Structure-activity relationships of N-substituted 4-(trifluoromethoxy) benzamidines with affinity for GluN2B-containing NMDA receptors

Article abstract of DOI:10.1016/j.bmcl.2013.12.087

GluN2B subtype-selective NMDA antagonists represent promising therapeutic targets for the symptomatic treatment of multiple CNS pathologies. A series of N-benzyl substituted benzamidines were synthesised and the benzyl ring was further replaced with various polycyclic moieties. Compounds were evaluated for activity at GluN2B containing NMDA receptors where analogues 9, 12, 16 and 18 were the most potent of the series, replacement of the benzyl ring with polycycles resulted in a complete loss of activity.

Full text of DOI:10.1016/j.bmcl.2013.12.087

Bioorganic & Medicinal Chemistry Letters 24 (2014) 828–830
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure–activity relationships of N-substituted
4-(trifluoromethoxy)benzamidines with affinity
for GluN2B-containing NMDA receptors
Corinne Beinat a, Samuel D. Banister a,b, Jane Hoban a, John Tsanaktsidis c, Athanasios Metaxas d,
Albert D. Windhorst d, Michael Kassiou a,b,e,
⇑
a School of Chemistry, The University of Sydney, NSW 2006, Australia
b Brain and Mind Research Institute, Sydney, NSW 2050, Australia
c CSIRO Materials Science & Engineering, Ian Wark Laboratory, Bayview Avenue, Clayton, Victoria 3168, Australia
d Department of Radiology & Nuclear Medicine, VU University Medical Centre, Amsterdam 1081HV, The Netherlands
e Discipline of Medical Radiation Sciences, University of Sydney, NSW 2006, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
GluN2B subtype-selective NMDA antagonists represent promising therapeutic targets for the symptom-
atic treatment of multiple CNS pathologies. A series of N-benzyl substituted benzamidines were synthes-
ised and the benzyl ring was further replaced with various polycyclic moieties. Compounds were
evaluated for activity at GluN2B containing NMDA receptors where analogues 9, 12, 16 and 18 were
the most potent of the series, replacement of the benzyl ring with polycycles resulted in a complete loss
of activity.
Received 26 November 2013
Revised 18 December 2013
Accepted 19 December 2013
Available online 28 December 2013
Keywords:
GluN2B
Ó 2013 Elsevier Ltd. All rights reserved.
NMDA
Amidine
CNS
Structure–activity relationships
N-Methyl-
D
-aspartate receptors (NMDARs) are hetero-
the hippocampus, olfactory bulb, and the cerebral cortex, with
minimal expression in the cerebellum.3 This differential anatomi-
cal distribution of GluN2A and GluN2B-comprised NMDARs allows
GluN2B-selective antagonists to exhibit the full therapeutic effects
of traditional non-selective NMDAR antagonists but with reduced
side-effect profiles, particularly undesirable locomotor effects,
due to reduced expression in the cerebellum.4
Early, non-selective GluN2B antagonists such as ifenprodil (1),
eliprodil (2), and haloperidol (3) (Fig. 1), were used to study
GluN2B-comprised NMDAR activity and were defined as first gen-
eration GluN2B antagonists. However, the off-target activity and
pharmacokinetic properties of these compounds limited their
therapeutic application. Researchers at Merck have since identified
(E)-N1-(benzyl)cinnamamidine (4, Ki = 9 nM vs [3H]1) to have
considerable affinity for GluN2B-containing NMDARs.5 As styryl
amidine 4 can potentially act as a Michael acceptor, a series of
non-styryl compounds were explored as a novel class of GluN2B
oligomeric ligand-gated cation channels widely distributed in the
CNS and activated by the major excitatory neurotransmitter,
L
-glutamate, in the presence of co-agonist glycine.1 NMDARs exist
as tetrameric or pentameric complexes of a glycine-binding GluN1
subunit with one or more glutamate-binding GluN2 (A, B, C, D)
subunits. NMDARs are essential to learning, memory, and informa-
tion processing within the human brain, and defects of NMDAR
function are implicated in multiple central nervous system (CNS)
disorders, including Parkinson’s disease, neuropathic pain, stroke
and major depressive disorder (MDD).1 Many historical NMDAR
antagonists were developed before the degree of receptor hetero-
geneity was known and have demonstrated therapeutic efficacy
against the aforementioned pathologies. However, the clinical
application of these drugs is restricted by side-effect profiles attrib-
utable to general NMDAR blockade, including hallucinations,
hypertension, dysphoria, and loss of co-ordination.2 Messenger
RNA (mRNA) mapping studies have since revealed a ubiquitous
distribution of the GluN2A subunit throughout the brain, whereas
the GluN2B subunit is confined to regions of the forebrain, such as
selective antagonists. Although benzamidine
5
displayed
weak affinity for the GluN2B subtype (Ki = 10,000 nM) against
[3H]ifenprodil, exploration of ring-substitution patterns identified
potent and selective GluN2B antagonists. The need for novel
therapeutic agents targeting GluN2B-containing NMDARs with
desirable pharmacokinetic and physiochemical properties
⇑
Corresponding author.
E-mail address: michael.kassiou@sydney.edu.au (M. Kassiou).
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.12.087

C. Beinat et al. / Bioorg. Med. Chem. Lett. 24 (2014) 828–830
829
HO
OH
N
OH
Cl
N
N
F
F
O
OH
Cl
1 (ifenprodil)
2 (eliprodil)
3
(haloperidol)
NH
NH
CF3
NH
N
H
N
H
N
H
F3CO
6
5
(BZAD01)
NH
4
Figure 1. NR2B-selective NMDA receptor antagonists.
NH·HCl
N
H2N
HCl (g), EtOH
0°C-RT, 18 h
R
OEt
N
R
H
F3CO
F3CO
F3CO
NEt3, EtOH
RT, 16 h
74%
9 19
-
7
8
·HCl
Scheme 1. Synthesis of amidines 9–19.
Table 1
Synthesis and binding affinities of amidines 9–19
Compound
R
Yield (%)
Ki (nM)
Compound
R
Yield (%)
Ki (nM)
6
9
10
11
12
13
2-(CF3)
2-(OMe)
2,3-(OMe)2
2,4-(OMe)2
2,5-(OMe)2
3,5-(OMe)2
72
65
89
90
68
70
8.09 1.01
2.09 0.37
303 64.5
16.2 9.19
4.29 1.57
22.9 9.43
14
15
16
17
18
3,4,5-(OMe)3
3,4-(OCH2O)
3-F
3,4-(F)2
3,5-(F)2
62
83
78
52
86
1140 530
26.3 4.25
2.86 1.04
8.25 3.51
3.58 0.78
The affinity of compounds for the ifenprodil binding site of GluN2B-containing NMDA receptors was determined through measurements of various concentrations of
unlabelled ligand to inhibit the binding of 5 nM [3H]ifenprodil in rat forebrain membranes.
F3CO
prompted us to explore derivatives of BZAD01 (6), which displayed
antagonist activity in an GluN2B FLIPR assay (IC50 = 47 nM), and
selectivity over GluN2A-, GluN2C-, and GluN2D-containing recep-
tors.6 In animal models of PD, previous work in our laboratory
has shown BZAD01 (6) to produce anti-parkinson effects and
improvements in Levodopa-induced dyskinesias when co-adminis-
tered with Levodopa therapy, validating the use of this scaffold for
further development as a potential therapeutic agents.7 Our inves-
tigations have also revealed BZAD01 to improve postural symme-
try and apomorphine induced rotation in 6-OHDA+Pf-lesioned
rats.8
The aims of the present work were to develop novel GluN2B
antagonists based on BZAD01 (6) for the potential treatment of
PD. A library of N-benzyl benzamidines was prepared from 4-
(trifluoromethoxy)benzonitrile and variously substituted benzyl
amines (Scheme 1, Table 1). The synthesis begain with conver-
sion of 4-(trifluoromethoxy)benzonitrile to its corresponding imi-
date hydrochloride salt 8 via a Pinner synthesis in excellent
yield. Imidate hydrochloride 8 was then reacted with a series
of substituted benzylamines to furnish amidines 6 and 9–18 in
good yields (52–90%). The prepared amidines 6 and 9–18 were
then evaluated for affinity at the ifenprodil binding site of
GluN2B-containing NMDA receptors by measuring the ability of
various concentrations of the unlabeled ligands to inhibit the
binding of 5 nM [3H]ifenprodil. The results are summarised in
Table 1.
NH
H
N
N
H
NH
F3CO
23
24
Figure 2. Polycyclic ‘cage’ amidines 23 and 24.
Replacement of the electron-withdrawing trifluoromethyl
group of BZAD01 with an electron-donating methoxy unit, as in
9, led to an improvement in GluN2B binding affinity. Disubstitu-
tion with the methoxy group revealed some noteworthy SAR, with
the greatest activity observed for the 2,4- and 2,5-dimethoxy-
substituted compounds (11 and 12, respectively). Introduction of
the methoxy substituent to the 3-position resulted in diminished
activity, as seen in compounds 10, 13 and 14, and methylenedi-
oxy-bridged 15. In contrast to trifluoromethyl or methoxy substi-
tution at the 3-position, introduction of one or more fluorine
atoms was generally tolerated, as demonstrated by compounds
16–18.
To provide greater detail into the steric and electronic require-
ments of GluN2B binding for the benzylic moiety of BZAD01, the
benzyl ring was replaced with the polycyclic ‘cage’ hydrocarbons
cubane and adamantane. Analogous to the preparation of benzami-
dines 6 and 9–18, cubylamidine 23 and adamantylamidine 24

830
C. Beinat et al. / Bioorg. Med. Chem. Lett. 24 (2014) 828–830
O
O
CO2Me
COOH
a-d
e-g
8 steps
NH2
19
MeO2C
21
22
20
Scheme 2. Reagents and conditions: (a) NaOH, MeOH, THF, rt, 16 h, 85%; (b) PhI(OAc)2, I2, PhMe, reflux, 16 h; (c) aq. NaOH, MeOH, THF, rt, 16 h, 80% over 2 steps; (d) n-BuLi,
THF, À78 °C, 1 h, then MeOH, rt, 1 h, 87%; (e) (COCl)2, À78 °C to rt, 1 h; (f) liq. NH2, À78 °C to rt, CH2Cl2, 2 h, 85%; (g) LiAlH4, THF, 0 °C to reflux, 15 h, 59%.
Acknowledgments
Table 2
Binding affinities and yields of amidines 23 and 24
C.B. acknowledges financial support provided by CSIRO through
the Australian Biotech Growth Partnerships Theme. This work was
supported by CTMM LeARN, work package 02N-101-01.
Compound
Yield (%)
Ki (lM)
23
24
65
58
0.15 0.01
4.81 2.16
Supplementary data
(Fig. 2) were synthesised via reaction of imidate salt 8 and the cor-
responding polycyclic amine.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2013.12.
087.
The synthesis of requisite aminomethylcubane (22) is shown in
Scheme 2. Cyclopentanone (19) was converted to dimethyl 1,4-
cubanedicarboxylate (20) in 19% over 6 steps using the procedure
reported by Bliese and Tsanaktsidis.9,10 Further manipulation of 20
yielded monofunctionalised cubanecarboxylic acid (21), which was
converted to the desired amine 22. 1-adamantanemethylamine
was synthesised from commercially available 1-adamantanecarb-
oxylic according to methods previously described in our
laboratory.11
These analogues were similarly tested for affinity at GluN2B-
containing NMDA receptors (Table 2). Replacement of the benzyl
ring with the cubane moiety resulted in affinity comparable to
the least potent benzylic derivatives, whereas incorporation of
the larger adamantane group reduced activity by several orders
of magnitude, suggesting that steric bulk is not tolerated in this
portion of the molecule. These findings also suggest a subtle elec-
tronic requirement for affinity at GluN2B containing NMDA
receptors.
References and notes
1. Beinat, C.; Banister, S.; Moussa, I.; Reynolds, A. J.; McErlean, C. S.; Kassiou, M.
Curr. Med. Chem. 2010, 17, 4166.
2. Kemp, J. A.; McKernan, R. M. Nat. Neurosci. 2002, 5, 1039.
3. Monyer, H.; Burnashev, N.; Laurie, D. J.; Sakmann, B.; Seeburg, P. H. Neuron
1994, 12, 529.
4. Boyce, S.; Wyatt, A.; Webb, J. K.; O’Donnell, R.; Mason, G.; Rigby, M.;
Sirinathsinghji, D.; Hill, R. G.; Rupniak, N. M. J. Neuropharmacology 1999, 38,
611.
5. Curtis, N. R.; Diggle, H. J.; Kulagowski, J. J.; London, C.; Grimwood, S.; Hutson, P.
H.; Murray, F.; Richards, P.; Macaulay, A.; Wafford, K. A. Bioorg. Med. Chem. Lett.
2003, 13, 693.
6. Claiborne, C. F.; McCauley, J. A.; Libby, B. E.; Curtis, N. R.; Diggle, H. J.;
Kulagowski, J. J.; Michelson, S. R.; Anderson, K. D.; Claremon, D. A.; Freidinger,
R. M.; Bednar, R. A.; Mosser, S. D.; Gaul, S. L.; Connolly, T. M.; Condra, C. L.;
Bednar, B.; Stump, G. L.; Lynch, J. J.; Macaulay, A.; Wafford, K. A.; Koblan, K. S.;
Liverton, N. J. Bioorg. Med. Chem. Lett. 2003, 13, 697.
7. Warraich, S. T.; Allbutt, H. N.; Billing, R.; Radford, J.; Coster, M. J.; Kassiou, M.;
Henderson, J. M. Brain Res. Bull. 2009, 78, 85.
8. Truong, L.; Allbutt, H. N.; Coster, M. J.; Kassiou, M.; Henderson, J. M. Brain Res.
Bull. 2009, 78, 91.
9. Bliese, M.; Tsanaktsidis, J. Aust. J. Chem. 1997, 50, 189.
10. Falkiner, M. J.; Littler, S. W.; McRae, K. J.; Savage, G. P.; Tsanaktsidis, J. Org.
Process Res. Dev. 2013, 1, 21.
Selected compounds 9, 12, 16, and 18 were identified as GluN2B
binders with greater potency than BZAD01 (6). Further evaluation
of activity of these compounds in 6-OHDA+Pf lesioned rats is cur-
rently under investigation, it is hoped that the further exploration
of these compounds will elaborate our understanding of GluN2B
pharmacology and potentially provide novel therapeutics for PD.
11. Gunosewoyo, H.; Guo, J. L.; Bennett, M. R.; Coster, M. J.; Kassiou, M. Bioorg. Med.
Chem. Lett. 2008, 18, 3720.