Conversion of human-selective PPARα agonists to human/mouse dual agonists: A molecular modeling analysis

Article abstract of DOI:10.1016/j.bmcl.2004.09.031

To understand the species selectivity in a series of α-methyl- α-phenoxy carboxylic acid PPARα/γ dual agonists (1-11), structure-based molecular modeling was carried out in the ligand binding pockets of both human and mouse PPARα. This study suggested that interaction of both 4-phenoxy and phenyloxazole substituents of these ligands with F272 and M279 in mouse PPARα leads to the species-specific divergence in ligand binding. Insights obtained in the molecular modeling studies of these key interactions resulted in the ability to convert a human-selective PPARα agonist to a human and mouse dual agonist within the same platform.

Full text of DOI:10.1016/j.bmcl.2004.09.031

Bioorganic & Medicinal Chemistry Letters 14 (2004) 6113–6116
Conversion of human-selective PPARa agonists to
human/mouse dual agonists: a molecular modeling analysis
a,
b
a
Minmin Wang,a,* Leonard L. Winneroski, Robert J. Ardecky, Robert E. Babine,
*
Dawn A. Brooks,a Garret J. Etgen,a Darrell R. Hutchison,a Raymond F. Kauffman,a
Aaron Kunkel,a Dale E. Mais,b Chahrzad Montrose-Rafizadeh,a Kathleen M. Ogilvie,b
Brian A. Oldham,a Mary K. Peters,a Christopher J. Rito,a
Deepa K. Rungta,b Allie E. Tripp,a Sarah B. Wilson,a Yanping Xu,a
Richard W. Zinka and James R. McCarthya,*
aLilly Research Laboratories, Eli Lilly & Company, Indianapolis, IN 46285, USA
bLigand Pharmaceuticals, 10275 Science Center Drive, San Diego, CA92121, USA
Received 11 August 2004; revised 10 September 2004; accepted 16 September 2004
Available online 26 October 2004
Abstract—To understand the species selectivity in a series of a-methyl-a-phenoxy carboxylic acid PPARa/c dual agonists (1–11),
structure-based molecular modeling was carried out in the ligand binding pockets of both human and mouse PPARa. This study
suggested that interaction of both 4-phenoxy and phenyloxazole substituents of these ligands with F272 and M279 in mouse PPARa
leads to the species-specific divergence in ligand binding. Insights obtained in the molecular modeling studies of these key interac-
tions resulted in the ability to convert a human-selective PPARa agonist to a human and mouse dual agonist within the same
platform.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
propanoic acid derivatives as another example of hu-
man-selective PPARa activators.4 During our
investigation of novel PPAR ligands for the treatment
of dyslipidemia and type 2 diabetes, a discrepancy in
the reduction of triglycerides in apoA-1 transgenic mice
was noticed for a series of phenoxy propionic acids with
significant difference in human and mouse PPARa bind-
ing.5,6 Here we report the results of our SAR study illus-
trating the ability to enhance mouse activity in this class
of ligands and an interpretation for SAR convergence in
human and mouse PPARa receptors using the molecu-
lar modeling studies. Recently, Miyachi and Uchiki re-
ported their observations on molecular determinant
leading to human-selective PPARa activation using wild
type and a point-mutated LBD.8
A reliable rodent model possessing the target receptor
for a human disease is often an important tool for drug
development. In certain cases, however, potential drugs
may target a human protein for which the correspond-
ing rodent protein differs thus making pre-clinical candi-
date evaluation difficult. The species-dependent
transactivation of some PPARa agonists has been re-
ported previously.1,2 For example, the classical PPAR
agonist WY-14643 and a recently disclosed ureidothio-
isobutyric acid derivative, GW9578 showed preferential
transactivation for rodent over human PPARa recep-
tor.2,3 In contrast, ETYA (5,8,11,14-eicosatetraynoic
acid) showed preferential transactivation for human
PPARa.1,2 Nomura et al. published substituted phenyl-
2. Chemistry
Keywords: PPARa agonist; Species selectivity; Structure-based design.
*
Corresponding authors. Tel.: +1 317 433 9621; fax: +1 317 276 5431
(M.W.); tel.: +1 317 277 2216; fax: +1 317 277 2035 (L.L.W.); tel.: +1
317 433 1345; fax: +1 317 276 5431 (J.R.M.); e-mail addresses:
The synthesis of compounds 1–11 was accomplished in a
convergent manner as outlined in Schemes 1 and 2. The
substituted toluene-4-sulfonic acid 2-(5-methyl-2-phen-
yl-oxazol-4-yl)-ethyl esters 12 were prepared in seven
wang_minmin@lilly.com;
lilly.com
lwinneroski@lilly.com;
jmccarthy@
0960-894X/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2004.09.031

6114
M. Wang et al. / Bioorg. Med. Chem. Lett. 14 (2004) 6113–6116
O
O
N
O
N
b
a
Br
Br
Cl
NOH
O
c, d, e
O
N
O
N
f, g
Ph
Br
OTos
OH
12
Scheme 1. Reagents and conditions: (a) Br-benzaldehyde, HOAc, HCl (74%); (b) POCl3, CHCl3 (72%); (c) KCN, KI, DMF (100%); (d) KOH, 2-
MeO–ethanol (60%); (e) BH3–THF, MeOH (72%); (f) Br–PhB(OH)2, Pd(OAc)2, PPh3 (95%); (g) (Tos)2O, Pyr, DMAP (95%).
O
steps from commercially available butanedione
monooxime according to the literature procedure.9,10
The substituted 3-(4-hydroxy-phenyl)-2-methyl-2-phen-
oxy-propionic acid ethyl esters 13 were prepared as pre-
viously described.6 Treatment of tosylates 12 with
phenol esters 13 in DMF with Cs2CO3, followed by
hydrolysis afforded ligands 1–11. The enantiomers of
phenols 13 were separated by chiral chromatography
using a Chiralpak AD 4.6 · 250mm column. The (À)-
enantiomer of 13 was utilized to prepare compound 2
while the (+)-enantiomer was utilized to prepare com-
pound 3.
O
HO
a
OEt
Br
+
OEt
O
R3
R3
O
O
OH
O
OEt
c,d
OEt
b
O
HO
BnO
13
R3
R3
CO2H
O
N
e, f
3. Results and discussion
O
R1
O
R2
The in vitro binding affinity and transactivation activity
of compounds 1–11 toward human and mouse PPARa
are summarized in Table 1. To explain the structure–
activity relationship in the mouse PPARa receptor, a
homology model was prepared using the crystal struc-
ture of human PPARa with GW409544 (PDB: 1k7l,
Fig. 1).7,11 The root-mean square deviation (RMSD)
1 - 11
R3
Scheme 2. Reagents and conditions: (a) Cs2CO3, DMF, 90ꢁC (90%);
(b) LDA, À78ꢁC, then p-benzoloxy benzaldehyde (70%); (c) TFAA,
Pyr., CH2Cl2 (quant.); (d) H2, 5–10% Pd/C, EtOAc (90%); (e) 12,
Cs2CO3, DMF, 55ꢁC (70%); (f) 5N NaOH, EtOH, reflux (>99%).
Table 1. In vitro binding IC50 and cotransfection EC50 datac in human and mouse PPAR receptor subtypes
a
b
CO2H
O
O
R1
N
O
R2
R3
Entry
R1
R2
R3
Enantiomer
hPPARad
mPPARad
IC50 (nM)a,c
EC50 (nM)b,c
IC50 (nM)a,c
EC50 (nM)b,c
1
H
H
H
S/R
S
42
21
6
4
12
5
2
1
2563 389
1543 136
6764 844
7020 3744
404 25
1625 197
1299 106
Eff < 20
2
H
H
H
3
H
H
H
R
1546 36
503 41
167 25
390 23
4
Ph
H
H
H
S/R
S/R
S/R
S/R
S/R
69
7
6
1
1
3
5
2343 175
173 11
5
Ph
H
H
Me
6
H
64
8
7
1396 167
1461 44
530 41
585 151
1041 192
151 22
7
Ph
H
H
Me
375 26
306 52
39
23
8
Ph
H
Me
9
H
tBuS/R
tBuS/R
tBuS/R
149 28
196 21
133 12
11
260 28
49
2
515 56
700 44
146 66
211 26
1167 152
91 15
10
11
Ph
H
H
Ph
5
a Concentration of test compound required to displace 50% of tritiated ligand.
b Concentration of test compound that produced 50% of the maximal reporter activity.
c n = 3.
d Tritium-labeled PPARa agonist 2-(4-{2-[3-(2,4-difluorophenyl)-1-heptylureido]ethyl}phenoxy)-2-methylbutyric acid was used as radioligand for
generating displacement curves and IC50 values.

M. Wang et al. / Bioorg. Med. Chem. Lett. 14 (2004) 6113–6116
6115
O
HO
NH
O
O
N
O
GW409544
Figure 1. Structure of GW409544.
˚
for the 267 Ca atom pairs is 0.25 A. The four differences
in the ligand binding pocket of mouse PPARa as com-
pared to the human receptor, I272F (helix 3), T279M
(helix 3), V324L (helix 5), and V332I (b-strand), are
on helixes 3 and 5, and the b-strand. The compounds
were docked into both human and mouse LBDs using
CDocker.12,13 The docking mode was selected according
to the molecular overlay with GW409544 as it appeared
in the crystal complex among the top 10 poses.
Figure 2. The proposed binding model for a-aryloxy-a-methylhydro-
cinnamic acid derivatives 1 (magenta) and 9 (yellow) in the human (A)
and mouse (B) PPARa LBD. GW409544 is shown in white.
3.1. Effect of the stereochemistry at the a-position of the
a-aryloxy-a-methylhydrocinnamic acid
homology model. The R3 substitution is buried in a
hydrophobic cluster of F272, F273, F351, L344, and
M355 in the mouse receptor. Increasing the size of R3
would enhance hydrophobic interaction with the mouse
receptor and therefore better affinity to the receptor.
However, for the human isoform, the c-CH3 of I272
interferes with the binding of the tert-butyl in 9, which
may contribute to the threefold decrease in IC50 on hu-
man PPARa. It was also found that when R3 is H as in
compound 1, the docked conformers in human and
mouse receptors are very different in the orientation of
hydrophobic tail piece (Fig. 2). Increasing the size of
the R3 group led to the convergence of docked confor-
mations in the two species. The t-BuPhO group resem-
bles the position of the side chain of GW409544. The
docked conformation for 9 in both human and mouse
PPARa LBD supports the in vitro results.
The enantiomer selectivity in human PPARa transacti-
vation activity was previously reported for a-alkoxy
substituted carboxylic acids.4,14 We chose compound 1
to investigate the enantiomer-dependent binding affinity
and transactivation activity on human PPARa. The (R)-
and (S)-enantiomers of 1 were made from the pure
enantiomers of 13.6 The racemic mixture 1 showed a
similar profile to the (S)-enantiomer 2 in both binding
and transactivation activities, therefore, only racemates
were tested in the study.
The docking of 2 and 3 into the human PPARa receptor
suggests that the (S)-enantiomer 2 assumes a similar
binding mode to GW409544 in which the a-methyl of
the a-aryloxy-hydrocinnamic acid head piece points to-
ward the space formed by helix 3 and the loop between
helix 10 and the AF2 helix. The carboxylic acid group
formed a good hydrogen bond network with Y464,
Y314, and S280 as GW409544. Ne of H440 was within
3.3. Effect of the substitution at the phenyloxazole
A comparison of compounds 1 and 5 showed that a sig-
nificant improvement in mouse PPARa binding and
transactivation activity was obtained by substitution of
the phenyloxazole ring with an additional phenyl ring
in the meta position (R2 in Table 1). A sixfold improve-
ment in mouse PPARa binding and a ninefold increase
in mouse PPARa functional activity was achieved for 5
with this modification. In contrast, substitution of the
phenyloxazole with a para phenyl group such as 4 (R1
in Table 1) led to a two- to threefold decrease in mouse
PPARa binding. The same trend was also observed
when R3 is methyl or tert-butyl (e.g., 7 and 10).
˚
2.8A distance to the phenyl ether oxygen of 2. However,
the methyl group on the (R)-enantiomer 3 would bump
F138 and H440 on helixes 5 and 10 assuming the same
binding mode. As a result, the (S)-enantiomers were
docked in the subsequent SAR rationalization study.
3.2. Effect of substitution at the aryloxy ring
Substitution of the para position of the aryloxy ring (R3
in Table 1) with lipophilic substituents, such as methyl,
or tert-butyl (6 and 9) led to a two- to fivefold increase in
mouse PPARa binding with a two- to eightfold increase
in functional activity relative to the nonsubstituted 1.
We successfully converted a human-selective PPARa
agonist to a more potent mouse PPARa agonist by the
addition of the tert-butyl group to provide 9.
F272 also plays an important role in understanding the
impact of R1 substitution on mouse PPARa binding.
While I272 in human PPARa allows the extension of
the R1 from H to phenyl, it is difficult to dock 4 into
the mouse PPARa receptor without changing the bind-
ing mode (the p-phenyl group would overlap with F272
otherwise). Instead, the m-phenyl (R2) of 5 pushes the
central phenyl ring out of its original position slightly
The impact of R3 on the binding to mouse PPARa can
be explained by the interaction with the side chain of
F272 according to the docking to the mouse PPARa

6116
M. Wang et al. / Bioorg. Med. Chem. Lett. 14 (2004) 6113–6116
2. Mukherjee, R.; Jow, L.; Noonan, D.; McDonnell, D. P. J.
Steroid Biochem. Mol. Biol. 1994, 51, 157.
causing minimal changes to the interaction with the
LBD. This rearrangement is also facilitated by the
T279M mutation in the mouse receptor.
3. Brown, P. J.; Winegar, D. A.; Plunket, K. D.; Moore, L.
B.; Lewis, M. C.; Wilson, J. G.; Sundseth, S. S.; Koble, C.
S.; Wu, Z.; Chapman, J. M.; Lehmann, J. M.; Kliewer, S.
A.; Willson, T. M. J. Med. Chem. 1999, 42, 3785.
4. Nomura, M.; Tanase, T.; Ide, T.; Tsunoda, M.; Suzuki,
M.; Uchiki, H.; Murakami, K.; Miyachi, H. J. Med.
Chem. 2003, 46, 3581.
5. Etgen, G.; Broderick, C.; Ldham, B.; Bean, J.; Bensch, W.;
Montrose, C.; Barr, R.; Zink, R.; Jett, D.; Liu, S.; Ogilvie,
K.; Ardecky, R.; Brooks, D.; Rito, C.; Xu, Y.; Mccarthy,
J.; Kauffman, R. Diabetes 2003, 52, A141.
6. Xu, Y.; Rito, C. J.; Bosley, J. R.; Brooks, D. A.;
Dominianni, S. J.; Etgen, G.; Hahn, P. J.; Wilson, S. B.;
Kauffman, R. F.; Montrose-Rafizadeh, C.; Peters, M.;
Shuker, A. J.; Tripp, A. E.; Winneroski, L. L.; Zink, R.;
McCarthy, J. R. J. Med. Chem. 2004, 47, 2422.
7. Xu, H. E.; Lambert, M. H.; Montana, V. G.; Plunket, K.
D.; Moore, L. B.; Collins, J. L.; Oplinger, J. A.; Kliewer,
S. A.; Gampe, R. T., Jr.; McKee, D. D.; Moore, J. T.;
Willson, T. M. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
13919.
Furthermore, the combination of R2 and R3 exhibit
greater mouse PPARa binding and transactivation
activity than either alone as exemplified in compounds
8 and 11. Compound 11 showed a 18-fold improvement
in both assays compared to 1, establishing an SAR con-
vergence for this series on both human and mouse
PPARa. When R3 changes from hydrogen to tert-butyl,
the binding modes in human and mouse converge
and the additional van der Waals interaction from R2
helps the ligand achieve the high affinity to both recep-
tor isoforms. The convergence of human and mouse
PPARa binding and transactivation potential within this
series translated into improved rodent in vivo potency.
Using triglycerides reduction as an indicator of in vivo
PPARa activation, compound 1 reduced triglycerides
by ꢀ39% at a dose of 3mg/kg in mice treated for 7days,
whereas compound 11 lowered triglycerides by ꢀ56% at
the same dose level as compare to the control vehicle.
8. Miyachi, H.; Uchiki, H. Bioorg. Med. Chem. Lett. 2003,
13, 3145.
9. Brooks, D. A.; Etgen, G. J.; Rito, C. J.; Shuker, A. J.;
Dominianni, S. J.; Warshawsky, A. M.; Ardecky, R.;
Paterniti, J. R.; Tyhonas, J.; Karanewsky, D. S.; Kauff-
man, R. F.; Broderick, C. L.; Oldham, B. A.; Montrose-
Rafizadeh, C.; Winneroski, L. L.; Faul, M. M.; McCarthy,
J. R. J. Med. Chem. 2001, 44, 2061.
10. Godfrey, A. G.; Brooks, D. A.; Hay, L. A.; Peters, M.;
McCarthy, J. R.; Mitchell, D. J. Org. Chem. 2003, 68,
2623.
11. Homology modeling. The homology model of mouse
PPARa LBD was constructed using Modeler as imple-
mented in Insight 2000.1 (Accelrys, Inc., San Diego, CA).
The side chains of the conserved residues in the ligand
binding pocket were manually adjusted to achieve the best
overlay with the crystal structure of hPPARa in 1k7l.
12. Vieth, M.; Hirst, J. D.; Kolinski, A.; Brooks, C. L., III J.
Comput. Chem. 1998, 19, 1612.
In summary, the molecular modeling analysis for the
SAR trends in human and mouse PPARa receptors
for a novel series of a-aryloxy-a-methylhydrocinnamic
acids was described. Two of the four differences in the
human PPARa ligand binding pocket (F272 and
M279) predominantly confer the noted species-specific
selectivity. The SAR indicated the species selectivity
can be overcome by 4-phenoxy and phenyloxazole sub-
stitutions of these ligands.
Acknowledgements
The authors thank Don Jett, John Osborne, Michael
Denney, Robert Brickley, and Lawrence Goodwin for
technical support.
13. Vieth, M.; Hirst, J. D.; Dominy, B. N.; Daigler, H.;
Brooks, C. L., III J. Comput. Chem. 1998, 19, 1623.
14. Liu, K. G.; Smith, J. S.; Ayscue, A. H.; Henke, B. R.;
Lambert, M. H.; Leesnitzer, L. M.; Plunket, K. D.;
Willson, T. M.; Sternbach, D. D. Bioorg. Med. Chem.
Lett. 2001, 11, 2385.
References and notes
1. Keller, H.; Devchand, P. R.; Perroud, M.; Wahli, W. Biol.
Chem. 1997, 378, 651.