Identification of Selective Dual ROCK1 and ROCK2 Inhibitors Using Structure-Based Drug Design

Article abstract of DOI:10.1021/acs.jmedchem.8b01098

A HTS campaign identified compound 1, an excellent hit-like molecule to initiate medicinal chemistry efforts to optimize a dual ROCK1 and ROCK2 inhibitor. Substitution (2-Cl, 2-NH₂, 2-F, 3-F) of the pyridine hinge binding motif or replacement with pyrimidine afforded compounds with a clean CYP inhibition profile. Cocrystal structures of an early lead compound were obtained in PKA, ROCK1, and ROCK2. This provided critical structural information for medicinal chemistry to drive compound design. The structural data indicated the preferred configuration at the central benzylic carbon would be (R), and application of this information to compound design resulted in compound 16. This compound was shown to be a potent and selective dual ROCK inhibitor in both enzyme and cell assays and efficacious in the retinal nerve fiber layer model after oral dosing. This tool compound has been made available through the AbbVie Compound Toolbox. Finally, the cocrystal structures also identified that aspartic acid residues 176 and 218 in ROCK2, which are glutamic acids in PKA, could be targeted as residues to drive both potency and kinome selectivity. Introduction of a piperidin-3-ylmethanamine group to the compound series resulted in compound 58, a potent and selective dual ROCK inhibitor with excellent predicted drug-like properties.

Full text of DOI:10.1021/acs.jmedchem.8b01098

Subscriber access provided by University of Virginia Libraries & VIVA (Virtual Library of Virginia)
Article
Identification of Selective Dual ROCK1 and ROCK2
Inhibitors Using Structure Based Drug Design
Adrian D Hobson, Russell A. Judge, Ana L Aguirre, Brian S Brown, Yifang Cui, Ping Ding, Eric Dominguez,
Enrico DiGiammarino, David A Egan, Gail M. Freiberg, Sujatha M Gopalakrishnan, Christopher M Harris,
Marie P Honore, Karen L Kage, Nicolas J Kapecki, Christopher Ling, Junli Ma, Helmut Mack, Mulugeta
Mamo, Stefan Maurus, Bradford McRae, Nigel S Moore, Bernhard K Mueller, Reinhold Mueller, Marian
T Namovic, Kaushal Patel, Steve D Pratt, C Brent Putman, Kara L Queeney, Kathy K Sarris, Lisa M
Schaffter, Vincent S. Stoll, Anil Vasudevan, Lei Wang, Lu Wang, William Wirthl, and Kimberly Yach
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b01098 • Publication Date (Web): 01 Nov 2018
Downloaded from http://pubs.acs.org on November 2, 2018
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 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 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.
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 89
Journal of Medicinal Chemistry
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 89
1
2
3
4
5
6
7
8
9
Identification of Selective Dual ROCK1 and ROCK2
Inhibitors Using Structure Based Drug Design
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
Adrian D. Hobson,*a Russell A. Judge,b Ana L. Aguirre,b Brian S. Brown,b Yifang Cui,c Ping Ding,a,d
Eric Dominguez,a Enrico DiGiammarino,b David A. Egan,b Gail M. Freiberg,b Sujatha M.
Gopalakrishnan,b Christopher M. Harris,a Marie P. Honore,b Karen L. Kage,b,e Nicolas J. Kapecki,b
Christopher Ling,b Junli Ma,b Helmut Mack,c Mulugeta Mamo,b,f Stefan Maurus,c Bradford McRae,a
Nigel S. Moore,a Bernhard K. Mueller,c,g Reinhold Mueller,c Marian T. Namovic,b Kaushal Patel,b Steve
D. Pratt,b C. Brent Putman,b Kara L. Queeney,a Kathy K. Sarris,b Lisa M. Schaffter,a Vincent Stoll,b Anil
Vasudevan,b Lei Wang,a Lu Wang,a William Wirthla and Kimberly Yacha.
a AbbVie Bioresearch Center, 381 Plantation Street, Worcester, MA 01605, United States
b AbbVie, Inc., 1 North Waukegan Road, North Chicago, IL 60064, United States
c AbbVie Deutschland GmbH & Co KG, Knollstrasse 50, 67061, Ludwigshafen, Germany
d Current address: DowDupont, 455 Forest Street, Marlborough, MA 01752, United States
e Current address: Altor Bioscience, 2810 North Commerce Parkway, Miramar, FL 33025, United States
f Current address: Novartis Institute for Biomedical Research, 4560 Horton St, Emeryville, CA 94608,
United States
g Current address: Swiss ReGen Therapeutics, 8404 Winterthur, Switzerland
KEYWORDS Rho kinase, ROCK1, ROCK2, PKA, Kinase Inhibitor, Structure-based drug design.
ACS Paragon Plus Environment

Page 3 of 89
Journal of Medicinal Chemistry
ABSTRACT: A HTS campaign identified compound 1, an excellent hit-like molecule to initiate
medicinal chemistry efforts to optimize a dual ROCK1 and ROCK2 inhibitor. Substitution (2-Cl, 2-
NH2, 2-F, 3-F) of the pyridine hinge binding motif or replacement with pyrimidine afforded compounds
with a clean CYP inhibition profile. Co-crystal structures of an early lead compound were obtained in
PKA, ROCK1 and ROCK2. This provided critical structural information for medicinal chemistry to
drive compound design. The structural data indicated the preferred configuration at the central benzylic
carbon would be (R) and application of this information to compound design resulted in compound 16.
This compound was shown to be a potent and selective dual ROCK inhibitor in both enzyme and cell
assays and efficacious in the retinal nerve fiber layer model after oral dosing. This tool compound has
been made available through the AbbVie Compound Toolbox. Finally, the co-crystal structures also
identified that aspartic acid residues 176 and 218 in ROCK2, which are glutamic acids in PKA, could be
targeted as residues to drive both potency and kinome selectivity. Introduction of a piperidin-3-
ylmethanamine group to the compound series resulted in compound 58, a potent and selective dual
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
ROCK
inhibitor
with
excellent
predicted
drug-like
properties.
INTRODUCTION
Rho-associated coiled-coil-containing protein kinase1 is a serine/threonine protein kinase that is more
commonly referred to simply as ROCK kinase. ROCK exists as two isoforms, ROCK12 (also known as
ROK3) and ROCK22 (also known as ROK4). In its activated form ROCK participates in numerous
signaling mechanisms including the phosphorylation of myosin light chain,5 the serine/threonine kinases
Lim kinase 1 and 2,6 adducin,7 and ezrin/radixin/moesin protein complex.8 This has resulted in ROCK
inhibition being pursued as a possible therapeutic for multiple diseases,9 most notably cerebral
vasospasm,10 glaucoma,11 neurodegenerative diseases12 and hypertension.13,14 This interest has resulted
in the discovery of multiple ATP-competitive ROCK inhibitors. For an excellent review of Rho
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 89
kinases and their therapeutic potential, see Feng,15 which highlights the physiological and biological
functions of rho kinases and the different chemical series that have been developed as ROCK inhibitors.
These inhibitors deploy various groups for the critical interaction with the hinge region that joins the N-
1
2
3
4
5
6
17, 18
and C-terminal portions of the conserved protein kinase domain, including indazole,16,
7
8
9
aminofurazan,19 pyridine,20, 21 pyrazole,22 and isoquinoline.23, 24, 25
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
Our interest was heightened by several publications demonstrating the effects of ROCK inhibitors in
areas of immunology including, effects on histamine-induced airflow obstruction,26 inhibition of
pulmonary eosinophilia, bronchoconstriction and airway hyperresponsiveness27 and anti-inflammatory
activities.28
Encouraged by this we conducted a high-throughput screening (HTS) campaign to identify novel ROCK
inhibitors. The HTS assay used ROCK2 enzyme (amino acids 11-552), with peptide substrate
phosphorylation monitored via 33P-labeled ATP. Approximately 700,000 compounds were screened,
and >3,100 of these produced inhibition of ROCK2 in concentration response testing. Herein we report
the discovery of an attractive hit series exemplified by compound 1 (Figure 1) and the subsequent
structure-activity relationship studies guided by critical structural information. This ultimately resulted
in the discovery of compound 58, a highly potent inhibitor of both ROCK1 and ROCK2 with broad
selectivity across the kinome.
RESULTS AND DISCUSSION
Mol wt
PSA
332
67
OH
Ph
O
1
N
H
cLogP
2.7
N
ROCK2 BEI29 20.2
Figure 1. Structure of HTS hit compound 1
ACS Paragon Plus Environment

Page 5 of 89
Journal of Medicinal Chemistry
Over 3,000 active hits were identified from a HTS campaign and prioritized using hit-like criteria29, 30, 31
and an affinity prediction plot.32 After structure and purity confirmation by liquid chromatography mass
spectrometry (LCMS) of the prioritized hits, compound 1 was selected for further investigation. In
addition to the hit-like properties exhibited by compound 1, it’s synthetic tractability was also very
appealing as it was afforded in just one step from 4-pyrid-4-ylbenzoic acid 2, which was offered in gram
quantity from Maybridge Chemical Co. Ltd33 and available through Ryan Scientific.34 This made it an
ideal candidate for follow up with focused libraries in the centralized high-throughput chemistry group
(HTC) at AbbVie. In addition there have been multiple reports on the successful development of
ROCK inhibitors with pyridine as the hinge binding motif.15, 21, 24, 35, 36, 37
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
To expand the SAR for 1, several libraries were produced by HTC. Using standard amide bond forming
conditions, commercially available 4-pyrid-4-ylbenzoic acid 2 was reacted with various benzylamines 3
to afford the desired compounds with the general formula 4 (Specific compounds 5 - 15 shown in Table
1)
Scheme 1. Synthetic scheme for HTC library follow up on compound 1
O
O
i.
OH
N
Ar
H
+
H2N
Ar
N
N
2
3
4
Reagents and conditions: i) PS-DCC, DIEA, DMA, microwave, 100 °C, 10 min.
The compounds were evaluated in enzyme activity assays for both ROCK1 and ROCK2, with PKA
used as a surrogate kinase to provide an early read on general kinome selectivity. Encouragingly, these
libraries afforded immediate SAR and identified many potent ROCK1 and ROCK2 inhibitors.
Compound 5 (Table 1) with the unsubstituted aryl ring resulted in a slight increase in the Binding
Efficiency Index (BEI) relative to the HTS hit while having no selectivity versus PKA. Introduction of
the 3-methoxy in compound 6 maintained the high BEI and increased the selectivity window over PKA
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 89
from 8 to almost 40-fold. Homologues 7 and 8 caused a slight drop in BEI while improving the
selectivity against PKA to 65- and 85-fold respectively. The 3-hydroxy analogue 9 produced the
highest BEI of nearly 28 while having a similar selectivity to the 3-methoxy compound. The larger 3-
OCF3 analogue 10 and the 2,3-dimethoxy compound 11 resulted in a loss of potency, with both
compounds being >1 M at all 3 kinases. Benzodioxolane compound 12, the ring constrained version
of 11, had a similar potency and selectivity profile to compound 6. Ring expansion to benzodioxepine
13 maintained the potency and BEI of the benzodioxolane at ROCK2 while affording a much larger
window of selectivity over PKA of 130. The final two analogues were designed to investigate the
tolerance for the introduction of solubilizing basic nitrogen groups to enable moderation of compound
properties, should this be necessary later in the project. Unfortunately while dimethylamine 14 and
morpholine 15 maintained the potency and selectivity profile of 6 this was with the penalty of a much
reduced BEI of 18.1 and 15.8 respectively.
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
Table 1. SAR for compounds 5 - 15
O
N
R
H
N
IC50 / M at 100 M ATP
ROCK2 PKA/ ROCK2
Cpd
R
BEI
Selectivity
ROCK1
ROCK2
PKA
0.55±0.0 0.12±0.00 0.94±
0.02
5
6
23.9
8
8
7
0.16±0.0 0.021±0.0 0.86±
07 0.43
O
24.0
39
7
ACS Paragon Plus Environment

Page 7 of 89
Journal of Medicinal Chemistry
O
1
2
3
4
5
6
7
8
7
8
0.78
0.52
0.019
4.4
0.11
0.118
0.003
1.3
7.4
10
20.9
20.0
27.9
15.8
65
85
36
5
O
OH
9
9
0.11
7.1
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
O
F
F
10
F
O
O
11
12
13
16.0±7.5
0.28
3.8±1.6
0.037
10
0.92
4.6
15.6
21.4
20.7
3
O
O
24
O
O
0.28
0.034
130
O
N
14
15
0.34
1.0
0.087
0.25
2.3
10
18.1
15.8
27
40
O
N
O
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In addition, these compound libraries enabled profiling in our HT-ADME panel of assays (Table 2).
This data showed that this series had good solubility, free fraction in plasma, and cellular permeability
as assessed using PAMPA.38 As predicted, compound 14 in which the dimethylamine was introduced,
displayed the highest solubility and free fraction. However, it was evident that the major liability was
their pan-inhibition of CYP enzymes, especially 3A4, predicting unwanted drug-drug interactions. It is
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 89
well known that 2,6-unsubstituted pyridines act as potent CYP inhibitors39, and so modification of this
motif would need to be addressed during the medicinal chemistry project.
1
2
3
4
5
6
Table 2. HT-ADME data for compounds 6, 7, 8, 12 and 14
7
8
9
CYP inhibition IC50 / M
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
Solubility /
PPB / %
free
PAMPA /
10-6 cms-1
Cpd
M
2C9
0.12
0.48
0.26
2.04
16.0
2D6
0.97
0.24
0.09
0.20
0.08
3A4
0.85
0.08
0.08
0.08
2.00
6
7
-
98.0
99.6
99.7
97.2
76.8
20.2
18.1
16.2
14.9
2.47
9.9
8
9.3
12
14
33.9
63.3
While the pragmatic amide bond-forming libraries afforded useful SAR, this was a one-dimensional
strategy, and a more rational design approach was desired as the project progressed. To this end several
crystallography campaigns were initiated using compounds from the HTC libraries. At the time, there
were no public crystal structures of ROCK2 and limited structures of ROCK1.40 In previous studies we
had successfully used PKA as a structure surrogate for ROCK kinases.41 Initial structures in this study
were therefore obtained in PKA. We and others (Akama42 and Boland43) later succeeded in obtaining
some structures in ROCK kinases. The structure of the early lead compound 12 was therefore obtained
in PKA, ROCK1 and ROCK2, affording a great opportunity to compare and contrast the 3 active sites.
ACS Paragon Plus Environment

Page 9 of 89
Journal of Medicinal Chemistry
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 2. Crystal structures of compound 12 in a) PKA (PDB 6E9L), b) Rock 1 (PDB 6E9W) and
c) Rock 2 (PDB 6ED6). All structures were obtained by co-crystallization.
In PKA (Figure 2a) the hinge interaction is with the backbone nitrogen of valine 124. The hydrophobic
core of the compound is sandwiched between alanine 71 and valine 58 above and leucine 174 below.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 89
The carbonyl of the amide makes a hydrogen bond with lysine 73, while the benzodioxine group forms
a large van der Waals surface interaction with the glycine rich loop which is situated above it. The
backbone nitrogen for phenylalanine 55 is within 3.5 Å of an oxygen of the dioxine which is accommo-
dated, but not likely to be a productive hydrogen bond. Phenylalanine 55 also stacks against the benzo-
dioxine group. For ROCK1 (Figure 2b) the binding is very similar. Here binding to the hinge is with
the backbone nitrogen of methionine 156. Again the residues of alanine 103 and valine 90 above and
leucine 205 below sandwich the compound core. The amide carbonyl forms a hydrogen bond to lysine
105 and the benzodioxine ring system is forming an extensive van der Waals interaction with the
glycine rich loop. Here, however, phenylalanine 87 is extended out from the glycine rich loop and does
not stack against the benzodioxine group. Differences in the region of PKA Lys79 to Thr89 (ROCK 1
resi-dues Glu111 to Phe121, ROCK 2 Glu 127 to Phe 137) likely allows for differential dynamics of the
gly-cine rich loop between PKA and ROCK1, in particular PKA Phe55 versus ROCK1 Phe87. The
glycine rich loop and PKA Phe55 are more constrained than the glycine rich loop of either ROCK1 or
ROCK2 and perhaps a driver of the observed selectivity of these compounds. For ROCK2 (Figure 2c)
the bind-ing mode is very similar to ROCK1. The hinge interaction is with the nitrogen backbone of
methionine 172. Again the amide carbonyl makes a hydrogen bond to lysine 121. Once again, the
benzodioxine ring system is forming an extensive van der Waals interaction with the glycine rich loop.
However, in the case of ROCK2, the glycine rich loop is even less constrained than that of ROCK1 with
additional flex-ibility observed in the neighbouring region of E127 to F137 which can further contribute
to potency dif-ferences observed between ROCK1 and ROCK2 allowing for more optimal van der
Waals contacts be-tween the glycine rich loop substituents (benzodioxine in this case) in ROCK2 than
ROCK1 or PKA.
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
Further in these structures there are acid residues nearby which could be reached by extension off the
compound core. In ROCK1 and ROCK2 these are both aspartic acids (160 and 202 in ROCK1 and 176
ACS Paragon Plus Environment

Page 11 of 89
Journal of Medicinal Chemistry
and 218 in ROCK2). In PKA these residues are both glutamic acids (128 and 171) and so targeting these
residues could provide some selectivity against PKA and AKT.
1
2
3
4
5
6
Table 3. SAR for compounds 16 and 17
7
8
9
O
O
N
H
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
N
IC50 / M at 100 M
CYP inhibition IC50 / M
PAMP
ATP
Solubili
ty / M % free
Cp Config
PPB /
A / 10-6
cms-1
d
uration
ROC
K1
ROCK2 PKA
2C9
0.64
-
2D6
3.62
-
3A4
0.08
-
0.017
±0.00
5
0.24
0.002±0
±0.1
.001
4
16
17
(R)
(S)
74.3
83.1
19.6
5.1±0.
1
2.4±0.9 >10
-
-
-
The structural data also suggested that the preferred configuration at the benzylic carbon would be (R).
To test this hypothesis compounds with both (R)-1-(3-methoxyphenyl)ethan-1-amine and (S)-1-(3-
methoxyphenyl)ethan-1-amine were synthesized. Gratifyingly, the two methyl isomers 16 and 17
confirmed the structural hypothesis of the preferred configuration (Table 3). While compound 16 with
the (R) configuration at the benzylic carbon was ten-fold more potent than the corresponding des-methyl
analogue 6, compound 17 with the (S)-configuration at the benzylic carbon was more than 3 logs less
potent.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 89
To address CYP inhibition, analogues incorporating substituted pyridine or heterocyclic replacements of
pyridine were designed. To secure these, a straightforward 2-step synthetic route was followed. The
desired pyridyl replacement was purchased as the 4-halogen analogue and reacted with 4-bromobenzoic
acid under palladium catalyzed conditions. The resultant carboxylic acid was then coupled with the
benzylamine component using standard conditions.
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
Table 4. SAR for compounds 18 - 25
O
O
N
H
R
CYP inhibition
Solu
bilit
y /
PAM
PPB / PA /
% free 10-6
cms-1
IC50 / M at 100 M ATP
IC50 / M
Cp
d
R
2C
9
ROCK1 ROCK2 PKA
2D6 3A4
M
0.34
0.11±0.0 0.012±0.
>2
0
H2N
18
±0.0 21.7 88.3
2
-
17.8 11.9
N
1
001
0.085±0. 0.014±0. 1.2±
18.
1
F
19
20
21
24.1 90.8
14.7
>20 >20
N
02
003
0.1
0.72±0.0 0.14±0.0
Cl
7.8
-
-
-
-
-
-
-
-
-
-
-
-
N
9
2
0.61±0.0 0.096±0. 5.7±
005 0.5
N
4
ACS Paragon Plus Environment

Page 13 of 89
1
2
3
4
5
6
Journal of Medicinal Chemistry
N
22
>10
>10
>10
-
-
-
-
-
-
-
0.05
F
0.011±0. 0.001±0.
7
8
23
8±0. 68.4 93.0
005
21.4 7.7
6.4
N
9
001
0.21
9.4
001
0.027
0.70
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
N
24
25
2.0 73.5 72.0
19.7 6.0 11.6 19.6
N
>2
N
N
-
50.5 66.9
12.5
>20 16.7
0
Introduction of 2-amino and 2-fluoro in compounds 18 and 19 resulted in a slight loss in potency at
ROCK2. However, these compounds now had a clean profile in the CYP inhibition assays. Introduction
of 2-chloro and 2-methyl in compounds 20 and 21 caused a much larger reduction in potency, while the
2,6-dimethyl analogue 22 was inactive. As a result, none of these compounds were advanced to HT-
ADME for screening. The most surprising result was 3-fluoro compound 23. Despite the lack of
groups ortho to the pyridine nitrogen to afford steric hindrance, this compound had much reduced
inhibition of both CYP2C9 and CYP3A4 while being the most potent ROCK2 inhibitor yet identified.
There was a concern over the lability and possible reactivity of the halo-pyridine analogues 19 and 20.
To address this both compounds were tested in our ALARM NMR assay44 and shown to be non-
reactive. The final two compounds were heterocyclic replacements of the pyridine. As expected
pyrimidine 24 displayed an improved CYP inhibition profile which was accompanied by a moderate
loss in potency at ROCK2. Disappointingly, introduction of the pyridazine in compound 25 was not
tolerated and resulted in essentially complete loss of potency at ROCK2.
Scheme 2. Synthetic route towards (benzyloxy)benzene analogues 16 - 25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 89
O
O
R
Hal
i.
OH
1
2
3
4
+
R
N
OH
N
Y
X
Br
Y
X
ii.
O
5
6
7
OMe
N
R
H
N
Y
X
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
Reagents and Conditions: i) Pd(PPh3)4, Cs2CO3, DME, DIPEA, 80 °C, 72 h; ii) (R)- or (S)-1-(3-
methoxyphenyl)ethan-1-amine, EDAC, HOBt, 0 °C to RT, O/N.
Compound 18 with the 2-aminopyridine hinge binding motif had shown good potency at ROCK2 and a
clean CYP inhibition profile. Data from the crystal structures suggested that substitution off the amino
group would be more tolerated in ROCK1 and ROCK2 than PKA. Such an approach could be used to
access the extended hinge driving selectivity versus PKA and also enabling moderation of compound
properties. To interrogate this hypothesis analogues were synthesized that introduced small alkyl and
alkoxy groups off the 2-amino group. The desired compounds were obtained in one-step from the
chloro-pyridine compound 20. Treatment of compound 20 with the chosen amine under palladium
catalyzed conditions afforded the desired compounds in 34 - 69% yield.
Scheme 3. Synthetic route towards (benzyloxy)benzene analogues 26 - 30
O
O
O
O
i.
N
R
N
H2N
R
H
H
+
Cl
HN
N
20
N
t
Reagents and Conditions: i) Pd2(dba)3, sodium 2-methylpropan-2-olate, BuXPhos, dioxane, N2, 80 °C,
O/N.
The most striking SAR from this series was the oblation of activity at PKA with all the compounds
having high micro molar IC50s (Table 5). Both compound 26 with an ethyl substituent and compound
27 with the methylene cyclopropyl substituent maintained respectable potency at both ROCK1 and
ROCK2 while displaying good selectivity versus PKA and only a small reduction in BEI. The
ACS Paragon Plus Environment

Page 15 of 89
Journal of Medicinal Chemistry
methylene propan-2-ol of compound 28 and the methoxy ethyl of compound 29 were less well tolerated
at ROCK1 and ROCK2. Compound 30, the homologue of 29, with the methoxy propyl group had a
similar overall profile to compounds 26 and 27.
1
2
3
4
5
6
7
8
9
Table 5. SAR for compounds 26 - 30
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
O
O
N
H
R
N
IC50 / M at 100 M ATP
ROCK2
Cpd
R
BEI
ROCK1
ROCK2
PKA
>10
N
H
26
27
28
29
30
0.65±0.01 0.084±0.02
18.8
17.8
15.3
16.4
17.0
N
H
0.40±0.04 0.068±0.002 8.0±0.3
OH
N
2.4±0.4
1.5±0.1
0.39±0.05
0.22±0.1
>10
>10
>10
H
O
N
H
N
H
O
0.58±0.05 0.075±0.02
Having identified suitable motifs to ameliorate the CYP inhibition, the project now turned its focus to
leveraging the structural data. This suggested that compounds with improved potency at ROCK1 and
ROCK2 could be obtained through targeting an interaction with the aspartic acids on the side of the
pocket (ROCK1 160, 202, ROCK2 176, 218).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 89
While many enantiomerically pure -substituted benzylamines are commercially available, the -
substituents are typically small alkyl or aryl in nature. As a result a flexible synthetic route that
provided the opportunity to introduce hydrophilic groups in a late synthetic step was investigated. To
reach the aspartic acids on the side of the pocket for ROCK1 and ROCK2, the structural data suggested
that a two carbon linker from the benzylic carbon would be the optimal spacer length for incorporation
of the amine. Fortunately many enantiomerically pure 3-amino-3-phenylpropanoic acids are available.
(R)-3-Amino-3-(3-methoxyphenyl)propanoic acid was identified as the key starting material as this
incorporated both the correct configuration at the benzylic carbon and the optimal 3-methoxyphenyl that
had been identified in the previous SAR studies.
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
Treatment of boc-(R)-3-amino-3-(3-methoxy-phenyl)-propionic acid 31 with butyl chloroformate
followed by subsequent reaction with sodium borohydride afforded alcohol 32. Treatment with mesyl
chloride efficiently afforded key intermediate 33 which was appropriately functionalized to enable the
introduction of nitrogen containing motifs that were hydrophilic in character. Finally, treatment with
HCl/dioxane afforded, after aqueous work-up or trituration, the desired enantiomerically pure
benzylamines 34 with hydrophilic groups attached from the benzylic carbon in high chemical purity
(Scheme 4). This synthetic route also provided a route to alcohol 35 which enabled additional
analogues to be synthesized.
Scheme 4. Synthetic route towards key intermediates for compounds 36 - 44.
OH
OH
OMs
i.
ii.
O
O
O
O
O
O
O
O
N
O
N
O
N
H
H
H
31
32
33
iii. iv.
iii.
R1 R2
N
OH
O
O
H2N
35
H2N
34
ACS Paragon Plus Environment

Page 17 of 89
Journal of Medicinal Chemistry
Reagents and Conditions: i) EDAC, HOBt, 0 °C - RT, O/N, 61%; ii) Na2CO3, DMSO, 110 °C, 34%; iii.
Pd(PPh3)2Cl2, Cs2CO3, Dioxane/H2O, 100 °C, 75%; iv. Chiral separation; v. TFA, DCM, 61%.
1
2
3
4
5
6
7
8
9
The first analogue compound 36 (Table 6) gave great encouragement being 15 times more potent than
compound 5, the corresponding compound without the ethyl pyrrolidine. Unfortunately this increase in
potency was likely due to the preferred (R)-configuration rather than a beneficial interaction with
ASP202 (ROCK1) and ASP218 (ROCK2). Compounds 37 and 38 maintained a very respectable BEI at
ROCK2 and the pyrrolidine and hydroxyl may impart benefits to drug-like properties. However
compounds 39 to 43 were all substantially less potent. The only compound of note was 43, which both
reinforced the positive view of the 3-fluoro pyridine motif identified in compound 23 and also showed
good selectivity versus PKA. To confirm the importance of the benzylic carbon configuration as a
potency driver, the (R)-methyl was incorporated with the propyloxy from compound 8 that drove
selectivity versus PKA. This afforded compound 44 with an impressive BEI at ROCK2 of 21.7 in
conjunction with over 600-fold selectivity versus PKA.
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 6. SAR for compounds 36 - 44
O
R2
R3
N
R1
H
N
PKA/
IC50 / M at 100 M ATP
ROCK2
BEI
Cpd
R1
R2
R3
ROCK2
ROCK1
ROCK2
PKA
Selectivity
N
36
37
H
H
H
0.046
0.014
0.008
0.003
0.065
0.095
21.1
20.7
9
N
OMe
37
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 89
38
39
H
OMe
OMe
0.039
0.139
0.006
0.36
1.0
22.7
16.7
60
32
1
2
3
4
5
6
7
8
OH
N
N
N
2-F
0.032
0.18
O
40
41
2-F
2-F
OnPr
F
0.55
0.24
>10
14.2
16.8
56
4
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
0.082
0.33
42
43
44
2-F
3-F
H
OnPr
OnPr
OnPr
1.6
0.72
0.35
0.14
>10
>10
15.8
16.8
21.7
28
OH
OH
290
670
Me
0.16±0.07
0.015±0.012 >10
In an attempt to improve the selectivity profile, the next synthetic effort was directed at targeting an
interaction with Asp160 (ROCK1) and Asp176 (ROCK2). Consideration of the structural data
suggested that the central aryl ring afforded the preferred sites for attachment, optimally from the meta-
position relative to the hinge binding pyridine. For compounds 46 – 49, it was possible to effect the
displacement of the 3-fluoro from the central aryl ring of 45 with the desired amine (Scheme 5).
Scheme 5. Synthetic route towards compounds 46 - 49
O
O
OH
OH
i.
Br
O
+
+
H2N
HO
F
B
F
N
N
OH
O
O
F
O
N
O
H
ii.
iii.
N
N
H
N
N
NH2
45
46
ACS Paragon Plus Environment

Page 19 of 89
Journal of Medicinal Chemistry
Reagents and Conditions: i) Pd(PPh3)4, Cs2CO3, DME, DIPEA, 80 °C, 72 h; ii) EDAC, HOBt, 0 °C to
RT, O/N. iii) N-ethyl-N-isopropylpropan-2-amine, diamine, DMSO, 110 °C, O/N.
1
2
3
4
5
6
7
Table 7. SAR for compounds 46 - 49
8
9
O
O
N
H
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
R
N
IC50 / M at 100 M ATP
PKA/ ROCK2
Cpd
R
Selectivity
ROCK1
ROCK2
PKA
0.008
±0.00
2
0.003±0.00 0.001±0.
N
46
8
NH2
1
001
NH2
N
H
47
48
0.19
0.021
0.012
1.0
48
76
N
NH2
0.095
0.87
H
0.021±0. 2.3±0.
003
NH2
N
49
0.31±0.03
110
H
1
Inspection of the crystal structure suggested that introduction of a 3-5 atom linker from the meta-
position (with respect to the hinge binding pyridine) of the central aryl ring should be suitable to
position a nitrogen in the vicinity of the ASP160 in ROCK1 and ASP176 in ROCK2 and present the
opportunity to make an interaction. The first analogue 46 introduced piperidin-4-ylmethanamine which
unfortunately resulted in the most potent compound across ROCK1, ROCK2 and PKA (Table 7). To
furnish compounds with more flexibility to secure an interaction with the aspartic acid 160 ROCK1 and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 89
176 ROCK2 the next analogues investigated varying lengths of acyclic diamines. Increasing the linker
length with compounds 47 (ethane-1,2-diamine) 48 (propane-1,3-diamine) and 49 (butane-1,4-diamine)
gave the desired increased selectivity over PKA which can be explained by looking at the structure of
compound 47 in PKA (Figure 3). Here the shorter linker of compound 47 allows a hydrogen bond to be
made in PKA with glutamic acid 128. As the linker length increases to 4 atoms in 49, binding here
would seem to be reduced. For ROCK1 and ROCK2, the linker length of compound 48 would seem to
interface the shorter aspartic acid residues in these proteins.
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 3. Crystal structure of compound 47 in the active site of PKA (PDB 6E99). Structure was
obtained by co-crystallization.
Having established that a 5-atom linker afforded compounds with the best selectivity, ring constrained
variants of butane-1,4-diamine were considered. For the design of these compounds emphasis was
placed on the calculated physicochemical properties of the resulting compounds based on medicinal
chemistry guidelines like drug-like properties,45 fraction sp3 carbon46 and 3/75.47 Cognizant of the CYP
inhibition liability of compound 49 due to the unsubstituted pyridine, an alternative hinge binding motif
was selected for these analogues. Piperidine was selected ahead of 2-aminopyridine, 2-fluoropyridine
and 3-fluoropyrdine because it increases both the PSA and Fsp3 to values associated with a safer
predicted toxicity profile.
ACS Paragon Plus Environment

Page 21 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
(a)
(b)
Figure 4. (a) Radar plot of calculated physicochemical properties and (b) predicted reduced risk
of toxicity, for compounds 1, 5, 6, 16, 44 and 58.
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
From the radar plot visualization (Figure 4) it is evident that only compound 58 has excellent predicted
drug-like properties including Fsp3. The radar plot data is also shown in tabular form below (Table 8).
In addition compound 1, the initial hit, and compound 58 have calculated PSA and cLogP that suggest a
reduced risk of toxicity based on the 3/75 rule. These clearly show that compound 58 has excellent
calculated drug-like properties.
Table 8. (a) Calculated physicochemical properties for compounds 1, 5, 6, 16, 44 and 58
Cpd CLogP PSA
HBA LogD 7.4 HBD Mol wt NAR NRB Fsp3
1
5
2.72
3.37
3.29
3.60
4.65
2.60
62.2
42.0
51.2
51.2
51.2
93.4
4
3
4
4
4
7
2.86
3.20
3.04
3.46
4.34
0.63
2
1
1
1
1
3
332.4
288.3
318.4
332.4
360.4
445.6
3
3
3
3
3
3
6
4
5
5
7
7
0.14
0.05
0.10
0.14
0.22
0.35
6
16
44
58
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 89
To facilitate these analogues an alternative synthetic route was required. The (R)-(+)-1-(3-substituted-
phenyl)ethylamine 51 was reacted with 4-bromo-2-fluorobenzoic acid 50 to afford intermediate 52.
Nucleophilic displacement of the fluorine of 52 with mono boc-protected diamines was efficiently
mediated using standard conditions and afforded the target compounds without the hinge binding motif
53. These diastereomeric mixtures were subjected to chiral chromatography to afford stereochemically
pure material. Palladium catalyzed coupling with 4-chloropyrimidine afforded the boc-protected
analogues, which upon treatment with TFA furnished the desired compound 58 (Scheme 6).
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
Scheme 6. Synthetic route towards compounds 54 - 58
O
F
O
F
O
O
OH
H2N
R
N
H
R
i.
+
Br
Br
50
ii.
51
52
O
O
N
O
O
N
H
R
N
R
H
iii. iv.
Br
N
N
N
Boc
N
H
53
H2N
58 R = Me
Reagents and Conditions: i) EDAC, HOBt, 0 °C to RT, O/N; ii) N-ethyl-N-isopropylpropan-2-amine,
diamine, DMSO 110 °C. O/N; iii) CH3CO2K, 4,4,4',4',5,5,5',5'-octamethyl-2,2'-bi(1,3,2-dioxaborolane),
4-chloropyrimidine hydrochloride, PdCl2(dppf)-DCM, Cs2CO3, K2CO3, dioxane, 65 °C, 3 h; iv) TFA.
Table 9. SAR for compounds 54 - 58
O
R2
N
H
R1
N
N
Solubility
IC50 / M at 100 M
PKA/ROCK2
Selectivity
PPB PAMPA
/ %
/ 10-6
Cpd
R1
R2
/ M
ATP
ACS Paragon Plus Environment

Page 23 of 89
Journal of Medicinal Chemistry
free
-
cms-1
ROCK1 ROCK2 PKA
1
2
3
4
5
6
7
8
3-
NH2
N
H
54
55
56
57
58
0.99
1.9
0.072
0.32
>10
>10
>10
0.88
3.3
140
31
-
-
OMe
9
3-
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
NH2
N
63.9
70.6
1.4
-
OnPr
3-
NH2
N
1.7
0.27
38
-
-
OnPr
3-
N
NH2
0.16
0.12
0.019
0.014
46
65.5
51.2
66.0
57.1
2.6
2.5
OMe
3-
N
NH2
230
OMe
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
Compound 54 (Table 9) was prepared as the pyrimidine analogue of compound 49 to establish the
concept of replacing the pyridine hinge binding motif. This was confirmed with both compounds
having very similar profiles for both potency and selectivity. Replacement of the butane-1,4-diamine
with pyrrolidin-3-ylmethanaminethat also incorporated the propyloxy group on the terminal aryl ring
was not promising. Both compounds 55 and 56 were not only less potent at ROCK2 but were also far
less selective versus PKA. Incorporation of piperidin-3-ylmethanamine in compounds 57 and 58 was
far more beneficial. (R)-isomer 57 was found to be a potent inhibitor of ROCK2 with modest selectivity
over PKA. However the (S)-isomer 58 retained this potency against ROCK1 and ROCK2 and
introduced more than a 200-fold selectivity versus PKA. The absolute configuration of compounds 55,
56, 57 and 58 was not determined. The configuration was assigned arbitrarily using the crystal
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 89
structures to predict which the most selective compound was. Further studies are ongoing to assign the
absolute stereochemistries of these 4 compounds. Additional profiling of compounds 57 and 58 showed
they have good solubility and free fraction, with not surprisingly moderate permeability as assessed by
PAMPA due to the primary amine.
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
To identify the optimal compound for in vivo studies 44 of the above described ROCK inhibitors were
screened in the AbbVie kinome panel that includes 80 kinases (Supplementary Information). This
highlighted compound 16 as the best candidate. It provides IC50 values >10 M for 63 of the kinases
and over 1 M for a further 10 kinases. Consistent with the project screening data it was highly potent
at both ROCK1 (IC50 1 nM) and ROCK2 (IC50 1 nM). It only inhibited two other kinases at the low
nanomolar level PKG1A (IC50 1 nM) and PKA (IC50 47 nM) showing this to be a highly selective dual
ROCK1 and ROCK2 inhibitor suitable for in vivo studies. The broad kinome profiling showed that
compounds 55, 56, 57 and 58 were the most selective against PKG1A, while compound 22, the hindered
2,6-dimethyl pyridine was shown to be inactive at all 80 kinases.
Table 10. Lead compounds in pMYPT1 ELISA cell assay
Cpd
6
pMYPT1 / IC50 M*
0.075
PAMPA / 10-6 cms-1
PSA / Å
51.2
20.2
16.2
14.9
2.5
8
0.961
51.2
12
14
16
0.45
60.5
2.23±1.3
0.012±0.01
54.5
19.6
51.2
ACS Paragon Plus Environment

Page 25 of 89
Journal of Medicinal Chemistry
17
18
19
24
25
27
30
44
58
49.8
0.173±0.03
0.116±0.09
0.415±0.24
7.83
-
51.2
77.2
51.2
64.1
64.1
63.3
72.5
51.2
93.4
1
2
3
4
5
6
7
8
-
14.7
19.7
12.5
5.1
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
0.963
0.802
9.2
0.056±0.01
0.094
18.8
2.5
* Errors shown for compounds tested multiple times.
A subset of these molecules were then progressed into a pMYPT1 ELISA, which is a ROCK dependent
cell-based assay (Table 10). Compound 6, the 3-methoxy analogue showed good potency while
compounds 8, 12 and 14, with the larger aryl substituents, were less potent. Compound 16, was the
most potent compound (IC50 0.012 M) and confirmed this as an excellent dual ROCK1 and ROCK2
tool compound. This compound has been made available through the AbbVie Compound Toolbox.48
Compound 17, the isomer of 16, was 3 logs less potent, consistent with the enzyme assay result.
Compounds 18, 19, 24, 27 and 30 in which the hinge binding motif was modified to reduce CYP
inhibition demonstrated an 8 - 15 fold shift in the cell assay. Cell data for pyridazine 25 was almost 10
M which agreed with the observed enzyme data and confirmed that this functional group was not
tolerated. Compounds 44 and 58 were shown to be excellent dual ROCK1 and ROCK2 inhibitors with
IC50s of 0.056 and 0.094 M respectively.
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 26 of 89
Table 11. Pharmacokinetics of compound 16 in C57BL6 mouse and Sprague Dawley rat
1
2
3
4
IV
IP
PO
5
6
7
8
9
Sample
Analyz
ed
T
Speci
es
AUC
CL
Tm Cmax
AUC
ug*h/
mL
Cmax
AUC
t1/2
hr
Vss
t1/2
hr
F
t1/2 ma
ug*h/m L/h/k
ax
hr
ug/m
L
ug/m
L
ug*h/ F %
mL
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
L/kg
%
hr
x
L
g
hr
SD
Rat
1. 0.6
Plasma 1.5
2.19
1.5
2.0
1.0 0.25 1.03
1.80
82
0.42
0.85
39
1
7
Plasma
Brain
-
-
-
-
-
-
-
-
0.8 0.25 2.36
1.2 0.25 1.92
2.30
2.63
-
-
-
-
-
-
-
-
-
-
SD
Rat
-
-
C57B
L6
2. 0.3
Plasma
Brain
-
-
-
-
-
-
-
-
1.5 0.42 1.66
1.8 0.33 0.99
2.08
1.78
-
-
0.84
-
1.42
-
-
-
0
3
Mous
e
-
-
All groups n = 3, dose of 10 mol/kg
Pharmacokinetic analyses (Table 11) of compound 16 in both C57BL6 mice and Sprague Dawley rats
all 10 mol/kg with group size n = 3, were performed using IV, IP and PO administration. The
compound has a reasonable half-life and good exposure in both plasma and brain. While the oral
bioavailability was moderately low it was sufficient to advance the compound for in vivo testing
Beneficial effects of ROCK inhibitors, like Fasudil and Y-27632, have been reported in a variety of
animal models of different neurodegenerative diseases.49 To confirm such potential in this hit series,
compound 16 was investigated in an optic nerve crush model50 in rats. In this model, compression of
the optic nerve leads to axon dissection followed by degeneration of the retinal nerve fiber layer (RNFL)
in the eye over time. Compound 16 was orally administered once daily at a dose of 10 mg/kg for 5
ACS Paragon Plus Environment

Page 27 of 89
Journal of Medicinal Chemistry
weeks and compared to a vehicle control. Subsequent immunohistochemical staining of the retina
revealed a significant increase in the number of ßIII tubulin positive fibers in compound 16 treated
animals indicative of a protection of the axons in the retinal nerve fiber layer by ROCK inhibition
(Figure 5). Further studies are ongoing to identify clear positive controls for this assay.51
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. Protection of the retinal nerve fiber layer with compound 16
CONCLUSIONS
Following a high-throughput screen compound 1, a 4-(pyridin-4-yl)benzamide, was identified as the
starting point for a medicinal chemistry project to identify a potent and kinome-selective dual ROCK1
and ROCK2 inhibitor. This compound was selected due to its excellent physicochemical properties and
synthetic accessibility. Parallel synthesis was employed to generate focused compound libraries that
furnished immediate SAR and most notably compound 12. Crystal structures were obtained for this
compound in PKA, ROCK1 and ROCK2. This provided critical structural information to the medicinal
chemistry team to drive compound design and enabled a unique comparison of the 3 binding sites with
the same bound inhibitor. Interrogation of the co-crystal structures indicated the preferred configuration
at the central benzylic carbon would be (R). This information was used to select monomers for focused
library design resulting in compound 16 with the (R)-configuration and compound 17 with the (S)-
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 89
configuration. Comparison of the potency of these two compounds clearly supported the structural
hypothesis that the (R)-configuration was preferred. Profiling of these early leads in a HT-ADME assay
panel showed the compounds were pan-inhibitors of CYP enzymes, especially 3A4, predicting
unwanted drug-drug interactions. Substitution (2-Cl, 2-NH2, 2-F, 3-F) of the pyridine or replacement
with pyrimidine afforded compounds with a clean CYP inhibition profile. Piperidine was selected
ahead of 2-aminopyridine, 2-fluoropyridine and 3-fluoropyrdine because it increases both the PSA and
Fsp3 to values associated with a safer predicted toxicity profile. Additional profiling of compound 16
showed it was a potent and selective dual ROCK1 and ROCK2 inhibitor in both enzyme and cell assays
and efficacious in the retinal nerve fiber layer model after oral dosing. This tool compound has been
made available through the AbbVie Compound Toolbox. Finally, the co-crystal structures also
identified that aspartic acid residues 176 and 218 in ROCK2, which are glutamic acids in PKA, could be
targeted as residues to drive both potency and kinome selectivity. Introduction of a piperidin-3-
ylmethanamine group to the compound series resulted in compound 58, a potent and selective dual
ROCK1 and ROCK2 inhibitor with excellent predicted drug-like properties including a significantly
reduced Cyp inhibition profile.
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
EXPERIMENTAL SECTION
Unless otherwise stated, reagents were purchased from Sigma Aldrich, Acros, Alfa Aesar or the Sigma
Aldrich Custom Packaged Reagent service. Reagent/reactant names given are as named on the
commercial bottle or as generated by IUPAC conventions or CambridgeSoft® ChemDraw Ultra 9.0.7.
Compound names are generated by IUPAC conventions or CambridgeSoft® ChemDraw Ultra 9.0.7.
For all final compounds purity was established by HPLC and purity confirmed to be ≥95%.
Analytical data is included within the procedures below, in the illustrations of the general procedures, or
1
13
19
in the tables of examples. Unless otherwise stated, all H, C and F NMR data were collected on a
Varian Mercury Plus 400 MHz or a Bruker AVIII 300 MHz or a Bruker 500 MHz instrument; chemical
ACS Paragon Plus Environment

Page 29 of 89
Journal of Medicinal Chemistry
shifts are quoted in parts per million (ppm). HPLC analytical data are either detailed within the
experimental or referenced to the table of LCMS and HPLC conditions, using the lower case letter in
high resolution mass spectra (HRMS) were obtained using an Agilent 6550 Q-TOF spectrometer with
an ESI source.
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
Chiral LC analysis was performed using Varian 218 LC pumps, a Varian CVM 500 with switching
valves and heaters for automatic solvent, column and temperature control. Detection methods included a
variable wavelength UV detector (Varian Prostar 320), an in-line polarimeter (PDR-chiral advanced
laser polarimeter, model ALP2002) used to measure qualitative optical rotation (+/-) and an evaporative
light scattering detector (ELSD) (a PS-ELS 2100 (Polymer Laboratories)). ELSD settings were as
follows: evaporator: 46 °C, nebulizer: 24 °C and gas flow: 1.1 SLM.
Chiral SFC analyses were performed using a Shimadzu Nexera UC SFCMS instrument. Detection
methods included a PDA detector (Model SPD-M20A), ESI (+/-) MS (model LCMS-2020), and an in-
line polarimeter (PDR-Separations advanced laser polarimeter, model 4G-ALP) used to measure
qualitative optical rotation (+/-).
Table 12. List of HPLC methods
Method HPLC Conditions
HPLC: A gradient of 10-100 % MeCN (A) and 0.1% trifluoroacetic acid in water (B) was
a
used, at a flow rate of 2.0mL/min (0-0.1 min 10% A, 0.1-2.6 min 10-100% A, 2.6-2.9 min
100% A, 2.9-3.0 min 100-10% A. 0.5min post-run delay).
HPLC: The gradient was 5-60% B in 0.75 min then 60-95% B to 1.15 min with a hold at
95% B for 0.75 min (1.3 mL/min flow rate). Mobile phase A was 10mM ammonium
acetate, mobile phase B was HPLC grade MeCN. The column used for the chromatography
b
ACS Paragon Plus Environment