Glycine amides as PPARα agonists

Article abstract of DOI:10.1016/j.bmcl.2010.04.019

The design, synthesis and pharmacological properties of a novel class of PPARα agonists is described. Compound 2 is a novel, potent and specific glycine amide with oral bioavailability in rodents. The compound is active in vivo and alters plasma lipids in

Full text of DOI:10.1016/j.bmcl.2010.04.019

Bioorganic & Medicinal Chemistry Letters 20 (2010) 3376–3379
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Glycine amides as PPAR
a
agonists
a
a
a
b
Klaus Urbahns a, , Michael Woltering , Susanne Nikolic , Josef Pernerstorfer , Hilmar Bischoff ,
*
Elke Dittrich-Wengenroth b, Klemens Lustig c
a Department of Medicinal Chemistry, Bayer Healthcare, Bayer-Schering Pharma, D-42096 Wuppertal, Germany
b Department of Cardiovascular Research, Bayer Healthcare, Bayer-Schering Pharma, D-42096 Wuppertal, Germany
c Department of Pharmacokinetics, Bayer Healthcare, Bayer-Schering Pharma, D-42096 Wuppertal, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 18 February 2010
Revised 6 April 2010
Accepted 7 April 2010
Available online 13 April 2010
The design, synthesis and pharmacological properties of a novel class of PPAR
a agonists is described.
Compound 2 is a novel, potent and specific glycine amide with oral bioavailability in rodents. The com-
pound is active in vivo and alters plasma lipids in hAPO-A1 transgenic mice after oral administration.
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
PPAR
Coronary heart disease
Human APO-A1 mice
Structure–activity relationship
Synthesis
In vivo activity
Coronary heart disease (CHD) is the leading cause of death in
industrialized countries and coronary events are among the top
ranking reasons for working disability in the western world. Lipid
disorders are well recognized as one of the leading risk factors for
the development of atherosclerosis, cardiovascular morbidity and
mortality. Peroxisome proliferator-activated receptors (PPAR’s)
are a family of nuclear hormone receptors that act as ligand-acti-
vated transcription factors and offer the potential to have benefi-
cial effects on cardiovascular disease.
NH by one position to furnish glycine amides A as novel PPARa
agonists. This structural modification leads to zwitterions and of-
fers synthetically convenient options for structural modifications
of R1 and R2 (Fig. 1).
Due to their zwitterionic nature we were expecting these com-
pounds to display favorable physicochemical properties like higher
polarity, lower binding affinity to human serum albumin, better
solubility and possibly an improved therapeutic window. We also
wanted to evaluate whether such modifications alter the struc-
ture–activity relationship (SAR) of the side chains.
PPARa is one member of this family regulating arteriosclerosis
relevant genes by binding to upstream promoter elements. PPAR
a
is highly expressed in liver and its activation leads to plasma tri-
glyceride decrease and high density lipoprotein (HDL) increase in
man. Fibrates, which are on the market since many years, have re-
CO2H
S
F
O
N
N
H
cently been shown to act via PPARa agonism, however, only in the
F
micromolar range and with low specificity. Thus, the idea of iden-
tifying potent and specific PPAR
a agonists has inspired many
1
groups, resulting more recently in new clinical compounds.1
In an earlier Letter, we have described a glycine amide lead
template that can be utilised to achieve different receptor specific-
R2
N
CO2H
O
ities against the known PPAR subtypes,
a
, b (d) and
c.
2 This lead is
O
derived from GW 9578 (1), a known PPAR
a
agonist which contains
R1
N
H
a urea moiety.3 In this study we explore shifting 1’s proximal urea
A
* Corresponding author at Present address: AstraZeneca, Bakewell Road, Lough-
borough, LE11 5RH, UK. Tel.: +44 15 09 64 49 78; fax: +44 15 09 64 55 69.
E-mail address: Klaus.Urbahns@astrazeneca.com (K. Urbahns).
Figure 1. Design of glycine amides (A). Shifting 1’s urea nitrogen by one position
leads to more hydrophilic zwitterions.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.04.019

K. Urbahns et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3376–3379
3377
O
O
OH
a)
O
Br
O
O
+
O
O
ii
iii
c), d), e), f)
i
O
b)
O
O
2
O
HN
iv
Figure 2. Synthesis of 2. Reagents and conditions: (a) Cs2CO3, DMF, 90 °C, 12 h, 30%; (b) 2-furfurylamine, 1,2-dichloroethane, NaB(OAc)3H, 5 h, rt, 72%; (c) THF/Et3N (20:1),
0.05 equiv (tBu)4N+IÀ, 1.5 equiv BrCH2COOEt, 3 h 60 °C, 100%; (d) EtOH, 3 equiv NaOH, 1 h 80 °C, 74%; (e) 2,4-dimethyl aniline (1.5 equiv), hydroxy-1H-benzotriazole
(1.3 equiv), EDC (1.3 equiv), 4-methyl morpholine (3 equiv), DMAP (cat.), 12 h, rt, 78%; (f) CH2Cl2/TFA (1:1), 2 h, rt, 82%.
The compounds were synthesized according to Figure 2 via
ether synthesis of iii using standard conditions. Reductive amina-
tion with furfuryl amine yielded the central building block iv. Sim-
ilar reductive aminations were carried out with a variety of
primary amines.
The SAR around the R2 position displayed clear steric and elec-
tronic requirements of the corresponding binding pocket (Table 1).
The furanyl methyl moiety in 2 was well tolerated, yielding a po-
tent and full PPAR
as compared to marketed fibrates like Gemfibrozil (100
brate (55 M), Bezafibrate (30 M) or Fenofibrate (25 M). Replac-
a
agonist with up to 1000-fold enhanced potency
lM), Clofi-
The resulting secondary amines were then coupled with
a
-bro-
l
l
l
mo ethyl acetate under phase transfer conditions. Selective sapon-
ification of the ethyl ester yielded the corresponding acetic acid,
which was then coupled with a set of primary amines. For exam-
ple, the use of 2,4-dimethyl aniline furnished 2 after deprotection
of the tert-Bu ester.4,5
ing the furfuryl moiety with an electron rich benzyl system (3, 5) or
a thiophene methyl system (4) resulted in a significant drop in
activity indicating rather tight SAR in this region. We also intro-
duced aliphatic side chains in order to probe whether substituents
corresponding to urea 1 would improve receptor affinity. Interest-
ingly, the introduction of such side chains resulted in a significant
loss of activity, indicating different SAR for the glycine amide series
Table 1
Structure–activity relationship of glycine amide PPAR
a agonists. Variation of the
tertiary amine residue R2. Ki values are medians of three dose–response curves.
Table 2
O
Structure–activity relationship of glycine amide PPAR
a agonists. Variation of the
R2
O
amide part R1. Ki values are medians of three dose–response curves.
O
OH
O
N
O
N
H
O
O
OH
R1
N
Compound
R2
h-PPARa EC50 (nM)
N
H
O
Compound
R2
h-PPARa EC50 (nM)
2
20
9
2000
3
4
1000
700
10
11
12
600
O
O
S
300
3000
MeO
MeO
F
F
5
900
13
1500
6
7
8
1000
14
15
16
>25,000
13,000
7000
N
10,000
>25,000
Cl

3378
K. Urbahns et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3376–3379
Table 3
91%
***
15
Direct comparison of selected in vitro parameters of 1 and 2
74%
***
58%
***
TC
[mmol/l]
2
1
29%
h-PPAR
h-PPPAR
h-PPARd (nM)
m-PPAR
m-PPAR
m-PPARd (nM)
Membrane affinity
HSA (lM)
a
c
(nM)
(nM)
20
100
800
6000
20
300
2000
480
10
1000
2000
10
2500
2000
13,500
1,3
5
a
c
(nM)
(nM)
Dose
[mg/kg]
Control
0.6
2
6
Solubility (mg/l)
Fmax (rat microsomes)
500
57%
100
73%
Figure 5. Investigation of 2 in the human APO-A1 transgenic mouse. The left hand
column indicates total cholesterol (TC) levels at day 1 (control), the right hand
column indicates total cholesterol levels after 7 days of oral treatment.
7000
5000
3000
2000
As the most potent compound in this series, compound 2 has been
selected for further characterization and direct comparison to 1
invitro(Table 3). In mice and humans, 2 displaysimprovedselectivity
1000
1000
700
500
c [µg/l]
to PPARc
and d subtypes over 1.5 Asexpected, the zwitterion 2 ismore
300
200
polar than urea 1, as demonstrated by its lower affinity to mem-
branes,6 human serum albumin (HSA)7 and its enhanced solubility.
Given the in vitro rat microsomal clearance data,8 we decided to
investigate 2’s pharmacokinetics in male NMRI mice and were
pleased to see bioavailability acceptable for oral dosing (F = 20%,
10 mg/kg, Fig. 3).9
We therefore investigated 2’s ability to reduce triglyceride lev-
els and elevate serum total cholesterol levels in human APO-A1
transgenic mice. These mice express the human APO-A1 gene un-
der the control of the natural APO-A1 promotor and have been
100
100
70
50
30
20
10
107
5
3
2
1
1
0.7
0.5
0.3
0
1
2
3
4
0
1
2
3
4
55
66
time [h]
t [h]
widely used in the evaluation of PPAR
wild-type mice, these transgenic species respond to PPAR
a
agonists. In contrast to
agonist
a
Figure 3. Pharmacokinetic profile of 2. Dose-normalised (1 mg/kg) plasma con-
centrations in male NMRI mice after oral (10 mg/kg, n = 3, s) and iv administration
(3 mg/kg, n = 3, j) in 20% PEG400.
treatment with increased HDL levels. As most of the serum choles-
terol exists in the form of HDL, we used serum total cholesterol as a
surrogate for HDL.10
Gratifyingly, 2 altered lipid parameters in this species in a dose-
dependent manner after 7 d once daily oral treatment (Figs. 4 and
5). It was also evident that 2 reduced triglyceride levels and en-
hanced serum total cholesterol levels. This is in line with observa-
1%
-3%
-20%
-63%
3.00
3
TG
[mmol/l]
2.50
tions made with other PPARa agonists in this species and confirms
2’s proposed mechanism of action.11
2.00
2
In summary, we have identified novel glycine amides as potent
1.50
and selective PPARa agonists, starting from urea 1. Modifying the
1.00
urea motif to a glycine amide markedly changed structure–activity
trends. An optimized derivative 212 is orally bioavailable and alters
triglyceride and cholesterol levels in rodents after oral
administration.
1
***
0.50
0.00
0
Dose
[mg/kg]
Control
0.6
2
6
References and notes
Figure 4. Investigation of 2 in the human APO-A1 transgenic mouse. The left hand
column of each pair indicates triglyceride (TG) levels at day 1 (control), the right
hand column indicates triglyceride levels after 7 days of oral treatment.
1. Recent reviews: Hansen, M. K.; Connolly, T. M. Curr. Opin. Invest. Drugs 2008, 9,
247; Semple, R. K.; Chatterjee, V. K. K.; O’Rahilly, S. J. Clin. Invest. 2006, 116, 581;
Chang, F.; Jaber, L. A.; Berlie, H. D.; O’Connell, M. B. Ann. Pharmacother. 2007, 41,
973; Liu, Y.; Reifel Miller, A. Drug Discovery Today 2005, 2, 165.
2. Weigand, S.; Bischoff, H.; Dittrich-Wengenroth, E.; Heckroth, H.; Lang, D.;
Vaupel, A.; Woltering, M. Bioorg. Med. Chem. Lett. 2005, 4619.
and may be suggesting an overall different orientation of the mol-
ecule in the binding site (6, 7 and 8).
3. Synthesis and characterisation of GW-9578: 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.; Kliewer, S. A.; Willson, T. M. J. Med.
Chem. 1999, 42, 3785; Improved synthesis: Ham, J.; Cho, S. J.; Ko, J.; Chin, J.;
Kang, H. J. Org. Chem. 2006, 71, 5781.
Keeping the furan substituent constant, we investigated a series
of amide substitutions in the R1 position (Table 2). Removal of 2’s
para methyl group resulted in a 100-fold loss of activity (9). Simi-
larly, shifting it into the meta-position (10) resulted a 30-fold loss,
highlighting the steric requirements of the binding pocket. Inter-
estingly, the 2,4-difluoro substitution pattern that confers potency
and seems essential in urea 1 doesn’t seem to have a beneficial ef-
fect in the glycine amide series (13). Changing the electronic nat-
ure of 13’s substituents renders a fivefold improvement (11).
Neither the heterocyclic pyridine 14 nor the benzylamides 15, 16
seemed to be effective.
4. For details of the syntheses and the assay procedures see: Urbahns, K.;
Woltering, M.; Nikolic, S.; Pernerstorfer, J.; Hinzen, B.; Dittrich-Wengenroth, E.;
Bischoff, H.; Hirth-Dietrich, C.; Lustig, K. (Bayer Health Care AG) WO 02/
028821, 2002; Chem. Abstr. 2002, 136, 294653.
5. The primary assay was performed as described.4 The different PPAR subtypes
are fusion proteins containing the ligand binding domain for PPAR
468), PPAR (aa 203–506) and PPARd (aa 139–442), respectively, fused to the
GAL4 DNA-binding domain (aa 1–147).
a (aa 167–
c
6. Membrane affinities were determined as described. Briefly, the reduction of
compound concentration after incubation with liposomal egg yolk lecithin
followed by ultracentrifugation was measured by HPLC: Loidl-Stahlhofen, A.;

K. Urbahns et al. / Bioorg. Med. Chem. Lett. 20 (2010) 3376–3379
3379
Hartmann, T.; Schottner, M.; Rohring, C.; Brodowsky, H.; Schmitt, J.; Keldenich,
J. Pharm. Res. 2001, 18, 1782.
Bean, J.; Montrose, C.; Mantlo, N.; Wagle, A. Mol. Pharmacol. 2005, 68, 763;
Mukherjee, R.; Locke, K. T.; Miao, B.; Meyers, D.; Monshizadegan, H.; Zhang, R.;
Search, D.; Grimm, D.; Flynn, M.; O’Malley, K. M.; Zhang, L.; Li, J.; Shi, Y.;
Kennedy, L. J.; Blanar, M.; Cheng, P. T.; Tino, J.; Srivastava, R. A. J. Pharmacol.
Exper. Ther. 2008, 327, 716; Van der Hoogt, C. C.; de Haan, W.; Westerterp, M.;
Hockstra, M.; Dallinga-Thie, G. M.; Romijn, J. A.; Princen, H. M. G.; Jukema, J.
W.; Havekes, L. M.; Rensen, P. C. N. J. Lipid Res. 2007, 48, 1763.
7. The binding constant to human serum albumin (HSA) was determined by
equilibrium dialysis: Loidl-Stahlhofen, A.; Schmitt, J.; Noller, J.; Hartmann, T.;
Brodowsky, H.; Schmitt, W.; Keldenich, J. Adv. Mater. 2001, 13, 1829; Sebille, B.;
Zini, R.; Madjar, C. V.; Thuaud, N.; Tillement, J. P. J. Chromatogr. 1990, 51, 531.
8. The prediction of hepatic clearance was performed according to Houston, J. B.;
Carlile, D. J.; Drug Metab. Rev. 1997, 29, 891. Briefly, compounds were incubated
11. Our compound was investigated in vivo according to Ref. 4. Briefly, plasma
samples were taken before and after oral u.i.d. 7 d treatment with 2 (vehicle:
Solutol HS15 + ethanol + brine (0.9%) = 1 + 1 + 8, 10 ml/kg) of male transgenic
hAPO-A1 C57BL/6 mice (n = 10/dose, Jackson Lab., Bar Harbor, ME).
Triglycerides and cholesterol levels were determined photometrically. In an
independent experiment, we confirmed that LDL plasma levels in these mice
were very low (<5%, HPLC) as expected, suggesting that the measurement of
total cholesterol can be used as a surrogate for HDL in this species.
(1 lM, 37 °C) with 0.2 mg/ml rat microsomal protein. Samples were taken at
seven time points within 30 min to calculate half-lives and extrapolate
clearance.
9. Pharmacokinetics of 2: A 10 mg/kg dose of 2 (suspension in 20% PEG 400, 5 ml/
kg) was given to male NMRI mice and plasma samples were analyzed at 7
different time points up to 6 h. cmax: 0.32 mg/ml, AUC: 0.37 mg h/l, tmax: 0.5 h.
In a second experiment 3 mg/kg were administered intravenously in the same
vehicle as a solution and plasma levels determined at nine different time
points, up to 6 h: Clearance: 5.35 l/h kg, t1/2: 2 h.
12. Characteristic data of 2: 1H NMR (200 MHz, CDCl3): d = 1.57 (s, 6H), 2.24 (s,
3H), 2.27 (s, 3H), 3.31 (s, 2H), 3.67 (s, 2H), 3.75 (s, 2H), 6.22–6.36 (m, 2H), 6.88
(m, 2H), 6.93–7.03 (m, 2H), 7.23 (d, 2H), 7.34–7.40 (m, 1H), 7.78 (m, 1H), 8.00
(s, br, 1H), 9,09 (s, 1H). MS (ESI): 451 [M+H]+, 901 [2 M+H]+.
10. PPAR agonists have been characterized in the human APO-A1 transgenic
mouse before: Singh, J. P.; Kaufmann, R.; Bensch, W.; Wang, G.; McClelland, P.;