An example of designed multiple ligands spanning protein classes: Dual MCH-1R antagonists/DPPIV inhibitors

Article abstract of DOI:10.1016/j.bmcl.2012.02.010

A ligand-based approach to identify potential starting points for a dual MCH-1R antagonist/DPPIV inhibitor medicinal chemistry program was undertaken. Potential ligand pairs were identified by analysis of MCH-1R and DPPIV in vitro data. A highly targeted

Full text of DOI:10.1016/j.bmcl.2012.02.010

Bioorganic & Medicinal Chemistry Letters 22 (2012) 2464–2469
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
An example of designed multiple ligands spanning protein classes: Dual MCH-1R
antagonists/DPPIV inhibitors
William T. Gattrell a, , Colin P. Sambrook Smith a, Alun J. Smith b
⇑
a Prosidion Ltd, Windrush Court, Watlington Road, Oxford OX4 6LT, United Kingdom
b Sygnature Discovery Ltd, BioCity, Pennyfoot Street, Nottingham NG1 1GF, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
A ligand-based approach to identify potential starting points for a dual MCH-1R antagonist/DPPIV inhib-
itor medicinal chemistry program was undertaken. Potential ligand pairs were identified by analysis of
MCH-1R and DPPIV in vitro data. A highly targeted synthetic effort lead to the discovery of pyridone
11, a dual MCH-1R antagonist/DPPIV inhibitor with selectivity over DPP8 and DPP9.
Ó 2012 Elsevier Ltd. All rights reserved.
Received 7 January 2012
Revised 2 February 2012
Accepted 3 February 2012
Available online 13 February 2012
Keywords:
Designed multiple ligands
Type II diabetes
MCH-1R
DPPIV
Type II diabetes, characterised by high blood glucose in the con-
text of insulin resistance, is due to a number of lifestyle and genetic
factors. Risk factors include poor diet, lack of physical activity,
obesity and paradoxically, prenatal under-nutrition. Obese individ-
uals with diabetes have poorer control of blood glucose levels; con-
versely, intentional weight loss is associated with reduced
mortality among individuals with diabetes.1 The majority of cur-
rent drug treatments are either weight neutral (Metformin, Acar-
bose, DPPIV inhibitors) or cause some weight gain (insulin,
Thiazolidinediones). An exception is GLP-1 analogues such as
Exanatide, however, they require administration by subcutaneous
injection.2 Thus there remains an unmet medical need for oral
anti-diabetics that give weight loss.
Oral dosing of MCH-1R antagonists is well documented to give
significant weight loss in rodent models. To date, few programs
have advanced far in the clinic, and validation in man remains elu-
sive.3 DPPIV inhibitors have been on the market for over 5 years,
and been shown to be well tolerated and weight neutral. The com-
bination of a MCH-1R antagonist and a DPPIV inhibitor would re-
quire co-formulation of two active ingredients, with the inherent
problems of achieving compatibility in the physicochemical, phar-
macokinetic and pharmacodynamic properties of the two active
ingredients. However, a single chemical entity combining MCH-
1R antagonism and DPPIV inhibition could circumvent such issues,
and may provide an oral anti-diabetic agent that gives significant
weight loss.
The term ‘designed multiple ligand’ (DML) was first proposed
by Morphy et al. as a generic phrase to describe compounds ratio-
nally designed to modulate multiple targets relevant to a disease.4
Examples of ligands designed to have activity toward unrelated
proteins (and within typical oral drug-like space) are rare. Ostensi-
bly, the disparity between the binding sites of unrelated proteins
may seem too great a technical challenge to identify drug-like li-
gands; nevertheless, examples have been reported.5 The undertak-
ing of such a seemingly daunting technical challenge is driven by
the potential for multi-target drugs to achieve greater clinical effi-
cacy over exquisitely selective drugs.6,7
A number of strategies have been proposed to identify chemical
starting points for DMLs. Linking of pharmacophores from selective
ligands, via a long chain is a common method exemplified. How-
ever, unless the individual pharmacophores are small or highly
overlapped, the resulting ligand will have molecular properties
outside typical drug-like space.8 Throughout medicinal chemistry,
common structural motifs in ligands reoccur. In designing a dual
ligand for a peptidase (DPPIV) and a Class A GPCR (MCH-1R) we
undertook a ligand-based approach to identify starting points for
a medicinal chemistry program. Figure 1 shows the structures of
FDA approved DPPIV inhibitors, and some literature MCH-1R
antagonists. Despite a lack of obvious structural similarity between
the ligands shown, we believed that a number of features of these
two targets may permit the design of dual ligands with drug-like
properties:
- Ligands for both targets display rich structural diversity, and
considerable SAR information is available from the literature.
⇑
Corresponding author.
E-mail address: wgattrell@yahoo.co.uk (W.T. Gattrell).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2012.02.010

W. T. Gattrell et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2464–2469
2465
F
F
O
N
F
O
H
N
O
N
N
NH2
N
N
N
N
N
N
H
O
N
HO
NH2
NC
N
NH2
F
F
F
Sitagliptin
Saxagliptin
Linagliptin
O
H
N
O
O
O
CF3
O
O
N
H
N
H
N
N
N
H
N
O
N
O
O
F
CO2H
Br
N
H
H
F
SNAP-7941
Early tool compound - Lundbeck
NGD-4715
Neurogen Deleopment candidate
AMG-076
Amgen Clinical Candidate
Figure 1. Structures of FDA approved DPPIV inhibitors (‘gliptins’) and some literature MCH-1H antagonists.
- Many potent DPPIV inhibitors are reported which possess low
molecular weight (and hence high ligand efficiency).
- Extensive X-ray crystallography data exists for inhibitors bound
to DPPIV which could aid in the design process.
met the criteria for synthesis.14 By investing considerable effort
in the design phase, and selecting only the highest quality ideas,
the resource intensive activities of synthesis and testing were
minimised.15
- X-ray crystallography has shown that DPPIV possesses a large
solvent exposed cavity proximal to the active site. This has
the potential to accommodate a significant part of the pharma-
cophore of a MCH-1R ligand.
Synthesis of the appropriate electrophiles was conducted as
outlined in Scheme 1. Alkylation of 5-hydroxy indan-1-one, fol-
lowed by ring expansion with sodium azide afforded the isoquino-
lin-1-one, with subsequent copper-mediated coupling with 2-
fluoro-5-iodo pyridine affording the final building block for 8.
Building blocks for 4, 6 and 9 were prepared using analogous cop-
per-mediated coupling; the precursors being commercially avail-
able or their synthesis outlined in the literature.16 Amides 5 and
7 were prepared starting from alkylation of the appropriate phenol,
followed by hydrolysis and amide coupling under standard condi-
tions. Finally, compounds 4 to 9 were prepared via SNAr reaction of
the electrophiles with the corresponding protected di-substituted
pyrolidine as outlined in Scheme 2.17
Ligand data was obtained for these targets from two sources.
Nexus is an in-house database containing data abstracted from
the medicinal chemistry literature; and in vitro data was extracted
from patents using proprietary techniques. In vitro data was col-
lated for ꢀ1500 MCH-1R antagonists, and ꢀ2500 DPPIV inhibitors.
Analysis identified a number of cases where MCH-1R and DPPIV li-
gands shared common structural motifs. Two examples of this, 1 &
2, and 2 & 3 are shown in Figure 2.
All three molecules comprise a common core containing a pyr-
olidine ring substituted on nitrogen with a heterocycle. MCH-1R
antagonist 1 (Ki 6 nM)9 has an amino group at the 3-position of
the pyrolidine, and likewise DPPIV inhibitor 2 (Ki 20 nM).10 On
the other hand, 2 has an additional 2,4,5-trifluorophenyl ring at
the 4-position, that fills the S1 pocket in DPPIV and is critical to
activity. Incorporation of the appropriately substituted phenyl ring
at the 4-position of the pyrolidine ring in 1 would furnish the po-
tential dual MCH-1R antagonist/DPPIV inhibitor 4.11 An indication
for the tolerance of a phenyl ring at this location in MCH-1R ligands
is gained from antagonist 3 (binding 25 nM)12 although it does lack
the amino group present in 1 and 2. Docking of 4 into a DPPIV crys-
tal structure suggested there was sufficient room to accommodate
the ligand.
The plethora of MCH-1R ligand data in the public domain
afforded the opportunity to explore many ideas for potential dual
ligands. MCH-1R antagonists believed to adopt a similar binding
mode to 1 and 3 were identified; from these, potential dual li-
gands were constructed virtually in a similar manner to 4. To pro-
vide compounds with drug-like properties, and to maximise the
potential for blood brain barrier penetration, limits on physio-
chemical properties were set at PSA 6 80 Å2, LogD 6 3.5 and
molecular weight 6 500.13 Out of 26 cores examined, only six
Compounds were screened for functional activity in CHO
cells stably expressing human MCH-1R receptor. Results are
shown in Table 1 for inhibition of agonist (h-MCH) response
at 5 lM. For the majority of analogues there was significant
inhibition of agonist response, and compounds with >40% inhi-
bition underwent full IC50 determination. Pleasingly, pyridone
4
showed excellent functional MCH-1R antagonism, IC50
0.14 M, whilst analogues 5 and 8 showed micromolar levels
of activity. Pyridone 4 also showed good DPPIV inhibition, IC50
1.65 M, although possessed no significant selectivity over
l
l
DPP8 and DPP9 (Table 2).
Encouraged by the generation of a dual MCH-1R antagonist/
DPPIV inhibitor, the effect of varying the substituent filling the
S1 pocket in DPPIV was probed. These were prepared in a similar
fashion to Scheme 2. Pyridone 10, the 2,4,5-trifluorophenyl ana-
logue of 4, showed comparable levels of functional MCH-1R antag-
onism, DPPIV inhibition and selectivity.18 Ring expansion of
pyrolidine 2 to the corresponding piperidine derivative is reported
to yield a more potent DPPIV inhibitor (Ki 6 nM).19 Thus the equiv-
alent piperidine analogue 11 was prepared. Gratifyingly, 11
showed submicromolar MCH-1R antagonism/DPPIV inhibition
(IC50 0.44
lM/0.35 lM, respectively) and selectivity over DPP8
and DPP9.20 The requirement of the S1 substituent for DPPIV inhi-

2466
W. T. Gattrell et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2464–2469
F
F
NH2
F
N
N
O
N
N
N
NH2
N
O
O
N
N
N
H
F3C
1
2
3
MCH-1R
antagonist
DPPIV
MCH-1R
antagonist
inhibitor
F
F
NH2
N
N
O
N
O
4
Dual MCH-1R antagonist/ DPPIV inhibitor?
Identfiy MCH-1R ligands with similar binding modes
Fragment & enumerate
Filter on properties
F
F
F
F
F
F
F
F
F
F
NH2
NH2
NH2
NH2
N
NH2
N
N
N
N
N
N
N
N
N
N
O
N
N
N
HN
O
HN
O
O
O
O
O
O
Figure 2. Outline of a ligand-based approach to identify potential dual MCH-1R antagonists/DPPIV inhibitors.
Table 1
Molecular properties and MCH-1R antagonism data for the initial set of analogues
Structurea
%Inhibition at 5
104
l
Mb
MCH-1R IC50
0.14
(lM)
PSA (Å2)
LogDd
c
Compd
F
O
N
N
F
4
72
1.9
O
N
NH2
O
F
O
N
H
N
F
5
43
3.8
80
2.8
N
NH2

W. T. Gattrell et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2464–2469
2467
Table 1 (continued)
c
Compd
Structurea
%Inhibition at 5
39
l
Mb
MCH-1R IC50
(
lM)
PSA (Å2)
LogDd
F
O
N
N
F
6
7
ND
62
2.4
3.5
N
NH2
F
O
35
ND
80
N
F
H
N
N
NH2
O
F
N
F
8
44
52
4.0
72
75
3.1
2.4
N
O
N
O
NH2
F
O
9
>3.0
N
F
N
N
N
NH2
a
All compounds were >95% pure by LCMS, and characterised by NMR and MS.
b
c
%Inhibition of control agonist response at 5
lM in CHO cells stably expressing human MCH-1R (conducted at Cerep).
N = 2.
ACDLogD at pH 7.4.
d
Table 2
In vitro data for some pyridone derivatives
Compd
Structure
IC50
(
lM)
LogDc
MCH-1R
0.02
DPPIVa
>100
DPP8a
DPP9a
>100
hERGb
ND
O
N
N
N
1
>100
-0.4
O
O
N
N
NH2
NH2
F
F
O
N
F
4
10
11
0.14
0.11
0.44
1.65
1.1
2.26
3.28
8.8
6.1
ND
1.9
2.1
2.2
F
O
N
N
N
F
2.6
2.2
O
O
N
NH2
F
O
N
N
F
0.35
>50
>50
NH2
F
a
b
c
For assay details see Ref. 5a; N = 2.
Performed at Essen Bioscience.
ACDLogD at pH 7.4; ND = not determined.
bition is underscored by pyrolidine 1, synthesised as a control,
Several pharmaceutical companies have invested heavily in the
which showed MCH-1R IC50 0.02
DPP8 and DPP9.
lM, but IC50 >100
l
M for DPPIV,
identification of viable MCH-1R antagonists, the main hurdle pre-
venting successful development of such compounds being cardio-

2468
W. T. Gattrell et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2464–2469
N
F
O
O
O
(iii)
NH
(i), (ii)
N
HO
O
O
R2
O
(vi)
(iv), (v)
O
O
O
HO
X
O
X
O
OH
N
N
H
Cl
(vi)
O
H
N
O
N
Cl
Scheme 1. Reagents and conditions: (i) Butylbromide, potassium carbonate, DMF, 85%; (ii) sodium azide, TFA, 80 °C, 12 h, 82%; (iii) DMF, 2-fluoro-5-iodo pyridine, potassium
carbonate, trans-N,N-dimethylcyclohexadiamine, 110 °C, 18 h, 35%; (iv) Alkyl bromide, potassium carbonate, DMF; (v) sodium hydroxide, methanol; (vi) Amine, HBTU,
triethylamine, DMF
F
F
N
(i), (ii)
F
R
X
+
HN
N
F
R
N
NH
NH2
O
O
Scheme 2. Reagents and conditions: (i) Diisopropylethylamine, DMSO, 150 °C, microwave, or potassium carbonate, DMF, 110 °C, microwave; (ii) TFA, DCM, room
temperature.
3.
A MCH-1R antagonist efficacious in reducing food intake in Cynomolgus
vascular safety, in particular as a result of hERG inhibition. The
hERG data obtained for compounds 4 and 10 (IC50 8.8 M and
6.1 M, respectively) suggest this chemotype represents a reason-
able starting point for lead optimisation.
Monkeys has recently been reported: Mihalic, J. T.; Chen, X.; Fan, P.; Chen, X.;
Fu, Y.; Liang, L.; Reed, M.; Tang, L.; Chen, J.; Jaen, J.; Li, L.; Dai, K. Bioorg. Med.
Chem. Lett. 2011, 21, 7001.
l
l
4. Morphy, R.; Kay, C.; Rankovic, Z. Drug Discovery Today 2004, 9, 641.
5. (a) Barba, O., Bradley, S. E., Fyfe, M. C. T., Hanrahan, P. E., Krulle, T. M., Procter,
M. J., Reynet McCormack, C., Schofield, K. L., Smyth, D. S., Stewart, A. J. W.,
Swain, S. A.; Widdowson, P. WO Patent, 2009034388, 2009.; (b) Toda, N.;
Kaneko, T.; Kogen, H. Chem. Pharm. Bull. 2010, 58, 273.
In summary, the desire to produce a new anti-diabetic with
weight loss in a single molecular entity drove an effort to design
a ligand with dual activity at two unrelated protein targets. A li-
gand-based approach was developed to identify starting points
for a dual MCH-1R antagonist/DPPIV inhibitor medicinal chemistry
program, with properties designed to maximise the potential for
CNS exposure. Through the highly targeted synthesis of only a
handful of molecules, potent dual MCH-1R antagonists/DPPIV
inhibitors were identified, with one compound having selectivity
over DPP8 and DPP9.
6. Hopkins, A. L. Nat. Chem. Biol. 2008, 11, 682.
7. Roth, B. L.; Sheffler, D. J.; Kroeze, W. K. Nat. Rev. Drug Disc. 2004, 3, 353.
8. For a recent example exploiting this feature see: Osborne, R.; Clarke, N.;
Glossop, P.; Kenyon, A.; Liu, H.; Patel, S.; Summerhill, S.; Jones, L. H. J. Med.
Chem. 2011, 54, 6998.
9. Armour, D. R., Galan, S. R. G., Lane, C. A. L., Lansdell, M. I., Mills, J. E., J.,
Sciammetta, N., Stupple, P. A. WO Patent 2008041090, 2008.
10. Backes, B. J.; Longenecker, K.; Hamilton, G. L.; Stewart, K.; Lai, C.; Kopecka, H.;
von Geldern, T. W.; Madar, D. J.; Pei, Z.; Lubben, T. H.; Zinker, B. A.; Tian, Z.;
Ballaron, S. J.; Stashko, M. A.; Mika, A. K.; Beno, D. W. A.; Kempf-Grote, A. J.;
Black-Schaefer, C.; Sham, H. L.; Trevillyan, J. M. Bioorg. Med. Chem. Lett. 2005,
2007, 17.
Acknowledgments
11. In house data suggested the lower molecular weight 2,5-difluorophenyl ring
would serve as a suitable group to fill the S1 pocket, and this group was
selected in the first instance.
12. Jiang, J.; Hoang, M.; Young, J. R.; Chaung, D.; Eid, R.; Turner, C.; Lin, P.; Tong, X.;
Wang, J.; Tan, C.; Feighner, S.; Palyha, O.; Hreniuk, D. L.; Pan, J.; Sailer, A. W.;
MacNeil, D. J.; Howard, A.; Shearman, L.; Stribling, S.; Camacho, R.; Strack, A.;
Van der Ploeg, L. H. T.; Goulet, M. T.; De Vita, R. Bioorg. Med. Chem. Lett. 2006,
16, 5270.
The author thanks Tim James and James Bell for molecular mod-
elling support, Ustav Bali for coordinating the in vitro screening,
Eleanor Curtis for assisting with the literature review, and Craig
Johnstone for review and discussion of the manuscript.
13. Hitchcock, S. A.; Pennington, L. D. J. Med. Chem. 2006, 49, 7559.
14. For example, the core of antagonist 3 was not selected for synthesis as the
corresponding potential dual ligand had calculated LogD 4.6.
15. Andersson, S.; Armstrong, A.; Bjore, A.; Bowker, S.; Chapman, S.; Davies, R.;
Donald, C.; Egner, B.; Elebring, T.; Holmqvist, S.; Inghardt, T.; Johannesson, P.;
Johansson, M.; Johnstone, C.; Kemmitt, P.; Kihlberg, J.; Korsgren, P.; Lemurell,
References and notes
1. Williamson, D. F.; Thompson, T. J.; Thun, M.; Flanders, D.; Pamuk, E.; Byers, T.
Diabetes Care 2000, 23, 1499.
2. Heine, R. J.; Van Gaal, L. F.; Johns, D.; Mihm, M. J.; Widel, M. H.; Brodows, R. G.
Ann. Intern. Med. 2005, 143, 559.

W. T. Gattrell et al. / Bioorg. Med. Chem. Lett. 22 (2012) 2464–2469
2469
M.; Moore, J.; Pettersson, J. A.; Pointon, H.; Pontén, F.; Schofield, P.; Selmi, N.;
Whittamore, P. Drug Discovery Today 2009, 14, 598.
16. Andres-Gil, J. I.; Alcazar-Vaca, M. J., Alvarez-Escobar, R. M., Oyarzabal, S. J.,
Dautzenberg, F. M., Macritchie, J., Simpson, D., Martinez Gonzalez, S. WO Patent
WO2007071646, 2007.
19. Pei, Z.; Li, X.; von Geldern, T. W.; Longenecker, K.; Pireh, D.; Stewart, K. D.;
Backes, B. J.; Lai, C.; Lubben, T. H.; Ballaron, S. J.; Beno, D. W. A.; Kempf-Grote, A.
J.; Sham, H. L.; Trevillyan, J. M. J. Med. Chem. 1983, 2007, 50.
20. 1-(2-Pyridyl)-3-amino-4-(2,4,5-trifluorophenyl)piperidine, the analogue of 11
lacking the benzyloxypyridone tail, is a reported DPPIV inhibitor (Ref. 17,
17. For synthesis of this and the other nucleophiles used, see Ref. 7 and Cox, J. M.;
Harper, B.; Mastracchio, A.; Leiting, B.; Roy, R. S.; Patel, R. A.; Wu, J. K.; Lyons, K.
A.; He, H.; Xu, S.; Zhu, B.; Thornberry, N. A. Bioorg. Med. Chem. Lett. 2007, 17,
4579.
18. A measured LogD of 3.4 (pH 7.4) was determined for compound 10, more than
a log unit higher than the calculated value.
compound 14, IC50 = 0.058 lM). Although it is not strictly correct to directly
compare this value with that for analogue 11, the data does suggest that the
benzyloxypyridone tail may reduce DPPIV potency. This could be due to the
requirement for positioning the terminal lipophilic aryl ring within the solvent
exposed cavity of the DPPIV protein.