The discovery of a structurally novel class of inhibitors of the type 1 glycine transporter

Article abstract of DOI:10.1016/j.bmcl.2009.04.035

The type 1 glycine transporter plays an important in regulating homeostatic glycine levels in the brain that are relevant to the activation of the NMDA receptor by the excitatory neurotransmitter glutamate. We describe herein the structure-activity relationships (SAR) of a structurally novel class of GlyT1 inhibitors following on a lead derived from high throughput screening, which shows good selectivity for GlyT1 and potent activity in elevating CSF levels of glycine.

Full text of DOI:10.1016/j.bmcl.2009.04.035

Bioorganic & Medicinal Chemistry Letters 19 (2009) 2974–2976
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The discovery of a structurally novel class of inhibitors of the type 1
glycine transporter
*
John A. Lowe III , Xinjun Hou, Christopher Schmidt, F. David Tingley III, Stan McHardy , Monica Kalman,
Shari DeNinno, Mark Sanner, Karen Ward, Lorraine Lebel, Don Tunucci, James Valentine
Pfizer Global Research and Development, Eastern Point Road, Groton CT 06340, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
The type 1 glycine transporter plays an important in regulating homeostatic glycine levels in the brain
that are relevant to the activation of the NMDA receptor by the excitatory neurotransmitter glutamate.
We describe herein the structure–activity relationships (SAR) of a structurally novel class of GlyT1 inhib-
itors following on a lead derived from high throughput screening, which shows good selectivity for GlyT1
and potent activity in elevating CSF levels of glycine.
Received 1 April 2009
Revised 9 April 2009
Accepted 10 April 2009
Available online 18 April 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Glycine
Transporter
Inhibitor
[3.1.0]-Azabicyclic
Schizophrenia
NMDA
Glycine plays a key role in eukaryotic physiology as both an
endogenous amino acid for protein synthesis and a neurotransmit-
ter. In its role as a neurotransmitter, glycine binds to the NR1 sub-
unit of the NMDA receptor, acting as an obligatory co-transmitter
in modulating receptor sensitivity for the endogenous agonist glu-
tamate. It also acts at strychnine-sensitive postsynaptic receptors
in the spinal cord to control motor coordination. In both these
roles, levels of glycine are controlled by glycine-uptake transport-
ers, which consist of two subtypes, GlyT1 and GlyT2.1 GlyT1 is
found on glial cells in the brain and is widely distributed through-
out the body, where it controls basal levels of glycine. GlyT2 is
found primarily in the spinal cord and brainstem, where it controls
levels of glycine following release by glycinergic neurons.
GlyT1 plays an important role not only throughout the body,
but also in controlling glycine levels in the vicinity of NMDA syn-
apses in order to prevent glycine levels from constantly saturating
the glycine binding site on the NR1 subunit of the NMDA receptor.
As a key player in mediating learning and memory, the NMDA
receptor has long attracted attention as a possible target for novel
therapeutic agents.2 In particular, one theory of the etiology of
schizophrenia hypothesizes that a deficit in NMDA-mediated neu-
F
O
O
N
N
OH
OH
O
O
O
2
1
O
Cl
OH
N
N
N
N
CH3
O
N
N
O
Cl
N
.2HCl
NH2
3
21
Figure 1. GlyT1 inhibitors discussed above.
rotransmission underlies many of the disorder’s defining symp-
toms.3 For example, a deficiency in NMDA transmission may be
responsible for the loss of executive functions such as decision-
making ability, memory loss, and decreased affect (personality)
that characterize the disease. Restoring NMDA function, however,
is not a simple matter of providing more glutamate, which can
be neurotoxic at high levels. As an alternative approach, elevating
* Corresponding author. Address: 28 Coveside Lane, Stonington, CT 06378, United
States. Tel.: +1 860 535 4283.
E-mail address: jal3rd@gmail.com (J.A. Lowe).
Present address: Southwest Research Institute, San Antonio, TX 78228, United
States.
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.04.035

J. A. Lowe et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2974–2976
2975
Cl
levels of glycine would augment endogenous glutamate transmis-
O
NH2
NH
HN
N
sion through NMDA receptors.4 This elevation could be achieved
through inhibition of GlyT1-mediated glycine uptake, and this
mechanism forms the basis for our proposed approach to a novel
antipsychotic agent described herein.
Cl
N
ArCHO
CH3
Na(OAc)3BH, DCE
49%
O
(iPr)2NEt
O
NH
While the atypical antipsychotic agents, exemplified by risperi-
done, olanzapine, quetiapine, ziprasidone, and aripiprazole, have
provided an important advance in therapy for schizophrenia, many
opportunities remain for continued improvement. For a novel ther-
apy to provide value, it should avoid the side-effect liabilities of the
currently available classes of antipsychotics while addressing some
of the symptom domains, such as negative and cognitive symp-
toms, poorly treated by these agents. GlyT1 inhibitors would fill
this latter role by augmenting executive function and memory to
address cognitive symptoms, as well as improving affect to address
negative symptoms.
CH3CN, reflux
88%
O
O
19
18
O
O
Cl
Cl
N
N
N
N
N
N
CH3
HCl
CH3
EtOAc
100%
.2HCl
NH2
O
NH
O
20
21
Several structurally diverse GlyT1 inhibitors have been
described previously (Fig. 1).5 The prototype GlyT1 inhibitor,
(R)-N[3-(40fluorophenyl)-3-(40phenylphenoxy)propyl]-sarcosine, is
designated ALX 5407, 1.6 NFPS, the racemic form of ALX 5407,
has been shown to augment downstream effects of NMDA trans-
mission, supporting the proposed role of GlyT1 in regulating
NMDA transmission.7 We recently reported an analogue of ALX
5407, (R)-N[3-phenyl-3-(40-(4-toluoyl)phenoxy)-propyl]sarcosine
((R)-NPTS), 2, as a suitable radioligand for a GlyT1 binding affinity
assay.8 In addition, we described more recently a series of indanyl
piperazine and propylamine GlyT1 inhibitors based on a hit from
high-throughput screening, exemplified by compound 3.9 Neither
series, however, provided a compound suitable for advancement
to development. Continued high-throughput screening using the
radioligand version of 2 led to the discovery of compound 21.
We report herein our follow-up of this lead compound and the
structure–activity relationships (SAR) for improving its permeabil-
Scheme 2. Preparation of HTS lead compound 21.
ity, as well as further structural modification for improving
clearance.
The preparation of the compounds described in this work is
shown in Schemes 1 and 2. In Scheme 1, following t-BOC protec-
tion of starting material 14,10 a series of straightforward steps
leads to the final analogues for testing via oxidation, reductive ami-
nation, acylation, de-protection, and reductive amination. This
chemistry lent itself to SAR elaboration using library design and
production, and the analogues reported herein reflect that versatil-
ity. The synthesis of the HTS lead, compound 21, is shown in
Scheme 2 starting from the commercially available 18, and follows
the chemistry shown in Scheme 1.
The in vitro assays used to characterize the compounds de-
scribed have been reported previously.8 The CSF glycine assay,
which will be described in more detail in a subsequent report,
takes advantage of the overflow of glycine into the CSF following
GlyT1 inhibition in the brain.
OH
Many previously described GlyT1 inhibitor series, such as those
we have previously disclosed as exemplified by compounds 2 and
3, suffer from drawbacks in their drug-like properties and conse-
quent pharmacokinetics and safety profiles. In seeking a new GlyT1
inhibitor series, we prioritized drug-like properties as the starting
point. We were therefore gratified when a high-throughput screen
of our compound file revealed a new series of GlyT1 inhibitors
exemplified by compound 21 (Table 1). In addition to its potent
GlyT1 inhibitory potency, 21 exhibits potent in vivo activity in
the CSF Glycine model. The excellent drug-like physical properties
of this compound, in particular its low molecular weight and lipo-
philicity, were exactly what we had prioritized. In addition, its
structure is very synthetically versatile, enabling rapid exploration
of SAR at the amide, benzyl side chain, template, and amino por-
tions of the molecule. Among the various surrogates for the 1,4-
diaminocyclohexane template, we discovered that a [3.1.0] azabi-
cyclic template provided good pharmacokinetic properties and im-
proved in vivo activity in the CSF Glycine model (Table 1). This
series, then, formed the basis for the remaining SAR investigation.
The first issue we addressed was the limited permeability in
compound 21, which is indicated by the low value of MDCK_AB_C
in Table 1. This parameter measures the transit of a molecule
across the cell membrane of MDCK cells, which are representative
of many cell types, and thus gives a good idea of both gut and
blood-brain barrier permeability. Increasing lipophilicity, as mea-
sured by cLog P, delivered improved permeability as expected, as
shown in compounds 4 and 5. Interestingly, further increases in
lipophilicity, as illustrated in compound 7, did not improve perme-
ability. The improved permeability in compounds 4 and 5 comes at
OH
1. (COCl)2, DMSO
H
H
(BOC)2O, Et3N
95%
H
H
TEA, CH2Cl2
N
2. Ar-CH2-NH2,
NaBH4, MeOH
N
H
O
O
100%
14
15
N
H
N
N
N
O
N
Ar
O
H3C
O
H3C
N
O
Ar
N
Cl
N
Ar
H
H
N
H
H
HCl, EtOAc
98%
H
H
CH3
N
N
O
O
N
H
(iPr)2NEt, DCE
88%
O
16
11-13
17
4, R1 = Et, Ar = 3-OCF3Ph
N
N
O
H3C
R1-CHO
5, R1 = CH2-cyclopropyl, Ar = 3-OCF3Ph
6, R1 = Me, Ar = 3-OCF3Ph
7, R1 = CH2-cyclopentyl, Ar = 3-OCF3Ph
8, R1 = Et, Ar = 3-ClPh
N
Ar
H
H
Na(OAc)3BH,
HOAC, DCE
9, R1 = Me, Ar = 3-ClPh
N
R1
10, R1 = Me, Ar = 3-Cl,4-FPh
11, R1 = H, Ar = 3-OCF3Ph
12, R1 = H, Ar = 3-ClPh
70-95%
4-10
13, R1 = H, Ar = 3-Cl, 4-FPh
Scheme 1. Preparation of compounds 4–17.

2976
J. A. Lowe et al. / Bioorg. Med. Chem. Lett. 19 (2009) 2974–2976
Table 1
Data for GlyT1 inhibitors
Compd No.
GlyT1 Ki (nM)
GlyT2 IC50 (nM)
CSF Glycine ED200 (mg/kg)
hMic t1/2 (min)
MDCK AB_C
HERG IC50 (nM)
cLog P
Mol. wt.
2
3
4
5
6
7
8
9
10
11
12
13
21
2.32 (2.08–2.57 n = 15)
(26.8 + 7.05)*
>10,000
NT
56.3 (25.6–124)
NT
1.8 (0.866–3.6)
0.8
0.97
4.7
2.9
8.6
3.5 (1.79–6.88)
>30
NT
NT
58
9.1
1.69
1.7
NT
NT
3.8
417
509
422
448
408
476
373
359
377
394
345
363
347
5.42
2.43
2.7
2.24 (1.60–3.14 n = 3)
2.32 (1.84–2.92 n = 8)
2.32 (1.49–3.60 n = 3)
3.82 (0.851–17.2 n = 3)
4.58 (3.03–6.94 n = 4)
11.0 (0.511–236 n = 2)
11.6 (7.34–18.3 n = 7)
15.2 (0.280–822 n = 4)
95.3 (31.4–289 n = 3)
109
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
NT
21.6
12.1
27.3
11.2
63.1
>120
80
>85.7
>120
>120
>120
12.1
15.7
13.9
7.55
12.4
14.1
13.2
1.7
2200
530
3800
460
NT
NT
8500
NT
1.9
3.82
2.11
1.58
1.73
1.45
1.14
1.28
1.75
3.2
0.9
0.5
NT
14,600
65,600
>10,000
1150
1.79 (1.59–2.01 n = 68)
3.9 (2.17–7.41)
NT—not tested; GlyT1 Ki values are given in nM units, with standard error of the mean and number of determinations indicated in parentheses. GlyT2 IC50 values are given in
nM units, for inhibition of [3H]-glycine uptake. CSF Glycine ED200 values represent the dose that doubles endogenous (baseline) levels of glycine in rat csf at a 90 min time
point, following subcutaneous administration of a test compound, with 95% confidence limits in parentheses. hMic t1/2—compound half-life in human microsomes given in
minutes. MDCK_AB_C—apparent permeability through MDCK cell membranes in units of 10–6 cm/s, corrected for background, at an initial concentration of 2 lM. HERG IC50
values given in nM units for block of HERG current in patch clamp measurement in HEK-293 cells. cLog P and Mol. wt. (molecular weight).
*
For compound 3, GlyT1 IC50 value in nM units.
a cost, however, in that the projected human metabolic stability
also decreases substantially. Table 1 indicates this metabolic sta-
bility in terms of the half-life of the compound when incubated
in vitro with human microsomes under conditions for metabolic
degradation by P450 enzymes. For example, compound 21 shows
good microsomal stability with a hMic t1/2 value of >120 min.
Increasing lipophilicity to improve permeability resulted in greatly
reduced hMic t1/2 values for, for example, compounds 4 (21.6 min)
and 5 (12.1 min). In general, we sought hMic t1/2 values of greater
than 60 min in order to provide good oral bioavailability. ADME in
silico modeling suggested that maintaining the tertiary amine
group in compounds 4 and 5 while reducing the molecular weight
and cLog P would be favorable for reducing microsomal clearance.
Although this balance of functionality and lipophilicity limits the
range of potential compounds, adjusting the aromatic substituent
on the benzyl side chain and using a small R group appended to
the endocyclic nitrogen was found to afford the desired micro-
somal stability, as shown in compounds 8, 9, and 10. Compound
10 in particular represents a favorable balance of potent GlyT1
inhibitory activity with good microsomal stability and permeabil-
ity, resulting in potent activity in the CSF Glycine in vivo assay. It
also shows that reduced lipophilicity decreases HERG activity
(e.g., relative to the more lipophilic compound 5), thus suggesting
another reason for maintaining the proper balance of lipophilicity.
The GlyT1 inhibitor approach to schizophrenia tests the hypoth-
esis that elevated synaptic levels of glycine will enhance NMDA
receptor-mediated signaling and thereby ameliorate a major con-
tributor to the symptomatic characteristics of the disease. The
compounds described in the present work, in particular compound
10, may prove applicable to testing this hypothesis clinically. Com-
pound 10 provides a balance of in vitro and in vivo potency with
favorable pharmacokinetic properties suggesting the potential for
good oral bioavailability in humans.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2009.04.035.
References and notes
1. Betz, H.; Gomeza, J.; Armsen, W.; Scholze, P.; Eulenburg, V. Biochem. Soc. Trans.
2006, 34, 55.
2. Coyle, J. T. Cell Mol. Neurobiol. 2006, 26, 365.
3. Millan, M. J. Psychopharmacology 2005, 179, 30.
4. Kinney, G. G.; Sur, C. Curr. Neuropharmacol. 2005, 3, 35.
5. Slassi, A.; Egle, I. Expert Opin. Ther. Patents 2004, 14, 201.
6. (a) 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.; Klitencik, M. A. Mol.
Pharmacol. 2001, 60, 1414; (b) Aubrey, K. R.; Vandenberg, R. J. Br. J. Pharmacol.
2001, 134, 1429.
7. Bergeron, R.; Meyer, T. M.; Coyle, J. T.; Greene, R. W. Proc. Natl. Acad. Sci. U.S.A.
1998, 95, 15730.
8. Lowe, J. A., III; Drozda, S. E.; Fisher, K.; Strick, C.; Lebel, L.; Schmidt, C.; Hiller, D.;
Zandi, K. S. Bioorg. Med. Chem. Lett. 2003, 13, 1291.
9. Lowe, J.; Drozda, S.; Qian, W.; Peakman, M. C.; Liu, J.; Gibbs, J.; Harms, J.;
Schmidt, C.; Fisher, K.; Strick, C.; Schmidt, A.; Vanase, M.; Lebel, L. Bioorg. Med.
Chem. Lett. 2007, 17, 1675.
10. Brighty, K. E.; Castaldi, M. J. Synth. Lett. 1996, 11, 1097.