3-Amino-1,5-benzodiazepinones: Potent, state-dependent sodium channel blockers with anti-epileptic activity

Article abstract of DOI:10.1016/j.bmcl.2008.01.123

A series of 3-amino-1,5-benzodiazepinones were synthesized and evaluated as potential sodium channel blockers in a functional, membrane potential-based assay. One member of this series displayed subnanomolar, state-dependent sodium channel block, and was orally efficacious in a mouse model of epilepsy.

Full text of DOI:10.1016/j.bmcl.2008.01.123

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1963–1966
3-Amino-1,5-benzodiazepinones: Potent, state-dependent
sodium channel blockers with anti-epileptic activity
Scott B. Hoyt,a,* Clare London,a Matthew J. Wyvratt,a Michael H. Fisher,a
Doreen E. Cashen,b John P. Felix,c Maria L. Garcia,c Xiaohua Li,a
Kathryn A. Lyons,a D. Euan MacIntyre,b William J. Martin,b Birgit T. Priest,c
McHardy M. Smith,c Vivien A. Warren,c Brande S. Williams,c
Gregory J. Kaczorowskic and William H. Parsonsa
aDepartment of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, RY 123-236, Rahway, NJ 07065-0900, USA
bDepartment of Pharmacology, Merck Research Laboratories, Rahway, NJ 07065, USA
cDepartment of Ion Channels, Merck Research Laboratories, Rahway, NJ 07065, USA
Received 20 December 2007; revised 29 January 2008; accepted 30 January 2008
Available online 7 February 2008
Abstract—A series of 3-amino-1,5-benzodiazepinones were synthesized and evaluated as potential sodium channel blockers in a
functional, membrane potential-based assay. One member of this series displayed subnanomolar, state-dependent sodium channel
block, and was orally efficacious in a mouse model of epilepsy.
Ó 2008 Elsevier Ltd. All rights reserved.
Approximately two dozen drugs are currently marketed
for the treatment of epilepsy, a common CNS disorder
that afflicts nearly 2% of the world’s population.1 Some
act by sodium or calcium channel blockade, while others
function at the GABAA receptor, or by mixed or un-
known mechanisms of action. Though a broad range
of treatment options is available, an estimated 30% of
patients do not respond to any current therapy.2 Addi-
tionally, in patients who do respond, existing drugs of-
ten elicit adverse effects such as sedation and neuronal
impairment.
Sodium channel blockers destined for the clinic must in-
hibit aberrant neuronal signaling while leaving normal
nerve functions intact. We believe that this can be
achieved, in large part, via state-dependent channel
block. Sodium channels are thought to exist in three
main conformational states: resting, open, and inacti-
vated. In healthy nerve and cardiac tissue, these chan-
nels exist predominantly in the resting state. During
seizure, on the other hand, the aberrant firing of high-
frequency action potential bursts causes sodium chan-
nels to accumulate in the inactivated state. Compounds
that selectively bind and stabilize that inactivated state
should inhibit aberrant signaling preferentially, thus
minimizing the potential for mechanism-based adverse
effects.
Epileptic seizures begin with the aberrant firing of action
potential bursts in the brain. The initiation and propa-
gation of these action potentials typically require the
opening of voltage-gated sodium channels (Nav1.x). Be-
cause they can inhibit action potential firing, sodium
channel blockers have been investigated as anti-epileptic
treatments. Weak blockers such as carbamazepine and
lamotrigine have demonstrated clinical anticonvulsant
activity, thereby providing validation for this approach
(Fig. 1).3,4
We recently reported the discovery of a structurally
novel series of benzazepinone sodium channel blockers.
Cl
O
H2N
Cl
O
O
OCF3
N
N
N
N
N
N
H
NH
H2N
NH2
Boc
Keywords: Sodium channel blocker; Nav1; Epilepsy.
*
Carbamazepine
Lamotrigine
1
Corresponding author. Tel.: +1 732 594 3753; fax: +1 732 594
5350; e-mail: scott_hoyt@merck.com
Figure 1. Sodium channel blockers.
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.01.123

1964
S. B. Hoyt et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1963–1966
Table 1. Effect of the R1 group on potency and off-target activity
F3C
H
N
O
O
N
O
a
O
F
b
N
NHBoc
NHBoc
N
H
N
N
H
N
R1
NH
H
Boc
2
3
Compound R1
hNav1.7
MK-0499
F3C
F3C
(IC50, nM) (% inhibition at 1, 10 lM)
O
O
O
N
N
c, d
6
409
142
0%, 25%
0%, 7%
NHBoc
H
Me
N
H
7
N
N
NH
Boc
4
5
8
47
49
81
36%, 62%
1%, 45%
20%, 65%
F3C
Scheme 1. Reagents and conditions: (a) KHMDS, CF3CH2OTf, THF,
0 °C (67%); (b) cyclopropylmethyl bromide, K2CO3, Bu4NI, THF,
100 °C (53%); (c) TFA, CH2Cl2; (d)N-Boc-D-phenylalanine, BOP,
i-Pr2NEt, CH2Cl2.
9
10
11
12
188
25
17%, 43%
14%, 84%
A benchmark in this class displayed potent, state-
dependent block of hNav1.7 in vitro (IC50 = 35 nM),
blocked spontaneous neuronal firing in vivo, and was
orally efficacious in a rat model of neuropathic pain
(compound 1, Fig. 1).5 Explorations in this general
class led us to examine SAR in a related series of 3-
amino-1,5-benzodiazepinones. These efforts culminated
in the discovery of a highly potent, state-dependent so-
dium channel blocker that shows good oral efficacy in
a rat model of epilepsy.
Table 2. Effect of the R2 group on potency and off-target activity
R2
N
O
O
F
N
H
N
NH
Boc
Target compounds were synthesized as shown in Scheme 1.
Benzodiazepinone 2, the enantiomer of a known com-
pound, was prepared in three steps from commercially
available materials via an established procedure.6 Treat-
ment of 2 with two equivalents of KHMDS and an alkyl
halide or triflate such as 2,2,2-trifluoroethane trifluoro-
sulfonate resulted in alkylation to yield 3. Heating of 3
in the presence of K2CO3 and an electrophile such as
cyclopropylmethyl bromide then effected alkylation of
the anilinic nitrogen to afford 4. Finally, exposure of 4
to standard conditions for N-Boc removal (TFA,
CH2Cl2) provided a crude TFA salt that could be cou-
pled with amino acids such as N-Boc-D-pheynylalanine
to afford target compound 5. Compounds listed in Ta-
bles 1–3 were synthesized according to these procedures
using the appropriate commercially available starting
materials. In several instances (19 and 27), the final
amide coupling step was performed using a non-com-
mercially available amino acid that had been synthesized
via the method of Schollkopf.7
Compound
R2
hNav1.7
(IC50, nM)
MK-0499
(% inhibition at
1, 10 lM)
9
13
14
Me
CH2CF3
i-Pr
49
101
102
1%, 45%
7%, 20%
7%, 33%
Table 3. Combined effects of R1 and R2 on potency and off-target
activity
R2
N
O
O
N
H
R3
N
NH
Boc
Compound R2
R3
hNav1.7 MK-0499
(IC50, nM) (% inhibition
at 10 lM)
5
13
15
16
17
18
19
20
21
22
23
24
25
26
27
CH2CF3 Ph
30
101
142
219
185
131
73
24%
19%
17%
18%
12%
11%
13%
6%
CH2CF3 2-F–Ph
CH2CF3 3-F–Ph
CH2CF3 4-F–Ph
Once synthesized, compounds were then tested for their
ability to block hNav1.7 sodium channels that had been
stably expressed in a HEK-293 cell line. The extent of
channel block was determined in a functional, mem-
brane potential-based assay that measures the fluores-
cence resonance energy transfer (FRET) between two
membrane-associated dyes. Specific details of the exper-
imental protocols employed have recently been de-
scribed.8 Compounds were also counterscreened at
several other ion channels targets. Because block of
hERG K+ channels has been associated with potentially
lethal ventricular arrhythmias, compounds were
screened in a binding assay that measures the displace-
CH2CF3 2,6-di-F–Ph
CH2CF3 2-CF3–Ph
CH2CF3 2-OCF3–Ph
CH2CF3 2,5-di-CF3–Ph 120
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
i-Pr
Ph
2-F–Ph
52
102
96
42%
32%
17%
29%
25%
21%
35%
3-F–Ph
4-F–Ph
196
97
2,6-di-F–Ph
2-CF3–Ph
2-OCF3–Ph
149
75

S. B. Hoyt et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1963–1966
1965
ment of 35S-labeled MK-0499, a known hERG K+ chan-
0.0
-0.5
-1.0
-1.5
A
nel blocker.9
3 nM Compound 5
As shown in Table 1, a variety of N-methyl benzodiaz-
epinones bearing a 2-fluorophenylalanine sidechain were
prepared. In this series, the impact of the R1 group on
potency was examined. Analogs with small R1 substitu-
ents such as 6 and 7 (R1 = H and Me) displayed only
moderate hNav1.7 block. As the size of R1 increased,
potency improved, reaching an optimum level in the tri-
fluoroethyl and methylenecyclopropyl analogs 8 and 9.
Compounds such as 10 and 11 that incorporated larger
R1 groups were correspondingly less potent. In a notable
exception to this trend, N-benzyl derivative 12 exhibited
best-in-series hNav1.7 block. Unfortunately, 12 was also
one of the most active compounds in the MK-0499
counterscreen. On balance, the methylenecyclopropyl
group seemed to provide the best combination of po-
tency and low MK-0499 activity, and was thus featured
in further studies.
0
200
400
600
800
Time (s)
100
80
60
40
20
0
B
With an optimized R1 group in hand, we next conducted
a limited scan at the lactam R2 position. To keep final
target molecular weights as low as possible, we focused
on smaller alkyl groups. As shown in Table 2, the N-
methyl, N-trifluoroethyl and N-isopropyl analogs 9,
13, and 14 exhibited comparable potencies in both the
hNav1.7 functional assay and the MK-0499 binding as-
say. Because they can hinder metabolic N-dealkylation,
the trifluoroethyl and isopropyl groups were thought to
offer potential pharmacokinetic advantages, and were
thus incorporated in subsequent designs.
1
10
100
[Compound 5] (nM)
Figure 2. Whole cell electrophysiology data for compound 5.
block on membrane potential is consistent with prefer-
ential block of channels in the inactivated state. At
À120 mV, essentially all hNav1.7 channels reside in the
resting state, suggesting that compound 5 affords little
binding to resting channels (Kr > 5 lM). At À70 mV,
an average of 41% of hNav1.7 channels reside in the
inactivated state (n = 107). Based on the fraction of
inactivated channels and the IC50 at À70 mV, the affin-
ity of compound 5 for inactivated channels (Ki) was cal-
culated to be 0.63 nM. By this measure, compound 5 is
one of the most potent small-molecule sodium channel
blockers described in the literature to date.
A final series of derivatives was prepared wherein the R3
group was varied (Table 3). Prior work had shown that a
lipophilic aromatic substituent was required at R3 for po-
tent hNav1.7 block. In the N-trifluoroethyl series, a num-
ber of mono- and difluorophenyl derivatives (compounds
13 and 15–17) were examined; all were found to be less
potent than simple phenyl analog 5. Because substitution
seemed best tolerated at the 2-position, the 2-CF3–Ph and
2-OCF3–Ph variants 18 and 19 were prepared; these too
were less potent than 5. Similar trends were observed in
the N-isopropyl series (compounds 21–27). In this set,
substituted derivatives 22–27 were typically 2- to 4-fold
less potent than the simple phenyl analog 21.
Compound 5 and several of its analogs were submitted
for rat pharmacokinetic (PK) determination.11 As
shown in Table 4, 5 exhibited a modest PK profile,
Table 4. Rat pharmacokinetic data for selected compounds
R2
Maximally potent blocker 5 was selected for further pro-
filing. For reasons outlined above, it was important to
ascertain whether 5 blocked hNav1.7 channels in a
state-dependent manner. Block of hNav1.7 channels by
compound 5 was therefore examined by whole cell elec-
trophysiology in stably transfected HEK-293 cells.10
Figure 2A shows the peak hNav1.7 current evoked by
20 ms depolarizations from a membrane potential of
À70 mV. The solid bar indicates bath application of
3 nM compound 5. The dose–response for inhibition
of hNav1.7 current by compound 5, applied at
À70 mV, is shown in Figure 2B (n = 2 for each concen-
tration). Fitting the data to the Hill equation yielded an
IC50 of 1.54 nM. In contrast, 1 lM compound 5 blocked
less than 10% of hNav1.7 currents when applied at a
membrane potential of À120 mV. The dependence of
O
O
N
N
H
R3
N
NH
Boc
a
b
F% AUCN Cmax ClP t1/2
c
d
Com- R2
pound
R3
5
18
26
27
CH2CF3 Ph
CH2CF3 2-CF3–Ph
5% 0.03
10% 0.05
9% 0.11
0.03 48
0.06 51
0.15 23
0.04 30
3.1
2.4
2.5
2.4
i-Pr
i-Pr
2-CF3–Ph
2-OCF3–Ph 4% 0.04
a (po, lM h/mpk).
b (lM).
c (mL/min/kg).
d (hours).

1966
S. B. Hoyt et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1963–1966
Parsons, W. H. Bioorg. Med. Chem. Lett. 2007, 17, 4630;
Williams, B. S.; Felix, J. P.; Priest, B. T.; Brochu, R. M.;
Dai, K.; Hoyt, S. B.; London, C.; Tang, Y. S.; Duffy, J. L.;
Parsons, W. H.; Kaczorowski, G. J.; Garcia, M. L.
Biochemistry 2008, 46, 14693.
and suffered both low bioavailability and a high clear-
ance rate (F = 5%, Clp = 48 mL/min/kg). Of the other
analogs studied, 26 displayed the best profile, offering
a moderate improvement over 5 (F = 9%, Clp = 23 mL/
min/kg, Cmax = 150 nM).
6. Lauffer, D. J.; Mullican, M. D. Bioorg. Med. Chem. Lett.
2002, 12, 1225.
7. Schollkopf, U. Tetrahedron 1983, 39, 2085.
8. Felix, J. P.; Williams, B. S.; Priest, B. T.; Brochu, R. M.;
Dick, I. E.; Warren, V. A.; Yan, L.; Slaughter, R. S.;
Kaczorowski, G. J.; Smith, M. M.; Garcia, M. L. Assay
Drug Dev. Tech. 2004, 2, 260.
Compound 5 proved highly efficacious in the mouse
maximum electroshock (MES) assay, a widely used pro-
tocol for assessing anticonvulsant activity.12 When
dosed orally at 3 mg/kg, 5 prevented shock induced to-
nic–clonic seizures in 90% of subjects (n = 10) at
30 min post-dosing. These results are broadly compara-
ble to those obtained with clinical standards such as car-
bamazepine (MES ED50 = 3.4 mg/kg) and lamotrigine
(MES ED50 = 2.2 mg/kg).13 Though brain levels of 5
were not determined, these initial results are promising,
and provide a basis for further investigation.
9. Wang, J.; Della Penna, K.; Wang, H.; Karczewski, J.;
Connolly, T. M.; Koblan, K. S.; Bennett, P. B.; Salata, J.
J. Am. J. Physiol. Heart Circ. Physiol. 2002, 284, H256.
10. Hamill, O. P.; Marty, A.; Neher, E.; Sakmann, B.;
Sigworth, F. J. Pflugers Arch. 1981, 391, 85. Procedure.
Sodium currents were examined by whole cell voltage
clamp using an EPC-9 amplifier and Pulse software
(HEKA Electronics, Lamprecht, Germany). Experiments
were performed at room temperature. Electrodes were fire-
polished to resistances of 1.5–4 MX. Voltage errors were
minimized by series resistance compensation (75–85%),
and the capacitance artifact was canceled using the
amplifier’s built-in circuitry. Data were acquired at
50 kHz and filtered at 10 kHz. The bath solution consisted
of 40 mM NaCl, 120 mM NMDG–Cl, 1 mM KCl,
2.7 mM CaCl2, 0.5 mM MgCl2, 10 mM NMDG–Hepes,
pH 7.4, and the internal (pipet) solution contained
110 mM Cs–methanesulfonate, 5 mM NaCl, 20 mM CsCl,
10 mM CsF, 10 mM BAPTA (tetra Cs salt), 10 mM Cs–
Hepes, pH 7.4. Liquid junction potentials were less than
4 mV and were not corrected for. Because whole cell
voltage clamp experiments are comparatively labor-inten-
sive, compound 5 is the only analog from this series to be
profiled using this technique.
11. Rat PK experiments were conducted as follows: test
compounds were typically formulated as 1.5 mg/mL solu-
tions in mixtures of PEG300/water or DMSO/PEG300/
water. Fasted male Sprague–Dawley rats were given either
a 1.0 mg/kg iv dose of test compound solution via a
cannula implanted in the femoral vein (n = 3) or a 3.0 mg/
kg po dose by gavage (n = 3). Serial blood samples were
collected at 5 (iv only), 15, and 30 min, and at 1,2,4,6, and
8 h post-dose. Plasma was collected by centrifugation, and
plasma concentrations of test compound were determined
by LC–MS/MS following protein precipitation with
acetonitrile.
In summary, we have identified a series of 3-amino-1,5-
benzodiazepinones that are potent blockers of voltage-
gated sodium channels. A benchmark compound from
this class exhibited state-dependent, subnanomolar
block of hNav1.7, and was orally efficacious in a mouse
model of epilepsy. Future work will focus on improving
pharmacokinetics in this series, and will be reported in
due course.
Acknowledgments
We thank Ramona Gray and Joe Leone for their out-
standing technical contributions to this work. We also
thank David Kaelin and Deborah Pan for their assis-
tance in proofreading this manuscript.
References and notes
1. For a recent review of the neurobiology of antiepileptic
drugs, see: Rogawski, M. A.; Loscher, W. Nat. Neurosci.
2004, 5, 553.
2. Perucca, E. J. Clin. Pharmacol. 1996, 42, 531.
3. Willow, M.; Catterall, W. A. Mol. Pharmacol. 1982, 22,
627; Goa, K. L.; Ross, S. R.; Chrisp, P. Drugs 1993, 46,
152; Brodie, M. J. Epilepsia 1994, 35, S41; Xie, X.;
Lancaster, B.; Peakman, T.; Garthwaite, J. Pflugers Arch.
1995, 430, 437.
12. The threshold for maximal electroshock seizures was
determined in male mice (C57BL6J mice from The
Jackson Laboratory; 14 weeks of age) using auricular
electrodes connected to an electroconvulsive device (Basile
57800) designed for inducing tonic–clonic activity in mice
and rats. For anticonvulsant testing, the following param-
eters were used: frequency = 100 Hz; pulse width = 0.7 ms;
shock duration = 0.5 s; current = 18 mA. Compound 5
was prepared in a vehicle consisting of 5% EtOH + 10%
Tween 80 + 85% water, sonicated to aid in solubilization,
then administered by gavage in a volume of 2 mL/kg.
Compound 5 was administered at 3 mg/kg po, and was
evaluated for anticonvulsant properties 30 min post-
dosage.
4. For a review of the medicinal chemistry of sodium channel
blockers, see: Anger, T.; Madge, D. J.; Mulla, M.; Riddall,
D. J. Med. Chem. 2001, 44, 115; For a review of the
biology of sodium channels, see: Ashcroft, F. M. Ion
Channels and Disease; Academic Press: San Diego, 2000.
5. Hoyt, S. B.; London, C.; Gorin, D.; Wyvratt, M. J.; Fisher,
M. H.; Abbadie, C.; Felix, J. P.; Garcia, M. L.; Li, X.;
Lyons, K. A.; McGowan, E.; MacIntyre, D. E.; Martin, W.
J.; Priest, B. T.; Ritter, A.; Smith, M. M.; Warren, V. A.;
Williams, B. S.; Kaczorowski, G. J.; Parsons, W. H. Bioorg.
Med. Chem. Lett. 2007, 17, 4630; Hoyt, S. B.; London, C.;
Ok, H.; Gonzalez, E.; Duffy, J. L.; Abbadie, C.; Dean, B.;
Felix, J. P.; Garcia, M. L.; Jochnowitz, N.; Karanam, B. V.;
Li, X.; Lyons, K. A.; McGowan, E.; MacIntryre, D. E.;
Martin, W. J.; Priest, B. T.; Smith, M. M.; Tschirret-Guth,
R.; Warren, V. A.; Williams, B. A.; Kaczorowski, G. J.;
13. Faigle, J. W.; Feldmann, K. F. In Antiepileptic Drugs;
Levy, R. H., Mattson, R. H., Meldrum, B. S., Eds.; Raven
Press: New York, 1995; pp 499–513; Miller, A. A.; Nobbs,
M. S.; Hyde, R. M.; Leach, M. J. Patent EP713703, 1996.