Synthesis and SAR of 4-(3-hydroxyphenylamino)pyrrolo[2,1-f][1,2,4]triazine based VEGFR-2 kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2004.12.079

A versatile synthesis of the suitably functionalized pyrrolo[2,1-f][1,2,4] triazine nucleus is described. SAR at the C-5 and C-6 positions of the 4-(3-hydroxy-4-methylphenylamino)pyrrolo[2,1-f][1,2,4]triazine template led to compounds with good in vitro potency against VEGFR-2 kinase. Glucuronidation of the phenol group is mitigated by incorporation of a basic amino group on the C-6 side chain of the pyrrolotriazine nucleus.

Full text of DOI:10.1016/j.bmcl.2004.12.079

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1429–1433
Synthesis and SAR of 4-(3-hydroxyphenylamino)pyrrolo-
[2,1-f][1,2,4]triazine based VEGFR-2 kinase inhibitors
Robert M. Borzilleri,a Zhen-wei Cai,a Christopher Ellis,a Joseph Fargnoli,b Aberra Fura,c
Tracy Gerhardt,b Bindu Goyal,a John T. Hunt,a Steven Mortillo,b Ligang Qian,a
John Tokarski,d Viral Vyas,c Barri Wautlet,b Xioping Zhenga and Rajeev S. Bhidea,*
aDepartments of Oncology Chemistry, Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, NJ 08543-4000, USA
bDepartments of Oncology Drug Discovery, Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton, NJ 08543-4000, USA
cDepartments of Metabolism and Pharmacokinetics, Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton,
NJ 08543-4000, USA
dDepartments of Structural Biology and Modeling, Bristol-Myers Squibb Pharmaceutical Research Institute, Princeton,
NJ 08543-4000, USA
Received 2 December 2004; revised 23 December 2004; accepted 30 December 2004
Available online 22 January 2005
Abstract—A versatile synthesis of the suitably functionalized pyrrolo[2,1-f][1,2,4]triazine nucleus is described. SAR at the C-5 and
C-6 positions of the 4-(3-hydroxy-4-methylphenylamino)pyrrolo[2,1-f][1,2,4]triazine template led to compounds with good in vitro
potency against VEGFR-2 kinase. Glucuronidation of the phenol group is mitigated by incorporation of a basic amino group on the
C-6 side chain of the pyrrolotriazine nucleus.
Ó 2005 Elsevier Ltd. All rights reserved.
Angiogenesis, the growth and expansion of blood vessels
from preexisting vasculature, is involved in a variety of
disease states that include diabetic retinopathy, rheuma-
toid arthritis, and tumor growth.1 Vascular endothelial
growth factors (VEGF) and their cognate receptors
(VEGFR), are critically involved in multiple processes
of angiogenesis that include increased vascular perme-
ability, endothelial cell (EC) migration, proliferation,
and survival.2 Of these receptors, VEGFR-2 is predomi-
nantly expressed on ECs and is involved in various as-
pects of tumor angiogenesis. Consequently, this
receptor tyrosine kinase has become an attractive target
for the treatment of multiple tumor types.3 Positive re-
sults from the clinical trials of Bevacizumab (AvastinTM),
a monoclonal antibody against VEGF, demonstrated
increased benefit in survival when used in combination
with cytotoxic agents.4 Accordingly, reduction in tumor
growth via disruption of the VEGF pathway is a viable
strategy for cancer therapy.
We recently reported the discovery of the pyrrolotri-
azine nucleus as a template for potent inhibitors of epi-
dermal growth factor receptor (EGFR) and VEGFR-2
(Fig. 1).5 In that preliminary report, the structure–activ-
ity relationships (SAR) focused on the effects of methyl
substitution at the 5-, 6-, and 7-positions of pyrrolotri-
azines containing different C-4 aniline substituents.
Herein we describe the expanded SAR studies of the
4-(3-hydroxyphenylamino)pyrrolotriazines on the inhi-
bition of VEGFR-2 kinase activity. This paper also
highlights a general synthesis of substituted C-5 and
C-6 pyrrolotriazines, which provided an opportunity
to improve the in vitro potency and metabolic stability
of these compounds.
Me
R5
HN
OH
4
5
N
R6
6
N
2
N
R7
Keywords: VEGFR-2; Angiogenesis; SAR; Kinase receptor.
*
Figure 1. General structure of the pyrrolo[2,1-f][1,2,4]triazine based
VEGFR-2 inhibitors reported previously.5
Corresponding author. Tel.: +1 609 252 4395; fax: +1 609 252
6601; e-mail: rajeev.bhide@bms.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.12.079

1430
R. M. Borzilleri et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1429–1433
Scheme 1. Synthesis of C-6 amides, ureas, carbamates. Reagents: (a) TosMIC, sodium hydride, DMSO, THF; (b) AlCl3, CCl3COCl, then NaOMe,
MeOH; (c) Ph2P(O)ONH2, NaH, DMF; (d) formamide, heat; (e) POCl3, heat; (f) aniline, heat; (g) LiOH; (h) EDCI, R6NH2; (i) diphenylphosphoryl
azide; (j) R6OH or R6NH2.
Pyrrolo[2,1-f][1,2,4]triazine derivatives bearing carbox-
amides and carbamates at the C-6 position were pre-
pared as shown in Scheme 1.6 b-Substituted acrylates
of general formula 1 were treated with tosylmethyl iso-
cyanide (TosMIC) in the presence of sodium hydride
to obtain 3,4-disubstituted pyrroles. Acylation at the
C-2 position of these disubstituted pyrroles with trichlo-
roacetyl chloride in the presence of aluminum chloride,
followed by treatment with sodium methoxide afforded
the tri-substituted pyrroles 2.7 The highly electron defi-
cient pyrrole nitrogen of 2 was difficult to aminate with
commercially available O-mesitylenesulfonylhydroxyl-
amine (MSH).5 However, N-amination occurred in high
yield with diphenyl phosphoryl hydroxylamine.8 The
aminated pyrrole intermediate was heated in formamide
to obtain the pyrrolotriazine nucleus 3. Treatment of 3
with phosphorous oxychloride at elevated temperature
afforded the hydrolytically-unstable chloroimidate inter-
mediate, which when treated with an appropriately
substituted aniline, provided compound 4. The ester
group at C-6 of 4 was easily manipulated to obtain var-
ious functional groups, which could be used as tethers
for incorporating the desired appendages. Thus, follow-
ing the hydrolysis of esters 4 with lithium hydroxide, the
resulting acids 5 were coupled with amines to obtain
amides 6. The acids 5 underwent Curtius rearrangement
when treated with diphenylphosphoryl azide in the pres-
ence of an appropriate alcohol or amine to afford carba-
mates 7 or ureas 8, respectively.
group (10) via a reduction-oxidation sequence. Bae-
yer–Villiger rearrangement of the aldehyde 10 occurred
upon treatment with mCPBA to provide phenol 11
(R6 = H). For compounds where R5 = i-propyl, elevated
temperatures were needed to effect this rearrangement.6
Alkylation of the C-6 hydroxyl group of 11 was
achieved either by Mitsunobu reaction or by treatment
with an alkyl halide in the presence of a base to afford
12. The phenol ether protecting group at C-4 of 12
was hydrolyzed with warm HCl, and the resulting amide
was converted to the target compounds 13 (R6 = alkyl)
as previously described in Scheme 1.
Our initial studies had revealed that the 3-hydroxy ani-
line at the C-4 position of the pyrrolotriazine ring af-
fords potent VEGFR-2 kinase inhibition,5 and was
therefore used throughout the SAR studies described
in this report. A methyl group para to the aniline nitro-
gen improved potency, however increasing the size of
the alkyl substituent led to a gradual decrease in potency
as shown in Table 1. Electron withdrawing groups (19,
20) did not improve the potency. Since the methyl group
afforded the most potent compound, it was retained dur-
ing the SAR studies focused at the C-5 and C-6 positions
of the pyrrolotriazine core.
Alkyl groups at the C-5 position in both C-6 ester and
C-6 amide series generally afforded the potent inhibitors
of VEGFR-2 kinase activity and VEGF-stimulated hu-
man umbilical vein endothelial cell (HUVEC) prolifera-
tion5 (Table 2). An alkoxy group at the C-5 position (22)
led to a 6-fold loss in biochemical potency, possibly due
to a hydrogen bond between the adjacent aniline NH
with the ether oxygen leading to an unfavorable orienta-
tion of the aniline. A bulky substituent, such as a tert-
butyl group at C-5 (24) reduced the biochemical potency
considerably. In the ester series, both n-propyl and
The ester group at the C-6 position of the pyrrolotri-
azine 3 was converted to the C-6 ether (13) as shown
in Scheme 2. Protection of the C-4 position of 3 as the
phenyl ether was carried out by conversion of the C-4
amide group to the chloroimidate (see Scheme 1) fol-
lowed by treatment with sodium phenoxide to give 9.
The ester group of 9 was converted to the aldehyde

R. M. Borzilleri et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1429–1433
1431
Scheme 2. Synthesis of C-6 ethers. Reagents: (a) POCl3; (b) sodium phenoxide; (c) LAH; (d) MnO2, toluene, heat; (e) mCPBA; (f) R6Br, NaH or R6–
OH, PPh3, DEAD; (g) HCl; (h) aniline, CH3CN, heat.
Table 1. SAR of the aniline ring
Table 2. SAR at C-5 position of pyrrolotriazine
R1
Me
HN
OH
R5
HN
Me
OH
R6
N
N
MeOOC
N
N
O
N
N
Compd
R1
VEGFR-2 IC50, nMa
Compd
R5
R6
VEGFR-2
IC50, nMa
HUVECb
IC50, nM
14
15
16
17
18
19
20
H
61
10
Me
Et
21
22
23
24
n-Pr
OEt
i-Pr
–OEt
–OEt
–OEt
–OMe
3
64
6
40
4
55
n-Pr
i-Pr
CF3
Br
190
380
370
33
14
310
t-Bu
NA
N
N
N
N
25
26
27
Me
40
90
18
10
35
8
H
a For assay conditions see Ref. 5.
N
H
n-Pr
i-Pr
i-propyl groups at the C-5 position afforded excellent
potency, however, in the metabolically more stable
amide series, only the i-propyl group maintained accept-
able potency and was used in the further SAR.
N
H
a For assay conditions see Ref. 5.
b Human umbelical vein endothelial cells stimulated by VEGF.
To better understand the binding mode of these inhibi-
tors, a binding model of 23 in the ATP site of
VEGFR-2 was built using the reported crystal structure
of the VEGFR-2 kinase domain.11 Figure 2 depicts
the proposed binding mode of 23 in the ATP binding
pocket, a model consistent with the reported binding
of similar kinase inhibitors.12
Key interactions include hydrogen bonds between the
N1 nitrogen of the pyrrolotriazine ring and the back-
bone NH of Cys919 as well as between the phenol OH
group and residues Asp1046 and Lys868 in protein
kinase ATP-binding pockets. These interactions could
potentially be exploited to design pharmacophores that
would exhibit better pharmacokinetic properties. The
C5-isopropyl group occupies a hydrophobic cleft cre-
ated by Leu840 and Leu1035 (not labeled in Fig. 2),
which also represents the pocket occupied by the ribose
ring of ATP. The buried hydrophobic surface areas of
both the ligand and protein in this region may account
for the increased binding affinity observed in this series
Figure 2. Binding model of 23 in ATP binding site of VEGFR-2
kinase.
relative to unsubstituted analogs.5 The ester group of
23 points toward surface-exposed protein and longer
appendages (27) at this position appear to be largely sur-
rounded by solvent. As such the C-6 position affords an
opportunity to incorporate moieties that may optimize
ADME properties without sacrificing potent binding.

1432
R. M. Borzilleri et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1429–1433
Preliminary in vitro stability testing revealed that early
leads suffered from high rates of in vitro metabolism
due to facile glucuronidation at the phenol oxygen.
Our SAR efforts, therefore, were directed at finding sub-
stitutions that would suppress the conjugation reaction.
The solvent exposed C-6 position of the pyrrolotriazine
was considered to be the most suitable for such a study.
Analogs with various appendages attached via different
functional groups such as amides, carbamates, ethers,
and ureas at the C-6 position were prepared and a pos-
sible interplay of substituents at C-5, and C-6 positions
on enzymatic activity and in vitro glucuronidation9 was
studied. It was observed that the compounds with neu-
tral groups (28–30, basic pKa < 6) exhibited high rates
of glucuronidation as shown in Table 3. However, intro-
duction of a basic amino group (calculated basic pKa
between 7 and 10) on an appendage attached to the C-6
position via different groups led to a substantial decrease
in the rate of glucuronidation. Thus, the in vitro glucu-
ronidation in human liver microsomes was reduced from
21% in 10 min for 29 to 1% in 10 min for 25. Similarly,
compounds 27, 31–34 also exhibited much lower rates of
glucuronidation compared compounds 28–30. At pres-
ent, it is not clear why a basic group distant from the
phenol group mitigates the conjugation reaction.
The effect of the size of the alkyl group at C-5 was mini-
mal (compare 25 and 27) but carbamates appeared to
have higher rates of conjugation compared to the
amides, ethers, and ureas (compare 29 with 30 and 31
with 32).10
Despite the very low rates of in vitro glucuronidation
and good overall metabolic stability in liver microsomes,
these compounds did not exhibit good in vivo pharma-
cokinetic properties. For example, when administered
orally to fasted mice at 50 mg/kg, 27 (in vitro metabolic
rate of 0.01 nmol/min/mg at 10 lM of substrate in liver
microsomes) demonstrated poor exposure, with
Cmax = 166 nM and AUC0–4 h = 267 nM h. When dosed
iv, high hepatic clearance and a short half-life of 30 min
was observed for this compound.
In summary, we report that compounds based on the
pyrrolo[2,1-f][1,2,4]triazine nucleus containing a 3-
hydroxyphenylamino at C-4 and various substituents
Table 3. Optimization of potency and the rate of glucuronidation
Me
OH
R5
HN
N
R6
N
N
Compd
R5
R6
VEGFR-2 IC50, nMa
HUVECb IC50, nM
NDd
Glucurdn,c nmol/min/mg
0.16
H
N
O
28
Me
150
Me
O
O
N
N
29
30
25
27
31
32
33
34
Me
i-Pr
Me
i-Pr
i-Pr
i-Pr
Me
i-Pr
30
7
20
ND
10
8
0.21
0.90
0.01
0.02
0.08
0.03
0.01
0.04
N
N
N
H
H
N
N
O
N
O
O
40
18
5
N
N
N
H
O
N
H
H
N
N
O
1
O
N
O
O
4
1
H
N
N
52
5
3
O
H
N
H
N
N
5
O
a For assay conditions see Ref. 5.
b Human umbelical vein endothelial cells stimulated by VEGF.
c Ref. 9.
d Not determined.

R. M. Borzilleri et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1429–1433
1433
N.; Fyfe, G.; Rogers, B.; Ross, R.; Kabbinavar, F. New
Engl. J. Med. 2004, 350, 2335–2342.
at C-6 of the pyrrolotriazine afford VEGFR-2 inhibi-
tors with potent cellular activity. Introduction of a
basic moiety at the C-6 position of these phenylamino
based inhibitors, while maintaining the biochemical
potency, led to a reduction in the in vitro rate of glucu-
ronidation of the phenol. Despite this success, the phar-
macokinetic properties of these inhibitors precluded
their further development. The versatile synthesis of
substituted pyrrolotriazines, the SAR elaborated in this
series and the understanding gained from modeling
studies, could be helpful in rapid SAR development
of a new series possessing improved pharmacokinetic
properties.
5. (a) Hunt, J. T.; Mitt, T.; Borzilleri, R.; Gullo-Brown, J.;
Fargnoli, J.; Fink, B.; Han, W.-C.; Mortillo, S.; Vite, G.;
Wautlet, B.; Wong, T.; Yu, C.; Zheng, X.-P.; Bhide, R. J.
Med. Chem. 2004, 47, 4054–4059; (b) 2.5 lM ATP
concentration was used in the enzyme assay.
6. For the detailed experimental see: Hunt, J. T.; Bhide, R.
S.; Borzilleri, R. M.; Qian, L. PCT Int. Appl. WO
2000071129.
7. An alternate synthesis of pyrroles 2 could be accomplished
using the published route: Suzuki, M.; Miyoshi, M.;
Matsumoto, K. J. Org. Chem. 1974, 39, 1980.
8. Klo¨tzer, W.; Baldinger, H.; Karpitschka, E.-M.; Knoflach,
J. Synthesis 1982, 592.
9. The rate of glucuronidation was determined under the
following conditions: substrate concentration, 20 lM;
protein concentration, 2 mg/mL; UDPGA, 5 mM;
pH 7.5 Tris buffer, 100 mM; magnesium chloride,
10 mM; alamethicine (in ethanol), 2.5 lg/mL. Incubations
were performed at 37 °C and were initiated by the addition
of UDPGA. 200 lL aliquots were taken at times 0, 10, and
20 min and the reaction terminated by the addition of
equal volume of acetonitrile. Following vortexing and
centrifugation, 20 lL aliquot was analyzed by HPLC-UV.
The percentage of compound remaining at each time point
was calculated and compared to the values at zero time
point. The rate of metabolism was calculated by deter-
mining the nanomoles of compound that disappeared and
dividing it by the time of incubation and the milligram
of protein.
Acknowledgements
Microanalyses, IR spectra and mass spectra were kindly
provided by the Bristol-Myers Squibb Department of
Analytical Research and Development. We are grateful
to Dr. Louis Lombardo for helpful discussions in the
preparation of this manuscript.
References and notes
1. (a) Folkman, J. Nat. Med. 1995, 1, 27–31; (b) Carmeliet,
P.; Jain, R. K. Nature 2000, 407, 249–257.
2. (a) Robinson, D. R.; Wu, Y. M.; Lin, S. F. Oncogene
2000, 19, 5548–5557; (b) Jakeman, L. B.; Winer, J.;
Bennett, G. L.; Altar, C. A.; Ferrara, N. J. Clin. Invest.
1992, 89, 244–253.
3. Kim, D. W.; Lu, B.; Hallahan, D. E. Curr. Opin. Invest.
Drugs 2004, 5, 597–604.
4. (a) Yang, J. C.; Haworth, L.; Sherry, R. M.; Hwu, P.;
Schwartzentruber, D. J.; Topalian, S. L.; Steinberg, S. M.;
Chen, H. X.; Rosenberg, S. A. New Engl. J. Med. 2003,
349, 427–434; (b) Hurwitz, H.; Fehrenbacher, L.; Nov-
otny, W.; Cartwright, T.; Hainsworth, J.; Heim, W.;
Berlin, J.; Baron, A.; Griffing, S.; Holmgren, E.; Ferrara,
10. 1H NMR (CD3OD, 300 MHz) of compound 27: d 7.80 (s,
1H), 7.55 (s, 1H), 7.01 (d, 1H, J = 8 Hz), 6.77 (s, 1H), 6.64
(d, 1H, J = 8 Hz), 3.60–3.47 (m, 3H), 3.29–3.26 (m, 2H),
3.12–3.10 (m, 2H), 3.08–2.89 (m, 2H), 2.04 (s, 3H), 1.99–
1.95 (m, 2H), 1.891.82 (m, 3H), 1.32 (d, 6H, J = 7 Hz).
11. McTigue, M. A.; Wickersham, J. A.; Pinko, C.; Showalter,
R. E.; Parast, C. V.; Tempczyk-Russell, A.; Gehring, M.
R.; Mroczkowski, B.; Kan, C.-C.; Villafranca, J. E.;
Appelt, K. Structure 1999, 7, 319.
12. Shewchuk, L.; Hassell, A.; Wisely, B.; Rocque, W.;
Holmes, W.; Veal, J.; Kuyper, L. J. Med. Chem. 2000,
43, 133.