Selective GlyT1 inhibitors: Discovery of [4-(3-fluoro-5- trifluoromethylpyridin-2-yl)piperazin-1-yl][5-methanesulfonyl-2-((S)-2,2, 2-trifluoro-1-methylethoxy)phenyl]methanone (RG1678), a promising novel medicine to treat schizophrenia

Article abstract of DOI:10.1021/jm100210p

The GlyT1 transporter has emerged as a key novel target for the treatment of schizophrenia. Herein, we report on the optimization of the 2-alkoxy-5-methylsulfonebenzoylpiperazine class of GlyT1 inhibitors to improve hERG channel selectivity and brain penetration. This effort culminated in the discovery of compound 10a (RG1678), the first potent and selective GlyT1 inhibitor to have a beneficial effect in schizophrenic patients in a phase II clinical trial.

Full text of DOI:10.1021/jm100210p

J. Med. Chem. 2010, 53, 4603–4614 4603
DOI: 10.1021/jm100210p
Selective GlyT1 Inhibitors: Discovery of [4-(3-Fluoro-5-trifluoromethylpyridin-2-yl)piperazin-1-yl]-
[5-methanesulfonyl-2-((S)-2,2,2-trifluoro-1-methylethoxy)phenyl]methanone (RG1678), a Promising
Novel Medicine To Treat Schizophrenia
Emmanuel Pinard,* Alexander Alanine, Daniela Alberati, Markus Bender, Edilio Borroni, Patrick Bourdeaux,
Virginie Brom, Serge Burner, Holger Fischer, Dominik Hainzl, Remy Halm, Nicole Hauser, Synese Jolidon, Judith Lengyel,
Hans-Peter Marty, Thierry Meyer, Jean-Luc Moreau, Roland Mory, Robert Narquizian, Mathias Nettekoven,
Roger D. Norcross, Bernd Puellmann, Philipp Schmid, Sebastien Schmitt, Henri Stalder, Roger Wermuth,
Joseph G. Wettstein, and Daniel Zimmerli
F. Hoffmann-La Roche Ltd., Pharmaceutical Division, CH-4070 Basel, Switzerland
Received February 17, 2010
The GlyT1 transporter has emerged as a key novel target for the treatment of schizophrenia. Herein, we
report on the optimization of the 2-alkoxy-5-methylsulfonebenzoylpiperazine class of GlyT1 inhibitors
to improve hERG channel selectivity and brain penetration. This effort culminated in the discovery of
compound 10a (RG1678), the first potent and selective GlyT1 inhibitor to have a beneficial effect in
schizophrenic patients in a phase II clinical trial.
Introduction
restoring NMDA receptor activity may represent a promising
novel avenue for the management of this disorder. As glycine
is an obligatory coagonist at the NMDA receptor complex,3
one approach to enhance NMDA receptor activity is to
elevate extracellular levels of glycine in the brain through
selective inhibition of glycine uptake mediated by the glycine
transporter-1 (GlyT1) which is coexpressed with the NMDA
receptor.4 Strong support for this approach comes from
clinical studies where glycine and D-serine, which are both
coagonists at the glycine site of the NMDA receptor and
sarcosine, a prototypical weak GlyT1 inhibitor, improved
positive, negative, and cognitive symptoms in schizophrenic
patients when added to conventional therapy.5 As sarcosine
has a poor pharmacokinetic profile and therefore is not an
optimal therapeutic agent, considerable effort has been fo-
cused on the development of improved selective GlyT1
inhibitors.6 The first reported set of examples were designed
by chemical modification of sarcosine and included com-
pounds 1,7 2,8 and 3 (Figure 1).9 However, it was soon realized
that many sarcosine derivatives suffered from a range of
undesirable effects including hypoactivity, ataxia, and reduced
respiratory activity.10,11 Consequently, these findings led a
large number of pharmaceutical companies to orient their
GlyT1 medicinal chemistry programs toward the identification
of non-sarcosine-based inhibitors. This broad drug discovery
effort successfully delivered a rich set of highly structurally
diverse non-amino-acid chemotypes as exemplified by the
selected compounds 4,12 5,13 and 614 (Figure 1). We have also
contributed to this field and described the optimization of the
non-sarcosine-based spiropiperidine 7.15
Schizophrenia is a severe and chronic mental illness with
prevalence estimates ranging from 0.5% to 1% of the popula-
tion. Symptoms of schizophrenia, which typically emerge
during adolescence or early adulthood, are broadly categori-
zed as positive, negative, or cognitive. The leading treatments
for this central nervous system (CNSa) disorder are the
atypical antipsychotics belonging to the family of dopamine
D2 and serotonin 5-HT2a receptor antagonists. These drugs
are efficacious in the management of positive symptoms;
however, they demonstrate minimal effects on negative symp-
toms and cognitive function. In addition, because of their
promiscuous pharmacology, these medicines have unsatisfac-
tory side effect profiles. There is therefore a clear need to
develop novel nondopaminergic/nonserotonergic approaches
that may enable the identification of more efficacious and
safer therapeutic agents.1 Over the past 2 decades, preclinical
and clinical evidence has accumulated suggesting that hypo-
function of glutamatergic N-methyl-D-aspartate (NMDA)
receptors may play an important role in the pathophysiology
of schizophrenia.2 Thus, therapeutic intervention aiming at
*To whom correspondence should be addressed. Phone: þ41 61 688
43 88. Fax: þ41 61 688 87 14. E-mail: emmanuel.pinard@roche.com.
a Abbreviations: Boc, tert-butyloxycarbonyl; CNS, central nervous
system; CYP450, cytochrome P450; DMF, N,N-dimethylformamide;
DMPK, drug metabolism and pharmacokinetic; EWG: electron with-
drawing group; FaSSIF, fasted state simulated intestinal fluid; FeSSIF,
fed state simulated intestinal fluid; GlyT1, glycine transporter 1; hERG,
human ether-a-go-go-related gene; HPLC, high pressure liquid chro-
matography; HTS, high-throughput screening; iv, intravenous; MNT,
micronucleus test; NMDA, N-methyl-D-aspartate; NMP, N-methylpyr-
rolidone; po, per os (oral); PAMPA, parallel artificial membrane permea-
tion assay; PPB, plasma protein binding; SAR, structure-activity
relationship; TBTU, 2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium
tetrafluoroborate; THF, tetrahydrofuran; THP, tetrahydropyranyl; TLC,
thin layer chromatography; TPGS, D-R-tocopheryl polyethylene glycol 1000
succinate.
More recently, we reported on the discovery of benzoylpi-
perazines as a further novel chemotype of GlyT1 inhibitors.16
Starting from high-throughput screening (HTS) hit 8 (GlyT1
EC50 = 15 nM), SAR studies accompanied by the early
assessment of molecular properties and metabolic stability
of the synthesized derivatives led to the identification of a
r
2010 American Chemical Society
Published on Web 05/21/2010
pubs.acs.org/jmc

4604 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Pinard et al.
Figure 1. A selection of published GlyT1 inhibitors.
highly promising subseries in which the morpholine and nitro
residues in positions 2 and 5 in 8 (Figure 1) were replaced
respectively by an alkoxy and a methyl sulfone group. Within
this 2-alkoxy-5-methylsulfonebenzoylpiperazine subseries,
highly potent, selective, and druglike GlyT1 inhibitors were
identified, demonstrating in particular excellent metabolic
stabilities, oral bioavailabilities, and no significant inhibition
of the major metabolizing CYP450 enzymes. Moreover, in an
in vivo microdialysis study in which the extracellular fluid in
the striatum of mice was sampled by means of an implanted
microdialysis probe, representative compound 9(GlyT1 EC50
=
16 nM) induced a robust and sustained increase level of
glycine after oral administration, providing evidence that 9
acts in vivo as a GlyT1 inhibitor. Furthermore, we have found
that our benzoylpiperazines were able to dose-dependently
reverse the hyperlocomotion induced in mice by (3R,4R)-3-
Figure 2. Dose-dependent inhibition of L-687,414-induced hyperlo-
comotionbycompound9in mice. Thedatarepresentmean horizontal
activity counts per group recorded over a 60 min time period. Error
bars indicate SEM (n = 8 per group): (##) p < 0.01 versus vehicle
(Veh) alone; (/) p < 0.05, (//) p < 0.01 versus L-687,414 alone.
amino-1-hydroxy-4-methylpyrrolidin-2-one (L-687,414),17
a
known selective and brain-penetrating glycine site NMDA
receptor antagonist. An ID50 (dose producing 50% inhibition
of L-687,414-induced hyperlocomotion) of 3 mg/kg was
measured for compound 9 after oral administration (Figure 2).
Such behavioral effect was observed as well with the potent
and brain penetrant published GlyT1 inhibitors we have
tested. For example, with sarcosine-based compound 2, an
ID50 of 0.6 mg/kg po was measured.18 In contrast, no effect
was seen with the selective GlyT2 inhibitor 4-benzyloxy-N-(1-
dimethylaminocyclopentylmethyl)-3,5-dimethoxybenzamide
(Org-25543),19 demonstrating the specificity of the L-687,414-
basedprocedure.18 The behavioraleffect observedwithGlyT1
inhibitors likely results from the increased extracellular level
of glycine in the brain and the subsequent displacement of the
NMDA glycine site antagonist from its binding site. This
L-687,414 based procedure thus represents a useful novel
functional method to detect the in vivo activity of GlyT1

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4605
inhibitors. In contrast to the microdialysis procedure, which is
a highly skilled, time-consuming assay requiring surgery on
the tested animal, the L-687,414 based method is compara-
tively straightforward and possesses sufficient throughput to
efficiently guide lead optimization efforts in GlyT1 inhibitor
series. We have thus utilized this procedure as a first-line
model for the in vivo screening of our benzoylpiperazines.
Despite its attractive in vitro and in vivo efficacy profile, 9
could not be developed further because of a pronounced
inhibitory activity measured at the human ether-a-go-go-related
gene (hERG) channel (IC50 = 0.6 μM, patch-clamp assay).
In addition, 9 and close analogues demonstrated a fairly low
CNS penetration with measured brain/plasma ratios typically
around 0.1.
culminated in the discovery of RG1678 (10a, Figure 3), the
first potent and selective GlyT1 inhibitor to have a beneficial
effect in schizophrenic patients as shown in a recent phase II
clinical trial.21
Chemistry
Before describing the synthesis of the GlyT1 inhibitors, we
will first discuss the preparation of several noncommercially
available benzoic acids as well as N-arylpiperazines and
N-heteroarylpiperazines, which were required as building
blocks. The benzoic acids 13a-l were prepared from the
known and readily available acid 1122 as shown in Scheme 1.
11 was first esterified to provide phenol ester 12 which was
subsequently alkylated either under Mitsunobu conditions or
by reaction with an alkyl halide or a triflate reagent.
The chiral acids 19 and 19(þ) were prepared from 2-fluoro-
benzoic acid 14 following the synthetic route described in
Scheme2. Upontreatment withchlorosulfonicacid, 14 under-
went a fully regioselective Friedel-Crafts reaction to provide
sulfonyl chloride 15 which was subsequently reduced to the
sulfinic acid 16. Reaction of 16 with methyl iodide delivered
the methylsulfone methyl ester derivative 17, which was then
saponified to the acid 18. Finally, the desired chiral acids 19
and 19(þ) were readily obtained upon heating 18 with,
respectively, (rac)- and (S)-1,1,1-trifluoropropan-2-ol23 in
the presence of a mild base such as potassium carbonate.
Importantly, this aromatic nucleophilic substitution reaction
We report here on our lead optimization effort in the
2-alkoxy-5-methylsulfonebenzoylpiperazine series,20 addres-
sing the hERG channel and brain penetration issues that
Figure 3. Structure of 10a (RG1678).
Scheme 1. Synthesis of Benzoic Acids 13a-la
a Reagents and conditions: (a) cat. H2SO4, MeOH, reflux, 8 days, 89%; (b) R1OH (1 equiv), di-tert-butylazodicarboxylate (1 equiv), triphenylpho-
sphine (1 equiv), THF, room temperature, 24 h, then 2 N NaOH (2 equiv), ethanol, 80 ꢀC, 1 h; (c) R1X (1.5 equiv), K2CO3 (2 equiv), acetone, 70 ꢀC, 22 h,
then 2 N NaOH (2 equiv), ethanol, 80 ꢀC, 0.5 h.
Scheme 2. Synthesis of Benzoic Acids 19 and 19(þ)a
a Reagents and conditions: (a) chlorosulfonic acid (10 equiv), 75 ꢀC, overnight, 76%; (b) sodium sulfite (7.5 equiv), water, room temperature, 2 h,
72%; (c) methyl iodide (3.5 equiv), potassium carbonate (3 equiv), room temperature, overnight, 59%; (d) LiOH (1.5 equiv), THF, water, room
temperature, 1 h, 85%; (e) rac- or (S)-1,1,1-trifluoropropan-2-ol (5 equiv), potassium carbonate (3 equiv), dimethylacetamide, 150 ꢀC, 2 h, 40-70%.

4606 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Pinard et al.
Scheme 3. Synthesis of N-Aryl-/Heteroarylpiperazines 21a-ha
a Reagents and conditions: (a) dimethylacetamide, 110 ꢀC, overnight; (b) dimethylacetamide, 150-180 ꢀC in microwave, 10-20 min; (c) potassium
carbonate(2 equiv), acetonitrile, reflux, 2 h; (d) tris(dibenzylideneacetone)dipalladium chloroform complex (0.01 equiv), 2-(dicyclohexylphosphino)biphenyl
(0.02 equiv), sodium tert-butoxide (1.4 equiv), toluene, 80 ꢀC, 12 h, (e) tris(dibenzylideneacetone)dipalladium (0.025 equiv), tri-o-tolylphosphine (0.1 equiv),
sodium tert-butoxide (2.8 equiv), dioxane, 100 ꢀC, 7 h; (f) 4 N HCl (4 equiv), dioxane, 80 ꢀC, overnight; (g) trifluoroacetic acid (5 equiv), dichloromethane,
reflux, 2-24 h.
Scheme 4. Synthesis of 2-Piperazin-1-yl-5-trifluoromethylpyrimidine 24a
a Reagents and conditions: (a) triethylamine (2.4 equiv), acetonitrile, room temperature, 3 h, 94%; (b) H2 (1 atm), 5% Pd/C, methanol, 60 ꢀC, 2 h, 92%.
proceeded without any erosion of the stereochemical integrity
of the chiral center.
Results and Discussion
Starting from 9, our efforts focused first on the optimiza-
tion of the alkoxy group. Replacement of the cyclopropyl-
methoxy moiety with a hydroxyl group (25) or with small
alkoxy substituents(26, 27) resultedina dramatic reduction of
GlyT1affinity (Table1). Pleasingly, the activitywas recovered
upon increasing the size or the lipophilic bulk of this
substituent as seen with the alkoxy (28-30), cycloalkoxy
(31-33), and fluoroalkoxy (34, 35) derivatives, which all
displayed GlyT1 activities in the low nanomolar range. An
abrupt drop of potency was, however, observed with more
extended substituents such as the cyclohexylmethoxy or ben-
zyloxy groups (36, 37) as well as with groups carrying polar
atoms (38, 39), suggesting that substituents at position 2 of the
benzoyl motif fit in a size-limited and lipophilic pocket of the
GlyT1 receptor. Gratifyingly, all active derivatives, in addi-
tion, displayed excellent selectivity versus the GlyT2 isoform
(Table 1).
The novel N-arylpiperazines or N-heteroarylpiperazines
21a-h were synthesized upon reaction of tert-butyl 1-piper-
azinecarboxylate with aryl or heteroaryl halides 20 followed
by cleavage of the Boc group under acidic conditions (Scheme 3).
Hartwig/Buchwald palladium-catalyzed amination conditions
were used for the least activated aryl and heteroaryl halides,
whereas noncatalyzed thermal conditions were employed for
the most activated ones.
Finally, novel 2-piperazin-1-yl-5-trifluoromethylpyrimidine
24 was readily constructed using Yamanaka’s methodology24
involving the condensation of commercial guanidine derivative
22 with (3-dimethylamino-2-trifluoromethylallylidene)dimethyl-
ammonium chloride salt (Scheme 4).
2-Alkoxy-5-methylsulfonebenzoylpiperazine derivatives
10a,b, 26-54, 56 were synthesized by employing an amidation
reaction between the benzoic acids 13a-m, 19, 19(þ) and
the N-arylpiperazines or heteroarylpiperazines 21a-p, 24
(Scheme 5). Phenol 25 was prepared by deprotection of
the O-benzyl substituted derivative 37 in the presence of
iodotrimethylsilane. 3-Methyl-5-trifluoromethylpyridin-2-yl
substituted compound 55 was obtained from 3-chloro deri-
vative 56 following a palladium-catalyzed cross-coupling
reaction using trimethylaluminum as the methyl donor
agent.25
Some of the most potent analogues were selected for further
evaluation atthe hERG channel and in the L-687,414-induced
hyperlocomotion assay in mice (Table 2). Isobutyloxy analo-
gue 29 displayed a similarly high activity at hERG channel
compared with initial lead compound9 (IC50 = 0.7 μM) and a
comparable in vivo activity (ID50 = 2.5 mg/kg). Interestingly,
exchange of the elongated alkoxy groups in 9 and 29 with

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4607
Scheme 5. Synthesis of 2-Alkoxy-5-methylsulfonebenzoylpiperazines: 10a,b and 25-56a
a Reagents and conditions: (a) TBTU (1.1 equiv), i-Pr2NEt (5 equiv), DMF, room temperature, overnight, 33-100%; (b) iodotrimethylsilane
(4 equiv), acetonitrile, 80 ꢀC, 48 h, 37%; (c) trimethylaluminium (3.6 equiv), tetrakis(triphenylphosphine)palladium (0.24 equiv), THF, 65 ꢀC, 68 h, 46%.
concomitant drop of in vivo activity (ID50 = 8 mg/kg for 30
and7 mg/kgfor34) was observed. Incontrast, the reductionin
hERG activity with the trifluoroisopropyloxy derivative 35
was found to be accompanied by an improvement in in vivo
efficacy reaching an ID50 of 1 mg/kg. In pharmacokinetics
studies, all compounds in Table 2 exhibited equivalent plasma
exposures after oral dosing but different brain concentrations
and therefore different brain/plasma ratios. Overall, in vivo
activity seemed to correlate quite well with the brain penetra-
tion, with the highest efficacy obtained with compound 35
which displayed a superior brain/plasma ratio (0.25) com-
pared to the other alkoxy derivatives examined (Table 2). In
summary, because of its beneficial effect on CNS penetration
plus in vivo and hERG profiles, the trifluoroisopropyloxy
group emerged as the best substituent among the set of alkoxy
residues investigated. This group thus constituted a highly
favored substituent at the 2-position in the benzoylpiperazine
class.
With the preferred trifluoroisopropyloxy group kept in
place at position 2, our effort next focused on the optimization
of the western aromatic region. In agreement with the pre-
liminary SAR results obtained on our initial HTS hit 8,16 the
introduction of electron-donating groups like methyl (40) or
methoxy (41) in the para position of the aromatic system
resulted in weakly active GlyT1 inhibitors. Higher affinities
were obtained with analogues carrying strong EWGs such as
nitrile (42), trifluoromethyl (43), methyl sulfone (44), or
methyl ketone (45) functions (Table 3). In particular, the best
in vitro results were seen with the nitrile (42) and trifluoro-
methyl (43) derivatives which showed respectively a GlyT1
affinity of 21and 30 nM. Keeping the p-trifluoromethyl group
in place, a significant improvement in GlyT1 potency was
detected upon the introduction of a small EWG like fluorine
in the ortho position as observed with compound 46 (13 nM).
Table 1. In Vitro Inhibitory Activity of 9 and 25-39 at hGlyT1 and
hGlyT2
EC50, μMa
compd
R1
hGlyT1
hGlyT2
9
CH2-c-Pr
H
Me
0.016
>10
9.80
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
>30
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
Et
n-Pr
0.59
0.034
0.010
0.070
0.018
0.012
0.008
0.028
0.044
0.39
i-Bu
i-Pr
c-C4H7
c-C5H9
c-C6H11
CH2CF3
CH(CH3)CF3
CH2-c-C6H11
benzyl
1.61
0.50
THP-4-yl
CH2CH2NMe2
>10
a EC50 values are the average of at least two independent experiments.
shorter substituents such as isopropyloxy (30), trifluoro-
ethoxy (34), and trifluoroisopropyloxy (35) induced a sig-
nificant decrease of hERG inhibitory activity (IC50 = 3.3, 2,
and 3 μM, respectively). In the case of 30 and 34, however, a

4608 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Pinard et al.
Table 2. Activity of 9, 29, 30, 34, 35 at the hERG Channel and in the L-686,414-Induced Hyperlocomotion Assay in Mouse
mouse
reversal of L-687,414-induced
hyperlocomotion in mouse ID50, mg/kgd
Cmax plasma,
ng/mLe
Cmax brain,
ng/mLe
2 c
˚
PSA, A
compd
hERG IC50, μMa
clogPb
brain/plasmae
9
0.6
0.7
3.3
2.0
3.0
2.64
3.13
2.51
2.64
2.77
80
80
80
80
80
3.0
2.5
8.0
7.0
1.0
3870
2900
3143
3000
3100
387
460
220
240
775
0.10
0.16
0.07
0.08
0.25
29
30
34
35
a Patch clamp assay. IC50 values are the average of at least two independent experiments. b Calculated lipophilicity.26 c Polar surface area calculated
with Moloc.27 d Compound given orally, n = 8. e Compound given orally at a dose of 10 mg/kg, n = 2.
However, introduction of a larger halogen atom such as
chlorine (47) at this position resulted in a sharp drop of
activity (200 nM). Further in vitro and in vivo characteriza-
tion revealed that the trifluoromethyl derivatives 43 and 46
had more favorable profiles than the corresponding nitrile
analogues 42 and 35. First, with brain/plasma ratios of about
1, 43 and 46 demonstrated a much-improved CNS penetra-
tion. Knowing that PSA might be an important parameter
that limits blood-brain barrier permeation, we hypothesized
that this result might be attributed to their lower polar sur-
face area (PSA = 59 A vs 80 A ). We applied the Roche
automatic SAR analyzer (ROSARA),28 a proprietary pro-
gram designed to study potential correlation between a given
property (i.e., brain penetration in this case) and numerous
calculated (e.g., clogP or PSA) or measured (e.g., GlyT1 EC50,
solubility or permeability) descriptors. In fact, ROSARA
revealed that the measured CNS penetration in our series
could be very well predicted by using solely PSA and clogP as
descriptors (Figure 4).
(IC50 = 80 nM). Similarly, a decrease of GlyT1 activity was
unfortunately observed with the pyrazine, pyrimidine, and
pyridazine derivatives 50-53. It is, however, interesting to
note that these more polar heteroaryl derivatives having clogP
values rangingfrom1.83 to 1.96were practicallydevoid of any
hERG activity (IC50 > 40 μM). Thus, the structure-hERG
activity relationship described in this compound class clearly
illustrates that the exchange of an aryl group by heteroaryl
systems can be a very powerful medicinal chemistry strategy
to control hERG affinity in chemical series. Starting from
pyridine derivative 48, our next step was to improve the
CNS penetration and in vivo potency while retaining the
excellent selectivity profile against the hERG channel. The
well-documented beneficial effect of fluorine atom insertion
on CNS penetration of biologically active molecules31 led us
to consider analogues to pyridine compound 48 incorporating
an extra fluorine atom. Additionally, we reasoned that this
modification would probably not affect the selectivity toward
the hERG channel, since a hydrogen to fluorine exchange in
molecules generally results in only a moderate increase of
lipophilicity.32
2
2
˚
˚
Inaddition totheir superior CNS penetration, the trifluoro-
methylated analogues 43 and 46 showed significantly lower
inhibitory activities at the hERG channel compared to the
The tolerance of the fluorine atom for GlyT1 activity
observed in the ortho aryl position in derivative 46 led us to
prepare the 3-F, 5-CF3-pyridyl substituted compound 54. As
we expected, 54 showed a very good affinity at GlyT1 trans-
porter (IC50 = 40 nM), comparable to the potency obtained
for the parent pyridine compound 48. Gratifyingly, as we
anticipated, 54 demonstrated an improved brain penetration
(brain/plasma: 0.5 vs 0.2 for 48) and consequently a superior
in vivo activity in the L-687,414 mouse assay with an ID50
reaching 1 mg/kg after oral administration. Moreover, as
desired, the introduction of a fluorine atom into 48 did not
impact negatively on the selectivity against the hERG channel
(500-fold selectivity, IC50 = 20 μM). In summary, the hydro-
gen to fluorine exchange performed at position 3 on the
pyridine ring of compound 48 delivered a potent GlyT1
inhibitor 54 displaying an excellent hERG selectivity com-
bined with a robust CNS penetration and in vivo potency. The
alternative exchanges examined at that position, hydrogen vs
methyl (55) and vs chlorine (56), met with much less success,
illustrating once again the unique beneficial effect a fluorine
group may have on compound properties. Indeed, the methyl
corresponding more polar nitrile analogues 42 and 35 (IC50
=
10 and 6.9 μM vs 1.2 and 3 μM). This result is quite remark-
able because it is well recognized that, on the contrary, a
reduction of compound lipophilicity in a chemical series is
often associated with a decrease of hERG channel affinity.29
We hypothesize that the high hERG inhibition observed with
the nitrile derivatives is due to a specific H-bonding interac-
tion of the cyano group with a key amino acid residue in the
pore region of the hERG channel.30 The superior in vitro and
brain/plasma ratios demonstrated by the trifluoromethyl
substituted derivatives prompted us to focus our subsequent
optimization effort specifically on this class of compounds.
Starting from 43, we set out to further improve the hERG
selectivity profile. Toward this end, analogues with increased
polarity, obtained by exchanging the western aromatic nu-
cleus of 43 with heteroaromatic systems such as pyridines 48
and 49, pyrazine 50, pyrimidines 51 and 52, and pyridazine 53,
were investigated (Table 3). Our strategy proved successful
because pyridine derivative 48, displaying a favorable GlyT1
affinity (IC50 = 37 nM), showed a 2-fold reduction in hERG
activity (IC50 = 20 μM) over the parent compound 43. With
compound 48, wethusachievedforthe firsttimeahighlevel of
selectivity against the hERG channel, reaching over 500-fold.
However, because of its increased polar surface area (PSA =
analogue 55 had a significantly lower GlyT1 affinity (IC50
80 nM) and the chloro derivative 56, while potent at GlyT1
(IC50 = 30 nM), showed a reduced in vivo activity (ID50
=
=
4 mg/kg) as well as a decreased selectivity vs the hERG
channel (300-fold), a result likely due to its higher lipophilicity
(clogP = 3.57 vs 3.00 for 54). Compound 54, existing as a
racemic mixture, was subsequently separated by chiral phase
HPLC to provide the two enantiomers 10a ((S)-enantiomer)
and 10b ((R)-enantiomer). 10a was found positively differ-
entiated over its enantiomer 10b by its higher in vitro GlyT1
2
˚
69 A ) and reduced lipophilicity (clogP = 2.82), 48 exhibited a
reduced CNS penetration (brain/plasma: 0.2 vs 1.1 for 43)
and accordingly, a lower in vivo activity (ID50=5 mg/kg)
in the L-687,414-induced hyperlocomotion assay. Shifting
the pyridinyl ring nitrogen to the adjacent position to afford
49 provided a compound with reduced GlyT1 potency

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4609
Table 3. In Vitro and in Vivo Profiles of 40-56 and 10a,b
a EC50 values are the average of at least two independent experiments. b Patch clamp assay. IC50 values are the average of at least two independent
experiments. c Calculated lipophilicity.26 d Polar surface area calculated with Moloc.27 e Compound given orally, n = 8. f Compound given orally at the
dose of 10 mg/kg, n = 2.

4610 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Pinard et al.
activity (IC50 = 30 nM) and superior in vivo efficacy in the
L-687,414-induced hyperlocomotion assay, reaching for the
first time an ID50 as low as 0.5 mg/kg po while hERG activity
still resided in a favorable range (IC50 = 17 μM). Additional
profiling revealed that 10a had an excellent selectivity profile
against the GlyT2isoform(IC50 > 30μM) and towarda panel
of 86 targets including transmembrane and soluble receptors,
enzymes, ion channels, and monoamine transporters (<41%
inhibition at 10 μM was measured for all targets).33 Although
10a, a nonbasic compound (pKa<2), showed low aqueous
solubility (1 μg/mL with crystalline material at pH 6.5), it
exhibited a satisfactory solubility profile in both fasted state
simulated intestinal fluid (FaSSIF) (20 μg/mL) and fed state
simulated intestinal fluid (FeSSIF) (60 μg/mL). 10a showed,
in addition, excellent membrane permeability as measured in
the parallel artificial membrane permeation assay (PAMPA)
(Pe = 3.2 ꢀ 10-6 cm/s).34 In vivo pharmacokinetic studies in
rat and monkey (Table 4) revealed that 10a had, in both
species, a low plasma clearance, an intermediate volume of
distribution, a good oral bioavailability (78% for rat, 56% for
monkey), and a favorable terminal half-life (5.8 h for rat, 6.4 h
for monkey). The plasma protein binding was high in the two
preclinical species (97%) and in human (98%). The CNS
penetration of 10a in rat (brain/plasma = 0.7) was better than
that in mouse (brain/plasma = 0.5). Finally, compound 10a
did not significantly inhibit the major drug metabolizing
CYP450 enzymes 3A4, 2D6, 2C9, 2C19, 1A2 (all IC50 values
above 24 μM) and was without activity in genotoxicity assays
(Ames and MNT tests).
Figure 4. Correlation of experimental vs calculated log BB (log of
brain/plasma ratio) using the Roche Automatic SAR analyzer
(ROSARA). The model relies solely on two descriptors: clogP and
PSA. The line of identity is shown in black. A linear regression leads
to the following parameters: n = 11; r2 = 0.875; q2 = 0.845.
Table 4. Pharmacokinetic Properties of 10a in Rat and Cynomolgus
Monkey
species
parameter
rath
cynomolgus monkeyh
iv dose (mg/kg)a
CL ((mL/min)/kg)b
Vdss (L/kg)c
2
0.5
The excellent data obtained with 10a prompted us to
evaluate the effect of this compound on the extracellular level
of glycine in rat striatum in a microdialysis study (Figure 5).
Gratifyingly, at an oral dose of 10 mg/kg, 10a produced a
robust and sustained increase of glycine levels (2.3-fold over
basal levels at 180 min). The lack of effect on the striatal
extracellular levels of D/L-serine (data not shown) indicated
that this compound did not affect other neutral amino acid
transporters and provided in vivo evidence of its selectivity for
the GlyT1 transporter.
4.32 ((25%)
3.58 ((14%)
3.62 ((9%)
1.98 ((8%)
2.8
po dose (mg/kg)d
Cmax (ng/mL)
3
721 ((13%)
8670 ((16%)
5.8 ((25%)
1
517 ((38%)
7230 ((2%)
6.4 ((8%)
5.5 ((64%)
56 ((3%)
AUC0-¥ (ng h/mL)e
T1/2 (h) terminal
3
T
max (h)
F (%)f
78 ((20%)
0.7
brain/plasma
PPB (% unbound)g
3
3
a iv formulations. Rat: NMP/hydroxypropyl γ-cyclodextrin. Monkey:
30% captisol. b Total plasma clearance. c Volume of distribution. d po
formulation. Tween-80, methylparaben, propylparaben, and hydroxy-
ethylcellulose. e Area under the curve extrapolated to infinity. f Oral
bioavailability. g Plasma protein binding. h n = 2. Mean values ((CV%).
Conclusion
In summary, chemical optimization in the class of benzoyl-
piperazines has led to the discovery of compound10a, a highly
Figure 5. Effect of 10a on the extracellular level of glycine in rat striatum at 10 mg/kg po measured by microdialysis. Data are the mean ( SEM
(n = 3).

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4611
potent and selective GlyT1 inhibitor exhibiting excellent
pharmacokinetic and in vivo efficacy profiles in nonhuman
species. On this basis, 10a advanced into clinical trials in
healthy subjects and in schizophrenic patients. F. Hoff-
mann-La Roche recently reported positive phase II results
with 10a for the treatment of negative symptoms, a core
clinical feature of schizophrenia where there are currently no
effective therapies. The complete biological characterization
of 10a will be reported in further detail in forthcoming
publications.
aqueous phase was extracted twice. The combined organic phases
were dried over sodium sulfate, then concentrated in vacuo to
provide the desired product as a light-yellow solid (13.1 g, 59%)
which was used in the next reaction without further purification.
1H NMR (300 MHz, CDCl3) δ 8.55 (dd, J = 6.6, 2.4 Hz, 1H), 8.13
(m, 1H), 7.36 (dd, J = 9.9, 9.0 Hz, 1H), 3.98 (s, 3H), 3.09 (s, 3H).
MS (EI) m/e: 250.1 (M þ NH4)þ.
2-Fluoro-5-methanesulfonylbenzoic Acid (18). To a solution of
2-fluoro-5-methanesulfonylbenzoic acid methyl ester 17 (18 g,
77.3 mmol) in THF (360 mL) and water (360 mL) was added
lithium hydroxide monohydrate (4.9 g, 116 mmol, 1.5 equiv).
The reaction mixture was stirred at room temperature for 1 h.
Then THF was distilled in vacuo and 1 N HCl was added to
reach pH 2. The mixture was extracted with diethyl ether (2 ꢀ
60 mL). The combined organic phases were dried over sodium
sulfate, then concentrated in vacuo. The crude material was
triturated with diisopropyl ether to provide, after filtration and
Experimental Section
Chemistry. General. All solvents and reagents were obtained
from commercial sources and were used as received. All reac-
tions were followed by TLC (TLC plates F254, Merck) or
LCMS (liquid chromatography-mass spectrometry) analysis.
Proton NMR spectra were obtained on Bruker 300 or 400 MHz
instrument with chemical shifts (δ in ppm) reported relative to
tetramethylsilane as internal standard. NMR abbreviations are
as follows: s, singlet; d, doublet; t, triplet; q, quadruplet; quint,
quintuplet; sext, sextuplet; hept, heptuplet; m, multiplet; br,
broadened. Purity was analyzed by reverse phase HPLC and for
specific compounds by elemental analysis. HPLC was per-
formed on Finnigan LTQ (Thermo Fisher Scientific) and Agi-
lent RRLC 1200 equipment. The column was Agilent XDB C15,
30 mm ꢀ 4.6 mm, 3.5 μm. Analytical conditions were as folllows:
Gradient used was from 5% to 95% acetonitrile in water contain-
ing 0.1% trifluoroacetic acid in 3 min. Flow was 4.5 mL/min. UV
detector was DAD, 190-400 nm. Sample solvent was in water/
acetonitrile (8/2). The UV detection was an averaged signal from
wavelengths of 190-400 nm. Elemental analyses were performed
by Solvias AG (Mattenstrasse, Postfach, CH-4002 Basel,
Switzerland). Column chromatography was carried out on silica
1
drying, the desired product as a white solid (14.4 g, 85%). H
NMR (300 MHz, DMSO-d6) δ 13.8 (br s, 1H), 8.36 (dd, J = 6.9,
2.4 Hz, 1H), 8.20 (m, 1H), 7.63 (dd, J = 10.5, 8.7 Hz, 1H), 3.30
(s, 3H). MS (EI) m/e: 235.9 (M þ NH4)þ.
rac-5-Methanesulfonyl-2-(2,2,2-trifluoro-1-methylethoxy)benzoic
Acid (19). A suspension containing 18 (0.20 g, 0.92 mmol), rac-
1,1,1-trifluoropropan-2-ol (0.40 mL, 4.58 mmol, 5 equiv), and
potassium carbonate (0.38 g, 2.75 mmol, 3 equiv) in dimethylace-
tamide (5 mL) was stirred at 150 ꢀC for 2 h, then cooled to room
temperature and concentrated in vacuo. The residue was diluted
with water (5 mL) and acidified with 1 N HCl to pH 2. The
precipitate was filtered, washed with water, and dried in vacuo to
afford the desired product as a white solid (0.11 g, 40%). 1H NMR
(300 MHz, CDCl3) δ8.63(d, J= 2.4 Hz, 1H), 8.13 (dd, J= 8.7, 2.4
Hz, 1H), 7.21 (d, J = 8.7 Hz, 1H), 4.94 (m, 1H), 3.09 (s, 3H), 1.66
(d, J = 6.5 Hz, 3H). MS (EI) m/e: 311.1 (M - H).
(S)-5-Methanesulfonyl-2-(2,2,2-trifluoro-1-methylethoxy)benzoic
Acid (19(þ)). Preparation followed the same procedure as for 19
from 18 and (S)-1,1,1-trifluoro-propan-2-ol23 (70% yield, white
solid). [R]2D0 þ16.9ꢀ (c1.01, MeOH). 1H NMR (300 MHz, CDCl3) δ
8.63 (d, J= 2.4 Hz, 1H), 8.13 (dd, J= 8.7, 2.4 Hz, 1H), 7.21 (d, J=
8.7 Hz, 1H), 4.94 (m, 1H), 3.09 (s, 3H), 1.66 (d, J = 6.5 Hz, 3H). MS
(EI) m/e: 311.1 (M - H). The optical purity was determined by chiral
HPLC of the corresponding methyl ester (obtained by refluxing acid
in methanol with the presence of catalytic amount of sulfuric acid)
using a Chiralpack AD column (flow 1 mL/min, pressure of 25 bar,
detection at 254 nm) and using heptane/isopropanol, 80:20, as
eluant: retention time of 22.9 min, ee of >99%.
1-(3-Fluoro-5-trifluoromethylpyridin-2-yl)piperazine (21e). A
mixture of 2,3-difluoro-5-trifluoromethylpyridine (21 g, 115 mmol),
tert-butyl 1-piperazinecarboxylate (23.1 g, 120 mmol, 1.05 equiv),
and potassium carbonate (31.7 g, 229 mmol, 2 equiv) in acetonitrile
(315 mL) was refluxed for 2 h. The reaction mixture was then cooled
to room temperature and filtered, and the filtrate was concentrated
in vacuo. The solid was dissolved in ethyl acetate (250 mL). The
solution was washed once with water (100 mL) and with 1% citric
acid solution (3 ꢀ 100 mL), dried over sodium sulfate, filtered, and
concentrated in vacuo to afford 4-(3-fluoro-5-trifluoromethylpyri-
din-2-yl)piperazine-1-carboxylic acid tert-butyl ester as a white
solid (37.5 g). This material was dissolved in dichloromethane
(376 mL), and trifluoroacetic acid (41.1 mL, 537 mmol, 5 equiv)
was added. The resulting mixture was refluxed for 16 h and then
cooled to room temperature and concentrated in vacuo. The residue
was diluted with water (450 mL), basified with 5 N NaOH,
and extracted with dichloromethane (3ꢀ50 mL). The combined
extracts were washed with brine (50 mL), dried over sodium sulfate,
filtered, and concentrated in vacuo to afford the desired compound
as a light-yellow solid (26.4 g, 93%). 1HNMR(300MHz, CDCl3) δ
8.22(d,J= 1.9 Hz, 1H), 7.37 (dd, J= 13.5, 2.0 Hz, 1H), 3.63 (t, J=
4.8 Hz, 4H), 2.99 (t, J= 5.1 Hz, 4H). MS (EI) m/e:250.2(Mþ H)þ.
rac-4-(3-Fluoro-5-trifluoromethylpyridin-2-yl)piperazin-1-yl]-
[5-methanesulfonyl-2-(2,2,2-trifluoro-1-methylethoxy)phenyl]-
methanone (54). General Procedure C. To a solution of 19
˚
gel 60 (32-60 mesh, 60 A) or on prepacked columns (Isolute
Flash Si). Mass spectra were recorded on an SSQ 7000 (Finnigan-
MAT) spectrometer for electron impact ionization. The purities
of final compounds 10a,b and 25-56 as measured by HPLC or
elemental analysis method were found to be above 95%.
Detailed Description. 5-Chlorosulfonyl-2-fluorobenzoic Acid
(15). To chlorosulfonic acid (238 mL, 3.5 mol, 10 equiv) cooled
to 0 ꢀC was added portionwise 2-fluorobenzoic acid 14 (50 g,
0.35 mol). After complete addition, the yellow solution was
allowed to warm to room temperature, then heated to 75 ꢀC
overnight. The reaction mixture was cooled to room tempera-
ture and then added dropwise to ice (1.5 L). The white pre-
cipitate was filtered, washed with water, and dried in vacuo to
afford the desired product as a white solid (63 g, 76%). 1H NMR
(300 MHz, CDCl3) δ 8.76 (dd, J = 6.3, 2.7 Hz, 1H), 8.29 (m,
1H), 7.47 (t, J = 9.3 Hz, 1H). MS (EI) m/e: 236.8(M - H).
2-Fluoro-5-sulfinobenzoic Acid (16). To a solution of sodium
sulfite (250 g, 1.98 mol, 7.5 equiv) in water (1 L) was added
portionwise during 20 min chlorosulfonyl-2-fluorobenzoic acid
15 (63 g, 0.264 mol). The reaction mixture was stirred at room
temperature for 2 h, then cooled to 0 ꢀC and acidified with 20%
sulfuric acid (∼230 mL) to pH 2. Water was evaporated, and
methanol (600 mL) was added. The mixture was stirred over-
night and filtered and the filtrate was then concentrated and
dried over sodium sulfate to afford the desired product as a
white solid (39 g, 72%). 1H NMR (300 MHz, DMSO-d6) δ 7.73
(dd, J = 7.6, 2.2 Hz, 1H), 7.36 (m, 1H), 6.95 (dd, J = 10.4, 8.2,
Hz, 1H). MS (EI) m/e: 203.0 (M - H).
2-Fluoro-5-methanesulfonylbenzoic Acid Methyl Ester (17).
To a solution 2-fluoro-5-sulfinobenzoic acid 16 (19.5 g, 95.3 mmol)
in DMF (166 mL) was added potassium carbonate (39.5 g,
286 mmol, 3 equiv) and then methyl iodide (20.8 mL, 333 mmol,
3.5 equiv). The reaction mixture was stirred at room temperature
overnight. Then DMF (100 mL) was removed in vacuo, and water
(100 mL) was added followed by ethyl acetate (100 mL). The

4612 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Pinard et al.
(14 g, 44.8 mmol) in DMF (140 mL) was added successively TBTU
(16.2 g, 49.3 mmol, 1.1 equiv), N-ethyldiisopropylamine (38.4 mL,
224.2 mmol, 5 equiv), and 21e (13.4 g, 53.8 mmol, 1.2 equiv). The
reaction mixture was stirred overnight at room temperature, then
was concentrated in vacuo, and the residue was dissolved with
ethyl acetate (150 mL). The mixture was washed with water (2 ꢀ
200 mL) and saturated sodium bicarbonate solution (2 ꢀ 150 mL).
The aqueous phase was extracted with ethyl acetate (1 ꢀ 100 mL).
The combined organic phases were dried over sodium sulfate,
filtered, and concentrated in vacuo. The crude material was taken
up in ethanol (150 mL), and the mixture was heated to 80 ꢀC and
then allowed to cool to room temperature and stirred for 3 h. The
precipitate was filtered, washed with a small amount of ethanol,
and dried in vacuo to afford the desired compound as a white solid
(19.9 g, 82%). 1H NMR (400 MHz, CDCl3) δ 8.25 (s broad, 1H),
7.98 (dd, J= 8.6, 2.4 Hz, 1H), 7.95 (br s, 0.3H), 7.91 (d, J=2.4Hz,
0.7 H), 7.43 (dd, J = 13.2, 1.6 Hz, 1H), 7.14 (d, J = 8.9 Hz, 0.7H),
7.10 (d, J = 8.9 Hz, 0.3 H), 4.89-4.75 (m, 1H), 4.00-3.84 (m, 2H),
3.81-3.68 (m, 2H), 3.67-3.54 (m, 2H), 3.48-3.28 (m, 2H), 3.07 (s,
3H), 1.6 (d, J = 6.5 Hz, 0.9H), 1.54-1.50 (m, 2.1 H) as a 7:3
mixture of two rotamers. MS (EI) m/e: 544.3 (M þ H)þ. HPLC:
retention time of 1.86 min, >98% pure.
(S)-4-(3-Fluoro-5-trifluoromethylpyridin-2-yl)piperazin-1-yl]-
[5-methanesulfonyl-2-(2,2,2-trifluoro-1-methylethoxy)phenyl]-
methanone (10a) and (R)-4-(3-Fluoro-5-trifluoromethylpyri-
din-2-yl)piperazin-1-yl][5-methanesulfonyl-2-(2,2,2-trifluoro-1-
methylethoxy)phenyl]methanone (10b). 54 (530 mg, 0.975 mmol)
was separated on a preparative Chiralpack AD column (flow of
35 mL/min, pressure of 15 bar, detection at 254 nm) using heptane/
isopropanol, 80:20, as eluant to afford (S)-enantiomer 10a (191 mg,
36% yield, first eluting product, retention time of 110 min) and
(R)-enantiomer 10b (197 mg, 37% yield, second eluting product,
retention time of 145 min).
AD column (flow of 1 mL/min, pressure of 25 bar, detection at
254 nm) using heptane/isopropanol, 80:20, as eluant: retention
time of 17.4 min, ee of >99%.
hGlyT1 Uptake Inhibition Assay. Mammalian CHO cells
(Flp-in-CHO) stably expressing hGlyT1b cDNA were main-
tained in monolayer culture at 37 ꢀC in humidified air with 5%
CO2 in nutrient mixture F-12 containing 10% fetal calf serum,
1% penicillin-streptomycin, 600 μg/mL hygromicin, and 1 mM
glutamine. On day 1 of the glycine uptake experiments, cells
were plated at a density of 40 000 cells/well in complete F-12
medium without hygromycin in 96-well culture plates. On day 2,
the medium was aspirated and the cells were washed twice with
uptake buffer (150 mM NaCl, 10 mM Hepes-Tris, pH 7.4, 1 mM
CaCl2, 2.5 mM KCl, 2.5 mM MgSO4, 10 mM (þ)-D-glucose).
Thereafter, the cells were incubated for 20 min at 22 ꢀC with the
tested compounds. A range of concentrations of the compounds
was used to generate data for calculating the concentration of
inhibitor resulting in 50% of the effect (i.e, EC50, or the
concentration of the competitor inhibiting glycine uptake of
50%). Each concentration effect was tested in quadruplicate.
Uptake was started by adding 60 nM [3H]glycine (15 Ci/mmol,
GE Healthcare) and 25 μM nonradioactive glycine. Nonspecific
uptake was determined with 10 μM ORG24598, a potent and
specific GlyT1 inhibitor.8 Plates were incubated with gentle
shaking, and the reaction was stopped by aspiration of the
mixture and washing three times with ice-cold uptake buffer.
After addition of scintillation liquid, the plates were shaken for 3
h and radioactivity was measured by liquid scintillation on a
Perkin-Elmer TopCount scintillation plate reader. The CPM value
for each quadruplicate of a concentration of competing compound
was averaged (y1). Then the percent specific binding was calculated
according to the equation [(y1 - nonspecific)/(total binding -
nonspecific)] ꢀ 100. An EC50 value, defined as the concentration of
the compound causing 50% inhibition of specific binding, was
calculated by linear regression analysis of the dose-response data
using an Excel based computer curve-fitting program.
hGlyT2 Uptake Inhibition Assay. The assay was performed
with mammalian cells, (Flp-in-CHO), transfected with hGlyT2
cDNA, using the same conditions as described for the hGlyT1
uptake assay with slight modifications. Uptake was started by the
addition of 200 nM [3H]glycine (without cold glycine). Nonspecific
uptake was determined with 5 μM ORG25543, a potent and specific
GlyT2 inhibitor,19 and the reaction was stopped after 30 min.
Reversal of L-687,414-Induced Hyperlocomotion in Mice.
Animals and Housing Conditions. Male NMRI mice (20-30 g)
supplied from Iffa Credo, Lyon, France, were housed in a
vivarium at controlled temperature (20-22 ꢀC) and under a
(12 h light)/(12 h dark) cycle (lights on at 6:00 a.m.). Mice were
allowed ad libitum access to food and water. The experimental
procedures used in the present study received prior approval
from the City of Basel Cantonal Animal Protection Committee
based on adherence to federal and local regulations. Behavioral
experiments were conducted between 8:00 a.m. and 2:00 p.m.
Formulation of L-687,414 and Test Substances. L-687,414
((3R,4R)-3-amino-1-hydroxy-4-methylpyrrolidin-2-one) was used
as its acetic acid salt. It was prepared following a synthetic route
developed at F. Hoffmann-La Roche.35 L-687,414 acetic acid salt
wasdissolvedin0.9%NaCl/0.3%Tween-80andadministered at a
dose of 50 mg/kg sc in a volume of 10 mL/kg body weight. Test
substances were dissolved in water/0.3% Tween-80 and admini-
stered orally in a volume of 10 mL/kg body weight.
10a:whitesolid,[R]2D0 -0.32ꢀ (c0.95, CHCl3).1HNMR(400MHz,
CDCl3) δ8.25 (s broad, 1H), 7.98 (dd, J= 8.6, 2.4 Hz, 1H), 7.95 (br s,
0.3H), 7.91 (d, J = 2.4 Hz, 0.7 H), 7.43 (dd, J = 13.2, 1.6 Hz, 1H),
7.14 (d, J = 8.9 Hz, 0.7H), 7.10 (d, J = 8.9 Hz, 0.3 H), 4.89-4.75
(m, 1H), 4.00-3.84 (m, 2H), 3.81-3.68 (m, 2H), 3.67-3.54 (m,
2H), 3.48-3.28 (m, 2H), 3.07 (s, 3H), 1.6 (d, J = 6.5 Hz, 0.9H),
1.54-1.50 (m, 2.1 H) as a 7:3 mixture of two rotamers. MS (EI)
m/e: 544.3 (M þ H)þ. Anal. Calcd for C21H20F7N3O4S: C, 46.41;
H, 3.71; N, 7.73; S, 5.90; F, 24.47. Found: C, 46.21; H, 3.63; N,
7.86; S, 6.07; F, 24.42. HPLC, analytical Chiralpack AD column
(flow 1 mL/min, pressure of 25 bar, detection at 254 nm) using
heptane/isopropanol, 80:20, as eluant: retention time of 17.4 min,
ee of >99%. HPLC: retention time of 1.86 min, >98% pure.
10b: white solid, [R]D20 þ0.7ꢀ (c 1, CHCl3). 1H NMR (400 MHz,
CDCl3) δ 8.25 (s broad, 1H), 7.98 (dd, J = 8.6, 2.4 Hz, 1H), 7.95
(brs, 0.3H), 7.91 (d, J = 2.4Hz, 0.7H), 7.43 (dd, J =13.2, 1.6Hz,
1H), 7.14 (d, J = 8.9 Hz, 0.7H), 7.10 (d, J = 8.9 Hz, 0.3 H),
4.89-4.75 (m, 1H), 4.00-3.84 (m, 2H), 3.81-3.68 (m, 2H),
3.67-3.54 (m, 2H), 3.48-3.28 (m, 2H), 3.07 (s, 3H), 1.6 (d,
J = 6.5 Hz, 0.9H), 1.54-1.50 (m, 2.1 H) as a 7:3 mixture of two
rotamers. MS (EI) m/e: 544.3 (M þ H)þ. HPLC, analytical
Chiralpack AD column (flow 1 mL/min, pressure of 25 bar,
detection at 254 nm) using heptane/isopropanol, 80:20, as eluant:
retention time of 22.0 min, ee of >99%. HPLC: retention time of
1.86 min, >98% pure.
Asymmetric Synthesis of 10a. Preparation followed procedure
C from 21e and 19(þ) (78% yield, white solid). [R]2D0 -0.32ꢀ
(c 0.95, CHCl3). 1H NMR (400 MHz, CDCl3) δ 8.25 (s broad,
1H), 7.98 (dd, J = 8.6, 2.4 Hz, 1H), 7.95 (br s, 0.3H), 7.91 (d, J =
2.4 Hz, 0.7 H), 7.43 (dd, J = 13.2, 1.6 Hz, 1H), 7.14 (d, J = 8.9
Hz, 0.7H), 7.10 (d, J = 8.9 Hz, 0.3 H), 4.89-4.75 (m, 1H),
4.00-3.84 (m, 2H), 3.81-3.68 (m, 2H), 3.67-3.54 (m, 2H),
3.48-3.28 (m, 2H), 3.07 (s, 3H), 1.6 (d, J = 6.5 Hz, 0.9H),
1.54-1.50 (m, 2.1 H) as a 7:3 mixture of two rotamers. MS (EI)
m/e: 544.3 (M þ H)þ. Anal. Calcd for C21H20F7N3O4S: C,
46.41; H, 3.71; N, 7.73; S, 5.90; F, 24.47. Found: C, 46.21; H,
3.63; N, 7.86; S, 6.07; F, 24.42. HPLC, analytical Chiralpack
Experiment. A computerized Digiscan 16 animal activity
monitoring system (Omnitech Electronics, Columbus, OH)
was used to quantify locomotor activity. Data were obtained
simultaneously from eight Digiscan activity chambers placed in
a soundproof room with a (12 h light)/(12 h dark) cycle.
Experiments were performed during the light phase between
6:30 a.m. and 5:00 p.m. Each activity monitoring chamber consisted
of a Plexiglas box (41 cm ꢀ 41 cm ꢀ 28 cm; W ꢀ L ꢀ H) with
sawdust bedding on the floor surrounded by invisible horizontal and

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12 4613
vertical infrared sensor beams. The chambers were divided by a
Plexiglas cross, providing each mouse with 20 cm ꢀ 20 cm of moving
space. Two animals per box were monitored simultaneously. Cham-
bers were connected to a Digiscan analyzer linked to a computer that
constantly collected the beam status information. The activity
detector operates by counting the number of times the beams change
from uninterrupted to interrupted status or vice versa. Records of
photocell beam interruptions for individual animals were taken every
5 min over the duration of the experimental session. Mice were first
treated with compounds administered po, and 15 min later, they
received an sc injection of 50 mg/kg of L-687,414. Mice were then
transferred from their home cages to the recording chambers for a
15 min habituation phase, allowing them free exploration of the new
environment. Horizontal activity was then recorded for a 60 min time
period. Data were analyzed using a Kruskal-Wallis ANOVA
followed by a Mann-Whitney t test. For dose-response experi-
ments the horizontal activity value for each group of animals at a
given dose of GlyT1 inhibitor (y1) was expressed as percent of
L-687,414-induced hyperlocomotion and calculated according to
the equation [(y1 - vehicle horizontal activity)/(L-687,414 hori-
zontal activity - vehicle horizontal activity)] ꢀ 100. ID50 values,
defined as doses of each compound producing 50% inhibition of
L-687,414-induced hyperlocomotion, were calculated by linear
regression analysis of the dose-response data using an Excel
based computer curve-fitting program.
Microdialysis Assay in Rat. Animals and Housing Conditions.
Adult male Sprague-Dawley rats weighing 220-250 g were
used. Rats were housed in groups of three in Makrolon cages
with sawdust bedding and were maintained under an artificial
(12 h light)/(12 h dark) cycle (light onset at 6 a.m.) in a
temperature and humidity controlled environment. Food and
water were available ad libitum. The experimental procedure
received prior approval from the City of Basel Cantonal Animal
Protection Committee based on adherence to federal and local
regulations on animal maintenance and testing.
Test Substance. Formulation. Compound 10a was suspended
in the vehicle shortly before the administration to the rats. The
composition of the vehicle was 10% methylcellulose, 15% vitamin
E TPGS, 0.18% methyl 4-hydroxybenzoate, and 0.025% propyl
4-hydroxybenzoate in demineralized water.
Surgery. At 30 min before anesthesia, the rat received 0.025 mg/kg
(sc) buprenorfine. The anesthesia was begun by placing the rat in
an anesthesia box with isoflurane, 3-5% (vol %). The rat was
then placed in a stereotaxic device equipped with dual mani-
pulators arms and an anesthetic mask, and the anesthesia was
maintained with isoflurane, 0.8-1.3 vol % (spontaneous respira-
tion; support gas oxygen/air, 2:1). The head was shaved and the
skin was cut along the midline to expose the skull from several
millimeters anterior and posterior to the bregma and lambda. A
small bore hole was made in the skull to allow the stereotaxical
insertion of the microdialysis probe (BAS5/4, carrying a 4 mm
polyacrylonitrile dialysis membrane, cutoff of 30 KD; BAS
Bioanalytical Systems Inc.) in the striatum (coordinates A 0.2 mm,
L2.9 mm, V7 mm). The probe was cemented into place using binary
dental cement. Once the cement was firm, the wound was closed
with silk thread for suture (Silkam) and the animal was removed
from the stereotaxic instrument and returned to its cage. After the
surgery, rats were treated with meloxicam, 1 mg/kg sc, and allowed
to recover for 3-4 days before the microdialysis experiment. The
body weight of the animals was measured before the surgery and in
the following days to monitor the recovery of the rat from surgery.
Microdialysis Experiment. On the day of the experiment, the
inlet of the implanted dialysis probe was connected to a micro-
perfusion pump (Carnegie Medicine) and the outlet was con-
nected to a fraction collector. The microdialysis probe was then
perfused with Ringer solution (146.5 mM NaCl, 2.7 mM KCl,
1.2 mM CaCl2, 0.85 mM MgCl2, 1.2 mM Na2HPO4, 0.27 mM
NaH2PO4, pH 7.4) at a constant flow rate of 0.5 μL/min, and
perfusates were collected in 20 min aliquots (volume of 10 μL).
During the experiment, the animal was free to move in its cage
and had free access to water and food. The levels of glycine and
other aminoacids in the dialysate became stable after 20-60 min
of perfusion. The first samples of perfusate were collected to
determine the baseline (pretreatment) levels of glycine. Then the
rat received either vehicle or 10a (dose of 10 mg/kg; oral gavage
of compound resuspended shortly before administration in the
vehicle, 10% methylcellulose, 15% vitamin E TPGS, 0.18%
methyl 4-hydroxybenzoate, propyl 4-0.025% hydroxybenzoate
in demineralized water (volume of administration, 10 mL/kg),
and further samples of perfusate were collected for 3 h. At the
end of the experiment, rats were disconnected and returned to
the animal house. Rats underwent a second microdialysis
experiment 36-40 h after the first experiment. Care was taken
to randomize the rats in the various experimental groups.
Analysis of Microdialysate. The levels of glycine and D/L-
serine in the perfusate were assayed by a modification of the
procedure of Donanzi et al.36 and Smit et al.37 using an HPLC
apparatus equipped with an electrochemical detector. Briefly,
the glycine and D/L-serine present in the perfusate were deriva-
tized using o-phthaldialdehyde (OPA), purified on a reverse
phase column and detected by use of an amperometric detector.
The derivatization procedure and the injection of the sample
were automated by use of a DIONEX ASI-100 refrigerated
autosampler unit. The derivatization procedure consisted of
mixing 20 μL of perfusate with 10 μL of derivatization reagent
(5 mg of OPA, 0.5 mL of sodium sulfite, 0.05 M, 0.5 mL of
ethanol, 9 mL of borate buffer at pH 10.4). This solution was
always prepared shortly before use and incubated for 4 min at
4 ꢀC. The OPA derivatives of glycine and D/L-serine were
separated on a YMC-Pack ODS-AQ (RP-C18) S-3 μm, 120 A
column (length 150 mm, diameter 4 mm). The composition of
the mobile phase was 0.1 M Na2HPO4 4H2O, 0.13 mM Na2ED-
TA H2O, and 10% methanol adjusted at pH 4.8 with orthopho-
3
3
sphoric acid. Glycine and D/L-serine were detected using an
ANTEC DECADE detector equipped with a VT-03 flow cell
and glassy carbon electrodes. Oxidation potential was 0.8 V.
Data Calculation. Data are expressed as picomoles of glycine
or D/L-serine measured in 10 μL/perfusate (this is the volume of
perfusate collected in 20 min of perfusion at 0.5 μL/min). For
each rat, glycine or D/L-serine levels measured in the first five
samples were averaged to provide the baseline. All glycine or
D/L-serine levels measured after vehicle or drug administration
in the same rat were then calculated as the percent of the mean
baseline and plotted vs time. 10a was tested in three rats, and the
results are expressed as the mean ( SEM.
Supporting Information Available: Preparation details of
compounds 25-53, 55, 56 and their intermediates; effect of 2
and selective GlyT2 inhibitor Org-25543 in L-687,414-induced
hyperlocomotion assay; hERG inhibition assay; rat and mon-
key pharmacokinetic measurements; HPLC trace of com-
pounds 10a,b and 25-56. This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) Conn, P. J.; Tamminga, C.; Schoepp, D. D.; Lindsley, C.
Schizophrenia: moving beyond monoamine antagonists. Mol.
Interventions 2008, 8, 99–107.
(2) Javitt, D. C. Glutamate and schizophrenia: phencyclidine,
N-methyl-D-aspartate receptors, and dopamine-glutamate interac-
tions. Int. Rev. Neurobiol. 2007, 78, 69–108.
(3) Danysz, W.; Parsons, C. G. Glycine and N-methyl-D-aspartate
receptors: physiological significance and possible therapeutic
applications. Pharmacol. Rev. 1998, 50, 597–664.
(4) (a) Sur, C.;Kinney, G. G. Glycinetransporter1inhibitors andmodu-
lation of NMDA receptor-mediated excitatory neurotransmission.
Curr. Drug Targets 2007, 8, 643–649. (b) Thomsen, C. Glycine
transporter inhibitors as novel antipsychotics. Drug Discovery Today:
Ther. Strategies 2006, 3, 539–545.
(5) Shim, S. S.; Hammonds, M. D.; Kee, B. S. Potentiation of
the NMDA receptor in the treatment of schizophrenia: focused

4614 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 12
Pinard et al.
on the glycine site. Eur. Arch. Psychiatry Clin. Nerurosci. 2008, 258,
16–27.
Inhibitors. Abstracts of Papers, 231st National Meeting of the American
Chemical Society, Atlanta, GA, March 26-30, 2006; American Chemi-
cal Society: Washington, DC, 2006; MEDI-407.
(6) (a) Javitt, D. C. Glycine transport inhibitors for the treatment of
schizophrenia: symptom and disease modification. Curr. Opin.
Drug Discovery Dev. 2009, 12, 468–478. (b) Bridges, T. M.; Williams,
R.; Lindsley, C. W. Design of potent GlyT1 inhibitors: in vitro and in
vivoprofiles. Curr. Opin. Mol. Ther. 2008, 10, 591–601. (c) Hashimoto,
K. Glycine transporter-1 inhibitors as novel therapeutic drugs for
schizophrenia. Cent. Nerv. Syst. Agents Med. Chem. 2007, 7, 177–
182. (d)Harsing, L. G.;Juranyi, Z.;Gacsalyi, I.;Tapolcsanyi, P.;Czompa,
A.; Matyus, P. Glycine transporter type-1 and its inhibitors. Curr. Med.
Chem. 2006, 13, 1017–1044. (e) Lindsley, C. W.; Wolkenberg, S. E.;
Kinney, G. G. Progress in the preparation and testing of glycine
transporter type-1 (GlyT1) inhibitors. Curr. Top. Med. Chem. 2006,
6, 1883–1896.
(7) Atkinson, B. N.; Bell, S. C.; De Vivo, M.; Kowalski, L. R.; Lechner,
S. M.; Ognyanov, V. I.; Tham, C.-S.; Tsai, C.; Jia, J.; Ashton, D.;
Klitenick, M. A. ALX 5407: a potent, selective inhibitor of the
hGlyT1 glycine transporter. Mol. Pharmacol. 2001, 60, 1414–1420.
(8) Brown, A.; Carlyle, I.; Clark, J.; Hamilton, W.; Gibson, S.;
McGarry, G.; McEachen, S.; Rae, D.; Thorn, S.; Walker, G.
Discovery and SAR of Org 24598, a selective glycine uptake
inhibitor. Bioorg. Med. Chem. Lett. 2001, 11, 2007–2009.
(9) Smith, G.; Ruhland, T.; Mikkelsen, G.; Andersen, K.; Christoffersen,
C. T.; Alifrangis, L. H.; Mork, A.; Wren, S. P.; Harris, N.; Wyman,
B. M.; Brandt, G. The synthesis and SAR of 2-arylsulfanyl-phenyl
piperazinyl acetic acids as GlyT-1 inhibitors. Bioorg. Med. Chem.
Lett. 2004, 14, 4027–4030.
(10) Egle, I.; Delaney, W.; Wang, Z.; Shumacher, R.; Hopper, A. T.;
Tehim, A.; Maddarod, S. Preparation of N-Propenylsarcosines as
Glycine Transport-1 Inhibitors (GlyT-1) for the Treatment of
Neurological and Neuropsychiatric Disorders. PCT Int. Appl.
WO 2002066456, 2002.
(11) Perry, K. W.; Falcone, J. F.; Fell, M. J.; Ryder, J. W.; Yu, H.; Love,
P. L.; Katner, J.; Gordon, K. D.; Wade, M. R.; Man, T.; Nomikos,
G. G.; Phebus, L. A.; Cauvin, A. J.; Johnson, K. W.; Jones, C. K.;
Hoffmann, B. J.; Sandusky, G. E.; Walter, M. W.; Porter, W. J.;
Yang, L.; Merchant, K. M.; Shannon, H. E.; Svensson, K. A.
Neurochemical and behavioral profiling of the selective GlyT1
(16) Pinard, E.; Alberati, D.; Borroni, E.; Fischer, H.; Hainzl, D.;
Jolidon, S.; Moreau, J. L.; Narquizian, R.; Nettekoven, M.;
Norcross, R. D.; Stalder, H.; Thomas, A. W. Discovery of ben-
zoylpiperazines as a novel class of potent and selective GlyT1
inhibitors. Bioorg. Med. Chem. Lett. 2008, 18, 5134–5139.
(17) Tricklebank, M. D.; Bristow, L. J.; Hutson, P. H.; Leeson, P. D.;
Rowley, M.; Saywell, K.; Singh, L.; Tattersall, F. D.; Thorn, L.;
Williams, B. J. The anticonvulsant and behavioral profile of
L-687,414, a partial agonist acting at the glycine modulatory site
on the N-methyl-D-aspartate (NMDA) receptor complex. Br. J.
Pharmacol. 1994, 113, 729–736.
(18) Alberati D.; Moreau J. L.; Mory R.; Pinard E.; Wettstein J. G.
Unpublished results. See Supporting Information file for the
effects measured with compound 2 and with the selective GlyT2
inhibitor 4-benzyloxy-N-(1-dimethylaminocyclopentylmethyl)-
3,5-dimethoxybenzamide (Org-25543) on the L-687,414-induced
hyperlocomotion in mice.
(19) Caulfield, W. L.; Collie, I. T.; Dickins, R. S.; Epemolu, O.;
McGuire, R.; Hill, D. R.; McVey, G.; Morphy, J. R.; Rankovic,
Z.; Sundaram, H. The first potent and selective inhibitors of the
glycine transporter type 2. J. Med. Chem. 2001, 44, 2679–2682.
(20) Jolidon, S.; Narquizian, R.; Nettekoven, M. H.; Norcross, R. D.;
Pinard, E.; Stalder, H. Preparation of Alkoxybenzoylpiperazines
as Inhibitors of Glycine Transporter 1 (GlyT-1). PCT Int. Appl.
WO 2005014563, 2005.
(21) On November 10, 2009, F. Hoffmann-La Roche Ltd. reported
positive phase II results with RG1678, a potential first-in-class
treatment for the negative symptoms of schizophrenia. See Investor
Update, November 10, 2009, at www.roche.com.
(22) Clark, M. T.; Coburn, R. A.; Evans, R. T.; Genco, R. J.
5-(Alkylsulfonyl)salicylanilides as potential dental antiplaque
agents. J. Med. Chem. 1986, 29, 25–29.
(23) Doswald, S.; Hanlon, S. P.; Kupfer, E. Asymmetric Reduction of
1,1,1-Trifluoroacetone by Saccharomyces cerevisiae. U.S. Patent
Appl. US009999, 2007.
(24) Yamanaka, H.; Takekawa, T.; Morita, K.; Ishihara, T. Prepara-
tion of novel β-trifluoromethyl vinamidinium salt and its synthetic
application to trifluoromethylated heterocycles. Tetrahedron Lett.
1996, 37, 1829–1832.
(25) Mangalagiu, I.; Benneche, T.; Undheim, K. Trialkylalanes in
palladium-catalyzed chemo- and regioselective alkylations. Tetra-
hedron Lett. 1996, 37, 1309–1312.
(26) Lipophilicity was calculated with the clogP, version 4.94, program
(BioByte Corp., Claremont, CA, http://www.biobyte.com).
(27) Gerber Molecular Design, Amden, Switzerland, http://www.moloc.ch.
(28) Fischer, H.; Kansy, M. Automated Generation of Multi-Dimen-
sional Structure Activity and Structure Property Relationships.
PCT Int. Appl. US 2007027632, 2007.
(29) Jamieson, C.; Moir, E. M.; Rankovic, Z.; Wishart, G. Medicinal
chemistry of hERG optimizations: highlights and hang-ups.
J. Med. Chem. 2006, 49, 5029–5046.
(30) For a recent review on the molecular basis of hERG channel
interaction, see the following: Thai, K. M.; Ecker, G. F. Predictive
models for hERG channel blockers: ligand-based and structure-
based approaches. Curr. Med. Chem. 2007, 14, 3003–3026.
(31) Shengguo, S; Adejare, A. Fluorinated molecules as drugs and imaging
agents in the CNS. Curr. Top. Med. Chem. 2006, 14, 1457–1564.
(32) Boehm, H.-J.; Banner, D.; Bendels, S.; Kansy, M.; Kuhn, B.;
Mueller, K.; Obst-Sander, U.; Stahl, M. Fluorine in medicinal
chemistry. ChemBioChem 2004, 5, 637–643.
(33) The panel consisted of targets included in the standard CEREP
high throughput profile screen complemented with enzymes
(phosphodiesterases, kinases, MAO-A, COMT, acetylcholinees-
terase): http://www.cerep.fr.
(34) Kansy, M.; Senner, F.; Gubernator, K. Physicochemical high
throughput screening: parallel artificial membrane permeation
assay in the description of passive absorption processes. J. Med.
Chem. 1998, 41, 1007–1010.
(35) Pinard, E.; Burner, S.; Cueni, P.; Montavon, F.; Zimmerli, D. A
short and efficient synthesis of the NMDA glycine site antagonist:
(3R,4R)-3-amino-1-hydroxy-4-methyl pyrrolidin-2-one (L-687,414).
Tetrahedron Lett. 2008, 49, 6079–6080.
(36) Donzanti, B. A.; Yamamoto, B. K. An improved and rapid method
for the isocratic separation of aminoacid neurotransmitters from
brain tissue and microdialysis perfusate. Life Sci. 1988, 43, 913–922.
(37) Smit, S.; Sharp, T. Measurement of GABA in rat brain micro-
dialysate using o-phthalaldialdehyde-sulphite derivatization and
high performance liquid chromatography with electrochemical
detection. J. Chromatogr. 1994, 652, 228–233.
inhibitors ALX5407 and LY2365109 indicate a preferential
action in caudal vs cortical brain areas. Neuropharmacology 2008,
55, 743–754.
(12) Dargazanli, G.; Estenne Bouhtou, G.; Magat, P.; Marabout, B.;
Medaisko, F.; Roger, P.; Sevrin, M.; Veronique, C. FR Patent
Appl. FR2838739, 2003.
(13) Wolkenberg, S. E. Glycine Transporter 1 (GlyT1) Inhibitors for the
Treatment of Schizophrenia. Abstracts of Papers, 236th National
Meeting of the American Chemical Society, Philadelphia, PA, August
17-21, 2008; American Chemical Society: Washington, DC, 2008;
MEDI-220.
(14) Lowe, J. A.; Hou, X.; Schmidt, C.; David, T., F.; McHardy, S.;
Kalman, M.; DeNinno, S.; Sanner, M.; Ward, K.; Lebel, L.;
Tunucci, D.; Valentine, J. The discovery of a structurally novel
class of inhibitors of the type 1 glycine transporter. Bioorg. Med.
Chem. Lett. 2009, 19, 2974–2976.
(15) (a) Pinard, E.; Ceccarelli, S. M.; Stalder, H.; Alberati, D. Discovery
of N-(2-aryl-cyclohexyl) substituted spiropiperidines as a novel
class of GlyT1 inhibitors. Bioorg. Med. Chem. Lett. 2006, 16,
349–353. (b) Ceccarelli, S. M.; Pinard, E.; Stalder, H.; Alberati, D.
Discovery of N-(2-hydroxy-2-aryl-cyclohexyl) substituted spiropiper-
idines as GlyT1 antagonists with improved pharmacological profile.
Bioorg. Med. Chem. Lett. 2006, 16, 354–357. (c) Alberati, D.;
Ceccarelli, S. M.; Jolidon, S.; Krafft, E. A.; Kurt, A.; Maier, A.; Pinard,
E.; Stalder, H.; Studer, D.; Thomas, A. W.; Zimmerli, D. Design and
synthesis of 4-substituted-8-(2-phenyl-cyclohexyl)-2,8-diaza-spiro-
[4.5]decan-1-one as a novel class of GlyT1 inhibitors: achieving
selectivity against the μ opioid and nociceptin/orphanin FQ peptide
(NOP) receptors. Bioorg. Med. Chem. Lett. 2006, 16, 4305–4310. (d)
Alberati, D.; Hainzl, D.; Jolidon, S.; Krafft, E. A.; Kurt, A.; Maier, A.;
Pinard, E.; Thomas, A. W.; Zimmerli, D. Discovery of 4-substituted-
8-(2-hydroxy-2-phenyl-cyclohexyl)-2,8-diaza-spiro[4.5]decan-1-one as
a novel class of highly selective GlyT1 inhibitors with improved meta-
bolic stability. Bioorg. Med. Chem. Lett. 2006, 16, 4311–4315. (e)
Alberati, D.; Hainzl, D.; Jolidon, S.; Kurt, A.; Pinard, E.; Thomas, A. W.;
Zimmerli, D. 4-Substituted-8-(1-phenyl-cyclohexyl)-2,8-diaza-spiro-
[4.5]decan-1-one as a novel class of highly selective GlyT1 inhibitors
with superior pharmacological and pharmacokinetic parameters. Bioorg.
Med. Chem. Lett. 2006, 16, 4321–4325. (f) Pinard, E.; Alberati, D.;
Borroni, E.; Ceccarelli, S. M.; Fischer, H.; Hainzl, D.; Jolidon, S.;
Moreau, J. L.; Stalder, H.; Thomas, A. W. Design, Synthesis and
Structure-Activity Relationship of N(1)-(2-Phenyl-cylohexyl)-4-amino-
piperidine Derivatives as Potent and Selective Glycine Reuptake