2,3-Diaminopyrazines as rho kinase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2009.12.012

Inhibition of rho kinase (ROCK) has been recognized as an important target for a number of diseases, including glaucoma. Herein we report SAR development around two hits from a kinase library that led to the discovery of the ROCK inhibitor compound 38. In vitro and in vivo analysis of this compound, including its effects in a monkey model of glaucoma will be discussed.

Full text of DOI:10.1016/j.bmcl.2009.12.012

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1137–1140
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
2,3-Diaminopyrazines as rho kinase inhibitors
a
a
a
a
b
Alan J. Henderson a, , Mark Hadden , Cheng Guo , Neema Douglas , Helene Decornez , Mark R. Hellberg ,
*
Andrew Rusinko b, Marsha McLaughlin b, Naj Sharif b, Colene Drace b, Raj Patil b
a AMRI, 26 Corporate Circle, PO Box 15098, Albany, NY 12212-5098, USA
b Alcon Laboratories Inc., 6201 South Freeway, Fort Worth, TX 76134-2099, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Inhibition of rho kinase (ROCK) has been recognized as an important target for a number of diseases,
including glaucoma. Herein we report SAR development around two hits from a kinase library that led
to the discovery of the ROCK inhibitor compound 38. In vitro and in vivo analysis of this compound,
including its effects in a monkey model of glaucoma will be discussed.
Received 20 October 2009
Revised 1 December 2009
Accepted 3 December 2009
Available online 5 December 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Rho kinase
ROCK
Glaucoma
Intraocular pressure
IOP
Rho kinase (Rho-associated coiled–coil-containing protein ki-
nase or ROCK) is a serine/threonine protein kinase that is a mem-
ber of the AGC family.1 There are two known isoforms (ROCK 1 and
ROCK 2) which have high sequence homology particularly in the
kinase domain (92%).2 Activation of ROCK leads to phosphorylation
of various target proteins, with myosin light chain (MLC) being one
of ROCK’s main substrates.3 Inhibition of ROCK has been proposed
as a mechanism for treating a number of diseases including renal
disease,4 hypertension,5 cancer,6 ischemia7 and glaucoma.8 The lat-
ter represented the focus of our program, as there was strong evi-
dence to suggest that ROCK inhibitors had the potential to lower
intraocular pressure (IOP) by increasing uveoscleral and trabecular
(conventional) outflow.9 Furthermore, phase I clinical data has
shown significant dose-dependant IOP lowering effects in humans
using a ROCK inhibitor.10 To date, the only marketed rho kinase
inhibitor (Fasudil, 1) is prescribed in Japan for the treatment of
cerebral vasospasm after hemorrhage in the subarachnoid space
leading to ischemia (Fig. 1).
NH
N
O S O
N
1
Figure 1. Structure of Fasudil.
N
N
NH2
N
N
N
NH2
N
N
N
N
OH
N
2
3
ROCK 1 IC50 = 4600 nM
ROCK 2 IC50 = 2500 nM
ROCK 1 IC50 = 4800 nM
ROCK 2 IC50 = 3100 nM
Figure 2. Hits obtained from focused kinase library.
Screening of a focused kinase library uncovered two aminopyr-
azine hits (2 and 3) that were of particular interest to us (Fig. 2).
Although both had fairly weak ROCK activity, docking of compound
3 into a ROCK 1 homology model suggested there was scope for
picking up additional binding interactions to improve the potency.
It was decided to focus on compound 3 as the initial hit, using a
structure-based approach towards improving the potency. An in-
house developed homology model (Fig. 3) suggested that the 4-
pyridyl substituent bound to the hinge region of the kinase. This
was somewhat validated by looking at the activity of other com-
pounds in the focused library, where non nitrogen-containing aryl
replacements for pyridine, as well as other pyridine isomers
resulted in a loss of activity. It was our belief that there was an
opportunity to enhance the binding efficiency in the hinge region
via modified pyridines, or heterocycles that had the potential to
pick up additional binding interactions. Furthermore, there
* Corresponding author. Tel.: +1 518 512 2305; fax: +1 518 512 2081.
E-mail address: Alan.Henderson@amriglobal.com (A.J. Henderson).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.12.012

1138
A. J. Henderson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1137–1140
Table 1
Effect of modifications of Ar group on ROCK 1 and ROCK 2 affinity
N
N
NH2
Ar
N
N
Compound
Ar
ROCK 115
IC50 (nM)
ROCK 215
IC50 (nM)
a
a
3
11
12
13
14
15
16
17
18
4-Pyridine
4600
2500
4-(2-Methyl)pyridine
4-(3-Methyl)pyridine
4-(2-Amino)pyridine
3-Pyrazole
4-(7-Aza)indole
4-Indazole
>100,000
120,000
25,000
37,000
1500
9600
2300
>10,000
>100,000
58,000
53,000
2900
530
2500
5-Indazole
6-Indazole
950
>10,000
a
IC50 values represent means of at least three determinations for each compound
reported in the table. Standard deviation generally <50%.
Figure 3. Docking model of 3 in the catalytic domain of ROCK 1. The activation loop
(orange), glycine rich loop (green), hinge (red) and the tail (yellow) are indicated in
the enzyme structure (PDB code: 2ETO).
nature of the pyridine or added steric bulk (methyl, amino; com-
pounds 11–13) were detrimental to potency. However, pyrazole
or bicyclic pyridine replacements where two nitrogens were avail-
able for binding (compounds 14–18) generally maintained the po-
tency (particularly against ROCK 2), although the substitution
pattern was important. The most effective substituent was 4-(7-
aza)indole 15, which led to a significant improvement in both
ROCK 1 and ROCK 2 potency.
We next turned our attention to the 2-amino substituent. Initial
attempts at replacing the primary amino group were disappoint-
ing. The monomethyl, dimethyl and diethyl analogs 19–21 pro-
vided no real improvement in potency whilst introduction of
polar functionality via an ethyl linkage (compounds 22–24) gener-
ally resulted in a drop in potency. However, introduction of a sat-
urated heterocyclic ring system markedly improved the potency
against ROCK 2 with the most potent compound 25 showing a
10-fold improvement in potency over the initial hit (Table 2).
Somewhat surprisingly, heterocyclic analogs 25 and 26 showed a
significant level of selectivity for ROCK 2 over ROCK 1 (18-fold
and 32-fold, respectively) suggesting the pocket accessible from
the 2-amino position differs between the two isoforms. Addition-
ally, the lack of selectivity seen for the diethyl analog 21 suggests
a level of conformational constraint is necessary within this pocket
to improve the ROCK 2 activity.
appeared to be a pocket accessible from the 2-amino substituent
that offered the potential of accessing previously unexplored
space. Finally, different amino groups to replace the 3-substituted
4-methylhomopiperidine provided an opportunity to improve
binding to the aspartic acid residue (D216) within the Asp-Phe-
Gly (DFG) motif.
A synthetic protocol was devised to allow for diversity of
analogs and rapid SAR development. 5-Aryl analogs could be
easily prepared from 2-amino-3,5-dibromopyrazine as shown in
Scheme 1. Regioselective SnAr displacement at the 3-position11
provided diaminopyrazines 5 which could be elaborated to the
final compounds 7 via a Suzuki coupling reaction.12 Alternatively,
the stannanes 7 could be formed13 and coupled with a suitable aryl
halide under Stille coupling conditions.12
For 2-substituted analogs additional steps had to be introduced
(Scheme 2). 2-Chloropyrazine was functionalized with a suitable
amine and dibrominated14 to form the aminodibromopyrazines
10. This was then progressed as in Scheme 1.
Table 1 summarizes the SAR of modifying the hinge-binding
pyridine ring. Simple substituents that affected the electronic
Finally, the 3-position was targeted for SAR development. The
most potent 2-amino substituent (pyrrolidine) was retained as
N
NH2
Br
N
NH2
N
NH2
a
b
Br
N
N
NR1R2
Br
N
NR1R2
Ar
4
5
6
Table 2
Effect of modifications of 20-R group on ROCK 1 and ROCK 2 affinity
d
c
R
N
N
N
N
NH2
NR1R2
Me3Sn
N
N
N
7
a
a
Scheme 1. Reagents and conditions: (a) HNR1R2, sealed tube, DMSO, 130 °C; (b)
ArB(OH)2/ArB(OR)2, Pd(dppf)Cl2, Na2CO3, DMF, H2O, 100 °C or ArB(OH)2/ArB(OR)2,
Pd(dppf)Cl2, K2CO3, DMSO, 100 °C; (c) (SnMe3)2, Pd(PPh3)4, toluene, reflux; (d) ArBr,
Pd(PPh3)4, toluene, reflux.
Compound
R
ROCK 1 IC50 (nM)
ROCK 2 IC50 (nM)
3
19
20
21
22
23
24
25
26
27
–NH2
4600
6900
6500
2500
2400
1700
6700
3000
52,000
45000
330
–NHMe
–NMe2
–NEt2
–NH(CH2)2OH
–NH(CH2)2OMe
–NMe(CH2)2OMe
1-Pyrrolidine
1-Piperidine
2600
42,000
10,000
48,000
6000
18,000
68,000
N
Cl
N
N
NR1R2
N
NR1R2
Br
a
b
N
Br
N
560
6300
8
9
10
1-Morpholine
a
Scheme 2. Reagents and conditions: (a) HNR1R2, sealed tube, DMSO, 130 °C; (b) N-
bromosuccinimide, DMSO, 0 °C to room temperature.
IC50 values represent means of at least three determinations for each compound
reported in the table. Standard deviation generally < 50%.

A. J. Henderson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1137–1140
1139
was the 4-pyridyl substituent in the 5-position (due to synthetic
0
-5
Compound 38
(300 ug)
accessibility). Removal of the N-methyl substituent from the
homopiperazine analog (compound 28) produced a highly signifi-
cant improvement in potency against both ROCK 1 and ROCK 2.
The NH-piperazine 29 resulted in similar potency as did the 4-
aminopiperidine 30 (although the N-methylated analog 31 was
not as well tolerated). Further homologation of the latter two com-
pounds led to 32 and 33 with the best combined activity (Table 3).
In order to further elaborate on the developing SAR, some addi-
tional compounds were synthesized to take advantage of the in-
creased potency. The initial hit 2 was revisited, as was the 4-(7-
aza)-indole replacement for the 4-pyridyl hinge binder (Table 4).
The homopiperazine analog of scaffold A (compound 34) produced
a modest improvement in potency. However, the piperazine analog
35 caused a highly significant improvement in potency (particu-
larly against ROCK 1) producing the best combined activity seen
in this series. The 4-aminopiperidine analog 36 did not, however,
have the beneficial effect seen in the 4-pyridyl series. Modifica-
tions to the hit compound 2 (scaffold B) also showed improve-
Compound 38
(600 ug)
-10
-15
-20
-25
-30
-35
-40
-45
HN1152 (300 ug)
0
1
2
3
4
5
6
Time after Dosing (hours)
Figure 4. IOP lowering effects of compound 38 and HN1152 in the lasered monkey
over a 6 h timeframe (n = 8–9). Baseline IOP was between 36.0 and 39.0 mmHg.
Compounds were administered as a single 30
l
L dose of a 1% solution (300
lg) or
two 30 L doses separated by at least 5 min (600
l
l
g).
ments. Introduction of substituted 2-amino groups had
markedly positive effect, particularly with the pyrrolidine 38.
a
The next step in the process was to evaluate compounds of
interest in a rabbit safety model of ocular dosing.16 A number of
compounds were tested for their effects on eye irritation, with
compound 38 showing the best profile. Only mild hyperemia (red-
dening of the eye) was observed with no discharge or swelling
apparent. This profile was very similar to one of the known rho ki-
1 IC50 = 55 nM, ROCK 2
IC50 = 34 nM) which had an excellent IOP lowering profile. Com-
pound 38 was then tested in a further battery of in vitro assays
Table 3
Effect of modifications of 3’-R group on ROCK 1 and ROCK 2 affinity
nase inhibitors, HN115217 (ROCK
N
N
N
R
N
where it was shown to have an acceptable profile (>2 lM inhibi-
tory activity against a panel of seven cytochrome P450’s,18 100%
remaining after 1 h in human liver microsomes19). Additionally,
compound 38 was subjected to a kinase selectivity panel where
it showed high levels of selectivity. Only 20 of 213 kinases tested
a
a
Compound
R
ROCK 1 IC50 (nM)
ROCK 2 IC50 (nM)
N
28
29
30
320
610
290
39
62
62
NH
NH
N
N
returned activity of greater than 50% inhibition at 10
compared favorably with HN1152 which showed activity of great-
er than 50% inhibition at 10 M against 91 of the 213 kinases.
lM. This
l
NH2
Based on the favorable in vitro data and the safety profile of
compound 38, it was decided to measure its efficacy in a model
for intraocular pressure lowering (lasered cymologous monkey).16
N
31
32
33
1800
130
250
28
NHMe
NH2
N
N
Compound 38 was topically dosed at 300
lg and 600 lg in a study
where HN1152 was used as a positive control (Fig. 4).
NHMe
150
49
Compound 38 showed a robust, highly efficacious effect in the
a
monkey, reducing the intraocular pressure by an average of 33%
at the 300 lg dose and 37% at the 600 lg dose. This compared very
IC50 values represent means of at least three determinations for each compound
reported in the table. Standard deviation generally < 50%.
well with HN1152 which produced an average reduction of 34%. It
was interesting to note that at the highest dose, compound 38 had
a sustained effect on the intraocular pressure that was maintained
after 6 h, whereas HN1152 returned to levels matching that of the
lower dose of compound 38. Furthermore, the magnitude of the
IOP lowering compares well with published data for other com-
pounds such as Y-39983 (14.3% maximum reduction in IOP for a
0.05% solution).20
Table 4
Effect of modifications of 30-R group on ROCK 1 and ROCK 2 affinity
N
N
R
N
N
N
N
R
HN
N
N
N
OH
In conclusion, SAR development around two hits from a focused
kinase library led to a number of ROCK inhibitors that had sub
500 nM activity against ROCK 1 and sub 100 nM activity against
ROCK 2. One of these compounds (compound 38) had a suitable
profile for in vivo determination of its effects on lowering intraoc-
ular pressure. The compound was found to be highly effective,
showing improved efficacy over HN1152 at the highest dose.
A
B
a
a
Compound Scaffold
R
ROCK 1 IC50
(nM)
ROCK 2 IC50
(nM)
N
34
35
36
A
A
A
170
34
21
17
89
NH
NH
N
N
References and notes
300
NH2
37
38
B
B
NMe2
1-Pyrrolidine
1400
300
300
78
1. Breitenlechner, C.; Gabel, M.; Engh, R.; Bossemeyer, D. Oncol. Res. 2004, 14, 267.
2. (a) Amino, M.; Fukata, Y.; Kaibuchi, K. Exp. Cell Res. 2000, 261, 44; (b) Riento, K.;
Ridley, A. J. Nat. Rev. Mol. Cell Biol. 2003, 4, 446.
3. Somlyo, A. P.; Somlyo, A. V. J. Physiol. 2000, 522, 177.
4. Sharpe, C. C.; Hendry, B. M. J. Am. Soc. Nephrol. 2003, 14, 261.
a
IC50 values represent means of at least three determinations for each compound
reported in the table. Standard deviation generally <50%.

1140
A. J. Henderson et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1137–1140
5. Doggrell, S. A. Exp. Opin. Invest. Drugs 2005, 14, 1157.
6. Aznar, S.; Pilar, F.-V.; Espina, C.; Lacal, J. C. Cancer Lett. 2004, 206, 181.
7. Hu, E.; Lee, D. Exp. Opin. Ther. Targets 2005, 9, 715.
8. Novak, G. D. Exp. Opin. Emerg. Drugs 2003, 8, 251.
9. Rao, P. V.; Deng, P. F.; Kumar, J.; Epstein, D. L. Invest. Ophthalmol. Vis. Sci. 2001,
42, 1029.
30 min at 23 °C. Reaction mixtures were then analyzed by fluorescence
polarization.
16. Hellberg, M. R.; Sallee, V. L.; McLaughlin, M. A.; Sharif, N. A.; Desantis, L.; Dean,
T. R.; Zinke, P. W. J. Ocul. Pharmacol. Ther. 2001, 5, 421.
17. HN1152 (also known as H-1152 or dimethyl-fasudil) has the structure shown
below.
10. Tanihara, H.; Inatani, M.; Honjo, M.; Tokushige, H.; Azuma, J.; Araie, M. Arch.
Ophthalmol. 2008, 126, 309.
NH
11. Vitse, O.; Laurent, F.; Pocock, T. M.; Benezech, V.; Zamik, L.; Elliott, K. R.; Subra,
G.; Portet, K.; Bompart, J.; Chapat, J. P.; Sunall, R. C.; Michel, A.; Bonnet, P. A.
Bioorg. Med. Chem. 1999, 7, 1059.
N
O S O
12. Miyaura, N.; Suzuki, A. J. Chem. Soc., Chem. Commun. 1979, 866.
13. Milstein, D.; Stille, J. K. J. Am. Chem. Soc. 1978, 100, 3636.
14. Jiang, B.; Cai-guang, Y.; Wen-Nan, X.; Wang, J. Bioorg. Med. Chem. 2001, 9,
1149.
N
18. The seven cytochrome P450’s tested were 1A2, 2B6, 2C8, 2C9, 2C19, 2D6 and
3A4.
15. ROCK 1 and ROCK 2 binding data was obtained using human recombinant
19. The percentage of parent compound remaining after 1 h was determined by
incubation with human liver microsomes for 60 min. The reported result is a
mean of two separate analyses performed in duplicate.
20. Tokushie, H.; Inatani, M.; Nemoto, S.; Sakai, H.; Katayama, K.; Uehata, M.;
Tanihata, H. Invest. Ophthalmol. Vis. Sci. 2007, 48, 3216.
rho kinase. Substrate and ATP concentrations used were 200 nM and 10 lM,
respectively, while the enzyme concentration was 3.96 ꢀ 10ꢁ3, units per well.
Enzyme reaction was allowed to proceed for 60 min at 23 °C (well volume of
20
lL) then 60 lL of the binding solution (IMAP kit provided by Molecular
Devices, Sunnyvale, CA) added to each well, and incubated for a further