Discovery of isoxazole voltage gated sodium channel blockers for treatment of chronic pain

Article abstract of DOI:10.1016/j.bmcl.2009.07.135

A series of novel isoxazole voltage gated sodium channel blockers have been synthesized and evaluated. Substitutions on the benzylic position of benzamide were investigated to determine their effect on Na_v1.7 inhibitory potency. The spirocyclobutyl substitution had the most significant enhancement on Na_v1.7 inhibitory activity.

Full text of DOI:10.1016/j.bmcl.2009.07.135

Bioorganic & Medicinal Chemistry Letters 19 (2009) 5334–5338
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of isoxazole voltage gated sodium channel blockers for treatment
of chronic pain
a
a
a
a
a
Pengcheng P. Shao a, , Feng Ye , Ann E. Weber , Xiaohua Li , Kathryn A. Lyons , William H. Parsons ,
Maria L. Garcia b, Birgit T. Priest b, McHardy M. Smith b, John P. Felix b, Brande S. Williams b,
Gregory J. Kaczorowski b, Erin McGowan c, Catherine Abbadie c, William J. Martin c, Daniel R. McMasters d,
Ying-Duo Gao d
*
a Department of Medicinal Chemistry, Merck Research Laboratories, Rahway, NJ 07065, USA
b Department of Ion Channels, Merck Research Laboratories, Rahway, NJ 07065, USA
c Department of Pharmacology, Merck Research Laboratories, Rahway, NJ 07065, USA
d Department of Molecular Modeling, Merck Research Laboratories, Rahway, NJ 07065, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of novel isoxazole voltage gated sodium channel blockers have been synthesized and evaluated.
Substitutions on the benzylic position of benzamide were investigated to determine their effect on Nav1.7
inhibitory potency. The spirocyclobutyl substitution had the most significant enhancement on Nav1.7
inhibitory activity.
Received 7 June 2009
Revised 27 July 2009
Accepted 29 July 2009
Available online 6 August 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Voltage gated sodium channel blocker
Neuropathic pain
Inflammatory pain
Amide replacement
Heterocycles
Isoxazole
In the previous Letter, we reported our approach to replace one
of the amide groups in CDA54 with heterocycles. Our initial results
showed that replacing the amide with aromatic heterocycles such
as oxazole and isoxazole resulted in significant loss of potency,
while oxazoline and isoxazoline (1) appear to be well tolerated.1
However, all the compounds examined at that point did not have
substitutions at the benzylic position of the remaining benzamide.
Since substitutions at benzylic position are expected to change the
free energy difference between ‘bent’ and ‘extended’ conforma-
tions (Fig. 1), this should have significant consequences on the pop-
ulation ratio between the two conformations and thus in the
inhibitory potency of the compound.
HN
N
F3CO
O
O
H2NO2S
CDA54
H
N
H
N
R1
R2
HN
R
F3CO
R2
R
O
R1
"extended"
R1, R2: H, Me, -(CH2)n-
O
N
O
O
Me
1
"bent"
The synthesis of isoxazole derivatives with substitutions on the
benzylic position of benzamide is illustrated in Scheme 1. Benzyl
amine derivatives such as intermediate III were conveniently syn-
thesized according to the procedure developed by Ellman’s group.2
Figure 1. Equilibrium between ‘extended’ and ‘bent’ conformations.
* Corresponding author. Tel.: +1 732 594 2955.
E-mail address: pengcheng_shao@merck.com (P.P. Shao).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.07.135

P. P. Shao et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5334–5338
5335
Other phenylcyclobutane amine derivatives with substitutions (F,
OH, and OMe) on the cyclobutane ring were synthesized according
to a previously published procedure.3 The isoxazole intermediate
such as VI was synthesized in a three-step sequence starting from
1-cyclopentenecarboxylic acid. TBAF-catalyzed Michael addition4
between nitromethane and 1-cyclopentenecarboxylic acid pro-
ceeded with good selectivity (8:1) and favored the desired trans
isomer. The minor, cis isomer was removed by flash chromatogra-
phy. Intermediate IV was treated with phenylisocyanate and TEA
to form the nitrile oxide, which underwent 1,3-dipolar cycloaddi-
tion with phenylacetylene in situ to form isoxazole V.5 Hydrolysis
of the methyl ester, followed by standard BOP-mediated amide
coupling afforded the final isoxazole compound 11.
nient method to construct these compounds (Scheme 2). 1,3-Dipo-
lar addition involving ethynyl boronic ester6 gave the boronic acid
derivative VIII in moderate yield after hydrolysis of the boronic es-
ter VII. Suzuki coupling reaction between iodide VI and boronic
acid VIII proceeded under mild conditions to provide the desired
coupling product IX in good yield. Hydrolysis of the methyl ester,
followed by standard amide coupling reaction yielded the desired
isoxazole compound 30.
Once synthesized, the compounds were evaluated for their abil-
ity to block the voltage gated sodium channels stably expressed in
a HEK-293 cell line. In the present Letter, only hNav1.7 activity was
evaluated. The extent of the channel block was determined in a
functional, membrane potential-based assay that measures the
fluorescence resonance energy transfer (FRET) between two mem-
brane-associated dyes.7 Compounds showing good potency in
In cases where acetylene derivatives were not readily available,
Suzuki coupling reaction utilizing available halides was a conve-
O2
a
O
S
b
c
F3C
S
F3C
N
H
NH2
O
N
I
II
III
d
e
g
f
HO2C
MeO2C
MeO2C
NO2
MeO2C
N
N
O
O
VI
IV
V
F3C
HN
O
N
O
11
Scheme 1. Reagents and conditions: (a) THF, t-butylsulfinamide (1 equiv), Ti(OEt)4 (2 equiv), 60 °C, 14 h (92%); (b) THF, 1-bromo-3-trifluoromethyl benzene (2 equiv), nBuLi
(1.6 M in hexane, 2 equiv), À78 °C (46%); (c) MeOH, HCl (concd, excess), 25 °C, 1 h (95%); (d) CH3NO2 (3 equiv), TBAF (1 M in THF, 0.1 equiv), 70 °C, 20 h (83%); (e) toluene,
phenylacetylene (5 equiv), phenylisocyanate (5 equiv), TEA (3 equiv), 120 °C, 3 h (72%); (f) MeOH, LiOH (10 equiv), 50 °C, 2 h (92%); (g) THF, intermediate III (1 equiv), Bop
(1.5 equiv), DIEA (4 equiv), 25 °C, 1 h (35%).
I
O
OH
O
a
OH
b, c
CF3
CF3
CF3
V
VI
d
e
f
HO
MeO2C
MeO2C
NO2
MeO2C
MeO2C
CF3
O
h
OH
N
N
N
B
O
B
O
O
O
IV
VII
VIII
OH
IX
g
HO
HO
HN
CF3
HO2C
CF3
O
N
N
O
O
X
30
Scheme 2. Reagents and conditions: (a) Toluene, THP (2 equiv), PPTS (catalytic), 60 °C, 16 h (95%); (b) THF, nBuLi (1.6 M in hexane, 1.05 equiv), I2 (1.1 equiv, as THF solution),
À78 °C to 0 °C; (c) acetone, 3 M HCl (excess) (72%—two steps); (d) toluene, intermediate IV (slow addition), 2-ethynyl-4,4,5,5-tetramethyl-1,3,2-dioxaborolane (1.2 equiv),
phenyl isocyanate (2 equiv), TEA (2 equiv), 115 °C, 12 h; (e) EtOAc, 3 M HCl (excess), 15 min (48%—two steps); (f) dioxane, intermediate VI (1.5 equiv), Pd(dppf)Cl2 (0.1 equiv),
K3PO4 (3 equiv), 40 ?C, 2.5 h (75%); (g) MeOH/H2O (5:1), LiOH (10 equiv), 25 °C, 5 h (82%); (h) THF, 1-phenylcyclobutanamine (1.2 equiv), Bop (1.2 equiv), DIEA (5 equiv),
25 °C, 1 h (65%).

5336
P. P. Shao et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5334–5338
Table 1
increased further, the potency started to decrease (7, 8). Interest-
ingly, benzylic substitution did not afford potency enhancement
Effect of benzylic substitutions on Nav1.7 potency
in the isoxazoline (9, 10) series. It is also noteworthy that com-
pounds 4–6 were free of MK-0499 binding activity (IC50 >10 lM).
This dramatic potency enhancement was rather surprising, par-
ticularly with cyclobutane (6) and cyclopentane (7) analogs. We
suspect that the potency of the cycloalkyl analogs is due in part
to a preference of these analogs for the bioactive ‘bent’ conforma-
tion. Molecular modeling provided strong support for this hypoth-
esis (Table 2). In the unsubstituted case (2), molecular mechanics
calculations9 show a nearly equal energy for the extended and bent
conformations. Substituting with either one methyl (3) or gem-di-
methyl groups (4) causes a small preference for the extended con-
formation (Fig. 2), while substituting with a spirocycloalkyl group
(5–8) causes a preference for the bent conformation (Figs. 3 and 4).
The preference for the ‘bent’ conformation reaches a maximum
with the spirocyclobutane compound (6), in coincidence with its
Nav1.7 blocking activity.
Me
HN
R1
HN
R2
or
N
Ph
N
O
Ph
R2
O
O
O
R1
Compound R1 R2
Structure
hNav1.7 MK-0499 (% inh MK-0499 (%
at 1
l
M or IC50
,
l
M)
inh at 10 lM)
2
3
4
5
H, H
H, CH3
CH3, CH3 Isoxazole
–
Isoxazole
Isoxazole
37%
1.7
0.43
0.44
35%
47%
45%
47%
Isoxazole
CH2CH2–
6
7
8
9
10
–(CH2)3– Isoxazole
–(CH2)4– Isoxazole
–(CH2)5– Isoxazole
0.12
0.22
0.58
37%
85%
ND
ND
ND
H, H
Isoxazoline 19%
–(CH2)3– Isoxazoline 8%
The torsional energy plots are shown in Figures 2–4.
Since spirocyclobutane substitution was optimal for enhancing
Nav1.7 potency, this structural feature was maintained in further
SAR studies. Although compound 6 is highly potent and selective
against the hERG channel, it has poor pharmacokinetic properties
with high clearance and low oral bioavailability (Table 5). In at-
tempts to address this issue, substitutions on the aromatic and
cyclobutane ring were explored. As shown in Table 3, a CF3 group
at the meta position resulted in decreased Nav1.7 potency and
increased MK-0499 binding potency (11). The effect of fluorine
substitution varies depending on the position of the substitution
(12–14). The para substitution enhanced Nav1.7 potency, while
ortho substitution caused a decrease in potency. Unfortunately,
the para-fluoro substitution also led to enhanced MK-0499 binding
activity (14). Next, we turned our attention to substitutions on the
cyclobutane ring. Introducing a polar hydroxyl group resulted in
complete loss of Nav1.7 activity (15). The methoxy group showed
a smaller effect with a twofold reduction of potency (16). Mono-
fluoro substitution had little effect on Nav1.7 potency (17, 18).
However, gem-difluoro substitution resulted in significant loss of
Nav1.7 activity (19). Compound 18 was evaluated for pharmacoki-
netic properties (Table 5). Unfortunately, its clearance rate was still
very high.
Table 2
Energies (in kcal/mol) and torsion angles (in degrees) of bent and extended
conformations of substituted N-benzylacetamides
R1, R2
Analog E (extended) Angle (extended) E (bent) Angle (bent)
H, H
H, Me
2
3
4
5
6
7
8
0.07
0
0
3.59
3.97
2.93
1.78
180
153
173
180
180
174
177
0
98
92
63
87
70
67
64
0.51
0.37
0
0
0
Me, Me
–(CH2)2–
–(CH2)3–
–(CH2)4–
–(CH2)5–
0
blocking Nav1.7 sodium channel were also counter screened
against several other ion channel targets, particularly hERG K+
channel, since block of this channel has been associated with
potentially lethal ventricular arrhythmias. A binding assay that
measures the displacement of 35S-labeled MK-0499, a known hERG
K+ channel blocker, was used as a high-throughput counter screen
assay for hERG K+ activity.8
Next we turned our focus to modification of the aromatic ring
attached to the isoxazole ring. As shown in Table 4, the effect of
lipophilic, electron-withdrawing substitutions on Nav1.7 potency
varies (20–22). However, these substituents generally cause strong
MK-0499 binding activity. The effect of a methoxy group is highly
dependent on its position (23–25). With ortho-methoxy substitu-
As shown in Table 1, substitution on the benzylic position of the
amide had a dramatic effect on Nav1.7 activity. The addition of one
methyl group (3) resulted in a small increase in potency. The gem-
dimethyl and cyclopropyl analogs (4, 5) showed significant
increase in potency. Maximal potency was reached with the cyclo-
butane compound (6) with an IC50 of 120 nM. When ring size was
Figure 2. Left: unsubstituted (black) versus mono-Me-substituted (enantiomer 1, red; enantiomer 2, blue). Right: unsubstituted (black) versus gem-dimethyl (green).

P. P. Shao et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5334–5338
5337
Figure 3. Left: unsubstituted (black) versus spirocyclopropyl (olive); Right: unsubstituted (black) versus spirocyclobutyl (light blue, two lines are from two cyclobutane ring
puckers).
Figure 4. Left: unsubstituted (black) versus spirocyclopentyl (orange, two lines are from two cyclopentane puckers). Right: unsubstituted (black) versus cyclohexyl (magenta,
two lines are from two cyclohexane puckers).
tion, the compound (23) was threefold less potent. para Substitu-
tion (25) had little effect on Nav1.7 potency. meta Substitution
(24) enhanced potency significantly. Compound 24 was one of
the most potent sodium channel blockers in the series with an
IC50 of 20 nM. The effect of a hydroxyl group also depends on sub-
stitution position (26–28). meta and para Substitution had little ef-
fect on potency. An ortho-hydroxyl group caused significant loss of
potency. However, the potency loss caused by ortho-hydroxyl sub-
stitution can be compensated by introducing a lipophilic group
such as OCF3 (29) or CF3 (30). It is interesting that the effect of
the CF3 group was highly dependent on its position with respect
to the hydroxyl group (30–34). When a racemic mixture of com-
pound 30 was separated on chiral HPLC to give a pair of enantio-
mers, the fast eluting isomer (30a) showed good Nav1.7 potency
and good selectivity against the hERG channel, and the slow elut-
ing isomer (30b) was completely inactive on the Nav1.7 channel.
Pyridine analogs showed good potency regardless of having elec-
tron donating (35) or withdrawing (36) substitutions. However,
the more polar pyridone analog (37) was completely inactive.
Compound 36 was highly selective against the hERG channel.
A selected group of compounds were further evaluated for
pharmacokinetic properties.10 As shown in Table 5, poor pharma-
cokinetic profile was a persistent issue with this series. One excep-
tion was compound 30. It had low clearance and good oral
bioavailability. The active enantiomer 30a was evaluated in the
rat spinal nerve ligation (SNL) model of neuropathic pain.11 It
showed comparable efficacy to CDA54 with 30% and 28% reversal
of mechanical allodynia at 2 and 4 h post a 10 mg/kg oral dose,
respectively, while CDA54 showed 35% reversal at 3 h after the
same oral dose.12
Table 3
Effect of benzylamide substitutions on Nav1.7 potency and MK-0499 binding activity
HN
N
Ar
O
O
R2
R1
Ar
Compound
R1 R2
hNav1.7 (% inh at
M or IC50 M)
MK-0499
(% inh at 10 lM)
1
l
,
l
6
11
12
13
14
15
16
17
18
19
H, H
H, H
H, H
H, H
H, H
H, OH
H, OMe
H, F
F, H
Ph
0.12
0.59
0.45
0.11
0.084
>5
0.38
0.18
0.16
41%
37%
89%
40%
64%
81%
3-CF3-Ph
2-F-Ph
3-F-Ph
4-F-Ph
Ph
Ph
Ph
Ph
Ph
51%
46%
59%
27%
F, F

5338
P. P. Shao et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5334–5338
Table 4
In summary, a series of isoxazole voltage gated sodium channel
blockers have been identified. Substitutions at the benzylic posi-
tion of the benzamide showed significant effects on Nav1.7 potency
with the most dramatic effect deriving from the spirocyclobutyl
substitution. Further SAR studies led to compound 30, which dis-
plays improved pharmacokinetic properties. One of the enantio-
mers, 30a, showed comparable in vivo efficacy in the rat SNL
neuropathic pain model as the previously reported voltage gated
sodium channel inhibitor CDA54.
Effect of aromatic ring substitutions on Nav1.7 potency and MK-0499 binding activity
HN
N
Ar
O
O
Compound
Ar
hNav1.7 (% inh at
M or IC50 M)
MK-0499
(% inh at 10
1
l
,
l
lM)
6
Ph
2-Cl-Ph
2-CF3-Ph
3-OCF3-Ph
2-OMe-Ph
3-OMe-Ph
4-OMe-Ph
2-OH-Ph
3-OH-Ph
0.12
0.18
0.74
0.37
0.39
0.020
0.18
44%
0.21
0.13
0.26
0.34
0.43
>10
0.30
1%
37%
76%
91%
55%
54%
62%
44%
54%
64%
15%
47%
45%
ND
Acknowledgments
20
21
22
23
24
25
26
27
28
29
30
30a
30b
31
32
33
34
The authors would like to thank Dr. Joseph Duffy for proof-read-
ing the manuscript. The authors would also like to thank Dr. Rich-
ard Hargreaves and Dr. Joseph Duffy for their helpful suggestions
on subtype selectivity issues. The authors were also grateful to
Dr. Samer Eid and Mr. Christopher J. Daley for their contribution
on Nav1.5 assay.
4-OH-Ph
2-OH-3-OCF3-Ph
2-OH-3-CF3-Ph
Enantiomer A
Enantiomer B
2-OH-4-CF3-Ph
2-OH-5-CF3-Ph
2-OH-6-CF3-Ph
4-OH-3-CF3-Ph
References and notes
ND
72%
ND
ND
ND
1. Shao, P. P.; Ye, F.; Weber, A. E.; Li, X.; Lyons, K. A.; Parsons, W. H.; Garcia, M. L.;
Priest, B. T.; Smith, M. M.; Felix, J. P.; Williams, B. S.; Kaczorowski, G. J.;
McGowan, E.; Abbadie, C.; Martin, W. J.; McMasters, D. R.; Gao, Y.-D. Bioorg.
Med. Chem. Lett. 2009, 19, 5329.
1%
0%
2. Ellman, J. A.; Owens, T. D.; Tang, T. P. Acc. Chem. Res. 2002, 35, 984.
3. Shao, P. P.; Ye, F. Tetrahedron Lett. 2008, 49, 3554.
4. Clark, J. H. Chem. Rev. 1980, 80, 429.
5. Grundmann, C. Synthesis 1970, 7, 344.
6. Geny, A.; Leboeuf, D.; Rouquie, G.; Vollhardt, K.; Peter, C.; Malacria, M.; Gandon,
V.; Aubert, C. Chem.—A Eur. J. 2007, 13, 5408.
7. 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. L.; Garcia, M. L. Assay Drug
Dev. Tech. 2004, 2, 260.
8. 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.
N
OMe
35
36
37
0.19
0.22
0%
54%
22%
ND
N
CF3
H
N
O
9. Torsional energy profiles of the substituted N-benzylacetamide model systems
in which the cyclopentane–phenylisoxazole was truncated to a methyl group
were calculated using MacroModel (Schrödinger, Portland, OR). The MMFFs
force field was employed with a distance-dependent dielectric of 4r. The
torsions were driven both forward and backward, and the lowest energy at
each dihedral angle was taken. For the spirocyclopentyl and spirocyclohexyl
analogs, two puckers of the carbocycle were used.
Table 5
Rat pharmacokinetic data for selected compounds
10. Rat PK experiments were conducted as follows: Test compounds were typically
formulated as 1.5 mg/mL solutions 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.
HN
Ar
O
N
O
R
d
Compound
R
Ar
F%a
AUCNb
Clpc
t1/2
6
18
22
24
27
29
30
H
F
Ph
Ph
3
12
46
2
0
6
0.024
0.056
0.20
0.011
0
58
86
89
79
293
35
1.3
1.7
2.6
1.6
11. Chaplan, S. R.; Bach, F. W.; Pogrel, J. W.; Chung, J. M.; Yaksh, T. L. J. Neurosci.
Methods 1994, 53, 55.
H
H
H
H
H
3-OCF3-Ph
3-OMe-Ph
3-OH-Ph
2-OH-3-OCF3-Ph
2-OH-3-CF3-Ph
12. (a) Shao, P.; Ok, D.; Fisher, M. H.; Garcia, M. L.; Kaczorowski, G. J.; Li, C.; Lyons,
K. A.; Martin, W. J.; Meinke, P. T.; Priest, B. T.; Smith, M. M.; Wyvratt, M. J.; Ye,
F.; Parsons, W. H. Bioorg. Med. Chem. Lett. 2005, 15, 1901; (b) Brochu, R. M.;
Dick, I. E.; Tarpley, J. W.; McGowan, E.; Gunner, D.; Herrington, J.; Shao, P.; Ok,
D.; Li, C.; Parsons, W. H.; Stump, G. L.; Regan, C. P.; Lynch, J. J., Jr.; Lyons, K. A.;
McManus, O. B.; Clark, S.; Ali, Z.; Kaczorowski, G. J.; Martin, W. J.; Priest, B. T.
Mol. Pharmacol. 2006, 69, 823.
0.16
0.88
0.92
0.06
1.9
49
9.5
N
CF3
36
H
36
0.1
135
1.51
a
b
c
Bioavailability.
Normalized AUC (po, lM h/mpk).
Clearance (ml/min/kg).
Half life (h).
d