Identification of novel series of pyrazole and indole-urea based DFG-out PYK2 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2012.10.039

Previous drug discovery efforts identified classical PYK2 kinase inhibitors such as 2 and 3 that possess selectivity for PYK2 over its intra-family isoform FAK. Efforts to identify more kinome-selective chemical matter that stabilize a DFG-out conformation of the enzyme are described herein. Two sub-series of PYK2 inhibitors, an indole carboxamide-urea and a pyrazole-urea have been identified and found to have different binding interactions with the hinge region of PYK2. These leads proved to be more selective than the original classical inhibitors.

Full text of DOI:10.1016/j.bmcl.2012.10.039

Bioorganic & Medicinal Chemistry Letters 22 (2012) 7523–7529
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Identification of novel series of pyrazole and indole-urea based DFG-out PYK2
inhibitors
Samit K. Bhattacharya a, , Gary E. Aspnes a, Scott W. Bagley a, Markus Boehm a, Arthur D. Brosius a,
⇑
Leonard Buckbinder c, Jeanne S. Chang d, Joseph Dibrino a, Heather Eng b, Kosea S. Frederick b,
David A. Griffith a, Matthew C. Griffor d, Cristiano R. W. Guimarães a, Angel Guzman-Perez a, Seungil Han d,
Amit S. Kalgutkar b, Jacquelyn Klug-McLeod a, Carmen Garcia-Irizarry a, Jianke Li a, Blaise Lippa a,
a
a
a
a
a
David A. Price a, , James A. Southers , Daniel P. Walker , Liuqing Wei , Jun Xiao , Michael P. Zawistoski ,
⇑
Xumiao Zhao a
a Worldwide Medicinal Chemistry, Pfizer Global Research and Development, 620 Memorial Drive, Cambridge, MA 02139, United States
b Pharmacokinetics, Dynamics and Metabolism, Pfizer Global Research and Development, 620 Memorial Drive, Cambridge, MA 02139, United States
c Cardiovascular and Metabolic Diseases Research Unit, Pfizer Global Research and Development, 620 Memorial Drive, Cambridge, MA 02139, United States
d Structural Biology & Biophysics, Pfizer Global Research and Development, Eastern Point Road, Groton, CT 06340 United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Previous drug discovery efforts identified classical PYK2 kinase inhibitors such as 2 and 3 that possess
selectivity for PYK2 over its intra-family isoform FAK. Efforts to identify more kinome-selective chemical
matter that stabilize a DFG-out conformation of the enzyme are described herein. Two sub-series of PYK2
inhibitors, an indole carboxamide–urea and a pyrazole–urea have been identified and found to have dif-
ferent binding interactions with the hinge region of PYK2. These leads proved to be more selective than
the original classical inhibitors.
Received 6 August 2012
Revised 2 October 2012
Accepted 8 October 2012
Available online 16 October 2012
Keywords:
PYK2
Ó 2012 Elsevier Ltd. All rights reserved.
FAK
DFG-out
Osteoporosis
Kinase inhibitor
BIRB-796
There is an unmet need in the osteoporosis field for new treat-
ment options beyond the current mainstay of anti-resorptive ther-
apeutics such as bisphosphonates, an anti-RANKL antibody, and
estrogen receptor modulators (SERMs).1 While these treatments
have vastly improved the quality of life of patients, they have sig-
nificant drawbacks. Bisphosphonates have safety concerns includ-
ing increased risk of cancer, atrial fibrillation, and musculoskeletal
pain. More recently, osteonecrosis of the jaw has emerged as a rel-
atively rare but devastating adverse effect.2 In addition, there is a
dosing inconvenience whereby after oral administration the pa-
tient is required to be seated in an upright posture for up to an
hour post dosing. Furthermore, from an efficacy standpoint, both
the bisphosphonates and SERMs are antiresorptive and not bone
anabolic. Therefore they do not restore bone in chronically osteo-
porotic patients.
The only currently approved anabolic agent, recombinant hu-
man parathyroid hormone,3 is limited to use in the most severely
afflicted patients because of concerns for osteosarcoma and its
requirement for daily injections.
Within the osteoporosis literature there have been recent pub-
lications disclosing the potential of proline-rich kinase 2 (PYK2) as
a molecular target for improving bone formation. PYK2 and focal
adhesion kinase (FAK) are non receptor tyrosine kinases and to-
gether constitute the FAK subfamily.4 FAK is ubiquitously ex-
pressed; however, PYK2 expression is highest in the brain and
hematopoietic system.
Initial studies using PYK2 knock-out female mice exhibited the
skeletal phenotype of higher bone mass and mineral density when
compared to wild type animals.5 The mice were viable and fertile.
Spurred on by these initial findings, small molecule modulators
were designed and tested for efficacy. They increased bone forma-
tion and protected against bone loss in ovariectomized rats, an
established model of postmenopausal osteoporosis. These studies
suggested that selective modulation of this kinase may offer an
⇑
Corresponding authors. Tel.: +1 617 551 3177.
E-mail address: samit.k.bhattacharya@pfizer.com (S.K. Bhattacharya).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.10.039

7524
S. K. Bhattacharya et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7523–7529
O
O
N
CF3
CF3
SO2Me
H
N
N
N
N
N
HN
R
N
N
H
HN
N
H
N
HN
O
N
N
O
O
NH
O
N
O
O
O
1, PF-431396
2 R =
S
4
3 R =
S
PYK2 kinase IC50 31 nM
FAK kinase IC50 1 nM
95 nM
1500 nM
43 nM
1400 nM
35 nM
17 µM
Figure 1. Representative PYK2 kinase inhibitors in literature.
[determined using scintillation proximity resonance (SPR)], NMR
experiments and ultimately an X-ray crystal structure demon-
strated that BIRB-796 was a specific ligand for PYK2 and that it sta-
bilized a ‘DFG-out’ conformation, despite its low affinity. Notably,
the crystal structure revealed a PYK2 specific binding mode with
significant differences from the published p38-BIRB796 (yellow,
Fig. 2, PDB code: 1KV2) interaction. The naphthyl ring is rotated
out of plane relative to p38-BIRB796, and the morpholino substitu-
ent in PYK2-BIRB796 (blue, Fig. 2, PDB code: 3FZS) complex does
not interact with the hinge residues. Orientation of the tolyl-tert-
butylpyrazole is well conserved.
Despite exquisite potency against p38, we chose to use BIRB-
796 as an initial lead for medicinal chemistry efforts to develop a
potent and selective DFG-out inhibitor of PYK2. It rapidly became
evident that the tolyl-pyrazole fragment provided significant bind-
ing energy (as published for p38 inhibition).11 Conversely, modifi-
cation of the naphthyl group was tolerated and the PYK2 specific
binding mode suggested selectivity and potency might be found
through replacement of that moiety. Several compound libraries
were targeted that aimed at replacement of the naphthyl group
while holding the tolyl-pyrazole and urea constant. Two urea sub-
series that emerged from this work were the indole carboxamides
(6) and pyrazoles (7) which provided improvements in both po-
tency and ligand efficiency (Fig. 3). Subsequent efforts were made
Figure 2. Binding modes of BIRB-796 bound to PYK2 (blue) and p38 (yellow).
opportunity to deliver an orally available anabolic therapy for the
treatment of osteoporosis.
These aforementioned small molecule modulators were classi-
cal kinase inhibitors that make a hinge contact with the enzyme.
The originally disclosed lead classical inhibitor, 1 (PF-431396)
was improved upon to arrive at analogs 2 and 3 in the same class
that had better intra-family isoform selectivity versus FAK.6
(Fig. 1).
Broad selectivity across a kinase panel is often difficult to
achieve with classical kinase inhibitors while inhibitors that stabi-
lize the ‘inactive’ DFG-out conformation (these compounds move
the DFG motif of the enzyme upon their binding and do not neces-
sarily require the hinge contact)7 are relatively more selective.8
Thus we invested efforts at uncovering chemistry substrates that
would be broadly defined as stabilizers of the DFG-out conforma-
tion. The current article describes these efforts that led to two
sub-series of DFG-out leads for PYK2 and their selectivity profiles
in a typical kinase selectivity panel. A recent publication has de-
scribed chemical matter (4) that is DFG-in and is remarkably FAK
selective.9
H
N
H
N
N
O
MW/clogD/TPSA
527/6.0/81
N
N
O
O
PYK2 kinase IC50 11.5 µM
LE 0.18
5, BIRB-796
Cl
O
HN
NH
H
N
H
N
H
N
H
N
N
N
N
HN
N
N
O
O
6
7
O
To look for leads that could help PYK2 adopt a DFG-out confor-
mation we conducted targeted screening of known DFG-out inhib-
itors in the literature including imatinib, BIRB-796 (5), and
sorafenib against PYK2 enzyme.5 As previously reported, 5 was
identified as an active, albeit weak, PYK2 inhibitor (imatinib was
inactive in the PYK2 assay).10 Slow-on/slow-off binding kinetics,
MW/clogD/TPSA
444/6.1/.97
MW/clogD/TPSA
540/6.7/104
PYK2 IC50 225 nM
LE 0.27
PYK2 IC50 173 nM
LE 0.24
Figure 3. Two urea sub-series.

S. K. Bhattacharya et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7523–7529
7525
R1
O
N
NH2
R2
H
N
NH
a, b, c
Cl
Cl
O
O
H
N
H
N
H
N
O
N
N
Cl
N
N
O
O
8
9
10
Scheme 1. (a) DMSO, DIPEA, 50 °C, 12 h, 81%; (b) 1 M NaOH, MeOH, 50 °C, ꢀ2 h, 86%; (c) R1R2NH, HATU, DIPEA, NMP, 50 °C, 4 h, ꢀ98%. HATU = O-(7-azabenzotriazol-l-yl)-
N,N,N0,N0-tetramethyluronium hexaflourophosphate.
Table 1
Biological data for indole-carboxamide urea series
Entry R1
R2
MW cLogD
PYK2 kinase
LEa3
PYK2 Cell IC50 HLM Clearance Dofetilide % inhibition FAK Kinase
(eLogD⁄) IC50 (nM)a
(nM)b
ml/min/kg
@ 10
lM
IC50 (nM)
31
3.8
N
10a
484
444
0.28 265
77
28
15
(<30 n = 2)a2
10b
10c
Me
H
H
5.2
1700
0.23 2780
(4.9)
6.0
(3310, 870)
78
iPr
472
0.27 530
93
3
7860
(5.3)
(73, 83)
(438, 642)
57 19
(n = 3)
N
10d
539
0.35
0.24 7660
53
91
HN
O
398 219
(n = 3)
10e
10f
Me
549
516
4
0.22 553
0.26 283
53
31
4
>10000
>10000
N
75 40
(n = 5)
H
5.9
122
⁄eLogD = measured LogD (at pH 7.4)
a
PYK2 fluorescence polarization (FP) protocol as described in Ref. 5 IC50 values represent the concentration to inhibit 50% of the phosphorylation of a peptide substrate
relative to vehicle control. Numbers indicate the IC50 values generated from individual 8-point concentration response relationships in triplicate. Where n > 3, the geometric
mean and standard deviation is reported.
a2
Data generated in PYK2 FP 8-point assay (described above) with a higher top dose (100
LE = ligand efficiency.
lM instead of 10 lM).
a3
b
PYK2 LI-COR cellular assay (protocol has been published).13 Numbers indicate IC50 values generated from 9-point concentration response relationships in quadruplicate.
Where n > 3, the geometric mean and standard deviation is reported.
c
Protocols for measuring half-lives in HLM and subsequent scaling to blood clearance have been published.14
d [3H]-Dofetilide binding assay as described.15 Numbers indicate the percentage inhibition of [3H]-dofetilide binding to the hERG protein stably expressed on HEK- 293 cells
following a 10 lM dose of test compound.
e FAK fluorescence polarization (FP) protocol as described in Ref.5 using FAK kinase domain construct as described.16 IC50 values represent the concentration to inhibit 50%
of the phosphorylation of a peptide substrate relative to vehicle control. Numbers indicate IC50 values generated from individual 8-point concentration response relationships
in triplicate.
to improve the potency and physicochemical properties of each
urea subseries in parallel.
Some additional efforts were directed towards replacement of
the indole motif. However, even conservative changes (e.g., imid-
azole/benzoxazole) greatly reduced activity (data not shown). Sub-
sequent solution of an X-ray co-crystal structure with 10c helped
to provide a rationale for the importance of the indole carboxamide
(see below).
Clearly the high molecular weight and lipophilicity of the indole
carboxamide sub-series were liabilities and we sought to mitigate
these through further optimization.12 Amide libraries were de-
signed in an effort to improve potency while reducing molecular
weight and/or lipophilicity (some examples shown in Table 1).
Template and library preparation is shown in Scheme 1. 7-Amino-
indole-2-carboxylic methyl ester (9) was coupled with troc-amino-
pyrazole (8) and the ester saponified to give an acid template
which was coupled to various amines with HATU.
Several compounds with improved properties were found. Ef-
forts to reduce molecular weight and improve ligand efficiency
led to entries 10a, b and c. While the simple methyl amide, 10b,
eroded PYK2 potency substantially, the corresponding isopropyl
indole amide 10c provided improved potency with reduced lipo-
philicity relative to the lead 6. Other efforts to reduce the lipophil-
icity and/or increase solubility of these indoles led to entries 10d,
10e and 10f. For the spiro-bis-pyrollidine 10d, while kinase po-
tency was good, it showed a marked loss in cellular potency most
probably because the basic amine affected cellular permeability.
The other two entries, 10e and 10f had varying potency in the ki-
nase assay, but similar cell potency to 10c which suggested that
these might be viable in vivo tools.
While optimization was ongoing in the indole subseries, we also
looked at the bis-pyrazole urea series. In the indole series, we had
found the tert-butyl-pyrazole-tolyl group provided optimal po-
tency and ligand efficiency (data not shown). Thus for the pyrazole
series, we targeted exploration off the NH-pyrazole vector to max-
imize lipophilic efficiency.17,18 We utilized library and singleton
chemistry to explore the 3-position of the NH-pyrazole (13,
Scheme 2). The desired ureas were synthesized by reacting the
common tolyl-pyrazole carbamate intermediate 11 with various
aminopyrazoles under thermal conditions, as noted in Scheme 2.
As shown in Table 2, attaining good PYK2 activity in the bis-pyr-
azole urea series was challenged with high molecular weight and
lipophilic chemical space. A pendant aryl group with a hydrogen
bond acceptor appeared critical for PYK2 potency, for reasons
which would become evident from X-ray data (vide infra). Subse-
quent modifications were directed towards decreasing lipophilicity
while maintaining an appropriate H-bond acceptor. The phenyl
group in 7 (or 13a) was replaced with pyridines and diazines in or-

7526
S. K. Bhattacharya et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7523–7529
der to lower the LogD. Addition of nitrogen ortho to the pyrazole
was deleterious for PYK2 kinase potency (13b and 13d), while
the meta-pyridyl isomer, 13c, was nearly equipotent. The corre-
sponding pyridone, 13e, had better physicochemical properties
but while PYK2 kinase potency was improved, cell potency was
substantially reduced, presumably due to permeability issues. Sim-
ilar to analogs 13b and 13d, addition of an ortho-nitrogen to give
the pyridazinone analog 13f led to reduced PYK2 potency. To re-
move the additional H-bond, the N-methylated pyridone 13g was
made but showed further loss in PYK2 kinase and cell potency.
A parallel effort was made to improve potency by increasing the
H bond acceptor basicity and/or removing a rotatable bond by
Table 2
Biological data for bis-pyrazole urea series
Entry
R
MW
cLogD
PYK2 kinase
IC50 (nM)a
LEb
PYK2 Cell
IC50 (nM)c
HLM Clearance
ml/min/kgd
Dofetilide
FAK kinase
IC50 (nM)f
(eLogD⁄)
%Inhibition
@ 10 (
l
M)e
1200 55
(n = 3)
13a
459
6.5
5.2
0.24
0.24
0.27
129
7
4770
O
N
3819
(3360, 4340)
13b
13c
13d
445
445
446
607
>10000
O
O
O
5.4
(4.9)
361 30
(n = 3)
524 191
(n = 3)
48
15
N
N
>10000
(n = 2)
5.0
N
H
N
O
3.75
(3.7)
141 28
(n = 3)
4.7 2.1
(n = 3)
l
M
M
13e
13f
13g
431
432
445
0.29
0.26
0.26
11.9
16
36
24
30
HO
O
N
N
3.6 1.1
(n = 3)
l
0.75
1350a2
N
4.1
(3.8)
429
(315, 585)
2.5 lM
(1.8, 3.5)
20
O
O
328
(316, 342)
13h
13i
13j
472
454
465
5.7
69
0.28
0.30
0.30
92
44
55
1
5.5
(4.5)
25 41
(n = 4)
236 33
(n = 3)
N
55
31
N
6.3
(5.2)
6.4 3.7
(n = 4)
98 78
(n = 4)
608
N
N
5.4
(4.7)
12.5 2.3
(n = 4)
62
(41, 94)
13k
466
0.30
0.26
149
109
12
16
473
N
637 407
(n = 3)
190 112
(n = 4)
13l
445
368
4.9
3.2
>10000
N
O
13m
HO
>100 l
Ma3
⁄eLogD = measured LogD (at pH 7.4)
a
PYK2 fluorescence polarization (FP) protocol as described in Ref. 5. IC50 values represent the concentration to inhibit 50% of the phosphorylation of a peptide substrate
relative to vehicle control. Numbers indicate IC50 values generated from individual 8-point concentration response relationships in triplicate. Where n>3, the geometric mean
and standard deviation is reported.
a2
Data generated in PYK2 FP 8-point assay (described above) with a higher top dose (100 lM instead of 10 lM).
Single point data
LE = ligand efficiency.
a3
b
c
PYK2 LI-COR cellular assay (protocol as described in Ref. 13). Numbers indicate IC50 values generated from 9-point concentration response relationships in quadruplicate.
Where n>3, the geometric mean and standard deviation is reported.
d
Protocols for measuring half-lives in HLM and subsequent scaling to blood clearance have been published (see: Ref. 14).
e
[3H]-Dofetilide binding assay as described in Ref. 15. Numbers indicate the percentage inhibition of [3H]-dofetilide binding to the hERG protein stably expressed on HEK-
293 cells following a 10 lM dose of test compound.
f
FAK fluorescence polarization (FP) protocol as described in Ref. 5 using FAK kinase domain construct as described in Ref. 16. IC50 values represent the concentration to
inhibit 50% of the phosphorylation of a peptide substrate relative to vehicle control. Numbers indicate IC50 values generated from individual 8-point concentration response
relationships in triplicate

S. K. Bhattacharya et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7523–7529
7527
NH2
H
a
1
H
N
N
HN
N
N
N
H
N
O
HN
R
N
N
O
N
3
O
R
13
11
12
Scheme 2. (a) K2CO3, THF, 80 °C, 16 h, ꢀ80%.
replacing the anisole with heterobicyclic groups (13h–k). These
analogs provided a 3–10ꢁ boost in PYK2 potency although they
are at the edges of desired physicochemical properties. 13l was
made in an effort to improve physical properties by conceptually
breaking the proximal ring of quinoline analog 13j. Though LogD
was reduced, the additional flexibility in 13l resulted in decreased
PYK2 potency and increased clearance (compared to 13j). Further
truncation to the hydroxymethyl compound 13m, led to near com-
plete loss of PYK2 activity, highlighting the need for an appropriate
H bond acceptor in this series. While entry 13k had the best lipo-
philic efficiency (LipE = 2.3)17 in this sub-series, entry 13j19 pro-
vided a lead (LipE = 1.5) that had an acceptable overall profile in
terms of PYK2 activity and ADME properties (as exemplified by
moderate clearance in human liver microsomes, moderate hERG
inhibition (as evidenced by dofetilide binding assay) and lack of
competitive DDI versus CYP3A4, CYP2D6, CYP2C9 or CYP1A2.
Crystallographic analysis on both the bis-pyrazole and the in-
dole carboxamide urea leads revealed different binding modes
for these two sub-series. The crystal structure of pyrazole 13a re-
vealed that the methoxy oxygen has a hydrogen bond interaction
with the amide backbone of Tyr505 in the hinge region, providing
a rationale for the apparently critical H-bond acceptors in analogs
13a–m. The pyrazole moiety is sandwiched by Met502 (gatekeeper
residue) and Phe568 (DFG motif). One of the pyrazole nitrogens
has a hydrogen bond (H-bond) interaction with Lys457. In contrast,
there was no interaction with the hinge region present in the co-
crystal structure of indole 10c in PYK2. The indole nitrogen is in-
volved in a water mediated H-bond interaction with Lys457 and
Glu474. The isopropyl amide moiety at C2 position of the indole
is involved in van der Waals contacts with Gly-rich loop.
agement for a path toward kinome selectivity without requiring
hinge interactions. This was realized, as in contrast to the classical
kinase inhibitor, 1, the indole carboxamide leads (entries 10c, 10e
and 10f, Table 1) were very selective versus FAK kinase. The bis-pyr-
azole urea leads were also selective against FAK but not as selective
as the indole leads. This trend of kinase selectivity was evidenced
when tested in a kinase selectivity panel at 1 lM. As expected the
indole-carboxamide urea series which does not make a hinge-con-
tact has a better selectivity profile over the bis-pyrazole urea series.
In general, both DFG-out lead series possessed improved kinome
selectivity in comparison to the classical kinase inhibitor 1 as
shown in the heat map below (Fig. 5).21 Although the starting lead
(BIRB-796) for this current work was a potent p38a inhibitor, both
the pyrazole and the indole series show selectivity for p38
a
(MAPK14). Similar to the trend of FAK selectivity, the indole carbox-
amides exhibit better selectivity than the pyrazoles.
The X-ray crystal data for 13a suggested one additional potential
avenue to optimize potency and permeability in the bis-pyrazole
series. The observed H-bond between the pyrazole-N and Lys457
highlighted potential tautomerism in the pyrazole-NH motif and
the possibility to eliminate a hydrogen-bond donor in order to im-
prove permeability. Computational investigation of the tautomeric
equilibrium in water revealed that the tautomer in the bound state
(wherethe NH of the pyrazole is proximalto the NH of the urea), sug-
gested by the hydrogen bond network in the crystal structure, is not
the most stable one in solution. An energy penalty of +1.85 kcal/mol
against the binding of 7 due to the tautomeric equilibrium was esti-
mated by a combination of conformational analysis in solution and
quantum mechanical calculations. These calculations suggested
that an alternative heterocycle with no associated tautomerism
could potentially have a 20-fold gain in potency if all contributions
to binding were kept the same. To test this hypothesis and remove
an H-bond to improve permeability, we made the isomeric pyrazole
analogs, 13n and 13o. Indeed the pyrazole isomer, 13n, where the
distal N (from the urea) is methylated leaving the other N unencum-
bered to make the critical interaction with Lys457 was potent while
Thus both sub-series do induce a DFG-out conformation of PYK2.
However as shown in Fig. 4, the indole sub-series does not make
any contact with the hinge region while the pyrazole analog with
an appropriate substituent is able to reach to the hinge backbone.20
The fact that indole carboxamide 10c was able to give good po-
tency without interaction with the hinge backbone gave us encour-
Figure 4. Crystal Structures of pyrazole 13a and indole 10c.

7528
S. K. Bhattacharya et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7523–7529
Figure 5. Kinase selectivity screen for representative analogs at 1 lM dose.
Table 3
Rat Pharmacokinetic Data for Selected Analogsa
b
Entry Compound Series
Route Dose t1=2
Clp (ml/min/kg) Vdss (L/kg)
Cmax (nM)
Calc Cmax free (nM) %F
PPB (fu)
0.00066
1
2
3
4
10e
10f
13j
13k
Indole urea
Indole urea
IV
1
2.2 0.9
4.1 2.6
3.2 1.3
4.8 0.9
0.4 0.06
PO
IV
30
1
38265 10396
621, 784
25
42
69 12
0.7 0.2
0.06
PO
30
1
1.6 (1.5, 1.7)
Pyrazole urea IV
2.0 (2.6, 1.5) 2.0 (1.8, 2.1)
0.15 (0.18, 0.12)
0.003
0.016
PO
Pyrazole urea IV
PO
30
n/a
61324 (60142, 62505) 184
68 (84, 52) 1.1
47
100 7.3 (3.7, 10.9)
n/a
a
Pharmacokinetic studies were conducted in male Wistar-Han rats (n = 2ꢂ3). In cases where two animals were used, individual pharmacokinetic parameters are also
depicted in parenthesis. Compounds were administered in 10% DMSO: 40% water :50% PEG400 for oral (PO) and intravenous (IV) studies.
b
Half-life t1=2 values represent the terminal elimination (b) phase.
the corresponding isomer 13o where the N closer to the urea is
methylated is devoid of activity pointing to the significance of pyra-
zole N- Lys457 interaction.
However it was intriguing that there was not much difference in
activity between the NH-pyrazole (7) and its N-methylated counter-
part (13n). Application of the physics-based scoring method MM-
GB/SA22 indicated that the active tautomer of 7 binds to PYK2 more
favorably than its N-methylated analog 13n by ꢂ2.2 kcal/mol. This
difference in binding energy is due to a water-mediated hydrogen
bond between the pyrazole-NH in 7 and the backbone carbonyl of
Phe568 (this water was explicitly included in the MM-GB/SA calcu-
lations). We hypothesize that the energy gain for 13n due to elimi-
nation of the tautomeric penalty is almost completely offset by its
less favorable interaction in the binding site compared to 7 and
hence the equivalent potencies of 7 and 13n. Unfortunately, further
attempts to replace the pyrazole with other 5-membered heterocy-
cles (e.g., oxazole, isoxazole, thiazole, oxadiazole etc.) where there
would be no tautomeric penalty did not yield better leads (data
not shown).
H
N
H
N
H
N
H
N
N
N
N
N
N
N
N
N
O
O
13o
PYK2 kinase IC >10 uM
13n
O
O
PYK2 kinase IC50 360 nM
LE 0.26
50
PYK2 cell IC50 381 nM

S. K. Bhattacharya et al. / Bioorg. Med. Chem. Lett. 22 (2012) 7523–7529
7529
Kath, J. C.; Roberts, W. G.; Smock, S. L.; Guzman-Perez, A.; Brown, T. A.; Li, M.
Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 10619.
We have previously reported on a functional osteogenesis assay
in hMSC cultures.4,8 The classical inhibitor 1 led to modest in-
creases in alkaline phosphatase activity but only when dosing
was restricted to days 4–7. Compound-treated cells appeared to
be stressed and the cultures did not mineralize, a finding that
was attributed to the off-target activity of 1. Incontrast, 13j and
other DFG-out analogs such as 13l showed a dose-dependent in-
crease in both alkaline phosphatase activity and mineralization
for a period of 14 days (as reported earlier),10,13 consistent with
their improved selectivity.
6. (a) Walker, D. P.; Zawistoski, M. P.; McGlynn, M. A.; Li, J. C.; Kung, D. W.;
Bonnette, P. C.; Baumann, A.; Buckbinder, L.; Houser, J. A.; Boer, J.; Mistry, A.;
Han, S.; Xing, L.; Guzman-Perez, A. Bioorg. Med. Chem. Lett. 2009, 19, 3253; (b)
Walker, D. P.; Bi, F. C.; Kalgutkar, A. S.; Bauman, J. N.; Zhao, S. X.; Soglia, J. R.;
Aspnes, G. E.; Kung, D. W.; Klug-McLeod, J.; Zawistoski, M. P.; McGlynn, M. A.;
Oliver, R.; Dunn, M.; Li, J. C.; Richter, D. T.; Cooper, B. A.; Kath, J. C.; Hulford, C.
A.; Autry, C. L.; Luzzio, M. J.; Ung, E. J.; Roberts, W. G.; Bonnette, P. C.;
Buckbinder, L.; Mistry, A.; Griffor, M. C.; Han, S.; Guzman-Perez, A. Bioorg. Med.
Chem. Lett. 2008, 18, 6071.
7. For a review on the field of DFG-out inhibitors, see: Liu, Y.; Gray, N. S. Nat.
Chem. Biol. 2006, 358.
8. Davis, M. I.; Hunt, J. P.; Herrgard, S.; Ciceri, P.; Wodlicka, L. M.; Pallares, G.;
Hocker, M.; Treiber, D. K.; Zarrinkar, P. P. Nat. Biotechnol. 2011, 29, 1046.
9. Allen, J. G.; Lee, M. R.; Han, C. Y.; Scherrer, J.; Flynn, S.; Boucher, C.; Zhao, H.;
O’Connor, A. B.; Roveto, P.; Bauer, D.; Graceffa, R.; Richards, W. G.; Babij, P.
Bioorg. Med. Chem. Lett. 2009, 19, 4924.
10. Han, S.; Mistry, A.; Chang, J. S.; Cunningham, D.; Griffor, M.; Bonnette, P. C.;
Wang, H.; Chrunyk, B. A.; Aspnes, G. E.; Walker, D. P.; Brosius, A. D.;
Buckbinder, L. J. Biol. Chem. 2009, 284, 13193.
11. Regan, J.; Pargellis, C. A.; Cirillo, P. F.; Gilmore, T.; Hickey, E. R.; Peet, G. W.;
Proto, A.; Swinamer, A.; Moss, N. Bioorg. Med. Chem. Lett. 2009, 19, 4924.
12. For references on optimal drug-like properties see: (a) Bickerton, G. R.; Paolini,
G. V.; Besnard, J.; Mureasan, S.; Hopkins, A. L. Nat. Chem. 2012, 4, 90; (b)
Gleeson, M. P.; Hersey, A.; Montanari, D.; Overington, J. Nat. Rev. Drug Disc.
2011, 10, 197; (c) Hughes, J. D.; Blagg, J.; Price, D. A.; Bailey, S.; DeCrescenzo, G.
A.; Devraj, R. V.; Elsworth, E.; Fobian, Y. M.; Gibbs, M. E.; Gilles, R. W.; Greene,
N.; Huang, E.; Krieger-Burke, T.; Loesel, J.; Wager, T.; Whiteley, L.; Zhang, Y.
Bioorg. Med. Chem. Lett. 2008, 4872; (d) Leeson, P. D.; Springthorpe, B. Nat. Rev.
Drug Disc. 2007, 881; (e) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P.
J. Adv. Drug Deliv. Rev. 1997, 23, 3.
13. Bonnette, P. C.; Robinson, B. S.; Silva, J. C.; Stokes, M. P.; Brosius, A. D.;
Baumann, A.; Buckbinder, L. J. Proteomics 2010, 73, 1306.
14. Obach, R. S. Drug Metab. Dispos. 1999, 27, 1350.
15. Greengrass, P. M.; Stewart, M.; Wood, C. M. PCT WO 2003021271, March 13,
2003.
16. Slack-Davis, J. K.; Martin, K. H.; Tighman, R. W.; Iwanicki, M.; Ung, E. J.; Autry,
C.; Luzzio, M. J.; Cooper, B.; Kath, J. C.; Roberts, W. G.; Parson, J. T. J. Biol. Chem.
2007, 282, 14845.
In order to evaluate the in vivo pharmacokinetic properties, a
few of the leads from both sub-series were advanced in rat oral
bioavailability studies as tabulated in Table 3.
Representatives from both urea subseries were orally available
with low plasma clearance (Clp) and low steady state distribution
volumes (Vdss). Despite excellent total plasma exposure (Cmax),
the high protein binding levels (displayed as fraction unbound in
rat, rPPB) as a consequence of their higher lipophilicity meant that
the free concentrations (calculated Cmax free) were quite low. High
doses of these compounds would be required for meaningful inhi-
bition of PYK2, and thus these leads were not advanced further.
In summary, we have described our efforts to convert the arche-
typal DFG-out p38 inhibitor BIRB-796 into two DFG-out lead series
(with and without hinge contacts) that are potent and selective for
PYK2. Compounds from these series showed functional osteogenic
activity and oral exposure in rats. The bis-pyrazole and indole car-
boxamide urea series leads displayed surprisingly different modes
in which they bind to PYK2 in a DFG-out conformation leading to
one sub-series that is more selective than the other.
References and notes
17. Edwards, M. P.; Price, D. A. Annu. Rep. Med. Chem. 2010, 45, 380. LipE values
reported here were calculated using cLogD (at pH 7.4).
18. Also referred to as LLE: Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug Disc. 2007,
881.
1. (a) Allen, J. G.; Fotsch, C.; Babij, P. J. Med. Chem. 2010, 4332; (b) Lewiecki, E. M.
Cleve. Clin. J. Med. 2009, 76, 457.
2. (a) Kanis, J. A.; Reginster, J.-Y.; Kaufman, J.-M.; Ringe, J.-D.; Adachi, J. D.;
Hiligsmann, M.; Rizzoli, R.; Cooper, C. Osteoporos. Int. 2012, 23, 213; (b) Khosla,
S.; Bilezikina, J. P.; Dempster, D. W.; Lewiecki, M.; Miller, P. D.; Neer, R. M.;
Recker, R. R.; Shane, E.; Shoback, D.; Potts, J. T. J. Clin. Endocrinol Metab. 2012,
2272.
3. Canalis, E.; Giustina, A.; Bilezikian, J. P. N. Engl. J. Med. 2007, 357, 905.
4. Lipinski, C. A.; Loftus, J. C. Expert Opin. Ther. Targets 2010, 14, 95.
5. Buckbinder, L.; Crawford, D. T.; Qi, H.; Ke, H.-Z.; Olson, L. M.; Long, K. R.;
Bonnette, P. C.; Baumann, A. P.; Hambor, J. E.; Grasser, W. A.; Pan, L. C.; Owen, T.
A.; Luzzio, M. J.; Hulford, C. A.; Gebhard, D. F.; Paralkar, V. M.; Simmons, H. A.;
19. Synthesis of 13j (PF-4594755) has been published. See Ref. 13.
20. The coordinates for crystal structures of PYK2 with pyrazole 13a and indole
10c have been deposited to PDB with accession codes 4H1 J and 4H1 M
respectively.
21. Kinase selectivity determined by a panel of biochemical kinase assays run in a
384-well streptavidin capture (flashplate) format in presence of a specific
biotinylated peptide and a mixture of both unlabeled and 33P-labeled ATP
(assay run at Km of ATP for each kinase).
22. Guimarães, C. R.; Cardozo, M. J. Chem. Inf. Model. 2008, 48, 958.