Discovery of 3-aryl-5-acylpiperazinyl-pyrazoles as antagonists to the NK3 receptor

Article abstract of DOI:10.1016/j.bmcl.2011.02.033

A series of 3-aryl-5-acylpiperazinyl-pyrazoles (e.g., 3a-b) initially identified through a high-throughput screening campaign using the aequorin Ca²⁺ bioluminescence assay as novel, potent small molecule antagonists of the G protein-coupled hum

Full text of DOI:10.1016/j.bmcl.2011.02.033

Bioorganic & Medicinal Chemistry Letters 21 (2011) 1991–1996
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of 3-aryl-5-acylpiperazinyl-pyrazoles as antagonists
to the NK3 receptor
Hamid R. Hoveyda a, , Marie-Odile Roy a, Sebastien Blanc a, Sophie Noël a, Joseph M. Salvino b, Mark A. Ator b,
⇑
Graeme Fraser a
a Euroscreen SA, 47, rue Adrienne Bolland, Gosselies, Belgium
b Cephalon Inc., Worldwide Discovery Research, 145 Brandywine Parkway, West Chester, PA 19380, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of 3-aryl-5-acylpiperazinyl-pyrazoles (e.g., 3a–b) initially identified through a high-throughput
screening campaign using the aequorin Ca2+ bioluminescence assay as novel, potent small molecule
antagonists of the G protein-coupled human tachykinin NK3 receptor (hNK3-R) is described. Preliminary
profiling revealed poor plasma and metabolic stability for these structures in rodents. Further optimiza-
tion efforts resulted in analogs with improved potency, stability, and pharmacokinetic properties as well
as good brain permeability, for example, compounds 26 and 42. Unexpected cytotoxicity was observed in
such N-Me pyrazole structures as compounds 41–42.
Received 25 December 2010
Revised 6 February 2011
Accepted 8 February 2011
Available online 13 February 2011
Keywords:
NK3 antagonist
Pyrazole
Ó 2011 Elsevier Ltd. All rights reserved.
Antipsychotic
The NK3 receptor (NK3-R) is a GPCR that belongs to the tachyki-
nin receptor family and is preferentially activated by neurokinin B
(NKB).1 NK3-R is expressed centrally in cortical regions and in basal
ganglia structures relevant to psychiatric disorders. Based on pre-
clinical studies, NK3-R has been found to modulate monoaminergic
and GABA and glutamate neurotransmission in the mammalian
brain, thus raising the speculation that NK3-R modulators may be
useful for treatment of psychological disorders such as schizophre-
nia.2 Discovery of osanetant (2) by Sanofi researchers in 1994, and
shortly thereafter, talnetant (1a) by GSK scientists fueled the
search for non-peptidic antagonists to hNK3-R3 and has subse-
quently resulted in the development of many variants of such
structures, especially in the case of talnetant.4
encouraged us to consider this series for further development.
However preliminary profiling results quickly revealed additional
challenges in terms of the instability of analogs 3a–b in rodent
plasma and hepatic microsomal stability assays (cf. Table 1), even
though in human plasma and microsomal incubations these
analogs proved significantly more stable (e.g., 3a and 3b human
liver microsomes (HLM) T1/2 = 55 and 71 min, respectively). In view
of the practical utility of rat animal models in the development
phase, we sought to evaluate whether such hNK3-R antagonist
structures were amenable to improvements in potency as well as
the stability profiles.
The majority of the five-membered heterocyclic acids required
for the present work were directly purchased. Conventional pyra-
zole synthesis through a 1,3-ketoester intermediate followed by
cyclization with hydrazine (or ammonia for the pyrrole acid pre-
cursor6 to 43) and subsequent saponification of the ester afforded
the pyrazole acid building blocks for analogs 22, 24, 27, 29, and 30.
The pyrazole sulfonyl chloride used to prepare analog 37 was made
using an in situ diazotization approach as outlined in Scheme 1
(step f).7 With the heterocyclic acid building blocks in hand, the
preparation of arylpiperazines was carried out using Buchwald
amination (step b, Scheme 2) for the phenylpiperazine analogs,
or via reductive amination (Scheme 3) for analogs 5–9. For struc-
tures with more elaborate ortho-alkoxy substituents (20–21), a
benzyl group protective strategy was employed (Scheme 2). The
syntheses were completed through amide coupling of the five-
membered heterocyclic acids (step f, Scheme 2). Methyl-substi-
tuted piperazines (38–40) and pyrazole ring bioisostere variants
(41–45) were also prepared as per Scheme 2. The versatility of
Compounds 3a–b (Fig. 1) were two of the initial hit structures
detected in our hNK3-R high-throughput screening campaign, using
the aequorin Ca2+ bioluminescence assay.5 Of the two, the ortho
anisyl analog 3b was superior in potency and was further profiled.
Thus 3b, as an initial hit, was found to be reasonably potent in
hNK3-R binding (Ki = 1.1
aequorin assay (IC50 = 480 nM), even though it displayed a >10-fold
right-shift in binding against rNK3-R (rNK3-R Ki = 12 M). Further-
more, 3b displayed a reasonable off-target selectivity profile, both
in terms of hNK-R subtypes (hNK1-R Ki = 6.7 M, hNK2-R
Ki = 12 M) as well as CYP P-450 isoforms (IC50 >15 M for 3A4,
lM) and in its cellular response in the
l
l
l
l
2D6, 1A2, 2C9, and 2C19). These considerations, together with
the structural novelty of such structures as hNK3 antagonists,2–4
⇑
Corresponding author.
E-mail address: hhoveyda@euroscreen.com (H.R. Hoveyda).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.02.033

1992
H. R. Hoveyda et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1991–1996
Cl
Cl
O
H
N
5
(S)
NH
R
N
O
N
3
O
N
N
N
(R)
R
N
N
O
1a: R= OH, Talnetant
1b
2: Osanetant
3a R = H
: R = Me (SB-222200)
3b
R = OMe
Figure 1. Structures of talnetant variants (1a–b), osanetant (2), and the initial hit structures (3a–b).
Table 1
Potency and stability SAR at Ring A
Ring A:
O
H
N
R1
OMe
N
N
6
N
N
3
4
5
4
Ring A
2
3
6
7
= 2-pyridyl
= 3-pyridyl
R2
3-4, 10-14
5
9
8 = 4-pyridyl
R1
R2
hNK3 IC50 (nM)
hNK3 Kib (nM)
RLM T1/2 (min)
%Plasma stabilityc
a
Compd
3a
3b
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
H
H
H
2100
480
130
NDd
1100
250
490
ND
ND
ND
ND
ND
ND
ND
ND
ND
190
ND
ND
ND
400
20
9
5
3
8
15
3
8
2-OMe
2-OMe
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
H
Cl
Cl
Cl
Cl
Cl
Cl
Cl
See diagram
See diagram
See diagram
See diagram
See diagram
3-OMe
4-OMe
2-CF3
2-CH3
2-Ph
2-Et
2-CN
2-OH
2-C(O)NH2
2-OMe, 5-Cl
cf. Scheme 2
cf. Scheme 2
130
46
ND
ND
ND
ND
ND
ND
ND
ND
ND
14
ND
ND
ND
0
5090
3810
1800
>10,000
>10,000
>10,000
>10,000
>10,000
>10,000
180
ND
ND
ND
ND
ND
ND
ND
12
ND
5
22
ND
ND
120
8
480
1300
>10,000
370
60
80
0
48
240
29
a
b
c
Inhibition of NKB-induced aequorin Ca2+ signal in CHO cells expressing recombinant hNK3-R.
Displacement of [3H]-labeled 1b (cf. Fig 1) from recombinant hNK3-R in CHO cells.
Ex vivo rat plasma stability reported as % intact compound at 1 h.
ND = no data.
d
reductive amination chemistry (Scheme 3) that includes ready
access to myriad benzaldehyde building blocks was exploited
using parallel synthesis to rapidly scan the potency SAR through
benzylpiperazine (e.g., 5) variants, with the most active analogs
thus discovered confirmed subsequently in the corresponding phe-
nylpiperazine (e.g., 4) structure as well.
For brevity each of the four rings in the lead structure is anno-
tated alphabetically (thus ‘Rings A–D’), as defined in the illustra-
tions that accompany Tables 1–4, and are used throughout the text.
In an early follow-up to the discovery of compound 3b as an
hNK3-R antagonist hit, we observed that para chloro-substitution
in Ring D (i.e., R2 = Cl in 4) further improved potency appreciably
both in terms of binding and cellular response (cf. Table 1).8 More-
over, 4 displayed a less right-shifted binding potency in rNK3-R
appeared to maintain the good off-target selectivity profile exhib-
ited by 3b, both in terms of hNK-R subtypes (hNK1-R: Ki = 1.8 M,
hNK2-R Ki = 8 M) as well as CYP P-450 profile (IC50 >15 M for
3A4, 2D6, 1A2, 2C9, and 2C19 isoforms). However, despite the over-
all improved in vitro bioactivity profile, 4 like 3b proved unstable,
especially in rat plasma and microsomal assays (cf. Table 1). Thus
it remained to establish whether the instability issues observed in
rodent assays in the more potent analog 4 could be improved. An
additional early key observation was that benzylpiperazine analogs
such as 5 were potent antagonists of hNK3-R (cf. Table 1), but not of
l
l
l
rNK3-R (Ki = 4
proved reasonably selective against hNK-R subtypes (hNK1-R:
Ki = 1.9 M, hNK2-R Ki = 19 M), but it was less attractive than con-
gener 4 in terms of its CYP P-450 profile with the IC50 values for
three key isoforms 3A4, 2D6, and C9 determined as 8, 3, and 3 M,
lM). In terms of the initial off-target selectivity, 5
l
l
(Ki = 250 nM) versus analog 3b (Ki = 12
lM). Finally, analog 4
l

H. R. Hoveyda et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1991–1996
1993
O
CO2H
NH
O
O
a
b-c
used to prepare precursors
R
to analogs 22, 24, 27, 29 and 30
R
CO2Et
R
N
O
O
O
CO2H
Br
e, c
d
CO2Et
Cl
Cl
N
H
Cl
O
S
O
Ph
O
Ph
g
f
MeO
O
N
CO2Et
S
Ph
N
N
N
N
H
Cl
N
H
NH
N
37
Scheme 1. Preparation of five-membered heterocyclic acids and sulfonyl chloride derivative. Reagents and conditions: (a) diethyloxalate, t-BuOK, toluene, rt, 8 h; (b)
hydrazine, EtOH, reflux, 20 h; (c) NaOH, water/EtOH (1:1), reflux, 3 h; (d) ethyl acetoacetate, THF; (e) NH3, acetic acid, 80 °C, 1 h; (f) concd HCl, 0 °C then CuCl2, AcOH, SO2 at
rt; (g) o-anisyl-piperazine, DIEA, MeCN.
O
R
O
R
H
N
O
N
f
20 R=
21 R=
a-e
N
N
N
NH
F
OH
NH2
Br
O
Cl
Scheme 2. Preparation of phenylpiperazine analogs. Reagents and conditions: (a) BnBr, K2CO3, 18-crown-6, MeCN, reflux, 20 h; (b) N-Boc-piperazine, X-Phos, Pd(OAc)2,
Cs2CO3, dioxane, reflux, 20 h; (c) 5% Pd/C, H2, rt, 2 h; (d) 2-chloroacetamide or 1-(bromomethyl)-3-fluorobenzene, K2CO3, 18-crown-6, MeCN, reflux, 20 h; (e) 4 M HCl/
dioxane, DCM, rt, 5 h; (f) pyrazole acid, HATU, NMM, DMF, 45 °C, 20 h.
O
H
Boc
N
5
6
7
8
9
R= 2-anisyl
R= 2-pyridyl
R= 3-pyridyl
R= 4-pyridyl
R= cyc -hexyl
N
N
O
b-c
N
a
N
R
N
R
R
Cl
Scheme 3. Preparation of analogs 5–9. Reagents and conditions: (a) NaBH(OAc)3, N-Boc-piperazine, DCE/DCM, rt, 20 h; (b) 4 M HCl/dioxane, DCM, rt, 5 h; (c) phenyl-
heterocycle-acid, HATU, NMM, DMF, 45 °C, 20 h.
respectively. Furthermore, 5 offered no advantages in terms of rat
plasma stability versus 4 and also displayed a short half-life of
8 min in the rat liver microsome assay (cf. Table 1). Thus a greater
preference was accorded to the phenylpiperazine analogs, exempli-
fied by 4, which also contain one less rotatable bond than 5,9
whereas the benzylpiperazine variants10 were utilized for a rapid
initial scan of the potency SAR through generation of focused
libraries.
Reduction of the aryl ring to cyclohexyl at Ring A resulted in an
inactive structure (e.g., 9 vs 5, Table 1), while reduction of Ring D
caused a significant (Pfourfold) right shift in potency (e.g., 35 vs
3b, Tables 1 and 2), indicating a preference for an aromatic Rings
A–D combination for good hNK3-R antagonist potency. Contrary
to the phenyl ring structures, pyridyl ring variants failed to furnish
potent antagonists when applied either at Ring A (6–8 vs 5, Table
1) or at Ring D (32–34 vs 4, Tables 1 and 2). However, exploration
of the impact of aromatic ring substitution in Rings A and D in
terms of hNK3-R antagonist potency SAR proved more fruitful. As
mentioned above, ortho anisyl Ring A had a significant (ꢀfourfold)
impact on potency as revealed by a comparison of IC50 values for
3a versus 3b (cf. Table 1). Indeed ortho substitution in Ring A
was typically preferred over meta or para analogs as exemplified
by the three anisyl Ring A regioisomers (4, 10–11). In terms of
the nature of ortho substituents, the potency SAR results did not
correlate either with the electronic or the lipophilicity properties
of the substituent. For example, both ortho methyl and ortho CF3
moieties resulted in inactive compounds (12, 13), yet both ortho
cyano (16) and ortho ethyl (15) Ring A analogs yielded compara-
tively active antagonists (cf. Table 1). Comparison between ortho
phenol (17) versus ortho anisyl (4) congeners and between ortho
methyl versus ethyl analogs (13 vs 15) seemed to suggest that
for optimal receptor interaction, the ortho Ring A side-chain must
be comprised of two heavy (non-hydrogen) atoms. Subsequently,
the latter structural feature was further investigated using a wide
range of ortho Ring A substituents bearing a rotatable bond (com-
prised of two non-hydrogen atoms), initially through the benzylpi-
perazine variants, as discussed above. The most active analogs
were eventually confirmed in the analogous phenylpiperazine
structures, such as with 20–21, hNK3-R IC50 <90 nM (Table 1).
The significant impact of the presence of a rotatable bond linker
at the ortho substituent in Ring A in terms of antagonist potency
was again apparent by comparing analogs 20 versus 14 and 21 ver-
sus 18. In each case the absence of a rotatable bond resulted in an
inactive analog (14, 18: IC50 >10 lM), in sharp contrast to the po-

1994
H. R. Hoveyda et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1991–1996
Table 2
Potency and stability SAR at Ring D
Ring D:
O
H
N
OMe
N
N
R2
2
N
6
3
Ring D
5
4
22
23-34
35
Compd
R2
hNK3 IC50 (nM)
hNK3 Ki (nM)
RLM T1/2 (min)
%Plasma stability
22
23
24
25
26
27
28
29
30
31
32
33
34
35
See diagram
2-F
3-F
4-F
4-CF3
4-OCF3
4-CH3
4-SO2Me
4-CN
2,4-DiCl
2-N
3300
1200
580
240
42
130
390
2500
208
205
ND
ND
ND
ND
4
5
5
ND
ND
8
45
ND
ND
ND
ND
ND
ND
ND
18
55
59
ND
ND
91
2500
1600
530
120
310
510
ND
700
260
3600
ND
5
3100
7100
7200
2000
ND
ND
ND
ND
3-N
4-N
See diagram
ND
2900
Table 3
Potency and stability SAR at Ring B
Ring B:
O
O
H
O
O
O
O
N
N
O
N
S
N
N
N
N
*
N
N
N
N
N
Ring B
O
36
37
38 (S)-Me
39 (R)-Me
40
R2
Compd
R2
hNK3 IC50 (nM)
hNK3 Ki (nM)
RLM T1/2 (min)
%Plasma stability
36
37
38
39
40
Cl
H
Cl
Cl
Cl
3100
7400
190
2900
93
ND
ND
370
ND
240
ND
ND
2
ND
147
ND
ND
100
ND
48
tent analogs bearing similar side-chains but attached through a
rotatable bond linker (20–21: IC50 <90 nM). In addition to im-
proved potency, the ca. tenfold improved rodent metabolic stabil-
ity in analog 21 (c log P = 2.8) versus 4 (c log P = 4.4)11 is consistent
with the reduced lipophilicity in 21 through introduction of the po-
lar acetamide side-chain. In addition, we observed that metabolic
blocking through chloro-substitution in Ring A para to the elec-
tron-rich ortho anisyl moiety also imparted significant metabolic
stability to the attendant molecule (cf. 19 RLM T1/2 = 120 min in Ta-
ble 1), although the rodent plasma instability issue remained unaf-
fected. Overall, these results demonstrated the utility of Ring A to
improve potency and rodent metabolic stability through either
reduction of lipophilicity and/or through metabolic blocking
strategies.
In terms of monosubstituted Ring D analogs, para substitution
was typically optimal for antagonist potency, as exemplified
through fluoro-substituted congeners (23–25, Table 2). In stark
contrast with Ring A, for Ring D a clear SAR was observed in that
the most potent analogs appeared to be confined to those possess-
ing lipophilic and electron-withdrawing substituents. For example,
para-CF3 analog 26 (IC50 = 42 nM) was nearly tenfold more potent
than the para toluyl congener (28, Table 2). Likewise, para-OCF3
analog 27 proved significantly more potent than 29–30 that bear
polar electron-withdrawing substituents (thus not lipophilic).
Interestingly, analogs 26–27 and 30 with para electron-withdraw-
ing substituents in Ring D displayed improved ex vivo plasma sta-
bility but not microsomal metabolic stability in rats. Moreover,
further reduction of electron density in phenyl Ring D through di-
chloro-substitution at both ortho and para phenyl ring positions
(31, Table 2) resulted in significant improvement in rat microsomal
half-life, albeit with little impact either on rodent plasma stability,
or on hNK3-R antagonist potency (vs 4). Thus collectively, the fore-
going results on aryl Rings A and D demonstrate that substitution
on these aromatic rings plays an important role in modulation of
potency and stability profiles.
Concerning the SAR for the piperazine ring (Ring B) and the
associated amide bond, the following serve as the key highlights
(cf. Table 3). Analog 36, wherein the piperazine ring was replaced
with a ketopiperazine ring, proved to be ꢀ24 times less potent
than 5; this established the importance of the sp3 hybridized basic

H. R. Hoveyda et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1991–1996
1995
Table 4
Potency and stability SAR at Ring C
Ring C:
O
O
O
N
N
O
Ring C
N
NH
H
N
N
O
N
N
41
42
43
N
O
O
O
N
N
O
R2
44
45
Compd
R2
hNK3 IC50 (nM)
hNK3 Ki (nM)
RLM T1/2 (min)
%Plasma stability
41
42
43
44
45
Cl
Cl
Cl
H
180
180
280
>5000
1100
540
370
350
ND
44
40
17
ND
17
100
100
100
ND
78
Cl
1800
Table 5
Comparative evaluation of the in vitro and in vivo profiles of analogs 1a, 2 versus 41–42
1a
2
41
42
Subtype selectivity, Kia (nM)
rNK3
hNK3
hNK2
hNK1
190
7.4
3700
>10,000
57
1.9
85
3700
540
10,000
2200
3400
370
13,000
1600
330
CYP P-450, IC50
3A4, 2D6, 1A2
2C9, 2C19
(lM)
4, 18, 5
2, 9
82%
1, 55, >100
11, 7
97%
12, 29, 21
3, 6
6.7%
20, 42, 34
4, 9
8.5%
% Cell viability (HepG2)b
Stability and protein binding (rat)
RLM, T1/2 (min)
60
23
44
40
Plasma stability (at 60 min)
% Plasma protein binding
100%
99.9
NA
99.4
100%
99.5
100%
99.8
Pharmacokinetics (rat)c
CL (mL/min/kg)
Vdss (L/kg)
T1/2 (min)
AUC oral dosing (ng min/mL)
%F
1.9
1.4
450
933,806
53
57
8.7
110
6856
17
13
7.3
3.3
310
90,501
53
4.9
270
ND
NA
Brain penetration (mouse)d
Brain:plasma ratio (at 10 min)
Brain concn (ng/g) at 10 min
0.07
340
0.03
30
0.88
830
0.64
1700
a
b
c
Radioligand binding assay.
Measured at 50 lM test concentration after 24 h.
Male Sprague–Dawley rats at 1 mg/kg iv and 3 mg/kg po in physiological saline with 9% 2-hydroxypropyl-b-cyclodextrin (HP-b-CD).
Male CD-1 mice at 5 mg/kg ip in physiological saline with 9% HP-b-CD.
d
nitrogen in piperazine for optimal receptor interactions. Replacing
the piperazine amide juncture with sulfonamide (37) resulted in
ꢀ15-fold right shift in potency (vs 3b, Table 1). Moreover, methyl
substitution on the piperazinyl ring provided further noteworthy
results summarized below. Thus methyl substitution at the carbon
adjacent to the amide piperazine nitrogen displayed a stereospe-
cific potency SAR with the (S)-Me enantiomer (38, Table 3) proving
to be ꢀ15-fold more potent than its antipode (39) and nearly equi-
potent to the reference analog 4 (Table 1). Importantly, analog 38
was found to be fully stable in rat plasma contrary to 4 that was
rather unstable in rat plasma. Nonetheless, the rat microsomal sta-
bility remained comparable between the latter analogs (RLM T1/2
2–3 min). The improved rat plasma stability in 38 may stem from
steric occlusion of the potentially hydrolyzable amide linkage.12
Interestingly, analog 40 (RLM T1/2 = 147 min) proved more robust
than congener 38 in terms of rat microsomal stability whereas
its rat plasma stability profile, while significantly better than 4,
was inferior to 38. Analog 40 was not pursued further due to infe-
rior plasma stability; nonetheless, these results made abundantly
clear that the piperazine ring modifications provide yet another
avenue for advanced lead optimization in the foregoing hNK3-R
antagonist series.
Principle features of the SAR for the five-membered heteroaryl
ring segment in the current lead structures are summarized in Ta-
ble 4. Replacing the pyrazole ring with an isoxazole ring resulted in
Peightfold loss of potency irrespective of the regioisomerism (cf.
44 vs 3b and 45 vs 4). Yet the 2-methyl-pyrrole bioisostere (43)
was merely twofold right-shifted in potency (vs 4, Table 1). More-
over, analog 43 proved devoid of rat plasma stability issues and
displayed moderately improved rat microsomal half-life of

1996
H. R. Hoveyda et al. / Bioorg. Med. Chem. Lett. 21 (2011) 1991–1996
17 min. The results for analog 43, together with the relatively good
hNK3-R antagonist potency in the N-methyl pyrazole regioisomers,
41–42, tend to suggest that the diminished potency in isoxazole
isosteres is unrelated to the absence of the hydrogen bond donor
present in an unsubstituted pyrazole ring such as 4. In fact, the best
analogs in terms of the overall potency and stability profiles turned
out to be the N-methyl pyrazole analogs, 41–42. The improved
plasma stability in N-methyl pyrazole analog 41 recalls the effect
of methyl piperazine substitution in 38 that was suggested to be
at least partially due to steric occlusion nearby the amide linker
(vide supra). While the hNK3-R antagonist potency in the latter
analogs only marginally differed from the reference analog 4, the
rat plasma and microsomal stability profiles in 41 and 42 were sig-
nificantly ameliorated. Thus both analogs proved to be completely
stable in rat plasma in addition to displaying T1/2 >40 min in the rat
microsomal assay.
A comparative evaluation of analogs 41–42 against talnetant
(1a) and osanetant (2) is provided in Table 5.13 Among all analogs,
a right shift was observed in terms of human versus rodent NK3-R
binding potency, with talnetant showing >25-fold shift vis-à-vis
ꢀ30-fold shift for osanetant and ꢀ10-fold shift for 41–42. In addi-
tion, for all four analogs the best NK-subtype selectivity is that of
hNK3:hNK2 with talnetant being the most selective (500-fold),
while osanetant and analogs 41–42 display a ꢀ20- to 50-fold selec-
tivity. In terms of hNK3:hNK1 selectivity the following trend is dis-
cernible: talnetant (>1000-fold) > osanetant (174-fold) > 41–42 (4-
to 7-fold). However, in terms of CYP P-450 profile, analogs 41–42
proved superior to talnetant or osanetant. In addition, 41 and 42
were cleared from systemic circulation in rats at a reasonably
low rate, 3- to 6-fold more rapidly than talnetant14 and signifi-
cantly superior to the essentially flow-limited clearance rate for
osanetant. Compounds 41–42 also displayed distribution volumes
5- to 10-fold that of total body water, and were present at 0.5% and
0.2% fraction unbound based on plasma protein binding data in
rats. Analog 42 was also found to be 53% orally bioavailable in
rat. Indeed both 41 and 42 were found to be highly CNS-penetrant,
superior in this respect to both talnetant and osanetant, both in rel-
ative (brain-to-plasma ratios >0.6) and in absolute terms (>800 ng/
g at 10 min). Overall these analogs displayed promising pharmaco-
kinetic and brain permeability profiles in rodents and as such ap-
pear to be viable entities for further lead development towards
CNS therapeutic agents.
doned in view of the cytotoxicity issue encountered in a subset of
these structures (namely 41 and 42) and further prompted by dis-
covery of other more promising lead series in this program, which
will be reported in due course.
Acknowledgment
We are grateful to Dr. Vladimir Trukhan and his team at the
ChemBridge Corporation, Moscow, Russia, for executing the syn-
thetic chemistry in support of this project.
References and notes
1. Spooren, W.; Riemer, C.; Meltzer, H. Nat. Rev. Drug Disc. 2005, 4, 967.
2. Smith, P. W.; Dawson, L. A. Recent Patents CNS Drug Disc. 2008, 3, 1.
3. Giardina, G. A. M.; Raveglia, L. F. Expert Opin. Ther. Patents 1997, 7, 307.
4. See for example: (a) Smith, P. W.; Wyman, P. A.; Lovell, P.; Goodacre, C.;
Serafinokwska, H. T.; Vong, A.; Harrington, F.; Flynn, S.; Bradley, D. M.; Porter,
R.; Coggon, S.; Murkitt, G.; Searle, K.; Thomas, D. R.; Watson, J. M.; Martin, W.;
Wu, Z.; Dawson, L. Bioorg. Med. Chem. Lett. 2009, 19, 837; (b) Elliott, J. M.;
Carling, R. W.; Chambers, M.; Chicchi, G. G.; Hutson, P. H.; Jones, A. B.; MacLeod,
A.; Marwood, R.; Mensese-Lorente, G.; Mezzogori, E.; Murray, F.; Rigby, M.;
Royo, I.; Russell, M. G. N.; Sohal, B.; Tsao, K. L.; Williams, B. Bioorg. Med. Chem.
Lett. 2006, 16, 5748; (c) Xiong, H.; Kang, J.; Woods, J. M.; McCauley, J. P., Jr.;
Koether, G. M.; Albert, J. S.; Hinkley, L.; Gadient, R. A.; Simpson, T. R. Bioorg.
Med. Chem. Lett. 2010. doi:10.1016/j.bmcl.2010.11.003.
5. Le Poul, E.; Hisada, S.; Mizuguchi, Y.; Dupriez, V. J.; Burgeon, E.; Detheux, M. J.
Biomol. Screen. 2002, 7, 57. Euroscreen has developed stably transfected cell
lines with plasmids encoding apoaequorin, the GPCR of interest (e.g., hNK3-R),
and, if needed, a recombinant G-protein to redirect receptor coupling towards
Ca2+ signaling. Upon receptor activation, the change in intracellular [Ca2+
]
causes a conformational change in the aequorin protein thereby eliciting a
quantifiable luminescent signal. This method is adapted to detect agonist,
antagonist or allosteric modulators of receptor activity.
6. Petruso, S.; Caronna, S.; Sferlazzo, M.; Sprio, V. J. Heterocycl. Chem. 1990, 27,
1277.
7. Bellemin, R.; Festal, D. J. Heterocycl. Chem. 1984, 21, 1017.
8. Given the good agreement between the radioligand binding and aequorin
functional assays, the latter set of results will be principally used in foregoing
SAR discussions.
9. Veber, D. F.; Johnson, S. R.; Cheng, H. Y.; Smith, B. R.; Ward, K. W.; Kopple, K. D.
J. Med. Chem. 2002, 45, 2615.
10. No appreciable solubility difference between 4 and 5 was detectable at pH 7,
although at pH 5 benzylpiperazine analog 5 was decidedly more soluble.
11. C log P values cited herein are calculated using ChemBioDraw (CambridgeSoft,
v.12).
12. Kerns, E. H.; Di, L. Drug-like Properties: Concepts, Structure Design and Methods:
From ADME to Toxicity Optimization; Academic Press: Burlington,
Massachusetts, 2008. Chapter 12.
13. Talnetant was made and characterized as per: (a) Giardina, G. A. M.; Sarau, H.
M.; Farina, C.; Medhurst, A. D.; Grugni, M.; Foley, J. J.; Raveglia, L. F.; Schmidt, D.
B.; Rigolio, R.; Vassallo, M.; Vecchietti, V.; Hay, D. W. P. J. Med. Chem. 1996, 39,
2281; Osanetant was purchased from Charnwood Molecular, UK. It was made
and characterized as per: (b) Giardina, G. A. M.; Grugni, M.; Rigolio, R.; Vassllo,
M.; Erhard, K.; Farina, C. Bioorg. Med. Chem. Lett. 1996, 6, 2307; (c) Chen, H. G.;
Chung, F.-Z.; Goel, O. P.; Kesten, J. S.; Knobelsdorf, J.; Lee, H. T.; Rubin, J. R.
Bioorg. Med. Chem. Lett. 1997, 7, 555.
14. For a report on the pharmacokinetic profile of talnetant, see Tables 2 and 3 in:
Sarau, H. M.; Griswold, D. E.; Potts, W.; Foley, J. J.; Schmidt, D. B.; Webb, E. F.;
Martin, L. D.; Brawner, M. E.; Elshourbagy, N. A.; Medhurst, A. M.; Giardina, G.
A. M.; Hay, D. W. P. J. Pharmacol. Exp. Ther. 1997, 281, 1303. The reported values
in Table 5 herein are in agreement with those in the cited reference.
15. For example: Martres, P.; Faucher, N.; Laroze, A.; Pineau, O.; Fouchet, M.-H.;
Potvain, F.; Grillot, D.; Beneton, V. Bioorg. Med. Chem. Lett. 2008, 18, 6251.
Compound 7 (GSK183390A) in the cited reference is depicted below and has
been reported to have a clean 7-day toxicology profile in rats with no safety
issues.
Despite these promising developments, however, high levels of
cytotoxicity (assessed in HepG2 cells at 50 lM after 24 h) were
encountered in 41 and 42 (cf. Table 5). These results were some-
what unexpected given the absence of any significant cytotoxicity
in several non-methylated pyrazole analogs such as 4, and clean
toxicology profile reported on related structures in the literature.15
Nonetheless, these results tend to implicate a potential role for N-
Me pyrazole as a toxicophore in such structures as 41–42.
We described herein a series of novel hNK3 antagonists that
proved amenable to optimization in terms of hNK3-R antagonist
potency and rodent stability profiles through several types of
structural modifications. Structures with good potency and oral
bioavailability that were CNS-penetrant were achievable in this
series, as exemplified by analog 42. Ultimately this work was aban-
O
HO2C
O
N
N
H
N
GSK183390A (Cpd 7)