Discovery of novel and orally active NR2B-selective N-methyl-d-aspartate (NMDA) antagonists, pyridinol derivatives with reduced HERG binding affinity

Article abstract of DOI:10.1016/j.bmcl.2007.08.039

Novel NR2B antagonists with an amide tether were found by an approach to avoid pharmacophoric similarity to dofetilide. Structure-activity relationship investigation led to N-[cis-4-hydroxy-4-(5-hydroxypyridin-2-yl)cyclohexyl]-3-henylpropanamide 14e as an

Full text of DOI:10.1016/j.bmcl.2007.08.039

Bioorganic & Medicinal Chemistry Letters 17 (2007) 5533–5536
Discovery of novel and orally active NR2B-selective
N-methyl-D-aspartate (NMDA) antagonists, pyridinol
derivatives with reduced HERG binding affinity
Makoto Kawai,a,* Hiroshi Nakamura,a Isao Sakurada,a Hirohisa Shimokawa,a
Hirotaka Tanaka,a Miyako Matsumizu,a Kazuo Ando,a Kazunari Hattori,a
Atsuko Ohta,a Seiji Nukui,a Atsushi Omurab and Mitsuhiro Kawamuraa
aDiscovery Chemistry, Pfizer Global Research & Development, Nagoya Laboratories, 5-2, Taketoyo, Aichi 470-2393, Japan
bPharmaceutical R&D, Pfizer Global Research & Development, Nagoya Laboratories, 5-2, Taketoyo, Aichi 470-2393, Japan
Received 14 June 2007; revised 30 July 2007; accepted 2 August 2007
Available online 22 August 2007
Abstract—Novel NR2B antagonists with an amide tether were found by an approach to avoid pharmacophoric similarity to dof-
etilide. Structure–activity relationship investigation led to N-[cis-4-hydroxy-4-(5-hydroxypyridin-2-yl)cyclohexyl]-3-henylpropana-
mide 14e as an orally active NR2B-subtype selective N-methyl-D-aspartate (NMDA) receptor antagonist with very weak HERG
(human ether-a-go-go related gene) binding (IC50 > 30 lM). This compound exhibited potent in vivo anti-allodynic activity in the
mouse partial sciatic nerve ligation (PSL) model (minimum effective dose = 10 mg/kg, po).
Ó 2007 Elsevier Ltd. All rights reserved.
N-Methyl-D-aspartate (NMDA)-type glutamate recep-
tors are widely distributed in the brain and spinal cord
and play critical roles in learning, memory, motor con-
trol, and pain transmission.1 Non-selective NMDA
receptor antagonists have shown pain relief in various
clinical and animal studies.2–4 However, the clinical
development of non-selective antagonists has been ham-
pered by unfavorable side-effects due to non-specific ac-
tions at analgesic doses.5–7
than non-selective NMDA antagonists in animal mod-
els.8 Moreover, it was recently reported that com-
pound 1 is significantly effective to reduce pain
intensity in patients with spinal cord injury and mono-
radiculopathy without significant adverse events
(Fig. 1).
9
We had already reported identification of compound 2,
a novel NR2B selective NMDA antagonist, starting
from CP-101,606 as a lead compound.10 Compound 2
has an improved profile over CP-101,606 (1) in terms
of pharmacokinetic (PK) variability and QT prolonga-
tion. Nevertheless, QT prolongation, which is believed
to cause lethal arrhythmia, was still a concern for com-
pound 2 because of its moderate HERG (human ether-
a-go-go related gene) current inhibitory activity
(iHERG IC20 = 1.1 lM). Potential risk of structure-re-
lated toxicity was another concern due to the anilide11
and hydroxyphenylpiperidine12 moieties. Thus, further
research was conducted to find structurally distinct
back-up development candidates with minimal HERG
activity.
Functional NMDA receptors consist of heteromeric
combinations of the NR1 subunits and one or more
subunits designated as NR2A, 2B, 2C, and 2D. Vari-
ous heteromeric NMDA receptor channels formed by
combinations of NR1 and NR2 subunits are known
to differ in neuronal function. NR2B-containing
NMDA receptors are expressed predominantly in the
forebrain and spinal cord.1 Therefore NR2B selective
antagonists are believed to have improved safety pro-
file. In fact, CP-101,606 (1), an NR2B-selective
NMDA antagonist, demonstrated wider safety profile
Most of the NR2B antagonists including 1 and 2 were
derived from ifenprodil (3). The pharmacophore of
Keywords: HERG; NMDA; NR2B; Pain; Solubility.
*
Corresponding author. E-mail: makoto.kawai@pfizer.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.08.039

5534
M. Kawai et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5533–5536
Figure 1. Structures of NR2B selective antagonists, dofetilide (4) and a nonbasic ifenprodil derivative 5.
Scheme 1. Reagents and conditions: (a) BnBr, K2CO3, acetone, reflux 95%; (b) i—BnNH2, Ti(OiPr)4, THF, rt; ii—NaBH4, methanol, 0 °C–rt;
iii—column chromatography, 43%; (c) 20% Pd(OH)2–C, methanol, H2 (4 atm), rt, 85%; (d) RCOOH, EDCI, HOBt, CH2Cl2, rt.
these NR2B antagonists consists of a basic aliphatic
amine at the center of the molecule and two aromatic
rings.13 HERG channel blockers represented by dof-
etilide (4) also have a similar pharmacophore consist-
ing of a basic nitrogen and lipophilic sites located at
a certain distance from the basic nitrogen.14,15 Espe-
cially the basic amine enhances affinity for the
HERG potassium channel. For example, compound
5 without a basic amine exhibited significantly weaker
HERG activity (>30 lM) than the (S,S)-isomer of
ifenprodil (3) (<1 lM). Therefore, we revisited non-
basic compounds previously synthesized in our
NR2B programs. Among them, compound 6a exhib-
ited extremely low HERG activity (IC50 > 30 lM)
and then was selected as a lead. Due to insufficient
NR2B activity (IC50 = 30 nM) and poor solubility
(<5 lg/ml at pH of the small intestine) of
compound 6a, our efforts were focused on solving
two issues.
acids to build a compound library of cyclohexyla-
mides 6.
Data of representative examples are summarized in
Table 1. As expected, HERG activities of all the com-
pounds were very low (>30 lM). However, replacement
of the benzene ring with heterocycles led to moderate or
complete loss of the affinity to the NR2B receptor.
Introduction of an amine moiety also had a deleterious
effect on the NR2B binding (6f and 6g). Therefore, the
dihydrocinnamic acid moiety was fixed for further
modification.
Next we shifted our efforts onto the phenol moiety. A
divergent synthetic route was developed to expedite ana-
logue synthesis (Scheme 2). Amidation of dihydrocin-
namic acid with trans-aminocyclohexanol 11 afforded
alcohol 12. Swern oxidation of 12 yielded ketone 13,
which was subsequently treated with dianions generated
from 4-bromophenol derivatives to provide cyclohexa-
nol derivatives 14 as a desired isomer by chromato-
graphic purification. Compounds 15 were finally
obtained by dehydration followed by hydrogenation.
This procedure worked well not only for phenols but
also for heterocycles.
To improve the aqueous solubility, the first attempt
was made to replace the benzene ring with various
heteroaromatics and to introduce a basic moiety at
the position away from the center of the molecule
by parallel synthesis. (Scheme 1) Benzylation of 4-
(4-hydroxyphenyl)cyclohexanone (7) with benzyl bro-
mide and potassium carbonate yielded ketone 8.
Reductive amination of 8 followed by chromato-
graphic purification to remove a cis-isomer afforded
amine 9. Deprotection of 9 afforded a key intermedi-
ate 10 which was coupled with various carboxylic
Table 2 summarizes the SAR results around cyclohex-
ane and cyclohexanol derivatives. This modification
did not affect the HERG activity as well. It was revealed
that intermediates 14(a–d) generally showed better solu-
bility and NR2B activity than the corresponding cyclo-

M. Kawai et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5533–5536
Table 1. NR2B IC50, HERG IC50, and solubility of heterocycles Table 2. NR2B IC50, HERG IC50, and solubility of phenols
5535
Compound
R
X
NR2B HERG Solubility
IC50
Compound
R
NR2B HERG Solubility
IC50
(nM)
(at pH 6.5,
lg/ml)
at pH 6.5e
a
b
IC50
(nM)
IC50
(lM)
(lM)
(lg/ml)
14a
6a
OH
H
19
30
>30
>30
20
6a
6b
6c
6d
6e
30 >30
>1000 >30
856 >30
<5
<5
14b
15b
OH
H
189 >10a
>5000 NDb
<5
<5
NTd
92
14c
15c
OH
H
25 >10a
22
722
>30
<5
96 >10c
9
14d
15d
OH
H
13
>30
17
NTc
59 >10a
>1000 NTd
NTd
a Not measured due to low solubility at 30 lM in the assay.
b Not detected due to low solubility at 10 lM in the assay.
c NT, not tested.
6f
>1000 >30
>1000 NTd
18
weaker NR2B activity probably due to its steric hin-
drance. Compound 14e with 3-pyridinol moiety did em-
body the best compound with desired profiles of both
potency and solubility (Table 3).
6g
NTd
a Measured as the IC50 value for displacement of tritiated racemic CP-
101,606 from the rat forebrain P2 membrane.
b Measured as the IC50 value for displacement of tritiated dofetilide
from human HERG channel expressed in HEK(human embryonic
kidney)293 cells.
Compound 14e16 exhibited potent analgesic activity in
the mouse partial sciatic nerve legation (PSL) model
(minimum effective dose = 10 mg/kg, po) and exhibited
much lower HERG current inhibitory activity (9% at
30 lM). In addition, this compound possesses a good
PK profile, high metabolic stability (>60 min), and bio-
availability (27%) (Table 4).
c Not measured due to low solubility at 30 lM in the assay.
d NT, not tested.
e Using a KH2PO4–Na2HPO4 buffer adjusted by KOH.
In summary, we identified a new lead NR2B antago-
nist by taking a strategy to avoid pharmacophoric
similarity to dofetilide in order to minimize HERG
current inhibitory activity. Further efforts to improve
solubility and NR2B activity of the lead have yielded
14e with potent analgesic activity and very weak
HERG current inhibitory activity.
hexane derivatives 6a and 15(b–d). Thus, instead of
cyclohexane derivatives, SAR around cyclohexanol
derivatives 14 was pursued further with other heterocy-
cles to improve solubility. Replacement with heterocy-
cles resulted in loss of potency except for 3-pyridinol
14e. Introduction of methyl group to 14e also led to
Scheme 2. Reagents and conditions: (a) PhCH2CH2COOH, EDCI, HOBt, CH2Cl2, rt, 85%; (b) Swern oxidation, 93%, À60 °C ꢀ rt; (c) ArLi (3
equiv), THF, À78 °C; (d) i—CF3COOH, CH2Cl2, rt; ii—10% Pd/C, MeOH, H2, rt.

5536
M. Kawai et al. / Bioorg. Med. Chem. Lett. 17 (2007) 5533–5536
Table 3. NR2B IC50, HERG IC50, and solubility of heterocyclic
analogues
Acknowledgments
The authors thank Dr. Rodney W. Stevens for his valu-
able suggestion.
References and notes
Compound
R
NR2B
IC50 (nM)
Solubility
(at pH 6.5,
lg/ml)
1. Parsons, C. G.; Danysz, W.; Quack, G. Drug News
Perspect. 1998, 11, 523.
2. Hughes, A. M.; Rhodes, J.; Fisher, G.; Sellers, M.;
Growcott, J. W. Br. J. Clin. Pharmacol. 2002, 53, 604.
3. Knox, D. J.; McLeod, B. J.; Goucke, C. R. Anaesthesia
Intensive Care 1995, 23, 620.
4. Max, M. B.; Byas-Smith, M. G.; Gracely, R. H.; Bennett,
G. J. Clin. Neuropharmacol. 1995, 18, 360.
5. Ilkjaer, S.; Dirks, J.; Brennum, J.; Wernberg, M.; Dahl, J.
B. Br. J. Anaesthesia 1997, 79, 600.
6. McCartney, C. J. L.; Sinha, A.; Katz, J. Anesthesia
Analgesia (Hagerstown, MD, United States) 2004, 98,
1385.
7. Nelson, K. A.; Park, K. M.; Robinovitz, E.; Tsigos, C.;
Max, M. B. Neurology 1997, 48, 1212.
8. Chenard, B. L.; Bordner, J.; Butler, T. W.; Chambers, L.
K.; Collins, M. A.; De Costa, D. L.; Ducat, M. F.;
Dumont, M. L.; Fox, C. B., et al. J. Med. Chem. 1995, 38,
3138.
9. Sang, C. N.; Jinga, L.; Wouden, J.; Saltarelli, M. D. Soc.
Neurosci. Abstr. 2003, 814.
10. Kawai, M., Bioorg. Med. Chem. Lett., in press.
11. Eiskjaer, H.; Schmiegelow, M.; Jespersen, B.; Tietze, I. N.;
Jensen, J. D.; Soerensen, S. S.; Thomsen, K.; Pedersen, E.
B. Eur. J. Clin. Pharmacol. 1991, 41, 547.
14e
14f
14g
14h
14i
14j
14k
14l
10
85
130
69
116
NT
NT
NT
NT
NT
NT
>5000
>5000
>5000
2440
4960
12. Usuki, E.; Van Der Schyf, C. J.; Castagnoli, N., Jr. Drug
Metab. Rev. 1998, 30, 809.
13. Kornberg, B. E.; Nikam, S. S.; Wright, J. L.; Kesten, S.
R.; Meltzer, L. T.; Coughenour, L.; Barr, B.; Serpa, K. A.;
McCormick, J. Bioorg. Med. Chem. Lett. 2004, 14, 1213.
14. Cavalli, A.; Poluzzi, E.; De Ponti, F.; Recanatini, M. J.
Med. Chem. 2002, 45, 3844.
15. Du, L.-P.; Tsai, K.-C.; Li, M.-Y.; You, Q.-D.; Xia, L.
Bioorg. Med. Chem. Lett. 2004, 14, 4771.
Table 4. Pharmacological profile of compound 14e
NR2B binding (IC50
HERG binding (IC50
16. Chemical data of compound 14e; 1H NMR (300 MHz,
DMSO-d6) d. 9.68 (br s, 1H), 8.03 (d, J = 2.6 Hz, 1H), 7.73
(d, J = 7.7 Hz, 1H), 7.45 (d, J = 8.6 Hz, 1H), 7.31–7.09 (m,
6H), 4.88 (s, 1H), 3.64–3.48 (m, 1H), 2.81 (t, J = 7.6 Hz,
2H), 2.34 (t, J = 7.6 Hz, 2H), 1.98–1.80 (m, 2H), 1.70–1.47
(m, 6H) ; MS (ESI) m/z 339.21 (MÀH+); Anal. Calcd for
C20H24N2O3: C, 70.56; H, 7.11; N, 8.23. Found: C, 70.19;
H, 7.11; N; 7.99.
)
10 nM
>30 lM
9%
)
HERG current inhibition at 30 lM
Solubility at pH 6.5
130 lg/ml
>60 min
27%
t1/2 in human liver microsomes
Bioavailability in mice
Mouse in vivo model (PSL)a
Minimum effective dose (po)
10 mg/kg
17. The method of PSL in mice is described in
WO2005035523.
a The mouse partial sciatic nerve ligation (PSL) model.17