Discovery of NR2B-selective antagonists via scaffold hopping and pharmacokinetic profile optimization

Article abstract of DOI:10.1016/j.bmcl.2019.02.017

Selective N-methyl-D-aspartate receptor subunit 2B (NR2B) antagonists show potential as analgesic drugs, and do not cause side effects associated with non-selective N-methyl-D-aspartate (NMDA) antagonists. Using a scaffold-hopping approach, we previously identified isoxazole derivative 4 as a potent selective NR2B antagonist. In this study, further scaffold hopping of isoxazole derivative 4 and optimization of its pharmacokinetic profile led to the discovery of the orally bioavailable compound 6v. In a rat study of analgesia, 6v demonstrated analgesic effects against neuropathic pain.

Full text of DOI:10.1016/j.bmcl.2019.02.017

Accepted Manuscript
Discovery of NR2B-selective antagonists via scaffold hopping and pharmaco-
kinetic profile optimization
Kosuke Anan, Moriyasu Masui, Aya Tazawa, Minoru Tomida, Yoshihiro Haga,
Masaharu Kume, Shoichi Yamamoto, Shunji Shinohara, Hiroki Tsuji, Shinji
Shimada, Shigenori Yagi, Nobuyoshi Hasebe, Hiroyuki Kai
PII:
S0960-894X(19)30091-5
https://doi.org/10.1016/j.bmcl.2019.02.017
BMCL 26297
DOI:
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
31 October 2018
25 January 2019
14 February 2019
Please cite this article as: Anan, K., Masui, M., Tazawa, A., Tomida, M., Haga, Y., Kume, M., Yamamoto, S.,
Shinohara, S., Tsuji, H., Shimada, S., Yagi, S., Hasebe, N., Kai, H., Discovery of NR2B-selective antagonists via
scaffold hopping and pharmacokinetic profile optimization, Bioorganic & Medicinal Chemistry Letters (2019), doi:
https://doi.org/10.1016/j.bmcl.2019.02.017
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Discovery of NR2B-selective antagonists via scaffold hopping and pharmacokinetic profile
optimization
Kosuke Anan,a* Moriyasu Masui,a Aya Tazawa,a Minoru Tomida,a Yoshihiro Haga,a Masaharu
Kume,a Shoichi Yamamoto,b Shunji Shinohara,d Hiroki Tsuji,d Shinji Shimada,d Shigenori Yagi,c
Nobuyoshi Hasebe,c and Hiroyuki Kaia
a Medicinal Chemistry Research Laboratory, Shionogi & Co., Ltd, 1-1 Futabacho, 3-chome, Toyonaka
561-0825, Japan
b
Drug Discovery & Disease Research Laboratory, Shionogi & Co., Ltd, 1-1 Futabacho, 3-chome,
Toyonaka 561-0825, Japan
c
Research Laboratory for Development, Shionogi & Co., Ltd, 1-1 Futabacho, 3-chome, Toyonaka
561-0825, Japan
d Shionogi Techno Advance Research Co., Ltd., 1-1 Futabacho, 3-chome, Toyonaka 561-0825, Japan
*Corresponding author. Tel.: +81 (6) 6331 5833; fax: +81 (6) 6332 6385. E-mail address:
kousuke.anan@shionogi.co.jp (K. Anan)
Keywords: NMDA receptor antagonist; scaffold hopping; analgesic activity
Abstract
Selective N-methyl-D-aspartate receptor subunit 2B (NR2B) antagonists show potential as
analgesic drugs, and do not cause side effects associated with non-selective N-methyl-D-aspartate
(NMDA) antagonists. Using a scaffold-hopping approach, we previously identified isoxazole
derivative 4 as a potent selective NR2B antagonist. In this study, further scaffold hopping of
isoxazole derivative 4 and optimization of its pharmacokinetic profile led to the discovery of the
orally bioavailable compound 6v. In a rat study of analgesia, 6v demonstrated analgesic effects
against neuropathic pain.
Text body
The N-methyl-D-aspartate (NMDA) receptor (NR), a glutamic acid receptor, is highly expressed
throughout the central nervous system.1 The NR is a heterotetrameric complex composed of the
three subunits NR1 (a-h), NR2 (A-D), and NR3 (A-B), and is activated by the amino acids glutamate
and glycine.2,3 The NMDA receptor subunit 2A (NR2A) is involved in memory formation and learning
acquisition, whereas excessive activation of the NMDA receptor subunit 2B (NR2B) results in
neurodegenerative cell death and transmission of pain related to brain ischemia.4 Studies
investigating mice overexpressing NR2B have revealed its important role in pain.5 Because of the
restricted distribution of NR2B to the forebrain and spinal cord dorsal horn, NR2B-selective
antagonists may reduce side effects associated with the use of non-selective antagonists.6 This
evidence points to the use of selective NR2B antagonists as a promising strategy in the development

of novel analgesics.
In a previous report, we introduced the use of a scaffold-hopping approach to develop a new
selective NR2B antagonist, by utilizing the structure of a known compound.7 Considering the
improved inhibitory activity against human ether-à-go-go-related gene (hERG) channels conferred
by the structure of the selective NR2B antagonist radiprodil (1)8 and compound 2 (Pfizer),9,10 we
designed the acetamide derivative 3 by hybridizing the side-chain in compound 1 and the
cyclohexanol ring in compound 2. Compound 3 showed high in vitro potency, but unfortunately
displayed low brain penetration. To improve brain penetration, this series was further optimized via
replacement of the amide in compound 3 with an isoxazole, and efforts to optimize the
pharmacokinetic profiles led to the discovery of the orally available brain penetrant compound 4,
which displayed analgesic activity in mice.
In this study, we designed the novel isoxazole derivative 5, in which the cyclohexanol ring in
compound 4 was converted via scaffold hopping to create a chemotype with improved
pharmacokinetics. Although isoxazole derivative 5 showed potent activity as expected, its
pharmacokinetic profile was unsatisfactory. Next, we developed the novel oxamide derivative 6, in
which the isoxazole ring was converted to an amide. Certain derivative compounds maintained high
activity, showed improved pharmacokinetic profiles, and exerted strong analgesic effects in a rat
model of neuropathic pain. Here, we describe our efforts to identify the novel isoxazole derivative 5
and oxamide derivative 6.
Figure 1. Summary of compounds described in previous reports (red) and in this report (blue)
The syntheses of pyrazoles 5a-c (Scheme 1) started with 3-ethynyl-1H-pyrazole (7), which was
treated with ethyl 2-chloro-2-hydroxyiminoacetate to produce ester 8. Hydrolysis of ester 8 following
amide coupling with corresponding amines 10 yielded pyrazoles 5a-c. Amine 10 was commercially
available or prepared from commercially available starting materials following standard
procedures.11

Scheme 1. Synthesis of pyrazoles 5a-c
Reagents and conditions: (a) Ethyl 2-chloro-2-hydroxyiminoacetate, NaHCO3, i-PrOH, room
temperature (rt) 3 days, yield: 76%; (b) 2M-NaOH aq., MeOH, rt 2 h, 75%; (c) 10·HCl, HOBt, DMAP,
WSCD·HCl, Et3N, DMF, rt 2 h, 69-90%.
The pyridones 5d-h (Scheme 2) were prepared from 5-ethynyl-2-methoxypyridine (11) via synthetic
methods similar to those depicted in Scheme 1. Removal of a methyl group from 14 yielded the
pyridones 5d-h.
Scheme 2. Synthesis of pyridones 5d-h
Reagents and conditions: (a) Ethyl 2-chloro-2-hydroxyiminoacetate, NaHCO3, i-PrOH, 40 °C 24 h,
80%; (b) 2M-NaOH aq., MeOH, rt 2 h, quant.,; (c) 10·HCl, HOBt, DMAP, WSCD·HCl, Et3N, DMF, rt
2 h, 91-97%; (d) (COCl)2, dimethylformamide (DMF, cat.), tetrahydrofuran (THF), reflux 2 h, then 10
·HCl, pyridine, THF, rt 3 h, 69-94%; (e) Trimethylsilyl chloride (TMSCl), NaI, MeCN, reflux 2 h,
20-91%.
The oxamide derivatives 6a-v were prepared according to the method depicted in Scheme 3. Amide
coupling of amine 15 with ethyl 2-chloro-2-oxoacetate followed by hydrolysis yielded the carboxylic
acid 16. Amine 15 was commercially available or synthesized according to reported procedures.12
Finally, amide coupling of 21 with anilines produced oxoamide derivatives 6a-v.

Scheme 3. Synthesis of oxamide 6a-v
Reagents and conditions: (a) ClCOCO2Et, Et3N, CH2Cl2, 0 °C 0.5-1 h, 80-99%; then 2M-NaOH aq.,
EtOH, rt 1.5 h, 77-99%; (c) R1-NH2, HOBt, DMAP, WSCD·HCl, Et3N, DMF, rt overnight, 46-91%.
Compounds were assessed via an NR2B-selective binding assay13 using the radiolabeled ligand
[3H]-ifenprodil, an NR1/NR2B-specific antagonist. Compounds active in this test were then
evaluated in a functional assay14 to assess the level of inhibition of Ca ion influx in
NR-overexpressing human embryonic kidney (HEK) 293 cells. A positive correlation between binding
and functional inhibitory potency was observed. Compounds 5a-f, in which the cyclohexanol ring in
compound 4 was substituted with a piperidine ring, were investigated (Table 1). All compounds
displayed potent binding and antagonistic activities, especially compound 5f, which was more potent
than compound 4 and radiprodil (1). Conversion of the cyclohexanol ring in compound 4 to a
piperidine ring had no significant influence on the activity of the compounds. The amide moiety
present in radiprodil (1) is converted to an isoxazole ring in compounds 5a-f, and considering both the
present results and those of previous reports,7,15,16 both the amide moiety and isoxazole ring are
assumed to be bioisosteres.
The rat pharmacokinetic studies of the compounds described in Table 1 (with the exception of 5b)
were conducted in male Sprague-Dawley (SD) rats with intravenous (iv) and oral administration of
0.5 and 1 mg/kg doses, respectively. Although the pyrazole-type compounds 5a and 5c showed high
brain partition coefficients (Kp), their total clearance and bioavailability were low. In contrast,
although the pyridine-type compounds 5d-f showed lower brain Kp than the pyrazoles, their total
clearance and bioavailability were acceptable. However, none of the compounds (5a and 5c-f) showed
comparable pharmacokinetic profiles to that of compound 4. Because radiprodil (1) also displayed a
high total clearance, these results suggest that the piperidine ring may be responsible for the poor
pharmacokinetic profiles displayed by the compound series 5.

Table 1. Structure-activity
derivatives 5a-f
relationships (SARs) of isoxazole
Compound
NR2B (bind)a
NR2B(Ca2+)b
CLtc
(mL/min/kg)
71.6
Fc
(%)
0.2
Brain
Kpd
3.89
NT
R1
R2
No
pIC50
7.1
pIC50
7.3
H
5a
5b
4-F
7.1
7.2
NT
NT
4-Cl
H
7.2
7.4
29.2
0.6
3.27
0.37
0.50
0.38
1.00
2.66
5c
7.2
7.2
19.3
18.4
18.4
9.3
5d
4-F
4-Cl
7.3
7.4
19.7
5e
7.6
8.0
5.4
5f
7.5
8.5
37.1
30.0
67.5
radiprodil (1)
4
7.0
7.9
4.0
a Inhibition of binding of [3H]-ifenprodil to rat brain membranes
b
Inhibition of Ca ion influx in mouse N-methyl-D-aspartate (NMDA) receptor-overexpressing
HEK293 cells
c
CLt = total clearance; F = bioavailability. Compounds were administered at doses of 0.5 mg/kg
(intravenous injection, iv) and 1.0 mg/kg (oral) using a cassette dosing method.
d
Brain Kp = total brain-to-plasma ratio. Plasma and brain samples were collected 0.5 h after iv
administration of a 0.5-mg/kg dose using a cassette dosing method.
NR2B, N-methyl-D-aspartate receptor subunit 2B; pIC50, negative log of the 50% inhibitory
concentration, -log(IC50); NT = not tested
To improve the pharmacokinetic profiles of compounds 5, the piperidine ring in compound 5f was
converted, resulting in compounds 5g and 5h (Table 2). Unfortunately, piperazine derivatives 5g and
5h showed reduced binding activity. However, a similar conversion was conducted on radiprodil (1),
yielding the benzyl piperazine 6a which showed reduced activity, and the phenylpiperazine 6b which
showed increased activity. In rat pharmacokinetic studies, 6a and 6b showed lower total clearance
than radiprodil (1). Furthermore, the benzoxazolone ring in 6b was converted to an indolone ring or a
dihydroquinolone ring, which showed high activity in our previous study.7 Compounds 6c and 6d
showed greatly decreased binding activity. From the piperazine derivatives described above, only 6b
showed high activity.

Table 2. SARs of piperidine linker moieties 5f-h and 6a-d
Compound
NR2B (bind)a
NR2B(Ca2+)b
CLtc
Fc
(%)
9.3
Brain
Kpd
0.38
NT
R3
X
Y
No
5f
pIC50
7.6
pIC50
8.0
(mL/min/kg)
CH2
N
CH2
CH2
None
CH2
CH2
5.4
NT
NT
37.1
6.8
< 6.0
< 6.0
7.5
NT
NT
NT
30.0
7.4
5g
5h
1
N
NT
NT
CH2
N
8.5
1.00
0.41
6.3
7.2
6a
N
None
7.8
8.1
NT
7.2
17.8
0.53
6b
6c
N
None
< 6.0
NT
NT
NT
N
None
< 6.0
NT
NT
NT
NT
6d
a Inhibition of binding of [3H]-ifenprodil to rat brain membranes
b
Inhibition of Ca ion influx in mouse N-methyl-D-aspartate (NMDA) receptor-overexpressing
HEK293 cells
c
CLt = total clearance; F = bioavailability. Compounds were administered at doses of 0.5 mg/kg
(intravenous injection, iv) and 1.0 mg/kg (oral) using a cassette dosing method.
d
Brain Kp = total brain-to-plasma ratio. Plasma and brain samples were collected 0.5 h after iv
administration of a 0.5-mg/kg dose using a cassette dosing method.
NR2B, N-methyl-D-aspartate receptor subunit 2B; pIC50, negative log of the 50% inhibitory
concentration, -log(IC50); NT = not tested
Phenyl piperazine 6b, which showed potent binding activity and an adequate pharmacokinetic
profile, was identified as the lead compound via a scaffold-hopping approach. We next investigated
R2 substituents in 6b (Table 3). Binding activity comparable to that of 6b was observed in the
compound with 4-CF3 (6g), whereas reduced binding activity was observed with 4-F (6f) and 4-MeO

(6h). Removal of the R2 substituent (6e) caused diminished binding activity. In addition, the
activities of 3,4-disubstituted compounds (6i, 6k-n) were comparable to those of 4-substituted
compounds, whereas the activity of the 3,5-disubstituted compound (6j) was reduced. Among
3,4-disubstituted compounds, those with 3-F-4-Cl (6l) and 3-F-4-CF3 (6m) displayed the most potent
antagonistic activities. In rat pharmacokinetic studies of these compounds, 6l displayed a suitable
pharmacokinetic profile. Total clearance of the compound with 3-F-4-MeO (6n) was extremely high,
and we therefore assume that the MeO group in 6n was the main metabolic point.
To improve pharmacokinetics profiles, the introduction of a methyl group into the piperazine ring
was investigated.17 Introduction of a methyl group with R configuration into the substituent R4 (6o)
resulted in binding activity comparable to that of 6b. Introduction of a methyl group with S
configuration into the substituent R4 (6p) caused diminished binding activity. Furthermore, when a
methyl group was introduced into the substituent R5, 6q (R configuration) showed high activity
whereas 6r (S configuration) displayed low binding activity. Compound 6o showed improved oral
bioavailability compared to 6b, and 6q showed slightly higher total clearance than 6b, and much
improved oral bioavailability and brain penetration. Therefore, a methyl group was introduced into
the piperazine ring with R2 (3,4-F2 and 3-F-4-Cl). As a result, compounds 6s-v displayed improved
pharmacokinetic profiles while maintaining high binding activity.

Table 3. SARs of phenylpiperazine derivatives 6b, e-v
Comp
NR2B (bind)a
NR2B(Ca2+)b
pIC50
CLtc
Fc
Brain
Kpd
ound
No
6b
6e
R2
R4
R5
pIC50
(mL/min/kg)
(%)
4-Cl
H
H
H
7.8
< 6.0
6.9
7.6
6.6
7.7
6.8
7.5
7.9
7.8
7.8
7.5
< 6.0
7.2
6.2
7.3
7.2
7.9
7.3
8.1
NT
6.9
8.2
6.7
8.0
7.0
8.0
9.0
9.0
7.8
8.2
NT
8.2
7.0
8.2
8.0
8.7
8.5
7.2
NT
17.8
NT
0.53
NT
H
H
4-F
H
H
11.1
15.1
NT
39.3
24.8
NT
NT
6f
4-CF3
H
H
2.99
NT
6g
6h
6i
4-MeO
3,4-F2
3,5-F2
3-Cl-4-F
3-F-4-Cl
3-F-4-CF3
3-F-4-MeO
4-Cl
H
H
H
H
4.0
13.0
NT
0.82
NT
H
H
NT
6j
H
H
18.3
5.6
24.6
28.2
8.7
1.24
0.79
0.65
3.17
0.95
NT
6k
6l
H
H
H
H
5.6
6m
6n
6o
6p
6q
6r
H
H
100.8
5.4
0.6
(R)-Me
(S)-Me
H
H
129.5
NT
4-Cl
H
NT
4-Cl
(R)-Me
(S)-Me
H
15.6
NT
52.6
NT
1.85
NT
4-Cl
H
3,4-F2
3,4-F2
3-F-4-Cl
3-F-4-Cl
(R)-Me
H
5.6
117.9
25.3
30.4
32.0
0.67
1.89
0.99
1.19
6s
(R)-Me
H
15.9
1.37
6.0
6t
(R)-Me
H
6u
6v
(R)-Me
a Inhibition of binding of [3H]-ifenprodil to rat brain membranes
b
Inhibition of Ca ion influx in mouse N-methyl-D-aspartate (NMDA) receptor-overexpressing
HEK293 cells
c
CLt = total clearance; F = bioavailability. Compounds were administered at doses of 0.5 mg/kg
(intravenous injection, iv) and 1.0 mg/kg (oral) using a cassette dosing method.
d
Brain Kp = total brain-to-plasma ratio. Plasma and brain samples were collected 0.5 h after iv
administration of a 0.5-mg/kg dose using a cassette dosing method.
NR2B, N-methyl-D-aspartate receptor subunit 2B; pIC50, negative log of the 50% inhibitory
concentration, -log(IC50); NT = not tested

Based on their favorable pharmacokinetic profiles, 6o, 6u, 6q, and 6v were subjected to an in vivo
analgesic study. Their analgesic effects were evaluated after oral administration based on the change
in pain threshold relative to allodynia (von Frey test) in a rat partial sciatic nerve ligation model.18
Mechanical allodynia was quantified by measuring the hind paw withdrawal response to von Frey
hair stimulation. The response before drug administration was measured (vehicle), and the
evaluation was carried out 3 h after oral administration of each compound (2.5, 5, 10, and 20 mg/kg).
The dose associated with an effect significantly different to that of the vehicle was considered the
minimum effective dose (MED). As shown in Table 4, all compounds showed a significant increase in
pain threshold at doses below 10 mg/kg. Among these compounds, 6v demonstrated a significant
analgesic effect at 2.5 mg/kg. Fortunately, the hERG inhibitory activity19 of 6v was lower than that of
the other compounds (6o, 6u, and 6q).
Table 4. Biological activity and human ether-à-go-go-related gene (hERG) inhibitory activity of
selected compounds
Compound
Analgesic effectsa
hERGb
No
6o
6q
6u
6v
MED (mg/kg)
Inhibition at 1uM (%)
10.0
10.0
5.0
36
25
29
10
2.5
a Data were obtained from 10 rats per group and expressed as the minimum effective dose (MED)
with antiallodynic effect; p < 0.01 vs the vehicle.
b % inhibition at 1 μM in human ether-à-go-go-related gene (hERG)-expressing HEK293 cells
In summary, the novel isoxazole derivative 5 was designed via a scaffold-hopping approach,
starting with the previously identified NR2B-selective antagonist 4. Although the compound series 5
exhibited high binding activity, pharmacokinetic profiles required improvement. Further
optimization of 5 through bioisosteric replacement and efforts to optimize the pharmacokinetic
profiles led to the discovery of 6v, which demonstrated analgesic effects in the neuropathic pain
model.

References and notes
1. Banke, T. G., and Traynelis, S. F. Nat. Neurosci. 2003, 6, 144–152.
2. McBain, C. J., and Mayer, M. L. Physiol. Rev. 1994, 74, 723–760.
3. Bendel, O., Meijer, B., Hurd, Y., and Von Euler, G. Neurosci. Lett. 2005, 377, 31–36.
4. Parsons, C. G., Danysz, W., Quack, G. Drug News Perspect. 1998, 11, 523-69.
5. Wei, F., Wang, G.-D., Kerchner, G. A., Kim, S. J., Xu, H.-M., Chen, Z.-F., Zhuo, M. Nat. Neurosci.
2001, 4, 164-9.
6. Gogas, K. R. Curr. Opin. Pharmacol. 2006, 6, 68–74.
7. Anan, K., Masui, M., Hara, S., Ohara, M., Kume, M., Yamamoto, S., Shinohara., Tsuji, H.,
Shimada, S., Yagi, S., Hasebe, N., Kai, H. Bio. Med. Chem. Lett. 2017, 27, 4194-4198.
8. Mony, L., Kew, J. N., Gunthorpe, M. J., Paoletti, P. Br. J. Pharmacol. 2009, 157, 1301-1317.
9. McCauley, J. Expert Opin. Ther. Patents. 2006, 16, 863-870.
10. Kawai, M., Kawamura, M., Sakurada, I., Morita, A., (Pfizer, Inc.), WO2005080317.
11. Zhou, Z-L, and Keana, J. F. W, J. Org. Chem. 1999, 64, 3763-3766.
12. Smith, B., Tsai, J., Chen, R., (Arena Pharmaceuticals, Inc.), WO2005016902.
13. Masui, M., Adachi, M., Mikamiyama, H., Matsumura, A., Tsuno, N., (Shionogi & CO., LTD.),
EP1988077A1, p 119.
14. Masui, M., Adachi, M., Mikamiyama, H., Matsumura, A., Tsuno, N., (Shionogi & CO., LTD.),
EP1988077A1, p 121.
15. (a) Kai, H., Nakayama, K., Tsuji, H., Masuko, M., Takase, A. J. Pesticide Sci. 1998, 23, 44-48.
(b) Kai, H., Ichiba, T., Miki, M., Takase, A., Masuko, M. J. Pesticide Sci. 1999, 24, 130-137.
(c) Kai, H., Ichiba, T., Tomida, M., Masuko, M. J. Pesticide Sci. 1999, 24, 149-155.
16. (a) Meanwell, N. A. J. Med. Chem. 2011, 54, 2529-2591.
(b) Zhiqing, L., Pingyuan, W., Haiying, C., Eric, A. W., Bing, T., Allan, R. B., Jia, Z. J. Med. Chem.
2017, 60, 4533-4558.
17. Eliezer, J. B., Arthur, E. K., Carlos, A. M. F. Chem. Rev. 2011, 111, 5215–5246.
18. Seltzer, Z., Dubner, R., Shir, Y. Pain. 1990, 43, 205–218.
19. Kai, H., Kameyama, T., Hasegawa, T., Oohara, M., Endoh, T., (Shionogi & CO., LTD.),
EP2399910A1, p 219.