Identification of a 5-HT4 receptor antagonist clinical candidate through side-chain modification

Article abstract of DOI:10.1016/j.bmcl.2005.01.039

Replacement of the N-butyl side-chain of lead 5-HT₄ receptor antagonist 2 with propanesulfonylpiperidinyl, morpholinyl, and piperazinyl groups led to higher affinity analogs 4-6. In vitro drug metabolism screens and cassette pharmacokinetic studies in the dog led to identification of the N-methylpiperazinyl analog (6b), which displayed pharmacokinetic, selectivity, and safety parameters sufficient for advancement to the clinic for the treatment of urinary incontinence.

Full text of DOI:10.1016/j.bmcl.2005.01.039

Bioorganic & Medicinal Chemistry Letters 15 (2005) 1697–1700
Identification of a 5-HT receptor antagonist clinical
4
candidate through side-chain modification
a,
Robin D. Clark, Alam Jahangir, Muzaffar Alam, Cynthia Rocha,
a
a
b, ,
*
b
*
Lin Lin,
Bodil Bjorner, Khanh Nguyen, Carole Grady, Timothy J. Williams, George Stepan,
b
b
b
c
c
c
Hai Ming Tang and Anthony P. D. W. Ford
c
a
Departments of Medicinal Chemistry, Roche Palo Alto, 3431 Hillview Avenue, Palo Alto, CA 94303, USA
Drug Metabolism and Pharmacokinetics, Roche Palo Alto, 3431 Hillview Avenue, Palo Alto, CA 94303, USA
b
c
Pharmacology, Roche Palo Alto, 3431 Hillview Avenue, Palo Alto, CA 94303, USA
Received 13 December 2004; revised 13 January 2005; accepted 18 January 2005
4
Abstract—Replacement of the N-butyl side-chain of lead 5-HT receptor antagonist 2 with propanesulfonylpiperidinyl, morpholi-
nyl, and piperazinyl groups led to higher affinity analogs 4–6. In vitro drug metabolism screens and cassette pharmacokinetic studies
in the dog led to identification of the N-methylpiperazinyl analog (6b), which displayed pharmacokinetic, selectivity, and safety
parameters sufficient for advancement to the clinic for the treatment of urinary incontinence.
Ó 2005 Elsevier Ltd. All rights reserved.
Advances in lead identification methodologies (medium
and high-throughput screening, molecular modeling,
etc.) have in many cases shifted the bottle neck in drug
discovery to the lead-optimization stage. High-affinity
leads are commonly identified early in medicinal chemis-
try programs; however, the achievement of acceptable
pharmacokinetic parameters, in particular oral bioavail-
ability, remains an often elusive and time-consuming
goal. Herein we describe the optimization of pharmaco-
kinetic properties through modification of side-chain
structure in a series of 5-HT4 receptor antagonists,
which allowed clinical candidate selection in a timely
manner.
Our initial clinical candidate, the aryl ketone sulamserod
3
(1), demonstrated good oral pharmacokinetics and
safety in Phase 1 clinical trials in man. However, due
to an unacceptable toxicity finding in the long-term
safety studies in the dog, which could not be easily moni-
tored in humans, further development of sulamserod
was not pursued. The search for a Ôsecond generationÕ
clinical candidate was centered on a related series based
on the lead benzodioxane carboxamide 2. The related
4
compound 3 (SB-207266, piboserod), has been studied
in the clinic for the treatment of irritable bowel syn-
5
6
drome, and atrial fibrillation.
The structure–activity relationships and potential clini-
cal utility of 5-HT receptor antagonists have been
reviewed. Our primary interest was in developing an
O
O
O
Cl
4
N
H
1
N
N
H N
O
NHSO Me
2
2
agent for the treatment of urinary incontinence (unsta-
ble bladder) on the basis of the apparent relationship be-
tween 5-HT receptor-mediated release of acetylcholine
O
O
1
2
O
4
2
,3
in detrusor smooth muscle and bladder contractility.
N
H
N
N
O
3
*
Corresponding authors. Tel.: +1 808 332 5449; fax: +1 808 332 0389
R.D.C.); e-mail addresses: robin.clark5@verizon.net; cynthia.
rocha@serono.com
(
Although compound 2 had high 5-HT receptor affinity
4
Present address: Instituto di Ricerche Biomediche ÔAntoine MarxerÕ,
RBM S.p.A., Via Ribes 1, 10010 Colleretto Giacosa (TO) Italy.
(pK 9.1), its generic (nonproprietary) structure, coupled
i
with poor pharmacokinetics (t1/2 < 1 h, 18% F in the rat)
0
960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.01.039

1698
R. D. Clark et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1697–1700
dictated structural modification. Previous work had
demonstrated steric tolerance in the piperidine N-substi-
tuent of 1; hence, this was an attractive position for
modification. A series of compounds 4–6 was selected
based on novelty, synthetic accessibility, and opportu-
nity for variation in the pendant piperidine and pipera-
zine rings.
with a general trend toward higher binding with in-
creased size/lipophilicity of the piperidine and piperazine
4-substituents, culminating in the cyclopentyl deriva-
tives 4d and 6g with binding constants of 10.6 and
10.8, respectively. However, there was a trend toward
decreased microsomal stability for these more lipophilic
analogues.
3
In addition to the liver microsomal data, pharmaco-
kinetic parameters determined using cassette dosing in
the dog were used to identify the most promising com-
pounds for further evaluation. Under the cassette
screening protocol, 7–12 compounds were coadminis-
tered orally or intravenously to two to four dogs per
dose route in a parallel study design. Each dog received
0.25–0.5 mg/kg/compound for a combined drug load of
O
N
H
O
O
N
S
O
N
O
X
4
X = CRR1,
X = O
X = NR
5
6
6
6 mg/kg/study. Serial blood samples were collected
through 24 h post dose from each animal and the result-
ing plasma for each dose route was pooled by time point
prior to quantitative analysis by LC–MS/MS. Pharma-
cokinetic parameters were determined using noncom-
partmental analysis. Selected parameters are presented
in Table 1. Under this protocol, many compounds dem-
onstrated good estimated oral bioavailability (%F >50%
in 12 of 17 shown) even though a wide range of in vivo
clearances were observed. The main differentiating fea-
tures were related to differences in persistence and abso-
lute systemic exposure (as shown with parameters t1/2
and Vz).
The series of compounds 4–6 were prepared as described
7
in Scheme 1. Treatment of 1,4-benzodioxan-5-carbonyl
chloride (7) with amine 8, followed by deprotection with
TFA afforded intermediate 9. Alkylation with chlorides
1
3
0 (prepared by reaction of commercially available
-chloropropanesulfonyl chloride with the requisite
piperidine, morpholine, N-alkylpiperazine, or Boc–
piperazine) furnished final compounds 4a–d, 5, and
6
b–i, as well as intermediate 11. Deprotection of the lat-
ter to 6a followed by derivatization of the piperazine
nitrogen gave analogues 6j–q. Analog 6b was more effi-
ciently prepared by reductive alkylation of 6a with
formaldehyde.
By using the in vitro–in vivo screening cascade during
the optimization process, a total of 74 compounds (not
all shown and some from a related series) were rapidly
screened for metabolic stability. Of these, roughly half
(47) advanced to the cassette PK testing in the dog
which identified nine compounds for advancement into
conventional, more time-consuming pharmacokinetic
studies.
Binding affinity for the human cloned 5-HT receptor,
4
dog and human liver microsomal degradation rates,
and dog pharmacokinetic parameters are presented in
Table 1. All analogs demonstrated high binding affinity
O
O
O
O
H N
2
N
On the basis of the totality of the data, the two closely
related analogues 6b–c appeared to have the most prom-
ise and of these, the N-methyl derivative 6b was chosen
for further development. This decision was based pri-
marily on the lower clearance and longer half life of
Cl
8
Boc
N
H
N
a,b
H
O
O
7
9
6
b relative to 6c.
O
O
N
O
O
Cl
S
1
0
X
When examined in a conventional pharmacokinetic
study in the dog (single-compound iv/po cross-over, ex-
tended to a 72 h collection period), 6b demonstrated a
multiphasic disposition with an initial distribution phase
of approximately 7 h. The inclusion of the late post-dose
time points lead to the improved ability to calculate the
terminal elimination t1/2 (18 h). Mean oral bioavailabil-
ity (n = 4) was 75 %F. When compared to the cassette
dosing regimen, the plasma half life was similar if calcu-
lated within the same time intervals but the shorter col-
lection time used in the screening protocol missed the
terminal elimination phase. Additionally, oral bioavail-
ability, though ranked as good, was underestimated. A
similar underestimate of the t1/2 was observed for the
N-acetyl and N-methanesulfonyl analogs 6j (t1/2 4.6 h)
and 6m (t1/2 3.6 h), respectively, upon oral administra-
tion to the dog under the conventional protocol. Thus
N
H
O
O
c
N
S
N
O
X
4
6
a-d X = CRR1 5 X = O
b-i (X = N-alkyl or aryl) 11 X = NBoc
O
O
N
H
O
O
N
S
b,d
N
O
N
R
6
6
a (R = H) 6b (R = Me)
j-q (R = acyl or sulfonyl)
Scheme 1. Reagents and conditions: (a) 8, TEA, dichloromethane; (b)
O, H
TFA, dichloromethane; (c) TEA, DMF, 80 °C; (d) for 6b: CH
2
2
,
Pd–C, ethanol; for 6j–q: acyl or sulfonyl halide, TEA,
dichloromethane.

R. D. Clark et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1697–1700
4
receptor ligands 4–6
1699
Table 1. SAR of 5-HT
O
O
O
O
O
O
N
H
O
O
N
H
N
S
N
S
N
N
R
N
O
O
4a-d
6a-q
R
R1
a
b
i
c
d
R
R
1
Mp (°C)
pK
dLM
hLM
DogPK
e
n
f
g
h
CL
Vz
t
1/2 (h)
%F
4
4
4
4
5
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
3
a
b
c
H
H
Me
165–168
107–108
124–125
160–165
192–195
92–94
9.7
9.9
160
440
460
405
16
2
14
110
101
250
2
12
7
13
27
33
––
7
0.5
1.5
1.4
0.5
0.7
0.5
87
100
15
––
76
33
46
94
9
Me
H
n-Pr
10.2
10.6
9.4
7
d
Spirocyclopentyl
––
12
8
––
0.6
34
––
1.0
37
6.6
a
b
c
d
e
f
H
9.8
9.6
2
11
4
Me
Et
186–189
220–224
243–245
233–236
271–272
245–256
220–221
20
13
110
22
65
53
210
90
56
50
3
3
12
7
2.2
3.6
5.7
3.6
10.3
10.5
10.5
10.6
10.8
10.3
9.6
6
13
23
14
––
37
27
5
3.1
2.8
3.0
n-Pr
i-Pr
0
7
17
12
22
55
42
39
11
7
7
72
––
42
44
65
51
64
––
90
82
––
––
54
71
i-Bu
––
8
––
––
g
h
i
Cyclopentyl
4-F-phenyl
2-Pyrimidinyl
Acetyl
9.3
1.5
0.3
0.7
0.2
2.9
0.7
0.7
1.3
1.0
7
i
173–177
204–208
148–150
>300
7
j
9.9
7
6
k
1
m
n
o
p
q
2-Furoyl
10.0
9.8
8
2
3-Picolinyl
Methylsulfonyl
n-Propylsulfonyl
i-Propylsulfonyl
4-F-phenylsulfonyl
1-Pyrolidinylsulfonyl
––
7
––
3
––
––
236–237
201–202
165–170
185–186
165–168
9.5
18
68
3
3
0.3
0.7
1.2
0.7
10.0
9.9
34
7
7
12
––
––
15
8
––
––
7
––
––
0.8
5.0
––
––
1.0
7.1
10.5
10.1
9.6
51
39
4
93
36
5
12
a
b
c
d
e
f
Melting point measured on a hot-stage apparatus of the hydrochloride (4a–d, 5, 6j–q) or dihydrochloride salt (6a–h) unless otherwise noted.
Determined in human cloned 5-HT receptors. Mean of two determinations.
4
Intrinsic clearance (lL/min/mg protein) in dog liver microsomes.
Intrinsic clearance (lL/min/mg protein) in human liver microsomes.
Number of compounds dosed in cassette to the dog. Dose of each compound was 0.25 mg/kg iv or 0.5 mg/kg po.
Plasma clearance (mL/min/kg).
g
h
i
Volume of distribution (L/kg).
Half-life upon iv administration.
Free base.
our experience confirms that cassette dosing is useful for
efficient ranking of related compounds rather than
affording absolute pharmacokinetic data. The latter
should be obtained from subsequent single-compound
studies.
pK of 8.95. 5-HT receptor antagonist activity was fur-
b 4
ther confirmed by inhibition of 5-HT-induced tachycar-
9
dia in the anesthetized vagotomized mini-pig (ID50
1.2 lg/kg, iv and 52lg/kg, id).
Profiling against standard receptor, ion channel, and up-
take system panels indicated that 6b was extremely selec-
tive (IC50 > 1 lM) with the exception of affinity for the
Compound 6b demonstrated good exposure and pro-
longed t1/2 in other species, including the mouse (t1/2
1
0
7
h), rat (t1/2 12 h) and mini-pig (t1/2 21 h). The result
central histamine H receptor where an IC of 100 nM
was observed. In theory an antagonist at the H receptor
1
50
in the rat was a dramatic improvement over the original
lead compound 2, which had a t1/2 of less than 1 h. Allo-
metric scaling predicted a t1/2 of 60 h in man. Metabo-
lism of 6b to 6a was observed in vivo, although in
higher species such as the dog this was a relatively minor
phenomenon (ratio of AUC of 6a to 6b = 0.2 following
1
could induce sedation if the antagonist crosses the
blood–brain barrier. However, 6b did not appear to pen-
etrate the CNS in the rat (as determined by quantitative
whole body autoradiography) and sedation was not ob-
served in any species examined, including man. The
safety profile (CYP inhibition, dog Purkinje fiber, Ames
test, etc.) of 6b was also unexceptional, as were toxicol-
ogy studies in the rat and dog; hence this compound was
chosen for clinical study in man.
8
an oral dose).
9
In the rat isolated esophagus, 6b was devoid of 5-HT
receptor agonist activity and was an antagonist with
4

1700
R. D. Clark et al. / Bioorg. Med. Chem. Lett. 15 (2005) 1697–1700
In Phase I clinical trial, 6b was found to have good
apparent oral bioavailability with a steady state plasma
t1/2 of >100 h. Thus the general pharmacokinetic prop-
erties of the molecule targeted by the preclinical lead
optimization effort (high oral bioavailability, suitable
systemic persistence to allow once-daily dosing) were
confirmed in the clinic.
DÕAgostino, G.; Vicini, D.; Tagliani, M.; Tonini, M. Br. J.
Pharmacol. 1996, 118, 1965.
. Clark, R. D.; Jahangir, A.; Flippin, L. A.; Langston, J. A.;
Leung, E.; Bonhaus, D. W.; Wong, E. H. F.; Johnson, L.
G.; Eglen, R. M. Bioorg. Med. Chem. Lett. 1995, 5,
3
4
2
119.
. (a) Gaster, L. M.; Joiner, G. F.; King, F. D.; Wyman, P.
A.; Sutton, J. M.; Bingham, S.; Ellis, E. S.; Sanger, G. J.;
Wardle, K. A. J. Med. Chem. 1995, 38, 4760; (b) Drugs
Future 2000, 25, 1323.
5. Houghton, L. A.; Jackson, N. A.; Whorwell, P. J.;
Cooper, S. M. Aliment. Pharmacol. Ther. 1999, 13,
In summary, modification of the side-chain of the initial
lead compound 2 led to 5-HT4 antagonists with in-
creased affinity and dramatically improved pharmacoki-
netic properties. Cassette pharmacokinetic dosing was
used in concert with microsomal stability studies to rap-
idly differentiate between numerous high-affinity com-
pounds and led to the selection of 6b as a clinical
candidate. Results of a Phase IIa study for the treatment
of urinary frequency and incontinence in women with
overactive bladder will be reported in due course.
1
437.
6
7
. GlaxoSmith Kline company website (www.gsk.com).
. Clark, R. D.; Jahangir, A. U.S. Patent 6,172,062, 2001;
Chem. Abstr. 2001, 132, 237098.
8
. The relatively low basicity (measured pK
methyl substituted piperazine nitrogen in 6b (compared to
the tertiary piperidine nitrogen with a pK of 8.5) could be
a
of 6.6) of the
a
a contributing factor to the low observed rate of N-
demethylation.
9
. Hegde, S. S.; Bonhaus, D. W.; Johnson, L. G.; Leung, E.;
Clark, R. D.; Eglen, R. M. Br. J. Pharmacol. 1995, 115,
References and notes
1
087.
1
. (a) 5HT4 receptors in the brain and periphery; Eglen, R.
M., Ed.; Springer: Berlin, 1998; (b) Langois, M.; Fisch-
meister, R. J. Med. Chem. 2003, 46, 319.
. (a) Ford, A. P. D. W.; Kava, S. M. See Ref. 1a; pp 171–
184; (b) Candura, S. M.; Messori, E.; Franceschetti, G. P.;
10. Receptors included: adenosine, alpha and beta adrenergic,
bradykinin, NE uptake, angiotensin-1, benzodiazepine,
dopaminergic, dopamine uptake, endothelin, GABA,
NMDA, muscarinic, neurokinin, NPY, opiate, sigma,
other serotonergic, and serotonin uptake.
2