1,5-Benzodioxepin derivatives as a novel class of muscarinic M3 receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2006.11.058

The structure-activity relationships of novel 1,5-benzodioxepin derivatives as muscarinic M₁-M₃ receptor antagonists are reported. Some of these compounds were found to possess high binding affinity for the muscarinic M₃ r

Full text of DOI:10.1016/j.bmcl.2006.11.058

Bioorganic & Medicinal Chemistry Letters 17 (2007) 925–931
1,5-Benzodioxepin derivatives as a novel class of muscarinic M3
receptor antagonists
Shuji Sonda,a,* Kenichi Katayama,c Masakazu Fujio,b Hiroshi Sakashita,b Kenichi Inaba,b
Kiyoshi Asanob and Toshiaki Akirab
aProcess Chemistry Laboratories, Pharmaceutical Technology Center, Mitsubishi Pharma Corporation, 14, Sunayama,
Kamisu, Ibaraki 314-0255, Japan
bPharmaceutical Research Division, Mitsubishi Pharma Corporation, 1000, Kamoshida-cho, Aoba-ku, Yokohama 227-0033, Japan
cPharmacovigilance and Quality Assurance Division, Mitsubishi Pharma Corporation, 1000, Kamoshida-cho, Aoba-ku,
Yokohama 227-0033, Japan
Received 21 August 2006; revised 15 November 2006; accepted 18 November 2006
Available online 22 November 2006
Abstract—The structure–activity relationships of novel 1,5-benzodioxepin derivatives as muscarinic M1–M3 receptor antagonists are
reported. Some of these compounds were found to possess high binding affinity for the muscarinic M3 receptor and potent effect on
rhythmic increase in bladder pressure in unanesthetized rats following oral administration. These compounds displayed selectivity
for the bladder over the salivary gland.
Ó 2006 Elsevier Ltd. All rights reserved.
Muscarinic receptors are expressed in peripheral tissues
and the central nervous system, which are classified into
the subtypes M1–M5.1 One of these receptor subtypes,
the muscarinic M3 receptor, which is distributed in the
airway, small intestine, urinary bladder, and secretory
glands, is believed to be a mediator of smooth muscle
contraction and glandular secretion, and its antagonists
are therefore thought to be potentially useful in the
treatment of disorders such as urinary incontinence
(UI), irritable bowel syndrome (IBS), and chronic
obstructive pulmonary disease (COPD). Unfortunately,
anti-cholinergic medication often brings about undesir-
able side effects such as tachycardia, dry mouth, consti-
pation, and blurred vision. For example, oxybutynin,
available for the treatment of overactive bladder, occa-
sionally causes dry mouth due to antagonism of the
muscarinic receptor in the salivary gland.
(diphenyl or phenylcycloalkyl) in their skeletons
(Fig. 1). We performed random screening to explore a
new class of compounds structurally different from these
muscarinic M3 receptor antagonists. We hypothesized
that these new compounds would be useful as anti-mus-
carinic drugs and would display different characteristics
from existing medicines for overactive bladder, IBS, and
COPD.
As a result, we discovered 1,5-benzodioxepin derivatives
have high affinity for the muscarinic M3 receptor.
In the present study, we attempted to develop the 1,5-
benzodioxepins as useful medicines for urinary inconti-
nence. In the present paper, we report the preparation
and structure–activity relationship (SAR) of 1,5-benzo-
dioxepin derivatives, their muscarinic M1–M3 receptor
binding affinities, and the effect of these derivatives on
overactive bladder and the salivary gland.
The muscarinic M3 receptor antagonists currently
known (oxybutynin, solifenacin,2 darifenacin,3 toltero-
dine,4,5 tiotropium,6 Banyu compound,7–9 KRP-
197)10,11 have relatively similar chemical structures
The preparation of the 1,5-benzodioxepin key interme-
diates 5a–c was performed from the starting materials
catechol, 4-fluorocatechol, or 4,5-dichlorocatechol as
shown in Scheme 1.12–14 The alkylation of catechols
with methyl chloroacetate gave compounds 1a–c, which
were then cyclized under basic conditions (60% NaH in
DMSO) to give 2a–c. Decarboxylation of these com-
pounds by treatment with refluxing HCl/2-propanol
Keywords: Muscarinic M1–M3 receptor; Antagonists; Urinary incon-
tinence; 1,5-benzodioxepin derivatives.
*
Corresponding author. Tel.: +81 479 46 7362; fax: +81 479 46
7364; e-mail: Sonda.Shuuji@ma.m-pharma.co.jp
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.11.058

926
S. Sonda et al. / Bioorg. Med. Chem. Lett. 17 (2007) 925–931
O
N
OH
O
N
OH
N
O
CONH2
Me
Oxybutynin
Darifenacin
Tolterodine
N
NH2
O
N
N
N
O
N
H
H
N
O
Me
OH
N
CONH2
F
F
Banyu compound
Solifenacin
KRP-197
Figure 1.
COOMe
O
COOMe
O
O
4
O
O
OH
OH
8
a
R6
R6
R6
7
5
COOMe
1a : R6 = H (61%)
1b : R6 = 4-F (82%)
1c : R6 = 4,5-di-Cl (39%)
2a : R6 = H
2b : R6 = 7-F (or/and 8-F)
2c : R6 = 7,8-di-Cl
O
O
O
O
CN
NH
c
d, e
f
O
R6
R6
N
R6
N
O
O
3a : R6= H (42%)
3b : R6= 7-F (68%)
3c : R6= 7,8-di-Cl (67%)
4a : R6= H
4b : R6= 7-F
4c : R6= 7,8-di-Cl
5a : R6= H (73%)
5b : R6= 7-F (36%)
5c : R6= 7,8-di-Cl (24%)
Scheme 1. Reagents: (a) ClCH2COOCH3, K2CO3, KI, MeOH (or 2-butanone); (b) 60% NaH, DMSO/toluene; (c) HCl/2-propanol; (d) KCN, H2O/
MeOH; (e) MeNH–HCl; (f) Red-Al/toluene.
afforded the 1,5-benzodioxepin-3-one derivatives 3a–c.
These compounds were treated with KCN and the
resulting cyanohydrines were reacted with Me2NH
hydrochloride to give compounds 4a–c, reduction of
which with Red-Al gave the desired key intermediates
5a–c.
The preparation of compounds 10a and b was per-
formed by cyanation of the ketone 3a followed by treat-
ment with pyrrolidine or piperidine hydrochloride. The
nitriles 10a and b were reduced and the resulting pri-
mary amines 11a and b were treated with a 2-phen-
ylethylbromide to provide compounds 12a and b,
respectively (Scheme 3).
The intermediates 5a, 5b (racemates), 5c, and 6a were
functionalized as shown in Scheme 2. The derivatives
6b, c, e, and 8g–i were synthesized by coupling of 5a with
aralkylhalide, alkylhalide in MeCN in the presence of
K2CO3. Reductive amination of 6a with HCHOaq–
NaBH3CN afforded the tertiary amine 6h.
Compound 4a was treated with Grignard reagents
derived from 4-cyclohexylbutylbromide and cyclo-
hexylmethyl bromide to give compounds 13a and 13b,
respectively (Scheme 4).
The binding affinity of 1,5-benzodioxepin derivatives for
the human muscarinic receptor subtypes (M1, M2, and
M3) expressed in insect Sf 9 cells was determined by
measuring their ability to inhibit the binding of 1-[N-
methyl-3H]scopolamine methyl chloride.15
The compounds 6a, d, f, g and 8a–f, j–m were synthe-
sized by condensation of 5a–c with a variety of carboxy-
lates followed by reduction of the amide group using
NaBH4–BF3Et2O (or borane–THF).

S. Sonda et al. / Bioorg. Med. Chem. Lett. 17 (2007) 925–931
927
R1
X
R1
CH3
N
a
O
O
O
N
b
X = Br, OTs
32-68%
H
6a
N
N
O
57%
c,d
R1
COX
6h
6a-g
41-76%
X = OH, Cl
R5
R5
(CH2)n
(CH2)n
c,d
O
O
COX
X
N
H
X = OH, Cl
R6
41-76%
e
N
R5
n=1,2
8a-i,l,m
(CH2)n
50-58%
X = Cl, Br, OMs, OTs
n = 1,2
OH
(CH2)n
X
X
O
O
O
f
g,h or i,,j
19-32%
O
NH
O
O
NH
N
HO
N
57-94%
n=1,2
mixture of cis/trans
X=O, CHOMe
9j, k
8j, k
Scheme 2. Reagents: (a) 5a, K2CO3, (KI)/MeCN; (b) 37% HCHOaq, NaBH3CN; (c) 5a–c, Bop-reagent (or DEPC, EDC, none)/THF (or AcOEt,
DMF); (d) NaBH4, BF3–Et2O (or BH3–THF)/THF, then HCl/EtOH; (e) 5a, K2CO3, (KI)/MeCN; (f) 5a, Bop-reagent/THF; (g) NaBH4/EtOH; (h)
NaBH4, BF3–Et2O; (i) HCl/THF; (j) BH3–THF, then HCl/EtOH.
O
O
CN
N
c
a, b
3a
(CH2)n
10a :n=1
10b :n=2
O
O
NH2
O
NH
d
N
N
(CH2)n
O
(CH2)n
11a :n=1 (56%)
11b :n=2 (34%)
12a :n=1 (17%)
12b :n=2 (64%)
Scheme 3. Reagents: (a) KCN, H2O, MeOH; (b) pyrrolidine–HCl or piperidine–HCl ; (c) Red-Al/toluene; (d) 2-phenylethylbromide, K2CO3/MeCN.
O
(CH2)n
b
a
(CH2)n
Br
N
O
13a : n = 3 (63%)
13b : n = 0 (42%)
Scheme 4. Reagents: (a) Mg, I2/THF; (b) 4a/THF.

928
S. Sonda et al. / Bioorg. Med. Chem. Lett. 17 (2007) 925–931
Table 1. Binding affinity of compounds to muscarinic M1, M2, and M3 receptors
4
R1
2
6
O
N
R2
R3
O
Compound
R1
R2
R3
Binding affinitya Ki (nM)
Ratio
M1
M2
M3
M1/M3
M2/M3
6a
6b
6c
6d
6e
6f
H
H
H
H
H
H
H
H
Me
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
N(CH3)2
220
220
260
200
NTb
540
380
NTb
90
48
20
18
11
12
13
3.6
—
23
20
—
4.5
2.7
2.5
0.8
—
4-Me
4-Cl
49
20
4-i-Pr
4-MeO
2,6-Di-Cl
2-F
44
56
NTb
49
>100c
23
2.1
2.1
—
6g
6h
41
NTb
19
>100c
H
12a
H
H
H
H
NTb
NTb
>100c
—
—
12b
Oxybutynin
NTb
1.0
NTb
8.1
>100c
0.78
—
1.3
—
10
a Each value is the mean from triplicate assays in a single experiment.
b Not tested.
c IC50 value.
The binding affinity of the parent compound 6a, its
derivatives 6b–h and 12a, b is presented in Table 1.
Compound 6a with the 2-phenylethylamine moiety at
the 3-position of the 1,5-benzodioxepin ring showed
moderate binding affinity for the muscarinic M1, M2,
and M3 receptors (Ki = 220, 90 nM, and M3 = 20 nM,
respectively).
this receptor does not have enough space for bulkier
alkylamines such as pyrrolidine or piperidine.
The M1–M3 receptor binding affinity of compounds
with cycloalkyl amine at the 3-position of 1,5-benzodi-
oxepin is summarized in Table 2. Cyclohexyl
derivative 8c obtained by saturation of the 2-phenyl-
ethylamine moiety of 6a has 8-fold higher binding
affinity for the M3 receptor than the parent
compound.
At first, we examined the influence of the substituent
group in the benzene ring of the 2-phenylethylamine moi-
ety of 6a. The introduction of a methyl group (6b) or a
chlorine atom (6c) at the 4-position of the benzene ring
had little effect on the binding affinity for the M1, M2,
and M3 receptors. Replacement with an isopropyl group
at this position (compound 6d) resulted in decreased affin-
ity for the M3 receptor compared with 6a. This decrease
was particularly dramatic when a methoxy group was
incorporated at the 4-position of the benzene ring (com-
pound 6e; IC50 > 100 nM). These results suggest that
the introduction of electron-donating groups at this
position is unfavorable for muscarinic M3 receptor
binding.
On the other hand, this transformation had almost no
effect on the binding affinity ratio for the M3 versus
M1–M2 receptors. The high affinity for the muscarinic
M1–M3 receptors was retained after replacement of the
cyclohexyl moiety of compound 8c with a five-mem-
bered ring (8b), a seven-membered ring (8d), or a 1-ada-
mantyl group (8e). This suggests that the corresponding
hydrophobic binding region of the muscarinic M1–M3
receptors has a common space favorable for a five- to
seven-membered cycloalkyl or 1-adamantyl group in
the aminomethyl moiety at the 3-position of the 1,5-ben-
zodioxepin and that this binding region prefers an ali-
phatic cycloalkyl ring to an aromatic ring (compounds
8a–e vs 6a).
Transformation of the dimethylamino group to 1-pyrro-
lidinyl (12a) or 1-piperidino (12b) led to drastic decrease
of the binding affinity for the M3 receptor
(IC50 > 100 nM). This finding suggests that the musca-
rinic M3 receptor has an interaction with the dimethyl-
amine part at the 3-position of 1,5-benzodioxepin
derivatives and that the corresponding binding site of
Meanwhile, insertion of an oxygen atom (8f) or a sulfur
atom (8h) at the 4-position of the cyclohexyl ring de-
creased the binding affinity for the muscarinic M3 recep-
tor compared with compound 8c.

S. Sonda et al. / Bioorg. Med. Chem. Lett. 17 (2007) 925–931
Table 2. Binding affinity of compounds to muscarinic M1, M2, and M3 receptors
929
O
R4
R5
7
(CH2)n
3
R6
N
8
O
1
2
Compound
n
R4
R5
R6
Binding affinity Kia (nM)
Ratio
M1
31
M2
12
M3
4.9
M1/M3
6.3
M2/M3
8a
1
NH
H
2.4
8b
2
2
2
2
2
2
2
2
2
2
2
2
2
0
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
CH2
None
H
H
H
H
H
H
H
H
H
H
7-F
21
2.4
3.3
6.4
7.2
4.4
3.9
5.3
6.1
7.6
24
0.7
2.5
0.7
0.4
0.5
0.6
1.1
2.1
0.7
0.9
1.2
0.55
—
8c
18
6.2
2.5
8d
12
2.0
2.7
8e
5.9
58
0.66
6.0
1.5
8f
11
8g
86
8.8
14
8h
52
7.2
6.8
8i
140
19
12
5.8
8jb
8kb
8lc
8m
13a
13b
1.9
2.9
6.6
8.6
8.6
8.2
—
18
1.8
2.1
36
5.1
4.2
7,8-Di-Cl
1.8
NTd
NTd
0.12
NTd
NTd
0.22
>100e
>100e
H
H
—
—
a Each value is the mean from triplicate assays in a single experiment.
b A mixture of cis/trans isomers.
c A mixture of enantiomers.
d Not tested.
e IC50 value.
Additionally, the presence of a double bond at the
3,4-position of the tetrahydrothiopyran (8i) caused a
2.7-fold loss in binding affinity for the muscarinic M1
receptor compared with 8h and increased the binding
affinity ratio for the M3 versus M1 receptor
(M1/M3 = 24). Likewise, the 2,6-dichlorophenyl (6f)

930
S. Sonda et al. / Bioorg. Med. Chem. Lett. 17 (2007) 925–931
Table 3. Inhibitory effects of oxybutynin and the compounds 8g, j, k, and l on rhythmic contraction of bladder and methacholine-induced salivation
Compound
ID50 (mg/kg, iv)
Rhythmic contractiona
Salivation/rhythmic contraction
Salivationb
Oxybutynin
0.038
0.012
0.0088
0.015
0.010
0.11
0.25
0.18
0.16
0.20
2.9
20
20
10
20
8g
8j
8k
8l
a The ID50 values represent the doses to reduce the atropine-sensitive rhythmic contraction by 50% in unanesthetized rats (n = 4).
b The ID50 values represent the 50% inhibitory doses on methacholine-induced salivation in anesthetized rats (n = 5).
stituents did not affect the binding affinity ratio for the
muscarinic M3 versus M1–M2 receptors.
We examined the inhibitory effect of intravenous injec-
tion of oxybutynin and compounds 8g, j, k, l on rhyth-
mic bladder contraction in unanesthetized rats16 and
methacholine-induced salivation in anesthetized rats17
(Table 3). The inhibitory effect of compounds 8g, j, k,
and l on bladder contraction was 10- to 20-fold more
potent than on the salivary gland (oxybutynin: 2.9-fold).
This result suggests that these compounds possess tissue
selectivity for the bladder over the salivary gland in rat.
The effect of oral administration (3 mg/kg) of compound
8k on rhythmic increase in the bladder pressure of
unanesthetized rats is presented in Figure 2. In oral
administration, this compound was found to have more
potent effect than oxybutynin.
Figure 2. Effects of 8k (3 mg/kg, po) and oxybutynin (3 mg/kg, po) on
the rhythmic increase in bladder pressure of unanesthetized rats. Each
symbol and bar indicates the mean SEM.
In summary, we prepared a novel class of 1,5-benzodi-
oxepin derivatives with high affinity for the muscarinic
M1–M3 receptors. These compounds were structurally
different from the previously known typical muscarinic
receptor M1–M3 antagonists.
and 2-fluorophenyl (6g) derivatives showed some selec-
tivity between the M3 and M1 receptors (M1/M3 =
20–23) as shown in Table 1. These results suggest that
factors at the ortho-position of a six-membered ring such
as thiopyran bearing a double bond or a 2,6-dichloro-
phenyl group of derivatives may be unfavorable for
muscarinic M1 receptor binding affinity.
SAR studies established that the secondary amine group
at the 3-position on 1,5-benzodioxepin was essential
moiety for muscarinic M1–M3 receptor binding affinity.
Specifically, it was observed that a number of 1,5-benzo-
dioxepin derivatives with a cycloalkyl group in the sec-
ondary amine moiety possessed high affinity for the
muscarinic M1–M3 receptor. One of these, the 1,5-ben-
zodioxepin derivative 8k, displayed high affinity for
the muscarinic M1–M3 receptors, had potent effect on
rhythmic increase in the bladder pressure of unanesthe-
tized rats, and displayed selectivity for the bladder over
the salivary gland.
In the next step, SAR development was focused on var-
iation of the secondary amine moiety. Methylation of
the secondary amine (6h) caused a drastic loss in the
binding affinity for the M3 receptor as shown in Table
1. Similarly, transformation of this amino group into a
methylene chain (13a and 13b) produced a notable de-
crease in the binding affinity (IC50 >100 nM) for the
muscarinic M3 receptor. These findings support the
hypothesis that the secondary amino group is a moiety
essential for binding affinity and acts as a hydrogen
bond donor to the corresponding binding site of the
muscarinic M3 receptor.
In addition, 1,5-benzodioxepin derivatives should pro-
vide useful anti-muscarinic drugs for disorders such as
irritable bowel syndrome (IBS) and chronic obstructive
pulmonary disease (COPD).
Finally, we performed introduction of halogen atoms
into the 1,5-benzodioxepin moiety. The introduction of
a fluorine atom at the 7-position of 1,5-benzodioxepin
(8l) somewhat increased the binding affinity for the mus-
carinic M3 receptor (Ki = 4.2 nM) compared with the
unsubstituted compound 8f. The 7,8-dichloro derivative
8m possessed 50-fold higher affinity (Ki = 0.22 nM) than
8f. Meanwhile, the incorporation of these halogen sub-
Acknowledgments
We thank Ms. M. Miyoshi and Ms. Y. Hattori for some
of the biological results. We also thank Dr. K. Adachi,
Dr. Y. Matsumoto, and Mr. K. Haga for helpful
discussion.

S. Sonda et al. / Bioorg. Med. Chem. Lett. 17 (2007) 925–931
931
12. Rooney, C. S.; Stuart, R. S.; Wasson, B. K.; Williams, H.
W. R. Can. J. Chem. 1975, 53, 2279.
References and notes
13. Tahara, T.; Masago, M.; Fuji, T.; Anami, N. Jpn. Kokai
Tokkyo Koho 1979, 7873, 581; . Chem. Abstr. 1979, 90,
23131r.
14. Sonda, S.; Katayama, K.; Ikebe, T.; Fujio, M.; Nakam-
ura, M. Jpn. Kokai Tokkyo Koho 2000, 169468.
15. The suspension of membranes from insect Sf 9 cells
expressing cloned human m1–m3 (NEN Life Science) was
incubated with tested compounds and 1-[N-meth-
yl-3H]scopolamine methyl chloride (final concentration
of 0.3 nM) at 25 °C for 1 h. Reactions were terminated by
rapid filtration through a Whatmann GF/B filter and the
radioactivity was counted by liquid scintillation counter.
Non-specific binding was defined in the presence of
1. Kubo, T.; Fukuda, K.; Mikami, A.; Maeda, A.; Takah-
ashi, H.; Mishina, M.; Haga, T.; Haga, K.; Ichiyama, A.;
Kangawa, K.; Minamino, N.; Matsuo, H.; Numa, S.
Nature 1986, 323, 411.
2. Ikeda, K.; Kobayashi, S.; Suzuki, M.; Miyata, K.;
Takeuchi, M.; Yamada, T.; Honda, K. Naunyn-Schmiede-
berg’s Arch. Pharmacol. 2002, 366, 97.
3. Alabaster, V. A. Life Sci. 1997, 60, 1053.
4. Nilvebrant, L.; Hallen, B.; Larsson, G. Life Sci. 1997, 60,
1129.
5. Gillberg, P. G.; Sundquist, S.; Nilvebrant, L. Eur.
J. Pharmacol. 1998, 349, 285.
6. Disse, B.; Speck, G. A.; Rominger, K. L.; Witek, T. J., Jr.;
Hammer, R. Life Sci. 1999, 64, 457.
atropine sulfate to give
a final concentration of
0.01 mM. The IC50 value was determined by non-linear
regression of the displacement curve, and the Ki value was
calculated according to the formula Ki = IC50/(1 + L/Kd),
where L is the concentration of radioligand and Kd is the
dissociation constant of the radioligand.
7. Mitsuya, M.; Mase, T.; Tsuchiya, Y.; Kawakami, K.;
Hattori, H.; Kobayashi, K.; Ogino, Y.; Fujikawa, Y.;
Satoh, A.; Kimura, T., et al. Bioorg. Med. Chem. 1999, 7,
2555.
8. Mitsuya, M.; Kobayashi, K.; Kawakami, K.; Satoh, A.;
Ogino, Y.; Kakikawa, T.; Ohtake, N.; Kimura, T.; Hirose,
H.; Sato, A.; Numazawa, T.; Hasegawa, T.; Noguchi, K.;
Mase, T. J. Med. Chem. 2000, 43, 5017.
9. Ogino, Y.; Ohtake, N.; Kobayashi, K.; Kimura, T.;
Fujikawa, T.; Hasegawa, T.; Noguchi, K.; Mase, T.
Bioorg. Med. Chem. Lett. 2003, 13, 2167.
10. Miyachi, H.; Kiyota, H.; Segawa, M. Bioorg. Med. Chem.
1999, 7, 1151.
16. We applied Kawashima’s method to this examination in
unanesthetized rats: Kawashima, K. Anitechs 1991, 3, 80.
17. Male Crj; CD rats (180–300 g) were anesthetized with
urethane (1.5 g/kg ip) and administered methacoline
(1 mg/kg sc). The cotton balls were placed in oral cavity
of rats for 5 min and the amounts of the secreted saliva
were estimated by the comparison between the weight of
the cotton ball before and after the insertion. The
compounds were administered intravenously 5 min prior
to the injection of methacoline.
11. Miyachi, H.; Kiyota, H.; Uchiki, H.; Segawa, M. Bioorg.
Med. Chem. 1998, 8, 1807.