Novel, druglike 1,7-disubstituted 2,3,4,5-tetrahydro-1H-benzo[b]azepine- based selective inhibitors of human neuronal nitric oxide synthase (nNOS)

Article abstract of DOI:10.1016/j.bmcl.2012.02.004

A novel class of 1,7-disubstituted 2,3,4,5-tetrahydro-1H-benzo[b]azepine derivatives was designed, synthesized and evaluated as human nitric oxide synthase (NOS) inhibitors. Structure-activity relationship studies based on various basic amine side chains

Full text of DOI:10.1016/j.bmcl.2012.02.004

Bioorganic & Medicinal Chemistry Letters 22 (2012) 2510–2513
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Novel, druglike 1,7-disubstituted 2,3,4,5-tetrahydro-1H-benzo[b]azepine-based
selective inhibitors of human neuronal nitric oxide synthase (nNOS)
Subhash C. Annedi a, , Jailall Ramnauth a, Michele Cossette a, Shawn P. Maddaford a, Peter Dove a, Suman
⇑
Rakhit a, John S. Andrews a, Frank Porreca b
a NeurAxon Inc., 2395 Speakman Drive, Suite # 1001, Mississauga, ON, Canada L5K 1B3
b Department of Pharmacology, University of Arizona, 1501 N. Campbell Ave., Tucson, AZ, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel class of 1,7-disubstituted 2,3,4,5-tetrahydro-1H-benzo[b]azepine derivatives was designed, syn-
thesized and evaluated as human nitric oxide synthase (NOS) inhibitors. Structure–activity relationship
studies based on various basic amine side chains attached at the 1-position of the 2,3,4,5-tetrahydro-1H-
benzo[b]azepine ring led to the identification of several potent and highly selective inhibitors (17, 18, 25,
( )-39, and ( )-40) of human neuronal NOS. The potential therapeutic application of one of these new
selective nNOS inhibitors (17) was demonstrated in an in vivo spinal nerve ligation model of neuropathic
pain, and various in vitro safety pharmacology studies such as the hERG K+ channel inhibition assay and
high throughput broad screen (minimal activity at 79 receptors/transporters/ion channels).
Ó 2012 Elsevier Ltd. All rights reserved.
Received 11 January 2012
Revised 31 January 2012
Accepted 1 February 2012
Available online 9 February 2012
Keywords:
2,3,4,5-Tetrahydro-1H-benzo[b]azepine
derivatives
Nitric oxide
Nitric oxide synthase
Selective neuronal nitric oxide synthase
inhibitors
Neuropathic pain
Nitric oxide (NO), an endogenous free radical gas that has diverse
roles in a number of physiological processes such as regulation of
blood pressure, neurotransmission and macrophage defense sys-
tems.1,2 The synthesis of NO is catalyzed by three different isoforms
of nitric oxide synthase (NOS), two constitutive forms in neuronal
cells (nNOS) and endothelial cells (eNOS), and an inducible form
NOS isoforms appears to be relatively conserved with 16 out of
18 residues within 6 Å being identical, which presumably explains
the difficulty in obtaining selective NOS inhibitors.9 Another major
challenge in developing selective NOS inhibitors with druglike
properties is to target the relatively polar and acidic active site of
NOS.10 Over two decades of extensive research resulted in the
development of several structural classes of peptidic and nonpep-
tidic selective NOS inhibitors targeting the arginine binding site or
BH4 site as well as dimerization inhibitors.7
The pharmacophore model adapted by our group that targets
the arginine binding site of the NOS enzyme contains a guanidine
isosteric group and a basic amine group, attached to a central aryl
scaffold as described earlier.11–13 The amidine group makes an
important bidentate interaction with the conserved glutamic acid
residue; whereas the basic amine is shown to improve potency
and selectivity among the NOS isoforms (1–3, Fig. 1).14–16 In our
continued search for new small molecule druglike selective nNOS
inhibitors,11–16 herein we report the synthesis and biological activ-
ity evaluations of a novel class of 1,7-disubstituted 2,3,4,5-tetrahy-
dro-1H-benzo[b]azepine-based17 thiophene amidine compounds
as selective nNOS inhibitors.
in macrophage cells (iNOS).3 NOS converts
L-arginine to L-citrulline
and NO utilizing NADPH and O2 as substrates, and flavinadenine
dinucleotide (FAD), flavin mononucleotide (FMN), tetrahydrobiop-
terin (BH4) and iron protoporphyrin (heme) as cofactors.4 Despite
the beneficial effects of NO, overproduction by specific isoforms
such as nNOS and iNOS has been shown to play a key role in several
disorders, including septic shock, arthritis, diabetes, ischemia–
reperfusion injury, pain and various neurodegenerative diseases
such as Parkinson’s disease, MS, and Alzheimer’s disease.5 However,
any NOS inhibitors to treat these conditions must avoid eNOS inhi-
bition due to its important role in regulating blood pressure.6 There-
fore, the development of selective NOS inhibitors is of considerable
interest from a therapeutic perspective.7
Early NOS inhibitors (based on the endogenous substrate,
L-arginine) has been instrumental in identifying a specific role for
NO, however they do not effectively distinguish among the NOS
The synthesis of all target compounds was carried out as shown
in schemes 1 through 3. Alkylation of 4 with various alkyl halides
in presence of a base provided 5–8 (Scheme 1), respectively. The
reduction of nitro group into the corresponding amine was carried
out under standard hydrogenation conditions to obtain the
isoforms.8 At the same time, the active sites of the three different
⇑
Corresponding author. Tel.: +1 416 673 8457; fax: +1 905 823 5836.
E-mail address: sannedi@neuraxon.com (S.C. Annedi).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.02.004

S. C. Annedi et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2510–2513
2511
H
N
N
R1
S
NH
a
d
NH
5
N
N
N
N
H
NH
S
N
H
N
N
N
H
R2
R
S
NH
2
1
21: R1 = NO2, R2 = Me
b
c
24: R = CO2Ph
25: R = H
22: R1 = NO2, R2 = CO2Ph
23: R1 = NH2, R2 = CO2Ph
e
H
N
S
Scheme 2. Reagents and conditions: (a) BH3ÁTHF, THF, reflux, 75%; (b) PhOCOCl,
CH2Cl2, rt, 82%; (c) Pd-C/H2, EtOH, rt, 79%; (d) methyl thiophene-2-carbimidothioate
hydroiodide, EtOH, rt, 53%; (e) NaOH, EtOH:H2O, 100 °C, 41%.
NH
N
N
under neutral conditions using NBS to obtain 7-bromo derivatives
30–32, respectively in excellent yields. Removal of the Boc-protect-
ing group in 32 under strong acidic conditions followed by reduc-
tive amination of the obtained amine (33) with formaldehyde
provided the N-methyl derivative 34. The bromine in 30–32 and
34 was converted into the corresponding amine by a Buchwald–
Hartwig amination reaction using LiHMDS as an ammonia surro-
gate to obtain 35, 36, 38, and 40, respectively.19 Removal of the
Boc-protecting group in 36 and 38 was carried out under acidic
conditions as described above to obtain the target compounds 37
and 39, respectively. No attempts were made to separate ( )-20,
( )-39, and ( )-40 at this time and were tested as racemates.
All compounds were converted into their corresponding dihy-
drochloride salts20 and inhibitory activities were measured against
all three human NOS isoforms (Table 1) as described earlier.15
Recombinant human nNOS, eNOS and iNOS were produced in
3
Figure 1. Recently reported selective nNOS inhibitors.
intermediate amines 9–12, respectively in excellent yields. Reduc-
tion of the amide group into the corresponding amine was carried
out using LiAlH4 to obtain 13–16, respectively. The free amines
were then coupled to methyl thiophene-2-carbimidothioate hyd-
roiodide18 to obtain the target compounds 17–20, respectively.
Reduction of the amide group in 5 into the corresponding amine
21 was carried out using borane in THF (Scheme 2). N-demethyla-
tion and concomitant protection of 21 was carried out with phe-
nylchloroformate to obtain the phenyl carbamate 22. Reduction
of the nitro group into the corresponding amine (23), followed
by coupling to methyl thiophene-2-carbimidothioate hydroiodide
provided 24. The phenyl carbamate in 24 was removed under
strong basic conditions to obtain the target compound 25.
The cyclic basic amine side chains were introduced by the
reductive amination of 2,3,4,5-tetrahydro-1H-benzo[b]azepine
(26) with various 6- and 5-membered ketones to obtain 27–29,
respectively (Scheme 3). Bromination (27–29) was carried out
a
b
N
N
H
N
O
R
n
n
26
N
O2N
O2N
a
b
R
27:
n = 2, R = Me
28: n = 2, R = Boc
N
H
N
R
O
O
29:
n = 1, R = Boc
4
5-8
H
N
Br
S
H
N
H2N
d
S
NH
N
N
e
NH
N
R
N
R
X
n
n
17-20
9-12: X = O
N
R
N
c
13-16: X = H2
R
35:
n = 2, R = Me
30: n = 2, R = Me
36: n = 2, R = Boc
37: n = 2, R = H
31:
n = 2, R = Boc
32: n = 1, R = Boc
c
c
5, 9, 13, 17: R =
7, 11, 15, 19: R =
, 6, 10, 14, 18: R =
N
N
c
d
38:
n = 1, R = Boc
33:
n = 1, R = H
34: n = 1, R = Me
39: n = 1, R = H
40:
N
n = 1, R = Me
, 8, 12, 16, 20: R =
N
Scheme 3. Reagents and conditions: (a) R = O, NaBH(OAc)3, AcOH, DCE, rt, 49–94%;
(b) NBS, DMF, rt, 83–100%; (c) 1 N HCl, MeOH, reflux, 79–93%; (d) formaldehyde,
NaBH(OAc)3, AcOH, DCE, rt, 90%; (e) (i) PtBu3, Pd2(dba)3, LiMDS, THF, 100 °C; (ii)
methyl thiophene-2-carbimidothioate hydroiodide, EtOH, rt, 48–85%.
Scheme 1. Reagents and conditions: (a) R-Cl, NaH, DMF, 49–94%; (b) Pd-C/H2,
EtOH, rt, 80–100%; (c) LiAlH4, THF, rt, 35–94%; (d) methyl thiophene-2-carbimido-
thioate hydroiodide, EtOH, rt, 36–83%.

2512
S. C. Annedi et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2510–2513
Table 1
Inhibition of human NOS enzymes by 1,7-disubstituted 2,3,4,5-tetrahydro-1H-benzo[b]azepine derivatives
H
N
S
NH
N
R
Compound
R
Human NOS IC50
eNOS
(l
M)a
Selectivity
nNOS
iNOS
eNOS/nNOS
iNOS/nNOS
0.44 (0.20–0.93)
0.12 (0.05–0.28)b
99.1 (31.8–307)
1.50 (0.73–3.07)
>100b,c
225
522
9
>833
62.6 (35.5–110.4)b
17
18
N
N
0.16 (0.11–0.22)
1.96 (1.17–3.24)
82.2 (29.5–228)
16.9 (4.84–59.2)
>100c
514
29
106
>51
N
19
56.3 (33.8–93.5)
30.8 (20.5–46.1)
N
( )-20
1.86 (0.79–4.36)
>100c
17
>54
H
N
25
35
0.10 (0.04–0.23)
2.96 (1.35–6.44)
47.1 (15.4–143)
20.0 (9.85–40.7)
>100c
471
7
>1000
0.5
1.34 (0.80–2.23)
N
37
2.38 (1.70–3.33)
0.33 (0.12–0.91)
66.2 (58.3-71.4)
75.0 (22.9–245)
>100c
28
>42
41
HN
HN
( )-39
13.5 (6.9–26.5)
227
( )-40
0.08 (0.02–0.24)
0.09 (0.05–0.19)
60.5 (34.0–107)
45.6 (8.6–242)
>100c
756
465
>1250
92
N
3d
25.7 (17–39)
Values reported in parentheses are 95% confidence intervals.
a
In a radiometric method, inhibitory activities were measured by the conversion of [3H]-
L L-citrulline.
-arginine into [3H]-
b
c
Inhibitory values are for rat NOS isoforms.
>100, Not active at the maximum tested concentration of 100
lM.
d
Compound 3 is previously reported selective nNOS inhibitor, included for comparison.
Baculovirus-260 infected Sf9 cells. In a radiometric method, inhibi-
tory activities were measured by the conversion of [3H]-
-arginine
into [3H]-
-citrulline. Most of the compounds in the series showed
sub-micromolar potency for nNOS and good to excellent selectivity
over the eNOS and iNOS isoforms. The most potent nNOS inhibitor
Due to the polar and acidic nature of the NOS active site, inhib-
itors that mimic the substrate -arginine are likely to be polar and
L
L
L
basic, which pose a major challenge (blood–brain barrier and cell
permeability issues) in designing selective NOS inhibitors to treat
CNS and related disorders.21 To assess the druglike characteristics
of these compounds, the physicochemical properties were calcu-
lated and shown to obey Lipinski’s rule of five (LogP <5, hydrogen
bond donors <5, hydrogen bond acceptors <10, molecular weight
<500).22 Overall, most of these compounds have ideal physico-
chemical properties required for drug candidates targeting CNS re-
lated disorders.
was ( )-40 (IC50 = 0.08 lM) with excellent selectivity for nNOS over
eNOS (756-fold) and iNOS (>1250-fold) isoforms among the current
series. Bulky basic amine substitutions (19, ( )-20, 36, and 37)
showed 4 to 30-fold weaker inhibitory activities against the nNOS
isoform when compared to the smaller basic amine side chains
(17, 18, 25, ( )-39, and ( )-40). Similarly poor selectivity for nNOS
over the eNOS and iNOS isoforms was observed with bulkier side
chains (19, ( )-20, 36, and 37) when compared to smaller basic
amine side chains (17, 18, 25, ( )-39, ( )-40) in contrast to the
previously reported 2-aminobenzothiazole-,11 indole-,12–14
quinoline-,15 and indoline-based16 selective nNOS inhibitors, where
bulky basic amine substitutions provided better selectivity for nNOS
over the eNOS and iNOS isoforms. N-demethylation of 2-(dimethyl-
amino)ethyl side chain (17) into 2-(methylamino)ethyl side chain
(25) resulted in an increased potency for nNOS (fourfold) and eNOS
(twofold) isoforms with improved selectivity for nNOS over eNOS
(twofold) and iNOS (>100-fold) isoforms. N-methylation of the pyr-
rolidine side chain (( )-40) resulted in an increased potency for the
nNOS isoform (fourfold) with better selectivity over the eNOS
(threefold) and iNOS (>30-fold) isoforms when compared to the
pyrrolidine side chain (( )-39).
To investigate the potential therapeutic application of these
selective nNOS inhibitors, 17 was selected for further profiling in
an in vivo neuropathic pain model (Fig. 2). The Chung or spinal
nerve ligation (SNL) model involves ligation of both the L5 and L6
spinal nerves of one side of the rat, which produce thermal hyper-
algesia and mechanical allodynia of the affected paw.23 The ther-
mal hyperalgesia was tested by the withdrawal latency to
application of radiant heat to the paw. Intraperitoneal administra-
tion of 17 at 30 mg/kg resulted in complete reversal of thermal
hyperalgesia at 30 min with activity lasting up to 120 min
(Fig. 2). After the positive in vivo data, 17 was tested in human re-
combinant HEK-293 cell lines24 for its ability to inhibit the hERG K+
channel considering the strong hERG activity with reference com-
pound 3 (hERG K+ channel: IC50 = 4.7
l
M).15 Compound 17 was
inactive (up to 30 M) in the conventional patch-clamp assay
l

S. C. Annedi et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2510–2513
2513
The cardiovascular safety and minimal side activity profile dis-
closed in this report makes 17 an ideal candidate for further evalu-
ation in CNS related disorders where excess NO produced by nNOS
plays a significant role.
30
20
10
0
17 (30 mg/kg, i.p.)
Acknowledgments
We are grateful to NoAb BioDiscoveries Inc. (Mississauga, ON,
Canada), Asinex Ltd (Moscow, Russia) for performing the human
NOS inhibition assays and Cerep (France) for hERG K+ channel as-
say and broad screen profile. We thank the Natural Sciences and
Engineering Research Council (NSERC) of Canada for providing an
undergraduate student research award (USRA) for M.C.
References and notes
BL
Post-SNL
30
60
90
120
Time (min) after injection
1. Mustafa, A. K.; Gadalla, M. M.; Synder, S. H. Sci. Signal. 2009, 2, 1.
2. Moncada, S.; Palmer, R. M. J.; Higgs, E. A. Pharmacol. Rev. 1991, 43, 109.
3. Alderton, W. K.; Cooper, C. E.; Knowles, R. G. Biochem. J. 2001, 357, 593.
4. Stuehr, D. J. Annu. Rev. Pharmacol. Toxicol. 1997, 37, 339.
5. Vallance, P.; Leiper, J. Nat. Rev. Drug Disc. 2002, 1, 939.
6. Huang, P. L.; Huang, Z.; Mashimo, H.; Bloch, K. D.; Moskowitz, M. A.; Bevan, J.
A.; Fishman, M. C. Nature 1995, 377, 239.
7. Maddaford, S.; Annedi, S. C.; Ramnauth, J.; Rakhit, S. Annu. Rep. Med. Chem.
2009, 44, 27.
8. Erdal, E. P.; Litzinger, E. A.; Seo, J.; Zhu, Y.; Ji, H.; Silverman, R. B. Curr. Top. Med.
Chem. 2005, 5, 603.
Figure 2. Compound 17 attenuates thermal hyperalgesia in the L5/L6 SNL model of
neuropathic pain. The mean values reported are from six animals.
against hERG K+ channel, suggesting that the compound would be
devoid of any cardiovascular liabilities associated with hERG K+
channel binding that are physiologically relevant.
To identify the off-target activity (side activity) of 17, it was
tested in 79 validated in vitro pharmacological assays (receptors/
transporters/ion channels) covering a broad range of targets (Cerep
broad screen).25 This cost effective high throughput profile is used
to identify the potential limitations and liabilities of the lead can-
didate associated with off target activities during the selection pro-
9. Haitao, J.; Huiying, L.; Marta´sek, P.; Roman, L. J.; Poulos, T. L.; Silverman, R. B. J.
Med. Chem. 2009, 52, 779.
10. Silverman, R. B. Acc. Chem. Res. 2009, 42, 439.
11. Patman, J.; Bhardwaj, N.; Ramnauth, J.; Annedi, S. C.; Renton, P.; Maddaford, S.
P.; Rakhit, S.; Andrews, J. S. Bioorg. Med. Chem. Lett. 2007, 17, 2540.
12. Maddaford, S.; Renton, P.; Speed, J.; Annedi, S. C.; Ramnauth, J.; Rakhit, S.;
Andrews, J.; Mladenova, G.; Majuta, L.; Porreca, F. Bioorg. Med. Chem. Lett. 2011,
21, 5234.
13. Renton, P.; Speed, J.; Maddaford, S.; Annedi, S. C.; Ramnauth, J.; Rakhit, S.;
Andrews, J. Bioorg. Med. Chem. Lett. 2011, 21, 5301.
14. Annedi, S. C.; Maddaford, S. P.; Mladenova, G.; Ramnauth, J.; Rakhit, S.;
Andrews, J.; Lee, D. K. H.; Zhang, D.; Porreca, F.; Bunton, D.; Christie, L. J. Med.
Chem. 2011, 54, 7408.
15. Ramnauth, J.; Speed, J.; Maddaford, S. P.; Dove, P.; Annedi, S. C.; Renton, P.;
Rakhit, S.; Andrews, J.; Silverman, S.; Mladenova, G.; Zinghini, S.; Nair, S.;
Catalano, C.; Lee, D. K. H.; Felice, M. D.; Porreca, F. J. Med. Chem. 2011, 54, 5562.
16. Annedi, S. C.; Ramnauth, J.; Maddaford, S. P.; Renton, P.; Rakhit, S.; Mladenova,
G.; Dove, P.; Silverman, S.; Andrews, J.; Felice, M. D.; Porreca, F. J. Med. Chem.
2012, 55, 943.
cess. Accordingly, 17 was tested at a single concentration of 10 lM
in 79 different pharmacological assays to identify the off target
activity. Compound 17 was found to be not active for most of the
targets tested (<70% inhibition at 73 of 79) and showed significant
inhibition only at the following targets: alpha-2 adrenergic recep-
tor (a2: 81%), human muscarinic receptor (M1: 81%, M2: 86%, M3:
91%, M4: 92%) and 5-HT1B (100%). Overall 17 possess an excellent
in vitro safety (side activity) profile.
In conclusion, a novel class of 1,7-disubstituted 2,3,4,5-tetrahy-
dro-1H-benzo[b]azepine-based thiophene amidine compounds
was designed, synthesized and shown to be selective inhibitors of
human nNOS. In this study, a 1,7-disubstituted 2,3,4,5-tetrahy-
dro-1H-benzo[b]azepine ring was employed as the central aromatic
linker and various basic amine side chains at the 1-position of the
ring were investigated. Compounds with smaller basic amine side
chains (17, 18, 25, ( )-39, and ( )-40) showed sub-micromolar po-
tency for nNOS isoform and excellent selectivity over the eNOS and
iNOS isoforms when compared to the bulkier basic amine side
chains (19, ( )-20, 36, and 37). The potential therapeutic applica-
tion of these compounds was demonstrated by the reversal of ther-
mal hyperalgesia in an in vivo model of neuropathic pain (with 17).
17. Vaquero, J. J.; Cuadro, A. M.; Herradon, B. Mod. Heterocycl. Chem. 2011, 4, 1865.
18. Bercot-Vatteroni, M. Ann. Chim. 1962, 7, 303.
19. Huang, X.; Buchwald, S. L. Org. Lett. 2001, 3, 3417.
20. A solution of the free base (1.0 equiv) in methanol was treated with 1 M HCl
solution in diethyl ether (3.0 equiv) dropwise at room temperature. The
resulting mixture was stirred for 10 min and concentrated to dryness. The
product was dried under reduced pressure to obtain the dihydrochloride salt as
a solid.
21. Ji, H.; Stanton, B. Z.; Igarashi, J.; Li, H.; Martasek, P.; Roman, L. J.; Poulos, T. L.;
Silverman, R. B. J. Am. Chem. Soc. 2008, 130, 3900.
22. Lipinski, C. A. J. Pharmacol. Toxicol. Methods 2000, 44, 235.
23. Kim, S. H.; Chung, J. M. Pain 1992, 50, 355.
24. Zhou, Z.; Gong, Q.; Ye, B.; Fan, Z.; Makielski, J. C.; Robertson, G. A.; January, C. T.
Biophys. J. 1998, 74, 230.
25. http://www.cerep.fr/.