Design and synthesis of novel and potent amide linked PPARγ/δ dual agonists

Article abstract of DOI:10.1016/j.bmcl.2007.10.047

A series of potent amide linked PPARγ/δ dual agonists (1a) has been discovered through rational design. In the ZDF rat model of type 2 diabetes, compound (R)-3-[4-(3-{1-[(5-chloro-1,3-dimethyl-1H-indole-2-carbonyl)-amino]-ethyl}-5-fluoro-phenoxy)-2-ethyl-

Full text of DOI:10.1016/j.bmcl.2007.10.047

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 17 (2007) 6744–6749
Design and synthesis of novel and potent amide
linked PPARc/d dual agonists
Qing Shi,* Emily J. Canada, Yanping Xu, Alan M. Warshawsky, Garret J. Etgen,
Carol L. Broderick, Cathleen K. Clutinger, Lynnie A. Irwin, Michael E. Laurila,
Chahrzad Montrose-Rafizadeh, Brian A. Oldham, Minmin Wang, Leonard L. Winneroski,
Chaoyu Xie, Jeremy S. York, Nathan P. Yumibe, Richard W. Zink and Nathan Mantlo
Eli Lilly and Company, Lilly Research Labs, Lilly Corporate Center, Indianapolis, IN 46285, USA
Received 15 July 2007; revised 14 October 2007; accepted 15 October 2007
Available online 18 October 2007
Abstract—A series of potent amide linked PPARc/d dual agonists (1a) has been discovered through rational design. In the ZDF rat
model of type 2 diabetes, compound (R)-3-[4-(3-{1-[(5-chloro-1,3-dimethyl-1H-indole-2-carbonyl)-amino]-ethyl}-5-fluoro-phen-
oxy)-2-ethyl-phenyl]-propionic acid (42) from this series has demonstrated glucose lowering efficacy comparable to the marketed
PPARc agonist rosiglitazone with less weight gain.
Ó 2007 Elsevier Ltd. All rights reserved.
The peroxisome proliferator-activated receptors (PPARs)
belong to the nuclear receptor superfamily of ligand-mod-
ulated transcription factors. There are three PPAR sub-
types, namely PPARa, c, and d. PPARs play a central
role in regulating the storage and catabolism of dietary
fats.1 PPARa agonists, such as fenofibrate, are effective
at lowering serum triglycerides and raising high-density
lipoprotein (HDL) cholesterol.1,2 The role of PPARc
has been extensively studied and is known to be involved
with glucose homeostasis, insulin sensitization, and fat
storage. PPARc agonists, such as rosiglitazone, 5-{4-[2-
(methyl-pyridin-2-yl-amino)-ethoxy]-benzyl}-thiazolidine-
2,4-dione, increase insulin sensitivity and have been ap-
proved for the treatment of type 2 diabetes.3 While not
as extensively studied as the other subtypes, the role of
PPARd has become clearer recently with the generation
of potent, selective ligands for this PPAR subtype. As
exemplified in studies with GW501516, {2-methyl-4-[4-
methyl-2-(4-trifluoromethyl-phenyl)-thiazol-5-ylmethylsul-
fanyl]-phenoxy}-acetic acid, PPARd activation appears to
increase fatty acid b-oxidation, insulin sensitivity, and
HDL cholesterol.4,5
associated cardiovascular disease while reducing the
weight gain typically observed with PPARc activation.
We hypothesized that a PPARc/d dual agonist could
effectively lower glucose and hemoglobin A1c levels
while improving the dyslipidemia commonly seen in dia-
betic patients by simultaneously activating both receptor
subtypes.6 In addition, we hypothesized that beneficial
effects could be achieved with a reduced weight gain pro-
file due to PPARd induced activation of fatty acid b-oxi-
dation and a reduced need to drive efficacy solely
through PPARc activation. Recently, we reported the
design and synthesis of a series of novel ether linked
PPARc/d dual agonists, analogs of compound 1b
(Fig. 1), and showed that a PPARc/d dual agonist with
a properly controlled c/d ratio can be efficacious on glu-
cose control while producing less weight gain than rosig-
litazone in animal models of diabetes.6,7
In an earlier effort to identify structurally novel PPAR
ligands, PPARc, d dual agonist 5 was discovered.
Through extensive SAR studies around 5, we identified
molecules with a wide range of PPARc/d in vitro activ-
ity. Herein, we report the design, synthesis, and biolog-
ical evaluation of these novel amide linked PPARc/d
dual agonists (1a, Fig. 1).
The primary goals of the present work were to develop a
PPAR agonist to treat type 2 diabetes and to prevent
A general synthetic scheme is illustrated in Scheme 1.
Nitriles 28 were hydrogenated in the presence of 5%
Pd/C or Raney nickel to give the corresponding amines
Keywords: PPAR; Diabetes; Dyslipidemia.
*
Corresponding author. Tel.: +1 317 277 9158; fax: +1 317 276
1417; e-mail: shi_qing@lilly.com
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.10.047

Q. Shi et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6744–6749
6745
R3
O
R2
R4
O
Y
OH
O
O
H
X
N
R4
O
OH
Cl
O
R1
O
1b
1a
X=Ar, heteroAr
R1=H, Me
R2=H, Me, F
R3=Me, Et
Y=CH2, O
R4=H, Me
Figure 1. Amide linked series 1a and ether linked compound 1b.
(3). Coupling 3 with a variety of aryl carboxylic acids
(4)8 and subsequent hydrolyses afforded carboxylic acids
5–24. Indole derivatives 25–33 were prepared in an anal-
ogous fashion starting from 3 and various indole-2-car-
boxylic acids.8
detrimental to all PPAR activity (11), the sterically
demanding phenoxy group (12) attenuated only
PPARc/d activities.
Compound 6 was identified as a potent and selective
PPARc/d dual agonist. Replacement of the CF3 substi-
tuent in 6 by Me, Cl, and OMe resulted in compounds
14, 15, and 17, which had significantly reduced PPARd
functional activities (EC50d = 286, 254, and 376 nM,
respectively). The OCF3 analog (16) showed attenuated
PPARc/d affinity (IC50c = 47 nM; IC50d = 54 nM).
Removing the CF3 group from 6 was detrimental to
PPARd potency (18, EC50d = 1878 nM).
Pure more active enantiomers of benzyl-substituted
compounds 42 and 43 were prepared as described in
Scheme 2. Commercially available 1,3-dibromo-5-flu-
oro-benzene 34 and 3-(2-ethyl-4-hydroxy-phenyl)-propi-
onic acid ethyl ester 358 were subjected to a 2,2,6,6,-
tetramethylheptane-3,5-dione accelerated Ullmann reac-
tion to give diaryl ether 36.9 Stille coupling of 36 and
tributyl (1-ethoxyvinyl) tin in the presence of dichlorobis
(triphenylphosphine) palladium (II) and subsequent
hydrolysis provided ketone 37. The ketone was reduced
enantioselectively to the corresponding alcohol (38)
using N,N-diethylaniline-borane and a catalytic amount
of (R)-2-methyl-oxazaborolidine (MeCBS).10 Treatment
of alcohol 38 with diphenyl phosphorazidate (dppa) and
1,8-diazabicyclo[5.4.0]undec-7-ene (DBU) afforded a
mixture of the desired azide (39) and a phosphate inter-
mediate, which was also converted to azide 39 using tri-
methylsilyl azide in the presence of fluoride anion.
Reduction of azide yielded amine 40. This enatiomeri-
cally enriched compound (R configuration, 67–85% ee)
was purified by chiral chromatography with a Chiralcel
OD column (eluent: 0.2% DMEA in MeOH) to give 40
as a single enantiomer (100% ee). Coupling amine 40
with 5-chloro-1,3-dimethyl-1H-indole-2-carboxylic acid8
or commercially available 5-chloro-3-methyl-benzo
[b]thiophene-2-carboxylic acid 41 and subsequent
hydrolyses provided carboxylic acids 42 and 43,
respectively.
Substitution on the spacer phenyl ring resulted in com-
pounds 19–23. Compared to parent compound 9
(EC50d = 22 nM), PPARd functional activity was im-
proved by 5-Me substitution (20, EC50d = 4 nM). The
5-F substituted compound (22) had comparable
PPARa, c, and d activities to 9. Substitution at C-4
and C-6 on the spacer phenyl ring (19, 21, and 23)
reduced PPARd functional activity.
Variation at the carboxylic acid headpiece influenced the
PPAR activity of these ligands. With the dihydrocin-
namic acid headpiece, 24 showed comparable PPARd
activity and reduced PPARa and c activities versus
fibrate-based 6.
To further reduce PPARa activity and increase PPARd
activity, we examined the tailpiece SAR through a set of
substituted indoles (25–33) using the dihydrocinnamic
acid headpiece (Table 2). Among the monosubstituted
indoles, 5-Cl substitution afforded the most potent
and selective PPARc/d dual agonist (26). The 5,6-di-
chloro-indole 30 showed attenuated PPARc activity.
With 5-Cl substitution, N-methylation (31) tended to
further reduce PPARa and c activity. Methylation at
C-3 (32) increased PPARa activity, maintained PPARc
activity, and significantly improved PPARd functional
activity. The combined trisubstituted indole compound
(33) proved to be a potent and selective PPARc/d dual
agonist as shown in Table 2, especially showed im-
proved d activity compared with substituted phenyl
compounds.
The final compounds were tested for PPAR binding
(IC50)11 and receptor transactivation (EC50)12 activities
(Tables 1–3). Based on the SAR results from our pre-
viously reported ether linked series (1b),2,3 we exam-
ined the effects of ortho substitution on the tailpiece
phenyl of 5 (Table 1). Substitutions of Me, Et, Cl, F,
Br, and OMe were all well tolerated for PPARc/d
activities (6–10 and 13). Fluorine (9) and methoxy
(13) substitution led to increased PPARa activities.
While a strong electron withdrawing group, CF3, was

6746
Q. Shi et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6744–6749
O
O
R2
F
R1
R3
i
Y
i
O
O
O
+
R2
NC
O
HO
Br
Br
2
34
35
F
O
R2
ii, iii
R1
R3
O
Y
ii, iii
O
O
+
X
OH
H2N
R2
O
Br
O
4
3
36
O
F
O
iv
R2
R1
5
Y
4
6
OH
O
H
X
N
R2
O
O
3
1
2
O
O
5 - 24
5: X= 4-CF3-phenyl, R1=H, Y=O, R2=Me
37
F
O
6: X= 2-Me-4-CF3-phenyl, R1=H, Y=O, R2=Me
7: X= 2-Et-4-CF3-phenyl, R1=H, Y=O, R2=Me
8: X= 2-Cl-4-CF3-phenyl, R1=H, Y=O, R2=Me
9: X= 2-F-4-CF3-phenyl, R1=H, Y=O, R2=Me
10: X= 2-Br-4-CF3-phenyl, R1=H, Y=O, R2=Me
11: X= 2-CF3-4-CF3-phenyl, R1=H, Y=O, R2=Me
12: X= 2-OPh-4-CF3-phenyl, R1=H, Y=O, R2=Me
13: X= 2-OMe-4-CF3-phenyl, R1=H, Y=O, R2=Me
14: X= 2-Me-4-Me-phenyl, R1=H, Y=O, R2=Me
15: X= 4-Cl-2-Me-phenyl, R1=H, Y=O, R2=Me
16: X= 2-Me-4-OCF3-phenyl, R1=H, Y=O, R2=Me
17: X= 4-OMe-2-Me-phenyl, R1=H, Y=O, R2=Me
18: X= 2-Me-phenyl, R1=H, Y=O, R2=Me
19: X= 2-F-4-CF3-phenyl, R1=6-Me, Y=O, R2=Me
20: X= 2-F-4-CF3-phenyl, R1=5-Me, Y=O, R2=Me
21: X= 2-F-4-CF3-phenyl, R1=4-Me, Y=O, R2=Me
22: X= 2-F-4-CF3-phenyl, R1=5-F, Y=O, R2=Me
23: X= 2-F-4-CF3-phenyl, R1=4-F, Y=O, R2=Me
24: X= 2-Me-4-CF3-phenyl, R1=H, Y=CH2, R2=H
v, vi
O
O
OH
38
F
O
vii
O
O
N3
39
F
O
Cl
OH
O
+
X
O
O
NH2
41
40
Cl
viii, ix
F
O
OH
H
Scheme 1. Reagents and conditions: (i) H2, HOAc, 5% Pd/C, rt, 60 Psi,
overnight or Raney nickel, 2 N NH3 in MeOH, 40 °C, 60 Psi,
overnight; (ii) EDC, HOBT, DIEA; (iii) LiOH, H2O, 1,4-dioxane.
N
X
O
O
42: X= MeN
43: X= S
Further modifications included substitution at the ben-
zylic carbon of the linker and evaluation of the benzo-
thiophene tailpiece (Table 3). We used the ethyl instead
of methyl substituted dihydrocinnamic acid headpiece
to hinder b-oxidation.6 Benzothiophene 43 exhibited a
similar in vitro profile compared with indole 42, both
being potent and selective PPARc/d dual agonists with
more relaxed c activity compared with compound 33.
Scheme 2. Reagents and conditions: (i) 2,2,6,6-tetramethyl-3,5-hep-
tanedione, Cs2CO3/NMP, CuCl, 120 °C, overnight; (ii) tributyl (1-
ethoxyvinyl) tin, Pd (PPh3)2Cl2, toluene; (iii) 1 N HCl; (iv) (R)-
MeCBS, N,N-diethylaniline-borane, toluene; (v) dppa, DBU, toluene;
(vi) TMSN3, TBAF (1 M in THF), 40 °C; (vii) Ph3P, THF, H2O; (viii)
EDC, HOBT, DIEA; (ix) LiOH, H2O, dioxane.
Compound 42 (ꢀ99% ee) was studied for in vivo eval-
uation in the male ZDF rat, a model of type 2 diabe-
tes.6 As shown in Figure 2, 42 dose-dependently
lowered plasma glucose levels in this model. Com-
pound 42 demonstrated 50% glucose normalization
(ED50) at 0.82 mg/kg compared to 1 mg/kg for rosig-
litazone. Moreover, compound 42 showed significantly
(p < 0.05) less weight gain than rosiglitazone at the

Q. Shi et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6744–6749
6747
11
12
Table 1. Binding IC50 and receptor transactivation EC50 data13 on human PPARs of compounds 5–24
Compound
hPPARa
IC50 (nM)
hPPARa
EC50 (nM)
hPPARc
IC50 (nM)
hPPARc
EC50 (nM)
hPPARd
IC50 (nM)
hPPARd
EC50 (nM)
5
3487
1763
1127
1230
640
1253
921
18
9
42
24
6
3
95
38
6
7
828
836
10
15
12
12
203
114
7
57
49
4
81
59
8
3
9
10
466
590
24
43
4
22
36
1594
10,313
1483
463
3
11
12
n.a.
706
1058
479
16
40
32
4
2870
409
30
13
14
329
1749
7154
4850
4200
n.b.
1042
2205
420
7
25
28
4
286
254
50
15
16
9
6
47
8
10
26
54
12
64
3
17
18
1504
2727
792
376
1878
96
20
5
125
24
19
20
188
313
544
877
12
17
8
37
43
5
4
86
21
22
331
728
6
754
22
28
3
48
23
24
384
1053
2888
n.a.
8
3
119
33
n.b.
n.b.
126
67
345
308
5
n.b.
Rosiglitazone
n.a.
11
12
Table 2. Binding IC50 and receptor transactivation EC50 data13 on human PPARs of compounds 25–33
R3
O
R2
O
OH
H
N
N
O
R1
Compound
R1
R2
R3
hPPARa
IC50 (nM)
hPPARa
EC50 (nM)
hPPARc
IC50 (nM)
hPPARc
EC50 (nM)
hPPARd
IC50 (nM)
hPPARd
EC50 (nM)
25
26
27
28
29
30
31
32
33
H
H
4-Cl
5-Cl
n.b
2997
2459
588
258
13
233
8
1088
25
12
3
1300
47
80
93
527
59
87
6
H
H
10,034
1132
7457
4865
4352
n.b.
H
H
6-Cl
348
32
3
H
H
5-Br
7-Br
5,6-di-Cl
5-Cl
2770
2404
1887
2735
558
4
H
H
37
95
98
27
35
158
238
135
24
9
H
H
5
Me
H
H
3
Me
Me
5-Cl
5-Cl
2104
n.b.
3
Me
2654
44
3
4
11
12
Table 3. Binding IC50 and receptor transactivation EC50 data13 on human PPARs of compounds 42 and 43
Compound
X
hPPARa
IC50 (nM)
hPPARa
EC50 (nM)
hPPARc
IC50 (nM)
hPPARc
EC50 (nM)
hPPARd
IC50 (nM)
hPPARd
EC50 (nM)
42
43
MeN
S
10,257
9336
2824
n.a.
39
50
116
163
5
5
1
7
same doses (Fig. 3). Additionally, compound 42 signif-
icantly lowered plasma triglycerides and plasma free
fatty acids at 3 mg/kg dose compared with the control
group (Figs. 4 and 5).
model, 42 demonstrated efficacy in lowering plasma glu-
cose levels comparable to rosiglitazone with a signifi-
cantly reduced weight gain side effect at the same
doses. These results further supported our initial
hypothesis that a PPARc/d dual agonist with a properly
controlled PPARc/d in vitro profile can reduce the
weight gain side effect associated with marketed PPARc
agonists.
In summary, we have presented the synthesis and
in vitro characterization of a novel series of potent
and selective PPARc/d dual agonists. In the ZDF rat

6748
Q. Shi et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6744–6749
Figure 2. Dose–response for effects of compound 42 and rosiglitazone
on glucose normalization in ZDF rats after 7 days of oral gavage
dosing.
Figure 5. Dose–response for effects of compound 42 on fatty acids in
ZDF rats after 7 days of oral gavage dosing.
Acknowledgment
Q.S. thanks James R. McCarthy for valuable discus-
sions regarding the research.
References and notes
1. Willson, T.; Brown, P.; Sternbach, D.; Henke, B. J. Med.
Chem. 2000, 43, 527.
2. Staels, B.; Dallongeville, J.; Auwerx, J.; Schoonjans, K.;
Leitersdorf, E.; Fruchart, J.-C. Circulation 1998, 98, 2088.
3. Lehmann, J.; Moore, L.; Smith-Oliver, T.; wilkison, W.;
Willson, T.; Kliewer, S. J. Biol. Chem. 1995, 270, 12953.
4. Lee, C.-H.; Olson, P.; Hevener, A.; Mehl, I.; Chong,
L.-W.; Olefsky, J.; Gonzalez, F.; Jungyeob, H.; Kang,
H.; Peters, J.; Evans, R. Proc. Natl. Acad. Sci. U.S.A.
2006, 103, 3444.
Figure 3. Dose–response for effects of compound 42 and rosiglitazone
on body weight in ZDF rats after 7 days of oral gavage dosing.
5. Oliver, W.; Shenk, J.; Snaith, M.; Russell, C.; Plunket, K.;
Bodkin, N.; Lewis, M.; Winegar, D.; Sznaidman, M.;
Lambert, M.; Xu, H.; Sternbach, D.; Kliewer, S.; Hansen,
B.; Willson, T. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,
5306.
6. Xu, Y.; Etgen, G. J.; Broderick, C. L.; Canada, E. J.;
Gonzalez, I.; Lamar, J.; Montrose-Rafizadeh, C.; Oldham,
B. A.; Osborne, J. J.; Xie, C.; Shi, Q.; Winneroski, L. L.;
York, J.; Yumibe, N.; Zink, R.; Mantlo, N. J. Med. Chem.
2006, 49, 5649.
7. Gonzalez, I. C.; Lamar, J.; Iradier, F.; Xu, Y.; Winnero-
ski, L. L.; York, J.; Yumibe, N.; Zink, R.; Montrose-
Rafizadeh, C.; Etgen, G. J.; Broderick, C. L.; Oldham, B.
A.; Mantlo, N. Bioorg. Med. Chem. Lett. 2007, 17, 1052.
8. Canada, E. J.; Gossett, L.; Mantlo, N.; Shi, Q.; Wang, M.;
Warshawsky, A. M.; Xu, Y. WO 2005054176.
9. Buck, E.; Song, Z.; Tschaen, D.; Dormer, P.; Volante, R.;
Reider, P. Org. Lett. 2002, 4, 1623.
10. Salunkhe, A.; Burkhardt, E. Tetrahedron Lett. 1997, 38,
1523.
11. IC50: Concentration of test compound required to displace
50% of tritiated ligand: tritium labeled PPARa/PPARd
agonist, 2-(4-{2-[3-(2,4-difluoro-phenyl)-1-heptyl-ureido]-
Figure 4. Dose–response for effects of compound 42 on plasma
triglycerides in ZDF rats after 7 days of oral gavage dosing.

Q. Shi et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6744–6749
6749
ethyl}-phenoxy)-2-methyl-butyric acid, and tritium
labeled PPARc agonist, 2-methyl-2-(4-{3-{propyl-(5-pyri-
din-2-yl-thiophene-2-sulfonyl)-amino}-propyl}-phenoxy)-
propionic acid, were used; n.b.: IC50 > 10 lM.
hPPARa was used to eliminate interference by endoge-
nous PPARc receptors in CV-1 cells; n.a.: EC50 > 10 lM.
13. Minimum significant ratio (MSR, smallest ratio of
affinities/potencies for statistical significance in assay
data) for PPARa, c, and d: IC50: 1.5, 2.6, 1.3; EC50:
2.5, 1.8, 2.9.
12. EC50: Concentration of test compound required to
produce 50% of the maximal reporter activity. Gal4-