Identification of agonists and antagonists of the human melanocortin-4 receptor from piperazinebenzylamines

Article abstract of DOI:10.1016/j.bmcl.2004.10.096

SAR studies of a series of piperazinebenzylamines resulted in identification of potent agonists and antagonists of the human melanocortin-4 receptor. Thus, the 1,2,3,4-tetrahydroisoquinolin-1-ylacetyl compound 12e and the quinolin-3-ylcarbonyl analogue 12l possessed K_i values of 6.3 and 4.5 nM, respectively. Interestingly, 12e was a full agonist with an EC ₅₀ value of 31 nM, and 12l was a weak partial agonist (IA = 17%) and functioned as an antagonist (IC₅₀ = 300 nM).

Full text of DOI:10.1016/j.bmcl.2004.10.096

Bioorganic & Medicinal Chemistry Letters 15 (2005) 833–837
Identification of agonists and antagonists of the human
melanocortin-4 receptor from piperazinebenzylamines
Joe A. Tran,a Joseph Pontillo,a Melissa Arellano,a Nicole S. White,a Beth A. Fleck,b
Dragan Marinkovic,a Fabio C. Tucci,a Marion Lanier,a Jodie Nelson,a John Saunders,a
Alan C. Fosterc and Chen Chena,*
aDepartment of Medicinal Chemistry, Neurocrine Biosciences, Inc., El Camino Real, San Diego, CA 92130, USA
bDepartment of Pharmacology, Neurocrine Biosciences, Inc., El Camino Real, San Diego, CA 92130, USA
cDepartment of Neurosciences, Neurocrine Biosciences, Inc., El Camino Real, San Diego, CA 92130, USA
Received 31 August 2004; revised 8 October 2004; accepted 12 October 2004
Available online 2 December 2004
Abstract—SAR studies of a series of piperazinebenzylamines resulted in identification of potent agonists and antagonists of the
human melanocortin-4 receptor. Thus, the 1,2,3,4-tetrahydroisoquinolin-1-ylacetyl compound 12e and the quinolin-3-ylcarbonyl
analogue 12l possessed Ki values of 6.3 and 4.5nM, respectively. Interestingly, 12e was a full agonist with an EC50 value of
31nM, and 12l was a weak partial agonist (IA = 17%) and functioned as an antagonist (IC50 = 300nM).
Ó 2004 Elsevier Ltd. All rights reserved.
The melanocortin-4 receptor (MC4R) has been associ-
ated with regulation of food intake and energy homeo-
stasis, therefore, MC4R agonists and antagonists could
be potentially used for treatment of obesity and cachex-
ia.1 Several small molecule MC4 agonists2 and antago-
nists3 from different chemical classes have been
discovered since the report of the first potent agonist
1a in 2002.4 In our previous paper we described that
by replacing the triazole moiety of our initial lead Tic-
(4-Cl)Phe dipeptide 2a (Ki = 270nM, EC50 = 80nM)5
with a basic nitrogen bearing an aliphatic side chain
we were able to produce a series of piperazinebenzylam-
ines exemplified by 2c (Ki = 6.4nM, EC50 = 4.7nM) as
potent and selective agonists of the human MC4 recep-
tor.6 Interestingly, when the triazole moiety of 2a was
replaced by a basic nitrogen bearing a small lipophilic
side chain, the binding affinity of these compounds in-
creased over 10-fold in many cases, however, their ago-
nist potency was barely changed. For example, the Ki
value (11nM) of 2b was almost 25-fold better than that
of 2a, but its EC50 value in stimulation of cAMP pro-
duction was only 360nM. By analyzing the initial data
we concluded that when a polar group, such as a hydro-
xyl and an amine was incorporated into the side chain,
agonist potency of these compounds was improved.
For example, 2c was a potent full agonist. To further
understand the structure–activity relationships as agon-
ists and antagonists of the melanocortin-4 receptor, a
series of benzylamines bearing an additional polar moi-
ety was synthesized. In addition, we surveyed a set of
different (4-Cl)Phe-amides to explore the role of the
1,2,3,4-tetrahydroisoquinolin-3-ylcarbonyl (Tic) group
of 2 in the activation of the MC4 receptor. Here we re-
port our findings of the structure–activity relationships
on this set of compounds (Fig. 1).
The synthesis of the piperazinebenzylamines 5 started
from 2-fluorobenzaldehyde 4 with a similar procedure
as previously described (Scheme 1).6 Compounds 12
were prepared from 20-fluoroacetophenone 7. Thus,
condensation of 7 with mono-Boc-protected piperazine
gave the corresponding 20-(piperazin-1-yl)acetophenone,
which was deprotected with TFA and coupled with (R)-
N-Boc-(4-Cl)Phe-OH (EDC/HOBt/DCM) to give the
acylpiperazine 8. After deprotection of 8, the resultant
free amine 9 was coupled with (R)-N-Boc-Tic-OH to
give the dipeptide 10. Reductive amination of 10 with
an alkylamine afforded, after Boc-deprotection with
TFA, the products 6a,b. Compound 9 was also coupled
with various carboxylic acids including N-Boc protected
Keywords: Melanocortin; MC4R; Agonists; Antagonist.
*
Corresponding author. Tel.: +1 858 617 7634; fax: +1 858 617
7998; e-mail: cchen@neurocrine.com
0960-894X/$ - see front matter Ó 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.10.096

834
J. A. Tran et al. / Bioorg. Med. Chem. Lett. 15 (2005) 833–837
S
N
N
Cl
Cl
Cl
N
N
HN
N
N
N
O
N
O
O
N
N
N
N
H
N
H
N
H
O
HN
O
NH
O
NH
2a
1a
2b
O
HN
Cl
Cl
N
N
Cl
O
NH
N
N
N
O
O
N
N
N
N
N
H
N
N
H
N
H
O
O
NH
O
2c
Figure 1. Some small molecule agonists of the human MC4 receptor.
1b
3
R1
R2
R3
fected with the human melanocortin-4 receptor as previ-
ously described.7 In the binding assay [125I]-NDP-MSH
was used as the radiolabeled ligand. For the functional
measurement, a-MSH was used to stimulate cAMP re-
lease in the antagonist assay, and as a standard for cal-
culation of intrinsic activity in the agonist assay.
Cl
N
N
O
F
Ref. 6
R3
O
N
N
H
O
HN
4: R3 = H
7: R3 = Me
5: R3 = H
Initial data from our previous report on piperazineben-
zylamines indicate that compounds with a N-alkyl
group bearing a heteroatom such as nitrogen and oxy-
gen display better agonist potency than those with a sim-
ple N-alkyl side chain. For example, the diamine 2c
(EC50 = 4.7nM) is a much more potent agonist than
the alkylamine 2b (EC50 = 360nM), while they possess
similar Ki values of 6.4 and 11nM, respectively. Despite
the difference in potency, these two compounds display
full efficacy with the intrinsic activity at the same level
of a-MSH. These results prompted us to postulate that
the diamine side chain of 2c is a good mimic of the
Arg-8 residue of the melanocortin peptides, which is
critical for receptor activation.8 Interestingly, while the
Scheme 1. Synthesis of compounds 5.
amino acids to give the intermediates 11, which were
subjected to a reductive amination with 3-aminopiperi-
dine, followed by Boc-deprotection where needed, to
give the desired products 12 (Scheme 2). The final com-
pounds 5, 6, and 12 were purified on a HPLC instrument
equipped with a mass detector, and the possible diaste-
reo-isomers of 6 and 12 were not separated by this puri-
fication method.
The competition binding experiments and the agonist
assay were performed using HEK293 cells stably trans-
Cl
Cl
Cl
O
O
N
O
a
b
N
O
N
N
N
N
NH
N
H
Boc
NHBoc
2
O
O
N
O
10
9
8
e
c,d
1
2
Cl
HN
R
R
Cl
Cl
O
N
NH
N
N
N
f,d
O
O
O
N
N
N
4
4
N
H
R
N
H
R
N
H
O
O
O
HN
11
12
6
Scheme 2. Reagents: (a) TFA/CH2Cl2, 98%; (b) (R)-N-Boc-Tic-OH/EDC/HOBt/Et3N; (c) R1R2NH/NaBH(OAc)3; (d) TFA/CH2Cl2; (e) R4COOH/
EDC/HOBt/Et3N; (f) 3-aminopiperidine/NaBH(OAc)3.

J. A. Tran et al. / Bioorg. Med. Chem. Lett. 15 (2005) 833–837
Table 1. SAR of benzylamines 5 and 1-phenethylamines 6 at the human MC4 receptora
835
1
2
R
R
Cl
N
N
3
R
O
N
N
H
O
HN
2b,5a-m,6a-b
Compound
R1NR2
R3
Ki (nM)b
EC50 (nM)c
2b
H
11
360
S
HN
HN
O
5a
5b
H
H
110
120
520
570
O
N
HN
HN
N
N
5c
5d
5e
H
H
H
160
370
270
570
640
360
HN
N
O
HN
N
H2N
HN
5f
H
89
390
N
5g
5h
5i
H
H
H
H
H
H
H
36
110
53
350
750
360
720
210
210
550
HN
N
COOEt
O
N
N
5j
240
78
N
NH
5k
5l
N
O
N
140
22
O
N
5m
N
HN
O
6a
6b
Me
Me
41
87
560
230
O
N
HN
a The human melanocortin receptor was stably transfected in HEK 293 cells.
b In the binding assay, [125I]NDP-MSH was used as the radiolabeled ligand. Key compounds (5a,b, 5d, 5f–i, 5k, 5m, 6a,b, 12d, and 12l) were measured
three times or more, and SEM of these measurements are less than 20% of the average.
c All compounds displayed intrinsic activity >80% of a-MSH.
guanidine functionality is highly basic (pKa > 12), and
its interaction with the melanocortin receptors is largely
believed to be a charge–charge attraction with a nega-
tively charged cage formed by three acidic residues, that

836
J. A. Tran et al. / Bioorg. Med. Chem. Lett. 15 (2005) 833–837
Table 2. SAR of substituted acetamides 12 at the human MC4
receptora
Arg-8 residue of a melanocortin peptide and the acidic
residues of the receptor is replaceable by hydrogen-
bonding interactions of a nonbasic functionality such
as an acylguanidine, or the triazole of 1a and 2a. The
acidic residues of the receptor such as Asp-122 and
Asp-126, which reside very closely on transmembrane
domain three of the hMC4R, most likely interact coop-
eratively with a ligand.
HN
Cl
NH
N
O
N
3
N
H
R
O
2c,12a-l
We first examined a set of alkylamines bearing a polar
group such as amine, ether, amide, and ester, which
are able to form hydrogen-bonding interaction, to ex-
plore the scope of the structure–activity relationship at
the benzylamine site. Unexpectedly, while these com-
pounds possessed moderate to good binding affinities
(KiÕs 22–370nM) and high intrinsic activity in stimula-
tion of cAMP release (>80%), none of the side chains
we selected significantly improved agonist potency of
these compounds (EC50Õs 210–750nM) from 2b (Table
1). Among them, the homopiperazine 5k showed a Ki
value of 78nM and EC50 of 210nM. The 4-pyridyl ana-
logue 5g had good binding affinity but only moderate
agonist potency (Ki = 36nM, EC50 = 365nM). The dia-
mine 5m (Ki = 22nM) exhibited the highest affinity of
this set of compounds in Table 1. Introduction of a
methyl group at the benzylic position of 5 only im-
proved binding affinity slightly. Thus, in comparison
with 5a (Ki = 110nM, EC50 = 520nM), the a-methyl
analogue 6a possessed a Ki value of 41nM and an
EC50 value of 560nM. Similarly, compound 6b was
slightly better in both binding affinity and agonist po-
tency than the benzylamine analogue 5b.
Compound
R4
Ki (nM)
EC50 (nM)
4.7
2c
6.4
HN
H2N
12a
47
(40%)
H2N
NH2
12b
12c
12d
24
120
10
440 (66%)
630
H2N
HN
330
12e
6.3
31
NH2
NH2
12f
12g
12h
52
56
49
400 (70%)
220 (66%)
(40%)
While the Tic-group seems to be essential for agonist
activity for small molecule MC4R ligands, limited infor-
mation of structure–activity relationship on this moiety
has been reported. A very recent publication shows that
the Tic-group of 1a can be replaced by other functional-
ities. For example, the N,N-dimethyl-b-alanine piperi-
dine 1b is a potent agonist (EC50 = 40nM).11 Similarly,
the phenylpiperazine 3 without the Tic-group has an
EC50 value of 15nM.12 A series of piperizinebenzylam-
ines with different amides including a- and b-amino
amides to replace the Tic-group was synthesized and
tested to explore the structure–activity relationship (12,
Table 2). While most of these compounds possessed
good binding affinity (Ki < 50nM), their agonist potency
was lower than that of the Tic-analogue 2c. Especially,
many analogues were not able to stimulate cAMP pro-
duction to the maximal level of a-MSH at high doses.
The 1,2,3,4-tetrahydroisoquinolin-1-ylacetyl compound
12e, however, possessed a Ki value of 6.3nM and an
EC50 of 31nM with the maximal cAMP stimulation at
the same level of a-MSH (Fig. 2).
N
NH2
NH2
N
12i
12j
12k
12l
34
33
30
4.8
(22%)
(31%)
(15%)
(17%)
N
a The human melanocortin receptor was stably transfected in HEK 293
cells.
is, Glu-90, Asp-122, and Asp-126 of the human melano-
cortin-4 receptor,9 recent data suggest that the basicity
of this guanidine is not essential for efficient interactions
of peptide ligands, and an acylguanidine analogue,
which is unlikely to be charged positively under physio-
logical conditions, retains the agonist activity of the cor-
responding guanidine compound.10 These results may
indicate that the charge–charge attraction between the
On the other hand, 12l with a quinolin-3-ylcarbonyl
moiety also exhibited high binding affinity
(Ki = 4.8nM), but stimulated very low level of cAMP re-
lease at high concentrations (maximal 17% of a-MSH at
a concentration of 10lM). Moreover, 12l was demon-
strated to be a functional antagonist. Thus, it dose-
dependently inhibited a-MSH-stimulated cAMP
production with an IC50 value of 300nM (Fig. 3).

J. A. Tran et al. / Bioorg. Med. Chem. Lett. 15 (2005) 833–837
837
such as 12l, only stimulated cAMP production to a min-
imal level (17% of a-MSH for 12l).
120
100
80
60
References and notes
40
α-MSH
12e
1. Goodfellow, V.; Saunders, J. Curr. Topic Med. Chem.
2003, 3, 855.
20
2. Richardson, T. I.; Ornstein, P. L.; Briner, K.; Fisher, M.
J.; Backer, R. T.; Biggers, C. K.; Clay, M. P.; Emmerson,
P. J.; Hertel, L. W.; Hsiung, H. M.; Husain, S.; Kahl, S.
D.; Lee, J. A.; Lindstrom, T. D.; Martinelli, M. J.; Mayer,
J. P.; Mullaney, J. T.; OÕBrien, T. P.; Pawlak, J. M.;
Revell, K. D.; Shah, J.; Zgombick, J. M.; Herr, R. J.;
Melekhov, A.; Sampson, P. B.; King, C.-H. R. J. Med.
Chem. 2004, 47, 744, and references cited therein.
3. (a) Vos, T. J.; Caracoti, A.; Che, J.; Dai, M.; Farrer, C. A.;
Forsyth, N. E.; Drabic, S. V.; Horlick, R. A.; Lamppu, D.;
Yowe, D. L.; Balani, S.; Li, P.; Zeng, H.; Joseph, I. B. J.
K.; Rodriguez, L. E.; Claiborne, C. F. J. Med. Chem.
2004, 47, 1602, and references cited therein; (b) Marsilje,
T. H.; Roses, J. B.; Calderwood, E. F.; Stroud, S. G.;
Forsyth, N. E.; Blackburn, C.; Yowe, D. L.; Miao, W.;
Drabic, S. V.; Bohane, M. D.; Daniels, J. S.; Li, P.; Wu,
L.; Patane, M. A.; Claiborne, C. F. Bioorg. Med. Chem.
Lett. 2004, 14, 3721.
4. Sebhat, I. K.; Martin, W. J.; Ye, Z.; Barakat, K.; Mosley,
R. T.; Johnston, D. B. R.; Bakshi, R.; Palucki, B.;
Weinberg, D. H.; MacNeil, T.; Kalyani, R. N.; Tang, R.;
Stearns, R. A.; Miller, R. R.; Tamvakopoulos, C.; Strack,
A. M.; McGowan, E.; Cashen, D. E.; Drisko, J. E.; Hom,
G. J.; Howard, A. D.; MacIntyre, D. E.; van der Ploeg, L.
H. T.; Patchett, A. A.; Nargund, R. P. J. Med. Chem.
2002, 45, 4589.
0
-11 -10
-9
-8
-7
-6
-5
-4
log concentration
Figure 2. Dose–response curves of cAMP production stimulated by
compound 12e and a-MSH in HEK 293 cells stably transfected with
the human melanocortin-4 receptor.
110
90
70
50
30
α-MSH
α-MSH (10 nM) + 12l
10
-10
-11 -10
-9
-8
-7
-6
-5
-4
log concentration
Figure 3. Dose–response curves of a-MSH-stimulated cAMP produc-
tion and inhibition by compound 12l.
5. Dyck, B.; Parker, J.; Philips, T.; Carter, L.; Murphy, B.;
Summers, R.; Hermann, J.; Baker, T.; Cismowski, M.;
Saunders, J.; Goodfellow, V. Bioorg. Med. Chem. Lett.
2003, 13, 3793.
6. Pontillo, J.; Tran, J. A.; Arellano, M.; Fleck, B.; Huntley,
R.; Marinkovic, D.; Lanier, M.; Nelson, J.; Parker, J.;
Saunders, J.; Tucci, F. C.; Jiang, W.; Chen, C. W.; White,
N. S.; Foster, A.; Chen, C. Bioorg. Med. Chem. Lett. 2004,
14, 4417.
Compound 12l was also tested at the other human mel-
anocortin subtypes and found to be very selective. Thus,
12l possessed Ki values of 630 and 600nM, respectively,
at the hMC3R and the hMC5R, and it only showed 38%
inhibition of a-MSH-stimulated cAMP production at
10lM concentration at the hMC1R. Similarly, 12e also
had weak binding affinity at the human MC3 receptor
(Ki = 1.9lM).
7. Nickolls, S. A.; Cismowski, M. I.; Wang, X.; Wolff, M.;
Conlon, P. J.; Maki, R. A. J. Pharmacol. Exp. Ther. 2003,
304, 1217.
8. Yang, Y.-k.; Fong, T. M.; Dickinson, C. J.; Mao, C.; Li,
J.-M.; Tota, M. R.; Mosley, R.; Van der Ploeg, L. H. T.;
Gantz, I. Biochemistry 2000, 39, 14900.
9. (a) Haskell-Luevano, C.; Cone, R. D.; Monck, E. K.;
Wan, Y. P. Biochemistry 2001, 40, 6164; (b) Yang, Y.;
Dickinson, C. J.; Haskell-Luevano, C.; Gantz, I. J. Biol.
Chem. 1997, 272, 2300; (c) Lu, D.; Vage, D. I.; Cone, R.
D. Mol. Endocrinol. 1998, 12, 592.
10. Cheung, A. W.-H.; Danho, W.; Swistok, J.; Qi, L.;
Kurylko, G.; Franco, L.; Yagaloff, K.; Chen, L. Bioorg.
Med. Chem. Lett. 2002, 12, 2407.
11. Ruel, R.; Herpin, T. F.; Iben, L.; Luo, G.; Martel, A.;
Mason, H.; Mattson, G.; Poirier, B.; Ruediger, E. H.; Shi,
D.; Thibault, C.; Yu, G.; Zimanyi, A.; Poindexter, G. S.;
Macor, J. E. Bioorg. Med. Chem. Lett. 2003, 13,
4341.
12. Xi, N.; Hale, C.; Kelly, M. G.; Norman, M. H.; Stec, M.;
Xu, S.; Baumgartner, J. W.; Fotsch, C. Bioorg. Med.
Chem. Lett. 2004, 12, 377.
In conclusion, we have designed and synthesized a series
of piperazinebenzylamines to study the structure–activ-
ity relationship as agonists of the melanocortin-4 recep-
tor. Attempting to mimic the functionality of the
diamine side chain of 2c by incorporating a heteroatom
in the N-alkyl group to enhance the possibility of hydro-
gen-bonding interactions with the receptor was not very
successful. However, we were able to find a replacement
of the Tic-group of 2c, thus, 12e with the 1,2,3,4-tetrahy-
dro-isoquinolin-1-ylacetyl moiety displayed good
agonist potency (EC50 = 31nM). In addition, the quino-
lin-3-ylcarbonyl derivative 12l was found to be a weak
partial agonist,13 but it functioned as an antagonist
(IC50 = 300nM) in inhibition of a-MSH-stimulated
cAMP release. These results demonstrate that the inter-
action of the Tic- or Tic-like group of these piperazine-
benzylamine compounds with the MC4 receptor is
crucial for receptor activation. All compounds in Table
1 with the Tic-group were able to fully activate the MC4
receptor, at least at a high concentration (10lM), but
many compounds without the Tic-moiety in Table 2,
13. An EC50 value of 610nM with intrinsic activity of 17%
was obtained on a dose–response curve.