Mechanistic Insight on the Mode of Action of Colletoic Acid

Article abstract of DOI:10.1021/acs.jmedchem.9b00187

The natural product colletoic acid (CA) is a selective inhibitor of 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1), which primarily converts cortisone to the active glucocorticoid (GC) cortisol. Here, CA's mode of action and its potential as a chemical tool to study intracellular GC signaling in adipogenesis are disclosed. 11β-HSD1 biochemical studies of CA indicated that its functional groups at C-1, C-4, and C-9 were important for enzymatic activity; an X-ray crystal structure of 11β-HSD1 bound to CA at 2.6 ? resolution revealed the nature of those interactions, namely, a close-fitting and favorable interactions between the constrained CA spirocycle and the catalytic triad of 11β-HSD1. Structure-activity relationship studies culminated in the development of a superior CA analogue with improved target engagement. Furthermore, we demonstrate that CA selectively inhibits preadipocyte differentiation through 11β-HSD1 inhibition, suppressing other relevant key drivers of adipogenesis (i.e., PPARγ, PGC-1α), presumably by negatively modulating the glucocorticoid signaling pathway. The combined findings provide an in-depth evaluation of the mode of action of CA and its potential as a tool compound to study adipose tissue and its implications in metabolic syndrome.

Full text of DOI:10.1021/acs.jmedchem.9b00187

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES
Article
Mechanistic insight on the mode of action of colletoic acid
Taotao Ling, Darcie J Miller, Walter Lang, Elizabeth Griffith, Adaris Rodríguez-Cortés, Ikbale El
Ayachi, Gustavo Palacios, Jaeki Min, Gustavo A. Miranda-Carboni, Richard E Lee, and Fatima Rivas
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.9b00187 • Publication Date (Web): 11 Jul 2019
Downloaded from pubs.acs.org on July 11, 2019
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 56
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Mechanistic Insight on the Mode of Action of Colletoic Acid
‡
Ϯ
‡
‡
‡
Taotao Ling, Darcie J. Miller, Walter H. Lang, Elizabeth Griffith, Adaris Rodriguez-Cortes, Ikbale
†
‡
‡
†
‡
El Ayachi, Gustavo Palacios, Jaeki Min, Gustavo Miranda-Carboni, Richard E. Lee and Fatima
Rivas*,‡
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
‡
Ϯ
Department of Chemical Biology and Therapeutics, Department of Structural Biology, St. Jude
Children’s Research Hospital. 262 Danny Thomas Place. Memphis, TN 38105-3678,
†
USA. Department of Medicine, The University of Tennessee Health Science Center, Memphis,
Tennessee 38163 USA.
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 56
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
ABSTRACT. The natural product colletoic acid (CA) is a selective inhibitor of 11β-hydroxysteroid
dehydrogenase type 1 (11β-HSD1), which primarily converts cortisone to the active glucocorticoid (GC)
cortisol. Here, CA’s mode of action and its potential as a chemical tool to study intracellular GC
signaling in adipogenesis is disclosed. 11β-HSD1 biochemical studies of CA indicated its functional
groups at C-1, C-4, and C-9 were important for enzymatic activity; an X-ray crystal structure of 11β-
HSD1 bound to CA at 2.6 Å resolution revealed the nature of those interactions, namely a close-fitting
and favorable interactions between the constrained CA spirocycle, and the catalytic triad of 11β-HSD1.
Structure-activity relationship studies culminated in the development of a superior CA analogue with
improved target engagement. Furthermore, we demonstrate that CA selectively inhibits preadipocyte
differentiation through 11β-HSD1 inhibition, suppressing other relevant key drivers of adipogenesis (i.e.
PPARγ, PGC-1α), presumably by negatively modulating the glucocorticoid signaling pathway. The
combined findings provide an in-depth evaluation of the mode of action of CA and its potential as a tool
compound to study adipose tissue and its implications in metabolic syndrome.
ACS Paragon Plus Environment
2

Page 3 of 56
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
INTRODUCTION
Intracellular cortisol levels are regulated by 11β-hydroxysteroid dehydrogenases 1 and 2 (11β-HSD),
while circulating cortisol is controlled by the hypothalamic-pituitary-adrenal axis.1-3 The present study
focused on studying the correlation of the natural product (+)-colletoic acid (CA, 1, Figure 1) with the
bidirectional metabolic enzyme 11β-HSD1, which primarily activates local glucocorticoid (GC) action
by converting cortisone to cortisol. 11β-HSD1 is highly expressed in metabolic tissue (i.e. adipose,
liver),4,5 and it can also exert context dependent regulation of cellular action in other tissue (muscle,
5
osteoblast, etc.). In humans, 11β-HSD1 expression is increased in acquired central obesity and
positively correlates to accumulation of subcutaneous and intra-abdominal fat, body mass index, and
insulin resistance, symptoms associated with metabolic syndrome (MetS) and type 2 diabetes mellitus
(
T2DM).6-9 Inhibition of intracellular cortisol production by blocking 11β-HSD1 ameliorates the risk
6
factors associated with MetS, T2DM, and fatty liver in preclinical studies. Therefore, several medicinal
chemistry campaigns have been launched to identify small heterocycle molecules to selectively inhibit
1
1β-HSD1.10-14 However, the limited molecular complexity presented by heterocycles and the modest
clinical performance have promoted the search for natural product inhibitors with rich structure
1
5
complexity. Despite the potential therapeutic relevance of this biological target, few natural products
have been reported as potent and selective inhibitors of this enzyme (3-4, Figure 1).16,17 For example,
carbenoxolone (CBX, 4) is a potent 11β-HSD1 inhibitor, but its poor isoform selectivity limits its use.17
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 56
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
Thus, natural products that can selectively inhibit 11β-HSD1 offer a unique opportunity to study
intracellular GC signaling. Selective and tracktable 11β-HSD1 inhibitors can help improve our
understanding of adipogenesis and its role in metabolic homeostasis, which is essential to develop new
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
therapeutic avenues for the effective treatment of MetS, TDM2 and related disorders. Previous studies
have demonstrated that CA is a potent 11β-HSD1 inhibitor with no observable 11β-HSD2 inhibition. 18,
1
9
CA is a member of the acorane natural products family, featuring a synthetically challenging all-
carbon spirocenter. A convergent and enantioselective total synthesis of CA and its first-generation
2
0
analogues were previously described by our group. However, the exact mode of action of CA and its
potential activity in functional cellular assays remained unknown. Thus, the main objective of our study
was to elucidate the mode of action of CA and to develop new potential chemical tools to study GC
signaling in adipocyte biology.
O
HO C
O
HO
2
R2
Me
Me
R1
OH
9
10
A
1
O
A
C
Me
B
OH
4
B H
D
7
R3
H H
iPr
1
. (+)- Colletoic acid (CA)
IC50= 13 nM microsomal assay (Ref. 18)
2
. Cortisone
Ki= 0.62 M biochemical assay (Ref. 19)
OAc
O
Me
CO2H
succinic acid
Me
H
H
Me
H
OMe
HO
Me
Me
Me
Me
A
O
A
E
D
D
Me
Me
Me
Me H
Me
O
3
. Hupehenol B
4. Carbenoxolone (CBX)
Ki= 0.062 M biochemical
assay (Ref. 17)
IC50= 15.3 M microsomal
assay (Ref. 16)
Figure 1. Representative natural product inhibitors of 11β-HSD1 and its native ligand.
Our streamlined synthetic strategy of the second generation of CA analogues successfully led to a
superior 11β-HSD1 inhibitor, compound 20. Simultaneously, we solved a single-crystal X-ray
diffraction structure of the 11β-HSD1-CA complex, providing to the best of our knowledge, the first X-
ACS Paragon Plus Environment
4

Page 5 of 56
Journal of Medicinal Chemistry
1
2
3
ray structure of a selective and potent natural product inhibitor bound to this enzyme. Our study also
provides novel insight into how CA inhibits preadipocyte differentiation via selective 11β-HSD1
inhibition in both murine and human preadipocytes cellular models. Through validation studies in the
murine preadipocyte model, this work expands the current 11β-HSD1 pharmacological toolkit by
providing a structurally distinct molecular scaffold, which can help explore the therapeutic potential of
modulating intracellular GC action via 11β-HSD1 in adipose tissue.
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
RESULTS
Chemistry. Molecular docking studies of CA with 11β-HSD1 were performed prior to synthetic
efforts, but provided limited information since the 11β-HSD1 cavity is flexible, and CA is relatively
small (data not shown). Therefore, a systematic chemical modification of the core of CA was carried out
to identify structure−activity relationship (SAR). The synthetic approach commenced with the total
synthesis of the enantiomer of the natural product (+)-CA, (ent-1) to enable cross-validation studies of
the combined chiral centers. It was envisioned that (−)-CA could arise from monoprotected 1, 3-
cyclohexadione via a series of functional group manipulations, namely Wittig mono-olefination, acid-
catalyzed ketone protection, deconjugation by treatment with DBU, hydrolysis of the ester followed by
coupling with (S)-4-benzyl-2-oxazolidinone and alkylation with a vinyl iodide moiety at C-1, and a
1
9
Heck-mediated cyclization reaction to afford the CA core system. This synthetic effort provided ent-1
in 19% yield over 21 steps from compound 21c as a diastereomerically pure compound (Figure 2A).
Next, to probe the influence of specific carbon centers (namely C-1, C-7 and C-8) of the natural product
(+)-CA, several analogues were generated (Figure 2B). Initial efforts to direct amidation of the
carboxylic acid under various conditions failed since rapid isomerization occurred at the C-1 center to
yield the corresponding epimer. The resultant epimer is arranged in close proximity to the C-9 hydroxyl
group, favorably facilitating an intramolecular nucleophilic cyclization to provide lactone 36 as the sole
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 56
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
product. To validate the identity of the epimer, base-mediated hydrolysis of lactone 36 was conducted,
and compound 7 was isolated with no traces of (+)-CA. Esterification of CA with
trimethylsilyldiazomethane in MeOH provided ester 5 in quantitative yield. Dihydroxylation of CA at 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
7
and C-8 was achieved upon treatment with catalytic osmium tetroxide and NMO, yielding compound
1
2 in 82% yield as a single stereoisomer, presumably due to the C-9 hydroxyl group directing effect.
However, the direct synthesis of the corresponding amide of CA, compound 6, was prevented by this C-
hydroxyl group. Thus, the C-9 hydroxyl group was oxidized with Dess–Martin periodinane, followed
9
by treatment with ammonium salt and HATU in the presence of Hünig’s base, providing compound 6 as
a single diastereoisomer (at C-9) after enone reduction with sodium borohydride.
Subsequently, C-4, C-8, C-9 and C-10 centers of CA were evaluated (Figure 2C). Epoxide 13a was
20
synthesized from the CA core in gram quantity as previously reported, which was subjected to a two-
step reaction sequence involving Swern oxidation [(COCl) , DMSO, and Et N] of the primary hydroxyl
2
3
group, followed by Pinnick oxidation (NaClO , t-BuOH, 2-Methyl-2-butene) to afford compound 13f.
2
Alternatively, compound 13a was directly submitted to TBSCl in the presence of imidazole, and
o
sequential treatment with thionyl chloride and pyridine at 0 C to afford the exocyclic olefin system.
Removal of the silyl group, and two-step oxidation of the primary alcohol resulted in compound 13e.
Acid-mediated epoxide opening occurred exclusively from the favored α-face and provided compound
1
3 in 81% yield. Then, the corresponding TBS-protected 13a was treated with lithium aluminum
hydride (LAH) at RT to provide a mixture of 10a and 14a in a 10:1 ratio. As expected, the aluminum ion
coordinates with the C-8 hydroxyl group of compound 13a, leading to hydride attack from the α-face,
which yielded compound 10a as the major product. Alternatively, reductive oxirane ring opening
1
9
conditions (PtO /H ) were used to obtain compound 14a as the major product. Compound 10a was
2
2
acetylated and exposed to two-step oxidation protocol (Swern/Pinnick), providing compound 11 in 77%
ACS Paragon Plus Environment
6

Page 7 of 56
Journal of Medicinal Chemistry
1
2
3
yield. Deprotection of the acetyl group followed by acid-promoted elimination led to the endocyclic
compound 10. Next, the C-9 hydroxyl group of 14a was selectively protected with an acetyl group, as
other evaluated protecting groups provided the thermodynamically more favorable endocyclic olefin
during the elimination step. Then, the protected intermediate was treated with thionyl chloride, followed
by TBAF to generate the exocyclic compound 14e. Swern/Pinnick protocol provided compound 14,
followed by acetyl deprotection to afford compound 15 in 81% yield (Figure 2C). Finally, to evaluate
the role of the substituent at C-4 and C-8, compounds 16i and 20q were generated. Following our
synthetic streamlined protocol, compound 16i, which was subjected to a series of protecting group
maneuvers to selectively eliminate the tertiary hydroxyl group at C-8, and ultimately afford compound
16l. The resultant compound was subjected to Swern/Pinnick oxidation and removal of the acetyl
protecting group, afforded compound 16 in 59% overall yield from 16i.
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
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 56
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
Me
A.
CO H
HO
OH
2
O
Me
9
O
O
1
16 Steps
1
4
5
Steps
Me
O
O
A
B
6
9% overall
28 %
1
N
(S)-4-Benzyl-2-
oxazolidinone
chiral auxiliary)
O
overall yield
yield
iPr
ent-1
21c
21i
iPr
(
Bn
O
CO H
HO
Me
CO H
O
2
2
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
B.
a. NH , MeOH
O
3
9
9
1
1
a. DMP
o
Me
5
60 C, 16 h
1
Me
81%
7
9
iPr
iPr
1:(+)-CA
3
6
iPr
b. HATU, NH Cl,
iPr EtNH
c. NaBH , MeOH,
4
OsO4 (cat)
b. i. NaOH(aq)
(CH3)3SiCHN2
2
o
o
ii. H+
MeOH, 0 C, 94%
NMO, 4 C
(aq)
4
8
2%
82%
,
(3 steps)
80%
HO
CO H
2
CO2H
1
HO
CONH2
HO
CO Me
2
HO
HO
Me
HO
8
9
1
1
Me
Me
7
Me
iPr
6
iPr
5
iPr
12
7
iPr
C.
a. TBSCl, imid.
b. SOCl , Pyr.
c. TBAF
O
R
O
CO H
OMe
2
HO
CO2H
+
e. H , MeOH
2
Me
HO
9
10
o
9 10
1
90 C
8
d. i. Swern oxid.
ii. Pinnick oxid.
0% (4 steps)
81%
13
i
Pr
13e iPr
iPr
7
1
3a: R= CH OH
2
Oxi.
92%
1
3f: R= CO H
2
a. Ac O/pyr
b. TBAF
0% (2 steps)
OH CO H
2
2
d. i. NaOH(aq)
ii. H+
90%
AcO CO H
2
OTBS
OH
10
9
Me
HO
Me
Me
c. i. Swern oxid.
ii. Pinnick oxid.
77% (2 steps)
1
0
10
iPr
HO
1
1
iPr
CH OH
10a
iPr
CO H
2
RO
OTBS
a. Ac O/pyr
b. SOCl , Pyr.
AcO
2
9
2
d. i. Swern oxid.
ii. Pinnick oxid.
HO
Me
2
9
8
c. TBAF
90% (4 steps)
e. NaOH(aq)
(81%, 3 steps)
iPr
4: R= Ac
15: R= H
HO
1
1
4e iPr
1
4a Pr
i
OH
OH a. i. TESCl, imid.
CO H
AcO
HO
2
HO
c. Swern/Pinnick [O]
d. NaOH(aq) /H+
ii. SOCl , Pyr.
2
Me
9
b. i. HCl(aq)
4
4
ii. TESCl, imid.
59% (overall yield)
HO
iii. Ac O/pyr.
iv.TBAF
2
16l Me
16 Me
16i Me
Figure 2. (A) Total synthesis of ent-1. (B) Synthesis of (+)-CA analogues derived directly from
synthetic 1. (C) Synthesis of CA analogues from advanced intermediates (Scheme S1-S5, Supporting
Information).
ACS Paragon Plus Environment
8

Page 9 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
Compounds 17, 18 and 19 were the products of chemical transformations of compound 20o (Figure
). Olefin reduction and 1,2-opening of the epoxide 20o, followed by protecting group maneuvers
3
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
provided compound 20l in 19% overall yield. Then 2-step oxidation of compound 20l, followed by
hydrolysis yielded compound 17 in 27% overall yield. Olefin reduction followed by 1,3 opening of the
epoxide 20o provided 20e in good yield, which was mono-acetylated to afford compound 22ff
quantitatively. Thionyl chloride elimination followed by silyl protecting group removal leads to
compound 22h. Then 2-step oxidation of the corresponding free alcohol to the carboxylic acid, and
consequent base treatment to remove the acetate group provided a mixture of compound 18 and 19 (2:1
ratio) in 85% overall yield. Finally, a series of chemical transformations and protecting group maneuvers
were conducted upon compound 20q to generate compound 20w (in 19% overall yield), which was
subjected to Swern/Pinnick protocol, followed by acetyl group deprotection, leading to compound 20 in
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
3
2% yield (3 steps).
OTBS
OH
O
OTBS
OAc
OH
HO
a. PtO2/H2(g),
AcOH, 2 h
HO
Me
c. i. Ac2O/Pyr.
Me
Me
b. DIBAL-H,
ii. SOCl , Pyr.
2
o
THF, 0 C
iii. TBAF
70% (3 steps)
2
0o
22h
80% ( 2 steps)
20e
a. PtO2/H2(g), AcOH
b. Ac2O/Pyr
c.TBAF
d. Swern/Pinnick [O]
e. NaOH (aq) /H+
1
8.6%
85%
(3 steps)
(
3 steps)
OH
OH CO2H
2
CO H
AcO
HO
Me
CO2H
d. Swern/
Pinnick [O]
Me
HO
Me
+ Me
e. NaOH(aq)
then H+
HO
17
27% (3 steps)
18
(2:1)
19
20l
OH
OH
HO
Me
HO
CO H
AcO
2
a. i. TESCl, imid,
ii. SOCl , Pyr.
b. i. HCl(aq)
1
c. i. Swern/Pinnick [O]
ii. NaOH (aq) /H+
9
2
1
4
Me
Me
4
HO
ii. TESCl, imid,
iii. Ac2O/Pyr
iv.TBAF
32% (3 steps)
ACS Paragon Plus Environment
20q
20w
20
9
19% (5 steps)

Journal of Medicinal Chemistry
Page 10 of 56
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 3. Synthesis of compounds 17-20 (Scheme S6, Supporting Information).
RESULTS AND DISCUSSION
Biochemical Evaluation of CA Analogues against Human 11β-HSD1. Although remarkable
binding affinity results have been reported with enriched 11β-HSD1 microsomal assays by other
groups,21-23 variability among batches prompted us to work with a highly robust purified protein system
measuring enzymatic activity as a function of cortisol concentration and NADPH fluorescence as a
readout.19, 22 Systematic SAR information was obtained by testing the generated CA analogues (Figure
4
) using a biochemical assay with a truncated 11β-HSD1 construct to determine relative inhibition
constant (K ) values (Table 1). Based on these findings, and the intricate arrangement of functional
i
groups of CA, it was anticipated that only (+)-CA (1), the natural enantiomer would have preference for
engagement with the topology of 11β-HSD1 to form a lock-and-key fit. To validate this hypothesis, ent-
1
was synthesized and evaluated first, and it showed at least 30-fold decrease in activity when compared
to (+)-CA as speculated. Next, the corresponding methyl ester 5 and amide 6 of CA were evaluated, and
both compounds display weaker biochemical activity, indicating a potential important role for the
carboxylic group at C-1. Compound 7, the epimer of CA at C-1 showed significant loss of activity (Ki=
354 μM), highlighting the role of the carboxylic acid and its stereochemistry, which provide unique
ACS Paragon Plus Environment
10

Page 11 of 56
Journal of Medicinal Chemistry
1
2
3
diastereofacial features to CA. Next, the role of the C-9 hydroxyl group as a potential hydrogen donor
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
was probed. Compound 8 with the acetylated hydroxyl group at C-9 displayed a decrease in activity
(K = 21 μM). However, after oxidation of the C-9 hydroxyl group, the resulting compound 9 showed
i
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
moderate activity (K = 5 μM). Evaluation of 3D molecular models of 9 indicated global conformational
i
changes of the A ring (Figure S3, Supporting Information) induced by the restricting enone system,
while this is a subtle change, the experimental data indicates a reduction in binding capacity to the
enzyme. The analogue compound 10 bearing the hydroxyl group at C-10 showed modest potency, but its
acetylated isomer 11, also lacking the olefin, showed a significant decrease in activity, implying the
hydroxyl group along with the conformation of the A ring play a role in the binding interactions with
11β-HSD1. Compound 12 was also less active than compound 10. The sp3 hybridization at C-8 distorts
the A ring, which adopts a twisted chair-like conformation in the endo-cyclic isomer, decreasing the
productive engagement of carboxyl group at C-1 with the target. Next, the exo-cyclic system was
interrogated by generating compounds 13-16, in which the A ring adopts a chair conformation with the
olefin in the pseudo equatorial position, while the C-9 substituent, and the carboxyl group are placed in a
relative similar position as CA. Compound 13 with an additional group at C-10 (methoxy group) showed
reduced activity, but in the same range as compound 12. Compound 14 showed improved activity
compared with compound 13, presumably due to lack of the bulky methoxy group, which negatively
affects the interactions required for the carboxylic group at C-1. Thus, compound 15 with the free C-9
hydroxyl group displayed better activity than compound 14 as it could interact with the catalytic triad
similarly to CA. The combined findings indicate the carboxylic acid at C-1, the C-9 hydroxyl and the A
ring conformation are important for effective binding to the target.
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 56
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
HO
CO H
2
CO H
2
HO
CO2Me HO
CONH2
HO
CO H AcO
2
CO H
2
HO
9
9
9
1
1
1
1
4
1
A
B
4
Me
Me
Me
Me
Me
Me
7
1
ent-1
K = 34M
LipE= 1.8
5
6
7
8
iPr
iPr
iPr
iPr
iPr
iPr
Ki= 21.2M iPr
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
i
K = 353.7M
i
K = 0.72M
LipE= 3.5
O
Ki>500M
K >500M
i
i
HO
CO2H
MeO
HO
AcO
CO H
AcO CO H
CO H
2
2
CO H
OH
2
CO H
2
2
HO
Me
12
9
10
10
8
10
Me
HO
8
8
Me
8
7
Me
4
4
13
14
9
1
1
10
HO
K = 130M
iPr
K = 5.15M iPr
iPr
Ki= 145M iPr
K = 12M
i
iPr
i
K = 289M
K = 8.23M
i
i
i
LipE= 2.8
OH
HO
CO2H Me
CO H
CO H
2
CO H
2
HO
CO2H
2
HO
CO H HO
2
1
4
9
1
4
Me
Me
Me
8
8
HO
1
9
4
4
18
20
1
7
K = 0.45M
i
K = 4.63M
15
i
Ki= 0.21M
LipE= 3.3
16
Ki= 57.6M
LipE= 2.2
iPr
Me
LipE= 1.93
K = 5M
LipE= 2.6
i
K = 24M
i
Figure 4. Biochemical activity (K ) evaluation of CA analogues: SAR study.
i
Finally, the size of the substituent group at C-4 was investigated. Compound 16 displays a small
methyl group substituent at C-4, exhibited five-fold less activity than compound 15, indicating the
isopropyl group was indeed also a contributing factor to the binding mode. This finding prompted us to
hypothesize that a large hydrophobic group at C-4 might be involved in favorable hydrophobic
interactions between the CA core and the binding site, and increase ligand residence time. Again, the
sp3 nature of the C-8 was validated with compound 17 as it was weakly active, but compound 18
showed improved activity. Compounds 19 and 20 displayed further improved K values compared to
i
CA. Gratifyingly, the superior compound 20 showed a three-fold increase of potency over CA.
Furthermore, these compounds (1, ent-1, 5-20) showed selectivity for 11β-HSD1 over its isoform
with no appreciable activity against 11β-HSD2 as shown in Table 1. The exquisite selectivity of these
compounds is relevant as 11β-HSD2 inhibition can lead to hypertension, hyperinsulinemia among other
17
side effects. The work illustrates how small chemical modifications to the core of CA influence its
binding ability, specifically, the presence and configuration of the carboxyl group at C-1, and the
ACS Paragon Plus Environment
12

Page 13 of 56
Journal of Medicinal Chemistry
1
2
3
influence of the hydroxyl groups at C-8, C-9, and C-10 on the conformational arrangement of the A ring.
The linear chemical synthesis of CA enantiomer, C-1, C-4, and C-8 to C-10 chemical manipulations
enabled a comprehensive SAR analysis against 11β-HSD1, leading to a superior CA analogue
compound 20 (Figure 4). The combined biochemical and cytotoxicity data for CA and its analogues is
described in Table 1, no cytotoxicity against the murine preadipocyte cell model (3T3-L1) under non-
differentiating conditions was observed at the tested concentrations for these compounds.
Table 1. Biochemical and cytotoxicity evaluation of CA analogues.
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
1
1β-HSD1
Ki (μM)
11β-HSD1 11β-HSD2
3T3-L1
Compound
(CA)
IC (μM)
IC (μM)
EC (µM)
5
0
50
50
1
0.73 ± 0.26
4.1 ± 0.4
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>300
>500
23.55 ± 3.3
ent-1
13.5 ± 2.2
80.3 ± 12.9
4
(CBX)
0.062 ± 0.038 0.41 ± 0.05 0.17± 0.05
5
6
7
8
9
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
354 ± 194
21.3 ± 4.4
5.15± 0.30
8.23± 0.63
289 ± 58.4
130 ± 57
>500
178 ± 37
42.4 ± 2.5
55.1 ± 4.2
>500
1
1
0
1
12
>500
1
1
1
3
4
5
145 ± 23.5
11.8 ± 1.9
5.4 ± 0.47
26.9 ± 4.8
57.6 ± 5.4
4.63± 0.57
0.27 ± 0.034
0.21 ± 0.022
0.28 ± 0.058
>500
71 ± 11.3
38.3 ± 3.3
162 ± 28.4
379 ± 35.5
27.2 ± 3.4
2.2 ± 0.3
1.34 ± 0.14
2.07 ± 0.05
16
1
1
1
7
8
9
20
PF-915275
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 56
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
Target engagement: Complex of human 11β-HSD1 construct with CA. Based on previous 11β-
HSD1 studies,22, 24 it was hypothesized that CA would bind to the catalytic site as a competitive ligand.
To test that hypothesis, co-crystallization and ligand-soaking strategies of CA with a construct of 11β-
HSD1 containing the coding sequence for the catalytic domain (residues 24–292), a single point
mutation (C272S), and a C-terminal truncation (residues 264–292) were conducted. These amino-acid
perturbations of the protein construct are remote from the ligand- and cofactor-binding sites (the
catalytic pocket) and should have minimum effects on the ligand-protein interaction as previously
described.19, 20 After several attempts, CA was soaked into pre-formed 11β-HSD1 crystals resulting in 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
2
.6-Å resolution structure of the 11β-HSD1-CA crystal complex (Table S3), with a root-mean-square
deviation of 0.008 Å for aligned Cα atoms and describes the binding mode of CA to the enzyme for the
first time. A global cartoon of the tetramer of 11β-HSD1 with CA bound near the cofactor NADP+ is
shown in Figure 5, highlighting the narrow but flexible cavity of the ligand entrance.
Figure 5. Cartoon model of the co-crystal structure of 11β-HSD1 tetramer with CA bound (6NJ7). The
right panel provides a close-up view of the active site cavity surface for chain B with CA and NADP+
rendered as sticks (yellow for CA and dark blue for cofactor carbons, respectively).
ACS Paragon Plus Environment
14

Page 15 of 56
Journal of Medicinal Chemistry
1
2
3
4
5
The arrangement of the catalytic residues in the 11β-HSD1 active site align similar to previous
reported crystal structures of 11β-HSD1, revealing a catalytic triad (Ser170, Tyr183, and Lys187) with a
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
hydrophobic cavity that facilitates the interactions of steroidal chemical scaffolds. The hydrophobic
nature of the binding site and the conformational plasticity of 11β-HSD1 becomes evident when
comparing open vs. closed states of the dimer with different ligands, clarifying why numerous small-
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
1
4
molecule inhibitors with little or no structural similarities have been reported. The substrate binding
pocket of 11β-HSD1 is shielded from polar solvent as it is destined for hydrophobic substrates
2
4
(
cortisone/cortisol). The cofactor NADP+ binds to 11β-HSD1 analogous to previous reports of binding
at the active site, residing just above CA, in which the ribose rings display a C2-endo conformation. The
nicotinamide ring of NADP+ is near the carboxylate of CA, and stabilization interactions occur because
of the proximity.
Figure 6. Close-up view of the crystal structure of the catalytic triad of 11β-HSD1 bound with CA as
determined from collected co-crystal structure data (6NJ7). (A) CA (yellow) bound to 11β-HSD1 chain
A, NADP+ is shown as dark blue. Hydrogen bonds are indicated by dashed lines. (B) Representative
rendering of electron density of CA. (C) View looking down of the catalytic pocket (residues 226, 227,
and 231 of the variable loop have been removed for clarity).
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 56
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
The ligand binding pocket is partially formed by the ‘variable loop’ and the C-terminal region of the
dimer subunit, and this distal portion of the pocket is dynamic and is only partially observed within the
crystal structure. The rigid anti-periplanar ring system of CA is fully appreciated in close-up images of
the stick representation (Figure 6C), which highlights the rich stereo-complexity generated by the all-
carbon spirocenter at C-5. The saturated five-membered B ring of CA, which can exist as a combination
of puckered conformations, appears to favor the envelope conformer; it is unlikely to interconvert with
the twist nonplanar form, which would require the simultaneous rearrangement of the cyclohexyl A ring,
leading to an energetically unfavorable isomer. The carboxyl group at C-1 of CA interacts directly with
both catalytic triad residues Tyr183 and Ser170, forming hydrogen bonds. Remarkably, the carboxylate
moiety of CA also interacts with the cofactor NADP+: a highly favorable π-stacking interaction occurs
between the negatively charged CA carboxylate and the positively charged nicotinamide moiety (Figure
S4, Supporting Information). Furthermore, the hydroxyl group at C-9 of CA is close enough to form an
intramolecular hydrogen bond with the carboxylate oxygen at C-1, potentially enhancing the stability of
the bound inhibitor conformation. The methyl group at the C-8 position of CA is near the more-
constrained region of the binding pocket, but there appears to be opportunities to introduce other
functional groups at the C-8 that could favorably interact with Tyr 177 at the dimer interface. The
electron density of the binding pocket complements the steric restrictions imposed by the compact
scaffold of CA. The stereochemical disposition of the terminal methyl groups of the isopropyl group at
C-4 are not clearly defined in the electron density, indicating freedom of rotation, which could
accommodate larger hydrophobic groups. This unexpected finding was experimentally validated by
compound 20 which shows superior inhibition (Table 1) due to the cyclohexyl group, suggesting that
bulkier hydrophobic alkyl group can further improve the potency. Finally, the methyl groups were
modelled in hydrophobic contact with the nonpolar residues Val180 and Leu126 due to its juxtaposition
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
16

Page 17 of 56
Journal of Medicinal Chemistry
1
2
3
as shown in Figure 6C, the predominantly hydrophobic amino acid residues that line the interface of the
dimer become visible.
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
In Vitro Cellular Evaluation of CA. The process of adipogenesis is highly regulated and it involves
well-defined cellular stages, namely a commitment stage guided by transcription factors (such as
CCAAT/enhancer-binding protein [C/EBP] family members, peroxisome proliferator-activated
receptors, and glucocorticoid receptor) at the onset of differentiation, and cellular remodeling stage that
is accompanied by lipid accumulation, leading to differentiation termination.25-27 Hypertrophic growth
involves the increase in size of the lipid droplet and an increase in protein content; such characteristics
are exaggerated in obesity.28-32 Several studies implicate GC signaling in both pre-adipocyte
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
1
6
commitment during differentiation and hypertrophic adipocyte growth. While 11β-HSD1 inhibitors
have shown anti-diabetic effects in vivo,12-14 such as PF-915275 and derivatives, they also showed
cytotoxicity against preadipocytes at the effective concentrations, presumably due to mitochondrial
damage (Table S2, Supporting Information).13b
13
Thus, the adipogenesis cellular assay was selected to study and validate the biological activity of CA.
First, the effects of CA on cell viability and cell cycle were evaluated in the murine 3T3-L1
preadipocyte model. CA showed no cytotoxicity at the tested concentrations (up to 1mM) with no
significant effects on cell cycle under non-differentiating or induction conditions (Figure S7), rendering
CA a valuable probe to study preadipocytes adipogenesis without cellular cytotoxicity. While we had
3
3
previous shown that CA inhibits the conversion of [ H]-cortisone to [ H]-cortisol in the LPS activated
1
9
murine J774.1 macrophage by inhibition of 11β-HSD1, it was important to assess whether CA would
also inhibit the glucocorticoid receptor (GR) signaling pathway via 11β-HSD1. In alignment to reported
studies of glucocorticoid effects in 3T3-L1 differentiation by Kim, et. al.,33a dehydrocorticosterone
(DHC), cortisone or dexamethasone were evaluated (Figure S7, Supporting Information). CA
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 56
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
significantly inhibits differentiation in the presence of all three compounds. For the presented successful
differentiation of the cells, established protocols were followed.33,34 For induction of the preadipocyte
murine 3T3-L1 and human preadipocytes, dehydrocorticosterone or cortisone (250 nM), and
hydrocortisone hemisuccinate (100 nM) were used respectively. Cells were treated with appropriate
controls (non-differentiating medium NDM for negative control, differentiating medium with DMSO for
positive control), or CA (50-100 µM) for 7 days for the murine cellular model, and 10-14 days for the
human cellular model. Oil Red O staining (ORO) results of the 3T3-L1 cell model treated with negative
control, positive control, or CA are shown in Figure 7 A-D respectively. CA treatment clearly inhibited
adipogenesis when compared to the differentiated control cells, and showed a significant decrease in
lipid droplet accumulation as quantified by ORO staining. A similar pattern of adipogenesis inhibition
was observed for compound 20 as expected and at lower concentrations (Figure S11, Supporting
Information). A neutral lipid (green) was used together with a mitochondria tracker (red) in a double
staining experiment, providing a closer visualization of the cellular morphology and the lipid droplets at
higher resolution (Figure 7 E-H). Unlike differentiated cells, CA treated cells displayed a staining
pattern similar to undifferentiated cells. A significant (p < 0.0001) increase of lipid droplets was
captured in those images for the positive control when compared to the negative control. However, for
CA treated cells no significant difference was observed when compared to the NDM control cells. CA
showed inhibition properties only when added at induction day. CA treatment after day 3 had little to no
significant effect on the course of adipogenesis, indicating CA suppresses adipogenesis by attenuating
GC signaling at the cell commitment stage. The adipocyte differentiation assay was also conducted with
human visceral preadipocytes (Figure 7 I-L) and human subcutaneous preadipocyte (Figure S12,
Supplemental Information). The cell doubling time is much slower, thus it took 10-14 days to observe
significant lipid accumulation. CA also demonstrated modest inhibition of adipogenesis when compared
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
18

Page 19 of 56
Journal of Medicinal Chemistry
1
2
3
to the controls. However, the CA treated visceral human preadipocytes that differentiated displayed
smaller lipid droplets than the corresponding differentiated control cells, indicating CA affected lipid
accumulation. Interestingly, a greater effect was observed for the subcutaneous preadipocytes than for
the visceral preadipocytes. Also, no quantifiable effect was observed in human preadipocytes when CA
was added at day 3 after induction, highlighting the adipogenesis inhibitory effects on early stage cell-
fate commitment.
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
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 56
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
Figure 7. Representative images of adipogenesis assay of murine 3T3-L1 (A-H), and visceral human
preadipocytes (I-L) after staining with Oil Red O (red) or lipid stain (green along with mitochondria red
tracker). (A and I) Non-differentiating medium (NDM)–treated control cells. (B and J) Differentiating
medium (DM)–treated control cells. (C and K) DM and CA (50 µM)-treated cells. (D and L) Relative
quantification of ORO. (E) NDM–treated control cells. (F) DM–treated control cells. (G) DM and CA
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
(50 µM)-treated cells. (H) Relative Mean FL intensity of lipid staining (green). Magnification 10× in A-
C and I-K, and 63× in E-G.
Next, the steady-state gene transcript levels relevant to the adipogenesis process were measured for
3
T3-L1 by using qRT-PCR, and the effects of CA inhibition on cell-fate and fatty acid accumulation
3
5
related genes for days 0, 1, 2, and 7 was determined (Figure 8). First CREB, C/EBP isoforms (α, β, ),
and S6K1 were evaluated as they are crucially involved in the early stage of adipocyte differentiation.
Upon induction of adipocyte differentiation, phosphorylation of CREB leads to activation of C/EBP
transcription, and an increase in C/EBPα, β and  mRNA levels is expected. Indeed, a modest increase
of the mRNA level of these genes was observed on day 1 and day 2 in the control differentiated cells,
while CA-treated cells showed a significant reduction of C/EBPα, β and  mRNA expression for day 1-
2, namely during induction, the critical period that determines cell-fate. A similar mRNA expression
reduction was recorded for the S6K1 gene for the CA-treated cells at day 1 and day 2. S6K1 has been
3
6
described as a key element of adipocyte differentiation commitment. During later stage of adipocyte
differentiation, Peroxisome Proliferator-activated Receptor γ (PPARγ) and PPARγ coactivator-1-alpha
(PGC-1α) promote accumulation of fatty acid deposits, leading to visible lipid dropplets.37-38 In the
differentiated 3T3-L1 cells, mRNA expression levels of these genes were significantly elevated by day 2
ACS Paragon Plus Environment
20

Page 21 of 56
Journal of Medicinal Chemistry
1
2
3
and plateau at day 7 (Figure 8). However, the mRNA expression level of the same genes was greatly
reduced in CA-treated cells. In fact, PGC-1α mRNA expression in the differentiated control cells was at
least ten-fold greater than in the CA-treated cells.
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 8. Differential gene expression during the 3T3-L1 adipogenesis assay on days 0, 1, 2 and 7. The
steady-state mRNA levels were analyzed according to the ΔCt method using β-actin, PP1A and PP1B as
housekeeping genes with their calculated geometric mean for each sample as standardized Ct [i.e.
3
5
2
(Ct,HK–Ct,target)−1 (Pfaffl et al., 2004)]. Specific primers sets of each gene were used to amplify
single products (Table S5, Supporting Information). The values represent the means ± SEM, n = 3.
Significant differences between the groups by using ANOVA’s one-way followed by post-test and are
indicated by asterisks; ****pꢀ<ꢀ0.0001; ***pꢀ<ꢀ0.001; **pꢀ<ꢀ0.01; *p<0.05; ns, not significant (p>0.05).
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 56
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
Expression of 11β-HSD1 mRNA was also reduced in the CA-treated cells when compared with the
differentiated control cells. Furthermore, a comparable pattern was observed for the mRNA expression
of Glucose Transporter 4 (Glut4) and Acetyl-CoA carboxylase beta (ACACB), both genes involve in
regulation of metabolism. Finally, the mRNA expression of terminal marker genes of adipogenesis,
Fatty Acid Binding Protein 4 (FABP4, also known as AP2), and Fatty Acid Synthase (FASN) were
evaluated. The mRNA expression levels in the CA-treated cells were at least 10-fold lower than in the
differentiated control cells. The combined data suggests that CA-treated cells are prevented from
committing to the differentiation program during induction, and the lipid formation program was
suppressed at the late stage of adipogenesis.
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
To confirm that gene expression changes correlated with protein expression level, western blot (WB)
analyses at days 3 and 7 were conducted in 3T3-L1 cell model. Indeed, congruency was observed
between RNA and protein expression, particularly for 11β-HSD1, PPARγ, and PGC-1α (Figure S13,
Supporting Information). No substantial 11β-HSD1, PPARγ, and PGC-1α expression differences were
detected between non-differentiating medium (NDM) treated control cells and differentiating medium
(DM) treated control cells at day 3 (Figure 9). However, significant increase of these protein expression
levels was recorded by day 7 in the DM treated cells. In the CA-treated cells, expression levels remain
comparable to those in NDM cells, suggesting that CA also effectively suppress the protein expression
of these genes, particularly PPARγ at the late stage of adipogenesis. Thus, CA could prevent adipocyte
3
8
hypertrophy via PPARγ suppression as described by Iwabu et. al. In addition, total GR expression
remained largely unchanged for day 3 and 7 under the different experimental conditions (Figure S13,
Supporting Information).
ACS Paragon Plus Environment
22

Page 23 of 56
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 9. CA-treated 3T3-L1 differentiating cells showed down-regulation of PGC-1α, 11β-HSD1, and
PPARγ expression by day 7. No significant change was detected for day 3. Representative Western blot
(30 μg of total protein and tubulin was used as the loading control) and quantification of day 3 and day 7
samples. NDM–treated control cells, DM–treated control cells, DM-treated with Rosiglitazone (Rosi)
and DM-treated with CA (50 µM)-treated cells. The values represent the means ± SEM, n = 3.
Significant differences between groups are indicated by asterisks; ***pꢀ<ꢀ0.0001; **pꢀ<ꢀ0.002; *pꢀ<ꢀ0.01;
ns, not significant (p>0.05).
The significant reduction of PGC-1α expression at the RNA and protein level for the CA treated cells
when compared to DM control cells led us to interrogate whether mitochondrial density was affected by
using live-cell imaging with a mitochondrial tracker (magenta) for 3T3-L1 cells during adipogenesis.
PGC-1α is a regulator of energy metabolism and promotes fatty acid oxidation, which stimulate
mitochondrial biogenesis to provide the required energy for lipid synthesis and storage.29,38 The mean
intensity of mitochondrial tracker (magenta) staining for each treatment group along with nuclear
Hoechst stain (blue) was captured with a confocal imaging system and it is shown in Figure 10. The
resulting findings clearly illustrated higher mitochondrial mass for the differentiated cells (Figure 10B),
while both the NDM (Figure 10A) and the CA-treated cells (Figure 10C) displayed less mitochondrial
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 56
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
mass as relative quantified (Figure 10D). The images confirmed that PGC-1α expression correlates with
mitochondrial mass, and that CA inhibits the required signaling for lipid deposit formation.
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 10. Live-cell images of 3T3-L1 adipogenesis assay stained with mitochondrial tracker (magenta)
and nuclear stain (blue). (A) NDM control cells. (B) DM–treated control cells. (C) DM–treated CA (50
μM). (D) Mean intensity of mitochondrial tracker. Imaged on a Marianas spinning disk confocal
imaging system at 63× magnification.
Although no detection of cytotoxicity upon CA treatment at the effective concentrations, protein
synthesis inhibition could affect cellular commitment without causing cell death. Protein synthesis and
cellular degradation are closely regulated processes that enable cells to adapt during fluctuating
environmental conditions. Thus, using a global protein synthesis assay, CA was evaluated for alterations
in intracellular protein homeostasis in a fibroblast cell model (Figure 11). The method is based on an
alkyne containing o-propargyl-puromycin probe, which stops translation by covalently binding with
nascent polypeptide chains. Then, the truncated polypeptides are rapidly turned over by the proteasome
39
and can be detected based on the subsequent click reaction with a fluorescent azide (red), and the
counter DNA stain picogreen (green) for overall evaluation of the cell. The protein synthesis inhibitor,
cycloheximide was used as the positive control. No significant changes were observed for CA treated
ACS Paragon Plus Environment
24

Page 25 of 56
Journal of Medicinal Chemistry
1
2
3
cells as compared to DMSO control. Several time points were evaluated, but no protein synthesis
inhibition was observed over time (data not shown).
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 11. Evaluation of protein synthesis by using a global protein synthesis assay. (A) Vehicle
(DMSO). (B) Cycloheximide (20 nM, positive control). (C) CA (50 μM). Nascent protein (red), DNA
(green) and merged representative images at 20× magnification with a Nikon C2 scanning confocal
microscope for 3 independent experiments.
To investigate the possibility that CA interacts directly with PPARγ and GR, both of which directly
influence the differentiation process, supporting assays were conducted. First the potential of CA to bind
or directly affect the GR activation was evaluated. A transiently transfected mammalian cellular model
4
0
using a GR response element–driven luciferase DNA construct was used. Cell were treated with CA or
vehicle in the presence of dexamethasone and CA treatment did not inhibit GR activation (Figure 12),
suggesting that CA does not directly bind to GR. This data indicates that CA does not affect GR
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 56
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
activation in non-GR dependent cellular models, and the observed adipogenesis inhibition is primarily
due to selective inhibition of 11β-HSD1.
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 12. Human GR reporter assay showed CA did not affect GR activation. Glucocorticoid receptor
(GR) functional activation. Transiently GR-transfected cells were induced (In) by treatment with 100
nM dexamethasone. Cells were treated with colletoic acid (CA) 1 h prior to GR induction. Control cells
were non-induced (Non-In). The values represent the means ± SEM, n = 3. Significant differences
between groups are indicated by asterisks; ****pꢀ<ꢀ0.0001.
Finally, the cellular PPARγ reporter assay kit41 was used to evaluate whether CA affects this
important transcriptional factor. Previous studies of transcriptional antagonists of PPARγ have
demonstrated adipogenesis inhibition in 3T3-L1 cells.27 The potent irreversible PPARγ antagonist
GW9662 was used as a positive control (Figure 13). CA showed no quantifiable PPARγ antagonistic
activity, indicating CA does not directly interact with PPARγ.
ACS Paragon Plus Environment
26

Page 27 of 56
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 13. Human PPARγ reporter antagonist assay. Transcriptional response of colletoic acid after
stimulation by rosiglitazone (20 nM) during PPARγ reporter assay. Irreversible PPARγ antagonist
GW9662 was used as a control (EC ꢀ=ꢀ2 nM). Colletoic acid did not show antagonistic effect in this
5
0
PPARγ reporter assay at the tested concentrations.
Physicochemical Properties and its Analogues. Lipophilic efficiency (LipE), also known as
lipophilic ligand efficiency, is a metric that normalizes lipophilicity relative to potency, aiding in
prioritizing lead compounds. Thus, LipE and in vitro ADME properties were evaluated for CA and its
active analogues (9, 18-20, Table S1, Supporting Information). Interestingly, CA displayed the best
LipE (3.5) while compound 20 (3.3) was comparable to CA. Overall the solubility properties of
compound 19 and 20 were acceptable and similar to CA, but further improvement would be necessary
for in vivo efforts. Metabolic stability was evaluated by microsomal assay using mouse (m) and human
(hs) liver microsomes: CA (m t1/2=1.33 h; hs t1/2=1.18 h), compound 18 (m t1/2=0.3 h; hs t1/2=1.29
h), compound 19 (m t1/2=0.27 h; hs t1/2=0.93 h) and compound 20 (m t1/2=0.13 h; hs t1/2=0.3 h)
showed rapid clearance in mice, while they were more stable in human microsomes. Conversely,
compound 9 (m t1/2=2.23 h; hs t1/2=6.8 h) showed the highest metabolic stability, indicating that the C-
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 56
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
9
hydroxyl group might be responsible for the observed clearance. However, human plasma stability
assay was remarkably high for all the tested compounds, CA, 9, 18-20 (t1/2> 48 h). A similar trend was
recorded in mouse plasma for this set of compounds except for CA, which showed poor mouse plasma
stability. Simulated gastric fluid stability assay was not optimal for CA (<2 h), but it was promising for
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
compounds 9, 18-20 (> 6 h). Thus, preliminary PK/PD studies of CA (Table S2, Supporting
Information) were conducted to test its in vivo potential. However, in vivo systemic unfavorable plasma
clearance was observed with a volume of distribution at steady-state (V ) was close to the predicted
SS
(Obs V 0.247 ± 0.169 L/kg, ACD Pred 0.33 L/kg), which indicates that further development of CA is
SS
required for in vivo investigations. However, the combined studies indicate there is ample opportunities
for improvement of the physicochemical properties of the CA scaffold using bioisostere strategies (C-9
hydroxyl group replacement), and/or prodrug formulation/liposome delivery systems. This study
demonstrated opportunity for CA’s medicinal chemistry optimization, suggesting its potenital utility in
animal models.
CONCLUSIONS
In summary, we describe the design, synthesis, and cellular evaluation of CA analogues as selective
inhibitor of 11β-HSD1, providing mechanistic insights on the mode of action of CA in adipocyte
biology. CA shows no cytotoxicity at the tested concentration, therefore it can serve as a selective
chemical probe to enable an in-depth dissection of the role of GC signaling during adipocyte
differentiation via pharmacological inhibition of 11β-HSD1. A comprehensive SAR analysis of CA
revealed a strong correlation between the functional groups at C-1, C-4, and C-9. We have solved the
first X-ray crystal structure of the 11β-HSD1-CA complex, highlighting the sterically demanding all-
carbon quaternary spirocycle, which leads to a tight binding complex via interactions of Ser170/Tyr183
with the carboxyl group at C-1 of CA, and interactions of the hydrophobic residues of the catalytic site
ACS Paragon Plus Environment
28

Page 29 of 56
Journal of Medicinal Chemistry
1
2
3
(Thr124, Leu126, and Val180) with the alkyl substituent at C-4 of CA, thus providing ligand
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
stabilization, which was congruent with our SAR studies. The pseudo-transition state of such binding
mode provides rational for the high selectivity of CA to 11β-HSD1. The combined efforts led to a
superior CA analogue, namely compound 20, which was developed by introducing a bulky cyclohexyl
group at C-4 and displayed a three-fold increase in potency than CA. CA did not inhibit de novo protein
synthesis, and under preadipocyte differentiating conditions reduced mitochondria biogenesis. We
demonstrate that CA inhibits adipocyte differentiation in both murine and human cellular models at the
early stage of adipogenesis. No cellular toxicity was observed during CA treatment, thus the decrease of
PPARγ, and Glut4 during adipogenesis can be solely attributed to 11β-HSD1 inhibition. The
demonstrated synthetic manipulations of the CA core render it now feasible for the development of
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
42
5c
fluorescent/radio probe for live-cell or in vivo imaging studies. Additional in-depth cellular studies
aimed at evaluating CA analogues in combination with other therapies are warranted, as attenuating
adipogenesis by selectively inhibiting 11β-HSD1 may represent an alternative approach to prevent
adipocyte dysfunction in MetS related disorders.
EXPERIMENTAL SECTION
General Chemistry. All reactions were carried out under argon atmosphere, and anhydrous
conditions, unless noted otherwise. Dry tetrahydrofuran (THF), toluene, diethyl ether (Et O), ethyl
2
acetate (EtOAc) and methylene chloride (CH Cl ) were obtained by passing commercially available pre-
2
2
dried, oxygen-free formulations through activated alumina columns. Acetonitrile (CH CN), dimethyl
3
sulfoxide (DMSO), and isopropanol were purchased in anhydrous form and used without further
1
purification. Yields refer to chromatographically and spectroscopically ( H NMR) homogeneous
materials, unless otherwise stated. Reagents of the highest available quality were purchased and used
without further purification, unless otherwise stated. Reactions were magnetically stirred and monitored
ACS Paragon Plus Environment
29