Pan-PPAR agonists based on the resveratrol scaffold: Biological evaluation and docking studies

Full text of DOI:10.1002/cmdc.201000360

DOI: 10.1002/cmdc.201000360
Pan-PPAR Agonists Based on the Resveratrol Scaffold: Biological Evaluation
and Docking Studies
Wei Li, Xinhua He, Weiguo Shi, Haoyan Jia, and Bohua Zhong*[a]
The peroxisome proliferator-activated receptors (PPARs) belong
to the nuclear hormone receptor superfamily and consist of
three receptor isoforms, PPARa, PPARg, and PPARd, which play
important roles in carbohydrate and lipid metabolism. PPARa
agonists such as fenofibrate, or the active metabolite fenofibric
acid (1), are effective at lowering serum triglycerides and rais-
ing high-density lipoprotein (HDL) cholesterol.[1] PPARg has
been identified as a key regulator for insulin sensitivity, glucose
homeostasis, and fat storage.[2] PPARd activation appears to in-
crease fatty acid b-oxidation, insulin sensitivity, and HDL cho-
lesterol. PPARd agonists have been shown to increase plasma
HDL cholesterol levels while decreasing LDL cholesterol and tri-
glycerides in obese and dyslipidemic rhesus monkeys.[3]
Considering the aforementioned beneficial pharmacological
effects of the PPARs, the concept of simultaneously activating
all PPAR subtypes with a single compound, that is, a pan-PPAR
agonist is extremely attractive, especially for the treatment of
metabolic syndrome, which consists of an accumulation of
metabolic and cardiovascular risk factors that cause a predis-
position to heart attack, stroke, heart failure, sudden cardiac
death, and certain cancers.[4] Simultaneous activation of all
PPAR subtypes might also decrease the occurrence of adverse
side effects.[4]
sveratrol (trans-resveratrol; 2), the most widely studied stil-
bene, is a phytoalexin produced naturally by several plants
when under attack by pathogens such as bacteria or fungi. Re-
sveratrol is found in the skin of red grapes and is a constituent
of red wine; indeed a lower risk of cardiovascular disease has
been observed among wine-drinking populations.[6] In mouse
and rat experiments, anticancer, anti-inflammatory, blood-
sugar-lowering, and other beneficial cardiovascular effects of
resveratrol have been reported.[7] Resveratrol protects the
heart and blood vessels by directly scavenging oxidants that
can cause lipid oxidation and by preventing oxidative-stress-in-
duced apoptosis of endothelial cells.[8] It may also help to pre-
vent heart damage after cardiac arrest. Reduced platelet aggre-
gation by resveratrol can decrease the risk of atherosclerosis.[9]
Resveratrol has also been demonstrated to decrease blood
lipid levels in animals.[10] Some naturally occurring or synthetic
resveratrol analogues have been shown to significantly activate
PPARa or to lower plasma lipid levels when fed to hamsters.[11]
These beneficial effects of resveratrol have attracted much
attention, and this compound may provide a scaffold for the
development of novel therapeutic drugs, which would proba-
bly produce synergistic pharmacological effects. In previous
studies, simple functional groups were introduced at the 4’-po-
sition, and methoxy groups were introduced at the 3- and 5-
positions of resveratrol.[12] From ongoing studies in our re-
search group, a series of phenoxyalkylcarboxylic acid deriva-
tives based on the resveratrol scaffold appear to have good
hypolipidemic activity in vivo, and compound 3 was found to
be the most potent. It was predicted that the hypolipidemic
activity of 3 may be due to the activation of the PPARa, and
the results of these studies will be reported elsewhere. In con-
tinuation of our drug discovery program on agents for the
treatment of metabolic diseases, and in order to gain more in-
sight on the structure–activity relationships, we discovered the
pan-PPAR agonists 6 and 9 by combining the resveratrol scaf-
fold with the fibrate head group and methylating the remain-
ing phenolic hydroxy groups. Interestingly, compound 6, bear-
ing two fibrate head groups, was found incidentally as the by-
product of the reaction. Herein we describe the synthesis and
in vitro evaluation of these compounds for PPAR transactiva-
tion. We also evaluated their in vivo biological activity and car-
ried out docking studies.
Stilbene-based components, or stilbenoids, have been sug-
gested to have many health benefits including antioxidant,
anti-inflammatory, antileukemic, antibacterial, antifungal, anti-
platelet aggregation, vasodilator, and antitumor activities.[5] Re-
[a] Dr. W. Li,+ X. He, W. Shi, H. Jia, Prof. B. Zhong
Beijing Institute of Pharmacology and Toxicology
27 Tai-Ping Road, 100850, Beijing (China)
Fax: (+86)1066931639
E-mail: bohuazhong@yahoo.com
[+] Current address: CAS Key Laboratory of
The synthesis of the target compounds are shown in
Schemes 1 and 2. Compared with 3, the target compound 6
has two fibrate head groups and one methoxy group. The fi-
brate head group was introduced at the 3- and 4’-positions of
resveratrol by alkylation of the phenolic hydroxy groups with
isobutyl-5-chloro-2,2-dimethylvalerate; the remaining phenolic
hydroxy group was methylated by methyl iodide. Alkaline hy-
drolysis of ester 5 gave the target compound 6. The target
Pathogenic Microbiology and Immunology (CASPMI)
Institute of Microbiology, Chinese Academy of Sciences
1 Beichen Xilu Road, Chaoyang District, 100101, Beijing (China)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000360.
ChemMedChem 2010, 5, 1977 – 1982
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1977

MED
Table 1. In vitro PPARa activation by tested compounds.
Compd
Fold PPARa activation[a]
3 mm
30 mm
3
9
1.53Æ0.11
2.03Æ0.28
1.67Æ0.11
1.52Æ0.11
2.36Æ0.31
2.42Æ0.24
fenofibric acid
[a] Fold activation at the indicated concentration relative to vehicle con-
trol; values are expressed as the mean ÆSE (n=3).
acid (1.67-fold) and 3 (1.53-fold). However, at 30 mm, com-
pound 9 demonstrated similar PPARa agonist activity (2.36-
fold activation) to that of fenofibric acid (2.42-fold) and again
higher potency than 3 (1.52-fold activation).
Comparative dose–response studies of 6 and 9 were further
performed with fenofibric acid (a PPARa agonist) toward
PPARa, rosiglitazone (a PPARg agonist) toward PPARg, and 2-
bromohexadecanoic acid (a PPARd agonist) toward PPARd. As
shown in Table 2, compounds 6 and 9 were identified as pan-
PPAR agonists, as they activate all three PPAR subtypes at simi-
lar doses and have balanced EC50 data for the three PPAR sub-
types, albeit with slight selectivity for PPARa. They act as weak
agonists of PPARg and PPARd relative to rosiglitazone and 2-
bromohexadecanoic acid, respectively. Compounds 6 and 9
have similar PPAR agonist activity and selectivity. Although 6,
9, and fenofibric acid activate PPARa in a similar manner,
PPARg and PPARd activation by 6 and 9 is ~5–10-fold more
potent than that of fenofibric acid, whereas fenofibric acid is a
dual activator of PPARa and
Scheme 1. Reagents and conditions: a) isobutyl-5-chloro-2,2-dimethylvaler-
ate, anhydrous K2CO3, DMF, 708C, 32 h, 62%; b) CH3I, anhydrous K2CO3, ace-
tone, reflux, 12 h, 73%; c) LiOH, MeOH/H2O, reflux, 48 h, then HCl, 54%.
compound 9 was afforded by hydrolysis of ester 8, which was
first provided by methylation of the 3- and 4’-positions of re-
sveratrol through a Williamson–Ether method and then alkyla-
tion of the phenolic hydroxy group at the 5-position with iso-
butyl-5-chloro-2,2-dimethylvalerate (Scheme 2). In previous
PPARg, with ~10-fold selectivity
for PPARa.[14] In general, com-
pounds 6 and 9 are more effec-
tive than fenofibric acid in their
ability to activate PPARs.
The ability of the compounds
to affect lipid/cholesterol ho-
meostasis was initially examined
in vivo by using the Triton WR-
1339 induced hyperlipidemic
mouse model, and the results
Scheme 2. Reagents and conditions: a) CH3I, anhydrous K2CO3, acetone, reflux, 12 h, 32%; b) isobutyl-5-chloro-2,2-
dimethylvalerate, anhydrous K2CO3, DMF, 708C, 24 h, 64%; c) LiOH, MeOH/H2O, reflux, 32 h, then HCl, 72%.
are listed in Table 3. Compound
6 showed greater activity than 3,
through replacement of the phe-
studies, 3,4’,5-trimethoxyresveratrol was obtained through the
Williamson–Ether method with methyl iodide.[5a,11,13] In our
study, 3,4’-dimethoxyresveratrol (7) was obtained by control-
ling the reaction conditions (molar ratios of reactants and reac-
nolic hydroxy group at the 3-position by a fibrate head group.
Out of the three compounds 3, 6, and 9, compound 9 was
identified as the most potent followed by 6; compounds 6
and 9 were more than twice as potent as fenofibric acid at a
1
tion time). All target compounds were confirmed by H and
13C NMR spectroscopy as well as mass spectrometry.
dose of 300 mgkgÀ1 body weight p.o. At
a dose of
100 mgkgÀ1 administered by i.p. injection, compound 6 and
fenofibric acid showed similar hypolipidemic activity, and com-
pound 9 was the most potent, lowering triglyceride levels by
80.1% (p<0.01) and total cholesterol by 66.7% (p<0.01). The
parent compound, resveratrol, did not show significant hypoli-
pidemic activity at the same dose.
The target compounds were evaluated for their activities as
PPAR agonists in U2OS cells. As shown in Table 1, fold activa-
tion values of PPARa by compounds 3 and 9 at 3 and 30 mm
were first obtained with vehicle control defined as 1-fold acti-
vation and fenofibric acid used as positive control. Compounds
3 and 9 showed PPARa agonist activity: at 3 mm, 9 (2.03-fold
activation) exhibited higher PPARa activation than fenofibric
Compounds 3, 6, and 9 were further tested for their hypoli-
pidemic activity in vivo in an alloxan-induced diabetic mouse
1978
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1977 – 1982

weight ratio), and serum bio-
chemical parameters in hyperli-
pidemic Balb/c mice fed a fat
Table 2. In vitro transactivation activity profile of tested compounds.
Compd
EC50 [mm][a] (activation [%][b]
)
PPARa
PPARg
PPARd
emulsion
at
a
dose
of
80 mgkgÀ1 body weight daily for
19 days. The results are listed in
Table 4. The most potent com-
6
9
6.87 (137.6)
8.46 (124.9)
8.51 (100)
ND[c]
19.37 (38.9)
11.94 (43.7)
104.72 (35.0)
0.03 (100)
ND
26.13 (74.6)
11.15 (78.2)
125.91 (55.4)
ND
fenofibric acid
rosiglitazone
2-bromohexadecanoic acid
pound
9 lowered triglyceride
ND
4.36 (100)
levels by 64.5% (p<0.01), and 6
lowered the triglycerides by
34.0% (p<0.05). The reference
compound fenofibric acid low-
ered the triglycerides by 50%
(p<0.01). The liver weight and
liver coefficient were significantly
increased in the model group
(p<0.05) compared with the
control group. Treatment with
fenofibric acid and compounds
[a] EC50: concentration that effects 50% maximal activity for a given compound; data represent the mean from
at least three independent experiments, each conducted in triplicate. [b] Percent maximal responses relative to
control were calculated with fenofibric acid at 10 mm as 100% for PPARa, rosiglitazone at 10 mm as 100% for
PPARg, and 2-bromohexadecanoic acid at 10 mm as 100% for PPARd. [c] ND: not determined.
Table 3. In vivo efficacy of compounds in acute hyperlipidemic mice.[a]
Compd
300 mgkgÀ1 p.o. (Triton)
100 mgkgÀ1 i.p. (Triton)
150 mgkgÀ1 p.o. (alloxan)
TG
TC
TG
TC
TG
GLU
3
6
9
47.1*
79.7**
86.1**
12.8
40.3*
61.8**
66.0**
9.4
52.2*
60.9**
80.1**
18.1
50.3*
47.0**
66.7**
11.2
59.7**
63.8**
67.2**
NT[b]
11.1
3.0
8.6
NT
13.2
6
and 9 further exacerbated
hepatomegaly, because the liver
weight and the liver coefficient
of mice in these groups were in-
creased relative to the model
group (p<0.01). The present
study supports the concept that
long-term treatment with PPARa
resveratrol
fenofibric acid
38.7*
30.6*
59.1*
46.2*
54.6**
[a] Values are expressed as percent decrease; TC: total cholesterol, TG: triglycerides, GLU: glucose; **p<0.01,
*p<0.05 relative to model group. [b] NT: not tested.
model at a dose of 150 mgkgÀ1 p.o. (Table 3). In this particular
mouse model, triglycerides and blood glucose levels are signif-
icantly higher than normal. The test compounds showed good
lowering of triglycerides, and are more potent than fenofibric
acid. Compound 9 was the most potent, lowering triglyceride
levels by 67.2% (p<0.01), followed by 6 (63.8%). Because this
animal model is mainly type 1 diabetic, none of the com-
pounds exhibited potent blood glucose lowering activity.
We next observed the effects of these compounds on body
weight, liver weight, liver coefficient (liver weight/body weight
ratio), spleen weight, spleen coefficient (spleen weight/body
agonists induces hepatomegaly in rodents. When administrat-
ed to rodents, fibrates produce a liver-specific response, result-
ing in peroxisomal proliferation, hepatomegaly, and ultimately
hepatocellular carcinoma.[15] This effect has been demonstrated
to be mediated by PPARa.[15] The magnitude of this response
appears to vary considerably among species. In non-human
primates and humans, PPARa agonists neither induce peroxi-
some proliferation nor the development of liver cancer.[16]
Indeed, fibrates have been prescribed extensively for more
than 30 years, and an increased risk of liver cancer has never
been reported in humans.[17] It is therefore likely that com-
Table 4. Effects of compounds on hyperlipidemic Balb/c mice fed a fat emulsion.[a]
Parameters
Compound
control
model
6
9
fenofibric acid
Body weight [g]:
Liver weight [g]:
Liver coefficient [%]:
Spleen weight [mg]:
Spleen coefficient [%]:
TG [mm]:
22.98Æ3.12
0.95Æ0.11
4.15Æ0.61
73.54Æ16.54
0.32Æ0.05
0.64Æ0.12
23.87Æ2.14
1.22Æ0.21#
5.10Æ0.41#
78.33Æ17.22
0.32Æ0.04
24.52Æ1.44
2.21Æ0.17**
9.02Æ0.65**
76.67Æ19.66
0.31Æ0.07
20.88Æ3.10
2.37Æ0.35**
11.37Æ0.58**
66.67Æ18.62
0.31Æ0.04
22.81Æ1.45
2.00Æ0.13**
8.78Æ0.54**
73.33Æ16.33
0.32Æ0.06
2.14Æ0.72##
1.41Æ0.26*
0.76Æ0.20**
1.07Æ0.27**
À50.0
À34.0
À64.5
TP [gLÀ1]:
62.89Æ2.12
6.99Æ0.98
43.62Æ5.21
22.33Æ2.25
57.50Æ6.62
63.55Æ2.96
7.36Æ0.80
62.95Æ1.90
6.93Æ0.39
62.45Æ1.44
7.56Æ0.61
63.67Æ2.50
6.39Æ0.66
BUN [mm]:
CREA [mm]:
ALT [ULÀ1]:
AST [ULÀ1]:
42.75Æ3.59
38.00Æ6.96##
78.83Æ7.52##
44.47Æ5.03
36.00Æ7.43##
71.33Æ8.64#
42.85Æ3.53
34.83Æ4.44##
74.50Æ4.85##
48.07Æ6.28
35.17Æ5.53##
75.17Æ6.82##
#
[a] TG: triglycerides, TP: total protein, BUN: blood urea nitrogen, CREA: creatinine, ALT: alanine transaminase, AST: glutamic-oxaloacetic transaminase; p<
0.05, ##p<0.01 relative to normal group; **p<0.01 relative to model group.
ChemMedChem 2010, 5, 1977 – 1982
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1979

MED
pounds
induce
6
and
9
would not
in
Table 6. Docking of compounds in PPARa, PPARg, and PPARd.
hepatotoxicity
humans, although this requires
further investigation. Compared
with the control group, mice in
the model group had significant-
ly increased serum alanine trans-
aminase (ALT; p<0.01) and glu-
tamic-oxaloacetic transaminase
(AST; p<0.01). Fenofibric acid
and compounds 6 and 9 failed
to decrease serum ALT (p>0.05)
and AST (p>0.05) levels relative
to the model group at the same
dose. No changes in spleen
Compd
Affinity [kcalmolÀ1
]
H-bond residues
PPARa
PPARg
PPARd
PPARa
PPARg
PPARd
6
À8.5
À8.4
À9.0
Thr283, Tyr314,
Tyr464, Gly335
Ser289, His323,
Tyr327, His449,
Tyr473
Thr289, Thr292,
Ala342, Asn343,
His449, Tyr473
9
À8.2
À7.7
À8.6
À8.2
À9.1
À8.4
Tyr279, Tyr314,
Tyr464
Ser342, His449,
Tyr473
Thr289
fenofibric acid
Ser280, Tyr314,
Tyr464
Ser289, His323,
Tyr327, His449,
Tyr473
His449, Tyr473
cludes the activation function helix 2 (AF-2 helix), which acti-
vates transcription if the acid head group of the PPAR agonist
forms a hydrogen bond network with the corresponding resi-
dues of the helix. The hydrophobic moiety of PPAR agonists
either interacts with the hydrophobic arm II (tail-up pocket) or
with the partly hydrophilic arm III (entrance region of the
ligand binding pocket, tail-down pocket). Some PPAR agonists
with a branched hydrophobic moiety are able to interact with
both hydrophobic arms.[18]
GW409544 is a full PPARa and PPARg agonist.[19] GW409544
was mainly docked into arms I and II of the binding pocket of
PPARa, while 6 was docked into arms I and III. The conforma-
tion of 9 was almost identical to that of GW409544, as shown
in Figure 1a. Unlike compounds 6 and 9, docking of fenofibric
acid shows the two phenyl rings as mainly docked into the
“benzophenone” pocket. This re-
weight, spleen coefficient, total protein (TP), blood urea nitro-
gen (BUN), or creatinine (CREA) were observed.
The antidiabetic effect of compounds was further evaluated
in KKAy mice, which exhibit obesity, insulin resistance, and
symptoms of type 2 diabetes. Compound 6 decreased blood
glucose levels by 41.5% after daily administration at
150 mgkgÀ1 for 21 days, and compound 9 similarly effected a
34.6% decrease, demonstrating the glucose-lowering efficacy
of both compounds in vivo; fenofibric acid lowered glucose
levels by 18.9%, as shown in Table 5. Compounds 6, 9, and fe-
nofibric acid all significantly decreased triglycerides (TG;
>30%). Moreover, the three compounds caused weight loss to
varying degrees; compound 6 effected 3.18 g weight loss per
mouse on average (p<0.05).
vealed that 6 and 9 may bind to
the receptor in a mode different
Table 5. In vivo efficacy of compounds in KKAy mice.[a]
from that of fenofibric acid. In
Compd
model
Dose [mgkgÀ1
]
GLU [mm]
TG [mm]
Body weight gain [g]
complexes of PPARa with com-
–
17.50Æ4.69
5.08Æ1.47
3.13Æ0.98*
0.4
pound 6 or fenofibric acid, hy-
drogen bonding interactions
were observed between the car-
boxylic acid group and the
amino acids Tyr314 and Tyr464,
which are considered key inter-
actions for activation of the re-
ceptor. Notably, the methoxy
group at the 5-position of com-
pound 9 makes hydrogen bond
interactions with Tyr314 and
Tyr464 and the carboxylic acid
group docked into arm II. The af-
6
9
150
10.24Æ3.60*
À3.18*
À41.5
À38.4
150
75
11.44Æ3.19*
À34.6
13.90Æ0.93
À20.6
2.89Æ0.59**
À43.1
4.63Æ0.94
À8.9
À1.92
À0.43
fenofibric acid
150
14.20Æ3.37
À18.9
2.92Æ0.12**
À42.5
À0.57
[a] **p<0.01, *p<0.05, compared with model group.
Molecular modeling (docking) studies were carried out to
gain more insight on the binding mode of the compounds
with PPARs. Proposed binding poses of compounds 6, 9, and
fenofibric acid are superimposed on the co-crystal structures
of the PPARa–GW409544 complex, the PPARg–rosiglitazone
complex, and the PPARd–GW2433 complex, as illustrated in
Figure 1. The affinities and hydrogen bonding residues are
listed in Table 6.
finities for PPARa obtained computationally are in agreement
with the activation activity in vitro.
Figure 1b shows the docking poses of compounds 6, 9, and
fenofibric acid in comparison with the co-crystallized binding
conformation of rosiglitazone. The acid head groups of 9 and
fenofibric acid, as well as the acid head group at the 3-position
of 6, make several specific hydrogen bonding interactions with
the key amino acid residues Ser289, His449, and Tyr473. Rosi-
glitazone, a PPARg agonist, binds in a U-shaped conformation
(arms I and II) while occupying only 40% of the ligand binding
The Y-shaped ligand binding pocket of PPAR can be divided
into three arms. Arm I consists mainly of polar residues and in-
1980
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1977 – 1982

Figure 1. a) From left to right: compounds 6 (yellow), 9 (green), and fenofibric acid (cyan) docked into the co-crystal structure of PPARa in complex with
GW409544 (purple) (PDB ID: 1K7L); b) From left to right: compounds 6 (yellow), 9 (green), and fenofibric acid (cyan) docked into the co-crystal structure of
PPARg in complex with rosiglitazone (purple) (PDB ID: 2PRG); c) From left to right: compounds 6 (yellow), 9 (green), and fenofibric acid (cyan) docked into
the co-crystal structure of PPARd in complex with GW2433 (purple) (PDB ID: 1GWX). Selected residues within 4 ꢁ of the bound ligand are shown in all panels.
The important residues involved are labeled. H-bond interactions are shown as dashed lines.
site.[14] Compound 6 and fenofibric acid occupy arms I and II of
the binding pocket, like rosiglitazone. Relative to fenofibric
acid, compound 6 occupies more of arm II of the binding
pocket of PPARg due to the long acid head group at the 4’-po-
sition, and compound 9 occupies three arms of the Y-shaped
pocket. Furthermore, both 6 and 9 are stabilized by hydropho-
bic interactions with the binding pocket. These favorable inter-
action profiles for 6 and 9 in the PPARg ligand binding domain
(LBD) may explain their greater activity in the PPARg transacti-
vation assay than fenofibric acid in vitro.
carboxylic acid group and the key amino acid residues His449
and Tyr473 (Figure 1c). Compound 9 forms a hydrogen bond
with Thr289, and no other key polar interactions were ob-
served between 9 and PPARd. GW2433 is a high-affinity ligand
for human PPARd, which fills all three arms of the Y-shaped
pocket of PPARd.[14] Compounds 6 and 9 docked into two arms
of the Y-shaped pocket, whereas fenofibric acid occupied only
arm I. Moreover, due to the long acid head group at the 3-po-
sition of 6, this compound occupies more of arm III of the
binding pocket than does GW2433. Furthermore, 6 and 9 are
stabilized by hydrophobic interactions with the binding
pocket. These may be the reasons why 6 and 9 exhibit more
In complexes of PPARd with compound 6 or fenofibric acid,
hydrogen bonding interactions were observed between the
ChemMedChem 2010, 5, 1977 – 1982
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1981

MED
[3] W. R. Oliver, J. L. Shenk, M. R. Snaith, Proc. Natl. Acad. Sci. USA 2001, 98,
5306–5311.
[4] J. Kasuga, D. Yamasaki, K. Ogura, M. Shimizu, M. Sato, M. Makishima, T.
Doi, Y. Hashimoto, H. Miyachi, Bioorg. Med. Chem. Lett. 2008, 18, 1110 –
1115.
potent activation of PPARd than fenofibric acid in vitro. The af-
finities for PPARd obtained computationally are generally in
agreement with the activation activity in vitro.
In conclusion, the pan-PPAR agonists 6 and 9, which are
based on the resveratrol scaffold, show significant lowering of
triglycerides and are more potent than fenofibric acid in two
acute in vivo hypolipidemic activity tests. Compound 9 low-
ered triglycerides by 64.5% (p<0.01) in hyperlipidemic Balb/c
mice fed a fat emulsion at a daily dose of 80 mgkgÀ1 body
weight for 19 days. Compound 6 decreased blood glucose
levels by 41.5% in KKAy mice after daily administration at a
dose of 150 mgkgÀ1 for 21 days. This activity was closely fol-
lowed by compound 9 (34.6%), demonstrating the good glu-
cose-lowering efficacy of both compounds in vivo. Docking
studies revealed that 6 and 9 dock into the active sites of
PPARa, PPARg, and PPARd. The conformations of 6, 9, and fe-
nofibric acid, and their interactions with amino acid residues in
the PPAR LBD may determine their differences in activity. Fur-
ther work is needed to evaluate the biological profile of this
class of ligands. Moreover, further research is required in order
to determine whether the compounds can produce synergistic
pharmacological effects by combining the pharmacophore
moieties with the natural scaffold. The studies reported herein
provide a good starting point for the development of novel
pan-PPAR agonists for the treatment of metabolic diseases.
[5] a) L. A. Stivala, M. Savio, F. Carafoli, P. Perucca, L. Bianchi, G. Maga, L.
Forti, U. M. Pagnoni, A. Albini, E. Prosperi, V. Vannini, J. Biol. Chem. 2001,
276, 22586–22594; b) Y. Kimura, H. Okuda, S. Arichi, Biochim. Biophys.
Acta Lipids Lipid Metab. 1985, 834, 275–278; c) E. Mannila, A. Talvitie, E.
Kolehmainen, Phytochemistry 1993, 33, 813–816; d) G. B. Mahady, S. L.
Pendland, L. R. Chadwick, Am. J. Gastroenterol. 2003, 98, 1440–1441;
e) L. L. Creasy, M. J. Coffee, J. Am. Soc. Hortic. Sci. 1988, 113, 230–234;
f) M. I. Chung, C. M. Teng, K. L. Cheng, F. N. Ko, C. N. Lin, Planta Med.
1992, 58, 274–276; g) Y. Inamori, M. Kubo, H. Tsujibo, M. Ogawa, Y.
Saito, Y. Miki, S. Takemura, Chem. Pharm. Bull. 1987, 35, 887–890; h) Y.
Schneider, B. Duranton, F. Gosse, R. Schleiffer, N. Seiler, F. Raul, Nutr.
Cancer 2001, 39, 102–107.
[6] a) D. M. Hegsted, L. M. Ausman, J. Nutr. 1988, 118, 1184–1189; b) S.
Renaud, M. de Lorgeril, Lancet 1992, 339, 1523–1526.
[7] J. A. Baur, D. A. Sinclair, Nat. Rev. Drug Discovery 2006, 5, 493–506.
[8] a) E. N. Frankel, A. L. Waterhouse, J. E. Kinsella, Lancet 1993, 341, 1103–
1104; b) Z. Ungvari, Z. Orosz, A. Rivera, N. Labinskyy, Z. Xiangmin, S.
Olson, A. Podlutsky, A. Csiszar, Am. J. Physiol. 2007, 292, H2417–H2424.
[9] Z. Wang, J. Zou, Y. Huang, K. Cao, Y. Xu, J. M. Wu, Chin. Med. J. 2002,
115, 378–380.
[10] H. Arichi, Y. Kimura, H. Okuda, K. Baba, M. Kozawa, S. Arichi, Chem.
Pharm. Bull. 1982, 30, 1766–1770.
[11] A. M. Rimando, R. Nagmani, D. R. Feller, W. J. Yokoyama, J. Agric. Food
Chem. 2005, 53, 3403–3407.
[12] C. S. Mizuno, G. Ma, S. Khan, A. Patny, M. A. Avery, A. M. Rimando,
Bioorg. Med. Chem. 2008, 16, 3800–3808.
[13] G. Z. Morris, R. L. Williams, M. S. Elliott, S. J. Beebe, Prostate 2002, 52,
319–329.
[14] T. Willson, P. J. Brown, D. D. Sternbach, B. R. Henke, J. Med. Chem. 2000,
43, 527–550.
Experimental Section
The details of compound synthesis, characterization data, biologi-
cal evaluation, and computational studies are available in the Sup-
porting Information.
[15] J. K. Reddy, R. Chu, Ann. N. Y. Acad. Sci. 1996, 804, 176–201.
[16] a) T. Hays, I. Rusyn, A. M. Burns, M. J. Kennett, J. M. Ward, F. J. Gonzalez,
J. M. Peters, Carcinogenesis 2005, 26, 219–227; b) J. M. Peters, R. C. Catt-
ley, F. J. Gonzalez, Carcinogenesis 1997, 18, 2029–2033; c) S. S. Lee, T.
Pineau, J. Drago, E. J. Lee, J. W. Owens, D. L. Kroetz, P. M. Fernandez-Sal-
guero, H. Westphal, F. J. Gonzalez, Mol. Cell. Biol. 1995, 15, 3012–3022;
d) J. A. Balfour, D. McTavish, R. C. Heel, Drugs 1990, 40, 260–290.
[17] A. Rubenstrunk, R. Hanf, D. W. Hum, J. C. Fruchart, B. Staels, Biochim.
Biophys. Acta Mol. Cell Biol. Lipids 2007, 1771, 1065–1081.
[18] P. Markt, R. K. Petersen, E. N. Flindt, K. Kristiansen, J. Kirchmair, G. Spitzer,
S. Distinto, D. Schuster, G. Wolber, C. Laggner, T. Langer, J. Med. Chem.
2008, 51, 6303–6317.
Acknowledgements
The authors are grateful to the management of the Bohua
Zhong research group for encouragement, and to the analytical
department for support.
[19] H. E. Xu, M. H. Lambert, V. G. Montana, K. D. Plunket, L. B. Moore, J. L.
Collins, J. A. Oplinger, S. A. Kliewer, R. T. Gampe, D. D. McKee, J. T.
Moore, T. M. Willson, Proc. Natl. Acad. Sci. USA 2001, 98, 13919–13924.
Keywords: metabolic diseases · molecular modeling · pan-
PPAR agonists · receptors · resveratrol
[1] P. H. Chong, B. S. Bachenheimer, Drugs 2000, 60, 55–93.
[2] a) C. Leyvraz, M. Suter, C. Verdumo, J. M. Calmes, A. Paroz, C. Darimont,
R. C. Gaillard, F. P. Pralong, V. Giusti, Diabetes Obes. Metab. 2010, 12,
195–203; b) N. K. Verma, J. Singh, C. S. Dey, Br. J. Pharmacol. 2004, 143,
1006–1013.
Received: August 24, 2010
Revised: September 21, 2010
Published online on October 25, 2010
1982
www.chemmedchem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1977 – 1982