Structure-guided design and synthesis of isoflavone analogs of GW4064 with potent lipid accumulation inhibitory activities

Article abstract of DOI:10.1016/j.bmcl.2018.10.021

Our group has previously reported a series of isoflavone derivatives with antidyslipidemic activity. With this background, a series of isoflavone analogs of GW4064 were designed, synthesized and evaluated the lipid-lowering activity of analogs. As a result, most of compounds significantly reduced the lipid accumulation in 3T3-L1 adipocytes and four of them (10a, 11, 15c and 15d) showed stronger inhibitory than GW4064. The most potent compound 15d exhibited promising agonistic activity for FXR in a cell-based luciferase reporter assay. Meanwhile, 15d up-regulated FXR, SHP and BSEP gene expression and down-regulated the mRNA expression of lipogenesis gene SREBP-1c. Besides, an improved safety profile of 15d was also observed in a HepG2 cytotoxicity assay compared with GW4064. The obtained biological results were further confirmed by a molecular docking study showing that 15d fitted well in the binding pocket of FXR and interacted with some key residues simultaneously.

Full text of DOI:10.1016/j.bmcl.2018.10.021

Accepted Manuscript
Structure-guided design and synthesis of isoflavone analogs of GW4064 with
potent lipid accumulation inhibitory activities
Rongmao Qiu, Guoshun Luo, Xuerong Cai, Linyi Liu, Mingqi Chen, Deying
Chen, Qidong You, Hua Xiang
PII:
S0960-894X(18)30823-0
https://doi.org/10.1016/j.bmcl.2018.10.021
BMCL 26080
DOI:
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
23 July 2018
1 October 2018
14 October 2018
Please cite this article as: Qiu, R., Luo, G., Cai, X., Liu, L., Chen, M., Chen, D., You, Q., Xiang, H., Structure-
guided design and synthesis of isoflavone analogs of GW4064 with potent lipid accumulation inhibitory activities,
Bioorganic & Medicinal Chemistry Letters (2018), doi: https://doi.org/10.1016/j.bmcl.2018.10.021
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Structure-guided design and synthesis of isoflavone analogs of GW4064 with
potent lipid accumulation inhibitory activities
Rongmao Qiua,b, Guoshun Luoa,b, Xuerong Caia,b, Linyi Liua,b, Mingqi Chenc, Deying
Chend, Qidong Youa,b, Hua Xianga,b*
aState Key Laboratory of Natural Medicines, Jiangsu Key Laboratory of Drug Design
and Optimization, China Pharmaceutical University, Nanjing 210009,China
bDepartment of Medicinal Chemistry, School of Pharmacy, China Pharmaceutical
University, Nanjing 210009, China
cLaboratory of Biology, School of Higher Vocational Education, China
Pharmaceutical University, 639 Longmian Avenue, Nanjing, 211198, PR China
dDepartment of Analytical Chemistry, School of Science, China Pharmaceutical
University, 639 Longmian Avenue, Nanjing, 211198, PR China
* Corresponding author. Tel.: +86 025 83271096; Fax: +86 025 83271096 (H.
Xiang).
E-mail addresses: xianghua@cpu.edu.cn (H. Xiang)
Abstract
Our group has previously reported a series of isoflavone derivatives with
antidyslipidemic activity. With this background, a series of isoflavone analogs of
GW4064 were designed, synthesized and evaluated the lipid-lowering activity of
analogs. As a result, most of compounds significantly reduced the lipid accumulation
in 3T3-L1 adipocytes and four of them (10a, 11, 15c and 15d) showed stronger
inhibitory than GW4064. The most potent compound 15d exhibited promising
agonistic activity for FXR in a cell-based luciferase reporter assay. Meanwhile, 15d
up-regulated FXR, SHP and BSEP gene expression and down-regulated the mRNA
expression of lipogenesis gene SREBP-1c. Besides, an improved safety profile of 15d
was also observed in a HepG2 cytotoxicity assay compared with GW4064. The
obtained biological results were further confirmed by a molecular docking study
showing that 15d fitted well in the binding pocket of FXR and interacted with some
key residues simultaneously.
Keywords: Dyslipidemia, GW4064, Isoflavone, Lipid accumulation, FXR agonist
Cardiovascular disease (CVD), which includes coronary heart disease, stroke and

peripheral vascular disease, is the leading cause of death and morbidity in the
developed countries. 1 As a disorder of plasma lipoprotein metabolism, dyslipidemia
plays a major role in the initiation and progression of CVD. The farnesoid X receptor
(FXR) is a member of the nuclear hormone receptor superfamily and senses the levels
of natural bile acids as its endogenous ligands.2-5 As a ligand activated transcription
factor, FXR is mainly expressed in the liver, intestine, kidney and adipose tissue.6
When physiologically activated by bile acids, FXR heterodimerizes with the retinoid
X receptor (RXR) and this dimer complex controls the expression of various target
genes that are involved in bile acid, cholesterol, triglyceride and lipoprotein
homeostasis in the liver and circulation.7-9 Activation of FXR by ligands results in
beneficial metabolic effects in various rodent models, such as glucose, triglyceride
and cholesterol lowering and insulin sensitization.10-12 Thus, FXR has been considered
as an important drug target for the treatment of dyslipidemia and other metabolic
syndrome after the deorphanization of FXR in 1999.8
Since then, intensive research on discovering FXR ligands has yielded numerous
synthetic steroidal and non-steroidal compounds.13 By introduction of an ethyl group
in the most potent FXR endogenous ligand chenodeoxycholic acid (CDCA), the first-
in-class FXR agonist OCA (Fig. 1A) was discovered and has recently received
approval for the treatment of primary biliary cirrhosis.14 Researchers at Glaxo-
SmithKline (GSK) previously reported a potent non-steroidal FXR agonist GW4064
(Fig. 1A) endowed with potential hypolipidemic activity in vivo.15 However, several
liabilities associated with GW4064 including high clearance, low bioavailability and a
potentially toxic stilbene pharmacophore limited its utility as a drug for
cardiovascular diseases.16 In spite of that, GW4064 has been successfully utilized as a
research tool and lead compound for exploring FXR mechanism of action and
pharmacology as well as identifying novel FXR agonists respectively. In an attempt to
improve the therapeutic profile of GW4064, it was recently reported by Eli Lilly that
a novel piperidinylisoxazole analog LY2562175 (Fig. 1A) possessed robust lipid
modulating properties and has been advanced to clinical evaluation in humans.17
On the other hand, isoflavones, the most important bioactive components of soy,

have been reported to have many biological activities including antiproliferative, anti-
inflammatory, anti-oxidant and antidyslipidemic effects.18-21 Particularly, the major
soy isoflavones, daidzein and genistein (Fig. 1B), exhibited lipid-lowering activity in
23
various mice models.22,
Moreover, several studies demonstrated that some
isoflavones like calycosin attenuate triglyceride accumulation and hepatic fibrosis via
activating FXR.24 Our group has been involving in the structure-modification of
isoflavones for a long time and has previously reported a series of antidyslipidemic
isoflavone derivatives.25, 26 The most potent compound XH601 (Fig. 1B) was found to
exhibit strong ability to improve the dyslipidemia in hamsters fed with high-fat diet
and is currently undergoing preclinical investigation. 27
Inspired by the promising research results mentioned above, a serious of novel
lipid-lowering compounds bearing isoflavone scaffold and isoxazole moiety were
designed, synthesized and evaluated for their lipid-lowering activity. In view of the
structure activity relationship of GW4064, we replaced the stilbene skeleton with
isoflavone scaffold aiming at identifying structurally distinct FXR agonists with
improved hypolipidemic activity and safety profile (Fig. 1C). The classical carboxylic
acid group was further derivatized as carboxyl bioisosteres like amide group to enrich
structure-activity relationship information.
A
COOH
O
O
COOH
O
O
N
N
N
Cl
Cl
Cl
Cl
COOH
Cl
HO
OH
N
H
R
R=H CDCA
R=Et OCA
GW4064
LY2562175
B
O
O
OH
OH
OH
HO
O
O
O
O
OH
O
O
O
HO
O
HO
O
XH601
Genistein
Calycosin
Daidzein
C
O
O
N
O
N
COOH
O
O
R
O
Cl
Cl
Cl
Cl
Cl
GW4064
target compounds

Fig.1. (A) Examples of FXR agonists reported in the literature. (B) Structures of some
antidyslipidemic isoflavones. (C) Design of the novel target compounds.
The preparation of designed compounds was depicted in Scheme 1. Initially, the
oxime 2 was formed by condensing hydroxylamine with the commercially available
aldehyde 1. Chlorination of 2 with NCS in DMF provided chloroxime 3 which was
subjected to cyclization with methyl isobutyrylacetate under basic conditions to yield
the isoxazole 4. After reduction of the ester 4 with DIBAL-H, the intermediate 6 was
obtained by chlorination of the resulting alcohol 5 with SOCl2. Resorcinol was
acylated by phenylacetic acids 7a-7d, followed by cyclization to give a series of
isoflavones 9a-9d.28 Then, target compounds 10a-10d were furnished from
isoflavones by nucleophilic substitution with intermediate 6. 10c was treated with
BBr3 to provide 11. Alkylation of phenol 11 with ethyl chloroacetate provided 12, and
then hydrolysis of the ester 12 afforded 13. A series of amides 15a-15d were
synthesized by condensation of different amines with 14, which was obtained by
hydrolysis of 10d. Details of the synthetic procedures and structural characterizations
were described in the supplementary material. The structures of all the target
1
13
compounds (10a-10d, 11-14, 15a-15d) were confirmed by H NMR, C NMR and
HRMS.
OH
N
OH
N
O
O
O
Cl
H
Cl
OH
N
CHO
N
N
CO2CH3
Cl
Cl
Cl
d
e
c
Cl
Cl
Cl
Cl
Cl
Cl
b
Cl
Cl
Cl
a
O
O
O
6
1
2
3
5
4
N
h
R
Cl
Cl
O
HO
O
HO
OH
10a-10d
HO
OH
HO
g
f
+
O
10a R = F
10b: R = NO2
10c: R = OCH3
10d: R = CO2CH3
:
O
R
O
R
R
9a - 9d
7a - 7d
8a - 8d
O
O
O
O
O
O
O
O
O
O
HO
O
O
O
O
O
k
N
N
N
j
i
N
O
O
O
Cl
Cl
Cl
Cl
OH
Cl
Cl
O
O
Cl
Cl
O
O
O
12
10c
11
13
O
O
O
O
O
O
O
15a
R
1 R2 = HCH3
15b: R1 R2 = HCH2CH3
15c: R1 2 = HCH(CH2)2
15d: R1 R2 = CH3CH3
:
l
N
N
O
m
N
O
O
O
O
Cl
Cl
Cl
Cl
O
O
Cl
Cl
R
O
OH
N
R2
R1
10d
14
15a-15d
Scheme1. Synthesis of 10a-10d, 11-14 and 15a-15d. Reagents and conditions: (a)
NH2OH·HCl, NaOH, EtOH, 90 oC; (b) NCS, DMF; (c) methyl isobutyrylacetate,

NaOCH3, CH3OH, THF; (d) DIBAL-H, THF; (e) SOCl2, DCM ; (f) BF3·Et2O, 80-
90°C; (g) BF3·Et2O, DMF, CH3SO2Cl; (h) K2CO3, KI, DMF, 60 oC; (i) BBr3, DCM,
RT; (j) ethyl chloroformate, K2CO3, KI, DMF, 60 oC; (k) Conc. HCl, 1,4-Dioxane,
100 oC; (l) Conc. HCl, 1,4-Dioxane, 100 oC; (m) EDCI, HOBt, HNR1R2, Et3N, DMF.
It is well-known that 3T3-L1 cell-based method is an efficacious tool for
screening lipid-lowering compounds. To examine cell viability, 3T3-L1 cells were
treated with various concentrations (5, 10, 20, and 40 μM) of target compounds and
GW4064 for 48 h. Then, the cell viability was measured using MTT assays. As shown
in Fig. 2, all of the tested compounds did not show significant cytotoxicity to the 3T3-
L1 cells even at 40 μM, except the phenolic hydroxyl containing compound 11. Some
of compounds, such as 15c, exhibited slight promotion on 3T3-L1 cells proliferation
at 5 μM. Due to of compounds did not obviously affect cell viability at 10 μM, the
nontoxic concentration of compounds selected for further biological investigation was
at 10 μM.
Fig.2. The effect of design compounds on cell viability of 3T3-L1 preadipocytes. The
cells were treated with various concentrations (5, 10, 20, and 40 μM) of synthesized
compounds and GW4064 for 48 h. The control group of cells was treated with 0.1%
DMSO. The results were expressed as the mean ± SD of three independent
experiments.
3T3-L1 preadipocytes were differentiated into mature adipocytes as previously
reported by our group.25, 26 To assess the effect of design compounds on lipid
accumulation in 3T3-L1 adipocytes, Oil Red O staining and its subsequent

quantification were performed to observe the intracellular lipid accumulation.29 The
biological evaluation results were depicted in Fig. 3A. All of the 12 design
compounds showed promising inhibitory effect on lipid accumulation in 3T3-L1
adipocytes at 10 μΜ. Among these compounds, four compounds 10a, 11, 15c and 15d
exhibited greater activities than GW4064. Compound 15d bearing N,N-
dimethylbenzamide group was identified as the best lipid accumulation inhibitor,
which reduced 37.7% lipid accumulation at 10 μΜ (**p < 0.01).
In order to further ascertain the inhibitory efficacy of compound 15d on lipid
accumulation, the mature 3T3-L1 adipocytes were treated with compound 15d at
different concentrations (10, 20, 40, and 80 μM). As shown in the Fig. 3B, compound
15d significantly reduced the intracellular lipid accumulation in a dosage-dependent
manner. It was worthy to note that at high concentration (80 μM), compound 15d
even reduced more than 50% lipid accumulation in 3T3-L1 adipocytes. The inhibition
of 15d on lipid accumulation was further confirmed by microscopic images in the Fig.
3C.
Fig.3. (A) Screening of design compounds’ inhibition on lipid accumulation in 3T3-
L1 adipocytes at 10 μM. (B) Absorbance of extracted Oil Red O at 492 nm in absence
and presence of different concentration of compound 15d. (C) Microscopic images of
3T3-L1 adipocytes stained with Oil Red O. The results were expressed as the mean ±
SD of two independent experiments (*p < 0.05, **p < 0.01 vs vehicle).

FXR agonists have proven their profound effects on plasma lipids in animal
models and have significant potential as therapeutic agents for the treatment of
dyslipidemia.15, 17 To test whether compound 15d is a FXR agonist, we performed a
cell-based transactivation assay using the Dual-Luciferase Reporter Assay System as
previously described.30 As seen in Fig. 4, compound 15d strongly activated the FXR
transactivity at 10 μM and showed all most the same potency as GW4064 did,
indicating its potent FXR agonistic activity.
Fig.4. Biological activity of compound 15d in FXR transactivation assay. Cells were
exposed to 10 μM of compound 15d and GW4064. The results were expressed as the
means ± SEM (n = 3).
Liver is an important organ which maintains lipid and glucose homeostasis in vivo.
Highly expressed in hepatocyte, FXR has been confirmed to regulate a series of target
genes, including short heterodimer partner (SHP), bile salt export pump (BSEP) and
sterol regulatory element-binding protein 1c (SREBP-1c). To further clarify the FXR
agonistic effect of 15d, we analyzed the expression levels of FXR and FXR
downstream genes in HepG2 cells as described previously.31 As shown in Fig. 5, 15d
treatment markedly induced the mRNA expression of FXR together with its target
genes SHP and BSEP, while down-regulated the expression of SREBP-1c. Results
confirmed that 15d was a truly FXR agonist and possessed potential lipid-lowering
activity through inhibition of expression of lipogenesis gene SREBP-1c.

Fig.5. Real-time PCR analysis of mRNA expression of FXR and FXR regulated genes
SHP, BSEP and SREBP-1c in HepG2 cells treated with compound 15d and GW4064
at 10 μM. Values are mean ± SEM (n = 3). (*p < 0.05, **p < 0.01, ***p < 0.001,
****p < 0.0001 vs vehicle).
It’s well known that several liabilities associated with GW4064 limited its utility
as a drug to treat FXR-mediated diseases. One of liabilities is that GW4064 contains a
potentially toxic stilbene pharmacophore.16 In order to examine the hepatotoxicity of
compound 15d, HepG2 cells viability was chosen for the studies. As shown in Fig 6,
15d did not cause significant cytotoxicity even at 80 µM. And interestingly, 15d
induced HepG2 cells proliferation below 20 µM. The results manifested that
compound 15d is more safety than GW4064.
Fig.6. The effect of 15d and GW4064 on HepG2 cells viability. The cells were treated
with various concentrations (5, 10, 20, 40 and 80 μM) of 15d and GW4064. The
control group of cells was treated with 0.1% DMSO. The results were expressed as

the mean ± SD of three independent experiments.
Analysis of the crystal structure of GW4064 suggested that hydrogen bonds formed
between the carbonyl acid group and Arg331 play important roles in binding to
FXR.32 To help rationalize the effect of compound 15d on FXR, we examined
computational molecular modeling simulations of compound 15d docked into the X-
ray structure of the FXR ligand binding domain (PDB ID: 3DCT). The docking study
was carried out using the Discovery Studio 3.0/CDOCKER protocol. Compound 15d
was optimized using Hyperchem v7.0. And the protein and ligand were optimized and
charged with CHARMm force field to perform docking. As shown in Fig. 7A, an
excellent superimposition of compound 15d over the structure of GW4064 was
observed in the binding pocket of FXR, which suggested that the designed
compounds exhibited the same binding patterns as that of GW4064. The docking
model in Fig. 7B showed that the isoxazole moiety appears to make two hydrogen
bonds with His447. And the isoflavone ring system provides conformational restraint
to the central linker section of the ligand and delivers the optimal geometry for the
amide substituent to establish the essential hydrogen bonds interaction with ARG331
and MET265.
Fig.7. (A) Superimposed poses of GW4064 (blue) with representative derivatives 15d
(yellow) binding to FXR. (B) Detailed interaction of 15d (yellow) with FXR. Red
dashed lines indicate potential hydrogen bonds to key residues.
FXR plays a critical role in the control of bile acid, lipid and glucose
metabolism, and has become a promising target for the treatment of dyslipidemia,
cholestatic disease and nonalcoholic steatohepatitis. Given the beneficial effects of
isoflavone, a serious of GW4064 analogs based on the isoflavone scaffold were

studied. Screening of all target compounds in Oil Red O staining assay identified
compound 15d as a potent inhibitor on lipid accumulation in adipocytes. Additionally,
15d exhibited transactivation activity and induced the expression of FXR, SHP and
BSEP, while down-regulating the expression of SREBP-1c. Although the effect of
15d in modulating FXR gene expression was less efficient than GW4064, 15d
showed less toxic in HepG2 cells. Overall our preliminary results indicated that 15d
could act as a promising lead compound for the design of novel FXR modulators
against dyslipidemia.
Supplementary data
Supplementary data associated with this article can be found in the online version
at
Acknowledgements
This work was supported by grants from the Natural Science Foundation of
China (NSFC, 81373279), the Twelfth Five-Year Plan Major Project of Candidate
Drugs (Ministry of National Science and Technology, 2012ZX09103101048) and
National Major Scientific and Technological Special Project for “Significant New
Drugs Development” (Ministry of National Science and Technology,
2018ZX09301042-004).
References and notes
1. Mozaffarian D, Benjamin EJ, Go AS, et al. Circulation. 2016, 133: 380.
2. Laffitte BA, Kast HR, Nguyen CM et al. J Biol. Chem. 2000, 275: 10638.
3. Wang H, Chen J, Hollister K, et al. Mol. Cell. 1999, 3: 543.
4. Parks DJ, Blanchard SG, Bledsoe RK, et al. Science. 1999, 284: 1365.
5. Makishima M, Okamoto A, Repa J, et al. Science.1999, 284: 1362.
6. Forman BM, Goode E, Chen J, et al. Cell. 1995, 81: 687.
7. Russell DW. Cell. 1999, 97: 539.
8. Cariou B, Staels B. Trends Pharmacol Sci. 2007, 28: 236.
9. Schaap FG, Trauner M, Jansen PL. Nat Rev Gastroenterol Hepatol. 2014, 11: 55.
10. Cariou B, van Harmelen K, Duran-Sandoval D, et al. J Biol Chem. 2006, 281:
11039.

11. Zhang Y, Lee FY, Barrera G, et al. Proc Natl Acad Sci. 2006, 103: 1006.
12. Ma K, Saha P K, Chan L, et al. J Clin Invest. 2006, 116: 1102.
13. Xu Y, J Med Chem. 2016, 59: 6553.
14. Pellicciari R, Fiorucci S, Camaioni E, et al. J Med Chem. 2002, 45: 3569.
15. Maloney PR, Parks D, Haffner CD, et al. J Med Chem. 2000, 43: 2971.
16. Sugihara K, Kitamura S, Sanoh S, et al. Toxicol Appl Pharmacol. 2000, 15: 46.
17. Genin MJ, Bueno AB, Agejas Francisco J, et al. J Med Chem. 2015, 58: 768.
18. Mukund V, Mukund D, Sharma V, et al. Crit Rev Oncol Hematol. 2017, 119: 13.
19. Szeja W, Grynkiewicz G, Rusin A. Curr Org Chem. 2017, 3: 218.
20. Gruca A, Krawczyk Z, Szeja W, et al. Molecules. 2014,19: 18558.
21. Choi MS, Jung UJ, Yeo J, et al. Diabetes Metab Res Rev. 2008, 24: 74.
22. Ae PS, Choi MS, Cho SY et al. Life Sciences. 2006, 79: 1207.
23. Wang S, Wang Y, Pan MH, et al. Food Funct. 2017, 11: 3831.
24. Duan X, Meng Q, Wang C, et al. Phytomedicine. 2017, 25: 83.
25. Zhao W, Sun J, Xiang H, et al. Bioorg Med Chem. 2011, 19: 3192.
26. Xiang H, Zhao W, Xiao H, et al. Bioorg Med Chem. 2010, 18: 3036.
27. Zhao MJ, Wang SS, Jiang Y, et al. Lipids Health Dis. 2017, 16: 85.
28. Wang WB, He Y, Xu P, et al. Bioorg Med Chem. 2015, 23: 4428.
29. Hao M, Li Y, Liu L, et al. Bioorg Med Chem. 2017, 25: 5506.
30. Kumar R, Ramachandran U, Raichur S, et al. J Med Chem. 2007, 42: 503.
31. Pellicciari R, Passeri D, De Franco F, et al. J Med Chem. 2016, 59: 9201.
32. Akwabi-Ameyaw A, Bass JY, Caldwell RD, et al. Bioorg Med Chem Lett. 2008,
18:4339.
Highlights:
 A series of isoflavone analogs of GW4064 have been designed and synthesized.
 The most potent compound 15d displayed lipid accumulation inhibitory activities through FXR.
 15d exhibited less toxicity than GW4064 in HepG2 cells.
 15d fitted well in the binding pocket of FXR.
Graphical Abstract