The synthesis and SAR of 2-arylsulfanylphenyl-1-oxyalkylamino acids as GlyT-1 inhibitors

Article abstract of DOI:10.1016/j.bmcl.2006.05.017

Elevation of glycine levels by inhibition of the glycine transporter-1 (GlyT-1) and activation of the NMDA receptor is a potential strategy for the treatment of schizophrenia. A novel series of 2-arylsulfanylphenyl-1-oxyalkyl amino acids have been identif

Full text of DOI:10.1016/j.bmcl.2006.05.017

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3981–3984
The synthesis and SAR of 2-arylsulfanylphenyl-1-oxyalkylamino
acids as GlyT-1 inhibitors
Garrick Smith,a,* Gitte Mikkelsen,a Jørgen Eskildsena,ꢀ and Christoffer Bundgaardb
aMedicinal Chemistry Research, H. Lundbeck A/S, 9 Ottiliavej, DK 2500 Valby, Denmark
bDrug ADME Research, H. Lundbeck A/S, 9 Ottiliavej, DK 2500 Valby, Denmark
Received 3 April 2006; revised 4 May 2006; accepted 4 May 2006
Available online 24 May 2006
Abstract—Elevation of glycine levels by inhibition of the glycine transporter-1 (GlyT-1) and activation of the NMDA receptor is a
potential strategy for the treatment of schizophrenia. A novel series of 2-arylsulfanylphenyl-1-oxyalkyl amino acids have been iden-
tified. The most prominent member of this series S-1-{2-[3-(3-fluoro-phenylsulfanyl)biphenyl-4-yloxy]ethyl}pyrrolidine-2-carboxylic
acid (38) is a potent GlyT-1 inhibitor (IC50 = 59 nM). In vitro and in vivo assessment of CNS exposure indicates this compound is a
likely substrate for active efflux transporters.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Glycine is a major neurotransmitter with both excit-
atory and inhibitory roles in the central nervous sys-
tem. Extracellular levels are in part governed by the
glycine transporters1 of which there are two subtypes,
the glycine transporter-1 (GlyT-1) and the glycine
transporter-2 (GlyT-2). Five GlyT-1 splice variants
are known (a,b,c,d,e), whilst two have been identified
for GlyT-2 (a,b). Blockade of the GlyT-1 elevates gly-
cine levels and enhances glutamatergic neurotransmis-
sion due to the coagonistic action of glycine at the
NMDA receptor2 which is thought to be of benefit
in schizophrenia.3
develop safe and effective therapies based on GlyT-1
blockade.5,6
We have previously reported the identification of the
biaryl thioether piperazine acetic acids,7 for example, 1
which were based on the first GlyT-1 sarcosine-based
inhibitors reported in the literature.8,9 In this letter, we
wish to report further optimisation of 1 in which we
replaced the piperazine linker with an oxyalkylamino
chain as shown by the generic structure 2. The objective
of this work was to improve further the potency against
GlyT-1 and explore the role of the amino acid fragment
in the interaction with GlyT-1.
Further understanding of the physiological role of the
GlyT-1 has come from gene deletion studies.4 Heterozy-
gote GlyT-1 +/À mice display behaviour characteristic
of enhanced NMDA receptor response, whilst homozy-
gote GlyT-1 À/À mice display neonatal lethality appar-
ently due to sustained activation of the inhibitory
glycine receptor. There is at present widespread interest
in the pharmaceutical industry in understanding how to
Three synthetic approaches were taken to prepare the
compounds of the general structure 2. Variations in
R1 and R2 were achieved in a parallel fashion as shown
O
O
R1
OH
OH
R3
N
N
O
R2
N
Keywords: Glycine transporter-1; GlyT-1 inhibitor; Amino acids;
Schizophrenia.
S
S
*
Corresponding author. Tel.: +45 3643 3226; fax: +45 3643 8237;
e-mail: gsm@lundbeck.com
R4
R5
O
ꢀ
Present address: ACADIA Pharmaceuticals; Medeon Science Park,
S-205 13 Malmo¨, Sweden.
2
1
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.05.017

3982
G. Smith et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3981–3984
in Scheme 1. N-Alkyl amino ethanol was used to alkyl-
ate tert-butyl bromoacetate derivative 3 under basic
conditions and the resulting alcohol 4 was reacted with
the appropriately substituted 2-(arylsufanyl)phenol10
under Mitsunobu conditions. Cleavage of the tert-butyl
ester gave the final product, 5. Examination of R3 was
achieved also in a parallel fashion as shown in Scheme
2. Bromo tert-butyl acetate 6 was reacted with tert-
butyldimethyl silyl (TBDMS) protected alcohols and
the secondary amine 7 was alkylated with methyl
iodide. Removal of the TBDMS group gave the alco-
hol 8. Coupling with the appropriately substituted 2-
(arylsufanyl)phenol under Mitsunobu conditions, fol-
lowed by cleavage of the tert-butyl ester, gave final
product 9.
In order to vary R4 in the S-proline series, a synthetic
sequence was employed as illustrated in Scheme 3. Cou-
pling of 2-methoxy-3-chlorophenyl thiol 10a with 3-flu-
oroiodobenzene gave access to chlorine substitution in
the 6 position. Otherwise appropriately substituted
halo-2-iodo-anisoles 10b11 were coupled with 3-fluoro-
benzenethiol using palladium catalysis to give the biaryl
sulfanyl derivatives 11. Removal of the methyl group
gave the corresponding 2-(3-fluorophenylsulfanyl)-phe-
nol derivatives 12 which were then treated with ethylene
carbonate to give the 2-arylsulfanylphenyloxyethyl alco-
hol 13. Activation to the triflate and then displacement
with D- or L-proline tert-butyl ester gave 14. In the case
of R4 being bromide, Suzuki coupling with aryl boronic
R5
R2
N
O
R2
O
O
OH
O
b,c
R1
a
S
Br
N
OtBu
HO
OtBu
R1
R1
3
4
5
Scheme 1. Reagents and conditions: (a) HN(R2)CH2CH2OH, N-ethyldiisopropylamine (DIPEA), DMF, 50 ꢁC, 16 h; (b) 2-(R5-Ph–S)–Ph–OH,
DEAD, PS–PPh3, THF, rt; (c) HCl/AcOH, rt.
R5
O
O
H
N
N
O
OtBu
b,c
O
O
O
OH
R3
a
d,e
Si
N
S
Br
R3
HO
OtBu
OtBu
R3
6
7
8
9
Scheme 2. Reagents and conditions: (a) NH2CH(R3)CH2OTBDMS, DIPEA, DMF, 50 ꢁC, 16 h; (b) MeI, DIEA, DMF, 50 ꢁC, 16 h; (c) Et3NÆ3HF,
AcCN; (d) 2-(R5-Ph–S)–Ph–OH, DEAD, PS–PPh3, THF, rt; (e) HCl/AcOH.
O
OH
Cl
SH
O
O
OH
a
b
S
S
S
d
c
10a
R4
R4
R4
F
F
F
13
O
12
11
I
e
R4
10b
N
N
OH
OtBu
O
O
O
O
S
S
f, g
R4
14
R4
F
F
15
Scheme 3. Reagents and conditions: (a) 3-fluoroiodobenzene, Pd2dba3, DPEPhos, KtBuO, toluene, 100 ꢁC, 90 min, 55%; (b) 3-fluorobenzenethiol,
KtBuO, Pd2dba3, toluene; (c) BBr3, toluene, 0 ꢁC, 16 h; (d) ethylene carbonate, K2CO3, DMF; (e) Tf2O, Et3N, D- or L-proline-OtBu; (f) where,
R4 = Br, Ar-B(OH)2, Pd(PPh3)2Cl2, K2CO3, DME, 85 ꢁC, 2 h; (g) HCl/AcOH, 16 h, rt.

G. Smith et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3981–3984
3983
Table 2. Effect of R1/R2 substitution on inhibition of [3H]glycine
uptake at the GlyT-1b transporter
acid and cleavage of the tert-butyl ester yielded final
product 15.
Cl
The prepared compounds were tested for inhibition of
uptake of [3H]glycine into cells transfected with the
human GlyT-1b transporter. The cells were pre-washed
twice with HBS (10 mM Hepes–Tris (pH 7.4), 2.5 mM
KCl, 1 mM CaCl2 and 2.5 mM MgSO4) and preincubat-
ed with the test compound for 6 min. Afterwards, 10 nM
of [3H]glycine was added, and the cells were incubated
for 15 min. The cells were washed twice in HBS. Scintil-
lation fluid was added, and the plates were counted on a
Trilux (Wallac) scintillation counter.
R1
S
R2
N
O
O
HO
Compound
R1
R2
GlyT-1b IC50 (lM)
26
7.3
28
29
30
31
32
Me
H
Me
Et
Me
(R)-CH2CH2CH2–
(S)-CH2CH2CH2–
Et
50
22
Replacement of the 2R,5S-dimethyl piperazine moiety
in compound 1 with a oxyethylamino linker resulted
initially in compounds with micromolar potency as list-
ed in Table 1. A slight advantage was observed for fluo-
rine in the meta position, (23) compared to the para
position (18) with other substitutions weaker in the para
position, (17, 19 and 21) or no potency in the case of
acetamide (20). With R5 as 3-F no beneficial effects were
observed with alkyl substitution in the 2 position of the
oxyethyl chain (24, 25, 26 or 27).
0.38
Table 3. Effect of R4 substitution on inhibition of [3H]glycine uptake
at the GlyT-1b transporter
S
F
2
1
O
N
Investigation of the amino acid fragment as listed in
Table 2 revealed that the N-methyl and N-ethyl alanine
analogues (28 and 30, respectively) showed considerably
reduced potency compared to the sarcosine analogue 22.
Intolerance of N-ethyl was also seen with the glycine
analogue 29. Surprisingly the ring closed S-proline ana-
logue 32 showed significantly improved potency.
R4
O
OH
Compound
R4
GlyT-1b IC50 (lM)
33
34
35
36
37
38
39
40
41
42
5-Cl
6-Cl
3-Cl
4-Cl
0.6
19
1.2
0.4
Within the S-proline series the effect of R4 substitution
in the central benzene ring was examined as listed in
Table 3. Introduction of chlorine displayed intolerance
for the 6 position (34) and the 3 position (35) was also
disfavoured, whereas both the 4 and 5 positions were
5-Phenyl
4-Phenyl
0.95
0.059
0.52
0.16
0.3
4-(3-Thiophene)
4-(4-MeO–Ph)
4-(3-MeO–Ph)
4-(4-Cl–Ph)
1.4
Table 1. Effect of R5 and R3 substitution on inhibition of [3H]glycine
uptake at the GlyT-1b transporter
tolerated (36 and 33, respectively). Further space for
substitution was observed with a 4-phenyl group (38)
significantly improving potency though not with the
isosteric 4-(3-thiophene) analogue 39 or in the 5 position
(37). Substitution on the additional 4-phenyl was also
tolerated with 4 and 3-methoxy groups (40 and 41),
whilst 4-chlorine was less favourable in (42).
R5
O
3OH
N
S
O
R
Compound 38 was found to be inactive against the
GlyT-2 at a concentration of 10 lM. Key compounds
were profiled in vitro and in vivo for their ability to pen-
etrate the CNS as summarised in Table 4. Caco-2 cells,
exhibiting tight and confluent monolayers, serve as a
model of cell permeability.12 Thus, transport studies
using Caco-2 cells may be applied as an initial indication
of blood–brain barrier (BBB) transport characteristics.
Compound 1 showed good Caco-2 cell properties with
no indication of active efflux and this was corroborated
with good CNS penetration in the mouse with a brain/
plasma (B/P) ratio13 of 0.54. Compounds 38 and 40 both
Compound
R5
R3
GlyT-1b IC50 (lM)
1
16
17
18
19
20
21
22
23
24
25
26
27
0.16
7.8
4-Cl
H
H
H
H
H
H
H
H
Et
4-MeO
4-F
4-Me
30
15
47
4-NHCOCH3
>50
9.2
3.9
5.1
11
4-tert-Butyl
3-Cl
3-F
3-F
3-F
Isopropyl
R-Me
S-Me
18
12
3-F
3-F
displayed high apparent permeability coefficients (Papp
in the Caco-2 assay. However, their transport ratios
)
8.9

3984
G. Smith et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3981–3984
Table 4. In vitro and in vivo assessment of CNS penetration for key
compounds
2. Bergeron, R.; Meyer, T. M.; Coyle, J. T.; Greene, R. W.
Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 15730.
3. Tsai, G.; Coyle, J. T. Annu. Rev. Pharmacol. Toxicol.
2002, 42, 165.
Compound Caco-2 Papp Transport ratio (b-a/a-b) B/P Mice
(·10À6 cm/s)
´
´
4. Araragon, C.; Lopez-Corcuera, B. Trends Pharmacol. Sci.
1
38
40
33.4
25.1
26.3
0.8
2.2
2.9
0.54
0.17
0.09
2005, 26, 283.
5. Hashimoto, K. Recent Patents on CNS Drug Discovery
2006, 1, 43.
6. Lechner, S. M. Curr. Opin. Pharmacol. 2006, 6, 75.
7. Smith, G.; Ruhland, T.; Mikkelsen, G.; Andersen, K.;
Christoffersen, C. T.; Alifrangis, L. H.; Mørk, A.; Wren,
P.; Harris, N.; Wyman, B. M.; Brandt, G. Bioorg. Med.
Chem. Lett. 2004, 14, 4027.
(basal–apical(b-a)/apical–basal(a-b)) of >2 in Caco-2
cells pointed towards a possible involvement of active ef-
flux from the brain.
8. 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. Mol. Pharmacol.
2001, 60, 1414.
9. Brown, A.; Carlyle, I.; Clark, J.; Hamilton, W.; Gibson,
S.; McGarry, G.; McEachen, S.; Rae, D.; Thorn, S.;
Walker, G. Bioorg. Med. Chem. Lett. 2001, 11, 2007.
10. Samoshin, V. V.; Kudryavstev, K. V. Tetrahedron Lett.
1994, 34, 7413.
11. Preparation of noncommercially available starting mate-
rials 10b was as follows: 5-chloro-2-iodo-anisole/5-bro-
mo-2-iodo-anisole were prepared from the respective
anilines using the Sandmeyer reaction employing (1)
NaNO2/H2SO4/H2O, 5–10 ꢁC, 30 min; (2) KI, rt, 1 h.
3-Chloro-2-iodo-anisole was prepared by treatment of
3-chloro-anisole with s-BuLi (1.3 M in cyclohexane)/
THF, À95 ꢁC , 1 h followed by addition of I2, 16 h, rt.
4-Bromo-2-iodo-anisole was prepared by treatment of
4-bromo-anisole with PhI(OAc)2, I2, CH3CN, N2, 16 h.
12. Lennernas, H. J. Pharm. Pharmacol. 1997, 49, 627.
13. Compounds were dosed at 5 mg/kg sc and plasma
and brain samples were taken after 70 min. Brains
were collected by cervical dislocation, homogenised
1:4 (w/w) with 50% acetonitrile/water solution using
an Ultraturrax homogenizer. The supernatant was
separated by centrifugation (2000g, 10 min). Plasma
and extracted brain supernatant were analysed by
LC–MS/MS.
As such, compounds 38 and 40 could thus be substrates
for active efflux transporters such as P-glycoprotein.
These in vitro data appeared to be predictive of in vivo
brain exposure where modest exposure in the CNS was
observed. Thus, 70 min following sc administration of
5 mg/kg, 38 and 40 displayed average brain exposures of
only 60 and 63 ng/ml, compared to plasma levels of 381
and 723 ng/ml for the two compounds (n = 2). This result-
ed in B/P ratios of 0.17 and 0.09 for 38 and 40,
respectively.
In summary, further optimisation of the 2-phenylsulfa-
nylphenyl piperazine acetic acid series leads to the
identification of the S-proline derivatives 38 and 40
with improved in vitro potency in blocking the GlyT-
1. In vitro and in vivo assessment of these compounds
showed that CNS utility of these compounds might be
diminished due to active efflux transporter activity rel-
ative to 1.
References and notes
´ ´
1. Aragon, C.; Lopez-Corcuera, B. Eur.J. Pharmacol. 2003,
479, 249.