Identification of a brain penetrant PDE9A inhibitor utilizing prospective design and chemical enablement as a rapid lead optimization strategy

Article abstract of DOI:10.1021/jm9015334

By use of chemical enablement and prospective design, a novel series of selective, brain penetrant PDE9A inhibitors have been identified that are capable of producing in vivo elevations of brain cGMP.

Full text of DOI:10.1021/jm9015334

7946 J. Med. Chem. 2009, 52, 7946–7949
DOI: 10.1021/jm9015334
Identification of a Brain Penetrant PDE9A
Inhibitor Utilizing Prospective Design and
Chemical Enablement as a Rapid Lead
Optimization Strategy
Patrick R. Verhoest,* Caroline Proulx-Lafrance,
Michael Corman, Lois Chenard, Christopher J. Helal,
Xinjun Hou, Robin Kleiman, Shenping Liu, Eric Marr,
Frank S. Menniti, Christopher J. Schmidt,
Michelle Vanase-Frawley, Anne W. Schmidt,
Robert D. Williams, Frederick R. Nelson, Kari R. Fonseca,
and Spiros Liras*
Figure 1. Comparison of key properties for sildenafil and 1 vs 2.
Neuroscience Chemistry, Pfizer Global Research and Development,
Eastern Point Road, Groton, Connecticut 06340
synapses, and ultimately lead to an improvement in cognitive
deficits observed in patients with Alzheimer’s disease.
The Pfizer neuroscience PDE platform represents a valu-
able opportunity to capitalize on the accumulated company
experience from the discovery of Sildenafil and subsequently
from the prosecution of the PDE10A program. We developed
a chemistry strategy for the prosecution of additional mem-
bers of the PDE family of interest to neuroscience. The
strategy was based on the principles of prospective design
and chemical enablement. Prospective design enables us
to probe in tandem multiple layers of hypotheses for lead
optimization based on knowledge. Chemical enablement
is the efficient bond disconnection that facilitates the syn-
thesis of libraries that probe multiple hypotheses. This strat-
egy enabled the rapid parallel optimization of multiple lead
options.
Internal screening efforts and in silico ADME assessment
yielded one brain penetrating chemical series of interest
depicted in Figure 1. The screening hit 1 possessed numerous
attractive features including high affinity for PDE9A (2.1
nM), low molecular weight (<300), high ligand efficiency
(LE = 0.53),10 and a brain to plasma ratio (B/P) of 1.2. This
lead was plagued by numerous liabilities such as high clear-
ance, poor selectivity across multiple PDEs (PDE1C 2.1 nM),
and low solubility.
To optimize clearance and retain high exposure in the CNS,
we focused our design plan on a set of key molecular proper-
ties identified as critical based on the PDE10A project con-
clusions. For example, the PDE10A candidate (1) achieved
appreciable brain exposure with a molecular weight of <400
and a modest N/O count. Other quinazoline PDE10A inhibi-
tors were able to achieve brain penetration with higher
TPSA11 (83) and lower ClogP. The key to their central access
was that unlike many CNS drugs, they do not possess a
strongly basic nitrogen, and they minimize molecular weight
and hydrogen bond donor count.
Incontrast, sildenafil, whichdemonstratesmodestabilityto
enter the CNS,12 is characterized by a higher molecular
weight, a high count of N/O atoms (resulting in an elevated
TPSA), and a hydrogen bond donor. Thus, the guiding
principles for aligning clearance, CNS disposition, and
safety13 were postulated to be maintaining molecular weight
under 430 and ClogP under 3.0, limiting the number of
H bonddonoratoms to1, and targetingTPSA inthe spectrum
between 50 and 100.
Received October 16, 2009
Abstract: By use of chemical enablement and prospective design, a
novel series of selective, brain penetrant PDE9A inhibitors have
been identified that are capable of producing in vivo elevations of
brain cGMP.
Members of the phosphodiesterase (PDEa) gene family are
widely and abundantly expressed in the brain.1 On the basis of
the physiological importance of cyclic nucleotide signaling in
the central nervous system (CNS) and the therapeutic poten-
tial for restoration of dysfunctional neuronal signaling cas-
cades via PDE inhibition, we launched an initiative to identify
potential therapeutic utility for several CNS disorders. An
attractive feature of the PDE neuroscience discovery platform
was the apparent drug-ability of the gene family.2 We have
previously disclosed the discovery of several novel classes of
selective PDE10A inhibitors,3 including the discovery of PF-
2545920 (1), a selective PDE10A inhibitor that has entered
clinical development for the treatment of schizophrenia.4
Here we report our subsequent efforts to identify and
optimize selective PDE9Ainhibitors. PDE9A is a high affinity
cGMP-specific PDE expressed in the brain.5 The high affinity
for cGMP (Km = 170 nM) relative to all other PDE enzymes
suggests a physiological role in modulation of cGMP in a
lower cytoplasmic cGMP concentration range than any other
PDE.6 Our main objective was to identify selective pharma-
cological tools that would allow us to establish confidence in
the biological hypothesis that PDE9A inhibition will treat
cognitivedeficits.7 It hasbeen reported thatcGMPcan reverse
Aβ induced deficits in long-term potentiation in hippocampal
slices8 and improve cognition in behavioral studies.9 We
hypothesize that inhibition of the enzyme will elevate cGMP
and improve synaptic transmission, stabilize vulnerable
*To whom correspondence should be addressed. For P.R.V.: phone,
860-686-2288; fax, 860-686-0013; e-mail, patrick.r.verhoest@pfizer.
com. For S.L.:, phone, 860-441-1739; fax, 860-715-8224; e-mail, spiros.
liras@pfizer.com.
a Abbreviations: PDE, phosphodiesterase; cGMP, cyclic guanosine
monophosphate; CNS, central nervous system; SBDD, structure based
drug design.
r
pubs.acs.org/jmc
Published on Web 11/17/2009
2009 American Chemical Society

Letter
The second objective was to develop a strategy for the rapid
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24 7947
delivery of tools with adequate exposure to trigger pharma-
codynamic effects. Pharmacodynamic effects were observed
at free plasma or CSF exposures that corresponded to
approximately 5 times the IC50 in previous projects targeting
PDE514 and PDE10A. Our aim was to maintain potency and
optimize free fraction, in vivo efficacy, and improve the
metabolic profile in parallel by reducing lipophilicity. We
expected thesemodifications wouldincreaseplasmaand brain
free fraction.15
Our final objective was to devise a synthesis plan that would
allow us to meet our design objectives with a minimum
number of design-synthesis cycles. Thus, we built an innova-
tive synthesis plan to probe in tandem a range of substitutions
for both the pyrazole carbocycle and the pyrimidinone benzyl
group. We developed a plate-based chemistry approach for
the parallel manipulation of two vectors which is depicted in
Scheme 1.
Figure 2. PDE9A inhibitor 2 (PDB 3JS1) bound and structural
differences with PDE1C (pink).
Scheme 1. Core Parallel Synthesis
The crystal structure of the lead inhibitor 2 (Figure 2)
showed that the pyrimidinone made a two-point hydrogen
bond interaction with the conserved glutamine (Gln-453)
similar to the known PDE5 inhibitors. The benzyl group is
situated in a lipophilic pocket formed by Val-417, Leu-421,
Phe-441, and Ala-452. Utilizing the PDE9A crystal struc-
ture of 2 and a homology model of PDE1C, we developed a
strategy for targeting the two residues that are distinct and
in proximity to the conserved glutamine.18 PDE9A con-
tains Tyr-424, a unique residue within the PDE family. In
PDE1C and most other PDEs this position contains
a phenylalanine. We envisioned that a ligand interacting
with the tyrosine would increase its selectivity for PDE9.
We anticipated that increasing selectivity through a polar
interaction would also improve our druglike properties.
The second PDE9A residue, Phe-441, which is an alanine
in PDE1C, could also provide an opportunity for an aro-
matic π-stacking interaction. Utilizing this knowledge, we
docked19 the 2400 compounds from our virtual library into
the PDE9A binding pocket with the pyrazolopyrimidine
core partially constrained as in the X-ray structure of 2.
These docked compounds were scored using an in-house
scoring method.20 We chose to synthesize the top ranking
compounds (500) based on scoring and visual inspections
by focusing on the compound’s ability to form a hydrogen
bond interaction with Tyr-424 and a π-stacking interaction
with Phe-441.
Synthetically, the dicyano electrophile 3 could be coupled
with a variety of hydrazines to afford pyrazoles. The remain-
ing cyano group could be oxidized in the presence of a large
number of functional groups with 30% peroxide in aqueous
ammonium hydroxide. Cyclization with esters proceeded in
dioxane with sodium hydride as the base followed by addition
of n-BuOH and heating to provide the desired pyrazolopyr-
imidines. This parallel protocolproduced 50% of the designed
molecules in high purity for testing.
The power of the chemical enablement allowed us to
consider making 89000 compounds based on the commer-
cially and internally available hydrazines and esters. With the
large potential pool of compounds to synthesize we developed
a strategy for filtering our monomers and potential products.
By filtering the hydrazine and ester monomers sets to MW <
220 and by excluding monomers with additional hydrogen
bond donors, we expected to synthesize molecules with desir-
able CNS penetration. As part of our monomer selection, we
also included all parallel enabled monomers, i.e., monomers
that could provide an additional handle for future library
chemistry and rapid SAR optimization. Many enabled mono-
mers contain a basic nitrogen that would aid in solubility. We
anticipated that the insertion of a basic center was likely to
invite additional issues such as PgP efflux, 2D6 metabolism,16
and interaction with HERG.17 However, the parallel chem-
istry enabled monomers would allow us to rapidly explore the
optimal basic pKa range to alleviate these problems.
From thiseffort weidentifieda key ester precursorshownin
6-8. The library containing this pyrrolidine benzyl group
afforded compounds with improved selectivity over PDE1C
(Table 1).
Table 1. PDE9A and PDE1C IC50 in nM
Filtering the monomers provided 3500 potential com-
pounds for synthesis. Selecting the final compounds based
on the physicochemical properties of ClogP < 3, TPSA < 110,
and MW < 430 provided 2400 compounds for consideration.
The pyrrolidine group of 8 improved selectivity 15-fold
compared to the initial lead 2. We acquired an X-ray crystal
structure of 7 bound in PDE9A. The design concept of

7948 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24
Verhoest et al.
The major advantage of selecting enabled monomers was
evidenced by the identification of 8. We were able to rapidly
synthesize 8 on scale and remove the benzyl group via hydro-
genation. This allowed us to utilize parallel chemistry for the
second time to conduct a reductive amination library with a
variety of aromatic and heteroaromatic aldehydes. This
monomer flexibility permitted us to rapidly explore the SAR
and to modulate the physicochemical properties of our in-
hibitors. The heteroaryls provided compounds with decreased
lipophilicity and reduced the pKa of the pyrrolidine.
The reductive amination library exemplified in Table 2
showed that this series has inherent PDE9A selectivity. On
average each compound was at least 10-fold selective. The
SAR of the aromatic ring was very flexible in that it tolerated
5- and 6-member heterocycles along with substitutents that
were electron withdrawing or donating. In addition, 10-member
bicyclic aromatic groups, such as 18, provided excellent potency
(2 nM) and selectivity >30Â. This is potentially a result of
a better edge-on interaction with Phe-456 of PDE9A.
Compound 18 exhibited the desired in vitro profile, trigger-
ing separation of the enantiomers. On the basis of the X-ray
structure (Figure 3), we determined which enantiomer would
possess the majority of the activity. We were able to separate 8
via preparative chiral chromatography. Following Scheme 2
we were able to prepare both enantiomers of 18.
Figure 3. PDE9A inhibitor 7 bound in PDE9A making key inter-
actions to Tyr-424 (PDB 3JSW).
Scheme 2a
Compound 19 (PF-4181366, Figure 4) showed PDE9A
affinity of 1.8 nM with >25Â selectivity. The less active enantio-
mer displayed a PDE9 IC50 of 50 nM. Compound 19 displayed
the in vitro profile we were targeting for a tool compound to build
our confidence in rationale for Alzheimer’s disease.
a (a) H2, Pd/C, EtOH (57%); (b) NaCNBH3, parallel chemistry.
Table 2. PDE9A and PDE1C IC50 in nM
Figure 4. Active enantiomer 19.
Compound 19 was dosed in rodents to understand the
potential for CNS penetration. By design, this compound
had a weakly basic pKa, minimal MW, and a limited
number of hydrogen bond donors which resulted in excel-
lent central exposure with no PgP liability. The brain to
plasma ratio in a rat was 1.4 and the free drug exposure
equaled the CSF exposure, suggesting 19 has no barriers to
CNS penetration.
Compound 19 elevated cGMP in the striatum of mice in a
dose responsive fashion (Figure 5). A 145% elevation oc-
curred at 3.2 mg/kg, and at 10 mg/kg a 215% elevation was
hydrogen bonding to Tyr-424 had come to fruition. The
pyrrolidine nitrogen of 7 hydrogen bonds to Tyr-424
through a water molecule, representing a slight deviation
from our projected docking model. In addition, the benzyl
group has multiple conformations and shows density in the
X-ray structure consistent with an edge on interaction with
Phe-456 and a π-stack interaction with Phe-441, just as
designed. The methyl group on the pyrrolidine partially
filled the lipophilic pocket that was occupied by the benzyl
group of 2.
Figure 5. Dose responsive increase in striatal cGMP with 19.

Letter
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 24 7949
observed. As predicted from experience with other PDE
targets, we saw a physiological effect at ∼5Â the PDE9 IC50
in the CSF.
Pathway during Hippocampal Synaptic Plasticity. J. Neurosci.
2005, 25 (29), 6887–6897.
(9) Prickaerts, J.; Steinbusch, H. W.; Smits, J. F. M.; De Vente, J.
Possible Role of Nitric Oxide-Cyclic GMP Pathway in Object
Recognition Memeory: Effects of 7-Nitroindazole and Zaprinast.
Eur. J. Pharmacol. 1997, 337 (2/3), 125–136.
(10) Hopkins, A. L.; Groom, C. R.; Alex, A. Ligand Efficiency: A
Useful Metric for Lead Selection. Drug Discovery Today 2004, 9
(10), 430–431.
(11) Ertl, P.; Rohde, B.; Selzer, P. Fast Calculation of Molecular Polar
Surface Area as a Sum of Fragment-Based Contributions and Its
Applications to the Prediction of Drug Transport Properties.
J. Med. Chem. 2000, 43 (20), 3714–3717.
(12) Walker, D. K.; Ackland, M. J.; James, G. C.; Muirhead, G. J.;
Rance, D. J.; Wastall, P.; Wright, P. A. Phamacokinetics and
Metabolism of Sildenafil in Mouse, Rat, Rabbit, Dog and Man.
Xenobiotica 1999, 29 (3), 297–310.
(13) Hughes, J. D.; Blagg, J.; Price, D. A.; Bailey, S.; DeCrescenzo,
G. A.; Devraj, R. V.; Ellsworth, E.; Fobian, Y. M.; Gibbs, M. E.;
Gilles, R. W.; Greene, N.; Huang, E.; Krieger-Burke, T.; Loesel, J.;
Wager, T.; Whiteley, L.; Zhang, Y. Physiochemical Drug Proper-
ties Associated with in Vivo Toxicological Outcomes. Bioorg. Med.
Chem. Let. 2008, 18 (17), 4872–4875.
(14) Govier, F.; Potempa, A.; Kaufman, J.; Denne, J.; Kovalenko, P.;
Ahuja, S. A Multicenter, Randomized, Double-Blind, Crossover
Study of Patient Preference for Tadalafil 20 mg or Sildenafil Citrate
50 mg during Initiation of Treatment for Erectile Dysfunction.
Clin. Ther. 2003, 25 (11), 2709–2723.
(15) Summerfield, S. G.; Read, K.; Begley, D. J.; Obradovic, T.;
Hidalgo, I. J.; Coggon, S.; Lewis, A. V.; Porter, R. A.; Jeffrey, P.
Central Nervous System Drug Deposition: The Relationship
between in Situ Brain Permeability and Brain Free Fraction.
J. Pharmacol. Exp. Ther. 2007, 322 (1), 205–213.
Overall, 19 exhibited all the attributes of a pharmacolo-
gical tool that could be used to build confidence in the
biological rationale of the target. Delivery of this com-
pound validated our lead optimization strategy which was
built upon knowledge management, prospective design,
and chemical enablement. Future publications will detail
the advancement of this series and identification of a PDE9i
clinical candidate.
Acknowledgment. We acknowledge the Pfizer ATG group
and the analytical chemistry group for their efforts.
Supporting Information Available: Chemistry, biology, and
pharmacokinetics experimental descriptions and data. This
material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) Menniti, F. S.; Faraci, S. W.; Schmidt, C. J. Phosphodiesterases in
the CNS: Targets for Drug Development. Nat. Rev. Drug Discovery
2006, 5 (8), 660–670.
(2) 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.
(3) Chappie, T. A.; Humphrey, J. M.; Allen, M. P.; Estep, K. G.; Fox,
C. B.; Lebel, L. A.; Liras, S.; Marr, E. S.; Menniti, F. S.; Pandit, J.;
Schmidt, C. J.; Tu, M.; Williams, R. D.; Yang, F. V. Discovery of a
Series of 6,7-Dimethoxy-4-pyrrolidylquinazoline PDE10A Inhibi-
tors. J. Med. Chem. 2007, 50 (2), 182–185.
(4) 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. Discovery of a Novel Class of
PDE10A Inhibitors and Identification of Clinical Candidate
2-[4-(1-Methyl-4-pyridin-4-yl-1H-pyrazol-3-yl)phenoxyl methyl]-
quinoline (PF-2545920) for the Treatment of Schizophrenia.
J. Med. Chem. 2009, 52 (16), 5188–5196.
(16) De Groot, M. J.; Ackland, M. J.; Horne, V. A.; Alex, A. A.;
Jones, B. C. Novel Approach to Predicting P450-Mediated
Drug Metabolism: Development of a Combined Protein and
Pharmacophore Model for CYP2D6. J. Med. Chem. 1999, 42 (9),
1515–1524.
(17) Murphy, S. T.; Case, H. L.; Ellsworth, E.; Hagen, S.; Huband, M.;
Joannides, T.; Limberakis, C.; Marotti, K. R.; Ottolini, A. M.;
Rauckhorst, M.; Starr, J.; Stier, M.; Taylor, C.; Zhu, T.; Blaser, A.;
Denny, W. A.; Lu, G.; Small, J.; Rivault, F. The Synthesis and
Biological Evaluation of Novel Series of Nitrile-Containing Fluor-
oquinolones as Antibacterial Agents. Bioorg. Med. Chem. Lett.
2007, 17 (8), 2150–2155.
(18) 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.
(19) Gehlhaar1, D. K.; Verkhivker1, G. M.; Rejto1, P. A.; Sherman, C.
J.; Fogel, D. R.; Fogel, L. J.; Freer, S. T. Molecular Recognition of
the Inhibitor AG-1343 by HIV-1 Protease: Conformationally
Flexible Docking by Evolutionary Programming. Chem. Biol.
1995, 2, 317–324.
(20) Marrone1, T. J.; Luty, B. A.; Rose, P. W. Discovering High-
Affinity Ligands from the Computationally Predicted Structures
and Affinities of Small Molecules Bound to a Target: A Virtual
Screening Approach. Perspect. Drug Discovery Des. 2000, 20, 209–230.
(5) Andreeva, S. G.; Dikkes, P.; Epstein, P. M.; Rosenberg, P. A.
Expression of cGMP-Specific Phosphodiesterase 9A mRNA in the
Rat Brain. J. Neurosci. 2001, 21 (22), 9068–9076.
(6) Fisher, D. A.; Smith, J. F.; Pillar, J. S.; St. Denis, S. H.; Cheng, J. B.
Isolation and Characterization of PDE9A, a Novel Human cGMP-
Specific Phosphodiesterase. J. Biol. Chem. 1998, 25 (273), 15559–
15564.
(7) Van Der Staay, F. J.; Rutten, K.; Barfacker, L.; DeVry, J.; Erb, C.;
Schroder, H. H.; Hendrix, M. The Novel Selective PDE9 Inhibitor
BAY 73-6691 Improves Learning and Memory in Rodents. Neu-
ropharmacology 2008, 55 (5), 908–918.
(8) Puzzo, D.; Vitolo, O.; Fabrizio, T.; Jacob, J. P.; Palmeri, A.;
Arancio, O. Amyloid-beta Peptide Inhibits Activation of the
Nitric Oxide/cGMP/cAMP Responsive Element-Binding Protein