Optimisation of a series of potent, selective and orally bioavailable GlyT1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2009.02.102

A series of heterocyclic sulfonamides have been developed which are potent and selective inhibitors of hGlyT1. SAR studies to optimise the in vitro and in vivo properties are described. Optimisation of the central scaffold resulted in cyclohexane sulfones

Full text of DOI:10.1016/j.bmcl.2009.02.102

Bioorganic & Medicinal Chemistry Letters 19 (2009) 2235–2239
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Optimisation of a series of potent, selective and orally bioavailable
GlyT1 inhibitors
Joanne L. Thomson, Wesley P. Blackaby, Andrew S. R. Jennings, Simon C. Goodacre, Andrew Pike,
*
Steve Thomas, Terry A. Brown, Alison Smith, Gopalan Pillai, Leslie J. Street, Richard T. Lewis
Department of Medicinal Chemistry, Merck Sharp and Dohme, Neuroscience Research Centre, Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of heterocyclic sulfonamides have been developed which are potent and selective inhibitors of
hGlyT1. SAR studies to optimise the in vitro and in vivo properties are described. Optimisation of the cen-
tral scaffold resulted in cyclohexane sulfones 28 and 29, which have good PK properties and show prom-
ise for further development.
Received 10 February 2009
Revised 23 February 2009
Accepted 24 February 2009
Available online 28 February 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Schizophrenia
Glycine transporter
GLYT1
Bioavailable
Inhibitor
NMDA
Pharmacokinetics
Glycine plays a major role as a neurotransmitter in the mamma-
lian central nervous system (CNS). In the forebrain, glycine acts as
a necessary co-agonist with glutamate at N-methyl-D-aspartate
viable method of potentiating NMDA receptor function in vivo
and hence ameliorating the negative symptoms of schizophrenia.
In this communication we describe the development of a series
of potent and selective inhibitors of GlyT1.14
(NMDA) receptors, thereby modulating NMDA-dependent excit-
atory neurotransmission.1–4 The concentration of glycine at the
synapse is regulated primarily by re-uptake via glycine transport-
ers, of which two, GlyT1 and GlyT2, have been characterized.2–6
Propyl sulfonamide 1, a potent and selective GlyT1 inhibitor,
evolved from ‘hit-to-lead’ optimisation of a compound identified
from an HTS screen of the Merck sample collection.14–16 This com-
pound exhibited excellent in vivo occupancy when dosed i.p., and
microdialysis studies showed that it increased levels of glycine in
both the hippocampus and cortex.16,17 However, this compound
was compromised by very low plasma exposures on p.o. dosing,
which correlated with high turnover in rat liver microsomes. Pre-
liminary studies indicated that oxidation of the propyl side chain
was a significant metabolic pathway. Attempts to find a more met-
abolically stable replacement for the propyl sulfonamide were hin-
dered by very tight SAR, which indicated a lack of tolerance to
substitution on the chain or replacement with more bulky substit-
uents. Intriguingly, compound 2, in which the propyl chain was
replaced with cyclopropyl methyl, retained good potency. This
led us to speculate that replacement of the cyclopropylmethyl moi-
These belong to the Na+/ClÀ dependent family, which includes
c-
aminobutyric acid, taurine and proline transporters.2 GlyT2, which
is localised largely in the brainstem and cerebellum, is responsible
for glycine uptake at glycinergic synapses. GlyT1, which is located
throughout the brain and is predominantly expressed by glial cells,
is involved in regulating glycine concentration in the vicinity of
NMDA receptor-expressing synapses.7 Hypofunction of NMDA
receptors has been implicated in the pathophysiology of schizo-
phrenia, with evidence coming from both pre-clinical models and
clinical data.8–13 For example, sarcosine, an endogenous inhibitor
of GlyT1, has been shown to provide significant improvements in
the positive, negative and cognitive symptoms of stable schizo-
phrenics when used as an adjunct to standard antipsychotic ther-
apy.10 This suggests that the pharmacological manipulation of
synaptic glycine concentration using a GlyT1 inhibitor may be a
ety with a small p-system, for example a five-membered aromatic
heterocycle, might be tolerated and afford an improved metabolic
profile. A diverse range of heterocyclic sulfonamides were synthes-
ised by treatment of the piperidine intermediate with the appro-
priate sulfonyl chloride in the presence of base (Scheme 1)
utilising a RAS strategy with commercially available sulfonyl
* Corresponding author at present address: Amgen Inc., One Kendall Square,
Building 1000, Cambridge, MA 02139, USA. Tel.: +1 617 444 5232.
E-mail address: riclewis@amgen.com (R.T. Lewis).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.02.102

2236
J. L. Thomson et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2235–2239
Table 2
F
F
O
Cl
O
Cl
Potency at GlyT1 and % microsomal turnover for pyrazoles 11–13
Cl
O
N
N
H
N
N
H
i
R
N
Cl
H
Cl
Cl
N
H
N
S
R
O
O
n
n
N
O
S O
i
Scheme 1. Reagents and condition: (i) RSO2Cl, Pr2NEt, DCM, rt.
N N
Compound
R1
n
hGlyT1 IC50 (nM)a
% Turnoverb
D
R
H
F
11
12
1
2
1
2.95
10.35
3.62
56
<5
38
70
25
Table 1
N
Potency at GlyT1 and % microsomal turnover for compounds 1–10
F
O
Cl
5
10
21
N
N
H
Cl
N
S
R
O
O
13
17
hGlyT1 IC50 (nM)a
% Turnoverb
D
a,b
Compound
R
See Table 1.
R
H
chlorides (Table 1).15 Methyl imidazole 4, which showed improved
potency and microsomal turnover relative to compound 1, emerged
as a lead for further optimisation. Further substitution on the imid-
azole ring was not tolerated, as demonstrated by compound 5.
Methyl pyrazole 3 showed good potency but higher turnover in
rat liver microsomes relative to compound 4, although it was still
an improvement over initial lead 1. Although pyridine 8 was po-
tent at the GlyT1 receptor, it was compromised by very high turn-
over in liver microsomes, which metabolism studies revealed was
due to N-oxidation. Other six-membered aromatic systems were
not tolerated (compounds 9 and 10). Concurrently, we investigated
azetidine as an alternative scaffold, which we anticipated might of-
fer more steric tolerance to variation in the propyl sulfonamide
side chain. The heterocyclic side chain modification gave potent
compounds in this series (Table 2). In particular, the azetidine ana-
logue of pyrazole 3, compound 11, was potent at the GlyT1 recep-
tor, although turnover in rat liver microsomes remained high;
metabolism studies indicated N-dealkylation to be a major route.
Replacement of the fluoropyridine moiety of piperidine 3 with a
cyclopropyl methyl group gave 12, with good potency at the GlyT1
receptor. Surprisingly, this compound displayed low turnover and
no apparent N-dealkylation in rat liver microsomes, in sharp con-
trast to the 3-fluoropyridine analogue 3. Combining these two
changes resulted in the potent methyl pyrazole 13. Microsomal
turnover in rat was moderate but this compound possessed accept-
able pharmacokinetics in rat (F = 48%; Clp = 46 ml/min/kg;
1
2
3
4
4.33
2
73
25
20
24
21
23
78
55
30
21
42
—
5.43
1.74
N N
N
N
5
6
31%@3
124.6
l
M
—
—
—
—
N
N
S
98
72
S
7
8
9
20%@3
11.55
l
M
—
—
—
Cl
99
96
99
—
N
Cl
Cl
O
O
38%@3
33%@3
l
M
M
—
—
—
—
R1
R1
N
N
Cl
Cl
H
H
i or ii
n
n
n
n
N
S
N
S
10
l
—
O
O
O
O
3 N
3 N
a
R2
See Ref. 15. For cases where <50% inhibition was achieved at 3
l
M, values are
N
N
1
N
NH
1
2
expressed as %inhibition.
2
b
Turnover of GlyT1 compounds (1 lM) in rat, dog and human liver microsomes.
All incubations were carried out at 37 °C for 15 min. Protein concentration = 0.5 mg/
ml; cosolvent = 0.99% MeCN + 0.01% DMSO.
Scheme 2. Reagents and conditions: (i) N,N-dimethylformamide dimethyl acetal,
toluene, reflux; (ii) NaH or K2CO3, R2I, DMF, rt.

J. L. Thomson et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2235–2239
2237
Table 3
Potency at GlyT1 and % microsomal turnover for triazoles 14–26
Cl
O
R1
N
Cl
H
n
n
N
O S O
R2
Compound
R1
R2
n
hGlyT1 IC50 (nM)a
% Turnoverb
D
R
H
14
1
0.1
88
24
32
N
N
N
N
15
16
17
1
1
1
131.7
184.9
870.8
—
—
—
—
—
—
—
—
—
N
N
N
N
N
N
N
N H
N
18
19
20
2
1
2
.32
.31
.65
88
15
6
34
5
43
22
20
N
N
F
F
N
N
N
N
N
17
N
N
N
N
N
21
22
1
2
2.58
65
—
44
—
45
—
N
N
N
N
28.22
N
N
23
2
501.6
314.3
—
—
—
N
F3C
24
25
2
2
—
—
—
—
—
—
N
N
N
N
13%@3
123.6
lM
N
N
26
2
—
—
—
N
N
N
a,b
See Table 1.

2238
J. L. Thomson et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2235–2239
Cl
Cl
Cl
O
O
O
CN
i-v
N
N
H
vi-viii
N
H
Cl
H
Cl
Cl
+
O
O
OH
O2S
O2S
N
N
N
N
N
N
27
28
29
Scheme 3. Reagents and conditions: (i) KHMDS, toluene, cyclopropylmethyl bromide, À78 °C; (ii) Lithium aluminium hydride, Et2O, À78 °C to rt; (iii) 2,4-dichlorobenzoyl
i
chloride, Pr2NEt, DCM, 0 °C to rt; (iv) HCl, THF, rt; (v) NaBH4, EtOH, rt; (vi) MsCl, pyridine, 0 °C; (vii) 1-methyl-1H-[1,2,3]triazole-4-thiol., NaBH4, EtOH, 50 °C; (viii) oxone,
acetone, water, reflux.
Table 4
Potency at GlyT1, microsomal turnover, rat pharmacokinetic properties and in vivo occupancy for compounds 28 and 29
Compound
Stereochemistry
IC50 (nM)a
% Turnoverb
H
Pharmacokineticsc
Clp (ml/min/kg)
Occ50 (rat) (mg/kg)
R
F (%)
t1/2 (h)
Vdss (l/kg)
28
29
Sulfonyl and amide trans
Sulfonyl and amide cis
1.89
5.44
23
10
28
28
59
105
1.8
1.1
16
34
2.3
2.6
2.2
3.4
a,b
See Table 1.
See Ref. 18002E
c
t
= 2.4 h) with excellent occupancy (Occ50 = 3.8 mg/kg p.o.) in the
ment with the triazole thiolate anion to give two sulfide
geometric isomers in a 1:1 ratio, which were subsequently oxi-
dised to the sulfones 28 and 29.19 As shown in Table 4, both iso-
mers showed excellent potency and occupancy in vivo with
acceptable pharmacokinetics in rat. Turnover was low to moderate
in both rat and human liver microsomes. SAR studies involving
½
mouse in vivo binding assay.16
Encouraged by these results, the analogous 1,2,4- and 1,2,3-tri-
azoles were targeted. These compounds were accessed by sulfon-
amide formation as before to give the unsubstituted triazoles,
which were subsequently methylated using N,N-dimethylformam-
ide dimethyl acetal (Scheme 2).
cyclohexane analogues will be the subject of
communication.
a subsequent
For the 1,2,3-triazoles, this method allowed access to the N2-
and N3-methylated products but N1-methylation occurred with
concomitant formylation at the 5-position. However, alkylation
with methyl iodide in the presence of base gave the N1 and N2 iso-
mers only, in approximately a 1:3 ratio. The N-ethyl analogues
were similarly prepared. The 1,2,3-triazoles were more potent than
the 1,2,4-triazoles; the N1-methylated isomers were consistently
the most active, (see Table 3). For example, cyclopropyl methyl
piperidine 18 showed excellent potency and occupancy in vivo
(Occ50 < 1 mpk (mouse)). This compound had reasonable pharma-
cokinetic properties in rat, with moderate clearance (39 ml/min/
kg), volume of distribution (2.2 l/kg) and bioavailability (27%) but
a fairly short half-life (1.1 h).18 Compound 18, like all the cyclopro-
pyl methyl triazole analogues, exhibited very high turnover in rat
liver microsomes due to N-dealkylation. In an attempt to mitigate
this pathway, the 2,2,2-trifluoroethyl analogue 23 was accessed by
alkylation with 2,2,2-trifluoroethyl trifluoro-methanesulfonate.
However, compound 23 showed poor potency compared with its
methyl and ethyl analogues. Interestingly, high turnover was not
seen with the corresponding fluoropyridine analogues. Thus, com-
pound 20 retained excellent potency and showed good in vivo
occupancy (Occ50 = 7.5 mg/kg (mouse)) with significantly de-
creased turnover in rat liver microsomes. However, this compound
had a very short half-life (0.3 h), low bioavailability (3%) and high
clearance in rat.18 Routes of metabolism studies carried out on 20
indicated oxidation of the piperidine ring and no observed N-deal-
kylation. It is clear from our studies that in optimising the meta-
bolic fate of the molecule, one needs to consider the gross
structure and not simply the moiety undergoing metabolic modifi-
cation; this is to be expected from the molecular recognition com-
ponent of P450 mediated oxidation. To address the oxidative
metabolism of the piperidine moiety of 18, cyclohexane analogues
28 and 29 were prepared, as shown in Scheme 3. The sulfone moi-
ety was constructed from the mesylate of alcohol 27 by displace-
In conclusion, the replacement of the cyclopropylmethyl group of
2 with heterocycles gave a novel series of potent and selective GlyT1
inhibitors. Compounds such as 20 and 13 displayed excellent po-
tency and in vivo occupancy upon oral dosing, with reduced meta-
bolic turnover compared with the initial lead. Cyclohexane
analogues 28 and 29 also showed improvements over the initial lead
and benefitedfroma lackof oxidativemetabolismon thecentral ring
which translated into improved pharmacokinetic properties in rat.
References and notes
1. Sur, C.; Kinney, G. G. Exp. Opin. Invest. Drugs 2004, 13, 515.
2. Bergeron, R.; Meyer, T. M.; Coyle, J. T.; Greene, R. W. Proc. Natl. Acad. Sci. U.S.A.
1998, 95, 15730.
3. Danysz, W.; Parsons, C. G. Pharmacol. Rev. 1998, 50, 597.
4. Johnson, J. W.; Ascher, P. Nature 1987, 325, 529.
5. Borowsky, B.; Mezey, E.; Hoffman, B. J. Neuron 1993, 10, 851.
6. Smith, K. E.; Borden, L. A.; Hartig, P. R.; Branchek, T.; Weinshank, R. L. Neuron
1992, 8, 927.
7. Zafra, F.; Gomeza, J.; Olivares, L.; Aragon, C.; Gimenez, C. Eur. J. Neurosci. 1995,
7, 1342.
8. Olney, J. W.; Newcomer, J. W.; Farber, N. B. J. Psychiat. Res. 1999, 33, 523.
9. Javitt, D. C.; Zylberman, I.; Zukin, S. R.; Heresco-Levy, U.; Lindenmayer, J. P. Am.
J. Psychiatry 1994, 151, 1234.
10. Tsai, G.; Lane, H. Y.; Yang, P.; Chong, M. Y.; Lange, N. Biol. Psychiatry 2004, 55,
452.
11. Heresco-Levy, U. Int. J. Neuropsychopharmacol. 2000, 3, 243.
12. Jentsch, J. D.; Roth, R. H. Neuropsychopharmacology 1999, 20, 201.
13. Leiderman, E.; Zylberman, I.; Zukin, S. R.; Cooper, T. B.; Javitt, D. C. Biol.
Psychiatry 1996, 39, 213.
14. Selectivity over hGLYT2 appears to be inherent in this class of inhibitor. For
example, compounds 3, 4, 12, 13, 18, and 20 were all assayed at >10
lM
against hGLYT2. See also Williams, J. B.; Mallorga, P. J.; Lemaire, W.; Williams,
D. L.; Na, S.; Patel, S.; Conn, J. P.; Pettibone, D. J.; Austin, C.; Sur, C. Anal. Biochem.
2003, 321, 31.
15. Blackaby, W. WO 2005094514, 2005.
16. (a) Zhao, Z.; Leister, W. H.; O’Brien, J. A.; Lemaire, W.; Williams Jr., D. L.;
Jacobson, M. A.; Sur, C.; Kinney, G. G.; Pettibone, D. J.; Tiller, P. R.; Smith, S.;
Hartman, G. D.; Lindsley, C. W.; Wolkenberg, S. E. Bioorg. Med. Chem. Lett. 2009,

J. L. Thomson et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2235–2239
2239
19(5), 1488–1491; (b) Wolkenberg, S. E.; Zhao, Z.; Wisnoski, D. D.; Leister, W.
H.; O’Brien, J. A.; Lemaire, W.; Williams Jr., D. L.; Jacobson, M. A.; Sur, C.;
Kinney, G. G.; Pettibone, D. J.; Tiller, P. R.; Smith, S.; Gibson; Ma, B. K.; Polsky-
Fisher, S. L.; Lindsley, C. W.; Hartman, G. D. Bioorg. Med. Chem. Lett. 2009, 19(5),
1492–1495.
17. Whitehead, K. J.; Manning, J. P.; Smith, C. G. S.; Bowery, N. G. Brain Res. 2001,
910, 192.
18. Determined in 6 male Sprague-Dawley rats, 3 dosed i.v. (0.1 mg/kg) and 3
dosed p.o (1 mg/kg).
19. Begtrup, M. Acta Chemica Scandinavica (1947–1973) 1972, 26, 1243.