Potent oxadiazole CGRP receptor antagonists for the potential treatment of migraine

Article abstract of DOI:10.1016/j.bmcl.2010.01.012

A pharmacophore model was built, based on known CGRP receptor antagonists, and this was used to aid the identification of novel leads. Analogues were designed, modelled and synthesised which incorporated alternative 'LHS' fragments linked via either an amide or urea to a privileged 'RHS' fragment commonly found in CGRP receptor antagonists. As a result a novel series of oxadiazole CGRP receptor antagonists has been identified and the subsequent optimisation to enhance both potency and bioavailability is presented.

Full text of DOI:10.1016/j.bmcl.2010.01.012

Bioorganic & Medicinal Chemistry Letters 20 (2010) 1368–1372
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Potent oxadiazole CGRP receptor antagonists for the potential treatment
of migraine
a
b
a
a
Paula L. Nichols a, , Jonathan Brand , Michael Briggs , Mathilde D’Angeli , Jennifer Farge ,
Stephen L. Garland c, Paul Goldsmith b, Rio Hutchings a, Ian Kilford a, Ho Y. Li a, David MacPherson a,
Fiona Nimmo a, Francis Dominic Sanderson d, Sanjeet Sehmi a, Nicola Shuker a, John Skidmore a,
Michael Stott a, Jennifer Sweeting a, Hasmi Tajuddin a, Andrew K. Takle a, Giancarlo Trani a, Ian D. Wall c,
Robert Ward a, David M. Wilson a, David Witty a
*
a Department of Medicinal Chemistry, Neurosciences Centre of Excellence for Drug Discovery, GlaxoSmithKline, New Frontiers Science Park, Third Avenue,
Harlow, Essex CM19 5AW, UK
b Department of Drug Metabolism and Pharmacokinetics, Neurosciences Centre of Excellence for Drug Discovery, GlaxoSmithKline, New Frontiers Science Park,
Third Avenue, Harlow, Essex CM19 5AW, UK
c Computational and Structural Chemistry, Molecular Discovery Research, GlaxoSmithKline, New Frontiers Science Park, Third Avenue, Harlow, Essex CM19 5AW, UK
d Early Development Group, Pharmaceutical Development, GlaxoSmithKline, New Frontiers Science Park, Third Avenue, Harlow, Essex CM19 5AW, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
A pharmacophore model was built, based on known CGRP receptor antagonists, and this was used to aid
the identification of novel leads. Analogues were designed, modelled and synthesised which incorporated
alternative ‘LHS’ fragments linked via either an amide or urea to a privileged ‘RHS’ fragment commonly
found in CGRP receptor antagonists. As a result a novel series of oxadiazole CGRP receptor antagonists has
been identified and the subsequent optimisation to enhance both potency and bioavailability is
presented.
Received 18 November 2009
Revised 2 January 2010
Accepted 4 January 2010
Available online 11 January 2010
Keywords:
Ó 2010 Elsevier Ltd. All rights reserved.
CGRP receptor antagonists
Oxadiazoles
Migraine
CGRP
Migraine is a chronic neurovascular disorder, which results in
episodic headache attacks lasting 4–72 h and is believed to affect
more than 10% of adults.1 The current standard treatment for mi-
graines is the class of 5-HT1B/1D receptor agonists known as the
triptans.2 However, triptans are contraindicated in patients with
cardiovascular disease because their mechanism of action involves
direct vasoconstriction.3 Therefore, a novel class of migraine drug
that did not result in direct vasoconstriction would offer a signifi-
cant therapeutic advantage.
demonstrated similar efficacy to the triptans following an IV infu-
sion but without any direct vasoconstriction.7
Following on from the successful proof of concept with IV
administered olcegepant, our programme aimed to identify a novel
orally bioavailable CGRP receptor antagonist for the treatment of
migraine. Further support for this came during the course of this
work when the first orally bioavailable CGRP receptor antagonist,
MK-0974, was reported to be efficacious in a Phase II clinical trial
for the treatment of acute migraine.8
Calcitonin gene-related peptide (CGRP) is a 37 amino acid neu-
ropeptide, present in the CNS and periphery,4 and is a potent vaso-
dilator.5 CGRP has been implicated in the pathophysiology of
migraine, as enhanced levels of CGRP are observed during migraine
attacks and an intravenous (IV) infusion of CGRP can induce mi-
graine-like headaches.6 Support for this theory initially came from
clinical trials of the CGRP receptor antagonist, olcegepant, which
To aid the design and identification of novel leads, a pharmaco-
phore model was built based on known CGRP receptor antagonists,
such as benzodiazepinone 1 and indazole 2.9,10 The ‘right-hand side’
(RHS) fragments were all highly conserved across these molecules
and they incorporated key hydrogen-bond acceptor/donor function-
ality. In addition, the linkers also included another hydrogen-bond
acceptor that was believed to be important for binding (Fig. 1).
In order to further improve our understanding of the CGRP
pharmacophore, novel analogues were designed, modelled and
synthesized. These analogues linked novel ‘left-hand side’ (LHS)
fragments, via an amide or a urea, to a range of privileged RHS
* Corresponding author.
E-mail address: paula.l.nichols@gsk.com (P.L. Nichols).
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.01.012

P. L. Nichols et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1368–1372
1369
H
N
F
O
F
N
F
O
N
O
H
H
N
N
N
N
N
N
N
O
N
H
O
N
2
1
LINKER
RHS
LHS
H
N
O
Y
HBA
N
( )n
X
N
Aromatic
group
X= CH2 or NH
Y= CH or N
O
Figure 1. CGRP pharmacophore schematic.
structures commonly found in known CGRP receptor antagonists.
The RHS in compounds 1 and 2 was not extensively explored due
to stability concerns therefore RHS A,11 which gave rise to com-
pounds with similar levels of potency, was routinely used. The sim-
plified LHS groups were designed to probe specific elements of the
CGRP pharmacophore, such as the necessity and positioning of cer-
tain aromatic lipophilic groups and hydrogen-bonding interactions
(Fig. 2).
From the analogues prepared, the simple biaryl system 3, with
an appropriately placed hydrogen-bond acceptor in the LHS frag-
ment, was identified as a weak CGRP receptor antagonist (fpKi
5.6).12
To follow up on this initial oxadiazole lead, a range of substi-
tuted-aryl analogues were prepared. Compounds 9a–t could be
prepared, for example, by treatment of the acid chlorides 5a–t with
hydrazine, followed by acylation gave the cyclisation precursors
7a–t. Subsequent microwave irradiation, in the presence of Bur-
gess reagent, afforded the 1,3,4-oxadiazoles 8a–t which then
underwent ester hydrolysis and amide coupling, with the privi-
leged RHS fragment A11 motif, to give analogue structures 9a–t
(Scheme 1).13
From the examples prepared various SAR trends were observed.
In general substitution on the aryl ring at the 2-, and 3-positions
was well tolerated but did not enhance potency, whereas substitu-
tion at the 4-position had a slightly detrimental effect on potency
(Table 1). However, the introduction of two substituents onto the
aromatic ring proved to be much more interesting, with 3,5-di-
chloro-substitution leading to a surprising increase in activity
(compound 9h, Table 1). Several other 3,5-substituent pairs were
explored and, in general whilst there was a requirement for a small
lipophilic group in the 3-position (e.g., Cl, Me), several groups of
varying size were well tolerated in the 5-position. Molecular mod-
elling was employed to aid our understanding of this SAR in rela-
tion to the pharmacophore, and this will be reported in a
separate communication.
N
O
O
N
O
O
N
O
N
O
N
N
NH
NH
N
N
4
3
fpKi 6.7, BEI 16.6
fpKi 5.6, BEI12.7
Following the identification of compound 3, further work to
optimise the hydrogen-bonding capacity led to the replacement
of the central methoxy-aryl core with a range of alternative 5-
and 6-membered heteroaryl systems. The most optimal of these
analogues was the 1,3,4-oxadiazole, as exemplified by compound
4, which demonstrated a significant improvement in potency (fpKi
6.7) and binding efficiency. It was postulated that the significantly
enhanced potency, observed with the 1,3,4-oxadiazole, was due to
the optimal hydrogen-bond accepting capacity of this heterocycle.
The most promising analogues were explored further and pro-
filed in vivo to understand their pharmacokinetic profile in the ro-
dent. Of particular interest was compound 9o which combined
high levels of permeability and an encouraging pharmacokinetic
profile, with clearance approaching 40% liver blood flow (rat IV,
1 mg/kg: Cl = 33 mL/min/kg, t1/2 = 0.5 h). This compound was also
explored in the rat to understand the oral exposure profile
(Fig. 3).
Interestingly, following oral dosing of 9o in a methylcellulose
formulation, the exposure was very limited. However, when dosed
in a solubilising formulation, the exposure was significantly en-
hanced (rat po, 10 mg/kg: Cmax = 0.7
l
M, Tmax = 1.5 h, dose-norma-
lised AUC = 6.3 min kg/L; estimated F = 20%14
)
(Fig. 3). This
suggested that exposure for this compound was probably limited
by solubility or dissolution rate.
Consequently the next phase of optimisation was focused on
improving oral bioavailability by enhancing the solubility of the
compounds. Therefore, a strategy of introducing basic substituents
onto the terminal aryl ring was employed (compounds 10a–h,
Table 2).
Figure 2. CGRP pharmacophore.

1370
P. L. Nichols et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1368–1372
R1
R1
O
O
R2
R3
i
ii
R2
R3
Cl
N
H
NH2
R5
R5
R4
5a-t
R4
6a-t
R1
R1
O
O
O
O
R2
R3
N
N
R2
R3
iii
OEt
OEt
N
H
N
H
O
R5
R5
R4
R4
7a-t
8a-t
H
N
R1
N
O
R2
R3
N
N
O
O
N
N
iv,v
O
N
NH
R5
N
R4
N
H
9a-t
RHS A
Scheme 1. Synthesis of aryl-oxadiazole analogues. Reagents and conditions: (i) hydrazine hydrate, ethanol, 80 °C; (ii) ethyl 3-chloro-3-oxopropanoate, Et3N, THF, rt; (iii)
burgess reagent, CH2Cl2,
lW, 100 °C; (iv) LiOH, THF, rt; (v) RHS A, EDC, HOBT, N-methyl morpholine, DMF, rt, 2–55% yield over five steps.
1.00
Table 1
Aryl-ring SAR
Rat 10
Rat 11
Rat 16
Rat 17
Rat 18
R1
R2
R3
R4
R5
CGRP fpKia
0.80
0.60
0.40
0.20
0.00
4
H
Cl
H
H
Cl
Cl
Cl
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
H
Cl
H
H
H
H
Cl
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
H
6.7b
6.3b
6.5c
6.0
9a
9b
9c
9d
9e
9f
9g
9h
9i
9j
9k
9l
9m
9n
9o
9p
9q
6.8
6.9b
6.1
H
Cl
Cl
F
Me
CF3
OMe
Cl
Cl
6.6
Cl
F
7.8d
6.9b
7.6c
7.6c
7.3
Me
CF3
OMe
F
Br
CF3
OMe
Ph
0
1
2
3
4
5
6
Time (h)
7.6c
7.9e
8.4f
8.1c
8.6g
Figure 3. Rat oral administration of 9o. Rats 16, 17 and 18 were dosed using 9o
formulated at a concentration of 2 mg/mL in 5% NMP, 80% PEG400 diluted to
volume with 20% Vitamin E TPGS in water. Rats 10 and 11 were dosed using 9o
formulated at a concentration of 2 mg/mL in 1% (w/v) aq methylcellulose.
Cl
Cl
Cl
9r
9s
9t
H
H
H
Cl
H
H
H
H
H
H
8.2h
8.2i
6.7b
N
N
N
O
O
O
Cl
N
O
N
O
RHS
Me
Ph
R1
10a-h
a
Values are mean of two experiments carried out in the absence of HSA, unless
O
N
O
otherwise indicated.
N
b
H
N
Values are mean of four experiments.
Values are mean of six experiments.
Values are mean of 12 experiments.
Values are mean of eight experiments.
Values are mean of 11 experiments.
Values are mean of five experiments.
Values are mean of 10 experiments.
c
N
N
NH
d
e
f
N
g
h
RHS A
RHS B
i
Values are mean of 73 experiments.

P. L. Nichols et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1368–1372
1371
Table 2
Optimisation of the basic substituents
R1
RHS
CGRP fpKia (À/+HSA)
FaSSIF15 sol.g
40
(l
g/ml)
A.M. Perm.h (nm/s)
120
CLb (mL/min/kg)
70
calcd CHI log D16
O
N
9r
A
A
8.2/6.9
1.7
—
N
N
10a
8.3b/7.6
210
<3
72
N
N
10b
10c
A
A
8.4c/7.7c
8.5d/7.8d
531
565
22
5
—
—
0.8
0.9
CHF2
10d
A
8.4b/7.1d
10
200
67
2.4
N
N
N
10e
10f
A
A
8.6/6.9
<1
220
<3
—
1.7
1.2
8.8d/7.0
390
52
N
N
10g
10h
A
B
8.3e/7.3f
7.9d/7.8
330
120
12
—
0.9
1.6
O
N
185
50
O
a
b
c
Values are mean of two experiments unless otherwise indicated.
Values are mean of eight experiments.
Values are mean of six experiments.
Values are mean of four experiments.
Values are mean of 12 experiments.
Values are mean of three experiments.
Value indicates the amount of compound which had dissolved in a fasted-state simulated intestinal fluid at the one hour time point.
Values were obtained using an in vitro artificial membrane permeability assay.
d
e
f
g
h
Gratifyingly the compound had a very encouraging pharmaco-
kinetic profile (rat IV, 1 mg/kg: Cl = 50 mL/min/kg; t1/2 = 1.9 h) and
demonstrated encouraging exposure when dosed orally using a
Pleasingly, replacement of the trifluoromethyl group in 9o with
a piperazine or a piperidine resulted in maintenance of potency
and a significant improvement in aqueous solubility (compounds
10a, 10b and 10c). However, these analogues generally had sub-
optimal metabolic stability and permeability was limited due to
the presence of the strongly basic centre.
standard methylcellulose formulation (rat
po, 10 mg/kg:
Cmax = 0.25 lM, Tmax = 3 h, dose-normalised AUC = 2.5 min kg/L;
estimated F = 14%14), albeit with a delayed absorption profile
(Fig. 4).
As a result, substituents with reduced basicity were targeted as
a means of achieving the optimum balance of solubility and per-
meability. The pyridyl analogue 10e maintained a good level of po-
tency but was not sufficiently basic to give the desired increase in
solubility, whereas the imidazole analogue displayed good levels of
solubility but was poorly permeable. Oxy-linked analogues such as
10g were, again, highly soluble but did not have the desired level of
permeability.
It was evident from an analysis of the series that good perme-
ability could be achieved with compounds where the calculated
CHI log D was greater than 1.7 but increasing this beyond 2.0 often
resulted in reduced solubility. In this regard, compound 10h which
incorporated an oxypropylamine LHS with the privileged RHS frag-
ment B,17 gave the most optimal balance of permeability and solu-
bility of all the analogues prepared. These optimal attributes were
believed to be due to basicity-reducing effect of the b-oxygen,
resulting in increased permeability whilst retaining some solubil-
ity, and the compound having an overall log D close to the desired
range. This compound was selected for further profiling in the
rodent.
In summary, a novel series of aryl-oxadiazole CGRP receptor
antagonists was designed using a pharmacophore model based on
previously reported CGRP receptor antagonists. The initial lead
was optimised via both modification of the central core and intro-
duction of suitable substituents in the LHS terminal aryl group. In
addition a strategy of balancing solubility, permeability and log D
Rat 10
Rat 11
Rat 12
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0
1
2
3
4
5
6
Time (h)
Figure 4. Rat oral administration of 10h, formulated at a concentration of 2 mg/mL
in 1% methylcellulose.

1372
P. L. Nichols et al. / Bioorg. Med. Chem. Lett. 20 (2010) 1368–1372
Hallermayer, G.; Wu, D.; Entzeroth, M.; Rudolf, K.; Engel, W.; Eberlein, W. Br.
J. Pharmacol. 2000, 129, 420.
8. Ho, T. W.; Mannix, L.; Fan, X.; Assaid, C.; Furtek, C.; Jones, C.; Lines, C.; Rapoport,
A. Neurology 2008, 70, 1304.
9. Williams, T. M.; Stump, C. A.; Nguyen, D. N.; Quigley, A. G.; Bell, I. M.;
Gallicchio, S. N.; Zartman, C. B.; Wan, B.-L.; Della Penna, K.; Kunapuli, P.; Kane,
S. A.; Koblan, K. S.; Mosser, S. D.; Rutledge, R. Z.; Salvatore, C.; Fay, J. F.; Vacca, J.
P.; Graham, S. L. Bioorg. Med. Chem. Lett. 2006, 16, 2595.
successfully delivered compounds with encouraging levels of oral
exposure in the rat. However, the delayed absorption profile ob-
served with 10h is not ideal as an acute migraine therapeutic. Conse-
quently further optimisation work aimed at improving the
bioavailability, in particular the Tmax, is requiredand will be reported
at a later date.
10. Luo, G.; Chen, L.; Degnan, A. P.; Dubowchik, G. M.; Macor, J. E.; Tora, G. O.;
Chaturvedula, P. V. WO2005/056550 A2.
11. Burgey, C. S.; Deng, Z. J.; Nguyen, D. N.; Paone, D. V.; Shaw, A. W.; Williams, T.
Acknowledgements
M. WO2004/092166 A2.
12. The functional pKi (fpKi) for antagonism of CGRP-stimulated increases in
intracellular cAMP was determined in HEK 293 cells expressing the human
recombinant receptor. Compounds which had encouraging potency in the
absence of HSA were then screened in the presence of HSA to give a better
prediction of the in vivo potency. For details of this assay, see: Nichols, P. L.;
Skidmore, J.; Ward, R. W.; Wilson, D. M. WO2009/000819 A1.
13. Scheme 1 details the most common route used to synthesise compounds of the
type 9a–t. Alternative syntheses can be found in Nichols, P. L.; Skidmore, J.;
Ward, R. W.; Wilson, D. M. WO2009/000819 A1.
The authors would like to thank the drug metabolism and phar-
macokinetic group for providing in vivo data. Thanks also go to the
pharmaceutical development group for the solubility analysis and
formulation work.
References and notes
14. F (%) was estimated, using data obtained with different animals in separate IV
and oral studies, using following calculation: F (%) = (AUC(0–t)/D po)/(AUC(0–
t)/D iv) Â 100.
1. Goadsby, P. J.; Lipton, R. B.; Ferrari, M. D. N. Eng. J. Med. 2002, 346, 257.
2. Silberstein, S. D. Lancet 2004, 363, 381.
3. Wakenfors, A.; Jarvius, M.; Ingemansson, R.; Edvinsson, L.; Malmsjo, M. J.
Cardiovasc. Pharmacol. 2005, 45, 476.
4. Poyner, D. R. Pharmacol. Ther. 1992, 56, 23.
5. Durham, P. L. N. Eng. J. Med. 2004, 350, 1073.
6. Lassen, L. H.; Haderslev, P. A.; Jacobsen, V. B.; Iversen, H. K.; Sperling, B.; Olesen,
J. Cephalalgia 2002, 22, 54.
7. (a) Olesen, J.; Diener, H.-C.; Husstedt, I. W.; Goadsby, P. J.; Hall, D.; Meier, U.;
Pollentier, S.; Lesko, L. N. Eng. J. Med. 2004, 350, 1104; (b) Doods, H.;
15. Galia, E.; Nicolaides, D.; Horter, D.; Lobenberg, R.; Reppas, C.; Dressman, J. B.
Pharm. Res. 1998, 15, 698.
16. CHI log D was used as a prediction of lipophilicity and was calculated using a
high-throughput HPLC method (chromatographic log D) which relates the
retention time to the lipophilicity (chrom log D = CHI Â 0.0857 À 2).
17. Burgey, C. S.; Stump, C. A.; Williams, T. M. WO2005/000807 A2.