Discovery of highly selective κ-opioid receptor agonists: 10α-Hydroxy TRK-820 derivatives

Article abstract of DOI:10.1016/j.bmcl.2017.06.017

κ-Opioid receptor agonists with high selectivity over the μ-opioid receptor are attractive targets in the development of drugs for pain and pruritus. We previously reported the synthesis of 10α-hydroxy TRK-820 (1). In this study, we elucidated the biologi

Full text of DOI:10.1016/j.bmcl.2017.06.017

Accepted Manuscript
Discovery of highly selective κ-opioid receptor agonists: 10α-hydroxy
TRK-820 derivatives
Shinya Suzuki, Yuji Sugawara, Riichiro Tsuji, Ryuji Tanimura, Chihiro
Kaneko, Natsumi Yuzawa, Mai Yagi, Koji Kawai
PII:
S0960-894X(17)30619-4
DOI:
http://dx.doi.org/10.1016/j.bmcl.2017.06.017
BMCL 25051
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
8 May 2017
1 June 2017
4 June 2017
Please cite this article as: Suzuki, S., Sugawara, Y., Tsuji, R., Tanimura, R., Kaneko, C., Yuzawa, N., Yagi, M.,
Kawai, K., Discovery of highly selective κ-opioid receptor agonists: 10α-hydroxy TRK-820 derivatives, Bioorganic
& Medicinal Chemistry Letters (2017), doi: http://dx.doi.org/10.1016/j.bmcl.2017.06.017
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Discovery of highly selective κ-opioid receptor agonists: 10α-hydroxy TRK-820 derivatives
Shinya Suzuki*, Yuji Sugawara, Riichiro Tsuji, Ryuji Tanimura, Chihiro Kaneko, Natsumi Yuzawa, Mai
Yagi, Koji Kawai
Pharmaceutical Research Laboratories, Toray Industries, Inc., 6-10-1 Tebiro, Kamakura, Kanagawa
248-8555, Japan
*Corresponding author:
Shinya Suzuki
Pharmaceutical Research Laboratories, Toray Industries, Inc., 6-10-1 Tebiro, Kamakura, Kanagawa
248-8555, Japan
Tel.: +81 467 32 9643
FAX: +81 467 32 9931
E-mail: Shinya_Suzuki@nta.toray.co.jp
Abstract
κ-Opioid receptor agonists with high selectivity over the µ-opioid receptor are attractive targets in the
development of drugs for pain and pruritus. We previously reported the synthesis of 10α-hydroxy
TRK-820 (1). In this study, we elucidated the biological properties of 1 and optimized its 6-acyl unit by
modifying our synthetic route. Among the 10α-hydroxy TRK-820 derivatives prepared, 26 showed the
most potent κ-opioid agonist activity (EC50 = 0.00466 nM) and excellent selectivity and 22 was the most
κ-selective agonist.

Keywords: µ-Opioid receptor, κ-Opioid receptor, Analgesic, Antipruritic, 4,5-Epoxymorphinan,
TRK-820
The µ-opioid receptor (MOR) is deeply involved in the analgesic effect of morphine, and thus various
MOR agonists are used as analgesics. However, MOR agonists cause drug dependence and respiratory
depression. Therefore, analgesics that avoid these adverse effects are strongly desired in the clinical
setting. The κ-opioid receptor (KOR) has attracted interest because its activation produces an analgesic
effect with minimal physical dependence and respiratory depression.1 Nalbuphine is a KOR agonist/MOR
antagonist used as an analgesic (Figure 1).2 In addition, KOR agonists are useful for treating pruritus. We
previously developed TRK-820 (nalfurafine hydrochloride: a 4,5-epoxymorphinan derivative that is a
KOR agonist/MOR partial agonist) as an antipruritic agent for uremic pruritus.3 TRK-820 is a strong
KOR agonist (EC50: 0.0082 nM) and moderate MOR agonist (EC50: 1.7 nM).4 Because these
4,5-epoxymorphinan KOR agonists affect also MOR to some extent, the intrinsic potential of a pure KOR
agonist as medication is not known. Pure KOR agonists are attractive targets and a challenging theme for
drug development. Accordingly, several pharmaceutical companies are conducting research on KOR
agonists; in particular, CR-845 is highly KOR-selective peptide that is currently in clinical trials for
treatment of pain by intravenous administration.5

Figure 1. Chemical structures of known κ-opioid receptor agonists.
We previously reported 10α-hydroxy TRK-820 (1), which was obtained as a degradation product in the
development of a TRK-820 formulation (Figure 2).6 We confirmed 1 has strong KOR agonist activity
(EC50: 0.025 nM) and excellent selectivity for KOR over MOR (44,000 fold). Therefore, we expected that
1 could serve as a starting point for developing a pure KOR agonist for use as an orally available drug. In
this letter, we report a synthetic method and the pharmacological properties of 10α-hydroxy TRK-820
derivatives.
Figure 2. Structure of 10α-hydroxy TRK-820 (1).
First, we tested the pharmacological effects of 1 in rats after partial sciatic nerve ligation (Figure 3).7 1
was administered 7 days after ligation. The anti-allodynic effect of 1 was evaluated at 30 min after
intravenous administration. Allodynia was assessed using the 50% withdrawal threshold measured
according to the up-down method with von Frey filaments. A significant reduction of allodynia was
observed in the 3 and 5 mg/kg groups compared with the vehicle group.

Figure 3. Anti-allodynic effect of 1 in rats after partial sciatic nerve ligation. Data are presented as mean
S.E.M. (n = 3-7). #: P < 0.05 (t-test); *: P < 0.025 (Shirley-Williams test, one side). i.v., intravenous.
We also tested the antipruritic effect of 1 on substance P (SP)-induced scratching behavior in mice
(Figure 4).8 1 was administered by intraperitoneal injection to mice, and 15 min later, SP (250 nmol/site)
was administered by intradermal injection into the nape of the neck. Immediately after the injection of SP,
the number of scratching behavior was recorded for 30 min by using MicroAct® (Neuroscience, Tokyo,
Japan). Scratching behavior was significantly inhibited in the 3 mg/kg groups compared with the vehicle
group.
Figure 4. Effects of 1 on substance P-induced scratching behavior in mice. Data are presented as mean
S.E.M. (n = 7–9). #: P < 0.05 (Welch’s test); *: P < 0.05 (Dunnett’s multiple comparisons). i.d.,
intradermal; i.p., intraperitoneal; PBS, phosphate-buffered saline.

In these in vivo experiments, 1 showed good anti-allodynic and antipruritic effects, both of which appear
promising for therapeutic use. To date, although no structure–activity relationship (SAR) study of
10α-hydroxy-4,5-epoxymorphinan has not been conducted, Portoghese and co-worker demonstrated that
the subtype selectivity of 4,5-epoxymorphinan could be controlled through the message-address
concept.9-11 Additionally, we reported that changing the address site of norbinaltorphimine (C6 position of
4,5-epoxymorphinan) to flexible a substituent based on the same concept afforded a KOR-selective
ligand.12 Therefore, we modified the 6-amide unit of 1 to clarify the SAR of
6-amide-10α-hydroxy-4,5-epoxymorphinans.
We have previously reported the synthesis of 1, which required multiple steps and introduction of the
10α-hydroxy group after the 6-acyl group.13 However, this route is not suitable for modifying the 6-amide
unit. Thus, we developed an effective intermediate for acylation of the 6-amino group in the last step.
First, we considered how to easily introduce the 10α-hydroxy group and then to derivatize the acyl
group in the last step. In our previous study, the 10-oxo compound was reduced with NaBH4 to
selectively give 10β-hydroxy compound, which was converted into the 10α-hydroxy compound by
mesylation and the subsequent SN2 inversion with sodium acetate and hydrolysis.13-14 Because of the
steric effect of the 17-cyclopropylmethyl group, the hydride’s approach from the α-face (front side) is
tightly restricted. Hence, we attempted to use ceric ammonium nitrate (CAN) as an oxidant. CAN induces
one-electron oxidization and provides a carbon radical. After the susequent one-electron oxidation of the
carbon radical, the resulting carbon cation reacts with water from the back side to provide a secondary
alcohol forms. Because of the steric hindrance of the tri-substituted benzylic carbon, we expected that
further oxidation would not occur. In our strategy, we selected naltalimide (TRK-130) as a starting
material.15-16 Thus, we established a kilogram-scale synthesis of TRK-130 and can easily derivatize the
6-amide part in the last step.
The 10α-hydroxy derivatives were synthesized by the route shown in Scheme 1. The phenolic hydroxyl
group of TRK-130 was protected with a benzyl group. Using CAN, oxidation of the 10α position
proceeded in moderate yield.17-19 Deprotection of the phthalimide by treatment with hydrazine
monohydrate gave amine 9. Two-step reductive amination by treatment with benzaldehyde followed by

formalin gave 10. Removal of the two benzyl groups in 10 gave N-methyl amine 11 as a common
intermediate. Finally, amidation with various carboxylic acids (A) or acyl chlorides (B) gave the desired
compounds.
Scheme 1. Synthetic route of 6-acyl derivatives.
To investigate of influence of the acyl (-C(=O)R) groups as the address moiety, we introduced various
phenyl-substituted acyl groups, focusing on sidechain length. Among the compounds with a C3 linker,
cinnamamide 18 and 3-phenylpropiolamide 19 showed KOR agonist activity and selectivity comparable
to those of 1, whereas 3-phenylpropanamide 15 showed moderate KOR agonist activity.
Phenylbutanamide 16 and phenoxyacetamide 17 also showed moderate KOR agonist activity.
Surprisingly, benzamide 13 and phenylacetamide 14 showed no agonist activity up to 100 nM toward
KOR and 30,000 nM toward MOR.20
Table 1. In vitro KOR and MOR agonist activities.

N
HO
OH
O
R
O
HO
N
Me
KOR
EC50 : nM
(Emax : %)
MOR
EC50 : nM
(Emax : %)
Yield of amidation
reaction (%)
No.
13
14
15
16
17
18
19
R
>100
(23)
>30,000
(44)
91%b)
91%b)
41%a)
56%a)
74%b)
40%a)
9%a)
>100
(18)
>30,000
(43)
1.04
(100)
18,300
(57)
7.51
(96)
12,900
(99)
O
5.04
(97)
>30,000
(17)
0.0222
(101)
209
(107)
0.0799
(102)
904
(95)
a) Synthesis method A
b)
Synthesis method B
KOR, -opioid receptor, MOR, -opioid receptor
We found that cinnamyl derivative 18 showed good KOR agonist activity in our initial screening.
Therefore, we next optimized the aryl part of 18. First, we introduced methyl (20), chloro (21), and
hydroxy (22) substituents at the para position (Table 2). These compounds showed the same or slightly
weaker KOR agonist activity as 1. On the other hand, whereas 20 and 21 showed moderate MOR agonist
activity, 22 showed markedly attenuated MOR agonist activity and greatly enhanced KOR selectivity. To
understand the SAR of the 10α-hydroxy derivatives listed on Tables 1 and 2, we constructed atomic
models of the MOR and KOR complexed with these derivatives.
Using these models, we investigated the interactions between the 10α-hydroxy group and the receptors.
In the crystal structure of the complex of MOR and the morphinan agonist BU72 (PBD ID 5c1m),21 there
is enough space to introduce the OH group at the 10α position of the bound agonist. The 10α-hydroxy

derivatives can bind to MOR in a manner similar to the corresponding unsubstituted parent molecules
without any clashes between the introduced OH group and the receptor atoms, and so these derivatives
retain their MOR agonist activity. The KOR and MOR have the same amino acid residues adjacent to the
10α position; thus, the 10α-hydroxy derivatives can also bind to KOR.22
We next examined why 22 has greatly attenuated MOR agonist activity. Comparison of the crystal
structures of the MOR complexed with a morphinan agonist BU72 (PBD ID 5c1m) and with a morphinan
antagonist β-FNA (PBD ID 4dkl) shows that the positions of transmembrane domain (TM) 1 are quite
different (Fig. 5).21, 23 In the MOR agonist binding form, the OH group of the sidechain of Tyr75 in TM1
and the nitrogen atom of the sidechain of His319 in TM7 form a hydrogen bond that keeps TM1 near the
ligand binding pocket shown as a green ribbon model (Fig. 6a). When 22 binds to the MOR, the 4-OH
group of 22 may form another hydrogen bond with His319, possibly breaking the hydrogen bond between
Tyr75 and His319 (Fig. 6b). We speculate that the loss of the hydrogen bond between TM1 and TM7
changes the position of TM1 from the agonist binding form to the antagonist binding form leading to the
disappearance of MOR agonist activity. On the other hands, there is enough space to locate the 4-OH
group of 22, and no influence on the relationship of TM1 and TM7 in KOR case (Fig. 6a, gray ribbon
model). For that reason, we consider that 22 showed strong KOR agonist activity.
Figure 5. Chemical structures of BU72 and β-FNA.

Figure 6a.
Figure 6b.
Figure 6. Docking model of 22 and MOR (green ribbon) or KOR (gray ribbon).
(6a) Ligand binding pocket viewed from the extracellular side. Extracellular loops are omitted from this
view.

(6b) Side view of the ligand binding pocket. The surface of the ligand binding pocket is shown in
blue mesh. Compound 22 is shown as a ball-and-stick model. The hydrogen bond between Tyr75 and
His319 in the MOR is shown as a red dotted line . The KOR (PDB ID 4dkl) superimposed on the MOR is
shown as a gray ribbon model.TM, transmembrane domain.
In addition, we replaced the benzene moiety with a heteroaromatic ring (Table 2). The KOR agonist
activity fluctuated according to the heteroatom position in pyridine, thiophene, and furan. In particular,
2-thienyl derivative 26 exhibited strong KOR agonist activity equal to that of TRK-820.
Table 2. In vitro KOR and MOR agonist activities.
In conclusion, we have demonstrated the biological properties of 1 revealing an analgesic and
antipruritic effect in mice. Based on these results, 1 has the possibility of being groundbreaking drug.
Moreover, we successfully developed a suitable synthetic route for 10α-hydroxy TRK-820 derivatives
and acquired information about the SAR of the 6-acyl unit in 10α-hydroxy TRK-820 derivatives. Among
them, 2-thienyl derivative 26 exhibited the highest KOR agonist activity and 22 showed the best KOR

selectivity. Further pharmacological investigations of 22 and 26 are in progress and will be reported in
due course.
References and notes
1. Riviere PJM. Br J Pharmacol. 2004;141:1331–1334.
2. Zeng Z, Lu J, Shu C, Chen Y, Guo T, Wu QP. Sci Rep. 2015;5:10927.
3. Kumada H, Miyakawa H, Muramatsu T, Ando N, Oh T, Takamori K, Nakamoto H. Hepatol Res.
2016;doi: 10.1111/hepr.12830.
4
Nakao A, Ikeda K, Kurokawa T, Togashi Y, Umeuchi H, Honda T, Okano K, Mochizuki H. The
Japanese Society of Neuropsychopharmacology. 2008;28:75–83.
5. Schteingart CD, Menzaghi F, Jiang G, Alexander RV, Sueiras-Diaz J, Spencer RH, et al., inventors;
Cara Therapeutics Inc, assignee. WO patent 2008057608, November 13, 2007.
6. Tuji R, Kawai K, Inada H, Suzawara Y, inventors; Toray Industries Inc, assignee. WO patent
2016152953A1, March 24, 2016.
7. Shir Y, Seltzer Z. Neurosci Lett. 1990;115:62–67.
8. Carlos B, Cynthia B, Oscar G F, Barry R K, Pharmacology Biochemistry and
Behavior.1989;34:491–495.
9. Portoghese PS, J Med Chem. 1991;34:1757–1762.
10. Portoghese PS, Sultana M, Takemori AE. J Med Chem. 1990;33:1714–1720.
11. Lipkowski AW, Tam SW, Portoghese PS. J Med Chem. 1986;29:1222–1225.
12. Nagase H, Hayakawa J, Kawamura K, Kawai K, Takezawa Y, Matsuura H, Tajima C, Endo T. Chem
Pharm Bull. 1998;46:366–369.
13. Horikiri H, Hirano N, Tanaka Y, Oishi J, Hatakeyama H, Kawamura K, et al. Chem Pharm Bull.
2004;52:664–669.
14. Horikiri H, Kawamura K. Heterocycles. 2004;63:865–870.
15. Fujimura M, Izumimoto N, Momen S, Yoshikawa S, Kobayashi R, Kanie S. J Pharmacol Exp Ther.
2014;350:543–551.

16. Izumimoto N, Kawai K, Kawamura K, Fujimura M, Komagata T, inventors; Toray Industries Inc,
assignee. WO patent 2004033457A1, October 8, 2003.
17. Gary A C, Spencer D Jr. Synthetic Communications. 2000;30:4513–4521.
18. Walter M, Gregory A N, Robert B, Robert C, Michael H, Ken S, Christopher A T. Tetrahedron
Letters. 2003;44:7381–7384.
19. Zezula J, Rice K C, Hudlicky T. Synlett. 2007;18:2863–2867.
20. As indicators of in vitro KOR and MOR agonist activities of compounds, we examined the inhibitory
effects on forskolin-induced cAMP accumulation in Chinese hamster ovary (CHO)-K1 and
CHO-dhfr(-) cells expressing the human KOR and MOR, respectively. Cellular concentration of
cAMP was measured using the LANCE cAMP 384 assay kit (Perkin Elmer), according to the
manufacturer’s instructions.
21. Weijiao H, Aashish M, Venkatakrishnan A J, Toon L, Evan N F, Adrian L S, Hideaki E K, Kathryn E
L, Thor S T, Ralf C K, Sebastien G, Peter G, Stephen M H, John R T, William I W, Jan S, Ron O D,
Brian K K. Nature. 2015;524:315–321.
22. Huixian W, Daniel W, Mauro M, Vsevolod K, Gye W H, Eyal V, Wei L, Aaron A T, Xi-Ping H,
Carroll F I, Mascarella S W, Westkaemper R B, Philip D M, Bryan L R, Vadim C, Raymond C S.
Nature. 2012;485:327–332.
23. Aashish M, Andrew C K, Tong S K, Foon S T, Jesper M M, Roger K S, Leonardo P, William I W,
Brian K K, Sebastien G. Nature. 2012;485:321–326.