Discovery and preclinical profiling of 3-[4-(morpholin-4-yl)-7H-pyrrolo[2,3-d]pyrimidin-5-yl]benzonitrile (PF-06447475), a highly potent, selective, brain penetrant, and in vivo active LRRK2 kinase inhibitor

Article abstract of DOI:10.1021/jm5014055

Leucine rich repeat kinase 2 (LRRK2) has been genetically linked to Parkinson's disease (PD) by genome-wide association studies (GWAS). The most common LRRK2 mutation, G2019S, which is relatively rare in the total population, gives rise to increased kinase activity. As such, LRRK2 kinase inhibitors are potentially useful in the treatment of PD. We herein disclose the discovery and optimization of a novel series of potent LRRK2 inhibitors, focusing on improving kinome selectivity using a surrogate crystallography approach. This resulted in the identification of 14 (PF-06447475), a highly potent, brain penetrant and selective LRRK2 inhibitor which has been further profiled in in vivo safety and pharmacodynamic studies.

Full text of DOI:10.1021/jm5014055

Article
pubs.acs.org/jmc
Discovery and Preclinical Profiling of 3‑[4-(Morpholin-4-
yl)‑7H‑pyrrolo[2,3‑d]pyrimidin-5-yl]benzonitrile (PF-06447475), a
Highly Potent, Selective, Brain Penetrant, and in Vivo Active LRRK2
Kinase Inhibitor
Jaclyn L. Henderson,† Bethany L. Kormos,† Matthew M. Hayward,∥ Karen J. Coffman,∥ Jayasankar Jasti,∥
Ravi G. Kurumbail,∥ Travis T. Wager,† Patrick R. Verhoest,† G. Stephen Noell,⊥ Yi Chen,‡ Elie Needle,‡
Zdenek Berger,‡ Stefanus J. Steyn,§ Christopher Houle,# Warren D. Hirst,‡ and Paul Galatsis*,†
‡
§
†Worldwide Medicinal Chemistry, Neuroscience Research Unit, and Pharmacokinetics, Dynamics, and Metabolism, Pfizer
Worldwide R&D, 610 Main Street, Cambridge, Massachusetts 02139, United States
⊥
#
∥Worldwide Medicinal Chemistry, Primary Pharmacology Group, and Drug Safety R&D, Pfizer Worldwide R&D, Eastern Point
Road Groton, Connecticut 06340, United States
S
* Supporting Information
ABSTRACT: Leucine rich repeat kinase 2 (LRRK2) has been
genetically linked to Parkinson’s disease (PD) by genome-wide
association studies (GWAS). The most common LRRK2
mutation, G2019S, which is relatively rare in the total population,
gives rise to increased kinase activity. As such, LRRK2 kinase
inhibitors are potentially useful in the treatment of PD. We herein
disclose the discovery and optimization of a novel series of potent
LRRK2 inhibitors, focusing on improving kinome selectivity using
a surrogate crystallography approach. This resulted in the
identification of 14 (PF-06447475), a highly potent, brain
penetrant and selective LRRK2 inhibitor which has been further
profiled in in vivo safety and pharmacodynamic studies.
of the pharmaceutical industry, resulting in the discovery of a
number of LRRK2 inhibitors4 with a range of CNS druglike
properties5,6 (Figure 1). Although the kinase domain of LRRK2
shares low overall sequence homology with other kinases, it has
been characterized as a serine−threonine kinase in the TKL
group of kinases and has highest homology to LRRK1 and the
receptor-interacting protein (RIP) kinases.3a,5 Because of the
chronic nature of PD and the aging patient population, an
excellent safety profile, potentially requiring a high degree of
kinase selectivity, is critical for a successful therapy in this area.
In this report we describe the discovery and optimization of a
series of pyrrolopyrimidines leading to a highly selective and
brain penetrant compound, which has proved to be a useful
tool in interrogating both central and peripheral LRRK2
biology.
INTRODUCTION
■
Parkinson’s disease (PD) is a debilitating neurodegenerative
disorder that affects over seven million people worldwide.
Despite extensive research, the underlying disease mechanisms,
and thus targets for therapeutic intervention, have remained
unclear.1 Recent genome-wide association studies (GWAS)
have identified several proteins with significant linkages to PD.2
Among these, leucine rich repeat kinase 2 (LRRK2) is a highly
promising target, as PD patients carrying LRRK2 mutations are
almost indistinguishable from idiopathic PD patients, indicating
that this may represent a common mechanism between
inherited and sporadic PD.2 Furthermore, a number of
additional genetic variants in LRRK2 have been identified as
having an increased PD risk, suggesting that it is important in
the cause and pathogenesis of PD in a greater proportion of
patients with this disease than previously believed.2
While there is growing evidence to support a central role of
LRRK2 dysfunction/dysregulation in the development of PD,
the complexity of LRRK2 has made elucidation of its
(patho)physiological role challenging.3 For example, the natural
substrates of the kinase and the protein-binding partners have
remained elusive. Nevertheless, the potential for disease
modification and the recent successes in the development of
kinase inhibitors for chronic illness have attracted the attention
RESULTS AND DISCUSSION
■
Our initial strategy was to conduct a high throughput screen
(HTS) of the Wyeth compound file using truncated, GST-
Special Issue: New Frontiers in Kinases
Received: September 12, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
because of its enhanced LRRK2 potency, and thus, further
optimization was initiated.
Lead compound 7 originated from a previous program
targeting the Janus kinase (JAK) family, and thus, crystal
structures of related compounds were available in house. It is
interesting to note that others have also identified lead matter
for their LRRK2 programs from JAK inhibitors,11 perhaps
confirming this family as a suitable surrogate in the absence of
LRRK2 crystallographic data. The available crystal structures
and literature data suggested the pyrrolopyrimidine was likely
an ATP-binding pocket two-point hinge binder, with the aryl
moiety oriented toward the DFG motif and the amine
occupying the ATP-ribose pocket. Synthetically, this hit offered
opportunities for parallel medicinal chemistry optimization with
the amine installation proceeding from the pyrimidyl chloride
through an SNAr reaction and the aryl group being accessed
through Suzuki cross-coupling reactions. Initial optimization
focused on improving both kinase selectivity and human liver
microsome (HLM) stability through optimization of the
lipophilic 3-methylpiperidine, summarized in Table 1. All
compounds prepared were predicted to be in good CNS
druglike space (CNS MPO5) and exhibited passive perme-
ability and P-gp efflux ratios congruent with high brain
availability.
Figure 1. Previously published LRRK2 inhibitors highlighting CNS
druglike scores5 (CNS druglike space: CNS MPO ≥ 4.0/6.0).
Although initial library work investigated a diverse set of both
primary and secondary amines, simple cyclic and acyclic
secondary amines proved optimal for LRRK2 potency. From
lead compound 7, it was found that the methyl group on the
amine substituent was not a requirement for activity, with
piperidine 9 and pyrrolidine 10 being close to equipotent.
Despite its small size, dimethylamino substitution (12) was one
of the more potent compounds, with reduced lipophilicity
affording improved clearance. Interestingly, azetidine analogue
11 lost 10× to 20× LRRK2 potency compared to the closely
related analogues 10 and 12, a finding that may be due to an
overall electronic effect on the core hinge-binding interaction.
Attempts to add additional polarity to the molecule were not
well tolerated; basic amines such as piperazine 15 ameliorated
LRRK2 potency, while ethers 16−18 failed to improve either
potency or microsomal stability. Morpholine 14 was the
exception, showing improvements in potency, clearance, and
selectivity over the piperidine analogues.
After this single round of optimization, we felt that both
dimethylamino 12 and morpholino 14 analogues were
sufficiently potent in the LRRK2 whole cell assay (WCA)
(pS935 end point) to warrant additional profiling. In particular,
we were interested in corroborating what at the time was a
relatively new assay by generating a second measure of LRRK2
potency in a physiologically relevant system.12 Testing these
two compounds in human PBMC lysates using the ActivX
KiNativ technology gave both an indication of LRRK2 potency
alongside a much wider selectivity assessment.13 While LRRK2
inhibition lined up well with the WCA (ActivX LRRK2 IC50:
12, 42 nM; 14, 15 nM), the selectivity profile of 14 particularly
differentiated it from the dimethylamino analogue (Figure 3
and Supporting Information for further details, Figure S1 and
Table S2). Of particular note in the off-target activity profiles
were the numerous hits from the STE20 kinase family (LOK,
SLK, and MST kinases) and the observation that, despite their
resemblance to previously reported JAK inhibitors, both
compounds were devoid of JAK activity.
tagged LRRK2 protein and LRRKtide as a surrogate kinase
substrate in a FRET assay. This resulted in a relatively large
number of hits (0.6% confirmed hits) which featured known
ATP hinge-binding scaffolds. Thus, a kinase selectivity panel
(KSS) was used in conjunction with ligand efficiency (LE),
lipophilic efficiency (LipE), and CNS desirability multi-
parameter optimization scores (CNS MPO)5 to triage the
initial hits. Because of its small size, high LE, and excellent
selectivity profile, imidazopyrimidine 6 stood out as a key hit of
this preliminary triage. Instead of running a second HTS to
interrogate the Pfizer compound file, we used a selection of
diverse hits from the Wyeth screen as templates to identify
additional chemical matter from a set of compounds derived
from (1) our internal kinase-targeted library,7 (2) compounds
that had activity in kinases with similar ATP-binding sites as
determined by alignment in PFAAT,8 and (3) virtual screening
based on similarity determined by atom-pairs9 and feature
trees.10 On the basis of the original imidazopyrimidine hit 6,
highly potent pyrrolopyrimidine 7 was identified (Figure 2).
While this compound does have affinity for several other
kinases, the overall selectivity profile was considered acceptable
In an effort to identify key opportunities to enhance
selectivity and to drive further understanding of LRRK2
Figure 2. Initial HTS hit leading to identification of pyrrolopyrimidine
scaffold.
B
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
e
Table 1. SAR around 3-Methylpiperidine
a
b
c
d
Biochemical assay, geometric mean (n ≥ 2). Cellular assay, geometric mean (n ≥ 2). Human liver microsomal clearance. Pfizer internal kinase
e
selectivity panel, quoted as number of kinases displaying >50% inhibition at 1 μM. nt: not tested.
binding interactions in the absence of a LRRK2 X-ray crystal
structure, we employed a surrogate crystallography approach
based on kinase similarity and crossover of compound activity.
LRRK2 only has ∼30% residue identity and ∼50% residue
similarity in the kinase domain compared to its closest
neighbors; however, its ATP-binding site residues have greater
similarity to other kinases. For example, mammalian STE20-like
protein kinase 3 (MST3) ATP-binding site residues are 73%
similar to those in LRRK2. As such, it was not surprising that
there was some off-target activity at the MST family of kinases
(vide infra) suggesting MST3 was a reasonable surrogate
crystallographic system for LRRK2.
As mentioned, the pyrrolopyrimidine scaffold is a well-known
two-point hinge binder in kinases, and this was confirmed by
the X-ray crystal structure of 14 with MST3 (Figure 4). The
pyrrole NH formed a H-bond (HB) with the backbone
carbonyl of Glu85, and N1 formed a HB with the backbone
NH of Leu87 (corresponding to Glu1948 and Ala1950 in
LRRK2, respectively). N3 was involved in a water-mediated HB
network to the backbone carbonyl of Leu87 and the side chain
of Asp94, which correspond to Ala1950 and Arg1957 in
LRRK2, respectively. The morpholine ring formed van der
Waals (VDW) interactions with the Gly-rich loop, including
Ile15, Gly16, Lys17, and Val23 (corresponding to Leu1885,
Gly1886, Asp1887, and Val1893 in LRRK2), and the oxygen of
C
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. Detailed profiling of 14: (A) potency and ADMET; (B) in vitro kinase selectivity (DiscoveRx KINOMEscan); (C) ex vivo kinase
selectivity (ActivX KiNativ).
LRRK2 (22), although this small change reduced selectivity;
however, addition of fluoro substitutions at other positions
around the ring significantly reduced LRRK2 potency (23 and
24), illustrating the compact nature of this pocket. In an
attempt to lower HLM clearance, we examined the effects of
various heterocycles both with and without the 3-cyano group.
Similar to the fluoro substitution, a pyridyl nitrogen in the 2-
position was tolerated in terms of LRRK2 potency (25);
however, other pyridyl regioisomers were significantly less
potent. The position of the nitrogen in the different pyridyl
regioisomers did not significantly affect the torsional profile or
conformation of the phenyl ring; however, a nitrogen in the 4-
or 5-position was oriented toward residues that create a
negative electrostatic potential in that part of the pocket (the
catalytic Asp147, the backbone carbonyl of Ala133, and the
carbonyl of the Asn134 side chainAsp2017, His1998, and
Asn1999 in LRRK2), which may lead to a loss in potency due
to electrostatic repulsion. In comparison to the cyanophenyl, a
majority of the simple heteroaromatics examined were
detrimental to LRRK2 potency with N-methylpyrazole 28
being one of the few analogues that maintained LRRK2
potency while improving human microsomal stability. Although
the whole cell potency was somewhat higher than 14, we felt
the KSS selectivity and improved pharmacokinetic properties of
pyrazole analogue 28 warranted further profiling. The N-methyl
group also provided a straightforward opportunity for radio-
labeling, which was of particular interest in terms of potential
PET ligand development. An ActivX selectivity screen revealed
that 28 was around 3× less potent at LRRK2 than 14 (50 nM
in ActivX) and shared similarly high selectivity (see Supporting
Information, Figure S2 and Tables S1 and S2). Examining the
X-ray crystal structure of 28 in MST3 (Figure 5) revealed
similar interactions to those seen with 14, including a water-
mediated HB between the pyrazole N and the conserved Lys38
(Lys1906 in LRRK2), mimicking the interaction of the
cyanophenyl with this residue. The development of radio-
labeled 28 will be disclosed in a subsequent publication
(manuscript in preparation).
Figure 4. X-ray crystal structure of 14 with MST3 (1.63 Å, PDB code
4U8Z).
the morpholine ring was in position to make a nontraditional
CH···O HB to Gly16. Such an interaction may contribute to
the modest boost in LRRK2 potency observed with the change
from piperidine 9 to morpholine 14 in this position.
As shown in Table 2, attempts to optimize the 3-
cyanophenyl substituent illustrated the important contribution
of this part of the molecule to both potency and selectivity. In
the X-ray crystal structure of 14 with MST3 (Figure 4), the
nitrile formed a charge-assisted HB with the conserved Lys38
(Lys1906 in LRRK2), which maintained its salt bridge with
Glu55 (Glu1920 in LRRK2) from the αC-helix. The phenyl
ring formed VDW interactions with the ceiling and floor of the
binding site. A comparison of 14 and 19 suggested the nitrile
interaction contributes a 3× to 4× potency improvement in
LRRK2. Surprisingly for an interaction with a residue that is
conserved across the kinome, it also appeared to be
contributing to enhanced selectivity. It has been shown in
other chemotypes that it is possible to modulate kinase binding
affinity and selectivity by varying steric and electrostatic
properties of substituents in this region despite the highly
conserved Lys residue.14 SAR efforts also indicated that the
interactions of this portion of the molecule with LRRK2 are
quite sensitive to both steric and electronic effects. An ortho-F
substituent in the 2-position of the aryl ring was tolerated by
Because of the lack of information around the role of the
MST kinases and any consequences in terms of compound
safety profile, we aimed to further optimize 14 to increase
selectivity versus this family.15 Of the residue differences
determined by sequence alignment of the ATP binding
pockets16 of LRRK2 and MST1−4 (Figure 6), four were
considered as potential areas to target for selectivity. Several of
D
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
e
Table 2. SAR around 3-Cyanophenyl
a
b
c
d
Biochemical assay, geometric mean (n ≥ 2). Cellular assay, geometric mean (n ≥ 2). Human liver microsomal clearance. Pfizer internal kinase
e
selectivity panel quoted as number of kinases displaying >50% inhibition at 1 μM. nt: not tested.
these residues have been targeted previously with different
chemotypes to drive LRRK2 selectivity over other kina-
ses.4d,11,14a,17 Our initial interest was drawn to the arginine
residue at H13 (Figure 6),16 as targeting this residue would be
expected to enhance selectivity not only over the MST kinases
but also across the kinome because of its low overall incidence
at that position: 3% of the kinome has an arginine at H13, while
16% has a positively charged amino acid. An arginine at the
H13 position is likely solvent exposed and quite flexible, thus
forming a HB or cation−π interaction with a ligand may restrict
the motion of the arginine side chain. This would cause a
conformational entropic penalty, which could cancel out any
gain in binding free energy achieved with a productive
interaction. However, interactions of this nature are prece-
E
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(MD) simulations were run on several LRRK2 homology
models. On the basis of the results of 10 ns simulations for each
homology model (60 ns total), Arg1957 may be involved in HB
interactions with nearby residues up to 68% of the simulation
time, restricting its motion and possibly orienting it in a
position for ligand interaction. Alternatively, it is possible that
Arg1957 interacts with one of the other LRRK2 protein
domains, which could also restrict its motion. These analyses
suggested the possibility that Arg1957 could occupy a position
whereby it could form either a HB or cation−π interaction with
a substituent occupying the 3-position of the morpholine. We
thus embarked on exploring substitution at this position
alongside potential morpholine isosteres. A summary of this
optimization is provided in Table 3.
Although 14 had generally good kinome selectivity, there was
significant off-target activity against the MST kinases as
evidenced by both the enzyme assay and ActivX data. We
were surprised to find that addition of a simple methyl group
(33) increased selectivity for LRRK2 to 25× over MST2,
although this was to the slight detriment of LRRK2 potency.
Selectivity versus MST4 was much higher than that for MST2
(and MST1; vide infra), which could potentially be explained
by the glycine at C1 allowing MST2 a greater degree of
conformational flexibility. Polar groups such as nitrile (34) or
CH2CN (35) did not significantly enhance selectivity over the
methyl analogue, suggesting that we were unsuccessful in
targeting a HB interaction with the arginine in these cases. We
also investigated heterocyclic substituents to contact the H13
arginine. Such analogues led to a substantial enhancement in
selectivity over MST2, our best example being oxadiazole 38,
which did not significantly lose LRRK2 potency over
Figure 5. X-ray crystal structure of 28 in MST3 (1.77 Å, PDB code
4W8D).
dented in the literature. For example, in the 1Q4L X-ray crystal
structure of GSK-3β,18a the inhibitor I-5 forms a HB interaction
with Arg141 (H13 residue) through a carboxylate group, and in
the 1Q5K X-ray crystal structure of GSK-3β,18b the inhibitor
AR-A014418 forms a cation−π interaction with Arg141
through a phenyl substituent. In both of these cases, it can
be seen that a residue, Glu137, is interacting with Arg141
through a salt bridge, restricting the motion of the arginine and
enabling productive interactions with the ligands without
paying a large conformational entropic penalty. To assess the
flexibility of the H13 arginine in LRRK2, molecular dynamics
Figure 6. Alignment of key residues in the ATP binding site for LRRK2 and MST kinases. Annotation of binding site residues was based on that
described in ref 16. Only binding site residues with differences between LRRK2 and the MST kinases are listed in the table; all others were identical:
purple, LRRK2 homology model using TAK-1 kinase as a template; green, corresponding residues from MST3 X-ray crystal structure, PDB code
4U8Z.
F
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
e
Table 3. Key SAR for LRRK2 vs MST Kinases
a
b
c
d
Biochemical assay, geometric mean (n ≥ 2). Cellular assay, geometric mean (n ≥ 2). Human liver microsomal clearance. Pfizer internal kinase
e
selectivity panel quoted as number of kinases displaying >50% inhibition at 1 μM. nt: not tested.
G
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
unsubstituted morpholine 14 but improved MST2 selectivity to
100×. LRRK2 potency did not improve with any of these
analogues, which could be due to an increase in desolvation
penalty, a conformational entropy penalty due to restricting the
flexibility of Arg1957, or due to lack of success in making a
productive interaction with the H13 arginine. The X-ray crystal
structure of 38 with MST3 (Figure 7) revealed similar
appropriate exposure levels could be achieved for initial in vivo
studies. Table 4 presents the oral bioavailability data for 14 in
the rat, dog, and primate.
Table 4. Oral Pharmacokinetic Data of 14 (5 mg/kg) in
Three Species
parameter
units
rat
dog
NHP
Cmax
ng/mL
181
0.50
629
6.22
1.83
41
15
25
Tmax
h
0.38
10
2.0
78
AUC
t1/2 (oral)
Vdss
ng·h/mL
h
0.78
0.92
0.6
2.27
1.06
3.5
L/kg
%
F
Despite extensive investigations over many years, there has
not yet been a robust model of LRRK2 mediated PD reported
in the literature; thus, efficacy measurements currently reflect in
vivo LRRK2 inhibition. Compound 14 was dosed to wild type,
BAC-transgenic WT LRRK2 and BAC-transgenic G2019S
LRRK2 mice. Inhibition was measured by changes in levels of
LRRK2 pS935 phosphorylation after compound treatment.
Robust dose−response relationships were observed in both
brain and kidney in all animals studied. Figure 8 shows the
Figure 7. X-ray crystal structure of 38 in MST3 (1.79 Å, PDB code
4W8E).
interactions to those seen with 14. In addition, the oxadiazole
substituent was involved in an intricate water-mediated HB
network with the protein. Oxadiazole N4 formed a bifurcated
water-mediated HB with the backbone NH of Lys17 in the Gly-
rich loop (Asp1887 in LRRK2). This water molecule was also
involved in a HB with the morpholine oxygen. O1 in the
oxadiazole ring formed a water-mediated HB with Tyr276, a
residue that is in an α-helix in an unconserved part of the kinase
domain. N2 in the oxadiazole ring was involved in a water-
mediated HB network to N3 of the pyrrolopyrimidine core and
Tyr86 (Leu1949 in LRRK2), the H5 residue. The MST3 H13
residue, Asp94, is approximately 4.5−5 Å away from the
oxadiazole ring, suggesting that the longer Arg1957 in LRRK2
could reach far enough to interact with this substituent.
With the morpholine ring appearing to be key for both
potency and selectivity, we also investigated a range of
morpholine isosteres (40−43); however, all lost significant
potency compared to 14 and also to some of the simpler ring
systems (e.g., piperidine 9, pyrrolidine 10, and azetidine 11). As
the morpholine isosteres are longer and/or orient the oxygen
atom differently than the morpholine substituent, this
suggested little flexibility in the position of the oxygen atom
to make optimal interactions in LRRK2.
As the potency, selectivity, and predicted human pharmaco-
kinetic (PK) properties of 14 were in line with our target
profile, we progressed with in vivo experiments to further assess
both the potency and pharmacokinetics. The majority of the
compounds in this series had passive permeability and P-gp
efflux ratios that were suggestive of compounds that should
penetrate the blood−brain barrier. Good brain availability in
rat, with close to equal distribution across compartments, was
observed for 14 (see Supporting Information, Figure S3).
However, despite high permeability and moderate clearance,
oral bioavailability was poor across species. Clearance in rodent,
which was mainly CYP mediated, could be saturated; thus,
Figure 8. Dose response study of 14 (administered subcutaneously) in
wild type mice measuring phosphorylation of S935 in brain and kidney
at 90 min after dose (mean SEM). Unbound drug IC50 values were
estimated to be 8 nM in brain and 11 nM in kidney.
results of a representative study in wild-type mice with the
unbound drug brain IC50 of ∼8 nM corresponding well with
the cell-based measurement from ActivX of 15 nM. Total
LRRK2 levels were not affected.
Phosphorylation of S935 is not considered to be directly
mediated by LRRK2, but rather dephosphorylation of this
residue seems to be an indication of inhibitor binding to
LRRK2. Recently a LRRK2 autophosphorylation site was
identified that may represent a more direct assessment of
LRRK2 inhibitor potency.19 Comparing the inhibition of the
two phosphorylation sites by 14 in G2019S mice suggested that
pS1292 may represent a somewhat more sensitive measure of
LRRK2 inhibition than the pS935 end point (Figure 9). It is
also interesting to note that 14 is ∼10× less potent in the BAC-
transgenic G2019S mice than WT mice at the pS935 end point.
One of the major concerns regarding chronic LRRK2
administration has been the effects on kidney, where LRRK2
expression is significant. There have been several reports of
kidney pathology observed in LRRK2 knock-out (KO) animals,
and it has to date been unclear if these effects would be
recapitulated by LRRK2 kinase inhibition in adulthood.20
A
number of cell based assays such as THLE cell viability,21
H
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 5. Summary of an Oral (b.i.d.) 14-Day Rat Exploratory
Toxicology Study at Day 11
dose
Cmax,u fold
fold
(mg kg−1 day−1
)
(nM)
Ceff AUCu (nM·h) Ceff
findings
6
20
60
86
6
25
70
843
4140
2
no treatment
related findings
380
12 no treatment
related findings
1050
11170
31 no treatment
related findings
CHEMISTRY
■
The syntheses of analogues 7−43 are illustrated in Scheme 1.
Beginning with commercially available 44, protection of the
Figure 9. Dose response study of 14 (administered subcutaneously) in
G2019S BAC-transgenic mice comparing the response of pS935 and
pS1292 phosphorylation to drug treatment in brain lysate (mean
SEM). IC50 values (based on unbound drug concentration) were
estimated to be the following: pS935, 103 nM; pS1292, 21 nM.
a
Scheme 1. General Synthetic Strategy
Ames, and in vitro micronucleus suggested 14 would have a
clean safety profile in vivo that would allow assessment of any
LRRK2 mediated safety risks. To this end, a 14-day rat
exploratory toxicology study was designed to maximize LRRK2
inhibition (target engagement) while minimizing the contribu-
tion of off-target kinase inhibition to any safety findings. PK
modeling, shown in Figure 10, predicted a b.i.d.-dosing
paradigm to give near 24 h coverage of the Ceff (set at the
unbound drug LRRK2 IC50 in brain of 15 nM) at the highest
dose. The actual exposures aligned well with the predicted
profile and support the maximization of target engagement in
conjunction with a blunting of Cmax exposure. This combination
allowed us to better understand the effects of LRRK2 kinase
inhibition without the potential for confounding off-target
toxicology. As a result, oral administration of 14 twice daily for
15 days was well tolerated at doses up to 60 mg kg−1 day−1 (31-
fold Cave,0−24h; 70-fold Cmax) with only minimal alterations in
RBC parameters, electrolytes, and food consumption without
concurrent effects on body weight (Table 5). Treated animal
kidneys appeared similar to control, as there were no alterations
in clinical chemistry or urinary biomarker end points suggesting
altered renal function. Additionally, no changes in the lung by
routine H&E evaluation and by staining, specifically for type 2
pneumocytes, were noted.22
a
Reagents and conditions: (a) NaH, SEM-Cl, THF, 0 °C; (b)
morpholine, DIPEA, n-BuOH, reflux; (c) boronic acid/ester, Pd-
(dppf)Cl2, K2CO3, EtOH, H2O, Δ; (d) TFA, MeOH (e) (3-
cyanophenyl)boronic acid, Pd(dppf)Cl2, K2CO3, DME, H2O, Δ; (f)
amine, DIPEA, n-BuOH, reflux.
pyrrole N with a SEM group generated 45. This compound
provided access to both points of diversity within the
pyrrolopyrimidine scaffold and could be used for single
compound preparation or for parallel library compound
generation. Suzuki cross-coupling reaction with 3-cyanophe-
nylboronic acid followed by deprotection afforded a key
intermediate, 47. SNAr displacement of this chloride with
amines in either singleton or library format was relatively facile
and afforded compounds 7−18, Table 1. Alternatively, 45 could
Figure 10. Predicted and measured (days 1 and 11) unbound plasma exposures (concentration plasma unbound (Cpu)) following oral dosing (10
and 30 mg/kg b.i.d.) of 14.
I
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
The 10 ns molecular dynamics simulations were run on each
homology model using Desmond Suite 201227 with the OPLS_2005
force field.28 The homology models were found to be stable over the
course of the simulations with a backbone rmsd of 2.5−3.0 Å from the
starting structure after 1 ns. HB interactions of Ser1954 and Asp1956
with Arg1957 were monitored over the course of the simulations.
Biology. LRRK2 Biochemical Assay. LRRK2 kinase activity was
measured using Lantha Screen technology from Invitrogen. GST-
tagged truncated LRRK2 from Invitrogen (catalog no. PV4874) or
mutant G2019S LRRK2 (Invitrogen catalog no. PV4881) was
incubated with fluorescein-labeled peptide substrate LRRKtide
(Invitrogen catalog no. PR8976A) in the presence of a dose response
of compound. Upon completion, the assay was stopped and detected
with a terbium labeled anti-phospho-ERM antibody (Invitrogen,
catalog no. PR8975A). The assay was carried out under the following
protocol: 3 μL of a working solution of substrate (233 nM LRRKtide,
117 μM ATP) prepared in assay buffer (50 mM HEEPES, pH 7.5, 3
mM MgCl2, 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 serial diluted by half-log in DMSO 11
times. Aliquots (3.5 μL) of the 100% DMSO dose response were
mixed with 46.5 μL of water, and then 1 μL of this mixture was added
to the 3 μL of substrate mix in the 384-well plate. The kinase reaction
was started with 3 μL of a working solution of LRRK2 enzyme at a
concentration of 4 μg/mL. The final reaction concentrations were 100
nM LRRKtide, 50 μM ATP, 1.7 μg/mL LRRK2 enzyme, and a
compound dose response with a top dose of 32 μM. The reaction was
allowed to progress at room temperature for 2 h or 90 min for the
mutant protein and then stopped with the addition of 7 μL of
detection buffer (20 mM Tris, pH 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 and 495 nm. The ratio of the 520 and 495 nm
emission was used to analyze the data.
Cellular Assay. For transfection and cell treatment, HEK 293 cells
were transiently transfected with full length LRRK2 using a DNA/
Lipofectamine 2000 ratio of 1:2.5 (μg:μL) in a 85−90% confluent
T175 flask. After 6 h the transfection medium was replaced with
growth medium (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 37 °C, 5% CO2. After compound treatment,
cells were lysed with 15 μL of 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 min at 4 °C and frozen at −80 °C for 30 min to ensure
complete lysis. An amount of 8 μL 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.
For ELISA, 384-well high binding plates (Greiner 781074) were
coated with 10 μL of mouse monoclonal anti-LRRK2 antibody
(Covance SIG-39840) at a final concentration of 10 μg/mL in sodium
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 h and
washed in PBS-Tween before receiving 8 μL of treated cell lysis. The
next day plates were washed and incubated with anti-LRRK2 phospho-
S935 antibody (Abcam, UDD210(12)) at 1:2000 in PBS-Tween for 1
h and then washed and incubated with an anti-rabbit-HRP antibody
(GE Healthcare NA9340V) for 1 h. Plates were washed one final time,
and signal was detected with SuperSignal ELISA Pico chemilumines-
cent substrate (Pierce 37069).
undergo the SNAr displacement with morpholine to produce
46. A Suzuki cross-coupling reaction with boronic acids or
esters followed by a similar deprotection then produced the
compounds shown in Table 2. Where the boronic acids or
esters were unstable, 46 could be treated with bis(pinacolato)-
diboron under palladium catalysis to produce a stable boronic
ester 48 (Scheme 2). With chemistry similar to that shown
above, a Suzuki cross-coupling with heterocyclic halides
followed by deprotection then yielded the target compounds.
a
Scheme 2
a
Reagents and conditions: (a) bis(pinacolato)diboron, Pd2(dba)3,
XPhos, dioxane, 95 °C; (b) heteroaryl bromide, Pd(PPh3)4, NaHCO3,
dioxane/water, 120 °C; (c) TFA, rt.
CONCLUSIONS
■
From a ligand efficient starting point, a series of LRRK2
inhibitors was identified using virtual screening. Optimization,
primarily through parallel synthesis, arrived at a potent and
selective LRRK2 inhibitor, which is highly brain penetrant. This
compound was studied further in rodent pharmacodynamic
models, and underwent two-week toxicological assessment.
Although the higher species PK is not suitable to continue
progression of this compound toward the clinic, its overall
properties make this an exceptional tool compound to study
the function of LRRK2.
EXPERIMENTAL SECTION
■
Expression and Purification of MST3. Human MST3 kinase
domain (residues 1−303; NP_001027467) containing His-tag and
thrombin site at the amino terminal region was codon optimized for
expression in E. coli (BL21-DE3 cells). The protein was purified and
crystallized according known published conditions.23 Briefly, the
protein was purified using Ni-NTA metal affinity resin and the purified
protein was digested overnight with thrombin (10 U of thrombin/mg
protein) to remove the tag, dialyzed against buffer containing 50 mM
Tris-HCl (pH 8.0), 100 mM NaCl, and 10% glycerol and concentrated
to at least 25 mg/mL.
MST3 Crystallization and Structure Determination. The
protein was mixed with 5 mM DTT and 1 mM 14 or 28 or 38 (all
prepared as 100 mM stocks in DMSO) and incubated at room
temperature for at least an hour. The samples were centrifuged and
crystallized by mixing 0.5 μL of the sample with 0.5 μL of the reservoir
solution (0.1 M Tris-HCl (pH 8.5), 180 mM MgCl2, 5 mM
manganese acetate, and 12−17% PEG3350) using sitting drop
vapor-diffusion technique. Crystals were obtained in 1−2 days, and
the crystals were frozen in liquid nitrogen using reservoir solution
containing 15% glycerol and 1 mM appropriate ligand. 3A7J was used
as a search model for determining the structure.
Computational Methods. LRRK2 homology models were
constructed with MOE24 using the AMBER99 force field.25 The
amino acid sequence of human LRRK2 was retrieved from UniProt26
and aligned to template structures using MOE’s Protein Align function
with default settings followed by manual editing of loop regions,
insertions, and deletions. The template X-ray crystal structures used
were TAK-1 (2EVA), MLK1 (3DTC), RET (2IVV), SRC (3DQW),
CLK1 (1Z57), and JAK3 (internal structure).
MST2/4 Biochemical Assay. MST2 and MST4 activity was
measured using the Z-Lyte assay kit from Invitrogen (Z-Lyte peptide 7
assay kit, catalog no. PV3180) using MST2 (STK3 catalog no.
PV4805) and MST4 (catalog no. PR7067A) supplied from Invitrogen.
The assays were run in an assay buffer containing 50 mM HEPES (pH
7.0 for MST2 and pH7.5 for MST4), 0.01% BRIJ-35, 10 mM MgCl2,
J
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
and 1 mM EGTA. Compounds were tested in dose response following
the kit protocols with the following final assay concentrations; 2 μM
Ser/Thr peptide 7, 75 μM ATP, and 18 nM MST2 or MST4. The
assay was stopped after 45 min for MST2 and after 60 min for MST4
using the kit stop reagent. Plates were then read using a fluorescence
plate reader measuring two emission wavelengths from one excitation
wavelength (excitation 400 nm; emission, 445 and 520 nm). Data were
analyzed and IC50 values determined based on percent phosphor-
ylation calculated from the emission ratio of 445/520 described in the
kit protocol.
pS1292 Assay. 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 20800g. Supernatant was used for Western blot analysis.
Vehicle (1 Mequiv of 1 N HCl, 10% NMP (1-methyl-2-pyrrolidone)
in 20% SBECD (β-cyclodextrin sulfobutyl ether) solution) was used as
a control.
In Vivo Studies. Four-month-old C57BL/6J or seven-month-old
LRRK2 Bac Tg G2910S mice were purchased from JAX Laboratories.
All mice were housed on a 12 h light/dark cycle and given free access
to food and drinking water as well as environmental enrichment. All
animal procedures and animal care methods were approved by Pfizer
Institutional Animal Care and Usage Committee. Compound 14 was
dissolved in vehicle (10% NMP, 1-methyl-2-pyrrolidone in 20%
SBECD, β-cyclodextrin sulfobutyl ether) and subcutaneously admin-
istered. Plasma and brain tissues were collected following 90 min of
compound administration and snap frozen for further PKPD
experiments. Rat studies utilized Wistar Han animals sourced from
Charles River Laboratories that were between 200 and 300 g and
between 8 and 11 weeks old at study start. For 14, doses were given by
oral gavage twice daily for 14 days using a vehicle comprising 10%
propylene glycol/20% PEG400/70% 9.5% methocellulose.
Chemistry. General Methods. All solvents and reagents were
used as obtained from the commercial vendors. Products were dried
under vacuum before being carried on to further reactions or
submitted for biological testing. All final products were purified by
preparative reverse phase HPLC using appropriate gradients of
acetonitrile and water with 0.05% formic acid as the counterion. For
syntheses referencing, procedures in other examples or methods
reaction conditions (length of reaction and temperature) may vary.
Purification conditions may vary between experiments: in general,
solvents and the solvent ratios used for eluents/gradients were chosen
to provide appropriate Rf or retention times. All tested compounds
were determined to be >95% pure by high-performance liquid
chromatography under the following conditions. Column: Waters
XBridge C18, 2.1 mm × 50 mm, 5 μm. Mobile phase A: 0.0375%
trifluoroacetic acid in water. Mobile phase B: 0.01875% trifluoroacetic
acid in acetonitrile. Gradient: 1−5% B over 0.60 min, then 5% to 100%
B over 3.40 min. Flow rate: 0.8 mL/min. Purity was calculated as a
percentage of total area at 254 nm. 1H NMR spectra were recorded on
a Bruker instrument operating at 400 MHz. Chemical shifts are
expressed in parts per million (ppm, δ) referenced to residual peaks
from the deuterated solvents employed (chloroform 7.25 ppm or
DMSO-d6 2.50 ppm). When peak multiplicities are reported, the
following abbreviations are used: s = singlet, d = doublet, t = triplet, m
= multiplet, br = broad, dd = doublet of doublets, dt = doublet of
triplets. Coupling constants are reported in hertz. The mass spectra
were obtained using liquid chromatography−mass spectrometry (LC−
MS) on an Agilent instrument using atmospheric pressure chemical
ionization (APCI) or electrospray ionization (ESI).
warmed to room temperature and allowed to stir for 3 h. The reaction
was quenched with saturated aqueous sodium chloride solution (250
mL), and the organic layer was dried over sodium sulfate, filtered, and
concentrated in vacuo. Silica gel chromatography (eluent, 10:1
petroleum ether/ethyl acetate) afforded the product as a white solid.
Yield: 8 g, 20 mmol, 57%. 1H NMR (400 MHz, DMSO-d6) δ 8.69 (s,
1H), 8.14 (s, 1H), 5.60 (s, 2H), 3.51 (t, J = 8 Hz, 2H), 0.82 (t, J = 8
Hz, 2H), −0.10 (s, 9H).
4-(5-Iodo-7-((2-(trimethylsilyl)ethoxy)methyl)-7H-pyrrolo-
[2,3-d]pyrimidin-4-yl)morpholine (46). Morpholine (2.45 g, 28.1
mmol) and N,N-diisopropylethylamine (6.63 g, 51.3 mmol) were
added to a solution of 4-chloro-5-iodo-7-((2-(trimethylsilyl)ethoxy)-
methyl)-7H-pyrrolo[2,3-d]pyrimidine (45) (10.5 g, 25.6 mmol) in n-
butanol (300 mL), and the reaction mixture was heated at reflux for 18
h, then concentrated under reduced pressure. Aqueous hydrochloric
acid (0.1 M, 100 mL) was added and the resulting solid was collected
by filtration, washed with water (20 mL), and dried under vacuum to
provide the product as a yellow solid. Yield: 8.0 g, 17 mmol, 66%.
LCMS m/z 461.2 [M + H+]. 1H NMR (400 MHz, DMSO-d6) δ 8.39
(s, 1H), 7.81 (s, 1H), 5.52 (s, 2H), 3.80−3.86 (m, 4H), 3.46−3.53 (m,
6H), 0.77−0.84 (m, 2H), −0.10 (s, 9H).
3-(4-Chloro-7H-pyrrolo[2,3-d]pyrimidin-5-yl)benzonitrile
(47). Step 1. To a stirred mixture of 4-chloro-5-iodo-7-((2-
(trimethylsilyl)ethoxy)methyl)-7H-pyrrolo[2,3-d]pyrimidine (45)
(8.2 g, 20 mmol), (3-cyanophenyl)boronic acid (3.2 g, 22 mmol),
and potassium carbonate (8.3 g, 60 mmol) in a mixture of 1,2-
dimethoxyethane and water (4:1 ratio, 250 mL) was added [1,1′-
bis(diphenylphosphino)ferrocene]dichloropalladium(II) (731 mg,
1.00 mmol). The reaction mixture was degassed and then charged
with nitrogen; this procedure was carried out a total of three times.
The reaction mixture was heated at reflux for 3 h, then cooled to room
temperature and diluted with saturated aqueous sodium chloride
solution (100 mL). The organic layer was dried over sodium sulfate,
filtered, and concentrated under reduced pressure. Purification via
silica gel column chromatography (eluent, 10:1 petroleum ether/ethyl
acetate) provided 3-(4-chloro-7-((2-(trimethylsilyl)ethoxy)methyl)-
7H-pyrrolo[2,3-d]pyrimidin-5-yl)benzonitrile as a yellow oil. Yield:
5.0 g, 12 mmol, 60%. 1H NMR (400 MHz, DMSO-d6) δ 8.75 (s, 1H),
8.13 (s, 1H), 8.00−8.02 (m, 1H), 7.84−7.92 (m, 2H), 7.68 (dd, J =
7.8, 7.8 Hz, 1H), 5.70 (s, 2H), 3.60 (dd, J = 8.0, 8.0 Hz, 2H), 0.86 (dd,
J = 8.0, 8.0 Hz, 2H), −0.08 (s, 9H).
Step 2. A solution of 3-(4-chloro-7-((2-(trimethylsilyl)ethoxy)-
methyl)-7H-pyrrolo[2,3-d]pyrimidin-5-yl)benzonitrile (47) (3.8 g, 9.9
mmol) in trifluoroacetic acid (25 mL) was stirred at room temperature
for 2 h. The reaction mixture was concentrated in vacuo and then
taken up in methanol (100 mL) and adjusted to pH >12 by addition of
solid potassium carbonate. Solvent was removed in vacuo, and the
residue was mixed with water (100 mL). The resulting solid was
isolated via filtration and washed with water, providing the product as
a white solid. Yield: 1.3 g, 5.1 mmol, 52%. 1H NMR (400 MHz,
DMSO): δ 13.50−12.50 (bs, 1H), 8.65 (s, 1H), 8.00 (s, 1H), 7.93 (s,
1H), 7.88 (d, J = 7.6 Hz, 1H), 7.82 (d, J = 7.6 Hz, 1H), 7.65 (t, J = 7.8
Hz, 1H). LCMS m/z 255.0 [M + H+].
4-(5-(4,4,5,5-Tetramethyl-1,3,2-dioxaborolan-2-yl)-7-((2-
(trimethylsilyl)ethoxy)methyl)-7H-pyrrolo[2,3-d]pyrimidin-4-
yl)morpholine (48). To
a solution of 4-(5-iodo-7-((2-
(trimethylsilyl)ethoxy)methyl)-7H-pyrrolo[2,3-d]pyrimidin-4-yl)-
morpholine (46) (500 mg, 1.09 mmol) and 4,4,5,5-tetramethyl-1,3,2-
dioxaborolane (543 mg, 4.24 mmol) in 1,4-dioxane (20 mL) were
added tris(dibenzylideneacetone)dipalladium(0) (99.7 mg, 0.109
mmol), triethylamine (439 mg, 4.34 mmol), and 2-dicyclohexylphos-
phino-2′,4′,6′-triisopropylbiphenyl (XPhos, 51.8 mg, 0.109 mmol),
and the reaction mixture was heated at 95 °C for 18 h. After cooling to
room temperature, the reaction mixture was diluted with water (100
mL) and extracted with ethyl acetate (3 × 50 mL). The combined
organic layers were washed with saturated aqueous sodium chloride
solution (100 mL), dried over sodium sulfate, filtered, and
concentrated in vacuo. Purification via silica gel chromatography
(gradient, 0−30% ethyl acetate in petroleum ether) afforded the
product as a yellow oil. Yield: 415 mg, 0.901 mmol, 83%. LCMS m/z
4-Chloro-5-iodo-7-((2-(trimethylsilyl)ethoxy)methyl)-7H-
pyrrolo[2,3-d]pyrimidine (45). A solution of 4-chloro-5-iodo-7H-
pyrrolo[2,3-d]pyrimidine (44) (9.8 g, 35 mmol) in tetrahydrofuran
(250 mL) was cooled to 0 °C and treated with sodium hydride (60%
in oil, 1.54 g, 38.5 mmol) in three portions. After the reaction mixture
had stirred at 0 °C for 1 h, 2-(trimethylsilyl)ethoxymethyl chloride
(6.4 g, 38 mmol) was added dropwise, and the reaction mixture was
K
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
461.3 [M + H+]. 1H NMR (400 MHz, CDCl3) δ 8.45 (s, 1H), 7.73 (s,
1H), 5.59 (s, 2H), 3.87−3.93 (m, 4H), 3.68−3.74 (m, 4H), 3.49−3.56
(m, 2H), 1.35 (s, 12H), 0.87−0.93 (m, 2H), −0.06 (s, 9H).
combined organic layers were washed with saturated aqueous sodium
chloride solution (100 mL), dried over sodium sulfate, filtered, and
concentrated in vacuo; purification via preparative thin layer
chromatography (eluent, 1:1 petroleum ether/ethyl acetate) afforded
6-(4-morpholino-7-((2-(trimethylsilyl)ethoxy)methyl)-7H-pyrrolo-
[2,3-d]pyrimidin-5-yl)picolinonitrile as a brown oil. Yield: 110 mg,
0.252 mmol, 57%. 1H NMR (400 MHz, CDCl3) δ 8.52 (s, 1H), 7.84−
7.93 (m, 2H), 7.74 (s, 1H), 7.59 (dd, J = 7.0, 1.2 Hz, 1H), 5.66 (s,
2H), 3.56−3.65 (m, 6H), 3.34−3.40 (m, 4H), 0.93 (dd, J = 8.3, 8.0
Hz, 2H), −0.05 (s, 9H).
Step 2. A solution of 6-(4-morpholino-7-((2-(trimethylsilyl)-
ethoxy)methyl)-7H-pyrrolo[2,3-d]pyrimidin-5-yl)picolinonitrile (110
mg, 0.252 mmol) in trifluoroacetic acid (3 mL) was stirred at room
temperature for 2 h. The reaction mixture was concentrated in vacuo,
taken up in acetonitrile (3 mL), and brought to a pH of >12 via
addition of solid potassium carbonate. After 30 min at room
temperature, the reaction mixture was filtered and concentrated in
vacuo. Purification via preparative HPLC (column, Phenomenex
Gemini C18, 8 μm; mobile phase A, ammonia in water, pH 10; mobile
phase B, acetonitrile; gradient, 10−50% B) afforded the product as a
white solid. Yield: 15.2 mg, 49.6 μmol, 20%. LCMS m/z 307.2 [M +
H+]. 1H NMR (400 MHz, DMSO-d6) δ 8.38 (s, 1H), 8.13 (dd, J = 8.1,
7.7 Hz, 1H), 7.96 (br d, J = 8 Hz, 1H), 7.89 (br d, J = 8 Hz, 1H), 7.86
(s, 1H), 3.50−3.55 (m, 4H), 3.19−3.24 (m, 4H).
Example of Method A: 5-(1-Methyl-1H-pyrazol-4-yl)-4-
(morpholin-4-yl)-7H-pyrrolo[2,3-d]pyrimidine (28).29 Step 1.
To a solution of 4-(5-iodo-7-((2-(trimethylsilyl)ethoxy)methyl)-7H-
pyrrolo[2,3-d]pyrimidin-4-yl)morpholine (46) (500 mg, 1.1 mmol)
and 1-methyl-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-pyra-
zole (272 mg, 1.31 mmol) in a mixture of ethanol and water (4:1, 10
mL) were added dichlorobis(triphenylphosphine)palladium(II) (41
mg, 58 μmol) and potassium carbonate (447 mg, 3.23 mmol). The
reaction mixture was degassed and purged with nitrogen; this
procedure was carried out a total of three times. It was then heated
at 100 °C for 18 h. After concentration in vacuo, the residue was
purified via chromatography on silica gel (eluent, 1:1 ethyl acetate/
petroleum ether) to provide 4-(5-(1-methyl-1H-pyrazol-4-yl)-7-((2-
(trimethylsilyl)ethoxy)methyl)-7H-pyrrolo[2,3-d]pyrimidin-4-yl)-
1
morpholine as a yellow solid. Yield: 200 mg, 0.48 mmol, 44%. H
NMR (400 MHz, DMSO-d6) δ 8.39 (s, 1H), 7.86 (s, 1H), 7.59 (s,
1H), 7.52 (s, 1H), 5.57 (s, 2H), 3.90 (s, 3H), 3.50−3.58 (m, 6H),
3.20−3.27 (m, 4H), 0.83 (dd, J = 8.0, 7.9 Hz, 2H), −0.09 (s, 9H).
Step 2. A solution of 4-(5-(1-methyl-1H-pyrazol-4-yl)-7-((2-
(trimethylsilyl)ethoxy)methyl)-7H-pyrrolo[2,3-d]pyrimidin-4-yl)-
morpholine (200 mg, 0.48 mmol) in trifluoroacetic acid (5 mL) was
stirred at room temperature for 2 h. The reaction mixture was
concentrated under reduced pressure to afford the product as a yellow
oil, which was then taken up in methanol (5 mL) and brought to a pH
of >12 via addition of solid potassium carbonate. The reaction mixture
was stirred for 30 min, filtered, and concentrated in vacuo. Purification
via preparative HPLC (column, Agella Venusil ASB C18, 5 μm; mobile
phase A, 0.225% formic acid in water; mobile phase B, acetonitrile;
eluent, 13% B) provided the product as a yellow solid. Yield: 90 mg,
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental details for compounds 7−13, 15−24, 26, 27, and
29−43c; figures detailing the tissue exposure (plasma, brain,
and kidney) vs time for 14; tables summarizing kinase
selectivity profiles for 12, 14, and 28. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
0.32 mmol, 67%. LCMS m/z 285.1 [M + H+]. H NMR (400 MHz,
DMSO-d6) δ 12.96 (br s, 1H), 8.47 (s, 1H), 7.85 (s, 1H), 7.57 (s, 1H),
7.52 (d, J = 2.5 Hz, 1H), 3.90 (s, 3H), 3.53−3.59 (m, 4H), 3.45−3.51
(m, 4H).
Accession Codes
PDB codes for MST3 with bound 14 (PF-06447475) is 4U8Z,
for 28 (PF-06454589) is 4W8D, and for 38 is 4W8E.
Example of Method B: 3-[4-(Morpholin-4-yl)-7H-pyrrolo[2,3-
d]pyrimidin-5-yl]benzonitrile (14).29 Step 1. A solution of 3-(4-
chloro-7-((2-(trimethylsilyl)ethoxy)methyl)-7H-pyrrolo[2,3-d]-
pyrimidin-5-yl)benzonitrile (47) (3.8 g, 9.9 mmol) in trifluoroacetic
acid (25 mL) was stirred at room temperature for 2 h. The reaction
mixture was concentrated in vacuo and then taken up in methanol
(100 mL) and adjusted to pH >12 by addition of solid potassium
carbonate. Solvent was removed in vacuo, and the residue was mixed
with water (100 mL). The resulting solid was isolated via filtration and
washed with water, providing 3-(4-chloro-7H-pyrrolo[2,3-d]pyrimidin-
5-yl)benzonitrile as a white solid. Yield: 1.3 g, 5.1 mmol, 52%. LCMS
m/z 255.0 [M + H+].
Step 2. Morpholine (871 mg, 10 mmol) and N,N-diisopropylethyl-
amine (2.6 g, 20 mmol) were added to a solution of 3-(4-chloro-7H-
pyrrolo[2,3-d]pyrimidin-5-yl)benzonitrile (2.5 g, 9.8 mmol) in n-
butanol (100 mL), and the reaction mixture was heated at reflux for 3
h. Solvents were removed in vacuo, and the residue was purified using
chromatography on silica gel (eluent, 1:1 ethyl acetate/petroleum
ether). Subsequent recrystallization from ethyl acetate and tert-butyl
methyl ether afforded the product as a white solid. Yield: 770 mg, 2.52
mmol, 26%. LCMS m/z 306.0 [M + H+]. 1H NMR (400 MHz,
DMSO-d6) δ 12.34 (br s, 1H), 8.41 (s, 1H), 7.99−8.02 (m, 1H), 7.89
(br d, J = 8 Hz, 1H), 7.76 (br d, J = 7.5 Hz, 1H), 7.71 (s, 1H), 7.68
(dd, J = 7.8, 7.8 Hz, 1H), 3.44−3.50 (m, 4H), 3.11−3.17 (m, 4H).
Example of Method C. 6-(4-Morpholino-7H-pyrrolo[2,3-
d]pyrimidin-5-yl)picolinonitrile (25). Step 1. To a solution of 6-
bromopyridine-2-carbonitrile (80 mg, 0.44 mmol) and 4-(5-(4,4,5,5-
tetramethyl-1,3,2-dioxaborolan-2-yl)-7-((2-(trimethylsilyl)ethoxy)-
methyl)-7H-pyrrolo[2,3-d]pyrimidin-4-yl)morpholine (48) (241 mg,
0.523 mmol) in 1,4-dioxane (2.5 mL) and water (0.5 mL) was added
tetrakis(triphenylphosphine)palladium(0) (51 mg, 44 μmol) and
sodium carbonate (140 mg, 1.32 mmol). The reaction mixture was
heated at 120 °C under microwave irradiation for 15 min, then diluted
with water (30 mL) and extracted with ethyl acetate (3 × 50 mL). The
AUTHOR INFORMATION
■
Corresponding Author
*E-mail: paul.galatsis@pfizer.com. Phone: 1-617-326-0709.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank Matt Griffor and Xi Song for their protein
production work for MST3.
ABBREVIATIONS USED
■
BAC, bacterial artificial chromosome; CL, clearance; CNS,
central nervous system; Cpu, concentration plasma unbound;
CYP, cytochrome P450; ER, efflux ratio; FRET, fluorescence
resonance energy transfer; GWAS, genome-wide association
studies; HB, H-bond; HLM, human liver microsome; HTS,
high throughput screen; JAK, Janus kinase; KO, knock-out;
KSS, kinase selectivity screen; LE, ligand efficiency; LipE,
lipophilic efficiency; LRRK2, leucine rich repeat kinase 2;
MPO, multiparameter optimization; NHP, non-human pri-
mate; PBMC, peripheral blood mononuclear cell; PD,
Parkinson’s disease; PET, positron emission tomography;
PFAAT, protein family alignment annotation tool; P-gp, P-
glycoprotein; PK, pharmacokinetic; RBC, red blood cell; RLM,
rat liver microsome; SEM, 2-(trimethylsilyl)ethoxymethyl;
THLE, transformed human liver epithelial cell line; WCA,
whole cell assay; WT, wild type; VDW, van der Waals
L
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(7) Xi, H.; Lunney, E. A. The design, annotation, and application of a
kinase-targeted library. In Chemical Library Design; Zhou, J. Z., Ed.;
Methods in Molecular Biology, Vol. 685; Humana Press: New York,
2011; pp 279−291
(8) Caffrey, D. R.; Dana, P. H.; Mathur, V.; Ocano, M.; Hong, E.-J.;
Wang, Y. E.; Somaroo, S.; Caffrey, B. E.; Potluri, S.; Huang, E. S.
PFAAT version 2.0: a tool for editing, annotating, and analyzing
multiple sequence alignments. BMC Bioinf. 2007, 8, 381.
REFERENCES
■
(1) (a) Lim, K.-L.; Zhang, C. W. Molecular events underlying
Parkinson’s diseasean interwoven tapestry. Front. Neurol. 2013, 4,
33. (b) Kurtis, M. M.; Martinez-Martin, P. Parkinson’s disease:
symptoms, unmet needs and new therapeutic targets. RSC Drug
Discovery 2013, 34, 3−25.
(2) (a) Labbe, C.; Ross, O. A. Association studies of sporadic
Parkinson’s disease in the genomic era. Curr. Genomics 2014, 15, 2−10.
(b) Lubbe, S.; Morris, H. R. Recent advances in Parkinson’s disease
genetics. J. Neurol. 2014, 261, 259−266. (c) Ross, O. A.; Soto-
Ortolaza, A. I.; Heckman, M. G.; Aasly, J. O.; Abahuni, N.; Annesi, G.;
Bacon, J. A.; Bardien, S.; Bozi, M.; Brice, A.; Brighina, L.; Van
Broeckhoven, C.; Carr, J.; Chartier-Harlin, M.-C.; Dardiotis, E.;
Dickson, D. W.; Diehl, N. N.; Elbaz, A.; Ferrarese, C.; Ferraris, A.;
Fiske, B.; Gibson, J. M.; Gibson, R.; Hadjigeorgiou, G. M.; Hattori, N.;
Ioannidis, J. P. A.; Jasinska-Myga, B.; Jeon, B. S.; Kim, Y. J.; Klein, C.;
Kruger, R.; Kyratzi, E.; Lesage, S.; Lin, C.-H.; Lynch, T.; Maraganore,
D. M.; Mellick, G. D.; Mutez, E.; Nilsson, C.; Opala, G.; Park, S. S.;
Puschmann, A.; Quattrone, A.; Sharma, M.; Silburn, P. A.; Sohn, Y. H.;
Stefanis, L.; Tadic, V.; Theuns, J.; Tomiyama, H.; Uitti, R. J.; Valente,
(9) Carhart, R. E.; Smith, D. H.; Venkataraghavan, R. Atom pairs as
molecular features in structure−activity studies: definition and
applications. J. Chem. Inf. Comput. Sci. 1985, 25, 64−73.
(10) (a) Rarey, M.; Dixon, J. S. Feature trees: a new molecular
similarity measure based on tree matching. J. Comput.-Aided Mol. Des.
1998, 12, 471−490. (b) Rarey, M.; Hindle, S.; Maass, P.; Metz, G.;
Rummey, C.; Zimmermann, M. Feature trees: theory and applications
from large-scale virtual screening to data analysis. In Pharmacophores
and Pharmacophore Searches; Langer, T., Hoffman, R. D., Eds.;
Methods and Principles in Medicinal Chemistry, Vol. 32; Wiley-VCH:
Weinheim, Germany, 2006; pp 81116
(11) 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. Discovery of selective
LRRK2 inhibitors guided by computational analysis and molecular
modeling. J. Med. Chem. 2012, 55, 5536−5545.
E. M.; van de Loo, S.; Vassilatis, D. K.; Vilarino-Guell, C.; White, L. R.;
̃
̈
Wirdefeldt, K.; Wszolek, Z. K.; Wu, R.-M.; Farrer, M. J. Association of
LRRK2 exonic variants with susceptibility to Parkinson’s disease: a
case-control study. Lancet Neurol. 2011, 10, 898−908.
(3) (a) Marin, I. The Parkinson disease gene LRRK2: evolutionary
and structural insights. Mol. Biol. Evol. 2006, 23, 2423−2433.
(b) Raquel Esteves, A.; Swedlow, R. H.; Cardoso, S. M. LRRK2, a
puzzling protein: insights into Parkinson’s disease pathogenesis. Exp.
Neurol. 2014, 261, 206−216.
(12) Delbroek, L.; Van Kolen, K.; Steegmans, L.; da Cunha, R.;
Mandemakers, W.; Daneels, G.; De Bock, P.-J.; Zhang, J.; Gevaert, K.;
De Strooper, B.; Alessi, D. R.; Verstreken, P.; Moechars, D. W.
Development of an enzyme-linked immunosorbent assay for detection
of cellular and in vivo LRRK2 S935 phosphorylation. J. Pharm. Biomed.
Anal. 2013, 76, 49−58.
(4) (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. Characterization of a selective inhibitor of the Parkinson’s disease
kinase LRRK2. Nat. Chem. Biol. 2011, 7, 203−205. (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. Chemoproteomics-based design of potent LRRK2-selective
lead compounds that attenuate Parkinson’s disease-related toxicity in
human neurons. ACS Chem. Biol. 2011, 6, 1021−1028. (c) 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.
GSK2578215A; a potent and highly selective 2-arylmethoxy-5-
substituent-N-arylbenzamide LRRK2 kinase inhibitor. Bioorg. Med.
Chem. Lett. 2012, 22, 5625−5630. (d) 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. Discovery of highly potent, selective, and brain-
penetrable leucine-rich repeat kinase 2 (LRRK2) small molecule
inhibitors. J. Med. Chem. 2012, 55, 9416−9433. (e) Galatsis, P.;
Henderson, J. L.; Kormos, B. L.; Han, S.; Kurumbail, R. G.; Wager, T.
T.; Verhoest, P. R.; Noell, G. S.; Chen, Y.; Needle, E.; Berger, Z.;
Steyn, S. J.; Houle, C.; Hirst, W. D. Kinase domain inhibition of
leucine rich repeat kinase 2 (LRRK2) using a [1,2,4]triazolo[4,3-
b]pyridazine scaffold. Bioorg. Med. Chem. Lett. 2014, 24, 4132−4140.
(5) Wager, T. T.; Hou, X.; Verhoest, P. R.; Villalobos, A. Moving
beyond rules: the development of a central nervous system
multiparameter optimization (CNS MPO) approach to enable
alignment of druglike properties. ACS Chem. Neurosci. 2010, 1,
435−449.
(13) Graves, L. M.; Litchfield, D. W. “Going KiNativ”: probing the
native kinome. Chem. Biol. 2011, 18, 683−684.
(14) (a) Troxler, T.; Greenidge, P.; Zimmermann, K.; Desrayaud, S.;
Druckes, P.; Schweizer, T.; Stauffer, D.; Rovelli, G.; Shimshek, D. R.
Discovery of novel indolinone-based, potent, selective and brain
penetrant inhibitors of LRRK2. Bioorg. Med. Chem. Lett. 2013, 23,
4085−4090. (b) Mologni, L.; Rostagno, R.; Brussolo, S.; Knowles, P.
P.; Kjaer, S.; Murray-Rust, J.; Rosso, E.; Zambon, A.; Scapozza, L.;
McDonald, N. Q.; Lucchini, V.; Gambacorti-Passerini, C. Synthesis,
structure−activity relationship and crystallographic studies of 3-
substituted indolin-2-one RET inhibitors. Bioorg. Med. Chem. 2010,
18, 1482−1496. (c) Sun, L.; Tran, N.; Tang, F.; App, H.; Hirth, P.;
McMahon, G.; Tang, C. Synthesis and biological evaluations of 3-
substituted indolin-2-ones: a novel class of tyrosine kinase inhibitors
that exhibit selectivity toward particular receptor tyrosine kinases. J.
Med. Chem. 1998, 41, 2588−2603.
(15) Radu, M.; Chernoff, J. The DeMSTification of mammalian ste20
kinases. Curr. Biol. 2009, 19, R421−R425.
(16) Xing, L.; Rai, B.; Lunney, E. A. Scaffold mining of kinase hinge
binders in crystal structure database. J. Comput.-Aided Mol. Des. 2014,
28, 13−23.
(17) (a) 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.
Discovery of a highly selective, brain-penetrant aminopyrazole LRRK2
inhibitor. ACS Med. Chem. Lett. 2013, 4, 85−90. (b) 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.
Discovery of highly potent, selective, and brain-penetrant amino-
pyrazole leucine-rich repeat kinase 2 (LRRK2) small molecule
inhibitors. J. Med. Chem. 2014, 57, 921−936.
(6) (a) Manning, G.; Whyte, D. B.; Martinez, R.; Hunter, T.;
Sudarsanam, S. The protein kinase complement of the human genome.
Science 2002, 298, 1912−1934. (b) Zhang, D.; Lin, J.; Han, J.
Receptor-interacting protein (RIP) kinase family. Cell. Mol. Immunol.
2010, 7, 243−249.
(18) (a) Bertrand, J. A.; Thieffine, S.; Vulpetti, A.; Cristiani, C.;
Valsasina, B.; Knapp, S.; Kalisz, H. M.; Flocco, M. Structural
M
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
characterization of the GSK-3β active site using selective and non-
selective ATP-mimetic inhibitors. J. Mol. Biol. 2003, 333, 393−407.
(b) Bhat, R.; Xue, Y.; Berg, S.; Hellberg, S.; Ormo, M.; Nilsson, Y.;
̈
Radesater, A.-C.; Jerning, E.; Markgren, P.-O.; Borgegård, T.; Nylof,
̈
̈
M.; Gimen
́
ez-Cassina, A.; Hernan
́
dez, F.; Lucas, J. J.; Díaz-Nido, J.;
Avila, J. Structural insights and biological effects of glycogen synthase
kinase 3-specific inhibitor AR-A014418. J. Biol. Chem. 2003, 278,
45937−45945.
(19) 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.; Burdick, 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. Ser1292
autophosphorylation is an indicator of LRRK2 kinase activity and
contributes to the cellular effects of PD mutations. Sci. Transl. Med.
2012, 4, 164ra161.
(20) (a) Ness, D.; Ren, Z.; AGardai, S.; Sharpnack, D.; Johnson, V. J.;
Brennan, R. J.; Brigham, E. F.; Olaharski, A. J. Leucine-rich repeat
kinase 2 (LRRK2)-deficient rats exhibit renal tubule injury and
perturbations in metabolic and immunological homeostasis. PLoS One
2013, 8, e66164. (b) 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.; van der Putten,
P. H.; Rovelli, G.; Shimshek, D. R. LRRK2 protein levels are
determined by kinase function and are crucial for kidney and lung
homeostasis in mice. Hum. Mol. Genet. 2011, 20, 4209−4223.
(c) Tong, Y.; Yamaguchi, H.; Giaime, E.; Boyle, S.; Kopan, R.;
Kelleher Raymond, J., 3rd; Shen, J. Loss of leucine-rich repeat kinase 2
causes impairment of protein degradation pathways, accumulation of
alpha-synuclein, and apoptotic cell death in aged mice. Proc. Natl.
Acad. Sci. U.S.A. 2010, 107, 9879−9884.
(21) Greene, N.; Aleo, M. D.; Louise-May, S.; Price, D. A.; Will, Y.
Using an in vitro cytotoxicity assay to aid in compound selection for in
vivo safety studies. Bioorg. Med. Chem. Lett. 2010, 20, 5308−5312.
(22) Fuji, R. Nonclinical safety studies of selective LRRK2 kinase
inhibitors. Presented at Seventh Annual Parkinson’s Disease
Therapeutics Conference, The New York Academy of Sciences,
October 24, 2013.
(23) Ko, T. P.; Jeng, W. Y.; Liu, C. I.; Lai, M. D.; Wu, C. L.; Chang,
W. J.; Shr, H. L.; Lu, T. J.; Wang, A. H. J. Structures of human MST3
kinase in complex with adenine, ADP, an dMn2+. Acta Crystallogr.
2010, 66, 145−154.
(24) Molecular Operating Environment (MOE), version 2013.08;
Chemical Computing Group Inc. (1010 Sherbooke St. West, Suite No.
910, Montreal, QC, H3A 2R7, Canada), 2013.
(25) Wang, J.; Cieplak, P.; Kollman, P. A. How well does a restrained
electrostatic potential (RESP) model perform in calculating conforma-
tional energies of organic and biological molecules? J. Comput. Chem.
2000, 21, 1049−1074.
(26) The Universal Protein Resource (UniProt). Nucleic Acids Res.
2008, 36, D190−D195.
(27) (a) Desmond Molecular Dynamics System, version 3.1; D. E.
Shaw Research: New York, NY, 2012. (b) Maestro-Desmond
Interoperability Tools, version 3.1; Schrodinger: New York, NY, 2012.
̈
(c) Bowers, K. J.; Chow, E.; Xu, H.; Dror, R. O.; Eastwood, M. P.;
Gregersen, B. A.; Klepeis, J. L.; Kolossvary, I.; Moraes, M. A.;
Sacerdoti, F. D.; Salmon, J. K.; Shan, Y.; Shaw, D. E. Scalable
algorithms for molecular dynamics simulations on commodity clusters.
Proceedings of the ACM/IEEE Conference on Supercomputing (SC06),
Tampa, FL, , November 11−17, 2006; ACM Press: New York, 2006.
(28) Banks, J. L.; Beard, H. S.; Cao, Y.; Cho, A. E.; Damm, W.; Farid,
R.; Felts, A. K.; Halgren, T. A.; Mainz, D. T.; Maple, J. R.; Murphy, R.;
Philipp, D. M.; Repasky, M. P.; Zhang, L. Y.; Berne, B. J.; Friesner, R.
A.; Gallicchio, E.; Levy, R. M. Integrated modeling program, applied
chemical theory (IMPACT). J. Comput. Chem. 2005, 26, 1752−1780.
(29) Compounds 14 (PF-06447475) and 28 (PF-06454589) are now
available from Sigma Aldrich.
N
dx.doi.org/10.1021/jm5014055 | J. Med. Chem. XXXX, XXX, XXX−XXX