Hit-to-lead optimization of 1,4-dihydroindeno[1,2-c]pyrazoles as a novel class of KDR kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.05.052

A series of 1,4-dihydroindeno[1,2-c]pyrazoles was prepared and evaluated for their enzymatic inhibition of KDR kinase. Computer modeling studies revealed the importance of attaching a basic side chain in predicting the binding mode of those compounds. Fur

Full text of DOI:10.1016/j.bmcl.2006.05.052

Bioorganic & Medicinal Chemistry Letters 16 (2006) 4371–4375
Hit-to-lead optimization of 1,4-dihydroindeno[1,2-c]pyrazoles as a
novel class of KDR kinase inhibitors
Jurgen Dinges,a,* Irini Akritopoulou-Zanze,a Lee D. Arnold,b Teresa Barlozzari,b
¨
Peter F. Bousquet,b George A. Cunha,b Anna M. Ericsson,b Nobuhiko Iwasaki,c
Michael R. Michaelides,a Nobuo Ogawa,c Kathleen M. Phelan,a Paul Rafferty,b
Thomas J. Sowin,a Kent D. Stewart,a Ryukou Tokuyama,c
Zhiren Xiaa and Henry Q. Zhanga
aGlobal Pharmaceutical Research and Development, Abbott Laboratories, 100 Abbott Park Road, Abbott Park, IL 60064-6101, USA
bAbbott Bioresearch Center, 100 Research Drive, Worcester, MA 01605-5314, USA
cHokuriku Seiyaku Co., Ltd, 37-1-1 Inokuchi, Katsuyama, Fukui, 911, Japan
Received 27 March 2006; revised 16 May 2006; accepted 16 May 2006
Available online 5 June 2006
Abstract—A series of 1,4-dihydroindeno[1,2-c]pyrazoles was prepared and evaluated for their enzymatic inhibition of KDR kinase.
Computer modeling studies revealed the importance of attaching a basic side chain in predicting the binding mode of those com-
pounds. Further investigation of structure-activity relationships led to 19, a lead compound with an acceptable selectivity profile,
activity in whole cells, and good oral efficacy in an estradiol-induced murine uterine edema model of VEGF activity.
Ó 2006 Elsevier Ltd. All rights reserved.
Multicellular organisms have to form new blood vessels
to maintain a sufficient supply of nutrients and oxygen
when they grow beyond a size of 1–2 mm. Initially blood
vessels are assembled from endothelial precursors (vas-
culogenesis) and are then expanded through angiogene-
sis (sprouting and intussusception).1 Angiogenesis is a
tightly controlled process, which is balanced by a num-
ber of angiogenic growth factors and endogenous inhib-
itors.2 While angiogenesis is an essential process, for
example, in embryonic development, the menstrual cy-
cle, and wound healing,3 its imbalance has been impli-
cated in various disease states such as rheumatoid
arthritis, stroke, psoriasis, and in cancer.4 At an early
stage, solid tumors depend on angiogenesis to grow, sur-
vive, and metastasize. Without sufficient vasculariza-
tion, however, they remain small and become necrotic
or apoptotic. Members of the family of vascular endo-
thelial growth factors (VEGFs), especially VEGF-A,
were identified as major regulators of these events.5
Expression of VEGF is stimulated by hypoxia and its
biological effects are mediated by the receptor tyrosine
kinases FLT1 (VEGFR1, Fms-like tyrosine kinase 1)
and, more importantly, KDR (VEGFR2, kinase insert
domain-containing receptor tyrosine kinase).6 KDR is
a transmembrane protein that is specifically expressed
in vascular endothelial cells. Upon binding of VEGF
to its extracellular ligand binding domain, KDR under-
goes dimerization and trans-autophosphorylation of its
intracellular kinase domain. This initiates a cascade of
downstream signaling events, which ultimately lead to
endothelial cell proliferation and migration. Its essential
role in tumor progression has made the inhibition of
KDR signaling a highly attractive target for therapeutic
intervention in cancer.7 In fact, small molecule KDR
inhibitors have been shown to inhibit the growth of solid
tumors derived from breast, colon, lung, and prostate in
tumor xenograft models. Clinical validation for the gen-
eral strategy can be derived from the recent FDA
approval of AvastinTM, a monoclonal antibody to
VEGF, for the treatment of colorectal cancer,8 as well
as from the advancement of ZD64749 and PTK78710
into phase III clinical trials.
High-throughput screening of our compound repository
identified the 1,4-dihydroindeno[1,2-c]pyrazole 1 (Fig. 1)
Keywords: Cancer; Receptor tyrosine kinase; KDR kinase; Inhibitor;
1,4-Dihydroindeno[1,2-c]pyrazoles.
*
as a reversible ATP-competitive inhibitor (IC50
0.93 lM) of KDR kinase. In this report, we describe
=
Corresponding author. Tel.: +1 847 935 1448; e-mail: jurgen.
dinges@abbott.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.05.052

4372
J. Dinges et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4371–4375
Figure 1. Structure of the high throughput screening hit 1 and
demonstration of pseudo-C2 symmetry in 2.
an improved synthesis of 1,4-dihydroindeno[1,2-c]pyraz-
oles and their hit-to-lead evaluation as KDR kinase
inhibitors.
Scheme 1. Reagents and conditions: (a) ethylene glycol, p-TsOH,
benzene, reflux (Dean–Stark trap); (b) i—n-BuLi, THF, ꢀ78 °C; ii—
DMF, THF, ꢀ78 °C to rt; (c) NaBH4, MeOH/THF (10:1), 0 °C to rt;
(d) p-TsOH, acetone/H2O (4:1), reflux; (e) MsCl, Et3N, THF, 0 °C; (f)
1-methylpiperazine, K2CO3, EtOH, 0 °C to rt; (g) i—NaH, phenyl
benzoate, benzene, reflux; ii—H2NNH2ÆH2O, HOAc, EtOH, reflux.
Compound 1 already represents the minimum required
pharmacophore for KDR activity. Without the
thiophene substituent in 3-position the compound is
inactive. Previous studies of 1,4-dihydroindeno[1,2-c]-
pyrazoles bound to the PDGF11 or CDK212 kinase ac-
tive sites indicated the possibility of multiple binding
modes based on the pseudo-C2 symmetry of those mol-
ecules. The situation becomes more obvious if the 3-thio-
phene substituent is replaced by a phenyl group, as
shown in 2. Using the crystal structure of KDR kinase,
this molecule then can be docked into the ATP binding
site in such a way that its N(1)-H acts as a hydrogen
bond donor to Glu 917 and its N(2) acts as a hydrogen
bond acceptor to Cys 919 in the hinge region of KDR
(mode A in Fig. 2). In this case, the phenyl group is
oriented toward the solvent. The tautomeric form of 2,
however, is able to bind in the opposite direction such
that the 3-phenyl substituent projects toward the
specificity pocket (mode B). To further investigate this
situation, we decided to attach a (4-methylpiperazin-
1-yl)methyl side chain to the core 2.
was directly reduced with sodium borohydride to give
the hydroxymethyl compound 6. After deprotection of
the keto functionality, the hydroxyl group in 7 was
mesylated to afford 8. Indanone 9 was then obtained
via nucleophilic substitution with 1-methylpiperazine,
using potassium carbonate as the base. The yield in
the subsequent formation of the tricyclic ring system
in 10 is very sensitive to the nature of the substituent
in 3-position. Therefore, we developed an improved
‘one-pot’ procedure,13 which is more amenable to
elucidate structure–activity relationships (SARs) in this
position. The 1,4-dihydroindeno[1,2-c]pyrazoles with
basic side chains in 30- or 40-position were produced in
analogy to Scheme 1, starting from 2-indanone and
the appropriate 4-methyl piperazinomethyl benzoate.14
The activity of the produced compounds to inhibit the
phosphorylation of a peptide substrate (biotin–Ahx–
AEEEYFFLFA–amide) by KDR kinase was assessed
in an HTRFÒ assay at 1.0 mM concentration of adeno-
sine 50-triphosphate (ATP).15 Table 1 shows that having
the polar side chain attached to the 5-position (11),
directed toward the ribose binding pocket, was not well
tolerated. Attachment in 6- (10) and 7-position (12) pro-
duced about equipotent compounds with a slight
improvement of activity compared to the parent
compound 2, and substitution in 30- (13) and 40-position
The general synthesis of 1,4-dihydroindeno[1,2-c]pyra-
zoles with basic side chains in the 5-, 6- or 7-position
is exemplified with the synthesis of compound 10
(Scheme 1). Indanone 3 was protected in the form of
its dioxolane 4. Halogen–lithium exchange followed by
formylation with DMF afforded aldehyde 5, which
Table 1. KDR inhibitory activity of parent compound 2 and its
(4-methylpiperazin-1-yl)methyl substituted derivatives 10–14
a
Compound
Substituent in position
KDR IC50 (lM)
2
11
10
12
13
14
No substituent
5
1.2
14.1
0.7
6
7
0.6
3.4
30
40
1.4
Figure 2. Diagram of compound 2 bound into active site of KDR
kinase (PDB 1VR2) in two orientations.
a Values are means of two experiments.

J. Dinges et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4371–4375
4373
Table 3. KDR inhibitory activity of a selected set of 1,4-dihydroin-
deno[1,2-c]pyrazoles with modifications in 6- and 7-position 27–35
(14) resulted in a slight loss of potency. Based on our
modeling studies, compound 12 cannot bind to KDR
in mode A because of severe clashes between the basic
side chain and the protein. On the other hand, com-
pound 14 is unable to bind in mode B because of the
same reason. However, both compounds have reason-
able potencies. Together with the fact that polar groups,
such as 4-methylpiperazine, will likely prefer the solvent-
accessible region of the active site, we therefore assume
that the position of the basic side chain will determine
the binding mode of our compounds. Overall, adding
a polar side chain improved the solubility of our com-
pounds and had a positive effect on their potency in
those cases where the substituent was attached in the
6- or 7-position. Both forms were considered for further
studies.
Compound R1
R2
R3
KDR
a
IC50 (nM)
S
N
27
28
29
30
31
32
33
34
35
H
1005
143
824
190
4081
665
190
108
147
N
N
H
H
H
H
H
H
H
H
N
S
S
S
S
S
S
S
S
N
N
Table 2. KDR inhibitory activity of a selected set of 1,4-dihydroin-
deno[1,2-c]pyrazoles with modifications in 3-position 15–26
N
O
N
N
N
Compound R1
R2
R3
KDR
a
IC50 (lM)
H
N
O
N
N
N
15
16
17
18
19
20
21
22
23
24
25
26
H
H
H
H
0.50
9.20
N
N
N
N
N
OH
N
N
N
N
N
N
N
S
N
1.05
N
a Values are means of two experiments.
0.40
S
S
Next, we examined the SAR of substituents in 3-posi-
tion of the 1,4-dihydroindeno[1,2-c]pyrazole system
(Table 2). Substitutions on the phenyl ring generally
resulted in a loss of potency, only the two analogs with
a hydroxyl group in 30- (not shown) and 40-position (15)
maintained activities comparable to the parent com-
pound 10. Replacing the phenyl group with six-mem-
bered aromatic heterocycles also proved to be
detrimental. For example, the 30-pyridyl derivative 16,
as the most active representative of this subseries,
showed about a 13-fold loss in KDR enzymatic potency.
Attempts to attach aliphatic residues in 3-position
rendered the compounds completely inactive. Only
thiophenes, even in comparison with other five-mem-
bered aromatic heterocycles, gave promising results.
As the comparison of the KDR IC50-values for
compounds 17 and 18 shows, attaching a thiophene in
its 30-position is better tolerated than attachment in 20-
position. A boost in potency, compared to the parent
compound 10, was then achieved by switching the basic
side chain from the 6- into the 7-position (19). In mod-
eling 19 into the active side of KDR, we note the prox-
imity of the thiophene sulfur relative to the sulfur of Cys
1045 of the enzyme. ‘Soft–soft’ molecular interactions
between the two sulfur atoms may provide a possible
N
H
0.18
N
S
N
N
N
N
N
N
N
N
H
H
H
H
H
H
H
2.84
N
N
N
N
N
N
N
N
S
7.48
S
16.64
5.47
N
S
S
12.54
>50.00
>50.00
S
S
a Values are means of two experiments.

4374
J. Dinges et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4371–4375
Table 4. Kinase inhibition profiles of 1,4-dihydroindeno[1,2-c]pyrazoles 19 and 35
a
IC50 (nM)
Compound
KDR
FLT1
FLT4
cKit
PDGFR
Tie2
FGFR
EGFR
LCK
cMET
19
35
178
148
970
601
1573
169
145
103
>50,000
>50,000
47,793
17,641
>50,000
34,355
>50,000
>50,000
>50,000
21,141
>50,000
>50,000
a Values were determined in HTRFÒ assays.
rationale for the preferred binding of the thiophene ana-
logs over other aromatic rings. We also speculate that,
depending on its position, the basic side chain could help
to orient the molecule toward this interaction. To fur-
ther expand on the SAR of the thiophene substituent,
we introduced an additional heteroatom into the system.
However, the assay results for compounds 20–22 dem-
onstrate that this change was not very well tolerated.
A ‘methyl-walk’ around the 20-thienyl group was carried
out to investigate the possibility of attaching larger sub-
stituents, which could reach into the hydrophobic spec-
ificity pocket (compounds 23–25). Here it was found
that, in comparison, the 50-position would be best suited
(23), but even the attachment of a methyl group already
resulted in about a 5-fold loss in potency. Annulation of
an additional phenyl ring as in 26 only produced inac-
tive compounds.
induced murine uterine edema model of VEGF activi-
ty15 to determine its in vivo efficacy and systemic oral
exposure. Here 35 achieved an ED50 of 20 mg/kg. Table
4 shows the inhibitory potencies of 35 and its parent
compound 19 against a panel of kinases. The com-
pounds are most active against the kinases of the VEG-
FR family as well as cKit, and overall display a
selectivity profile that is unique, compared to other
known classes of KDR inhibitors.
In summary, 1,4-dihydroindeno[1,2-c]pyrazoles were
identified as potent KDR inhibitors. Attachment of a
basic side chain allowed the prediction of the binding
mode of these compounds into the active site of KDR,
and affords improved pharmacological properties. Fur-
ther hit-to-lead optimization studies resulted in lead
compound 19, which shows an acceptable selectivity
profile, activity in whole cells, and good efficacy when
dosed orally in a primary in vivo efficacy model. Fol-
low-up work in this series will be presented in due time.
Concluding that a 30-thiophene substituent is optimal
for achieving potent KDR activity, we focused on the
optimization of the basic side chain (Table 3). Elimina-
tion of the methylene spacer in compound 17 (Table 2)
had no effect on the KDR activity (27), while extending
this tether to an ethyl group (19 vs 28) slightly gained
potency but then turned out to be detrimental in the
in vivo efficacy model. Opening of the 1-methylpiper-
azine ring and elimination of one of the basic nitrogens
was unfavorable as demonstrated with 29. Extension of
the N-alkyl group, exemplified with 30, was tolerated
but did not result in any significant improvement.
Reduction of the basicity of either nitrogen by introduc-
tion of a neighboring carbonyl functionality (31 and 32)
showed that the internal nitrogen has the more signifi-
cant contribution to the overall potency of the inhibi-
tors, but also that both basic nitrogens are required
for optimum potency. While considering aromatic sys-
tems, we identified several five-membered heterocycles
of interest. The 1H-1,2,3-triazol substituted compound
33 almost retained the potency of 19, although being a
lot less basic. Changing to its 1H-1,2,4-triazol isomer
34 then even led to an improvement of in vitro activity.
This find was of great interest to us because through
modulating the overall basicity of our compounds, we
were able to reduce the volume of distribution in their
PK profile. Unfortunately, we found that the in vivo effi-
cacy decreased with the decreasing pKa values of our
KDR inhibitors. As a consequence, the imidazole-con-
taining compound 35 was produced and showed a
KDR IC50 of 147 nM. To determine its whole cell activ-
ity, 35 (KDR cell IC50 = 293 nM) was evaluated in an
enzyme linked immunosorbent assay (ELISA) for its
ability to inhibit VEGF-induced phosphorylation of
KDR in full length human KDR-transfected NIH3T3
cells.15 The compound was then tested in an estradiol-
References and notes
1. Carmeliet, P.; Jain, R. K. Nature 2000, 407, 249.
2. Tonini, T.; Rossi, F.; Claudio, P. P. Oncogene 2003, 22,
6549.
3. Risau, W. Nature 1997, 386, 671.
4. Folkman, J. Nat. Med. 1995, 1, 27.
5. Ferrara, N. Semin. Oncol. 2002, 29, 10.
6. Jansen, M.; de Witt Hamer, P. C.; Witmer, A. N.; Troost,
D.; van Noorden, C. J. F. Brain Res. Rev. 2004, 45, 143.
7. Paz, K.; Zhu, Z. Front. Biosci. 2005, 10, 1415.
8. Hurwitz, H.; Fehrenbacher, L.; Cartwright, T.; Hains-
worth, J.; Heim, W.; Berlin, J.; Baron, J.; Griffing, S.;
Holmgren, E.; Ferrara, N.; Fyfe, G.; Rogers, B.; Ross, R.;
Kabbinavar, F. N. Eng. J. Med. 2004, 23, 2335.
9. Wedge, S. R.; Ogilvie, D. J.; Dukes, M.; Kendrew, J.;
Chester, R.; Jackson, J. A.; Boffey, S. J.; Valentine, P. J.;
Curwen, J. O.; Musgrove, H. L.; Graham, G. A.; Hughes,
G. D.; Thomas, A. P.; Stokes, E. S. E.; Curry, B.;
Richmond, G. H. P.; Wadsworth, P. F.; Bigley, A. L.;
Hennequin, L. F. Cancer Res. 2002, 62, 4645.
10. Thomas, A. L.; Morgan, B.; Drevs, J.; Unger, C.;
Wiedemann, B.; Vanhoefer, U.; Laurent, D.; Dugan, M.;
Steward, W. P. Semin. Oncol. 2003, 30, 32.
11. Ho, C. H.; Ludovici, D. W.; Maharoof, U. S. M.; Mei, J.;
Sechler, J. L.; Tuman, R. W.; Strobel, E. D.; Andraka, L.;
Yen, H. K.; Leo, G.; Li, J.; Almond, H.; Lu, H.; DeVine,
A.; Tominovich, R. M.; Baker, J.; Emanuel, S.; Grunin-
ger, R. H.; Middleton, S. A.; Johnson, D. L.; Galemmo,
R. A. J. Med. Chem. 2005, 48, 8163.
12. Furet, P.; Meyer, T.; Strauss, A.; Raccuglia, S.; Rondeau,
J. M. Bioorg. Med. Chem. Lett. 2002, 12, 221.
13. A mixture of the indanone 9, phenyl benzoate (1 equiv),
and sodium hydride (2.5 equiv) in benzene was heated to
reflux for 2.5 h to give a dicarbonyl intermediate. The

J. Dinges et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4371–4375
4375
reaction was quenched by addition of 50% aq HOAc and
the reaction mixture was evaporated in high vacuum. For
the ring closure, the crude residue was taken up in ethanol,
hydrazine (1.5 equiv) and HOAc (3 equiv) were added and
the mixture was refluxed for 4 h.
15. The conditions for all biological assays are described in:
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.
D.; Stoll, V. S.; Tapang, P.; Wishart, N.; Davidsen, S. K.;
Michaelides, M. R. J. Med. Chem. 2005, 48, 6066.
14. Sairam, P.; Puranik, R.; Kelkar, A. S.; Sasikiran, S.;
Veerender, M.; Parvathi, A. Synth. Commun. 2003, 33,
3597.