Design of potent PPARα agonists

Article abstract of DOI:10.1016/j.bmcl.2007.03.015

Computational analysis of the ligand binding pocket of the three PPAR receptor subtypes was utilized in the design of potent PPARα agonists. Optimum PPARα potency and selectivity were obtained with substituents having van der Waals volume around 260. Compound 6 had a PPARα potency of 0.002 μM and a selectivity ratio to PPARγ and PPARδ of 410 and 2000, respectively.

Full text of DOI:10.1016/j.bmcl.2007.03.015

Bioorganic & Medicinal Chemistry Letters 17 (2007) 3198–3202
Design of potent PPARa agonists
Per Sauerberg,a,* John P. Mogensen,a Lone Jeppesen,a Paul S. Bury,a Jan Fleckner,a
Grith S. Olsen,a Claus B. Jeppesen,a Erik M. Wulff,a Pavel Pihera,b
Miroslav Havranek,b Zdenek Polivkab and Ingrid Petterssona
a
˚
Novo Nordisk A/S, Novo Nordisk Park, 2760 Maløv, Denmark
bRE&D VUFB, Podebradska 56/186,180 66 Praha 9, Czech Republic
Received 12 February 2007; revised 6 March 2007; accepted 7 March 2007
Available online 12 March 2007
Abstract—Computational analysis of the ligand binding pocket of the three PPAR receptor subtypes was utilized in the design of
potent PPARa agonists. Optimum PPARa potency and selectivity were obtained with substituents having van der Waals volume
around 260. Compound 6 had a PPARa potency of 0.002 lM and a selectivity ratio to PPARc and PPARd of 410 and 2000,
respectively.
ꢀ 2007 Elsevier Ltd. All rights reserved.
The peroxisome proliferator-activated receptors (PPA
Rs) are transcription factors belonging to the nuclear
receptor super family. The marketed insulin sensitizers
(rosiglitazone and pioglitazone) and the lipid lowering
fibrates (e.g., fenofibrate and clofibrate) are PPARc
and PPARa agonists, respectively. There is no drug
available targeting the third PPARd receptor, but sev-
eral reports suggest PPARd to be involved in lipid
metabolism as well as in glucose homeostasis.1,2
Alignment of the amino acids of the LBDs of the
PPAR receptors combined with docking studies into
the X-ray structures of the three receptors identified
five amino acids in the lipophilic pocket: 264, 266,
281, 284, and 348, Figure 1. Of these five amino acids,
position 264 had the largest difference in size, with
PPARa being the smallest. The idea was to keep the
core-structure of our compounds constant and only
change the substituents predicted to interact with posi-
tion 264.
The currently available fibrates were identified as lipid
lowering agents two decades ago using animal models.
The mechanism of the fibrates was not known at time.
The present knowledge shows us that the marketed
fibrates are PPARa agonists with rather low potency,
which might limit the clinical efficacy.
Our structural starting point was the PPARpan-agonist
NNC 61-3058, Figure 2, recently described in Sauerberg
et al.3 In vitro transactivation assays using the LBDs of
hPPARa, c, and d were used to evaluate the designed
and synthesized ligands.4 NNC 61-3058 had a PPARa
potency of 0.4 lM (EC50) compared to the EC50 of fen-
ofibric acid and clofibric acid of 32 and 3 lM, respec-
tively, Table 1.
The present investigation was aimed at identifying
potent and selective PPARa agonists to be used as
research tools and potential lipid lowering drugs.
The designed compounds were synthesized as outlined
in Scheme 1, starting from commercialy available substi-
tuted benzaldehydes, or substituted benzaldehydes
synthesized according to Scheme 2. Using Horner–Em-
mons reaction conditions the aldehydes were converted
to the corresponding allylic esters, which were reduced
to the alcohols using DIBAL. Mitsunobu coupling to
the (S)-2-ethoxy-3-(4-hydroxy-phenyl)-propionic acid
ethyl ester5 followed by hydrolysis gave the desired
compounds 1–14.12
The aim was further to investigate if it was possible to
utilize the difference in a single amino acid in the lipo-
philic pocket of the ligand binding domain (LBD) of
the three PPAR receptors to design selective PPARa
agonists.
Keywords: PPARa; Agonist; Docking.
*
Corresponding author. Tel.: +45 44434858; fax: +45 44434547;
e-mail: psa@novonordisk.com
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.03.015

P. Sauerberg et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3198–3202
3199
At the PPARd receptor only smaller substituents (3–8)
increased the potency compared to the un-substituted
analogue (1), Table 1. The most potent compound (3)
with CF3 as substituents was approximately 10 times
more potent than 1. Larger side chains led, as expected,
to PPARd inactive compounds (10–14), probably due to
repulsive steric interactions with the tryptophan.
O
O
O
O
N
NNC 61-4424
The meta-substituents also improved the PPARc po-
tency significantly. A potency optimum was observed
with isopropoxy (7), which was 100-fold more potent
than the un-substituted starting point (1). Again, larger
substituents gave less potent PPARc agonists, although
even phenylethyl (13) was more potent than 1.
PPARγIle281
PPARα Ile
PPARδVal
PPARγGly284
PPARαCys
PPARδArg
PPARa with the smallest amino acid (Ala) in position
264 tolerated substituents the best. The most potent
compound was obtained with OCH2CF3 substituents
(6), which was 800 times more potent than 1. Smaller
(5) and larger (7) substituents were approximately
2–10 times less potent than 6, suggesting favorable inter-
actions. Compound 6 also displayed the highest PPARa
selectivity with a potency ratio to PPARc and PPARd of
410 and 2000, respectively.
PPARγHis266
PPARα Leu
PPARδGln
PPARγMet348
PPARαIle
PPARδVal
PPARγPhe264
PPARα Ala
PPARδ Trp
To further understand the SAR, various physical–chem-
ical properties (MW, van der Waals area, van der Waals
volume, and logP) of the substituents were mapped
against the PPARa potency. The van der Waals volumes
turned out to explain the relationship the best, Figure 3.
This reflects the difference in size of the amino acid in
the part of the binding pocket in which the substituted
phenyl ring is predicted to bind. The physical–chemical
properties for the di-meta substituted phenyl ring were
calculated with MOE.8 The PPARa potency was im-
proved when the volume of the substituted benzene
was increased up to 266 (6) which corresponded to
two OCH2CF3 substituents. As the volume of the sub-
stituents increased, the potency decreased.
Figure 1. The crystal structure of the ligand binding domain of the
PPARc receptor crystallized with NNC 61-4424 (yellow) and with
NNC 61-3058 (magenta) docked into the active site. The amino acids
˚
within 5 A from the biphenyl ring in NNC 61-3058 are listed together
with the corresponding amino acids in the PPARa and PPARd
receptors. His323, His449, and Tyr473 are also displayed.
NNC 61-3058 was docked into the binding domain of
PPARc crystallized with NNC 61-4424 (pdb code
1KNU).4 In order to design PPARa selective agonists,
amino acids with difference in size or charge were iden-
tified. Nearly all of the amino acids identified were neu-
tral, Figure 1, so focus was therefore placed on the size
of the amino acids. The amino acid in position 264
showed a difference in size with PPARa being the small-
est (Ala), then PPARc (Phe) and PPARd the largest
(Trp). Docking studies with NNC 61-30586 suggested
that selective PPARa ligands could be obtained by
removing the p-phenyl in NNC 61-3058 and introducing
substituents (R) in meta-positions of compound 1 lead-
ing to a series of analogues, Figure 2 and Table 1.7 Com-
pound 1 was 2–4 times less potent on the PPAR
receptors than the lead compound NNC 61-3058, but
the PPARa selectivity was retained (see potency ratios
Table 1).
In conclusion, computational analysis of the ligand
binding pocket of the three PPAR receptor subtypes
was utilized in the design of potent PPARa agonists.
Analogues with substituents of different van der Waals
volume gave PPARa agonists with varying potency
and selectivity. Optimum PPARa potency and selectivity
were obtained with substituents having van der Waals
volume around 260.
Compound 6 had a PPARa potency of 0.002 lM and a
selectivity ratio to PPARc and PPARd of 410 and 2000,
R1=
COOH
O
O
R1
R1
O
R1
O
R
R
2-14
1
NNC 61-3058
Figure 2. Graphic illustration of the design of selective PPARa agonists.

3200
P. Sauerberg et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3198–3202
Table 1. In vitro transactivation data for compounds 1–14 and standard compounds
R
O
COOH
R
O
Compound
Substituent R
hPPARa
EC50, lMa
hPPARc
hPPARd
Potency ratio
hPPARc
hPPARa
% maxb
EC50, lM
% maxc
EC50, lM
9.6
18
20
1.4
13
7.9
% maxd
hPPARd
hPPARa
NNC 61-3058
1
0.5
1.6
158
169
121
195
186
197
149
207
30
2.0
8.0
4.1
0.4
1.2
0.2
0.8
0.08
0.2
0.3
0.19
3.1
3.0
5.1
1.0
29
118
96
235
177
139
74
4
5
19
11
H
Br
2
0.3
58
70
14
21
2
67
70
3
CF3
OMe
OEt
0.02
0.06
0.02
0.002
0.004
0.02
0.02
0.05
0.5
4
100
79
171
124
171
116
117
71
217
395
2000
1500
200
750
60
5
9
6
OCH2CF3
OiPr
Ph
81
90
4.0
6.0
4.0
410
20
12e
14
4
7
8
80
9
10
OcPen
OBz
135
121
108
66
107
104
84
15
3
—
—
—
—
—
—
30
18
11
12
C„CPh
CH=CHPh
CH2CH2Ph
Ph-4-tBu
6
—
—
0.09
0.03
91
90
19
14
33
182
0.07e
2e
13
14
118
110
100
43
—
—
14
13
4.1
44
22
0
WY14.643
Rosiglitazone
6
—
—
0.2
100
7
0.05
aCompounds were tested in five concentrations ranging from 0.01 to 30 lM in at least two independent experiments. Data represent mean, with SEM
15%. EC50’s were not calculated for compounds producing transactivation lower than 25% at 30 lM. bFold activation relative to maximum
activation obtained with WY14643 (approx. 20-fold corresponded to 100%), with crosiglitazone (approx. 120-fold corresponded to 100%), and with
dcarbacyclin (approx. 250-fold corresponded to 100%). eThese potency ratios may not be comparable to the rest as partial efficacy was obtained on
either PPARa or PPARc.
R
R
(EtO)2P(O)CH2COOEt
R
DIBAL
H
OH
R
NaH
THF
0oC
Toluene
-70oC
R
R
COOEt
R
O
COOEt
O
NaOH
EtOH
O
HO
COOH
O
R
Benzene
0oC
Bu3P
ADDP
R: H, CF3, OMe and OBz
Scheme 1. Synthesis of target compounds from commercially available benzaldehydes.
R2
OH
R1
Br
palladium
cross coupling
O
R1Br
O
H
O
K2CO3
R2
O
Br
HO
O
O
H
H
R1
H
R2: Ph, tBuPh. From Boronic acids (Suzuki).
CHCHPh, CCPh. (Sonogashia)
R1: iPr, cPen, Et, CH2CF3.
Scheme 2. Synthesis of substituted benzaldehydes.

P. Sauerberg et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3198–3202
3201
ꢀ1 was assigned and the molecule was minimized with the
OPLS_2005 force field. NNC 61-3058 was docked with
Glide version 4.0 using the SP mode and the van der
Waals radii of the ligand atoms were scaled by 0.8.
7. Simple para-substituted analogues, for example, allyloxy
and benzyloxy, were PPARpan-agonists (EC50 (% max) a:
0.6 lM (180%), c: 1.6 lM (117%), d: 14 lM (186%) and a:
8.4 lM (231%), c: 1.3 lM (116%), d: 17 lM (377%),
respectively).
8. MOE ver. 2005.06. Chemical Computing Group Inc.;
www.chemcomp.com.
9. Two analogues, 5 and 7, were tested for pharmacokinetic
properties in rats. Both compounds exhibited excellent
behavior after oral dosing (2.5 mg/kg): Fpo = 86% and
95%, T1/2po = 130 min. and 145 min., Cmax = 6155 ng/ml
and 1845 ng/ml, respectively.
10. Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren, T.
A.; Klicic, J. J.; Mainz, D. T.; Repasky, M. P.; Knoll, E.
H.; Shaw, D. E.; Shelley, M.; Perry, J. K.; Francis, P.;
Shenkin, P. S. J. Med. Chem. 2004, 47, 1739.
11. Halgren, T. A.; Murphy, R. B.; Friesner, R. A.; Beard, H.
S.; Frye, L. L.; Pollard, W. T.; Banks, J. L. J. Med. Chem.
2004, 47, 1750.
Figure 3. The relationship between in vitro PPARa potency and the
van der Waals volume of the di-meta substituted phenyl ring.
respectively. The selective PPARa agonist 6 could be a
useful tool to dissect the pharmacology of PPARa med-
iated effects.9
12. Experimental procedure for synthesis of (E)-(S)-3-(4-{3-
[3,5-bis-(2,2,2-trifluoro-ethoxy)-phenyl]-allyloxy}-phenyl)-
2-ethoxy-propionic acid (6): (a) to a solution of 3,5-
dihydroxybenzaldehyde (2.0 g, 14.5 mmol) in DMF
(35 mL) were added potassium carbonate (11.0 g,
80.0 mmol) and 1,1,1-trifluoro-2-iodoethane (33.3 g,
160 mmol). The reaction mixture was heated in a sealed
reactor at 50 ꢁC for 7 days. The mixture was filtered and
the filtrate was washed with ethyl acetate. The filtrate was
washed with water and the organic phase isolated. The
aqueous phase was extracted once more with ethyl acetate.
The combined organic phases were dried (MgSO4),
filtered, and concentrated in vacuo. The residue was
purified by flash chromatography eluting with toluene to
give 906 mg (18%) of 3,5-bis-(2,2,2-trifluoro-ethoxy)-benz-
aldehyde. 1H NMR (CDCl3, 300 MHz) d: 4.43 (q, 4H),
6.85 (t, 1H), 7.15 (d, 2H), 9.95 (s, 1H). (b) Sodium hydride
(60% in oil, 750 mg, 18 mmol) was added at 0 ꢁC to a
solution of triethyl phosphonoacetate (2.67 g, 12 mmol) in
dry THF (60 mL). After stirring at 0 ꢁC for 15 min, a
solution of 3,5-bis-(2,2,2-trifluoro-ethoxy)-benzaldehyde
(1.85 g, 6 mmol) in dry THF (10 mL) was added, and
the mixture was stirred for additional 45 min. Water
(100 mL) was added and the aqueous phase was extracted
with ethyl acetate (100 mL). The combined organic phases
were washed with water (3· 100 mL), dried (MgSO4),
filtered, and concentrated in vacuo to give 2.2 g of crude
(E)-3-[3,5-bis-(2,2,2-trifluoro-ethoxy)-phenyl]-acrylic acid
ethyl ester. (c) A 1 M solution of DIBAL-H in toluene
(30 mL, 30 mmol) was added dropwise at ꢀ70 ꢁC to a
stirred solution of crude (E)-3-[3,5-bis-(2,2,2-trifluoro-
ethoxy)-phenyl]-acrylic acid ethyl ester (2.2 g, 6 mmol) in
dry THF (100 mL) and the mixture was stirred for 30 min.
Methanol (20 mL) was added, and the reaction mixture
was poured into 0.1 N HCl (600 mL). The mixture was
extracted with ethyl acetate (2· 200 mL), dried (Na2SO4)
and evaporated to give a crude oil. The crude product was
crystallized from ethyl acetate and heptanes to give (E)-3-
[3,5-bis-(2,2,2-trifluoro-ethoxy)-phenyl]-prop-2-en-1-ol as
colorless crystals: 510 mg. (d) Under an atmosphere of
nitrogen, 1,10-(azodicarbonyl) dipiperidine (380 mg,
1.5 mmol) and tributylphosphine (303 mg, 1.6 mmol) were
added to a solution of (E)-3-[3,5-bis-(2,2,2-trifluoro-eth-
oxy)-phenyl]-prop-2-en-1-ol (250 mg, 0.75 mmol) and (S)-
2-ethoxy-3-(4-hydroxy-phenyl)-propionic acid ethyl ester
(180 mg, 0.75 mmol) in dry THF (20 mL) at 0 ꢁC. The
Acknowledgments
The technical assistance of R. Burgdorf, P.S. Klifforth,
B. Metzler, A. Ravn, and O. Larsson is greatly
appreciated.
References and notes
1. Oliver, W., Jr.; Shenk, J. L.; Snaith, M. R.; Russell, C. S.;
Plunket, K. D.; Bodkin, N. L.; Lewis, M. C.; Winegar, D.
A.; Sznaidman, M. L.; Lambert, M. H.; Xu, H. E.;
Sternbach, D. D.; Kliewer, S. A.; Hansen, B. C.; Willson,
T. M. Proc. Natl. Acad. Sci. 2001, 98, 5306.
2. Tanaka, T.; Yamamoto, J.; Iwasaki, S.; Asaba, H.;
Hamura, H.; Ikeda, Y.; Watanabe, M.; Magoori, K.;
Ioka, R. X.; Tachibana, K.; Watanabe, Y.; Uchiyama, Y.;
Sumi, K.; Iguchi, H.; Ito, S.; Dio, T.; Hamakubo, T.;
Naito, M.; Auwerx, J.; Yanagisawa, M.; Kodama, T.;
Sakai, J. Proc. Natl. Acad. Sci. 2003, 100, 15924.
3. Sauerberg, P.; Bury, P. S.; Mogensen, J. P.; Deussen, H.-
J.; Pettersson, I.; Fleckner, J.; Nehlin, J.; Frederiksen, K.
S.; Albrektsen, T.; Din, N.; Svensson, L. S.; Ynddal, L.;
Wulff, E. M.; Jeppesen, L. J. Med. Chem. 2003, 46, 4883.
4. Sauerberg, P.; Pettersson, I.; Jeppesen, L.; Bury, P. S.;
Mogensen, J. P.; Wassermann, K.; Brand, C. L.; Sturis, J.;
Wo¨ldike, H. F.; Fleckner, J.; Andersen, A.-S. T.; Mor-
tensen, S. B.; Svensson, L. A.; Rasmussen, H. B.;
Lehmann, S. V.; Polivka, Z.; Sindelar, K.; Panajotova,
V.; Ynddal, L.; Wulff, E. M. J. Med. Chem. 2002, 45, 789.
5. Ebdrup, S.; Pettersson, I.; Rasmussen, H. B.; Deussen, H.-
J.; Jensen, A. F.; Mortensen, S. B.; Fleckner, J.; Pridal, L.;
Nygaard, L.; Sauerberg, P. J. Med. Chem. 2003, 46, 1306.
6. Glide 4.0 docking. The structure manipulations were
performed in Maestro version 7.5 (Maestro 7.0, Schro¨-
dinger, LLC, New York, NY, 1999–2005). Glide calcula-
tions were performed with Impact version 4.0.10,11 The
crystal structure of the PPARc receptor crystallized with
NNC 61-4424 (1KNU) was used for docking.4 NNC 61-
3058 was built in Maestro version 7.5, a formal charge of

3202
P. Sauerberg et al. / Bioorg. Med. Chem. Lett. 17 (2007) 3198–3202
reaction mixture was stirred for 30 min and then poured
dissolved in ethanol (10 mL) and 1 N sodium hydroxide
(1.5 mL, 4.4 mmol) was added. The mixture was heated
slightly to obtain a clear solution and then stirred at room
temperature for 1.5 h. The ethanol was evaporated in
vacuo and the mixture was acidified to pH 1 with 1 N
hydrochloric acid. The product was isolated by extraction
with ethyl acetate (2· 40 mL). The combined organic
phases were dried (MgSO4), filtered, and evaporated to
give 130 mg (91%) of the title compound as an oil. 1H
NMR (CDCl3, 300 MHz) d: 1.20 (t, 3H), 2.97 (dd, 1H),
3.10 (dd, 1H), 3.41–3.53 (m, 1H), 3.55–3.68 (m, 1H), 4.05
(dd, 1H), 4.35 (q, 4H), 4.67 (d, 2H), 6.35–6.48 (m, 2H),
6.60–6.70 (m, 3H), 6.87 (d, 2H), 7.15 (d, 2H).
into water. The mixture was extracted with ethyl acetate
(50 mL), dried, and evaporated in vacuo. The product was
purified by column chromatography, eluting with heptane/
ethyl acetate (3:1), to give 300 mg (73%) of (E)-(S)-3-(4-{3-
[3,5-bis-(2,2,2-trifluoro-ethoxy)-phenyl]-allyloxy}-phenyl)-
2-ethoxy-propionic acid ethyl ester.1H NMR (CDCl3,
300 MHz) d: 1.15 (t, 3H), 1.22 (t, 3H), 2.95 (d, 2H), 3.30–
3.40 (m,1H), 3.55–3.67 (m, 1H), 3.97 (t, 1H), 4.15 (q, 2H),
4.33 (q, 4H), 4.65 (d, 2H), 6.32–6.48 (m, 2H), 6.55–6.70
(m, 3H), 6.85 (d, 2H), 7.15 (d, 2H). (e) (E)-(S)-3-(4-{3-[3,5-
bis-(2,2,2-trifluoro-ethoxy)-phenyl]-allyloxy}-phenyl)-2-
ethoxy-propionic acid ethyl ester (300 mg, 0.54 mmol) was