Structure - Antitussive activity relationships of naltrindole derivatives. Identification of novel and potent antitussive agents

Article abstract of DOI:10.1021/jm701440h

We have previously reported antitussive effects of naltrindole (NTI), a typical δ opioid receptor antagonist, in a rat model. The ED₅₀ values of NTI by intraperitoneal and peroral injections were 104 μg/kg and 1840 μg/kg, respectively, comparable to those of codeine. Codeine, one of the most reliable centrally acting antitussive drugs, has μ agonist activity and thus the same side effects as morphine, e.g., constipation, dependency, and respiratory depression. Because NTI is a δ opioid antagonist, its derivatives have potential as highly potent antitussives, free from the μ opioid agonist side effects. We attempted to optimize the NTI derivatives to develop novel antitussive agents. On the basis of the studies of structure - antitussive activity relationships of alkyl substituted NTI derivatives, we designed NTI derivatives with extra ring fused structures. As a clinical candidate, we identified a highly potent new compound, (5R,9R,13S,14S)-17- cyclopropylmethyl-6,7-didehydro-4,5-epoxy-5′,6′-dihydro-3-methoxy- 4′H-pyrrolo[3,2,1-ij]quinolino[2′,1′:6,7]morphinan-14-ol (5b) methanesulfonate (TRK-850) which was effective even by oral administration (ED₅₀ 6.40 μg/kg).

Full text of DOI:10.1021/jm701440h

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
4404
J. Med. Chem. 2008, 51, 4404–4411
Structure-Antitussive Activity Relationships of Naltrindole Derivatives. Identification of Novel
and Potent Antitussive Agents
Satoshi Sakami,† Masayuki Maeda,† Koji Kawai,† Takumi Aoki,† Kuniaki Kawamura,† Hideaki Fujii,†,§ Ko Hasebe,†
Mayumi Nakajima,† Takashi Endo,† Shinya Ueno,† Tsuyoshi Ito,† Junzo Kamei,‡ and Hiroshi Nagase*,†,§
Pharmaceutical Research Laboratories, Toray Industries, Inc., 6-10-1 Tebiro, Kamakura, Kanagawa 248-8555, Japan, and Department of
Pathophysiology & Therapeutics, School of Pharmacy and Pharmaceutical Sciences, Hoshi UniVersity, 4-41, Ebara 2-chome,
Shinagawa-ku, Tokyo 142-8501, Japan
ReceiVed NoVember 14, 2007
We have previously reported antitussive effects of naltrindole (NTI), a typical δ opioid receptor antagonist,
in a rat model. The ED50 values of NTI by intraperitoneal and peroral injections were 104 µg/kg and 1840
µg/kg, respectively, comparable to those of codeine. Codeine, one of the most reliable centrally acting
antitussive drugs, has µ agonist activity and thus the same side effects as morphine, e.g., constipation,
dependency, and respiratory depression. Because NTI is a δ opioid antagonist, its derivatives have potential
as highly potent antitussives, free from the µ opioid agonist side effects. We attempted to optimize the NTI
derivatives to develop novel antitussive agents. On the basis of the studies of structure-antitussive activity
relationships of alkyl substituted NTI derivatives, we designed NTI derivatives with extra ring fused structures.
As a clinical candidate, we identified a highly potent new compound, (5R,9R,13S,14S)-17-cyclopropylmethyl-
6,7-didehydro-4,5-epoxy-5′,6′-dihydro-3-methoxy-4′H-pyrrolo[3,2,1-ij]quinolino[2′,1′:6,7]morphinan-14-
ol (5b) methanesulfonate (TRK-850) which was effective even by oral administration (ED50 6.40 µg/kg).
Chart 1. Structures of NTI and Its Derivatives with an Extra
Fused Ring
Introduction
Three types of opioid receptors (µ, δ, and κ) are now well
established based not only on pharmacological studies but also
on molecular biological investigations,1–3 and a number of
nonpeptide opioid receptor ligands have been developed as either
drug candidates or pharmacological tools.4–6 The antitussive
effects of the µ and κ opioid agonists have also been well
recognized.7–9 In addition, the δ opioid receptor may counteract
the antitussive processes that are mediated by the µ and κ opioid
receptor.10,11 On the basis of these observations, we previously
reported that naltrindole (NTI)a (Chart 1), a typical δ opioid
antagonist,12–14 exerted a marked and long-lasting antitussive
effect in mice and rats.15 We also demonstrated that NTI
suppressed the cough reflex mainly by functioning as a δ opioid
antagonist and that the antitussive effect of NTI resulted from
the antagonism of the δ opioid receptor-mediated internal µ and
κ opioid inhibitory system for the antitussive process.15,16 These
results directly supported the feasibility of the development of
δ opioid antagonists as antitussives.
In contrast to our research results, SmithKline Beecham has
described a δ opioid agonist with an antitussive effect.17 This
discrepancy can be explained hypothetically by the observation
that the δ1 and δ2 pharmacological subtype receptors18 affect
the cough response through opposing mechanisms. The δ2
receptor agonist may suppress the cough reflex in the same
manner as the µ or κ receptor agonists do, whereas the δ1 agonist
inhibits the cough suppressive effects of the µ or κ receptor
agonists.19 Alternatively, the discrepancy in the pharmacological
effects attributed to the opioid receptor subtypes, as described
above, may in fact arise from activation of opioid receptor
heterodimers.20,21
NTI was designed and synthesized based on a message-
address concept based on the endogenous δ opioid peptide
enkephalin, with the structure (Tyr-Gly-Gly-Phe-X).12 The
morphinan structure, the pyrrole ring, and the benzene ring of
NTI mimic 1Tyr-2Gly as a message, 3Gly as a spacer, and 4Phe
as an address, respectively. We proposed that indolomorphinan
is a good starting structure to synthesize new and potent
antitussive products by conserving both the δ antagonist and
antitussive activities of NTI. Therefore, we investigated the
structure-antitussive activity relationships among the NTI
derivatives to define their structural determinants in order to
design novel and improved analogues.
Rationale
Centrally acting drugs must penetrate the intractable mem-
brane barriers, such as the blood-brain barrier, in order to be
effective in vivo. The physicochemical requisites for the
compounds capable of passing through this barrier are low
molecular weight, high n-octanol/water distribution coefficient
characteristics, low numbers of hydrogen bond donors, and so
on.22–24 Although systemically injected NTI itself fully sup-
presses the cough reflex, we considered that the augmentation
of the hydrophobic properties of NTI would enable a compound
* To whom correspondence should be addressed. Phone: +84-3-5791-
6372. Fax: +84-3-3442-5707. E-mail: nagaseh@pharm.kitasato-u.ac.jp.
†
Toray Industries, Inc.
‡
Hoshi University.
Current affiliation is School of Pharmacy, Kitasato University, 5-9-1
§
Shirokane, Minato-ku, Tokyo 108-8641, Japan.
a Abbreviations: CHO, Chinese hamster ovary; DAMGO, [D-Ala2,
MePhe4, Gly-ol6]enkephalin; DPDPE, [D-Phe2,5]-enkephalin; HEK, human
embryonic kidney; MVD, mouse vas deferens; NTI, naltrindole.
10.1021/jm701440h CCC: $40.75
2008 American Chemical Society
Published on Web 07/19/2008

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Structure-AntitussiVe ActiVity of Naltrindole DeriVatiVes
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15 4405
Scheme 1. Synthesis of Alkyl Substituted NTI Derivativesa
Table 1. Antitussive Effects of Alkyl Substituted NTI Derivatives on
the Capsaicin-Induced Coughs in Rats (ip)
compound R1
R2
R3
antitussive activity, ED50 (µg/kg)a,b
3a (NTI)
H
H
H
104 (20.3-553)
1830 (890-3820)c
63.9 (29.4-139)
204 (41.8-999)
14.8 (4.09-53.8)
6.35 (2.05-19.6)
1.79 (0.67-4.89)
3b
3c
3d
3e
3f
Me
H
Me n-Pr
Me Me
Me Et
H
Me
H
H
H
7′-Me
7′-Me
a Reagents and conditions: (a) MeSO3H, EtOH, reflux; (b) R2OTs,
n-Bu4NHSO4, NaOH, water, benzene, 50 °C.
a ED50 values; the dose which reduces the number of coughs to 50% vs
control, are expressed as mean (N ) 8). Figures in parentheses indicate
95% confidence limits. b Each compound was administered ip route unless
otherwise noted. c Administered perorally.
previous paper.15 Test compounds were administered by ip or
po routes, and the difference in the cough number for 3 min
between the pre- and postdrug injection was examined.
Opioid receptor antagonistic activities were evaluated using
electrically stimulated mouse vas deferens (MVD) prepara-
tions.27 Morphine,[D-Phe2,5]-enkephalin(DPDPE),andU-50,488H
were used as µ, δ, and κ agonists, respectively. The antagonist
potencies are expressed as pA2 values.
Opioid receptor binding affinities were determined by dis-
placement of radioligands from guinea pig brain membranes.28
[3H][D-Ala2,MePhe4,Gly-ol6]enkephalin ([3H]DAMGO), [3H]-
NTI, and [3H]U-69,593 were used as µ, δ, and κ radioligands,
respectively.
Opioid receptor agonist activities were evaluated by
[35S]GTPγS binding assay, which were carried out using
membrane preparations from transfected CHO (δ and µ) or
HEK-293 (κ) cells that constitutively expressed the respective
human opioid receptor type.29–32
Substituted analogues of NTI were compared to the unsub-
stituted parent compound (Table 1). NTI (3a) showed a
remarkable antitussive activity (ED50 104 µg/kg, 95% confidence
limits 20.3-533 µg/kg) when administered by ip route, but, in
contrast, an oral administration of the compound resulted in
significantly lower antitussive activity (ED50 1830 µg/kg, 95%
confidence limits 980-3820 µg/kg). This difference in antitus-
sive activity between ip and po administration suggested that
NTI would be poorly absorbed perorally.
to more easily pass through the blood-brain barrier and thus
potentiate its activity in vivo. We therefore prepared a series of
compounds with these features and focused on NTI derivatives
with hydrophobic substituents to identify the structural require-
ments as potent antitussive agents.
The antitussive activity of these compounds 3a-f revealed
interesting structure-activity relationships (Table 1). Simulta-
neous alkyl substitution at the 3-hydroxy group, the 1′ position,
and the 7′ position of NTI significantly improved the antitussive
activity (3e,f). Among these NTI derivatives, compound 3f
showed excellent antitussive activity, however, its δ receptor
antagonist activity was decreased in an opioid receptor antago-
nism test using mouse vas deferens (MVD) preparations (vide
infra).
On the basis of these results, we hypothesized that substitution
at the 1′ and 7′ positions with hydrophobic substituents would
result in improved permeability through the intractable mem-
brane barriers, and as for the hydrophobic substituents, compact
functionalities might be superior for avoiding a decrease of δ
opioid receptor antagonist activity. This working hypothesis led
us to design new types of NTI analogues. We considered that
connection of the 1′ with 7′ position by an alkyl chain may
possibly satisfy this hypothesis. Consequently, we synthesized
additional ring-fused derivatives (4-6) (Chart 1), which have
hydrophobic structures with minimum steric hindrance.
Chemistry
Transformation of the 3-hydroxy group of NTI to the methyl
ether (3b) slightly increased the antitussive activity, but the effect
of this transformation was modest. In contrast, incorporation
of a methyl group at the 1′ position of NTI decreased antitussive
activity (3c). However, simultaneous alkyl substitution at both
the 3-hydroxy group and the 1′ position drastically improved
the antitussive activity (3d). Moreover, introduction of a third
alkyl substituent at the 7′ position further enhanced the potency
of compounds 3e and 3f by factors of 16 and 58, respectively,
as compared to the reference compound. From these results,
we presumed that the substitution of alkyl groups around the
indole substructure increased the hydrophobicity of these
compounds, which was supported by calculated logP values,33
and improved their permeability through the blood-brain barrier
and, as a consequence, the antitussive activities were improved.
Among these NTI derivatives, compound 3f showed excellent
antitussive activity, but its δ opioid receptor antagonist activity was
decreased in the MVD test; its pA2 values of δ, κ, and µ receptors
were 6.69, 6.58, and 6.67, respectively. Because NTI is a highly
Substituted NTI derivatives were synthesized as described
in Scheme 1. Compounds 3a-c13,25 and 3g were obtained under
the conditions of the Fischer indole synthesis from naltrexone
(1a) or 3-O-methylnaltrexone (1b) and the corresponding aryl-
hydrazines 2a-c.12 The indole nitrogens of 3b and 3g were
alkylated with appropriate alkylating agents to give 3d and 3e
or 3f, respectively.
The additional ring-fused derivatives 4-6 were also synthesized
using the Fischer indole synthesis from naltrexone (1a) or 3-O-
methylnaltrexone (1b) with arylhydrazines 9a-c, which were
derived from the corresponding cyclic aryl amines 8a-c (Scheme
2). The cyclohexene ring-fused derivative 7 was also prepared from
hydrazine 10 using the same conditions (Chart 2).26 O-Methylation
of 6a was conducted to give compound 6b.
Results and Discussion
The antitussive effect was evaluated in in vivo studies using
the rat capsaicin-induced cough model as described in our

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
4406 Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15
Sakami et al.
Scheme 2. Synthesis of NTI Derivatives with an Extra Fused Ringa
a Reagents and conditions: (a) NaNO2, HCl, EtOH, water, 0 °C, then Na2S2O4, NaOH, reflux; (b) MeSO3H, EtOH, reflux; (c) MeI, K2CO3, DMF, rt.
Chart 2. Structures of Compound 7 and Its Starting Material 10
between the 6′ and 7′ positions, also exhibited more potent
antitussive activity than NTI.
The NTI derivatives with an extra fused ring (4, 5a-b, 6a)
exhibited high δ antagonist potency and high δ receptor
selectivity (Table 2). The 5- and 6-membered ring fused
derivatives 4 and 5a showed especially high δ antagonist
potency. Transformation of the 3-hydroxy group of compound
5a into a methoxy group (5b) somewhat decreased antagonist
potency, but the overall δ antagonist potency remained high.
Compound 6b, with a 7-membered fused ring and a methoxy
group at the 3 position, exhibited poor antagonist potency.
Despite this drawback, compound 6b retained high antitussive
activity. The antitussive activity of 6b may be induced by its
active metabolites rather than from 6b itself.
Compounds 4, 5a-b, and 6a also showed higher δ receptor binding
affinity than compound 3f by opioid receptor binding affinity assay
(Table 3). In the case of 3-OMe derivatives, compound 5b (propylene
bridge) showed higher δ receptor binding affinity than compound 6b
(butylene bridge). With respect to 3-OH analogues, the order of δ
receptor binding affinities was 4 (ethylene bridge) > 5a (propylene
bridge) > 6a (butylene bridge). These results strongly suggested that
δ receptor binding affinities would be correlated with bulkiness around
the indole substructure and that compact fused rings would be
important for exhibiting higher δ receptor binding affinity.
We examined the opioid receptor agonist activities of selected
compounds at a high compound concentration of 1 µM using
[35S]GTPγS binding assay (Table 3). All of the tested com-
pounds were inactive toward µ receptor, whereas some com-
pounds exhibited weak δ or κ receptor agonist activities. The
3-OMe analogues showed increased δ receptor agonist activities
compared with the parent compound NTI or the corresponding
3-OH derivatives, but their potencies were still weak (3f, 5b,
6b). NTI exhibited weak κ receptor agonist activity and
compounds 3f, 5a, 6a, and 7 revealed the same or slightly higher
κ agonist potency than that of NTI.
selective and potent δ receptor antagonist, we speculated that the
bulkiness around the indole substructure may decrease the δ
receptor binding affinity, resulting in a decrease in δ receptor
antagonist activity. To examine this speculation, we evaluated
opioid receptor binding affinity of selected compounds (Table 3).
Binding affinity data of 3a34 and 3b25 were also cited from
references for the purpose of comparison. Coop et al. reported that
introduction of a methyl group to the 3-hydroxy group of 3a
increased δ selectivity in binding assays (Table 3, 3a-b).25
Compound 3f, which possesses additional alkyl substituents around
the indole substructure of compound 3b, showed a significant
decrease in δ receptor binding affinity and somewhat of a decrease
in δ selectivity in comparison with compounds 3a and 3b. The
antitussive activity of 3f may be induced by its active metabolites
(for example 3-OH analogue) rather than from 3f itself.
On the basis of these observations, we hypothesized that
compounds possessing a compact hydrophobic moiety around
the indole substructure would show high antitussive activity
without loss of δ receptor selectivity. As the connection of the
1′ position with the 7′ position by an alkyl chain is a possible
strategy to satisfy this hypothesis, we synthesized the additional
ring-fused derivatives 4-6 (Scheme 2).
Table 2 shows the results obtained from the evaluation of
NTI analogues possessing an extra fused ring structure. Com-
pounds with an alkylene bridge connecting the 1′ position and
the 7′ position (4, 5a-b, 6a-b) dramatically improved anti-
tussive activity compared with NTI. Particularly, the 3-methoxy
analogue 5b showed extremely high antitussive activity (its ED50
was 30-fold lower than that of NTI). As shown in Figure 1,
compound 5b decreased the number of coughs in a dose-
dependent manner. This potency increment requires further
investigation for these compounds, but the enhanced perme-
ability through the blood-brain barrier of these compounds,
which is assumed to result from increased hydrophobicity,33
could be counted as a key factor. Compound 7, another type of
extra ring fused NTI derivative possessing a butylene bridge
Among these compounds, the 6-membered fused-ring
derivative 5b, whose methanesulfonic acid salt is TRK-850,
showed excellent pharmacological properties. We evaluated
the antitussive effect of 5b by oral administration. As shown
in Table 2, 5b exhibited highly potent antitussive effect even
by oral administration, with an ED50 value of 6.40 µg/kg
(95% confidence limit: 1.54-26.5 µg/kg), which is ∼280
times smaller than that of peroral NTI. Increased hydropho-
bicity of compound 5b, which might improve its permeability

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Structure-AntitussiVe ActiVity of Naltrindole DeriVatiVes
Journal of Medicinal Chemistry, 2008, Vol. 51, No. 15 4407
Table 2. Antitussive Effecta and Opioid Antagonist Activityb of NTI and Its Analogues with an Extra Fused Ring Structure
antagonist activity
e,h
compd
R1
H
n
antitussive activityc,d,e ED50 (µg/kg)
9.14(4.23-19.8)
agonist
dose ratioe,f
conc (nM)g
pA2
4
1
DPDPE (δ)
50 (38-67)
30
9.2 (9.1-9.3)
7.8 (7.5-8.0)
8.1 (7.9-8.2)
9.10 (8.98-9.24)
7.90 (7.60-8.16)
7.56 (7.03-7.94)
8.20 (7.76-8.52)
j (j-6.75)
7.17 (6.94-7.35)
8.50 (8.16-8.83)
7.74 (7.22-8.04)
7.29 (j-7.87)
6.53 (j-7.58)
j (j-6.94)
morphine (µ)
U-50,488H (κ)
DPDPE (δ)
2.8 (2.1-3.7)
4.5 (3.5-5.9)
5a
5b
6a
6b
7
H
2
2
3
3
8.58(2.33-31.7)
38.5 (29.4-52.6)
3.39 (2.19-5.33)
2.08 (1.32-3.59)
17.0 (6.76-34.5)
0.92 (0.61-1.37)
2.48 (1.87-3.24)
4.18 (2.44-7.69)
1.55 (1.16-2.10)
1.19 (0.83-1.75)
1.03 (0.77-1.39)
0.76 (0.58-0.98)
1.58 (0.85-3.33)
5.28 (4.02-6.80)
0.3 (0.20-0.43)
1.06 (0.75-1.50)
30
100
10
morphine (µ)
U-50,488H (κ)
DPDPE (δ)
morphine (µ)
U-50,488H (κ)
DPDPE (δ)
morphine (µ)
U-50,488H (κ)
DPDPE (δ)
morphine (µ)
U-50,488H (κ)
DPDPE (δ)
Me
H
3.43(0.93-12.6)
6.40(1.54-26.5)i
12.3(3.92-38.3)
9.68(1.98-47.4)
39.5(12.8-121)
Me
10
7.77 (j-8.37)
7.63 (7.48-7.76)
j (j-j)
100
morphine (µ)
U-50,488H (κ)
5.78 (j-6.70)
a Evaluated using rat capsaicin-induced cough model. b Evaluated using electrically stimulated mouse vas deferens (MVD) preparations. c Each compound
was administered ip route unless otherwise noted. d ED50 values; the dose which reduces the number of coughs to 50% vs control, are expressed as mean
(N ) 8). e Figures in parentheses indicate 95% confidence limits. f Dose ratio, the ratio of agonist concentrations that elicit equal responses in the absence
and presence of the antagonist at increasing concentrations, are expressed as mean (N ) 4). g Concentration of competitive antagonist. h pA2 ) -log{
[B]/(dose ratio - 1)}. [B]: concentration of competitive antagonist (nM). i Administered perorally. j Not calculated.
Table 3. Opioid Receptor Binding Affinitiesa and Agonist Activitiesb of NTI Analogues
binding affinity
[
35S]GTPr.S bindingc
compd
Ki (nM)
selectivity
% responsed at 1 µMe
[3H]NTI (δ)
[3H]DAMGO (µ)
[3H]U-69,593 (κ)
7.44
30.4
3911
4650
7.99
15.1
3290
40.0
3343
52.0
µ/δ
κ/δ
δ
µ
κ
3a
3af
3bg
3f
0.135
0.217
6.57
12.7
27
2510
>3670
23.2
94
124
382
>17
79
90
>46
50
55
140
595
21
27
22
42
18
12
8.2
-3
5
12
h
47
26
-2
47
20
45
19
h
1
0
h
21
3
26
8
32
11
33
218
4
0.295
0.673
79.0
2.26
274
5a
5b
6a
6b
7
60.6
1
>3670
-1
-1
-1
-1
113
>3670
200
>13
32
6.32
a Evaluated by displacement of radioligands from guinea pig brain membranes. Data are expressed as mean (N ) 2). b Evaluated using [35S]GTPγS
binding assay. Data are expressed as mean (N ) 2). c Carried out using membrane preparations from transfected CHO (δ and µ) or HEK-293 (κ) cells that
constitutively expressed the respective human opioid receptor type. d Response relative to DPDPE (δ), DAMGO (µ), or U-69,593 (κ) response. EC50 values
of DPDPE, DAMGO, and U-69,593 were 16, 1.7, and 11 nM, respectively. e Concentration of each tested compound. f Binding affinity data were from ref
34. g Binding affinity data were from ref 25. h Not tested.
through the blood-brain barrier, could result in potent
antitussive effects at extremely low doses (3.43 µg/kg, ip;
6.40 µg/kg, po). As compound 5b also showed weak or no
opioid agonist activities, especially for µ and κ opioid
receptors, δ opioid receptor antagonist activity of 5b would
mainly induce its antitussive effect. However, it cannot be
denied that the antitussive effects may be derived from weak
opioid receptor agonist activities.
compound 5b would be derived from the δ receptor
antagonist activity.
Conclusions
In the present study, we designed NTI derivatives possessing
hydrophobic substituents based on the hypothesis that increased
hydrophobicity might improve their permeabilities through the
blood-brain barrier and thereby increase their antitussive
activities compared with that of NTI itself. Although introduc-
tion of some alkyl groups into NTI skeleton increased the
antitussive activities of these derivatives, it simultaneously
decreased the δ opioid receptor antagonist activity. Introduction
of an extra fused ring, a compact size moiety, at the indole ring
We also evaluated the effect of (-) TAN-67 (selective δ1
receptor agonist)35,36 on cough reflex using the rat capsaicin-
induced cough model. Administration of (-) TAN-67 (10
mg/kg, sc) markedly increased the number of coughs.37 The
results also suggested that antitussive activity induced by