Discovery and Optimization of Non-bile Acid FXR Agonists as Preclinical Candidates for the Treatment of Nonalcoholic Steatohepatitis

Article abstract of DOI:10.1021/acs.jmedchem.0c01065

Farnesoid X receptor (FXR) plays a key role in bile acid homeostasis, inflammation, fibrosis, and metabolism of lipid and glucose and becomes a promising therapeutic target for nonalcoholic steatohepatitis (NASH) or other FXR-dependent diseases. The phase III trial results of obeticholic acid demonstrate that the FXR agonists emerge as a promising intervention in patients with NASH and fibrosis, but this bile acid-derived FXR agonist brings severe pruritus and an elevated risk of cardiovascular disease for patients. Herein, we reported our efforts in the discovery of a series of non-bile acid FXR agonists, and 36 compounds were designed and synthesized based on the structure-based drug design and structural optimization strategies. Particularly, compound 42 is a highly potent and selective FXR agonist, along with good pharmacokinetic profiles, high liver distribution, and preferable in vivo efficacy, indicating that it is a potential candidate for the treatment of NASH or other FXR-dependent diseases.

Full text of DOI:10.1021/acs.jmedchem.0c01065

Subscriber access provided by UCL Library Services
Article
Discovery and Optimization of Non-Bile Acid FXR Agonists as
Preclinical Candidates for the Treatment of Nonalcoholic Steatohepatitis
Junyou Li, Mengqi Liu, Yazhou Li, Dandan Sun, Zhihao Shu, Qian Tan, Shimeng
Guo, Rongrong Xie, Lixin Gao, Hongbo Ru, Yi Zang, Hong Liu, Jia Li, and Yu Zhou
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c01065 • Publication Date (Web): 29 Sep 2020
Downloaded from pubs.acs.org on October 1, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the
course of their duties.

Page 1 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery and Optimization of Non-Bile Acid FXR Agonists as
Preclinical Candidates for the Treatment of Nonalcoholic
Steatohepatitis
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Junyou Li†,‡, ∥ , Mengqi Liu†,‡, ∥ , Yazhou Li , Dan-dan Sun , Zhihao Shu , Qian Tan , Shimeng
§,†
†
†
†
§,†
†
†
†
†,‡
†,‡
†,‡,#,*
and Yu Zhou†,‡,*
Guo , Rongrong Xie , Lixin Gao , Hongbo Ru , Yi Zang , Hong Liu , Jia Li
†State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese Academy
of Sciences, 555 Zu Chong Zhi Road, Shanghai, 201203, China.
‡
University of Chinese Academy of Sciences, No.19A Yuquan Road, Beijing 100049, China.
§
School of Chinese Materia Medica, Nanjing University of Chinese Medicine, Nanjing 210046,
China
#Open Studio for Druggability Research of Marine Natural Products, Pilot National Laboratory for
Marine Science and Technology (Qingdao), 1 Wenhai Road, Aoshanwei, Jimo, Qingdao, 266237,
China
∥These authors contributed equally
*
corresponding authors
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ABSTRACT
FXR plays a key role in bile acid homeostasis, inflammation, fibrosis, metabolism of lipid and
glucose and becomes a promising therapeutic target for NASH or other FXR dependent diseases. The
phase III trial results of OCA demonstrate the FXR agonists emerge as a promising intervention in
patients with NASH and fibrosis, but this bile acid derived FXR agonist brings severe pruritus and
an elevated risk of cardiovascular disease for patients. Herein, we reported our efforts in the discovery
of a series of non-bile acid FXR agonists and 36 compounds were designed and synthesized based on
the structure-based drug design and structural optimization strategies. Particularly, compound 42 is a
highly potent and selective FXR agonist, along with good pharmacokinetics profiles, high liver
distribution and preferable in vivo efficacy, indicating it is a potential candidate for the treatment of
NASH or other FXR dependent diseases.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ACS Paragon Plus Environment

Page 3 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
1
. INTRODUCTION
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Nonalcoholic steatohepatitis (NASH), an inflammatory manifestation of nonalcoholic fatty liver
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
disease (NAFLD), has become a leading cause of liver transplantation and liver associated death,1
which is characterized by excessive lipid accumulation, hepatocellular injury, inflammation and
progressive fibrosis.2,3 A growing body of evidences demonstrate that patients histologically
diagnosed NASH, especially those with varying degrees of fibrosis, are at higher risk of cirrhosis and
4
liver related mortality. However, hitherto there is no approved pharmacologic option to target the
underlying pathophysiology, symptoms and complications of NASH. Thus, the new effective
interventions are urgently needed for the therapy of NASH.
The farnesoid X receptor (FXR) is a ligand-activated nuclear receptor transcription factor which
5,6
is highly expressed in the liver, kidney and intestine, and bile acids are its endogenous ligands. FXR
is a key metabolic regulator of bile acid homeostasis, inflammation, metabolism of lipid and glucose.7-
9
It engages a feedback loop that reduces hepatic bile acid biosynthesis via a small-heterodimer partner
1
0
(
SHP) induction, resulting in repression of cytochrome P450 CYP7A1. FXR also down-regulates
CYP7A1 by promoting the secretion of fibroblast growth factor 15 (FGF15) in mouse or FGF19 in
human which is secreted from the intestine and returns to the liver through enterohepatic circulation
1
1
to provide a feedback inhibition of bile acid synthesis. Specifically, activation of FXR induces SHP
expression thus suppressing the transcription of liver X receptor (LXR) induced sterol regulatory
element binding protein1c (SREBP-1c), which further inhibits the synthesis of fatty acids and
triglycerides.12-14 Moreover, activated FXR reduces the proliferation and activation of hepatic stellate
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
15
cells (HSCs) to protect liver from inflammation and fibrosis. These multifaceted roles of FXR in
bile acid metabolism make it become a promising therapeutic target for NASH. 16,17
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Recently, many FXR agonists have been developed to treat NASH in different clinical trials. As
a bile acid derived FXR agonist, OCA (1, Figure 1) emerged as a promising intervention in patients
with NASH and fibrosis and the results from the latest phase III trial confirmed that the fibrosis
improvement endpoint was achieved by 23% patients in 25 mg OCA group compared to 12% in
placebo group.17,18 However, more than half of patients treated with 25 mg OCA suffered from severe
1
6
pruritus (the most common adverse event), as well as dyslipidemia, which are probably due to the
over-expression of TGR5 in opioid dependent manner stimulated by OCA or its metabolites.19-21
Additionally, OCA has also been found to induce adverse effects against serum lipid profiles in
patients, such as elevating LDL cholesterol and modestly reducing HDL, which may bring an elevated
2
2
risk of cardiovascular disease. Therefore, researchers took more interest in developing new selective
non-bile acid FXR agonists (Figure 1). GW4064 (2) from Glaxo-SmithKline was firstly reported as
a non-bile acid FXR agonist and has been broadly utilized as a tool compound for scientists to explore
FXR mechanism of action and pharmacology, but its poor pharmacokinetics profiles (high clearance
and low bioavailability) and potential toxicities (a potential toxic stilbene pharmacophore and
stilbene-mediated UV light instability) preclude any further studies.23,24 As a result, several
conformationally optimized analogs were designed and were shown to be potent FXR agonists, such
as compound GSK2324 (3, 102% efficacy relative to GW4064) by replacing stilbene moiety with a
fused binary sketon, but it still lacked the necessary physicochemical properties and PK profiles for
further progress.24-29 To solve these problems, researchers at Gilead replaced the vinyl bond with a
ACS Paragon Plus Environment

Page 5 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
heterocycle and obtained GS-9674 (4) for the treatment of primary sclerosing cholangitis (PSC, Phase
III trial) and NASH (Phase II trial). The researchers from Eli Lilly replaced the potentially toxic
stilbene pharmacophore with a piperidine linker to obtain a partial agonist LY2562175 (5, also known
as TERN-101, 41% efficacy relative to GW4064) which has a great improvement on overall
druggability and was ultimately advanced to the clinic trial for evaluation in humans.30,31 Similarly,
after introducing a bicyclic tropane-substituted benzothiazole carboxylic acid moiety into a
trisubstituted isoxazole scaffold, researchers at Novartis discovered a full agonist LJN452 (6, 92%
efficacy relative to GW4064) that has highly potent in vivo activities in rodent models, and has been
advanced into phase II clinical trials in patients with NASH and primary biliary cholangitis (PBC).32
Although great progress in hand, hitherto no drug has been approved for the therapy of NASH by
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
FDA.
O
O
Cl
O
OH
O
COOH
O
H
N
N
Cl
Cl
Cl
Cl
H
N
HO
OH
COOH
H
1
(Obeticholic acid ,OCA)
2 (GW4064)
3 (GSK2324)
Cl
O
O
O
N
O
O
O
N
COOH
N
S
N
N
COOH
N
COOH
HO
Cl
Cl
Cl
Cl
F3CO
N
N
N
F
4
(GS-9674)
5 (LY2562175, TERN-101)
6 (LJN452)
Figure 1. Representative FXR agonists
To address these unmet clinical needs, we initiated our efforts to discover safer and more
effective FXR agonists for the treatment of NASH or other FXR dependent diseases. We have mainly
focused on the development of novel non-bile acid FXR agonists to improve the efficacy, safety and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
tolerability of the current molecules. Herein, a series of novel non-bile acid FXR agonists that feature
fluorinated linker ring have been designed, synthesized and tested for their biological efficacy on the
basis of the structure-based drug design strategy and structural optimization efforts from the important
scaffold cores of lead compounds GW4064 and LY2562175. To our pleasure, we found several
compounds have potent activities and ideal efficacies on FXR. Specifically, compound 42 has high
potency (EC = 8.7 nM) and more desirable efficacy (80.4% relative to GW4064) on FXR with no
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
5
0
activation on TGR5, indicating it may not have severe pruritus side effect as OCA. Futhermore, it has
good pharmacokinetics and excellent liver distribution profiles, simultaneously it also shows better
effects in several preclinical NASH models than OCA and LY2562175, which demonstrates that it
shows the potential to be a good candidate for further explorations.
2
2
. RESULTS AND DISCUSSION
.1 Rational Design of Novel Non-Bile Acid FXR Agonists
To help rationalize the design of non-bile acid FXR agonists, we built a computational molecular
mode from two key lead compounds GW4064 (2) and LY2562175 (5) to explore which regions of
the molecule could be modulated by medicinal chemistry SAR efforts. The docking model between
GW4064 and LY2562175 in Figure 2a showed that the isopropyl from GW4064 or cyclopropyl
group from LY2562175 fills the hydrophobic pocket formed by side chains of F284, L287, F461 and
W454, while the substituted phenyl rings establish hydrophobic contacts with M328, F329, M365
and I352. The aryl carboxylic acid moieties efficiently fill the narrow portion of the hydrophobic
pocket and deliver the optimal geometry for the carboxylic acid substituent to form an important
electrostatic interaction with R331 residue. Besides, a weak H-bond between the isoxazole nitrogen
ACS Paragon Plus Environment

Page 7 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
and H447 was also detected, in which the N-N distance is about 3.6 Å and 3.3 Å for the two lead
compounds. The binding model indicated that except for keeping the important interactions with the
hydrophobic pocket around isoxazole scaffold and the important electrostatic interaction with R331
residue from carboxylic acid moiety, prolonging the length of molecule to enhance H-bond
interaction between the isoxazole nitrogen and H447 maybe bring an improvement in potency and
selectivity of FXR. Recently, selectively fluorinating for bioactive compounds has become a pratical
and efficient strategy in the design and discovery of new drugs to improve their ADME properties,
potencies and physicochemical characteristics.35-40 Therefore, we have attempted to introduce a
fluorine atom and simultaneously prolonged a carbon atom distance for the piperidine ring linker of
lead compound LY2562175, expecting the new scaffold can better fill the whole binding pocket of
FXR. Based on the above, 36 new compounds (7-42) were synthesized and their biological activities
on FXR were evaluated (Tables 1-3).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(a)
(b)
Figure 2. (a) Model of GW4064 (gray) and LY2562175 (cycan) docked into FXR ligand binding domain from X-
ray structure of PDB 3DCT; (b) the binding mode of compound 42. The figures were generated using PyMol
(http://www.pymol.org/).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
.2 Chemistry
The synthesis of key intermediates 45a and 45b is described in Scheme 1. Independently
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
epoxidation of the starting materials 43a and 43b by Corey-Chaykovsky reaction gave the
intermediate 44a and 44b, followed by a nucleophilic addition reaction with HF in pyridine solution
to conveniently afford 45a and 45b, respectively. The synthesis of halogenated aromatic esters
fragments is described in Scheme 2. 49a-b were synthesized from commercially available material
4
6 by nucleophilic substitution reaction. A Friedel-Craft reaction of material 47 and followed by an
esterification led to intermediate 49c. Cyclization of commercially available aldehyde 48 got
intermediate 49d. Intermediates 52a-b were prepared in two steps operations from commercially
available materials 50a-b, which were treated with NaSCN and bromine using Hugerschoff reaction
to afford 51a-b, followed by a Sandmeyer reaction using tert-butyl nitrite to yield intermediates 52a
and 52b, respectively.
The synthesis of compounds 7-38 is described in Scheme 3. Independently treating the
commercially available aldehydes 53a-h with hydroxylamine hydrocloride afford oxime
intermediates 54a-h, followed by a chlorination with NCS to give 55a-h, respectively. Next,
cycloaddition of 55a-h with methyl 3-cyclopropyl-3-oxopropanoate can prepare the desired isoxazole
scaffolds 56a-h, which are reduced by LiAlH to give the corresponding intermediates 57a-h,
4
respectively. Further bromination lead to intermediates 58a-h, which undergo an alkylation with 45a
to result in the intermediates 59a-h, respectively. Thereafter, deprotection of 59a-h yields the key
amines 60a-h, which go through a cross-coupling reaction with corresponding halogenated aromatic
esters and further hydrolysis reaction to afford the desired compounds 7-38. The synthesis of
ACS Paragon Plus Environment

Page 9 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
compounds 39-42 is described in Scheme 4. Alkylation of 58f with 45b results in the intermediate
5
9i, followed by deprotection to yield the key amine 60i, which undergoes a cross-coupling reaction
with corresponding halogenated aromatic esters and further hydrolysis reaction to afford the desired
compounds 39-42, respectively.
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1. Synthesis of Intermediates 45a and 45ba
OH
a
b
O
N
Boc
N Boc
O
N
Boc
F
43a
44a
45a
a
b
F
O
N
Boc
O
N
Boc
HO
N
Boc
43b
44b
45b
a
t
o
Reagents and conditions: (a) Me SI, BuOK, DMSO, rt, 30 min,90-95%; (b) 70% HF in pyridine, CH Cl , -15 C
3
2
2
to rt, 4 h, 60-65%.
Scheme 2. Synthesis of Intermediates 49a-d, 52a-ba
S
Br
47
d
O
O
OH O
O
O
R
e
c
N
Br
Br
X
H
Br
48
46
49a: X=N-methyl, R=CH3
4
4
4
9b: X=N-ethyl, R=CH3
9c: X=S, R=CH3
9d: X=O, R=CH CH3
2
O
O
R1
S
g
S
N
f
O
O
H N
2
H N
Br
2
N
OMe
R1
R1
O
50a-b
51a-b
52a: R =H
1
5
2b: R =F
1
a
o
Reagents and conditions: (c) NaH, DMF, CH I or CH CH I, -10 C to rt, 1 h, 95-99%; (d) AlCl , CH Cl , Oxalyl
3
3
2
3
2
2
o
chloride, CH OH, -25 C to rt, 45%; (e) HBF -Et O, CH Cl , ethyl diazoacetate, H SO , rt, 2 h, 72%; (f) NaSCN,
3
4
2
2
2
2
4
t
Bromine, AcOH, rt, 24 h, 68-75%; (g) BuONO, CuBr , CH CN, rt, 24 h, 70-75%.
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 3. Synthesis of Compounds 7-38a
OH
N
OH
N
O
O
O
N
OH
Cl
COOMe
N
b
d
a
c
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
R2
R2
54a-h
R2
55a-h
R2
R2
53a-h
56a-h
57a-h
O
N
Boc
O
N
R1
NH
O
N
O
N
N
Br
N
O
g
f
h,i
O
O
F
F
Cl
F
Cl
-9,13,17,18
COOH
R2
R2
6
6
6
6
0a: R =2,6-dichloro
0b: R =2-F,6-Cl
0c: R =2-F
0d: R =2-Cl
60e: R =2-CF3
2
0f: R =2-OCF3
R2
2
7
59a-h
58a-h
2
2
X
COOH
2
R1 =
6
6
2
2
0g: R =3-OCF3
7
9
: X=C
: X=N
8
60h: R =4-OCF3
2
COOH
COOH
10: R =2,6-dichloro, X=N-methyl
2
4: R =2,6-dichloro, X=N-ethyl
COOH
X
1
2
1
7:X=N-ethyl
j,i
h,i
15: R =2,6-dichloro, X=O
2
18:X=S
16: R =2,6-dichloro, X=S
2
N
N
NH
19: R =2-F,6-Cl, X=N-methyl
2
2
0: R =2-F, X=N-methyl
2
1
3
2
1: R =2-Cl, X=N-methyl
2
COOH 22: R =2-CF , X=N-methyl
2
3
11: R =2,6-dichloro, R =H
2
3
23: R =2-OCF , X=N-methy
2 3
S
O
12: R =2,6-dichloro, R =F
2
3
24: R =3-OCF , X=N-methyll
2 3
25: R =4-OCF , X=N-methyl
2 3
N
N
32: R2=2-F,6-Cl, R3=F
O
N
N
X
O
R3
3
3
3
3: R =2-F, R =F
4: R =2-Cl, R =F
5: R =2-CF , R =F
2 3
O
26: R2=2-F,6-Cl, X=N-ethyl
27: R =2-OCF , X=N-ethyl
F
2
3
F
2
3
2
3
3
28: R =2-F, X=S
2
R2
36: R =2-OCF , R =F
2
3
3
29: R =2-CF X=S
11,12,32-38
1
0,14-16,19-31
2
3,
3
7: R =3-OCF , R =F
R2
2
3
3
30
: R2=2-OCF3, X=S
3
8: R =4-OCF , R =F
2
3
3
31: R =3-OCF X=S
2
3,
a
o
o
Reagents and conditions: (a) NH OH·HCl, NaOH, EtOH/H O, 0 to 60 C, 2 h, 95-99%; (b) NCS, DMF, 40 C, 2
2
2
o
h, 90-98%; (c) TEA ,EtOH, Methyl 3-cyclopropyl-3-oxopropanoate, 0 C to rt, 12 h, 40%-55%; (d) LiAlH , THF,
4
o
o
-
10 C to rt, 0.5 h, 86-95%; (e) CBr , PPh , THF, 0 C to rt, 2 h, 85-90%; (f) 45a, 18-crown-6, potassium tert-
4 3
o
butoxide, THF, rt,12 h, 65-80%; (g) TFA, DCM, 0 C to rt, 1 h, 90-95%; (h) corresponding halogenated aromatic
o
o
esters, Pd dba , X-Phos, Cs CO , toluene, 110 C,10 h, 50-65%; (i) 2N KOH, THF, CH OH, 60 C, 6 h, 75-87%;
2
3
2
3
3
o
(j) corresponding halogenated aromatic esters, DIPEA, NMP, 60 C, 5 h, 85-90%.
ACS Paragon Plus Environment

Page 11 of 89
Journal of Medicinal Chemistry
1
2
3
4
Scheme 4. Synthesis of Compounds 39-42a
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
N
F3CO
COOH
O
N
F
X
O
O
h,i
39: X=N-methyl
40: X=N-ethyl
41: X=S
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Br
N
O
N
N
F3CO
f
g
F3CO
O
F3CO
O
NH
N
Boc
O
F
j,i
N
F
COOH
F3CO
S
N
O
6
0i
58f
59i
N
F
F
42
a
o
Reagents and conditions: (f) 45b, 18-crown-6, potassium tert-butoxide, THF, rt, 12 h, 70%; (g) TFA, DCM, 0 C
o
to rt, 1 h, 90%; (h) corresponding halogenated aromatic esters, Pd dba , X-Phos, Cs CO , toluene, 110 C,10 h, 56-
2
3
2
3
o
o
61%; (i) 2N KOH, THF, CH OH, 60 C, 6 h, 79-84%; (j) 52b, DIPEA, NMP, 60 C, 5 h, 90%.
3
2
.3 FXR Activity Evaluations for All Compounds and Their SAR Discussion
Two assays were designed to identify FXR bioactivity for different compounds, thereby ensuring
the concentration for 50% of maximal effect (EC ) values of FXR. Firstly, an in vitro ligand-induced
5
0
FXR coactivator interaction assay was based on AlphaScreen technology that measures the binding
ability of compounds to FXR for mediating the recruitment of its transcriptional co-activator using
human recombinant FXR ligand binding domain (LBD) protein and biotinylated co-activator peptide,
the steroid receptor coactivator-1 (SRC1). The second method relating to a FXR-luciferase reporter
assay was constructed using the human FXR encoding plasmid and the FXRE-luciferase reporter
plasmid co-transfected HEK293T cells to measure the compound modulating transcriptional activity
of FXR. After forming a heterodimer with retinoid X receptor (RXR) by ligand mediating co-activator
recruitment, FXR will bind to the cis-acting element on FXRE-luciferase plasmid, leading to the
downstream expression of a luciferase reporter. Accordingly, the FXR transcriptional activity can be
confirmed by quantification of luciferase fluorescence intensity. The reported FXR agonists OCA
(
1), GW4064 (2) and LY2652175 (5) were selected as the positive controls, and the results were
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
shown in Tables 1-3.
After introducing a fluorine atom and simultaneously prolonging a carbon atom distance for the
piperidine ring linker of lead molecule, we have firstly screened different aryl carboxylic acid
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
moieties on the R position to identify which aryl carboxylic acid moiety is the best geometry for this
1
binding pocket to form the important electrostatic interaction with R331 and seven compounds (7-
1
3) were synthesized. As shown in Table 1, introducing benzonic acid (7), naphthoic acid (8) or
picolinic acid (9) is harmful to the agonistic activity of FXR whatever in molecular level or cellular
level assay, displaying weaker potencies compare to the lead compounds GW4064 and LY2652175,
but stronger than OCA. Further explorations showed that the indolecarboxylic acid moiety (10) turned
out to be a much better choice for this molecule scaffold. Compounds 11-13 showed weak activities,
probably because the aryl carboxylic acid moiety can’t fill the binding pocket well to form a proper
electrostatic interaction with R331. Inspired by these results, we have further introduced some
bioisosteres of indolecarboxylic acid scaffold, including N-substituted indolecarboxylic acid (14),
benzofuran-3-carboxylic acid (15) and benzothiophene-3-carboxylic acid (16), and excellent
activities could be afforded as expected. To investigate the importance of the position of carboxylic
acid, compounds 17 and 18 with 2-carboxylic acid were synthesized separately and the results showed
indole-2-carboxylic acid could also offer high potency on FXR in molecular and cellular level
evaluations.
ACS Paragon Plus Environment

Page 13 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
Table 1. Modification of the R Group and Their Agonistic Activities on FXRa
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
O
N
R1
N
Cl
O
F
Cl
Molecular level
Cellular level
Compd.
7
R1
EC （µM）
EC （µM）
5
0
50
COOH
1.97±0.05
1.10±0.40
3.71±0.53
0.18±0.01
0.10±0.01
0.44±0.04
COOH
8
9
N
COOH
COOH
1
0
0.22±0.04
4.29±1.09
9.55±1.14
2.81±0.36
0.40±0.08
0.0058±0.0004
0.33±0.03
N
COOH
COOH
S
N
11
S
N
1
1
1
2
3
4
0.74±0.20
F
COOH
COOH
0.91±0.11
N NH
0.06±0.006
N
COOH
15
3.10±0.92
0.020±0.002
O
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
COOH
1
6
2.89±0.51
0.096±0.007
S
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
COOH
COOH
17
18
0.32±0.06
1.95±0.21
0.0043±0.0006
0.14±0.01
N
S
OCA (1)
-
0.290±0.01
/b
0.71±0.06
0.012±0.001
0.004±0.0003
GW4064 (2)
LY2562175 (5)
-
-
0.11±0.02
aMean values of three independent experiments; EC50 values are shown as mean ± SD. Control assays were
b
performed with OCA, GW4064 and LY2562175. Untested
a
Table 2. Modification of the R and R Group and Their Agonistic Activities on FXR
1
2
R1
O
N
N
O
F
R2
Molecular level
Cellular level
Compd.
R1
R2
EC （µM）
EC （µM）
5
0
50
COOH
F
Cl
1
9
0.42±0.14
0.0052±0.0005
0.065±0.009
0.037±0.005
N
COOH
F
20
1.09±0.12
0.80±0.08
N
COOH
Cl
2
1
N
ACS Paragon Plus Environment

Page 15 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
COOH
F3C
2
2
2
3
0.42±0.002
0.49±0.15
1.73±0.16
3.39±0.20
0.28±0.05
1.11±0.06
6.23±0.02
2.04±0.60
2.12±0.49
3.91±0.09
0.035±0.003
0.0013±0.0002
0.77±0.03
N
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
COOH
F3CO
N
COOH
24
N
F3CO
COOH
2
2
2
2
5
6
7
8
16.77b
N
OCF3
COOH
F
Cl
0.0076±0.0007
0.0015±0.0002
0.23±0.01
N
COOH
F3CO
N
COOH
F
S
COOH
F3C
29
0.047±0.005
0.0010±0.0002
1.24±0.13
S
COOH
F3CO
3
3
0
1
S
COOH
S
F3CO
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
COOH
COOH
COOH
COOH
COOH
COOH
COOH
S
N
F
Cl
3
3
2
3
0.36±0.07
4.67±0.12
5.36±0.16
3.98±0.13
1.72±0.55
7.21±0.09
12.19±0.19
0.011±0.002
0.27±0.01
0.14±0.72
0.18±0.01
0.0028±0.0003
2.73±1.63
44.66b
F
F
F
F
F
F
F
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
S
N
F
S
N
Cl
34
S
N
F3C
3
3
3
3
5
6
7
8
S
N
F3CO
S
N
F3CO
S
N
OCF3
aMean values of three independent experiments; EC50 values are shown as mean ± SD. Control assays were
b
performed with OCA , GW4064 and LY2562175. Tested in one independent experiment due to weak agonism
activity.
With these important findings obtained, we have kept indolecarboxylic acid moiety at R position
1
and subsequently varied R substituents to explore which aryl groups could offer more proper
2
hydrophobic contacts with M328, F329, M365 and I352 (as compounds 19-27, Table 2). Among
them, most compounds have excellent biological activities on FXR, especially compounds 19, 23, 26
ACS Paragon Plus Environment

Page 17 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
and 27. This four compounds offer a comparable or even better potency than lead compound
LY2562175, and is far superior to lead compound OCA and GW4064, in which the results indicated
that introducing a fluorine, trifluoromethyl or trifluoromethoxy group into 2-position of phenyl group
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
of R is beneficial to the improvement of potency of FXR. Hence, we have attempted to replace
2
indolecarboxylic acid at R position of compounds 20, 22, 23 and 24 with benzothiophene-3-
1
carboxylic acid to give compounds 28-31 and the results demonstrated that 2-trifluoromethoxyphenyl
group at R position is the best choice. Thereafter, the benzothiazole carboxylic acid, a
2
pharmacophore from phase II FXR agonist LJN452, was further introduced to design compounds 32-
38, in which 2-trifluoromethoxyphenyl group at R position also showed the best biological activitiy,
2
and the EC of compound 36 reaches 0.0028 ± 0.0003 µM at the cellular level evaluations.
5
0
During our studies, the scientists from Novartis reported a new non-bile acid FXR agonist
LJN452 (6) with a bicyclic [3.2.1] tropane linker as the key component responsible for the exceptional
32
potency. We thus kept the optimal substituents at R and R position on the basis of foregoing SAR
1
2
investigations and then replaced the fluorine-substituted piperidine linker by the fluorine-substituted
bicyclic [3.2.1] tropane linker to design compounds 39-42. As we expected, all four compounds
displayed excellent biological efficacies on both molecular and cellular FXR assays (Table 3).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 3. Optimization of the R Group and Their Agonistic Activities on FXRa
1
O
N
R1
N
O
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
F3CO
F
Molecular level
Cellular level
Compd.
R1
EC （µM）
EC （µM）
5
0
50
COOH
3
4
4
4
9
0
1
2
0.78±0.09
0.56±0.12
0.85±0.19
2.52±0.33
0.0004±0.00007
0.0005±0.00002
0.0007±0.00009
0.0087±0.0023
N
COOH
N
COOH
S
COOH
S
N
F
aMean values of three independent experiments; EC50 values are shown as mean ± SD. Control assays were
performed with OCA, GW4064 and LY2562175.
3
4
It was reported that full activation of FXR may lead to some side effects. Because metabolic
diseases (such as NASH) usually require a stable long-term therapy, thus full activation on FXR is
not necessary for beneficial effects, indicating moderate to high efficacy may be the best choice for
3
3
the treatment of NASH. Therefore, we selected parts of active compounds (10, 12, 14, 17, 18, 19,
26, 32 and 42) and examined their efficacies relative to GW4064. The results showed their efficacies
range from 60.73% to 89.98% (Table 4), indicating these compounds are moderate to high potent
partial FXR agonists, which may offer better regulatory effects for FXR-related diseases, as reported
ACS Paragon Plus Environment

Page 19 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
the efficacy is too low to work well, and full activation of FXR may lead to various side effects.33,34
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 4. Parts of Compounds and Their Efficacy Relative to GW4064
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compd.
10
Cellular level EC (µM)
Efficacy relative to GW4064 (%)
71.42±3.95
5
0
0.0058±0.0004
0.74±0.20
1
1
1
1
2
4
7
8
77.24±6.58
0.06±0.006
89.98±4.49
0.0043±0.0006
0.14±0.01
60.73±1.11
79.08±3.84
19
0.0052±0.0005
0.0076±0.0007
0.011±0.002
0.0087±0.0023
70.00±4.28
2
3
4
6
2
2
62.39±2.15
72.00±6.45
80.43±8.00
2
.4 In Vivo Liver Distribution of Compounds
Compounds with excellent target organs distribution commonly have improved
pharmacodynamics on target and reduced whole body side effects.41,42 In view of high expression of
FXR in the liver, we have attempted to investigate the selectivity of compounds against its target
organs, especially the distribution of compounds in liver. Thus, seven compounds and lead compound
LY2562175 (5) were selected to determine their concentration in plasma and liver at 1 h and 4 h after
oral administration at the dosage of 3 mg/kg in SD rats, respectively. The results summarized in Table
5
3
showed that compared to the control lead compound LY2562175 (5), most compounds (10, 26, 32,
9 and 42) displayed more excellent drug distribution in liver except compounds 17 and 19.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table 5. Mean Plasma and Liver Concentrations of Compounds After 3 mg/kg Oral Administration
in SD ratsa
Plasma
Liver
Ratio：
liver/plasma
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compd.
10b
Time (h)
conc. (ng/mL) conc. (ng/mL)
1
4
1
4
1
4
1
4
1
4
1
4
1
4
1
4
1.7±0.1
91.8±1.8
56.6±3.3
29.7±13.1
11.7±3.7
13.8±3.3
6.8±2.9
9.0±0.3
269.5±125.26
72.0±24.1
6.3±1.6
17c
19c
12.6±5.2
185.8±117.6
87.2±60.2
12.4± 4.1
58.5±33.3
24.0±13.0
49.0±32.0
240.8±98.1
94.8±31.6
12.7±1.3
368.4±89.2
656.3±127.4
772.0±219.2
1552.0±443.3
928.9±156.9
130±19.9
8.2±5.0
32.1±12.9
21.9±13.8
6.7±1.1
2
3
3
4
6c
2c
9b
2d
10.1±1.6
10.3±0.5
14.7±2.7
2.2±0.7
26.4±2.3
384.5±36.1
2342.6±392.9
612.3±313.9
121±18.4
1136.6±485.8
165.1±115.0
19.3±1.2
4.0±0.9
6.3±0.5
LY2562175 (5)b
52.1±23.5
153±18.4
13.28±0.7
a
b
Male SD rats. p.o.: 3 mg/kg − DMSO/Tween 80/0.5% MC (1/2/97, v/v/v). 1h: n=2, 4 h: n=2 animals per group.
c1 h: n=3, 4 h: n=3 animals per group. 1 h: n=3, 4 h: n=2 animals per group.
d
More particularly, compound 10 displayed the highest liver distribution ratio (the concentration
ratio of liver and plasma is 56.6 ± 3.3 at 1 h and 29.7 ± 13.1 at 4 h, respectively) and compound 42
had the optimal drug concentration in liver with 2342.6 ± 392.9 ng/mL and 612.3 ± 313.9 ng/mL at
1
h and 4 h after oral administration of 3 mg/kg of compounds in SD rats, respectively. The high
ACS Paragon Plus Environment

Page 21 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
liver/plasma ratio of 10 showed heavy accumulation in the liver, along with active FXR-binding
indicating substantial expression of FXR-targeted genes and presumably facilitating the subsequent
function of the pathway on inhibiting steatosis, inflammation, fibrosis in the liver. For similar reasons,
the high liver and plasma concentration of compound 42 may bring a good in vivo efficacy.
Table 6. The Assay of FXR Selectivitiesa
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compd.
TGR5-EC (μM)
LXR-EC (μM)
RXR-EC (μM)
5
0
50
50
1
4
0
2
>100
>200
>200
>200
>200
>200
-
>100
>200
LY2562175 (5)
OCA (1)
1.497±0.256
>200
0.0237±0.005
>200
INT-777b
0.0958±0.014
-
T0901317c
-
-
0.201±0.016
-
-
9
-cis-Retinoic acidd
0.0469±0.002
a
b
c
Mean values of three independent experiments; EC values are shown as mean ± SD. control of TGR5. control
5
0
d
of LXR. control of RXR.
2.5 In Vitro FXR Selectivity Study of Compounds
Recent years, it has become clearer that FXR is appropriately classified as a nuclear bile acid
receptor.5,43 FXR forms a heterodimer with the RXR to modulate expression of target genes in the
same way of LXR which is the first isotype found in the liver,44,45 and the FXR–RXR heterodimer
can also respond to ligands for RXR. One of the complicated factors in studying FXR biology is that
bile acids, such as chenodeoxycholic acid (CDCA) and cholic acid, are also ligands for pregnane X
receptor (PXR), constitutive androstane receptor (CAR; also known as NR1I3), vitamin D receptor
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
46
(
VDR) and the G protein-coupled receptor TGR5 (also known as GPBAR1). In light of the excellent
results of molecular and cellular activity of these compounds on FXR and target organs distribution,
two representative compounds 10 and 42 were chosen for the extensive preclinical profiling across
nuclear receptor and GPCR assays. Results demonstrated the high selectivity of compounds 10 and
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
2 for FXR with no significant off-target activity (> 1000-fold selectivity for FXR, Table 6). We also
conducted a cAMP biological assay of our FXR agonists on bile acid activated GPCR and TGR5
Table 6). Significantly, both compounds 10 and 42 showed no measurable activity on TGR5 (> 100
(
μM), offering a high selectivity for FXR, whereas the control lead compound 5 displayed weak
activation of TGR5 (EC = 1.494 μM). This results showed potential advantages of these compounds
5
0
originated from non-bile acid FXR agonists, because several investigations proposed that the
activation of TGR5 may be the reason for pruritus caused by bile acids.19,20
2
.6 In Vivo Pharmacokinetic Study of Compounds
It is instructive to study compounds equipped with good FXR activity and promising liver
targeting ability on their pharmacokinetic profiles in SD rats. Thus, we have investigated the
pharmacokinetic profiles for several compounds, such as compounds 10, 19, 26 and 42. The results
indicated that all four selected compounds had good pharmacokinetic profiles compared to the lead
compound LY2562175 (5, Table 7). Among them, compounds 10, 19 and 26 displayed higher drug
exposure than that of LY2562175 (5) after I.V. administration, but maybe higher clearance of
compound 10 or relative worse absorbtion of compounds 19 and 26 resulted in much lower
bioavailability in SD rats after oral administration. However, as shown in Table 7, compound 42
showed the best pharmacokinetic profiles in SD rats, to be specific, compound 42 had a larger drug
ACS Paragon Plus Environment

Page 23 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
exposure (1930 vs 201 h·ng/mL) and slower plasma clearance rate (8.54 vs 82.6 mL/min/kg) than
LY2562175 (5) after I.V. administration. For the P.O. administration, compound 42 displayed
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
larger exposure (AUC0−t = 1863 ± 82 ng/L*h) and a nearly ten-fold long terminal half-life (t = 20.4
1/2
h) compared to LY2562175 (5). The high oral bioavailability of 42 (F = 32.2%) in rats indicates that
it has appropriate gastrointestinal permeability. On account of high permeability, high drug exposure
in plasma and the low clearance, compound 42 was considered to have attractive drug-like properties
for further in vivo biological investigations. Therefore, considering the results of liver distribution,
selectivity of FXR and pharmacokinetic profiles, compounds 10 and 42 were selected for further
investigation to explore their therapeutic potential against NASH or other FXR dependent diseases.
Table 7. Mean PK Parameters of Selected Compounds in SD Ratsa
1
0b
19c
26c
42b
LY2562175 (5)c
0.25±0
Tmax (h), p.o.
0.25±0
117±37
0.25±0
0.25±0
117±45
0.38±0.18
590±236
Cmax (ng/mL), p.o.
t (h), p.o.
47.0±11.9
283±194
2.22±0.27
195±47
1.84±0.24 2.31±1.87 0.98±0.42 20.4±11.7
1
/2
AUC (h·ng/mL), p.o.
105±25
190±27
130±55
404±53
103±15
611±25
1863±82
0
-t
AUC (h·ng/mL), i.v.
1930±335
201±21
0
-t
MRT0-∞ (h), i.v.
0.39±0.04 0.59±0.13 0.51±0.07 1.70±0.44
88.2±12.9 41.0±4.69 26.4±0.73 8.54±1.77
2.05±0.44 1.46±0.43 0.81±0.13 0.84±0.10
0.35±0.03
82.6±8.51
1.74±0.30
32.4
CL (mL/min/kg), i.v.
V (L/kg)
ss
F (%)
18.4
10.7
5.64
32.2
aMale SD rats. iv: 1 mg/kg − DMSO/EtOH/PEG300/NaCl (2.5/7.5/40/50, v/v/v/v). p.o.: 3 mg/kg − DMSO/0.5%
b
c
HPMC (2.5/97.5, v/v). p.o.: n=2, iv: n=3 animals per group. p.o.: n=3, iv: n=3 animals per group.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2
.7 Molecular Modelling Study of Compounds
To study the structural basis of the potent FXR activity attained through the introduction of longer
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
distance fluorine-substituted piperidine or tropane ring linkers, we have initiated a molecular
modelling study for compound 42 docked into the reported X-ray structure of FXR agonist GW4064
binding domain (PDB 3DCT). The docking model is shown in Figure 2b, and ligand binding domain
is in well accordance with that of the published X-ray co-crystal structure of GW4064. Besides, the
cyclopropyl group on the isoxazole scaffold of compound 42 can also fill the hydrophobic pocket
formed by side chains of F284, L287, F461 and W454, and the substituted phenyl moieties can
establish hydrophobic contacts with M328, F329, M365 and I352. The aryl carboxylic acid moieties
also efficiently fill the narrow portion of the hydrophobic pocket, and deliver the optimal geometry
for the carboxylic acid substituent to form an important electrostatic interaction with R331 residue.
Meanwhile, a much stronger H-bond between the isoxazole nitrogen and H447 than that of GW4064
can be detected, in which the N-N distance is about 3.1 Å (vs 3.6 Å for GW4064). Moreover, an
additional π-π interaction between isoxazole ring and TRP 454 was detected, and introduction of
fluorine atom might form a new fluorine-π interaction between fluorine atom and 2-trifluorophenyl
group of compound 42. These important interactions maybe provide an important and necessary
guarantee for the remarkable boost in potency on FXR.
2
.8 In Vivo Anti-Fibrotic Effects of Compound 10 on TAA-Induced Rats
To further investigate the potential of NASH treatment, we evaluated the abilities of selected
compounds in improving hepatic injury, fibrosis and other pathophysiologic features of NASH in
three distinct rodent models. Considering the multifactorial and multistage features of
ACS Paragon Plus Environment

Page 25 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
NAFLD/NASH, it is challenging for any single preclinical model to imitate the full spectrum of
human NASH pathophysiology.47
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
In chemical induced murine model, hepatotoxins are sufficiently concentrated in the liver and
4
8
cause liver injury with necrosis while prolonging treatment results in fibrosis. Used for decades of
years, chemical model is thought to be stable and effective for investigating late-stage NASH
endpoint fibrosis in relatively short time. Therefore, the first model in our studies is a chemical-
induced rat model (thioacetamide, TAA) in which liver fibrosis developed by intraperitoneal injection
of TAA three times a week for 8 weeks. Compound 10 (20 mg/kg) was intragastric administrated
daily from week 4 to 8. Meanwhile, we used OCA (1) and LY2562175 (5) as positive controls. The
results displayed that control OCA (1) significantly caused compensatory liver hypertrophy in this
model, while treatment with compound 10 showed no obvious influence on liver to body ratio
compared to vehicle group (Figure 3A). All of treating groups had no obvious effect on body weight
(Figure S1A). Serum ALP was reduced substantially in rats treated with compound 10 (Figure 3B),
indicating remission of liver injury which was likewise achieved by OCA (1) and LY2562175 (5).
The hematoxylin & eosin (H&E) and sirius red stained sections of liver revealed that typical bridging
fibrosis and moderate vacuolar degeneration (ballooning) occurred in TAA induced rat model
(Figure 3C). Rats treated with compound 10 showed statistically significant reduction of the fibrotic
areas (Figure 3C-D) in comparison with those of vehicle group, and OCA (1) exhibited reduced sirius
red positive areas similarly. Hepatic mRNA levels of profibrotic genes Col1a1 were dramatically
decreased in compound 10 and LY2562175 (5) treating groups (Figure 3E) and that levels of α-sma
were significantly decreased in rats treated with control LY2562175 (5) and OCA (1, Figure S1B).
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Compound 10, LY2562175 (5) and OCA (1) treating groups all showed tendency to decrease content
of hepatic hydroxyproline (Figure S1C). These findings indicated that compound 10 could
significantly reduce liver enzymes (ALP) and ameliorate liver fibrosis at the dose of 20 mg/kg in
TAA model to a mildly larger extent than the pharmacological effects displayed by OCA (1) and
LY2562175 (5).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Compound 10 ameliorates liver injury and fibrosis in TAA-induced rat model. (A) liver weight, liver per
body weight ratio. (B) serum markers ALP. (C) representative images of Sirius Red staining (upper) and H&E
staining (lower) of liver (control n = 4, others n= 8 mice per group). Scale bars, 400 μm. (D) area of collagen
quantified by Sirius Red staining. (E) quantitative PCR of profibrogenic markers Col1a1 in liver tissue.
Measurements were taken from distinct samples. Data are presented as the mean ± SEM; n = 8 male C57/BL6J
mice per group, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001 compound versus vehicle (#p ≤ 0.05, ##p
ACS Paragon Plus Environment

Page 27 of 89
Journal of Medicinal Chemistry
1
2
3
4
≤
0.01, ###p ≤ 0.001, and ####p ≤ 0.0001 control versus vehicle) using one-way ANOVA with Dunnett’s test. Con,
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
C57BL/6J mice control; Veh, TAA model. TAA, thioacetamide; ALP, alkaline phosphatase.
2.9 In Vivo Efficacy of Compounds 10 and 42 Against Fibrosis on CCl -Induced Mice
4
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
On account of the efficacious improvement of hepatic fibrosis and injury brought by compound
1
0 and positive control OCA (1) in TAA-induced rat model, we have further evaluated their effects
in a carbon tetrachloride (CCl ) induced mice model by a head to head comparative trial. This model
4
has become one of the most commonly used experimental model for inducing toxin-mediated liver
fibrosis and produces highly reproducible results. Mice suffered from liver fibrosis by intraperitoneal
injection of CCl three times a week for 2 weeks, followed by compound treatment for 6 weeks (8
4
weeks totally). Consequently, the membrane permeability in mitochondria, endoplasmic reticulum,
and plasma membrane was lowered, and generalized hepatic damage occured. Serum samples were
collected at 3 and 6 weeks after intragastric administrated compound 10 and positive control OCA
(1, 6 mg/kg, 20 mg/kg and 60 mg/kg). Markers of liver damage aspartate aminotransferase (AST),
alanine aminotransferase (ALT) were substantially elevated in CCl injection group (Vehicle)
4
compared to control group. Mice treated with compound 10 for 3 weeks showed a dose dependent
tendency of reduction in ALT and AST relative to vehicle. After 6 weeks treatment, compound 10
(60 mg/kg) could significantly reduce AST to the similar degree of 20 mg/kg OCA treating group
(Figure 4A-B), reflecting liver injury recovery. The percentage of sirius red-positive areas within
liver sections was significantly lower in 10 and OCA (1) treated mice in a dose-dependent way
relative to vehicle mice (Figure 4C&E), demonstrating the amelioration of perisinusoidal and
portal/periportal fibrosis. Mice treated with 10 (20 mg/kg) also displayed a reduction on hepatic
hydroxyproline content (Figure 4D), a marker of collagen deposition, which was equivalent to the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 89
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
clinincal drug OCA (1).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4. Compound 10 dose-dependently improves liver injury and fibrosis in CCl -induced mouse model. (A)
4
serum markers of hepatic injury AST and (B) ALT (control n = 9, others n = 10 male C57/BL6J mice per group).
(C) Sirius Red staining of liver (control n = 4, others n = 10 male C57/BL6J mice per group). Scale bars, 200 μm.
(D) hepatic hydroxyproline content (control n = 7, others n = 10 male C57/BL6J mice per group). (E) area of
collagen quantified by Sirius Red staining. Data are presented as the mean ± SEM; *p ≤ 0.05, **p ≤ 0.01, ***p ≤
0.001, and ****p ≤ 0.0001 compound versus vehicle (#p ≤ 0.05, ##p ≤ 0.01, ###p ≤ 0.001, and ####p ≤ 0.0001
control versus vehicle) using one-way ANOVA with Dunnett’s test. Con, C57BL/6J mice control; Veh, CCl model.
4
CCl , Carbon tetrachloride; AST, aspartate transaminase; ALT, alanine transaminase.
4
ACS Paragon Plus Environment

Page 29 of 89
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Compound 42 attenuates liver injury and fibrosis in CCl -induced mouse model. (A) serum ALT and (B)
4
AST. (C) Sirius Red staining of liver. Scale bars, 200 μm. (D) hepatic hydroxyproline content. (E) area of collagen
quantified by Sirius Red staining. Data are presented as the mean ± SEM; vehicle n = 8, 42 (2mg/kg) n=10, others
n=9 male C57/BL6J mice per group, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, and ****p ≤ 0.0001 compound versus
vehicle (#p ≤ 0.05, ##p ≤ 0.01, ###p ≤ 0.001, and ####p ≤ 0.0001 control versus vehicle) using one-way ANOVA
with Dunnett’s test.
Encouraged by the promising liver distribution (Table 5) and pharmacokinetic profiles (Table
7) of compound 42, we have employed the CCl -induced mouse model to evaluated its anti-fibrotic
4
and anti-damage effects. Very importantly, treatments with compound 42 at a dose of 20 mg/kg for
6
weeks markedly reduced serum AST and ALT levels (Figure 5A-B), indicating the improvement
ACS Paragon Plus Environment