New scaffolds in the development of Mu opioid-receptor ligands

Article abstract of DOI:10.1016/S0960-894X(03)00194-X

A new class of μ selective receptor antagonists has been developed using a combinatorial approach based on previously reported Dmt-Tic dipeptide ligands. Modified tetrahydroisoquinoline (Tiq) residues were reacted with different electrophiles in order to

Full text of DOI:10.1016/S0960-894X(03)00194-X

Bioorganic & Medicinal Chemistry Letters 13 (2003) 1585–1589
New Scaffolds in the Development of Mu Opioid-Receptor Ligands
Daniel Page,a,* Natalie Nguyen,a Sylvain Bernard,a Martin Coupal,b Mylene Gosselin,b
a
Julie Lepage,b Lynda Adamb and WilliamBrown
aDepartment of Chemistry, AstraZeneca R&D Montreal, 7171 Frederick-Banting, Saint-Laurent, Quebec, Canada H4S 1Z9
bDepartment of Molecular Pharmacology, AstraZeneca R&D Montreal, 7171 Frederick-Banting, Saint-Laurent, Quebec,
Canada H4S1Z9
Received 2 December 2002; accepted 23 January 2003
Abstract—A new class of m selective receptor antagonists has been developed using a combinatorial approach based on previously
reported Dmt-Tic dipeptide ligands. Modified tetrahydroisoquinoline (Tiq) residues were reacted with different electrophiles in
order to create novel molecules that would mimic the original dipeptide. A specific class of thioureas bearing basic pyrrolidine
residues were shown to give good binding affinities. Further alkylation of the pyrrolidine ring with benzyl derivatives also proved to
increase the m binding affinity. In addition, it was demonstrated that m binding was enhanced by the presence of polar groups
around the benzyl ring having hydrogen-bonding character (donor/acceptor). This new class of ligands represents a novel scaffold
in the development of opioid analogues.
# 2003 Elsevier Science Ltd. All rights reserved.
The development of selective agonists and antagonists
toward the d, m and k-opioid receptors has always been
slowed down by the lack of knowledge about the
geometry of their binding pocket. While most opioid
agonists1 have mainly been used in pain-relieving pro-
cesses,2 the opioid antagonists3 could have potential
applications in the treatment of different drug addic-
tions, such as cocaine.4 However, their complete
mechanisms of action are not fully understood since the
direct evaluation of the ligand-binding site interactions
for such G-coupled protein receptors is rather compli-
cated. One way in which to investigate the structure of
these receptors is to look at their binding pocket with a
specific class of ligands even though the structural
diversity of the reported opioid ligands could suggest a
relative degree of plasticity for these receptors.
pharmacophore have been reported such as N-alkyl-
ations,8 substitutions around the aromatic ring of
Tic9,10 or the replacement of Tic by various heteroaro-
matic residues,11,12 which all produced altered pharma-
cological profiles for d and m. Recent studies involving
C-terminus modifications demonstrated that the inser-
tion of large hydrophobic residues could further alter
the bioactivity of these ligands, producing ligands with
dual activity profiles at the opioid receptors.13À17 More
specifically, previous work reported by our group14
showed that the insertion of t-butyl or phenyl groups
with a urea/thiourea linker at the C-terminus could
result in ligands possessing d or m agonism.
Some modifications around the Dmt moiety of the
pharmacophore have also been reported. A few repor-
ted substitutions around the Dmt (or Tyr) aromatic ring
were shown to lead mostly in decreased activity.18
Another more drastic change involved the substitution
of the Dmt N-terminal amine by a methyl group19,20
and other hydrophobic substituents21 that, when incor-
porated into peptides, still managed to produce ligands
with potent opioid antagonist properties. However, to
our knowledge, no complete substitution of the Dmt
moiety has been put forward yet. The introduction of
different nuclei could be a starting point for the devel-
opment of new types of opioid analogues and could give
us further insight into the binding domain of opioid
receptors. Incorporation of the C-terminus substituents
mentioned above14 could also benefit those ligands in
One particular class of reported opioid ligands was that
containing the Dmt-Tic5 (Dmt: 20,60-dimethyltyrosine;
Tic: 1,2,3,4-tetrahydroisoquinoline-3-carboxylic acid)
pharmacophore. This synthetic dipeptide was the direct
result of decades of studies around truncated enkepha-
lin peptides. The introduction of the Dmt moiety in
various opioid ligands produced different analogues
with enhanced binding affinities toward mostly the d
and m receptors.5À7 Several modifications around this
*Corresponding author. Tel.: +1-514-832-3200; fax: +1-514-832-
3232; e-mail: daniel.page@astrazeneca.com; http://www.astrazeneca.
ca/E/montreal
0960-894X/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0960-894X(03)00194-X

´
D. Page et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1585–1589
1586
Interesting hits were obtained particularly in the
thiourea class of compounds (Table 1). The presence of
a 2-tetrahydrofuran (9) moiety resulted in improved
binding affinities for all three opioid receptors. When
the corresponding isomers (10 and 11) were synthesized,
the R enantiomer was shown to bring the d binding
affinity to 53 nM, but the analogous urea derivative (12)
lost substantial d binding affinity. Subsequent replace-
ment of the tetrahydrofuran by a pyrrolidine ring (13)
increased the m-binding affinity while slightly decreasing
d. Since this pyrrolidine moiety introduced a basic
nitrogen, it appeared to be a more promising scaffold in
terms of possible chemistry, and was therefore used as
an anchor for the subsequent synthesis of our new class
of Tiq derivatives.
Scheme 1. (i) (a) Isobutyl chloroformate, NMM, DME, 0 ꢀC, 5 min;
(b) NaBH4, H2O, 0 ꢀC, 15 min; (ii) PPh3, DEAD, DPPA, THF, 0 ꢀC to
rt, 4 h; (iii) H2, 10% Pd/C, MeOH, 12 h; (iv) RNCO, TEA, DCM, rt, 2
h; (v) 1M HCl/AcOH, rt, 1 h.
terms of increased binding affinities. Therefore, we
report herein the synthesis and pharmacological
characterization of a new series of ligands based on the
original Dmt-Tic pharmacophore where C-terminus
modified tetrahydroisoquinoline (Tiq) residues were
reacted with different electrophiles used as potential
Dmt replacements.
Different analogues were prepared according to Scheme
3. The C-terminus modified Tiq ureas 7 or 8 were reac-
ted with the 2-Boc-pyrrolidine isothiocyanate that was
generated in situ fromthe corresponding 2- R-(amino-
methyl)-1-Boc-pyrrolidine and thiophosgene to form
compounds 14 and 15. Removal of the Boc groups
resulted in amines 16 and 17, which were then either
alkylated or acylated to give products 18–24 (Table 2).
A combinatorial synthesis approach was used in order
to generate an initial library of about 500 compounds.
The C-terminus modified Tiq derivative could easily be
prepared from commercially available Boc-l-Tic (1)
(Scheme 1). The carboxylic acid was reduced to the
corresponding alcohol (2) using a previously reported
mixed anhydride procedure.22 The alcohol (2) was then
transformed to the azide (3) using standard Mitsunobu
conditions. Catalytic hydrogenation afforded the corre-
sponding amine (4), which was reacted directly with
either t-butylisocyanate or phenyl isocyanate to obtain
the Boc-protected urea derivatives 5 and 6. Removal of
the nitrogen protecting-group under acidic conditions
afforded the final tetrahydroisoquinoline derivatives 7
and 8.
Fromthese results, it was determined that any substitu-
tion on the pyrrolidine nitrogen was detrimental to d
binding but an aromatic substitution (20) increased
m-binding affinity (Ki=23.2 nM). Furthermore, it
appeared that the basic nitrogen was required to retain
m-binding affinity as derivatives 22 and 23 bearing the
benzoate and benzyl urea moieties respectively, lost m
affinity. This observation is consistent with most of the
previously reported opioid ligands. However, the d and
k binding affinities were comparable to those observed
for the receptor m. It was observed that the chirality of
the pyrrolidine did not seemto influence greatly the m
binding since the corresponding S isomer of 24 showed
an almost identical m affinity of 31.3 nM. These results
prompted us to focus on this specific class as m ligands.
The Tiq derivative bearing a phenyl urea was chosen as
a template since our previous work14 revealed this group
to be a preferred motif for m-binding. This Tiq-pyrrol-
idine derivative was reacted with different aldehydes
under reductive amination conditions, affording
The t-butyl Tiq analogue (7) was then reacted with a
variety of commercially available electrophiles such as
acids, acid chlorides, isocyanates, isothiocyanates and
aldehydes to generate libraries of dipeptide analogues
that included scaffolds such as aromatic, heteroaro-
matic, alkanes or cycloalkanes (Scheme 2). Most of the
compounds generated this way could be obtained in
>85% purity after simple washings and without any
further purification. Their opioid receptor binding pro-
files were then determined. All tests were done at least in
triplicate.
Table 1. Binding affinities of generated hits to opioid receptors
Compd
X
Y
*
m Ki (nM)
d Ki (nM)
k Ki (nM)
9
O
O
O
O
NH
S
S
S
O
S
R/S
R
S
R
R
685Æ127
303Æ26
>1000
308Æ12
52.7Æ3.6
>1000
481Æ84
511Æ72
>1000
>1000
>1000
10
11
12
13
Scheme 2. (i) RCOCl, TEA, DCE, rt, 24 h; (ii) RCOOH, HATU,
DIPEA, DMF, rt, 24 h; (iii) RNCO/RNCS, TEA, DCE, rt, 24 h; (iv)
.
RCHO, THF, BH3 pyr, rt, 24 h.
392Æ94
649Æ28
85.5Æ5.4
159Æ19

´
D. Page et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1585–1589
1587
Table 3. Mu binding affinities of tetrahydroisoquinoline-pyrrolidine
derivatives
Compd
R
m Ki (nM) d Ki (nM) k Ki (nM)
DAMGO
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
0.53Æ0.01
31.1Æ3.7
11.7Æ1.5
5.2Æ0.9
290Æ82
—
—
140Æ19
113Æ17
509Æ28
—
—
—
—
—
—
—
—
—
—
—
—
—
>1000
—
Ph
2-OH-Ph
3-OH-Ph
—
194Æ52
Scheme 3. (i) 2-R-(aminomethyl)-1-Boc-pyrrolidine, CSCl2, DIPEA,
rt, 1 h; (ii) 1 M HCl/AcOH, rt, 1 h; (iii) RCHO, NaBH(OAc)3, AcOH,
THF, rt, 12 h; (iv) RCOCl, TEA, DCM, rt, 1 h; (v) RNCO, TEA,
DCM, rt, 1 h.
4-OH-Ph
6.0Æ0.2
530Æ79
2-OMe-Ph
3-OMe-Ph
4-OMe-Ph
2-CN-Ph
3-CN-Ph
4-CN-Ph
2-F-Ph
3-F-Ph
4-F-Ph
8.8Æ0.4
109Æ15
34.3Æ8.5
22.8Æ6.2
14.8Æ3.3
50.4Æ9.4
33.7Æ7.8
40.5Æ6.7
56.0Æ10.1
26.2Æ4.6
10.3Æ1.2
62.8Æ9.4
22.7Æ1.3
24.0Æ3.9
31.1Æ6.0
2.7Æ0.8
—
—
—
—
—
—
—
—
—
—
—
—
—
compounds 25–46 (Scheme 3).23 The results are sum-
marized in Table 3. The d and k binding affinities were
included for compounds having Ki values for m <10 nM
in order to assess the relative selectivity.
4-NHAc-Ph
2-Naphthalene
3-Furan
3-Thiophene
2-Pyridine
2,3-di-OMe-Ph
2-OH, 3-OMe-Ph
Fromthese results, it was shown that the introduction
of polar groups, especially those with hydrogen bonding
character (donor/acceptor) such as OMe and OH,
greatly increased the binding affinities toward the m
receptor. Position 4 of the aromatic ring seemed speci-
fically sensitive to hydrogen-bond donating groups,
such as those in compounds 27 and 37 bearing phenol
and N-acetate groups, which possessed binding affinities
of 6.0 and 10.3 nM, respectively. The best ortho sub-
stituent was the OMe group (compound 28, Ki=8.8
nM). Other aromatic systems such as naphthalene (38),
113Æ12
263Æ19
—
—
127Æ9
39.5Æ12.5
543Æ140
—
—
130Æ9
6.8Æ1.8
3-OH, 4-OMe-Ph 17.7Æ4.2
3,4-di-OH-Ph
2-OMe, 4-OH-Ph
21.4Æ1.4
1.1Æ0.1
furan (39), thiophene (40), and pyridine (41) did not
greatly influence the binding. It seems that anything
planar, but not too bulky, with an aromatic-type char-
acter can be accommodated in the binding pocket of the
receptor. The best result was obtained with a combi-
nation 2-OMe and 4-OH groups (46), increasing the m
binding affinity to 1.1 nM.
Table 2. Binding affinities of tetrahydroisoquinoline-pyrrolidine
derivatives to opioid receptors
a
a
Compd
R1
R2
m Ki (nM) d Ki (nM) k Ki (nM)
Dmt-Tic
14
16
17
18
19
20
21
22
945Æ211
>1000
1.8Æ0.6
330Æ41
159Æ19
554Æ12
>1000
>1000
>1000
>1000
310Æ21
>1000
>1000
480Æ15
>1000
405Æ67
>1000
303Æ39
t-Bu
t-Bu
Ph
t-Bu
t-Bu
t-Bu
t-Bu
t-Bu
Boc
H
H
Me
t-Bu
Bn
Ac
Bz
85.5Æ5.4
38.2Æ4.7
176Æ13
137Æ7
>1000
23.2Æ3.9
508Æ44
443Æ32
551Æ40
>1000
>1000
>1000
23
24
t-Bu PhNHCO
Ph Bn
395Æ55
Figure 1. GTPg[35S] binding reversals of 500 nM of DAMGO by
compound 26 (~), compound 42 (!) and compound 46 (^). Full
agonistic properties of DAMGO (&) are also shown and represent
31.1Æ3.8
205Æ22
aAbbreviations: Ac: acetyl; Bn: benzyl; Boc: tert-butoxycarbonyl;
t-Bu: tert-butyl; Bz: benzoyl; Me: methyl; Ph: phenyl.
100% relative Emax
.

´
D. Page et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1585–1589
1588
All the compounds having binding affinities less than 15
nM were tested in GTPg[35S] assays24 to see whether
they activated the m receptor. However, no activation
was observed for any of the compounds up to a con-
centration of 1mM. Compounds 26, 42, and 46 were
therefore tested in competition assays against the m
agonist DAMGO (DAMGO: Tyr-DAla-Gly-[NMe-
Phe]-NH(CH2)2-OH) to assess their potential antagonist
effects (Fig. 1). They were shown to block the effect of
DAMGO with K0 values25 of 21, 126 and 6 nM, respec-
tively.
14. Page, D.; Naismith, A.; Schmidt, R.; Coupal, M.;
Labarre, M.; Gosselin, M.; Bellemare, D.; Payza, K.; Brown,
W. J. Med. Chem. 2001, 44, 2387.
15. Balboni, G.; Guerrini, R.; Salvadori, S.; Bianchi, C.;
Rizzi, D.; Bryant, S. D.; Lazarus, L. H. J. Med. Chem. 2002,
45, 713.
16. Balboni, G.; Salvadori, S.; Guerrini, R.; Negri, L.; Gian-
nini, E.; Jinsmaa, Y.; Bryant, S. D.; Lazarus, L. H. J. Med.
Chem. 2002, 45, 5556.
17. Schiller, P. W.; Weltrowsak, G.; Bolewska-Pedyczak,
E.; Nguyen, T. M.-D.; Lemieux, C.; Chung, N. N. In Pep-
tides 1996, Proceedings of the Twenty-Fourth European Pep-
tide Symposium; Ramage, R., Epton, R., Eds.; Mayflower
Scientific: Kingstwinford, UK, 1998; p 785.
18. Santagada, V.; Caliendo, G.; Severino, B.; Perissutti, E.;
Ceccarelli, F.; Giusti, L.; Mazzoni, M. R.; Salvadori, S.;
Temussi, P. A. J. Peptide Sci. 2001, 7, 374.
19. Lu, Y.; Weltrowska, G.; Lemieux, C.; Chung, N. N.;
Schiller, P. W. Bioorg. Med. Chem. Lett. 2001, 11, 323.
20. Lu, Y.; Nguyen, T. M.-D.; Weltrowska, G.; Berezowska,
I.; Lemieux, C.; Chung, N. N.; Schiller, P. W. J. Med. Chem.
2001, 14, 3048.
21. Schiller, P. W.; Lu, Y.; Weltrowska, G.; Berezowska, I.;
Wilkes, B. C.; Nguyen, T. M.-D.; Chung, N. N.; Lemieux, C.
In Peptides: The Wave of the Future, Proceedings of the Second
International and Seventeenth American Peptide Symposium;
Lebl, M., Houghten, R. A., Eds.; American Peptide Society:
San Diego, 2001; pp 676–678.
These molecules represent new scaffolds in the develop-
ment of m opioid-receptor antagonists. It is noteworthy
to mention that even though some partial m-agonism
was previously observed with this Tiq urea scaffold,14 it
is clear in this study that the compounds produced were
antagonists, suggesting a different mode of binding. It
has been previously demonstrated that subtle changes to
these types of molecules could greatly alter their phar-
macological profiles, converting them from antagonists
to agonists. Work is currently underway in our labora-
tories to explore further the SAR of this series.
Acknowledgements
22. Rodriguez, M.; Llinares, M.; Doulut, S.; Heitz, A.; Mar-
tinez, J. Tetrahedron Lett. 1991, 32, 923.
The authors are thankful to Dr. Ralf Schmidt and Dr.
Chris Walpole for helpful discussions.
23. All products were purified by reversed phase HPLC on
Luna C-18 column (250Â21.2 mm) using a gradient of 30–
80% CH3CN/H2O containing 0.1% TFA and were isolated as
their corresponding TFA salts in 50–65% yields. All products
gave satisfactory analytical characterization showing purity
>95% as determined by HPLC using a Zorbax C-18 column
(l=215, 254 and 280 nm). Selected analytical characteriza-
tion: Compound 26: 1H NMR (400 MHz, CD3OD) d 7.30–
6.80 (m, 13H), 4.78 (m, 1H), 4.51 (d, J=12.69 Hz, 1H), 4.16–
3.92 (m, 5H), 3.42 (m, 1H), 3.20–3.08 (m, 3H), 2.96–2.92 (m,
1H), 2.25 (m, 1H), 1.96 (m, 3H); MS (MH+): 530.2. Anal.
calcd (%) for C30H35N5O2S; C, 56.85; H, 5.35; N, 10.11;
Found(%): C, 56.87; H, 5.34; N, 10.30; HPLC k0: 4.35. Com-
pound 42: 1H NMR (400 MHz, CD3OD) d 7.29–6.95 (m,
12H), 4.80 (m, 1H), 4.51 (d, J=12.30 Hz, 1H), 4.26–4.16 (m,
2H), 3.93 (m, 1H), 3.85 (s, 3H), 3.79 (s, 3H), 3.75 (m, 1H), 3.52
(m, 1H), 3.21 (m, 1H), 3.10 (dd, J=5.08, 15.62 Hz, 1H), 2.95
(d, J=16.01 Hz, 1H), 2.25 (m, 1H), 1.97 (m, 3H); MS
(MH+): 574.3. Anal. calcd (%) for C32H39N5O3S: C, 56.30;
H, 5.62; N, 9.43; Found (%): C, 56.31; H, 5.53; N, 9.49;
References and Notes
1. Kaczor, A.; Matosiuk, D. Curr. Med. Chem. 2002, 9, 1567.
2. Beedle, A. M.; Zamponi, G. W. Drug Develop. Res. 2002,
54, 118.
3. Kaczor, A.; Matosiuk, D. Curr. Med. Chem. 2002, 9, 1591.
4. Menkens, K.; Bilsky, E.; Wild, K.; Portoghese, P. S.; Reid,
L.; Porreca, F. Eur. J. Pharm. 1992, 219, 345.
5. Lazarus, L. H.; Bryant, S. D.; Cooper, P. S.; Guerrini, R.;
Balboni, G.; Salvadori, S. Drug Discovery Today 1998, 3, 284.
6. Bryant, S. D.; Salvadori, S.; Cooper, P. S.; Lazarus, L. H.
Trends Pharmacol. Sci. 1998, 19, 42.
7. Schiller, P. W.; Weltrowska, G.; Berezowska, I.; Nguyen,
T. M.-D.; Wilkes, B. C.; Lemieux, C.; Chung, N. N. Biopoly-
mers (Peptide Science) 1999, 51, 411.
8. Salvadori, S.; Balboni, G.; Guerrini, R.; Tomatis, R.;
Bianchi, C.; Bryant, S. D.; Cooper, P. S.; Lazarus, L. H. J.
Med. Chem. 1997, 40, 3100.
HPLC k0: 5.77. Compound 46: H NMR (400 MHz, CD3OD)
1
9. Page, D.; McClory, A.; Mischki, T.; Schmidt, R.; Butter-
worth, J.; St-Onge, S.; Labarre, M.; Payza, K.; Brown, W.
Bioorg. Med. Chem. Lett. 2000, 10, 167.
10. Santagada, V.; Balboni, G.; Caliendo, G.; Guerrini, R.;
Salvadori, S.; Bianchi, C.; Bryant, S. D.; Lazarus, L. H.
Bioorg. Med. Chem. Lett. 2000, 10, 2745.
11. Balboni, G.; Salvadori, S.; Guerrini, R.; Bianchi, C.; San-
tagada, V.; Calliendo, G.; Bryant, S. D.; Lazarus, L. H. Pep-
tides 2000, 21, 1663.
12. Tourwe, D.; Cauwenberghe, S. V.; Vanommeslaeghe, K.;
Mannekens, E.; Geerlings, P.; Toth, G.; Peter, A.; Csombos, J.
In Peptides: The Wave of the Future, Proceedings of the Second
International and Seventeenth American Peptide Symposium;
Lebl, M., Houghten, R. A., Eds.; American Peptide Society:
San Diego, 2001; p 683.
d 7.31 (m, 3H), 7.25–7.19 (m, 9H), 6.98 (m, 1H), 6.41 (s, 1H),
6.35 (dd, J=1.36, 8.10 Hz, 1H), 4.75 (d, J=15.82 Hz, 1H),
4.36 (m, 1H), 4.17 (m, 2H), 3.96 (m, 1H), 3.78 (s, 3H), 3.74 (m,
1H), 3.48 (m, 1H), 3.20 (m, 1H), 3.12 (dd, J=5.08, 15.62 Hz,
1H), 2.95 (d, J=15.43 Hz, 1H), 2.26 (m, 1H), 1.98 (m, 3H);
MS (MH+): 560.2. Anal. calcd (%) for C31H37N5O3S: C,
54.57; H, 5.21; N, 9.25; Found(%): C, 54.56; H, 5.21; N, 9.08;
HPLC k0: 3.41.
24. The reversal of DAMGO-induced stimulation of
GTPg[35S] binding was used to assay the antagonist properties
of the compounds. Membranes were combined with approxi-
mately 0.2 nM GTPg[35S], 500 nM of DAMGO, and various
concentrations of the antagonist compounds. The assay was
performed in 50 mM Hepes, 20 mM NaOH, pH 7.4, 5 mM
MgCl2, 100 mM NaCl, 1 mM EDTA, 1 mM DTT, 0.1% BSA,
15 mM GDP. After 1 h, the bound radioactivity was deter-
mined by filtration. Control and stimulated GTPg[35S] binding
was determined in absence and presence of 30 mM of
13. Salvadori, S.; Guerrini, R.; Balboni, G.; Bianchi, C.; Bry-
ant, S. D.; Cooper, P. S.; Lazarus, L. H. J. Med. Chem. 1999,
42, 5010.

´
D. Page et al. / Bioorg. Med. Chem. Lett. 13 (2003) 1585–1589
1589
DAMGO. Values of DAMGO EC50 and Emax were obtained
from a 4-parameter logistic curve fits of percent stimulated
GTPg[35S] binding vs log(molar ligand), solving for EC50,
Emin, Emax and hill slope.
Cheng and Prusoff (Biochem. Pharmacol. 1973, 22, 3099)
equation:
IC50Antagonist
.
1 þ ð½Agonist usedŠ=EC50 AgonistÞ
25. K0 values were determined for each antagonist using the