Design, development and evaluation of novel dual PPARδ/PPARγ agonists

Article abstract of DOI:10.1016/j.bmcl.2012.11.060

Type 2 diabetes is at epidemic proportions and thus development of novel pharmaceutical therapies for improving insulin sensitivity has become of paramount importance. The objectives of the current study were to develop novel dual PPARγ/δ agonists without

Full text of DOI:10.1016/j.bmcl.2012.11.060

Bioorganic & Medicinal Chemistry Letters 23 (2013) 873–879
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design, development and evaluation of novel dual PPARd/PPAR
c
agonists
Symon Gathiaka a, , Gayani Nanayakkara b, , Tracey Boncher c, Orlando Acevedo a, Johnathon Wyble b,
Sagar Patel a, Akash Patel a, Mary Elizabeth Shane b, Blake Bonkowski c, Jason Wieczorek c, Yinghui Rong d,
Kevin Huggins d, Forest Smith e, Rajesh H. Amin b,
⇑
a Dept. of Chemistry and Biochemistry, Auburn University, Auburn, AL 36849, USA
b PPAR and Metabolic Research Lab, Dept. of Pharmacal Sciences, Harrison School of Pharmacy, Auburn University, Auburn, AL 36849, USA
c Dept. of Pharmaceutical Sciences, Ferris State University, Big Raids, MI 49307, USA
d Dept. of Nutrition and Food Sciences, Auburn University, Auburn, AL 36849, USA
e Dept. of Pharmacal Sciences, Harrison School of Pharmacy, Auburn, AL 36849, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 27 July 2012
Revised 7 November 2012
Accepted 12 November 2012
Available online 3 December 2012
Type 2 diabetes is at epidemic proportions and thus development of novel pharmaceutical therapies for
improving insulin sensitivity has become of paramount importance. The objectives of the current study
were to develop novel dual PPAR
c
/d agonists without the deleterious side effects associated with full
agonists. Docking simulations of 23 novel compounds within the ligand binding domain of
/d were performed using AutoDock Vina which consistently reproduced experimental binding
PPAR
PPAR
c
c
poses from known PPAR agonists. Comparisons were made and described with other docking programs
AutoDock and Surflex-Dock (from SYBYL-X). Biological evaluation of compounds was accomplished by
transcriptional promoter activity assays, quantitative PCR gene analysis for known PPARc/d targets as
well as in vitro assays for lipid accumulation and mitochondrial biogenesis verses known PPAR agonists.
We found one (compound 9) out of the 23 compounds evaluated, to be the most potent and selective dual
Keywords:
Peroxisomal proliferator activating receptor
(PPAR)
AutoDock vina
Transciptional activity
Adiposity
PPARc/d agonist which did not display the deleterious side effects associated with full PPARc agonists.
Diabetes
Ó 2012 Elsevier Ltd. All rights reserved.
Type 2 diabetes mellitus is at epidemic proportions and is asso-
ciated with excessive cardiovascular morbidity and mortality. Both
impaired insulin secretion and resistance contribute to the devel-
opment of the disease. The thiazolidinediones (TZDs) represent a
class of drugs which improve whole body insulin sensitivity.1 The
beneficial effects of TZDs are attributed to the activation of the nu-
agonists for this class of receptors do not have a significant impact
upon improving insulin sensitivity.4 Consequently, because diabe-
tes and metabolic syndrome are associated with defects in glucose
oxidation and lipid metabolism, the concept of discovering dual
agonists which can activate both PPARb/d and PPAR
c simulta-
neously, has emerged as a potential therapeutic target for improv-
clear receptor class of transcription factors; PPARc (Peroxisomal
ing outcomes of diabetes. Thus, the goal of the current study was
Proliferator Activating Receptor). However, TZDs are associated
with adverse cardiovascular events mostly due to development
of peripheral edema and weight gain.2
to rationally discover new dual PPAR
without the deleterious side effects of full PPAR
accomplish this task a joint computational, synthesis and gene
regulation study of novel dual PPAR and PPARd agonists were
c
and PPARd agonists
c
agonists. To
The PPAR family of nuclear receptors constitute three members;
c
PPAR-
a
, b/d, and
c
that are highly conserved in mammals, each
carried out. Twenty-three nontraditional compounds have been
designed and synthesized as potential agonists. Docking simula-
tions of the proposed compounds within the ligand-binding do-
forming a functional heterodimeric complex with 9-cis retinoic acid
receptor (RXR). PPAR b/d activation is associated with improving
overall circulating cholesterol levels (HDL, LDL and triglycerides)
and is ubiquitously expressed throughout the body. Recently, it
has been found that overexpression of PPAR b/d in skeletal muscle
improves the glycolytic muscle fiber type in animal models and
thus improves circulating glucose and fatty acid levels.3 However,
main of PPAR
c
and PPARd have been performed. Compounds
/d as
found to be having high and low binding activity on PPAR
c
determined by the docking studies were evaluated for biological
significance. In vitro cellular assays were utilized to test transcrip-
tional regulation of both PPARc and PPARd targets as well as for
adiposity and mitochondrial biogenesis.
We studied 23 compounds by docking calculations (as de-
scribed further below) and predicted that compound 9 (Fig. 1) to
⇑
Corresponding author.
E-mail address: rha0003@auburn.edu (R.H. Amin).
These authors contributed equally to the manuscript.
possess the highest binding affinity in both PPAR
c and PPARd
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.11.060

874
S. Gathiaka et al. / Bioorg. Med. Chem. Lett. 23 (2013) 873–879
The general synthesis of compounds 9 and 3-121 are depicted
in Schemes 1 and 2.
F3C
N
OH
Scheme 1: Treatment of 4-(aminomethyl)benzoic acid 2 with
Fmoc chloride, Na2CO3, dioxane, water, at 0 °C provided 3. Nucleo-
philic substitution of 3 with 2-chlorotrityl chloride resin, D-IPEA,
COMU, methanol, DCM, and DMF provided intermediate 4. Subse-
quent treatment of 4 with 3,5-bis(trifluoromethyl) benzyl bromide,
DIPEA, sodium hydride and DCM, followed by deprotection of
Fmoc using DCM, DMF, Piperidine in a 1:1:2 ratio, yielded 5. The
Fmoc protected amino propyl bromide intermediate 6 was created
by reacting 3-bromopropylamine hydrobromide with Fmoc chlo-
ride, sodium carbonate, dioxane, water, at 0 °C. The intermediate
compound 6 was reacted with 5, DIPEA, sodium hydride, DCM,
DMF to generate 7. 4-Methylbenzyl bromide, DIPEA, sodium hy-
dride, DCM, DMF were reacted with 7 for 12 h before subsequent
treatment with DCM, DMF, Piperidine in a 1:1:2 ratio to remove
the Fmoc and generate 8. The last step involved to generate the fi-
nal compound 9 was to clip it from the resin using 90% TFA and
DCM for 1.5 h to dissociate resin from final compound 9.
CF3
O
H
N
CH3
Compound 9
O
O
HN
Scheme 2: Treatment of aminoacetaldehyde diethyl acetal 1
with 2-mesitylenesulfonyl chloride, 10% NaOH, CH2Cl2 for 24 h,
provided 2. Nucleophilic substitution of 4-(triflouromethyl)benzyl
bromide, NaH, DMF, under nitrogen, with 2 yielded 3. The last step
required to generate 3-121 was reacting 3 with HBr solution (33%
acetic acid), phenol, ethyl acetate, for 24 h, to create 3-121.
Computional methods: AutoDock45 AutoDock Vina6 and Surflex-
Dockfrom Sybyl-X7 were used to dock known and proposed
CF3
Compound 3-121
Figure 1. Compound 9 (top) and 3-121 (bottom) structures.
binding pockets. Compound 3-121 (Fig. 1) was found to have low
binding affinity for PPAR
c
and average affinity for PPARd.
O
O
a
O
O
HO
NH2
HO
HN
(2)
O
(3)
O
b
c, d
O
O
O
O
NH
R1
HN
Cl
(5)
CF3
(4)
CF3
R2
O
f
e
O
N
H
Br
O
N
O
R
2
NH
CF3
(7)
(6)
CH3
CF3
CH3
O
N
H
HO
i
O
g, h
N
O
N
CF3
H
N
(9)
R2
CF3
(8)
CF3
CF3
Compound 9
Scheme 1. Reagents and conditions: (a) Fmoc, Na2CO3, dioxane, water, at 0 °C; (b) 2-chlorotrityl chloride resin, DIPEA, COMU, methanol, DCM, DMF; (c) 3,5-bis
(trifluoromethyl) benzyl bromide, DIPEA, NaH, DCM; (d) DCM, DMF, piperidine, 1:1:2; (e) 3-bromopropylamine hydrobromide, Fmoc chloride, Na2CO3, dioxane, water at 0 °C;
(f) (8), DIPEA, NaH, DCM, DMF; (g) 4-methylbenzyl bromide, DIPEA, NaH, DCM, DMF; (h) DCM, DMF, piperidine, 1:1:2; (i) 90% TFA, DCM.

S. Gathiaka et al. / Bioorg. Med. Chem. Lett. 23 (2013) 873–879
875
(1)
(2)
(3)
O
O
O
O
O
O
a
b
CF3
NH
S
N
S
NH2
O
O
O
O
c
O
O
NH
CF3
(3-121)
Scheme 2. Reagents and conditions: (a) 2-mesitylenesulfonyl chloride, 10% NaOH,
CH2Cl2 24 h; (b) 4-(trifluoro-methyl) benzyl bromide, NaH, DMF; (c) HBR solution of
33% acetic acid, phenol, EtoAC.
Figure 2. The ligand binding domain of PPAR
c
bound with an agonist. Fig-
ure adapted from Zoete et al.1
agonists of PPAR
protein–ligand structures were derived from a reported 1.95 Å
crystal structure of PPAR
(PDB ID: 3ET3)8and a 2.00 Å crystal
structure of PPARd (PDB ID: 3GZ9).9 In the case of PPAR
, the pres-
ent model included the active site and all residues within 15 Å of it.
Clipped residues were capped with acetyl or N-methylamine. The
c and PPARd. Initial Cartesian coordinates for the
AutoDock details: The Lamarckian genetic algorithm17 was cho-
sen to search for the best conformers. The rigid roots of each ligand
were defined automatically and the amide bonds were made non-
rotatable. The docking process used for AutoDock mirrored that of
a recent study by Perryman and McCammon.18 For example, 100
conformers were considered for each compound. The population
size was set to 150, and the individuals were initialized randomly.
c
c
reduced PPAR
c
model consisted of ꢀ2800 atoms and 162 residues
out of 275 residues. The PPARd model used the protein in its en-
tirety. The protein targets were prepared for molecular docking
simulation by removing water molecules and bound ligands. Auto-
Dock Tools (ADT)10 was used to prepare and analyze the docking
simulations for the AutoDock and AutoDock Vina programs,
whereas Sybyl-X was used to prepare, carry out runs, and analyze
the Surflex-Dock simulations. All ligands were constructed using
PyMOL11 with subsequent geometry optimizations carried out
using the semi empirical method PDDG/PM312–14 and the BOSS
program.15 Polar hydrogens were added, and in the case of
AutoDock, Gasteiger charges were assigned. Nonpolar hydrogen
was subsequently merged. The protonation state of the ligands
was adjusted to the species assumed predominant at physiological
pH, specifically, carboxylic acid moieties were deprotonated.
Conjugate gradient minimizations of the systems were performed
using MCPRO15 and GROMACS.16 A grid was centered on the
catalytic active site region and included all amino acid residues
within a box size set at x = y = z = 70 grid points and 26 Å for
AutoDock and AutoDock Vina, respectively. AutoGrid 4 was used
to produce grid maps for AutoDock calculations where the search
space size utilized grid points of 0.375 Å.
The maximum number of energy evaluation was set to 9 Â 1010
.
The docking parameters used were: maximum number of genera-
tions was 9 Â 104, maximum number of top individuals that auto-
matically survived set was 1, mutation rate of 0.02, crossover rate
of 0.8, step sizes were 2 Å for translations, 50° for quaternions, and
50° for torsions, a cluster tolerance of 2 Å was employed, and the
maximum number of iterations in the pseudo-Solis-and-Wets-
minimization/local search was increased to 3000.
Surflex-Dock details: Following the receptor’s preparation, all
hydrogen bonds were added and the side chain amides in all Asn
and Gln oriented to maximize hydrogen bonding. Minimization
was performed for 100 iterations using the AMBER FF99 force
field.19 Prior to docking runs, the Surflex docking algorithm20 re-
quired the generation of a protomol,21an idealized representation
of the binding site that defines the search area. For this purpose,
the prepared receptor files were loaded and the protomol genera-
tion constructed based on protein residues that constitute the ac-
tive site, decreasing the threshold and increasing bloat
parameters values to 0.25 and 2 Å, respectively. The threshold
has a default value of 0.50 and decreasing this number increases
the volume while the bloat on the other hand has a default of 0
and increasing this value inflates the protomol. During the docking
runs, all other adjustable parameters were left at their default
values.
AutoDock Vina details: Standard flexible protocols of AutoDock
Vina using the iterated local search global optimizer1 algorithm
were employed to evaluate the binding affinities of the molecules
and interactions with the receptors. All ligands and active site res-
idues, as defined by the box size used for the receptors, were set to
be rotatable. Calculations were carried out with the exhaustiveness
of the global search set to 100, number of generated binding modes
set to 20, and maximum energy difference between the best and
the worst binding modes set to 5. Following completion of the
docking search, the final compound pose was located by evaluation
of AutoDock Vina’s empirical scoring function where the confor-
mation with the lowest docked energy value was chosen as the
best.
In silico: Design of small molecules that bind to a biological tar-
get have made great advancements in technique recently.22 Of par-
ticular interest, virtual screening methods that dock ligands into a
receptor, allow for a large number of compounds to be compared
quickly, including PPAR-c 23
However, virtual screening methods
.
often neglect important statistical and chemical contributions in
favor of computational efficiency.24 To examine the accuracy of
the present docking methods, rosiglitazone (PDB ID: 2PRG),25

876
S. Gathiaka et al. / Bioorg. Med. Chem. Lett. 23 (2013) 873–879
Table 1
versus crystal structure binding poses are given in the Supplemen-
tary data (Figs. S1–S9) and show that AutoDock Vina reproduced
the controls significantly better than the other docking programs.
Consequently, AutoDock Vina was used to dock and interpret data
for the current compounds.
Predicted binding affinities (kcal/mol) for proposed agonists in PPAR
c
and PPARd
PPAR PPARd
c
Molecule
Affinity
Molecule
Affinity
9
8
7
3-91
4-20
4
3-95
5
À12.0
À11.8
À11.1
À11.0
À10.2
À10.0
À9.9
À9.7
À9.6
À9.4
À9.4
À9.3
À9.2
À9.2
À8.8
À8.8
À8.4
À8.3
À7.9
À7.7
À7.5
À7.1
À7.0
À9.2
9
4-20
8
3-75
3-71
3-91
5
7
4
3-95
2
6
À10.8
À10.6
À10.6
À10.4
À10.2
À10.2
À10.1
À9.8
À9.6
À9.5
À9.3
À9.2
À9.2
À9.0
À8.9
À8.6
À8.6
À8.3
À8.2
À7.7
À7.4
À7.1
À6.9
À12.2
Biological assays: To validate the specificity of the compounds
towards activation of PPARc/d targets, we tested the compounds
capacity to bind to select PPARc/d Peroxisome Proliferator Re-
sponse Elements (PPRE). The PPRE are unique sites located in the
promoter region where PPARs bind and transcriptionally activate
the target genes. AP2-PPRE is the PPARc target involved in adipo-
cyte growth and differentiation. PDK4–PPRE is a PPARd target in-
volved in regulating energy metabolism in the cell. These PPREs
were inserted into a PGL3 luciferase vectors containing CMV pro-
moters (ADDGENE) by PCR. These vectors were co-transfected with
6
3-71
3-65
3-75
4-21
4-23
3-137
2
3
RXR, PPARc/PPARd and b-galactosidase vectors into HEK-293 cells
3-65
4-21
3-121
4-23
3-115
1
3-137
3-133
4-19
3-81
using Lipofectamine 2000 (Invitrogen). Relative light units (RLU)
were measured using a Glomax Luminometer (Promega). Data
were standardized to b-galactosidase activity using ONPG
(Promega).
We then validated the physiological effects of our compounds
upon lipid accumulation; 3T3L1 adipocytes were cultured and
treated with compounds for a period of 6 days. Lipid accumulation
was determined by oil red-o staining, and measured the absor-
bance spectrophotometrically at a wavelength of 510 nm and stan-
dardized to total protein concentrations. Data from lipid
accumulation studies were compared to cells treated with rosiglit-
3
3-115
1
3-121
3-133
4-19
3-81
Rosi. (PPAR
c
)
D321 (PPARd)
indeglitazar (PDB ID: 3ET3),8 and D321 (PDB ID: 3GZ9)26 in PPAR
c
azone (full PPARc agonist).
Recently it has been found that PPARd agonists (GW0742) in-
duce mitochondrial biogenesis.27 Therefore to better understand
the level of PPARd agonism of our compounds, C2C12 skeletal mus-
cle cells were cultured and were treated with the compounds for
4 days. Changes associated with markers for mitochondrial biogen-
esis were determined by quantitative real time PCR.
and PPARd have been calculated using AutoDock, AutoDock Vina,
and Surflex-Dock and the resultant poses compared to the crystal
structures. The flexible protocols of AutoDock Vina yielded
computed.
RMSD values 0.88, 0.83, and 0.07 Å for rosigilitazone, indeglita-
zar, and D321, respectively, for all atoms between the controls and
the docked conformations. AutoDock’s computed RMSD values
were 1.87, 1.40 and 1.07, while Surflex-Dock’s were 1.16, 1.81
and 0.70 Å for the three agonists. Illustrations of the calculated
Twenty-three proposed synthetic agonists have been docked
with both PPARc and PPARd using above mentioned docking
simulation programs to elucidate the interactions with active site
Figure 3. Predicted images of compounds 9 and 3-121 bond patterns. (A) Compound 9 bound to the active site of PPARc (top) and PPARd (bottom) with key residues shown.
(B) Compound 3-121 bound to the active site of PPAR (top) and PPARd (bottom) with key residues shown.
c

S. Gathiaka et al. / Bioorg. Med. Chem. Lett. 23 (2013) 873–879
877
residues. The PPAR ligand binding domain resembles a large
Y-shaped cavity starting from the entrance and extending into
Arm I and Arm II pockets (Fig. 2).28 Arm I is substantially polar
while the entrance and Arm II are primarily hydrophobic. Out
of the 23 compounds studied, the calculations predicted compound
phenyl ring is revealed to accept a NÀHÁ Á Á
p
hydrogen bond of ca.
2.69 Å from His449. It is interesting to note that this interaction
holds the ring moiety in position making 9 avoid contact with
Tyr473 which is ca. 4.8 Å away. This residue is crucial to the
stabilization of the AF2 helix H12 which allows the binding of
co-activators that lead to the activation of the genes responsible
for adipogenesis.29 Our in silico data were confirmed by transcrip-
tional assays which demonstrate that compound 9 minimally acti-
vates the AP2-PPRE (Fig. 4A). The results appear consistent with
the present lipid accumulation assays, where compound 9 has
negligible effects upon lipid accumulation in adipocytes (Fig. 4B).
Furthermore compound 9 marginally induced the expression of
genes associated with lipid accumulation and synthesis such as
9 to possess the highest binding affinity for both PPAR
c and
PPARd (Table 1). Inspection of the in silico PPAR /9 complex
c
suggests that the carboxylate group of compound 9 forms a
hydrogen bond at a distance of 2.02 Å with the entrance residue
Glu343 while the 2° amine donates a tighter hydrogen bond of
1.85 Å with the polar Ser289 residue of Arm I (Fig. 3A). The nonca-
nonical XÀHÁ Á Á
p hydrogen bond has been shown to be of great
importance in proteins (ref) and 9’s trifluoromethyl disubstituted
AP2 dose curve
A
15000
10000
5000
0
Rosi
3-121
9
3-91
-2
0
2
Drug concentration [log (μM)]
B
Control
Rosiglitazone (10μM)
3-91
3-121
9
4-23
Lipid accumulation
C
**
200
150
100
50
Control
Rosiglitazone
3-91
3-121
9
**
***
4-23
0
0
10
20
30
40
50
60
Drug concentration (μM)
Figure 4. PPARc biological studies. (A) AP2-PPRE dose–response curve. Data from luciferase assays shows that compound 3-91 and rosiglitazone (10 lM) have strong binding
affinity for the AP2-PPRE. However compounds 9 and 3-121 show less binding affinity for the AP2-PPRE. (B) Compound 3-91 induces lipid accumulation comparable to
rosiglitazone in 3T3-L1 adipocytes, as demonstrated by the oil red-o stain (red color). Compounds 9, 3-121 and 4-23 induce less lipid accumulation. (C) Lipid accumulation
levels were quantitated by measuring the absorbance at a wavelength of 510 nm and graphically represented as a percent change from vehicle treated cells (control).⁄P <0.05,
⁄⁄P <0.005, ⁄⁄⁄P <0.0001.

878
S. Gathiaka et al. / Bioorg. Med. Chem. Lett. 23 (2013) 873–879
PPARd in circulation.28 The 2° amine also donates a H-bond of
PDK4 dose vs reponse curve
A
B
2.87 Å to the Sulfur atom of the polar residue Cys285 in Arm I.
Furthermore, the methyl substituted phenyl ring in compound 9
located between the entrance and Arm II and surrounded by the
hydrophobic residues Leu330, Leu339 and Phe368 stabilize the li-
gand binding domain. The biological significance of these modifica-
tions were determined by luciferase assays (Fig. 5A) which
demonstrate that compound 9 binds to PDK4–PPRE, a direct PPARd
target. We also found an increase of PDK4 gene expression in adi-
pocytes (Fig. 6). Furthermore, an increased gene expression of
mitochondrial markers was observed in Figure 5B. Interestingly,
compound 9 significantly increased cytochrome oxidase c 4-2
subunit which is involved in inducing mitochondrial respiration
and biogenesis in response to ischemia in the heart.30 This suggests
a possible mechanism by which compound 9 can offer cardio
protection against ischemic injury.
15000
10000
5000
0
Rosi
GW0742
3-121
9
3-91
-2
0
2
Drug concentration [log(μM)]
Mitochondrial markers
8
6
4
2
0
COX4-1
To further validate the significance of the AutoDock Vina mod-
eling of PPARc/d agonism, we tested another compound, 3-121.
Out of the 23 compounds tested, compound 3-121 was ranked in
the lower end of the relative binding affinities for both receptors
(Table 1). Our analysis predicted that the ether oxygen atoms of
3-121 accept two hydrogen bonds of ca. 2.17 and 2.86 Å from the
backbone NH group of the polar residue Glu343 in the entrance re-
ddd
COX 4-2
NRF-1
TFAM
bbb
ccc
ddd
bbb
dd
ccc
b
c
a
a
a
gion of PPARc (Fig. 3B). The calculations do not predict binding
with any of the polar residues located in Arm I; this appears con-
sistent with the lipid accumulation studies which demonstrate
an insignificant induction of adiposity with compound 3-121 com-
Con
GW 0742
3-121
9
3-91
Figure 5. PPARd biological studies. (A) PDK4–PPRE dose–response curve. Data from
luciferase assays shows that compound 9 and GW0742 have strong binding affinity
for the PDK4–PPRE. However compounds 3-91 and 3-121 show less binding affinity
for the PDK4–PPRE. (B) Compound 9 and GW0742 increases the gene expression of
mitochondrial markers in C2C12 skeletal muscle cells as demonstrated by quanti-
tative RT-PCR. Data are means SEM from 3 independent experiments. Values were
set fold change from control. ⁄P <0.05, ⁄⁄P <0.005, ⁄⁄⁄P <0.0001.
pared to full PPARc agonists (Fig. 4). The docking studies also pre-
dicted 3-121 to bind weakly to PPARd due to the absence of the
conserved hydrogen bonding. These binding predictions correlate
well with results from the luciferase assays which indicates that
3-121 did not activate the PPARc target; AP2 or the PPARd target
PDK4.
Compound 3-91 was predicted to display high PPAR
capacity because it possesses the same binding mode as the full
PPAR agonist rosiglitazone. This was demonstrated in our docking
analysis which predicted that the PPAR receptor forms two
hydrogen bonds with the isoindole-dione substituent group of
3-91, the hydrogen on His323 and a carbonyl oxygen, and the
hydrogen on His449 and the nitrogen at distances of 2.38 and
2.66 Å, respectively (Supplementary data S10). In addition, the
OH group of Ser289 also forms a hydrogen bond of 2.57 Å with
the N of the isoindole-dione substituent group. This binding mode
is consistent with known full agonists, for example, rosiglitazone.31
c activation
c
FATP, CD36 and DGAT (Fig. 6). However, compound 9 elevated the
expression of genes associated with fatty acid oxidation (MCAD,
HSL and ATGL) and insulin sensitivity (adiponectin) (Fig. 6).
Docking studies of our compounds on PPARd shows that the
carboxylate group of the receptor forms a hydrogen bond of
3.03 Å with Thr289 in Arm I which is part of the His323, His449
and Tyr473 hydrogen bond network. This site is specifically in-
volved in the carboxylate group of fatty acids and eicosanoic acids
(the endogenous ligands) binding, which are natural ligands for
c
8
Control
Rosi
GW-0742
3-121
9
PPARγ/δ targets
iii
6
4
2
0
iii
eee
eee
eee
hh
gg
a
bb
hh
hh
ff
dd
aa
dd
ff
cc
ggg
bb
bb
b
cc
ff
g
h
c
a
d
a
c
DGAT FABP FATP
CD36 ADPN
HSL
MCAD ATGL PDK4
lipogenesis
fatty acid oxidation
PPARγ
PPARδ
Figure 6. MRNA analyses of PPAR
c
/d targets. Rosiglitazone, GW0742, 3-121 and 9 (10
lM dose for all compounds), were applied to 3T3-L1 adipocytes to determine effects on
PPAR /d target mRNA expression levels. Compound 9 increased the gene expression of PPAR
c
c
targets MCAD, HSL, ATGL and adiponectin. Furthermore compound 9 also
increased the gene expression of PPARd target, PDK4. Values represent the fold change from control mRNA expression levels, where a, b, d, g, h; P <0.05, aa, bb, cc, dd, ee, ff, gg,
hh; P <0.005, eee, ggg, iii; P <0.0001.

S. Gathiaka et al. / Bioorg. Med. Chem. Lett. 23 (2013) 873–879
879
Our cellular assays which tested the biological effects of compound
3-91 on lipid accumulation verified it to be a strong PPAR activa-
tor and thus induced comparable levels of lipid accumulation to
rosiglitazone. However 3-91 does not form substantial hydrogen
bonds with PPARd despite having a larger binding affinity relative
to 3-121.
References and notes
c
1. Amin, R. H.; Mathews, S. T.; Camp, H. S.; Ding, L.; Leff, T. Am. J. Physiol.
Endocrinol. Metab 2009, 298, E28.
2. Lago, R. M.; Singh, P. P.; Nesto, R. W. Lancet 2007, 370, 1129.
3. Wang, Y. X.; Zhang, C. L.; Yu, R. T.; Cho, H. K.; Nelson, M. C.; Bayuga-Ocampo, C.
R.; Ham, J.; Kang, H.; Evans, R. M. PLoS Biol. 2004, 2, e294.
4. Dimopoulos, N.; Watson, M.; Green, C.; Hundal, H. S. FEBS Lett. 2007, 581, 4743.
5. Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew, R. K.; Goodsell, D.
S.; Olson, A. J. J. Comput. Chem. 2009, 30, 2785.
6. Trott, O.; Olson, A. J. J. Comput. Chem. 2010, 31, 455.
7. Huang, Y.; Hickey, R. P.; Yeh, J. L.; Liu, D.; Dadak, A.; Young, L. H.; Johnson, R. S.;
Giordano, F. J. FASEB J. 2004, 18, 1138.
8. Artisa, D. R.; Lina, J. J.; Zhanga, C.; Wanga, W.; Mehraa, U.; Perreaultb, M.; Erbeb,
D.; Krupkaa, H. I.; Englanda, B. P.; Arnolda, J.; Plotnikova, A. N.; Marimuthua, A.;
Nguyena, H.; Willb, S.; Signaevskyc, M.; Kralc, J.; Cantwella, J.; Settachatgulla,
C.; Yana, D. S.; Fonga, D.; Oha, A.; Shia, S.; Womacka, P.; Powella, B.; Habetsa,
G.; Westa, B. L.; Zhanga, K. Y. J.; Milburna, M. V.; Vlasukb, G. P.; Hirtha, K. P.;
Nolopa, K.; Bollaga, G.; Ibrahima, P. N.; Tobinb, J. F. Proc. Natl. Acad. Sci. U.S.A.
2009, 106, 262.
9. Connors, R. V.; Wang, Z.; Harrison, M.; Zhang, A.; Wanska, M.; Hiscock, S.; Fox,
B.; Dore, M.; Labelle, M.; Sudom, A.; Johnstone, S.; Liu, J.; Walker, N. P.; Chai, A.;
Siegler, K.; Li, Y.; Coward, P. Bioorg. Med. Chem. Lett. 2009, 19, 3550.
10. Morris, G. M.; Huey, R.; Lindstrom, W.; Sanner, M. F.; Belew, R. K.; Goodsell, D.
S.; Olson, A. J. J. Comput. Chem. 2009, 30, 2785.
11. Xue, W.; Cai, L.; Tan, Y.; Thistlethwaite, P.; Kang, Y. J.; Li, X.; Feng, W. Am. J.
Pathol. 1995, 177, 97.
12. Repasky, M. P.; Chandrasekhar, J.; Jorgensen, W. L. J. Comput. Chem. 2002, 23,
1601.
13. Tubert-Brohman, I.; Guimarães, C. R. W.; Repasky, M. P.; Jorgensen, W. L. J.
Comput. Chem. 2003, 25, 138.
14. Tubert-Brohman, I.; Guimarães, C. R. W.; Jorgensen, W. L. J. Chem. Theory
Comput. 2005, 1, 817.
15. Jorgensen, W. L.; Tirado-Rives, J. J. Comput. Chem. 2005, 26, 1689.
16. Hess, B.; Kutzner, C.; Spoel, D. v. d.; Lindahl, E. J. Chem. Theory Comput. 2008, 4,
435.
17. Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart, W. E.; Belew, R. K.;
Olson, A. J. J. Comput. Chem. 1998, 19, 1639.
18. Perryman, A. L.; McCammon, J. A. J. Med. Chem. 2002, 45, 5624.
19. Lindorff-Larsen, K.; Piana, S.; Palmo, K.; Maragakis, P.; Klepeis, J. L.; Dror, R. O.;
Shaw, D. E. Proteins 1950, 2010, 78.
In summary we have developed a series of novel dual PPARd/
agonists which spans from full agonists for PPAR as well as PPARd.
Binding studies: Our computational analysis allows us to predict
compounds which will have strong binding affinity for PPAR or
PPARd. Interestingly, our docking studies indicate that compound
9 possess the highest binding affinity for both PPAR and PPARd
c
c
c
c
out of the 23 compounds studied. A detailed analysis of compound
9 shows that it forms hydrogen bonds with Leu228, Ser289 and
Glu295 in PPAR
viously mentioned, these sites are essential for both PPAR
c
binding pocket and with Thr289 in PPARd. As pre-
and
c
PPARd activity. Furthermore we have found that compound 9 is a full
PPARd agonists as demonstrated by our luciferase and gene expres-
sion studies which includes mitochondrial markers in muscle cells.
Adiposity: In regards to PPARc activation, our goals were to de-
velop compounds which will lack the deleterious side effects found
with current TZDs, including weight gain and the eventual heart
failure.32 We have found that compound 9 does not interact with
Tyr473 of the PPAR
ization of the AF2 helix H12 and has been shown to lead to adipo-
genesis.29 These conformational changes of the PPAR
binding
pocket by compound 9 selectively activates PPAR target genes
c binding pocket, a residue crucial to the stabil-
c
c
to have negligible effects on lipid accumulation while activating
other genes involved in regulating fatty acid oxidation(HSL, ATGL
and MCAD).
In conclusion, we have found that compound 9 is a potent
PPARd activator. Although our docking studies indicate that com-
20. Jain, A. N. J. Med. Chem. 2003, 46, 499.
pound 9 is a full PPAR
compound 9 lacks the deleterious effect of lipid accumulation
found with full PPAR agonist rosiglitazone. These findings will al-
c agonist our biological assays indicate that
21. Ruppert, J.; Welch, W.; Jain, A. N. Protein Sci. 1997, 6, 524.
22. Acevedo, O.; Ambrose, Z.; Flaherty, P. T.; Aamer, H.; Jain, P.; Sambasivarao, S. V.
Curr. Pharm. Des. 2012, 18, 1199.
23. Lewis, S. N.; Bassaganya-Riera, J.; Bevan, D. R. PPAR Res. 2009, 2010, 861238.
24. Klebe, G. Drug Discovery Today 2006, 11, 580.
25. Nolte, R. T.; Wisely, G. B.; Westin, S.; Cobb, J. E.; Lambert, M. H.; Kurokawa, R.;
Rosenfeld, M. G.; Willson, T. M.; Glass, C. K.; Milburn, M. V. Nature 1998, 395,
137.
26. Connors, R. V.; Wang, Z.; Harrison, M.; Zhang, A.; Wanska, M.; Hiscock, S.; Fox,
B.; Dore, M.; Labelle, M.; Sudom, A.; Johnstone, S.; Liu, J.; Walker, N. P.; Chai, A.;
Siegler, K.; Li, Y.; Coward, P. Bioorg. Med. Chem. Lett. 2009, 19, 3550.
27. Miglio, G.; Rosa, A. C.; Rattazzi, L.; Grange, C.; Camussi, G.; Fantozzi, R. Br. J.
Pharmacol. 2012.
28. Zoete, V.; Grosdidier, A.; Michielin, O. Biochim. Biophys. Acta 2007, 1771, 915.
29. Lewis, S. N.; Bassaganya-Riera, J.; Bevan, D. R. PPAR Res. 2010, 2010, 861238.
30. Fukuda, R.; Zhang, H.; Kim, J. W.; Shimoda, L.; Dang, C. V.; Semenza, G. L. Cell
2007, 129, 111.
31. Gampe, R. T., Jr.; Montana, V. G.; Lambert, M. H.; Miller, A. B.; Bledsoe, R. K.;
Milburn, M. V.; Kliewer, S. A.; Willson, T. M.; Xu, E. Mol. Cell 2000, 5, 545.
32. Nesto, R. W.; Bell, D.; Bonow, R. O.; Fonseca, V.; Grundy, S. M.; Horton, E. S.; Le
Winter, M.; Porte, D.; Semenkovich, C. F.; Smith, S.; Young, L. H.; Kahn, R.;
American Heart Association; American Diabetes Association Circulation 2003,
108, 2941.
c
low us to focus upon further development of compound 9 deriva-
tives and studying the effects of these compounds upon insulin
signaling and metabolism in diabetic animal models.
Clinically: Current findings indicate that full PPAR
c agonists
(Rosiglitazone) increase the risk of adverse cardiovascular events,
mostly due to the increase in fluid retention. To explain this rosig-
litazone has been found to increase the epithelial sodium channel
in the collecting duct of the kidney, resulting in development of
edema.33 Consequently our novel dual PPAR-
c/PPAR-d agonist will
be tested for the impact of compound 9 upon the expression of epi-
thelial sodium channels in the collecting ducts of normal and dia-
betic mice.
Supplementary data
33. Vallon, V.; Hummler, E.; Rieg, T.; Pochynyuk, O.; Bugaj, V.; Schroth, J.;
Dechenes, G.; Rossier, B.; Cunard, R.; Stockand, J. J. Am. Soc. Nephrol. (JASN)
2009, 20, 721.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.
11.060.