Novel class of LIM-kinase 2 inhibitors for the treatment of ocular hypertension and associated glaucoma

Article abstract of DOI:10.1021/jm901226j

The discovery of a pyrrolopyrimidine class of LIM-kinase 2 (LIMK2) inhibitors is reported. These LIMK2 inhibitors show good potency in enzymatic and cellular assays and good selectivity against ROCK. After topical dosing to the eye in a steroid induced mouse model of ocular hypertension, the compounds reduce intraocular pressure to baseline levels. The compounds also increase outflow facility in a pig eye perfusion assay. These results suggest LIMK2 may be an effective target for treating ocular hypertension and associated glaucoma.

Full text of DOI:10.1021/jm901226j

J. Med. Chem. 2009, 52, 6515–6518 6515
DOI: 10.1021/jm901226j
structures other thanthe trabecular meshwork.7 Evenwith the
multiple therapies available, many patients experience signifi-
cant side effects or fail to adequately control IOP, resulting in
disease progression and the need for invasive surgery to
control IOP.8,9 It is reported that 50-75% of patients are
unable to control their glaucoma with topical medical treat-
ment.10 Consequently, new and more effective therapies are
needed.
Novel Class of LIM-Kinase 2 Inhibitors for the
Treatment of Ocular Hypertension and
Associated Glaucoma
Bryce A. Harrison,*,† N. Andrew Whitlock,§
Michael V. Voronkov,† Zheng Y. Almstead,† Kun-jian Gu,†
Ross Mabon,† Michael Gardyan,† Brian D. Hamman,‡
Jason Allen,‡ Suma Gopinathan,^ Beth McKnight,§
Mike Crist,§ Yulian Zhang,† Ying Liu,†
Increasing conventional outflow facility through the trabe-
cular meshwork is a promising approach to treat open-angle
glaucoma, as this mechanism would directly reverse the causes
of ocular hypertension. One approach to increasing outflow
facility is to relax the trabecular meshwork cells through
disruption of the actin cytoskeleton. To this end, several
ROCK inhibitors have been developed that show some pro-
mise in reducing IOP in the clinic.11However, the reported
compounds are associated with significant side effects, espe-
cially conjunctival hyperemia, which may limit their usefulness.
LIM-kinases (LIMK1 and LIMK2) are downstream from
ROCK in the pathways that regulate the polymerization of
actin filaments. LIM kinases phosphorylate and deactivate
cofilin, a protein that depolymerizes actin filaments.12,13 In-
hibition of LIMK in the trabecular meshwork could promote
depolymerizationof the actinfilaments, resulting inrelaxation
of the tissue, increased outflow facility, and reduced IOP.
Similar to reported ROCK knockout mice,14-16 LIMK2
knockout mice had lower IOP when compared to their wild
type littermates.17 As the LIM-kinases are downstream of
ROCK in the signaling pathway, inhibition of these targets
could be associated with fewer side effects.
Herein we report the discovery of a series of novel com-
pounds that inhibit LIMK218 and their use in lowering IOP in
animal models. Our objective was to develop potent inhibitors
of LIMK2 with selectivity against ROCK that could be
incorporated into an aqueous formulation for topical applica-
tion to the eye. Compounds were assayed in a primary screen
for in vitro inhibition of cofilin phosphorylation by LIMK2
and further evaluated for inhibition of cofilin phosphoryla-
tion in pig trabecular meshwork cells using Western blot
analysis. Several highly potent compounds were identified in
Lawrence F. Courtney,† Billie Key,‡ Julia Zhou,‡
Nita Patel,‡ Phil W. Yates, Qingyun Liu,‡
Alan G. E. Wilson,^ S. David Kimball,† Craig E. Crosson,
Dennis S. Rice,§ and David B. Rawlins*,†
†Department of Medicinal Chemistry, Lexicon Pharmaceuticals, 350
Carter Road, Princeton, New Jersey 08540, ‡Department of
Pharmaceutical Discovery, ^Department of Drug Metabolism,
Pharmacokinetics, and Toxicology and §Department of
Ophthalmology, Lexicon Pharmaceuticals, 8800 Technology Forest
Place, The Woodlands, Texas 77381, and Department of
Ophthalmology, Medical University of South Carolina, 167 Ashley
Avenue, Charleston, South Carolina 29425
Received August 17, 2009
Abstract: The discovery of a pyrrolopyrimidine class of LIM-
kinase 2 (LIMK2) inhibitors is reported. These LIMK2 inhibitors
show good potency in enzymatic and cellular assays and good
selectivity against ROCK. After topical dosing to the eye in a
steroid induced mouse model of ocular hypertension, the com-
pounds reduce intraocular pressure to baseline levels. The com-
pounds also increase outflow facility in a pig eye perfusion assay.
These results suggest LIMK2 may be an effective target for treating
ocular hypertension and associated glaucoma.
Open-angle glaucoma is a progressive neurodegenerative
disease of the inner retina and optic nerve head. One signifi-
cant risk factor in glaucoma is increased intraocular pressure
(IOPa). High IOP can contribute to optic nerve head damage,
resulting in vision loss and blindness. The IOP is regulated by
the fluid dynamics of aqueous humor in the anterior chamber.
Aqueous humor is produced by the ciliary body and flows
through the anterior chamber before exiting primarily
through the trabecular meshwork.1 In ocular hypertensive
patients, conventional outflow facility through the trabecular
meshwork has decreased because of structural changes in the
tissue.2
Table 1. Aryl Analogues of Compound 1
It is estimated that 2.2 million people in the U.S. suffer from
open-angle glaucoma.3 All approved therapies to treat glau-
coma are directed at lowering IOP. Carbonic anhydrase
inhibitors, β-blockers, and R-2 adrenergic receptor agonists
decrease aqueous humor production,4-6 and prostaglan-
din analogues increase uveoscleral outflow facility through
Compd
R
LIMK2 IC50 (nM)
1
H
2-Me
800
11500
650
5a
5b
5c
5d
5e
5f
5g
*To whom correspondence should be addressed. For B.A.H.: phone,
609-466-6078; fax, 609-466-3562; e-mail, bharrison@lexpharma.com.
For D.B.R.: phone, 609-466-5564; fax, 609-466-3562; e-mail, drawlins@
lexpharma.com.
a Abbreviations: IOP, intraocular pressure; ROCK, Rho kinase;
LIMK, LIM-kinase; HPMC, hydroxypropylmethylcelluose; SEM,
standard error of the mean; HATU, O-(7-azabenzotriazol-1-yl)-N,N,
N0,N0-tetramethyluronium hexafluorophosphate.
3-Me
4-Me
1300
520
3-OMe
4-OMe
3-OPh
3-Br
1500
64
300
r
2009 American Chemical Society
Published on Web 10/15/2009
pubs.acs.org/jmc

6516 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Scheme 1. Synthesis of Amide Analogues of Compound 1
Harrison et al.
Table 2. Ureas and Cyanoguanidines
LIMK2
IC50 (nM)
Compd
X
Y
R
Scheme 2. Synthesis of Urea and Cyanoguanidine Analogues
of Compound 1
5b
5d
CH
CH
CH
N
O
O
O
O
O
O
Me
OMe
Br
650
520
300
92
5g
11a
11b
11c
12
Me
OMe
Br
N
140
38
N
N
NCN
Br
11
Table 3. Piperazine Substitution
the cell based screen. A formulation screen was conducted to
identify compounds suitable for topical instillation in a steroid-
induced mouse model of ocular hypertension.19
Through high-throughput screening and elaboration of the
resulting hits, we identified an active chemotype typified by
pyrrolopyrimidine compound 1 (Table 1). This compound
was an attractive lead due to having submicromolar activity
and a modular structure amenable to rapid determination of
SAR. Analogues of this compound were made to optimize
potency and improve properties for topical dosing.
Amide analogues of 1 were synthesized as described in
Scheme 1. Nucleophilic substitution of chloropyrrolopyrimi-
dine 220 with ethyl isonipecotate 3, followed by saponification
of the ester, gave acid 4. Coupling of acid 4 with anilines using
HATU or other coupling reagents afforded amides 5. In
addition to amides, urea and cyanoguanidine analogues of 1
were prepared (Scheme 2). Pyrrolopyrimidine 2 was reacted
with substituted Boc-piperazines 6, followed by Boc deprotec-
tion, to give amines 7. Amines 7 could be converted to ureas 8
by reaction with isocyanates, purchased commercially or
prepared from anilines and triphosgene. Alternatively, amines
7 could be converted to cyanoguanidines 9 by reaction
with phenyl carbamimidates 10, prepared from anilines and
diphenyl-N-cyanocarbonimidate.
Compd
X
R1
R2
R3
LIMK2 IC50 (nM)
11c
12
13
14
15
16
17
18
19
20
O
H
H
H
H
38
11
NCN
O
H
H
H
H
H
H
(S)-Me
H
3.9
0.9
37
NCN (S)-Me
(R)-Me
NCN (S)-i-Pr
H
O
H
H
4.2
3.9
133
16
NCN Me (trans)
H
Me
H
O
O
Me (cis)
-CH2CH2-
Me
= R1
H
H
=R1
NCN -CH2-
76
Table 4. Pyrrolopyrimidine Substitution
SAR around the aryl ring of 1 was quickly established as
illustrated by 5a-g (Table 1). Meta substitution with a variety
of substituents (5b, 5d, 5f, and 5g) gave improved potency
compared with the unsubstituted parent, while para substitu-
tion (5c and 5e) reduced activity and ortho substitution (5a)
almost completely eliminated it. This trend was consistent
throughout the work on the project. Especially potent in the
meta position were aryl ethers. For example, phenyl ether 5f
had an IC50 of 64 nM. However, these aryl ethers had very
Compd
R1
R2
R3
LIMK2 IC50 (nM)
14
Me
H
H
H
H
H
H
Cl
0.9
8.1
21aa
21b
21c
21d
H
Cl
H
H
1.2
4.2
Me
H
H
>1000
a Racemate.

Letter
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21 6517
Table 5. Aryl Analogues with Cell-Based Activity
Table 6. Profile of Selected Compounds
IC50 (nM)
Δ IOP
( SEMa
(mmHg)
Compd
LIMK 2
LIMK 1
ROCK 1
ROCK 2
14
0.9
1.5
3.2
1.2
0.5
3.7
6.6
3.2
260
310
51
2700
2200
280
3.7 ( 0.4
22e
22f
22j
3.6 ( 0.6
3.6 ( 0.6
490
1100
a Reduction in intraocular pressure at 1 h postdose in a dexamethsone
induced ocular hypertensive mouse following topical instillation of 5 μL
of 1 mg/mL xanthan gum based suspension of compound, compared
to vehicle dosed mice (14: n = 30, vehicle n = 33; 22f: n = 13, vehicle
n = 14; 22j: n = 18, vehicle n = 17).
LIMK2 Cellular % inh
at 10 nMa
Compd
X
R
IC50 (nM)
14
NCN Br
NSO2Me Br
0.9
6.8
1.6
2.3
1.8
1.5
3.2
1.9
1.2
1.1
1.2
3.5
5.5
0.8
3.0
3.0
3.4
1.7
2.9
50
48
49
0
22a
22b
22c
22d
22e
22f
22g
22h
22i
22j
22k
22l
NCN
O
Cl
Cl
NCN
NCN
O
F
39
63
51
12
18
10
61
0
Br, 4-F
Br, 4-F
CF3
CN
NCN
NCN
NCN
O
t-Bu
OCONMe2
NCN
O
CONH-i-Pr
CONH-i-Pr
0
22m NCN
CONH(CH2)2OH
CONH(CH2)2OH
CONHCH(CH2OH)2
CONH(CH2)2NMe2
SO2NH-i-Pr
9
22n
22o
22p
22q
22r
O
0
O
27
0
NCN
NCN
NCN
0
22
Figure 1. Change in IOP in a dexamethsone induced ocular hyper-
tensive mouse following topical instillation of 3 μL of a 1 mg/mL or
0.1 mg/mL HPMC based solution of 22j or of a 2.5 mg/mL solu-
tion of timolol (vehicle, n = 10; 0.3 μg of 22j, n = 10; 3 μg of 22j, n =
9; timolol, n = 10).
SO2NH(CH2)2OH
a Inhibition of cofilin phosphorylation in pig trabecular meshwork
cells as determined by Western blot analysis. Compounds tested in
triplicate at 100, 10, and 1 nM. For clarity, only results at 10 nM are
shown.
potency. Substitution at R2 (21c) was tolerated, while halogen
substitution at R3 (21d) was not.
poor solubility which limited their use in aqueous topical
formulations. Another substitution at the meta position of the
aryl ring that showed good potency was bromine (5g). Ana-
loguesofarylbromide5ghadsufficientsolubilitytoformulate
for in vivo assays and were readily accessible, facilitating
further SAR development.
We next turned our attention to the piperidine ring. Re-
placement of the piperidine with piperazine to give ureas
11a-c (Table 2) improved potency 4- to 8-fold over the parent
analogues 5b, 5d, and 5g. For bromide 11c, modification of
the urea to give cyanoguanidine 12 gave an additional 3-fold
boost in potency.
The piperazine ring allowed us to easily introduce further
substitution in an effort to improve potency (Table 3). The
best results were achieved by introducing a group in the S
configuration at R1. Thus, an (S)-methyl group at the R1
position gave a 10-fold increase in potency in the urea (13) and
cyanoguanidine series (14). The enantiomeric (R)-methyl
group at R1 (15) did not boost potency. Larger groups at
R1, such as (S)-i-Pr (16), were also fairly welltolerated, as were
trans-dimethyl substitution at R1 and R3 (17) and bridging
between R1 and R2 (19). However, cis-dimethyl substitution
at R1 and R2 (18) and bridging between R1 and R3 (20)
resulted in reduced potency.
The SAR around the pyrrolopyrimidine is summarized in
Table 4. The methyl group at R1 (14) was responsible for a
significant increase in potency compared to the unsubstituted
compound 21a. The methyl group could be replaced with a
chloride or other small group as in 21b with little loss of
With potency established by 13 and 14, multiple aryl
analogues of ureas, cyanoguanidines, and sulfonylguanidines
were prepared with the goal of optimizing selectivity, solubi-
lity, and activity in cells (Table 5). We discovered that a wide
range of groups were tolerated at the meta position of the aryl
ring, including halides, alkyl groups, carbamates, amides, and
sulfonamides (22a-r). The compounds were further differ-
entiated in the cell based assay. Compounds were tested at
three concentrations (100, 10, and 1 nM) with the goal of
identifying compounds exhibiting >50% inhibition of cofilin
phosphorylation in pig trabecular meshwork cells at 10 nM.
Four compounds (14, 22e, 22f, and 22j) were identified that
met these criteria.
The four compounds active in the cellular assay were also
tested for in vitro inhibition of LIMK1, ROCK1, and
ROCK2 (Table 6). The compounds generally showed com-
parable activity between LIMK1 and LIMK2, and 14, 22e,
and 22j showed >100-fold selectivity for LIMK2 against
ROCK1 and >900-fold selectivity against ROCK2. Com-
pound 22f was less selective, only 16-fold selective against
ROCK1 and 88-fold against ROCK2.
Compounds 14, 22f, and 22j were selected for in vivo
experiments and were formulated as suspensions using a
xanthan gum based vehicle. These were then assayed in the
dexamethasone induced ocular hypertensive mouse model.19
A single 5 μL drop of 1 mg/mL formulation of each com-
pound was administered to one eye of an ocular hypertensive
mouse, and IOP was measured after 1 h. At this 5 μg dose, 14,

6518 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 21
Harrison et al.
References
(1) Weinreb, R. N.; Toris, C. B.; Gabelt, B. T.; Lindsey, J. D.; Kaufman,
P. L. Effects of prostaglandins on the aqueous humor outflow
pathways. Surv. Ophthalmol. 2002, 47 (Suppl. 1), S53–S64.
(2) Toris, C. B.; Koepsell, S. A.; Yablonski, M. E.; Camras, C. B.
Aqueous humor dynamics in ocular hypertensive patients.
J. Glaucoma 2002, 11, 253–258.
(3) Friedman, D. S.; Wolfs, R. C.; O’Colmain, B. J.; Klein, B. E.;
Taylor, H. R.; West, S.; Leske, M. C.; Mitchell, P.; Congdon, N.;
Kempen, J. Prevalence of open-angle glaucoma among adults in
the United States. Arch. Ophthalmol. 2004, 122, 532–538.
(4) Coakes, R. L.; Brubaker, R. F. The mechanism of timolol in
lowering intraocular pressure in the normal eye. Arch. Ophthalmol.
1978, 96, 2045–2048.
(5) Toris, C. B.; Gleason, M. L.; Camras, C. B.; Yablonski, M. E.
Effects of brimonidine on aqueous humor dynamics in human eyes.
Arch. Ophthalmol. 1995, 113, 1514–1517.
(6) Lippa, E. A.; Carlson, L. E.; Ehinger, B.; et al. Dose response and
duration of action of dorzolamide, a topical carbonic anhydrase
inhibitor. Arch. Ophthalmol. 1992, 110, 495–499.
(7) Gabelt, B. T.; Kaufman, P. L. Prostaglandin F2R increases uveoscl-
eral outflow in the cynomolgus monkey. Exp. Eye Res. 1989, 49,
389–402.
Figure 2. Change in outflow facility in a perfused pig eye upon
infusion of 100 nM 14, 100 nM 22j, or 10 μM 23 (vehicle, n = 6; 14,
n = 6; 22j, n = 5; 23, n = 7).
22f, and 22j lowered IOP by 3.6-3.7 mmHg relative to vehicle
dosed control animals (Table 6). This reduction in IOP
effectively represents a return to the baseline IOP measured
in normotensive animals.21 Systemic exposure on topical
dosing was below the limits of quantitation for 22j (<18 nM).
Compound 22j was reformulated as 1 and 0.1 mg/mL
solutions using a cosolvent/HPMC based vehicle and com-
pared to a marketed formulation of the β-blocker timolol in
the ocular hypertensive mouse model at 1, 2, and 4 h postdose
(Figure 1). With this formulation, 22j showed a dose depen-
dent response with maximal efficacy at 1-2 h (3 μg dose,
Δ IOP = -3.1 mmHg). While 22j did not have the same
maximal effect as timolol (Δ IOP = -5.2 mmHg) at 2 h, the
effect at other time points and the duration of action were
similar.
To investigate the mechanism of action of the LIMK2
inhibitors, 14 and 22j were also examined in a pig eye perfu-
sion assay (Figure 2). In this experiment, a solution of the
compound is infused at a constant pressure into a mounted
anterior chamber of an enucleated pig eye. With fixed pres-
sure, changes in inflow volume are directly attributable to
changes in outflow facility. Consistent with the proposed
mechanism of action for LIMK2 inhibitors, 14 and 22j
increased outflow facility by about 30% at 100 nM. This
result compared favorably with the effects of clinical ROCK
inhibitor Y-39983 (23)11 which increased outflow facility by
28% at 10 μM.
(8) Schwartz, K.; Budenz, D. Current management of glaucoma. Curr.
Opin. Ophthalmol. 2004, 15, 119–126.
(9) Reyes, E.; Izquierdo, N. J.; Blasini, M. Adverse drugs reactions
associated with glaucoma medications. Bol.;Asoc. Med. P. R.
1997, 89, 51–55.
(10) Lewis, T. (Principal Author). Optometric Clinical Practice Guide-
line, Care of Patient with Open Angle Glaucoma: Reference Guide
for Clinicians. Updated Aug 17, 2002. http://www.aoa.org/docu-
ments/CPG-9.pdf (accessed Sep 28, 2009).
(11) Tanihara, H.; Inatani, M.; Honjo, M.; Tokushige, H.; Azuma, J.;
Araie, M. Intraocular pressure-lowering effects and safety of
topical administration of a selective ROCK inhibitor, SNJ-1656,
in healthy volunteers. Arch. Ophthalmol. 2008, 126, 309–315.
(12) Arber, S.; Barbayannis, F. A.; Hanser, H.; Schneider, C.; Stanyon,
C. A.; Bernard, O.; Caroni, P. Regulation of actin dynamics
through phosphorylation of cofilin by LIM-kinase. Nature 1998,
393, 805–809.
(13) Yang, N.; Higuchi, O.; Ohashi, K.; Nagata, K.; Wada, A.; Kangawa,
K.; Nishida, E.; Mizuno, K. Cofilin phosphorylation by LIM-kinase
1 and itsrole in Rac-mediated actin reorganization. Nature 1998, 393,
809–812.
(14) Lindsey, J. D.; Weinreb, R. N. Elevated intraocular pressure and
transgenic applications in the mouse. J. Glaucoma 2005, 14, 318–320.
(15) John, S. W.; Anderson, M. G.; Smith, R. S. Mouse genetics: a tool
to help unlock the mechanisms of glaucoma. J. Glaucoma 1999, 8,
400–412.
(16) Whitlock, N. A.; Harrison, B.; Mixon, T.; Yu, X.-Q.; Wilson, A.;
Gerhardt, B.; Eberhart, D. E.; Abuin, A.; Rice, D. S. Decreased
intraocular pressure in mice following either pharmacological or
genetic inhibition of ROCK. J. Ocul. Pharmacol. Ther. 2009, 25,
187–194.
(17) Whitlock, A; Rice, D. S.; et al. et al. Manuscript in preparation.
(18) We are aware of only one other report of LIM kinase inhibitors in
the literature: Ross-Macdonald, P.; de Silva, H.; Guo, Q.; Xiao,
H.; Hung, C.-Y.; Penhallow, B.; Markwalder, J.; He, L.; Attar,
R. M.; Lin, T.; Seitz, S.; Tilford, C.; Wardwell-Swanson, J.;
Jackson, D. Identification of a nonkinase target mediating cyto-
toxicity of novel kinase inhibitors. Mol. Cancer Ther. 2008, 7, 3490–
3498.
(19) Whitlock, A.; McKnight, B.; Rice, D. S. Increased Intraocular
Pressure in Mice Treated with Dexamethasone. Abstracts ofPapers,
2008 Annual Meeting of the Association for Research in Vision and
Ophthalmology, Ft. Lauderdale, FL; Association for Research in Vision
and Ophthalmology: Rockville, MD, 2008; Abstract 2100/A211.
(20) West, R. A. 4-Hydroxypyrrolo[2,3-d]pyrimidine: Mannich reac-
tion. J. Org. Chem. 1961, 26, 4959–4961.
In conclusion, we report the discovery of a novel class of
LIMK2 inhibitors and their optimization for topical ocular
dosing. Treatment of ocular hypertensive mice results in a
significant reduction in IOP, and the compounds increase
outflow facility in pig eye perfusion assays. These results
suggest that LIMK2 may be a valid target for the treatment
of glaucoma. Further elaboration and development of
LIMK2 inhibitors for the treatment of glaucoma will be
reported later.
(21) The average IOP for normotensive mice in our experiments was
13.9 ( 0.9 mmHg. The average IOP for the dexamethasone induced
ocular hypertensive mice was 17.7 ( 1.1 mmHg, representing a
3.8 mmHg increase from baseline.
Supporting Information Available: Experimental details and
analytical data of the compounds. This material is available free
of charge via the Internet at http://pubs.acs.org.