Kinase domain inhibition of leucine rich repeat kinase 2 (LRRK2) using a [1,2,4]triazolo[4,3-b]pyridazine scaffold

Article abstract of DOI:10.1016/j.bmcl.2014.07.052

Leucine rich repeat kinase 2 (LRRK2) has been genetically linked to Parkinson's disease (PD). The most common mutant, G2019S, increases kinase activity, thus LRRK2 kinase inhibitors are potentially useful in the treatment of PD. We herein disclose the str

Full text of DOI:10.1016/j.bmcl.2014.07.052

Bioorganic & Medicinal Chemistry Letters 24 (2014) 4132–4140
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Kinase domain inhibition of leucine rich repeat kinase 2 (LRRK2)
using a [1,2,4]triazolo[4,3-b]pyridazine scaffold
a
a
b
b
Paul Galatsis a, , Jaclyn L. Henderson , Bethany L. Kormos , Seungil Han , Ravi G. Kurumbail ,
Travis T. Wager a, Patrick R. Verhoest a, G. Stephen Noell c, Yi Chen d, Elie Needle d, Zdenek Berger d,
Stefanus J. Steyn e, Christopher Houle f, Warren D. Hirst d
⇑
a Worldwide Medicinal Chemistry, Pfizer Worldwide R&D, 610 Main St., Cambridge, MA 02139, United States
b Worldwide Medicinal Chemistry, Pfizer Worldwide R&D, Eastern Point Rd., Groton, CT 06340, United States
c Primary Pharmacology Group, Pfizer Worldwide R&D, Eastern Point Rd., Groton, CT 06340, United States
d Neuroscience Research Unit, Pfizer Worldwide R&D, 610 Main St., Cambridge, MA 02139, United States
e Pharmacokinetics, Dynamics, and Metabolism, Pfizer Worldwide R&D, 610 Main St., Cambridge, MA 02139, United States
f Drug Safety R&D, Pfizer Worldwide R&D, Eastern Point Rd., Groton, CT 06340, 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 genetically linked to Parkinson’s disease (PD). The most
common mutant, G2019S, increases kinase activity, thus LRRK2 kinase inhibitors are potentially useful
in the treatment of PD. We herein disclose the structure, potential ligand–protein binding interactions,
and pharmacological profiling of potent and highly selective kinase inhibitors based on a triazolopyrid-
azine chemical scaffold.
Received 16 July 2014
Accepted 18 July 2014
Available online 30 July 2014
Keywords:
Leucine rich repeat kinase 2
LRRK2
Ó 2014 Elsevier Ltd. All rights reserved.
Kinase inhibitors
Triazolopyridazine
Parkinson’s disease (PD) is the most common movement disor-
der and the second most common neurodegenerative disorder
after Alzheimer’s disease (AD).1 Significant excitement has been
generated by recent genome-wide association studies (GWAS) that
linked the PARK8 mutation, encoding the leucine rich repeat kinase
2 (LRRK2) protein, to PD.2 LRRK2 is a very large protein (2527
amino acids; 286 kDa MW) comprised of multiple domains
(Fig. 1).3 The most common LRRK2 mutation is G2019S (GS) which
is in the activation loop (DFG motif-DYG in LRRK2) of the kinase
domain.4 This mutation has been reported to increase the kinase
activity of LRRK2, thus the identification of potent, selective, brain
penetrant kinase inhibitors could dampen this hyperactivity and
be of value in the treatment of PD.5
LRRK2 linkage to PD is a very recent discovery and a significant
amount of LRRK2 biology relating to PD has yet to be elucidated.6
One key gap is our lack of understanding of the exact role of LRRK2
in PD. In an effort to gain insight into its PD role, various groups
have explored LRRK2 biology with re-purposed kinase inhibitors.7
Other groups sought to identify novel chemical tools with a goal of
improving kinase specificity.8 Utilization of modestly selective tool
compounds to probe the biology of LRRK2 can be problematic in
that it generates data that is difficult to interpret, as one needs to
de-convolute on-target from off-target pharmacology.9 Although
the discovery of potent and selective LRRK2 kinase inhibitors
may be challenging and require significant investment of
resources, they should allow for a robust understanding of the bio-
logical consequences of LRRK2 inhibition. Thus, in our efforts to
develop LRRK2 kinase inhibitors, potency, selectivity, in vivo effi-
cacy and brain penetration were all tracked in an effort to generate
Abbreviations: AD, Alzheimer’s disease; ANK, ankyrin repeat domain; ARM,
armadillo repeat domain; AUC, area under the curve; BA, brain availability; BBB,
blood brain barrier; BCRP, breast cancer related protein; BI, brain impairment; CSD,
Cambridge Structural Database; CNS, central nervous system; COR, C-terminal of
ROC domain; ER, efflux ratio; GS, LRRK2 G2019S mutation; GWAS, genome-wide
association studies; HLM, human liver microsome; HTS, high throughput screening;
KO, knockout; Kp,uu, unbound brain to unbound plasma concentration ratio; KSS,
kinase selectivity screening; LE, ligand efficiency; LipE, lipophilic ligand efficiency;
LRR, leucine rich repeat domain; LRRK2, leucine rich repeat kinase 2; MDR1, multi-
drug resistance protein 1 (human P-gp); MPO, multi-parameter optimization; PD,
Parkinson’s disease; PDB, protein data bank; PK/PD, pharmacokinetic/pharmaco-
dynamic; ROC, ras of complex domain; SAR, structure activity relationship; WCA,
whole cell assay based on pS935 readout; WD40, WD40 repeat domain; WT, wild
type.
⇑
Corresponding author. Tel.: +1 617 395 0709.
E-mail address: paul.galatsis@pfizer.com (P. Galatsis).
http://dx.doi.org/10.1016/j.bmcl.2014.07.052
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

P. Galatsis et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4132–4140
4133
Figure 1. Multi-domain structure of LRRK2.
and found to be roughly equipotent with the wt isoform (within
3Â). For all compounds tested in this series, we found a 5–10Â
rightward shift in potency in going from the cell free to whole cell
assays10 despite the good potency at cellularly relevant ATP
concentrations and excellent passive permeability.
50 µM ATP LRRK2 (WT) IC50 = 40 nM
LE = 0.39; LipE = 3.6; CNS MPO = 4.82
1 mM ATP LRRK2 (WT) IC50 = 173 nM
1 mM ATP LRRK2 (GS) IC50 = 82 nM
HLM CLIA,S = 159 mL/min/kg
OMe
OMe
N
N
N
N
S
P
app AB = 12 x 10-6 cm/sec
While our hit-to-lead efforts were ongoing, we were very inter-
ested in better understanding the potential binding interactions of
these compounds. In the absence of LRRK2 crystallographic infor-
mation, we employed a surrogate crystallography approach based
on kinase similarity and cross-over of compound activity. Though
LRRK2 only has ꢀ30% residue identity and ꢀ50% similarity in the
overall kinase domain to its closest neighbors, the residues in its
ATP-binding site pocket have greater similarity to a number of
other kinases. For instance, tyrosine kinase 2 (Tyk2) ATP-binding
site residues are 74% similar to those in LRRK2. In addition, there
was some cross-over activity of this series of triazolopyridazine
compounds with the JAK family of kinases (vide infra), suggesting
Tyk2 is a reasonable surrogate crystallographic system for LRRK2.
With crystals of Tyk2 readily available from a previous project,
we pursued soaking studies of 15 with Tyk2 to see if it could act as
a model system and provide some insight into the potential bind-
ing interactions. Figure 3 shows the X-ray crystal structure of 15 in
Tyk2 (PDB ID: 4PY1) highlighting several protein-ligand interac-
tions. Compounds of this scaffold appear to make a single point
interaction with kinase hinge via the C1 N-atom of the triazole moi-
ety, with the Me-pyrrazole occupying the ribose pocket (towards
solvent) and not a position adjacent to gatekeeper, in contrast to
a binding mode previously suggested for this chemotype.8i
Having identified optimal C3 substituents, we next focused
attention on the C6-S substituent as it was a potential metabolic
soft spot and safety liability (potential for quinone formation).
Table 2 provides a sampling of substituents explored at this posi-
tion. Compounds 17 and 18 provided evidence that the two meth-
oxy groups of the phenyl ring were having a synergistic effect, in
that, independently they provided weak potency to the scaffold
but together, as in 1, resulted in a dramatic increase in kinase inhi-
bition. This is somewhat in contrast to the SAR reported by Elan,
whereby they demonstrated LRRK2 potency with solely a meta
substituted aryl at this position.8i The binding pose (Fig. 3) allows
one to speculate if an interaction of the two methoxys with the P-
loop and floor of the ATP site is required for good potency
(observed with tofacitinib). This bioactive conformation could be
favored by the ortho-OMe twisting the aryl ring orthogonal to
the plane of the triazolopyridazine core. In support of this hypoth-
esis, the ortho-OMe was found to be required, whereas, small sub-
stituents could be tolerated at the meta-position, for example, 19
and 20. While 20 maintained good predicted brain availability, this
change did increase clearance compared to 15. Conversely, 19
showed a modest decrease in clearance but now had the potential
for P-gp efflux susceptibility. Clearance could be greatly improved
as 23 demonstrated, by modulating the potential phenyl metabolic
soft spots, but could not be coupled with potency.
MDR1 ER (BA/AB) = 1.77
KSS = 0/39 hits at 1 µM
1
Figure 2. Data summary of HTS hit 1.
the most efficient inhibitors for potential treatment of PD. In this
Letter, we disclose our optimization of a triazolopyridazine scaffold
targeting the LRRK2 kinase domain.
A high throughput screen (HTS) of about 750,000 compounds
allowed us to identify multiple scaffolds including triazolopyrid-
azine 1 as a potent inhibitor of LRRK2, which bears high similarity
to a hit recently described by scientists at Elan (Fig. 2).8i It exhib-
ited double-digit nM potency in a LANTHA screen assay at the
ATP KM (50 l
M) of LRRK2.10 Shifting the ATP concentration to
one more physiologically relevant (1 mM) resulted in a modest
rightward shift in potency for both the WT and the GS mutation.10
Along with its desirable in vitro potency, 1 exhibited physicochem-
ical properties consistent with CNS drug space (e.g. high LE, LipE
and good CNS MPO score).11 This observation was buttressed by
in vitro data showing excellent kinase selectivity,12 high passive
permeability, and a lack of P-gp transporter-mediated membrane
asymmetry (good potential for brain availability). The HLM clear-
ance was found to be high, as one might predict for a compound
susceptible to an O-demethylation metabolic clearance pathway.
As part of our hit-to-lead approach, strict adherence to design-
ing compounds with an eye to physicochemical properties was fol-
lowed. Given that treating PD would involve a long-term, chronic
regimen, minimizing daily dose and maximizing the safety profile
were critical attributes that could be readily achieved, at the design
stage, using this strategy. The CNS MPO score was a relatively
straightforward yardstick to monitor our ability to align drug-like
properties and all compounds presented in this Letter met or
exceeded the cut-off for CNS drug-like space.11c In examining the
sources of potential diversity within the di-substituted triazolo-
pyridazine scaffold, 4 points of diversity were readily apparent;
the 2 substituent moieties (at C3 and C6), the S-linker atom, and
the hinge interaction core itself.
Our SAR optimization began with the C3 substituent. Table 1
provides a summary of some of the compounds prepared in the
evaluation of this position. While small alkyls (7), saturated het-
ero-cycles (8–9), and 6-membered ring heteroaromatics (2–6)
showed varying degrees of potency, the 5-membered ring hetero-
aromatics (11–16) proved to be optimal for this diversity vector. In
particular, methyl-pyrazole 15 became an obvious standout. Not
only does this compound have improved potency (and LipE) but
it also is the only compound that significantly reduced clearance
whilst maintaining potency. Adding a methylene spacer to this
moiety (10) resulted in a dramatic loss in potency. While not for-
mally presented in the tables, G2019S IC50 values were determined
We were concerned the S-linker may be a metabolic liability
and explored options for its replacement. While all of the com-
pounds examined were predicted to be in good CNS space
(MPO > 4), all linker replacements examined (ether 24, amino 25,

4134
P. Galatsis et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4132–4140
Table 1
SAR data for substitution at C3 of [1,2,4]triazolo[4,3-b]pyridazine
N
N
N
OMe
R
N
S
MeO
a
e
Compd
R
LRRK2 IC50 (nM)
LE/LipEb
LRRK2 WC IC50 (nM)c
HLMd
159
Papp
P-gpf
1
173
0.36/3.10
1380
12
1.77
N
N
2
3
496
242
0.33/3.84
0.35/4.36
1310
736
104
86
20
21
1.47
1.27
4
5
6
349
880
66
0.34/4.21
0.32/4.58
0.35/4.51
nt
77
nt
26
nt
1.25
nt
N
N
nt
N
N
440
137
17
1.77
OMe
7
8
2290
2489
0.33/3.36
0.31/4.32
nt
nt
146
28
17
32
0.70
1.51
O
9
12,600
0.26/3.49
nt
nt
7.4
4.46
N
N
N
10
11
12
13
32,000
804
91
0.23/2.97
0.33/4.09
0.38/4.38
0.38/4.76
nt
nt
30
nt
nt
4330
722
660
30
20
26
1.16
1.66
1.65
N
O
195
71
N
S
95
S
N
N
N
14
15
16
224
64
0.35/4.57
0.38/5.32
0.33/4.65
2760
387
116
42
19
34
16
2.17
2.22
1.80
N
N
N
323
1260
106
N
a
b
c
d
e
f
Geometric mean of n P 2 for wild-type LRRK2 at 1 mM ATP.
Calculated from 1 mM ATP IC50 using clogP values.
Geometric mean of n P 2 for whole cell pS935 determinations in wild-type LRRK2.
Human liver microsomal clearance (mL/min/kg).
Passive permeability (AB Â 10À6 cm/s).
MDR1 Efflux Ratio (BA/AB). nt = not tested.

P. Galatsis et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4132–4140
4135
>90°).8i Though we agree that the dihedral angle is important for
interactions of the phenyl substituents with the pocket, we also
believe the C–X–C angle (X = S, O, NH or CH2) is important in
obtaining optimal interactions with the Gly-rich loop and the rest
of the binding pocket. Based on an analysis of structures in the
Cambridge Structural Database (CSD)13 using Mogul14, the average
aromatic C–S–aromatic C angle is ꢀ103°, whereas when O, NH or
CH2 are substituted for S, the average angles are 120°, 128° and
114°, respectively. In addition, the indicated binding pose (Fig. 3)
revealed a favorable S-S non-bonded interaction between the
ligand and the Met978 gatekeeper, which is conserved in LRRK2.
The S-linker alternatives could be generating steric clashes with
the gatekeeper Met.
The final feature of this scaffold subjected to SAR analysis was
the core framework and the effect this may have on hinge binding
(Table 4). Adding methyl substituents (as in compounds 32 and 33)
resulted in a dramatic loss of potency. Consistent with the binding
pose (Fig. 3), Me or larger substituents at C7 would cause a clash
with the Met1947 gatekeeper residue and substituents at C8 would
severely clash with the Glu1948 backbone carbonyl at the hinge
(LRRK2 numbering).
Figure 3. Crystal structure of Tyk2 in complex with 15 (4PY1).
methylene 27, hydroxymethyl 28, keto 29, difluoromethyl 30, and
NMe 31) did not show any improvement upon the S-linker
(Table 3). It has been proposed, based on homology modeling, that
the thioether dihedral (ꢀ90°) provides the requisite trajectory for
this substituent compared to O, NH, or CH2 linkers (dihedral
Additional support for the bioactive conformation depicted in
Figure 3 can be inferred from 34. Substituting the pyridazine
N-atom for a C-atom resulted in a significant decrease in potency
Table 2
SAR data for substitution at C6-S of [1,2,4]triazolo[4,3-b]pyridazine
N
N
N
R1
N
S
R2
a
c
e
Compd
R1
R2
LRRK2 IC50 (nM)
LE/LipEb
na
LRRK2 WC IC50 (nM)
HLMd
164
Papp
P-gpf
OMe
17
1070g
nt
nt
7.3
3.8
1.51
18
19
20
21
2110g
64
na
189
28
1.72
3.87
1.97
1.60
MeO
NC
Cl
OMe
0.38/5.65
0.39/4.47
0.37/2.73
2070
245
19
N
N
N
N
N
N
N
N
OMe
Cl
78
113
72
20
10
373
1800
Cl
OMe
22
23
1350
0.30/2.66
0.26/4.43
nt
nt
221
<8
15
21
1.76
1.50
OMe
12,635
N
N
N
MeO
N
a
b
c
d
e
f
Geometric mean of n P 2 for wild-type LRRK2 at 1 mM ATP.
Calculated from 1 mM ATP IC50 using clogP values.
geometric mean of n P 2 for whole cell pS935 determinations in wild-type LRRK2.
human liver microsomal clearance (mL/min/kg).
passive permeability (AB Â 10À6 cm/s).
MDR1 Efflux Ratio (BA/AB). nt = not tested.
g
geometric mean of n > 2 for wild-type LRRK2 at KM ATP.

4136
P. Galatsis et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4132–4140
Table 3
Linker-atom substitution of [1,2,4]triazolo[4,3-b]pyridazine
N
N
N
OMe
R
N
X
MeO
a
c
e
Compd
R
X
LRRK2 IC50 (nM)
LE/LipEb
LRRK2 WC IC50 (nM)
HLMd
168
Papp
P-gpf
24
–O–
236
0.35/3.35
1740
nt
15
1.26
25
26
–NH–
–O–
22,265g
549
na
250
nt
2.47
nt
2.78
nt
0.33/4.78
nt
nt
nt
nt
nt
nt
N
N
N
N
N
N
N
N
N
N
N
N
27
28
29
30
31
–CH2–
–CHOH–
–CO–
6202
9124
1211
763
0.27/3.41
0.25/5.00
0.30/4.65
0.30/4.62
0.24/2.65
18
<8
60
28
13
19
14
16
22
14
1.59
7.14
2.02
1.46
1.58
–CF2–
–NMe–
21,858
a
b
c
d
e
f
Geometric mean of n P 2 for wild-type LRRK2 at 1 mM ATP.
Calculated from 1 mM ATP IC50 using clogP values.
Geometric mean of n P 2 for whole cell pS935 determinations in wild-type LRRK2.
Human liver microsomal clearance (mL/min/kg).
Passive permeability (AB Â 10À6 cm/s).
MDR1 Efflux Ratio (BA/AB). nt = not tested.
g
geometric mean of n > 2 for wild-type LRRK2 at KM ATP.
relative to 15. This could be attributed to the greater rotational
flexibility of the C6 substituent provided by the loss of a key non-
bonding electron-electron repulsion present in 15.
kinases that have a larger amino acid side chain at that position,
thus giving rise to the loss of selectivity.
Having identified 15 as a potent LRRK2 inhibitor with high
kinome selectivity from internal profiling (0/39 hits observed at
Modifying the scaffold core by manipulating the N-atom substi-
tution pattern resulted in mixed results. Most of the changes gave
rise to a loss in potency relative to 1, however, 35 and 36, the imi-
dazo-pyridazine and pyrrazolopyrimidine cores, respectively, gen-
erated analogs that were equipotent to 15. While these analogs
maintained the predicted good brain availability (Papp and P-gp
data), 35 and 36 did not provide any improvement in HLM clear-
ance. Whereas, 15 did not exhibit significant activity in our internal
1
l
M), we sought to further extend our kinome analysis of this
compound. Evaluating 15 in the DiscoverRx Kinomescan panel
(Fig. 4) at 1 M resulted in a selectivity score of S(35) = 0.008
l
(number of non-mutant kinases with % control < 35/number of
non-mutant kinases tested where positive control = 0% and
DMSO = 100%). This was the result of inhibiting, in addition to
LRRK2 (18% of control), Tyk2 (19% of control) and JAK3 (29% of con-
trol) for a total of 3 out of 392 unique kinases.
KSS panel (i.e., >50 % inhibition at 1 lM), both 35 and 36 exhibited
a 9/39 hit rate. While the overlap of these nine kinase hits for the
compounds was not identical, the dramatic change in promiscuity
was not anticipated. The binding pose (Fig. 3) indicated the N2-
atom points towards the backbone carbonyl of Ala1950, which,
due to its small side chain, may be more flexible to accommodate
a water-mediated H-bond to this N-atom (observed in crystallo-
graphic data). The C–H at C2 of 35 and 36 will likely form a non-tra-
ditional H-bond to the backbone carbonyl of Ala1950, as well as
We continued to profile 15 and, in particular, were very inter-
ested in its brain availability (BA). The in vitro permeability and
P-gp ER data were predictive of good BA (Table 1). Figure 5 shows
the free brain and plasma drug concentrations vs. time plot from
which the rat BA for 15, dosed at 10 mg/kg, was determined.15
The plot illustrates the instantaneous equilibration between brain
and plasma at each time point. The AUC(0-Tlast)-based Kp,uu of 0.14
indicated there was 7Â brain impairment (BI) in the rat. This

P. Galatsis et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4132–4140
4137
Table 4
Core variation on [1,2,4]triazolo[4,3-b]pyridazine
N
A
A
A
A
OMe
A
N
S
N
MeO
a
c
e
Compd
Core
LRRK2 IC50 (nM)
LE/LipEb
LRRK2 WC IC50 (nM)
HLMd
56
Papp
P-gpf
N
N
32
>32,000
>32,000
1461
69
0.23/2.12
nt
13
2.57
1.83
1.56
1.10
2.07
2.39
1.04
N
N
N
N
N
33
34
35
36
37
38
0.23/2.12
0.31/3.07
0.38/4.31
0.38/4.18
0.33/3.13
0.32/2.84
nt
38
20
17
12
10
nt
N
N
N
nt
57
N
N
400
488
5240
5850
156
97
N
N
N
N
N
N
N
N
64
N
N
525
221
110
N
N
N
661
17
39
40
41
1145
115
0.31/2.16
0.35/4.12
0.35/3.98
nt
206
10
13
10
1.51
1.00
1.52
N
N
N
945
759
>300
>300
N
N
N
N
110
a
b
c
d
e
f
Geometric mean of n P 2 for wild-type LRRK2 at 1 mM ATP.
Calculated from 1 mM ATP IC50 using clogP values.
Geometric mean of n P 2 for whole cell pS935 determinations in wild-type LRRK2.
Human liver microsomal clearance (mL/min/kg).
Passive permeability (AB Â 10À6 cm/s).
MDR1 Efflux Ratio (BA/AB). nt = not tested.
apparent in vitro/in vivo disconnect was further investigated. We
had already shown that 15 was not a P-gp substrate (Table 1). BCRP
(breast cancer related protein) is another transporter expressed at
the BBB that could also contribute to this BI. Submitting 15 to our
mouse BCRP assay led to the conclusion it was a substrate (Table 5).
Interestingly, the BA found for 15 stands in contrast to the BA
found for 35 (Kpu,u = 0.42) or 36 (Kpu,u = 0.27) and could be poten-
tially rationalized by the observation they were neither a P-gp or
BCRP substrates. It is unclear whether this rat BI would translate
to humans as it is known that the level of transporter expression
at the BBB is species dependent.16
revealed kidney and lung phenotypes.17 We decided to conduct a
PK/PD study with 15 where it was orally dosed in mice at 0, 30,
and 300 mg/kg/day for 14 days (5 males/dose). Endpoints included
a full hematologic and clinical chemistry profile on day 14 along
with a macroscopic examination at necropsy and a microscopic
examination of selected tissues (kidney, liver, heart, lung, spleen,
and thymus). Table 6 summarizes the data generated from that
study indicating the exposures achieved and the findings observed.
The result supported the highly selective nature of this compound,
in that no treatment related findings were observed at high
exposures.
Although brain impaired, we felt the excellent kinase selectivity
profile of 15 would prove useful for probing the peripheral safety
liabilities of LRRK2 kinase inhibition as LRRK2 KO animals have
We were disappointed to learn of the results from an in vivo
determination of the pS935 endpoint18 in brain and kidney. As
one might expect for a compound that is brain impaired, 15

4138
P. Galatsis et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4132–4140
Table 6
Safety findings from a multi-dose, 14 day mouse study
Dose
(mg/kg)
Cmax
(ng/mL)
AUC(0–24)
(ng h/mL)
Findings
0
30
300
—
200
12,870
—
350
20,900
No treatment-related findings
No treatment-related findings
No treatment-related findings
Cl
N
N
O
Cl
42
H2N
Cl
N
H
R2
NH2NH2
Cl
Cl
N
N
N
CDI
N
N
O
N
O
HN
N
H
43
R2
HN
N
NH
45
NH2
44
1. R2CO2H
2. POCl3
POBr3
Figure 4. TREEspot™ visualization of kinome selectivity for 15 at 1 lM.
H
Cl
1000.0
Cl
Mean Cb,u (nM)
Suzuki
R2
N
N
N
N
Mean Cp,u (nM)
Br
N
N
100.0
N
46
N
47
SNAr
10.0
1.0
R1
N
N
N
R2
0
2
4
6
N
48
Time (h)
Figure 5. Plot of rat free brain and plasma vs time for 15.
Scheme 1. Synthetic route for preparing the triazolopyridazines.
behaved similar to LRRK2-IN-1,8a in that no effect on the pS935
levels was found even after 14 days at 300 mg/kg/day. More sur-
prising was a similar observation for the kidney. Despite the highly
perfused nature of the kidney and the determination that 15 exhib-
ited a Ckidney/Cplasma ꢀ 2, no change in pS935 levels were observed
at any dose.
Synthetically, triazolopyridazines 48 could be prepared in a
straightforward manner from commercially available 42.8i,21
Scheme 1 outlines routes for this conversion. Exposure of 42 to
hydrazine afforded 44 which could be converted in 2 steps to 46
by amide formation followed by ring closure. Alternatively, the
overall conversion of 42 to 46 could be conducted by displacement
of the chloride with the hydrazoic acid followed by ring closure via
43. To potentially enable library protocols, dihalide 47 could be
prepared from 44 by ring closure and treatment of the resulting tri-
azinone 45 with POBr3. Using the appropriate nucleophiles
resulted in the conversion of 46 to 48. The generation of 48 from
47 could be achieved by sequential Suzuki cross-coupling and SNAr
reaction through the intermediary 46.
Table 5
Mouse BCRPa data for selected compounds
b
Compd
Induced (Dox)
Non-induced
Ratio of ratiosc
AB (Â10À6
)
BA (Â10À6
)
ER
AB (Â10À6
)
BA (Â10À6
)
ER
15
35
36
10
25
29
39
28
27
3.9
1.1
0.9
33
24
29
28
24
25
0.9
1
0.9
4.5
1.1
1.1
a
b
c
BCRP transfected cells (n = 3) using prazosin as positive control and quinidine as negative control.
BCRP induced with doxycycline.
Substrate defined as having a ratio of ratios significantly higher than 1, calculated by ER(induced)/ER(non-induced).

P. Galatsis et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4132–4140
4139
Mountain, J. L.; Goldman, S. M.; Tanner, C. M.; Langston, J. W.; Wojcicki, A.;
Eriksson, N. PLoS Genet. 2011, 7, e1002141.
3. Anand, V. S.; Braithwaite, S. P. FEBS J. 2009, 276, 6428.
4. (a) Tsika, E.; Moore, D. J. Curr. Neurol. Neurosci. Rep. 2012, 12, 251; (b) Kramer,
T.; LoMonte, F.; Goering, S.; Amombo, O.; Marlyse, G.; Schmidt, B. ACS Chem.
Neurosci. 2012, 3, 151.
5. (a) Smith, W. W.; Pei, Z.; Jiang, H.; Dawson, V. L.; Dawson, T. M.; Ross, C. A. Nat.
Neurosci. 2006, 9, 1231; (b) Covy, J. P.; Giasson, B. I. Biochem. Biophys. Res.
Commun. 2009, 378, 473; (c) Nichols, R. J.; Dzamko, N.; Hutti, J. E.; Cantley, L. C.;
Deak, M.; Moran, J.; Bamborough, P.; Reith, A. D.; Alessi, D. R. Biochem. J. 2009,
424, 47; (d) Gilsbach, B. K.; Ho, F. Y.; Vetter, I. R.; van Haastert, P. J. M.;
Wittinghofer, A.; Kortholt, A. PNAS 2012, 109, 10322.
6. (a) Sen, S.; Webber, P. J.; West, A. B. J. Biol. Chem. 2009, 284, 36346; (b) Berger,
Z.; Smith, K. A.; LaVoie, M. J. Biochem. 2010, 49, 5523; (c) Devine, M. J.;
Kaganovich, A.; Ryten, M.; Mamais, A.; Trabzuni, D.; Manzoni, C.; McGoldrick,
P.; Chan, D.; Dillman, A.; Zerle, J.; Horan, S.; Taanman, J.-W.; Hardy, J.; Marti-
Masso, J.-F.; Healey, D.; Schapira, A. H.; Wolozin, B.; Bandopadhyay, R.;
Cookson, M. R.; van der Brug, M. P.; Lewis, P. A. PLoS One 2011, 6, e22489; (d)
Taymans, J.-M.; Vancraenenbroeck, R.; Ollikainen, P.; Beilina, A.; Lobbestael, E.;
De Maeyer, M.; Baekelandt, V.; Cookson, M. R. PLoS One 2011, 6, e23207.
7. Lee, B. D.; Shin, J.-H.; VanKampen, J.; Petucelli, L.; West, A. B.; Ko, H. S.; Lee, Y.-
I.; Maguire-Zeiss, K. A.; Bowers, W. J.; Federoff, H. J.; Dawson, V. L.; Dawson, T.
M. Nat. Med. 2010, 16, 998.
8. (a) Deng, X.; Dzamko, N.; Prescott, A.; Davies, P.; Liu, Q.; Yang, Q.; Lee, J.-D.;
Patricelli, M. P.; Nomanbhoy, T. K.; Alessi, D. R.; Gray, N. S. Nat. Chem. Biol. 2011,
7, 203; (b) Ramsden, N.; Perrin, J.; Ren, Z.; Lee, B. D.; Zinn, N.; Dawson, V. L.;
Tam, D.; Bova, M.; Lang, M.; Drewes, G.; Bantscheff, M.; Bard, F.; Dawson, T. M.;
Hopf, C. ACS Chem. Biol. 2011, 6, 1021; (c) Zhang, J.; Deng, X.; Choi, H. G.; Alessi,
D. R.; Gray, N. S. Bioorg. Med. Chem. Lett. 2012, 22, 1864; (d) Reith, A. D.;
Bamborough, P.; Jandu, K.; Andreotti, D.; Mensah, L.; Dossang, P.; Choi, H. G.;
Deng, X.; Zhang, J.; Alessi, D. R.; Gray, N. S. Bioorg. Med. Chem. Lett. 2012, 22,
5625; (e) Chen, H.; Chan, B. K.; Drummond, J.; Estrada, A. A.; Gunzner-Toste, J.;
Liu, X.; Liu, Y.; Moffat, J.; Shore, D.; Sweeney, Z. K.; Tran, T.; Wang, S.; Zhao, G.;
Zhu, H.; Burdick, D. J. J. Med. Chem. 2012, 55, 5536; (f) Estrada, A. A.; Liu, X.;
Baker-Glenn, C.; Beresford, A.; Burdick, D. J.; Chambers, M.; Chan, B. K.; Chen,
H.; Ding, X.; DiPasquale, A. G.; Dominguez, S. L.; Dotson, J.; Drummond, J.;
Flagella, M.; Flynn, S.; Fuji, R.; Gill, A.; Gunzner-Toste, J.; Harris, S. F.; Heffron, T.
P.; Kleinheinz, T.; Lee, D. W.; Le Pichon, C. E.; Lyssikatos, J. P.; Medhurst, A. D.;
Moffat, J. G.; Mukund, S.; Nash, K.; Scearce-Levie, K.; Sheng, Z.; Shore, D. G.;
Tran, T.; Trivedi, N.; Wang, S.; Zhang, S.; Zhang, X.; Zhao, G.; Zhu, H.; Sweeney,
Z. K. J. Med. Chem. 2012, 55, 9416; (g) Garofalo, A. W.; Adler, M.; Aubele, D. L.;
Brigham, E. F.; Chian, D.; Franzini, M.; Goldbach, E.; Kwong, G. T.; Motter, R.;
Probst, G. D.; Quinn, K. P.; Ruslim, L.; Sham, H. L.; Tam, D.; Tanaka, P.; Truong, A.
P.; Ye, X. M.; Ren, Z. Bioorg. Med. Chem. Lett. 2013, 23, 1974; (h) Chan, B. K.;
Estrada, A. A.; Chen, H.; Atherall, J.; Baker-Glenn, C.; Beresford, A.; Burdick, D. J.;
Chambers, M.; Dominguez, S. L.; Drummond, J.; Gill, A.; Kleinheinz, T.; Le
Pichon, C. E.; Medhurst, A. D.; Liu, X.; Moffat, J. G.; Nash, K.; Scearce-Levie, K.;
Sheng, Z.; Shore, D. G.; Van de Poel, H.; Zhang, S.; Zhu, H.; Sweeney, Z. K. ACS
Med. Chem. Lett. 2013, 4, 85; (i) Franzini, M.; Ye, X. M.; Adler, M.; Aubele, D. L.;
Garofalo, A. W.; Gauby, S.; Goldbach, E.; Probst, G. D.; Quinn, K. P.; Santiago, P.;
Sham, H. L.; Tam, D.; Truong, A.; Ren, Z. Bioorg. Med. Chem. Lett. 2013, 23, 1967;
(j) Estrada, A. A.; Chan, B. K.; Baker-Glenn, C.; Beresford, A.; Burdick, D. J.;
Chambers, M.; Chen, H.; Dominguez, S. L.; Dotson, J.; Drummond, J.; Flagella,
M.; Fuji, R.; Gill, A.; Halladay, J.; Harris, S. F.; Heffron, T. P.; Kleinheinz, T.; Lee, D.
W.; Pichon, C. E. L.; Liu, X.; Lyssikatos, J. P.; Medhurst, A. D.; Moffat, J. G.; Nash,
K.; Scearce-Levie, K.; Sheng, Z.; Shore, D. G.; Wong, S.; Zhang, S.; Zhang, X.; Zhu,
H.; Sweeney, Z. K. J. Med. Chem. 2014, 57, 921.
Figure 6. 19Dose response of 15 in G2019S transgenic mice showing pS935 and
pS1292 readouts for brain (A) and kidney (B) at 90 min (mean SEM is shown, each
endpoints based on data from three transgenic mice).
Modest changes in key PD endpoints, measured at 90 min, could
be observed when 15 was dosed in G2019S transgenic mice
(Fig. 6).19 Dosing 15 up to 300 mg/kg achieved an unbound brain
concentration of 1809 nM and exhibited about a 20% decrease in
the phosphorylation of the S935 and S129220 residues (Fig. 6A).
Kidney showed a decrease in phosphorylation of ꢀ20% for S935
and ꢀ40% for S1292 (Fig. 6B).
In summary, an HTS campaign identified triazolopyridazines as
having potency at LRRK2. This scaffold was subjected to SAR opti-
mization to not only improve target potency but to also maintain
good predicted brain availability using passive permeability and
P-gp efflux as leading indicators. Surrogate crystallography
enabled the identification of a binding pose that was consistent
with the observed SAR. We identified 15 as a tool compound exhib-
iting a good in vitro/cellular profile. While it was not a P-gp sub-
strate, it was found to be a substrate for BCRP which appears to
have limited its in vivo effectiveness. Our focus has now shifted
to the other HTS hits and we will disclose those data in due course.
9. Luerman, G. C.; Nguyen, C.; Samaroo, H.; Loos, P.; Xi, H.; Hurtado-Lorenzo, A.;
Needle, E.; Noell, G. S.; Galatsis, P.; Dunlop, J.; Geoghegan, K. F.; Hirst, W. D. J.
Neurochem. 2014, 128, 561.
10. LRRK2 kinase activity was measured using LANTHA Screen technology from
Invitrogen. GST-tagged truncated LRRK2 from Invitrogen (Cat # PV4874) or
mutant G2019S LRRK2 (Invitrogen cat
# PV4881) was incubated with
fluorescein-labeled peptide substrate LRRKtide (Invitrogen cat # PR8976A),
in the presence of a dose response of compound. Upon completion, the assay
was stopped and detected with
antibody (Invitrogen, cat # PR8975A). The assay was carried out under the
following protocol: 3 L of a working solution of substrate (233 nM LRRKtide,
117 M ATP) prepared in assay buffer (50 mM HEEPES, pH 7.5, 3 mM MgCl2,
a terbium labeled anti-phospho-ERM
l
l
Acknowledgments
with 2 mM DTT and 0.01% Brij35 added fresh) was added to a low volume
Greiner 384-well plate. The compound dose response was prepared by
diluting compound to a top concentration of 3.16 mM in 100% DMSO and
We thank John Knafels for crystal soaking and freezing
experiments.
serial diluted by half-log in DMSO 11 times. Aliquots (3.5
DMSO dose response were mixed with 46.5 water then
mixture was added to the 3 L substrate mix in the 384-well plate. The
kinase reaction was started with 3 L of a working solution of LRRK2 enzyme
at a concentration of 4 g/mL. The final reaction concentrations were 100 nM
LRRKtide, 50 ATP, 1.7 g/mL LRRK2 enzyme and compound dose
response with a top dose of 32 M. The reaction was allowed to progress at
room temperature for two hours or 90 minutes for the mutant protein and
then stopped with the addition of 7 L of detection buffer (20 mM Tris pH
l
L) of the 100%
lL
1 lL
of this
l
References and notes
l
l
1. (a) Jellinger, K. A. Mov. Disord. 2012, 27, 8; (b) Poulopoulos, M.; Levy, O. A.;
Alcalay, R. N. Mov. Disord. 2012, 27, 831.
l
M
l
a
l
2. (a) Satake, W.; Nabalyashi, Y.; Mizuta, I.; Hirota, Y.; Ito, C.; Kubo, M.;
Kawaguchi, T.; Tsunoda, T.; Watanabe, M.; Takeda, A.; Tomiyama, H.;
Nakashima, K.; Hasegawa, K.; Obata, F.; Yoshikwa, T.; Kawakami, H.; Sakoda,
S.; Yamamoto, M.; Hattori, N.; Murata, M.; Nakamura, Y.; Toda, T. Nat. Genet.
2009, 41, 1303; (b) Chang, X.-L.; Mao, X.-Y.; Li, H.-H.; Zhang, J.-H.; Li, N.-N.;
Burgunder, J.-M.; Peng, R.; Tan, E.-K. Am. J. Med. Genet. Part B 2011, 156, 334; (c)
Do, C. B.; Tung, J. Y.; Dorfman, E.; Kiefer, A. K.; Drabant, E. M.; Francke, U.;
l
7.6, 0.01% NP-40, 0.02% NaN3, 6 mM EDTA with 2 nM terbium labeled anti-
phospho-ERM). After an incubation of 1 h at room temperature, the plate was
read on an Envision with an excitation wavelength of 340 nm and a reading
emission at both 520 nm and 495 nm. The ratio of the 520 nm and 495 nm
emission was used to analyze the data. LRRK2 whole cell assay method:

4140
P. Galatsis et al. / Bioorg. Med. Chem. Lett. 24 (2014) 4132–4140
Transfection and cell treatment—HEK 293 cells were transiently transfected
with full length LRRK2 using a DNA:Lipofectamine 2000 ratio of 1:2.5 ( g: L)
16. (a) Uchida, Y.; Ohtsuki, S.; Katsukura, Y.; Ikeda, C.; Suzuki, T.; Kamiie, J.;
l
l
Terasaki, T. J. Neurochem. 2011, 117, 333; (b) Ito, K.; Uchida, Y.; Ohtsuki, S.;
Aizawa, S.; Kawakami, H.; Katukura, Y.; Kamiie, J.; Terasaki, T. J. Pharma. Sci.
2011, 100, 3939.
in a 85–90% confluent T175 flask. After 6 hours the transfection media was
replaced with growth media (DMEM+10%FBS) and incubated overnight. The
following day cells were harvested and plated into clear tissue culture treated
384-well plates at 10,000 cells/well and allowed to incubate overnight. The
next day cells were treated with a dose response of compound for 90 min at
17. (a) Miklavc, P.; Ehinger, K.; Thompson, K. E.; Hobi, N.; Shimshek, D. R.; Frick, M.
PLoS One 2014, 9, e84926; (b) Ness, D.; Ren, Z.; Gardi, S.; Sharpnack, D.;
Johnson, V. J.; Brennan, R. J.; Brigham, E. F.; Olaharski, A. J. PLoS One 2013, 8,
e66164; (c) Herzig, M. C.; Kolly, C.; Persohn, E.; Theil, D.; Schweizer, T.; Hafner,
T.; Stemmelen, C.; Troxler, T. J.; Schmid, P.; Danner, S.; Schnell, C. R.; Mueller,
M.; Kinzel, B.; Grevot, A.; Bolognani, F.; Stirn, M.; Kuhn, R. R.; Kaupmann, K.;
ver der Putten, P. H.; Rovelli, G.; Shimshek, D. R. Hum. Mol. Genet. 2011, 20,
4209.
37 °C, 5% CO2. After compound treatment, cells were lysed with 15
lL lysis
buffer (Lantha Lysis buffer (Invitrogen, PV5598) with protease inhibitor
(Sigma P2714), 0.1% SDS, 1 mM Na3PO4, 17.5 mM Na2H2P2O7, 25 mM NaF,
1 mM PMSF) for 30 minutes at 4 °C and frozen at À80 °C for 30 min to ensure
complete lysis. 8 lL of the lysed cells was then transferred to ELISA plates
coated with LRRK2 capture antibody and blocked with BSA and allowed to
incubate overnight at 4 °C. ELISA—384-well high binding plates (Greiner
18. Dzamko, N.; Deak, M.; Hentati, F.; Reith, A. D.; Prescott, A. R.; Alessi, D. R.;
Nichols, R. J. Biochem. J. 2010, 430, 405.
781074) were coated with 10
(Covance SIG-39840) at
bicarbonate buffer at pH 9.5 and incubated overnight at 4 °C. Plates were
then blocked with 1% BSA in PBS-Tween (0.05%) for 1 hour and washed in
PBS-Tween before receiving 8 lL of treated cell lysis. The next day plates
l
L of mouse monoclonal anti-LRRK2 antibody
19. Custom generated antibody against pS1292 LRRK2 was used. Specificity
against pS1292 was verified using pS1292A. Brains and kidneys were
homogenized in lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 1%
Triton-X100, 5% Glycerol, 10 mM PPA, 20 mM NaF, 2 mM Na3VO4, 2 mM EGTA,
2 mM EDTA), incubated on ice for 30 min and centrifuged at 4° C for 10 min at
20,800 g. Supernatant was used for western blot analysis. Vehicle (1 Meq of 1 N
HCl, 10% NMP (1-methyl-2-pyrrolidone) in 20% SBECD (beta-cyclodextrin
sulfobutyl ether) solution) was used as a control.
20. Sheng, Z.; Zhang, S.; Bustos, D.; Kleinheinz, T.; Le Pichon, C. E.; Dominguez, S. L.;
Solanoy, H. O.; Drummond, J.; Zhang, X.; Ding, X.; Cai, F.; Song, Q.; Li, X.; Yue,
Z.; van der Brug, M. P.; Durdick, D. J.; Gunzner-Toste, J.; Chen, H.; Liu, X.;
Estrada, A. A.; Sweeney, Z. K.; Scearce-Levie, K.; Moffat, J. G.; Kirkpatrick, D. S.;
Zhu, H. Sci. Transl. Med. 2012, 4(164r), a161.
a
final concentration of 10 g/mL in sodium
l
were washed and incubated with anti-LRRK2 phospho-S935 antibody
(Abcam, UDD210(12)) at 1:2000 in PBS-Tween for 1 hour then washed and
incubated with an anti-Rabbit-HRP antibody (GE Healthcare NA9340V) for 1
hour. Plates were washed one final time and signal was detected with
SuperSignal ELISA Pico Chemiluminescent Substrate (Pierce 37069).
11. (a) Meanwell, N. A. Chem. Res. Toxicol. 2011, 24, 1420; (b) Edwards, M. P.; Price,
D. A. Ann. Rep. Med. Chem. 2010, 45, 381; (c) Wager, T. T.; Hou, X.; Verhoest, P.
R.; Villalobos, A. ACS Chem. Neurosci. 2010, 1, 435.
12. Kinome selectivity screening (KSS) against a panel of 39 kinases, chosen to
broadly cover the kinome
13. Groom, C. R.; Allen, F. H.; Henderson, S. The Cambridge Structural Database
(CSD). In: Reference Module in Chemistry, Molecular Sciences and Chemical
Engineering, Elsevier, Waltham, MA, USA, 2013.
14. Bruno, I. J.; Cole, J. C.; Kessler, M.; Luo, J.; Motherwell, W. D. S.; Purkis, L. H.;
Smith, B. R.; Taylor, R.; Cooper, R. I.; Harris, S. E.; Orpen, A. G. J. Chem. Inf.
Comput. Sci. 2004, 44, 2133.
21. (a) Collins, I.; Castro, J. L.; Street, L. J. Tetrahedron Lett. 2000, 41, 781; (b) Street,
L. J.; Sternfeld, F.; Jelley, R. A.; Reeve, A. J.; Carling, R. W.; Moore, K. W.;
McKernan, R. M.; Sohal, B.; Cook, S.; Pike, A.; Dawson, G. R.; Bromidge, F. A.;
Wafford, K. A.; Seabrook, G. R.; Thompson, S. A.; Marshall, G.; Pillai, G. V.;
Castron, J. L.; Atack, J. R.; MacLeod, A. M. J. Med. Chem. 2004, 47, 3642; (c) Thiel,
O. R.; Achmatowicz, M. M.; Reichelt, A.; Larsen, R. D. Angew. Chem., Int. Ed.
2010, 49, 8395; (d) Heinelt, U.; Wehner, V.; Herrmann, M.; Schoenafinger, K.;
Steinhagen, H.; Scheiper, B. Patent Application 2009097970. 2009.; (e) Aldrich,
L. N.; Lebois, E. P.; Lewis, L. M.; Nalywajko, N. T.; Niswender, C. M.; Weaver, C.
D.; Conn, P. J.; Lindsley, C. W. Tetrahedron Lett. 2009, 50, 212.
15. For
a review of blood brain barrier penetration see: Reichel, A. Chem.
Biodiversity 2009, 6, 2030.