Identification of novel, orally bioavailable spirohydantoin CGRP receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2006.09.045

A rapid analogue approach to identification of spirohydantoin-based CGRP antagonists provided novel, low molecular weight leads. Modification of these leads afforded a series of nanomolar benzimidazolinone-based CGRP receptor antagonists. The oral bioavai

Full text of DOI:10.1016/j.bmcl.2006.09.045

Bioorganic & Medicinal Chemistry Letters 16 (2006) 6165–6169
Identification of novel, orally bioavailable spirohydantoin
CGRP receptor antagonists
Ian M. Bell,a,* Rodney A. Bednar,b John F. Fay,b Steven N. Gallicchio,a
Jerome H. Hochman,c Daniel R. McMasters,d Cynthia Miller-Stein,c Eric L. Moore,b
Scott D. Mosser,b Nicole T. Pudvah,c Amy G. Quigley,a Christopher A. Salvatore,b
Craig A. Stump,a Cory R. Theberge,a Bradley K. Wong,c C. Blair Zartman,a
Xu-Fang Zhang,a Stefanie A. Kane,b Samuel L. Graham,a
Joseph P. Vaccaa and Theresa M. Williamsa
aDepartment of Medicinal Chemistry, Merck Research Laboratories, West Point, PA 19486, USA
bPain Research, Merck Research Laboratories, West Point, PA 19486, USA
cDrug Metabolism, Merck Research Laboratories, West Point, PA 19486, USA
dMolecular Systems, Merck Research Laboratories, West Point, PA 19486, USA
Received 3 August 2006; revised 12 September 2006; accepted 14 September 2006
Available online 5 October 2006
Abstract—A rapid analogue approach to identification of spirohydantoin-based CGRP antagonists provided novel, low molecular
weight leads. Modification of these leads afforded a series of nanomolar benzimidazolinone-based CGRP receptor antagonists. The
oral bioavailability of these antagonists was inversely correlated with polar surface area, suggesting that membrane permeability was
a key limitation to absorption. Optimization provided compound 12, a potent CGRP receptor antagonist (Ki = 21 nM) with good
oral bioavailability in three species.
ꢀ 2006 Elsevier Ltd. All rights reserved.
Migraine is a chronic, highly disabling disorder that af-
flicts more than 10% of adults and is often characterized
by severe, unilateral headache.1 The pathophysiology is
thought to involve dilation of cranial blood vessels and
activation of the trigeminovascular system.2 The pre-
ferred treatment for acute migraine is the triptan class
of 5-HT1B/1D receptor agonists, although these com-
pounds are contraindicated for patients with cardiovas-
cular disease because they are potent vasoconstrictors.1
A number of lines of evidence have implicated the neu-
ropeptide calcitonin gene-related peptide (CGRP) as a
key player in migraine pathology.3 Moreover, the po-
tent, intravenously administered CGRP receptor antag-
onist BIBN4096BS (olcegepant) demonstrated clinical
proof of concept for the acute treatment of migraine
headache.4 BIBN4096BS achieved comparable efficacy
to that seen historically for triptans and no serious
adverse effects or changes in hemodynamics were ob-
served.5 Therefore, a CGRP receptor antagonist suitable
for oral administration may represent a significant ad-
vance in acute migraine therapy.
Our program to develop orally bioavailable, small
molecule CGRP receptor antagonists began with the
identification of benzodiazepinone 1 (Fig. 1, CGRP
Ki = 4800 nM) via high throughput screening.6 Initial
work focused on replacement of the tetralin spirohydan-
toin fragment with a variety of piperidinyl privileged
structures and led to nanomolar antagonists, including
some with modest oral bioavailability.6,7
O
Me
N
O
NH
O
O
N
H
N
H
N
H
N
Keywords: CGRP receptor antagonists; CGRP; Migraine;
Spirohydantoin.
1
*
Figure 1. High throughput screening lead.
Corresponding author. E-mail: ian_bell@merck.com
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.09.045

6166
I. M. Bell et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6165–6169
In a complementary approach, we sought to replace the
benzodiazepinone moiety in 1 using a rapid analogue
strategy, as illustrated in Figure 2. Briefly, three spiro-
hydantoin templates (two tetralins and one indane) were
chosen to display potential benzodiazepinone replace-
ments from different vectors with respect to the hydan-
toin pharmacophore. The templates were derivatized
using several linkers (e.g., X = NH(CO)NH, CONH,
SO2NH, NHCO) to attach a wide variety of fragments
(generally with Mw < 200).
strated that the potency essentially resided in the
(R)-enantiomer (5), while the (S)-enantiomer (6) was
significantly less active. The preference for (R)-ind-
anylspirohydantoin was consistent in this series of antag-
onists, therefore only the (R)-enantiomers of the
analogues in Table 2 will be discussed (vide infra).
Replacement of the benzoxazolinone in lead 3 with benz-
imidazolinone gave compound 7, which was equipotent
with 3 in the binding assay but about 4-fold more potent
in the cell-based assay. Substitution at the benzimidazoli-
none nitrogen led to the methyl analogue 8 and phenyl
analogue 9, both of which exhibited enhanced intrinsic
potency (Table 1). Combining the modifications in com-
pounds 5 and 7 afforded benzimidazolinone 10 (Table
2), which possessed similar affinity to 5 for the CGRP
receptor (Ki = 43 nM).
Screening of such analogues yielded a number of struc-
turally diverse antagonists with Ki < 100 lM, including
representatives of the three templates shown in Figure
2 and a variety of linkers. Compounds 2 and 3 (Table
1), in which a simple benzoxazolinone was substituted
for the benzodiazepinone in 1, were among the most po-
tent leads and were selected for further evaluation. The
tetralin 2 (Ki = 607 nM) exhibited improved intrinsic
potency relative to 1 combined with reduced molecular
weight, while indane 3 offered a further boost in potency
(Ki = 309 nM). A small potency advantage for indanes
versus tetralins was consistently observed in this series
of compounds so the remaining discussion will focus
on indanyl analogues.
The unsubstituted benzimidazolinone nitrogen in com-
pound 10 could be derivatized with a wide variety of
substituents with no loss in potency. A subset of these
modifications, in which this position is substituted with
a series of cyclic groups, is shown in Table 2. The most
potent of these analogues were the thiazole 13 and iso-
thiazole 15, but a range of aryl, heteroaryl, and saturat-
ed rings were well tolerated. In general, there was a good
correlation between the Ki values obtained in the radio-
ligand binding assay and the IC50 values for functional
antagonism, with the latter value being 3- to 5-fold less
potent than the former. Because a range of substituents
on the benzimidazolinone were consistent with good
potency, we reasoned that this may be an ideal site to
Methylation of the spirohydantoin moiety provided 4
(Table 1), which exhibited improved potency compared
with 3 in the radioligand binding assay and as a
functional antagonist in the cell-based assay
(IC50 = 1400 nM).9 Resolution of the racemic 4 demon-
O
O
O
NH
NH
NH
X
O
N
O
N
N
O
H
H
X
H
X
= benzodiazepinone replacement
X = linker
Figure 2. Rapid analogue strategy.
Table 1. CGRP antagonist activity of benzoxazolinone and benzimidazolinone compounds
O
A
O
NH
X
O
N
N
R
O
n
N
H
Compound
A
X
n
R
Stereoa
CGRP Kib,c (nM)
cAMP
b,d
IC50 (nM)
2
3
4
5
6
7
8
9
—
—
—
—
—
H
O
O
O
O
O
N
N
N
2
1
1
1
1
1
1
1
H
RS
RS
RS
R
607 35 (3)
309 61 (6)
103 23 (4)
48 5 (3)
3300 (2)
2900 (2)
1400 (2)
350 (2)
—
H
Me
Me
Me
H
S
1050 230 (3)
280 100 (4)
78 11 (4)
RS
RS
RS
690 (1)
530 180 (3)
430 (2)
Me
Ph
H
H
87 35 (6)
a Stereochemistry of spirohydantoin.
b Mean value standard deviation, where appropriate; number of replicates in parentheses.
c Ki values for competition with 125I-hCGRP determined using membranes from HEK293 cells stably expressing cloned human CLR/RAMP1 (line
E10).8
d Inhibition of CGRP-induced cAMP production in the E10 cell line.8

I. M. Bell et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6165–6169
Table 2. CGRP antagonist activity of benzimidazolinone compounds
6167
O
A
O
NH
N
O
N
N
O
N
H
Me
Compound
A
H
CGRP Ki (nM)a
cAMP IC50 (nM)a
10
43 25 (7)
120 (2)
180 (2)
78 23 (3)
24 (2)
F
11
12
13
14
15
16
17
18
19
20
29 11 (5)
21 8 (7)
8.7 2.5 (5)
20 4 (5)
6.8 1.7 (7)
15 6 (4)
10 3 (9)
25 4 (5)
12 1 (5)
9.5 5.8 (8)
17 2 (4)
F
N
S
N
H
O
61 3 (3)
39 (2)
N
S
54 (2)
N
MeO
F3C
N
50 (2)
N
N
110 (2)
64 (2)
N
N
N
N
N
50 23 (3)
64 (2)
O
H
21
S
O
a See footnote to Table 1 for details.
modify physical properties in an attempt to modulate
the pharmacokinetic profile of such compounds.
solution, pKa, and metabolism.12 Perhaps because the
compounds in Table 3 were closely related in structure,
some of the variability in these other parameters was
minimized and a trend related to PSA was observed
more clearly. For example, pKa was probably not a ma-
jor factor since none of the compounds were very basic.
In terms of metabolism, the bioavailable compounds in
Table 3 had low clearance but other compounds with
similar clearance were not orally bioavailable (e.g., 17
and 19). Thus, low clearance may be necessary but not
sufficient for oral bioavailability in these analogues.
Table 3 summarizes pharmacokinetic data determined
in rats for compounds 11–21. The compounds are pre-
sented in order of increasing calculated polar surface
area10 (PSA) and it is apparent that there is an inverse
correlation between this parameter and oral bioavail-
ability (F). For example, analogues 11–13, which had
the lowest PSA values, had good oral bioavailability
(F = 20–36%). On the other hand, analogues 19–21,
which had the highest PSA values, exhibited essentially
no bioavailability (F = 0–1%). This behavior is consis-
tent with oral absorption being limited by membrane
permeability.10
It is clear that compounds must possess sufficient solu-
bility in order to be orally bioavailable.13 However,
most of these analogues had poor aqueous solubility
and there was no clear relationship with bioavailability.
Compounds 12 and 13 (F = 20–36%) had extremely low
solubility (<1 lg/mL at pH 7.4), while compounds 20
It should be noted that oral bioavailability is also depen-
dent on other variables, including solubility, rate of dis-

6168
I. M. Bell et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6165–6169
Table 3. Pharmacokinetic data for benzimidazolinone CGRP antagonists
2
c
iv t1/2 (h)
c
Vdss (L/kg)
d
PSAa (A )
Fb (%)
Clc (mL/min/kg)
Papp (·10À6 cm/s)
˚
Compound
11
12
13
14
15
16
17
18
19
20
21
117
125
127
128
131
134
137
138
140
146
158
24
36
20
5
6.5
1.6
2.0
0.9
1.0
1.2
2.8
1.1
1.1
0.5
1.1
3.1
3.4
3.7
9.0
60
1.4
15
11
13
4
0.47
0.54
0.43
2.0
1
—
—
—
—
8
13
0
11
0.84
1.8
7.5
23
2
0.50
0.64
0.95
0.80
1
11
46
0
4
0.3
0
31
a Polar surface area calculated by method of Clark10 and was an average of the values for 8–10 energy-minimized conformers.
b Oral bioavailability in rats following dosing (10 mpk) in 1% methylcellulose.
c Pharmacokinetic parameters in rats following iv dosing (2 mpk) in DMSO.
d Passive permeability through CACO-2 monolayer determined using 20 lM test compound in the presence of 5 lM cyclosporin.11
and 21 (F = 0%) were more soluble by orders of magni-
tude (>100 lg/mL at pH 7.4). It appeared that these
analogues possessed minimal solubility but that it could
be sufficient for reasonable oral absorption. Passive per-
meability determined on a limited number of analogues
using CACO-2 monolayers was generally consistent
with the inverse correlation between F and PSA (see
Table 3). Thus, those compounds with good permeabil-
ity (Papp > 10 · 10À6 cm/s) were orally bioavailable
(F P 20%) while those with lower permeability
(Papp < 10 · 10À6 cm/s) were poor (F 6 5%). Interesting-
ly, compounds 14, 19, and 20 exhibited poor oral bio-
availability but had CACO-2 Papp values that are
often consistent with good absorption.12 Apparently,
the low solubility of analogues in this series led to a
more stringent requirement for good intestinal
permeability.
The relationship between oral bioavailability and polar
surface area seen in Table 3 was observed generally for
related spirohydantoins. Figure 3 shows a plot of oral
bioavailability in rats against PSA for compounds in
this structural class. For such analogues, it appears that
the probability of good oral bioavailability is much bet-
2
˚
ter for compounds with PSA < 130 A than for those
˚
2
compounds with PSA > 140 A , consistent with pub-
lished data.10 The recognition that membrane perme-
ability was a key limitation to oral absorption in this
structural class, combined with the use of PSA as a read-
ily calculated surrogate for permeability, facilitated the
design of orally bioavailable spirohydantoin CGRP
antagonists.
40
30
20
10
0
Compound 12, which possessed the best combination of
oral bioavailability in rats and functional potency in this
series, was selected for more detailed pharmacokinetic
characterization. A summary of these data is shown in
Table 4. In analogy with the observations in rats, 12
was orally bioavailable in dogs (F = 83%) and rhesus
(F = 29%), and exhibited low plasma clearance in all
three species. Importantly, compound 12 was not an
inhibitor of CYP3A4, CYP2C9, or CYP2D6
(IC50 > 10 lM for these enzymes).
110
120
130
140
150
160
170
PSA (Å2)
Figure 3. Oral bioavailability in rats (%) versus calculated polar
2
surface area (A ) for spirohydantoin CGRP antagonists. Oral
˚
The synthesis of compound 12, which is representative
of the methodology used to prepare the compounds
described herein, is shown in Scheme 1.14
bioavailability and polar surface area determined as described in
Table 3. The graph represents the results for 39 benzoxazolinone and
benzimidazolinone spirohydantoin analogues.
Table 4. Pharmacokinetic data for compound 12
Species
po dosea (mpk)
iv doseb (mpk)
F (%)
iv t1/2 (h)
Cl (mL/min/kg)
Vdss (L/kg)
Rat
Dog
10
2
2
36
83
29
1.6
4.8
9.6
3.4
1.7
0.47
0.69
0.31
0.5
0.5
Rhesus
2
0.85
a Compound dosed in 1% methylcellulose.
b Compound dosed in DMSO.

I. M. Bell et al. / Bioorg. Med. Chem. Lett. 16 (2006) 6165–6169
6169
H
N
oral absorption and consideration of calculated polar
surface area values facilitated the design of orally bio-
available compounds. These studies led to compound
12, which combined good potency with excellent phar-
macokinetics in three species.
O
O
N
H
a
d,e
O
O
HN
O
NH
Acknowledgments
N
Ot-Bu
N
O
Me
The authors thank Douglas Mitchell for synthesis of sev-
eral compounds; Theodore Detwiler, Elizabeth Landis,
Yvonne Leonard, Maria Stranieri, and Audrey Wallace
for animal dosing; Matt Zrada for solubility determina-
tions; Susan Crathern for CYP inhibition studies; the
MRL West Point analytical chemistry, mass spectrome-
try, and NMR spectroscopy groups for analytical and
spectroscopic data; and Chris Burgey, Shane Roller,
Hal Selnick, and Mike Wood for helpful discussions.
b
f,g
O
N
NH
O
N
O
N
O
O2N
N
Ot-Bu
Me
c
h
O
N
References and notes
NH
O
N
O
N
O
H2N
N
1. Goadsby, P. J.; Lipton, R. B.; Ferrari, M. D. N. Eng. J.
Med. 2002, 346, 257.
OH
Me
2. Edvinsson, L. Pharmacol. Toxicol. 2001, 89, 65.
3. Goadsby, P. J. Drugs 2005, 65, 2557.
i
4. Olesen, J.; Diener, H.-C.; Husstedt, I. W.; Goadsby, P. J.;
Hall, D.; Meier, U.; Pollentier, S.; Lesko, L. M. N. Eng. J.
Med. 2004, 350, 1104.
5. Iovino, M.; Feifel, U.; Yong, C.-L.; Wolters, J.-M.;
Wallenstein, G. Cephalalgia 2004, 24, 645.
6. 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.
7. Burgey, C. S.; Stump, C. A.; Nguyen, D. N.; Deng, J. Z.;
Quigley, A. G.; Norton, B. R.; Bell, I. M.; Mosser, S. D.;
Salvatore, C. A.; Rutledge, R. Z.; Kane, S. A.; Koblan, K.
S.; Vacca, J. P.; Graham, S. L.; Williams, T. M. Bioorg.
Med. Chem. Lett. 2006, 16, 5052.
O
N
O
NH
N
O
N
N
O
N
H
Me
12
Scheme 1. Synthesis of compound 12. Reagents and conditions: (a)
NaH, tert-butyl bromoacetate, DMF, 0 ꢁC, 27%; (b) 2-bromopyridine,
Cu, CuCl, KOAc, py, 100 ꢁC, 84%; (c) HCl, EtOAc, 0 ꢁC, 100%; (d)
MeNH3Cl, KCN, MeOH, H2O, 0 ꢁC to rt, then HCl, Et2O, 75%; (e)
KOCN, AcOH, H2O, then HCl, H2O, 79%; (f) 90% HNO3, 60%; (g)
ChiralPak AD, 90:10 CH3CN/MeOH, second major peak is (R)-
enantiomer; (h) H2, Pd/C, EtOAc, MeOH, 100%; (i) EDC, HOBT,
DIEA, DMF, 89%.
8. Bell, I. M.; Stump, C. A. WO 2006/029153.
9. In related analogues, methylation of the imide nitrogen of
the spirohydantoin led to a significant loss of potency
(data not shown).
10. Clark, D. E. J. Pharm. Sci. 1999, 88, 807.
11. Prueksaritanont, T.; Meng, Y.; Ma, B.; Leppert, P.;
Hochman, J.; Tang, C.; Perkins, J.; Zrada, M.;
Meissner, R.; Duggan, M. E.; Lin, J. H. Xenobiotica
2002, 32, 207.
12. Sun, D.; Yu, L. X.; Hussain, M. A.; Wall, D. A.; Smith, R.
L.; Amidon, G. L. Curr. Opin. Drug Discov. Devel. 2004, 7,
75.
13. Martinez, M. N.; Amidon, G. L. J. Clin. Pharmacol. 2002,
42, 620.
Most of the carboxylic acid intermediates containing
benzimidazolinone or benzoxazolinone moieties were
synthesized using standard alkylation or arylation chem-
istry, such as the sequential derivatization of both benz-
imidazolinone nitrogens illustrated in Scheme 1. The
spirohydantoins were derived from 2-indanone via
Bucherer–Bergs methodology, followed by nitration
with fuming nitric acid. Separation of the enantiomers
of the 5-nitroindane intermediate was achieved using a
ChiralPak AD column, to provide the active (R)-enan-
tiomer.15 Following reduction to the corresponding ani-
line, EDC-mediated coupling with the carboxylic acid of
interest provided the final compound.
14. All final compounds were characterized by 1H NMR,
HPLC, and HRMS. Additional synthetic details are
provided in: Bell, I. M.; Gallicchio, S. N.; Theberge, C.
R.; Zhang, X.-F.; Stump, C.; Zartman, C. B. WO 2004/
082605.
In conclusion, a novel series of spirohydantoin-based
CGRP receptor antagonists was identified using a rapid
analogue strategy. In this series of compounds, mem-
brane permeability was apparently a key limitation to
15. Absolute stereochemistry established by X-ray crystallo-
graphy on a derivative of the corresponding aniline.