Synthesis and evaluation of an orally available “Y”-shaped biaryl peroxisome proliferator-activated receptor δ agonist

Article abstract of DOI:10.1016/j.bmc.2018.06.044

In this study, we designed and synthesized several novel “Y”-shaped biaryl PPARδ agonists. Structure-activity relationship (SAR) studies demonstrated that compound 3a was the most active agonist with an EC₅₀ of 2.6 nM. We also synthesized and evaluated enantiospecific R and S isomers of compound 3a to confirm that R isomer (EC₅₀ = 0.7 nM) shows much more potent activity than S isomer (EC₅₀ = 6.1 nM). Molecular docking studies between the PPAR ligand binding domain and enantiospecific R and S isomers of compound 3a were performed. In vitro absorption, distribution, metabolism, excretion, and toxicity (ADMET) and in vivo PK profiles show that compound 3a possesses superior drug-like properties including good bioavailability. Our overall results clearly demonstrate that this orally administrable PPARδ agonist 3a is a viable drug candidate for the treatment of various PPARδ-related disorders.

Full text of DOI:10.1016/j.bmc.2018.06.044

Bioorganic&MedicinalChemistryxxx(xxxx)xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Synthesis and evaluation of an orally available “Y”-shaped biaryl
peroxisome proliferator-activated receptor δ agonist
Dong-Su Kima,f, Jaehwan Leeb,f, Ashwini M. Londhed,e, Tara Man Kadayata, Jeongmin Jooa,
Hayoung Hwanga, Kyung-Hee Kima, Ae Nim Paed,e, Jungwook China,⁎, Sung Jin Choa,⁎
,
⁎
Heonjoong Kangb,c,
a
New Drug Development Center, Daegu-Gyeongbuk Medical Innovation Foundation, Daegu 41061, Republic of Korea
b
The Center for Marine Natural Products and Drug Discovery, School of Earth and Environmental Sciences, Seoul National University, NS-80, Seoul 151-747, Republic of
Korea
c
Research Institute of Oceanography, Seoul National University, NS-80, Seoul 151-747, Republic of Korea
Convergence Research Center for Diagnosis, Treatment and Care System of Dementia, Korea Institute of Science and Technology, PO Box 131, Cheongryang, Seoul 130-
d
650, Republic of Korea
e
Division of Bio-Medical Science & Technology, KIST School, Korea University of Science and Technology, Seoul 02792, Republic of Korea
A R T I C L E I N F O
A B S T R A C T
Keywords:
Nuclear receptors
In this study, we designed and synthesized several novel “Y”-shaped biaryl PPARδ agonists. Structure-activity
relationship (SAR) studies demonstrated that compound 3a was the most active agonist with an EC50 of 2.6 nM.
We also synthesized and evaluated enantiospecific R and S isomers of compound 3a to confirm that R isomer
(EC50 = 0.7 nM) shows much more potent activity than S isomer (EC50 = 6.1 nM). Molecular docking studies
between the PPAR ligand binding domain and enantiospecific R and S isomers of compound 3a were performed.
In vitro absorption, distribution, metabolism, excretion, and toxicity (ADMET) and in vivo PK profiles show that
compound 3a possesses superior drug-like properties including good bioavailability. Our overall results clearly
demonstrate that this orally administrable PPARδ agonist 3a is a viable drug candidate for the treatment of
various PPARδ-related disorders.
PPARδ
Structure-activity relationship (SAR)
ADMET
PK
1. Introduction
fiber conversion to enhance endurance.17,18 Thus, diverse strategies for
the regulation of lipid metabolism by modulating PPARδ have been
Peroxisome proliferator-activated receptors (PPARs) have recently
received significant attention due to their function as transcription
factors that regulate gene expression patterns of biological processes.1–3
These nuclear receptors comprise three subtypes including PPARα,
PPARδ and PPARγ that have tissue-specific expression and functions in
vivo.4–10 Among these subtypes, PPARδ has been implicated in various
metabolic disorders including diabetes, obesity, atherosclerosis, and
cancer.11 The major functions of PPARδ include reproductive cell ex-
pression, regulation of neuronal cell differentiation in the central ner-
vous system (CNS), and anti-inflammatory effects.12–14 In addition,
PPARδ appears to play an important role in adipocyte differentiation
and lipid metabolism.15,16 Recent studies show that PPARδ is also in-
volved in the generation of new mitochondria in muscle and muscle
considered for the treatment of obesity and various metabolic diseases.
GW501516 and GW610742, discovered and developed by
GlaxoSmithKline, are the most well-known PPARδ agonists
(Figure 1a).19,20 They are quite effective and specific PPARδ agonists
and may be useful for the treatment of hyperlipidemia and type 2
diabetes.21,22 However, novel PPARδ agonists are still required. To this
extent, a variety of scaffolds for PPARδ agonist have been devised.
Recently, we developed a “Y”-shaped GW501516 derivative 1 that
shows high potency and selectivity for hPPARδ (Figure 1b).23 Another
“Y”-shaped PPARδ -specific agonist 2 was devised by Evans et al.24
Based on these works, we developed a new “Y”-shaped PPARδ agonist,
compound 3a, bearing a distinctive biaryl core structure in place of the
phenylthiazole (Figure 1c). In this study, we describe the synthesis of
⁎
Corresponding authors at: New Drug Development Center, Daegu-Gyeongbuk Medical Innovation Foundation, Daegu 41061, Republic of Korea (J. Chin and S.J.
Cho). The Center for Marine Natural Products and Drug Discovery, School of Earth and Environmental Sciences, Seoul National University, NS-80, Seoul 151-747,
Republic of Korea (H. Kang).
E-mail addresses: jwchin@dgmif.re.kr (J. Chin), sjcho@dgmif.re.kr (S.J. Cho), hjkang@snu.ac.kr (H. Kang).
These authors contributed equally to this work.
f
https://doi.org/10.1016/j.bmc.2018.06.044
Received 10 May 2018; Received in revised form 29 June 2018; Accepted 30 June 2018
0968-0896/©2018ElsevierLtd.Allrightsreserved.
Pleasecitethisarticleas:Kim,D.-S.,Bioorganic&MedicinalChemistry(2018),https://doi.org/10.1016/j.bmc.2018.06.044

D.-S. Kim et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxx–xxx
Fig. 1. (a) Well-known PPARδ agonists; (b) previously reported “Y”-shaped
PPARδ agonists; (c) devised “Y”-shaped biaryl PPARδ agonists.
Scheme 1. Synthetic route for “Y”-shaped biaryl PPARδ agonist 3a. Reagents
and conditions: (a) SOCl2, DMF, DCM, rt, 12 h, 93%; (b) 4-mercapto-2-me-
thylphenol, Cs2CO3, ACN, rt, 12 h, 72%; (c) TBDMSCl, imidazole, DMF, rt, 12 h,
72%; (d) benzyl bromide, lithium diisopropylamide, THF, −78 °C to rt, 4 h,
56%; (e) TBAF, THF, rt, 1 h, 98%; (f) methyl 2-bromoacetate, Cs2CO3, ACN, rt,
12 h, 92%; (g) 1 M LiOH, THF/H2O, rt, 15 min, 99%
this novel “Y”-shaped biaryl compound that has high potency and se-
lectivity for PPARδ (Figure 1c). Its structure–activity relationship
(SAR), in vitro ADMET, and in vivo PK profiles were also evaluated.
Additional molecular docking studies of the best compound in this
series, 3a, were demonstrated to describe the potential binding mode
with the target protein.
in 3g < 2-fluoro-5-chlorophenyl in 3h (EC50 = 18.0, 12.0, 9.1, 5.8,
5.3 nM, respectively) (entries 4–8). A 4-fluoro-1,1'-biphenyl group at
the R2 position with respective substituents such as 3,4,5-tri-
fluorophenyl in 3i, 2-fluoro-5-(trifluoromethyl)phenyl in 3j, and 2-
fluoro-5-chlorophenyl in 3k groups at R1 displayed low activities
(EC50 = 65.0, 38.0, 25.0 nM) (entries 9–11). However, when the R1
group was changed to phenyl in 3l or 2,5-difluorophenyl in 3m groups,
better activities (EC50 = 7.0, 4.0 nM) were observed (entries 12–13). As
we predicted, compound 3n functionalized with 3,4,5-trifluorophenyl
group at R1 and biphenyl group at R2 position showed low activity with
an EC50 of 86.0 nM (entry 14). Subsequently, changing the R1 group in
3n to 2-fluoro-5-chlorophenyl in 3o or 2-fluoro-5-(trifluoromethyl)
phenyl in 3p lowered the EC50 to 26.0 and 10.0 nM, respectively (en-
tries 15–16). Interestingly, replacing the 3,4,5-trifluoro-1,1'-biphenyl
group at the R2 position with various R1 substituents such as phenyl in
3q, 2-fluoro-5-chlorophenyl in 3r, 2,5-difluorophenyl in 3s, and 2,5-
dichlorophenyl in 3t groups resulted in excellent PPARδ activities
overall (EC50 = 4.5, 6.9, 6.3, and 9.5 nM, respectively) (entries 17–20).
This SAR analysis shows that several of the “Y”-shaped biaryl com-
pounds that we synthesized are excellent PPARδ agonists with com-
pound 3a being the best among them.
Next, enantiospecific synthesis of compound 3a was attempted
(Scheme 2). Initially, (R)- and (S)-9a were isolated from a racemic
mixture of compound 9a using chiral HPLC (Chiralpak AD-H,
Hex:IPA = 90:10). The enantiospecific structures were confirmed by X-
ray crystallography of camphanoyl group incorporated (R)-9a gener-
ated by reaction of (R)-9a with (1S)-(-)-camphanic chloride (see SI).
Further SN2 reactions of (R)- and (S)-9a with methyl 2-bromoacetate,
followed by hydrolysis were carried out to provide (R)- and (S)-3a,
respectively. hPPARδ transactivation assays comparing (R)- and (S)-3a
demonstrate that (R)-3a has a superior EC50 of 0.7 nM, whereas that of
(S)-3a was only 6.1 nM. These results demonstrate that the R-isomer
(R)-3a shows much better activity than the S-isomer (S)-3a as a PPARδ
agonist.
2. Results and discussion
2.1. Chemistry
The synthesis of compound 3a is described as the represented
scheme for synthesizing the series of “Y”-shaped biaryl PPARδ agonists
in this manuscript (Scheme 1).25 Briefly, the chlorination of 4a with
SOCl2 with a catalytic amount of DMF generates 5a (93% yield). The
SN2 type reaction of 5a with 4-mercapto-2-methylphenol affords 6a
which undergoes protection with TBDMSCl to yield 7a. Benzylation of
7a was then carried out by treatment with LDA (2 equiv) at −78 °C,
followed by slow addition of benzyl bromide into the reaction mixture
to provide 8a (56% yield). Deprotection of 8a with TBAF gives 9a
which undergoes a SN2 reaction with methyl 2-bromoacetate to pro-
duce 10a. Finally, hydrolysis of 10a in the presence of 3 equiv of
aqueous 1 M LiOH leads to formation of the target product, 3a (99%
yield).
Based on the above synthetic scheme (Scheme 1), diverse “Y”-
shaped biaryl PPARδ agonists 3a-t were successfully synthesized, and
their activities were evaluated (Table 1). SAR analysis, using a trans-
activation assay, shows that introducing a phenyl group at R1 and a 4-
(trifluoromethyl)-1,1'-biphenyl group at the R2 position resulted in a
potent and selective hPPARδ agonist, 3a (EC50 = 2.6 nM) (Table 1,
entry 1). When the R2 position in compound 3a was replaced with a 2-
(4-(trifluoromethyl)phenyl) pyridyl group, similar activity (compound
3b, EC50 = 3.2 nM) was observed (entry 2). Interestingly, when the R1
position in compound 3b was changed to 3,4,5-trifluorophenyl group, it
showed less activity (3c, EC50 = 56.0 nM) than compound 3b (entry 3).
When a 3-(trifluoromethyl)-1,1'-biphenyl group was introduced at the
R2 position, the activities of the corresponding compounds 3d-h in-
creased in the order phenyl compound 3d < 3,4,5-trifluorophenyl in
3e < 2,5-difluorophenyl in 3f < 2-fluoro-5-(trifluoromethyl)phenyl
2

D.-S. Kim et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxx–xxx
2.2. Docking results
Table 1
In vitro activities of compounds 3a-ta
We next performed in silico docking studies for (R)- and (S)-3a, and
GW501516 using the X-ray crystal structure of PPARδ protein (PDB ID:
1Y0S).26 Discovery Studio 2017 R2 client software was utilized for this
in silico modeling study.27 Ligand docking was performed using the
-CDOCKER method implemented in Discovery Studio. Images were
rendered using pymol software (www.pymol.org). The docked pose of
the three molecules show good overlap with each other (Figure 2A).
GW501516 occupies two pockets of the binding cavity (Figure 2B). For
the (R)- and (S)-3a isomers, the phenyl group at the R1 position ex-
tended into the third pocket inside binding cavity demonstrating its Y-
shaped binding as compared to GW501516.
GW501516 and (R)- and (S)-3a have a similar polar group (acetic
acid moiety) that has four hydrogen bond interactions with His449,
His323, and Tyr473 (Figure 2C-E). In addition to these polar contacts,
(R)-3a (–CDOCKER score 50.34) showed a strong network of non-polar
contacts with the ligand binding domain (LBD). The 2-methyl phenyl
group interacts with Ile326, Phe327, Cys285, His323, and His449 by
hydrophobic interactions. The phenyl group at the R1 position makes
hydrophobic contact with Leu330 and ILE333. The trifluoromethyl-
(phenyl) phenyl group is strongly packed by hydrophobic interactions
with Ile249, Leu255, Val281, Arg284, Leu339, Val341, and Val348. The
S-isomer (S)-3a shows an interaction pattern similar to (R)-3a with a
–CDOCKER score of 45.80 (Figure 2D). The only difference between
(R)- and (S)-3a isomer binding is the phenyl (R1-position) group or-
ientation inside the LBD (Figure 2B). Despite this confirmation differ-
ence, the phenyl (R1-position) group in both isomers showed hydro-
phobic contact with the same amino acid residues, Ile333 and Leu330.
In the case of GW501516, in addition to its hydrogen bonds, it has
hydrophobic contact with Trp264, Val281, Arg284, Cys285, Thr289,
His323, Phe327, Leu330, Ile333, Leu339, Val341, Val348, His449, and
a –CDOCKER score of 53.03 (Figure 2E).
Along with the –CDOCKER energy, we have also calculated the
binding energy of these molecules (Table 2). As compared to
GW501516 (binding energy −335.06) and the (S)-3a isomer (binding
energy −325.37), (R)-3a has greater binding energy of −343.60.
Overall, these data suggest that (R)-3a has a secure network of polar
and nonpolar contacts and Y-shaped binding inside LBD that contribute
to its high binding energy and potent activity towards PPARδ protein.
2.3. ADMET and PK
We next evaluated the in vitro ADMET of compound 3a, including
cytochrome P450 (CYP) inhibition, liver microsomal metabolic stabi-
lity, parallel membrane permeability (PAMPA) and human ether-à-go-
go-related gene (hERG) assay (Table 3). 10 µM of compound 3a ex-
hibited 35.5% moderate inhibition only for CYP2C9, but did not mea-
surably inhibit the other CYP subtypes by 0–7.3%. Almost 100% of
remarkable metabolic stability in microsomes, including human mi-
crosomes, was also observed. When we used PAMPA to confirm the
permeability of 3a, excellent result of 9.54 × 10−6 cm/s was observed.
In hERG inhibition screening, compound 3a was also found to exhibit
the proper IC50 value with > 10 uM. These results indicated that
compound 3a did not show any significant problem in in vitro ADMET
profiling assays.
We also conducted in vivo PK studies of compound 3a in rats
(Table 4). Intravenous delivery of compound 3a showed that the
compound had low plasma clearance (CL = 4.0 mL/min/kg), a mod-
erate volume of distribution (Vss = 1.0 L/kg), and an extended half-life
(t1/2 = 3.6 h). Based on these results, compound 3a shows acceptable
bioavailability (F% = 32.6) for the following in vivo efficacy evalua-
tion.
(Continue on next page)
3

D.-S. Kim et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxx–xxx
Aesar (Ward Hill, MA, USA). Solvents were obtained from Fisher
Scientific (Hampton, NH, USA) or Aldrich and were used without fur-
ther purification unless noted otherwise.
a
Compounds were screened for agonist activity on PPARs in transiently
transfected CV-1 cells. EC50 value is the molar concentration of the test com-
pound that afforded 50% of the maximal reporter activity. bTransactivation
assay of all compounds shows 8–12-fold induction.
4.2. 4-(chloromethyl)-4′-(trifluoromethyl)-1,1′-biphenyl (5a)
To
a solution of [(4′-(trifluoromethyl)-[1,1′-biphenyl]-4-yl)me-
thanol (4a, 553 mg, 2.19 mmol) in anhydrous CH2Cl2 (20 mL) were
added SOCl2 (391 mg, 3.29 mmol) and DMF (32.1 mg, 0.44 mmol) at
0 °C. The reaction mixture was then allowed to warm to toom tem-
preature. After stirring the reaction mixture at room temperature for
12 h, the reaction mixture was diluted with water and extracted with
ethyl acetate for 3 times. The combined organic layer was dired over
MgSO4, filtered, and concentrated. The obtained residue was purified
by MPLC using hexane/ethyl acetate as the eluant to give the desired
product 5 as a white solid (553 mg, 93%). 1H NMR (400 MHz, CDCl3) δ
7.71–7.66 (m, 4H), 7.58 (d, 2H, J = 8.2 Hz), 7.50 (d, 2H, J = 8.2 Hz),
4.55 (s, 2H).
4.3. 2-methyl-4-(((4′-(trifluoromethyl)-[1,1′-biphenyl]-4-yl)methyl)thio)
phenol (6a)
To a stirred solution of 4-(chloromethyl)-4′-(trifluoromethyl)-1,1′-
biphenyl (5a) and 4-mercapto-2-methylphenol (448 mg, 3.19 mmol) in
CH3CN (10 mL) was added Cs2CO3 (1.30 g, 3.99 mmol). After stirring
the reaction mixture at room temperature for 12 h, the solvent was
evaporated from the mixture, and the mixture was poured into water
(30 mL) prior to extracting with ethyl acetate for 3 times. The combined
organic layer was then dried over anhydrous MgSO4, filtered, and
concentrated. The obtained residue was purified by MPLC using
hexane/ethyl acetate as the eluant to give the desired product 6a as a
white solid (714 mg, 72%). 1H NMR (400 MHz, CDCl3): δ 7.69–7.65 (m,
4H), 7.50 (d, 2H, J = 8.1 Hz), 7.28 (d, 2H, J = 8.08 Hz), 7.13 (s, 1H),
7.07 (d, 1H, J = 8.2 Hz), 6.73 (d, 1H, J = 8.2 Hz) 4.77 (s, 1H), 4.02 (s,
2H), 2.18 (s, 3H).
Scheme 2. Enantioselective synthesis of compound 3a. Reagents and condi-
tions: (a) methyl 2-bromoacetate, Cs2CO3, ACN, rt, 2 h; (b) 1 M LiOH, THF/
H2O, rt, 2 h.
3. Conclusion
In conclusion, we successfully developed a novel “Y”-shaped biaryl
PPARδ agonist 3a (EC50 = 2.6 nM) which has high potency and se-
lectivity. In addition, molecular docking studies between the PPARδ
ligand binding domain and the enantiospecific R and S isomers of
compound 3a suggested that the R1 phenyl moiety extends into a third
pocket of the PPARδ LBD, which is lacking in GW501516. Compound
3a was also shown to have considerable in vitro ADMET and in vivo PK
profiles for the subsequent development. These results clearly demon-
strate that our novel PPARδ agonist 3a is a promising drug candidate
possibly for the treatment of PPARδ-related disease. Further in-
vestigation of the effects of compound 3a on PPARδ-related diseases is
warranted.
4.4. tert-butyldimethyl(2-methyl-4-(((4′-(trifluoromethyl)-[1,1′-biphenyl]-
4-yl)methyl)thio)phenoxy)silane (7a)
To a stirred solution of 2-methyl-4-(((4′-(trifluoromethyl)-[1,1′-bi-
phenyl]-4-yl)methyl)thio)phenol (6a, 1.2 g, 3.20 mmol) in anhydrous
DMF (15 mL) were added imidazole (436 mg, 6.41 mmol) and tert-
butyl-dimethylsilyl chloride (531 mg, 3.53 mmol). The reaction mixture
was stirred at room temperature for 12 h. After completion of the re-
action, the reaction mixture was concentrated under reduced pressure
and diluted with ethyl acetate and distilled water. The combined or-
ganic layer was dried with MgSO4, filtered, and concentrated under
reduced pressure. The obtained residue was purified by MPLC using
hexane/ethyl acetate as the eluant to give the desired product 7a as a
white solid (714 mg, 72%). 1H NMR (400 MHz, CDCl3): δ 7.69–7.65 (m,
4H), 7.49 (d, 2H, J = 7.9 Hz), 7.29 (d, 2H, J = 7.9 Hz), 7.14 (s, 1H),
7.05 (s, 1H, J = 8.2 Hz), 6.67 (s, 1H, J = 8.2 Hz), 4.03 (s, 2H), 2.15 (s,
3H), 1.01 (s, 9H), 0.20 (s, 6H).
4. Experimental
4.1. General
All reactions were performed in oven- and flame-dried glassware.
Air- and moisture-sensitive reagents and solvents were transferred via
syringes or cannulae and were introduced into the reaction vessel
through a rubber septum. All experiments were carried out under a N2
atmosphere in dried glassware. 1H nuclear magnetic resonance (NMR,
300 and 400 MHz) and 13C NMR (75 and 100 MHz) spectra were re-
corded for solutions in CDCl3 and MeOD-d4. Chemical shifts (δ) are
expressed in parts per million (ppm) downfield from the internal tet-
ramethylsilane or with reference to residual CHCl3 in CDCl3. Mass
spectra were measured in positive electrospray ionization (ESI) mode
on LCMS-2020 system (Shimadzu, Tokyo, Japan). Column chromato-
graphy was performed using a CombiFlash Rf system with RediSep Rf
(Teledyne Isco, Lincoln, NE, USA). For the final compounds, further
purification was performed by preparative HPLC on Kinetex 5 μm bi-
phenyl, 100 Å (GX-281 HPLC system, Gilson, Middleton, WI, USA;
column tube 250 mm × 21.2 mm i.d.), with ACN/H2O as eluent. The
purity of the target compounds was determined to be > 95% by ana-
lytical HPLC using dual different wavelength UV detector. Starting
materials were obtained from Aldrich (St. Louis, MO, USA) or Alfa
4.5. tert-butyldimethyl(2-methyl-4-((2-phenyl-1-(4′-(trifluoromethyl)-
[1,1′-biphenyl]-4-yl)ethyl)thio)phenoxy)silane (8a)
A
stirred solution of tert-butyldimethyl(2-methyl-4-(((4′-(tri-
fluoromethyl)-[1,1′-biphenyl]-4-yl)methyl)thio)phenoxy)silane (7a,
978 mg, 2.00 mmol) in anhydrous THF (20 mL) was cooled to −78 °C in
dryice-acetone bath under N2 atmosphere. 2.0 M lithium diisopropyla-
mide (2.00 mL, 4.00 mmol) was added dropwise to reaction mixture at
−78 °C. The reaction mixture was stirred at −78 °C for 30 min.
Subsequently, benzyl bromide (411 mg, 2.40 mmol) was added to the
reaction mixture, dropwise at −78 °C. After the mixture was stirred at
same temperature for further 30 min, the reaction mixture was warmed
4

D.-S. Kim et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxx–xxx
Fig. 2. Binding orientation of (R)-3a, (S)-3a and GW501516 inside PPARδ ligand binding domain (LBD). (A) Binding superpose of (R)-3a (cyan), (S)-3a
(magenta) and GW501516 (blue) inside ligand binding domain (LBD) of PPAR delta protein. (B) The binding cavity is showed with gray color. Molecular docking
model of (R)-3a (C), (S)-3a (D) and GW501516 (E) in LBD. Protein represented as the orange helix. His449, His323, and Tyr473 amino acids involved in hydrogen
bond interactions highlighted as an orange stick. Hydrogen bonds showed in green dash format. Amino acids involved in hydrophobic contacts represented as an
orange line.
Table 2
Table 3
Docking energies.
In vitro absorption, distribution, metabolism, excretion, and toxicity (ADMET)
profiles of 3a.
Protein
Compounds
EC50 (nM)
-CDOCKER Energy
Binding Energy
CYP inhibition assay (% inhibition at
1A2
1.5
2C9
35.5
2C19
4.3
2D6
0
3A4
7.3
10 μM)a
hPPARδ
GW501516
(R)-3a
(S)-3a
2.0
0.7
6.1
53.03
50.34
45.80
−335.06
−343.60
−325.37
MS (% of remaining)b
Human
99.8
Dog
> 100
Rat
> 100
Mouse
> 100
PAMPA (cm/s)c
9.54 × 10−6
> 10
to room temperature. The mixture was quenched by adding sat. NH4Cl
solution. The mixture was diluted by ethyl acetate and the organic layer
was dried with MgSO4, filtered, and concentrated under reduced pres-
sure. The obtained residue was purified by MPLC using hexane/ethyl
acetate as the eluant to give the desired product 8a as an yellow oil
(649 mg, 56%). 1H NMR (400 MHz, CDCl3): δ 7.69–7.63 (m, 4H), 7.44
(d, 2H, J = 8.2 Hz), 7.23 (d, 2H, J = 8.2 Hz), 7.20 (d, 2H, J = 7.0 Hz),
7.17–7.12 (m, 1H), 7.06–6.99 (m, 4H), 6.61 (d, 1H, J = 8.2 Hz), 4.27
(dd, 1H, J = 9.0, 6.0 Hz), 3.29 (dd, 1H, J = 13.8, 6.0 Hz), 3.20 (dd, 1H,
J = 13.8, 9.0 Hz), 2.09 (s, 3H), 0.99 (s, 9H), 0.18 (s, 6H).
hERG, IC50 (μM)d
a
Cytochrome P450 (CYP450) enzyme activity was measured by incubating
CYP450 enzyme isotype-specific substrates and liver microsomes in the pre-
sence of the test compound for 15 min. Data are expressed as percent inhibitory
activity relative to the absence of compound.
b
MS = microsomal stability. The test compound was incubated with mi-
crosomes for 30 min and the percentage of the remaining compound was
measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS).
c
The effective permeability after 5 h was calculated as described in
Supporting Information. The measurements reported are the mean of four de-
terminations.
4.6. 2-methyl-4-((2-phenyl-1-(4′-(trifluoromethyl)-[1,1′-biphenyl]-4-yl)
ethyl)thio)phenol (9a)
d
Human ether-à-go-go-related gene (hERG) channel binding assay was
performed using predictor hERG fluorescence polarization assay; IC50 values
were calculated with non-linear regression fit with four parameters.
a,b For two three assays, the measurements reported are the means of dupli-
cations. The standard deviation is not shown as the two assays were found to be
reproducible on repeated testing.28,29
To a stirred solution of tert-butyldimethyl(2-methyl-4-((2-phenyl-1-
(4′-(trifluoromethyl)-[1,1′-biphenyl]-4-yl)ethyl)thio)phenoxy)silane
(8a, 649 mg, 1.12 mmol) in THF (20 mL) was added 1 M tetra-
butylammonium fluoride (2.80 mL, 2.80 mmol) at room temperature.
The reaction mixture was sirred for 1 h. After completion of the
5

D.-S. Kim et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxx–xxx
Table 4
In vivo rat pharmacokinetic (PK) parameters of compound 3a.a
iv (1 mg/kg)
AUC all
(h*ng/mL)
3833.5
CL (mL/min/kg)
4.0 1.7
Tmax (h)
4.5 1.0
Vss
(L/kg)
1.0
t1/2
(h)
3.6
1568.3
0.3
1.2
po (10 mg/kg)
a
Cmax
(ng/mL)
2325.0
AUC all
(h*ng/mL)
12,483.9
F(BA)
(%)
32.6
730.5
4776.5
Rats (n = 4) were dosed with 1 mg/kg iv and 10 mg/kg po; parameters were calculated from composite mean plasma concentration–time data. Data are
expressed as the means
SD. Formulation included ethanol 10%, dimethylsulfoxide (DMSO) 10%, Cremophor EL 10%, water 50%, and PEG400 30%. AUC = area
under the curve; CL = clearance; Vss = apparent volume of distribution; T1/2 = half life.
reaction, the reaction mixture was diluted with water and extracted
with ethyl acetate for 3 times. The combined organic layer was dired
over MgSO4, filtered, and concentrated. The obtained residue was
purified by MPLC using hexane/ethyl acetate as the enuant to give the
desired product 9a as an ivory solid (509 mg, 98%). 1H NMR (400 MHz,
CDCl3) δ 7.69–7.64 (m, 4H), 7.45 (d, 2H, J = 8.2 Hz), 7.23 (d, 2H,
J = 8.2 Hz), 7.21 (d, 2H, J = 6.9 Hz), 7.18–7.13 (m, 1H), 7.05–7.02 (m,
4H), 6.62 (d, 1H, J = 7.8 Hz), 4.83 (s, 1H), 4.27 (dd, 1H, J = 8.8,
6.1 Hz), 3.29 (dd, 1H, J = 13.8, 6.1 Hz), 3.20 (dd, 1H, J = 13.8,
9.0 Hz), 2.14 (s, 3H).
4.9. 2-(2-methyl-4-((2-phenyl-1-(6-(4-(trifluoromethyl)phenyl)pyridin-3-
yl)ethyl)thio)phenoxy)acetic acid (3b)
1H NMR (300 MHz, CDCl3): δ 8.15 (d, 1H, J = 1.9 Hz), 7.94 (d, 2H,
J = 8.2 Hz), 7.75 (m, 1H), 7.23–7.16 (m, 3H), 7.08 (d, 2H, J = 7.0 Hz),
7.02 (s, 1H), 6.98 (m, 1H), 6.54 (d, 1H, J = 8.4 Hz), 4.56 (s, 2H), 4.31
(m, 1H), 3.38 (m, 2H), 2.15 (s, 3H).
4.10. 2-(2-methyl-4-((1-(6-(4-(trifluoromethyl)phenyl)pyridin-3-yl)-2-
(3,4,5-trifluorophenyl)ethyl)thio)phenoxy)acetic acid (3c)
1H NMR (300 MHz, CDCl3): δ 8.12–8.08 (m, 3H), 7.96 (br s, 1H),
7.83–7.77 (m, 3H), 7.01 (s, 1H), 6.95 (d, 1H, J = 8.4 Hz), 6.75 (t, 2H,
J = 6.9 Hz), 6.60 (d, 1H, J = 8.3 Hz), 4.65 (s, 2H), 4.26 (m, 1H), 3.28
(m, 2H), 2.15 (s, 3H).
4.7. Methyl 2-(2-methyl-4-((2-phenyl-1-(4′-(trifluoromethyl)-[1,1′-
biphenyl]-4-yl)ethyl)thio)phenoxy)acetate (10a)
To a solution of 2-methyl-4-((2-phenyl-1-(4′-(trifluoromethyl)-[1,1′-
biphenyl]-4-yl)ethyl)thio)phenol (9a, 460 mg, 0.99 mmol) in CH3CN (7
mL) was added Cs2CO3 (968 mg, 2.97 mmol) followed by methyl bro-
moacetate (303 mg, 1.98 mmol). After stirring the reaction mixture at
room temperature for 12 h, it was diluted with water and extracted with
ethyl acetate for 3 times. The combined organic layer was dried over
MgSO4, filtered, and concentrated. The obtained residue was purified
by MPLC using hexane/ethyl acetate as the eluant to give the desired
product 10a as a colorless oil (487 mg, 92%); 1H NMR (400 MHz,
CDCl3): δ 7.69–7.64 (m, 4H), 7.45 (d, 2H, J = 8.2 Hz), 7.23 (d, 2H,
J = 8.2 Hz), 7.18 (d, 2H, J = 7.2 Hz), 7.17–7.13 (m, 1H), 7.07–7.02 (m,
4H), 6.53 (d, 1H, J = 9.0 Hz), 4.60 (s, 2H), 4.28 (dd, 1H, J = 8.8,
6.2 Hz), 3.77 (s, 3H), 3.28 (dd, 1H, J = 13.8, 6.1 Hz), 3.19 (dd, 1H,
J = 13.8, 8.8 Hz), 2.17 (s, 3H).
4.11. 2-(2-methyl-4-((2-phenyl-1-(3′-(trifluoromethyl)-[1,1′-biphenyl]-4-
yl)ethyl)thio)phenoxy)acetic acid (3d)
1H NMR (300 MHz, CDCl3): δ 7.79 (s, 1H), 7.72 (d, 1H, J = 7.6 Hz),
7.58–7.50 (m, 2H), 7.45 (d, 2H, J = 8.1 Hz), 7.25–7.13 (m, 5H),
7.08–7.00 (m, 4H), 6.56 (d, 1H, J = 9.0 Hz), 4.63 (s, 2H), 4.30 (m, 1H),
3.25 (m, 2H), 2.18 (s, 3H).
4.12. 2-(2-methyl-4-((1-(3′-(trifluoromethyl)-[1,1′-biphenyl]-4-yl)-2-
(3,4,5-trifluorophenyl)ethyl)thio)phenoxy)acetic acid (3e)
1H NMR (300 MHz, CDCl3): δ 7.80 (s, 1H), 7.73 (d, 1H, J = 7.6 Hz),
7.60–7.53 (m, 2H), 7.48 (d, 2H, J = 8.1 Hz), 7.21 (d, 2H, J = 8.2 Hz),
7.09 (s, 1H), 7.08 (s, 1H), 6.63 (t, 2H, J = 7.8 Hz), 6.59 (d, 1H,
J = 8.0 Hz), 4.56 (s, 2H), 4.21 (m, 1H), 3.15 (m, 2H), 2.19 (s, 3H).
4.8. 2-(2-methyl-4-((2-phenyl-1-(4′-(trifluoromethyl)-[1,1′-biphenyl]-4-
yl)ethyl)thio)phenoxy)acetic acid (3a)
4.13. 2-(4-((2-(2,6-difluorophenyl)-1-(3′-(trifluoromethyl)-[1,1′-
biphenyl]-4-yl)ethyl)thio)-2-methylphenoxy)acetic acid (3f)
To a stirred solution of methyl 2-(2-methyl-4-((2-phenyl-1-(4′-(tri-
fluoromethyl)-[1,1′-biphenyl]-4-yl)ethyl)thio)phenoxy)acetate
(10a,
1H NMR (300 MHz, CDCl3): δ 7.78 (s, 1H), 7.72 (d, 1H, J = 7.6 Hz),
7.58–7.51 (m, 2H), 7.45 (d, 2H, J = 8.2 Hz), 7.29 (d, 2H, J = 8.2 Hz),
7.11–7.07 (m, 3H), 6.76 (t, 2H, J = 7.7 Hz), 6.58 (d, 1H, J = 8.0 Hz),
4.64 (s, 2H), 4.45 (m, 1H), 3.30 (m, 2H), 2.19 (s, 3H).
34 mg, 0.070 mmol) in THF/H2O (15 mL/5 mL) at 0 °C was added
dropwise 1 M LiOH (2.46 mL, 2.46 mmol). The reaction mixture was
then stirred at room temperature for 15 min. After completion of the
reaction, the reaction mixture was acidified with 1 N HCl to give a pH of
5, and extracted with ethyl acetate for 3 times. The combined organic
layer was dried over MgSO4, filtered, and concentrated. The obtained
residue was purified by MPLC using dichloromethane/methanol as the
eluant to give the desired product 3a as a white solid (425 mg, 99%); 1H
NMR (400 MHz, MeOD-d4): δ 7.75 (d, 2H, J = 8.3 Hz), 7.70 (d, 2H,
J = 8.4 Hz), 7.50 (d, 2H, J = 8.2 Hz), 7.26 (d, 2H, J = 8.2 Hz),
7.18–7.10 (m, 4H), 7.08–7.05 (m, 3H), 6.68 (d, 1H, J = 8.4 Hz), 4.64
(s, 2H), 4.38 (dd, 1H, J = 9.2, 6.0 Hz), 3.27 (dd, 1H, J = 13.8, 6.0 Hz),
3.19 (dd, 1H, J = 13.7, 9.2 Hz), 2.16 (s, 3H); 13C NMR (100 MHz,
MeOD-d4): δ 171.1, 156.2, 144.4, 141.8, 138.8, 137.9, 136.3, 132.6,
128.9, 128.7 (q, J = 31.9 Hz), 128.6, 127.7, 127.3, 126.9, 126.4, 125.8,
125.3 125.2 (q, J = 3.6 Hz), 124.4 (q, J = 269.2 Hz) 111.1, 64.6, 55.4,
41.7, 14.8; HRMS (ESI): calcd: for [M−H]- C30H25F3O3S, 521.1404;
found, 521.1371.
4.14. 2-(4-((2-(2-fluoro-6-(trifluoromethyl)phenyl)-1-(3′-
(trifluoromethyl)-[1,1′-biphenyl]-4-yl)ethyl)thio)-2-methylphenoxy)acetic
acid (3g)
1H NMR (300 MHz, CDCl3): δ 7.79 (s, 1H), 7.72 (d, 1H, J = 7.6 Hz),
7.58–7.51 (m, 2H), 7.45–7.41 (m, 3H), 7.31–7.26 (m, 3H), 7.14–7.06
(m, 3H), 6.55 (d, 1H, J = 8.1 Hz), 4.68 (s, 2H), 4.48 (m, 1H), 3.45 (m,
2H), 2.17 (s, 3H).
4.15. 2-(4-((2-(2-chloro-6-fluorophenyl)-1-(3′-(trifluoromethyl)-[1,1′-
biphenyl]-4-yl)ethyl)thio)-2-methylphenoxy)acetic acid (3h)
1H NMR (300 MHz, CDCl3): δ 7.79 (s, 1H), 7.72 (d, 1H, J = 7.6 Hz),
6

D.-S. Kim et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxx–xxx
7.58–7.50 (m, 2H), 7.45 (d, 2H, J = 8.1 Hz)), 7.31 (m, 2H), 7.12–7.05
(m, 4H), 6.84 (m, 1H),6.56 (d, 1H, J = 9.0 Hz), 4.63 (s, 2H), 4.51 (m,
1H), 3.39 (m, 2H), 2.18 (s, 3H).
4.25. 2-(4-((2-(2-chloro-6-fluorophenyl)-1-(3′,4′,5′-trifluoro-[1,1′-
biphenyl]-4-yl)ethyl)thio)-2-methylphenoxy)acetic acid (3r)
1H NMR (300 MHz, CDCl3): δ 7.34 (d, 2H, J = 8.2 Hz), 7.29 (d, 2H,
J = 8.3 Hz), 7.16–7.06 (m, 6H), 6.84 (m, 1H), 6.56 (m, 1H), 4.68 (s,
2H), 4.50 (m, 1H), 3.38 (m, 2H), 2.18 (s, 3H).
4.16. 2-(4-((1-(4′-fluoro-[1,1′-biphenyl]-4-yl)-2-(3,4,5-trifluorophenyl)
ethyl)thio)-2-methylphenoxy)acetic acid (3i)
1H NMR (300 MHz, CDCl3): δ 7.53–7.49 (m, 2H), 7.43 (d, 2H,
J = 8.2 Hz), 7.18 (d, 2H, J = 8.2 Hz), 7.13–7.09 (m, 4H), 6.61 (m, 3H),
4.65 (s, 2H), 4.19 (m, 1H), 3.16 (m, 2H), 2.19 (s, 3H).
4.26. 2-(4-((2-(2,6-difluorophenyl)-1-(3′,4′,5′-trifluoro-[1,1′-biphenyl]-
4-yl)ethyl)thio)-2-methylphenoxy)acetic acid (3s)
1H NMR (300 MHz, CDCl3): δ 7.34 (d, 2H, J = 8.3 Hz), 7.28 (d, 2H,
J = 8.3 Hz), 7.15–7.06 (m, 5H), 6.76 (t, 2H, J = 7.8 Hz), 6.57 (m, 1H),
4.65 (s, 2H), 4.42 (m, 1H), 3.28 (m, 2H), 2.18 (s, 3H).
4.17. 2-(4-((2-(2-fluoro-6-(trifluoromethyl)phenyl)-1-(4′-fluoro-[1,1′-
biphenyl]-4-yl)ethyl)thio)-2-methylphenoxy)acetic acid (3j)
1H NMR (300 MHz, CDCl3): δ 7.52–7.49 (m, 2H), 7.42–7.38 (m,
3H), 7.29–7.25 (m, 2H), 7.13–7.06 (m, 5H), 6.55 (m, 1H), 4.63 (s, 2H),
4.47 (m, 1H), 3.44 (m, 2H), 2.17 (s, 3H).
4.27. 2-(4-((2-(2,6-dichlorophenyl)-1-(3′,4′,5′-trifluoro-[1,1′-biphenyl]-
4-yl)ethyl)thio)-2-methylphenoxy)acetic acid (3t)
1H NMR (300 MHz, CDCl3): δ 7.35–7.31 (m, 4H), 7.21 (d, 2H,
J = 8.0 Hz), 7.15–7.03 (m, 5H), 6.54 (d, 1H, J = 8.2 Hz), 6.57 (m, 1H),
4.63 (s, 2H), 4.59 (m, 1H), 3.54 (m, 2H), 2.17 (s, 3H).
4.18. 2-(4-((2-(2-chloro-6-fluorophenyl)-1-(4′-fluoro-[1,1′-biphenyl]-4-
yl)ethyl)thio)-2-methylphenoxy)acetic acid (3k)
1H NMR (300 MHz, CDCl3): δ 7.51–7.49 (m, 2H), 7.44 (d, 2H,
J = 8.0 Hz), 7.28 (d, 2H, J = 8.2 Hz), 7.16–7.05 (m, 6H), 6.84 (m, 1H),
6.55 (m, 1H), 4.65 (s, 2H), 4.51 (m, 1H), 4.39 (m, 2H), 2.17 (s, 3H).
4.28. Procedure for the chiral separation of racemic 9a
Separation condition: Two enantiomers of (R)-9a (45 mg) and (S)-9a
(45 mg) were isolated from racemic mixture of compound 9a (90 mg)
employing semiprep chiral HPLC column (chiralpak AD-H). Hexane and
isopropyl alcohol were used as a mobile phase in 90:10 ratio and 3 mL/
min flow rate was used.
4.19. 2-(4-((1-(4′-fluoro-[1,1′-biphenyl]-4-yl)-2-phenylethyl)thio)-2-
methylphenoxy)acetic acid (3l)
1H NMR (300 MHz, CDCl3): δ 7.52–7.49 (m, 2H), 7.39 (d, 2H,
J = 8.2 Hz), 7.21–7.08 (m, 9H), 7.03 (d, 2H, J = 7.0 Hz), 6.57 (d, 1H,
J = 8.2 Hz), 4.63 (s, 2H), 4.28 (m, 1H), 4.25 (m, 2H), 2.18 (s, 3H).
4.29. Procedure for the X-ray crystallography of (R)-9a
Crystallization condition: Camphanoyl group incorporated (R)-9a
was generated by reaction of (R)-9a with (1S)-(-)-camphanic chloride.
Camphanoyl group introduced (R)-9a was crystallized as needle type
under non-polor condition (Hexane/diethyl ether/ethyl acetate = 18/
3/1).
X-ray crystallography: A crystal of compound (R)-9a was coated with
epoxy glue in order to prevent spontaneous liberation of molecules
from the specimen under ambient conditions. The data collection was
carried out at 296 K on a Bruker SMART Apex II X-ray diffractometer
with a graphite-monochromated MoKα radiation (λ = 0.71073 Å). All
calculations in the structural solution and refinement were performed
using the Bruker SHELXTL crystallographic software package.30 The
structure was solved by the direct method31 and refined by the full-
matrix least-squares method followed by difference Fourier maps. All
the non-hydrogen atoms were refined anisotropically; all the hydrogen
atoms were put into calculated positions with the isotropic thermal
parameters.
4.20. 2-(4-((2-(2,6-difluorophenyl)-1-(4′-fluoro-[1,1′-biphenyl]-4-yl)
ethyl)thio)-2-methylphenoxy)acetic acid (3m)
1H NMR (300 MHz, CDCl3): δ 7.51–7.49 (m, 2H), 7.39 (d, 2H,
J = 8.2 Hz), 7.27–7.25 (m, 2H), 6.75 (t, 2H, J = 7.77 Hz), 6.58 (d, 1H,
J = 8.2 Hz), 4.64 (s, 2H), 4.44 (m, 1H), 3.29 (m, 2H), 2.18 (s, 3H).
4.21. 2-(4-((1-([1,1′-biphenyl]-4-yl)-2-(3,4,5-trifluorophenyl)ethyl)thio)-
2-methylphenoxy)acetic acid (3n)
1H NMR (300 MHz, CDCl3): δ 7.58–7.56 (m, 2H), 7.48–7.42 (m,
4H), 7.34 (m, 1H), 7.18 (d, 2H, J = 8.2 Hz), 7.09 (m, 2H), 6.62 (m, 3H),
4.65 (s, 2H), 4.20 (m, 1H), 3.16 (m, 2H), 2.19 (s, 3H).
4.22. 2-(4-((1-([1,1′-biphenyl]-4-yl)-2-(2-chloro-6-fluorophenyl)ethyl)
thio)-2-methylphenoxy)acetic acid (3o)
1H NMR (300 MHz, CDCl3): δ 7.56–7.54 (m, 2H), 7.45–7.39 (m,
4H), 7.33–7.28 (m, 3H), 7.09–7.04 (m, 4H), 6.85 (m, 1H), 6.54 (m, 1H),
4.61 (s, 2H), 4.49 (m, 1H), 3.40 (m, 2H), 2.17 (s, 3H).
Acknowledgments
This work was supported by the National Research Foundation of
Korea (NRF) funded by the Korea government (Ministry of Science &
ICT) (NRF-2017R1D1A1B03030196, NRF-2017R1C1B1005599, NRF-
2017M3A9G7073088 and 2015M1A5A1037572).
4.23. 2-(4-((1-([1,1′-biphenyl]-4-yl)-2-(2-fluoro-6-(trifluoromethyl)
phenyl)ethyl)thio)-2-methylphenoxy)acetic acid (3p)
1H NMR (300 MHz, CDCl3): δ 7.56 (d, 2H, J = 7.3 Hz), 7.46–7.40
(m, 5H), 7.33–7.26 (m, 3H), 7.14–7.04 (m, 3H), 6.55 (m, 1H), 4.61 (s,
2H), 4.48 (m, 1H), 3.44 (m, 2H), 2.17 (s, 3H).
A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at https://doi.org/10.1016/j.bmc.2018.06.044.
4.24. 2-(2-methyl-4-((2-phenyl-1-(3′,4′,5′-trifluoro-[1,1′-biphenyl]-4-yl)
ethyl)thio)phenoxy)acetic acid (3q)
References
1H NMR (300 MHz, CDCl3): δ 7.34 (d, 2H, J = 8.2 Hz), 7.21–7.13
(m, 7H), 7.07 (m, 2H), 7.02 (m, 2H), 6.56 (m, 1H), 4.65 (s, 2H), 4.27
(m, 1H), 3.25 (m, 2H), 2.18 (s, 3H).
1. Poulsen LLC, Siersbæk M, Mandrup S. Semin Cell Dev Biol. 2012;23:631–639.
2. Meinke PT, Wood HB, Szewczyk JW. Nuclear Hormone Receptor Modulators for the
Treatment of Diabetes and Dyslipidemia. In: Wood A, ed. Annu. Rep. Med. Chem.
7

D.-S. Kim et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxx–xxx
Academic Press; 2006:99–126.
17. Wang Y-X. Cell Res. 2010;20:124–137.
3. Issemann I, Green S. Nature. 1990;347:645–650.
4. Monsalve FA, Pyarasani RD, Delgado-Lopez F, Moore-Carrasco R. Mediat. Inflamm.
2013;2013:1–18.
18. Narkar VA, Downes M, Yu RT, et al. Cell. 2008;134:405–415.
19. Sznaidman ML, Haffner CD, Maloney PR, et al. Bioorg. Med. Chem. Lett.
2003;13:1517–1521.
5. Ahmadian M, Suh JM, Hah N, et al. Nat. Med. 2013;99:557–566.
6. Vernochet C, Peres SB, Davis KE, et al. Mol. Cell. Biol. 2009;29:4714–4728.
7. Tontonoz P, Spiegelman BM. Annu. Rev. Biochem. 2008;77:289–312.
8. He W, Barak Y, Hevener A, et al. Proc. Nat. Acad. Sci. USA. 2003;100:15,712–15,717.
9. Kersten S, Seydoux J, Peters JM, Gonzalez FJ, Desvergne B, Wahli W. J. Clin. Invest.
1999;103:1489–1498.
20. Willson TM, Brown PJ, Sternbach DD, Henke BR. J. Med. Chem. 2000;43:527–550.
21. Cox RL. Proc. Nat. Acad. Sci. USA. 2017;114:3284–3285.
22. Billin AN. Expert Opin. Inv. Drug. 2008;17:1465–1471.
23. Ham J, Hwang H, Kim E, et al. Eur. J. Med. Chem. 2012;53:190–202.
24. Wu C-C, Baiga TJ, Downes M, et al. Proc. Nat. Acad. Sci. USA.
2017;114:E2563–E2570.
10. Costet P, Legendre C, Moré J, Edgar A, Galtier P, Pineau T. J. Biol. Chem.
25. Lee S-J, Samala M, Woo SY, et al. ACS Omega. 2018;3:1970–1976.
26. Takada I, Yu RT, Xu HE, et al. Mol. Endocrinol. 2000;14:733–740.
27. Dassault Systèmes BIOVIA, Discovery Studio Modeling Envi-ronment, Release 4.
5(Version), San Diego: Dassault Systèmes, 2016.
1998;273:29,577–29,585.
11. Tyagi S, Gupta P, Saini A, Kaushal C, Sharma SJ. Adv. Pharm. Tech. Res.
2011;2:236–240.
12. Tan NS, Michalik L, Noy N, et al. Gene. Dev. 2001;15:3263–3277.
13. Peters JM, Lee SST, Li W, et al. Mol. Cell. Biol. 2000;20:5119–5128.
14. Krémarik-Bouillaud P, Schohn H, Dauça M. J. Chem. Neuroanat. 2000;19:225–232.
15. Takahashi S, Tanaka T, Kodama T, Sakai J. Pharmacol. Res. 2006;53:501–507.
16. Evans RM, Barish GD, Wang Y-X. Nat. Med. 2004;10:355–361.
28. Kim J, Chin J, Im CY, et al. Eur. J. Med. Chem. 2016;120:338–352.
29. Kim J, Woo SY, Im CY, et al. J. Med. Chem. 2016;59:10,209–10,227.
30. Bruker, SHELXTL Structure Analysis Software Package, Karlsruhe, Germany.
31. Sheldrick GM. Acta Cryst. 1990;A46:467.
8