Novel cinnoline-based inhibitors of LRRK2 kinase activity

Article abstract of DOI:10.1016/j.bmcl.2012.11.021

Leucine rich repeat kinase 2 (LRRK2) has been implicated in the pathogenesis of Parkinson's disease (PD). Inhibition of LRRK2 kinase activity is a therapeutic approach that may lead to new treatments for PD. Herein we report the discovery of a series of cinnoline-3-carboxamides that are potent against both wild-type and mutant LRRK2 kinase activity in biochemical assays. These compounds are also shown to be potent inhibitors in a cellular assay and to have good to excellent CNS penetration.

Full text of DOI:10.1016/j.bmcl.2012.11.021

Bioorganic & Medicinal Chemistry Letters 23 (2013) 71–74
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel cinnoline-based inhibitors of LRRK2 kinase activity
Albert W. Garofalo a, , Marc Adler b, Danielle L. Aubele a, Simeon Bowers a, Maurizio Franzini a,
⇑
Erich Goldbach c, Colin Lorentzen c, , R. Jeffrey Neitz a,à, Gary D. Probst a, Kevin P. Quinn c, Pam Santiago c,
Hing L. Sham a, Danny Tam d, Anh P. Truong a, Xiaocong M. Ye a, Zhao Ren d
a Department of Medicinal Chemistry, Elan Pharmaceuticals, 180 Oyster Point Blvd, South San Francisco, CA 94080, United States
b Department of Molecular Design, Elan Pharmaceuticals, 180 Oyster Point Blvd, South San Francisco, CA 94080, United States
c Department of Drug Metabolism and Pharmacokinetics, Elan Pharmaceuticals, 180 Oyster Point Blvd, South San Francisco, CA 94080, United States
d Department of Assay Development, Elan Pharmaceuticals, 180 Oyster Point Blvd, South San Francisco, CA 94080, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Leucine rich repeat kinase 2 (LRRK2) has been implicated in the pathogenesis of Parkinson’s disease (PD).
Inhibition of LRRK2 kinase activity is a therapeutic approach that may lead to new treatments for PD.
Herein we report the discovery of a series of cinnoline-3-carboxamides that are potent against both
wild-type and mutant LRRK2 kinase activity in biochemical assays. These compounds are also shown
to be potent inhibitors in a cellular assay and to have good to excellent CNS penetration.
Ó 2012 Elsevier Ltd. All rights reserved.
Received 9 September 2012
Revised 31 October 2012
Accepted 7 November 2012
Available online 16 November 2012
Keywords:
LRRK2
Leucine rich repeat kinase 2
Parkinson’s disease
Kinase inhibitors
Since the identification of the Park8 gene,1 evidence has been
mounting that implicates LRRK2 in the pathogenesis of Parkinson’s
disease (PD). At least five causal mutations in LRRK2 have been
identified.2 Of those, the G2019S mutation is the most prevalent
amino acid substitution being found in ꢀ5% of hereditary and
1–2% of idiopathic PD patients.3,4 This prevalence rises to much
higher levels in certain ethnic populations. The G2019S mutation
is located in the kinase domain of LRRK2 and has been observed
to cause an increase in kinase activity.5–7 This increased kinase
activity leads to broad neurotoxicity in both cellular and in vivo
models thus providing strong experimental evidence that supports
LRRK2 as a potential therapeutic target for PD.
We employed a kinase-inhibitor-focused screen of our in-house
library as an expedient way to find LRRK2 inhibitor chemotypes.
The screen used an HTRF assay measuring the inhibition of phos-
phorylation of LRRKtide.8 A series of 4-aminoquinoline-3-carbox-
amides (e.g., 1 and 2) were identified as LRRK2 inhibitors (Fig. 1).
The in vitro potency of 2 was encouraging and the large difference
in potency accompanying changes in substitution at C4 suggested a
manipulable SAR. However, compounds of this type are known
CSF1R inhibitors9 and in-house studies indicated that they were
substrates for P-gp and had poor CNS uptake. We were therefore
pressed to incorporate unique structural changes that would im-
part selectivity and improve the DMPK characteristics.
One approach that was examined involved replacing the quino-
line core with cinnoline. We reasoned that introduction of nitrogen
at C2 might improve CNS penetration by providing an additional
intramolecular hydrogen bond with the amide.10 These com-
pounds were generally prepared as shown in Scheme 1. Addition
of cyanoacetamide to the diazonium salt of 3-iodoaniline (3) affor-
ded diazene 4. AlCl3-promoted cyclization in toluene, in a sealed
tube, at 150 °C produced cinnoline 5 which was converted to chlo-
roamide 7 using a 3-step sequence.11 Addition of isopropylamine
followed by a Suzuki coupling gave the target cinnoline 11.
LRRK2 in vitro assay results from the quinoline inhibitors sug-
gested that substitution at C4 with small alkyl amines was pre-
ferred. The biochemical potencies for several analogs with
differing C4 alkyl substitution in the cinnoline series are listed in
Table 1. In general, wild-type potencies paralleled G2019S poten-
cies, although with slightly lower IC50 values. A notable exception
is compound 20 which exhibited better potency against the mu-
tant enzyme. Interestingly, homologation by a single carbon, com-
pound 16, reversed this selectivity. In vitro potency was greatest
with small, lipophilic amines at this position and the introduction
of a chiral center positively impacted the biochemical potency (cf.
13 and 15). Compounds 12 and 13 are two of the most potent com-
pounds that we prepared. We had previously observed a eudisimic
⇑
Corresponding author.
E-mail address: albert.garofalo@elan.com (A.W. Garofalo).
Present address: Novartis Institutes for Biomedical Research, Emeryville, CA
94608, United States.
à
Present address: Small Molecule Discovery Center, Univ. of California-San
Francisco, San Francisco, CA 98158, United States.
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.11.021

72
A. W. Garofalo et al. / Bioorg. Med. Chem. Lett. 23 (2013) 71–74
homology to LRRK2 as judged by the BLAST score (1e-25, with
31% identities). However, the 3DTC contained several ligand
induced conformational changes that were undesirable. These in-
cluded a collapsed P-loop and a partially disordered DFG sequence.
The P-loop was remodeled using homologous structures of kinases
with ATP analogues. The DYG sequence was built to resemble PKC-
iota (pdb ID: 3A8X), one of the few pdb kinase structures that con-
tain a DYG sequence.13 Docking of compound 11 into this model
shows a two-point hinge interaction with N1 of the cinnoline
accepting a hydrogen bond from A1950 and the carboxamide
donating a hydrogen bond to E1948 (Fig. 2). This model shows
the C7 substituent pointing out toward the solvent channel. Based
on this model several analogs with varying substitution at C6 and
C7 were examined (Table 2). We observed only small differences in
in vitro potency with identical substituents at either C6 or C7 as
would be expected for this region of the molecule (cf. 27 and 25,
28 and 29, 30 and 31).
NH
O
NH
O
5
8
4
6
NH2
NH2
3
2
7
N
N
S
S
1
2
O
O
O
O
LRRK2 IC50 = 781 nM
LRRK2 IC50 = 35 nM
Figure 1. Initial screening hits.
ratio (R/S) of 10 for chiral amines in the quinoline series (data not
shown) and prepared only the (R)-enantiomers in the cinnolines.
This seemed reasonable given the similar presumed hinge binding
orientations of both series (vide infra).
We developed a homology model of LRRK2 based on MLK1 (pdb
ID: 3DTC).12 At the time, the 3DTC structure had the closest
NH2
O
O
a, b
c
NH2
N
I
N
NH2
I
NH2
N
N
I
CN
3
4
5
OH
O
Cl
N
O
d
e, f
NH2
g
OH
NH2
+
N
N
I
N
I
8
6
7
NH
O
NH
O
B(OH)2
NH2
h
NH2
+
N
N
N
I
N
MeO2S
10
11
9
MeO2S
Scheme 1. Reagents and conditions: (a) NaNO2, 6 N HCl; (b) NaOAc, NCCH2CONH2 (78%, 2 steps); (c) AlCl3, toluene, 150 °C (95%); (d) KOH, dioxane, 110 °C; (e) SOCl2; (f)
NH4OH (45%, 3 steps); (g) EtOH, 140 °C (47%); (h) Pd(PPh3)4, K2CO3, dioxane, H2O (70%).
Table 1
Inhibitory activity against wild-type (wt) and mutant G2019S LRRK2
R
H
N
R
H
N
N
N
O
N
O
NH2
NH2
N
N
S
A
B
O
O
Compound
Series
R
LRRK2 (wt) IC50
(lM)
LRRK2 (G2019S) IC50 (lM)
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
A
A
A
A
A
A
A
A
A
A
A
A
B
B
B
B
Isopropyl
(R)-sec-Butyl
(R)-1-Cyclopropylethyl
Cyclobutyl
Cyclopropylmethyl
Ethyl
0.015
0.001
0.001
0.010
0.017
0.046
0.023
0.049
0.017
0.119
0.142
0.965
0.121
0.145
0.015
0.002
0.027
0.004
0.007
0.019
0.027
0.055
0.032
0.052
0.053
0.056
0.298
1.495
0.162
0.264
0.016
0.006
Tetrahydropyran-4-yl
Cyclopropyl
(+/À)-Tetrahydrofuran-3-yl
Methyl
H
(R)-1-Phenylethyl
3-Hydroxypropyl
3-Hydroxyethyl
Isopropyl
(R)-1-Cyclopropylethyl
Enzyme activity was determined using GST-LRRK2 (970–2527), GST-(G2019S) LRRK2 (970–2527), and LRRKtide in the presence of 100 lM ATP.

A. W. Garofalo et al. / Bioorg. Med. Chem. Lett. 23 (2013) 71–74
73
Table 3
Cellular potencies
Compound
LRRK2 (G2019S) EC50 (lM)
11
12
13
25
26
28
34
0.218
0.073
0.070
0.242
0.062
0.109
0.290
Table 4
Inhibitory activity against CSF1R and PDGFR
a
Compound
CSF1R IC50
(l
M)
PDGFRa IC50 (lM)
13
14
11
33
0.017
<0.040
0.012
0.054
0.022
0.127
0.262
0.022
Figure 2. LRRK2 homology model with compound 9 docked.
Table 2
Inhibitory differences with substitution at C6 and C7
CNS distribution of compounds was evaluated in both the Mdr1
a/b (À/À) and P-gp competent wild-type FVB mice to obtain in vivo
P-gp efflux and brain uptake results. Gratifyingly, good to excellent
CNS penetration was observed at 5 min following a 1 mg/kg IV
dose. Penetration was assessed from the brain concentration to
plasma concentration ratios (Kp) in the wild-type mice and was
0.76, 1.50 and 11.4 for compounds 14, 26 and 33, respectively.
Negligible in vivo P-gp efflux was also measured and was consis-
tent with values obtain in vitro from MDR1-MDCK transfected
cells.
NH
O
NH2
R
N
N
Compound
R
LRRK2 (wt)
LRRK2 (G2019S)
IC50 M)
IC50
(
lM)
(l
27
25
28
29
30
31
32
33
34
6-(Pyridin-4-yl)
7-(Pyridin-4-yl)
6-(1H-Pyrazol-4-yl)
7-(1H-Pyrazol-4-yl)
6-(1-Methylpyrazol-4-yl)
7-(1-Methylpyrazol-4-yl)
7-(4-Methylpiperazin-1-yl)
7-(4-Morpholinophenyl)
7-(Thiazol-4-yl)
0.030
0.015
0.001
0.012
0.378
0.009
0.053
0.009
0.014
0.066
0.016
0.005
0.018
0.425
0.012
0.059
0.012
0.020
We selected several potent compounds for selectivity screening
against a panel of 40 kinases.15 Unfortunately we were disap-
pointed to find that most of the compounds tested were fairly pro-
miscuous kinase inhibitors. Inhibition of both CSF1R and PDGFR
a
was of particular concern and dose responses were determined
for these kinases to confirm the observations of the broad selectiv-
ity screen (Table 4).16 Another clear result from this screen was
that analogs substituted at C6 as opposed to C7 were much more
promiscuous inhibitors.
In summary, we have designed and synthesized a novel series of
cinnoline carboxamides that are potent and CNS penetrant LRRK2
inhibitors. A homology model of LRRK2 binding was developed
and an inhibitor binding mode postulated based on this model.
Close interactions indicated by the model were tested by the prep-
aration of specific analogs. However, the inhibitors presented here
exhibited less than desirable kinase specificity and as a result
further medicinal chemistry efforts have been suspended.
Additionally, we tested a few N-methylamides as well as substi-
tution at C8 (Fig. 3). Secondary amides were substantially less po-
tent, likely due to a steric interaction with the methionine
gatekeeper. As expected substitution at C8 was not tolerated due
to the proximity of this position with the hinge.
We were interested in obtaining a cellular measurement of the
ability of these compounds to inhibit LRRK2 kinase activity. It has
been demonstrated that inhibition of LRRK2 kinase decreases phos-
phorylation of S935 and that this decreased phosphorylation can be
used as a measure of LRRK2 kinase activity.14 Our cellular assay em-
ployed HEK293 cells stably transfected with LRRK2 (G2019S). Sev-
eral compounds exhibited better than 100 nM potency and we
observed good correlation with in vitro results (Table 3).
Acknowledgments
The authors thank Wayman Chan (Elan) for performing difficult
purifications. We would also like to acknowledge the work of
NH
O
NH
O
NH
O
N
H
N
H
NH2
N
N
N
N
N
N
O
O
O
O S
N
S
35
LRRK2(wt) IC50 = 0.412 µM
36
LRRK2(wt) IC50 = 0.168 µM
37
LRRK2(wt) IC50 = 3.199 µM
µ
LRRK2(G2019S) IC50 = 2.512 M
µ
µ
LRRK2(G2019S) IC50 = 0.418
M
LRRK2(G2019S) IC50 = 0.247
M
Figure 3. Lower in vitro potencies observed with N-methylamides and methyl substitution at C8.

74
A. W. Garofalo et al. / Bioorg. Med. Chem. Lett. 23 (2013) 71–74
7. Jaleel, M.; Nichols, R. J.; Deak, M., et al Biochem. J. 2007, 405, 307.
8. Homogeneous time-resolved fluorescence (HTRF) assay using GST-LRRK2.
Detection was done using biotinylated LRRKtide, Eu(K)-Ab (p-Moesin)/SA-
XL665 measured at 620 and 665 nm.
Pui Seto, Monica You, Jeanne Baker, and Kevin Tanaka for the pro-
duction and purification of antibodies used in our assays.
9. Smalley, T. L., Jr.; Chamberlain, S. D.; Mills, W. Y., et al Bioorg. Med. Chem. Lett.
2007, 17, 6257.
Supplementary data
10. It has been speculated that intramolecular hydrogen bonds can account for the
high permeability of compounds with non-optimal physicochemical
properties. See: Alex, A.; Millan, D. S.; Perez, M., et al Med. Chem. Commun.
2011, 2, 669.
11. Scott, D. A.; Dakin, L. A.; Del Valle, D. J., et al Bioorg. Med. Chem. Lett. 2011, 21,
1382.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.
11.021.
12. Hudkins, R. L.; Diebold, J. L.; Tao, M., et al J. Med. Chem. 2008, 51, 5680.
13. Takimura, T.; Kamata, K.; Fukasawa, K., et al Acta Crystallogr. D. Biol. Crystallogr.
2010, D66, 577.
References and notes
1. Funayama, M.; Hasegawa, K.; Kowa, H., et al Ann. Neurol. 2002, 51, 296.
2. Greggio, E.; Cookson, M. R. ASN Neurochem. 2009, 1, 13.
3. Healy, D. G.; Falchi, M.; O’Sullivan, S. S., et al Lancet Neurol. 2008, 7, 583.
4. Latourell, J. C.; Sun, M.; Lew, M. F., et al BMC Med. 2008, 6, 32.
5. West, A. B.; Moore, D. J.; Biskup, S., et al Proc. Nat. Acad. Sci. 2005, 102, 16842.
6. Greggio, E.; Jain, S.; Kingsbury, A., et al Neurobiol. Dis. 2006, 23, 329.
14. Dzamko, N.; Deak, M.; Hentati, F., et al Biochem. J. 2010, 430, 405.
15. Compounds were profiled at Invitrogen in
Supplementary data).
a single-point assay (see
16. Dose responses for CSF1R and PDGFR
a were generated in-house using a
peptide-based TR-FRET assay in the presence of 1 mM ATP.