Dibenzazepines and dibenzoxazepines as sodium channel blockers

Article abstract of DOI:10.1016/j.bmcl.2014.11.025

We have identified two related series of dibenzazepine and dibenzoxazepine sodium channel blockers, which showed good potency on Na_v1.7 in FLIPR-based and electrophysiological functional assays.

Full text of DOI:10.1016/j.bmcl.2014.11.025

Bioorganic & Medicinal Chemistry Letters 25 (2015) 43–47
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Dibenzazepines and dibenzoxazepines as sodium channel blockers
⇑
Stephen M. Lynch, Laykea Tafesse , Kevin Carlin, Parijat Ghatak, Donald J. Kyle
Discovery Research, Purdue Pharma L.P., 6 Cedarbrook Drive, Cranbury, NJ 08512, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
We have identified two related series of dibenzazepine and dibenzoxazepine sodium channel blockers,
which showed good potency on Nav1.7 in FLIPR-based and electrophysiological functional assays.
Ó 2014 Elsevier Ltd. All rights reserved.
Received 25 September 2014
Revised 4 November 2014
Accepted 6 November 2014
Available online 14 November 2014
Keywords:
Neuropathic pain
Sodium channel
Nav1.7
Dibenzazepine
Dibenzoxazepine
Neuropathic pain is a debilitating condition that affects an esti-
mated 3–4.5% of the global population.1 It is prevalent in patients
suffering from metabolic disorders such as diabetes, infectious dis-
eases such as HIV and shingles and as a side effect from both phys-
ical traumas and various cancer chemotherapies. The chronic
discomfort associated with neuropathic pain is caused by lesions
or dysfunctions in the somatosensory pathways of either the cen-
tral or peripheral nervous system and is often characterized by
allodynia, hyperalgesia and dysesthesia.2,3 The persistent sensation
of pain commonly leads to a dramatic decrease in quality of life
and loss of workforce productivity. Front line treatment for neuro-
pathic pain currently consists of antidepressants, anticonvulsants
and analgesics. Unfortunately there remains a high unmet medical
need in this field since only an estimated 40–60% of patients
achieve partial relief with these current pharmacologic treatments,
some of which impart deleterious CNS-related side effects.4
A compelling body of evidence has implicated the role of volt-
age gated sodium channels as a contributor to neuropathic pain.5,6
Voltage gated sodium channels are believed to be important for
establishing and maintaining firing patterns in primary afferent
neurons. Blockage of these channels in dysfunctional neurons
could therefore inhibit action potential firing and reduce pain sen-
sitivity. The weak sodium channel blocker carbamazepine (1) has
demonstrated clinical efficacy in the treatment of neuropathic pain
which has validated the concept of this therapeutic approach
(Fig. 1).6,7
To date, nine voltage-gated sodium channel subtypes (Nav1.1–
Nav1.9) have been identified in mammals. These nine channels
are unique in their voltage-dependent properties as well as the
location of their expression in primary tissue.5,6 Nav1.7 has been
of particular interest to researchers based on human genetic stud-
ies which revealed a definitive link between SCN9A, which encodes
Nav1.7, and neuropathic pain. Individuals possessing
a loss-
of-function mutation in this gene display a complete indifference
to pain while maintaining otherwise standard nerve function.8,9
The enormous therapeutic potential of Nav1.x blockers for the
treatment of pain has led to a rapidly expanding body of literature
describing a variety of proof-of-concept molecules.7,10,11 Previous
research in our lab has resulted in several series of sodium channel
blockers and culminated with the discovery of 2-[4-(4-chloro-2-
fluorophenoxy)phenyl]-pyrimidine-4-carboxamide (PPPA, 2),
a
broad-spectrum state-dependent blocker which was efficacious
in multiple models of chronic pain.12–14 As part of a broad scaffold
hopping strategy toward the discovery of novel chemotypes, we
envisioned placing a tether between the A and B-rings of PPPA
giving rise to ring constrained compounds such as 4 (Fig. 1).15
The resultant dibenzazepine would bear a strong resemblance to
carbamazepine. This report describes the synthesis and prelimin-
ary SAR of these new sodium channel blockers.
The synthesis of all dibenzazepine analogs commenced with
commercially available 5H-dibenz[b,f]azepine (5) and is outlined
in Scheme 1. Regioselective mono-chlorination followed by
mono-bromination according to the method of Stachulski provided
key intermediate 7a.16 The bromide was converted to the boronic
ester 8a under palladium catalysis without effecting the chloride
⇑
Corresponding author. Tel.: +1 609 409 5837; fax: +1 609 409 6936.
E-mail address: laykea.tafesse@pharma.com (L. Tafesse).
http://dx.doi.org/10.1016/j.bmcl.2014.11.025
0960-894X/Ó 2014 Elsevier Ltd. All rights reserved.

44
S. M. Lynch et al. / Bioorg. Med. Chem. Lett. 25 (2015) 43–47
resulted in diaryl ether 18. Stirring under a hydrogen atmosphere
at ambient pressure and Raney nickel catalysis induced nitro group
reduction followed by intramolecular imine formation and subse-
quent imine reduction to afford 19a in good yield. Alkylation to
form 19b and 19c proceeded smoothly. And conversion to the diols
21a–c followed the usual procedure. Amide 21d was synthesized
in mild fashion from ester 21c using ammonia in MeOH at ambient
temperature. The related, regioisomeric dibenzoxazepine 26 was
synthesized using a similar strategy which is outlined in Scheme 4.
Horner–Wadsworth–Emmons olefination of 6-bromopyridine-
2-carboxaldehyde (27) followed by Sharpless asymmetric
dihydroxylation of the resulting cinnamate gave rise to diol 28
(Scheme 5). After conversion to the amide, Suzuki coupling of 29
with boronic ester 20b provided compound 30.
Once synthesized, the compounds were tested for their ability
to block Nav1.7 in a FLIPR based membrane potential assay.19 In
order to confirm activity in a more physiologically relevant
environment, compounds showing reasonable potency were also
evaluated using electrophysiology (EP) experiments.20 Data from
these studies are reported in Table 1.
NH2
N
NH2
N
R
O
O
N
carbamazepine (1)
4
cyclize
N
C
NH2
O to N
swap
NH2
N
N
B
O
R
O
O
N
F
A
Cl
2
3
PPPA ( )
Figure 1. Structure of carbamazepine (1) and comparison to a scaffold derived from
A–B ring cyclization of PPPA (2).
Compound 9a was moderately potent in our FLIPR assay and
appeared very promising during EP testing. The corresponding
N-methyl analog (9b) showed increased FLIPR potency, but its EP
Ki could not be measured due to poor solubility under the assay
conditions. Although the kinetic solubility of N-acetylated analog
9d was improved, a significant reduction in potency was also
observed. Interestingly, the N-ethyl analog 9c was not active at
moiety. Suzuki coupling with 6-bromopyridine-2-carboxamide
provided 9a. The azepine nitrogen of 9a proved quite resistant to
functionalization without affecting the adjacent amide nitrogen.
On the other hand, 7a was readily alkylated or acetylated to pro-
vide 7b–d, respectively. These bromides were converted to amides
9b–d using a sequence analogous to that used for 7a.
To synthesize dibenzazepine 13, 5 was first brominated, then
converted to the corresponding iodide using the procedure of
Buchwald (Scheme 2).17 N-Methylation provided compound 10.
The iodide was trifluoromethylated using methyl chlorodifluoroac-
etate and potassium fluoride in the presence of stoichiometric cop-
per to afford 11 in good yield.18 After regioselective bromination,
palladium catalyzed boronic ester synthesis gave 12. Diol 15 was
prepared from 6-bromopyridine-2-carboxaldehyde (14) via a Wit-
tig reaction followed by Sharpless asymmetric dihydroxylation
with AD-mix-alpha. Boronic ester 12 and diol 15 were joined via
Suzuki coupling to provide 13.
concentrations up to 10 lM. We suspected the pyridyl amide motif
present in 9 as a contributing factor in the limited solubility. Previ-
ous experiments in our lab have demonstrated that replacement of
the C-ring aryl amide with a linear diol can provide analogs which
display superior solubility while maintaining similar potency. Sim-
ilarly, we had some knowledge that trifluoromethyl A-ring substi-
tution could be advantageous, given the proposed lipophilic
binding pocket.8 Therefore we prepared and evaluated compound
13 which possessed these features. This compound had similar
FLIPR potency as 9b, but was much more soluble, especially at
acidic pH, which allowed for determination of its EP activity.
Since we were also interested in incorporating substitutions
within the 2 carbon A-B ring tether, we decided to explore modifi-
cations to the dibenzazepine scaffold. Substitution on the tether
would be more easily accomplished if a heteroatom were present
The synthesis of dibenzoxazepines 21a–d is illustrated in
Scheme 3. SNAr reaction of salicylaldehyde 17 and nitrophenol 16
Br
a
b
Cl
Cl
N
R
N
H
N
H
5
6
R = H; 7a
c
d
e
7b
R = Me;
7c
R = Et;
R = Ac; 7d
O
B
NH2
O
f
N
g
O
Cl
Cl
N
N
R
R
8a
9a
R = H;
R = H;
8b
9b
R = Me;
R = Et; 8c
8d
R = Me;
R = Et; 9c
9d
R = Ac;
R = Ac;
Scheme 1. Synthesis of dibenzazepines 9a–d. Reagents and conditions: (a) N-chlorosuccinimide, SiO2, CHCl3, rt, 15 h (34%); (b) N-bromosuccinimide, SiO2, CHCl3, rt, 0.5 h
(75%); (c) NaH, MeI, DMF, rt, 0.5 h (99%); (d) NaH, EtI, DMF, rt, 0.5 h (99%); (e) acetyl chloride, DMAP, toluene, 100 °C, 24 h (90%); (f) bis(pinacolato)diboron, Pd(dppf)Cl2, KOAc,
1,4-dioxane, 90 °C, 15 h (64–79%); (g) 6-bromopyridine-2-carboxamide, Pd(PPh3)2Cl2, aq. Na2CO3, DME/EtOH, 85 °C, 1 h (67–78%).

S. M. Lynch et al. / Bioorg. Med. Chem. Lett. 25 (2015) 43–47
45
a-c
d
I
F3C
N
N
H
N
11
5
10
O
B
OH
OH
O
N
e,f
g
F3C
F3C
N
N
12
13
h, i
H
OH
OH
Br
N
Br
N
O
14
15
Scheme 2. Synthesis of dibenzazepine 13. Reagents and conditions: (a) N-bromosuccinimide, SiO2, CHCl3, rt, 0.5 h (34%); (b) trans-N,N0-dimethylcyclohexane-1,2-diamine,
CuI, NaI, 1,4-dioxane, 100 °C, 15 h; (c) NaH, MeI, DMF, rt, 1 h (93%, 2 steps); (d) ClF2CO2Me, KF, CuI, DMF, 120 °C, 8 h (80% yield, 80% purity); (e) N-bromosuccinimide, SiO2,
CHCl3, rt, 0.5 h (87%); (f) bis(pinacolato)diboron, Pd(dppf)Cl2, KOAc, 1,4-dioxane, 90 °C, 15 h (71%); (g) 15, Pd(PPh3)2Cl2, aq. Na2CO3, DME/EtOH, 85 °C, 1.5 h (33%); (h)
CH3(Ph3)P+Brꢀ, n-BuLi, THF, 0 °C to rt (81%); (i) AD-mix-alpha, t-BuOH/H2O (1:1), rt, 15 h (83%).
F3C
NO2
O
R
O
Br
N
F3C
NO2
F
Br
O
a
b
H
F3C
H
O
HO
18
17
R = H; 19a
19b
R = CH2CO2Me; 19c
16
c
Br
R = Me;
d
O
B
R
R
OH
N
N
N
e
O
f
F3C
F3C
OH
O
O
R = H; 21a
21b
R = H; 20a
20b
R = Me;
R = CH2CO2Me; 21c
21d
R = Me;
R = CH2CO2Me; 20c
g
R = CH2CONH2;
Scheme 3. Synthesis of dibenzoxazepines 21a–d. Reagents and conditions: (a) K2CO3, DMF, 80 °C, 2 h (54%); (b) Ra-Ni, H2 (1 atm), EtOH, rt, 24 h (74%); (c) NaH, MeI, DMF, rt,
1 h (87%); (d) NaH, methyl bromoacetate, DMF, rt, 1 h; (e) bis(pinacolato)diboron, Pd(dppf)Cl2, KOAc, 1,4-dioxane, 90 °C, 15 h (68%); (f) 15, Pd(PPh3)2Cl2, aq. Na2CO3, DME/
EtOH, 85 °C, 1.5 h (33%); (g) 7.0 N NH3 in MeOH, rt, 15 h (36%).
O
F3C
H
N
O
O
H
Br
F3C
a,b
c
H
O
F3C
O
NO2
22
23
24
Br
O
B
N
N
O
OH
OH
d,e
O
f
N
F3C
F3C
O
25
26
Scheme 4. Synthesis of dibenzoxazepine 26. Reagents and conditions: (a) BBr3, CH2Cl2, -78 °C to rt (90%); (b) 4-bromo-1-fluoro-2-nitrobenzene, Cs2CO3, DMF, 80 °C, 5 h
(47%); (c) Ra-Ni, H2 (1 atm), EtOH, rt, 48 h (41%); (d) NaH, MeI, DMF, rt, 1 h (91%); (e) bis(pinacolato)diboron, Pd(dppf)Cl2, KOAc, 1,4-dioxane, 90 °C, 15 h (63%); (f) 15,
Pd(PPh3)2Cl2, aq. Na2CO3, DME/EtOH, 85 °C, 1.5 h (34%).

46
S. M. Lynch et al. / Bioorg. Med. Chem. Lett. 25 (2015) 43–47
a,b
c
OH
OEt
OH
NH2
Br
N
Br
N
H
Br
N
HO
28
HO
29
O
O
O
27
OH
NH2
N
d
N
F3C
HO
O
O
30
Scheme 5. Synthesis of dibenzoxazepine 30. Reagents and conditions: (a) triethyl phosphonoacetate, NaH, THF, 0 °C to rt, 15 h (56%); (b) AD-mix-alpha, t-BuOH/H2O (1:1), rt,
15 h (65%); (c) 7.0 N NH3 in MeOH, rt, 15 h (99%) (d) 20b, Pd(PPh3)2Cl2, K2CO3, 1,4-dioxane/H2O (5:1), 85 °C, 1.5 h (21%).
Table 1
FLIPR and electrophysiological potencies, solubility measurements and calculated AlogP98 values for dibenzazepine and dibenzoxazepine analogsa
Compound
FLIPR
Nav1.7 IC50
Electrophysiology
Nav1.7 Ki ( M)
Solubilityb
pH 1.2
Solubilityb
pH 6.8
Calcd
AlogP98
(lM)
l
1
9a
9b
9c
9d
13
21a
21b
21d
26
30
nd
32.7 11.5
0.65 0.12
nd
75
0.2
0
74
0
0
0
7
0.1
12
0.4
1
6
0.8
2.68
4.87
5.08
5.43
4.23
5.13
3.90
4.25
3.11
4.25
3.24
1.96 0.04
0.84 0.04
>10
7.62 0.80
1.08 0.11
2.57 0.06
0.37 0.05
1.80 0.79
1.03 0.07
1.05 0.07
nd
0.2
12
60
80
70
80
73
70
3.03 1.26
1.03 0.23
nd
0.44 0.09
nd
0.56 0.09
0.92 0.32
nd = not determined.
a
Mean SEM (standard error of the mean), n P 3.
b
lM.
at the point of functionalization. In addition, our limited SAR led us
to believe that nitrogen substitution larger than methyl might not
be tolerated at the azepine nitrogen (cf. 9b vs 9c, 9d). Taken
together, these two points prompted us to synthesize the dib-
enzoxazepines 21a and 21b which possesses a nitrogen within
the tether and an oxygen as the diaryl linker (Scheme 3). In partic-
ular, the N-methyl analog 21b proved to be a potent blocker of
Nav1.7. Since it was unclear which position of the tether was most
favorable for substitution, we also prepared and evaluated the reg-
ioisomeric dibenzoxazepine 26 (Scheme 4). While slightly less
potent in FLIPR, this compound appeared to be comparable to
21b in our EP assay and presented an improved solubility profile.
The lipophilicity of the dibenzazepines and dibenzoxazepine
analogs was estimated by calculating the atomic based partition
coefficient (AlogP98) for each compound. The same value of 4.25
was generated for compounds 21b and 26. While this number pre-
dicts modest lipophilicity, we were interested in studying the
effect of incorporating additional polarity within this scaffold. To
this end, we prepared compounds 21d and 30 which each con-
tained a supplemental amide at differential locations. In each case,
the additional amide imparted a drop of ꢁ1log unit in the pre-
dicted AlogP98 value. In particular, compound 30 was quite attrac-
tive since its potency was within twofold of 21b despite the added
polarity. Although 21d was fivefold less active than its methyl con-
gener 21b, we found it promising that the amide was tolerated in
the central region of the core where it disrupts the inherent amphi-
philic nature of analogs such as 30.
may be attributed to reactive metabolites derived from arene oxi-
dation and 10,11-epoxide formation.16 Given the structural simi-
larities between our scaffold and carbamazepine, we were
curious to assess how their in vitro metabolic stabilities would
compare. As summarized in Table 2, all new analogs possessed
greater metabolic stabilities than carbamazepine upon incubation
in both human and rat liver microsomes.21 The dibenzoxazepine
derivatives were slightly more stable than the dibenzazepine 13.
The combination of aryl substitution and removal of the C10–C11
double bond almost certainly play a role in the observed improve-
ment in stability.
In conclusion, using a scaffold hopping strategy, we have iden-
tified two new related series of sodium channel blockers. Com-
pounds in these classes showed good potency for Nav1.7 in
FLIPR-based and electrophysiological assays and good drug-like
properties. Future studies will focus on continued SAR exploration
and assessment of series selectivity versus other sodium channel
subtypes.
Table 2
In vitro metabolic stability measurements in human and rat liver microsomes for
selected compounds
Compound
Human liver microsomesa
Rat liver microsomesa
1
38
79
98
83
91
94
37
72
94
91
74
93
13
21b
21d
26
30
Carbamazepine is known to undergo extensive oxidative
metabolism in vivo. Its clinical administration has been associated
with dose limiting adverse events such as hepatotoxicity which
a
% compound remaining.

S. M. Lynch et al. / Bioorg. Med. Chem. Lett. 25 (2015) 43–47
47
14. Yang, J.; Gharagozloo, P.; Yao, J.; Ilyin, V. I.; Carter, R. B.; Nguyen, P.; Robledo, S.;
Woodward, R. M.; Hogenkamp, D. J. J. Med. Chem. 2004, 47, 1547.
References and notes
15. Sun, H.; Tawa, G.; Wallqvist, A. Drug Discovery Today 2012, 17, 310.
16. Elliott, E.-C.; Regan, S. L.; Maggs, J. L.; Bowkett, E. R.; Parry, L. J.; Williams, D. P.;
Park, B. K.; Stachulski, A. V. J. Med. Chem. 2012, 55, 9773.
17. Klapars, A.; Buchwald, S. L. J. Am. Chem. Soc. 2002, 124, 14844.
18. Mulder, J. A.; Frutos, R. P.; Patel, N. D.; Qu, B.; Sun, X.; Tampone, T. G.; Gao, J.;
Sarvestani, M.; Eriksson, M. C.; Haddad, N.; Shen, S.; Song, J. J.; Senanayake, C.
H. Org. Process Res. Dev. 2013, 17, 940.
19. CHO-Nav1.7 cells were seeded cells at 25 K/well in tissue culture treated 96-
well plates (Costar 3904) with complete media and incubated in a 37 °C/5%CO2
humidified incubator overnight (ca. 16 h.) The Membrane Potential Dye (MP
Kit R8042) was reconstituted with 30–35 ml assay buffer (HBSS Buffer
supplemented with 20 mM Hepes; pH 7.40 with NaOH and 2.0 mM
1. Global Industry Analysts, Inc., Report, January 10, 2011.
2. Cohen, S. P.; Mao, J. BMJ 2014, 348, f7656.
3. Smith, H. S.; Kara, S. D.; Argoff, C. E. Eur. Neurol. Rev. 2012, 7, 72.
4. Dworkin, R. H.; O’Conner, A. B.; Backonja, M.; Farrar, J. T.; Finnerup, N. B.;
Jensen, T. S.; Kalso, E. A.; Loeser, J. D.; Miaskowski, C.; Nurmikko, T. J.; Portenoy,
R. K.; Rice, A. S. C.; Stacey, B. R.; Treede, R.-D.; Turk, D. C.; Wallace, M. S. Pain
2007, 132, 237.
5. Dib-Hajj, S. D.; Cummins, T. R.; Black, J. A.; Waxman, S. G. Annu. Rev. Neurosci.
2010, 33, 325.
6. Eijkelkamp, N.; Linley, J. E.; Baker, M. D.; Minett, M. S.; Cregg, R.; Werdehausen,
R.; Rugiero, F.; Wood, J. N. Brain 2012, 135, 2585.
7. Kemp, M. I. Prog. Med. Chem. 2010, 49, 81.
8. Dib-Hajj, S. D.; Yang, Y.; Black, J. A.; Waxman, S. G. Nat. Rev. Neurosci. 2013, 14,
49.
Probenecid) (Invitrogen P36400). The cell media was removed and 100 ll
probenecid-free assay buffer with 90 mM KCl was added and incubated at RT
for 45 min. Buffer was removed and the plate was washed with probenecid-
9. Cox, J. J.; Reimann, F.; Nicholas, A. K.; Thornton, G.; Roberts, E.; Springell, K.;
Karbani, G.; Jafri, H.; Mannan, J.; Raashid, Y.; Al-Gazali, L.; Hamamy, H.;
Valente, E. M.; Gorman, S.; Williams, R.; McHale, D. P.; Wood, J. N.; Gribble, F.
M.; Woods, C. G. Nature 2006, 444, 894.
10. Theile, J. W.; Cummins, T. R. Front. Pharmacol. 2011, 2, 1.
11. Nantermet, P. G.; Henze, D. A. Annu. Rep. Med. Chem. 2011, 46, 19.
12. Ilyin, V. I.; Pomonis, J. D.; Whiteside, G. T.; Harrison, J. E.; Pearson, M. S.; Mark,
L.; Turchin, P. I.; Gottshall, S.; Carter, R. B.; Nguyen, P.; Hogenkamp, D. J.;
Olanrewaju, S.; Benjamin, E.; Woodward, R. M. J. Pharmacol. Exp. Ther. 2006,
318, 1083.
free assay buffer before adding 100
to the cell plate. This was incubated at rt for 45 min before acquiring data on
ll Membrane Potential (MP) dye 1ꢂ drug
the FLIPR. Test compounds were diluted 100 fold in MP dye (i.e., 2.5
2 mM + 250 l MP dye yields 20
consisted of a 7 s baseline read of cell plate, followed by the addition of 50
l
l
l
l
M assay concentration). The FLIPR assay
ll
3ꢂ agonist and a 1 min read during which time the antagonists were added.
20. Assay was run according to the protocol described. In: Shao, B. PCT Int. Appl.
WO 2013072758.
21. Rat and human liver microsomes were obtained from commercial sources.
Compound (1 lM) was incubated with 0.5 mg/mL of appropriate microsomes
at 37 °C for 30 min. Compound was quantified at t = 0 min and t = 30 min and
results reported as the percentage of compound remaining post incubation.
13. Shao, B.; Victory, S.; Ilyin, V. I.; Goehring, R. R.; Sun, Q.; Hogenkamp, D.;
Hodges, D. D.; Islam, K.; Sha, D.; Zhang, C.; Nguyen, P.; Robledo, S.;
Sakellaropoulos, G.; Carter, R. B. J. Med. Chem. 2004, 47, 4277.