Discovery of [11C]MK-8193 as a PET tracer to measure target engagement of phosphodiesterase 10A (PDE10A) inhibitors

Article abstract of DOI:10.1016/j.bmcl.2015.05.080

Phosphodiesterase 10A (PDE10A) inhibition has recently been identified as a potential mechanism to treat multiple symptoms that manifest in schizophrenia. In order to facilitate preclinical development and support key proof-of-concept clinical trials of novel PDE10A inhibitors, it is critical to discover positron emission tomography (PET) tracers that enable plasma concentration/PDE10A occupancy relationships to be established across species with structurally diverse PDE10A inhibitors. In this Letter, we describe how a high-throughput screening hit was optimized to provide [¹¹C]MK-8193 (8j), a PET tracer that supports the determination of plasma concentration/PDE10A occupancy relationships for structurally diverse series of PDE10A inhibitors in both rat and rhesus monkey.

Full text of DOI:10.1016/j.bmcl.2015.05.080

Bioorganic & Medicinal Chemistry Letters xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of [11C]MK-8193 as a PET tracer to measure target
engagement of phosphodiesterase 10A (PDE10A) inhibitors
b
a
b
b
b
Christopher D. Cox a, , Eric D. Hostetler , Broc A. Flores , Jeffrey L. Evelhoch , Hong Fan , Liza Gantert ,
Marie Holahan b, Waisi Eng b, Aniket Joshi b, Georgia McGaughey c, Xiangjun Meng b, Mona Purcell b,
Izzat T. Raheem a, Kerry Riffel b, Youwei Yan d, John J. Renger e, Sean M. Smith e, Paul J. Coleman a
⇑
a Discovery Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
b Imaging, Merck Research Laboratories, West Point, PA 19486, USA
c Chemical Modeling & Informatics, Merck Research Laboratories, West Point, PA 19486, USA
d Structural Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
e Neuroscience, Merck Research Laboratories, West Point, PA 19486, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Phosphodiesterase 10A (PDE10A) inhibition has recently been identified as a potential mechanism to
treat multiple symptoms that manifest in schizophrenia. In order to facilitate preclinical development
and support key proof-of-concept clinical trials of novel PDE10A inhibitors, it is critical to discover posi-
tron emission tomography (PET) tracers that enable plasma concentration/PDE10A occupancy relation-
ships to be established across species with structurally diverse PDE10A inhibitors. In this Letter, we
describe how a high-throughput screening hit was optimized to provide [11C]MK-8193 (8j), a PET tracer
that supports the determination of plasma concentration/PDE10A occupancy relationships for struc-
turally diverse series of PDE10A inhibitors in both rat and rhesus monkey.
Received 26 April 2015
Revised 22 May 2015
Accepted 26 May 2015
Available online xxxx
Keywords:
PDE10A
Positron emission tomography
PET tracer
Ó 2015 Elsevier Ltd. All rights reserved.
Phosphodiesterase 10A
Imaging
Schizophrenia
Huntington’s disease
Following discovery of the phosphodiesterase 10A (PDE10A)
enzyme in 1999, rapid progress has been made in understanding
its physiological localization and function using genetic, biologic
and pharmacologic techniques.1 Pioneering work by a group at
Pfizer suggested that PDE10A inhibition may be an effective new
mechanism to ameliorate the positive symptoms of schizophrenia,
characterized most prominently by hallucinations and delusions.2
Subsequent efforts from groups across the pharmaceutical industry
have provided preclinical data suggesting that PDE10A inhibitors
may also be effective in addressing the cognitive impairment and
negative symptoms of schizophrenia, with a differentiated side-ef-
fect profile relative to currently marketed agents.3 PDE10A inhibi-
tors have also been proposed as potential therapeutic agents to
treat Huntington’s disease.4 Consistent with the significant thera-
peutic potential for PDE10A inhibition, no less than 20 companies
have filed patent applications on novel chemical matter for
PDE10A,3b and at least seven of those companies have initiated
clinical trials.5
Not surprisingly, based on a general agreement that positron
emission tomography (PET) tracers can be critical components of
CNS drug discovery efforts,6 many companies and academic labs
have pursued the development of PET tracers as target engagement
biomarkers to establish plasma concentration/PDE10A occupancy
relationships in preclinical efficacy studies and assist in dose-selec-
tion for clinical studies.7 Multiple PDE10A tracers have provided
promising results in preclinical species, and preliminary human
data for several of these indicate that they should indeed allow
assessment of target engagement in the human brain and enable
occupancy/efficacy relationships to be established in key Phase 2
proof-of-concept studies.7d,g–i Preliminary evidence has been
presented that suggests PDE10A PET tracers may also be valuable
as diagnostic biomarkers in the early stages of Huntington’s
disease.8 In this report, we describe the discovery of [11C]MK-
8193 (8j), a PDE10A PET tracer derived from a Merck high-
throughput screening (HTS) hit, and show its utility for measuring
target engagement of PDE10A inhibitors in the rat and rhesus mon-
key brain.
Our initial interest in PDE10A was spurred by the pioneering
work at Pfizer where highly potent and selective inhibitors such
⇑
Corresponding author. Tel.: +1 (215) 652 2411; fax: +1 (215) 652 7310.
E-mail address: chris_cox@merck.com (C.D. Cox).
http://dx.doi.org/10.1016/j.bmcl.2015.05.080
0960-894X/Ó 2015 Elsevier Ltd. All rights reserved.
Please cite this article in press as: Cox, C. D.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.05.080

2
C. D. Cox et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
Table 1
as MP-10 were discovered and employed to probe the biology of
PDE10A inhibition in vivo.9 Our first attempts at PET tracer devel-
opment were based on incorporation of radioisotopes onto the MP-
10 scaffold to provide ligands such as 1a and 1b that were studied
in vitro and in vivo. As others have reported,7a–c we found that the
slow kinetics, high non-displaceable binding, and the potential to
form brain-penetrant metabolites with 1a and 1b made these
analogs less than ideal for PET tracer development;10 however, this
work was instrumental in illustrating that a useful PDE10A PET tra-
cer was highly feasible given the density and localization of
PDE10A.
Initial SAR exploration of HTS hit 5
2
NH
O
3
4
O
N
1
O
N
R1
O
1
4
2
N
N
R2
5
3
N
N
O
O
6
7
Compound
R1
R2
PDE10A Ki (nM)
6a
6b
6c
6d
6e
6f
6g
6h
6i
6j
6k
6l
6m
6n
6o
7
2-Me
3-Me
4-Me
4-F
H
H
H
H
H
H
H
H
1124
109
87
280
123
265
49
Me
N
R
N
N
N
4-Cl
4-CN
4-Ph
4-OMe
N
N
O
O
61
64
N
N
R = 11CH3
3,4-DiOMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
4-OMe
—
H
1a
MP-10
1-Cl
2-Cl
3-Cl
4-Cl
3-OMe
3-Me
—
171
163a
45
1b R = CH2CH218F
>1000a
39
Rather than try to optimize the MP-10 series into a suitable tra-
cer as other groups have reported,7b,c,f we instead elected to
progress a novel scaffold identified from our ongoing lead-finding
efforts. To that end, we had earlier embarked on a broad effort to
identify novel PDE10A inhibitors that included HTS, ALIS
(Automated Ligand Identification System),11 and fragment screen-
ing, and our laboratory has recently disclosed novel series of
PDE10A inhibitors derived from hits identified by each of these
methods: 2, 3, and 4, respectively (Fig. 1).12 An additional hit iden-
tified from HTS was quinazolinone 5, a compound that stood out as
having good potency with moderate ligand binding efficiency (LBE)
and reasonable calculated properties (cLogP = 3.9, PSA = 75).
Initial SAR efforts in the quinazolinone series made use of the
one-pot procedure illustrated in Scheme 1 for exploration of the
15
13
a
Represents data from a single determination.
northeastern (blue) and western (red) aryl groups. Though many
additional compounds were made and tested, the progression of
SAR in Table 1 highlights the salient features.13 Compounds 6a–c
illustrate that whereas methyl substitution at the 2-position of
the northeastern aryl is not tolerated, methyl substituents at the
3- or 4-positions are slightly potency enhancing, with a small
advantage provided by the latter. Compounds 6d–i indicate that
whereas 4-F, 4-Cl, and 4-CN are detrimental to potency relative
to 4-Me, 4-OMe and 4-Ph are potency enhancing, and disubstitu-
tion at the 3- and 4-positions is not advantageous over the mono-
substituted analog (cf. 6h ? 6i). Holding the R1 group constant as
OMe and walking a chlorine atom around the western aryl ring
at the R2 position (6j–m) indicates that 3-substitution is optimal,
with 6l demonstrating a potency of 45 nM. Changing R2 to a more
polar 3-OMe group maintained potency, but 3-Me provided the most
potent analog, 6o, with PDE10A Ki = 15 nM (LBE = 0.33). We also
explored aryl rings other than phenyl in the northeastern region,
but most compounds led to a significant reduction in potency; one
notable exception was indole 7 that, without the potency-enhancing
R2 substituent at the 3-position on the western aryl, provided a
14 nM inhibitor (LBE = 0.32). Though increases in activity were mod-
est to this point, 6o and 7 provided a nearly a 10-fold increase in
potency over the HTS hit 5 while maintaining acceptable LBE and
physical properties, thus providing us with inspiration to continue
SAR exploration in other regions of the molecule.
MeO
MeO
N
NH
N
O
Me
N
N
N
N
H
N
N
N
N
N
H2N
Me
2
(HTS)
3
(ALIS)
PDE10A Ki = 94 nM
LBE = 0.27
PDE10A Ki = 90 nM
LBE = 0.39
Me
O
Cl
Cl
N
O
N
N
N
N
O
With the optimal western and northeastern substitutions in
hand, we then focused our attention on the phthalimide and linker
region. Decreasing the linker length to one carbon (8a) or increas-
ing it to three carbons (8b) were both detrimental to potency
(Table 2), as were aliphatic substitutions on the alkyl chain (data
4
(fragment)
5
(HTS)
PDE10A Ki = 8.6 µM
PDE10A Ki = 130 nM
LBE = 0.31
LBE = 0.57
Figure 1. Hits identified by internal screening at Merck.
O
R1
O
O
O
R1
N
O
R2
OH
H2N
H2O
N
R2
N
N
NH2
a
R3
O
O
R3
Scheme 1. Reagents and conditions: (a) P(OPh)3, pyridine, 100 °C in a sealed tube; then add aniline and heat to 100 °C (yields: 25–75%).
Please cite this article in press as: Cox, C. D.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.05.080

C. D. Cox et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
3
Table 2
PDE10A inhibitor in this series with Ki = 150 pM (LBE = 0.37).
Installation of a 1-bromo substituent to provide 8k was potency
neutral with respect to the 1-Cl analog 8c, but provided a conve-
nient handle for installation of aryl substituents via cross-coupling
chemistry. Among the multiple 1-aryl groups installed, the 2-thia-
zole (8l) and the 4-pyridyl (8m) stood out as two of the most inter-
esting in terms of potency and drug-like properties.
Additional SAR exploration of leading quinazolinone 6h
OMe
O
N
O
1
N
N
n
2
O
R3
At this point in our exploration of the quinazolinones, we noted
that although the pharmacokinetics of leading compounds was
uniformly poor, with clearance in rat being near hepatic blood
flow, the rest of the profile looked favorable as illustrated in
Table 3 for the most potent analogs described thus far, 8j and
8m. These compounds have: (1) improved LBE’s relative to HTS
hit 5, (2) HPLC LogDs in a drug-like range,14 (3) high selectivity
for PDE10A over the other PDEs; (4) moderate plasma protein
binding; (5) no Pgp efflux liability in human or rat, and (6) very
good cellular permeability. These characteristics, along with being
highly amenable to radiolabel incorporation at the phenolic methyl
ether, made them ideal candidates to investigate as potential PET
tracers. At this point, two additional compounds were rationally
designed as potential PET tracers for in vivo analysis: (1) com-
pound 9, incorporating a nitrogen at the 8-position of the quina-
zolinone core, was made in an effort to further reduce the
lipophilicity of 8j, and (2) compound 10, the most potent analog
identified within this series, which was made based on indole 7,
but incorporating an additional nitrogen to form an indazole sys-
tem which lowered lipophilicity relative to the corresponding
indole analog.
3
4
8
Compound
n
R3
PDE10A Ki (nM)
8a
8b
8c
8d
8e
8f
8g
8h
8i
0
2
1
1
1
1
1
1
1
1
1
1
1
H
H
1-Cl
2-Cl
1,4-DiCl
1-Me
1-OH
1-OMe
1-OEt
1-OiPr
1-Br
1-(2-Thiazole)
1-(4-Pyridyl)
>1000a
219
7.7
30
6.3
6.1
20
2.0
1.0
8j
8k
8l
0.15
8.4a
1.8
8m
0.072
a
Represents data from a single determination.
not shown). Most attempts to remove or alter the phthalimide
group were also unsuccessful, though some efforts toward that
end were fruitful and will be the subject of a future publication.
However, investigating substitution around the aryl (green) ring
of the phthalimide provided the key breakthrough in potency we
were seeking. Compounds 8c–d indicate that substitution at the
1-position was more favorable than the 2-position, with the 1-Cl
analog demonstrating single-digit nanomolar potency for the first
time (8c). 1,4-Dichloro substitution provided no significant
improvement over monosubstitution (8e vs 8c) but significantly
increased lipophilicity. Altering 1-Cl to 1-Me was potency neutral
(8f), but the corresponding phenol 8g resulted in a 3-fold loss in
activity. However, using the phenol as a synthetic handle to exam-
ine various alkoxy groups at the 1-position illustrated that increased
potency can be had by increasing the size of the substituent (8h–j),
with the isopropyl analog 8j providing the first sub-nanomolar
The four compounds illustrated in Table 3 were radiolabeled
with 11C and studied in vivo in rhesus monkey to assess brain
uptake and specific binding to PDE10A. Somewhat surprisingly,
compound 8j which, based on in vitro data looked less promising
than the others based on its intermediate potency and higher
HPLC LogD,15 was the standout in terms of in vivo profile with high
brain uptake, rapid washout, and a very large apparent signal.16
A
baseline PET image of [11C]8j in rhesus monkey brain is shown in
Figure 2A. As expected based on the distribution of PDE10A in
the brain, highest tracer uptake was observed in the striatum with
very low tracer retention in all other brain regions. Striatal uptake
in rat brain was lower than in rhesus monkey, but rapid washout in
regions devoid of PDE10A result in a useful specific signal (Fig. 3A).
Table 3
In vitro characterization of leading tracer candidates
N
N
OR
OR
O
R
O
N
O
N
O
N
N
N
N
N
O
O
O
Me
X
N
N
N
O
O
O
8j
8m
10
R = CH3
R = CH3; X = CH
R = CH3
R = 11CH3; X = CH
R = 11CH3
R = 11CH3
[
11C]-10
[
11C]-8j
[
11C]-8m
9
[
R = CH3; X = N
11C]-9
R = 11CH3; X = N
8j
9
8m
10
PDE10A Ki (nM)
LBE
0.15
0.37
3.6
0.37
0.35
2.8
0.072
0.36
2.8
>29,000
93
0.025
0.37
3.4
>100,000
98
HPLC LogD (pH 7)
PDE selectivitya
>60,000
>27,000
Protein binding (human) (%)
Pgp (B-A/A-B)
97
61
28
83
61
39
61
31
61
27
Permeability (Â10À6 cm/s)
a
PDE selectivity is calculated by the following formula: PDEX Ki/PDE10A Ki wherein PDEX is the PDE other than PDE10A with the most potent Ki.
Please cite this article in press as: Cox, C. D.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.05.080

4
C. D. Cox et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
The synthetic route utilized to provide [11C]8j is highlighted in
1.0
0.5
0
Scheme 2.18 The 1-isopropoxyphthalimide intermediate 13 was
made via a three-step procedure starting with 11, and then cou-
pled via the one-pot procedure described in Scheme 1 to provide
benzyl-protected phenol 14. Removal of the benzyl group with
TMSBr in TFA with thioanisole provided the labeling precursor
15. Synthesis of [11C]8j was carried out in good yield with high
B
Striatum
A
specific activity (>1 Ci/
Cs2CO3 in DMF.
l
mol) and purity using [11C]CH3I and
With a robust route available to access quantities of [11C]8j on
demand, we proceeded to study its utility for evaluating target
engagement of a leading series of PDE10A inhibitors in rat and rhe-
sus monkey brain. Figure 2 shows a summed PET image of [11C]8j
in rhesus monkey both at baseline and after a 15 mpk iv dose of the
PDE10A inhibitor THPP-1.12a The markedly reduced tracer
retention in the striatum after THPP-1 administration indicates
the ability of [11C]8j to determine PDE10A occupancy for this
structurally diverse PDE10A inhibitor. Similar results were
observed in the rat brain at baseline and after a 6 mpk i.v. dose
of THPP-1, as illustrated in Figure 3. A forthcoming publication
Figure 2. Summed PET images (30–90 min) of [11C]8j in rhesus monkey brain
overlaid on MRI under baseline conditions (A) and after administration of THPP-1
(B). The baseline PET study shows high tracer retention (red color) in the striatum.
Tracer binding in the striatum is blocked after administration of THPP-1. Scale is in
SUV (standardized uptake value).
B
A
will demonstrate the usefulness of
[
11C]8j for determining
Striatum
quantitative plasma exposure/PDE10A occupancy relationships in
both rat and rhesus monkey striatum, as well as provide data
that support the expectation that [11C]8j will also provide a
useful specific signal in humans.19,20
Figure 3. Summed PET images (0–90 min) of [11C]8j in rat brain under baseline
conditions (A) and after administration of THPP-1 (B). The baseline PET study shows
high tracer retention (red color) in the striatum. Tracer binding in the striatum is
blocked after administration of THPP-1. Scale is in SUV (standardized uptake value).
In contrast to the excellent in vivo rhesus monkey data for
[
11C]8j, [11C]8m displayed high non-specific binding and a possible
brain-penetrant metabolite in rhesus, while [11C]9 had a smaller
specific signal relative to [11C]8j, despite having similar potency
with more drug-like physicochemical properties (lower HPLC
LogD and greater free fraction in plasma). The most potent com-
pound, [11C]10, provided a very large signal in vivo, but was slow
to reach steady-state, making occupancy determinations less cer-
tain. We also studied several 11C-labeled analogs from the THPP
series (derived from HTS hit 2) in vivo in rhesus monkeys that
had in vitro profiles nearly identical to that of 8j; however, these
tracers uniformly displayed poor brain uptake for reasons that
are not completely understood. The uncertainty of translation from
in vitro predictions to in vivo behavior underscores the advantages
of having an embedded imaging chemist on a project team with
resources to study multiple promising compounds in vivo.17
Importantly for our preclinical development program, a tritiated
version of 8j, [3H]8j, was found to be a useful tracer for establishing
a high-throughput, ex vivo occupancy assay in the rat, and future
publications from our group will demonstrate how this assay
allowed us to drive lead optimization efforts based on maximizing
PDE10A occupancy in the brain while minimizing plasma exposure.
It is important to note that, as illustrated in Figures 2 and 3, [11C]8j
is able to determine PDE10A occupancy of compounds from series
that are structurally diverse from the tracer itself. In fact, during
the course of our program, we have investigated the ability of
[
11C]8j and [3H]8j to determine plasma/occupancy correlations
across many structurally diverse series of PDE10A inhibitors,
including the THPPs, leads derived from ALIS hit 3 and fragment
OH
a
O
O
O
O
O
O
O
OH
Cl
O
O
N
O
HO
N
b,c
O
NH3
O
O
O
12
13
11
O
O11CH3
O
OR
OH
O
O
N
NH2
N
N
O
d
f
O
N
N
N
14 R = OBn
15 R = OH
O
O
e
[
11C]8j
Scheme 2. Reagents and conditions: (a) dioxane, triethylamine, 50 °C (65% yield); (b) Cs2CO3, iPrI, CH3CN, 40 °C; (c) TFA, DCM (75% for 2 steps); (d) P(OPh)3, pyridine, 100 °C
in a sealed tube; then p-OBn aniline (78% yield); (e) TMSBr, TFA, thioanisole, 0 °C (64% yield); (f) [11C]CH3I, Cs2CO3, DMF, 45 °C, 3 min.
Please cite this article in press as: Cox, C. D.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.05.080

C. D. Cox et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
5
References and notes
1. Chappie, T. A.; Helal, C. J.; Hou, X. J. Med. Chem. 2012, 55, 7299.
2. (a) Menniti, F. S.; Chappie, T. A.; Humphrey, J. M.; Schmidt, C. J. Curr. Opin.
Invest. Drug 2007, 8, 54; (b) Schmidt, C. J.; Chapin, D. S.; Cianfronga, M. L.;
Corman, M. L.; Hajos, M.; Harms, J. F.; Hoffman, W. E.; Lebel, L. A.; McCarthy, S.
A.; Nelson, F. R.; Proulx-LaFrance, C.; Majchrzak, M. J.; Ramirez, A. D.; Schmidt,
K.; Seymour, P. A.; Siuciak, J. A.; Tingley, F. D., III; Williams, P. R.; Verhoest, P. R.;
Menniti, F. S. J. Pharmacol. Exp. Ther. 2008, 325, 681.
3. (a) Grauer, S. M.; Pulito, V. L.; Navarra, R. L.; Kelly, M. P.; Kelley, C.; Graf, R.;
Langen, B.; Logue, S.; Brennan, J.; Jiang, L.; Charych, E.; Engerland, U.; Liu, F.;
Marquis, K. L.; Malamas, M.; Hage, T.; Comery, T. A.; Brandon, N. J. J. Pharmacol.
Exp. Ther. 2009, 331, 574; (b) Kehler, J.; Nielsen, J. Curr. Pharm. Des. 2011, 17,
137; (c) Smith, S. M.; Uslaner, J. M.; Cox, C. D.; Huszar, S. L.; Cannon, C. E.;
Vardigan, J. D.; Eddins, D.; Toolan, D. M.; Kandebo, M.; Yao, L.; Raheem, I. T.;
Schreier, J. D.; Breslin, M. J.; Coleman, P. J.; Renger, J. J. Neuropharmacology
2013, 64, 215; (d) Nawrocki, A. R.; Rodriguez, C. G.; Toolan, D. M.; Price, O.;
Henry, M.; Forrest, G.; Szeto, D.; Keohane, C. A.; Pan, Y.; Smith, K. M.; Raheem, I.
T.; Cox, C. D.; Hwa, J.; Renger, J. J.; Smith, S. M. Diabetes 2014, 63, 300.
4. (a) Kleiman, R. J.; Kimmel, L. H.; Bove, S. E.; Lanz, T. A.; Harms, J. F.; Romegialli,
A.; Miller, K. S.; Willis, A.; Des-Etages, S.; Kuhn, M.; Schmidt, C. J. J. Pharmacol.
Exp. Ther. 2011, 336, 64; (b) Giampà, C.; Laurenti, D.; Anzilotti, S.; Bernardi, G.;
Menniti, F. S.; Fusco, F. R. PLoS One 2010, 5, e13417.
Phe719
Gln716
Tyr683
Figure 4. X-ray of 8j in the PDE10A enzyme active site. Key residues and hydrogen-
5. The following companies have publically announced clinical trials with their
associated PDE10A inhibitors: Pfizer (PF-254590), Amgen (AMG 579), Roche
(RO 5545965), Lundbeck (Lu AF 11167), Takeda (TAK-063), En Vivo/Forum
(EVP-6308), and Omeros (OMS824 and OMS643762).
bonding interactions are highlighted.
6. Honer, M.; Gobbi, L.; Martarello, L.; Comley, R. Drug Discovery Today 2014, 19,
1936.
hit 4, as well as MP-10 itself; in every case, the tracer allowed us to
determine plasma/PDE10A occupancy relationships and enable
cross-compound, cross-series and cross-species comparisons.
The ability of [11C]8j to determine occupancy across multiple
series of PDE10A inhibitors can be readily explained by examining
the X-ray crystal structure of 8j bound to the PDE10A catalytic core
domain as shown in Figure 4.21 As expected, the tracer binds in the
active site and forms direct interactions with key active site resi-
dues: One phthalimide oxygen is engaged in a hydrogen bonding
interaction (dN–O = 3.2 Å) with the conserved Gln 716, the phthal-
7. (a) Tu, Z.; Fan, J.; Li, S.; Jones, L. A.; Cui, J.; Padakanti, P. K.; Xu, J.; Zeng, D.;
Shoghi, K. I.; Perlmutter, J. S.; Mach, R. H. Bioorg. Med. Chem. 2011, 19, 1666; (b)
Andrés, J.-I.; De Angelis, M.; Alcàzar, J.; Iturrino, L.; Langlois, X.;
Dedeurwaerdere, S.; Lenaerts, I.; Vanhoof, G.; Celen, S.; Bormans, G. J. Med.
Chem. 2011, 54, 5820; (c) Celen, S.; Koole, M.; Ooms, M.; DeAngelis, M.; Sannen,
I.; Cornelis, J.; Alcazar, J.; Schmidt, M.; Verbruggen, A.; Langlois, X.; VanLaere,
K.; Andrés, J. I.; Bormans, G. NeuroImage 2013, 82, 13; (d) Van Laere, K.; Ahmad,
R. U.; Hudyana, H.; Dubois, K.; Schmidt, M. E.; Celen, S.; Bormans, G.; Koole, M.
J. Nucl. Med. 2013, 54, 1285; (e) Hwang, D.-R.; Hu, E.; Rumfelt, S.;
Easwaramoorthy, B.; Castrillon, J.; Davis, C.; Allen, J. R.; Chen, H.; Treanor, J.;
Abi-Dargham, A.; Slifstein, M. Nucl. Med. Biol. 2014, 41, 343; (f) Ooms, M.; Celen,
S.; Koole, M.; Langlois, X.; Schmidt, M.; DeAngelis, M.; Andrés, J. I.; Verbruggen,
A.; Van Laere, K.; Bormans, G. Nucl. Med. Biol. 2014, 41, 695; (g) Plisson, C.;
Weinzimmer, D.; Jakobsen, S.; Natesan, S.; Salinas, C.; Lin, S.-F.; Labaree, D.;
Zheng, M.-Q.; Nabulsi, N.; Marques, T. R.; Kapur, S.; Kawanishi, E.; Saijo, T.;
Gunn, R. N.; Carson, R. E.; Rabiner, E. A. J. Nucl. Med. 2014, 55, 595; (h) Barret, O.;
Thomae, D.; Tavares, A.; Alagille, D.; Papin, C.; Waterhouse, R.; McCarthy, T.;
Jennings, D.; Marek, K.; Russell, D.; Seibyl, J.; Tamagnan, G. J. Nucl Med. 2014,
55, 1297; (i) Kehler, J.; Kilburn, J. P.; Estrada, S.; Christensen, S. R.; Wall, A.;
Thibblin, A.; Lubberink, M.; Bundgaard, C.; Brennum, L. T.; Steiniger-Brach, B.;
Christoffersen, C. T.; Timmermann, S.; Kreilgaard, M.; Antoni, G.; Bang-
Andersen, B.; Nielson, J. J. Nucl Med. 2014, 55, 1513; (j) Harada, A.; Suzuki, K.;
Miura, S.; Hasui, T.; Kamiguchi, N.; Ishii, T.; Taniguchi, T.; Kuroita, T.; Takano,
A.; Stepanov, V.; Halldin, C.; Kimura, H. Nucl. Med. Biol. 2015, 42, 146.
8. (a) Ahmad, R.; Bourgeois, S.; Postnov, A.; Schmidt, M. E.; Bormans, G.; Van
Laere, K.; Vandenberghe, W. Neurology 2014, 82, 279; (b) Russell, D. S.; Barret,
O.; Jennings, D. L.; Friedman, J. H.; Tamagnan, G. D.; Thomae, D.; Alagille, D.;
Morley, T. J.; Papin, C.; Papapetropoulos, S.; Waterhouse, R. N.; Seibyl, J. P.;
Marek, K. L. JAMA Neurol. 2014, 71, 1520.
imide aryl group engages in a
p-stacking interaction with Phe
719, and the quinazolinone ring fills the PDE10A ‘selectivity
pocket’ with the nitrogen at the 1-position of the ring engaged in
a
hydrogen-bonding interaction (dN–O = 2.7 Å) with Tyr-683.
Notably, all reported PDE10A inhibitors have been found to bind
in this same region of the enzyme and make at least two of these
three key interactions, nicely explaining the utility of 8j to enable
occupancy measurement across many diverse series of PDE10A
inhibitors.
Based on the results of in vitro and in vivo studies, some of
which were shown herein, as well as additional favorable charac-
teristics such as safety, selectivity, pharmacokinetics, metabolism
profile, and pharmaceutical properties, [11C]8j was selected for
development as a target engagement biomarker to support devel-
opment of PDE10A inhibitors and identified as [11C]MK-8193.
Forthcoming reports will provide a more detailed characterization
of the in vitro and in vivo behavior of this tracer, and demonstrate
how it was used to drive our lead optimization program.
9. Verhoest, P. R.; Chapin, D. S.; Corman, M.; Fonseca, K.; Harms, J. F.; Hou, X.;
Marr, E. S.; Menniti, F. S.; Nelson, F.; O’Connor, R.; Pandit, J.; Proulx-LaFrance,
C.; Schmidt, A. W.; Schmidt, C. J.; Suiciak, J. A.; Liras, S. J. Med. Chem. 2009, 52,
5188.
10. There has recently been some debate as to the tractability of radiolabeled
analogs of MP-10 for clinical use; see: Lin, S.-F.; Lebaree, D.; Chen, M.-K.;
Holden, D.; Gallezot, J.-D.; Kapinos, M.; Teng, J.-K.; Najafzadeh, S.; Plisson, C.;
Rabiner, E. A.; Gunn, R. N.; Carson, R. E.; Huang, Y. Synapse 2015, 69, 86.
11. ALIS is Merck’s proprietary affinity selection-mass spectrometry (AS-MS)-
based system; see: Huang, X.; Cheng, C. C.; Fischmann, T. O.; Duca, J. S.; Yang,
Y.; Richards, M.; Shipps, G. W. ACS Med. Chem. Lett. 2012, 2, 123.
Acknowledgments
The authors would like to thank the West Point NMR and Mass
Spectrometry facilities for structure elucidation support, Rodney A.
Bednar and Wei Lemaire (In Vitro Pharmacology) for PDE potency
measurements, Cristian Salinas (Imaging) for figure preparation,
and Hua Su (Chemical Modeling & Informatics) for assistance with
refining and submitting the X-ray structure of 8j to the PDB.
Additionally, the following colleagues are gratefully acknowledged
for their valuable contributions that led to development approval
for [11C]MK-8193: Jing Li (Discovery Process Chemistry), Nicole
Buist and Joy Fuerst (Discovery Pharmaceutical Sciences), Bennett
Ma and Somang H. Kim (Drug Metabolism, Pharmacokinetics and
Disposition), and Joseph Salata (Safety Assessment). This work
was funded by Merck.
12. (a) Raheem, I. T.; Breslin, M. J.; Fandozzi, C.; Fuerst, J.; Hill, N.; Huszar, S.;
Kandebo, M.; Kim, S. H.; Ma, B.; McGaughey, G.; Renger, J. J.; Schreier, J. D.;
Sharma, S.; Smith, S. M.; Uslaner, J.; Yan, Y.; Coleman, P. J.; Cox, C. D. Bioorg.
Med. Chem. Lett. 2012, 22, 5903; (b) Shipe; William, D.; Abeywickrema, P. D.;
Barrow, J. C.; Bednar, R.; Bruno, J.; Cofre, V.; Coleman, P. J.; Collin, D.; Cox, C. D.;
Fandozzi, C.; Getty, K.; Huszar, S.; Hutson, P.; Kemmerer, A.; Krishnan, R.;
Lemaire, W.; Ma, B.; McGaughey, G.; Munshi, S.; Nguyen, S.; Nolt, M. B.;
Parmentier-Batteur, S.; Raheem, I. T.; Sharik, S. S.; Sharma, S.; Smith, S. M.;
Theberge, C.; Uslaner, J.; Yan, Y. Abstracts of Papers, 247th National Meeting of
the American Chemical Society, Dallas, TX; American Chemical Society:
Washington, DC, 2014. MEDI-20; (c) Cox, C. D. Abstracts of Papers, 248th
National Meeting of the American Chemical Society, San Francisco, CA; American
Chemical Society: Washington, DC, 2014. MEDI-273.
13. All PDE10A inhibitory data reported in this manuscript is the average of two or
more determinations unless otherwise noted. Experimental details on the
assay used to determine the Ki of PDE10A inhibition can be found in: Smith, S.
Please cite this article in press as: Cox, C. D.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.05.080

6
C. D. Cox et al. / Bioorg. Med. Chem. Lett. xxx (2015) xxx–xxx
M.; Uslaner, J. M.; Cox, C. D.; Huszar, S. L.; Cannon, C. E.; Vardigan, J. D.; Eddins,
17. The Amgen group has reported a high-throughput LC-MS/MS technique that
allowed them to predict in vivo behavior of potential PDE10A PET tracers based
on in vitro assays and led them to the discovery of a useful PET ligand, thus
offering an alternative, experimental way to increase confidence in tracer
discovery: Hu, E.; Biorn, C.; Lester-Zeiner, D.; Cho, R.; Rumfelt, S.; Kunz, R. K.;
Nixey, T.; Michelsen, K.; Miller, S.; Shi, J.; Wong, J.; Della Puppa, G. H.; Able, J.;
Talreja, S.; Hwang, D.-R.; Hitchcock, S. A.; Porter, A.; Immke, D.; Allen, J. R.;
Treanor, J.; Chen, H. J. Med. Chem. 2012, 55, 4776.
D.; Toolan, D. M.; Kandebo, M.; Yao, L.; Raheem, I. T.; Schreier, J. D.; Breslin, M.
J.; Coleman, P. J.; Renger, J. J. Neuropharmacology 2013, 64, 215.
14. Though we use cLogP calculations in a predictive sense when designing
analogs, we relied on measured HPLC LogD values (at pH 7) as better measure
of true lipophilicity for this series because HPLC LogD correlated better with
measured shake-flask LogP than did cLogP; for instance, whereas the cLogP of
8j is 4.6, the HPLC LogD is 3.6 and the shake-flask LogP is 3.19. In general,
cLogP and HPLC LogD show the same stack-ranking of compounds, but the
absolute values of the cLogP measurements were misleadingly high.
18. For full synthetic details, see: Cox, C. D.; Flores, B. A.; Hostetler, E.; Fan, H. US
Patent 8,846,000, 2014.
15. Though HPLC LogD’s were used to monitor the optimization process, we
believe shake-flask LogP is a better predictor of suitability for PET tracer
19. Hostetler, E. D.; Fan, H.; Joshi, A.; Zeng, Z.; Eng, W.; Gantert, L.; Holahan, M.;
Meng, X.; Miller, P.; O’Malley, S.; Purcell, M.; Riffel, K.; Salinas, C.; Williams, M.;
Smith, S. M.; Coleman, P. J.; Cox, C. D.; Flores, B. A.; Raheem, I. T.; Cook, J. J.;
Evelhoch, J. L. for submission to Mol. Imaging Biol.
20. Scientists at Molecular Neuroimaging, LLC (New Haven, CT) have recently
reported human clinical data with [18F]MNI-659, a very close analog of [11C]8j,
and found it to be an excellent radiotracer for quantifying PDE10A in humans,
consistent with our predictions; see Refs. 6,8b.
development. With
a LogP of 3.19, 8j is predicted to have lipophilicity
supportive of PET tracer development; however, it is more lipophilic than the
other analogs in Table 3.
16. Based on this promising data, the 18F fluoroethyl analog of 8j was also made
and studied in vivo; despite showing favorable attributes, we favored the
advantages of 11C over 18F which include lower radiation exposure for patients
and the ability to image the same patient multiple times per day to facilitate
time-course occupancy studies.
21. Coordinates have been deposited with RCSB Protein Data Bank (PDB) under the
deposition number 4ZO5.
Please cite this article in press as: Cox, C. D.; et al. Bioorg. Med. Chem. Lett. (2015), http://dx.doi.org/10.1016/j.bmcl.2015.05.080