Cyclic urea derivatives as potent NK1 selective antagonists. Part II: Effects of fluoro and benzylic methyl substitutions

Article abstract of DOI:10.1016/j.bmcl.2005.10.072

A series of novel five-membered urea derivatives as potent NK₁ receptor antagonists is described. The effects of substitution of a 4-fluoro group at the phenyl ring and the introduction of an α-methyl group at the benzylic position to improve p

Full text of DOI:10.1016/j.bmcl.2005.10.072

Bioorganic & Medicinal Chemistry Letters 16 (2006) 1065–1069
Cyclic urea derivatives as potent NK1 selective antagonists.
Part II: Effects of fluoro and benzylic methyl substitutions
Ho-Jane Shue,a Xiao Chen,a John H. Schwerdt,a Sunil Paliwal,a,* David J. Blythin,a
Ling Lin,a Danlin Gu,a Cheng Wang,a Gregory A. Reichard,a, Hongwu Wang,a
John J. Piwinski,a Ruth A. Duffy,b Jean E. Lachowicz,b Vicki L. Coffin,b Amin A. Nomeir,c
Cynthia A. Morgan,b Geoffrey B. Vartyb and Neng-Yang Shiha
aChemical Research Department, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
bCNS Biology Department, Schering-Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
cDepartment of Drug Metabolism and Pharmacokinetics, Schering-Plough Research Institute,
2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
Received 1 September 2005; revised 19 October 2005; accepted 21 October 2005
Available online 14 November 2005
Abstract—A series of novel five-membered urea derivatives as potent NK1 receptor antagonists is described. The effects of substi-
tution of a 4-fluoro group at the phenyl ring and the introduction of an a-methyl group at the benzylic position to improve potency
and duration of in vivo activity are discussed. Several compounds with high affinity and sustained in vivo activity were identified.
Ó 2005 Elsevier Ltd. All rights reserved.
Substance P (SP) is a member of the tachykinin family of
neuropeptides that acts primarily through the NK1 recep-
tor. SP has been associated with numerous pathological
conditions in the central nervous system (CNS) and
peripheral tissues including pain, inflammation, depres-
sion, anxiety, and emesis.1 Hence, an antagonist of the
NK1 receptor has potential therapeutic use in the treat-
ment of a variety of central and peripheral diseases.
(GFT) assay which measures the potency of compounds
to antagonize an NK1-receptor mediated CNS effect
(54% inhibition of foot-tapping at 1 mg/kg po following
a 2 h pretreatment time).
F3C
O
NH
N
Recently, we reported a novel series of cyclic urea deriv-
atives as potent and selective NK1 receptor antagonists
that are orally active and have good CNS penetration.2
Preliminary SAR evaluation in this series led to the
identification of a five-membered unsubstituted urea
analogue 1 (SCH 388714) as a lead compound. SCH
388714 has single digit nanomolar binding affinity for
the NK1 receptor (Ki = 6 nM), excellent brain penetra-
tion (brain/plasma ratio of 18 in gerbil and 4 in rat),
and good oral in vivo activity in the gerbil foot-tapping
O
H
CF3
1
F
R
F3C
F3C
O
O
NH
NH
N
R1
N
R1
O
O
CF3
CF3
2
3 (R = H)
4 (R = F)
Keywords: NK1 antagonist; Sustained in vivo activity; Improved
potency.
*
Corresponding author. Tel.: +1 908 740 2603; fax: +1 908 740
7305; e-mail: sunil.paliwal@spcorp.com
However, the duration of in vivo activity for SCH
388714 was short, giving only 12% inhibition of foot-
tapping at 1 mg/kg po following a 4 h pretreatment time.
Present address: Department of Chemistry, Pfizer Global Research
and Development, 2800 Plymouth Road, Ann Arbor, MI 48105,
USA.
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.10.072

1066
H.-J. Shue et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1065–1069
In this paper, we describe the optimization of SCH
388714 to increase the duration of in vivo activity, while
further improving the NK1 receptor binding affinity and
maintaining good brain penetration.
Reduction of hydantoin 9 with lithium aluminum hy-
dride and aluminum chloride mixture gave racemic urea
compound 2a.6,7 Separation of hydantoin 9 by HPLC
on a Chiralcel ODÒ column followed by reduction affor-
ded chiral urea compounds 2b and 2c. The absolute con-
figurations of 2b and 2c were assigned based on the
established chiral synthesis.8 The racemic N-substituted
urea compounds 2d–g were prepared from ketone 8 by
methods as previously described for the corresponding
non-fluoro compounds.2
In an effort to improve the duration of action, we sought
to block the potential sites of metabolism at the phenyl
ring and the benzylic position. To prevent the hydroxyl-
ation of the phenyl ring, a fluoro group was placed at
the 4-position (2), and a methyl group was introduced
at the benzylic position to block the oxidation at that
site (3,4).3
The in vitro NK1 binding and in vivo NK1 agonist-in-
duced GFT inhibition data for 4-fluorophenyl cyclic
ureas (2a–g) are listed in Table 1. As shown in Table
1, the racemic five-membered fluorophenyl urea ana-
logue 2a exhibited good NK1 affinity (Ki = 4.5 nM).
Similar to phenyl urea series, the activity of 4-fluorophe-
nyl urea analogues also resided mostly in the R-isomer,
for example, compound 2c (R-isomer, Ki = 2.7 nM) ver-
sus compound 2b (S-isomer, Ki = 89 nM). In compari-
son to the lead compound 1, the fluoro analogue 2c
showed a two-fold better binding affinity. Moreover,
compound 2c displayed an improved in vivo activity at
the 4 h time point compared to 1 (55% inhibition at
1 mg/kg for 2c vs 12% for 1), and it also maintained
good GFT activity at the 6 h time point (51% inhibi-
The 4-fluorophenyl cyclic urea analogues (2a–c) of
compound 1 were prepared by the synthetic route
illustrated in Scheme 1.4 Treatment of 2-bromo-40-fluo-
roacetophenone 5 with sodium formate in the presence
of aqueous ethanol afforded compound 6. Direct
assembly of 8 via O-alkylation of 6 with 3,5-bis(triflu-
oromethyl)benzyl bromide was unsuccessful and result-
ed in the decomposition of 6. In order to decrease the
reactivity of compound 6, the ketone group was pro-
tected as an oxime to give compound 7. The O-alkyl-
ation of alcohol-oxime 7 proceeded smoothly and
subsequent deprotection afforded the ketone 8. Subjec-
tion of ketone 8 to Bucherer–Bergs conditions provided
the hydantoin 9.5
¨
Table 1. NK1 receptor binding affinity and GFT inhibition for
compounds 2a–g
Br
HO
a
b
e
O
ONa
F
F
F
+
O
H
O
5
7
6
F3C
F
O
NH
HO
O
F3C
c, d
N
O
O
F
CF3
R1
N
O
CF3
Compounda R1
GFTb (% inh)
8
b
NK1 Ki (nM)
F
F
t = 2 h t = 4 h t = 6 h
2a
–H
–H
–H
4.5
NTc
NTc
NTc
F3C
F3C
O
O
g
O
NH
NH
2b (S)
89
NTc
NTc
NTc
N
N
O
O
H
H
CF3
CF3
2a
9
2c (R)
2.7
6.3
22
22
55
0
51
f, g
O
N
NTc
F
F
2d
F3C
F3C
2e
2f
1.6
0.9
0.4
2
0
3
0
0
0
0
O
O
NH
NH
N
N
H
O
O
H
CF3
CF3
NH
2b
2c
NTc
NTc
Scheme 1. Reagent and conditions: (a) 85% aq EtOH, 80 °C, 5 h, 85%;
(b) methoxylamine hydrochloride, Et3N, EtOH, 85 °C, 2.5 h, 97%; (c)
NaH, THF, 3,5-bis(trifluoromethyl)benzyl bromide, rt, 1.5 h, 98%; (d)
6 N aq HCl 1,4-dioxane, 100 °C, 48 h, 56%; (e) KCN, (NH4)2CO3,
EtOH/H2O (1:1), 60 °C, 48 h, 73%; (f) separation of isomers by HPLC
on a Chiralcel ODÒ column eluting with CH3CN; (g) LiAlH4/AlCl3,
ether, 0 °C, 15 min then rt, 18 h, 82%.
N
2g
a Unless defined as (R) or (S), the compounds in the table are racemic.
b See Refs. 9–12.
c NT, not tested.

H.-J. Shue et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1065–1069
1067
tion). The lower activity of 2c at the 2 h time point could
F3C
F3C
Ph
Ph
OH
O
be due to its slower absorption. Based on the SAR of
phenyl urea series, the substitutions at the less hindered
NH were subsequently explored in the 4-fluorophenyl
series. We found that the neutral groups such as pyran
(compound 2d) retained good binding (Ki = 6 nM) but
did not give sustained GFT activity (0% inhibition at
4 h). Substitutions with basic amine side chains (2e–g)
improved in vitro activity and in some cases (2f and
2g) sub-nanomolar binding affinities were observed;
however, the in vivo GFT activity was significantly re-
duced. The latter may be due to the poor pharmacoki-
netic profile of the amine side chain containing
compounds. Overall, the unsubstituted analogue 2c
was the best compound in this series, exhibiting good
NK1 binding affinity, excellent brain penetration
(brain/plasma ratio 16 in gerbil) and bioavailability
(93% in rat), and improved duration of oral GFT activ-
ity compared to compound 1.
a
b, c
O
CF3
10
CF3
11
Ph
NH
F3C
F3C
O
O
O
NH
O
N
H
N
O
O
H
CF3
CF3
12a
12b
d
Ph
NH
Ph
NH
F3C
F3C
O
O
e
N
H
N
O
O
CF3
CF3
R1
3a
3e-g
As previously reported that the introduction of a meth-
yl group at the benzylic position improves binding
affinity and duration of in vivo activity in the 3-benzyl-
oxy-2-phenylpiperidine NK1 antagonist ether series,3
we next explored a series of benzylic methyl urea deriv-
atives. The synthetic route of preparation of these com-
pounds (3a–g) is illustrated in Scheme 2. Alkylation of
(R)-a-methyl 3,5-bis(trifluoromethyl)benzyl alcohol8
with the triflate of the 2-hydroxyacetophenone in the
presence of 2,6-di-tert-butyl-4-methyl-pyridine afforded
the ketone 11 with a good yield (70%). Hydantoin for-
mation followed by subsequent separation of the mix-
ture of isomers by HPLC on a Chiralcel ODÒ column
provided the pure hydantoin isomers 12a and 12b.
Reduction of 12a with lithium aluminum hydride/alu-
minum trichloride produced the desired urea com-
pound 3a. The assignment of absolute configuration
of the diastereomer 3a was made based on the estab-
lished chiral synthesis.8 Similarly hydantoin 12b was
converted to urea compound 3b. An attempt to prepare
the alkyl-substituted urea analogues 3e–g by alkylation
of 12a followed by treatment with lithium aluminum
hydride/aluminum trichloride resulted only in partial
reduction to the hydroxyl-urea compound due to steric
hindrance. However, reduction of 12a to 3a followed
by alkylation provided the compounds 3e–g. The com-
pounds 3c and 3d were prepared in a similar manner
to compounds 3a and 3b starting from (S)-a-methyl
3,5-bis(trifluoromethyl)benzyl alcohol in place of
(R)-a-methyl 3,5-bis(trifluoromethyl)benzyl alcohol.
4-fluorophenyl analogues 4a–d were synthesized by
the synthetic sequences shown in Scheme 2 for phenyl
analogues.
Ph
NH
F3C
O
d
12b
N
O
H
CF3
3b
Scheme 2. Reagent and conditions: (a) 2-hydroxyacetophenone, 2,6-
di-tert-butyl-4-methyl-pyridine, Tf2O, ClCH2CH2C1, rt, 3 h, then 10,
80 °C, 4 h, 70%; (b) KCN, (NH4)2CO3, EtOH/H2O (1:1), 60 °C, 48 h,
73%; (c) separation of isomers by HPLC on a Chiralcel ODÒ column
eluting with 4% hexane/96% CH3CN, 12a (59%) and 12 b (41%); (d)
LiAlH4/AlCl3, ether, 0 °C, 15 min then rt, 18 h, 62%; (e) NaH, alkyl
halide, DMF, 0 °C to rt, 18 h, 30–35%.
3c, Ki = 93 nM). The latter result is in agreement with
the molecular modeling analysis which showed (Fig. 1)
that the energy-minimum conformation of compound
3a superimposed almost perfectly with that of the ori-
ginal lead compound 1 when aligned with the ether
oxygen and the three ring centroids. The exact overlap
of energy-minimized conformations of 3c and 1 could
not be achieved. Similar folding conformations be-
tween two aryl rings as observed for 3a have been
reported to be the bioactive conformations in the liter-
ature.3,13 The significantly improved binding affinity of
3a compared to the original lead compound 1 (0.6 vs
6 nM) may stem from the favorable conformational
constraints caused by the a-methyl benzylic substitu-
tion; additionally, the benzylic methyl group could
participate in a hydrophobic interaction with the
receptor. A similar benzylic methyl effect was first seen
in the 3-benzyloxy-2-phenylpiperidine NK1 antagonist
ether series.3
The biological data for the benzylic methyl urea ana-
logues are shown in Table 2. Analogous to the
non-benzylic methyl series the (S) configuration at
the quaternary chiral center is less preferred to the
(R) configuration, for example, compound 3a
(Ki = 0.6 nM) versus compound 3b (Ki = 41 nM). At
the benzylic methyl center the (R) absolute stereo-
chemistry is required for high-affinity NK1 binding
(e.g., 1R,4R-isomer, 3a, Ki = 0.6 nM vs 1S,4R-isomer,
In addition to the sub-nanomolar binding affinity, com-
pound 3a displayed excellent oral bioavailability (100%
in rat), retained good brain penetration (brain/plasma
of 10 in gerbil and 4.4 in rat), and demonstrated much
more sustained GFT activity (57% and 48% inhibition
at 4 and 6 h, respectively) than 1. To see an effect of
increasing lipophilicity, some N-alkylated analogues

1068
H.-J. Shue et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1065–1069
Table 2. NK1 receptor binding affinity and GFT inhibition for compounds 3a–g and 4a–d
R
1
F3C
4
O
NH
N
O
CF3
R1
a
NK1 Ki (nM)
Compound
R
R1
GFTa (% inh)
t = 2 h
t = 4 h
t = 6 h
3a (1R, 4R)
3b (1R, 4S)
3c (1S, 4R)
3d (1S, 4S)
3e (1R, 4R)
3f (1R, 4R)
3g (1R, 4R)
4a (1R, 4R)
4b (1R, 4R)
4c (1R, 4R)
4d (1R, 4R)
–H
–H
–H
–H
–H
–H
–H
–F
–F
–F
–F
–H
–H
0.6
41
93
35
57
48
NTb
NTb
NTb
86
NTb
NTb
NTb
62
NTb
NTb
NTb
31
–H
–H
22% inh at 3 lM
–CH3
0.5
1.7
3.1
1.5
3.6
0.8
0.4
–CH2CH3
–CH(CH3)2
–H
45
46
NTb
NTb
41
NTb
40
NTb
44
–CH3
–CH2CH3
–CH2C(O)NH2
65
57
63
9
14
0
20
NTb
NTb
a See Refs. 9–12.
b NT, not tested.
In conclusion, we found that both the 4-fluoro substitu-
tion at the phenyl ring and the introduction of the a-
methyl group at the benzylic position improved duration
of in vivo activity. In addition, a-methyl substitution
provided a significant enhancement in binding affinity.
Several compounds (e.g., 3a, 3e, 4a, and 4b)15 were iden-
tified which retained good brain penetration and demon-
strated improved potency and duration of oral in vivo
activity compared to the original lead compound 1.
Figure 1. Overlay of the energy-minimum conformations of com-
pound 1 and its (R)-a-methyl analogue 3a.14 Oxygen atoms are colored
red; N, blue; F, green; C atoms of 1, orange and C of 3a, blue.
References and notes
1. (a) Duffy, R. A. Expert Opin. Emerg. Drugs 2004, 9, 9; (b)
Seward, E. M.; Swain, C. J. Expert Opin. Ther. Pat. 1999,
9, 571.
were explored. The N-methyl derivative 3e maintained
good binding affinity and GFT activity. The increase
in size to ethyl (3f) and isopropyl (3g) groups resulted
in loss of vitro activity, suggesting that the big lipophilic
groups may not be tolerated at the R1 position.
2. Shue, H.-J.; Chen, X.; Shih, N.-Y.; Blythin, D. J.; Paliwal,
S.; Lin, L.; Gu, D.; Schwerdt, J. H.; Shah, S.; Reichard, G.
A.; Piwinski, J. J.; Duffy, R. A.; Lachowicz, J. E.; Coffin,
V. L.; Liu, F.; Nomeir, A. A.; Morgan, C. A.; Varty, G. B.
Bioorg. Med. Chem. Lett. 2005, 15, 3896.
3. Swain, C. J.; Baker, W. R.; Cascieri, M. A.; Chicchi, G.;
Forrest, M.; Herbert, R.; Keown, L.; Ladduwahetty, T.;
Luell, S.; Maclntyre, D. E.; Metzger, J.; Morton, S.;
Owens, A. P.; Saowski, S.; Watt, A. P. Bioorg. Med.
Chem. Lett. 1997, 7, 2959.
We next evaluated the effect of 4-fluoro substitution in
the benzylic methyl series. Similar to the phenyl benzylic
methyl compound 3a, the corresponding 4-fluorophenyl
benzylic methyl analogue 4a also displayed good in vivo
duration of activity (Table 2). The N-methyl derivative
4b also retained a good GFT profile even though the
binding affinity was slightly decreased. Placement of
the ethyl group (4c) and a polar methyl amide group
(4d) on the nitrogen improved the binding affinity in
the 4-fluorophenyl series but the in vivo activity was sig-
nificantly reduced. Overall, the 4-fluorophenyl ana-
logues in the benzylic methyl series (e.g., 4a, 4b)
demonstrated comparable in vivo activities to the phenyl
benzylic methyl analogues (3a, 3e) despite slightly inferi-
or binding affinities.
4. For more experimental details see: Shih, N.-Y.; Shue,
H.-J.; Reichard, G. A.; Paliwal, S.; Blythin, D. J.;
Piwinski, J. J.; Xiao, D.; Chen, X. PCT Int. Appl. WO
0144200, 2001.
5. Bucherer, H. T.; Lieb, V. A. J. Prakt. Chem. 1934, 141, 5.
¨
(Weinheim) 1990, 326, 331.
¨
6. Knabe, J.; Buch, H. P.; Biwersi, J. Arch. Pharm.
7. Reduction with lithium aluminum hydride reagent only
was not clean. A partial reduction of hydantoin was
observed and reduction of trifluoromethyl group to
difluoromethyl was noticed. On the other hand, the
reduction with lithium aluminum hydride and alumi-
num chloride was clean, providing the desired product.

H.-J. Shue et al. / Bioorg. Med. Chem. Lett. 16 (2006) 1065–1069
1069
8. Reichard, G. A.; Stengone, C.; Paliwal, S.; Mergelsberg,
I.; Majmundar, S.; Wang, C.; Tiberi, R.; McPhail, A. T.;
Piwinski, J. J.; Shih, N.-Y. Org. Lett. 2003, 23, 4249.
9. NK1 assay: binding data are the average of two or three
independent determinations. Receptor binding assays were
performed on membrane preparations from CHO cells in
which recombinant human NK1 receptors were expressed.
[3H]-Sar-Met Substance P was used as the ligand for the
NK1 assay, at concentrations rear the experimentally
derived Kd value. Ki values were obtained using the Cheng
and Prusoff equation.
11. Cascieri, M. A.; Macleod, A. M.; Underwood, D.; Shiao,
L.-L.; Ber, E.; Sadowski, S.; Yu, H.; Merchant, K. J.;
Swain, C. J.; Strader, C. D.; Fong, T. M. J. Biol. Chem.
1994, 269, 6587.
12. Rupniak, N. M. J.; Williams, A. R. Eur. J. Pharmacol.
1994, 265, 179.
13. (a) Natsugari, H.; Ikeura, Y.; Kamo, I.; Ishimaru, T.;
Ishichi, Y.; Fujishima, A.; Tanaka, T.; Kasahara, F.;
Kawada, M.; Doi, T. J. Med. Chem. 1999, 42, 3982; (b)
Harrison, T.; Williams, B. J.; Swain, C. J.; Ball, R. G.
Bioorg. Med. Chem. Lett. 1994, 4, 2545.
10. The NK1 agonist GR73632 (3 pmol in 5 ll) was adminis-
tered centrally to female Mongolian gerbils via icv injec-
tion. Immediately following recovery from the anesthesia,
gerbils were placed into clear Plexiglas boxes for 5 min, and
the duration of foot tapping was measured. Foot tapping
was defined as rhythmic, repetitive tapping of the hind feet.
NK1 antagonists were administered orally in 0.4% meth-
ylcellulose in distilled water at a dose of 1 mg/kg (unless
otherwise stated) at various pretreatment times prior to
injection of GR73632. Data are expressed as a percent
decrease (% inhibition) in the amount of time spent foot
tapping compared to vehicle-treated controls.
14. The energy-minimum conformations were generated
using MCMM conformational search approach in
MacroModel method as implemented in Schro¨dinger
modeling package. 5000 steps of conformational search
each followed by energy minimization using MMFF94s
force field parameters were carried out for every
compound.
15. In contrast to most known NK1 antagonists which
carry basic center, the most efficacious urea
antagonists are neutral, for example, compound
4b (clogP = 5.1; solubility in pH 7.4 buffer =
3 lM).