Synthesis and evaluation of novel tricyclic benzo[4.5]cyclohepta[1.2]pyridine derivatives as NMDA/NR2B antagonists

Article abstract of DOI:10.1016/j.bmcl.2009.07.028

A novel series of annulated tricyclic compounds was synthesized and evaluated as NMDA/NR2B antagonists. Structure-activity development was directed towards in vitro optimization of NR2B activity and selectivity over the hERG K⁺ channel. Preferr

Full text of DOI:10.1016/j.bmcl.2009.07.028

Bioorganic & Medicinal Chemistry Letters 19 (2009) 5132–5135
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and evaluation of novel tricyclic benzo[4.5]cyclohepta[1.2]pyridine
derivatives as NMDA/NR2B antagonists
a
b
b
a
Charles J. McIntyre a, , John A. McCauley , Bohumil Bednar , Rodney A. Bednar , John W. Butcher ,
David A. Claremon a, Michael E. Cunningham b, Roger M. Freidinger a, Stanley L. Gaul b, Carl F. Homnick a,
Ken S. Koblan b, Scott D. Mosser b, Joseph J. Romano a, Nigel J. Liverton a
*
a Department of Medicinal Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
b Department of Molecular Pharmacology, Merck Research Laboratories, West Point, PA 19486, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel series of annulated tricyclic compounds was synthesized and evaluated as NMDA/NR2B antago-
nists. Structure–activity development was directed towards in vitro optimization of NR2B activity and
selectivity over the hERG K+ channel. Preferred compounds were subsequently evaluated for selectivity
Received 19 May 2009
Revised 25 June 2009
Accepted 2 July 2009
Available online 10 July 2009
in an
a1-adrenergic receptor binding counter-screen and a cell-based assay of NR2B activity.
Ó 2009 Elsevier Ltd. All rights reserved.
Keyword:
NR2B heterocycle
In the central nervous system (CNS), the ligand-gated N-methyl-
-aspartate (NMDA) ion channel plays a critical role in the
In the CNS, small-molecule annulated tricyclics are established
D
primarily as antidepressant and antipsychotic agents.5 The recent
identification of a new class of tricyclic NR2B inhibitors6 (1, NR2B
Ki 19 nM, Fig. 1) inspired us to investigate a similar series of
compounds possessing a central carbocyclic ring (2, Fig. 1).7
Synthetic studies were initiated with the dual primary objectives
of identifying compounds with optimized NR2B binding activity
and minimized human ether-a-go-go (hERG)8 K+ channel activity.
SAR in this series was developed by making modifications to the
pendant substituents R1 and R2 as well as the central ring component
of the tricycle. Selected compounds were further evaluated for both
maintenance of normal synaptic activity and plasticity. Located in
the postsynaptic membrane, NMDA ion channels are activated by
strong membrane depolarization coupled with presynaptic release
of the neurotransmitter glutamate. Opening allows for calcium
and sodium ion flux into the cell.1 Over-activation of the NMDA
ion channel complex leads to cellular inundation by sodium and cal-
cium ions and this phenomenon has been implicated in a number of
diseasestates. Studiesof NMDAantagonistsin clinicaltrialsrevealed
that non-specific blockade of the NMDA ion channel by agents such
as ketamine (for pain) or memantine (Alzheimer’s disease) were
attended by deleterious side effects such as dysphoria and hallucina-
tion.2 A subtype-specific NMDA ion channel antagonist could afford
agents with an improved therapeutic index. NMDA ion channel
subtypes are constructed of a heteromultimeric assembly of NR1
subunits (a–h) and at least one of four NR2 subunits (a–h). Expres-
sion of the NR2B subunit is delimited to the forebrain in mammals
and has been linked to many neurological disorders including
parkinsonism, cerebral ischemia, psychosis or neuropathic pain.3
A small-molecule allostericinhibitor of theNMDA/NR2Bion channel
complex has the potential to give an improved therapeutic index
over non-specific NMDA antagonists thereby allowing for the devel-
opment of new treatment therapies for a number of neurological
diseases.4
selectivity versus the a
1-adrenergic9 receptor and for activity in a
cell-based assay of NR2B antagonism. Herein we wish to report the
development of an efficient, versatile synthesis of novel tricyclic
compounds having intrinsic inhibitory activity versus the NR2B
subtype of the NMDA ion channel complex.
R2
R1
Ph
5
CN
CN
N
N
N
H
10
O
11
2
1
NR2B Ki 19 nM
hERG IP 431 nM
* Corresponding author.
E-mail address: cjmcintyc@yahoo.com (C.J. McIntyre).
Figure 1. Tricyclic lactam lead (1) and benzo[4.5]cyclohepta[1.2] pyridine target
compounds (2).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.07.028

C. J. McIntyre et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5132–5135
5133
CO2H
Br
Ph
Ph
CO2CH3
CH3
CN
Ph
CO2H
CH3
Ph
v
i, ii
iii, iv
N
N
N
CH3
N
HCl
HCl
HCl
6
3
4
5
O
HO
O
vi
Ph
viii
Ph
vii
Ph
CN
N
CN
N
N
Br
9
8
7
Scheme 1. Reagents and conditions: (i) 50% NaOH (aq), EtOH, reflux, 24 h; (ii) HCl (aq), 100%; (iii) SOCl2, reflux, 0.75 h; (iv) MeOH, reflux, 0.5 h, 100%; (v) KOtBu, hexane, 0 °C,
0.75 h, 2-bromobenzaldehyde; (vi) polyphosphoric acid, 200 °C, 85%; (vii) CuCN, DMF, 120 °C, 24 h, 90%; (viii) NaBH4, EtOH, reflux, 10 min, 77%.
(Scheme 1). Hydrolysis of the nitrile (3) followed by esterification of
the resulting carboxylic acid (4) gave (5) in high yield.
Table 1
In vitro evaluation of compounds 8 and 9a
Treatment of the 2,3,5-trisubstituted pyridine ester (5) with
potassium-t-butoxide in hexane followed by reaction with 2-bromo-
benzaldehyde gave the Aldol product, 6 which was heated in poly-
phosphoric acid to afford the annulated tricyclic intermediate, 7.
Application of the RosenmundÀvon Braun aromatic cyanation
reaction12 gave a high yield of the aryl-nitrile 8. Reduction of the
ketone (8) with sodium borohydride gave the initial target com-
pound, rac-9.13
Subsequent in vitro evaluation of rac-9 showed it possessed
excellent activity (Ki 2 nM) when tested in an assay for NR2B bind-
ing (Table 1).14 Furthermore, only modest NR2B activity was
observed in the ketone intermediate 8 (Ki 1000 nM), thus suggest-
ing the importance of the pendant hydroxyl functionality for activ-
ity in this structural class. Resolution of rac-9 was accomplished by
Compound
NR2B binding Kib (nM)
hERG IPc (nM)
8
1000
2
540
1
882
5100
10,000
590
rac-9
(+)-9
(À)-9
a
Each entry represents the mean of three experiments.
b
Inhibition of 3H-[(E)-N1-(2-methoxybenzyl)-cinnamamidine] binding to
hNR1a/NR2B receptors expressed in Ltk-cells.14
Inhibition of MK-499 binding to hERG in HEK293 cells.16
c
Initially, a classic synthetic approach10 utilizing the known 3-cy-
ano-2-methyl-5-phenylpyridine11 (3) intermediate, was employed
for the constructionof the appropriately substituted tricyclic system
Table 2
Chemical modification of the central carbocycle of compound rac-9 and resulting NR2B and hERG activitiesa
Reaction conditions
HOAc, reflux, 24 h
Yield (%)
Number
Product
NR2B binding Kib (nM)
hERG IPc (nM)
653
AcO
Ph
41
10
10
CN
CN
N
H3CO
Ph
60% NaH, CH3I, MeCN, 15 min
71
48
11
12
4800
56
660
N
HO
Ph
Ph
Ph
5% Pd on BaSO4, H2 (balloon), EtOH, 72 h
Zn (2 equiv), HOAc, reflux, 24 h
3564
CN
CN
CN
N
N
N
26
55
13
14
2
65
Zn (16 equiv), HOAc, reflux, 48 h
77
154
a
Each entry represents the mean of three experiments.
b
c
Inhibition of 3H-[(E)-N1-(2-methoxybenzyl)-cinnamamidine] binding to hNR1a/NR2B receptors expressed in Ltk-cells.14
Inhibition of MK-499 binding to hERG in HEK293 cells.16

5134
C. J. McIntyre et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5132–5135
O
CO2H
Cl
O
Cl
iii
i
ii
Cl
Cl
CO2CH3
CH3
N
B
N
Br
N
CN
N
r
HCl
15
17
18
16
R
AcO
HO
Ar
Cl
vi
Cl
iv
v
CN
N
CN
N
CN
N
R = OAc 21
20
19
vii
R = OH 22a-i
Scheme 2. Reagents and conditions: (i) KOtBu, THF, 0 °C, 0.75 h, 2-bromobenzaldehyde, 72%; (ii) polyphosphoric acid, 200 °C, 24 h, 82%; (iii) CuCN, DMSO, 150 °C, 6 h; (iv)
NaBH4, EtOH, reflux, 10 min, 87% (two steps); (v) Ac2O, HOAc, reflux, 24 h, 58% (vi) ArB(OH)2, Pd2(dba)3, Cs2CO3, Cy3P, dioxane, 100 °C, 24 h, 53À88%; (vii) K2CO3 (aq), MeOH,
25 °C, 30 min, 48À93%.
chiral preparative HPLC.15 Unfortunately, NR2B and hERG16 activi-
ties tracked together such that the enantiomer with the greatest
NR2B activity, (À)-9 (Ki 1 nM), was also the most potent hERG K+
channel blocker (IP 590 nM). Additional synthetic modifications
of rac-9 were initiated to better define the SAR around the central
ring in an effort to achieve better separation of NR2B and hERG
activities.
The synthetic strategy outlined in Scheme 1 was useful in that it
allowed for either derivitization of the pendant hydroxyl group or a
step-wise reduction at carbons 5, 10 and 11 of the central ring (Fig. 1,
Table 2). For instance compound rac-9 was refluxed in acetic acid to
give the acetyl ester 10. Likewise the methyl ether (11) was prepared
by submitting rac-9 to Williamson etherification conditions. Selec-
tive reduction of the alkene in the presence of the benzylic hydroxyl
on the central carbocycle was accomplished by hydrogenating rac-9
in the presence of Rosenmund’s catalyst (5% Pd/BaSO4) to give the
saturated hydroxycarbocycle derivative 12. Reduction with zinc
(2 equiv) in refluxing acetic acid gave the deoxygenated product,
13. On extended refluxing with a large excess of zinc the fully
reduced and deoxygenated carbocycle, 14 resulted.
central ring was both reduced and deoxygenated as in 14.
Examples 12À14 suggest the importance of the pendant hydroxyl
for achieving separation between NR2B and hERG activities.
The synthesis summarized in Scheme 1 gave NR2B active
compounds and allowed for limited SAR development of the
central carbocycle as illustrated in Table 2. This route did however
necessitate the early installation of the pendant aryl group in the 3-
position of the pyridyl ring and this made subsequent development
of SAR in this region of the molecule difficult. It was envisaged that
a 3-halo substituent on the pyridine would allow for late-stage
Suzuki17 palladium mediated cross-couplings with boronic acids
or amines thus affording synthetic products with both electronic
and steric variation at the 3-pyridyl position of the tricyclic struc-
ture (Scheme 2).
Scheme 2 illustrates this alternative synthesis of annulated tri-
cylics starting from the known 2,3,5-trisubstituted pyridine (15).18
A key synthetic design feature concerned the utilization of a chlo-
HO
Assessment of the acetyl ester (10) revealed a fivefold decrease
in NR2B activity and a sevenfold increase in hERG activity versus
rac-9. NR2B activity was even further diminished in the case of
the methyl ether derivative 11. Reduction of the alkene with reten-
tion of the pendant hydroxyl gave a derivative with reduced NR2B
activity (12) while the deoxygenated derivative (13) showed NR2B
activity comparable to rac-9. No advantage resulted when the
Het
i, ii
18
CN
N
23 a-f
Scheme 3. Reagents and conditions: (i) amine, Pd2(dba)3, rac-BINAP, NaOtBu,
toluene, 100 °C, 24 h, 23–44%; (ii) NaBH4, EtOH, reflux, 10 min, 48–60%.
Table 3
Aryl group variation of 22 a–i and resulting NR2B and hERG activitiesa
Table 4
Heterocycle analogs 23 a–f and resulting in vitro activitiesa
Ar
Number
NR2B binding Kib (nM)
hERG IPc (nM)
Het
Number
NR2B binding Kib (nM)
hERG IPc (nM)
o-Tolyl
m-Tolyl
p-Tolyl
o-MeO-phenyl
m-MeO-phenyl
p-MeO-phenyl
o-Fluorophenyl
m-Fluorophenyl
p-Fluorophenyl
22a
22b
22c
22d
22e
22f
22g
22h
22i
6
4
10
13
9
1351
1100
1698
5455
1534
1612
4445
4968
512
Pyrrolidine
Piperidine
Piperidine
Piperidine
Homopiperidine
Morpholine
4-Me-piperidine
4-Me-piperazine
23a
720
15,500
10,000
20,000
20,000
3000
20,000
5300
42,000
rac-23b
(+)-23b
(À)-23b
23c
43
35,000
17
8
428
16
23
3
23d
196
23e
83
23f
>16,000
a
a
Each entry represents the mean of three experiments.
Each entry represents the mean of three experiments.
b
b
Inhibition of 3H-[(E)-N1-(2-methoxybenzyl)-cinnamamidine] binding to
Inhibition of 3H-[(E)-N1-(2-methoxybenzyl)-cinnamamidine] binding to
hNR1a/NR2B receptors expressed in Ltk-cells.14
hNR1a/NR2B receptors expressed in Ltk-cells.14
Inhibition of MK-499 binding to hERG in HEK293 cells.16
Inhibition of MK-499 binding to hERG in HEK293 cells.16
c
c

C. J. McIntyre et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5132–5135
5135
rine atom as the appropriate halogen in the 5-position of the pyr-
idine ring in 15. This allowed the RosenmundÀvon Braun cyana-
tion reaction (17?18) to proceed with the bromide while
retaining the heteroaryl-chloride for subsequent participation in
cross-coupling reactions (20?21).
ries) for providing scaled-up batches of synthetic intermediates
and Joan S. Murphy (Merck Research Laboratories) for high resolu-
tion mass spectral analysis of all final compounds.
References and notes
Data for the products obtained via this chemistry is illustrated
in Table 3. Although this set of compounds possessed pendant aryl
rings with diverse electronic and steric topology, testing did not
reveal the emergence of any clear SAR trends regarding NR2B
and hERG activities. There remained, however, a tendency for
NR2B and hERG activities tracking together as in 22h and 22i.
A third synthetic iteration on the tricyclic motif involved the
introduction of a weakly basic functionality into the 3-position of
the fused pyridine heterocycle. Scheme 3 shows the two-step elab-
oration of intermediate 18 using Buchwald19 amination chemistry
to afford tertiary amine analogs (23a–f).
Derivatives possessing weakly basic cyclic amine in the 3-pyridyl
position (Table 4) were not as active against NR2B as their aryl coun-
terparts (Tables 1 and 3). Pyrrolidine (23a), homopiperidine (23c),
morpholine (23d) and N-methylpiperazine (23f) derivatives exhib-
ited poor activity when evaluated in the NR2B binding assay
(Ki 196À16,000 nM). Modest NR2B activity was observed for the 4-
methylpiperidine (23e, Ki 83 nM) derivative. The piperidine (rac-
23b) analog, however, showed both good NR2B activity (Ki 43 nM)
and an absence of hERG ion channel activity (IP 10,000 nM). Further-
more, chiral preparative HPLC resolution of rac-23b gave an active
NR2B compound ((À)-23b, 17 nM) with greater than 1000-fold
selectivity over hERG (IP 20,000 nM).
1. Kennedy, M. B. Science 2000, 290, 750.
2. Muir, K. W. Curr. Opin. .Pharmacol. 2006, 6, 53.
3. Loftis, J. M.; Janowsky, A. Pharmacol. Ther. 2003, 97, 55.
4. (a) Chazot, P. L. Curr. Med. Chem. 2004, 11, 389; (b) Gogas, K. R. Curr. Opin.
Pharmacol. 2006, 6, 68; (c) Layton, M. E.; Kelly, M. J., III; Rodzinak, K. J. Curr. Top.
Med. Chem. 2006, 6, 697; (d) McCauley, J. A. Exp. Opin. Ther. Pat. 2005, 15, 389.
5. (a) Baldessarini, R. J. In Goodman s the Pharmacological Basis of Therapeutics;
Hardman, J. G., Limbird, L. E., Eds., 10th ed.; McGraw-Hill: New York, 2001; pp
451–468. 486–490; (b) Eisendrath, S. J.; Lichtmacher, J. E. In Current Medical
Diagnosis & Treatment; Tierney, L. M., Jr., McPhee, S. J., Papadakis, M. A., Eds.,
45th ed.; Lange Medical Books/McGraw-Hill: New York, 2006; pp 1058–1061.
1071; (c) Dubroeucq, M.-C.; Stutzmann, J.-M.; Manfré, F.; Capet, M.; Böhme, G.
A.. In Mcguire, J. L., Ed.; Pharmaceuticals Classes, Therapeutic Agents, Areas of
Application; Wiley: Weinheim, 2000; Vol. 2, pp 588–605. 610–620.
6. Details concerning the synthesis and structure–activity-development that lead
to the discovery of 1 will be the subject of a future communication.
7. McCauley, J. A.; Butcher, J. W.; Claremon, D. A.; Liverton, N. J.; McIntyre, C. J.;
Romano, J. J. WO03084931; Chem. Abstr. 2003, 139, 323441.
8. (a) Vaz, R. J.; Li, Y.; Rampe, D.. In Progress in Medicinal Chemistry; King, F. D.,
Lawton, G., Eds.; Elsevier: Amsterdam, 2005; Vol. 43, pp 1–18; (b) Pearlstein,
R.; Vaz, R.; Rampe, D. J. Med. Chem. 2003, 46, 2017.
9. Calzada, B. C.; De Artiñano, A. A. Pharmacol. Res. 2001, 44, 195.
10. (a) Villani, F. J.; Daniels, P. J. L.; Ellis, C. A.; Mann, T. A.; Wang, K.-C. J. Heterocycl.
Chem. 1971, 8, 73; (b) Brenner, D. G.; Halczenko, W.; Shepard, K. L. J. Heterocycl.
Chem. 1982, 19, 897.
11. Christian, J.; Loebering, H. G.; Trinkl, K. H. Synthesis 1977, 5, 326.
12. Wagner, R. B.; Zook, H. D. Synthetic Organic Chemistry; Wiley & Sons: New York,
1953.
13. All final compounds gave a satisfactory 1H NMR, C, H and N combustion
analysis and high resolution mass spectral data. Representative
characterization data for rac-9: 1H NMR 400 MHz (CDCl3) d 8.80 (d, J = 2 Hz,
1H); 8.26 (s, 1H); 8.04 (d, J = 8 Hz, 1H); 7.69–7.40 (m, 9H); 5.39 (d, J = 4 Hz,
1H); 2.65 (d, J = 4 Hz, 1H). Anal. Calcd for C21H14N2OÁ0.35 H2O: C, 79.65; H,
4.68; N, 8.85. Found: C, 79.59; H, 4.46; N, 8.84. HRMS for C21H15N2O (M+H)+,
calcd: 311.1179, found: 311.1173.
Two compounds, (À)-9 and (À)-23b, were subsequently as-
sessed in both an
a
1-adrenergic receptor20 counter-screen assay
and a NR2B Ca2+ influx cell-based assay. Results for the
a1-adren-
ergic receptor assay ((À)-9 Ki 5,000 nM, (À)-23b Ki > 21,000 nM)
showed that the pendant piperidine analog had fourfold greater
selectivity than the phenyl compound. Consistent with the NR2B
binding assay results, comparison of (À)-9 and (À)-23b in a
NR2B Ca2+ influx cell-based assay7 showed a threefold loss in activ-
ity ((À)-9 IC50 9 nM, (À)-23b IC50 27 nM) for the piperidine
derivative.
In summary, a new class of NMDA/NR2B inhibitors were discov-
ered and their SAR investigated. Presence of a hydroxyl substituent
on carbon-5 of the central ring and replacement of a pendant aryl
with a piperidine heterocycle were necessary elements for achiev-
ing separation of NR2B and hERG activities. Compound (À)-23b
possesses a balanced profile with good NR2B activity and selectiv-
14. (a) Kiss, L.; Cheng, G.; Bednar, B.; Bednar, R. A.; Bennett, P. B.; Kane, S. A.;
McIntyre, C. J.; McCauley, J. A.; Koblan, K. S. Neurochem. Int. 2005, 46, 453; (b)
Mosser, S. D.; Gaul, S. L.; Bednar, B.; Koblan, K. S.; Bednar, R. A. J. Assoc. Lab.
Automat. 2003, 8, 54.
15. Chiral preparative HPLC chromatography was conducted using a ChirapakÒ AS
250 Â 4.6 mm column, 70% hexane (0.1% diethylamine)/30% ethanol, flow
1.0 mL/min, absorbance 310 nm.
16. Butcher, J. W.; Claremon, D. A.; Connolly, T. M.; Dean, D. C.; Karczewski, J.;
Koblan, K. S.; Kostura, M. J.; Liverton, N. J.; Melillo, D. G., PCT Int. Appl. WO
2002005860, 2002.
17. (a) Old, D. W.; Wolfe, J. P.; Buchwald, S. L. J. Am. Chem. Soc. 1998, 120, 9722; (b)
Littke, A. F.; Fu, G. C. Angew. Chem., Int. Ed. 1998, 37, 3387.
18. Marcoux, J.-F.; Marcotte, F.-A.; Wu, J.; Dormer, P. G.; Davies, I. W.; Hughes, D.;
Reider, P. J. J. Org. Chem. 2001, 66, 4194.
19. Wolfe, J. P.; Buchwald, S. L. Angew. Chem., Int. Ed. 1999, 38, 2413.
20.
a
1-Adrenergic binding assays were conducted according to: Greengrass, P.;
ity over hERG and
a1-adrenergic receptors.
Bremner, R. Eur. J. Pharmacol. 1979, 55, 323.
Acknowledgements
The authors wish to thank Jean-François Marcoux (Merck
Process Research) and Garry R. Smith (Merck Research Laborato-