From rigid cyclic templates to conformationally stabilized acyclic scaffolds. Part I: The discovery of CCR3 antagonist development candidate BMS-639623 with picomolar inhibition potency against eosinophil chemotaxis

Article abstract of DOI:10.1016/j.bmcl.2007.11.067

Conformational analysis of trans-1,2-disubstituted cyclohexane CCR3 antagonist 2 revealed that the cyclohexane linker could be replaced by an acyclic syn-α-methyl-β-hydroxypropyl linker. Synthesis and biological evaluation of mono- and disubstituted propyl linkers support this conformational correlation. It was also found that the α-methyl group to the urea lowered protein binding and that the β-hydroxyl group lowered affinity for CYP2D6. Ab initio calculations show that the α-methyl group governs the spatial orientation of three key functionalities within the molecule. α-Methyl-β-hydroxypropyl urea 31 with a chemotaxis IC₅₀ = 38 pM for eosinophils was chosen to enter clinical development for the treatment of asthma.

Full text of DOI:10.1016/j.bmcl.2007.11.067

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 576–585
From rigid cyclic templates to conformationally stabilized
acyclic scaffolds. Part I: The discovery of CCR3 antagonist
development candidate BMS-639623 with picomolar inhibition
potency against eosinophil chemotaxisq
Joseph B. Santella, III, Daniel S. Gardner, Wenqing Yao, Chongsheng Shi,
Prabhakar Reddy, Andrew J. Tebben, George V. DeLucca, Dean A. Wacker,
Paul S. Watson, Patricia K. Welch, Eric A. Wadman, Paul Davies,
Kimberly A. Solomon, Dani M. Graden, Swamy Yeleswaram, Sandhya Mandlekar,
Ilona Kariv, Carl P. Decicco, Soo S. Ko, Percy H. Carter and John V. Duncia*
Bristol-Myers Squibb Company, R&D, PO Box 4000, Princeton, NJ 08543-4000, USA
Received 20 July 2007; revised 16 November 2007; accepted 19 November 2007
Available online 22 November 2007
Abstract—Conformational analysis of trans-1,2-disubstituted cyclohexane CCR3 antagonist 2 revealed that the cyclohexane linker
could be replaced by an acyclic syn-a-methyl-b-hydroxypropyl linker. Synthesis and biological evaluation of mono- and disubsti-
tuted propyl linkers support this conformational correlation. It was also found that the a-methyl group to the urea lowered protein
binding and that the b-hydroxyl group lowered affinity for CYP2D6. Ab initio calculations show that the a-methyl group governs
the spatial orientation of three key functionalities within the molecule. a-Methyl-b-hydroxypropyl urea 31 with a chemotaxis
IC50 = 38 pM for eosinophils was chosen to enter clinical development for the treatment of asthma.
Ó 2007 Elsevier Ltd. All rights reserved.
Eotaxin1a is a chemotactic cytokine which binds to
CCR3 (CC chemokine receptor 3), the dominant func-
tional chemokine receptor found on eosinophils. It is be-
lieved that eotaxin, secreted by bronchial epithelial/
endothelial cells, triggers eosinophils to migrate into
the lungs of allergic asthmatic patients. Once there, the
eosinophils release major basic protein, membrane-de-
rived lipid mediators, proteins, and other toxic sub-
stances. The result is bronchial mucosal damage which
is thought to give rise to the clinical features of asth-
ma–airway obstruction and bronchial hyperresponsive-
ness (for a more detailed etiology, see Supplementary
material, Figure 1, electronic edition). Thus, a small
molecule antagonist of the CCR3 receptor might be an
effective drug for the treatment of asthma.1b
We have previously described the discovery of acyclic2
and cyclic3 CCR3 antagonists 1 and 2, respectively, both
of which exhibit nanomolar binding potency4 as shown
in Table 1. While the potency of both classes of inhibi-
tors, as measured by the binding inhibition, levels out
at 1 nM, the potency as measured by chemotaxis inhibi-
tion can be in the double-digit picomolar range.5 Thus,
we see that cyclohexyl-containing compounds 2b and 2a
are 300- and 700-fold more potent inhibitors of eosino-
phil chemotaxis than are 1b and 1a, respectively, even
though they all have about the same binding IC50 value
of 1–2 nM.5
Keywords: CCR3 antagonist; BMS-639623; Development candidate;
Ab initio; Eosinophil chemotaxis; Asthma; Conformational analysis;
Acyclic scaffold.
q
This work was originally presented by J.V.D. at Balticum Organicum
Syntheticum 2004, BOS2004, a biennial international conference on
organic synthesis, June 23, 2004, Riga, Latvia: www.BOS06.ttu.ee
and link to 2004 abstracts.
We hypothesized that the superior chemotaxis inhibition
potency of cyclic inhibitor 2 relative to the acyclic 1
stems from the conformational rigidity of the trans-
*
Corresponding author. Tel.: +1 609 252 3123; e-mail: john.
duncia@bms.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.11.067

J. B. Santella et al. / Bioorg. Med. Chem. Lett. 18 (2008) 576–585
Table 1. Binding and chemotaxis IC50 values for compounds 1 and 2
577
Acyclic Linker
N
Cyclic Linker
N
F
F
y
d
i
g
i
p
r
p
o
l
f
R3
3
R3
3
NH
NH
N
H
O
N
H
O
2
1
a,b
CCR3 IC50 (nM)
a,b
Chemotaxis IC50 (nM)
a
Compound
R3
CYP2D6 IC50 (lM)
PB % free
1a
1b
2a
2b
Ac
2.5 1.2
0.6 0.4
2.0 0.7
1.0 0.5
25
3.2 1.6
0.5
—
Tetc
Ac
1.6 0.7
0.06
0.2
0.6 0.6
—
6.6 1.5
0.034 0.019
0.010 0.010
Tet
a See Ref. 3 for binding and chemotaxis assays.
b Values without standard deviations represent a single determination.
c Tet = 1-methyl-tetrazol-5-yl.
1,2-disubstituted cyclohexane ring. To test this hypothe-
sis, we sought to rigidify 1 by introducing substituents
on the propyl linker to favor only a few low energy con-
formations. Conformational analysis of the cyclohexane
ring dictates the diequatorial placement of the substitu-
ents as shown in Newman projection 3 (Fig. 2). Dissec-
tion of the ring of 2 leads to an a,b-disubstituted propyl-
containing compound, existing in the three low energy
conformers, 4, 5, and 6. Conformer 4 is of lowest energy
and thus most favored since it only has two gauche
interactions instead of three as in conformers 5 and 6.
Conformer 4 also mimics the 60° relationship between
the CH2-piperidine and the urea which is found in the
cyclohexyl template of 2. We assume this 60° relation-
ship is critical for potency. Redrawing Newman projec-
tion 4 yields syn-disubstituted propylurea 7. In this very
na¨ıve analysis we assumed that there were no other
interactions besides the nonbonding gauche interac-
tions. If the piperidine nitrogen is not protonated in
the receptor, then H-bonding might occur between the
piperidine nitrogen and one of the urea nitrogens to
form a six-membered ring. This intramolecular H-bond-
ing can take place only in conformers 4 and 6 and not in
5. Thus conformer 4 is again the most favored for hav-
ing both the least number of gauche interactions and for
potentially having a stabilizing intramolecular H-bond.
o
o
cut here along
Piperidine
dotted line
H
H
R
NH
H
H
CH2Pip
urea
60º
N
O
H
H
H
3
2
H
H
urea
H
R1
H
R2
R1
R2
H
CH2Pip
urea
4
R2
CH2Pip
CH2Pip
What effect would only one substituent have on the
number of low energy conformers? This would simplify
our target since there would be only one chiral center on
the propyl linker instead of two. Substitution of R2 = H
in 4 or 7 (Fig. 2) leads to structure 8 or 11, respectively.
For this monosubstituted propyl linker, there are two
low energy conformers, 8 and 9, when only nonbonding
gauche interactions are considered. Of the two, only
conformer 8 mimics the 60° relationship between the
CH2-piperidine and the urea. Being in the right confor-
mation part of the time should make the monosubsti-
tuted linker slightly less potent than the disubstituted
linker where only the correct conformer 4 is preferred.
However, potentially intramolecular H-bonding can
take place between the piperidine nitrogen and the urea
in conformer 8, but not in 9. Thus as was the case for the
disubstituted linker, a single conformer 8 is favored
when only one R1 substituent is present.
60º
urea
R1
H
6
5
R3
R2
H
N
H
N
N
7
O
R1
F
minus R2
R1
H
urea
H
H
H
H
H
CH2Pip
H
CH2Pip
CH2Pip
60º
R1
R1
urea
urea
H
H
10
8
9
R3
H
N
H
N
N
O
R1
F
We were extremely delighted to find that the addition of
a simple methyl group a to the urea increases the chemo-
taxis inhibition potency anywhere from 3-fold (12 vs 1b)
to 50-fold (26 vs 25) (Table 2). Unfortunately, the
11
Figure 2. Conformational analysis of compound 2 leading to struc-
tures 7 and 11. ‘X’ denotes unfavorable gauche interactions.

578
J. B. Santella et al. / Bioorg. Med. Chem. Lett. 18 (2008) 576–585
Table 2. a-Methyl-substituted CCR3 antagonists and their CCR3 binding, chemotaxis, CYP2D6 inhibition, and PB activities
R3
H
N
H
N
N
O
R5
R1
F
R1
R3
R5
CYP2D6 IC50 (lM)
PB % free
a,b
Binding IC50 (nM)
a,b
Chemotaxis IC50 (nM)
b
Compound
1b
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
H
H
Tetc
Tet
Tet
Tet
Tet
Tet
Tet
Tet
Tet
Tet
Ac
0.6 0.4
1.3 0.3
1.5 1.0
1.2 0.4
1.0 0.3
2.0 1.3
2.6 0.6
2.6 0.4
1.1 0.2
0.7 0.4
1.2 0.4
2.7 0.8
2.6 1.4
1.5 0.2
1.6 0.8
1.7 1.0
3.2 1.6
1.1
1.6 0.7
0.4
14
0.6 0.6
Me
H
H
Me
11
0
1.0 0.4
<0.03
—
Me
H
Me
Et
0.2
1.3
8
—
4
Me
H
Et
i-Pr
0.02 0.01
0.30
0.8
4.6
0
Me
H
i-Pr
(CH3)2COH
0.021 0.02
0.304
1.6
22
2
15
—
0
Me
H
(CH3)2COH
Ac
<0.100
3
17
2.2
Me
H
Ac
Ac
Ac
0.2 0.03
—
1.3
4
Tet
Tet
Pyraz
Pyraz
37.7
2.1 0.3
2.8
—
13
1
Me
H
Ac
Pyrazc
Pyraz
0.02 0.1
5
0.1
Me
3.0
1
a See Ref. 3 for binding and chemotaxis assays.
b Values without standard deviations represent a single determination.
c Tet = 1-methyltetrazol-5-yl; pyraz = pyrazol-1-yl.
methyl group mimics the cyclohexyl template somewhat
since it also seems to slightly increase binding to cyto-
chrome P450 CYP2D6, a problem often encountered
in the cyclohexyl series (Table 1, compounds 2a, 2b).
Protein binding had been a problem in the unsubstituted
acyclic series. However, we now find that serum protein
binding decreases, with % free fraction going from 0.6%
for 1b to 11% free for its a-methyl analog 12. We had
hypothesized that this a-methyl group would disrupt
H-bonding between the urea and serum protein. We
see increases in free fraction for other compound cou-
plets as well, albeit smaller. For the bispyrazoles 25
and 26, however, it remains the same at 1%. Meta-sub-
stitution proved optimal on the phenylurea (SAR not
shown). 3,5-Disubstitution (dubbed as the ‘3,5-effect’)
further increases chemotaxis inhibition potency up to
50-fold (compare 14, 16, 18, 20, 24 with 12). Both of
these trends were also seen in the unsubstituted propyl
linker series.2 These aryl substituents most likely bind
to two different receptor subsites effectively locking the
phenylurea in place. None of the compounds in Table
2 were advanced due to selectivity problems with bind-
ing to CYP2D6 or other 7TM receptors, high protein
binding, or poor PK (data not shown).
Before we embarked on making disubstituted propyl
linkers, we needed to know in the monosubstituted ser-
ies if we were working with the more potent linker enan-
tiomer, since one was more potent than the other in the
cyclic series.3 In Table 4 when X = H2, there is a 2- and a
3-fold difference in binding potencies in the isomeric
pairs of piperidines 22/22a and 12/12a. However, the dif-
ference in chemotaxis IC50 values for isomers 22/22a is
300-fold! In the more rigid X = O molecules, the differ-
ence in binding potencies in the isomeric pair of piperid-
ineamides 29/29a is even more pronounced.
Piperidineamides, such as 29, were not pursued any fur-
ther because of undetectable free compound in blood
serum. Compounds 29/29a, in which there is no basic
nitrogen, display lower affinity for CYP2D6 compared
to their counterparts 12/12a.
As discussed earlier, disubstituted propyl linkers 7
should be more potent than monosubstituted 11 when
only nonbonding gauche interactions are considered in
the conformational analysis (i.e., no intramolecular H-
bonding). Thus, we embarked to see what effect
R2 5 H would have on potency. We chose R2 = OH
since it could easily be introduced (see Chemistry) while
at the same time it would decrease lipophilicity and thus
possibly CYP2D6 binding. We find that adding
R2 = OH to the propyl linker does increase chemotaxis
inhibition potency as seen when comparing compounds
12 and 31 (27-fold, Table 5), both of which contain a
monosubstituted phenylurea. However, for the other
cases, all of which contain the potency-enhancing 3,5-
disubstituted phenylurea, the chemotaxis inhibition po-
tency remains about the same. Apparently, the ‘3,5-ef-
fect’ maximizes potency to such an extent that an
What happens if the a-methyl is enlarged? With longer
a-alkyl substituents, the chemotaxis IC50 seems to be-
come even more potent (Table 3). However, this increase
in potency is at the expense of increased CYP2D6 bind-
ing. Longer alkyl substituents also exhibit decreased
metabolic stability in human liver microsomes (pro-
jected human clearance6) possibly due to the greater
number of alkyl positions which can be oxidatively
metabolized by CYPs.

J. B. Santella et al. / Bioorg. Med. Chem. Lett. 18 (2008) 576–585
579
Table 3. a-Alkyl-substituted CCR3 antagonists and their CCR3 binding, chemotaxis, and CYP2D6 IC50 values together with projected human
clearance
OH
H
N
H
N
N
O
R1
F
N
N
N
N
Compound R1
Binding IC50 (nM) Chemotaxis IC50 (nM) CYP2D6 IC50 (lM) CL (projected human clearance)6 mL/min/kg
a,b
a,b
b
19
20
27
28
H
1.1 0.2
0.7 0.4
2.4 0.2
0.3
<0.1
0.1
22
17
12
12
15
17
Me
Et
8.2
1.0
n-Pr 1.5 0.6
0.011
a See Ref. 3 for binding and chemotaxis assays.
b Values without standard deviations represent a single determination.
Table 4. Enantiomeric pairs of CCR3 antagonists with their binding, chemotaxis, and CYP2D6 activities
R3
H
N
H
N
N
O
R5
R1
R2
X
F
R1
R2
R3
R5
CYP2D6 IC50 (lM)
a,b
Binding IC50 (nM)
a,b
Chemotaxis IC50 (nM)
b
Compound
X
22
H2
H2
H2
H2
O
Me
H
H
Ac
Ac
H
Ac
2.7 0.8
5.8 1.1
0.2 0.03
60
1.3
3.5
22a
12
Me
H
Ac
Me
H
Tetc
Tet
Tet
Tet
1.3 0.3
4.0 0.7
1.1
>1
0.2
—
0.4
0.2
12a
29
Me
H
H
Me
H
H
4.2 1.2
37% at 0.3 lM
4.4
>100
29a
O
Me
H
a See Ref. 3 for binding and chemotaxis assays.
b Values without standard deviations represent a single determination.
c Tet = 1-methyltetrazol-5-yl.
additional R2 = OH group has little or no effect on che-
motaxis inhibition (compare 22 and 30, 14 and 34, 16
and 35, 18 and 36, 26 and 37). The other possibility is
that chemotaxis inhibition potency is completely in line
with our earlier prediction that both the monosubsti-
tuted (i.e., 11) and disubstituted (i.e., 7) linkers should
be equipotent when intramolecular H-bonding interac-
tions are invoked. Thus couplet 12 and 31 would be
the only exception. This would mean that in compound
12, intramolecular H-bonding does not occur for some
odd reason, since it is much less potent than its 3,5-
disubstituted counterparts 14, 16, and 18. There is noth-
ing radically different structurally in compound 12 from
the other more potent monosubstituted linker com-
pounds to prevent intramolecular H-bonding. Thus the
first explanation seems to be more accurate: the potency
differences amongst the monosubstituted linkers arise
from the presence or absence of 3,5-disubstitution on
the phenylurea while intramolecular H-bonding between
the urea and piperidine is very minimal to nonexistent.
Thus in summary, introduction of a b-substituent such
as R2 = OH increases potency unless the 3,5-disubsti-
tuted phenylurea is present.
(17-fold), 16 and 35 (3-fold), 18 and 36 (3-fold), 26
and 37 (2-fold), 40 and 41 (2-fold). We also observe
the same order of increasing potency for CYP2D6 with
increasing size of the R1 substituent: 45 > 44 > 31. Pro-
tein binding either becomes slightly better or stays about
the same with the addition of a R2 = OH group. The lin-
ker enantiomer of 30 was synthesized where both R1 and
R2 are syn and b (30a) to doublecheck whether we were
still working with the more potent diastereomer. We find
that 30 is more potent than 30a in both binding (3-fold)
and chemotaxis inhibition (>17-fold).
A wide variety of R2 groups are tolerated and yield
potent compounds. However, none could be ad-
vanced except for R2 = OH. For example, carbamate
derivatives of the OH group, such as 48, are all po-
tent, but suffer from instability in human liver micro-
somes. Replacement of R2 with an amino group leads
to dibasic molecule 49 which exhibits potent CYP2D6
inhibition. Acetamide 50 exhibits poor oral absorp-
tion in the mouse. Dimethylamino analog 51 is
unstable to human liver microsomes, most likely
due to N-demethylation. Finally, the a,b-dimethyl
analog 52 (which is a direct mimic of 2b except for
two carbons missing from the cyclohexyl ring) has
good potency, but it is extremely unstable in human
liver microsomes with a high intrinsic clearance value
of 20.0 mL/min/kg.
The addition of a polar R2 = OH group to our delight
always weakened CYP2D6 binding affinity as seen in
the following compound comparisons: 22 and 30 (5-
fold), 12 and 31 (6-fold), 32 and 33 (6-fold), 14 and 34

580
J. B. Santella et al. / Bioorg. Med. Chem. Lett. 18 (2008) 576–585
Table 5. a-Methyl-b-hydroxy-substituted CCR3 antagonists 7 and their CCR3 binding, chemotaxis, CYP2D6 inhibition, and PB activitiesa
R3
R2
H
N
H
N
N
7
O
R5
R1
F
Compound R1
R2
R3
R5
Bind. IC50 (nM) Chemotaxis IC50 (nM) CYP-2D6 IC50 (lM) PB % free
a,b
a,b
b
21
22
30
30a
1b
12
31
32
33
14
34
16
35
18
36
26
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
H
H
Ac
Ac
Ac
Ac
H
Ac
Ac
1.2 0.4
2.7 0.8
0.8 0.3
2.7 1.4
0.6 0.4
1.3 0.3
0.3 0.03
1.7 0.3
1.8 0.7
1.2 0.4
0.5
3
2.2
1.3
7.0
0
Me
Me
Med OHd
H
OH
0.2 0.03
0.06 0.05
>1
4
Ac
4
Ac
16
—
1
H
H
Tetc
Tet
Tet
Tet
Tet
Tet
Tet
Tet
Tet
Tet
Tet
3.2 1.6
1.1
1.6 0.7
0.4
2.6
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Et
H
H
11
16
0
OH
H
H
Br
0.04 0.01
—
0.05
0.5
3.1
OH
H
Br
0
Me
Me
Et
<0.03
<0.03
0.02 0.01
0.013
0.02 0.02
0.02 0.01
0.1
0.2
3.5
8
OH
H
8
2.0 1.3
1.8 0.6
2.6 0.4
1.0 0.6
0.8
2.5
4
OH
H
Et
5
i-Pr
i-Pr
1.6
4.7
2
OH
H
5
Pyrazc Pyraz 1.7 1.0
3.0
5.9
1
OH
OH
OH
H
Pyraz
H
Thiazc Thiaz
Triazc
Triaz
Oxazc
Isoxc
H
Pyraz
Imid
2.6
0.3
0.05
1.0
2
3.3
0.8
—
—
—
—
—
—
—
—
—
—
—
—
—
—
6
0.8
—
—
Triaz
Triaz
Oxaz
Isox
Tet
2.0 0.6
1.4 0.7
2.7 1.2
1.1 0.6
1.0 0.2
0.6 0.2
0.4c
0.5
0.9
OH
OH
OH
OH
—
—
0.5
1.5
0.1
—
0.9
0.5
n-Pr OH
i-Pr OH
i-Bu OH
H
Tet
Tet
—
—
H
0.3
0.6
H
Tet
Tet
1.2 0.4
0.5 0.1
1.2 0.6
0.6 0.2
0.5c
—
0.87
Me
Me
Me
Me
Me
OCONHMe
H
4.0
NH2
NHAc
NMe2
Me
H
Tet
Tet
—
—
0.4 0.2
3.8
1.4
H
Et
Tet
Tet
<0.03
0.064
H
1.8 0.2
2.2
a See Ref. 3 for binding and chemotaxis assays.
b Values without standard deviations represent a single determination.
c See Ref. 7 Tet = 1-methyltetrazol-5-yl; pyraz = pyrazol-1-yl; imid = 1-methylimidazol-2-yl; thiaz = thiazol-2-yl; triaz = triazol-1-yl; oxaz = oxazol-
2-yl; isox = isoxazol-3-yl.
d R1 and R2 are both syn and b and constitute the enantiomeric linker of 30.
Table 6. Oral bioavailability of BMS-639623 (31) in the mouse, rat,
dog, cyno, and chimp
Because 31 has poor binding and chemotaxis inhibition
potency in the mouse (IC50 = 31, 870 nM, respectively),
Species
Dose mg/kg po
% F
Cmax (nM)
AUC (nM h)
it was dosed intranasally in a Ascaris suum challenge
model in the cyno (cyno eosinophil chemotaxis
IC50 = 0.15 nM). The cyno study was a cross-over de-
sign in which five animals were dosed orally with either
vehicle or 31 at 5 mg/kg b.i.d. at 2 h before and 10 h
after A. suum aerosol challenge. Twenty-four hours after
challenge, the animals underwent bronchoalveolar la-
vage and the eosinophil number was determined. In
three animals, 31 reduced the allergen-dependent eosin-
ophilia by 65%, 78%, and 82% while in two animals
there was no response. Trough plasma levels of the com-
pound in all animals were at or above the theoretical to-
tal concentrations required to inhibit in vitro cyno
eosinophil chemotaxis at the IC90 (5 nM) levels. After
passing preclinical toxicology and cardiovascular assess-
ment, compound 31 was nominated for clinical
development.
Mouse
Rat
Dog
10
50
10
10
2
57
45
17
19
16
1542
3170
926
3151
21341
2740
1055
639
Cyno
Chimp
561
173
Based on its excellent potency, weak CYP2D6 inhibi-
tion, and reasonable free fraction, compound 31
(BMS-639623) was selected for further evaluation. It
was orally bioavailable in five species (Table 6) as well
as selective against other 7TM receptors and ion chan-
nels. In addition to human eosinophil chemotaxis inhibi-
tion, 31 also showed activity in another functional assay,
namely eotaxin-stimulated calcium flux in eosinophils4
where it exhibited an IC50 of 0.87 nM ( 0.41 (n = 6)).

J. B. Santella et al. / Bioorg. Med. Chem. Lett. 18 (2008) 576–585
581
Computer modeling. Further support to our conforma-
12
10
8
tional correlation between the cyclohexyl template and
the mono- and disubstituted propyl linkers comes from
ab initio calculations8a done on the four model systems
shown in Figure 3. We find that as previously predicted,
the a-methyl-b-hydroxypropyl linker (model 4) and the
a-methylpropyl linker (model 3) share the same lowest
energy conformation about bond #1 (/ = À 60, Fig. 3;
conformation 4, or 8, Fig. 2) and mimic the cyclohexyl
template’s lowest energy conformation (model 1,
/ = À 60, Fig. 3; conformation 3, Fig. 2). The lowest
energy conformers of the substituted acyclic linkers are
favored over the other two conformers by 1–2 kcal/mol.
6
4
2
0
-150
-100
-50
0
50
100
150
What the ab initio calculations further show is that the
a-methyl group mimics the cyclohexyl’s influence on
the conformation about bond #2, the a-carbon–urea
bond. In Figure 4 we see that the lowest energy con-
former is the same amongst models 1, 3, and 4 which
is favored over the other rotamer by 1–2 kcal/mol. This
most favored conformation relieves 1,3-allylic strain if
one lends double-bond character to the urea’s NH–CO
bond.9 Thus, the influence on conformation by the a-
methyl group is profound: not only does it influence
the orientation of the piperidine and urea relative to
one another via the lowest energy conformer about
bond #1, but it also influences the orientation of the
urea about bond #2. Given the rigidity of the urea, this
in turn influences the spatial orientation of the urea ‘tail
Dihedral Angle (degrees)
Figure 4. Dihedral drive around the connecting urea C–N bond.
Model 1 is in green, model 2 in red, model 3 in blue, and model 4 in
purple. The inset structures are an overlay of the lowest energy
structures for the corresponding local minimum.
end’, that is, the phenyltetrazole of 31. Thus, the spatial
orientation of three key functional groups, the piperidine,
the urea, and the urea tail piece, is governed by a single
methyl group!
Chemistry. The compounds in Tables 2–5 were synthe-
sized by a slight modification of the method of Davies10a
as exemplified in Scheme 1. Starting materials 53 and 60
were synthesized as described previously.3 Upon Mi-
chael-type addition of 56–55, the intermediate enololate
may be optionally quenched with the chiral Davis10b
oxaziridine to yield hydroxyl substituted propyl linkers
as per the method of Davies.10a
N
N
N
N
H
H
H
HO
1
H
1
1
1
2
2
O
O
2
O
2
O
HN
HN
HN
HN
A scalable synthesis of the monomethyl substituted pro-
pyl linkers starting from L-aspartic acid derivative 62 is
shown in Scheme 2. The method of Beaulieu11 as shown
in Scheme 3 was used to make 700 g of 31 starting from
D-alanine. Scheme 4 shows the synthesis of 49 and 50 via
an adaptation of the method of Davies.12 Scheme 5
shows a synthesis of 51 starting from a protected D-thre-
onine. The a,b-dimethylpropyl linker of compound 52
was synthesized using the methodology of Davies.13
Schemes 6–10 describe the synthesis of the key interme-
diates for the substituted phenylureas in Tables 2, 3 and
5. 3,5-Bis-acetylaniline was synthesized by the method of
Ulrich,14 followed by conversion to phenylurea as in
Scheme 6.
HN
HN
HN
HN
Model 1
Model 2
Model 3
Model 4
10
8
6
4
5
4
6
2
0
Conclusion. We have converted cyclic molecules into
acyclic ones via conformational analysis which has
yielded molecules exhibiting binding affinities in the sin-
gle digit nanomolar range. These binding affinities trans-
late into a pharmacological effect, namely the in vitro
picomolar inhibition of human eosinophil chemotaxis,
the inhibition of eotaxin-induced Ca2+ mobilization,
and the in vivo inhibition of allergen-induced eosino-
philia in the cyno. It remains to be seen whether in hu-
man clinical trials, compound 31 will alleviate the
symptoms of asthma via the in vivo inhibition of eosino-
phil chemotaxis. We have also shown that the syn-a,b-
-150
-100
-50
0
50
100
150
Dihedral Angle (degrees)
Figure 3. Energetic profile of the central dihedral about bond #1 of the
model systems. Model 1 is in green, model 2 in red, model 3 in blue,
and model 4 in purple. Energies are reported as relative to the global
minimum of each model. Model 1 is only evaluated in the torsional
range of À180 to 0 because of the geometrical constraints of the
cyclohexyl ring. The inset numbers denote geometries equivalent to the
structures shown in Figure 2.

582
J. B. Santella et al. / Bioorg. Med. Chem. Lett. 18 (2008) 576–585
F
F
a
HO
R
N
R
NH
O
55
O
54
53
58
Ph
Li
b
56
N
Ph
(OH)
H
(OH)
H
F
F
Ph
F
Ph
N
N
N
N
N
Ph
c
Ph
R
O
d
57
R
(OH)
H
(OH)
H
F
N
N
H
N
H
N
N
e
NH2
N
N
N
N
O
R
N
R
H
N
59
N
61
PhO
O
60
Scheme 1. Reagents and conditions: (a) PyBOP, Et3N, CH2Cl2, rt, 24 h, quant; (b) À78 °C, THF, 56, followed by optional S-(+)-camphor-
sulfonyloxaziridine quench to introduce OH group, 53–65%; (c) B2H6, THF, rt, 90%; (d) Pd(OH)2, H2, MeOH, AcOH, rt, 55–93%; (e) 60,
acetonitrile, 25 °C.
NH-BOC
NH-BOC
NH-BOC
CH2OMs
b
MeO2C
a
MeO2C
62
MeO2C
CH2OH
CO2H
64
63
c
NH-BOC
NH-BOC
NH-BOC
CH2I
MeO(Me)N-OC
MeO2C
65
MeO2C
d
e
CH3
66
CH3
67
f
F
1. g, 53
NH-BOC
CH3
68
OHC
urea
N
NH2
2. h
formation
as usual
CH3
69
Scheme 2. Reagents and conditions: (a) i—B2H6, THF, 0 °C, ii—HOAc, MeOH, 76%; (b) MsCl, Et3N, Et2O, rt, 70%; (c) LiI, THF, ultrasound, 92%;
(d) Pd(OH)2, H2, MeOH, quant; (e) i—LiOH, ii—HN(Me)OMe–HCl, BOP, lutidine, 66%; (f) i—DIBAL, À78 °C, ii—acetone, iii—tartaric acid aq
88%; (g) NaB(OAc)3H, 80–90%; (h) TFA, CH2Cl2, quant.
O
c
O
a
O
d
b
N
N
N
N
73
H
NH2
F
HO
N
HO
72
71
70
F
e
OH
OH
N
NH2
74
75
f
F
N
N
OH
N
H
N
H
N
N
N
31: BMS-639623
O
Scheme 3. Reagents and conditions: (a) i—BnBr, K2CO3, EtOH, rt; (ii) LAH, reflux, 24 h, 72%; (b) DMSO, Py–SO3, 10–15 °C, 2 h, 96% crude; (c)
CH2Br2, THF, n-BuLi, À55 °C, 95%, 6:1 mixture of diastereomers; (d) 53, EtOH, reflux, 20 h, two fractions isolated: 37% of an 8:1 mixture of
diastereomers and 36% of a 6:1 mixture of diastereomers; (e) Pd(OH)2, H2, MeOH, HOAc, 24h, 82%, 81%, after R-mandelic acid recrystallization; (f)
60, CH3CN, rt, 20 h, 86%, HPLC = 98.7%.
disubstituted propyl linker is isosteric with a trans-1,2-
disubstituted cyclohexane. Ab initio minimized struc-
tures of 2b and 31 are overlapped in Figure 5. The
two compounds are conformationally indistinguishable
with regard to (1) overlap of the cyclohexyl of 2b and
the syn-a-methyl-b-hydroxy-propyl linker of 31, (2) the

J. B. Santella et al. / Bioorg. Med. Chem. Lett. 18 (2008) 576–585
583
Ph
N
Ph
Ph
Ph
Ph
Ph
OH
N3
a
b
O
N
O
N
Ph
N
H
76
Ph
O
O
O
79
O
78
77
c, d
Ph
Ph
Ph
N
O
NH
N
H
N
H
N
Ph
f, g
e
N
HO
O
O
NH
O
NH
N
O
N
Ph
O
NH2
O
O
82
Ph
O
O
O
h, i
81
80
O
F
F
HN
HN
H
N
H
N
j, k
H
N
H
N
N
N
j
83
50
49
N
N
N N
N
N N
N
Scheme 4. Reagents and conditions: (a) i—n-BuLi; ii—tert-Butyl crotonate; iii—S-(+)-camphor-sulfonyloxaziridine, THF, 39%; (b) DPPA, DEAD,
NaN3, THF, 90%; (c) H2, 10% Pd–C, EtOAC/MeOH, 50 psi, 1 h, 90%; (d) Ph(CO)Cl, Et3N, THF, 93%; (e) H2, Pd(OH)2, MeOH; (f) 60, Et3N, DMF,
66% for steps e and f; (g) TFA, CH2Cl2; (h) 53, PyBop, Et3N, 20% for steps g and h; (i) B2H6, THF, 40%; (j) H2, Pd(OH)2, MeOH, 50 psi, 69%; (k)
Ac2O, Et3N, CH2Cl2, 98%.
F
Boc
O
O
HN
c, d
g, h
F
a, b
N
OH
HO
F
O
Ph
H
O
Ph
85
HN
HN
Boc
Boc
86
84
e, f
HN
Boc
H
F
Boc
N3
HN
i
H
N
N
N
N
O
O
87
88
N
N
N N
F
NH2
N
H
N
H
N
H
N
H
N
N
N
j
O
89
51
N
N
N
N
N N
N N
Scheme 5. Reagents and conditions: (a) N,O-dimethylhydroxylamine hydrochloride, PyBop, Et3N, 70%; (b) DIBAL-H, CH2Cl2, À78 °C, 78%; (c)
53, Na(OAc)3BH, CH2Cl2, 81%; (d) H2, 10% Pd–C, MeOH, 53% yield, 85:15 anti/syn, 17% isolated yield of anti; (e) MeSO2Cl, Et3N, CH2Cl2; (f)
NaN3, DMF, 60 °C, 45% for steps e and f; (g) H2, Pd(OH)2, MeOH, quant; (h) phenyl 3-ethyl-5-(1-methyl-1H-tetrazol-5-yl)phenylcarbamate,
CH3CN, 86%; (i) TFA, CH2Cl2 63%; (j) formaldehyde, NaCNBH3, CH3COOH (1 drop), CH3CN, HPLC purification, 47%.
O
NO2
NO2
O
N
NH
NO2
NH2
NH2
NO2
NH2
b
a
b,c
to phenyl-
carbamate
as in
a
X
X
N
N
X
X
F
F
N
Br
93
Br
N
N
N
N
N
Br
Br
N
N
N
Scheme 6
90
91: X=CH, N
N
N
92
95
12%
96
94
75%
Scheme 6. Reagents and conditions: (a) imidazole/DMSO or NaH/
1,2,4-triazole/DMF, 100 °C, 24 h, 60–80%; (b) i—Pd(OH)2, EtOAc/
MeOH, H2; ii—2,6-lutidine, phenylchloroformate,THF/CHCl3, 25 °C.
Scheme 7. Reagents and conditions: (a) i—AcOH, H2SO4, NaNO2;
ii—Cu2O, EtOH, 80%; (b) pyrazole, CuI, K2CO3, NO2Ph, D; (c) H2,
Pd–C, MeOH, EtOAc, 76%.
identical orientation of the benzylpiperidine and urea
substituents relative to one another in both 2b and 31
and (3) the a-methyl- (31) and cyclohexyl-induced (2b)
conformation of the urea and consequently the direc-
tional positioning of the phenyl tetrazole.
(a rigidified b-amino acid) and syn-a,b-disubstituted b-
amino acids induce a helix when incorporated into a
peptide chain.15,16 It has been suggested17 that the he-
lix-inducing properties of both arise from their shared
stabilized conformation, similar to what we propose in
Figure 2.
Others working in the field of peptide chemistry have
found that both trans-2-aminocyclohexylcarboxylic acid

584
J. B. Santella et al. / Bioorg. Med. Chem. Lett. 18 (2008) 576–585
HO
The question we now leave for the medicinal chemist is
whether the 16,000 compounds in the literature contain-
ing the syn-(1,3-diamino-2-hydroxy-1-methyl)propyl
substructure present in 31 can be converted into appro-
priately substituted cyclohexane-containing molecules
with retention of biological activity.
O
H
N
O
a
b
OH
O
O2N
N
O2N
O2N
97
98
H
O
99
O
N
OH
N
c
to phenyl-
carbamate
as in
d,e,f
OH
Scheme 6
B
W
Z
?
N
N
A = CH2-
B = OH
X
Y
N
N
N
H2N
O2N
O
101
100
N
A
16,000 literature compounds
Scheme 8. Reagents and conditions: (a) DMSO, TFAA, TEA,
CH2Cl2, À65 °C, 82%; (b) aminoacetaldehyde-dimethylacetal, P2O5,
H2SO4, 77%; (c) H2NOH–HCl, pyridine, EtOH; (d) TMS–acetylene,
NaOCl, TEA, CH2Cl2; (e) K2CO3, MeOH; (f) H2, Pd–BaSO4, MeOH,
EtOAc, 49%.
Acknowledgments
O
O
OH
O
O
O
We thank the members of the Separations Group, Al
Mical, Dauh-Rurng Wu, and Kathy Rathgeb for purify-
ing the mother liquors of 75 during the kilo-scale synthe-
sis of 31. We thank Srinivas Tummala, Ehrlic Lo, and
Steve Wang of the Process Research Dept. for perform-
ing the kilo-scale hydrogenation of 74. We also thank
the members of the Chemical Synthesis Group, Chris
Teleha, Gary Cain, Sergei Bobkov, Jan Hytrek, Chris
Tabaka, and George Emmett, for the large-scale prepa-
ration of 60 employed in the kilo-scale synthesis of 31.
We also thank Sarah Traeger for NMR studies on 31
and 2b.
O
d
a,b,c
NO2
e
NO2
103
NH2
104
NH2
105
N
N
N
N
N
N
HO2C
102
N
N
N
N
N
N
f
g,h
to phenyl-
NH2
106
carbamate
as in
NH2
107
Scheme 6
N
N
N
N
N
N
N
N
Scheme 9. Reagents and conditions: (a) ClCOCOCl, CH2Cl2; (b)
CH3NH2, THF; (c) Tf2O, NaN3, CH3CN, 0 °C, 72%–caution, use
blast shield; (d) H2, 10% Pd–C, MeOH, EtOAc, THF, quant; (e) MeLi,
THF, 18%; (f) H2, 10% Pd–C, MeOH, THF, concd HCl, cat., 92%; (g)
LAH, THF; (h) H2, 10% Pd–C, MeOH, concd HCl, cat., 33%.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2007.11.067.
Br
NO2
a,b,c,d
e
NO2
OEt
f
NO2
NH2
111
108
CO2H
109
110
N
N
N
N
N
N
N
N
N
References and notes
N
N
N
h
g
O
1. For reviews on eotaxin, CCR3 receptors, and eosinophils
in asthma, see (a) Corrigan, C. Curr. Opin. Investig. Drugs
2000, 1, 321; Strek, M. E.; Leff, A. R. In Asthma; Barnes,
P. J., Grunstein, M. M., Leff, A. R., Woolcock, A. J., Eds.;
Lippincott-Raven: Philadelphia, 1997; Owen, C. Pulm.
Pharm. Ther. 2001, 14, 193; Erin, E. M.; Williams, T. J.;
Barnes, P. J.; Hansel, T. T. Curr. Drug Targets-Inflamm.
Allergy 2002, 1, 201; Bertrand, C. P.; Ponath, P. D. Exp.
Opin. Invest. Drugs 2000, 9, 43; Elsner, J.; Escher, S. E.;
Forssmann, U. Allergy 2004, 59, 1243; (b) For other small
molecule CCR3 antagonists in the literature, see Ref. 3
and citations therein.
2. Varnes, J. G.; Gardner, D. S.; Santella, J. B.; Duncia, J.
V.; Estrella, M.; Watson, P. S.; Clark, C. M.; Ko, S. S.;
Welch, P.; Covington, M.; Stowell, N.; Wadman, E.;
Davies, P.; Solomon, K.; Newton, R. C.; Trainor, G. L.;
Decicco, C. P.; Wacker, D. Bioorg. Med. Chem. Lett.
2004, 14, 1645.
Br
to phenyl-
i,g
carbamate
as in
NH2
112
NH2
114
NO2
113
Scheme 6
N
N
N
N
N
N
N
N
N
N
N
N
Scheme 10. Reagents and conditions: (a) NBS, TFA, H2SO4, 72%; (b)
ClCOCOCl, CH2Cl2; (c) CH3NH2, THF, quant; (d) Tf2O, NaN3,
CH3CN, 0 °C, 35%–caution, use blast shield; (e) (CH2@CH)Sn(Bu)3,
Pd(PPh3)4, THF, reflux; (f) H2, 10% Pd–C, MeOH; (g) SnCl2, EtOH,
reflux, 2 h, quant; (h) (CH2@C(OEt))Sn(Bu)3, Pd(PPh3)4,
Pd(PPh3)2Cl2, Tol, reflux, 1 h, quant; (i) 1 N HCl, dioxane, rt,
15 min, quant.
3. De Lucca, G. V.; Kim, U.-T.; Vargo, B. J.; Duncia, J. V.;
Santella, J. B., III; Gardner, D. S.; Zheng, C.; Liauw, A.;
Wang, Z.; Emmett, G.; Wacker, D. A.; Welch, P. K.;
Covington, M.; Stowell, N. C.; Wadman, E. A.; Das, A.
M.; Davies, P.; Yeleswaram, S.; Graden, D. M.; Solomon,
K. A.; Newton, R. C.; Trainor, G. L.; Decicco, C. P.; Ko,
S. S. J. Med. Chem. 2005, 48, 2194.
Figure 5. Overlapped ab initio minimized structures of compounds 2b
(gray) and 31 (green).

J. B. Santella et al. / Bioorg. Med. Chem. Lett. 18 (2008) 576–585
585
4. For a description of the CCR3 binding, the human
eosinophil chemotaxis, and the intracellular Ca2+ flux
assays, see Ref. 3.
5. For a discussion of the disconnect between the binding
and chemotaxis IC50 values, see Ref. 3.
conformation for each model was then used as the starting
structure for torsion 2. The energetics of torsion 2 were
evaluated as above except torsion 1 was constrained to its
starting value. The Monte-Carlo and dihedral drive
calculations were conducted with the conformational
search and dihedral drive modules of MacroModel8b
using the MMFF94s force field in water. Ab initio energies
were calculated using Jaguar.8c; (b) MacroModel, version
9.1, Schro¨dinger LLC, New York, NY, 2005.; (c) Jaguar,
version 6.5, Schro¨dinger LLC, New York, NY, 2005.
9. Hoffmann, R. W. Chem. Rev. 1989, 89, 1841.
10. (a) Bunnage, M. E.; Chernega, A. N.; Davies, S. G.;
Goodwin, C. J. J. Chem. Soc. Perkin Trans. 1 1994, 2373;
(b) Davis, F. A.; Weismiller, M. C.; Murphy, C. K.;
Thimma Reddy, R.; Chen, B.-C. J. Org. Chem. 1992, 57,
7274.
11. Beaulieu, P. L.; Wernic, D. J. Org. Chem. 1996, 61, 3635.
12. Burke, A. A.; Davies, S. G.; Hedgecock, C. J. R. Synlett
1996, 621.
13. Davies, S. G.; Walters, I. A. S. J. Chem. Soc. Perkin Trans.
1 1994, 1129.
14. Ulrich, P. J. Med. Chem. 1984, 27, 35.
15. Lee, M.-r.; Raguse, T. L.; Schinnerl, M.; Pomerantz, W.
C.; Wang, X.; Wipf, P.; Gellman, S. H. Org. Lett. 2007, 9,
1801, and references therein.
16. Seebach, D.; Abele, S.; Gademann, K.; Guichard, G.;
Hintermann, T.; Jaun, B.; Matthews, J. L.; Schreiber, J. V.
Helv. Chim. Acta 1998, 81, 932, and references therein.
17. Gung, B. W.; Zhu, Z. J. Org. Chem. 1997, 62, 6100.
6. In vivo clearance was projected from in vitro oxidative
metabolism in human liver microsomes using the t1/2
method, as described by Obach, R. S.; Baxter, J. G.;
Liston, T. E.; Silber, B. M.; Jones, B. C.; MacIntyre, F.;
Rance, D. J.; Wastall, P. J. Pharm. Exp. Ther. 1997, 283, 46.
7. The binding assay was carried out using 125I-labeled
eotaxin and CHO cells stably transfected with a gene
encoding a chimeric receptor, consisting of the intracellu-
lar domain of human CCR2 with the extracellular and
transmembrane domains of human CCR3. This variation
was found to give results nearly identical to those obtained
using the native CCR3 receptor. Details may be found in
US 2005/0123972, published on June 9, 2005.
8. (a) The starting geometry for model 1 was the lowest
energy conformation found in 1000 steps of Monte-Carlo
conformational searching. Models 2–4 were derived from
the low energy model 1 structure via deletion of the
appropriate atoms and minimization to the closest local
minimum. Intermediate geometries were calculated by
dihedral driving where the dihedral is constrained to the
indicated angle and the remainder of the molecule allowed
to relax. Final energies were calculated by a single point ab
initio calculation at the B3LYP/6-31G** level of theory
with a water solvation model. The resulting low energy