Design, synthesis, and structure-activity relationship of podocarpic acid amides as liver X receptor agonists for potential treatment of atherosclerosis

Article abstract of DOI:10.1016/j.bmcl.2005.06.100

A series of podocarpic acid amides were identified as potent agonists for Liver X receptor α and β subtypes, which are members of a nuclear hormone receptor superfamily that are involved in the regulation of a variety of metabolic pathways including chole

Full text of DOI:10.1016/j.bmcl.2005.06.100

Bioorganic & Medicinal Chemistry Letters 15 (2005) 4574–4578
Design, synthesis, and structure–activity relationship
of podocarpic acid amides as liver X receptor agonists
for potential treatment of atherosclerosis
Weiguo Liu,* Steve Chen, James Dropinski, Lawrence Colwell, Michael Robins,
Michael Szymonifka, Nancy Hayes, Neelam Sharma, Karen MacNaul, Melba Hernandez,
Charlotte Burton, Carl P. Sparrow, John G. Menke and Sheo B. Singh*
Merck Research Laboratories, P.O. Box 2000, Rahway, NJ 07065, USA
Received 17 May 2005; revised 23 June 2005; accepted 29 June 2005
Available online 24 August 2005
Abstract—A series of podocarpic acid amides were identified as potent agonists for Liver X receptor a and b subtypes, which are
members of a nuclear hormone receptor superfamily that are involved in the regulation of a variety of metabolic pathways including
cholesterol metabolism. We recently reported podocarpic acid anhydride and imide dimers as potent LXR agonists. Through
parallel organic synthesis, we rapidly identified a series of new podocarpate leads with stable structures exemplified by adamantyl-
and phenylcyclohexylmethyl-podocarpic acid amides (14 and 18). Compound 18 exhibited LXRa/b 50/20 nM (binding affinity) and
33.7/35.3-fold receptor inductions. Synthesis, SAR, and biological activities of new podocarpate analogs are discussed.
Ó 2005 Elsevier Ltd. All rights reserved.
It has been well-established in the medical community
that higher level of cholesterol is a major risk factor
associated with the occurrence of coronary heart disease
and stroke. Pharmacologic intervention through effi-
cient regulation of cholesterol biosynthesis, metabolism,
acquisition, and transport in mammalian cells forms the
scientific base for drug therapy of atherosclerosis. Liver
X Receptors (LXR) are members of a superfamily of
nuclear hormone receptors represented by two subtypes,
LXRa and LXRb.1–3 Oxysterols have been identified as
endogenous ligands for both subtypes.4–7 These recep-
tors have been shown to play significant roles in choles-
terol homeostasis.8 Upon ligand binding, LXRs form
heterodimers with the retinoid X receptor (RXR), which
recruit a variety of co-activators and activate the expres-
sion of a number of genes involved, directly or indirect-
ly, in cholesterol and fatty acid metabolism including
ABCA1, a cholesterol transporter protein. It has been
shown that non-steroidal LXR agonists cause increased
expression of ABCA1 and raise the HDL levels in mice.
ABCA1 mediates the efflux of cholesterol out of the cells
and onto the ApoA1 protein of HDL particles. There-
fore, LXR agonists are expected to provide an opportu-
nity for the development of drugs to increase reverse
cholesterol transport and thus decrease the burden of
atherosclerosis.8
In a previous recent account, we reported the identifica-
tion of acetyl podocarpic acid anhydride (1) and imide
dimer (2) as potent agonists of LXR a and b recep-
tors.9,10 Monomeric natural product, podocarpic acid
(3), and various differently linked dimeric compounds
were either significantly less active or inactive. While
imide 2 exhibited good overall activity including in vivo
activity, it lacked the required physical and pharmacoki-
netic (PK) properties to be considered for further devel-
opment. During the SAR study of dimers, we observed
that unsymmetrical dimers with stable linkages retained
some LXR receptor-binding activity, albeit much weak-
er. It also became clear to us that all the active dimers
were linked through at least one carboxyl functional
group to the second molecule of the podocarpate. We
reasoned that if we take one molecule of podocarpic
acid and couple it through ester, amide, and imide
bond with readily available alcohols and amines, we
would generate a large number of podocarpic esters,
amides, and imides. By changing the shape, size,
Keywords: Atherosclerosis; LXR agonist; Cardiovascular; Cholesterol
lowering; Podocarpic acid.
*
Corresponding authors. Tel.: +732 594 3222; fax:+ 732 594 6880;
e-mail addresses: Weiguo_liu@merck.com; sheo_singh@merck.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.06.100

W. Liu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4574–4578
4575
polarity, functionality, and stereochemistry of the amino
donors, we would be able to fine-tune the SAR of these
analogs toward obtaining potent LXR agonists. The
discovery, synthesis, and SAR of selected esters (4–7),
amides (8–23), imide (24–25), and C-linked diketone
(26) are described.
prepared over 250 analogs of podocarpic acid predomi-
nated by amides together with a few esters, selected imi-
des, and C-linked diketones. Polar amides (including
O- and N-containing heterocycles) and esters with or
without linkers were uniformly inactive. Representative
examples of compounds are given in Table 1. The car-
boxy amide and the methyl ester did not exhibit any
binding activity. Lipophilic amides and esters exhibited
variable binding activity. Benzyl ester (4) exhibited bind-
ing activities with binding IC50 value of ꢀ2.8 lM against
both receptors. The allyl ester (5) was inactive, whereas
the esters with four (e.g., 6) and five (e.g., 7) carbon
chains showed modest activity. In the amide series,
phenyl amide was completely inactive, whereas benzyl
amide did not exhibit any binding affinity for a-receptor
but exhibited binding affinity for b-receptor
(IC50 = 3.9 lM). The extension of the chain length by
one methylene group (e.g., phenethyl derivative, 8) led
to improvement of activities against both receptors (Ta-
ble 1). However, p-hydroxylation of the phenethyl amide
lost activity completely. Amides of amino acids with
tert-butyl or allyl ester exhibited binding IC50 ranging
from 1 to 5 lM but only rarely exhibited any activity
in the functional HTRF assays. Amides with bulky alkyl
groups with or without phenyl group exhibited some of
the best activities. While a number of non-aromatic
amides (e.g., 9–13) exhibited sub lM to 5 lM IC50 val-
ues, adamantyl amides (10–11) exhibited better activities
and were pursued further. The binding affinities of the
primary (10) and the secondary (11) adamantyl amides
were essentially identical, exhibiting IC50 values of
ꢀ0.10 and ꢀ0.18 lM for a and b LXR receptors, respec-
tively, displaying about 2-fold preference for the a-re-
ceptor. Shortening of the chain length (e.g., 14) or
movement of the nitrogen to the adjacent carbon (e.g.,
15) did not have any significant impact on the binding
affinity. The ester substitution at C-9 of adamantyl ring
(e.g., 16) retained most of the binding affinity but hydro-
lysis of the ester to carboxylic acid lost all of the affinity
(data not shown). Monohydroxy substitutions at C-5,
C-7, or C-8 of the adamantyl ring resulted in inactive
(IC50 > 10-50 lM) compounds. Likewise, reduction of
the amide keto group of 10 to methylene lost the activity
(a/b binding IC50 7.04/8.50 lM). Substitution of podo-
carpic acid part of 10 with 13-hydroxy 13-iodo, 13-bro-
mo, 7-keto retained only 5–10% of the activity (data not
shown).
OAc
OH
OH
13
O
O
O
O
H
H
H
H
1
O
NH
7
4
O
H
OH
3
OAc
OH
1
2
Most of the primary and secondary amides presented in
Table 1 were synthesized by the reaction of primary and
secondary amines with unprotected podocarpic acid (3),
either by activation of HOBt and EDAC in CH2Cl2 or
Bop reagent in DMF in excellent yield. Synthesis of
some of the hindered secondary amides and cyclic
amides required protection of the phenolic group as a
benzyl ether10 and activation by either of the two re-
agents in DMF, followed by removal of benzyl ether
by catalytic hydrogenolysis (Scheme 1). Esters were syn-
thesized by acid chloride activation, followed by reac-
tion with appropriate alkyl halides. Imides were
synthesized by reacting benzyl protected podocarpic
amide (27)9,10 with appropriate acid chlorides (Scheme
2).
The diketone 26 was prepared in good yield by acylation
of deprotonated podocarpic methyl ketone (28)10 with
adamantyl acid chloride, followed by hydrogenation
(Scheme 3).
Deprotonation of the podocarpic amide (18) with sodi-
um hydride, followed by reaction with diethyl phosphor-
ochloridate, produced phosphonate (29), which upon
metal ammonia reduction provided 13-deoxy podocar-
pic amide (19) in >70 overall yield (Scheme 4).
Like dimeric podocarpates, all compounds were first
evaluated in LXRSPA a- and b-binding assay using
{3-chloro-4-(3-(7-(2,3-ditritio-propyl)-3-trifluoromethyl-
6-(4,5)-isoxazolyl) propylthio)-phenyl acetic acid, 30,
followed by measurement of in vitro agonist activity in
a coactivator association assay using recombinant ste-
roid receptor coactivator 1 (SRC1) protein and recombi-
nant human LXRa and b ligand-binding domains in a
homogeneous time-resolved fluorescence (HTRF) as-
say.11 Cell-based transactivation (TA) assay was used
to evaluate functional agonist or antagonist activities
of compounds using a chimeric LXR construct in
HEK-293 cells. This assay uses fusion proteins with
the yeast Gal4 DNA-binding domain connected to the
hinge region and the LBD domain of either LXR human
receptors and has been previously described.9,11 We
The binding affinity of these compounds correlated well
with the coactivator association activity against the a-re-
ceptor but not against the b-receptor. The three best
compounds of the series 10, 11, and 16 exhibited EC50
values of 0.23, 0.14, and 0.76 lM, respectively, against
a-receptor in coactivator association HTRF assay and
were full agonists. These compounds showed 22.6-,
29.5-, and 14.3-fold maximal induction, respectively, of
the LXR a-receptor in the transactivation assay. Despite
lack of stimulation against the b-receptor in the HTRF
assay, these compounds exhibited 24.4-, 19.4-, and 19.7-
fold TA induction, respectively, against the b-receptor.
In contrast to dimeric series, the adamantyl imide (24)
and C-linked diketone (26) derivatives had lost signifi-
cant activities in all assays.

4576
W. Liu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4574–4578
Table 1. Podocarpic acid ester, amides, imide, and C-linked diketone derivatives and their LXR activities
Compound R-podocarpate
SPA binding
(IC50, lM)
Coactivator association HTRF
assay (EC50, lM)
Transactivation
(EC50, lM)
TA max fold
induction
LXRa
LXRb
0.002
LXRa
0.002
0.001
>15
LXRb
0.002
0.001
>15
LXRa
LXRb
LXRa LXRb
1
2
4
5
6
7
8
Acetyl anhydride dimer
Imide dimer
Ph CH2O
0.002
0.001
2.8
0.010
<0.003
NA
0.010
<0.003
NA
60
24
54
30
0.001
2.9
NA
NT
NT
NT
NA
NT
NT
NT
CH2@CHCH2O
CH2@CHCH2CH2O
CH2@CHCH2CH2CH2O
PhCH2CH2NH
>15
>15
>50
>50
>50
>50
>50
>50
>50
NT
NT
NT
NT
5.06
3.8
3.85
5.0
NT
40% at 10a
NT
28% at 10a
2.93
2.22
5.2
10.5
11.5
9
3.1
4.12
0.19
3.1
>50
NA
NA
2.2
2.8
NH
NH
1
10
0.10
0.23
>50
0.86
50% at 10a
22.6
24.4
5
9
7
NH
11
12
13
14
15
16
17
18
19
20
21
22
23
0.09
0.35
0.27
0.14
0.13
0.03
0.30
0.05
0.20
0.49
0.20
0.75
0.36
0.18
0.31
0.20
0.17
0.10
0.53
0.19
0.02
0.13
0.42
0.18
0.50
0.37
0.14
>50
136% at 10a 49% at 10a
29.5
11.7
12.7
41.2
2.6
19.4
7.1
0.24 (40%)b >50
63% at 10a
13% at 10a
78% at 10a
81% at 10a
3.80
NH
NH
0.28
>50
1.70
23.3
26.8
2.3
NH
0.22
0.22 (27%)b 130% at 3a
NH
NH
0.32
>50
0.90
0.83
O
0.76
>10
>50
>50
>50
>50
51% at 3a
14.3
47.5
33.7
18.9
21.7
22.8
5.6
19.7
44.5
35.3
15
EtO
N
0.32
430% at 10a 139% at 10a
0.05
164% at 3a
98% at 10a
42% at 10a
90% at 3a
40% at 10a
21% at 10a
95% at 1a
275% at 1a
111% at 3a
NH
12-deoxy-18
0.17 (49%)
0.36
NH
5.5
0.07
0.07 (31%)b 92% at 1a
22.1
4.4
N
0.43
1.5
175% at 3a
144% at 3a
N
0.15
0.1 (25%)b
24.2
32.1
N
(continued on next page)

W. Liu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4574–4578
4577
Table 1 (continued)
Compound
R-podocarpate
SPA binding
(IC50, lM)
Coactivator association
HTRF assay (EC50, lM)
Transactivation
(EC50, lM)
TA max fold
induction
LXRa
LXRb
LXRa
LXRb
LXRa
LXRb
LXRa
LXRb
NH
O
24
25
26
30
0.29
0.42
1.67
0.035
0.29
0.12
1.01
0.025
1.5 (38%)b
>50
20% at 10a
NA
5.1
1.7
O
NH
0.17
>50
>50
NA
NA
NA
NA
CH2
O
>50
62% at 10a
13% at 10a
5.4
4.4
Control compound
0.035
0.016
1.79
1.08
20
35
a The % activation did not plateau.
b Partial agonist; NT (not tested); NA (not active at 10 lM); 30 {3-chloro-4-(3-(7-(2,3-ditritio-propyl)-3-trifluoromethyl-6-(4,5)-isoxazolyl) propyl-
thio)-phenyl acetic acid)}.
OBn
OH
OBn
OH
i-ii
i-iii
H3C
HO
HO
H
H
H
H
>60-90%
O
O
O
O
O
3
Bn ether
28
26
v
v
vi
iv or v
vi
Scheme 3. Reagents: (i) NaHMDS, THF, Adamantyl chloride; (ii) H2,
Pd/C, 45 psi.
OH
OH
OEt
OEt
P
RHN
N
H
H
OH
O
H
O
O
O
Cyclic amides
Amides
i
ii
Scheme 1. Reagents: (i) TMSCHN2, MeOH, PhH; (ii) Cs2CO3, BnBr,
DMF; (iii) K-t-BuO, DMSO, H3O+; (iv) HOBt, EDAC, CH2Cl2,
primary amines (v) Bop reagent, DMF, primary or secondary or cyclic
amines; (vi) H2, 10% Pd/C, 45 psi.
O
O
O
H
H
NH
NH
NH
18
29
19
OBn
OH
Scheme 4. Reagents: (i) NaH, THF, diethyl phosphorochloridate;
(ii) diethyl ether, liquid NH3, Li, À78 °C.
i-ii
H
N
R
H2N
H
H
design elements, such as variation of the length of the
linker and introduction of constraints leading to the syn-
thesis of primary, secondary, and cyclic amides. The
SAR of this class of compounds was very complex and
clustered in three groups. The most interesting com-
pounds of the group are presented in Table 1. Of these,
the primary amide 18 and the cyclic amide 21 were most
potent and exhibited a/b binding IC50 values of 0.05/
0.02 and 0.20/0.18 lM, respectively. While the primary
aromatic amide 18 exhibited potent activity (EC50
0.05 lM) in the coactivator association assay against
the a-receptor, it was completely inactive against the
b-receptor. In contrast, the cyclic aromatic amide 21
was a full agonist (EC50 0.07 lM) against the a-receptor
O
O
O
27
Imides
Scheme 2. Reagents: (i) NaHMDS, THF, acid chlorides, (ii) H2, Pd/C,
45 psi.
One of the other bulky aliphatic cyclic amides that is
worth mentioning is 17 that exhibited a/b binding IC50
of 0.30/0.19 lM, coactivator association EC50 of 0.32/
> 50 lM, and a/b TA fold induction of 47.5/44.5.
The phenethyl lead was explored further by coupling of
commercially available aromatic amines with various

4578
W. Liu et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4574–4578
and a partial agonist (EC50 0.07 lM) against the b-re-
ceptor (maximal stimulation observed was only 31%
for the b-receptor). The primary and secondary amides
(e.g., 18 and analogs, 20) were consistently devoid of
the b-receptor HTRF activity and exhibited selectivity
for a-receptor, while all cyclic amides (e.g., 21–23)
exhibited b-receptor HTRF activity and did not show
any selectivity. The amide 18 was a potent activator
and exhibited 33.7-fold activation of the a-receptor
and 35.3-fold activation of the b-receptor. Cyclic amide
21 was also a potent activator, providing 92% and 95%
max activation of a and b receptors in TA assays at
1 lM affording 22.8- and 22.1-fold inductions,
respectively.
amide 18 was evaluated for in vivo efficacy in C57Bl/
6J male mice (fed Chow diet) at 30 mg/kg twice daily
by oral administration for 7 days. Cholesterol and tri-
glyceride levels were measured. Unlike imide 2, this
compound did not show any effect on the lipid levels
potentially due to its lower (ꢀ50-fold) intrinsic potency
(compared to 2) and lower exposure level.
In summary, in this paper we have described podocarpic
amides as potent agonists of LXR, defined broad SAR,
and extended our initial studies of dimers.
References and notes
Of these two most active leads, amide 18 was selected for
further SAR studies. First, N-methylation caused 10- to
15-fold loss of activities. Replacement of the cyclohexyl
ring with cyclopropyl, cyclobutyl, cyclopentyl rings and
dimethyl group, substitution in the aromatic ring with
m- and p-OMe, m- and p-CO2Me uniformly resulted in
the reduction of binding affinity by 2- to 20-fold but sub-
stitution with m- or p-CO2H lost affinity completely. The
elimination of the phenolic group (e.g., 19) and the con-
version to imide analog (25) resulted in a reduction of
binding affinity and agonist activities.
1. Willy, P. J.; Umesono, K.; Ong, E. S.; Evans, R. M.;
Heyman, R. A.; Mangelsdorf, D. J. Gene Dev. 1995, 9,
1033.
2. Miyata, K. S.; McCaw, S. E.; Patel, H. V.; Rachubinski,
R. A.; Capone, J. P. J. Biol. Chem. 1996, 271, 9189.
3. Shinar, D. M.; Endo, N.; Rutledge, S. J.; Vogel, R.;
Rodan, G. A.; Schmidt, A. Gene 1994, 147, 273.
4. Janowski, B. A.; Willy, P. J.; Devi, T. R.; Falck, J. R.;
Mangelsdorf, D. J. Nature 1996, 383, 728.
5. Lehmann, J. M.; Kliewer, S. A.; Moore, L. B.; Smith-
Oliver, T. A.; Oliver, B. B.; Su, J.-L.; Sundseth, S. S.;
Winegar, D. A.; Blanchard, D. E.; Spencer, T. A.; Wilson,
T. M. J. Biol. Chem. 1997, 272, 3137.
6. Janowski, B. A.; Grogan, M. J.; Jones, S. A.; Wisely, G.
B.; Kliewer, S. A.; Corey, E. J.; Mangelsdorf, D. J. Proc.
Natl. Acad. Sci. U.S.A. 1999, 96, 266.
7. Fu, X.; Menke, J. G.; Chen, Y.; Zhou, G.; MacNaul, K.
L.; Wright, S. D.; Sparrow, C. P.; Lund, E. G. J. Biol.
Chem. 2001, 276, 38378.
8. Repa, J. J.; Mangelsdorf, D. J. Nat. Med. 2002, 8, 1243.
9. Sparrow, C. P.; Baffic, J.; Lam, M.-H.; Lund, E. G.;
Adams, A. D.; Fu, X.; Hayes, N.; Jones, A. B.; Macnaul,
K. L.; Ondeyka, J.; Singh, S.; Wang, J.; Zhou, G.; Moller,
D. E.; Wright, S. D.; Menke, J. G. J. Biol. Chem. 2002,
277, 10021.
10. Singh, S. B.; Ondeyka, J. G.; Liu, W.; Chen, S.; Chen, T.
S.; Li, X.; Bouffard, A.; Dropinski, J.; Jones, A. B.;
McCormick, S.; Hayes, N.; Wang, J.; Sharma, N.;
MacNaul, K.; Hernandez, M.; Chao, Y.-S.; Baffic, J.;
Lam, M.-H.; Burton, C.; Sparrow, C. P.; Menke, J. G.
Bioorg. Med. Chem. Lett. 2005, 15, 2824.
Movement of the nitrogen one carbon (e.g., 23) or con-
straining the flexibility (e.g., 22) of the cyclic amides led
to the loss of activity, indicating that both position of
the nitrogen and slight flexibility are important for bind-
ing affinities and agonist properties. Derivatization of
the phenolic group as ether or ester (except acetate)
was detrimental for the activity in all series.
Before in vivo studies, Male SD rats PK studies of 14
and 18 were conducted. Rats were dosed at 1 mg/kg
intravenous (N = 2) in EtOH:PEG400:H2O (20:30:50
v/v/v, 1 mg/mL) and 2 mg/kg oral (N = 3) by gavage.
The compound was recovered from blood by acetoni-
trile precipitation and analyzed by LC-MS/MS with
external standard. Amide 14 afforded iv AUC of
1.03 lM.h.kg/mg, Clp 42 mL/min/kg, Vdss of 2.67 L/
kg and t1/2 of 1.15 h; and po AUC 0.01 lM.h.kg/mg,
Cmax 0.01 lM and F 1.1%. Amide 18 gave iv AUC of
0.47 lM.h.kg/mg, high Clp 82 mL/min/kg, high Vdss
of 9.5 L/kg and t1/2 of 1.8 h; and po AUC
0.04 lM.h.kg/mg, Cmax 0.03 lM and F 13.5%. The
11. Menke, J. G.; Macnaul, K. L.; Hayes, N. S.; Baffic, J.;
Chao, Y.-S.; Elbrecht, A.; Kelly, L. J.; Lam, M.-H.;
Schmidt, A.; Sahoo, S.; Wang, J.; Wright, S. D.; Xin, P.;
Zhou, G.; Moller, D. E.; Sparrow, C. P. Endocrinology
2002, 143, 2548.