Discovery of a series of 6,7-dimethoxy-4-pyrrolidylquinazoline PDE10A inhibitors

Article abstract of DOI:10.1021/jm060653b

A papaverine based pharmacophore model for PDE10A inhibition was generated via SBDD and used to design a library of 4-amino-6,7-dimethoxyquinazolines. From this library emerged an aryl ether pyrrolidyl 6,7-dimethoxyquinazoline series that became the focal

Full text of DOI:10.1021/jm060653b

182
J. Med. Chem. 2007, 50, 182-185
Letters
Discovery of a Series of
6,7-Dimethoxy-4-pyrrolidylquinazoline PDE10A
Inhibitors†
Thomas A. Chappie,* John M. Humphrey, Martin P. Allen,
Kimberly G. Estep, Carol B. Fox, Lorraine A. Lebel,
Spiros Liras, Eric S. Marr, Frank S. Menniti,
Jayvardhan Pandit, Christopher J. Schmidt, Meihua Tu,
Robert D. Williams, and Feng V. Yang
CNS DiscoVery and Experimental Medicinal Sciences, Pfizer Global
Research and DeVelopment, Eastern Point Road,
Groton, Connecticut 06340
ReceiVed June 1, 2006
Abstract: A papaverine based pharmacophore model for PDE10A
inhibition was generated via SBDD and used to design a library of
4-amino-6,7-dimethoxyquinazolines. From this library emerged an aryl
ether pyrrolidyl 6,7-dimethoxyquinazoline series that became the focal
point for additional modeling, X-ray, and synthetic efforts toward
increasing PDE10A inhibitory potency and selectivity versus PDE3A/
B. These efforts culminated in the discovery of 29, a potent and selective
brain penetrable inhibitor of PDE10A.
Figure 1. Cocrystal structure of 1 bound to the PDE10A catalytic
site. Important PDE10A binding residues are highlighted in magenta.
The phosphodiesterases (PDEs) are a family of widely
expressed enzymes that hydrolyze the 3′-5′ phosphodiester
bond of the ubiquitous intracellular messenger molecules cyclic
adenosine monophosphate (cAMP) and cyclic guanosine mono-
phosphate (cGMP) to effect signal termination. The PDEs are
encoded by 21 different genes in humans, with over 60 isoforms
expressed as a result of alternative splicing patterns.1 Although
considerable variation is possible among the full-length enzymes,
the catalytic domains share common functionality and similar
modular structures with several conserved amino acid residues.
Key residues in the catalytic core form a hydrophobic pocket
that accommodates the planar nucleotide in a spatial orientation
that facilitates hydrogen bonding to an absolutely conserved
glutamine residue.2 Variations in the amino acid sequence at
the catalytic domain and elsewhere give rise to unique substrate
specificities, catalytic efficiencies, and pharmacologies and
permit subdivision of the PDEs into 11 distinct families. Given
the unique role of each PDE in regulating cyclic nucleotide
signaling in various organ systems, it is not surprising that
research in this area has produced marketed drugs for a variety
of disease states.3-7 Critical to the continued success of such
efforts is the ability to identify and exploit the subtle structural
differences within the catalytic cores of the different PDE
isoforms.
Figure 2. PDE10A schematic pharmacophore model.
PDE3A and 3B (Ki ) 17, 279, and 417 nM, respectively). In
studies employing papaverine as a pharmacological probe in
wild type and PDE10A knock out mice, we have found evidence
to suggest that PDE10A inhibitors may be effective for the
treatment of psychosis in schizophrenia.14-16 To advance this
theory, we initiated a chemistry effort aimed at designing more
potent and selective pharmacological tools to further our
understanding of PDE10A physiology and the therapeutic
potential of PDE10A inhibitors. This search for a new class of
PDE10 inhibitors was based on the structure of papaverine.
The recently reported tetrahydroisoquinolinyl dimethoxy-
quinazoline derivative (1, Figure 1),17 derived from a file screen
hit, was an attractive starting point for these studies due to its
reasonable potency versus PDE10A (Ki ) 25 nM) and structural
similarity to papaverine. In addition, chemical design would be
greatly facilitated by access to the X-ray crystal structure of
compound 1 bound within the catalytic subunit of PDE10A.18
According to this structure, the inhibitor 1 forms two key
interactions. In the first of these, a bidentate hydrogen-bonding
interaction between Gln-716 and the two methoxy groups of
the quinazoline ring, in which the methoxy groups share a single
N-H proton donated by the glutamine residue, anchors the
ligand within the active site. In the second essential interaction
the planar quinazoline ring occupies a hydrophobic region
defined by the residues Phe-719 and Phe-686 (Figure 2). In
doing so, the quinazoline achieves a π-stacking interaction with
PDE10A is the single member of a dual substrate (cGMP/
cAMP) family first reported in 1999.8-11 The enzyme is heavily
expressed by the medium spiny neurons of the mammalian
striatal complex,12,13 making it an intriguing target for the
treatment of diseases of the central nervous system. Work from
our laboratories shows that papaverine, a naturally occurring
plant alkaloid and smooth muscle relaxant, is a relatively potent
inhibitor of PDE10A, exhibiting moderate cross-reactivity versus
† PDB ID code: 208H.
* To whom correspondence should be addressed. Phone: (860) 715-
5142. Fax: (860) 686-1059. E-mail: thomas.a.chappie@pfizer.com.
10.1021/jm060653b CCC: $37.00 © 2007 American Chemical Society
Published on Web 12/22/2006

Letters
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 2 183
Table 1. PDE Kia
Scheme 1. Synthesis of 8-29a
#
X )
PDE10A
PDE3A
PDE3B
>100
<100
>100
>100
2
2-CF3
2-CH3
2-OCH3
2-OCH3
(S)-ent
2-Cl
2-CN
4-CN
4-OCF3
4-t-Bu
4-OEt
811
326
240
641
<100
<100
<100
>100
3
4
5
6
7
8
9
10
11
198
130
56
185
100
54
<100
<100
214 (4×)
739 (4×)
720 (7×)
161 (3×)
<100
<100
471 (8×)
1820 (10×)
1890 (19×)
328 (6×)
a Reagents and conditions: (a) (i) ArOH, DEAD, PPh3, THF, rt, 16 h;
(ii) TFA, CH2Cl2, rt, 4 h, 66-99%; (b) 4-chloro-6,7-dimethoxyquinazoline,
K2CO3, PhCH3, 80 °C, 16 h, 45-99%.
a Reported in nM. Selectivity vs PDE3A/B in parenthesis.
the “upper” residue, Phe-719, and an edge-on interaction with
Phe-686 located at the bottom of the pocket. Additional
favorable hydrophobic binding interactions occur between the
aryl portion of the tetrahydroisoquinoline fragment of 1 and
the protein, while the sulfonamide group is exposed to solvent.
Information obtained from the X-ray structure of bound 1
and from modeling studies with papaverine enabled construction
of a schematic PDE10A “pharmacophore model” that we would
use to design new compounds (Figure 2). The model is
comprised of three parts. First, a dimethoxy component is
employed to retain the bidentate Gln-716 hydrogen bond
interaction observed in 1. Second, a fused biaryl ring system
attached to the dimethoxy component occupies the hydrophobic
clamp defined by Phe-719 and Phe-686. Last, an aryl group is
appropriately attached to mimic the substituent aryl rings of
papaverine and 1, thereby increasing affinity through nonspecific
hydrophobic interactions with the protein (Figure 2).
Using this information, new inhibitor design efforts began
in parallel to work on 1. In advance of the chemical synthesis,
we prepared a virtual library of >2000 6,7-dimethoxyquinazo-
lines substituted at C4 with various aryl-bearing amine com-
ponents. The library was filtered by calculated properties
predictive of good ADME, including molecular weight, cLogP,
and TPSA, to afford some 300 compounds for actual library
synthesis. The synthetic work simply required coupling of
4-chloro-6,7-dimethoxyquinazoline19-21 with the filtered set of
aryl-bearing amines identified from the virtual library (Scheme
1, step b). HPLC purification afforded the desired products in
moderate yield. Upon analysis of the products, a collection of
chiral 4-pyrrolidyl-6,7-dimethoxyquinazolines emerged as prom-
ising leads due to moderate PDE10A potency and discernible
SAR (compounds 2-8, Table 1). Although potency and
selectivity were compromised in relation to 1, the series provided
many advantages with respect to physical chemical properties
needed for a CNS-based target. Also evident from this library
was that (i) the greatest PDE10A inhibitory activity resided with
the R-isomer; (ii) PDE3A and PDE3B inhibitory activity was
present and considerable, and (iii) as indicated by compound
8, achievement of PDE10A selectivity over PDE3A/B appeared
possible.
Figure 3. Cocrystal structure of 11 (cyan) bound to the PDE10A
catalytic site. Important residues are highlighted in PDE10A (magenta)
and PDE3A/B (yellow/green).
analogs were prepared manually beginning with commercial N-t-
Boc-(S)-3-hydroxypyrrolidine. Inversion at C3 and concurrent
substitution with variously substituted phenols was accomplished
via the Mitsunobu reaction and provided, after deprotection, the
3-aryloxy pyrrolidines in acceptable yield.22 Reaction with
4-chloro-6,7-dimethoxyquinazoline proceeded efficiently, pro-
viding the para-substituted target compounds 8-11 (Scheme
1).
At this point, X-ray diffraction studies and computational
modeling assumed a prominent role in our efforts to understand
PDE10/PDE3 selectivity. Paramount to these efforts was our
structure elucidation of the complex formed between the
catalytic domain of PDE10A and ligand 11. Coordinates from
these studies were overlaid with those provided by an in-house
homology model of the catalytic domains of PDE3A/B to
highlight regional differences between the enzymes and to
generate ideas for designing selective compounds (Figure 3).
Overlay of PDE3A/B homology model with PDE10 X-ray
structure reveals that a synthetically accessible difference exists
in the region proximal to the aryl ether appendage of bound
ligand. Specifically, the small residues Gly-718 and Ala-722
of PDE10 are replaced by the larger amino acids Ser (G718S)
and His (A722H), respectively, in both PDE3A and PDE3B.
Steric hindrance thus appeared to play a determinant role in
the observed PDE3B selectivity gains with ligand 11 versus
less selective ligands. The observation that analogs possessing
para-substitution patterns on the aryloxy ring achieved greater
selectivity and potency than ortho-substituted analogs guided
us in our next design efforts, which are shown in Table 2.
A diverse set of 3-aryloxy and 3-heteroaryloxy pyrrolidine
analogs were synthesized according to the route of Scheme 1.
Comparable potency and selectivity observed with 8 versus
papaverine warranted development of an efficient enantio-
selective synthesis of this chemotype. Thus, a small set of new

184 Journal of Medicinal Chemistry, 2007, Vol. 50, No. 2
Table 2. PDE Kia
Letters
Figure 4. Overlaid crystal structures of 21 (magenta) and 29 (orange)
in PDE10A. PDE3A/B residues yielding unfavorable steric interactions
are included as overlays.
ited >38-fold selectivity for PDE10 when assessed versus a
panel of 60 different CNS-associated receptors, enzymes, and
ion channels.
X-ray structures of PDE10A-inhibitor complexes with
compounds 21 and 29 reveal nearly identical binding modes
for each (Figure 4). Evident from these images is that increased
potency of this subset, relative to the mono-aryl derivatives, is
likely due to enhanced lipophilic interactions between the
napthyl/quinoxaline ring and underlying hydrophobic residues
(Ala-722, Val-723, Phe-719) of the enzyme. Also overlaid on
the PDE10 structure in Figure 4 are the two PDE3A/B residues,
Ser and His, to which we attribute reduced PDE3 affinity. For
the sake of clarity, we have shown rotamers of these residues
that are turned away from the bound ligand; however, it is
understood that free rotation of the His and Ser sidechains would
serve to reduce the effective size of the PDE3 binding pocket
relative to PDE10.
In light of its attractive in Vitro properties, 29 was further
investigated for evidence of the ability to inhibit PDE10A in
ViVo. In previous studies we noted that administration of
papaverine to mice causes a rapid elevation in levels of cGMP
and a concomitant increase in the phosphorylation of the
downstream transcription factor cAMP response element binding
protein (CREB) in the striatal regions of the brain.14 These
effects are consistent with the high level of expression of
PDE10A in this brain region13 and the ability of PDE10A to
metabolize both cGMP and cAMP.8-11 That the effect of
papaverine is due specifically to inhibition of PDE10A is
confirmed by the finding that papaverine has no effect in mice
in which the gene for PDE10A has been disrupted (i.e., PDE10A
knockout mice).15 Thus, we investigated the effects of 29 on
striatal cGMP and CREB phosphorylation in mice.
a Reported in nM. Selectivity vs PDE3A/B in parenthesis.
These new compounds were designed to assess the effects of
other aryl substitution patterns and pyrridyl ring systems on PDE
inhibition. In general, the data indicates that steric bulk in the
meta-position is detrimental to activity versus all three enzymes
(compounds 12-14), whereas relatively smaller groups in a
meta-/para-disubstitution pattern provided no improvements in
PDE10 potency and selectivity (15 and 16). 5-Cloropyrimidine
17 provides potency equivalent to papaverine but without
significant selectivity.
Gains in both potency and selectivity were finally realized
in a subset of these compounds (21-29) wherein the simple
monocyclic aryl ethers were replaced with larger aromatic- and
heteroaromatic-bicyclic ethers. As evident from a comparison
of 21 and 22, the 2-napthyl group offers far greater PDE10
potency over the 1-napthyl group (Ki ) 12 nM vs 210 nM,
respectively) and improved selectivity over PDE3A/B. Methoxy
substitution did not enhance the selectivity (21 vs 23 and 24).
Various other quinoline and isoquinoline ring systems were
examined in anticipation of improved PDE10 potency as was
observed with the pyrimidine 17. For the most part, however,
these analogs exhibited PDE10 potencies comparable with 21
and reduced selectivities spanning a range from ∼8× to ∼20×.
In a particularly favorable discovery, (R)-6,7-dimethoxy-4-(3-
(quinoxalin-2-yloxy)pyrrolidin-1-yl)quinazoline 29 proved to be
the most potent (4 nM) and selective (53-68-fold) PDE10
inhibitor of this quinazoline campaign and, furthermore, exhib-
Compound 29 administered subcutaneously (32 mg/kg)
caused a dose-dependent increase in striatal cGMP level after
30 min (Supporting Information). This effect of 29 was evident
for at least 90 min after administration and was completely
attenuated at 240 min after administration (Figure 5). Compound
29 also caused an increase in striatal CREB phosphorylation
over a time course similar to that for the elevation of striatal

Letters
Journal of Medicinal Chemistry, 2007, Vol. 50, No. 2 185
(4) Young, R. A; Ward, A. Milrinone. A Preliminary Review of its
Pharmacological Properties and Therapeutic Use. Drugs 1988, 36
(2), 158-192.
(5) Liu, Y.; Shakur, Y.; Yoshitake, M.; Kambayashi, J.-L. Cilostazol
(Pletal): A Dual Inhibitor of Cyclic Nucleotide Phosphodiesterase
Type 3 and Adenosine Uptake. CardioVasc. Drug ReV. 2001, 19 (4),
369-386.
(6) Wang, P.; Billah, M. M. Phosphodiesterase-4 (PDE4) as a Target
for Anti-Inflammatory Drug Discovery: Current Status and Future
Direction. Recent Res. DeV. Life Sci. 2003, 1 (2), 275-290.
(7) Zhu, J.; Mix, E.; Winblad, B. The Antidepressant and Anti-
Inflammatory Effects of Rolipram in the Central Nervous System.
CNS Drug ReV. 2001, 7 (4), 387-398.
(8) Loughney, K.; Snyder, P. B.; Uher, L.; Rosman, G. J.; Ferguson,
K.; Florio, V. A. Isolation and Characterization of PDE10A, A Novel
Human 3′-5-Cyclic Nucleotide Phosphodiesterase. Gene 1999, 234,
109-117.
Figure 5. The time course of effect of 29 (32 mg/kg, s.c. in 40%
â-cyclodextrin) on striatal cGMP and phosphorylated CREB levels in
mice.
(9) Soderling, S. H.; Bayuga, S. J.; Beavo, J. A. Isolation and
Characterization of a Dual-Substrate Phosphodiesterase Gene Fam-
ily: PDE10A. Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 7071-7076.
(10) Fujishige, K.; Kotera, J.; Michibata, H.; Yuasa, K.; Takebayashi, S.;
Okumura, K.; Omori, K. Cloning and Characterization of a Novel
Human Phosphodiesterase that Hydrolyzes both cAMP and cGMP
(PDE10A). J. Biol. Chem. 1999, 274 (26), 18438-18445.
(11) Weston, M. C.; Holst, L. S.; Degerman, E.; Manganiello, V. C.
cAMP/cGMP Dual-Specificity Phosphodiesterases. In Handbook of
Cell Signalling, Volume 2; Dennis, R. A. B. a. E. A., Ed.; Academic
Press: New York, 2003; Vol. 2, pp 441-446.
cGMP (Figure 5). These results indicate that 29 inhibits
PDE10A in the brain after systemic administration. Compound
29 is somewhat more potent than papaverine on a mg/kg basis.
Notably, the duration of action of 29 is considerably longer than
that of papaverine, comparing present results with those reported
in Siuciak et al.14 Thus, given the favorable in Vitro potency
and selectivity and in ViVo activity profile, 29 represents an
important new pharmacological tool to investigate the physiol-
ogy and therapeutic utility of PDE10A.
(12) Fujishige, K.; Kotera, J.; Omori, K. Striatum- and Testes-Specific
Phosphodiesterase PDE10A. Isolation and Characterization of a Rat
PDE10A. Eur. J. Biochem. 1999, 266, 1118-1127.
(13) Seeger, T. F.; Bartlett, B.; Coskran, T. M.; Culp, J. S.; James, L. C.;
Krull, D. L.; Lanfear, J.; Ryan, A. M.; Schmidt, C. J.; Strick, C. A.;
Varghese, A. H.; Williams, R. D.; Wylie, P. G.; Menniti, F. S.
Immunohistochemical Localization of PDE10A in the Rat Brain.
Brain Res. 2003, 985, 113-126.
(14) Siuciak, J. A.; Chapin, D. S.; Harms, J. F.; Lebel, L. A.; James, L.
C.; McCarthy, S. A.; Chambers, L. K.; Shrikehande, A.; Wong, S.
K.; Menniti, F. S.; Schmidt, C. J. Inhibition of the Striatum-
Enrichment Phosphodiesterase PDE10A: A Novel Approach to the
Treatment of Psychosis. Neuropharmacology 2006, 51 (2), 386-
396.
(15) Siuciak, J. A.; McCarthy, S. A.; Chapin, D. S.; Fujiwara, R. A.; James,
L. C.; Williams, R. D.; Stock, J. L.; McNeish, J. D.; Strick, C. A.;
Menniti, F. S.; Schmidt, C. J. Genetic Deletion of the Striatum-
Enriched Phosphodiesterase PDE10A: A Novel Approach to the
Treatment of Psychosis. Neuropharmacology 2006, 51 (2), 374-
385.
In summary, X-ray data from the ligand 1/PDE10A catalytic
site complex provided a schematic pharmacophore model that
was used to guide efforts toward a new series of inhibitors
illustrated by 11. With further X-ray data and computational
modeling, differences in the PDE10A and PDE3A/B catalytic
sites were identified and exploited to give compounds with
enhanced PDE10A potency and selectivity. The culmination of
this work led to 29, a potent (4 nM) and selective (>38-fold vs
PDEs and CNS related targets) compound that increases cGMP
and pCREB levels in mouse striatum in a dose- and time-
dependent manner. Compound 29 is currently being used as a
tool to further elucidate PDE10A inhibitory activity on brain
function.
Acknowledgment. We thank S. Simons, K. Fennel, and T.
Subashi for cloning and expression of the recombinant protein
construct used in the crystallography experiments.
(16) Lebel, L. A.; Menniti, F. S.; Schmidt, C. J. Use of Selective PDE10
Inhibitors for the Treatment of Central Nervous System Disorders.
Eur. Pat. Appl. 1 250 923 A2, Oct. 23, 2002.
(17) Allen, M. P.; Chappie, T. A.; Humphrey, J. M.; Liras, S. Tetra-
hydroisoquinoline Derivatives of Quinazoline and Isoquinoline.
US20050182079 A1, Aug. 18, 2005.
(18) Pandit, J. Crystal Structure of 3′,5′-Cyclic Nucleotide Phospho-
diesterase (PDE10A) and Uses Thereof. US20050202550 A1, Sept.
15, 2005.
(19) Rocco, S. A; Barbarini, J. E.; Rittner, R. Synthesis of Some
4-Anilinoquinazoline Derivatives. Synthesis 2004, 3, 429-435.
(20) Alexandre, F.; Berecibar, A.; Wrigglesworth, R.; Besson, T. Novel
Series of 8H-Quinazolino[4,3-b]quinazoline-8-ones Via Two
Niementowski Condensations. Tetrahedron 2003, 59 (9), 1413-1419.
(21) Liu, X.; Narla, R. K.; Uckun, F. Organic Phenyl Arsenic Acid
Compounds with Potent Antileukemic Activity. Bioorg. Med. Chem.
Lett. 2003, 13 (3), 581-583.
(22) Nagahara, T.; Yokoyama, Y.; Inamura, K.; Komoriya, S.; Yamaguchi,
H.; Hara, T.; Iwanoto, M. Dibasic (Amindinoaryl)propanoic Acid
Derivatives as Novel Blood Coagulation Factor Xa Inhibitors. J. Med.
Chem. 1994, 37, 1200-1207.
Supporting Information Available: Combinatorial experimen-
tal conditions, experimental crystal structure data, computational
models, and chemical and biological experimental details. This
material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) Manallack, D. T.; Hughes, R. A.; Thompson, P. E. The Next
Generation of Phosphodiesterase Inhibitors: Structural Clues to
Ligand and Substrate Selectivity of Phosphodiesterases. J. Med.
Chem. 2005, 48 (10), 3449-3462.
(2) Card, G. L.; England, B. P.; Suzuki, Y.; Fong, D.; Powell, B.; Lee,
B.; Luu, C.; Tabrizizad, M.; Gillete, S.; Ibrahim, P. N.; Artis, D. R.;
Bollag, G.; Milburn, M. V.; Kim, S.; Schlessinger, J.; Zang, K. Y. J.
Structural Basis for the Activity of Drugs that Inhibit Phospho-
diesterases. Structure 2004, 12, 2233-2247.
(3) Setter, S. M.; Iltz, J. L.; Fincham, J. E.; Campbell, R. K.; Baker, D.
E. Phosphodiesterase 5 Inhibitors for Erectile Dysfunction. Ann.
Pharmacol. 2005, 39 (7/8), 1286-1295.
JM060653B