Design, synthesis, and biological evaluation of novel, highly active soft rock inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.5b00308

ROCK1 and ROCK2 play important roles in numerous cellular functions, including smooth muscle cell contraction, cell proliferation, adhesion, and migration. Consequently, ROCK inhibitors are of interest for treating multiple indications including cardiovas

Full text of DOI:10.1021/acs.jmedchem.5b00308

Subscriber access provided by NEW YORK UNIV
Article
Design, Synthesis and Biological Evaluation
of Novel, Highly Active Soft ROCK Inhibitors.
Sandro Boland, Arnaud Bourin, Jo Alen, Jacques Geraets, Pieter Schroeders,
Karolien Castermans, Nele Kindt, Nicki Boumans, Laura Panitti, Silke Fransen,
Jessica Vanormelingen, Jean Marie Stassen, Dirk Leysen, and Olivier Defert
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 21 Apr 2015
Downloaded from http://pubs.acs.org on April 21, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Design, Synthesis and Biological Evaluation of
Novel, Highly Active Soft ROCK Inhibitors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
*
,†
†
†
†
†
Sandro Boland, Arnaud Bourin, Jo Alen, Jacques Geraets, Pieter Schroeders, Karolien
†
†
†
†
†
Castermans, Nele Kindt, Nicki Boumans, Laura Panitti, Silke Fransen, Jessica
†
†
‡
†
Vanormelingen, Jean Marie Stassen, Dirk Leysen and Olivier Defert
AUTHOR ADDRESS:
†
Amakem Therapeutics, Agoralaan A bis, 3590 Diepenbeek, Belgium
‡
CSD Farmakem, Elvire Boelensstraat 7, 9160 Lokeren, Belgium
KEYWORDS: RhoꢀKinase, ROCK, SAR, Soft Drug
GRAPHICAL TOC:
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT : ROCK1 and ROCK2 play important roles in numerous cellular functions,
including smooth muscle cell contraction, cell proliferation, adhesion and migration.
Consequently, ROCK inhibitors are of interest for treating multiple indications including
cardiovascular diseases, inflammatory and autoimmune diseases, lung diseases and eye diseases.
However, systemic inhibition of ROCK is expected to result in significant side effects. Strategies
allowing reduced systemic exposure are therefore of interest. In a continuing effort towards
identification of ROCK inhibitors, we here report the design, synthesis and evaluation of novel
soft ROCK inhibitors displaying an ester function allowing their rapid inactivation in the
systemic circulation. Those compounds display subꢀnanomolar activity against ROCK and
strong differences of functional activity between parent compounds and expected metabolites.
The binding mode of a representative compound was determined experimentally in a singleꢀ
crystal Xꢀray diffraction study. Enzymes responsible for inactivation of these compounds once
they enter systemic circulation are also discussed.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ACS Paragon Plus Environment

Page 3 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
TEXT:
Introduction
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
Rhoꢀassociated, coiled coil containing protein kinases (ROCKs), also known as RhoꢀKinases,
1
are serine/threonine kinases belonging to the AGC family. In humans, ROCKs are represented
by two isoforms, namely ROCK1 and ROCK2. These isoforms share an overall homology of
6
5%, while being 95% homologous in the kinase domain. ROCK plays an important role in
numerous cellular processes including smooth muscle cell contraction, cell proliferation,
2
adhesion and migration. Consequently, ROCK inhibitors are of potential interest for the
3
treatment of multiple indications including hypertension, pulmonary arterial hypertension
4
5ꢀ7
(
PAH), inflammatory and autoimmune diseases
such as asthma, Chronic Obstructive
8
Pulmonary Disease (COPD), Inflammatory Bowel Disease (IBD), and eye diseases (e.g.
glaucoma, ageꢀrelated macular degeneration (AMD)). The ROCK inhibitor Fasudil is used in
Japan since 1995 for the treatment of cerebral vasospasm. More recently, Fasudil has shown
9
clinical efficacy on PAH. However, systemic inhibition of ROCK induces a pronounced
1
0
decrease in blood pressure, and thus reduces the therapeutic window of ROCK inhibitors.
Additionally, ROCK inhibitors were recently shown to induce a reversible reduction in
11
lymphocyte counts. As a consequence, systemic ROCK inhibition should preferably be
avoided.
Topical application, which results in lower systemic exposure, is a common strategy in the
recent development of ROCK inhibitors. The field of eye diseases has received particular
attention, and several ROCK inhibitors have been clinically tested as intraocular pressure (IOP)ꢀ
8
lowering agents for topical treatment of glaucoma. Those include Yꢀ39983 (Novartis, Basel,
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
12
Switzerland), Ripasudil (Kꢀ115, Kowa Company Ltd, Nagoya, Japan), ARꢀ12286 & Rhopressa
(
ARꢀ13324, Aerie Pharmaceuticals, Inc., Bedminster, USA) and ATS907 (Altheos, Inc., South
San Francisco, USA). Ripasudil was recently approved in Japan as GLANATEC® ophthalmic
solution (0.4%) for the treatment of glaucoma and elevated IOP; and is currently undergoing
Phase II trials for treatment of diabetic retinopathy accompanying diabetic macular edema.
Additionally, Yꢀ27632, a reference ROCK inhibitor, has delivered promising results with respect
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
3
to corneal wound healing and several cases of Fuchs corneal dystrophy were successfully
1
4,15
treated by administration of Yꢀ27632 eye drops.
The disclosed structures of some of those
ROCK inhibitors are presented in figure 1. Besides ophthalmology, lung diseases were also
investigated as potential applications, and ROCK inhibitors optimized towards inhaled delivery
1
1
and reduced systemic exposure have been proposed for treatment of PAH. Even when used
topically, ROCK inhibitors are not devoid of side effects. In the context of ophthalmology, a
common adverse event resulting from topical administration of ROCK inhibitors is mild to
severe conjunctival hyperemia (redness) due to mechanismꢀbased smooth muscle cell relaxation
8,16,17
in conjunctival blood vessels.
Consequently, ROCK inhibitors with an improved therapeutic
window would still be of high interest.
Figure 1. Structure of the ROCK inhibitors Yꢀ27632, Yꢀ39983, Fasudil and Ripasudil
4
ACS Paragon Plus Environment

Page 5 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Soft drugs are biologically active compounds that are designed to undergo metabolic inactivation
1
8,19
by controlled conversion of the parent molecule into a predictable, nontoxic metabolite.
While soft drugs do not necessarily show fast inactivation, they can be optimized as locally
acting compounds, so that they retain sufficient stability in their target organ, while being rapidly
inactivated once entering systemic circulation. Such inactivation usually occurs in blood, but can
also be achieved in liver. A series of 3ꢀ[2ꢀ(aminomethyl)ꢀ5ꢀ[(pyridinꢀ4ꢀyl)carbamoyl]phenyl]
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
2
0
benzoates has recently been reported as the first example of soft ROCK inhibitors in literature.
As a first proof of concept, 1 (Figure 2) demonstrated in vivo effect in rabbits as IOPꢀlowering
20
agent, without causing noticeable conjunctival hyperemia. Additionally, AMA0076, a soft
ROCK inhibitor derived from this series, effectively reduces IOP in ocular normotensive and
2
1
acute hypertensive rabbits without causing distinct conjunctival hyperemia. This compound has
undergone phase IIa clinical trials as an IOPꢀlowering agent for the treatment of glaucoma. We
herein report the design, evaluation and characterization of 4ꢀ(aminomethyl)ꢀNꢀ(pyridinꢀ4ꢀyl)ꢀ
3[3ꢀ(phenylcarbamoyl)phenyl] benzamide derivatives displaying improved onꢀtarget activity and
controllable inactivation kinetics.
Figure 2. Structure of 1, a soft ROCK inhibitor with IOP lowering efficacy.
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Results
Our previous work on 3ꢀ[2ꢀ(aminomethyl)ꢀ5ꢀ[(pyridinꢀ4ꢀyl)carbamoyl]phenyl] benzoates sought
to confer soft drug properties to an existing ROCK inhibitor scaffold. This goal was achieved
through the introduction of ester groups that are hydrolyzed in plasma; yielding carboxylic acid
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
0
metabolites with markedly decreased functional activity. Molecular docking of such ester
2
2
derivatives inside the ATPꢀbinding site of ROCK1 (Figure 3), however suggested that the onꢀ
target activity of this series of compounds could still be improved. Indeed, the newly introduced
ester function was positioned in close vicinity of the carboxyl side chain of residue Asp216,
potentially causing some degree of electrostatic repulsion. This finding suggested displacement
of the ester group towards another part of the scaffold and replacement of the ester by an amide.
Additionally, ester derivatives such as 1 had their side chain within contact distance of Phe120.
This observation was noteworthy, since Phe120 is present in both ROCK1 and ROCK2, but is
absent from all other kinases from the AGC family, with the exception of DMPK (dystrophia
myotonicaꢀprotein kinase). The simple introduction of an aromatic structure on that position
therefore provided the possibility to form specific πꢀstacking interactions, as well as an
opportunity to further increase the selectivity of the chemical series. Combined together, those
observations suggested that a benzamide analog such as 2 (Figure 3) could be an interesting
platform for the development of further soft ROCK inhibitors, as it would interact favorably with
both Asp216 (formation of an additional Hꢀbond) and Phe120 (πꢀstacking interactions).
Interestingly, a number of readily available aniline building blocks display an ester function,
thereby offering rapid exploration of structureꢀactivity or structureꢀproperty relationships.
6
ACS Paragon Plus Environment

Page 7 of 60
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 3. Molecular modeling of compound 1 (C in cyan) in the ATPꢀbinding site of ROCK1
(PDB 2ETR; green). The ester function lies in close vicinity (3.8 Å) of the carboxylic acid side
chain from Asp216. Additionally, the oxolanꢀ2ꢀyl moiety of 1 closely contacts the side chain of
Phe120 (orange residue). The binding mode suggests replacement of the ester by amide and
combination with an aromatic ring to form πꢀπ interactions with Phe120. The resulting structure
2 is shown in grey. Position of Yꢀ27632 in structure 2ETR is shown for reference (magenta).
A number of ROCK inhibitors structurally derived from 2 were obtained by adapting the
2
0
procedure described for the synthesis of earlier soft ROCK inhibitors. Compounds were
prepared in ca. 3ꢀ5 steps, starting from the reported advanced intermediate 3 (Scheme 1).
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 1: (a) Pd(dppf)Cl , Na CO , H O, DMF, 100 °C; 16h; (b) T3P, DMAP, DCM, 0°C
ꢀ
2
2
3
2
rt, 16h; (c) LiOH, ACN/water, rt, 2h; (d) ROH, T3P, DMAP, DCM, 0°C
DCM, rt, 4h.
ꢀ
rt, 16h; (e) HCl(g),
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A biphenyl structure, displaying either a carboxylic acid function or an amino function allowing
further coupling, is introduced via Pdꢀcatalyzed Suzuki reaction, using the commercially
available 3ꢀcarboxyphenylboronic acid or 3ꢀaminophenylboronic acid, to obtain respectively
intermediates 4 and 5. Amide formation with selected anilines or benzoic acids can be achieved
through various methods to activate the acid. T3P used at reduced temperatures proved to be the
most effective coupling agent amongst the different conditions tested. At this stage, the
intermediate can either be treated with HCl gas, yielding the methyl ester as final compound, or
hydrolyzed via saponification with 1M aqueous LiOH solution. The resulting carboxylic acid can
then be further coupled with various alcohols, allowing modulation of the ester group. As before,
8
ACS Paragon Plus Environment

Page 9 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the NꢀBoc group in the resulting esters is deprotected with HCl gas, to yield the final
compounds. Alternatively, the carboxylic acid intermediate can be deprotected and isolated, as
this product corresponds to the expected metabolite resulting from enzymatic hydrolysis of the
candidate soft ROCK inhibitors.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Activity of initial derivatives and corresponding metabolites.
Cpd
R
IC50
IC50
IC ROCK2,
50
EC MLCꢀPP
t1/2 Plasma
50
a
b
ROCK1
ROCK2
200µM ATP (nM)
(nM)
(min)
(
nM)
NT
NT
NT
NT
<1
(nM)
54 ± 16
3.8 ± 1.4
2.5 ± 0.5
<1
Y-27632
NA
NA
NA
H
NT
197 ± 50
49 ± 16
31 ± 8
125 ± 18
144 ± 56
4.1 ± 1.0
11 ± 2
NT
630
210
94
>120
NT
Ripasudil
1
2
6
7
8
9
<5
22
>120
<5
3ꢀCO
Me
1.1 ± 0.5
1.1 ± 0.5
<1
79
2
3ꢀCH
CO
Me
<1
220
3.0
<5
2
2
4ꢀCO
4ꢀCH CO
3ꢀCO
3ꢀCH CO
4ꢀCO
4ꢀCH CO
Me
<1
>120
<5
2
Me
<1
<1
31
2
2
6M
7M
8M
9M
2
H
NT
NT
NT
NT
6.6 ± 0.6
1.5 ± 0.5
3.3 ± 0.8
1.8 ± 0.6
5900
1300
6600
3600
>120
>120
>120
>120
H
H
NT
2
2
H
NT
2
NT
2
2
a
b
:
Mean of at least 3 exp. : Mean of at least 2 exp. NA: Not applicable. NT: Not tested.
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Firstly, we evaluated 2, our generic template used for the development of additional soft
2
3
ROCK inhibitors, in a radiometric assay measuring phosphorylation of a peptidic substrate.
This compound encouragingly displayed a subꢀnanomolar IC50 value under standard assay
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
32
conditions, which involve presence of 1 nM enzyme and 1 µM PꢀATP. Such activity was
superior to Ripasudil and largely superior to Yꢀ27632. While Yꢀ27632 appeared relatively potent
in this assay (54 nM) compared to literature data (IC50 100ꢀ200 nM), its activity remained
consistent throughout independent experiments. One reason behind this observation could be the
low ATP concentration, which is under K and favors lower IC50 values. The encouraging
m
potency of 2 was confirmed by strong functional activity in the myosin light chain
2
4
phosphorylation (MLCꢀPP) assay. Compound 2 also displayed high stability in human plasma
(>90% remaining after 2 h). Consequently, hydrolysis of the added ester function should be
considered the only plausible mechanism of inactivation in plasma for upcoming soft drug
candidates. Addition of simple methyl esters on positions 3 and 4 was tolerated and yielded
compounds 6-9, which all displayed lowꢀnanomolar or subꢀnanomolar activity against both
ROCK1 and ROCK2 (Table 1). As the measured IC50 values were similar to or lower than
formal ROCK concentration (1 nM), comparison under standard assay conditions was actually
judged unreliable. However, as compounds were designed to occupy the ATPꢀbinding site of
ROCK, we reasoned that the inhibition they cause would be competitive with respect to ATP. A
followꢀup experiment was therefore carried out in presence of higher ATP concentration (200
µM) to reduce apparent inhibition and provide a more useful comparison. As expected, a clear
IC shift was observed for most compounds, confirming their ATPꢀcompetitive behavior. The
50
resulting data indicated that substitution on position 4 yielded stronger ROCK inhibitors, as both
8
and 9 improved over 1 and 2. Most notably, 8 retained an impressive IC of 4 nM when facing
50
10
ACS Paragon Plus Environment

Page 11 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
the increased ATP concentration. On the other hand, 6 and 7 displayed a moderate loss of
activity (4 to 5ꢀfold) with respect to 2. Evaluation in presence of 200 µM ATP was also more in
line with the functional EC50 values observed in the MLCꢀPP, with 8 and 9 inducing the
strongest inhibition of MLC phosphorylation and appearing more active than Yꢀ27632 or
Ripasudil. Stability in human plasma further differentiated 8 from the other ester derivatives, as
the former had t1/2 > 120 min, while the latter were all degraded within minutes.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The carboxylic acids resulting from ester hydrolysis, which represent the expected metabolites of
compounds 6-9, were also evaluated. While those acids retained single digit nanomolar activity
in the enzymatic ROCK2 assay, they nonetheless displayed a significant decrease of functional
activity in the MLCꢀPP assay, thereby validating ester hydrolysis as a viable inactivation
mechanism for the design of soft ROCK inhibitors. This loss of activity could be attributed, to
some extent, to reduced membrane permeability of the metabolites, as can be seen by comparing
2
5
8
and 8M (Table 2). This difference is understandable, as metabolites, which are zwitterions at
physiological pH, display decreased lipophilicity and increased topological polar surface area
(TPSA) compared to their parent compounds. In this respect, it should be noted that 8 already
displayed reduced permeability compared to 2, possibly because of its already increased TPSA.
After evaluation of this first group of analogs, we had therefore identified potent ROCK
inhibitors displaying different profiles, which prompted us for the optimization of 6, 8 and 9 as
separate subꢀseries. 7 was not subjected to further optimization, in view of the weaker inhibition
of ROCK2 (IC ) in presence of 200 µM ATP, of the lower functional activity (EC ), and of the
50
50
smaller differences between parent compound and carboxylic acid metabolite.
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 2. Cacoꢀ2 permeability, logD and TPSA for selected compounds
ꢀ6
ꢀ1 a
b
Cacoꢀ2 Papp (10 cm s )
logD (pH 7.4)
1.96*
TPSA (Ų)
99
2
8
2.3
0.8
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2.74 (exp) /1.97*
ꢀ0.56 (exp) /1.20*
125
8M
0.08
139
a
b
:
Mean of two experiments. : Mean of 3 experiments. *: Calculated (Instant J. Chem 6.3.3);
The aim of optimization of compound 6 was to increase functional activity while modulating its
stability. Indeed, while rapid inactivation in plasma is per se of interest for soft drug design, the
possibility of modulating compound stability according to the needs of a specific application or
target organ is also of high value. The hydrolysis rate of an ester by esterases such as
acetylcholinesterase is known to be the combined result of intrinsic compound reactivity, affinity
2
6
for the esterase and catalytic efficiency for a particular ester, and those combined factors can
potentially be used to modulate compound softness. In order to explore structureꢀactivity and
structureꢀproperty relationships, we first opted for modification of the amide linker, modification
of the ester side chain, and introduction of fluorine on the benzoic acid moiety (Table 3). The
corresponding metabolites were also synthesized.
Several of the resulting compounds were very strong ROCK2 inhibitors and, as discussed above,
could only be differentiated when evaluated against a higher ATP concentration. Increasing the
size of the alkyl chain was at best tolerated (6 vs. 12), but led in some occasions to a drop in onꢀ
target activity (6 vs. 11). Those observations were consistent with the MLCꢀPP assay, as
compounds with the highest onꢀtarget activity tended to have the highest effect on MLC
phosphorylation. Variation of ester moiety had a drastic effect on stability in human plasma.
12
ACS Paragon Plus Environment

Page 13 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. Structureꢀactivity and structureꢀproperty relationships for analogs of 6 and their
corresponding metabolites.
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
1
2
Cpd
6
R
R
X
IC50 ROCK2
nM)
IC50 ROCK2,
EC50 MLCꢀPP
t
1/2 Plasma
Expected
Metabolite
6M
a
b
(
200µM ATP (nM)
125 ± 18
NT
(nM)
(min)
Me
iꢀPr
H
H
C=ONH
C=ONH
C=ONH
C=ONH
C=ONH
C=ONH
C=ONH
C=ONH
NHC(=O)
NHC(=O)
NHC(=O)
NHC(=O)
C=ONH
C=ONH
C=ONH
C=ONH
NHC(=O)
NHC(=O)
C=ONH
1.1 ± 0.5
7.0 ± 0.6
27 ± 12
2.6 ± 0.5
1.6 ± 0.6
2.1 ± 0.5
5.5 ± 0.6
3.3 ± 0.5
<1
79
150
180
77
<5
1
0
1
34
<5
6M
1
nꢀBut
secꢀBut
Me
H
NT
6M
12
H
NT
16
6M
1
3
4
4ꢀF
4ꢀF
4ꢀF
4ꢀF
H
46 ± 12
NT
26
18
13M
13M
13M
13M
17M
17M
19M
19M
21M
21M
NA
1
nꢀPr
75
<5
15
(S) secꢀBut
(R) secꢀBut
Me
72 ± 17
NT
50
39
16
17
18
19
79
46
2.1± 0.8
NT
17
64
nꢀPr
H
1.8 ± 0.5
<1
15
<5
Me
4ꢀF
4ꢀF
2ꢀF
2ꢀF
H
1.4 ± 0.9
2.2 ± 1.4
8.8 ± 2.1
41 ± 14
NT
6.8
26
53
20
(CH
2
)
2
OMe
<1
<5
2
1
2
Me
1.2 ± 0.5
1.4 ± 0.5
6.6 ± 0.6
2.0 ± 0.6
<1
11
<5
2
cPen
H
43
9
6M
5900
3900
2000
8900
>120
>120
>120
>120
>120
1
3M
17M
H
4ꢀF
H
NT
NA
H
NT
NA
1
2
9M
1M
H
4ꢀF
2ꢀF
<1
NT
NA
c
H
3.5 ± 0.7
NT
>10000
NA
a
b
c
:
Mean of at least 3 exp. : Mean of at least 2 exp. Highest tested concentration. NA: Not
applicable. NT: Not tested.
Replacement of a methyl ester by nꢀpropyl ester resulted in substantially faster degradation (13
vs. 14 or 17 vs. 18). Esters synthesized from secondary alcohols were of particular interest for
modulating stability. Indeed, the isopropyl ester 10 initially led to a significant increase of
plasma stability. However, further enlargement of this group towards secꢀbutyl ester (12)
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
resulted in faster degradation. In 2ꢀF or 4ꢀF series, secꢀbutyl esters (15, 16) or cycloalkyl esters
(22) were also hydrolyzed in plasma. Reversal of the amide linker resulted in increased activity,
but also in increased stability (6 vs. 17; 13 vs. 19). The latter could again be easily varied by
changing the ester chain (18, 20). Fluorine atoms allowed further modulation of activity and
plasma stability (e.g. 6 vs. 13 or 21), depending on their position and on the amide linker that
was used.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Ultimately, this optimization exercise allowed the identification of derivatives displaying
improved cellular activity and covering a range of plasma stabilities. As examples, 17, 19 and 21
all displayed EC values under 20 nM, while their plasma t ranged from >60 min to <5 min.
5
0
1/2
At this point, we sought to experimentally verify the binding mode of soft ROCK inhibitors in
ROCK, so to provide further rationale for the observed SAR and additional insights regarding
compound optimization. Several soft ROCK inhibitors, including 8 and 13, were selected for
crystal soaking, based on their onꢀtarget potency, cellꢀbased activity and sufficient solubility.
The best ROCK2 coꢀcrystals were obtained with compound 13, allowing XRD data collection at
a resolution of 2.93 Å. In the resulting structure (Figure 4), ROCK2 adopts a bilobal architecture
characteristic of other members of the eukaryotic protein kinase family. As seen in other crystal
22
structures of ROCK1 and ROCK2, the protein crystallized as a dimer, with four monomers in
the asymmetric unit. The amino acid residues forming the active (ATPꢀbinding) site and the
ligand were well defined in the electron density map. The interpreted Xꢀray diffraction data
shows a clear binding mode, as well as the orientation and conformation of the bound ligand. In
line with our expectations, the part of the inhibitor that is similar to Yꢀ27632 occupies the ATPꢀ
binding site in a fashion that is almost identical to this reference compound. The additional
structural elements of 13 however form numerous interactions which contribute to the increased
14
ACS Paragon Plus Environment

Page 15 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
onꢀtarget activity. The biphenyl structure that was already present in previous soft ROCK
inhibitors sits under the Pꢀloop (Glycineꢀrich loop) and forms a πꢀcation interaction with Lys121.
2
2
Comparison to the crystal structure of a ROCK1ꢀYꢀ27632 complex (2ETR), reveals that the
increased bulkiness of 13 did not result in a repulsion of the Pꢀloop, that instead moved in
direction of the inhibitor, forming additional contacts. As intended, the additional benzamide
structure engages the side chains of Phe103 and Phe136 in πꢀstacking interactions. Based on a
distance of 3.5 Å between donor and acceptor, a total of six hydrogen bonds are observed
between ROCK2 and 13. In addition to the hydrogen bonds already observed in the ROCK1ꢀYꢀ
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
22
2
7632 complex, the newly introduced carbamoyl function forms two hydrogen bonds with
residues Phe103 and Asp232, as suggested by molecular modeling studies. The last hydrogen
bond involves the ester carbonyl and the side chain of Thr235, in three of the four ROCK2
monomers present in the unit cell. However, the benefits of such an interaction are probably
lessened by electrostatic repulsion with the neighboring side chain of Asp214. This observation
might explain why esters on position 4 display stronger onꢀtarget activity than on position 3.
While the methyl ester of 13 appears to take a slightly different conformation depending on the
monomers, position of the terminal methyl group is not clearly defined. In all cases, this ester
nonetheless sits in a region that is solventꢀexposed, explaining why further addition of
hydrophobic side chains does not result in improved activity.
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Crystal structure of a complex between ROCK2 and 13. A) Comparison of the
ROCK2ꢀ13 (protein in green, 13 in cyan) and ROCK1ꢀYꢀ27632 (protein in grey, compound in
magenta) complexes. Movement of the Pꢀloop above the arylamide structure is visible. B) Focus
on the newly added structural elements, and on the additional hydrogen bonds formed by 13.
Optimization of 9 focused on modulation of the stability profile via introduction of secondary
alcohols, since the starting compound already possessed strong onꢀtarget activity and functional
activity (Table 4). In contrast to the analogs of 6, several derivatives with larger alcohol chains
such as 24ꢀ26 or 29, retained subꢀnanomolar IC50 values, even when facing a higher ATP
concentration. Activity in the MLCꢀPP assay was however reduced by the presence of polar
functions, as demonstrated by 30 (Nꢀmethylꢀpiperidinꢀ4ꢀyl). As for carboxylic acid metabolites,
this weaker EC50 value likely results from lower membrane permeability. As observed with
analogs of 6, plasma stability was initially increased for esters from secondary alcohols (23), but
decreased once cyclic alcohols were used (26ꢀ30). Interestingly, we noted a significant difference
16
ACS Paragon Plus Environment

Page 17 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of plasma stability between 24 and 25, the (S) and (R) enantiomers of the secꢀbutyl ester, while
chirality did not affect the hydrolysis of 15 and 16.
Table 4. Structureꢀactivity relationships for analogs of 9 and their common metabolite.
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
Cp
d
R
IC50
IC ROCK2,
EC MLCꢀ
t1/2 Plasma
Expected
50
50
a
b
ROCK2
200µM ATP
(nM)
11 ± 2
<1
PP (nM)
(min)
Metabolite
(
nM)
<1
9
Me
iꢀPr
31
37
23
31
79
24
26
36
<5
94
17
>120
7
9M
9M
9M
9M
9M
9M
9M
9M
23
24
<1
(S) secꢀBut
(R) secꢀBut
THPꢀ4ꢀyl
cꢀBut
<1
<1
25
26
27
28
<1
<1
<1
<1
1.3 ± 0.5
<1
1.4 ± 0.6
1.3 ± 0.6
<1
6
cꢀPen
7
29
THPꢀ3ꢀyl
NꢀMeꢀpiperidinꢀ4ꢀ
yl
<1
6
3
0
<1
4.6± 1.9
NT
330
<5
9M
9M
H
1.8 ± 0.6
3600
>120
NA
a
b
:
Mean of at least 3 exp. : Mean of at least 2 exp. NA: Not applicable. NT: Not tested.
In view of the observed differences in plasma stability, we sought to characterize the enzyme(s)
responsible for hydrolysis of the soft ROCK inhibitors in systemic circulation. Therefore, we
assessed the effect of various esterase inhibitors on hydrolysis rate (Figure 5). In a first step 13
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
was incubated in human whole blood in presence of DTNB (arylesterase inhibitor), EDTA
(
PON1 inhibitor), PMSF (generic serine protease / serine esterase inhibitor), BNPP
carboxylesterase inhibitor) or tacrine (cholinesterase inhibitor). BNPP had no effect on the
(
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
hydrolysis rate of 13, consistent with literature indicating little or no carboxylesterase activity in
2
7
human plasma. DTNB and EDTA did not significantly alter the hydrolysis rate of 13 either.
This finding is in line with the absence of aryl ester or lactone, which represent the preferred
2
8
substrates of those enzymes, in 13. In opposition, addition of PMSF or tacrine both prevented
any significant degradation of 13 over the course of the experiment (120 min), suggesting that
cholinesterase activity is responsible for the hydrolysis of 13. In a second step, the effect of
tacrine on the stability of additional derivatives of 6 and 9 spanning a range of hydrolysis rates
(t1/2 in human plasma from <5 min to 53 min) was assessed by measuring the residual amount of
compound after 60 min incubation. We observed that addition of 5 µM tacrine was sufficient to
completely prevent hydrolysis of all but the most labile derivatives. The minor degradation that
is still found for 20 and 26 was attributed to the residual cholinesterase activity still present at
this inhibitor concentration, as the limited time of the experiment does not allow significant
chemical hydrolysis.
Optimization of 8 was initially limited to a small number of derivatives, in an attempt to
determine if plasma stability could be modulated. While they maintained onꢀtarget (ROCK2),
and functional (MLCꢀPP) activity, introduction of nꢀpropyl ester (31) or reversal of the amide
linker (32) did not result in faster degradation in human plasma (Table 5).
18
ACS Paragon Plus Environment

Page 19 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Control
DTNB
Tacrine
BNPP
EDTA
PMSF
1
00
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
7
5
50
2
5
0
0
40
80
120
Time (min)
b
1
03
1
01
9
8
100
1
00
88
8
2
Control
ꢀM Tacrine
7
5
2
5
0
5
0
5
42
2
3
1
3
2
0
0
13
19
20
24
26
28
Compound
Figure 5. a) Effect of five esterase inhibitors on the hydrolysis rate of 13 in human whole blood.
b) Effect of tacrine on the hydrolysis rate of selected soft ROCK inhibitors in human plasma.
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 5. Structureꢀactivity relationships for analogs of 8 and the corresponding metabolites.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Cpd
R
X
IC50
IC50 ROCK2,
EC50 MLCꢀ
t
1/2 Plasma
Expected
a
b
ROCK2
200µM ATP (nM)
PP (nM)
(min)
Metabolite
(nM)
8
Me
nꢀPr
Me
H
C=ONH
C=ONH
NHC(=O)
C=ONH
C=ONH
<1
4.1 ± 1.0
NT
3.0
8.3
>120
>120
>120
NT
8M
8M
3
1
<1
32
<1
3.1 ± 1.1
NT
5.2
32M
NA
8M
3.3 ± 0.9
1
6600
3600
3
2M
H
NT
NT
NA
a
b
:
Mean of at least 3 exp. : Mean of at least 2 exp. NA: Not applicable. NT: Not tested.
Being faced with the high stability of those derivatives in plasma, we reasoned their conversion
to the carboxylic acid metabolite might still occur in liver. Indeed, liver tissue has high esterase
activity, resulting from the presence of carboxylesterase 1 (CES1, also known as liver
carboxylesterase) and to a lesser extent carboxylesterase 2 (CES2), which are both absent from
human plasma. In order to validate that hypothesis, the stability of 8 in presence of human or
rabbit (New Zealand White, NZW) hepatocytes was first investigated. As expected, 8 was
rapidly degraded (t =12 min for both human and rabbit hepatocytes) under the tested
1
/2
conditions, with only limited compound (2.4% ± 1.3% in human hepatocytes, 0.4% ± 0.3% in
rabbit hepatocytes) remaining after 120 min of incubation. In the meantime, the formation of an
equivalent amount of 8M (corresponding to 99.3 ± 0.8 % of the parent compound in human
hepatocytes) was observed, confirming that 8 is degraded through ester hydrolysis.
20
ACS Paragon Plus Environment

Page 21 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 6. Pharmacokinetic data for 8
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Parameter
Study
Value
t1/2 human plasma
t1/2 rabbit plasma
Plasma incub.
Plasma incub.
Hepatocyte incub.
Hepatocyte incub.
> 120 min
> 120 min
t1/2 vs. human hepatocytes
t1/2 vs. rabbit hepatocytes
12 min
12 min
Plasma Cmax for 8 (iv, rabbit, 0.1 mg/kg)
Plasma Cmax for 8M
Rabbit PK
Rabbit PK
Rabbit PK
Rabbit PK
803 nM (1 min)
255 nM (10 min)
5090 nM min
21500 nM min
AUClast for 8 (iv, rabbit, 0.1 mg/kg)
AUClast for 8M
As a next step, 8 was submitted to a pharmacokinetic study in rabbit. Firstly, we confirmed
that 8 was essentially as stable in rabbit plasma as in human plasma (t1/2 > 120 min). NZW
rabbits received 8 intravenously (0.1 mg/kg) and plasma samples were taken at time points
ranging from 1 to 360 minutes after administration. The resulting pharmacokinetic parameters
are summarized in table 6. The concentration of 8 was highest for the 1 min time point (803 ±
2
18 nM) and was markedly reduced at the 5 min and 10 min time points (176 ± 18 nM and 57 ±
1
nM respectively). A slower, but still rapid decay (t1/2 = 22 min) was subsequently observed. 8
remained above level of quantitation (8 nM) until the 30ꢀ60 min time points. Formation of 8M
was already apparent at the 1 min time point, and plasma concentration of 8M was highest (255
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
nM ± 34 nM) after 10 min. 8M remained above level of quantitation (35 nM) until the 120 min
time point and was still detectable after 240 min. In view of the rapid elimination of 8 and of the
concomitant formation of 8M, we considered that an equilibrium in plasma / tissue distribution
of 8 could not be reached in this experiment. The distribution volume and clearance parameters
were therefore not calculated. From a soft drug perspective, the collected data nonetheless
demonstrate in vivo formation of the expected carboxylic acid metabolite and higher (4.2ꢀfold)
exposure for 8M than for 8. Thus, we concluded that derivatives displaying the ester moiety on
position para, such as 8, can qualify as soft drugs, in spite of their apparent stability in plasma.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The selectivity of 8 was investigated against a panel of 335 kinases (Reaction Biology Corp).
Compound was tested at 10 nM and 100 nM, against an ATP concentration of 10 µM, with the
highest levels of inhibition found for ROCK1 and ROCK2. A selectivity score was attributed to
8
by calculating the fraction of “offꢀtarget” kinases that are hit (>50% inhibition) at the tested
2
9
concentration, in a fashion similar to Karaman et al. (Ambit selectivity score). The S
for
(100 nM)
8
was 0.045. This score should be put in perspective with the subꢀnanomolar IC of 8 against
50
ROCK, and reflects selectivity of at least 100ꢀfold vs. other kinases. Most of the offꢀtarget
activities identified at 100 nM were located within the AGC kinase family, with the strongest
inhibition being found for novel PKCs (PKCε, PKCη, PKCδ and PKCθ), MSK1, PRK1, PRK2
and PRKX. Interestingly, PRK2 mediates some of the cellular effects of the reference ROCK
3
0
31
inhibitors Yꢀ27632 and Fasudil, suggesting that this “offꢀtarget” activity might actually be
beneficial. The S(10 nM) score of 8 was 0.006. This score is more in line with the criterion of a 10ꢀ
fold ratio between K values that was initially defined by Karaman et al., and positions 8 as a
d
nicely selective ROCK inhibitor. PKCη and PRK2 were, besides ROCKs, the only kinases
inhibited (> 50%) targeted in presence of 10 nM of 8.When tested in the same panel, 8M
22
ACS Paragon Plus Environment

Page 23 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
displayed a S(100 nM) score of 0.012, and did not inhibit any kinase that wasn’t previously affected
by 8. The offꢀtarget activities of 8M were in general weaker, with the exception of MSK1 & 2.
Results from the selectivity panel are illustrated in Figure 6.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Kinase selectivity data for 8 (left) and 8M (right). Both compounds were tested at a
concentration of 100 nM, against 10 µM ATP. Red dots denote >90% inhibition of kinase
activity, while orange dots indicate 50ꢀ90% inhibition. Figure drawn with Kinase Mapper
(Reaction Biology Corporation).
In addition to MLC phosphorylation, selected soft ROCK inhibitors were evaluated in further
2
cellꢀbased assays. Rho and ROCK play key roles in cell migration and motility. ROCK
inhibition by Yꢀ27632 was shown to increase migration of several cell lines, including T3T
3
0,32
fibroblast, human tenon fibroblasts (HTF) or epithelial (T84, Cacoꢀ2) cell lines.
Among
others, we chose to follow migration of human dermal fibroblast (HDFꢀ1) cells as a measure for
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
cellular functional activity of the soft ROCK inhibitors, in an inꢀhouse implemented wound assay
3
3
previously used on vascular endothelial cell lines. In a first step, cytotoxic effects on the HDFꢀ
cell line were assessed using both a viability assay (WSTꢀ1) and apoptosis assay (Caspaseꢀ3/7
induction). The cytotoxic concentration 50 (CC ), corresponding to a 50% decrease in viability
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
50
after 24h, was 46 µM for 8, 45µM for 13 and 29 µM for 28. Such values are in line with the
induction of apoptosis in the HDFꢀ1 cells by those compounds. Indeed, 8 and 13 both showed a
significant induction of caspase 3/7 activity as of 60 µM (24 h treatment), while induction of
apoptosis could be observed as of 40 µM for 28 (significant at 100 µM). In a second step,
migration of HDFꢀ1 was followed for 8 h, in presence of nonꢀcytotoxic concentrations of the
ROCK inhibitors. 8, 13 and 28 all increased cell migration of the HDFꢀ1 in a concentration
dependent manner, as compared to the equivalent DMSO controls after wounding of the
monolayer (Figure 7). 8 was most efficient, since significantly increased migration was observed
as of 0.001 µM (Figure 7A). 13 showed an induction of migration as of 0.01 µM whereas 28
only significantly stimulated migration at 0.1 µM (Figures 7C & 7D). The corresponding
metabolites did not display significant effects on cell migration until at least 1 µM (illustrated for
8M, Figure 7), in line with their lower activity in MLCꢀPP assay. Similar data could be collected
over additional cell lines, prompting the evaluation of compounds in in vivo preclinical models
3
4
35
of corneal wound healing or AMD. Those results are discussed in separate dedicated papers.
24
ACS Paragon Plus Environment

Page 25 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Cpd 8
Cpd 8M
1
00
80
100
80
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
6
0
60
40
40
vehicle control
.001ꢀM ****
0.01ꢀM *
0
vehicle control
2
0
0
20
0
0.01ꢀM
0.1ꢀM
1ꢀM
0
1
.1ꢀM ***
ꢀM ***
0
2
4
6
8
0
2
4
6
8
hours after wounding
hours after wounding
Cpd 13
Cpd 28
1
00
100
8
6
4
2
0
0
0
0
0
80
60
40
20
0
vehicle control
0
.001ꢀM
vehicle control
0.01ꢀM
0.1ꢀM *
0.01ꢀM *
0
1
.1ꢀM ****
ꢀM ****
1ꢀM *
0
2
4
6
8
0
2
4
6
8
hours after wounding
hours after wounding
Figure 7. Effect of 8, 8M, 13 and 28 in a HDFꢀ1 migration assay. All three parent compounds
show a doseꢀdependent effect on migration, which is significant with respect to vehicle as of
0
.001 µM (8), 0.01 µM (13) or 0.1 µM (28). In contrast, 8M does not differ from the vehicle
control at 0.01, 0.1 or 1 µM.
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Discussion:
For several years, ROCK has been identified as a potentially attractive target for the treatment of
3
multiple indications including cardiovascular diseases, inflammatory and autoimmune
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5ꢀ7
8
diseases, and eye diseases. Unfortunately, the potent effect of ROCK inhibitors on blood
pressure has hampered their development for indications where systemic exposure is either
required or likely. Consequently, development of ROCK inhibitors was often focused on topical
applications resulting in low systemic exposure, with the field of eye diseases receiving
particular attention. Even so, the side effects of locally applied ROCK inhibitors (e.g.
conjunctival hyperemia in case of administration as eye drops) remain a matter of debate. In this
context, soft drugs represent an attractive approach for minimizing ROCKꢀassociated side
effects, by reducing systemic exposure to ROCK inhibitors.
While our previous work started with the introduction of an ester moiety onto an existing
scaffold, we here sought to optimize the onꢀtarget and functional activity of our ROCK inhibitor
series. Our efforts were largely successful, as the resulting compounds stand among the strongest
ROCK inhibitors reported to date, retaining low nanomolar IC values even when facing high
5
0
ATP concentrations. Moreover, activity in a cellꢀbased MLC phosphorylation assay was
demonstrated, and an important difference of functional activity was observed between the
parent compounds and their carboxylic acid metabolites, thereby validating ester hydrolysis as a
viable strategy for the design of soft ROCK inhibitors. In spite of those achievements, a second
look at our results reveals that the design and optimization of soft ROCK inhibitors (and, likely,
soft kinase inhibitors in general) cannot be summarized into the mere addition of an ester
function to an existing scaffold. To start with, it is likely that the difference of functional activity
observed in the MLCꢀPP assay results at least in part from a difference of cell permeability
26
ACS Paragon Plus Environment

Page 27 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
between the parent compound and its carboxylic acid metabolite. Playing upon differences in
physicochemical properties to modulate compound efficacy is a wellꢀknown strategy in
medicinal chemistry, which has been used extensively – but in the opposite direction – for the
design of prodrugs. This strategy appears well suited to the current compound series, which
displayed from the start acceptable ꢀbut not excellentꢀ permeability. However, it can also be seen
as a doubleꢀedged sword, as we observed that even introduction of a moderately polar functional
group on the parent compound could result in weaker EC50 values in the MLCꢀPP assay.
Optimization of soft ROCK inhibitors relying on differences of permeability should therefore be
done carefully. Another interesting observation is the diversity of profiles resulting from the
addition of ester groups to parent molecules. Large differences in hydrolysis rates were indeed
observed between compounds, depending not only on the nature of the ester group, but also on
its position. This situation is easily understandable, as hydrolysis of the ester group does not
result from chemical instability, but rather from metabolism by esterases, as demonstrated by
esterase inhibition experiments. Consequently, recognition as an esterase substrate is a
fundamental step in the inactivation of the soft drug. It is likely that some ester functions
allowing a high inactivation rate when put on one position on one scaffold are still unable to
provide a sufficient rate when placed on another position or scaffold. The striking difference of
stability of compound 8 in plasma or in presence of hepatocytes also poses the question of
stability in different target organs. Esterases are indeed expressed in different levels, depending
on organ, or even tissue. For instance, differences in the expression of carboxylesterase 1, which
preferentially hydrolyzes esters from small alcohol groups and carboxylesterase 2, which favors
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
3
6
37
hydrolysis of esters from bulkier alcohol groups, are found between liver, lung or intestine.
Such differences in plasma and organ stability might ultimately influence the potential
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 60
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
indications of a soft ROCK inhibitor. For instance, compounds with relatively fast degradation
rate in plasma, such as 13 or 18, appear more adapted for indications where there is a significant
risk of high systemic exposure following administration. Oral administration for treatment of
intestinal diseases and administration through inhalation for treatment of lung diseases represent
potential examples of such applications. At the other side of the spectrum, compound 8 appears
quite stable in plasma, and might essentially be degraded in liver, as indicated by hepatocyte and
PK data. This compound could however be of particular interest for sustained release strategies,
where the limited amount of drug that can be encapsulated into a matrix imposes the use of a
very potent compound. The soft drug approach implemented within this compound could then be
used to minimize the effects of an initial burst release occurring after administration, which
could otherwise still lead to an unacceptable systemic concentration.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Conclusion
We have here presented the design, synthesis and evaluation of a novel series of soft ROCK
inhibitors; more precisely benzoic or phenacetic ester analogues of 4ꢀ(aminomethyl)ꢀNꢀ(pyridinꢀ
4ꢀyl)ꢀ3[3ꢀ(phenylcarbamoyl) phenyl] benzamide (2). The resulting compounds display low
nanomolar or subꢀnanomolar onꢀtarget activity, as well as functional activity in cellꢀbased assays
including MLC phosphorylation and wound healing assay. Upon entering systemic circulation,
these inhibitors should be rapidly inactivated by blood or liver esterases, thereby reducing the
risk of systemic ROCKꢀassociated side effects. The stability profile of those compounds can be
modulated by altering the nature and position of the ester moiety they contain, suggesting that
they might cover a range of potential indications for which local administration is applicable.
Further in vivo evaluation of compounds resulting from this work will be presented in due time.
28
ACS Paragon Plus Environment

Page 29 of 60
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Experimental section
Molecular modeling
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
2
D compound structures were drawn in Instant JChem v6.3.3. The protonation state of the
molecules was adjusted based on predicted pKa values, and the resulting structures were
exported as SMILES strings. Initial 3D structures were generated with openbabel, using the
obgen and obminimize tools and the mmff94 forcefield. 3D structures were manually inspected
before further modeling. Ligand files for molecular docking were then produced with
AutodockTools 1.5.4.
The receptor model used for docking was initially based on the ROCK2 crystal structure 2ETR.
Receptor files were later updated once the crystal structure of compound 10 with ROCK2 had
been obtained. Whenever necessary, missing residues or loops were rebuilt with Yasara
Structure, using the YASARA2 forcefield and the available homology modeling, model
refinement and template forcing routines. The docking grids were generated with AutodockTools
1
.5.4 and Autogrid4. Default parameters were used, except for grid step size, which was reduced
from 0.33 A to 0.25 A in order to obtain a more accurate docking grid.
Molecular docking was carried out using Autodock 4.2. Default parameters were mostly used.
However, we observed that docking quality was increased by adapting some of the parameters
based on the number of rotatable bonds (nrot) in the ligand with the following changes. The
genetic algorithm population size was 150 (default) for nrot <9 or 150 + (25 x (nrot ꢀ 8)) for nrot
≥
9. The maximal number of energy evaluations per run was 1000000 + (nrot x 250000) for nrot
29
ACS Paragon Plus Environment