Design, synthesis and biological characterization of selective LIMK inhibitors

Article abstract of DOI:10.1016/j.bmcl.2015.07.009

Abstract Inhibitors of LIM kinases are considered of interest for several indications, including elevated intraocular pressure (IOP), cancer, or infection by HIV-1. LX-7101 (Lexicon Pharmaceuticals) was advanced to Phase-I clinical trials as an IOP-lowering agent for treatment of glaucoma. We here discuss the design, synthesis and evaluation of LIMK inhibitors based on a pyrrolopyrimidine scaffold, which represent close analogs of LX-7101. Exploration of structure-activity relationships revealed that many of such compounds, including LX-7101, cause potent inhibition of LIMK1 and LIMK2, and also ROCK2 and PKA. Molecular variations around the various structural elements of LX-7101 were attempted. Substitution on position 6 of the pyrrolopyrimidine scaffold led to the identification of LX-7101 analogs displaying good selectivity versus ROCK, PKA and Akt.

Full text of DOI:10.1016/j.bmcl.2015.07.009

Bioorganic & Medicinal Chemistry Letters xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design, synthesis and biological characterization of selective LIMK
inhibitors
a
a
a
a
a
Sandro Boland a, , Arnaud Bourin , Jo Alen , Jacques Geraets , Pieter Schroeders , Karolien Castermans ,
Nele Kindt a, Nicki Boumans a, Laura Panitti a, Jessica Vanormelingen a, Silke Fransen a,
Sarah Van de Velde b, Olivier Defert a
⇑
a Amakem Therapeutics N.V., Agoralaan Abis, 3590 Diepenbeek, Belgium
b Laboratory of Ophthalmology, KU Leuven, 3000 Leuven, Belgium
a r t i c l e i n f o
a b s t r a c t
Article history:
Inhibitors of LIM kinases are considered of interest for several indications, including elevated intraocular
pressure (IOP), cancer, or infection by HIV-1. LX-7101 (Lexicon Pharmaceuticals) was advanced to Phase-I
clinical trials as an IOP-lowering agent for treatment of glaucoma. We here discuss the design, synthesis
and evaluation of LIMK inhibitors based on a pyrrolopyrimidine scaffold, which represent close analogs of
LX-7101. Exploration of structure–activity relationships revealed that many of such compounds, includ-
ing LX-7101, cause potent inhibition of LIMK1 and LIMK2, and also ROCK2 and PKA. Molecular variations
around the various structural elements of LX-7101 were attempted. Substitution on position 6 of the
pyrrolopyrimidine scaffold led to the identification of LX-7101 analogs displaying good selectivity versus
ROCK, PKA and Akt.
Received 22 May 2015
Revised 30 June 2015
Accepted 2 July 2015
Available online xxxx
Keywords:
Kinase inhibitor
LIMK
ROCK
SAR
Ó 2015 Published by Elsevier Ltd.
LX-7101
The LIM kinase (Lin-11/Isl-1/Mec-3 domain-containing protein
kinase) family consists of two members: LIM kinase 1 (LIMK1)1
and LIM kinase 2 (LIMK2).2 LIMK1 and LIMK2 are closely related
proteins containing two N-terminal LIM domains, a PDZ domain
with two nuclear export signals and one C-terminal kinase domain.
The LIM domains are protein-binding domains that are frequently
found among cytosolic proteins interacting with the actin
cytoskeleton.3 Consistent with the presence of these LIM domains,
both LIM kinases have been shown to influence the architecture of
the actin cytoskeleton by regulating the activity of the cofilin
family proteins cofilin1, cofilin2 and destrin.4 Regulation of actin
reorganization and contractility allows cells to control their shape,
movement, division and secretion. An important group of effectors
regulating this process is composed of the small GTPases from the
Rho Family. Active Rho GTPases will allow signal transmission
towards Rho-associated, coiled-coil containing protein kinases
(ROCK1 and ROCK2) and/or p21-activated protein kinases (PAK,
PAK2 and PAK4).5 Both PAKs and ROCKs can phosphorylate and
activate LIMKs, which will in turn influence the architecture of
the actin cytoskeleton via their action on cofilin1, cofilin2 and des-
trin. Given the position of LIMK in biological cascades, LIMK
inhibitors have been considered of interest for several indications,
including cancer,5 elevated intraocular pressure (IOP) and
glaucoma,6,7 pulmonary hypertension.8 Recent research has also
highlighted a potential role of LIMK in viral diseases.9 Indeed, inhi-
bition of actin polymerization affects human immunodeficiency
virus type 1 (HIV-1) binding and entry into host cells. It has also
been shown that LIMK1 is activated by HIV-1 in order to initiate
viral infection.10 Depletion of ROCK1 or inhibition of LIMK1 by a
synthetic peptide reduced the release or retroviral particles, as well
as cell–cell transmission events.11
In spite of those multiple potential indications, the number of
reported LIMK inhibitor series remains modest, in comparison to
more common kinase targets. The most potent examples of LIMK
inhibitors reported to date are based on pyrrolopyrimidines6,7,12,13
or 2-aminothiazole scaffolds.14–17 A series of non-competitive
(Type III) LIMK inhibitors was also recently reported.18 From a clin-
ical perspective, LX-71017 (Lexicon Pharmaceuticals, Fig. 1) is to
date the only LIMK inhibitor that reached clinical trials. LX-7101
was evaluated in a Phase-I trial as an IOP-lowering agent for treat-
ment of glaucoma. This compound however retains non-negligible
activity against ROCKs.6 As ROCK inhibitors themselves represent
IOP-lowering agents,19 it is therefore unclear whether the observed
effects of LX-7101 result from activity against LIMKs, ROCKs, or
both targets. Selective LIMK inhibitors of sufficient stability and
⇑
Corresponding author.
E-mail address: sandro.boland@amakem.com (S. Boland).
http://dx.doi.org/10.1016/j.bmcl.2015.07.009
0960-894X/Ó 2015 Published by Elsevier Ltd.
Please cite this article in press as: Boland, S.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.07.009

2
S. Boland et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
Boc-protection of the 4-aminomethyl piperidine was required
prior to the final coupling reaction.
O
O
H2N
N
H
O
N
Variations around the amine and carbamate moieties are sum-
marized in Table 1. Compounds were evaluated against LIMK1 and
2, ROCK2 and PKA. Compounds generally appeared more potent
against LIMK2 than against LIMK1. This finding is in line with pre-
vious reports regarding LX-7101 and structurally related com-
pounds.6,7 Replacement of the carbamate moiety of LX-7101 by a
methyl ester did not result in significant loss of on-target potency
against LIMK1 or LIMK2, as illustrated by 1. However, both com-
pounds displayed strong activity against ROCK2, and also PKA.
Methylation of the amine function was well tolerated with
respect to LIMK activity (LX-7101 vs 2, or 1 vs 3). However, both
2 and 3 retained significant activity against ROCK2 or PKA
Removal of the methylene between the amine and the piperidine
ring resulted in compounds 4 and 5. Those modifications were
again tolerated by LIMK1, but resulted in a significant loss of
potency versus LIMK2. As a result, 4 or 5 did not display improved
selectivity, in spite of their reduced potency against PKA and
ROCK2. Introduction of short spacers between the phenyl and ester
moieties (6–7) did not eliminate activity against PKA and ROCK2.
Displacement of the methyl ester on position 4 (8) was rapidly
put as second priority, as it reduced potency against LIMK2 while
maintaining (ROCK2) or even increasing (PKA) off-target activities.
Modulation around the alkyl ester (9–12) had some minor impact
on the IC50 values for LIMK, ROCK and PKA, but did not result in
compounds that could be considered selective.
In kinases, modifications of the hinge-binding motif often have
a profound impact on potency and selectivity. In 2004, Aronov and
Murko published a 5-point pharmacophore for kinase frequent hit-
ters, wherein four points involved interaction with the kinase
hinge region.22 Pyrrolopyrimidines represent a very common
kinase scaffold, with over 1100 references in literature. This
includes over 110 references wherein the pyrrolopyrimidine dis-
plays a six-membered, nitrogen-containing heterocycle on position
4 (see Supporting information). We believed this situation illus-
trated, besides a potentially crowded IP space, the overall promis-
cuity of this scaffold. Consequently, we investigated replacement
of the pyrrolopyrimidine as a potential strategy to improve selec-
tivity (Table 2).
N
N
N
N
H
Figure 1. Structure of the LIMK inhibitor LX-7101.
solubility for formulation and administration as eye drops could
reportedly not be identified by Lexicon.7 Selectivity between
LIMK and ROCK can however be achieved, as demonstrated by a
recent series of bis-aryl urea derivatives incorporating the
pyrrolopyrimidine scaffold but displaying strong selectivity versus
ROCK.13 Herein, we report the synthesis and evaluation of new
analogs of LX-7101, alongside with the identification and biological
characterization of a selective LIMK inhibitor. Our synthetic efforts
were focused on modulation of the amine, of the pyrrolopyrim-
idine scaffold, and on replacement of the carbamate moiety of
LX-7101. The latter was attempted with the purpose of optimizing
interactions with residues located around the Glycine-rich loop or
beyond, which are not conserved between LIMKs and ROCKs. Ester
chains were favored in our design, as they left open the possibility
of a soft drug approach for avoiding ROCK-associated side effects,20
in case the ROCK activity observed with LX-7101 could not be
eliminated.
Compounds were prepared in 2–5 steps starting from commer-
cially available piperidine derivatives (Fig. 2). Kinase hinge-binding
motifs were introduced to the piperidine derivatives through SNAr
reaction of aryl chloride which occurred selectively on the piperi-
dine nitrogen atom.21 When the aryl groups were not introduced
at the last step of the synthesis, isopropanol was replaced by
DMSO/water as solvent, providing better solubility of the starting
materials, allowing higher temperature and thereby accelerating
the reaction. Moreover, hydrolysis of the methyl ester occurred
in situ under these conditions, reducing the number of steps for
some of the compounds. Otherwise, the carboxylic acid could be
generated by a saponification reaction using aqueous NaOH. The
amine function was methylated via a reductive amination with
formaldehyde. Amide formation with selected anilines could be
achieved preferably at low temperature in DCM using T3P as
The simple, non-substituted pyrrolopyrimidine 13 was first
made for comparison. While this compound was essentially similar
to 3 with respect to LIMK1, it displayed reduced activity against
activating agent. For the compounds LX-7101,
1 and 27, a
Figure 2. Overview of compound synthesis. (a) ArCl, DIPEA, DMSO/water (1:1), 140 °C, 16 h; (b) ArNH2, T3P, DMAP, DCM, À15 °C, 16 h; (c) DCM/TFA (10:1), rt, 48 h; (d) ArCl,
DIPEA, iPrOH, 120 °C, 48 h; (e) CH2O, NaBH3CN, MeOH, rt, 2 h; (f) NaOH, EtOH/water (1:1), reflux, 16 h; (g) Boc2O, NaOH (1 M)/dioxane (1:2), 0 °C–rt, 2 h.
Please cite this article in press as: Boland, S.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.07.009

S. Boland et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
3
Table 1
Structure–activity relationships for variation around the amine and carbamate moieties of LX-7101
O
R2
R1
N
H
N
N
N
H
N
Compound
R1
R2
IC50 LIMK1 (nM)
IC50 LIMK2 (nM)
IC50 ROCK2 (nM)
IC50 PKA (nM)
LX-7101
1
2
3
4
5
6
7
8
9
10
11
12
CH2NH2
CH2NH2
CH2NMe2
CH2NMe2
NH2
3-OCONMe2
3-CO2Me
3-OCONMe2
3-CO2Me
3-CO2Me
3-CO2Me
3-OCH2CO2Me
3-CH2CO2Me
4-CO2Me
3-CO2n-Pr
3-CO2i-Pr
24
16
19
27
18
37
38
57 12
41
50 13
41 11
50 13
7.6 1.4
3
5
3
5
4
5
6
1.6 0.8
<1
<1
1.5 0.7
7.4 2.6
10
<1
40
8.9 1.6
100 16
63 6.8
7.9 2.0
56 11
7.2 0.7
41 6.2
53 36
137 23
16
4
<1
<1
2.6 0.5
24
7
6
81 18
NMe2
12
<1
<1
4
150 28
5.4 1.3
7.0 1.9
2.7 0.5
5.8 0.8
22 3.2
24 2.6
2.2 0.5
CH2NMe2
CH2NMe2
CH2NMe2
CH2NMe2
CH2NMe2
CH2NMe2
CH2NMe2
8
7.7 2.9
2.1 0.4
2.2 0.9
4.6 1.7
<1
3-CO2sec-Bu
3-CO2CH2C„CH
3
LIMK2 and stronger activity against both ROCK2 and PKA.
However, this effect was dependent on the amine displayed on
the piperidine moiety, as no real difference could be observed
between 4 and 17 in terms of PKA inhibition. The less common
pyrazolopyrimidine and pyrazolopyridine isosters 14 and 15 had
decreased potency against the four tested kinases, and therefore
did not display selectivity towards LIMK. Replacement of the orig-
inal scaffold by a 6-oxo-pyrrolidinopyrimidine resulted in 16,
which was essentially inactive or weakly active against all four
kinases. A series of compounds representing more drastic changes
on the hinge-binding moiety was also investigated. Simplification
of the pyrrolopyrimidine into a pyrimidine resulted in a complete
loss of activity (18), which was not rescued by presence of a
6-amino (19) or 2-amino (20) group. Addition of a carboxamide
was also attempted, by analogy to reported aminothiazoles.14
Compound 21 displayed some degree of selectivity for LIMK, but
displayed too little on-target potency to be considered useful.
As we could identify no suitable replacement for the pyrrolopy-
rimidine structure, we investigated the effect of substitutions on
this scaffold (Table 3). Small substitutions on position 5 such as
cyano or fluoro did not result in more selective compounds (23
or 24 vs 13). However, the mere addition of a methyl group on
position 6 resulted in 25, which displayed encouraging selectivity
against both ROCK2 and PKA. This newfound selectivity was fur-
ther increased by combining methyl groups in positions 5 and 6
(26) and could be fully transferred to LX-7101 (27 vs LX-7101).
This effect of methyl groups on selectivity appeared of special
interest to us. Indeed, introduction of a 6-Me was previously
reported as being tolerated in structurally related LIMK2 inhibitors,
but the corresponding compound was not evaluated against
ROCK.6 The 5,6-dimethylpyrrolopyrimidine fragment only appears
in a small fraction of the references describing pyrrolopyrimidine
derivatives as kinase inhibitors and was not reported among ana-
logs of LX-7101.6,7,12 However, the 6-methyl and 5,6-dimethyl sub-
stitution patterns were previously reported to favor selectivity for
Akt (PKB) over PKA for structurally related pyrrolopyrimidines.23
This suggested that such substitution patterns might still allow,
or even favor activity of the compounds against Akt and prompted
us to evaluate the activity of LX-7101, 26 and 27 against Akt1.
Interestingly, LX-7101 displayed potent inhibition of Akt1 under
the tested conditions (IC50 < 1 nM), which to our knowledge was
previously unreported. Meanwhile, neither 26 nor 27 displayed
significant inhibition of Akt1 under the same conditions
(IC50 > 1000 nM for both compounds).
Docking simulations provide
a likely explanation for the
observed selectivity of 27. As expected, the preferred binding mode
(Fig. 3) positions 27 in the ATP-binding site of LIMK2, with the
pyrrolopyrimidine moiety acting as kinase hinge binder, while
the phenyl group occupies a vacant space under the Glycine-rich
loop. However, its orientation is opposite to the one seen for struc-
turally related compounds in PKA or PKB,21 with position 6 facing
the solvent-exposed front pocket, instead of the back pocket and
gatekeeper residue. This leaves position 6 freely available for sub-
stitution in LIMK, while little or no space is available in the
PKA/PKB orientation. The orientation proposed in LIMK is unlikely
to be allowed in PKA, PKB or ROCK, as access to the front pocket is
prevented by a C-terminal loop that is typical of AGC kinases. With
respect to other interactions, the pyrrolopyrimidine moiety forms
two classical hydrogen bonds with the hinge region (I408, back-
bone). Other polar contacts involve the carbamate moiety (F341,
F342, backbone NH), the amine moiety (N456, side chain, D469,
side chain) and the amide moiety (K360, side chain). Refinement
of the complex structure proposed through docking also suggests
the possibility of a water-mediated hydrogen bond with the
gatekeeper residue T405. Interestingly, the piperidine moiety is
markedly rotated with respect to the plane defined by the
pyrrolopyrimidine structure, a conformation that is likely favored
by the 5-Me substituent. Several of our findings are in line with
those recently reported for bis-aryl urea LIMK inhibitors.13
However, some diverging observations can be made with respect
to the effect of methyl groups. We here found that a 5-Me substi-
tution on the pyrrolopyrimidine had little or no effect on LIMK1 (3
vs 13, 4 vs 17), but could affect selectivity versus ROCK2 and PKA,
depending on the amine present on the piperidinyl moiety. On the
other hand the effect of a 6-Me substitution on the pyrrolopyrim-
idine appeared less clear-cut than in bis-aryl urea series, as full
selectivity against ROCK2 could only be achieved through a
5,6-dimethyl substitution pattern. Those differences in SAR
might result from diverging conformational constraints between
the two series. In particular, the presence of a 5-Me group should
Please cite this article in press as: Boland, S.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.07.009

4
S. Boland et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
Table 2
Structure–activity relationships for variation around the kinase scaffold of LX-7101
O
R1
O
N
H
O
N
Ar
Compound
R1
Ar
IC50 LIMK1 (nM)
IC50 LIMK2 (nM)
1.5 0.7
IC50 ROCK2 (nM)
8.9 1.6
IC50 PKA (nM)
*
N
3
CH2NMe2
27
5
24
6
N
H
N
*
N
N
13
14
15
16
CH2NMe2
CH2NMe2
CH2NMe2
CH2NMe2
23 16
13 1.8
1.8 0.5
62 25
1.1 0.2
5.1 0.9
3.8 0.6
1130 360
N
H
N
*
282 38
65 35
193 16
274 66
1860 330
N
N
H
N
*
11
2
N
N
N
H
*
N
N
1000 350
>10,000a
O
N
N
*
H
17
18
19
20
NH2
NH2
NH2
NH2
31
7
9.6 3.2
67 7.5
>1000a
69 11
>1000a
>1000
N
H
N
*
>1000a
>10,000a
4560 670
>10,000a
N
N
N
*
945 275
>1000a
2060 530
416 70
N
NH2
*
>1000a
N
H2N
N
*
O
N
21
22
NH2
NH2
239 44
>10,000a
494 114
>10,000a
>1000a
>1000a
>1000a
>1000a
N
N
H
*
O
N
N
H
N
a
Highest tested concentration.
influence preferential orientation of the piperidine moiety and, in
turn, of the amine and carbamoylphenyl side chains. Meanwhile,
the bis-aryl urea series displays a para-substituted phenyl struc-
ture instead of the piperidine, which favors a more linear shape
that would be less influenced by presence of a 5-Me group. A com-
parison between the binding modes of the two chemical series is
presented as Supporting information.
At this stage, we had identified a clear possibility to obtain
selective LIMK inhibitors. Subsequent work on 5,6-dimethyl
pyrrolopyrimidines however had to take into consideration several
patent applications, which focus on other kinase targets. Those
applications do not exemplify compounds such as 26, 27, or the
corresponding substitution pattern, but contain broad disclo-
sures.25 Ironically, one of such documents contained granted
claims to compounds described as Akt inhibitors, even though
we could demonstrate selectivity against Akt for 26 and 27. The
results of our medicinal chemistry efforts towards further selective
LIMK inhibitors will be reported in an upcoming publication.
As we had identified compounds with improved selectivity
against ROCK2 and PKA, we chose to further investigate whether
or not this newfound selectivity was reflected in functional assays.
Cofilin is one of the protein substrates of LIMKs, and cofilin
Please cite this article in press as: Boland, S.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.07.009

S. Boland et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
5
Table 3
Structure–activity relationships for substitutions on pyrrolopyrimidine
O
R1
N
N
H
R2
N
N
R3
R4
N
H
Compound
R1
R2
R3
R4
IC50 LIMK1 (nM)
IC50 LIMK2 (nM)
IC50 ROCK2 (nM)
10
IC50 PKA (nM)
<1
LX-7101
3
CH2NH2
OCONMe2
CO2Me
CO2Me
CO2Me
CO2Me
CO2Me
CO2Me
OCONMe2
Me
Me
H
C„N
F
H
Me
Me
H
H
H
H
24
27
3
5
1.6 0.8
1.5 0.7
4
CH2NMe2
CH2NMe2
CH2NMe2
CH2NMe2
CH2NMe2
CH2NMe2
CH2NH2
8.9 1.6
1.8 0.5
2.5 0.4
5.7 1.7
652 227
>1000a
24
6
13
23
24
25
26
27
23 16
25 10
13
2
1.1 0.2
<1
<1
332 106
>1000a
>1000a
4.0 0.6
2.8 0.7
8.0 1.9
1.6 0.5
1.4 0.6
H
20
23
26
22
7
8
6
4
Me
Me
Me
>1000a
a
Highest tested concentration.
Table 4
Cell-based activity of LX-7101, 26 and 27
Compound
EC50 MLC-PPa (nM)
EC50 cofilin-PPa
LX-7101
26
27
2300
8.7
NTb
9.0
>10,000b
>10,000b
a
Data represents the average of two experiments, each run in triplicate.
Highest concentration tested. NT: not tested.
b
Figure 3. Proposed binding mode of 27 in LIMK2, as suggested by docking.
phosphorylation therefore represents a cellular marker of LIMK
activity.4 The effects of LX-7101 and 27 on cofilin phosphorylation
were evaluated in Hela cells, using a SurefireÒ assay developed by
Perkin Elmer. Under the tested conditions, both compounds dis-
played comparable efficacy, in the nanomolar range, which is in
line with the comparable on-target activity of the compounds
(Table 4). Monitoring of MLC phosphorylation24 represented a log-
ical counterscreen for ROCK inhibition, since MLC is phosphory-
lated by ROCKs, but not LIMKs. In this assay, LX-7101 had a clear
concentration-dependent effect (EC50 = 2300 nM). Several non-se-
lective inhibitors such as 2, 3 or 6 also affected MLC phosphoryla-
tion in the low micromolar or sub-micromolar range. In contrast,
26 and 27 did not affect MLC phosphorylation at the highest tested
concentrations, confirming the selectivity observed in on-target
assays. Based on those results, selectivity of LX-7101 versus
ROCK appears higher in terms of functional activity than in terms
of on-target potency. This observation appears in line with the
higher Km of LIMKs for ATP, which would result in a weaker com-
petition at physiological ATP concentrations.7
Figure 4. IOP-lowering efficacy of LX-7101 and 27 in ocular normotensive NZW
rabbits.
the non-selective LX-7101 displayed a significant IOP reduction
at time points ranging from 1 h to 6 h post administration.
However, some animals showed clear conjunctival hyperemia
(redness). Meanwhile, animals receiving the more selective 27 at
an equivalent concentration did not display a significant IOP reduc-
tion (vs vehicle) at any time point.
LX-7101 and 27 were further evaluated in vivo for their IOP-
lowering effects (Fig. 4). Ocular normotensive New Zealand
White (NZW) rabbits were used. Both compounds were formulated
as 0.5% w/v solutions in a PEG400/Water (1:1) vehicle, adjusted to
pH 7.0 and applied topically as 40
l
l eye drop (200
lg/eye drop).
The finding that a selective LIMK inhibitor such as 27 did not
All animals (n = 5 for each experiment) received the test compound
in one eye and placebo (vehicle) in the contralateral eye. Both stud-
ies were done blinded. Under such conditions, animals receiving
display IOP-lowering efficacy obviously came as a surprise.
Indeed, LIMK inhibitors have in multiple occasions been reported
as IOP-lowering agents in several patent applications12,16 and
Please cite this article in press as: Boland, S.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.07.009

6
S. Boland et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
journal articles.7,8,14 However, those results were generated in ani-
mal models featuring a higher IOP baseline than for NZW rabbits.
For instance, a baseline IOP of 28 mmHg was reported for rats used
in the evaluation of bis-aryl urea LIMK inhibitors;13 and LX-7101
was evaluated in a dexamethasone-induced ocular hypertensive
mouse model.7 This point might be of importance, because ROCK
inhibitors (and by extension LIMK inhibitors) normally target the
trabecular outflow of aqueous humor;26 which is a pressure-de-
pendent contribution.27 Consequently, LIMK inhibitors might not
be able to display their full efficacy in NZW ocular normotensive
rabbits which have a baseline IOP around 12 mmHg.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2015.07.
009.
References and notes
1. Mizuno, K.; Okano, I.; Ohashi, K.; Nunoue, K.; Kuma, K.; Miyata, T.; Nakamura,
T. Oncogene 1994, 9, 1605.
2. Okano, I.; Hiraoka, J.; Otera, H.; Nunoue, K.; Ohashi, K.; Iwashita, S.; Hirai, M.;
Mizuno, K. J. Biol. Chem. 1995, 270, 31321.
3. Bach, I. Mech. Dev. 2000, 91, 5.
In spite of such considerations, the observation remains that
LX-7101 displays efficacy in normotensive NZW rabbits while its
very close and LIMK-selective analog 27 does not. In terms of
in vitro activity, the compounds have comparable effects against
LIMK1, LIMK2 and cofilin phosphorylation. However, 27 and LX-
7101 differed significantly in their off-target activity profile.
While 27 displays clear selectivity vs ROCK2, LX-7101 inhibits
ROCK in enzymatic and cell-based assays. With ROCK inhibitors
representing known IOP-lowering agents,18 it is tempting to con-
clude that at least part of the observed IOP-lowering effects of
LX-7101 result from ROCK inhibition. Indeed, the EC50 we observe
for LX-7101 in the MLC-phosphorylation assay is comparable to its
reported Cmax in aqueous humor during ocular PK experiments
(1500 nM).7 Further, the observation that some of the animals
receiving LX-7101 displayed conjunctival hyperemia, which is a
typical side effect of most ROCK inhibitors, also supports that
hypothesis.
In conclusion, we have here presented the synthesis and biolog-
ical evaluation of a series of LIMK inhibitors based on the
pyrrolopyrimidine scaffold. Those compounds represent close ana-
logs of LX-7101, Lexicon’s candidate for the treatment of glaucoma.
We here showed that many of such derivatives, as well as LX-7101
display nanomolar potency against LIMK1 and LIMK2, and also
ROCK2 and PKA. Such non-selective derivatives, including LX-
7101 could inhibit ROCK at low-micromolar concentrations in a
cell-based assay. Exploration of structure–activity relationships
resulted in the identification of 5,6-dimethyl-pyrrolopyrimidines
derivatives with improved selectivity. However, elimination of
side-activities also abolished IOP-lowering efficacy in ocular nor-
motensive NZW rabbits. Our findings suggest that at least part of
the IOP-lowering effects of LX-7101 could be attributed to ROCK
inhibition rather than LIMK inhibition. Additionally, it is likely that
LIMK inhibitors should preferentially be tested in animal models
with high baseline IOP. Further evaluation of selective LIMK-in-
hibitors should allow a better separation between LIMK-associated
and ROCK-associated effects in in vitro or in vivo assays than with
LX-7101.
4. Bernard, O. Int. J. Biochem. Cell Biol. 2007, 39, 1071.
5. Manetti, F. Med. Res. Rev. 2012, 32, 968.
6. Harrison, B. A.; Whitlock, N. A.; Voronkov, M. V.; Almstead, Z. Y.; Gu, K. J.;
Mabon, R.; Gardyan, M.; Hamman, B. D.; Allen, J.; Gopinathan, S.; McKnight, B.;
Crist, M.; Zhang, Y.; Liu, Y.; Courtney, L. F.; Key, B.; Zhou, J.; Patel, N.; Yates, P.
W.; Liu, Q.; Wilson, A. G.; Kimball, S. D.; Crosson, C. E.; Rice, D. S.; Rawlins, D. B.
J. Med. Chem. 2009, 52, 6515.
7. Harrison, B. A.; Almstead, Z. Y.; Burgoon, H.; Gardyan, M.; Goodwin, N. C.;
Healy, J.; Liu, Y.; Mabon, R.; Marinelli, B.; Samala, L.; Zhang, Y.; Stouch, T. R.;
Whitlock, N. A.; Gopinathan, S.; McKnight, B.; Wang, S.; Patel, N.; Wilson, A. G.;
Hamman, B. D.; Rice, D. S.; Rawlins, D. B. ACS Med. Chem. Lett. 2015, 6, 84.
8. Dai, Y. P.; Bongalon, S.; Tian, H.; Parks, S. D.; Mutafova-Yambolieva, V. N.;
Yamboliev, I. A. Vasc. Pharmacol. 2006, 44, 275.
9. Manetti, F. Drug Discovery Today 2012, 81.
10. Vorster, P. J.; Guo, J.; Yoder, A.; Wang, W.; Zheng, Y.; Xu, X.; Yu, D.; Spear, M.;
Wu, Y. J. Biol. Chem. 2011, 286, 12554.
11. Wen, X.; Ding, L.; Wang, J. J.; Qi, M.; Hammonds, J.; Chu, H.; Chen, X.; Hunter,
E.; Spearman, P. J. Virol. 2014, 88, 6906.
12. Burgoon, H. A.; Goodwin, N. C.; Harrison, B. A.; Healy, J. P.; Liu, Y.; Mabon, R.;
Marinelli, B.; Rawlins, D. B.; Rice, D. S.; Whitlock, N. A. WO/2009/131940, 2009.
13. Yin, Y.; Zheng, K.; Eid, N.; Howard, S.; Jeong, J.-H.; Yi, F.; Guo, J.; Park, C. M.;
Bibian, M.; Wu, W.; Hernandez, P.; Park, H.; Wu, Y.; Luo, J.-L.; LoGrasso, P. V.;
Feng, Y. J. Med. Chem. 2015, 58, 1846.
14. Wrobelski, S. T.; Lin, S.; Leftheris, K.; He, L.; Seitz, S. P.; Lin, T. A.; Vaccaro, W.
WO/2006/084017, 2006.
15. Ross-Macdonald, P.; de Silva, H.; Guo, Q.; Xiao, H.; Hung, C. Y.; Penhallow, B.;
Markwalder, J.; He, L.; Attar, R. M.; Lin, T. A.; Seitz, S.; Tilford, C.; Wardwell-
Swanson, J.; Jackson, D. Mol. Cancer. Ther. 2008, 7, 3490.
16. Goodwin, N. C.; Rawlins, D. B. WO/2011/091204, 2011.
17. He, L.; Seitz, S. P.; Trainor, G. L.; Tortolani, D.; Vaccaro, W.; Poss, M.; Tarby, C.
M.; Tokarski, J. S.; Penhallow, B.; Hung, C. Y.; Attar, R.; Lin, T. A. Bioorg. Med.
Chem. Lett. 2012, 22, 5995.
18. Goodwin, N. C.; Cianchetta, G.; Burgoon, H. A.; Healy, J.; Mabon, R.; Strobel, E.
D.; Allen, J.; Wang, S.; Hamman, B. D.; Rawlins, D. B. ACS Med. Chem. Lett. 2015,
6, 53.
19. Mandell, K. J.; Kudelka, M. R.; Wirostko, B. Rev. Ophthalmol. 2011, 6, 611.
20. Boland, S.; Defert, O.; Alen, J.; Bourin, A.; Castermans, K.; Kindt, N.; Boumans,
N.; Panitti, L.; Van de Velde, S.; Stalmans, I.; Leysen, D. Bioorg. Med. Chem. Lett.
2013, 23, 6442.
21. Caldwell, J. J.; Davies, T. G.; Donald, A. J. Med. Chem. 2008, 51, 2147.
22. Aronov, A. M.; Murcko, M. A. J. Med. Chem. 2004, 47, 5616.
23. Blake, J. F.; Kallan, N. C.; Xiao, D.; Xu, R.; Bencsik, J. R.; Skelton, N. J.; Spencer, K.
L.; Mitchell, I. S.; Woessner, R. D.; Gloor, S. L.; Risom, T.; Gross, S. D.; Martinson,
M.; Morales, T. H.; Vigers, G. P.; Brandhuber, B. J. Bioorg. Med. Chem. Lett. 2010,
20, 5607.
24. Schröter, T.; Griffin, E.; Weiser, A.; Feng, Y.; LoGrasso, P. Biochem. Biophys. Res.
Commun. 2008, 374, 356.
25. (a) Davies, T. G.; Garrett, M. D.; Boyle, R. G.; Collins, I. WO/2007/125321; (b)
Berdini, V.; Boyle, R. G.; Saxty, G.; Walker, D. W.; Woodhead, S. J.; Wyatt, P. G.;
Caldwell, J.; Collins, I.; Da Fonseca, T. F.; Donald, A. WO/2006/046024.
26. Kopczynski, C. C.; Epstein, D. L. J. Ocul. Pharmacol. Ther. 2014, 30, 85.
27. Goel, M.; Picciani, R. G.; Lee, R. K.; Bhattacharya, S. K. Open Ophthalmol. J. 2010,
4, 52.
Acknowledgment
The authors wish to thank Dr. Jean-Marie Stassen for his contri-
bution and advice during the review of this manuscript.
Please cite this article in press as: Boland, S.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.07.009