Multiparameter optimization in CNS drug discovery: Design of pyrimido[4,5-d]azepines as potent 5-hydroxytryptamine 2C (5-HT2C) receptor agonists with exquisite functional selectivity over 5-HT2A and 5-HT2B receptors

Article abstract of DOI:10.1021/jm5003292

A series of 4-substituted pyrimido[4,5-d]azepines that are potent, selective 5-HT_2C receptor partial agonists is described. A rational medicinal chemistry design strategy to deliver CNS penetration coupled with SAR-based optimization of selectivity and agonist potency provided compounds with the desired balance of preclinical properties. Lead compounds 17 (PF-4479745) and 18 (PF-4522654) displayed robust pharmacology in a preclinical canine model of stress urinary incontinence (SUI) and no measurable functional agonism at the key selectivity targets 5-HT_2A and 5-HT_2B in relevant tissue-based assay systems. Utilizing recent advances in the structural biology of GPCRs, homology modeling has been carried out to rationalize binding and agonist efficacy of these compounds.

Full text of DOI:10.1021/jm5003292

Article
pubs.acs.org/jmc
Multiparameter Optimization in CNS Drug Discovery: Design of
Pyrimido[4,5‑d]azepines as Potent 5‑Hydroxytryptamine 2C (5-HT2C)
Receptor Agonists with Exquisite Functional Selectivity over 5‑HT2A
and 5‑HT2B Receptors
R. Ian Storer,*,†,# Paul E. Brennan,†,∞ Alan D. Brown,†,# Peter J. Bungay,§,# Kelly M. Conlon,‡
Matthew S. Corbett,∥ Robert P. DePianta,∥ Paul V. Fish,†,× Alexander Heifetz,⊥ Danny K. H. Ho,†
Alan S. Jessiman,† Gordon McMurray,‡,# Cesar Augusto F. de Oliveira,∥ Lee R. Roberts,† James A. Root,‡
Veerabahu Shanmugasundaram,∥ Michael J. Shapiro,∥ Melanie Skerten,† Dominique Westbrook,‡
Simon Wheeler,† Gavin A. Whitlock,† and John Wright‡
†Discovery Chemistry, ‡Discovery Biology, and §Pharmacokinetics, Dynamics and Metabolism, Sandwich Laboratories, Pfizer Global
Research and Development, Ramsgate Road, Sandwich, Kent CT13 9NJ, United Kingdom
∥Pfizer Worldwide Research and Development, Eastern Point Road, Groton, Connecticut 06340, United States
⊥Evotec (UK) Ltd., 114 Innovation Drive, Milton Park, Abingdon, Oxfordshire, OX14 4RZ, United Kingdom
S
* Supporting Information
ABSTRACT: A series of 4-substituted pyrimido[4,5-d]-
azepines that are potent, selective 5-HT2C receptor partial
agonists is described. A rational medicinal chemistry design
strategy to deliver CNS penetration coupled with SAR-based
optimization of selectivity and agonist potency provided
compounds with the desired balance of preclinical properties.
Lead compounds 17 (PF-4479745) and 18 (PF-4522654)
displayed robust pharmacology in a preclinical canine model of
stress urinary incontinence (SUI) and no measurable func-
tional agonism at the key selectivity targets 5-HT2A and 5-
HT2B in relevant tissue-based assay systems. Utilizing recent
advances in the structural biology of GPCRs, homology
modeling has been carried out to rationalize binding and agonist efficacy of these compounds.
A recent industry-wide search for potent and selective 5-
HT2C agonists led to the discovery of lorcaserin (2) (APD-356)
which gained FDA approval in 2012 for the treatment of
obesity (Figure 1).4 At the time of carrying out the studies
described herein, vabicaserin (3) (SCA-136) was also in clinical
trials for the treatment of schizophrenia.5 In addition, multiple
other small molecule 5-HT2C agonists have been reported to be
INTRODUCTION
■
Sertonin, 5-hydroxytryptamine (5-HT) (1), is the endogenous
agonist of at least 14 receptor subtypes, classified into seven
families, 5-HT1−7. The 5-HT2 class has three members 2A, 2B,
and 2C. Although 5-HT2A and 5-HT2B receptors are known to
also be expressed peripherally, the expression of 5-HT2C
receptors is believed to be restricted to the central nervous
system (CNS). The 5-HT2C receptor has been viewed as an
attractive drug target for many years with potential application
for the treatment of a number of medical conditions including
obesity, psychiatric disorders, sexual dysfunction, and urinary
incontinence.1 More recently, selectivity over agonism of 5-
HT2A and 5-HT2B receptor subtypes has become a major
imperative; 5-HT2A agonists are known to cause hallucinations
and drive adverse cardiovascular (CV) effects,2 while 5-HT2B
agonists have been associated with chronic irreversible cardiac
valvulopathy and pulmonary hypertension, as illustrated by the
market withdrawal of the unselective serotonergic agonist Fen-
Phen in 1997.3
Figure 1. Selected 5-HT2C agonists.
Received: March 1, 2014
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
in early clinical development or undergoing preclinical
evaluation.6
Previously, Pfizer has disclosed several selective 5-HT2C
receptor agonists,7 including compounds containing a
pyrimido[4,5-d]azepine template that led to the selection of
compound 4 as a first generation lead from this series (Figure
2).8 Although compound 4 delivered favorable levels of efficacy
and amines (5−11) led to improvements in both potency at 5-
HT2C and selectivity over both 5-HT2A and 5-HT2B (Table 1).
Interestingly, these selectivity improvements were also broadly
coupled with an unexpected decrease in Emax of agonism
efficacy at 5-HT2C. Additionally, compounds containing
aliphatic 4-amino substituents (8−11), were generally subject
to P-glycoprotein (P-gp) mediated efflux as measured using an
in vitro MDCK cell line transfected with P-gp.11 As efflux by P-
gp at the blood−brain barrier has been well characterized as a
mechanism that restricts entry of drugs into the CNS, it was
recognized that in vivo functional efficacy of compounds acting
as P-gp substrates may be limited by impairment of CNS
penetration.12 The in vitro MDCK-MDR1 cultured cell
monolayer assay has been demonstrated to be able to
distinguish P-gp substrates from non-P-gp substrates by
measuring efflux ratio (ER), whereby compounds with ER >
2.5 are considered to be significant substrates of P-gp.11,13
Compound 7 (R = MeO, non-P-gp substrate with low Emax
of 22%) and compound 8 (R = MeHN, P-gp substrate with
moderate Emax of 40%) exhibited the requisite levels of subtype
selectivity of >100-fold over 5-HT2B and presented different
levels of agonist Emax. As a result, both 7 and 8 were tested in an
established human efficacy correlated preclinical in vivo canine
model of stress urinary incontinence (SUI). This peak urethral
pressure (PUP) model was used to measure dose-dependent
increases in urethral tone relative to prototype lead compound
4 (Figure 3a).10 Refer to the Supporting Information for a
detailed description of the PUP model and its translation to
human disease. This dog model was used to inform on in vivo
efficacy for the 5-HT2C mechanism and as a translatable efficacy
surrogate for a range of diseases including SUI, obesity, erectile
dysfunction, and psychotherapeutic disorders.6h
Figure 2. Compound 4 showed measurable 5-HT2B agonism.8a
and safety in preclinical studies, a weak signal for 5-HT2B
agonism was observed in in vitro human colon tissue studies at
high concentration.9
Considering potential safety risk implications of chronic
dosing of a compound with even weak 5-HT2B agonism, further
medicinal chemistry effort was focused toward design of a CNS
penetrant 5-HT2C agonist candidate molecule with improved
selectivity over 5-HT2B. At the same time it was necessary to
retain good preclinical in vivo efficacy determined using a
canine model of stress urinary incontinence (SUI) for clinical
dose-prediction.10 This article describes the compound design
rationale and results of these studies.
RESULTS AND DISCUSSION
■
Designing Increased Subtype Selectivity: 4-Substi-
tuted Pyrimidines. Initial work focused on the incorporation
of small heteroatom-containing groups at the synthetically
enabled 4-position of the pyrimidine to target directional polar
electrostatic interactions. On the basis of previous knowledge of
structure−activity relationships (SARs) for related templates, it
was postulated that such modifications would retain agonist
activity at 5-HT2C and lead to enhanced receptor subtype
selectivity. In accordance with the hypothesis, the introduction
of small heteroatom linked substituents such as alkoxyethers
Compound 7 did not exhibit increases in PUP to a degree
viewed as biologically meaningful in comparison to PUP
increases elicited at clinically relevant exposures of duloxetine
and reboxetine, despite being freely CNS penetrant, as assessed
by measuring unbound concentration in plasma and
cerebrospinal fluid (CSF/free plasma ratio of 0.6, suggesting
Table 1. 4-Substituted Pyrimidine Derivatives
5-HT2C
EC50, nM (Emax
5-HT2C
binding Ki, nM
5-HT2B
EC50, nM (Emax
5-HT2A
EC50, nM (Emax
MDCK-MDR1
ab
,
ac
,
ad
,
e
compd
R
log D
)
)
)
ER
4
H
0.2
190 {164, 224} (75%), n = 19 160 {135, 209},
>10000 (41%), n = 14 3400 (51%), n = 12
1.2
n = 19
f
f
f
f
5
6
7
HO
H2N
CH3O
−0.1 1460 {1204, 1685} (12%),
n = 4
NT
NT
NT
NT
NT
3.0
f
f
0.1
>10000 (20%), n = 4
958 {458, 6483},
n = 2
NT
1.0
51 {29, 68} (22%), n = 6
36 {24, 53}, n = 5
>10000 (<10%), n = 4 >10000 (<10%),
n = 12
1.1
8
CH3HN
0.2
0.7
0.8
0.5
39 {26, 60} (40%), n = 13
>10000 (<20%), n = 3
31 {21, 45}, n = 8
29 {9, 87}, n = 3
233 {25, 255}, n = 3
20 {11, 34}, n = 6
>10000 (<10%), n = 4 <10%, n = 12
10.0
5.0
f
f
f
f
9
CH3CH2HN
(CH3)2N
NT
>10000 (<10%), n = 3 NT
247 (66%), n = 4 NT
NT
10
11
154 {70, 339} (56%), n = 3
33 {28, 39} (92%), n = 3
2.5
O(CH2CH2)2N
3.5
a
b
EC50 and Ki determinations are the geometric mean, and 95% confidence intervals are indicated in braces. Recombinant, stably expressed human
5-HT2C CHO K1 cell line, fluorescence imaging plate reader (FLIPR) assay.8b Data generated using a 5-HT2B-expressing human colon tissue
c
preparation. Data generated using a 5-HT2A-expressing dog femoral artery tissue preparation.8a Efflux ratio (ER) determined by ratio of apparent
d
e
permeability in basolateral to apical/apical to basolateral directions MDCK-MDR1 BA/AB. Substrates of P-gp are defined by ER ≥ 2.5.11 NT = not
f
tested.
B
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. (a) Effects of compounds 4, 7, and 8 on peak urethral pressure (PUP) as a percentage change relative to control in female anesthetized
dogs (n = 4, 2, and 4 for compounds 4, 7, and 8 respectively; administered iv to steady state by constant infusion; for compound 8 PUP data
represents 4 of 7 animals investigated, three “nonresponding” animals were excluded; for explanation see Supporting Information): compound 4
CSF/free plasma (0.5); compound 7 CSF/free plasma (0.6); compound 8 CSF/free plasma (0.1). Because of the reproducibility of urethral
profilometry within individual dogs, small changes (>5% change in PUP) have been routinely shown to be statistically significant, although such a
small change in urethral tone would be considered biologically irrelevant. As such a biologically meaningful response is considered of greater
relevance, in this regard experiments have been carried out using the clinical agents duloxetine and S,S-reboxetine to investigate the change in dog
PUP associated with the free plasma concentration which elicits a significant reduction in stress incontinence episodes (and associated increase in
urethral pressure) in controlled clinical trials. These studies have shown that a PUP increase of >20% in the dog correlates with clinically relevant
exposures and efficacy in humans for both compounds.10,14 (b) Recombinant, stably expressed human 5-HT2C CHO K1 cell line, FLIPR agonist
dose−response curves illustrating differences in efficacy Emax
.
sufficient CNS exposure). The lack of relative effect was likely
due to low functional agonism (Emax = 22%, Figure 3b),
rendering this compound a weak partial 5-HT2C agonist in vivo;
it is plausible that this compound could act as a functional
antagonist in the presence of a full agonist in vivo, although this
has not been investigated. Compound 8 exhibited greater
agonist efficacy (Emax = 40%, Figure 3b) and was able to elicit
an increase in PUP but only at high plasma concentrations and
also exhibited appreciable variability between animals, with
some individuals not responding to compound treatment.
Compound 8 was also a P-gp substrate (ER = 10) and likely to
have restricted CNS penetration (CSF/free plasma ratio of
0.1), requiring higher plasma concentrations to achieve
sufficient occupancy of 5-HT2C receptors in the CNS. Although
these investigations halted the progression of both compounds
7 and 8, they provided key information to guide future design
objectives by defining an acceptable target efficacy Emax of >40%
in the FLIPR assay and confirmed the need for ER of <2.5 in
the MDCK-MDR1 assay.
Table 2. Comparison of Physicochemical Properties with
Guidelines for CNS Permeability16
property
CNS physicochem guide compd 7 compd 8
molecular weight (Da)
log P
<450
1−3
7.5−10.5
<90 Å
<5
269
2.6
8.9
47
4
268
1.7
8.9
50
4
pKa (neutral or basic)
TPSA (Å2)
sum of N + O atoms
H-bond acceptors
H-bond donors
MPO desirability17
<6
4
4
<2
1
2
a
4−6
5.6
5.1
a
CNS MPO 0−6 with 6 being most desirable. MPO 4−6 consistent
with passive CNS permeability.
hydrogen bond acceptor (HBA) distances of ∼2.5 and ∼4.6 Å
have been described as recognition features. As shown in Figure
4, compound 8 has HBA distances of 2.4, 4.1, and 4.6 Å,
suggesting that this pattern of hydrogen bonding could be of
importance for P-gp binding.18
Designing Selective Agonists with Improved CNS
Exposure. Following conclusions from the in vivo SUI studies,
design efforts concentrated on improving CNS penetration for
analogues based on the highly selective efficacious agonist lead
compound 8. Multiple design guidelines and models for
predicting CNS penetration have been proposed in the
literature, mostly focusing on physicochemical properties that
relate to permeability across the blood−brain barrier (BBB).15
Table 2 compares an overview of suggested physicochemical
space for CNS penetration against compounds 7 and 8 and
illustrates that these compounds generally fall within good
property space for CNS permeability. It was therefore
hypothesized that the low levels of CNS penetration observed
for 8 were mostly driven by P-gp mediated efflux as opposed to
low passive permeability (PAMPA 10 × 10−6 cm/s), as
supported by the high efflux ratio noted in the MDCK-MDR1
assay. In an effort to design molecules with reduced propensity
for P-gp efflux, literature pharmacophore models for P-gp were
considered. In particular, the importance of intramolecular
It was further hypothesized that the additional electron
density conferred onto ring N-1 by the π-electron donation
from the C-4 NHMe group would increase the HBA ability of
N-1. It is believed that this led to an increased affinity for P-gp
for compound 8 relative to the freely CNS penetrant
unsubstituted prototype compound 4. These molecular dipole
effects were further supported by measured basic pKa values for
the pyrimidine system (Figure 5) which highlight a more basic
pyrimidine pKa for compound 8 (pKa2 of 5.7) compared to
compound 4 (pKa2 of 2.2). For this analysis, pKa1 refers to the
most basic position (azepine NH) and pKa2 refers to the second
most basic pKa for the compound (pyrimidine). These pKa
values were measured via pH titration analysis.
In addition to testing synthesized compounds in the MDCK-
MDR1 assay, in silico models of MDCK-MDR1 ER (cMDR
BA/AB: further details given in Supporting Information) and
cpKa (ACD/Labs) were used prospectively to assess designs in
advance of synthesis.7a,19 This proved to be a powerful tool to
guide medicinal chemistry design and helped ensure that the
C
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
majority of new targets synthesized had minimal P-gp mediated
efflux (Figure 6).
The two positions neighboring N-1 were also considered for
design variation (R1, R2, and R3, Table 3) to decrease P-gp
recognition of the molecules. Two strategies were imple-
mented: (i) introduction of neighboring electron withdrawing
fluorine atoms to reduce electron density on N-1 and (ii) steric
eclipsing of N-1 to disrupt interaction with P-gp (Figure 6). On
the basis of predicted cpKa and cMDR values from computa-
tional models, several compounds that were expected to exhibit
reduced P-gp efflux (targeting cMDR < 3) were synthesized
and profiled (Table 3).
Pleasingly, all compounds (12−18) showed lower P-gp ER
in the MDCK-MDR1 assay (ER of 2.3−4.0) relative to
compound 8 (Table 3), commensurate with our in silico
predictions and increasing the probability of achieving good in
vivo CNS penetration. In addition, all compounds possessed
high affinity for 5-HT2C with Ki values ranging from 9 to 101
nM and excellent levels of functional selectivity over 5-HT2B.
However, dramatic effects on 5-HT2C agonist efficacy profiles
were observed with small structural changes effectively causing
interconversion from agonism to antagonism. This was most
clearly demonstrated by the enantiomeric methyl pairs 12
(agonist)/13 (antagonist) and R-enantiomer 16 (antagonist)/
S-enantiomer 17 (agonist). At the time of this work, these
differences in functional pharmacology were not predicted or
understood and have been the focus of further structural
modeling, as described below.
Figure 4. Compound 8: distances between hydrogen bond acceptors
(HBA), modeled using Macromodel with OPLS-2005 force field in
conjunction with GB/SA solvation model.
Figure 5. Introduction of a 4-amino substituent increases electron
density on ring N-1 as illustrated by measured second pKa values.
From these designs two compounds achieved suitable target
profiles to trigger further assessment. Compound 17 (EC50
=
Figure 6. Use of cMDR and ACD/cpKa models to prospectively assess medicinal chemistry designs.
D
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 3. Potency and Selectivity Profiles for 4-Methylaminopyrimidine Derivatives
5-HT2C
5-HT2C
5-HT2B
abd
, ,
cMDR
(cpKa)
MDR1
ER
ab
,
ac
,
compd
R1, R2
H, H
R3
log D
EC50 nM (Emax
)
Ki, nM (binding)
EC50, nM (Emax)
8
H
H
H
H
H
0.2
0.8
0.8
1.1
0.8
0.7
0.7
0.8
39 {26, 60}, n = 13, (40%)
112 {95, 134}, n = 4, (71%)
>10000, n = 4, (<10%)
>10000, n = 7, (<10%)
>10000, n = 3, (<10%)
>10000, n = 3, (<10%)
21 {12, 36}, n = 8, (67%)
16 {11, 22}, n = 8, (58%)
31 {21, 45}, n = 8
38 {30, 48}, n = 6
30 {16, 57}, n = 6
50 {48, 53}, n = 3
102 {70, 146}, n = 3’
80 {25, 255}, n = 3
15 {11, 22}, n = 8
8 {6, 11}, n = 10
>10000 (<10%), n = 4
∼1000 (26%), n = 4
>10000 (<10%), n = 3
>10000 (<10%), n = 4
>10000 (<10%), n = 3
6.2 (6.4)
3.8 (5.9)
3.8 (5.9)
2.7 (6.1)
2.8 (6.1)
2.8 (6.5)
2.8 (6.5)
2.8 (3.6)
10.0
4.0
4.0
2.4
2.6
2.3
2.7
2.4
12
13
14
15
16
17
18
CH3, H ent.1
CH3, H ent.2
CH3, CH3
cPr
e
f
H, H
CH3 (R)
NT
e
H, H
CH3 (S)
>10000 (<10%), n = 6
>10000 (<10%), n = 6
F, F
H
a
b
EC50 and Ki determinations are the geometric mean, and 95% confidence intervals are in braces. 5-HT2B EC50 generated using human colon tissue
c
d
e
assay (n = 4−8). Tested up to 10 μM. Tested up to 30 μM. Absolute stereochemistry determined by VCD (vibrational circular dichroism) (see
Supporting Information for full details). NT = not tested.
f
Figure 7. Compounds 17 and 18 showed no agonist effects in either (a) 5-HT2A or (b) 5-HT2B tissue preparations (n = 6−14).
10 nM, Emax = 67%) and compound 18 (EC50 = 17 nM, Emax
=
found to have binding affinities of IC50 > 2.0 μM (>100-fold
selectivity vs 5-HT2C EC50) except for the 5-HT2A (360 nM,
antagonist), 5-HT2B (67 nM, antagonist), 5-HT1A (500 nM,
functional activity not tested), and 5-HT6 (280 nM, functional
activity not tested) receptors. Compound 18 was found to
exhibit binding affinities of IC50 > 2.0 μM except for the 5-
HT2A (119 nM, antagonist), 5-HT2B (110 nM, antagonist), 5-
HT1B (610 nM, functional activity not tested), and β2 (1400
nM, antagonist) receptors. The safety risk posed by this level of
activity in these assays was considered to be low based on
previous in-house experience, providing no immediate barrier
to further progression of these compounds toward the clinic.
Pharmacokinetic data were generated in vivo in rat for
compound 17 and in dog for compounds 17 and 18 (plasma
exposure during PUP studies). Following single intravenous
administration of compound 17 to rat, systemic clearance was
high (104 mL min−1 kg−1, approximately equal to hepatic blood
flow), consistent with the low in vitro metabolic stability in rat
liver microsomes observed for this compound. However,
metabolic stability of compounds 17 and 18 in human in
vitro systems was higher than in rat (Table 4). For compound
17, scaling of intrinsic clearance using the well-stirred model of
hepatic clearance21 using data from HLM (5 mL min−1 kg−1)
and human hepatocytes (4.7 mL min−1 kg−1) was found to be
consistent with low hepatic CL in human (approximately 25%
of hepatic blood flow). For compound 18, scaling from HLM
(2.8 mL min−1 kg−1) and human hepatocytes (6.7 mL min−1
kg−1) was found to be consistent with low/moderate hepatic
CL in human (approximately 13−30% of hepatic blood flow).
%) both exhibit promising levels of in vitro potency, agonism,
and efflux ratios of <3 in the MDCK-MDR1 assay. Importantly,
these two compounds showed no measurable functional effects
in tissue based assays for 5-HT2A and 5-HT2B (Figure 7).
As a result, both compounds (17 and 18) were advanced to
further in vitro and in vivo pharmacokinetic, pharmacology, and
safety studies.
Synthetic Chemistry. Compounds were prepared by
reaction of amidines 20, derived from the corresponding
nitriles 19, with ketoesters 21 (Scheme 1).20 The resulting
pyrimidones 22 were activated for reaction at the C-4 position
by formation of the corresponding C-4 triflates 23. These
versatile reactive intermediates enabled late stage variations of
the 4-position via nucleophilic displacement with amines to
yield 24. Parallel compound synthesis was used to fully explore
SAR at this position. Upon nucleophilic displacement, Boc
deprotection was carried out by treatment with HCl in dioxane
to yield target compounds 25. This route was applied to the
syntheses of compound analogues and to the multigram scale
synthesis of lead compounds 17 and 18.
Lead Compound Profiles. Compounds 17 and 18
exhibited metabolic stability in both human liver microsomes
(HLM) and human hepatocytes consistent with low predicted
clearance in humans, weak CYP450 enzyme inhibition, and
moderate passive permeability (Table 4).
Compounds 17 and 18 were tested for off-target
pharmacology against 90 receptors, enzymes, and ion channels
at CEREP (Bioprint) and within Pfizer. Compound 17 was
E
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 1. Synthesis of 4-Methylaminopyrimidine
a
Compounds 17 and 18
Figure 8. Relative effects of 17 and 18 versus lorcaserin (2) and
vabicaserin (3) on peak urethral pressure (PUP) as a percent change
relative to control in female anaesthetized dogs (n = 2, 3, 5, and 4 for
compounds 17, 18, lorcaserin, and vabicaserin, respectively; iv
infusion): compound 17 CSF/free plasma (0.25, n = 2). PUP increase
of >20% correlated with human efficacy.10
previously with clinically relevant mechanisms.10,14a,b On the
basis of this, compound 17 was predicted to exhibit a free
plasma Ceff of ∼70 nM and 18 a free plasma Ceff of ∼40 nM,
comparing favorably with leading literature compounds such as
lorcaserin (2) (free Ceff = 200 nM) and vabicaserin (3) (free
Ceff = 300 nM) as determined in this model.
The in vitro metabolic clearances in human of compounds
17 and 18 were scaled to a predicted systemic plasma clearance
from intrinsic clearance in HLM. These estimates were then
applied with a physiologically based pharmacokinetic (PBPK)
model to simulate predicted human pharmacokinetic profiles
using GastroPlus software (see Supporting Information for
details). Results of simulations suggested that daily oral doses in
human of approximately 160 mg of 17 and 25 mg of 18 would
be required to achieve Ceff at trough as defined above in the
PUP model.
a
Reagents and conditions. (i) For compound 17: (a) commercial
amidine sourced; (b) NaOMe, MeOH, 23 °C, 16 h, 74%; (c) Tf2O,
pyridine, CH2Cl2, 0 °C, 30 min; (d) CH3NH2, CH3CN, 23 °C, 75 h,
87% over two steps; (e) HCl, dioxane, 23 °C, 1 h, 95%. (ii) For
compound 18: (a) NH4Cl, (CH3)3Al, toluene, 85 °C, 17 h; (b)
NaOMe, MeOH, 23 °C, 39 h, 20% over two steps; (c) Tf2O, pyridine,
CH2Cl2, 0−23 °C, 2 h; (d) CH3NH2, CH3CN, 23 °C, 20 h, 60% over
two steps; (e) HCl, dioxane, 23 °C, 3 h, 75%.
Table 4. In Vitro Profiles: Physicochemistry and ADME
property
compd 17
compd 18
Physicochemical Properties
molecular weight (Da)
clogP
282
2.9
0.7
8.6
50
304
2.1
0.8
8.7
50
log D at pH 7.4
pKa1 (most basic)
TPSA (Å2)
Homology Modeling To Rationalize Affinity and
Agonist Efficacy of Key Compounds. Modeling has been
undertaken to rationalize the structural basis for 5-HT2C affinity
and agonism for this series. Published site directed mutagenesis
(SDM) data and crystal structures of the β2-adrenergic
PAMPA (10−6 cm/s)
11
11
ADME Profiles
HLM, Clint (μL min−1 mg−1
RLM, Clint (μL min−1 mg−1
DLM, Clint (μL min−1 mg−1
)
)
16
9
329
26
114
n/a
9.5
62
receptor were used to create in silico models of active and
)
active
hhep, Clint (μL min−1 (106 cells)−1
)
3.4
47
inactive receptor conformations, referred to as 5-HT2C
and
5-HT2Cinactive, respectively (Figure 9a).22 The model of 5-
hppb (%)
active
HT2C
was based on the crystal structure of the active
dppb (%)
51
61
conformation of the β2-adrenergic receptor (β2AR, PDB entry
3SN6), and the model of 5-HT2Cinactive was based on the crystal
structure of inactive conformation of the same receptor (β2AR,
PDB entry 2RH1). Flexible-docking was then applied to predict
the binding modes of compounds with 5-HT2C. Full details of
the SDM data, modeling, and docking procedures are outlined
in the Supporting Information.
CYP1A2, inh 10 μM (%)
CYP2C9, inh 10 μM (%)
CYP2D6, inh 10 μM (%)
CYP3A4, inh 10 μM (%)
<10
<10
17
<10
<10
<10
<10
11
An evaluation of compounds 17 and 18 in the in vivo canine
PUP model of SUI demonstrated a robust dose-dependent
effect (Figure 8); although it is clear that the maximum
efficacious effect has not been defined for compounds 17 and
18, it appears plausible that both compounds would behave in a
similar manner to the literature 5-HT2C agonists lorcaserin and
vabicaserin. Data from the canine PUP model were used to
predict the efficacious plasma concentrations (Ceff) for
treatment of SUI in human based on plasma exposures eliciting
a minimum 20% increase in dog PUP, this being defined
The proposed binding mode of prototype lead compound 4
in 5-HT2Cinactive is shown in Figure 9b. In this inactive state, it is
suggested that compound 4 can generate three notable
interactions with the key residues of TM3 and TM6. The
basic azepine of compound 4, known to be essential for series
activity, forms a salt bridge with D1343.32 and a cation−π
interaction with the W1303.28 residue. An additional aromatic
interaction is also predicted between F3286.52 and compound 4.
inactive
The binding site of 5-HT2C
is proposed to be shallower
F
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 9. Predicted binding poses of compounds 4 and 18 in 5-HT2C. Binding site residues identified from SDM data are highlighted in yellow, and
inactive
active
proline residues are colored in green: (a) docking pose of compound 4 in 7TMD of 5-HT2C
(ligand colored green) and 5-HT2C
(ligand
colored orange); (b) magnified view of docking pose of compound 4 in 5-HT2Cinactive; (c) magnified view of docking pose of compound 4 in 5-
HT2Cactive; (d) docking pose of compound 18 in 5-HT2C
.
active
Figure 10. Predicted binding poses of compounds 12 (a), 13 (b), 17 (c), and 16 (d) in 5-HT2Cactive. Key binding site residues are highlighted in
yellow (b, d). The position of the clashes between protein residues and the ligands are marked by a red dotted line.
active
than that of 5-HT2C
because of residues from TM3 and
entering deeper into the binding site and compensating via
interactions with various residues, including W3246.48, the
reported transmission switch residue.23,24 Compound 4 is
believed to interact with W3246.48, positioning the benzyl group
in front of P3266.50 (Figure 9a) and restraining the intracellular
TM6 forming stabilizing interhelical interactions in the 5-
23
inactive
HT2C
binding site. It is proposed that these interhelical
interactions are broken in the active conformation of the 5-
HT2C receptor, which is stabilized instead by agonist molecules
G
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
half of TM6 in the active conformation. The binding mode is
shown in Figure 9c, suggesting that compound 4 generates six
interactions with the key residues of TM3 and TM6. In
particular, D1343.32 forms a salt bridge with the key basic
azepine amine of compound 4. An additional nonclassical
hydrogen bond is predicted between S1383.36, and the CH2
benzylic linker of the ligand and residue F3276.55 forms a face-
to-face π−π stack with the aromatic pyrimido[4,5-d]azepine
core of compound 4. The benzyl group of compound 4 also
forms both face-to-face and edge-to-face π−π interactions with
W3246.48 and F3286.52, respectively. This proposed ability of
enantiomers compounds 13 and 16 behave as antagonists at the
5-HT2C receptor.
CONCLUSIONS
■
In summary, a series of 4-methylaminopyrimido[4,5-d]azepines
has been identified as potent and highly subtype selective 5-
HT2C receptor partial agonists. Preferred examples 17, 2-
benzyl-9-(S)-methyl-4-(methylamino)-5,6,8,9-tetrahydro-7H-
pyrimido[4,5-d]azepine (PF-4479745) and 18, 2-[difluoro-
(phenyl)methyl]-N-methyl-6,7,8,9-tetrahydro-5H-pyrimido-
[4,5-d]azepin-4-amine (PF-4522654) exhibited no measurable
agonism of either 5-HT2A or 5-HT2B receptor, offering
preclinical differentiation from both the previous candidate in
this series and notable 5-HT2C agonists from the literature.
These compounds also display robust in vivo pharmacology in
a canine peak urethral pressure model of SUI, reflecting their
degree of 5-HT2C agonist efficacy and CNS penetration. They
also demonstrate potential for attractive human pharmacoki-
netic profiles together with in vitro safety properties. Homology
modeling based on recent advances in GPCR structural biology
has been carried out to propose binding modes for ration-
alization of the affinity and agonist efficacy of key compounds
at the 5-HT2C receptor. In order to make these compounds
readily accessible for use by the broader scientific community,
compound 4 (PF-3246799) and compound 17 (PF-4479745)
have been made commercially available.30
compound 4 to interact simultaneously with both TM3 and
active
TM6 in 5-HT2C
bridges these two TM helices, increasing
active
the overall stability of 5-HT2C
and promoting activation.
This mode of action has also been observed for other agonists
of class A GPCRs.25 Furthermore, the importance of these four
key residues (D1343.32, S1383.36, W3246.48, and F3286.52) for
binding of 4-substituted pyrimido[4,5-d]azepines is supported
26
by the published SDM data, where mutations of D1343.32
,
27
S1383.36
,
conserved W3376.48 in 5-HT2B (W3246.48 in 5-
HT2C),28 and F3416.52 in 5-HT2B or F3406.52 in 5-HT2A
(F3286.52 in 5-HT2C) have all shown dramatic effects on ligand
affinity and potency.29 In summary, this model suggests that
compound 4 functions as an agonist, as it can form a larger
active
number of interactions with 5-HT2C
relative to 5-
active
HT2Cinactive, six with 5-HT2C
(Figure 9c) compared to
three with 5-HT2C inactive (Figure 9b).
These hypotheses have been further extended to lead
compound 18, which is believed to form similar interactions
to compound 4 but also an additional hydrogen bond between
the C4-NHCH3 substituent and Q3316.55 and a halogen bond
between difluorine linker and S1383.36 (Figure 9d). These
additional interactions with TM3 and TM6 offer a possible
explanation of why the agonist EC50 is improved for compound
18 compared to prototype lead compound 4.
EXPERIMENTAL SECTION
■
The syntheses of lead compounds and intermediates 17, 18, 21a, 22a,
22b, 23a, 22b, 24a, and 24b are described below as representatives of
the series. Detailed experimental procedures for the syntheses of all
compounds, characterization data, procedures for the biological assays,
and molecular modeling are in the Supporting Information.
1
Materials and Methods. H nuclear magnetic resonance (NMR)
spectra were in all cases consistent with the proposed structures.
Characteristic chemical shifts (δ) are given in parts per million
downfield from tetramethylsilane using conventional abbreviations for
designation of major peaks: s, singlet; d, doublet; t, triplet; m,
multiplet; br, broad. The mass spectra (m/z) were recorded using
either electrospray ionization (ESI) or atmospheric pressure chemical
ionization (APCI). The following abbreviations have been used for
common solvents: CDCl3, deuterochloroform; D6-DMSO, deuter-
odimethylsulfoxide; CD3OD, deuteromethanol; THF, tetrahydrofuran.
All solvents were reagent grade and, when necessary, were purified and
dried by standard methods. Concentration of solutions after reactions
and extractions involved the use of a rotary evaporator operating at a
reduced pressure of ∼20 Torr. Organic solutions were dried over
anhydrous magnesium sulfate or anhydrous sodium sulfate.
As described above, dramatic effects on 5-HT2C agonist
efficacy profiles were observed with small structural changes
effectively causing conversion from agonism to antagonism.
This was most clearly demonstrated by the enantiomeric
methyl pairs 12 (agonist)/13 (antagonist) and R-enantiomer
16 (antagonist)/S-enantiomer 17 (agonist). The proposed
active
binding mode of compounds 12, 13, 16, and 17 in 5-HT2C
is shown in Figure 10. It was predicted that compounds 12 and
inactive
17 can be accommodated by both 5-HT2Cactive and 5-HT2C
conformers of the receptor but preferentially bind to 5-
active
HT2C
because of a larger number of favorable interactions
(analogous to compounds 4 and 18). On the other hand, the
HPLC, unless indicated otherwise, was performed by one of the
following methods.
opposite enantiomers, compounds 13 and 16, respectively, can
inactive
be accommodated by just the 5-HT2C
and not by 5-HT2C
protein conformer
A: column, Sunfire C18 4.6 mm × 50 mm; mobile phase A, 0.05%
formic acid in water; mobile phase B, 0.05% formic acid in
acetonitrile.
B: column, Xterra 4.6 mm × 50 mm; mobile phase A, 0.05%
ammonia in water; mobile phase B, 0.05% ammonia in
acetonitrile.
C: column, Luna C8 4.6 mm × 50 mm; mobile phase A, 10 mM
ammonium acetate in water; mobile phase B:, 10 mM
ammonium acetate in acetonitrile.
D: column, C18 4.6 mm × 50 mm; mobile phase A, 0.1% formic
acid in water; mobile phase B, 0.1% formic acid in acetonitrile.
E: column, XBridge C18 4.6 mm × 150 mm; mobile phase A,
0.1% TFA in water; mobile phase B, 0.1% TFA in acetonitrile.
active
because of severe clashes of ligand
atoms with the binding site residues (distance of <2 Å). In the
case of compound 13 (Figure 10b) a severe clash is predicted
between the chiral methyl and S1383.36 (this clash does not
exist in its enantiomer compound 12). In the case of compound
16 (Figure 10d) the methylated azepine core has a different
predicted conformation compared to the core of compounds
17 and 18 (other potential conformations of the azepine core
of compound 16 would lose the essential salt bridge between
the basic nitrogen center and D1343.32). This conformation of
compound 16 causes the molecule to prefer a flipped binding
conformation that causes the NHCH3 substituent to clashes
with V3547.39. These modeling predictions rationalize why
compounds 12 and 17 act as agonists and their respective
2-Benzyl-9-(S)-methyl-4-(methylamino)-5,6,8,9-tetrahydro-
7H-pyrimido[4,5-d]azepine (17). A solution of HCl in dioxane
(4M, 4 mL) was added to a stirred solution of tert-butyl 2-benzyl-9-
H
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
methyl-4-(methylamino)-5,6,8,9-tetrahydro-7H-pyrimido[4,5-d]-
azepine-7-carboxylate (24a) (495 mg, 1.3 mmol) in CH2Cl2 (10 mL)
at room temperature. The reaction mixture was stirred at room
temperature for 1 h. The solvent was removed under reduced pressure
to yield the crude product (17) as a solid HCl salt (410 mg, 95%
yield).
A small amount of impurity (∼5%) was evident by NMR, so the
material was purified by preparative HPLC chromatography (see
Supporting Information for full details of this preparative separation).
This method provided good separation to yield the desired product,
177 mg, 99.2% ee of colorless oil free base. Absolute stereochemical
assignment was obtained via application of vibrational circular
dichroism (see Supporting Information for full details).
(2.1 g, 12 mmol). The reaction mixture was stirred at room
temperature for 16 h, then quenched by addition of water (3 mL)
and evaporated to dryness. The residue was partitioned between water
(20 mL) and EtOAc (3 × 20 mL) and then separated. The aqueous
layer was washed with EtOAc (3 × 10 mL). The combined organics
were dried over magnesium sulfate, filtered, and concentrated under
reduced pressure to yield the crude product as a solid. The material
was purified by flash column chromatography, eluting with CH2Cl2/
EtOAc, 100:0 to 70:30, to yield the product as an inseparable mixture
of regioisomers (∼10:1) as a yellow foam (22a) (2.74 g, 74% yield).
The mixture of isomers was taken on to the subsequent reaction
without further separation.
Major Isomer. 1H NMR (400 MHz, CDCl3) δ: 1.26 (d, J = 7.3 Hz,
3H), 1.43 (s, 9H), 2.68−2.91 (m, 1H), 3.01−3.20 (m, 2H), 3.31−3.43
(m, 1H), 3.45−3.81 (m, 3H), 3.93 (s, 2H), 7.22−7.39 (m, 5H). HPLC
tR = 2.94 min, ELSD >95% purity as a mixture (∼10:1) of two isomers.
LRMS (APCI) m/z 370 [M + H]+.
1H NMR (600 MHz, CDCl3) δ: 1.32 (d, J = 7.1 Hz, 3H), 2.48−2.55
(m, 1H), 2.56−2.65 (m, 1H), 2.30−2.39 (m, 2H), 2.97 (s, 3H), 2.98−
3.06 (m, 2H), 3.07−3.15 (m, 1H), 4.05 (s, 2H), 4.06 (br s, 1H), 7.16−
7.22 (m, 1H), 7.27−7.34 (m, 2H), 7.45−7.49 (m, 2H). 13C NMR
(150 MHz, CDCl3) δ: 16.2, 27.3, 28.6, 44.7, 45.9, 47.6, 52.4, 111.9,
126.0, 128.1, 129.4, 139.5, 161.0, 165.6, 169.1. HPLC (2 min), tR =
0.42 min, ELSD >95% purity. HPLC (12 min), tR = 5.40 min, >95%
purity. LRMS (ESI, APCI) m/z 283 [M + H]+. HRMS (EI) m/z: calcd
tert-Butyl 2-[Difluoro(phenyl)methyl]-4-oxo-3,4,5,6,8,9-hex-
ahydro-7H-pyrimido[4,5-d]azepine-7-carboxylate (22b). Step 1:
Amidine Formation. Ammonium chloride (1.6 g, 29 mmol) was added
to a solution of 2,2-difluorophenylacetonitrile (4.4 g, 29 mmol) in
toluene (25 mL). The resulting suspension was cooled to 0 °C under
nitrogen. Trimethylaluminum (2 M in toluene, 14.5 mL, 29 mmol)
was added dropwise with stirring at 0 °C over 10 min. The reaction
mixture was warmed to room temperature over 15 min, then heated to
85 °C with stirring for 17 h. The mixture was then cooled to room
temperature and then poured into a slurry of silica gel (140 g) in
CH2Cl2 (150 mL). The slurry was stirred for 5 min and then filtered,
washing thoroughly with methanol. The filtrate was concentrated
under reduced pressure to provide 2,2-difluorophenylamidine (20b) as
a dark brown residue (4.2 g) which was used in the subsequent
reaction without further purification.
20
for C17H23N4 [M + H]+ 283.1917; found 283.1916. [α]D −6.07°.
2-[Difluoro(phenyl)methyl]-N-methyl-6,7,8,9-tetrahydro-5H-
pyrimido[4,5-d]azepin-4-amine (18). A solution of HCl in EtOAc
(2.0 N, 25 mL, 50 mmol) was added to a solution of tert-butyl 2-
[difluoro(phenyl)methyl]-4-(methylamino)-5,6,8,9-tetrahydro-7H-
pyrimido[4,5-d]azepine-7-carboxylate (24b) (1.4 g, 3.4 mmol) in
CH2Cl2 (100 mL). The mixture was stirred at room temperature for 3
h. The reaction was quenched by addition of saturated aqueous
sodium bicarbonate and extracted into CH2Cl2. The combined
organics were dried over magnesium sulfate, filtered, and concentrated
under reduced pressure to provide the crude product as a brown gum.
Purification of the crude material by flash column chromatography,
eluting with CH2Cl2/MeOH/NH3, 30:1:0.2 to 10:1:0.2, afforded the
product (18) as a white solid (750 mg, 2.5 mmol, 75% yield, 99.2%
Step 2: Pyrimidinone Formation. A freshly prepared solution of
NaOMe (sodium 2.2 g, 96 mmol in methanol 25 mL) was added to a
solution of 1-tert-butyl 4-ethyl 5-oxoazepine-1,4-dicarboxylate (21b)
(see patent WO2006029154, example 1a, p 34) (8.7 g, 30.7 mmol)
and 2,2-difluorophenylamidine (20b) (4.2 g, 29 mmol) in MeOH
(135 mL) at 0 °C. The mixture was allowed to warm to room
temperature with stirring for 39 h. The reaction was quenched with
water (10 mL), acidified with HCl (2 N), and extracted with EtOAc (3
× 50 mL). The combined organics were dried over magnesium sulfate,
filtered, and concentrated under reduced pressure to provide the crude
product as an orange gum (11 g). The residue was purified by flash
column chromatography, eluting with CH2Cl2/EtOAc, 100:0 to 70:30,
to afford the title compound (22b) as a pale yellow solid (2.2 g, 20%
yield over two steps).
1
ee). H NMR (300 MHz, CDCl3) δ: 2.23 (br s, 1H), 2.50−2.68 (m,
2H), 2.88−3.12 (m, 9H), 4.78 (br s, 1H), 7.32−7.47 (m, 3H), 7.68−
7.79 (m, 2H). HPLC (6 min) tR = 1.96 min, ELSD 95% purity. HPLC
(25 min) tR = 9.47 min, UV >95% purity. LRMS (ESI, APCI) m/z 305
[M + H]+. HRMS (EI) m/z: [M + H]+ calcd for C16H19F2N4
305.1572; found 305.1568.
1-tert-Butyl-4-ethyl-6-methyl-5-oxoazepine-1,4-dicarboxy-
late (21a). Solutions of boron trifluoride diethyletherate (3.9 mL, 31
mmol) in 5 mL of Et2O and ethyl diazoacetate (4.17 g, 37) in 5 mL of
Et2O were added dropwise to a solution of the 4-piperidinone (6.0 g,
30 mmol) in diethyl ether (20 mL) at −50 °C, maintaining an internal
temperature below −30 °C. The reaction mixture was stirred at −50
°C for 1 h, then warmed to −20 °C for 1 h. Acetone (10 mL) was
added to quench any remaining ethyl diazoacetate, and the resulting
mixture was stirred for 10 min. Water and solid potassium carbonate
were added to bring the pH to 8.5. Then the solution was diluted with
EtOAc (300 mL). The aqueous layer was removed and the organics
were washed with saturated aqueous sodium bicarbonate (2 × 100
mL), brine (100 mL), then dried over MgSO4, filtered, and
concentrated under reduced pressure to provide the crude product
as a solid (8.0 g, 100% yield). Analysis (NMR) showed this to be a
mixture (∼10:1) of regioisomers in favor of the desired product. This
material was taken on for pyrimidone formation (22a) without further
purification.
1H NMR (400 MHz, CDCl3) δ: 1.42 (s, 9H), 2.83−2.92 (m, 2H),
2.92−2.99 (m, 2H), 3.48−3.61 (m, 4H), 7.22−7.51 (m, 3H), 7.59−
7.63 (m, 2H). HPLC (6 min) tR = 3.06 min, ELSD >95% purity.
t
LRMS (ESI) m/z 391 [M − H]−; m/z 336 [M + H − Bu]+.
tert-Butyl 2-Benzyl-9-(rac)-methyl-4-{[(trifluoromethyl)-
sulfonyl]oxy}-5,6,8,9-tetrahydro-7H-pyrimido[4,5-d]azepine-7-
carboxylate (23a). Triflic anhydride (1.54 g, 0.89 mL, 5.5 mmol) was
added to a solution of the tert-butyl 2-benzyl-9-methyl-4-oxo-
3,4,5,6,8,9-hexahydro-7H-pyrimido[4,5-d]azepine-7-carboxylate (22a)
(1.83 g, 5.0 mmol) in CH2Cl2 at 0 °C under N2. The resulting solution
was stirred for 30 min at 0 °C, then partitioned between 5% aqueous
citric acid and EtOAc. The organic layer was washed with saturated
aqueous sodium bicarbonate, dried over magnesium sulfate, filtered,
and concentrated under reduced pressure to yield an orange gum of
23a (2.5 g, 100% yield). This material was taken on to the next
Major Isomer. 1H NMR (400 MHz, CDCl3) δ: 1.14 (d, J = 7.2 Hz,
3H), 1.20−1.35 (m, 5H), 1.42 (s, 9H), 1.98−2.08 (m, 2H), 2.70−3.06
(m, 2H), 3.71−3.85 (m, 2H), 4.10−4.29 (m, 2H). HPLC (6 min) tR =
3.03 min, ELSD >95% purity as a mixture (∼10:1) of two isomers.
LRMS (ESI) m/z 322 [M + Na]+, m/z 298 [M − H]−.
tert-Butyl 2-Benzyl-9-(rac)-methyl-4-oxo-3,4,5,6,8,9-hexahy-
dro-7H-pyrimido[4,5-d]azepine-7-carboxylate (22a). A solution
of NaOMe (3.8M, 7.9 mL, 30 mmol) was added to a solution of keto
ester, 1-tert-butyl 4-ethyl-6-methyl-5-oxoazepine-1,4-dicarboxylate
(21a) ∼10:1 of regioisomers (3.0 g, 10 mmol) in MeOH (40 mL)
at room temperature followed by addition of the 2-phenylacetamidine
1
reaction without further purification. H NMR (400 MHz, CDCl3) δ:
1.30 (d, J = 7.3 Hz, 3H), 1.47 (s, 9H), 2.78−3.04 (m, 2H), 3.26−3.41
(m, 1H), 3.43−3.76 (m, 4H), 4.19 (s, 2H), 7.12−7.47 (m, 5H). HPLC
(6 min) tR 4.09 min, ELSD >95% purity. LRMS (ESI, APCI) m/z 446
t
[M + H − Bu]+.
tert-Butyl 2-[Difluoro(phenyl)methyl]-4-{[(trifluoromethyl)-
sulfonyl]oxy}-5,6,8,9-tetrahydro-7H-pyrimido[4,5-d]azepine-7-
carboxylate (23b). Triflic anhydride (1.3 mL, 7.5 mmol) was added
dropwise to a solution of the tert-butyl 2-[difluoro(phenyl)methyl]-4-
I
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
oxo-3,4,5,6,8,9-hexahydro-7H-pyrimido[4,5-d]azepine-7-carboxylate
(22b) (2.2 g, 5.6 mmol) and pyridine (0.8 mL, 9.9 mmol) in CH2Cl2
(25 mL) at 0 °C. The resulting solution was allowed to warm to room
temperature and stirred for 2 h. The reaction mixture was quenched by
addition of water, then partitioned between 5% aqueous citric acid and
EtOAc. The organic layer was washed with saturated aqueous sodium
bicarbonate, dried over magnesium sulfate, filtered, and concentrated
under reduced pressure to yield the crude product (23b) as a yellow
gum. This material was used in the subsequent reaction without
further purification. HPLC (6 min) tR = 3.86 min, ELSD >95% purity.
AUTHOR INFORMATION
Corresponding Author
*Phone: +44 (0)1304 641854. E-mail: ian.storer@pfizer.com.
■
Present Addresses
#R.I.S., A.D.B., P.J.B., G.M.: Pfizer Neusentis, The Portway
Building, Granta Park, Cambridge, CB21 6GS, United
Kingdom.
∞P.E.B.: Structural Genomics Consortium, Target Discovery
Institute, Nuffield Department of Medicine, Oxford, OX3 7BN,
United Kingdom.
t
LRMS (ESI, APCI) m/z 468 [M + H − Bu]+.
tert-Butyl 2-Benzyl-9-(S)-methyl-4-(methylamino)-5,6,8,9-
tetrahydro-7H-pyrimido[4,5-d]azepine-7-carboxylate (24a). A
solution of methylamine (2 M, 12.4 mL, 25 mmol) in THF was added
dropwise to a stirred solution of triflate tert-butyl 2-benzyl-9-methyl-4-
{[(trifluoromethyl)sulfonyl]oxy}-5,6,8,9-tetrahydro-7H-pyrimido[4,5-
d]azepine-7-carboxylate (23a) (2.48 g, 4.95 mmol) in acetonitrile (150
mL) at room temperature. The resulting mixture was stirred at room
temperature for 75 h. The mixture was concentrated under reduced
pressure, and then CH2Cl2 (50 mL) and a saturated aqueous sodium
bicarbonate (50 mL) were added. The mixture was separated and the
aqueous layer washed with CH2Cl2 (3 × 50 mL). The combined
organics were dried over magnesium sulfate, filtered, and concentrated
under reduced pressure to yield the crude product as an orange solid.
Purification by flash column chromatography, eluting with pentane/
EtOAc, 100:0 to 70:30, gave the desired product (24a) as a yellow
gum that crystallized on standing (1.64 g, 87% yield). NB: The minor
methyl regioisomer that had been carried through previous steps was
separated from the desired product via this column chromatography.
The desired racemic product was separated into individual
enantiomers by preparative chiral column chromatography (see
Supporting Information for full details of this preparative separation).
This separation yielded 480 mg, 99.7% ee of the intermediate en route
to the less potent enantiomer (16) after deprotection and yielded 495
mg, 99.7% ee of the intermediate en route to the more potent
enantiomer (17) after deprotection.
×P.V.F.: UCL School of Pharmacy, 29-39 Brunswick Square,
London, WC1N 1AX, United Kingdom.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We acknowledge the Primary Pharmacology Group for
screening data, Hau Gao for the cMDR efflux model, and
support from Laure Hitzel and Juliette Love in the Separation
and Analytical Services group for chiral analysis and preparative
chromatography. We also thank Jianmin Sun for coordinating
the collection of HRMS data on selected compounds.
ABBREVIATIONS USED
■
ADME, absorption, distribution, metabolism, and excretion;
Boc, tert-butyloxycarbonyl; BBB, blood−brain barrier; Ceff,
efficacious concentration; CHO, Chinese hamster ovary; CL,
clearance; CSF, cerebrospinal fluid; CNS, central nervous
system; CV, cardiovascular; CYP, cytochrome P450; DCM,
dichloromethane; DLM, dog liver microsome; Dof, dofetilide;
DOI, 2,5-dimethoxy-4-iodoamphetamine; dppb, dog plasma
protein binding; ER, efflux ratio; FDA, Federal Drug Agency;
FLIPR, fluorescence imaging plate reader; hep, hepatocytes;
hERG, human ether-a-go-go-related gene; hppb, human plasma
protein binding; HLM, human liver microsome; log D,
partition coefficient between octanol and water at pH 7.4;
MDCK, Madin−Darby canine kidney; MDR1, multidrug
resistance gene 1; PAMPA, parallel artificial membrane
permeability assay; PDB, Protein Data Bank; P-gp, P-
glycoprotein; ppb, plasma protein binding; PUP, peak urethral
pressure; RLM, rat liver microsome; SAR, structure−activity
relationship; SDM, site directed mutagenesis; SUI, stress
urinary incontinence; Tf, triflate (trifluoromethanesulfonate);
TM, transmembrane; TPSA, topological polar surface area;
VCD, vibrational circular dichroism
1H NMR (400 MHz, CDCl3) δ: 1.23 (d, J = 7.2 Hz, 3H), 1.42 (s,
9H), 2.40−2.51 (m, 1H), 2.97 (d, J = 6.0 Hz, 3H), 3.12−3.21 (m,
1H), 3.21−3.39 (m, 1H), 3.40−3.51 (m, 1H), 3.58−3.90 (m, 2H),
4.01 (s, 2H), 4.50−4.67 (m, 1H), 7.13−7.20 (m, 1H), 7.21−7.30 (m,
2H), 7.38−7.44 (m, 2H). HPLC (6 min), tR = 1.96 min, ELSD >95%
purity. LRMS (ESI, APCI) m/z 383 [M + H]+.
tert-Butyl 2-[Difluoro(phenyl)methyl]-4-(methylamino)-
5,6,8,9-tetrahydro-7H-pyrimido[4,5-d]azepine-7-carboxylate
(24b). A solution of methylamine (2.0 M, 22 mL, 44 mmol) was
added to a solution of tert-butyl 2-[difluoro(phenyl)methyl]-4-
{[(trifluoromethyl)sulfonyl]oxy}-5,6,8,9-tetrahydro-7H-pyrimido[4,5-
d]azepine-7-carboxylate (23b) (3.1 g, 5.6 mmol) in acetonitrile (120
mL) at room temperature with stirring. The mixture was stirred at
room temperature for 20 h. The solvent was evaporated under reduced
pressure to provide the crude title compound. Purification of the crude
material by flash column chromatography, eluting with pentane/
EtOAc, 4:1 to 2:1, afforded the product (24b) as a white solid (1.4 g,
60% yield). 1H NMR (300 MHz, CDCl3): 1.43 (9H, br s), 2.56−2.65
(m, 2H), 3.00 (d, J = 5.9 Hz, 3H) 3.06−3.10 (m, 2H), 3.50−3.80 (m,
4H), 4.78 (br s, 1H), 7.35−7.45 (m, 3H), 7.70−7.80 (m, 2H). HPLC
(6 min) tR = 3.19 min, ELSD >95% purity. LRMS (ESI, APCI) m/z
405 [M + H]+.
REFERENCES
■
(1) (a) Lee, J.; Jung, M. E.; Lee, J. 5-HT2C receptor modulators: a
patent survey. Expert Opin. Ther. Pat. 2010, 20, 1429−1455.
(b) Smith, B. M.; Thomsen, W. J.; Grottick, A. J. The potential use
of selective 5-HT2C agonists in treating obesity. Expert Opin. Invest.
Drugs 2006, 15, 257−266. (c) Rosenzweig-Lipson, S.; Dunlop, J.;
Marquis, K. L. 5-HT2C receptor agonists as an innovative approach for
psychiatric disorders. Drug News Perspect. 2007, 20, 565−571.
(d) Wacker, D. A.; Miller, K. J. Agonists of the serotonin 5-HT2C
receptor: preclinical and clinical progression in multiple diseases. Curr.
Opin. Drug Discovery Dev. 2008, 11, 438−445. (e) Mbaki, Y.; Ramage,
A. G. Investigation of the role of 5-HT2 receptor subtypes in the
control of the bladder and the urethra in the anesthetized female rat.
Br. J. Pharmacol. 2008, 155, 343−356.
ASSOCIATED CONTENT
■
S
* Supporting Information
Synthetic procedures, characterization, and purity analysis of all
compounds; selected spectra for key compounds; details of the
VCD determination of stereochemical assignment of com-
pounds 16 and 17; detailed descriptions of computational
modeling; procedures for the functional, binding, and tissue
based assays; detailed in vivo methodology. This material is
available free of charge via the Internet at http://pubs.acs.org.
(2) (a) Kaumann, A. J.; Levy, F. O. 5-Hydroxytryptamine receptors
in the human cardiovascular system. Pharmacol. Ther. 2006, 111, 674−
706. (b) Vollenweider, F. X.; Vollenweider-Scherpenhuyzen, M. F. I.;
J
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Babler, A.; Vogel, H.; Hell, D. Psilocybin induces schizophrenia-like
psychosis in humans via a serotonin-2A agonist action. NeuroReport
1998, 9, 3897−3902. (c) Nichols, D. E. Hallucinogens. Pharmacol.
Ther. 2004, 101, 131−181. (d) Villalon, C. M.; Centurion, D.
Cardiovascular responses produced by 5-hydroxytriptamine: a
pharmacological update on the receptors/mechanisms involved and
therapeutic implications. N-S Arch. Pharmacol. 2007, 376, 45−63.
(3) (a) Abenhaim, L.; Moride, Y.; Brenot, F.; Rich, S.; Benichou, J.;
Kurz, X.; Higenbottam, T.; Oakley, C.; Wouters, E. Appetite-
suppressant drugs and the risk of primary pulmonary hypertension.
N. Engl. J. Med. 1996, 335, 609−616. (b) Roth, B. L. Drugs and
valvular heart disease. N. Engl. J. Med. 2007, 356, 6−9.
(4) (a) Fleming, J. W.; McClendon, K. S.; Riche, D. M. New obesity
agents: lorcaserin and phentermine/topiramate. Ann. Pharmacother.
2013, 47, 1007−1016. (b) Smith, B. M.; Smith, J. M.; Tsai, J. H.;
Schultz, J. A.; Gilson, C. A.; Estrada, S. A.; Chen, R. R.; Park, D. M.;
Prieto, E. B.; Gallardo, C. S.; Sengupta, D.; Dosa, P. I.; Covel, J. A.;
Ren, A.; Webb, R. R.; Beeley, N. R. A.; Martin, M.; Morgan, M.;
Espitia, S.; Saldana, H. R.; Bjenning, C.; Whelan, K. T.; Grottick, A. J.;
Menzaghi, F.; Thomsen, W. J. Discovery and structure−activity
relationship of (1R)-8-chloro-2,3,4,5-tetrahydro-1-methyl-1H-3-benza-
zepine (lorcaserin), a selective serotonin 5-HT2C receptor agonist for
the treatment of obesity. J. Med. Chem. 2008, 51, 305−313.
Kimizuka, T.; Kimura, Y.; Hatanaka, K.-i.; Naitou, Y.; Wanibuchi, F.;
Sakamoto, S.; Tsukamoto, S.-i. Synthesis and structure−activity
relationships of a series of substituted 2-(1H-furo[2,3-g]indazol-1-
yl)ethylamine derivatives as 5-HT2C receptor agonists. Bioorg. Med.
Chem. 2008, 16, 1966−1982.
(7) (a) Brennan, P. E.; Whitlock, G. A.; Ho, D. K. H.; Conlon, K.;
McMurray, G. Discovery of a novel azepine series of potent and
selective 5-HT2C agonists as potential treatments for urinary
incontinence. Bioorg. Med. Chem. Lett. 2009, 19, 4999−5003.
(b) Andrews, M. D.; Green, M. P.; Allerton, C. M. N.; Batchelor, D.
V.; Blagg, J.; Brown, A. D.; Gordon, D. W.; McMurray, G.; Millns, D.
J.; Nichols, C. L.; Watson, L. Design and synthesis of piperazinylpyr-
imidinones as novel selective 5-HT2C agonists. Bioorg. Med. Chem. Lett.
2009, 19, 5346−5350. (c) Allerton, C. M. N.; Andrews, M. D.; Blagg,
J.; Ellis, D.; Evrard, E.; Green, M. P.; Liu, K. K. C.; McMurray, G.;
Ralph, M.; Sanderson, V.; Ward, R.; Watson, L. Design and synthesis
of pyridazinone-based 5-HT2C agonists. Bioorg. Med. Chem. Lett. 2009,
19, 5791−5795. (d) Siuciak, J. A.; Chapin, D. S.; McCarthy, S. A.;
Guanowsky, V.; Brown, J.; Chiang, P.; Marala, R.; Patterson, T.;
Seymour, P. A.; Swick, A.; Iredale, P. A. CP-809,101, a selective 5-
HT2C agonist, shows activity in animal models of antipsychotic activity.
Neuropharmacology 2007, 52, 279−290. (e) Kalgutkar, A. S.; Dalvie, D.
K.; Aubrecht, J.; Smith, E. B.; Coffing, S. L.; Cheung, J. R.; Vage, C.;
Lame, M. E.; Chiang, P.; McClure, K. F.; Maurer, T. S.; Coelho, R. V.,
Jr.; Soliman, V. F.; Schildknegt, K. Genotoxicity of 2-(3-chlorobenzy-
loxy)-6-(piperazinyl)pyrazine, a novel 5-hydroxytryptamine2C recep-
tor agonist for the treatment of obesity: role of metabolic activation.
Drug Metab. Dispos. 2007, 35, 848−858. (f) Kalgutkar, A. S.; Bauman,
J. N.; McClure, K. F.; Aubrecht, J.; Cortina, S. R.; Paralkar, J.
Biochemical basis for differences in metabolism-dependent genotox-
icity by two diazinylpiperazine-based 5-HT2C receptor agonists. Bioorg.
Med. Chem. Lett. 2009, 19, 1559−1563. (g) Fish, P. V.; Brown, A. D.;
Evrard, E.; Roberts, L. R. 7-Sulfonamido-3-benzazepines as potent and
selective 5-HT2C receptor agonists: hit-to-lead optimization. Bioorg.
Med. Chem. Lett. 2009, 19, 1871−1875. (h) Liu, K. K. C.; Cornelius,
P.; Patterson, T. A.; Zeng, Y.; Santucci, S.; Tomlinson, E.; Gibbons, C.;
Maurer, T. S.; Marala, R.; Brown, J.; Kong, J. X.; Lee, E.; Werner, W.;
Wenzel, Z.; Vage, C. Design and synthesis of orally-active and selective
azaindane 5-HT2C agonist for the treatment of obesity. Bioorg. Med.
Chem. Lett. 2010, 20, 266−271. (i) Liu, K. K. C.; Lefker, B. A.;
Dombroski, M. A.; Chiang, P.; Cornelius, P.; Patterson, T. A.; Zeng,
Y.; Santucci, S.; Tomlinson, E.; Gibbons, C. P.; Marala, R.; Brown, J.
A.; Kong, J. X.; Lee, E.; Werner, W.; Wenzel, Z.; Giragossian, C.;
Chen, H.; Coffey, S. B. Orally active and brain permeable proline
amides as highly selective 5-HT2C agonists for the treatment of obesity.
Bioorg. Med. Chem. Lett. 2010, 20, 2365−2369.
(8) (a) Andrews, M. D.; Fish, P. V.; Blagg, J.; Brabham, T. K.;
Brennan, P. E.; Bridgeland, A.; Brown, A. D.; Bungay, P. J.; Conlon, K.
M.; Edmunds, N. J.; af Forselles, K.; Gibbons, C. P.; Green, M. P.;
Hanton, G.; Holbrook, M.; Jessiman, A. S.; McIntosh, K.; McMurray,
G.; Nichols, C. L.; Root, J. A.; Storer, R. I.; Sutton, M. R.; Ward, R. V.;
Westbrook, D.; Whitlock, G. A. Pyrimido[4,5-d]azepines as potent and
selective 5-HT2C receptor agonists: design, synthesis, and evaluation of
PF-3246799 as a treatment for urinary incontinence. Bioorg. Med.
Chem. Lett. 2011, 21, 2715−2720. (b) Andrews, M. D.; Blagg, J.;
Brennan, P. E.; Fish, P. V.; Roberts, L. R.; Storer, R. I.; Whitlock, G. A.
Preparation of pyrimido[4,5-d]azepine derivatives as 5-HT2C agonists.
WO2008117169, 2008.
(9) (a) Borman, R. A.; Tilford, N. S.; Harmer, D. W.; Day, N.; Ellis,
E. S.; Sheldrick, R. L. G.; Carey, J.; Coleman, R. A.; Baxter, G. S. 5-
HT2B receptors play a key role in mediating the excitatory effects of 5-
HT in human colon in vitro. Br. J. Pharmacol. 2002, 135, 1144−1151.
(b) Baxter, G. S.; Murphy, O. E.; Blackburn, T. P. Further
characterization of 5-hydroxytryptamine receptors (putative 5-HT2B)
in rat stomach fundus longitudinal muscle. Br. J. Pharmacol. 1994, 112,
323−331. (c) Huang, X.-P.; Setola, V.; Yadav, P. N.; Allen, J. A.;
Rogan, S. C.; Hanson, B. J.; Revankar, C.; Robers, M.; Doucette, C.;
Roth, B. L. Parallel functional activity profiling reveals valvulopath-
(5) Ramamoorthy, P. S.; Beyer, C.; Brennan, J.; Dunlop, J.; Gove, S.;
Grauer, S.; Harrison, B. L.; Lin, Q.; Malberg, J.; Marquis, K.;
Mazandarani, H.; Piesla, M.; Pulicicchio, C.; Rosenzwieg-Lipson, S.;
Sabb, A.-M.; Schechter, L.; Stack, G.; Zhang, J. Discovery of SCA-136,
a novel 5-HT2C agonist, for the treatment of schizophrenia. Abstracts of
Papers, 231st National Meeting of the American Chemical Society,
Atlanta, GA, U.S., March 26−30, 2006; American Chemical Society:
Washington, DC, 2006; MEDI-021.
(6) (a) Monck, N. J. T.; Kennett, G. A. 5-HT2C ligands: recent
progress. Prog. Med. Chem. 2008, 46, 281−390. (b) Stahl, S. M.; Lee-
Zimmerman, C.; Cartwright, S.; Morrissette, D. A. Serotonergic drugs
for depression and beyond. Curr. Drug Targets 2013, 14, 578−585.
(c) Heal, D. J.; Gosden, J.; Smith, S. L. A review of late-stage CNS
drug candidates for the treatment of obesity. Int. J. Obes. 2013, 37,
107−117. (d) Yang, H. Y.; Tae, J.; Seo, Y. W.; Kim, Y. J.; Im, H. Y.;
Choi, G. D.; Cho, H.; Park, W.-K.; Kwon, O. S.; Cho, Y. S.; Ko, M.;
Jang, H.; Lee, J.; Choi, K.; Kim, C.-H.; Lee, J.; Pae, A. N. Novel
pyrimidoazepine analogs as serotonin 5-HT2A and 5-HT2C receptor
ligands for the treatment of obesity. Eur. J. Med. Chem. 2013, 63, 558−
569. (e) Rosenzweig-Lipson, S.; Comery, T. A.; Marquis, K. L.; Gross,
J.; Dunlop, J. 5-HT2C agonists as therapeutics for the treatment of
schizophrenia. Handb. Exp. Pharmacol. 2012, 213, 147−165.
(f) Sargent, B. J.; Henderson, A. J. Targeting 5-HT receptors for the
treatment of obesity. Curr. Opin. Pharmacol. 2011, 11, 52−58.
(g) Cho, S. J.; Jensen, N. H.; Kurome, T.; Kadari, S.; Manzano, M.
L.; Malberg, J. E.; Caldarone, B.; Roth, B. L.; Kozikowski, A. P.
Selective 5-hydroxytryptamine 2C receptor agonists derived from the
lead compound tranylcypromine: identification of drugs with
antidepressant-like action. J. Med. Chem. 2009, 52, 1885−1902.
(h) Bishop, M. J.; Nilsson, B. M. New 5-HT2C receptor agonists.
Expert Opin. Ther. Pat. 2003, 13, 1691−1705. (i) Lacivita, E.;
Leopoldo, M. Selective agents for serotonin2C (5-HT2C) receptor.
Curr. Top. Med. Chem. (Sharjah, United Arab Emirates) 2006, 6, 1927−
1970. (j) Nilsson, B. M. 5-Hydroxytryptamine 2C (5-HT2C) receptor
agonists as potential antiobesity agents. J. Med. Chem. 2006, 49, 4023−
4034. (k) Ahmad, S.; Ngu, K.; Miller, K. J.; Wu, G.; Hung, C.-p.;
Malmstrom, S.; Zhang, G.; O’Tanyi, E.; Keim, W. J.; Cullen, M. J.;
Rohrbach, K. W.; Thomas, M.; Ung, T.; Qu, Q.; Gan, J.; Narayanan,
R.; Pelleymounter, M. A.; Robl, J. A. Tricyclic dihydroquinazolinones
as novel 5-HT2C selective and orally efficacious anti-obesity agents.
Bioorg. Med. Chem. Lett. 2010, 20, 1128−1133. (l) Shimada, I.; Maeno,
K.; Kondoh, Y.; Kaku, H.; Sugasawa, K.; Kimura, Y.; Hatanaka, K.-i.;
Naitou, Y.; Wanibuchi, F.; Sakamoto, S.; Tsukamoto, S.-i. Synthesis
and structure−activity relationships of a series of benzazepine
derivatives as 5-HT2C receptor agonists. Bioorg. Med. Chem. 2008,
16, 3309−3320. (m) Shimada, I.; Maeno, K.; Kazuta, K.-i.; Kubota, H.;
K
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
ogens are potent 5-hydroxytryptamine2B receptor agonists: implica-
tions for drug safety assessment. Mol. Pharmacol. 2009, 76, 710−722.
(10) Conlon, K.; Christy, C.; Westbrook, S.; Whitlock, G.; Roberts,
L.; Stobie, A.; McMurray, G. Pharmacological properties of 2-((R-5-
chloro-4-methoxymethyl-indan-1-yl)-1H-imidazole (PF-3774076), a
novel and selective a1A-adrenergic partial agonist, in in vitro and in
vivo models of urethral function. J. Pharmacol. Exp. Ther. 2009, 330,
892−901.
(11) Feng, B.; Mills, J. B.; Davidson, R. E.; Mireles, R. J.; Janiszewski,
J. S.; Troutman, M. D.; de Morais, S. M. In vitro P-glycoprotein assays
to predict the in vivo interactions of P-glycoprotein with drugs in the
central nervous system. Drug Metab. Dispos. 2008, 36, 268−275.
(12) (a) Schinkel, A. H. P-Glycoprotein, a gatekeeper in the blood−
brain barrier. Adv. Drug Delivery Rev. 1999, 36, 179−194. (b) Begley,
D. J. ABC transporters and the blood−brain barrier. Curr. Pharm. Des.
2004, 10, 1295−1312.
L. G.; Schioth, H. B.; Pierce, K. L.; Premont, R. T.; Lefkowitz, R. J.;
Lefkowitz, R. J.; Shenoy, S. K.; Rosenbaum, D. M. High-resolution
crystal structure of an engineered human β2-adrenergic G protein-
coupled receptor. Science (Washington, DC, U. S.) 2007, 318, 1258−
1265.
(23) (a) Gether, U.; Lin, S.; Ghanouni, P.; Ballesteros, J. A.;
Weinstein, H.; Kobilka, B. K. Agonists induce conformational changes
in transmembrane domains III and VI of the β2 adrenoceptor. EMBO
J. 1997, 16, 6737−6747. (b) Farrens, D. L.; Altenbach, C.; Yang, K.;
Hubbell, W. L.; Khorana, H. G. Requirement of rigid-body motion of
transmembrane helixes for light activation of rhodopsin. Science
(Washington, DC, U. S.) 1996, 274, 768−770.
(24) Trzaskowski, B.; Latek, D.; Yuan, S.; Ghoshdastider, U.;
Debinski, A.; Filipek, S. Action of molecular switches in GPCRs
theoretical and experimental studies. Curr. Med. Chem. 2012, 19,
1090−1109.
(25) Kim, T. H.; Chung, K. Y.; Manglik, A.; Hansen, A. L.; Dror, R.
O.; Mildorf, T. J.; Shaw, D. E.; Kobilka, B. K.; Prosser, R. S. The role of
ligands on the equilibria between functional states of a G protein-
coupled receptor. J. Am. Chem. Soc. 2013, 135, 9465−9474.
(26) Abul Muntasir, H.; Takahashi, J.; Rashid, M.; Ahmed, M.;
Komiyama, T.; Hossain, M.; Kawakami, J.; Nashimoto, M.; Nagatomo,
T. Site-directed mutagenesis of the serotonin 5-hydroxytryptamine 2C
receptor: identification of amino acids responsible for sarpogrelate
binding. Biol. Pharm. Bull. 2006, 29, 1645−1650.
(27) Canal, C. E.; Cordova-Sintjago, T. C.; Villa, N. Y.; Fang, L.-J.;
Booth, R. G. Drug discovery targeting human 5-HT2C receptors:
residues S3.36 and Y7.43 impact ligand-binding pocket structure via
hydrogen bond formation. Eur. J. Pharmacol. 2011, 673, 1−12.
(28) Janoshazi, A.; Deraet, M.; Callebert, J.; Setola, V.; Guenther, S.;
Saubamea, B.; Manivet, P.; Launay, J.-M.; Maroteaux, L. Modified
receptor internalization upon coexpression of 5-HT1B receptor and 5-
HT2B receptors. Mol. Pharmacol. 2007, 71, 1463−1474.
(29) Rashid, M.; Manivet, P.; Nishio, H.; Pratuangdejkul, J.; Rajab,
M.; Ishiguro, M.; Launay, J.-M.; Nagatomo, T. Identification of the
binding sites and selectivity of sarpogrelate, a novel 5-HT2 antagonist,
to human 5-HT2A, 5-HT2B and 5-HT2C receptor subtypes by molecular
modeling. Life Sci. 2003, 73, 193−207.
(30) Compound 4 (PF-3246799, catalog no. PZ0229) and
compound 17 (PF-4479745, catalog no. PZ0032) are available from
Sigma Aldrich.
(13) Doan, K. M. M.; Humphreys, J. E.; Webster, L. O.; Wring, S. A.;
Shampine, L. J.; Serabjit-Singh, C. J.; Adkison, K. K.; Polli, J. W.
Passive permeability and P-glycoprotein-mediated efflux differentiate
central nervous system (CNS) and non-CNS marketed drugs. J.
Pharmacol. Exp. Ther. 2002, 303, 1029−1037.
(14) (a) Klarskov, N.; Scholfield, D.; Soma, K.; Darekar, A.; Mills, I.;
Lose, G. Evaluation of the sensitivity of urethral pressure reflectometry
(UPR) and urethral pressure profilometry (UPP) to detect
pharmacological augmentation of urethral pressure, using [S,S]-
reboxetine. J. Urol. 2008, 179, 521−523. (b) Wakenhut, F.; Allan, G.
A.; Fish, P. V.; Jonathan Fray, M.; Harrison, A. C.; McCoy, R.; Phillips,
S. C.; Stobie, A.; Westbrook, D.; Westbrook, S. L.; Whitlock, G. A. N-
[(3S)-Pyrrolidin-3-yl]benzamides as novel dual serotonin and
noradrenaline reuptake inhibitors: Impact of small structural
modifications on P-gp recognition and CNS penetration. Bioorg.
Med. Chem. Lett. 2009, 19, 5078−5081. (c) Fish, P. V.; Harrison, A.;
Wakenhut, F.; Whitlock, G. A. A case history on the challenges of
central nervous system and dual pharmacology drug discovery. RSC
Drug Discovery Ser. 2011, 4, 267−286. (d) Zinner, N.; Scholfield, D.;
Soma, K.; Darekar, A.; Grant, L.; Mills, I. A phase 2, 8-week, multi-
centre, randomized, double blind, placebo controlled, parallel group
study evaluating the efficacy, tolerability and safety of [S,S]-reboxetine
(PNU-165442G) for stress urinary incontinence. J. Urol. 2008, 179,
569−570.
(15) Van de Waterbeemd, H.; Camenisch, G.; Folkers, G.; Chretien,
J. R.; Raevsky, O. A. Estimation of blood−brain barrier crossing of
drugs using molecular size and shape, and H-bonding descriptors. J.
Drug Targeting 1998, 6, 151−165.
(16) Hitchcock, S. A.; Pennington, L. D. Structure−brain exposure
relationships. J. Med. Chem. 2006, 49, 7559−7583.
(17) Wager, T. T.; Hou, X.; Verhoest, P. R.; Villalobos, A. Moving
beyond rules: the development of a central nervous system
multiparameter optimization (CNS MPO) approach to enable
alignment of druglike properties. ACS Chem. Neurosci. 2010, 1,
435−449.
(18) Penzotti, J. E.; Lamb, M. L.; Evensen, E.; Grootenhuis, P. D. J. A
computational ensemble pharmacophore model for identifying
substrates of p-glycoprotein. J. Med. Chem. 2002, 45, 1737−1740.
(19) Gao, H.; Yao, L.; Mathieu, H. W.; Zhang, Y.; Maurer, T. S.;
Troutman, M. D.; Scott, D. O.; Ruggeri, R. B.; Lin, J. In silico
modeling of nonspecific binding to human liver microsomes. Drug
Metab. Dispos. 2008, 36, 2130−2135.
(20) Moriya, T.; Oki, T.; Yamaguchi, S.; Morosawa, S.; Yokoo, A.
Studies of seven-membered heterocyclic compounds containing
nitrogen. IX. The synthesis of 5-ethoxycarbonyl-1-azacycloheptan-4-
one and its derivatives. Bull. Chem. Soc. Jpn. 1968, 41, 230−231.
(21) Yang, J.; Jamei, M.; Yeo, K. R.; Rostami-Hodjegan, A.; Tucker,
G. T. Misuse of the well-stirred model of hepatic drug clearance. Drug
Metab. Dispos. 2007, 35, 501−502.
(22) Cherezov, V.; Rosenbaum, D. M.; Hanson, M. A.; Rasmussen, S.
G. F.; Thian, F. S.; Kobilka, T. S.; Choi, H.-J.; Kuhn, P.; Weis, W. I.;
Kobilka, B. K.; Stevens, R. C.; Takeda, S.; Kadowaki, S.; Haga, T.;
Takaesu, H.; Mitaku, S.; Fredriksson, R.; Lagerstrom, M. C.; Lundin,
L
dx.doi.org/10.1021/jm5003292 | J. Med. Chem. XXXX, XXX, XXX−XXX