Novel, potent, selective, and brain penetrant phosphodiesterase 10A inhibitors

Article abstract of DOI:10.1016/j.bmcl.2018.12.029

Herein we report the discovery of a novel series of phosphodiesterase 10A inhibitors. Optimization of a HTS hit (17) resulted in potent, selective, and brain penetrant 23 and 26; both exhibited much lower clearance in vivo and decreased volume of distribu

Full text of DOI:10.1016/j.bmcl.2018.12.029

Accepted Manuscript
Novel, potent, selective, and brain penetrant phosphodiesterase 10A inhibitors
Hervé Geneste, Karla Drescher, Clarissa Jakob, Loïc Laplanche, Michael Ochse,
Maricel Torrent
PII:
S0960-894X(18)30976-4
https://doi.org/10.1016/j.bmcl.2018.12.029
BMCL 26198
DOI:
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
6 November 2018
12 December 2018
13 December 2018
Please cite this article as: Geneste, H., Drescher, K., Jakob, C., Laplanche, L., Ochse, M., Torrent, M., Novel, potent,
selective, and brain penetrant phosphodiesterase 10A inhibitors, Bioorganic & Medicinal Chemistry Letters (2018),
doi: https://doi.org/10.1016/j.bmcl.2018.12.029
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Bioorganic & Medicinal Chemistry Letters
Novel, potent, selective, and brain penetrant phosphodiesterase 10A inhibitors
Hervé Geneste,a,* Karla Drescher,a,§ Clarissa Jakob,b Loïc Laplanche,c Michael Ochse,a Maricel Torrent,b
aAbbVie Deutschland GmbH & Co. KG, Neuroscience Research, D-67008 Ludwigshafen, Germany; bAbbVie, Structural Biology, IL 60064-6101, U.S.A; cAbbVie
Deutschland GmbH & Co. KG, DMPK, D-67008 Ludwigshafen, Germany.
dedicated to the memory of Professor Jacques Coste, ENSC Montpellier, France
ARTICLE INFO
ABSTRACT
Article history:
Received
Revised
Herein we report the discovery of a novel series of phosphodiesterase 10A inhibitors.
Optimization of a HTS hit (17) resulted in potent, selective, and brain penetrant 23 and 26; both
exhibited much lower clearance in vivo and decreased volume of distribution (rat PK) and have
thus the potential to inhibit the PDE10A target in vivo at a lower efficacious dose than the
Accepted
Available online
reference
compound
©2018 Elsevier Ltd. All rights reserved.
WEB-3.
Keywords:
phosphodiesterase
PDE10A
WEB-3
Inhibitors of phosphodiesterase 10A (PDE10A) offer a promising
therapeutic approach for the treatment or prevention of
neurological and psychiatric disorders, in particular
schizophrenia and related disorders. Activity in several
antipsychotic models has been shown with papaverine,1 the first
extensively profiled pharmacological tool compound for this
target. These models suggest that PDE10A inhibition has the
classical antipsychotic potential. Moreover inhibition of PDE10A
reverses sub-chronic PCP-induced deficits in attentional set-
shifting in rats suggesting that PDE10A inhibitors might alleviate
cognitive deficits associated with schizophrenia.2
In the past decade, multiple PDE10A inhibitors have been
advanced to clinical trials for the treatment of neurological
disorders and psychiatric diseases but none has yet reached the
market.7
We present here the discovery of novel, selective and brain
penetrant PDE10A inhibitors (Fig. 2) depicting a WEB-3-like
binding mode. The new PDE10A inhibitors are PDE-selective,8a
metabolically stable, Ames8b negative and cover a different CNS
druglike space (clogP < 3, CNS MPO score ≥ 5.0) compared to
formerly published related structural series.9
Identification and optimization of PDE10A inhibitors are highly
supported by X-ray structures: three main binding modes are
reported and illustrated in the literature3 by papaverine, MP-10
and WEB-3. Most lead optimization programs targeting PDE10A
inhibitors started either from papaverine4 or from MP-10.5
However in many cases PDE isoenzyme selectivity was a
challenge. WEB-3,6 out of the imidazo[1,5-a]-pyrido[3,2-
e]pyrazine chemical series, depicted a novel binding mode into
the PDE10A pocket. However its mutagenic potential (Ames
positive) and its sub-optimal metabolic stability had to be
addressed.6
Two different approaches starting from WEB-3 were investigated
to come to proprietary PDE10A inhibitors: the thiazolopyrazine10
and cinnoline11 structural series were thus identified (Fig. 1).
Figure 1. WEB-36 and AbbVie proprietary series (thiazolopyrazine (I & II)
& cinnoline (III) derivatives)
As for most PDE10A inhibitors, H-bond interaction with Gln726
is critical.3 In the case of WEB-3 (Fig. 1), additional interactions
include water-mediated H-bonds with Tyr524 and Asp674 as
well as -stacking interactions with Phe729, Phe696 and Ile692
(hydrophobic clamp).6
______
*Corresponding author:
Tel.: +49-621-589-3423; e-mail: herve.geneste@abbvie.com

N
N
N
iii
ii
i
NH2
Cl
NHBoc
S
COOH
S
S
H
12
11
N
10
Br
N
O
O
O
N
N
N
iv, v
vi
N
N
N
S
S
S
H
H
H
13
14
15
N
Br
S
N
Ar(Het)
N
N
N
vii
viii
N
ix
N
N
N
S
N
N
S
N
5
16
3 - 8
Scheme 1. Synthesis10 of thiazolopyrazine derivatives: (i) DPPA, TEA, t-
BuOH, 90°C, 12h, 40%; (ii) HCL/dioxane, MeOH, rt, 2h, 80%; (iii) Ac2O,
Et3N, DCM, rt, 3h, 70%; (iv) Br2, CHCl3, rt, 12h, 66%; (v) Raney Ni, DEA,
EtOH, rt, 12h, 78%; (vi) 4-methyl-1H-imidazole, (1R,2R)-N1,N2-
dimethylcyclohexane-1,2-diamine, CuI, Cs2CO3, DMF, 90°C, 12h, 20%; (vii)
P2O5/POCl3, sealed tube, 160°C, 12h, 33%; (viii) NBS, CH3CN, 20°C, 1h,
15%; (ix) (Het)ArB(OH)2, Pd(Ph3P)4, NaHCO3, PhMe/EtOH: 4/1, 100°C or
PdCl2(dppf), Cs2CO3, dioxane/H2O: 4/1, 100°C.
Figure 2. Thiazolopyrazine derivatives 1-9
Replacement of the pyridine moiety by a thiazole was
accompanied by both a diminution of LLE13 due to the reduced
in-vitro potency (IC50, 5-fold less compared to WEB-3; LLE 3.1)
and higher lipophilicity (clogP 3.7 to 4.0 for 1); the vector of the
substituents off the thiazole is less optimal than the vector off the
parent pyridine in WEB-3 (not shown). Me or H in place of the
OMe substituent (on position 2) retained the same level of in
vitro potency but allowed improved LLE (due to reduced clogP:
3.6 and 3.1 for 2 and 3 resp.; LLE 3.5 and 3.9 resp.), reduced
TPSA (both 43 vs 52 for WEB-3) and higher permeability14 (Papp
68-89 10E-6 cm/s, Table 1). In particular 3 depicted single digit
microsomal stability (hmCl, Table 1) and a remarkably high CNS
MPO score of 6.0.15 Unexpectedly reversing the thiazole ring (9)
or removing the substituent on position 2 (4) led to decreased
PDE10A biological activity (data not shown).
In our first approach the pyridine moiety of WEB-3 was replaced
by a bioisosteric thiazole (1, Fig. 2). This single change was
expected to impact the binding mode of the new tricyclic
molecule (compared to WEB-3). Even in the most conservative
scenario (namely, that the imidazolopyrazine portion of the core
remains in a similar position, see Fig. 3), the OMe group of 1
(shown in salmon) would be redirected more towards residues
Met703 and Tyr683 (shown in purple); this outcome would be
slightly less favorable compared to what is observed for WEB-3
(cyan).12
F729
Table 1. In vitro potency,a metabolic stability,b and permeability14 of
thiazolopyrazine derivatives.
Microsomal stability
mCl, µL/min/mg
MDCK-MDR1
Compds
(Fig. 2)
hPDE10A
IC50, nM
Papp
ER
Q726
(10E-6 cm/s)
human
rat
Y524
WEB-3
17
83
87
115
125
639
43
26
35
33
57
89
68
66
73
50
56
0.6
0.7
1.1
0.7
0.5
0.7
1
1
2
3
5
6
7
8
I692
16
4
6
45
16
3
Y693
M713
F696
7
2
12
8
20
2
3
2
ahPDE10A IC50 were measured in in vitro enzymatic reaction using
recombinant proteins and are means of at least two experiments (< 1µM).10
Test - retest are within max. 2-fold.
Figure 3. Comparison of 1 (model, salmon) docked into the PDE10A protein
(green) and WEB-3 (X-ray PDB 3LXG, cyan) bound to the PDE10A binding
site.12 Notice the shift in binding and specific location of OMe tail.
bMicrosomal intrinsic clearance values were determined according to
literature procedures16 and are means of at least two experiments.
Analogs 3-8 (Fig. 2, cluster 2) were prepared10 starting from the
commercially available thiazole-5-carboxylic acid 10. The
thiazole carbamate 11 obtained after Curtius rearrangement was
deprotected, acetylated and finally brominated (14) after
degradation of the dibrominated intermediate (not shown). The
imidazole moiety was introduced by copper catalyzed coupling to
afford 15, which was then cyclized to form the tricyclic central
core. Radical bromination yielded the halogenated precursor for
the Suzuki coupling.
In order to break the planarity and thus to avoid the potential risk
of mutagenicity,9 analogs 5-8 which carry a larger residue in
position 8 were prepared (Table 1). Very sensitive SAR were
observed when varying the C8-substituent: IC50 vary from 43 nM
to > 600nM for regioisomeric Me-pyridines while retaining the
favorable DMPK profile of 3 (moderate microsomal stability,
high Papp and low efflux ratio value supporting passive transport,
Table 1). The replacement of n-Pr with large aromatic rings
forces a shift in binding mode compared to WEB-3 (Fig. 5). The
added aromatic rings can no longer be properly accommodated
under residue Leu685. Instead, each of the new analogs is forced
to find
a novel binding spot, with various degrees of

Figure 5. A) Model of 8 (white) docked into a model of the PDE10A protein
(cyan). B) Comparison of 8 (white) and 29 (orange) manually docked into the
PDE10A protein (green).12
complementarity with their respective microenvironments
(dependent upon the proximity/distance from/to residue
Leu685).13 For example, Me at the ortho position (7) is still well
tolerated because it is still away from Leu685 (Fig. 4). However,
moving the Me substituent to the meta position (5), or even
further out (to the para position, 6), led to decreased PDE10A
biological activity.
8 constituted the front runner of this series; it combined low
clogP (2.2), high TPSA (86), high LLE (5.5), single digit
microsomal stability (human and rat, Table 1) and high
permeability (Papp 56 10E-6 cm/s). However the PDE2A
selectivity was sub-optimal (Table 2).
Table 2. PDE selectivity profilea. (thiazolopyrazine series)
Compds
(Fig. 2)
hPDE2A hPDE3A
IC50, nM IC50, nM
hPDE10A
IC50, nM
hPDE4D2
IC50, nM
L685
F729
17
83
87
115
125
43
2160
4890
3010
8500
16500
7070
7840
7850
7470
4940
5540
7850
1780
4220
4270
13500
2980
WEB-3
1
2
3
5
7
8
Q726
1860
1310
20
I692
M713
ahPDE10A, 2A, 3A & 4D28a IC50 were measured in in vitro enzymatic
reaction using recombinant proteins and are means of at least two
experiments (< 1µM).10 Test - retest are within max. 2-fold.
Y693
Y524
F696
As no path was clear how to diminish PDE2A affinity while
retaining or improving PDE10A potency,17 we then turned our
attention to a related chemical series which enclosed a cinnoline
central core (Fig. 6).
Figure 4. Comparison of 7 (purple) docked into the PDE10A protein (green)
and WEB-3 (PDB: 3LXG, cyan) bound to the PDE10A binding site.12 Notice
the shift in binding as pyridine can no longer fit right beneath Leu685 (shown
in magenta).
The 2-Me-pyridin-3-yl (in 7) was thus preferred because it
afforded an in vitro hPDE10A potency below 50nM, and an
improved profile compared to WEB-3 (improved LLE (4.7) and
in vitro DMPK properties (mCl & Papp), Table 1). A benzamide
moiety was also tolerated (in 8), as it was predicted to make a H-
bond interaction with the side chain of proximal residue Ser581
(Fig. 5A).12
5A
F729
Q726
S581
F696
Figure 6. Cinnoline derivatives 17-31
This second approach was suggested by the structural analogy of
a HTS hit (17, cinnolines, Fig. 1) which combined good
PDE10A in vitro potency (84nM, Table 3) and high selectivity
towards PDE isoforms PDE4D2 & PDE2A (Table 4). N-
methylation (position 3, 18) impacted the CNS MPO score very
favorably (from 3.8 to 5.0).
5B
F729
H525
Q726
The HTS hit 17 was expected to be a good WEB-3 mimetic given
the similarity in the overall 6,6,5 core structure with nitrogen
acceptors at positions 2 and 6. The observed reduction in potency
was attributed to the lack of elaboration of the core. However,
upon crystal structure analysis, it was revealed that 17 has a
different binding mode which is rotated 90° with respect to
WEB-3 (Fig. 7). 17 binds in the same plane as WEB-3, making 2
H-bonds to Gln726 via the oxygens of the two methoxy groups,
F696

O
O
O
N
N
one water mediated H-bond to Tyr524, and π-stacking
interactions to Phe729 (above the plane).
R
R
R
N
ii
i
N
N
R´
R´
R´
32
34
33
OTf
(Het)Ar
O
N
N
N
R
R
N
N
R
N
v
iv
iii
N
N
N
N
N
N
R´
R´
R´
19-31
36
35
Scheme 2. Synthesis11 of cinnoline derivatives: (i) diethyl carbonate, Na,
130°C; (ii) methylhydrazine, EtOH, 90°C; (iii) NaNO2, HCl, 0°C to rt; (iv)
Tf2O, TEA, THF, -78°C to rt; (v) (Het)ArB(OH)2, Pd(Ph3P)4, NaHCO3,
PhMe/EtOH: 4/1, 100°C or PdCl2(dppf), Cs2CO3, dioxane/H2O: 4/1, 100°C.
Starting from 18, bulky residues (s. the discussion above for the
thiazolopyrazine series) were introduced in position 1 in order to
break the planarity and thus avoid potential mutagenicity.9 We
first focused our efforts on the optimization of the substitution
pattern of the central core. A series of 2-Cl-phenyl derivatives
carried by various 6,8-disubstituted cinnoline tricycles (19-22)
was prepared;18 we observed up to a 80-fold difference in
PDE10A potency depending on the substituents: 19 vs 22.
Important learnings were gathered by one to one comparison of
these four close analogs: a fluorine on position 6 is detrimental
for in vitro PDE10A potency (19 far less potent than its 6-OMe
analog 21) whereas both F and OMe are well tolerated on
position 8 (compare 21 to 22), possibly reflecting a combination
of electronic and steric effects (i.e. electron-donating better than
electron-withdrawing substituents, and superior complementarity
of 6-OMe over F in a mixed lipophilic environment). Moreover,
the OMe is nicely filling the pocket (VdW contacts and general
hydrophobic interaction, not shown). Both 21 and 22 showed
moderate PDE10A in vitro potency, high microsomal stability
(human, hmCl <5 µL/min/mg) but high instability towards rat
microsomes (rmCl > 200 µL/min/mg, Table 3).
Figure 7. Superposition of WEB-3 (PDB: 3LXG, fushia) with 17 (green,
PDB: 6MSA) bound to hPDE10a showing the 90° rotation of the core in the
plane with respect to WEB-3. Key interacting residues are shown except for
Phe729 which contributes π-stacking interactions from above the plane.
It can be seen from the overlay of the crystal structure of 23 and
WEB-3 (Fig. 8), that elaboration of 17 at the 1 and 3 positions
sucessfully rotates the core to be more closely aligned with
WEB-3. However, the addition of the bulky aromatic ring at
position 1 and the methoxy group at position 6 results in an
overall shift of the molecule; the methyl pyrimidine is then
positioned in a pocket just past Leu675: this enables the ring to
adopt an orientation nearly perpendicular to the core while the
methyl substituent fits into a small hydrophobic pocket between
Phe696 and Tyr524. The oxygen of the 6-OMe makes the H-
bond to Gln726 and the nitrogen at position 2 makes a water
mediated H-bond to Tyr524.
Table 3. In vitro potency,a metabolic stability,b and permeability14 of
cinnoline derivatives. (17-22)
Microsomal stability
mCl, µL/min/mg
MDCK-MDR1
Compds
(Fig. 6)
hPDE10A
IC50, nM
Papp
ER
(10E-6 cm/s)
human
26
rat
WEB-3
17
18
19
20
21
17
84
171
1790
298
25
35
33
63
60
36
46
54
44
0.6
1
0.8
0.4
0.7
0.5
0.8
28
387
4
5
411
194
22
23
a,bcf. table 1.
In terms of selectivity towards other PDE isoforms high
selectivity was achieved (>800-fold, Table 4) versus PDE3A and
4D2; unexpectedly, whereas the PDE2A selectivity was sub-
optimal for 21 it was high for 22 (Table 4).
Figure 8. Superposition of WEB-3 (PDB: 3LXG, fushia) with 23 bound to
hPDE10a (teal, PDB: 6MSC). Note that the core of 23 is shifted for the OMe
to make the hydrogen bond to Q726; and in order to accommodate the
methylpyrimidine in the pocket beneath L675, the methyl fits into the
hydrophobic pocket formed by F696, I692, Y524 and H525.
Table 4. PDE selectivity profilea. (cinnoline series)
Analogs 19-31 (Fig. 6, cluster 5) were prepared11 starting from
the corresponding commercially available phenylacetonitrile 32.
A pyrazole moiety was first built by intramolecular cyclization
with methylhydrazine of the nitrile 34 and an acetate
functionality introduced after benzylic metalation. Formation of
the tricyclic cinnoline core could then be completed in presence
of sodium nitrite (compound 35). Finally the hydroxy group was
activated to the triflate (36) which allowed Suzuki type coupling
reactions to afford 19-31.
Compds
(Fig. 6)
hPDE2A hPDE3A
IC50, nM IC50, nM
hPDE10A
IC50, nM
hPDE4D2
IC50, nM
17
84
171
25
23
21
83
5
2160
>100000
8500
1110
5540
>20000
>20000
>20000
>20000
9540
WEB-3
17
18
21
22
23
24
26
>20000
>20000
>20000
8520
111900
23600
923
12900
3520
5390
1320
14200
8920
27
105
6130

28
11
4270
profile 23 and 26 were advanced to rat PK (Table 6). Their
improved microsomal stability (compared to WEB-3 data)
translated to a much lower clearance in vivo. Their decreased
clogP also translated to a decreased volume of distribution (Vss).
26 exhibited a shorter half-life but an improved bioavailability
versus both 23 and WEB-3. While WEB-3 showed a higher
brain/plasma ratio, both 23 and 26 showed around 10-fold
increased dose-normalized free brain concentration at 1 hour post
intraperitoneal administration; 23 and 26 have thus the potential
to inhibit the PDE10A target in vivo at a lower efficacious dose
acf. table 2.
Analogs of 21 and 22 were prepared with a focus on lowering the
lipophilicity; for both 21 and 22, clogP was unfavorable (>4)
leading to low LLE (3.4 and 3.1 resp.). Published SAR in related
structural series9 allowed us to expedite optimization by limiting
our efforts to a short list of Me-pyridine analogs (23, 24, and
26).19 23 & 26 combined high CNS MPO score (>5.5), high LLE
(5.2 and 5.5, resp.), moderate microsomal stability (h and r, table
5) and high permeability (Table 5). According to modeling,
electron-withdrawing substituents on the phenyl ring of the
tricyclic core (25) are less well tolerated than electron-donating
groups in that ring (23 and 24), diminishing the pi-stacking and
pi-facing strength of the core (not shown).
than
WEB-3.
Table 6. Pharmacokinetic profile in rat of WEB-3, 23 & 26
2321
2621
2.8 / 75 / 5.0
5.7
WEB-3
clogP / TPSA / pKa
CNS MPO
LLE / LigElit
3.7 / 52 / 6.7
5.2
2.5 / 66 / 4.2
5.5
Table 5. In vitro potency,a metabolic stability,b and permeability14 of
cinnoline derivatives (23-31).
4.1 / 0.53
0.003 / 0.007
1.7
5.2 / 0.44
0.097 / 0.094
1.7
5.5 / 0.45
0.032 / 0.068
0.8
fu pl / fu br
t1/2 (h) a
VSS (L/kg) a
ClP (L/h/kg) a
Foral (%) b
Microsomal stability
mCl, µL/min/mg
MDCK-MDR1
ER
(10E-6 cm/s)
Compds
(Fig. 6)
hPDE10A
IC50, nM
Papp
human
rat
2.4
1.0
0.4
23
24
25
26
27
28
29
30
21
83
1590
5
105
11
555
535
1800
6
4
1
2
8
7
4
6
53
68
29
55
0.8
0.7
0.6
1
2.6
1.2
1.0
19
12
42
c
C (ng/ml, tot.) plasma / brain
72 / 118
0.10 / 0.40
1.67 / 3.81
145 / 93
7 / 5
741 / 90
24 / 6
d
C (ng/ml, free) plasma / brain
Brain : Plasma Ratio c
0.64 / 0.62
0.12 / 0.26
2
5
4
27
164
8
57
38
62
1.1
9.6
0.8
(tot/free)
aAfter intravenous administratione
bAfter oral administratione
cAfter intraperitoneal administratione, at 1 hour post-dosing
dDose normalized
31
a,bcf. table 1.
eiv, po, ip dose: 2 mg·kg-1 WEB-3 & 23; 1 mg·kg-1 26.
A few more analogs were prepared (Fig. 6, Table 5):
In summary, novel phosphodiesterase 10A inhibitors were
disclosed. In particular, optimization of a HTS hit (17) resulted in
potent, selective and brain penetrant 23 and 26; both exhibited
much lower clearance in vivo and decreased volume of
distribution (rat PK) and have thus the potential to inhibit the
PDE10A target in vivo at a lower efficacious dose than WEB-3.
-
examplarily a non-pyridine derivative was synthesized (28
carries a diMe-thiazole in position 1): compared to 22, LLE
(4.6) and CNS MPO score (4.8) were improved; rmCl
though lowered remained sub-optimal and higher compared
to the one measured for 26 (Table 5); in this series, clogP
was a good predictor of rmCl (clogP 22 4.5, 23 3.3, 26 2.8);
unlike what was observed in the thiazolopyrazine series (8),
a benzamide in position 1 was detrimental for PDE10A in
vitro potency (29, IC50 555nM, Table 5); indeed, to avoid an
intramolecular steric clash between the OMe and the
benzamide groups of 29, the benzamide moiety moves out,
but becomes suboptimally close to residue His525 (Fig. 5B).
The penalty results in much lower affinity for 29 compared
to 8;
finally, we confirmed (with 25) the negative impact on the
affinity for PDE10A of the diF-substitution in presence of a
pyridyl moiety in position 1; as discussed for 19 (above),
replacing electron-donating moieties (OMe) by electron-
withdrawing groups (F) disfavors pi-pi-interactions;
our attempt to break the planarity with a cycloalkyl (30,
cyclopropyl) or branched alkyl (31, isobutyl) residue was
not further pursued: despite increased fsp3, lower MW and
retained high CNS MPO score (4.7 – 5.0), this approach was
-
Declaration of Interest
The authors declare the following competing financial interest(s):
The authors are current or former employees of AbbVie, and may
own company stocks; the design, study conduct, and financial
support for this research were provided by AbbVie. AbbVie
participated in the interpretation of data, review, and approval of
the publication. $K.D. is a former AbbVie employee.
-
-
Acknowledgments
We would like to thank Ferdinand Regner and Sonja Triebel and
the Analytical Chemistry group in Ludwigshafen. Use of the
IMCA-CAT beamline 17-ID at the Advanced Photon Source was
supported by the companies of the Industrial Macromolecular
Crystallography Association through a contract with Hauptman-
Woodward Medical Research Institute. Use of the Advanced
Photon Source was supported by the U.S. Department of Energy,
Office of Science, Office of Basic Energy Sciences, under
Contract No. DE-AC02-06CH11357.
detrimental in term of PDE10A in vitro potency (IC50
>
500nM, Table 5). The diminished potency here might be
reflective of the limited room available; in particular
branched/substituted
aliphatic
groups
might
not
accommodate this crowded region of the binding pocket.
References and notes
When comparing to other tricyclic PDE10A inhibitors [such as
imidazo[1,5-a]quinoxalines9a
and
benzo[e]imidazo[5,1-
1. Siuciak, J.A.; McCarthy, S.A.; Chapin, D.S.; Martin, A. N.
Psychopharmacol. 2008, 197, 115.
2. Rodefer, J.S; Murphy, E.R.; Baxter, M.G. Eur. J. Neurosci. 2005,
4, 1070.
c][1,2,4]triazines9b], the 3H-pyrazolo[3,4-c]cinnolines (23 & 26)
presented here have lowered clogP (2.5 and 2.8, resp.).20
Furthermore based on their balanced and favorable in vitro
3. Kehler, J.; Kilburn, J.P. Exp. Opin Ther. Pat. 2009, 19, 1715.

4. U.S. Patent 2005/0202550 A1, 2005. Chem. Abstr. 2005, 143,
301347.
21. analytical data 23 ESI MS (HRMS, m/z): calcd for C17H15FN5O
324.13, found 324.13 [M + H+ ]; 1H-NMR (400 MHz, d6-DMSO)
δ (ppm) = 8.76 (s, 1H), 7.36 (d, J= 4.8 Hz, 1H), 7.55 (d, J=4.8
Hz, 1H), 7.32 (dd, J=l 1.6, 2.4 Hz, 1H), 6.75 (dd, J=l 1.6, 2.4 Hz,
1H), 4.50 (s, 3 H), 4.15 (s, 3 H), 2.20 (s, 3 H); 13C-NMR (126
MHz, DMSO) δ 166.48, 164.47, 160.83, 160.72, 150.26, 146.69,
142.96, 139.35, 135.80, 135.63, 127.19, 122.51, 122.40, 107.38,
99.61, 99.38, 97.86, 97.66, 57.55, 35.91, 17.10; 26 ESI MS
(HRMS, m/z): calcd for C18H18N5O2 336.15, found 336.15 [M +
H+ ]; 1H-NMR (400 MHz, d6-DMSO) δ (ppm) = 8.68-8.66 (m,
1H), 7.94-7.92 (m, 1H), 7.48-7.45 (m, 1H),6.89 (d, J= 2.0 Hz 1
H), 6.39 (d, J= 2.0 Hz 1 H), 4.45 (s, 3 H), 4.08 (s, 3 H), 3.70 (s, 3
H), 2.39 (s, 3 H); kinome profiling (23 & 26): IC50 > 10µM for
>80 kinases, ATP compet. binding; mini-Ames Test8b (23 & 26):
negative; hERG (QPatch, HEK293, human): 23 (IC50 17.3 µM),
26 (IC50 >30 µM).
5. 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.
6. Hoefgen, N.; Stange, H.; Schindler, R.; Lankau, H.-J.; Grunwald,
C.; Langen, B.; Egerland, U.; Tremmel, P.; Pangalos, M.N.;
Marquis, K.L.; et al. J. Med. Chem. 2010, 53, 4399.
7. Advanced candidates include (a) MP-10 (PF-2545920, Pfizer).5
(b) OMS-824 (Omeros), Cutshall, N.S.; Onrust, R.; Rohde, A.;
Gragerov, S.; Hamilton, L.; Harbol, K.; Shen, H.R.; McKee, S.;
Zuta, C.; Gragerova, G.; Florio, V.; Wheeler, T.N.; Gage, J.L.
Bioorg. Med. Chem. Lett. 2012, 22, 5595. (c) AMG-579 (Amgen),
Hu, E.; Chen, N.; Bourbeau, M.P.; Harrington, P.E.; Biswas, K.;
Kunz, R.K.; Andrews, K.L.; Chmait, S.; Zhao, X.; Davis, C.; Ma,
J.; Shi, J.; Lester-Zeiner, D.; Danao, J.; Able, J.; Cueva, M.;
Talreja, S.; Kornecook, T.; Chen, H.; Porter, A.; Hungate, R.;
Treanor, J.; Allen, J.R. J. Med. Chem. 2014, 57, 6632. (d) PBF-
999 (Palobiofarma), ClinicalTrials.gov Identifier: NCT02907294.
(e) TAK-063 (Takeda), Kimura, H.; Taniguchi, T.; Med. Chem.
Rev. 2017, 52, 3. (f) RG-7203 (Roche) and FRM-6308 (Forum )
Thomson Reuters Cortellis Web site,
https://cortellis.thomsonreuterslifesciences.com.
8. (a) In particular selectivity towards PDE2A, 3A & 4D2 is critical;
for that reason these 3 PDEs were selected as our entry selectivity
panel. PDE2 is known to play a role in schizophrenia, PDE3
inhibition might induce CV side effects and we had PDE4 cross
activity with earlier structural series; (b) The in vitro Salmonella
Bacterial Reverse Mutation Assay or Mini-Ames Test is a widely
used short-term genotoxicity assay for the detection of gene
mutations. The plate incorporation assay uses histidine-dependent
Salmonella tester strains TA98 and TA100 and is carried out in
the absence and presence of an exogenous metabolic activation
system (S9). A test article that produces a response with the
highest increase equal to or exceeding twice the vehicle control
value and with a concentration-related increase is considered
mutagenic.
9. (a) Malamas, M.S.; Ni, Y.-K.; Erdei, J.; Stange, H.; Schindler, R.;
Lankau, H.-J.; Grunwald, C.; Fan, K.Y.; Parris, K.; Langen, B.; et
al. J. Med. Chem. 2011, 54, 7621. (b) Malamas, M.S.; Stange, H.;
Schindler, R.; Lankau, H.-J.; Grunwald, C.; Langen, B.; Egerland,
U.; Hage, T.; Ni, Y.; Erdei, J.; et al. Bioorg. Med. Chem. Letters
2012, 22, 5876.
10. The preparation of the building blocks mentioned, typical
procedures, and assay setup are described in WO 2014041175,
2014. Chem. Abstr. 2014, 160, 487023, and references cited
therein.
11. The preparation of the building blocks mentioned, typical
procedures, and assay setup are described in WO 2014027078,
2014. Chem. Abstr. 2014, 160, 341156, and references cited
therein.
12. (a) Models were generated using Schrodinger software, Release
2018-1, and figures were prepared using the PyMOL Molecular
Graphics System, Version 2.0 Schrödinger, LLC. (b) Li, J.; Chen,
Y.-L.; Deng, Y.-L.; Zhou, Q.; Wu, Y.; Wu, D.; Luo, H.-B. Front.
Chem. 2018, 6, 167.
13. LLE (or LipE) = −log Ki − clogP. (a) Leeson, P. D.;
Springthorpe, B. Nat. Rev. Drug Discov. 2007, 6, 881. (b)
Hopkins, A. L.; Groome, C. R.; Alex, A. Drug Discov. Today
2004, 9, 430. LigE (or LE) Bembenek, S.D.; Tounge, B.A.;
Reynolds, C.H. Drug Discov. Today 2009, 14, 278.
14. MDCK-MDR1 human P-gp transfected cell line - Papp = apparent
permeability, compound flux across cell monolayer in Transwell
system, either in direction apical-to-basolateral (AB) or vice versa
(BA); ER = Efflux Ratio = Ratio of Papp values BA/AB; Passive
perm = Papp AB in the presence of a pan-transporter inhibitor.
15. Wager, T.T.; Hou, X.; Verhoest, P.R.; Villalobos, A. ACS Chem.
Neurosci. 2010, 1, 435.
16. Obach, S. Drug Metab. Dispos. 1999, 27, 1350.
17. PDE2 and PDE10 are quite conserved overall, with high residue
conservation around the binding site: Omori, K.; Kotera, J. Circul.
Res. 2007, 100, 309.
18. This effort could be expedited by published SAR in related
structural series.9b
19. Synthesis of the 2-Me-4-pyridine analog of 22 failed.
20. s. front runners 969a & 1069a (clogP ≥ 4.3), 539b (clogP 3.3); logP
are calculated using Biobyte calculator.

To create your abstract, type over the instructions in the
template box below.
Graphical Abstract
Fonts or abstract dimensions should not be changed or altered.
Novel, potent, selective, and brain penetrant
phosphodiesterase 10A inhibitors
Leave this area blank for abstract info.
Hervé Geneste*, Karla Drescher, Clarissa Jakob, Loïc Laplanche, Michael Ochse, Maricel Torrent

Ms. No.: BMCL-D-18-01566
Title: Novel, potent, selective, and brain penetrant
phosphodiesterase 10A inhibitors
Corresponding Author: Dr. Hervé Geneste
Authors: Karla Drescher, Clarissa Jakob, Loïc Laplanche,
Michael Ochse, Maricel Torrent
Highlights
-
-
novel PDE10 inhibitors
optimization supported by x-ray crystal structure
analysis and molecular modeling
potent, selective, and brain penetrant
improved PK profile (rat): lower clearance in vivo
and decreased volume of distribution (compared
to WEB-3)
-
-
22.