The role of the side chain in determining relative δ- and κ-affinity in C5′-substituted analogues of naltrindole

Article abstract of DOI:10.1021/jm020997b

The role of the side chain in 5′-substituted analogues of naltrindole has been further explored with the synthesis of series of amides, amidines, and ureas. Amidines (8, 13) had greatest selectivity for the K receptor, as predicted from consideration of the message-address concept. It was also found that an appropriately located carbonyl group, in ureas (10) and amides (7), led to retention of affinity and antagonist potency at the δ receptor.

Full text of DOI:10.1021/jm020997b

3
14
J . Med. Chem. 2003, 46, 314-317
Th e Role of th e Sid e Ch a in in Deter m in in g Rela tive δ- a n d K-Affin ity in
C5′-Su bstitu ted An a logu es of Na ltr in d ole
†
‡
§
†
,†
Shannon L. Black, Andrew R. J ales, Wolfgang Brandt, J ohn W. Lewis, and Stephen M. Husbands*
Department of Pharmacy and Pharmacology, University of Bath, Bath, BA2 7AY, UK, School of Chemistry, University of
Bristol, Bristol, BS8 1TS, and Institute of Plant Biochemistry, Weinberg 3, D-06120 Halle, Germany
Received J uly 30, 2002
The role of the side chain in 5′-substituted analogues of naltrindole has been further explored
with the synthesis of series of amides, amidines, and ureas. Amidines (8, 13) had greatest
selectivity for the κ receptor, as predicted from consideration of the message-address concept.
It was also found that an appropriately located carbonyl group, in ureas (10) and amides (7),
led to retention of affinity and antagonist potency at the δ receptor.
In tr od u ction
The precise role of the kappa (κ) opioid receptor has
yet to be well established. However, interactions at this
receptor are known to create a number of physiological
1
effects, including analgesia, prevention of neurodegen-
2
eration in stroke and cerebral ischemia models, and
protection against epileptic convulsions.2 There is also
significant interest in the role of κ-agonists in cocaine
abuse, and in particular the findings that κ-agonists can
,3
4
block many of cocaine’s behavioral effects. While having
pharmacological effects of their own, κ-antagonists have
mostly been used as pharmacological tools to help define
the role of the κ-opioid system.
In recent years a number of κ-opioid antagonists have
been discovered, greatly extending our understanding
of the requirements for selective binding to this receptor.
5
-7
The majority of these ligands (e.g., 1 and 2)
are
8
related to the δ-antagonist naltrindole (3), as is the
most well-known and studied κ-antagonist, norBNI (4).
9
An exception is the recently reported 3-hydroxyphenyl-
piperidine, J DTic (5).10 κ-Selectivity in the ligands
related to naltrindole is reportedly derived from the
1
1
determining κ-selectivity in naltrindole derivatives hav-
ing both basic and nonbasic side chains and to this end
synthesized series of amidines, amides, and ureas. We
here report the synthesis and preliminary pharmaco-
logical characterization of these ligands and suggest
that while the presence of a basic group within the side
chain normally results in κ-selectivity, an appropriately
located carbonyl group can lead to retention of substan-
tial δ-antagonist activity and loss of selectivity.
presence of a basic, or cationic, group in the side chain.
At the κ-receptor this group is believed to interact with
Glu297, present at the top of transmembrane region 6,
while it is thought that negative interactions could occur
between the side chain and residues corresponding to
Glu297 in the µ and δ receptors.12 These are Lys303 (µ)
and Trp284 (δ) and may reduce binding affinity to the
µ- and δ-receptors through electrostatic repulsion
(Lys303) and steric hindrance (Trp284). Whatever the
Syn th esis. The amidines (8) were prepared as re-
exact binding interactions, it is accepted that an ap-
propriately located basic or cationic group is both
required, and sufficient, for κ-selectivity. In the series
6
6,11
ported previously for 8d . Thus amine (6)
was treated
with the appropriate ethyl alkylimidate hydrochloride,
itself prepared from the alkylnitrile by treatment with
HCl in dry ethanol. The products were purified by
preperative TLC on silica gel plates. Yields ranged from
5
of amidines reported by Olmsted et al. it is interesting
to note that selectivity for the κ-receptor is also related
to the length of alkyl chain side chain attached to the
amidine group. Thus the n-butyl analogue is more
selective than its n-propyl or ethyl congeners. We were
interested in exploring the role of lipophilic binding in
1
5 to 24% with a major side product for 8b, 8c, and 8d
being the equivalent amide (7a -c: 20-30%) (Scheme
). The formation of amides in the reaction appeared to
1
be due to hydrolysis of the alkylimidate. The ureas (10)
were obtained by treating amine (6) with the appropri-
*
Corresponding author. Tel: (44) 1225 383103. Fax: (44) 1225 826
1
14. E-mail: S.M.Husbands@bath.ac.uk.
16
ate isocyanate (Scheme 1). This led mainly to the
†
University of Bath.
University of Bristol.
Institute of Plant Biochemistry.
‡
desired ureas, but also gave approximately 10% of the
urethane derivatives (9). Base-promoted hydrolysis to
§
1
0.1021/jm020997b CCC: $25.00 © 2003 American Chemical Society
Published on Web 12/10/2002

Brief Articles
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 2 315
Sch em e 1a
a
(
i) RC(dNH‚HCl)OEt, EtOH; (ii) RNCO, CH2Cl2/MeOH (10:1), rt, 5 h; (iii): K2CO3 (5 equiv), MeOH/H2O (9:1), rt, 12 h; (iv): PhSH,
HBr, MeOH; (v) H2NR (vi) BOP, NEt3, CH2CH2, rt, 24 h; (vii) EDCl, DMAP, NEt3, RNH2, CH2Cl2, rt, 12 h.
Ta ble 1. Antagonist Activities in [35S] GTPγS Assays in
1
0 with potassium carbonate gave an overall yield of
Human Recombinant Receptors in CHO Cells
approximately 50%. Preparation of the reverse amidines
Ke (nM) ( SEM
(13) utilized the thiophenyl imidic ester (12) prepared
6
,11
µ-CHO
membranes membranes membranes
compd DAMGO DPDPE U69,593
δ-CHO
κ-CHO
by treating nitrile (11)
with thiophenol in methanol
under acidic conditions (Scheme 1).1 Addition of the
3
n
µ/κ δ/κ
appropriate amine to crude 12 gave the required
amidines that were purified by preparative thin-layer
chromatography, in 20-30% yield. Also isolated were
disubstituted amidines (14) in 5-10% and unreacted
nitrile (10%). Amide (16a ) was prepared from the known
7
7
7
8
8b
8c
8
8
1
10b
10c
1
1
1
1
1
16c
16d
16e
a
b
c
a
4.94 ( 1.08 0.38 ( 0.05 0.48 ( 0.27
3.17 ( 0.34 0.30 ( 0.10 0.35 ( 0.23
3.37 ( 1.22 0.23 ( 0.04 0.46 ( 0.14
3.78 ( 0.68 1.79 ( 0.69 0.21 ( 0.04
4.70 ( 1.34 1.77 ( 0.25 0.24 ( 0.03
4.21 ( 1.61 1.89 ( 0.33 0.18 ( 0.06
5.33 ( 0.63 3.31 ( 0.54 0.17 ( 0.05
14.73 ( 0.83 5.23 ( 0.13 0.32 ( 0.02
1.60 ( 0.15 0.65 ( 0.02 2.47 ( 0.20
1.63 ( 0.12 0.53 ( 0.08 1.52 ( 0.16
1.79 ( 0.32 1.04 ( 0.18 1.71 ( 0.16
3.19 ( 0.25 4.41 ( 0.79 0.05 ( 0.004
5.61 ( 0.28 3.83 ( 0.40 0.21 ( 0.03
2.04 ( 0.62 5.83 ( 0.42 0.37 ( 0.06
6.86 ( 0.93 6.95 ( 0.86 0.29 ( 0.08
4.40 ( 0.74 2.99 ( 0.22 0.73 ( 0.04
2.70 ( 0.31 1.21 ( 0.05 0.17 ( 0.03
2.78 ( 0.21 5.15 ( 0.25 0.26 ( 0.02
0.94 ( 0.12 6.20 ( 0.49 0.28 ( 0.04
4
4
5
6
4
4
5
5
4
4
4
4
5
5
5
5
6
5
8
10
9
7
18
20
23
31
46
1
1
0.5
9
7
11
20
16
acid (15)20 by BOP promoted coupling14 with heptyl-
d
e
0a
amine in 83% yield (Scheme 1). Under the same condi-
tions the benzyl and arylalkylamines (9b-e) gave <10%
yield. Improved, although still low, yields of ∼20% were
0.6 0.3
1
1
0.3
0.6
88
18
16
24
4
7
20
22
1
5
obtained in these cases with the use of EDCI/DMAP.
3a
3b
3c
6a
6b
64
27
6
24
6
16
10
3
Resu lts a n d Discu ssion
Opioid antagonist activity was determined by stimu-
3
5
lation of [ S]GTPγS in cloned human opioid receptors
transfected into Chinese hamster ovary (CHO) cells
1
7,18
(
[
Table 1).
None of the compounds stimulated
4
,
18.9 ( 1.8
4.42 ( 0.38 0.04 ( 0.004
484 113
3
5
S]GTPγS binding for any type of opioid receptor but
norBNI
were found to be antagonists of the selective agonists,
DAMGO (µ), Cl-DPDPE (δ), and U69,593 (κ) (Table 1).
The ligands were also evaluated in binding assays in
CHO cells transfected with cloned human opioid recep-
functional assay data for some of the compounds (e.g.,
d and 13a ). Previously an inverse relationship between
8
1
7
3
selectivity in binding and selectivity in functional
tors (Table 2). The displaced radioligands were [ H]-
potency has been noted for related compounds.10
3
3
DAMGO (µ), [ H]-Cl-DPDPE (δ), and [ H]-U69,593 (κ).
Binding affinity data for 8d has previously been re-
It is apparent from previous reports by, in particular,
6
11
ported. Differences were noted between the binding and
Portoghese’s group, that affinities for the three opioid

3
16 J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 2
Brief Articles
Ta ble 2. Receptor Binding to Recombinant Human Opioid
Receptors Transfected into CHO Cells
has a C5′-side chain of 9 atoms in total length, only one
longer than in 7c, yet 16a is a κ-selective antagonist
while 7c has slightly higher affinity for the δ-receptor
than for κ. This suggested that location of the amide
bond in the side chain could play a significant role in
determining δ-affinity. It was hypothesized that the
carbonyl of 7 might be hydrogen bonding to a residue
on the δ-receptor, resulting in the retention of affinity.
With a series of ureas (10) it was possible to test this
hypothesis. The urea moiety is basic, and thus should
be detrimental to δ-affinity, but also contains a carbonyl
group in the same location as amides 7, which if the
hypothesis was correct, would enhance δ-affinity. The
ureas did in fact retain δ-antagonist potency and affinity
with Keδ’s of 0.5-1 nM. Particularly striking is the
comparison of the ureas with the amidines (8). For
example, urea 10a and amidine 8c have the same
overall length of side chain and thus any change in
profile must be due to the nature of the basic group. In
the GTPγS functional assay 8c is 23-fold and 11-fold
selective for κ over µ- and δ-receptors, respectively, urea
Ki (nM) ( SEM
µ
δ
κ
compd [ H]-DAMGO [ H]-DPDPE [3H]U69,593 na µ/κ δ/κ
3
3
7
7
7
8
8
8
8
8
1
1
1
1
1
1
1
1
1
1
1
4
a
b
c
a
b
c
d
e
0a
0b
0c
3a
3b
3c
6a
6b
6c
6d
6e
,
22.97 ( 11.10 3.12 ( 0.50
1.57 ( 0.80
0.85 ( 0.40
0.68 ( 0.30
0.29 ( 0.10
0.28 ( 0.10
0.25 ( 0.10
0.30 ( 0.20
1.39 ( 0.14
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
15
36
2
4
30.97 ( 0.10
31.59 ( 0.59
3.70 ( 0.34
7.67 ( 0.64
46 11
77 74
94 78
163 110
730 127
34 14
22.32 ( 2.27 21.38 ( 1.32
26.35 ( 0.46 21.76 ( 1.50
40.79 ( 6.32 27.56 ( 2.11
219.16 ( 84.47 38.22 ( 4.91
47.40 ( 7.07 20.10 ( 4.29
b
37.64 ( 14.41 2.43 ( 0.46 12.32 ( 1.29
3
0.2
4.80 ( 0.91
13.58 ( 4.06
13.48 ( 0.54
2.60 ( 0.54
2.25 ( 0.16
5.29 ( 0.27
6.33 ( 0.40
8.13 ( 2.67
1.60 ( 0.28
1.44 ( 0.04
5.61 ( 0.37
0.8 0.4
2
8
18 12
10
3
4
11
6
0.3
3
25.29 ( 4.06 17.02 ( 7.19
56.62 ( 7.69 7.33 ( 1.13
1
3
1
2
10
4
61.91 ( 6.03 70.15 ( 33.93 21.89 ( 7.11
43.95 ( 8.27 11.19 ( 2.86 10.33 ( 0.66
23.53 ( 7.32
3.53 ( 1.24
2.21 ( 0.35
6.18 ( 0.47
2.11 ( 1.04
0.20 ( 0.05
37.05 ( 9.45 59.45 ( 4.82
4.64 ( 0.58
21.0 ( 5.0
7.41 ( 0.59
5.7 ( 0.9
2
105 28
1
0a is slightly (2-4-fold) selective for the δ-receptor over
norBNI
µ and κ. In binding, this change in selectivity was even
more pronounced with 8c being >100-fold selective for
κ- over µ- and δ-receptors, while 10a was 5-fold and 15-
fold selective for δ- over κ- and µ-receptors, respectively.
Thus replacement of the amidine moiety by the urea
group results in a significant shift in selectivity from κ
to δ. When considered in conjunction with the data for
amides 7 and 16 this appears to confirm that the
carbonyl group is a key binding motif in retaining
δ-activity. It is interesting to note that the δ-selective
fluorogenic affinity label PNTI (17) has been postulated
a
All values are the average of two experiments, each carried
b
out in triplicate. Data from ref 6.
receptors do not show large changes on alteration of the
side chain group in C5′-substituted analogues of nal-
trindole. Changes in antagonist activity in vitro appear
to be quite subtle, and this is also found in the current
study. Nearly all the compounds synthesized had sub-
nanomolar Ke’s for the κ-receptor and low nanomolar
Ke’s at the µ and δ-receptors. However, the effect of the
small changes observed on antagonist selectivity were
significant. Thus selectivities in the GTPγS assay
ranged from 64-fold and 88-fold selectivity for κ over µ
and δ receptors (13a ) to 2-4-fold selectivity for δ over
µ and κ receptors (10a ). For amidines (8), lengthening
the R group caused some reduction in antagonist
potency at µ and δ-receptors, with no change at κ,
resulting in increased κ-selectivity. An identical effect
was observed in the binding assays, with 8d proving
the most κ-selective compound in this assay.
For the reverse-amidines (13) an increase in length
of chain had little effect on µ or δ-antagonist activity in
the GTPγS assay, but slightly lowered κ-antagonist
activity and hence selectivity. In this assay 13a was, in
fact, the most potent and selective κ-antagonist tested
within these series. Amides (7a -c, 16a -e) were also
found to be potent antagonists of the κ-opioid receptor,
again having Ke’s in the subnanomolar range. At the µ
receptor, the amides typically had low nanomolar Ke’s,
resulting in selectivity of 1 order of magnitude for κ. At
neither the κ nor µ receptors was there any consistent
SAR with the length of side chain, location of the amide
bond, or type of side chain (aliphatic versus aromatic)
having any significant influence on the antagonist
activity of these ligands. It was at the δ-receptor that
the greatest interest was found. Amides 7a -c possessed
subnanomolar antagonist potency for the δ-receptor,
resulting in no κ/δ selectivity, whereas amides 16a -e
had low nanomolar Ke’s at δ, meaning they were
somewhat κ selective. The higher δ-antagonist potency
of amides (7) compared to amides (16) was somewhat
surprising, especially when comparing 16a to 7c. 16a
to interact with Lys214.19 Similarly, preliminary mo-
lecular modeling studies (data unpublished) suggest the
possibility of a hydrogen bond between the carbonyl of
the urea moiety and Lys214. Thus, rather than the side
chain having a negative interaction with Trp284, this
positive interaction could lead to the retention in affinity
observed experimentally.
Con clu sion s
The influence of side chain length and lipophilicity
appears to vary between series. Thus an increase in
chain length was detrimental to the selectivity of
amidines (13a -c) yet beneficial for 8a -d . While it is
clear that in most circumstances the presence of a basic
side chain, attached to C5′ of the naltrindole nulcleus,
leads to κ-selective antagonists, it is now apparent that
an appropriately positioned carbonyl group can negate
this effect and result in retention of affinity for the
δ-receptor. Furthermore it is suggested that this effect
may be the result of hydrogen bonding between the
carbonyl group and Lys214 of the δ-receptor.

Brief Articles
J ournal of Medicinal Chemistry, 2003, Vol. 46, No. 2 317
Ack n ow led gm en t. This work was supported by
(9) Portoghese, P. S., Lipkowski, A. W., and Takemori, A. E.
Binaltorphimine and nor-binaltorphimine, potent and selective
κ-opioid receptor antagonists. Life Sci. 1987, 40, 1287-1292.
(10) Thomas, J . B., Atkinson, R. N., Rothman, R. B., Fix, S. E.,
Mascarella, S. W., Vinson, N. A., Xu, H., Dersch, C. M., Lu, Y.-
F., Cantrell, B. E., Zimmerman, D. M., and Carroll, F. I.
Identification of the first trans-(3R,4R)-dimethyl-4-(3-hydroxy-
phenyl)piperidine derivative to possess highly potent and selec-
tive opioid κ receptor antagonist activity. J . Med. Chem. 2001,
44, 2687-2690.
(11) Stevens, W. C., J ones, R. M., Subramanian, G., Metzger, T. G.,
Ferguson, D. M., and Portoghese, P. S. Potent and selective
indolomorphinan antagonists of the kappa-opioid receptor. J .
Med. Chem. 2000, 43, 2759-2769.
(12) Portoghese, P. S. From models to molecules: Opioid receptor
dimers, bivalent ligands, and selective opioid receptor probes.
J . Med. Chem. 2001, 44, 2259-2269.
(13) Baati, R., Gouverneur, V., and Mioskowski, C. An improved
method for the preparation of amidines via thiophenylimidic
esters. Synthesis 1999, 6, 927-929.
(14) Castro, B., Dormoy, J . R., Evin, G., and Selve, C. Reactifs de
couplage peptidique IV (1) - L′hexafluorophosphate de benzo-
triazolyl N-oxytrisdimethylamino phosphonium (BOP). Tetra-
hedron Lett. 1975, 14, 1219-1222.
(15) Desai, M. C., and Stephens Stramiello, L. M. Polymer-bound
EDC (P-EDC) - a convenient reagent for formation of an amide
bond. Tetrahedron Lett. 1993, 34, 7685-7688.
(16) Satchell, D. P. N., and Satchell, R. S. Acylation by ketenes and
isocyanates. A mechanistic comparison. Chem. Soc. Rev. 1975,
4, 231-250.
(17) Toll, L., Berzetei-Gurske, I. P., Polgar, W. E., Brandt, S. R.,
Adapa, I. D., Rodriguez, L., Schwartz, R. W., Haggart, D.,
O’Brien, A., White, A., Kennedy, J . M., Craymer, K., Farrington,
L., and Auh, J . S. Standard binding and functional assays
related to (NIDA) Medications Development Division testing for
potential cocaine and narcotic treatment programs. NIDA Res.
Monogr. Series. 1998, 178, 440-466.
(18) Traynor, J . R., and Nahorski, S. R. Modulation by mu-opioid
agonists of guanosine-5′-O-(3-[S-35]thio)triphosphate binding
to membranes from human neuroblastoma SH-SY5Y cells. Mol.
Pharmacol. 1995, 47, 848-854.
NIDA Grants DA 00254 and DA 07315 and ligand
3
5
binding and [ S]GTPγS assays carried out by NIDA-
OTDP.
Su p p or tin g In for m a tion Ava ila ble: Procedures for the
synthesis of all compounds, including spectroscopic data. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Refer en ces
(
(
(
1) J ones, R. M., and Paterlini, M. G. κ-Opioid receptors: recent
advances and implications for drug design. Curr. Opin. Drug
Discovery Dev. 1998, 1, 175-182.
2) Tortella, F. C., and DeCoster, M. A. Kappa opioids: Therapeutic
considerations in epilepsy and CNS injury. Clin. Neuropharm.
1
994, 17, 403-416.
3) De Sarro, G., Trimarchi, G. R., Sinopoli, S., Masuda, Y., and De
Sarro, A. Anticonvulsant effects of U54494A and U50488H in
genetically epilepsy-prone rats and DBA/2 mice: A possible
involvement of glycine/NMDA receptor complex. Gen. Pharma-
col. 1993, 24, 439-437.
(
4) (a) Archer, S., Glick, S. D., and Bidlack, J . Cyclazocine revisited.
Neurochem. Res. 1996, 21, 1369-1373. (b) Heidbreder, C. A.,
Babovic-Vuksanovic, D., Shoaib, M., and Shippenberg, T. S.
Development of behavioural sensitisation to cocaine: Influence
of kappa opioid receptor agonists. J . Pharmacol. Exp. Ther. 1995,
2
75, 150-163. (c) Spealman, R. D., and Bergman, J . Opioid
modulation of the discriminative stimulus effects of cocaine:
Comparison of mu, kappa and delta agonists in squirrel monkeys
discriminating low doses of cocaine. Behav. Pharmacol. 1994,
5
, 21-31.
(
(
(
(
5) Olmsted, S. L., Takemori, A. E., and Portoghese, P. S. A
remarkable change of opioid receptor selectivity on the attach-
ment of a peptidomimetic κ address element to the δ antagonist,
naltrindole. J . Med. Chem. 1993, 36, 179-180.
6) J ales, A. R., Husbands, S. M., and Lewis, J . W. Selective κ-opioid
antagonists related to naltrindole. Effect of side-chain spacer
in the 5′-amidinoalkyl series. Bioorg. Med. Chem. Lett. 2000,
(19) Le Bourdonnec, B., El Kouhen, R., Poda, G., Law, P. Y., Loh, H.
H., Ferguson, D. M., and Portoghese, P. S. Covalently induced
activation of the δ opioid receptor by a fluorogenic affinity label,
7′-(Phthalaldehyde-carboxamido)naltrindole (PNTI). J . Med.
Chem. 2001, 44, 1018-1020.
1
0, 2259-2261.
7) J ones, R. M., Hjorth, S. A., Schwartz, T. W., and Portoghese, P.
S. Mutational evidence for a common kappa antagonist binding
pocket in the wild-type kappa and mutant mu[K303E] opioid
receptors. J . Med. Chem. 1998, 41, 4911-4914.
8) Portoghese, P. S. The design of δ-selective opioid receptor
antagonists. Il Farmaco 1993, 48, 243-251.
(20) Portoghese, P. S., and Olmsted, S. L. US Patent 5457208, 1995.
J M020997B