Cyclopentane-based human NK1 antagonists. Part 2: Development of potent, orally active, water-soluble derivatives

Article abstract of DOI:10.1016/j.bmcl.2006.06.044

The synthesis and optimization of a cyclopentane-based hNK1 antagonist scaffold 3, having four chiral centers, will be discussed in the context of its enhanced water solubility properties relative to the marketed anti-emetic hNK1 antagonist EMEND (Aprepitant). Sub-nanomolar hNK1 binding was achieved and oral activity comparable to Aprepitant in two in vivo models will be described.

Full text of DOI:10.1016/j.bmcl.2006.06.044

Bioorganic & Medicinal Chemistry Letters 16 (2006) 4504–4511
Cyclopentane-based human NK1 antagonists. Part 2:
Development of potent, orally active, water-soluble derivatives
Laura C. Meurer,a,* Paul E. Finke,a Karen A. Owens,a Nancy N. Tsou,a
Richard G. Ball,a Sander G. Mills,a Malcolm MacCoss,a Sharon Sadowski,b
Margaret A. Cascieri,b Kwei-Lan Tsao,c Gary G. Chicchi,c Linda A. Egger,d
Silvi Luell,d Joseph M. Metzger,d D. Euan MacIntyre,d Nadia M. J. Rupniak,e
Angela R. Williamse and Richard J. Hargreavese
aDepartment of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
bDepartment of Molecular Pharmacology, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
cDepartment of Immunology, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
dDepartment of Pharmacology, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
eNeuroscience Centre, Merck Sharp and Dohme, Terlings Park, Eastwick Rd, Harlow, EssexCM20 2QR, UK
Received 28 April 2006; revised 10 June 2006; accepted 12 June 2006
Available online 10 July 2006
Abstract—The synthesis and optimization of a cyclopentane-based hNK1 antagonist scaffold 3, having four chiral centers, will be
discussed in the context of its enhanced water-solubility properties relative to the marketed anti-emetic hNK1 antagonist EMENDÒ
(Aprepitant). Sub-nanomolar hNK1 binding was achieved and oral activity comparable to Aprepitant in two in vivo models will be
described.
Ó 2006 Elsevier Ltd. All rights reserved.
The human neurokinin-1 receptor (hNK1) is a mem-
ber of the seven-transmembrane G-protein coupled
family of receptors and is primarily associated with
sensory neurons in the periphery and specific areas
of the CNS. The natural ligand for the NK1 receptor
is the tachykinin peptide substance P (SP) which has
been implicated in the pathophysiology of a diverse
range of conditions, including asthma, inflammatory
bowel disease, pain, psoriasis, migraine, movement
disorders, cystitis, schizophrenia, emesis, and nausea.1
The potent, orally active hNK1 receptor antagonist
1 (Aprepitant) (hNK1 IC50 = 0.09 nM)2 was recently
approved by the FDA for the treatment of both mod-
erate and highly emetic chemotherapy-induced nausea
and vomiting (CINV).3,4 During clinical trials mor-
pholine 1 was found to have good bioavailability
upon oral administration, however, a parenteral for-
mulation was not possible due to its low aqueous sol-
ubility (0.2 lg/mL in isotonic saline, final pH 8.2).5
While the development of a soluble N-phosphoryl
pro-drug has been disclosed,5 the identification of a
suitably soluble parent entity for both oral and paren-
teral use still remained an important need. In addition
to other morpholine,6 piperidine,7 and cyclohexane8
core-based compounds that had been reported, the
preceding manuscript described the synthesis and some
preliminary hNK1 binding results for a novel cyclo-
pentane-based scaffold 2.9 Herein, we report on our
subsequent investigation of the modified five-mem-
bered scaffold 3 having an a-methyl on the benzyl
ether and a 4-fluoro on the phenyl as present in struc-
ture 1, as well as additional structure–activity relation-
ship (SAR) studies at the C-3 position. The effects of
these modifications in terms of additional improve-
ments to the hNK1 binding affinity, enhanced aqueous
solubility compared to 1, and the demonstration of
efficacy in both peripheral and CNS-based in vivo
models are described herein.10
Keywords: Neurokinin-1: substance P; NK1 antagonist: cyclopentane-
based structure.
*
Corresponding author. Tel.: +1 732 594 4291; fax: +1 732 594
5790; e-mail: laura_meurer@merck.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.06.044

L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4504–4511
CF3
4505
OH
CF3
CF3
(+)
6
Me
1
Me
CF3
CF3
CF3
O
F
O
CH O
O
O
O
O
O
1
3
a
b
4 : 1
(+)
+
OH
H
2
2
N
3
(
3
F
7
N
R1
F
)
N
4
n
CH O
O
F
R1
F
HN
O
3
R2
N
NH
R2
F
5
(+)
CH3O
O
1 (Aprepitant)
2
3
2a: R1 = H, R2 = CH2CONH2, n = 0
2b: R1 = Me, R2 = CH2CONH2, n = 0
OH
OH
c,d
e,f
(+)
6
+
R
R
F
F
O
O
O
O
The initial evaluation of hNK1 binding affinity for sev-
eral derivatives of the cyclopentane scaffold 2, having
three contiguous chiral centers on the cyclopentane ring,
revealed an unexpected result. Unlike the established
six-membered piperidine and morpholine core structures
which favored a cis configuration between the ether and
phenyl, in the case of the five-membered cyclopentane
scaffold better hNK1 affinity was achieved with the
1,2-trans-2,3-trans configuration rather than the antici-
pated 1,2-cis-2,3-trans arrangement. In addition, it was
shown that a more basic nitrogen moiety could be
appended to the cyclopentane scaffold at C-3; thus, en-
hanced water solubility was achieved compared to the
relatively non-basic morpholine structures (pKa < 3.5
in 1:1 methanol/water).11 This feature was deemed criti-
cal for development of an intravenous formulation and
spurred our further investigation of these cyclopen-
tane-based structures. Incorporation of an (R)-a-meth-
active (R)-salt series
inactive (S)-salt series
a: R = (R)-salt
b: R = (S)-salt
c: R = H
8a: [α]D = -8.86
8c: [α]D = -17.2
8d: [α]D = -28.0
9b: [α]D = +10.7
9c: [α]D =+18.4
9d: [α]D = +28.8
d: R = CH
3
OH
OH
c,d
(+)
7
+
e,f
R
R
F
F
O
O
O
O
active (R)-salt series
inactive (S)-salt series
10a: [α]D = +83.9
10c: [α]D = +126.4
10d: [α]D = +133.4
11b: [α]D = -80.6
11c: [α]D =-108.0
11d
Scheme 1. Reagents: (a) see Refs. 9,13,14 (50–60% from 4); (b)
NaBH4, MeOH (6 (77%) and 7 (19 %)); (c) NaOH, MeOH (98%); (d)
(S)-(ꢀ)- or (R)-(+)-PhCH(Me)NH2; (e) HCl (aq), ether (100%); (f) HCl
(g), MeOH (>95%).
ylbenzyl ether and
a
4-fluorophenyl onto the
morpholine scaffold had been shown to afford equiva-
lent hNK1 binding, but dramatically enhanced in vivo
activity compared to the original morpholine struc-
ture.2,11 Similar modifications of cyclopentane 2 as indi-
cated in structure 3 were followed by a more in-depth
investigation of the SAR at C-3.
A first attempt at installing the a-methylbenzyl ether
moiety on racemic alcohol 6 using a-methylbenzyl bro-
mide (10-fold excess) with NaH in DMF gave mostly
elimination to the styrene and a 20–40% yield of a 1:1
racemic mixture of the trans ethers 15 and 16
(Ar = 3,5-bis-(trifluoromethyl)phenyl). Fortunately, use
of resolved 8d and a 2-fold excess of racemic trichloro-
acetamidate 14 (derived from ketone 12 via alcohol 13)
in the presence of triflic acid (10–20 mol%) in 1:1 cyclo-
hexane/dichloromethane afforded an improved 70–90%
yield of the non-racemic 15 and 16 as an easily separable
1:1 mixture (Scheme 2).18 In the cis series, alkylation of
10d provided a 24% and 34% yield, respectively, of the
non-racemic a-methyl isomers 17 and 18 with 40%
recovery of the unreacted alcohol. Thus, resolution of
all four stereocenters was achieved.
The synthesis12,13 of methyl 3-keto-2-(4-fluorophe-
nyl)cyclopentane carboxylate (5) starting from commer-
cially available 4-fluorobenzaldehyde (4) was based on a
literature procedure14 and was described in the preced-
ing manuscript in the des-fluoro series (Scheme 1).9
Reduction of the ketone 5 with sodium borohydride
afforded a separable 4:1 mixture of the racemic trans
and cis alcohols 6 and 7. Hydrolysis to the individual hy-
droxy acids allowed for their resolution as either the (R)-
a-methylbenzylamine salts 8a and 10a or (S)-a-methylb-
enzylamine salts 9b and 11b. After conversion to the free
acids 8c–11c, re-esterification provided the two non-ra-
cemic 1,2-trans methyl esters 8d and 9d, and the two cor-
responding 1,2-cis esters 10d and 11d. While the
absolute stereochemistry for the more potent trans enan-
tiomer (see below) had previously been tentatively as-
signed as in structure 8, the trans salt 9b was found to
be amenable to a single crystal X-ray structure analysis
and the assignments for the three cyclopentane centers
were unequivocally determined.15 Subsequent to this
work, an improved route to racemic 616 and a stereospe-
cific synthesis of 8d17 were developed.
The initial in vitro results for racemic 15–18 (hNK1
binding assay IC50 = 100, 3, >100, and 8 nM, respective-
ly, see below) revealed that there was a significant differ-
ence in binding affinity related to the a-methyl isomers
and that the lower Rf diastereomers in both cases pos-
sessed the greater activity. Examination of models and
by analogy to the established morpholine series, as well
1
as analysis of the H NMR and TLC data (both more
1
potent ethers 16 and 18 had similar H NMR a-methyl
shifts and were both the lower Rf diastereomer) resulted
in the assignment of the (R)-methyl stereochemistry for

4506
L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4504–4511
NH
H3C
Ar
O
I
O
OH
O
Cl3C
O
H2N
a
CF3
CF3
CF3
b
a
22a-d
or
O
+
O
N
12
F
CF3
CF3
CF3
13
X
R1
Br
H
14
t-Bu-O
24a-d:
25b:
X = H2N-
X = t-Bu-O-
26b-33b: X = R'R"N-
H3C
Ar
H3C
Ar
b,c,d
OH
O
O
c or d
+
N
N
e
f
34a
22a
+
+
N
N
H
N
F
H
F
15 (higher Rf isomer)
F
16 (lower Rf isomer)
H
CH3O
O
O
CH3O
O
CH3O
O
8d
N
N
35a
(36a)
22a
22b
(N)
N
N
(N)
Cl
H3C
Ar
H3C
Ar
H
O
O
OH
d
N
N
N
Cl
g
Me
37b
38b
+
N
F
F
F
CH3O
O
CH3O
O
CH3O
O
H
N
17 (higher Rf isomer)
10d
18 (lower Rf isomer)
N
N
Cl
h
Me
R1
N
22b
22b
+
+
Scheme 2. Reagents: (a) NaBH4, THF (97%); (b) NaH, Cl3CCN, Et2O
(98%); (c) ( )-a-methylbenzyl bromide (10 equiv), NaH (10 equiv),
DMF (20–40%); (d) 14, triflic acid (10–20% cat.), 1:1 DCM/cyclohex-
ane (from 8d, 15 (36%) and 16 (40%); from 10d, 17 (24%), 18 (34%),
and recovered 10d (40%)).
O
NH
N
O
(R) or (S)
(R) or (S)
39a (S)
39b (S)
39b' (R)
(39b")(S)
i
Br
Br
O
O
N
N
H(Me)
H(Me)
both the cis and trans series.18 As expected, the 1,2-trans
diastereomer 16 (and compounds derived from 16) pos-
sessed significantly better binding affinity for hNK1 than
1,2-cis 18; thus, the former series was more extensively
investigated as illustrated in Schemes 3 and 4 (although
similar chemistry was also carried out in the 1,2-cis
series).
Ac
O
i
HN
N
40b
+
N
22a,b
then j
Me
O
N
H
N
Ac
NH2
N
N
i
H
N
41b,c
N
N
H
Cl
Cl
R1
22b,c
22b
+
N
N
then k
N
H
O
O
H3C
Ar
H3C
Ar
NH2
N
HN
i
H
N
+
42b
MeO
O
O
Me
O
N
H
then k
a
b,c,d
e
f
16
Scheme 4. Reagents and conditions: (a) DIPEA, MeCN, (24a, 62%;
24b, 87%; 24c, 65%; 24d, 57%; 25b, 93% (at 50 °C)); (b) TFA; (c) oxalyl
chloride, DMF (cat.), DCM; (d) R0R00NH, DCM (40–80% from 25b);
(e) NaCNBH3, HOAc, DCE (59%); (f) DIPEA, DCM (35a, 37%; 36a,
44%); (g) K2CO3, DMF (22%); (h) DIPEA, DCM (75%); (i) DIPEA,
MeCN, rt (for 39, 90 °C) (50–75%); (j) HNMe2 (82%); (k) xylenes,
140 °C (41b, 45%; 41c, 34%; 42b, 36%).
N
F
F
C
HO
O
19
O
20
H3C
Ar
H3C
Ar
O
O
g
CF3
Ar =
N
N
CBz
R1
21a-d
F
F
H
R1
CF3
Hydrolysis of the methyl ester 16 provided the acid 19
which underwent the Curtius rearrangement to the iso-
cyanate 20 (Scheme 3). Reaction with benzyl alcohol
afforded the CBz protected C-3 amine 21a. Alkylation
with methyl or ethyl iodide or 2-methoxyethyl bromide
afforded the CBz protected alkyl amines 21b–d. Remov-
al of the CBz by hydrogenation provided the alkyl
amines 22a–d which were then utilized in Scheme 4.
Reductive amination of 22d with formaldehyde also
afforded the tertiary amine 23d. Similar treatment of
18 afforded the corresponding 1,2-cis amines (structures
not shown).
22a-d
a: R1 = H
b: R1 = Me
c: R1 = Et
h
(22d)
N
d: R1 = (CH2)2OMe
Me
OMe
23d
Scheme 3. Reagents and conditions: (a) NaOH, MeOH, 60 °C (98%);
(b) Oxalyl chloride, DMF (cat.), DCM; (c) NaN3, acetone/water,
ꢀ10 °C; (d) toluene, 80 °C; (e) BnOH, DIPEA, DMAP (cat.), toluene,
100 °C (75–85% from 19); (f) MeI, EtI or ClCH2CH2OMe, NaH,
DMF (22b, 93%; 22c, 90%; 22d, 80%); (g) H2, Pd/C, EtOH (>95%); (h)
COH2, H2, 10% Pd/C, MeOH (>90%).

L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4504–4511
4507
Table 1. Structures, solubility, and pKa values for selected 1,2-trans cyclopentane derivatives with in vitro hNK1 binding affinities and in vivo
SYVAL and gerbil foot-tapping inhibition activities
CF3
H3C
CF3
O
chiral
N
F
R1 R2
Compound
R1
—
R2
—
NK1a IC50
Solubilityc
SYVALe
Gerbilg
(nM) (SEM, n)b
(mg/mL) [pKa]d
ED50 @ 1 h
ED90 @ 24 h
(or % I)f
ED50 @ 5 min
ED50 @ 24 h
(or % I)h
1
0.09
<0.0005
0.008 mg/kg
1.8 mg/kg
0.36 mg/kg
0.33 mg/kg
22a
22b
22c
22d
23d
H–
H–
H–
H–
Me–
–H
–Me
1.3 (0.9,3)
0.25 (0.13,3)
0.25 (0.05,3)
0.25 (0.08,3)
0.52 (0.29, 4)
–Et
–(CH2)2OCH3
–(CH2)2OCH3
O
O
O
O
24a
24b
24c
–H
0.40 (0.12, 3)
0.19 (0.02, 4)
0.24 (0.02, 4)
0.09
58% @ 0.1 mg/kg
25% @ 1 mg/kg
65% @ 0.3 mg/kg
28% @ 3 mg/kg
H2N
H2N
H2N
–Me
–Et
0.015
0.1 mg/kg
1.1 mg/kg
57% @ 0.3 mg/kg
29% @ 3 mg/kg
24d
26b
–(CH2)2OCH3
–Me
0.17 (0.04,2)
0.19 (0.02, 5)
<.005
H2N
Me
O
O
0.041 [pKa = 5.3]
66% @ 0.1 mg/kg
30% @ 1 mg/kg
N
H
Me
27b
–Me
0.39 (0.06, 3)
>6.7
66% @ 0.1 mg/kg
36% @ 1 mg/kg
N
Me
O
O
28b
29b
–Me
–Me
0.30 (0.003,3)
0.68 (0.07,3)
0.037
>6.2
MeO
MeO
N
H
34% @ 0.1 mg/kg
18% @ 1 mg/kg
1.3 mg/kg
ndi
N
Me
O
HO
30b
–Me
0.28 (0.01,3)
>10
66% @ 0.1 mg/kg
12% @ 1 mg/kg
N
Me
O
31b
32b
–Me
–Me
0.43 (0.05,3)
0.38 (0.03,3)
0.8
60% @ 0.1 mg/kg
40% @ 1 mg/kg
0.88 mg/kg
nd
N
O
0.46
34% @ 0.1 mg/kg
46% @ 1 mg/kg
1.0 mg/kg
nd
N
O
(continued on next page)

4508
L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4504–4511
Table 1 (continued)
Compound
R1
R2
NK1a IC50
(nM) (SEM, n)b
Solubilityc
(mg/mL) [pKa]d
SYVALe
Gerbilg
ED50 @ 1 h
ED90 @ 24 h
(or % I)f
ED50 @ 5 min
ED50 @ 24 h
(or % I)h
O
33b
–Me
0.95 (0.48,3)
>9
4.3
N
N
Me
N
NH
34a
35a
36a
37b
38b
–H
0.55 (0.17,3)
0.60 (0.20,3)
0.70 (0.20,3)
1.4 (0.10,3)
0.27 (0.05,3)
51% @ 0.1 mg/kg
44% @ 1 mg/kg
1.2 mg/kg
nd
N
–H
N
–H
0.20
N
N
–Me
–Me
<0.005
>7.7
N
53% @ 0.1 mg/kg
26% @ 1 mg/kg
O
N
H
O
(S)
(S)
39a
39b
H
0.34 (0.23,3)
0.20 (0.05,3)
HN
O
–Me
0.35
nd
44% @ 1 mg/kg
0.61 mg/kg
nd
HN
39b0
39b00
40b
41b
41c
–Me
–Me
–Me
–Me
–Et
0.25 (0.03,3)
0.14 (0.04,2)
0.41 (0.05,3)
0.16 (0.03,3)
0.16 (0.03,3)
0.23 (0.04, 3)
>7
25% @ 0.1 mg/kg
33% @ 1 mg/kg
O
O
(R)
(S)
HN
N
0.038
74% @ 0.1 mg/kg
48% @ 1 mg/kg
0.2 mg/kg
nd
Me
H
N
0.54 [pKa = 6.2]
0.072
49% @ 0.1 mg/kg
44% @ 1 mg/kg
nd
>3 mg/kg
O
HN
N
N
N
0.08 mg/kg
0.29 mg/kg
0.06 mg/kg
0.67 mg/kg
HN
N
0.06
24% @ 0.1 mg/kg
18% @ 1 mg/kg
0.22 mg/kg
68% @ 3 mg/kg
HN
H
N
42b
N
–Me
<0.007 [pKa = 4.7]
0.028 mg/kg
0.63 mg/kg
0.46 mg/kg
0.47 mg/kg
O
HN
a Displacement of [125I] labeled SP from the cloned hNK1 receptor expressed in CHO cells.22
b Data are an average of 2–5 independent replicate titrations.23
c Aqueous solubilities were measured in 0.1 M sodium acetate/acetic acid buffer at pH 5 (average of n = 2).19
d Titrations were preformed in 1:1 methanol/water.20
e See Ref 2 for protocol. Results are an average of three animals per data point.
f In an initial screening, compounds were dosed po at 0.1 mg/kg at 1 h prior to challenge and at 1 mg/kg at 24 h. For compounds that were not
subsequently titrated, % I are given at either 1 h and/or 24 h.
g See Ref. 25 for protocol.
h In an initial screening, compounds were dosed iv at 0.3 mg/kg 5 min prior to challenge and at 3 mg/kg at 24 h. For compounds that were not
subsequently titrated, % I are given at either 1 h and/or 24 h.
i No data.
The key intermediate amines 22a–d (or analogous 1,2-
cis amines) were then used to prepare a variety of
derivatives (Scheme 4).9,13 Alkylation with iodoaceta-
mide provided the glycinamide derivatives 24a–d.
Additional glycinamide derivatives in the N-Me series
were investigated via the t-Bu ester 25b with the prep-

L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4504–4511
4509
aration of several N-alkyl or N,N-dialkyl amide deriv-
atives (26b–33b, see Table 1). The imidazole analog
34a was prepared by reductive amination of 22a utiliz-
ing 2-imidazolecarboxaldehyde and sodium cyano-
borohydride. Alkylation of 22a with the 3- or 4-
picolyl chloride afforded the 3- and 4-pyridylmethyl
analogs 35a and 36a. The pyrazine 37b was prepared
by reaction of amine 22b with chloropyrazine. The
4-imidazolylacetamide 38b was prepared by acylation
of amine 22b with 4-imidazolylacetyl chloride. Several
C-3 aminomethylene linked heterocycles (39–42) which
are more closely related to 1 were prepared by previ-
ously described methods.9
pressed in CHO cells.22,23 Comparison of the previous
structure 2 compounds with these (R)-a-methylbenzyl
ether analogs indicated that a 3- to 7-fold enhance-
ment in binding affinity was realized, whereas the
hNK1 affinities in the morpholine series were essen-
tially unchanged with this modification (hNK1
IC50 = 0.09 nM for both).2 Examination of simple
models indicated that the (R)-methyl in the cyclopen-
tane series better reinforces the phenyl/phenyl interac-
tion for optimal pharmacophore fit, while the (S)-
methyl forces the two phenyls apart. The introduction
of the 4-fluoro moiety had minimal or no effect in
binding potency for either the morpholine2 or cyclo-
pentane scaffolds (data not shown). While the amine
intermediates 22a–d and the tertiary amine 23d were
moderately potent (hNK1 IC50 = 1.3, 0.25, 0.25,
0.25, and 0.52 nM, respectively), off-target issues were
a concern with these fully basic compounds, as had
been seen with the original piperidine and morpholine
compounds.7,11 The initial glycinamide compounds
24a and 24b were as potent as the above amines
and, more importantly, showed enhanced binding
affinity compared to the des-methyl analogs (hNK1
IC50 = 0.95 nM vs 0.40 nM for 24a and 2.7 (racemic)
vs 0.19 nM for 24b).9 In general, the tertiary N-Me
analogs were more potent than the secondary N-H
compounds but the larger N-alkyl analogs did not
provide further potency improvement (24c and 24d,
hNK1 IC50 = 0.24 and 0.17 nM). However, there was
a trend toward lower potency with increasing overall
size of the alkylated glycinamides (24b–32b) and with
the dibasic piperazine 33b. The lactams 39b and 39b0
were equipotent with the glycinamides with little dif-
ference seen between the (R)- and (S)-lactam stereo-
chemistry (hNK1 IC50 = 0.25 and 0.20 nM) or
between the NH and N-Me lactams (hNK1
IC50 = 0.20 and 0.14 nM). While the imidazolinone
40b appeared slightly less potent, the basic, aromatic
derivatives 34–37 were not suitable replacements for
the amide moiety. Interestingly, the best activity was
found in the derivatives analogous to triazolinone 1,
in particular the triazole 41b and the triazolinone
42b (hNK1 IC50 = 0.16 and 0.23 nM, respectively).
The activity seen in the latter cases also translated
into excellent in vivo activity (see below). Although
there was a general preference for moderately sized,
polar moieties at the C-3 position, the wide variety
of functionality tolerated here suggests that either
the receptor is somewhat disordered in the sub-do-
main which interacts with this site of the molecule
or that there is a fairly large, open space around the
C-3 position. The hNK1 selectivity over hNK2
(IC50 > 1 lM) and hNK3 (IC50 > 100 nM) was also
maintained in this cyclopentane scaffold for all deriva-
tives (see Ref. 2 for hNK2 and hNK3 binding
protocols).
Aqueous solubilities of selected compounds were deter-
mined at pH 5 (0.1 M sodium acetate/acetic acid buff-
er)19 and pKa determinations were done by
potentiometric titrations in 1:1 methanol/water.20 As
demonstrated in the preceding manuscript,9 a significant
enhancement in water solubility with the cyclopentyl
scaffold 2 was realized due to the increased basicity of
various C-3 nitrogen derivatives (i.e., 2a, pKa = 6.6
and solubility = 0.8 mg/mL at pH 5) compared to mor-
pholine 1 (pKa < 3.5 and solubility < 0.0005 mg/mL at
pH 5)19,20 in which the nitrogen is both benzylic and
part of the morpholine ring. Enhanced water solubility
was also seen in many of these modified cyclopentyl
structures, depending again in large part on the basicity
of the C-3 nitrogen moiety as well as the size and extent
of N-alkylation. For the glycinamide 24a, having a free
NH and a primary amide, the solubility (0.090 mg/mL)
was significantly lower than that of the initial structure
2a, although for the analogous N-Me derivatives the sol-
ubilities were comparable (0.015 mg/mL for 24b
(pKa = 5.3) vs 0.016 mg/mL for 2b (pKa = 5.8)). The ef-
fect of the larger 2-methoxyethyl ether moiety of 24d
gave a further 3-fold loss. Interestingly, sequential alkyl-
ation of the terminal amide improved solubility (see Ta-
ble 1) with 27b, 29b, and 30b giving excellent solubilities
(>6 mg/mL). This effect may be attributed to decreased
crystallinity and/or increased solvation of the protonat-
ed amine salt due to the lack of an internal hydrogen
bond with the N-H of the primary or secondary amides
(compare 24b, 26b, and 27b). This effect was not seen
with the (S)-lactam 39b in that the N-Me 39b00 was 10-
fold less soluble as would have been expected.
For the aromatic derivatives, the aqueous solubilities
correlated with the literature pKas of the parent hetero-
cycles, with the order of solubilities being 34a and
38b > 36a > 41b, c > 37b (literature21a pKas = 7 (imidaz-
ole), 5.2 (pyridine), 2.3 (triazole), and 0.65 (pyrazine),
respectively). Among the non-aromatic heterocycles
more closely related to 1, imidazolinone 40b had good
solubility (0.54 mg/mL; measured pKa = 6.2). The solu-
bility of the triazolinone analog 42b was disappointingly
poor (<0.007 mg/mL), consistent with the reduced
basicity of the C-3 nitrogen (measured pKa = 4.7),
although the solubility of its HCl salt in unbuffered
water was 3.5 mg/mL with a final pH of 2.5.
The 3- to 10-fold enhancement in hNK1 affinity seen in
the previous manuscript for the 1,2-trans series com-
pared to the 1,2-cis arrangement also carried over to
these more elaborate derivatives described above and
essentially followed the same SAR.9 For example, the
1,2-cis-2,3-trans version of the triazole 41b was 6-fold
The binding affinities were determined by the displace-
ment of [125I]-SP from human NK1 receptor stably ex-

4510
L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4504–4511
less potent (hNK1 IC50 = 1.0 nM vs 0.16 nM). In addi-
tion, the cis derivatives were generally also less soluble;
thus, the 1,2-trans series was preferentially investigated
and primarily discussed herein.
0.04, and 0.33 mg/kg at the 5 min, 4 and 24 h time
points, respectively.2 From the glycinamide series, 24b
indicated good CNS activity with ED50s of 0.1 and
1.1 mg/kg at the early and 24 h time points. Similar
24 h results were also seen with the more soluble N,N-
dialkyl derivatives 31b and 32b. The best heterocycles
in the SYVAL assay were 41b and 42b, and in the gerbil
assay afforded inhibition at 5 min (ED50 = 0.06 and
0.46 mg/kg) and at 24 h (0.67 and 0.47 mg/kg, respec-
tively) comparable to that of the clinical compound 1.
The C-3 substituted amino compounds in Table 1 dis-
playing low nanomolar binding affinity on the hNK1
receptor (IC50 < 0.5 nM) were further evaluated in vivo.
As a measure of peripheral inhibition of NK1 receptor
activation after oral administration, the SYVAL24 assay
was used.2 This assay quantified the ability of a test
compound to block plasma extravasation into the
esophagus of guinea pigs driven by the sensorotoxin-
elicited release of endogenous SP. Initial screens were
run at a dose of 0.1 mg/kg orally at 1 h prior to iv chal-
lenge with resiniferatoxin and at 1.0 mg/kg orally at 24 h
prior to challenge to determine their duration of action.
For active compounds (% I > 75%), titrations were car-
ried out to obtain an ED50 value at 1 h and an ED90 val-
ue at 24 h. As a reference, the reported SYVAL ED50
and ED90 results for 1 dosed po were 0.008 mg/kg at
1 h and 1.8 mg/kg at the 24 h time point, respectively.2
Incorporation of an (R)-a-methyl at the benzyl ether
and the 4-fluoro on the phenyl in the cyclopentyl scaf-
fold 3 was accomplished and provided a 3- to 7-fold
enhancement in activity over analogous structure 2
derivatives. N-H and N-alkyl-amino substitution at C-
3 was explored with the preparation of unsubstituted
and substituted acetamides as well as methylene spaced
aromatic and non-aromatic heterocycles. Potent hNK1
binding was demonstrated with most of the modifica-
tions at C-3, although a moderately basic moiety was re-
quired for improved aqueous solubility. In addition,
excellent in vivo activity and 24 h duration were demon-
strated in two different assays of NK1 inhibition. The
two best binding compounds 41b and 42b were found
to be potent in these in vivo assays of peripheral and
CNS inhibition of the hNK1 receptor comparable to
the current clinical compound 1. Triazole 41b also
achieved a greater than 100-fold enhancement in aque-
ous solubility at pH 5. These results led to additional
investigations of the trans ether/phenyl motif in this
and other scaffolds.26,27
In the primary glycinamide series, 24a gave moderate
1 h oral activity; although in the 24 h assay the activity
had decreased significantly (58% at 0.1 mg/kg, 25% at
1 mg/kg, respectively). N-substitution on the terminal
amide generally resulted in the same moderate SYVAL
activity at the 1 and 24 h time points, with only 32b
approaching an ED90 of 1 mg/kg at 24 h. The lactam
derivatives 39b and 39b0 showed moderate results in
the 24 h assay but 39b0 was only marginally active in
the 1 h screen indicating slower absorption than the less
basic glycinamides. The imidazoles 34a and 38b had
activity in SYVAL equivalent to the glycinamides, while
the less potent pyridines and pyrazine were not studied
further. In contrast, the very potent triazole 41b was
highly active in the SYVAL assay at the early time point
(ED90 = 0.08 mg/pk) and at 24 h (ED50 = 0.29 mg/kg)
was significantly better than both the glycinamide 32b
and morpholine 1, while also achieving some aqueous
solubility (0.072 mg/mL). Although possessing about
equipotent hNK1 affinity, the in vivo activityofthe N-eth-
yl derivative 41c was significantly decreased. The imid-
azolinone 40b showed very encouraging solubility, but
the hNK1 affinity and SYVAL activity were diminished.
The triazolinone 42b achieved an ED50 in SYVAL of
0.028 mg/kg at the early time point, but more important-
ly, an ED90 of 0.63 mg/kg at 24 h. Unfortunately, 42b had
attenuated aqueous solubility (<0.007 mg/mL).
References and notes
1. (a) Quatara, L.; Maggi, C. A. Neuropeptides 1998, 32, 1;
(b) Rupniak, N. M. J.; Kramer, M. S. Trends Pharmacol.
Sci. 1999, 20, 485; (c) Dando, T. M.; Perry, C. M. Drugs
2004, 64, 777.
2. Hale, J. J.; Mills, S. G.; MacCoss, M.; Finke, P. E.;
Cascieri, M. A.; Sadowski, S.; Ber, E.; Chicchi, G. G.;
Kurtz, M.; Metzger, J.; Eiermann, G.; Tsou, N. N.;
Tattersall, F. D.; Rupniak, N. M. J.; Williams, A. R.;
Rycroft, W.; Hargreaves, R.; MacIntyre, D. E. J. Med.
Chem. 1998, 41, 4607, and references therein.
3. Navari, R. M.; Reinhart, R. R.; Gralla, R. J.; Kris, M. G.;
Hesketh, P. J.; Khojasteh, A.; Kindler, H.; Grote, T. H.;
Pendergras, K.; Grunberg, S. M.; Carides, A. D.; Gertz, B.
J. N. Engl. J. Med. 1999, 340, 190.
4. Pendergrass, K.; Hargreaves, R.; Petty, K. J.; Carides, A.
D.; Evans, J. K.; Horgan, K. J. Drugs Today 2004, 40, 853.
5. Hale, J. J.; Mills, S. G.; MacCoss, M.; Dorn, C. P.; Finke,
P. E.; Budhu, R. J.; Reamer, R. A.; Huskey, S. W.; Luffer-
Atlas, D.; Dean, B. J.; McGowan, E. M.; Feeney, W. P.;
Chiu, S. L.; Cascieri, M. A.; Chicchi, G. G.; Kurtz, M. M.;
Sadowski, S.; Ber, E.; Tattersall, F. D.; Rupniak, N. M. J.;
Williams, A. R.; Rycroft, W.; Hargreaves, R.; Metzger, J.
M.; MacIntyre, D. E. J. Med. Chem. 2000, 43, 1234.
6. Harrison, T.; Owens, A. P.; Williams, B. J.; Swain, C. J.;
Williams, A.; Carlson, E. J.; Rycroft, W.; Tattersall, F. D.;
Cascieri, M. A.; Chicchi, G. G.; Sadowski, S.; Rupniak,
N. M. J.; Hargreaves, R. J. J. Med. Chem. 2001, 44, 4296.
7. (a) Swain, C. J. Prog. Med. Chem. 1998, 35, 57; (b) Mills,
S. G.; MacCoss, M.; Cascieri, M. A.; Sadowski, S.; Patel,
Selected compounds were also evaluated in a model
measuring inhibition of central NK1 receptors in the
gerbil, whose receptors have human-like pharmacology.
In this assay, icv injection of an exogenous NK1 agonist
into the brain of gerbils elicits a stress response (foot
tapping) for which the duration can be measured and
its inhibition quantified.25 The normal screen was per-
formed at a dose of 0.3 mg/kg iv immediately prior
(5 min) to agonist injection into the brain or at 3 mg/
kg iv 24 h prior to NK1 receptor stimulation in order
to obtain a measure of in vivo duration. As a reference,
the reported ED50s in this CNS assay for 1 were 0.36,

L. C. Meurer et al. / Bioorg. Med. Chem. Lett. 16 (2006) 4504–4511
4511
S.; Chapman, K. L.; Hutson, P. H. Bioorg. Med. Chem.
Lett. 1995, 5, 599.
8. Mills, S. G.; MacCoss, M.; Underwood, D.; Shah, S. K.;
Finke, P. E.; Miller, D. J.; Budhu, R. J.; Cascieri, M. A.;
Sadowski, S.; Strader, C. D. Bioorg. Med. Chem. Lett.
1995, 5, 1345.
9. Finke, P. E.; Meurer, L. C.; Levorse, D. A.; Mills, S. G.;
MacCoss, M.; Sadowski, S.; Cascieri, M. A.; Tsao, K. -L.;
Chicchi, G. G.; Metzger, J. M.; MacIntyre, D. E. Bioorg.
Med. Chem. Lett. 2006, 16, (preceding manuscript page).
10. For a preliminary report on this work, see Meurer, L. C.;
Finke, P. E.; MacCoss, M.; Mills, S. G.; Tsou, N.; Ball, R.
G.; Sadowski, S.; Cascieri, M. A.; Metzger, J. M.;
Eiermann, G.; MacIntyre, D. E.; Rupniak, N. M. J.;
Williams, A.; Hargreaves, R. J. American Chemical
Society 219th National Meeting, San Francisco, Califor-
nia, March 26–31, 2000.
18. Subsequent to this work, the use of (S)-14 and HBF4 in
place of triflic acid was shown to give predominantly net
inversion of the (S) imidate to afford the (R) ether 16, thus,
the designated (R) stereochemistry was confirmed. Jeff T.
Kuethe, unpublished results.
19. Solubilities of selected compounds as the free amines
were determined by HPLC after sonication in buffer at
1 mg/mL and filtration. The solubility for
1
(<0.0005 mg/mL) was determined in 0.1 M potassium
hydrogen phosphate buffer at pH 5, while the solu-
bilities for selected cyclopentane compounds were
determined in 0.1 M sodium acetate/acetic acid buffer
adjusted to pH 5. The reported solubilities are an
average of n = 2.
20. The reported pKa values were obtained in 1:1 methanol/
water. The presence of 50% MeOH in a titration has the
effect of lowering the pKa of an amine between 0.3 and
0.9 U (average value around 0.5) as compared to its value
in water.21b
21. (a) Albert, A.; Serjeant, E. P. The Determination of
Ionization Constants; Springer: New York, NY, 1984,
154–155; (b) see page 35.
22. Cascieri, M. A.; Ber, E.; Fong, T. M.; Sadowski, S.;
Bansal, A.; Swain, C. J.; Seward, E. M.; Frances, B.;
Burns, D.; Strader, C. D. Mol. Pharmacol. 1992, 42, 458.
23. The assay-to-assay replicate variability for the hNK1
binding assay was usually less than 2-fold from the
average of 2–6 independent determinations.
11. Hale, J. J.; Mills, S. G.; MacCoss, M.; Shah, S. K.; Qi, H.;
Mathre, D. J.; Cascieri, M. A.; Sadowski, S.; Strader, C.
D.; MacIntyre, D. E.; Metzger, J. M. J. Med. Chem. 1996,
39, 1760.
12. Representative experimental details and spectral data have
been described in reference 13. All mass spectral and H
1
NMR data were consistent with the assigned structures.
13. Finke, P. E.; MacCoss, M.; Meurer, L. C.; Mills, S. G.;
Caldwell, C. G.; Chen, P.; Durette, P. L.; Hale, J. J.;
Holson, E.; Kopka, I.; Robichaud, A. U.S. Patent
5,750,549, 1998; Chem. Abstr. 1997, 127, 17433.
14. Baker, W.; Leeds, W. G. J. Chem. Soc. 1948, 974.
15. The structure of salt 9b is shown below (ORTEP3, 20%
ellipsoids) with the crystallographic atom-numbering
scheme. Crystallographic data (excluding structure factors)
for the structure in this paper have been deposited with the
Cambridge Crystallographic Data Centre as supplementary
publication numbers CCDC [603506]. Copies of the data
can be obtained, free of charge, on application to CCDC, 12
Union Road, Cambridge CB2 1EZ, UK [fax: +44 (0)1223-
336033 or e-mail:deposit@ccdc.cam.ac.uk].
24. Acronym for resiniferatoxin-induced SY stemic VAscular
Leakage.
25. Rupniak, N. M. J.; Williams, A. R. Eur. J. Pharmacol.
1994, 265, 179.
26. (a) Finke, P. E. National Medicinal Chemistry Sympo-
sium, Madison, WI. June 27–July 1, 2004; (b) Finke, P. E.;
Meurer, L. C.; Mills, S. G.; MacCoss, M.; Qi, H. U.S.
Patent 6,479,518, 2002; Chem. Abstr. 2002, 136, 279139x.
27. The use of the trans ether/phenyl configuration was
subsequently applied to other six-28 and seven-membered
ring scaffolds.29
C17
O1
28. (a) Owens, S. N.; Seward, E. M.; Swain, C. J.; Williams,
B. J. U.S. Patent 6,458,830 B1, 2001; Chem. Abstr. 2000,
133, 252310x and 2000, 133, 266729c; (b) Nelson, T. D.;
Rosen, J. D.; Smitroviych, J. H.; Payack, J.; Craig, B.;
Matty, L.; Huffman, M. A.; McMamara, J. Org. Lett.
2005, 7, 55; (c) Dirat, O.; Elliott, J. M.; Jelly, R. A.;
Jones, A. B.; Reader, M. Tetrahedron Lett. 2006, 47,
1295.
29. Elliott, J. M.; Carlson, E. J.; Chicchi, G. G.; Dirat, O.;
Dominguez, M.; Gerhard, U.; Jelly, R.; Jones, A. B.;
Kurtz, M. M.; Tsao, K. L.; Wheeldon, A. Bioorg. Med.
Chem. Lett. 2006, 16, 2929.
C18
C16
C1
C2
C20
C13
C15
N1
C7
C3
C19
C14
C8
C5
C4
C6
F1
C11
C9
O2
C12
C10
O3
16. Desai, R. C.; Cicala, P.; Meurer, L. M.; Finke, P. E.
Tetrahedron Lett. 2002, 43, 4569.
17. Kuethe, J. T.; Wong, A.; Wu, J.; Davies, I. W.; Dormer, P.
G.; Welch, C. J.; Hillier, M. C.; Hughes, D. L.; Reider, P.
J. J. Org. Chem. 2002, 67, 5993.
.