH772 all at 3 mg/kg orally dosed once per day potently inhibited melanoma growth
(by 54ꢀ90 % for primary tumor and 71ꢀ96 % for metastasis growth) as depicted by treatment to control ratios
(Fig. 7A). At 3 mg/kg, the naphthaleneꢀ1ꢀcarboxamides showed the equivalent or better inhibition of primary
and metastatic tumor growth that 100 mg/kg PTK787 could achieve. AAL993, Sunitinib and Sorafenib had to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 47
22
1
2
3
4
5
6
7
8
9
be dosed at between 50ꢀ100mg/kg to attain similar efficacy while SU6668 and Vandetanib were substantially
less potent (Figure 7 and supplementary figure 5). The naphthaleneꢀ1ꢀcarboxamides were well tolerated, with
no obvious effects on the wellꢀbeing or behavior of the mice. Body weight and spleen weight was unaffected by
the administration of the naphthaleneꢀ1ꢀcarboxamides compared to a healthy naive mouse (data not shown). At
the end of the experiment blood was taken from the animals to look for any cellular abnormalities. Blood
chemistry showed that mice with high tumor burden (vehicle treated animals) showed lower hematocrit levels
and red blood cell counts than compound treated animals probably due to the fact that the tumor was full of
leaky blood vessels providing an avenue for loss of blood components. Hematocrit levels and red blood cell
counts returned to normal non tumor bearing naïve mouse quantities in the naphthaleneꢀ1ꢀcarboxamide treated
animal blood, presumably due to the compounds efficacy and reduction in abnormal vasculature
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(Supplementary Fig. 6 for BAW2881 and BFH772, others not shown). Other blood cell parameters measured
(white blood cells, platelets) were normal (Supplementary Fig.6 and data not shown).
Metastases were taken on day 21 and a protein lysate prepared from which TIEꢀ2 levels as a measure of
endothelial cell content, was assessed by ELISA (Fig. 7B and C). In line with the reduced tumor volumes, the
naphthaleneꢀ1ꢀcarboxamides significantly reduced TIEꢀ2 levels for BAW2881 and 4. We ruled out any direct
effects of the naphthaleneꢀ1ꢀcarboxamides on proliferation rate of the B16 melanoma cells by incubating
compounds for 48 h with tumor cells and measuring BrdU incorporation (Supplementary figure 7). These data
suggest that the naphthaleneꢀ1ꢀcarboxamides rather than affecting tumor cells, targeted the tumor vasculature
thus starving the tumor of nutrients and stunting further growth.
Consistent with antiꢀangiogenic activity, 4, and as previously described PTK78715, significantly reduced the
vascular permeability of B16 melanoma cervical metastases, as determined by DCEꢀMRI21 (Fig7DꢀF). After 3ꢀ
days of daily treatment there was a highly significant decrease in Ktrans compared to baseline, of 51 % and 61 %
for PTK787 and 4 respectively. Furthermore, the tumour plasma volume (Vp) and the parameter normally
ACS Paragon Plus Environment

Page 23 of 47
Journal of Medicinal Chemistry
23
1
2
3
utilized in the clinic to assess tumour bloodꢀvessel permeability (iAUEC), were also significantly reduced by 4
4
5
6
by approximately 65 %, while PTK787 showed a nonꢀsignificant trend for inhibition with decreases of 33ꢀ39 %.
7
8
9
We further performed dose response studies with BFH772 and 4 in the B16 melanoma model to estimate the
minimal efficacious dose. In addition, since compound 4 had a halfꢀlife of just over 10 h, we were interested to
investigate whether dosing on alternate days versus daily treatment impacted tumor growth (Figure 8).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 8: Effect of BFH772 (A) and compound 4 (B) on primary tumor growth rate and cervical lymphnode metastasis weight in the
B16 melanoma model. Treatment was started at day 7 when the primary B16 tumor was already established after assignment of
female C57Bl6 to groups of 6ꢀ10 animals with similar mean and range of tumor size. Primary tumor was measured with computer
assisted imaging software. Vehicle was 100 % PEG 200. n=10 animals per group. BFH772 and compound 4 were 6 animals per group.
BFH772 was dosed daily from treatment start. Compound 4 was dosed daily or once every 2 days from day 7. Significance
determined by a 1ꢀway ANOVA post Dunnets test, where *p<0.05, compared to vehicleꢀtreated mice.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 47
24
1
2
3
4
5
6
7
8
9
Both compounds inhibited primary and metastatic tumor growth at 3 mg/kg and BFH772 at 1 mg/kg reached
statistical significance with regard to the metastasis weight. Interestingly, 4 when applied every 2 days at 3
mg/kg attained a statistically significant inhibition of both primary and metastatic tumor weight, the latter
readout comparable to 3 mg/kg daily (Figure 8).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
BN472 mammary carcinoma
A second orthotopic tumor model using BN472 mammary carcinoma cells implanted into the breast pads of rats
was performed with BAW2881 and 4. Both compounds substantially inhibited tumor growth with BAW2881
showing a weaker efficacy than 4 with a T/C of 35.5 % versus 5.6 % respectively. In order to achieve the same
level of inhibition, PTK787 had to be dosed at 100 mg/kg compared to 3 mg/kg for 4 (Fig. 9A) while AAL993
(100 mg/kg) was reported to inhibit tumor growth with a T/C value of 0.3716 and Vandetanib (30 mg/kg) , a T/C
of 0.35. There were no significant effects on rat bodyꢀweight for any of the treatments (results not shown).
ACS Paragon Plus Environment

Page 25 of 47
Journal of Medicinal Chemistry
25
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Figure 9: (A) – BN472 tumor growth curves over 14 days
with compounds treated qd,. Vehicle was PEG300. Female
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
BrownꢀNorway rats were randomized when tumor volume
was approximately 200 mm3 before treatment started.
Experiment performed twice. n= 7ꢀ8 animals per group. (B) –
Histology of tumors after treatment with KDR inhibitors
PTK787 and cpd 4. Perfusable blood vessels were
highlighted by shortly injecting the animal with Hoechst dye
1 min before sacrifice (left panel of photomicrographs). 10
micron sections were stained with Ki67 to highlight tumor
proliferation (middle column) or caspase 3 to observe tumor
apoptosis (right column). Bar is 100 micron. All 3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 47
26
1
2
3
4
5
6
7
compounds caused a significant inhibtion of tumourꢀgrowth
compared to vehicle by dayꢀ14 of treatment (p<0.05, 2ꢀway
RMꢀANOVA).
8
9
Animals were injected intravenously with Hoechst dye shortly before sacrifice to highlight perfusable blood
vessels. Upon sacrifice, the tumors were excised and cryosections cut to visualize the Hoechst highlighted
vasculature, tumor proliferation with Ki67 staining and apoptosis with caspase 3 (Fig. 9B). Vehicle treated
tumors had a high blood vessel density which was substantially reduced in the PTK787 treated animals, and
further depleted in the 4, and BAW2881 treatment groups (Fig. 9B left column, vehicle, PTK787 and 4 shown).
By targeting the vasculature in the tumor with the VEGFR2 kinase inhibitors, the environment is predicted to
become hypoxic leading to less proliferation and more apoptosis of the tumor cells. Indeed, less tumor
proliferation was observed in the PTK787 treated animals, a feature which was augmented with the
naphthaleneꢀ1ꢀcarboxamide series (middle column). Likewise, apoptosis of tumor cells was especially
promoted upon treatment with 4 (right column).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A further study with compound 4 applied at 3 mg/kg/day, 1 mg/kg/day and 1 mg/kg every two days to rats
bearing BN472 tumors revealed T/C ratios of 0.005, 0.028 and 0.176 respectively (data not shown). This
compound was therefore highly effective even when applied every other day.
Discussion and conclusion
This paper describes the design and synthesis of a new and potent, highly selective series of VEGFR2 inhibitors.
Structure based morphing of the rather unselective FLT3 inhibitor AST487 led to the 6ꢀ(pyrimidinꢀ4ꢀyloxy)ꢀ
naphthaleneꢀ1ꢀcarboxamides, a class of potent VEGFR2 inhibitors being devoid of residual FLT3 activity. The
design was guided by the XꢀRay structure of AAL993 complexed to VEGFR2. Herein the pyridine structure
ACS Paragon Plus Environment

Page 27 of 47
Journal of Medicinal Chemistry
27
1
2
3
4
5
6
7
8
9
element of AAL993 was replaced by an aminopyrimidine as we have observed for the hinge binding motif for
AST487. Subsequently, the anthranilic acid core of AAL993 was substituted by a naphthyl moiety. These
modifications maintained all the favorable interactions between inhibitor and the ATP binding pocket of
VEGFR2 and led to the BAW2881 series. In addition, to understand the selectivity profile of the 6ꢀ(aminoꢀ
pyrimidinꢀ4ꢀyloxy)ꢀnaphthaleneꢀ1ꢀcarboxamides, BAW2881 was docked into the respective homology model
of the kinase domain of FLT3 where it sterically clashed with the “gate keeper” residue. The binding hypothesis
based on the XꢀRay structure of AAL993 bound to VEGFR2 explains the established SAR around the hinge
binding moiety of the 6ꢀ(pyrimidinꢀ4ꢀyloxy)ꢀnaphthaleneꢀ1ꢀcarboxamides: 2ꢀAminoꢀ (eg BAW2881), 6ꢀaminoꢀ
(eg compound 9) or 6ꢀhydroxymethylꢀpyrimidines (eg BFH772) are highly potent inhibitors for VEGFR2. In
contrast, bulkier hinge binders (compound 13) or 2ꢀaminoꢀ6ꢀchloroꢀpyrimidines (compound 2) are weaker
inhibitors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In cellular assays, these compounds have low nanomolar potency against VEGFR2 and exhibit very little
activity towards other kinases with the exception of RET and PDGFR. Taking into account both cellular and
biochemical assays, other potential kinases targeted were RAF, a common mutation in many cancers, and
weaklier TIEꢀ2 and cꢀKIT. VEGFR2, TIEꢀ2 and PDGFR are well established antiꢀangiogenesis targets by
disrupting the vasculature and surrounding pericytes.4,22 RET mutations are often found in thyroid cancers23
while cꢀKIT mutations are associated with GIST24 indicating that these tumor types would be potential
indications for the naphthaleneꢀ1ꢀcarboxamide series.
The high selectivity of our inhibitors gives them a potential advantage over the multiꢀkinase inhibitors in that
side effects should be reduced. Vandetanib, Sorafenib and Sunitinib target a broader spectrum of kinases and
although efficacious for cancers of the thyroid, kidney, GIST and liver, are associated with severe side effects
such as QT prolongation, cutaneous rash, and handꢀandꢀfoot skin reaction, necessitating drug interruption for
periods of time to allow the patient to recover. Although some of these side effects are not life threatening, they
do impair quality of life and raise the potential for tumors to regrow during the cessation of treatment.25 On the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 47
28
1
2
3
4
5
other side, there is a growing body of evidence from the clinic that VEGFR2 therapy can rapidly lead to
resistance as the tumor microenvironment starts to depend upon other pathways.26 How tumor cells
6
7
8
9
communicate with the stroma and vasculature is not fully understood. However, it is appreciated that interfering
with the cross talk between these with multiꢀkinase inhibitors, influences the tumor microenvironment and that
this may be a route to overcome drug resistance despite associated side effect profiles of these inhibitors.27 The
fact that there are numerous multiꢀkinase inhibitors successfully treating cancer patients would suggest that
there is a benefit to targeting a broad spectrum of signaling pathways.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The naphthaleneꢀ1ꢀcarboxamide compounds have the advantage that they are potent. They show reasonable
pharmacokinetic exposure and are able to inhibit VEGFꢀinduced vascular leakage, VEGFꢀinduced chamber
models and rodent orthotopic tumor models at doses as low as 3 mg/kg per day. BFH772 and compound 4 due
to their pharmacokinetic profile, stood out as the best two of the series, the latter compound due to its halfꢀlife
and exposure, able to show tumor weight inhibition when dosed every two days. As predicted by their mode of
action, the compounds target tissue vasculature as observed with TIEꢀ2 ELISA levels in the tissue chamber
model, and vasculature by Hoechst perfusion in a tumor model. This ultimately leads to a lower tissue weight
growing around the implanted chamber, and an increase in tumor cell death accompanied by lower amounts of
proliferation in the tumor models respectively.
Neoangiogenesis also plays an important role in other, nonꢀcancerous diseases. Amongst them are inflammatory
skin disorders such as psoriasis. In this context, a strong inhibitory effect of BAW2881 in in vitro and in vivo
psoriasis preclinical models has been demonstrated; highlighting the fact that VEGF receptor kinase inhibitors
might be useful for the treatment of patients with inflammatory skin disorders.11c Furthermore, the inhibitory
effect of BFH772 in preclinical models of skin inflammation has also been described.11d
Our findings show that this new series of naphthaleneꢀ1ꢀcarboxamides display potent, p.o. active, and wellꢀ
tolerated inhibition of VEGFꢀmediated responses. With the caveats described above for selective kinase
inhibitors versus multiꢀkinase drugs, our compounds have the potential to be suitable, not only for cancers
ACS Paragon Plus Environment

Page 29 of 47
Journal of Medicinal Chemistry
29
1
2
3
4
5
dependent on VEGF for their vascularization, but also for other diseases where VEGFꢀmediated angiogenesis
plays a crucial pathogenic role.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment