N′,2-Diphenylquinoline-4-carbohydrazide based NK3 receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2006.08.086

A new class of potent NK₃R antagonists based on the N′,2-diphenylquinoline-4-carbohydrazide core is described. In an ex vivo assay in gerbil, the lead compound 2g occupies receptors within the CNS following oral dosing (Occ₉₀ 30 mg/k

Full text of DOI:10.1016/j.bmcl.2006.08.086

Bioorganic & Medicinal Chemistry Letters 16 (2006) 5748–5751
N0,2-Diphenylquinoline-4-carbohydrazide based
NK3 receptor antagonists
Jason M. Elliott,a,* Robert W. Carling,a Mark Chambers,a Gary G. Chicchi,c
Peter H. Hutson,b A. Brian Jones,a Angus MacLeod,a Rose Marwood,b
Georgina Meneses-Lorente,a Elena Mezzogori,a Fraser Murray,b Michael Rigby,b
Inmaculada Royo,d Michael G. N. Russell,a Bindi Sohal,a Kwei Lan Tsaoc
and Brian Williamsa
aDepartment of Medicinal Chemistry, Merck Sharp & Dohme Research Laboratories, The Neuroscience Research Centre,
Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
bDepartment of In Vivo Neuroscience, Merck Sharp & Dohme Research Laboratories, The Neuroscience Research Centre,
Terlings Park, Eastwick Road, Harlow, Essex CM20 2QR, UK
cDepartment of Immunology, Merck Research Laboratories, Rahway, NJ 07065, USA
dCIBE-Merck Research Laboratories, Merck, Sharp and Dohme de Espan˜a, c/ Josefa Valca´rcel, 38 Madrid 28027, Spain
Received 26 July 2006; revised 17 August 2006; accepted 19 August 2006
Available online 6 September 2006
Abstract—A new class of potent NK3R antagonists based on the N0,2-diphenylquinoline-4-carbohydrazide core is described. In an
ex vivo assay in gerbil, the lead compound 2g occupies receptors within the CNS following oral dosing (Occ90 30 mg/kg po; plasma
Occ90 0.95 lM) and has good selectivity and promising PK properties.
ꢀ 2006 Elsevier Ltd. All rights reserved.
The mammalian tachykinins and their receptors have
been extensively studied,1 with the majority of drug dis-
covery work focused on substance P and its preferred
receptor NK1 and, to a lesser extent, the NKA/NK2R sys-
tem. However, attention has recently turned to NKB and
the NK3 receptor, which have now been implicated in a
range of conditions, including nociception, inflammation,
cough and schizophrenia.2 This paper describes our devel-
opment of a novel series of orally active non-peptide
NK3R antagonists which are able to occupy receptors
within the CNS.
A number of non-peptide based hNK3R antagonists
have been reported, including a series based on quino-
line cores, of which Talnetant3 (1, hNK3R IC50
2.4 0.8 nM)4 is a leading example. The SAR leading
to the (S)-N-(1-phenylpropyl)carboxamide group at
the 4-position of this series has been described.3 Howev-
er, the possibility that this group could be replaced by an
achiral phenylhydrazide (i.e., 2) has not previously been
considered.
Ph
Et
NH
Ph
R
N
O
O
NH
R'
Ph
Quinoline-4-carboxylic acids were easily prepared by
Pfitzinger reactions of appropriately substituted, com-
mercially available isatins with aryl ketones under either
acid or base catalysis (Scheme 1).6 Coupling to phen-
ylhydrazines could be carried out under a variety of pep-
tide coupling conditions. As expected, coupling occurs
exclusively to the more reactive hydrazine 2-nitrogen.
Compounds with electron-withdrawing acyl or carba-
mate groups on the hydrazine 1-nitrogen were easily
OH
N
Ph
N
2
1
Talnetant
Keywords: NK3; Schizophrenia; N0,2-Diphenylquinoline-4-carbohydrazide.
*
Corresponding author. Tel.: +44 1279 440000; fax: +44 1279
440390; e-mail: jason@elliottj25.fsnet.co.uk
0960-894X/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.08.086

J. M. Elliott et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5748–5751
5749
O
for alkyl, aryl, and carbonyl containing N-substituents.
In particular, the introduction of a second phenyl group
gave very high affinity (2d, hNK3R IC50 0.85 nM). How-
ever, further exploration of this compound showed that
it had poor pharmacokinetic stability (incubation with
rat liver microsomes: 2d, 65% turnover after 15 min.)
which made it unattractive as a lead. It seemed possible
that this liability was a consequence of the two electron-
rich N-phenyl rings; indeed, replacing one ring with an
electron-withdrawing group improved metabolic stabili-
ty. In particular, carbomethoxy substitution gave an
optimal combination of good affinity and metabolic sta-
bility (2g, hNK3R IC50 8.8 nM; incubation with rat liver
microsomes: 21% turnover after 15 min.). Larger carba-
mate groups led to reductions in affinity (2h–2i).
R'
CO2H
R'
O
Y
X
O
X
i or ii
N
N
H
Y
43-83%
Z
Ph
CO2Me
R
N
N
N
iii or iv-v
vi
O
O
NH
CO2Me
OMe
R'
X
N
Ph
N
Y
27-85%
3
Scheme 1. Reagents and conditions: (i) HCl–H2O (Concd), AcOH,
reflux; (ii) KOH, EtOH–H2O, reflux; (iii) PhN(R)NH2, EDC, HOBT,
Et3N; (iv) (COCl)2, DMF, CH2Cl2; (v) PhNHNH2, Et3N, CH2Cl2; (vi)
RCOCl, PhMe, reflux.
With the hydrazine N-substituent fixed as carbome-
thoxy, we next explored the effect of the substituent at
C-3 (Table 1, 2j–2m). Groups at this position improved
affinity, but it is interesting to note that methoxy (2g),
methyl (2k), and amino (2l) were preferred over hydrox-
yl (2m) in this series, in contrast to the series which led to
Talnetant (1).3
prepared by first coupling the quinoline carboxylic acid
to phenylhydrazine followed by treatment with an
appropriate acyl chloride or chloroformate. This gave
acylation exclusively on the hydrazine 1-nitrogen pro-
vided the reaction was carried out under neutral condi-
tions (toluene, reflux); addition of base resulted in
variable amounts of a second, unwanted acylation of
the hydrazine 2-nitrogen (e.g., 3). An alternative ap-
proach was the preparation of 1,1-disubstituted hydra-
zine 4 and coupling to give the target molecules in a
single step (Scheme 2). In practice, 4 was less reactive
than phenyl(alkyl)hydrazines, resulting in poor yields
under mild peptide bond formation conditions, but reac-
tion occurred smoothly with preformed quinoline-4-car-
bonyl chlorides. An example of this approach was the
synthesis of the 3-hydroxyquinoline 2m via its O-ben-
zyl-protected analog.
The effect of aromatic substitution was next explored by
introduction of a single fluorine at all available positions
(Table 2). In many cases, this led to losses in affinity rel-
ative to the unsubstituted compound 2g, although fluo-
rination at the meta-position of the N-phenyl ring (4b)
or the quinoline 5-position (4g) was tolerated. Only fluo-
rination at the quinoline 8-position (4j) offered a mar-
ginal improvement in affinity, but introduction of
larger substituents (4k–4m) was not tolerated.
The ability of compounds to occupy NK3 receptors
within the CNS is critical for any potential drug for
the treatment of psychiatric conditions such as schizo-
phrenia. We therefore required an assay to assess this
property for our lead compounds. We decided to target
a biochemical (rather than behavioral) assay to directly
measure receptor occupancy. As with other NK recep-
tors, species differences in NK3R pharmacology make
rats or mice inappropriate for study.1 However, the ger-
bil NK3 receptor is close to the human receptor, show-
ing very similar antagonist affinities (e.g., 2g, NK3R
IC50s: human, 8.8 nM; gerbil, 4.5 nM; rat, 66 nM). We
surveyed a range of NK3R radioligands for a combina-
tion of affinity, low non-specific binding, and high brain
penetration but found none suitable for an in vivo bind-
ing assay. Instead, we developed an ex vivo binding as-
say in which occupancy of novel ligands was measured
in gerbil cortex and striatum using [3H]-senktide, a pep-
tidic agonist (derivatized fragment 6–11 of substance P)
as the radioligand. Routinely, compounds were dosed
orally and occupancy was measured 45 min. post dosing
with collection of plasma for drug level determination.8
In this assay, Talnetant (1) had an Occ90 dose of 75 mg/
kg po and a plasma Occ90 of 3.8 lM. Under the same
conditions, the high affinity antagonist 2d achieved
70% occupancy at 10 mg/kg po (plasma concentration
0.04 lM). This was an encouraging result as it demon-
strated that compounds in this class were active in vivo.
However, while the low plasma exposure needed for
Our initial studies focused on optimizing the hydrazide
N-substituent. With the quinoline C-3 substituent fixed
as methoxy, a series of analogs were prepared (Table
1, 2a–2i). We were gratified to find that these showed
promising hNK3R affinity, with considerable tolerance
Ph
Ph
CO2Me
Ph
CO2Me
i, ii
iii, iv
87%
NH
N
N
NH2
HN
98%
NH2.HCl
CO2CH2Ph
4
Ph
CO2Me
N
O
NH
CO2H
vii, viii
33%
OR
OR
N
N
R = CH2Ph
R = H
R = H
R = CH2Ph
iii
v, vi
2m
Scheme 2. Reagents and conditions: (i) PhCH2OCOCl, NaHCO3,
K2CO3, EtOAc, H2O, 0–5 ꢁC; (ii) MeOCOCl, PhMe 100 ꢁC; (iii) H2
(40 psi), Pd–C, MeOH; (iv) HCl, MeOH; (v) PhCH2Br, NaI, THF,
reflux; (vi) NaOH, MeOH, H2O; (vii) (COCl)2, DMF, CH2Cl2; (viii)
compound 4, Et3N, CH2Cl2.

5750
J. M. Elliott et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5748–5751
Table 1.
hence bioavailability for 2d. A more interesting lead
was the less labile carbamate 2g which was orally active
with an Occ90 of 30 mg/kg po and a plasma Occ90 of
0.95 lM.
R
N
O
NH
In addition to good stability in vitro in microsomal incu-
bations, 2g had promising PK properties (rat: F, 43%;
t1/2, 4.6 h; dog: F, 36%; t1/2, 3 h). We felt it was impor-
tant to verify at an early stage that the presence of an
acylated hydrazide in 2g did not lead to the formation
of potentially toxic hydrazine metabolites. Both in vitro
and in vivo we found no evidence of oxidation or cleav-
age of the hydrazine N–N bond, or release of a hydra-
zine. Moreover, Ames testing9 of 2g was negative in
both the presence and absence of liver microsomal
enzymes.
R'
N
hPXR Responseb
a
Compound
R
R0
hNK3 IC50
(nM)
(%Rif. at 10 lM)
1
Talnetant
2.4 0.8
62
2a
2b
2c
2d
2e
2f
H
Me
OMe
OMe
OMe
32
57
16
4
2
7
Et
Ph
OMe 0.85 0.24
OMe 210 25
68
49
COMe
COEt
Cellular functional NK3R antagonist activity of 2g was
measured in inositol phosphate generation studies using
CHO/hNK3R cells in response to eledoisin or senktide
stimulation.10 Eledoisin caused a concentration-depen-
dent increase in inositol phosphate generation in these
cells with an EC50 of 5.7 0.1 nM (n = 2). In a Schild
analysis of its antagonist behavior, 2g caused a concen-
tration-dependent rightward shift in the EC50 of eledoi-
sin with no diminution of the maximal agonist response
(Kb of 11 nM, slope of 1.0), indicative of competitive
antagonism. When titrated versus an approximate
EC50 concentration (3 nM) of senktide, 2g inhibited ino-
sitol phosphate generation with an IC50 of 28 6 nM
(n = 2). Similarly, 2g (2.8 lM) was able to completely
block senktide-induced (30 nM) Ca2+ mobilization in
CHO/hNK3R cells. Compound 2g demonstrated no
measurable agonist activity for inositol phosphate
OMe
22
3
2g
2h
2i
CO2Me OMe
CO2Et OMe
8.8 5.1
60 22
CO2iBu OMe 190 20
2j
CO2Me
H
91 18
8.4 0.4
44
56
64
2k
2l
CO2Me Me
CO2Me NH2
CO2Me OH
10
39
2
4
2m
a Displacement of [125I] labeled neurokinin B from the cloned hNK3
receptor expressed in CHO cells. Data are means SD (n = 3 or
more).4
b Increase in hPXR activation in HepG2 cells transiently transfected
with hPXR (% of 10 lM rifampicin positive control).7
occupancy was a reflection of very good affinity and
availability to the receptor, the dose required was indic-
ative of the problems of poor metabolic stability and
Table 2.
3'
4'
2'
F
CO2Me
N
O
NH
5
OMe
6
7
F
2''
3''
N
8
F
4''
hPXR Responseb
a
hNK3 IC50 (nM)
’
Compound
Substitution
(%Rif. at 10 lM)
4a
4b
4c
4d
4e
4f
20-F
300-F
400-F
2 -F
300-F
400-F
5-F
53 26
7.4 4.8
33 33
—
54
—
—
69
54
55
62
62
75
88
87
87
110 48
22
24
11
38
28
6
5
3
4
5
4g
4h
4i
6-F
7-F
4j
8-F
4.1 0.3
430 93
4k
4l
8-Br
8-CN
8-CF3
35% 11% inh. at 1 lM
27% 15% inh. at 1 lM
4m
a Displacement of [125I] labeled neurokinin B from the cloned hNK3 receptor expressed in CHO cells. Data are means SD (n = 3 or more).4
b Increase in hPXR activation in HepG2 cells transiently transfected with hPXR (% of 10 lM rifampicin positive control).7

J. M. Elliott et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5748–5751
5751
generation or Ca2+ mobilization at concentrations up to
10 lM.
8. Test compounds were administered po as suspensions in
0.5% methyl cellulose. After 45 min, animals were
humanely culled and trunk blood collected into EDTA-
coated tubes. Plasma was obtained by centrifugation of
whole blood at 3000 rpm, at room temp. Cortex and
striatum were removed, then homogenized in 20 volumes
50 mM Tris, pH 7.4, containing 4 mM MnCl2, 100 lg/ml
bacitracin, 1 lM phosphoramidon, and 1 mM EDTA.
1.8 ml of the resulting homogenate was incubated with
200 ll [3H]-senktide (3 nM final) for 30 min. Samples were
harvested over GF/C filters presoaked in 0.1% PEI/0.5%
Triton, using 50 mM Tris, pH 7.4, as a wash buffer. Filters
were each dissolved in 10 ml Ecoscint overnight. Radio-
activity present in samples was determined on a Beckman
LS6500 counter. Percent occupancy was calculated as
follows:- 100 ꢀ (sample dpm ꢀ mean non-specific dpm/
mean total dpm minus mean non-specific dpm). Plasma
drug levels were determined by LC–MS/MS following
protein precipitation.
Further profiling of 2g showed that it had good selectiv-
ity over hNK1R and hNK2R (IC50s >1 lM) as well as a
panel of other receptors and ion channels, including the
hERG ion channel (Ki > 8 lM), blockade of which can
lead to QT interval prolongation and severe side effects.
2g showed low levels of CYP450 inhibition (human liver
microsomes, IC50s: 2C9, 2D6, >30 lM; 3A4, 12 lM),
but caused CYP3A4 induction in human hepatocytes
at high concentrations (77% of positive control at
20 lM),13 suggesting a risk for drug–drug interactions.
A key step in the main pathway for CYP3A4 induction
is activation of the hPXR nuclear receptor;7 2g was
found to activate hPXR in vitro (HepG2 cells transient-
ly transfected with hPXR; 49% of 10 lM rifampicin po-
sitive control). The substitutions reported here were
shown to have no effect on this activation (Tables 1
and 2); this was a critical issue which required
resolution.
9. No 2-fold or greater increases in revertants relative to
solvent control in a microbial mutation assay using 5
Salmonella typhimurium and Escherichia coli test strains
either with or without metabolic activation by liver
microsomal enzymes from rats pretreated with
xenobiotics.
In summary, we have shown that 2-phenyl quinolines
substituted at C-4 with a phenylhydrazide are potent
hNK3R antagonists which are capable of occupying
receptors within the CNS following oral dosing. The
lead compound 2g showed good selectivity and promis-
ing PK properties. The further development of this ser-
ies to improve the in vivo profile while addressing the
remaining issues of CYP induction and weak CYP3A4
inhibition will be reported in the following paper.
10. Agonist-stimulated inositol phosphate generation in
CHO/hNK3R cells was measured as a modification of a
previously reported protocol.11 All steps were conducted
in a 37 ꢁC incubator with 5% CO2 unless otherwise noted.
Cells (5 · 104/well) were plated overnight in 150 ll of
growth medium in 96-well plates. The following day, the
wells were washed once with loading medium (inositol-free
medium (ICN #1642954) supplemented with 0.02% bovine
serum albumin, 2 mM L-glutamine, 70 mM Hepes, pH
7.5, and 1· HT supplement (100·, Gibco #11067-030)),
followed by the addition of 1 lCi of [3H]-myo-inositol
(NEN #NET114A) in 150 ll loading medium. Inositol
phosphate generation was measured the following day as
follows. LiCl (10 mM) was added to each well for 15 min.
The monolayers were then pre-incubated with antagonists
for 30 min. followed by a 30 min. challenge with agonist
(eledoisin or senktide) after which the medium was
removed and the cells were lysed for 1 h in 60 ll of
10 mM formic acid at 25 ꢁC. A 25 ll aliquot of each lysate
was incubated with 1 mg RNA Binding Y Si (2–5 lm)
SPA beads (Amersham #RPNQ0013) in Optiplates for
2 h. with shaking at 25 ꢁC. The resulting [3H]-inositol
phosphates were quantitated using a Packard Topcount.
NK3 receptor-mediated Ca2+ mobilization was measured
by FLIPR (Fluorescence Imaging Plate Reader)
analysis.12
Acknowledgment
We thank Gwendolyn Ward and John DeLuca for
carrying out the Ames Assay on 2g.
References and notes
1. Almeida, T. A.; Rojo, J.; Nieto, P. M.; Hernandez, M.;
Martin, J. D.; Candenas, M. L. Curr. Med. Chem. 2004,
11, 2045.
2. Spooren, W.; Riemer, C.; Meltzer, H. Nat. Rev. Drug
Discov. 2005, 4, 967.
3. Giardina, G. A. M.; Raveglia, L. F.; Grugni, M.; Sarau,
H. M.; Farina, C.; Medhurst, A. D.; Graziani, D.;
Schmidt, D. B.; Rigolio, R.; Luttmann, M.; Cavagnera,
S.; Foley, J. J.; Vecchietti, V.; Hay, D. W. P. J. Med.
Chem. 1999, 42, 1053.
4. NK3 binding studies were conducted as a modification
of a previously reported protocol5 incubating 3 · 104
CHO cells stably expressing human NK3 receptors with
30 pM [125I] labeled neurokinin B for 1 h. at 25 ꢁC
followed by filtration over GF/C filter plates pre-soaked
in 0.5% BSA.
11. Brandish, P. E.; Hill, L. A.; Zheng, W.; Scolnick, E. M.
Anal. Biochem. 2003, 313, 311.
12. Lynch, K. R.; O’Neill, G. P.; Liu, Q.; Im, D. S.; Sawyer,
N.; Metters, K. M.; Coulombe, N.; Abramovitz, N.;
Figueroa, D. J.; Zeng, Z.; Connolly, B. M.; Bai, C.;
Austin, C. P.; Chateauneuf, A.; Stocco, R.; Greig, G. M.;
Kargman, S.; Hooks, S. B.; Hosfield, E.; Williams, D. L.,
Jr.; Ford-Hutchinson, A. W.; Caskey, C. T.; Evans, J. F.
Nature 1999, 399, 789.
13. Increase in human CYP3A4 mRNA in primary
cultured human hepatocytes measured by TaqMan
PCR. Data are expressed as percentage of the
5. Sadowski, S.; Huang, R. R. C.; Fong, T. M.; Marko, O.;
Cascieri, M. A. Neuropeptides 1993, 24, 317.
positive control, rifampicin, which was used at
a
´
6. Buu-Ho¨ı, P.; Hoan, K.; Xuong, D. J. Org. Chem. 1950, 15,
511.
7. Gao, Y.D.; Olson, S.H.; Balkovec, J.M.; Zhu, Y.; Royo,
I.; Mosley, R.T.; Ismail, A.J.; Hartley, D. Drug Metabo-
lism and Disposition, submitted for publication.
concentration of 10 lM. Meneses-Lorente, G.; Patti-
son, C.; Guyomard, C.; Chesne, C.; Heavens, R. P.;
Watt, A. P.; Sohal, B. Drug Metabolism and
Disposition, submitted for publication.