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 metabolic stability

Article abstract of DOI:10.1016/j.bmcl.2006.05.058

A novel class of 4-aryl-8-(2-hydroxy-2-phenyl-cyclohexyl)-2,8-diaza-spiro[4.5]decan-1- ones have been discovered and developed as potent and selective GlyT1 inhibitors. The molecules are devoid of activity at the GlyT2 isoform and display excellent select

Full text of DOI:10.1016/j.bmcl.2006.05.058

Bioorganic & Medicinal Chemistry Letters 16 (2006) 4311–4315
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 metabolic stability
a
b
c
c
c
`
Daniela Alberati, Dominik Hainzl, Synese Jolidon, Eva A. Krafft, Anke Kurt,
Axel Maier,c Emmanuel Pinard,c Andrew W. Thomasc,* and Daniel Zimmerlic
aDiscovery Biology, F. Hoffmann-La Roche Ltd, Pharmaceuticals Division, CH-4070 Basel, Switzerland
bSafety Technical Services, F. Hoffmann-La Roche Ltd, Pharmaceuticals Division, CH-4070 Basel, Switzerland
cDiscovery Chemistry, F. Hoffmann-La Roche Ltd, Pharmaceuticals Division, CH-4070 Basel, Switzerland
Received 17 May 2006; accepted 17 May 2006
Available online 6 June 2006
Abstract—A novel class of 4-aryl-8-(2-hydroxy-2-phenyl-cyclohexyl)-2,8-diaza-spiro[4.5]decan-1- ones have been discovered and
developed as potent and selective GlyT1 inhibitors. The molecules are devoid of activity at the GlyT2 isoform and display excellent
selectivities against the l-opioid receptor as well as the Nociceptin/Orphanin FQ peptide (NOP) receptor. In particular these novel
compounds 4 as well as the 4-substituted-8-(2-phenyl-cyclohexyl)-2,8-diaza-spiro[4.5]decan-1-one 3 show improved metabolic sta-
bility and pharmacokinetic profiles in rodents compared to previous triazaspiropiperidine series 1 and 2. We have also identified
within these diazaspiropiperidine series a key relationship between reducing basicity of the piperidine nitrogen and reducing hERG
affinity.
Ó 2006 Elsevier Ltd. All rights reserved.
GlyT1 is a selective transporter of the neurotransmitter
glycine that is localized in the brain in close vicinity with
the NMDA receptor for which glycine is an obligatory
co-agonist.1,2 In vitro experiments have recently shown
that GlyT1 inhibition results in an elevation of glycine lev-
els with a consequent enhancement of NMDA receptor
activity.3 This observation strongly suggests a role for
GlyT1 inhibitors in the treatment of CNS disorders like
schizophrenia where NMDA-R hypofunction is believed
to be involved.4 Additional support for this approach in
the treatment of schizophrenia comes from clinical studies
where sarcosine,5 a prototypical weak GlyT1 inhibitor,
improved positive, negative and cognitive symptoms in
schizophrenic patients, when administered together with
risperidone. As a result, considerable effort has been fo-
cused on the development of selective GlyT1 inhibitors.6
as a novel class of GlyT1 inhibitors which all display
excellent selectivity against the GlyT2 isoform.
H
N
H
N
O
O
N
N
HO
N
N
(1'R 2'R or 1'S 2'S)
(1'R 2'R or 1'S 2'S)
2
1
H
H
O
N
O
N
4
2'
HO
N
1'
N
4
3 (or diast.)
(1'R 2'R or 1'S 2'S)
(1'R 2'R or 1'S 2'S)
We have recently described the discovery of N-(2-aryl-
cyclohexyl) substituted spiropiperidines 17, 28 and 39
The main liability of 1 was identified early whereupon
only very low selectivity against the l-opioid receptor
and the Nociceptin/Orphanin FQ peptide (NOP) recep-
tor was achievable. However, we successfully addressed
these issues in a focused programme culminating in the
Keywords: GlyT1; GlyT2; NMDA; Schizophrenia; Transporter;
Glycine; Spiropiperidine.
*
Corresponding author. Tel.: +41 61 688 50 48; fax: +41 61 688 87
14; e-mail: andrew.thomas@roche.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.05.058

4312
D. Alberati et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4311–4315
discovery of 2 and 3. Since we had already demonstrated
significant improvement towards our desired pharmaco-
logical properties within the triazaspiro series by
performing the structural changes 1 to 2 (addition of a
20-OH group) and also within the diazaspiro series 1–3
(replacement of the 4-N with CH group), herein we wish
to report on our successful efforts to identify another
new class of GlyT1 inhibitors 4 by the incorporation
of both of the fore-mentioned chemical mutations into
one series.10
The synthetic strategy to access the target molecules 4
closely follows our previous work and was designed to
prepare a (1:1:1:1) mixture of four diastereoisomers with
a 1,2-cis relationship (Scheme 1). In analogy, the synthe-
sis commenced with an efficient opening of the commer-
cially available cyclohexene epoxide, however this time,
with the 4-substituted diazaspiropiperidine 511 to afford
the a-aminoalcohols 6 in excellent yield. Installation of
the appropriately substituted aryl residue (R1) with
excellent stereoselectivity was efficiently achieved
through a two-step process involving oxidation of 6 to
the a-aminoketone 7 by SO3–pyridine complex, fol-
lowed by reaction with the appropriate aryl-lithium
(R1-Li) reagent formed in situ.
H
O
N
H
O
O
N
a)
OH
+
N
HN
R2
6
(rac 5)
R2
H
H
O
O
N
N
In order to aid interpretation of the biological results,
where possible we endeavoured to separate the diastere-
oisomers by the implementation of chiral-phase HPLC
(Chiralpak ADÒ) and did not routinely test the
(1:1:1:1) mixture of four stereoisomers formed in the
reaction sequence. This was not always possible and
Table 1 shows the activity for the pure or the mixture
of stereoisomers with the fastest eluting component des-
ignated a and the slowest eluting designated d.
R1
b)
c)
R1 Br
O
HO
N
N
(see Table 1)
8-25
(1'R 2'R or 1'S 2'S)
R2
R2
7
Scheme 1. Synthesis of compounds 8–25. Reagents and conditions: (a)
EtOH, reflux, 95–99%; (b) SO3–pyridine complex, DMSO, TEA,
DCM, rt, 58–76%; (c) i—R1Br, BuLi, THF, ꢀ78 °C to rt, 45–90%.
Table 1. In vitro inhibitory activity at the GlyT1 transporter for compounds 8–25
H
O
N
R1
HO
N
R2
8-25
(1'R 2'R or 1'S 2'S)
R1
R2
GlyT1 EC50 (lM)
a
Compound
a
b
c
d
8
9
4-MeO–Ph
3-MeO–Ph
4-MeO–Ph
3-MeO–Ph
4-F–Ph
H
H
F
F
H
H
F
F
F
F
F
F
F
F
F
F
F
F
4.48b
>30b
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
3.300
0.158
1.181
2.599
1.128
0.127
0.07
0.158
5.301
0.825
5.301
0.646
0.069
0.073
0.129d
3-F–Ph
3-F–Ph
0.158
0.095
0.596
6.251
1.953
2.599
0.077
0.110
2.34
0.127
0.474
1.769
0.118
2-F–Ph
3-Cl–Ph
4-CN–Ph
4-CF3–Ph
4-MeSO2–Ph
4-Me–Ph
3-Me–Ph
2-Me–Ph
4-t-Bu–Ph
2-CF3O–Ph
4-Imidazole–Ph
11.33
>30c
>30c
1.313b
18.646b
0.42
>30
0.437
0.132
0.069
2.638
0.596
>30
10.71
0.110
>30
17.690
>30c
0.646c
1.394d
7.345
>30
27.544
0.175
a Radiometric assay using [3H]-glycine8.
b Mix of a, b, c and d.
c Mix of c and d.
d Mix of a and b.

D. Alberati et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4311–4315
4313
Table 3. Microsome stability data for compounds 1, 12c, 14b, 15d, 26,
and 27
For exploration of our preliminary SAR (Table 1), we
initially focused on derivatives where R2 = H or F, since
these were found to be optimal in the previous series 1, 2
and 3. Overall the SAR of this class of molecules 4 did
not follow the same SAR pattern previously established
with the series 1–3. For example, although only tested as
a mixture, compounds 8 and 9 bearing p-OMe and m-
OMe groups, respectively, were surprisingly found to
be inactive in sharp contrast to their potent congeners
in the triazaspiropiperidine series 1 and 2 which had
EC50 in the nanomolar range. Much improved activity
was obtained for R2 = F, with 10d and 11a,c also show-
ing affinity in the nM range. We then explored the effects
of the favoured halogen substituents at R1 with com-
pounds 12–16 each demonstrating a dramatic increase
to low nanomolar affinity at GlyT1. Electron-withdraw-
ing groups such as those evident in 17–19 provided only
inactive compounds. Methyl substituents were also well
tolerated at the p- and o-positions where 22 even provid-
ed three active diastereoisomers. Intriguingly, m-Me
substitution as in 21 was not tolerated in this series.
a
Compound
GlyT1
EC50 (lM)
CLint
Mouse
microsomes
Human
microsomes
1
26
0.026
0.024
0.061
0.069
0.077
0.070
240b
499b
36
68
78
23
28
11
44
27
15d
14b
12c
39
16
13
a CLint, intrinsic, lL/min/mg protein.
b Rat data.
However, we endeavoured to retain the aryl moiety in
the diazaspiropiperidine series and we were delighted
to observe a significant improvement in metabolic stabil-
ity of nearly all related derivatives in the diazaspiro ser-
ies 3 and 4 compared to the triazaspiro series 1 and 2
(Table 3). For example, the predicted maximum achiev-
able bioavailability (MAB), in human microsomes, for 1
was only ꢁ18%, whereas effecting the N to CH replace-
ment resulted in a significant improvement of the MAB
to ꢁ45% for 27.
Before moving further ahead with this novel diazaspiro-
piperidine class 4 it was essential to establish superiority
over the previous series 1–3 and we therefore selected
12a–d as a focus for further evaluation. Initially we set
out to demonstrate that the introduction of an addition-
al 20-OH group together with an N to CH mutation
would completely remove any off-target liabilities ob-
served in 1. Indeed, pleasingly these efforts were reward-
ed when we could demonstrate excellent selectivity over
GlyT2, NOP and most importantly we had completely
removed activity at the l-opioid receptor as shown in
Table 2.
This trend was also observed in a single dose in vivo
pharmacokinetic study in mice and rats for this series
where a much lower in vivo clearance was recorded
(Table 4). This resulted in an increase of oral bioavail-
ability (F) from 8% to 27%. In addition, we had also
achieved greater brain penetration for 27 compared to 1.
In the 20-OH series a significant improvement of MAB
was also observed by examining the pair 26 and 12c
where an improvement in predicted MAB, in human
microsomes, from 14% to ꢁ55% was observed. In gener-
al, it was not possible to increase further the metabolic
stability within these subseries of the diazaspiropiperi-
dines 3 and 4 with all derivatives in Table 1 generally
displaying comparable microsomal stability.
As a result, of these encouraging results, we proceeded
to explore the key liabilities identified in the previous
series in much more detail. A key impediment in the gen-
eral class of triazaspiropiperidines 1 and 2 has been the
non-optimal metabolic stability. This was vastly im-
proved upon in the triazaspiropiperidine series by
replacement of the R2 aryl residue with an alkyl
substituent.8
After achieving our desired in vitro pharmacological
profile with suitable pharmacokinetic properties in vitro
and in vivo, we then proceeded to further develop these
Table 2. In vitro inhibitory activity at the GlyT1 and the GlyT2
transporters and potency in inhibiting the NOP and l receptors for
compounds 1–3, 12a–d
F
H
N
H
N
a
b
c
O
O
Compound
EC50 (lM)
NOP IC50 (lM) l IC50 (lM)
GlyT1 GlyT2
N
N
HO
N
N
1
0.026
0.044
0.232
0.481
0.129
0.07
12
72
6
15
0.15
2.7
2
3d
12d
12ab
12c
12d
25
>10
>10
>10
>10
>10
3.92
(1'R 2'R or 1'S 2'S)
(1'R 2'R or 1'S 2'S)
>30
—
24
26
1
>10
>10
>10
F
H
>30
—
O
N
H
O
0.825
N
a Radiometric assay using [3H]-glycine.8
HO
N
b Displacement of [3H]NOP in membranes prepared from permanently
transfected HEK293 cells expressing hNOP receptors.12
c Displacement of [3H]naloxone in membranes prepared from BHK
cells transiently expressing hl receptors.12
N
(1'R 2'R or 1'S 2'S)
(1'R 2'R or 1'S 2'S)
12c (or diast.)
d Mixture of four diastereoisomers (see Table 1).
27 (or enant.)

4314
D. Alberati et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4311–4315
Table 4. Selected pharmacokinetic parameters for compounds 1 and
27
The synthesis of 28 was efficiently achieved as outlined
in Scheme 2. The 3,4-epoxytetrahydropyran 30 was pre-
pared according to literature precedent,13 and then
reacted with 29 to give a mixture of regioisomeric
a-aminoalcohols 31 and 32. The major product in the
reaction was the desired diastereoisomer 32 which was
then separated by chromatography on silica gel, oxi-
dized to the a-aminoketone 33 in good yield by SO3–
pyridine complex. Subsequent reaction with phenyl
lithium afforded the final product 28 in 59% yield. Pleas-
ingly, compound 28 apart from showing no activity at
the hERG K channel also demonstrated good affinity
(91 nM) at the GlyT1 transporter and was inactive at
the GlyT2 transporter.
Compound
F (%)
CLa
Vss (L/kg)
B/P
1b
27b
27c
8
120
63
7.3
4.3
8.9
1.6
5.2
4.1
27
99
a CL, total clearance, mL/min/kg protein.
b Measured in rat.
c Measured in mouse, B/P = brain to plasma ratio; Vss = volume of
distribution.
Table 5. In vitro inhibitory activity at the hERG K channel for
compounds 12c, 27 and 28
a
b
Compound
hERG IC50 (lM)
pKa
In conclusion, replacement of the sp2 nitrogen atom at the
position 4 of the imidazolidinone ring in our original ser-
ies 1 with a sp3 carbon followed by introduction of a 20-OH
groupresultedin a novel and potentclass of 4-aryl-8-(2-hy-
droxy-2-phenyl-cyclohexyl)-2,8-diaza-spiro[4.5]- decan-
1-one GlyT1 inhibitors. This diazaspiropiperidine series
display, as we had anticipated, high selectivities against
the l- and NOP-opioid receptors. In general, the diazaspi-
ropiperidine series are superior to the triazaspiropiperi-
dine series where a significant increase in microsomal
stability was achievable and related into an improved
PK profile in vivo in mouse and rat. We have also identi-
fied a key relationship between reducing basicity of the
piperidine nitrogen and reducing hERG affinity which al-
lowed us to design derivatives such as 28 free of activity at
the hERG K channel. The subsequent paper will describe
additional pharmacological improvements within the
diazaspiropiperidine family of GlyT1 inhibitors in the
development of another novel class of 4-substituted-8-
(1-phenyl-cyclohexyl)-2,8-diaza-spiro[4.5]decan-1-one
GlyT1 inhibitors.
27
1.400
1.800
>24.000
8.85
8.54
6.89
12c
28
a Inhibition of hERG K channel determined by whole-cell patch-clamp
experiments on a transfected CHO cell line.
b Determined by potentiometric titration in a MeOH/water mixture at
rt with the GLpKa instrument from Sirius Analytical Instruments.
molecules. However, we were concerned with the poten-
tial liability for blocking the hERG K channel where
overlapping pharmacophoric elements within our spiro-
piperidine class of molecules were known to exist. When
tested, as expected, low micromolar activity was ob-
served for 27 (Table 5) at the hERG K channel and
we believed that we could reduce these effects by modu-
lation of basicity of these compounds. Indeed, the intro-
duction of an additional b-oxygen relative to a nitrogen
atom is known to reduce its basicity and as expected 12c
and congeners show a slightly reduced effect at the
hERG K channel in line with the reduction of basicity
of the piperidine nitrogen. In order to reduce even fur-
ther the effects at the hERG K channel, we introduced
a second b-oxygen atom, this time within the cyclohexyl
ring, as shown in the tetrahydropyran compound 28,
which gratifyingly had the desired outcome of completely
removing activity at the hERG K channel.
Acknowledgments
We thank Serge Burner and David Vanguard for
additional technical assistance. Dr. Holger Fischer and
H
H
H
O
N
O
N
O
N
b)
O
OH
OH
O
+
N
N
N
O
O
F
33 75%
F
(1'R 2'R or 1'S 2'S)
(1'R 2'R or 1'S 2'S)
F
32 54%
31 19%
c)
34 59%
a)
H
H
N
O
N
O
O
+
O
HO
HN
N
30
O
(R) 29
F
F
(1'R 2'R or 1'S 2'S)
28
Scheme 2. Synthesis of compound 28. Reagents and conditions: (a) EtOH, reflux, 73%; (b) SO3–pyridine complex, DMSO, TEA, DCM, rt, 75%; (c)
PhLi, THF, ꢀ78 °C to rt, 59%.

D. Alberati et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4311–4315
4315
Expert Opin. Ther. Patents 2004, 14, 201; (c) Hashimoto,
K. Recent Pat. CNS Drug Disc. 2006, 1, 43.
7. (a) Ceccarelli, S. M.; Pinard, E.; Stalder, H. WO Patent
2005040166, 2005; Chem. Abstr. 2005, 142, 447121; (b)
Pinard, E.; Ceccarelli, S. M.; Stalder, H.; Alberati, D.
Bioorg. Med. Chem. Lett. 2006, 16, 349.
Bjoern Wagner are thanked for the measurement of pKa
values. Dr. Simona Ceccarelli is also thanked for helpful
discussions during the preparation of the manuscript
and Dr. Geo Adam is thanked for helpful discussions
at the onset of this work.
8. Ceccarelli, S. M.; Pinard, E.; Stalder, H.; Alberati, D.
Bioorg. Med. Chem. Lett. 2006, 16, 354.
9. 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. Bioorg. Med. Chem. Lett.
2006, 16, preceding paper.
10. Alberati, D.; Ceccarelli, S. M.; Jolidon, S.; Pinard, E.;
Thomas, A. W. US Patent 2005154001, 2005; Chem.
Abstr. 2005, 143, 115462.
11. Krafft, E. A.; Kurt, A.; Maier, A.; Thomas, A. W.;
Zimmerli, D. Synthesis 2005, 3245.
References and notes
1. (a) Eulenburg, V.; Armsen, W.; Betz, H.; Gomeza, J.
Trends Biochem. Sci. 2005, 30, 325; (b) Gomeza, J.; Ohno,
K.; Betz, H. Curr. Opin. Drug Discov. Devel. 2003, 6, 675.
2. Danysz, W.; Parsons, C. G. Pharmacol. Rev. 1998, 50, 597.
3. Bergeron, R.; Meyer, T. M.; Coyle, J. T.; Greene, R. W.
Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 15730.
4. For a recent review, see: Millan, M. J. Psychopharmacol-
ogy 2005, 179, 30.
5. Tsai, G.; Lane, H.-Y.; Yang, P.; Chong, M.-Y.; Lange, N.
Biol. Psychiatry 2004, 55, 452.
12. Wichmann, J.; Adam, G.; Roever, S.; Hennig, M.;
Scalone, M.; Cesura, A. M.; Dautzenberg, F. M. Eur. J.
Med. Chem. 2000, 35, 839.
6. For recent reviews, see: (a) Sur, C.; Kinney, G. G. Expert
Opin. Investig. Drugs 2004, 13, 515; (b) Slassi, A.; Egle, I.
13. Berti, G.; Catelani, G.; Ferretti, M.; Monti, L. Tetrahe-
dron 1974, 30, 4013.