Small-molecule phosphodiesterase probes: Discovery of potent and selective CNS-penetrable quinazoline inhibitors of PDE1

Article abstract of DOI:10.1039/c4md00113c

PDE1 is a family of calcium-activated, dual substrate phosphodiesterases expressed in both the CNS and periphery that play a role in the integration of intracellular calcium and cyclic nucleotide signaling cascades. Exploration of the potential in targeting this family of enzymes to treat neuropsychiatric disorders has been hampered by a lack of potent, selective, and brain penetrable PDE1 inhibitors. To identify such compounds we used high-throughput screening, structure-based design, and targeted synthetic chemistry to discover the 4-aminoquinazoline 7a (PF-04471141) and the 4-indanylquinazoline 27 (PF-04822163) each of which are PDE1 inhibitors that readily cross the blood brain barrier. These quinazoline-based PDE1-selective inhibitors represent valuable new tools to study the biological processes regulated by PDE1 and to begin to determine the potential therapeutic utility of such compounds to treat neuropsychiatric disorders.

Full text of DOI:10.1039/c4md00113c

Volume 5 Number 9 September 2014 Pages 1257–1424
MedChemComm
Broadening the field of opportunity for medicinal chemists
www.rsc.org/medchemcomm
ISSN 2040-2503
CONCISE ARTICLE
John M. Humphrey, Frank S. Menniti et al.
Small-molecule phosphodiesterase probes: discovery of potent and
selective CNS-penetrable quinazoline inhibitors of PDE1

MedChemComm
View Article Online
View Journal | View Issue
CONCISE ARTICLE
Small-molecule phosphodiesterase probes:
discovery of potent and selective CNS-penetrable
quinazoline inhibitors of PDE1†
Cite this: Med. Chem. Commun., 2014,
5, 1290
a
John M. Humphrey, Eddie Yang,a Christopher W. am Ende,a Eric P. Arnold,a
*
Jenna L. Head,a Stephen Jenkinson,b Lorraine A. Lebel,a Spiros Liras,c
Jayvardhan Pandit,a Brian Samas,a Felix Vajdos,a Samuel P. Simons,a
a
a
d
*
Artem Evdokimov, Mahmoud Mansour and Frank S. Menniti
PDE1 is a family of calcium-activated, dual substrate phosphodiesterases expressed in both the CNS and
periphery that play a role in the integration of intracellular calcium and cyclic nucleotide signaling
cascades. Exploration of the potential in targeting this family of enzymes to treat neuropsychiatric
disorders has been hampered by a lack of potent, selective, and brain penetrable PDE1 inhibitors. To
identify such compounds we used high-throughput screening, structure-based design, and targeted
synthetic chemistry to discover the 4-aminoquinazoline 7a (PF-04471141) and the 4-indanylquinazoline
27 (PF-04822163) each of which are PDE1 inhibitors that readily cross the blood brain barrier. These
quinazoline-based PDE1-selective inhibitors represent valuable new tools to study the biological
processes regulated by PDE1 and to begin to determine the potential therapeutic utility of such
compounds to treat neuropsychiatric disorders.
Received 11th March 2014
Accepted 15th May 2014
DOI: 10.1039/c4md00113c
www.rsc.org/medchemcomm
inhibition by various cell components.3 Signal transduction
cascades mediated by the cyclic nucleotides and regulated by
Introduction
the PDEs are diverse and include neurotransmission, metabo-
lism, smooth muscle contraction, odorant, taste and visual
responses, gene activation, insulin response, locomotion, cell
proliferation and cell death among others.4 Thus, pharmaco-
logical manipulation of these signaling cascades through PDE
inhibition is of considerable therapeutic interest.4,5
The cyclic nucleotide phosphodiesterases (PDEs) are a group of
enzymes that regulate intracellular signaling through metabolic
inactivation of the second messengers cyclic guanosine mono-
phosphate (cGMP) and cyclic adenosine monophosphate
(cAMP). The PDEs are encoded by 21 genes that are grouped
into eleven different subfamilies. The division is based on
primary sequence homology that gives rise to family specic
substrate specicity, composition of the N-terminal regulatory
domain, and sensitivity to pharmacological inhibitors.1,2 PDE
isoform activity is itself regulated through differential tissue
expression, cell compartmentalization and activation or
The present work focuses on PDE1, which has been exten-
sively studied but for which there is still a deciency of potent
and selective pharmacological tools.4,6 PDE1 was rst isolated in
1970 from bovine7 and rat brain,8 with the dening character-
istic of stimulation of enzyme activity by calcium/calmodulin.6
PDE1 isoenzymes are encoded by three genes, PDE1A, PDE1B,
and PDE1C. The N-terminal regions contain the regulatory
calcium/calmodulin binding sites that activate the enzyme to
increase Vmax. Despite high structural homology within the
C-terminal catalytic regions, the three PDE1 isoforms differ in
substrate affinity which suggests a level of functional differen-
tiation. Whereas each isoform hydrolyzes cGMP with compa-
rable affinity, their affinities for cAMP differ by as much as two
orders of magnitude with the rank order PDE1C > PDE1B >
PDE1A.6 Further indication that the isoforms mediate distinct
functions is indicated by the differential distribution
throughout the body. For example, while each appears to be
present throughout the brain, the expression level of PDE1B in
the striatum is much higher than that of PDE1A or PDE1C.
aNeuroscience Medicinal Chemistry, Pzer World Wide Research and Development,
Eastern Point Road, Groton, CT 06340, USA. E-mail: john.m.humphrey@pzer.com
bSafety Pharmacology CoE, Pzer World Wide Research and Development, 10646
Science Center Drive, San Diego, CA 92121, USA
cPzer World Wide Research and Development, 620 Memorial Drive, Cambridge, MA
02139, USA
dPharmaceuticals, Inc. One Davol Square, Suite 105, Providence, RI 02903, USA.
E-mail: fmenniti@mnemosynepharma.com
† Electronic supplementary information (ESI) available: Broad panel
ligand–receptor binding data, PDE1 assay data with standard error, select
pharmacokinetic data for 7a and 27, X-ray crystal structure experimental
information for 7a and 27, PDE1B1 enzyme purity data and binding assay
protocol, and complete synthesis experimental details. The crystal structure of
inhibitor 7a bound to the PDE10A catalytic subunit has been deposited to the
RCSB protein data bank with the entry code 4NPV. See DOI: 10.1039/c4md00113c
1290 | Med. Chem. Commun., 2014, 5, 1290–1296
This journal is © The Royal Society of Chemistry 2014

View Article Online
Concise Article
MedChemComm
Conversely, PDE1C is to be particularly highly expressed in the
olfactory epithelium, heart and bladder.9,10
Given the regulation of enzymatic activity by calcium, PDE1
appears to be an integration point for intracellular calcium and
cyclic nucleotide signaling cascades. In addition to differential
tissue distribution, PDE1 isoforms assume distinct roles in CNS
function. PDE1B knock-out mice have a phenotype consistent
with high striatal expression, including increased locomotor
activity and increased phosphorylation of striatal PKA
substrates in response to dopamine agonists.11 This data
suggests the possible use of PDE1B inhibitors in the treatment
of attention decit hyperactivity disorder or Parkinson's
disease. Other evidence associating PDE1A gene polymorphism
to remission with antidepressant treatment suggests a thera-
peutic utility in the treatment of depression.12 Recent reports
suggest potential for selective PDE1 inhibitors as cognition
enhancers13 and in the treatment of Huntington's disease.14
To further investigate the role of PDE1 in the regulation of
brain function and to begin to explore therapeutic utility, we
undertook an effort to discover potent, selective, and brain
penetrable inhibitors of PDE1 that would enable isolation of
biological effects mediated by PDE1 from those linked to other
PDE subfamilies. Our approach, which included high
throughput screening, structure based design, and parallel and
directed synthesis, yielded a series of 4-aminoquinazoline and
4-benzylquinazoline PDE1 inhibitors that readily cross the
blood brain barrier. The highlighted compounds from this
effort, aminoquinazoline 7a (PF-04471141) and indanylquina-
zoline 27 (PF-04822163), are among the most potent and
selective small molecule inhibitors of PDE1 reported to date .15‡
Fig. 1 Quinazoline hits 1 and 2 from a high throughput screen.
Fig. 2 PDE10A inhibitor 3 and key interaction with invariant glutamine
residue.
Results
For the purposes of chemical library screening and SAR devel-
opment we used recombinant full-length human PDE1B as a
representative of the PDE1 family. This effort began with the
identication of the two resynthesized quinazoline le screen
hits 1 and 2 (PDE1B IC50 219 nM and 130 nM, respectively,
Fig. 1). These compounds immediately attracted our attention
because of their structural resemblance to known PDE10A
inhibitors16 and physical properties predictive of good central
exposure.17,18 In addition, our prior work on PDE10A inhibitors
afforded intimate knowledge of the binding mode of the 6,
7-dimethoxyquinazolines through a crystal structure of PDE10A
with bound inhibitor 3.16 Key interactions included a bidentate
hydrogen bonding interaction between a PDE-invariant gluta-
mine residue and the 6,7-dimethoxy groups (Fig. 2). Based on
the similar structures of PDE10A and PDE1B, and on the pres-
ence of a 6,7-dimethoxy array in the screen hits 1 and 2, we
supposed that a similar binding mode was in effect for PDE1.
Beginning with the HTS hits, we set about SAR exploration in
two ways. We rst assessed quinazoline substituent effects
utilizing library synthesis to expand the di-n-propyl amine series
(1).19,20 The templates utilized for these library efforts include
Fig. 3 Quinazolines employed in parallel synthesis, and quinazoline
synthesis from anthranilic acid.
the ten 4-chloroquinazoline templates 4a–j (Fig. 3), which were
obtained from our in-house sample bank, commercial
suppliers, and chemical synthesis. The synthesized 4-chlor-
oquinazolines were generally prepared from commercial
anthranilic acid derivatives through the Niementowski quina-
zoline synthesis according to established methods21,22 with the
exception of 4b which was prepared via the route shown in
Fig. 8. The templates were then coupled with a broad set of
amines in a parallel format (Fig. 3).
The amines for the libraries were sampled from cyclic and
acyclic aliphatic amines, benzylic amines, and amines bearing
‡ Compound 7a, also known as PF-04471141 or its hydrochloride salt
PF-04471141-01, is now commercially available from Sigma Aldrich.
This journal is © The Royal Society of Chemistry 2014
Med. Chem. Commun., 2014, 5, 1290–1296 | 1291

View Article Online
MedChemComm
Concise Article
tethered aryl and heteroaryl groups. The data from this study
revealed that potent PDE1B inhibition with respect to the qui-
nazoline ring was favored with 7,8-dimethoxy functionalization,
and that a third methoxy group in the 6-position was
neither detrimental nor required for binding. Library-derived
compounds from each of the amine classes exhibited reason-
able potency (35–1000 nM) with the most potent analog being
that derived from 6,7,8-trimethoxyquinazoline and 3-amino-
pentane (7a, PDE1B IC50 ¼ 35 nM, Fig. 3). The nature of the
amino substituent in this series was judged to be relatively non-
critical for PDE1 inhibitory potency, in contrast to the case of
PDE10A which strongly prefers amine substituents with asso-
ciated aryl groups.16
In our second approach to SAR expansion we launched a
search for quinazoline replacements that would yield additional
active compounds. For example, good PDE10A potency is
available from quinazoline, phthalazine, and cinnoline
templates (compounds 8–10, Fig. 4a).23 However, in the PDE1B
series potent inhibition was only available from the quinazoline
template (11), with phthalazine, cinnoline, quinoline and iso-
quinoline templates each affording inactive compounds (12–15,
0% inhibition at 1 mM, Fig. 4b).
Fig. 5 X-ray structure the of PDE1B/7a complex (top) and schematic
representation (RCSB accession code 4NPV).
The X-ray crystal structure of 7a bound to the catalytic
subunit of PDE1B provided an explanation for the quinazoline
structural requirement. Relative to similar PDE10A inhibitors,
in PDE1B the quinazoline is rotated slightly (counter-clockwise
˚
vs. Fig. 2) such that the quinazoline 7- and 8-methoxy groups which is 2.8 A from a crystallographic water molecule. This
accept a bifurcated H-bond from the side chain nitrogen of the water molecule is in turn coordinated by hydrogen bonds to the
˚
PDE1 Gln-421 residue, rather than the 6- and 7-methoxy groups side chain phenol of Tyr-222 (2.7 A), to the carbonyl of Asp-370
˚
in PDE10A inhibitors (Fig. 5). This rotation positions the key (2.6 A), and to the catalytic Zn atom via a second well-ordered
quinazoline N1 atom to accept a hydrogen bond from the crystallographic water molecule. This network of water-medi-
imidazole side chain of His-373, which is absent in PDE10A. ated hydrogen bonds is likely important for quinazoline
The requirement for a quinazoline core in this series is further potency. This contention is reinforced by the observation that
explained by the hydrogen bonding pattern at quinazoline-N3, both the cinnoline and quinoline templates, which retain N1
but lack the hydrogen bond acceptor at position 3, show no
activity against PDE1B (Fig. 4b). The C4-amine donates a
hydrogen bond to a crystallographic water molecule but does
not appear to make productive interactions with the PDE1B
catalytic subunit. The 6-methoxy group likewise contributes no
productive protein–ligand interactions with the subunit and is
instead oriented towards solvent.
Synthesis of the 4-benzylquinazoline series was generally
accomplished by enolate addition of methyl aryl acetates (16) to
the 4-chloroquinazolines to give the esters 17. Subsequent
saponication/decarboxylation yielded the nal targets 18 as
shown in Fig. 6. This sequence could be conveniently accom-
plished in one vessel by quenching the enolate reaction mixture
with aqueous sodium hydroxide to induce saponication/
decarboxylation. For indanes such as 21b and 23b (Fig. 7), better
results were obtained through isolation of the ester adduct (i.e.
21a and 23a) prior to saponication. An oxygen-free environ-
ment was necessary during saponication of the indane deriv-
ative 23a to minimize quinazoline decomposition via the
alcohol 26, which occurs under strongly basic conditions in air.
This oxygen free environment was maintained by holding the
reaction at a steady reux during the saponication or by con-
ducting the saponication rapidly via microwave irradiation in
Fig. 4 Quinazoline vs. cinnoline, phthalazine, quinoline and iso-
quinoline structures.
1292 | Med. Chem. Commun., 2014, 5, 1290–1296
This journal is © The Royal Society of Chemistry 2014

View Article Online
Concise Article
MedChemComm
bromine or uorine yielded potency losses in the indane set
(23b vs. 24 and 25), possibly indicating the presence of a
hydrophobic pocket appropriately sized to accommodate chlo-
rine (see modeling discussion below).
The synthesis of the most potent compound of this
campaign, target 27 (PDE1B IC50 ¼ 2.4 nM), is shown in Fig. 8.
Oxidation of the nitrobenzaldeyde derivative 28 to the corre-
sponding carboxylic acid 29 and subsequent nitro group
reduction afforded the anthranilic acid derivative 30.
Conversion into quinazolone 31 was followed by POCl3
treatment to yield the requisite 4-chloroquinazoline 4b. The
chloroindane methyl ester 34 was prepared by direct chlori-
nation of the indane acid 32 (ref. 24–26) and selective
recrystallization of 33 (see ESI†) followed by esterication.
Coupling with the quinazoline 4b was accomplished via the
enolate addition/decarboxylation procedure (vide supra) to
give racemic 23b, and chiral HPLC resolution afforded the
nal product 27. X-ray crystallography was used to establish
absolute stereochemical conguration as the (S) congura-
tion as shown (see ESI†).
Fig. 6 Synthesis of 4-benzylquinazoline derivatives.
Limited attempts to crystallize derivative 27 bound to the
PDE1B catalytic subunit were not successful, but computer
modeling studies summarized in Fig. 9 provide a reasonable
description of possible protein–ligand interactions. The qui-
nazoline ring likely forms the same interactions as described for
the X-ray structure of bound 7a. In this orientation the chlorine
atom is well positioned to occupy the hydrophobic pocket
dened by Leu-388 and Phe-392. Signicantly greater affinity for
27 over its enantiomer (IC50 ¼ 2 nM vs. 458 nM) suggest a key
role for the chiral center in projecting the chlorine atom
towards this pocket.
Fig. 7 SAR of PDE1B inhibition by 4-benzylquinazoline inhibitors.
degassed solvent. Efforts to avoid this decomposition by
employing HCl or H2SO4 catalyzed ester hydrolysis were not
successful due to the instability of the quinazoline products to
strong acid. Pure samples of the indanes 23b and 27 were found
to be stable for several months during normal bench-top
storage.
SAR development of the benzyl series is summarized in
Fig. 7. Removal of the chlorine atom and the benzylic methyl
group is detrimental to binding (cf. 19, 20 and 2), and structure
rigidication yields potency improvements as evidenced by the
indane derivative 21b. As anticipated based on 4-amino-quina-
zoline SAR and the X-ray structure, transposition of the quina-
zoline 6-methoxy group to the 8-position afforded a signicant
potency boost (2 vs. 22). Replacement of the chlorine atom with Fig. 8 Synthesis of benzylquinazoline 27.
This journal is © The Royal Society of Chemistry 2014
Med. Chem. Commun., 2014, 5, 1290–1296 | 1293

View Article Online
MedChemComm
Concise Article
receptors (ESI†)31 and does not inhibit the binding of dofetilide
to the hERG ion channel. Compound 7a has good stability in
human liver microsomes (<6.5 mL minꢀ1 Kgꢀ1), somewhat
elevated rat microsomal clearance (45 mL minꢀ1 Kgꢀ1), and is
rapidly absorbed into the systemic circulation aer subcuta-
neous (sc) administration in both mouse and rat (Tmax ¼ 0.25 h).
The sc rodent plasma half-life of 7a is 0.25–0.66 h and it is
moderately bound to plasma proteins with unbound fractions
(fu) of 10–18% (Table 3). The compound readily penetrates the
blood brain barrier as indicated by brain/plasma (B/P) ratios of
0.3–0.4 that remained constant until the compound was elimi-
nated. Concentrations of 7a in the cerebrospinal uid (CSF)
were similar to that free in plasma (CSF/free plasma ¼ 0.9),
further indicating that the compound readily equilibrates into
brain aer systemic administration. When dosed orally to fed
rats (50 mpk), 7a achieved a maximal plasma concentration at
0.25 h of 3333 ng mLꢀ1 (10.9 mM), and expressed a terminal half-
life of 5.1 h (see additional data in ESI†) (Table 2).
Compound 27 presents a more complex structure, possess-
ing a single chiral center and requiring a more challenging
synthesis. The non-basic compound does not form a hydro-
chloride salt, and is sparingly water soluble. Other physical
properties are consistent with excellent brain penetrability (MW
¼ 340, c log P ¼ 4.8, TPSA ¼ 44). The compound is a more
potent inhibitor of PDE1 than 7a and likewise does not offer
appreciative PDE1 isoform selectivity. It exhibits >105ꢁ selec-
tivity for PDE1B and 36ꢁ selectivity for PDE1A over other PDE
isoforms. Broad receptor and enzyme proling (ESI†) reveals
Fig. 9 Computer generated model of benzylquinazoline 27 docked
within the active site of PDE1B (top), and schematic representation.
The quinazoline 7a presents a simple structure with favor-
able attributes for CNS penetration (MW ¼ 305, c log P ¼ 4.2,
TPSA ¼ 66).27 The compound forms a highly water soluble
hydrochloride salt, has good cell permeability, and is not an
efflux substrate. The potencies of 7a for inhibition of recombi-
nant full length and catalytic domain PDE1A, PDE1B, and
PDE1C varied by less than 2-fold, which in addition to the
crystal structure data suggests that 7a is a catalytic-site
competitive inhibitor. Compound 7a exhibits >100-fold selec-
tivity for PDE1B and >30-fold selectivity for PDE1A over other
PDE isoforms (Table 1),28–30 showed little activity when screened
against a wide panel of biologically relevant enzymes and
Table 2 Human PDE inhibition and PDE1B selectivity for compound
27
PDE
IC50 (nM)
Fold selectivity
PDE1A1
PDE1B1
PDE1C1
PDE2A1
PDE3A1
PDE4D3
PDE5A1
PDE6A
PDE7B
PDE8B
PDE9A1
PDE10A1
PDE11A4
2
2.4
7
0.8ꢁ
1ꢁ
2.9ꢁ
5895
>30 000
7620
>30 000
>29 500
>29 800
3467
>30 000
252
2456ꢁ
>12 500ꢁ
3175ꢁ
>12 500ꢁ
>12 292ꢁ
>12 416ꢁ
1444ꢁ
>12 500ꢁ
105ꢁ
Table 1 Human PDE inhibition and PDE1B selectivity for compound 7a
PDE
IC50 (nM)
Fold selectivity
8257
3440ꢁ
PDE1A1
PDE1B1
PDE1C1
PDE2A1
PDE3A1
PDE4D3
PDE5A1
PDE6A
PDE7B
PDE8B
PDE9A1
PDE10A1
PDE11A4
118
35
36
>30 000
>30 000
>23 600
9010
3ꢁ
1ꢁ
Table 3 Select PK data for 7a and 27 in rat and mousea
1ꢁ
>857ꢁ
>857ꢁ
>674ꢁ
257ꢁ
264ꢁ
>857ꢁ
>857ꢁ
>857ꢁ
138ꢁ
108ꢁ
Compd Species Doseb Tmax (h) t1/2
fu
B/P
CSF/Pu
7a
7a
27
Mouse
Rat
Rat
32
10
10
0.25
0.25
0.5
2.0 h
0.66 h 0.10
—
0.18
0.28
0.4
—
0.9
1.7
9260
0.028 3.4
>30 000
>30 000
>30 000
4820
a
mpk ¼ mg Kgꢀ1; sc ¼ subcutaneous; t1/2 ¼ half-life; fu ¼ fraction
Doses were administered subcutaneously; units are mg Kgꢀ1
unbound; B/P ¼ brain/plasma; Pu ¼ unbound fraction in plasma.
b
.
3780
Extensive additional data may be found in the ESI.
1294 | Med. Chem. Commun., 2014, 5, 1290–1296
This journal is © The Royal Society of Chemistry 2014

View Article Online
Concise Article
MedChemComm
that the 27 interacts with serotonin and adenosine receptors
(5HT2B IC50 ¼ 250 nM; adenosine A3 IC50 ¼ 370 nM) and is an
inhibitor of cytochrome P450 2C19 (IC50 ¼ 870 nM). The
compound is highly bound to plasma proteins (fu ¼ 2.8%) and
when dosed orally at 50 mpk in the rat it achieves plasma Tmax
of 274 ng mLꢀ1 (806 nM) at 0.5 h (Table 3) with a terminal half-
life of 5.5 h. Plasma exposure appears to plateau between the
50 mg Kgꢀ1 and 200 mg Kgꢀ1 oral doses, possibly due to solu-
bility limited absorption. While nearly 3ꢁ greater plasma
exposures were achieved with sc dosing, a similar plateau event
was observed between the 50 mg Kgꢀ1 and 200 mg Kgꢀ1 doses
(see additional data in ESI†). The compound is readily absorbed
into the CNS (B/P ¼ 3.4) and CSF concentrations were slightly
higher than that in free plasma (CSF/free plasma ¼ 1.7), further
illustrating unimpeded brain penetration.
Notes and references
1 J. A. Beavo, Physiol. Rev., 1995, 75, 725.
2 M. Conti and J. Beavo, Annu. Rev. Biochem., 2007, 76, 481.
3 C. Mehats, C. B. Andersen, M. Filopanti, S. L. C. Jin and
M. Conti, Trends Endocrinol. Metab., 2002, 13, 29.
4 C. Lugnier, Pharmacol. Ther., 2006, 109, 366.
5 F. S. Menniti, W. S. Faraci and C. J. Schmidt, Nat. Rev. Drug
Discovery, 2006, 5, 660.
6 A. Bender, in Cyclic Nucleotide Phosphodiesterases in Health
and Disease, ed. J. A. Beavo, S. H. Francis and M. D.
Houslay, CRC press, 2006, p. 35.
7 Y. Cheung, Biochem. Biophys. Res. Commun., 1970, 38, 533.
8 S. Y. R. Kakiuchi, Biochem. Biophys. Res. Commun., 1970, 41,
1104.
9 C. Yan, A. Z. Zhao, J. K. Bentley, K. Loughney, K. Ferguson
and J. A. Beavo, Proc. Natl. Acad. Sci. U. S. A., 1995, 92, 9677.
10 C. Yan, J. Bentley, W. Sonnenburg and J. Beavo, J. Neurosci.,
1994, 14, 973.
11 T. M. Reed, D. R. Repaske, G. L. Snyder, P. Greengard and
C. V. Vorhees, J. Neurosci., 2002, 22, 5188.
12 M.-L. Wong, F. Whelan, P. Deloukas, P. Whittaker,
M. Delgado, R. M. Cantor, S. M. McCann and J. Licinio,
Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 15124.
13 A. Blokland, F. S. Menniti and J. Prickaerts, Expert Opin.
Ther. Pat., 2012, 22, 349.
14 S. Gupta and B. Sharma, Eur. J. Pharmacol., 2014, 732, 111.
15 A series of potent PDE1 inhibitors has been described with
one example achieving a PDE1 IC50 of 0.07 nM. While
members of the series can achieve good selectivty for PDE1
over PDE5 and vice versa, full selectivity data vs. the other
PDE families (PDE2, PDE4, and PDE's 6-11) was not
reported. See H.-S. Ahn, et al., J. Med. Chem., 1997, 40,
2196–2210.
Conclusions
In conclusion, we have disclosed a new class of selective, CNS
penetrable quinazoline PDE1 inhibitors discovered through SAR
development of two le screen hits. In a combination of parallel
and traditional organic synthesis, and with the aid of X-ray
crystallography and molecular modeling, structural renement
of the lead compounds led to the aminoquinazoline 7a (PF-
04471141) and the indanylquinazoline 27 (PF-04822163). These
compounds are among the most potent and selective inhibitors
of PDE1 reported to date. Pharmacokinetic investigations in
rodents indicate that each compound achieves systemic
concentrations in excess of their IC50 values. Thus, compounds
7a and 27 offer considerable potential as chemical probes to
further investigate biological processes of the CNS impacted by
PDE1 function, and to assess the potential of pan-PDE1 inhibi-
tion as a therapy for the treatment of neuropsychiatric illness.
16 T. A. Chappie, J. M. Humphrey, M. P. Allen, K. G. Estep,
C. B. Fox, L. A. Lebel, S. Liras, E. S. Marr, F. S. Menniti,
J. Pandit, C. J. Schmidt, M. Tu, R. D. Williams and
F. V. Yang, J. Med. Chem., 2007, 50, 182.
17 T. T. Wager, R. Y. Chandrasekaran, X. Hou, M. D. Troutman,
P. R. Verhoest, A. Villalobos and Y. Will, ACS Chem. Neurosci.,
2010, 1, 420.
18 T. T. Wager, X. Hou, P. R. Verhoest and A. Villalobos, ACS
Chem. Neurosci., 2010, 1, 435.
19 Y. Takase, N. Watanabe, H. Adachi, K. Kodama, H. Ishihara,
T. Saeki and S. Souda, WO9507267, Esai Co, L., 1995.
20 R. Pamukeu and G. Piazza, US5990117, Cell Pathways, I,
USA, 1999.
Abbreviations
(ꢂ)
Racemic compound
Phosphodiesterase
Cyclic adenosine monophosphate
Cyclic guanosine monophosphate
Central nervous system
Crystallized
Long term potentiation
Long term depression
N-Chlorosuccinimide
Phencyclidine
PDE
cAMP
cGMP
CNS
cryst
LTP
LTD
NCS
PCP
PK
Pharmacokinetics
21 Y. Takase, T. Saeki, M. Fujimoto and I. Saito, J. Med. Chem.,
1993, 36, 3675.
rt
Room temperature
sc
Subcutaneous
22 Y. Takase, T. Saeki, N. Watanabe, H. Adachi, S. Souda and
I. Saito, J. Med. Chem., 1994, 37, 2106.
TFA
Triuoroacetic acid
23 J. M. Humphrey, T. A. Chappie and S. Liras, US0019975A1,
Pzer Inc, USA, 2006, p. 32.
24 C. E. Katz and J. Aube, J. Am. Chem. Soc., 2003, 125, 13948.
25 A. H. Mermerian and G. C. Fu, J. Am. Chem. Soc., 2005, 127,
5604.
Acknowledgements
We thank Stacey L. Becker and Frederick R. Nelson for gener-
ation of pharmacokinetic data.
This journal is © The Royal Society of Chemistry 2014
Med. Chem. Commun., 2014, 5, 1290–1296 | 1295

View Article Online
MedChemComm
Concise Article
26 T. Aono, S. Kishimoto, Y. Araki and S. Noguchi, Chem.
Pharm. Bull., 1977, 25, 3198.
C. J. Schmidt, J. A. Suiciak and S. Liras, J. Med. Chem.,
2009, 52, 5188.
27 T. T. Wager, R. Y. Chandrasekaran, X. Hou, M. D. Troutman, 30 C. J. Schmidt, D. S. Chapin, J. Cianfrogna, M. L. Corman,
P. R. Verhoest, A. Villalobos and Y. Will, ACS Chem. Neurosci.,
1, 420.
28 J.A.Siuciak,D.S.Chapin,J.F.Harms,L.A.Lebel,S.A.McCarthy,
L. Chambers, A. Shrikhande, S. Wong, F. S. Menniti and
C. J. Schmidt, Neuropharmacology, 2006, 51, 386.
29 P. R. Verhoest, D. S. Chapin, M. Corman, K. Fonseca,
M. Hajos, J. F. Harms, W. E. Hoffman, L. A. Lebel,
S. A. McCarthy, F. R. Nelson, C. Proulx-LaFrance,
M. J. Majchrzak, A. D. Ramirez, K. Schmidt, P. A. Seymour,
J. A. Siuciak, F. D. Tingley III, R. D. Williams,
P. R. Verhoest and F. S. Menniti, J. Pharmacol. Exp. Ther.,
2008, 325, 681.
J. F. Harms, X. Hou, E. S. Marr, F. S. Menniti, F. Nelson, 31 Broad receptor proling was accomplished by Cerep, Inc.,
R. O'Connor, J. Pandit, C. Proulx-LaFrance, A. W. Schmidt,
Redmond, WA.
1296 | Med. Chem. Commun., 2014, 5, 1290–1296
This journal is © The Royal Society of Chemistry 2014