Phenotypic Discovery of SB1501, an Anti-obesity Agent, through Modulating Mitochondrial Activity

Article abstract of DOI:10.1002/cmdc.202100062

Obesity has become a pandemic that threatens the quality of life and discovering novel therapeutic agents that can reverse obesity and obesity-related metabolic disorders are necessary. Here, we aimed to identify new anti-obesity agents using a phenotype-based approach. We performed image-based high-content screening with a fluorogenic bioprobe (SF44), which visualizes cellular lipid droplets (LDs), to identify initial hit compounds. A structure-activity relationship study led us to yield a bioactive compound SB1501, which reduces cellular LDs in 3T3-L1 adipocytes without cytotoxicity. SB1501 induced the expression of gene products that regulate mitochondrial biogenesis and fatty acid oxidation in 3T3-L1 adipocytes. Daily treatment with SB1501 improved the metabolic states of db/db mice by reducing body fat mass, adipose tissue mass, food intake, and increasing glucose tolerance. The anti-obesity effect of SB1501 may result from perturbation of the PGC-1α–UCP1 regulatory axis in inguinal white adipose tissue and brown adipose tissue. These data suggest the therapeutic potential of SB1501 as an anti-obesity agent via modulating mitochondrial activities.

Full text of DOI:10.1002/cmdc.202100062

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
1
2
3
4
5
6
7
8
9
Phenotypic Discovery of SB1501, an Anti-obesity Agent,
through Modulating Mitochondrial Activity
Ala Jo+,[a] Mingi Kim+,[a] Jong In Kim,[b] Jaeyoung Ha,[c] Yoon Soo Hwang,[a] Hyunsung Nam,[a]
Injae Hwang,[b] Jae Bum Kim,[b] and Seung Bum Park*[a, c]
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Obesity has become a pandemic that threatens the quality of
life and discovering novel therapeutic agents that can reverse
obesity and obesity-related metabolic disorders are necessary.
Here, we aimed to identify new anti-obesity agents using a
phenotype-based approach. We performed image-based high-
content screening with a fluorogenic bioprobe (SF44), which
visualizes cellular lipid droplets (LDs), to identify initial hit
compounds. A structure-activity relationship study led us to
yield a bioactive compound SB1501, which reduces cellular LDs
in 3T3-L1 adipocytes without cytotoxicity. SB1501 induced the
expression of gene products that regulate mitochondrial bio-
genesis and fatty acid oxidation in 3T3-L1 adipocytes. Daily
treatment with SB1501 improved the metabolic states of db/db
mice by reducing body fat mass, adipose tissue mass, food
intake, and increasing glucose tolerance. The anti-obesity effect
of SB1501 may result from perturbation of the PGC-1α–UCP1
regulatory axis in inguinal white adipose tissue and brown
adipose tissue. These data suggest the therapeutic potential of
SB1501 as an anti-obesity agent via modulating mitochondrial
activities.
Introduction
safety concerns due to unwanted adverse effects.[9,10] As an
alternative approach, small-molecule mitochondrial uncouplers
have gained attention as anti-obesity agents. Mitochondrial
uncouplers show significant weight-loss effects by allowing
protons to enter the mitochondrial matrix without generating
ATP, thereby increasing the oxidation of circulating lipids and
the cellular glucose uptake to compensate for the total amount
of ATP.[11] These observations highlight mitochondrial uncou-
plers as attractive therapeutic agents for reversing obesity. Even
though certain molecules might be specific to mitochondria,
only a few have anti-obesity effects with a suitable therapeutic
window.[12] For example, the small molecule 2,4-dinitrophenol is
the best-studied uncoupler, but its clinical use was banned due
to toxicity issues.[13] Therefore, there remains a need for new
chemical entities (NCEs) that present therapeutic potential as
anti-obesity agents with improved safety profiles.
This study aims to discover new anti-obesity agents using a
phenotype-based approach, which focuses on phenotypic
changes in cells or organisms as an outcome of perturbation in
multiple interconnected pathways.[14] This approach can be
more powerful when diseases have high biological hetero-
geneity or their therapeutic targets are not thoroughly defined.
Thus, phenotype-based drug discovery has emerged as a
promising approach for identifying NCEs as candidates for first-
in-class drugs.[15]
In recent decades, obesity has become a pandemic that
threatens the quality of life. Obesity is positively associated with
the risk of metabolic diseases, such as type 2 diabetes,
dyslipidemia, hypertension, cardiovascular diseases, and various
cancers.[1–4] Currently, over one-third of adults worldwide are
obese, and the rate of obesity has doubled in 73 countries since
1980.[5] The consistent increase in the obese population
constitutes an enormous medical and social burden in the 21st
century. Thus, the demand for novel therapeutic agents that
can reverse obesity and obesity-related metabolic disorders
continues to grow.
At present, five drugs have been approved by the U.S. Food
and Drug Administration (FDA) as anti-obesity agents in
patients with a body mass index (BMI) �27 kg/m2 and at least
one obesity-related metabolic disorder or patients with BMI
�30 kg/m2.[6,7] These drugs focus primarily on reducing energy
intake by targeting pathways in the central nervous system
(CNS) or preventing dietary fat absorption by inhibiting gastric
and pancreatic lipases in the case of orlistat.[8] Still, they pose
[a] Dr. A. Jo,+ M. Kim,+ Y. S. Hwang, H. Nam, Prof. S. B. Park
CRI Center for Chemical Proteomics
Department of Chemistry, Seoul National University
Seoul 08826 (Korea)
Here, we performed an image-based high-content screening
for quantitative monitoring of cellular lipid droplets (LDs) with a
fluorogenic LD bioprobe (SF44).[16] We discovered a new
bioactive compound, SB1501, which contains a benzopyran-
substituted 1,2,3-triazole as a new small-molecule that reduces
cellular LD accumulation. We demonstrated that SB1501
significantly reduces cellular LDs in 3T3-L1 adipocytes differ-
entiated from murine embryonic fibroblast and HeLa human
cervical cancer cells. We also observed the reduction of fat
accumulation in adipose tissues and body weight along with
E-mail: sbpark@snu.ac.kr
[b] Dr. J. I. Kim, Dr. I. Hwang, Dr. J. B. Kim
CRI Center for Adipocyte Structure–Function
School of Biological Sciences, Seoul National University
Seoul 08826 (Korea)
[c] J. Ha, Prof. S. B. Park
Department of Biophysics and Chemical Biology
Seoul National University
Seoul 08826 (Korea)
[+] These authors contributed equally to this work.
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/cmdc.202100062
ChemMedChem 2021, 16, 1104–1115
1104
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
improved glucose tolerance in a db/db mouse model. We
further demonstrated that SB1501 might manifest its anti-
obesity effect by perturbing the PGC-1α-UCP1 regulatory axis.
In summary, we identified the therapeutic potential of SB1501
as an anti-obesity agent.
Using SF44 that turns on in the hydrophobic environment of
LDs, we robustly monitored the homeostasis of cellular LDs in
live cells without any fixation or washing steps (Figure S1 in the
Supporting Information). These distinct advantages of SF44
allowed the non-invasive measurement of LD homeostasis with
excellent practicality compared to other commercial LD probes.
We then performed this high-content screening against an
approximately 3000-member in-house drug-like poly-hetero-
cycle library constructed using a privileged substructure-based
diversity-oriented synthesis (pDOS) strategy.[20] Our image-based
screening resulted in the discovery of the initial hit compound
8 containing a benzopyranyl 1,2,3-triazole skeleton, which
reduced cellular LDs down to 67% in HeLa cells at the 10 μM
concentration compared to DMSO-treated cells without signifi-
cant toxicity (Figure 1a, b and Table 1).
Next, we pursued a structure-activity relationship (SAR)
study of our initial hit compound 8 by designing and synthesiz-
ing approximately 30 analogs based on the benzopyranyl 1,2,3-
triazole scaffold (Figure 1c). We systematically introduced the
R1, R2, and R3 substituents on the benzopyranyl 1,2,3-triazole
core skeleton (Table 1). We first introduced various aliphatic
and aromatic substituents via the copper-mediated click
reaction to explore the substituent effects at the R3 group on
1
2
3
4
5
6
7
8
9
Results and Discussion
Discovery of SB1501 and its structure–activity relationship
study
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Obesity is a metabolic condition involving excess fat accumu-
lation in adipose tissues. Under obesity, adipocytes store
excessive energy in the form of triglycerides in LDs, resulting in
an expansion of cell sizes.[17] One of the most significant
phenotypic changes in adipocytes under obesity is the enlarge-
ment of cellular LDs caused by fat accumulation.[18] As we
envisioned the phenotype-based discovery of novel small-
molecule modulators that reduce cellular fat accumulation, we
first established cell-based high-throughput phenotypic screen-
ing with the fluorogenic LD-staining dye (SF44) for quantitative
monitoring of cellular LDs and lipid metabolism in HeLa cells.[19]
Table 1. Structure–activity relationship study of benzopyranyl 1,4-disubstituted-1,2,3-triazole analogs by measuring lipid droplet (LD) organelle count with
SF44 and cell viability in HeLa human cervical cancer cells.
Cmpd
R1
R2
R3
LD count[a] [%]
Cell viab.[b] [%]
1
2
3
4
5
6
7
8
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
n-heptyl
cyclohexyl
cyclohexylmethyl
benzyl
2-phenylethyl
3-pyidyl
phenyl
4-methoxyphenyl
4-trifluoromethylphenyl
4-cyanophenyl
4-bromophenyl
2-methoxyphenyl
2-trifluoromethylphenyl
2-cyanophenyl
2-chlorophenyl
3-methoxyphenyl
3-trifluoromethylphenyl
3-cyanophenyl
121
86
121
101
87
106
113
112
97
119
115
99
90
110
67 (14)[d]
103
96 (92)[e]
97
9
10
11
12[c]
13
14
15
16
17
18
19
20
21
22
23
24
25
96
95
101
115
47 (8)[d]
110
106 (106)[e]
115
107
110
95
98
83
97
87
98
90
129
90
97
3-fluorophenyl
2-hydroxyphenyl
2-aminophenyl
2-(N-methyl)aminophenyl
2-(N,N-dimethyl)aminophenyl
2-nitrophenyl
71 (17)[d]
66 (11)[d]
115
111 (103)[e]
103 (99)[e]
102
116
107
82
107
2,5-dimethoxyphenyl
52 (15)[d]
101 (110)[e]
26
27
methyl
2-methoxyphenyl
2-methoxyphenyl
77
110
103
fluoromethyl
fluoromethyl
103
[a] Lipid droplet organelle count [%] of each compound (10 μM) was monitored by image-based high throughput assay in HeLa human cervical cancer cells
for 1 day. [b] Cell viability assay of each compound (10 μM) in HeLa cells for 1 day. [c] SB1501. [d] Lipid droplet organelle count [%] of each compound
(40 μM). [e] Cell viability assay of each compound (40 μM) in HeLa cells for 1 day.
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1105
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Figure 1. Discovery of SB1501 from phenotypic screening. a) Image-based high-throughput screening against an in-house compound library. Lipid droplet
(LD) organelle count [%] of all compounds (10 μM for 24 h) was monitored in HeLa cells. Data were normalized to a DMSO control. Initial hit compound 8
reduced cellular LDs down to 67% compared to what was found in DMSO-treated cells without significant cytotoxicity. b) Structure of initial hit 8 and SB1501
(12). c) Structural modifications of benzopyranyl 1,4-disubstituted-1,2,3-triazoles. R1 =methyl, fluoromethyl; R2 =methyl, fluoromethyl, ester group; R3 =n-
heptyl, cyclohexyl, cyclohexylmethyl, benzyl, 2-phenylethyl, 3-pyridyl, phenyl, and phenyl with various functional groups. d) Monitoring cellular LDs by
fluorescence microscopy. Cellular LDs and nuclei were stained with SF44 and Hoechst 33342, respectively, after treatment with DMSO, oleic acid (5 μM),
serum-free medium (starvation) or SB1501 (10 μM) for 24 h in HeLa cells. Scale bar: 5 μm. e) Quantification of cellular LDs from the images in (d). The scores
were determined by multiplying the number of LD per cell and mean LD intensity. Each value was normalized to the DMSO control. f) Determination of the
EC50 values of initial hit 8 and SB1501(12) in HeLa cells after treatment for 24 h. g) Cytotoxicity of SB1501 in HeLa cells. A cell viability assay was performed
after SB1501 treatment for 24 and 48 h. All data are presented as mean � SD; ***p<0.001 followed by one-way Dunnett’s test.
the 1,2,3-triazole ring. Analogs bearing alkyl moieties, such as n-
heptyl, cyclohexyl, cyclohexylmethyl, benzyl, and 2-phenylethyl
groups (1–5), at the R3 position did not show any significant LD
reducing activity. Then, we introduced (hetero)aryl groups (7–
11) at the R3 position, especially aryl moieties containing various
substituents, such as CF3, CN, and halogen, at the para position
of the phenyl ring to examine the importance of substituents.
Still, we did not find any analogs with higher potency than 4-
methoxyphenyl compound 8. Therefore, we examined these
substituents’ positional effects by synthesizing a series of
analogs (12–18) with all possible positional isomers (ortho and
meta). Independent of the position, we did not observe a
cellular LD reduction in analogs possessing electron-withdraw-
ing groups. To our pleasant surprise, compound 12 containing
an electron-donating methoxy substituent at the ortho position
of the phenyl group showed the best efficacy by reducing LDs
by 47% in HeLa cells at 10 μM (Figure 1b, d and e).
Furthermore, the incorporation of hydroxy (20) and amino
groups (21) at the ortho position of the phenyl and 2,5-
dimethoxy phenyl moieties (25) at the R3 position also showed
reductions in cellular LDs, but these compounds did not show
higher efficacy than compound 12. Also, under a higher
concentration (40 μM), compound 12 displayed the most drastic
reduction in LD organelle count (8%) without cytotoxicity, while
compounds 20, 21, and 23 showed 17, 15, and 11% reductions,
respectively (Table 1).
Keeping the 2-methoxyphenyl group at the R3 position, we
modified the substituents of compound 12 at the R1 and R2
positions by introducing
a functional handle (26) and
fluoromethyl group (27), but we observed no significant LD-
reducing activities. Lastly, the bioisosteric replacement of
triazole with isoxazole in the core skeleton was performed, but
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1106
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
the resulting compound (28) showed significant cytotoxicity
(Figure S2).
1
2
To sum up, our image-based high-throughput screening
identified the initial hit compound 8 containing benzopyranyl
1,2,3-triazole skeleton, which reduced the cellular LDs in HeLa
cells without significant cytotoxicity. Based on our SAR study of
initial hit compound 8, we discovered that the lead compound
12 containing ortho-positioned methoxy group on the phenyl
ring at the R3 position, dimethyl groups at the R1 and R2 position
on benzopyran, and triazole branch between two moieties,
showed the best efficacy in the cellular LD reduction. Com-
pound 12 efficiently reduced cellular LDs in a dose-dependent
manner in HeLa cells and showed the improved half-maximal
effective concentration (EC50) of 10.02 μM (24 h) compared to
that of initial hit compound 8 without significant cytotoxicity in
HeLa cells for 24 h and 48 h (Figure 1f, g). Therefore, we named
this lead compound 12 as SB1501 and pursued its further
biological evaluations.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
SB1501 reduces fat accumulation in 3T3-L1 adipocytes
We evaluated the LD-reducing activity of SB1501 in differ-
entiating 3T3-L1 adipocytes. 3T3-L1 adipocytes originated from
embryonic mouse fibroblasts are widely used to study mature
adipocytes in vitro.[21] Upon induction of differentiation, 3T3-L1
preadipocytes start to express the transcriptional regulators of
adipogenesis, which results in significant LD maturation via the
accumulation of excess fat into LDs (Figure 2a).[22] As an
excessive fat accumulation within adipocyte’s LDs is closely
associated with pathological conditions of obesity and obesity-
related metabolic diseases, targeting the adiposity of adipo-
cytes has been considered as a promising therapeutic approach.
To determine whether SB1501 can reduce fat accumulation in
3T3-L1 adipocytes, we treated 3T3-L1 adipocytes with SB1501
along the differentiation process and monitored the subse-
quent phenotypic changes. Under the continuous exposure of
SB1501 to the differentiating 3T3-L1 adipocytes for 8 days, 3T3-
L1 adipocytes displayed a significant reduction of cellular fat
accumulation at day 8 monitored by Oil Red O staining
(Figure 2b), which was confirmed by the biochemical quantifica-
tion of cellular triglycerides (TGs; Figure S3a). Besides, SB1501
reduced the mRNA and the protein expression level of
peroxisome proliferator-activated receptor-gamma (PPARγ),
regulating adipogenesis and lipogenesis (Figure S4a, b). As it
has been known that the inhibition of adipogenesis reduces fat
accumulation in adipocytes, we examined whether SB1501
inhibits the onset of differentiation at the induction stage of
differentiation.[23,24] When we treated SB1501 for 2 days at the
early stage of differentiation (day 0–day 2), we didn’t observe
the reduction in fat accumulation in 3T3-L1 adipocytes (Fig-
ure 2b). Moreover, SB1501 showed LD-reducing effects in differ-
entiated C2 C12 skeletal muscle cell line (Figure S4c). These
data imply that SB1501 may not perturb the onset of crucial
transcriptional regulators at the early stage of 3T3-L1 adipocyte
differentiation. Instead, SB1501 affects fat accumulation and LD
maturation in adipocytes.
Figure 2. SB1501 reduces the size and fat accumulation of LDs in 3T3-L1
adipocytes. a) Schematic diagram of the 3T3-L1 differentiation process. b)
Monitoring of fat accumulation within 3T3-L1 adipocytes after treatment of
differentiating cells with SB1501 (40 μM) for 2 days (D0–D2) or 8 days (D0–
D8). Oil Red O staining was performed on day 8. c) Representative
fluorescence images of 3T3-L1 adipocytes stained with SF44. Cells were
treated with SB1501 (40 μM) for different durations after the induction of
differentiation. D0-D2; 2-day treatment at the early stage of differentiation,
D6–D8; 2-day treatment at the late stage of differentiation, D0–D8; 8-day
treatment along the whole differentiation. The fluorescence images were
taken on day 8 after staining LDs with SF44. Scale bars: 15 μm. Quantification
of the d) mean area and e) fluorescence intensities of LDs from the images
in (c). Each dot represents a single cell. All data are presented as mean � SD;
***p<0.001 followed by one-way Dunnett’s test. ns, not statistically
significant.
To investigate its underlying mechanism of action, we
treated SB1501 at different treatment times after the induction
of differentiation. We observed phenotypic changes in differ-
entiating 3T3-L1 adipocytes by monitoring cellular LDs with
live-cell fluorescent microscopy after SF44 staining (Figure S3b).
Upon long-term treatment of SB1501 (days 0–8), differentiated
3T3-L1 adipocytes showed a significant reduction in the size of
multilocular LDs, compared to DMSO-treated cells as a negative
control (Figure 2c). Quantitative image analysis further revealed
that the size of every single LD and the fat accumulation within
LDs were reduced upon 8-day treatment with SB1501 (Fig-
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1107
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
ure 2d, e). In contrast, short-term 2-day treatment of SB1501 at
the early stage of differentiation (day 0 to day 2) did not display
significant changes in the phenotype of LDs. In contrast, cells
treated with SB1501 at the late stage of differentiation for
2 days (day 6 to day 8) showed more efficient LD-reducing
effects (Figure 2c–e). Moreover, the LD-reducing effects of
SB1501 showed a positive correlation with the duration of
SB1501 treatment (Figure S3c, d). Collectively, our phenotypic
observation suggested that SB1501 would reduce the size of
LDs and fat accumulation via the perturbation of lipid
metabolism in mature adipocytes.
Mechanism of action study of SB1501
1
2
3
4
5
6
7
8
9
LDs are vital subcellular organelles that regulate lipid metabo-
lism in response to cellular energy demands. LD size and its
energy storage capacity are tightly controlled by the balance
among lipogenesis, lipolysis, and fatty acid oxidation (FAO)
(Figure 3a).[25,26] Excessive energy sources are transformed into
TGs and transferred into LDs by lipogenesis. Under energy-
deficient conditions, lipolysis facilitates the hydrolysis of TGs
into free fatty acids and glycerol, and fatty acids (FAs) released
from LDs are utilized in mitochondrial FAO for ATP
generation.[27] In fact, both pathways have been considered as
therapeutic targets to overcome obesity and obesity-related
metabolic diseases.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Figure 3. Mechanism of action study of SB1501 in 3T3-L1 adipocytes. a) Schematic diagram of cellular LD regulation by lipogenesis, lipolysis, and
mitochondrial fatty acid oxidation (FAO). b) Quantitative analysis of mRNA expression in 3T3-L1 adipocytes. qRT-PCR was performed after 3T3-L1 adipocytes
were treated with SB1501 (40 μM) for 8 days. Fold expressions of each gene were normalized to DMSO control. c) Lipolysis assay in 3T3-L1 adipocytes upon
SB1501 treatment. Differentiating 3T3-L1 adipocytes were treated with SB1501 (40 μM) for 8 days. Glycerol release was measured by using a lipolysis assay kit
at day 8 with or without isoproterenol (ISO) stimulation (100 nM). d) Schematic diagram of lipolysis and its modulation with lipolysis inducer (ISO) and
inhibitor (atglistatin). ISO induces lipolysis thereby increasing glycerol release. Atglistatin reduces glycerol release by inhibiting the activity of lipase. e)
Monitoring glycerol release in 3T3-L1 adipocytes after treatment with ISO (100 nM, 2 h), atglistatin (1 μM, 6 h), or co-treatment of atglistatin with SB1501
(40 μM, 6 h). f) Effect of SB1501 and atglistatin on cellular LD accumulation. LDs of 3T3-L1 preadipocytes were monitored after treatment with SB1501 (20 μM),
atglistatin (100 nM), or co-treatment of atglistatin with SB1501 for 24 h. g) Western blot images of PGC-1α and proteins that regulate mitochondrial biogenesis
and the FAO pathway. Differentiating 3T3-L1 adipocytes were treated with SB1501 (40 μM) for 8 days. h) Representative immunofluorescence images of 3T3-
L1 adipocytes stained with UCP1 (green) and MitoTracker Red (red). Scale bar: 15 μm. Quantified fluorescence intensity of i) UCP1 and j) MitoTracker Red from
the imaged in (h). Each dot represents a single cell. All data are presented as mean � SD; *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001 followed by
unpaired Student’s t-test (b, i, j) and one-way Dunnett’s test (e, f). ns: not statistically significant.
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1108
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
To understand the mechanism of action underlying the LD-
blot analysis revealed that the 8-day treatment of SB1510
increased the protein levels of PGC-1α, COX IV, and CPT1 in
3T3-L1 adipocytes (Figure 3g). In contrast, SB1501 reduced the
protein levels of fat-specific protein 27 (FSP27) in a dose-
dependent manner consistent with inhibition at the mRNA
level. Notably, the expression of mitochondrial uncoupling
protein 1 (UCP1) was upregulated upon SB1501 treatment in a
dose-dependent manner as a downstream effect of PGC-1α
upregulation (Figure 3g). UCP1, a downstream protein of PGC-
1α, is the gold standard protein marker for adaptive thermo-
genesis in adipocytes.[45,46] UCP1 is localized in the inner
mitochondrial membrane and functions to uncouple oxidative
phosphorylation. The induction of UCP1 increases cellular
energy expenditure by dissipating fatty acids as heat, which
leads to reduced fat accumulation.[47] Consistent with the data
above, immunofluorescence analysis further demonstrated the
upregulation of UCP1 along with its co-localization of Mito-
Tracker upon SB1501 treatment (Figure 3h–j). Together, these
data suggest that SB1501 increases mitochondrial biogenesis
and FAO with upregulation of PGC-1α-UCP1 regulatory axis,
which leads to reduced fat accumulation and demand for lipid
storage capacity in 3T3-L1 adipocyte.
1
2
3
4
5
6
7
8
9
reducing effect of SB1501, we examined whether SB1501
influences the expression levels of regulatory genes for lipid
metabolism in differentiating 3T3-L1 adipocytes. We first
examined the mRNA expression level of peroxisome prolifer-
ator-activated receptor-gamma coactivator-1-alpha (PGC-1α)
and its downstream proteins involved in mitochondria bio-
genesis and FAO. PGC-1α is a transcriptional co-activator and
one of the main regulators of energy metabolism in mammalian
cells.[28] PGC-1α regulates metabolic signaling pathways, includ-
ing adaptive thermogenesis, mitochondria biogenesis, and
FAO.[29–32] As shown in Figure 3b, we observed the elevated
mRNA levels of regulatory proteins for mitochondrial biogenesis
(PGC-1α, PPARα, and TFAM) and FAO (UCP1, CPT1, CD36, and
LPL), compared to those of DMSO-treated cells. The mRNA level
of Fsp27 was also reduced, which mediates the fusion of cellular
LDs in response to increased cellular lipid accumulation.[33]
SB1501 significantly suppressed the mRNA level of resistin, an
adipokine known to promote obesity-related insulin
resistance.[34] In contrast, the mRNA level of adiponectin
(AdipoQ), which enhances insulin sensitivity,[35] was not signifi-
cantly changed. Meanwhile, the mRNA level of SREBP-1, a
transcription factor that facilitates lipogenesis, showed little
difference compared to DMSO-treated cells, but the mRNA level
of SCD1 was somewhat decreased.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
SB1501 improves mitochondrial function
Next, to investigate whether SB1501 induces lipolysis, we
performed a lipolysis assay in 3T3-L1 adipocytes.[36] Since the
pharmacological induction of lipolysis results in the breakdown
of TGs into FAs and the decrease of TG accumulation within
We then investigated the correlation between the LD-reducing
activity of SB1501 and mitochondrial morphology in 3T3-L1
adipocytes. Cellular energy expenditure is closely linked to the
regulation of mitochondria fusion and fission dynamics.[48]
Especially, in vitro studies with 3T3-L1 cells revealed the
correlation between TG accumulation and mitochondrial
dynamics.[49,50] PGC-1α is known to regulate both mitochondrial
biogenesis and morphology along with its downstream regu-
lators. To investigate whether SB1501 perturbs mitochondrial
morphology through PGC-1α upregulation, we examined the
architectural changes of mitochondria in differentiated 3T3-L1
adipocytes upon treatment with SB1501 via live-cell imaging.
As shown in Figure 4a, high-resolution imaging of mitochondria
and LDs in live cells showed a significant increase in mitochon-
dria fusion along with the decreased LD accumulation upon 8-
day treatment of SB1501. Quantitative analysis of images
revealed that the average volume of the mitochondrial network
was increased in SB1501-treated cells, compared to the DMSO-
treated cells, while the number of mitochondrial networks per
cell remained the same (Figure 4b, c). SB1501 also increased
mitochondrial fusion in a time-dependent manner in 3T3-L1
adipocytes (Figure S6) as well as in HeLa cells (Figure S7), which
is consistent with its LD-reducing effect.
LDs,
a direct effect on adipocyte lipolysis could be of
therapeutic interest for reducing obesity and improving insulin
resistance.[37–39] Therefore, we hypothesized that SB1501 might
increase the release of glycerol by inducing lipolysis similar to
the effect of isoproterenol (ISO), a β-adrenergic receptor agonist
that activates hormone-sensitive lipases (Figure 3d).[40] The
treatment of differentiated 3T3-L1 adipocytes with ISO (100 nM)
drastically enhanced the released amount of glycerol, compared
to the DMSO-treated group (basal lipolysis, Figure 3c). However,
the SB1501-treated group showed a reduction in the glycerol
release in a dose-dependent manner with or without ISO
stimulation (Figure 3c). In contrast, atglistatin, a lipolysis inhib-
itor that blocks TG degradation,[41] reduced the glycerol release
in differentiated 3T3-L1 adipocytes while significantly elevating
the cellular LD accumulation (Figure 3e, f). Besides, the co-
treatment of atglistatin with SB1501 reversed the fat-accumulat-
ing effect of atglistatin (Figure 3f and S5). These data showed
that SB1501 reduces the cellular LD accumulation via a distinct
mechanism of action different from lipolysis inducer (ISO) and
inhibitor (atglistatin).
Next, we further investigated whether SB1501 perturbs the
expression of regulatory proteins for mitochondrial biogenesis
and FAO. Several studies have revealed that the induction of
mitochondrial biogenesis and FAO can reverse fat accumula-
tion, which may contribute to overcoming obesity-related
metabolic disorders.[42–44] Based on our mRNA level analysis, we
examined the protein expression of PGC-1α and its downstream
proteins involved in mitochondria biogenesis and FAO. Western
We then performed bioenergetic studies using a Seahorse
XF Analyzer to assess the effect of SB1501 on mitochondrial
function in differentiating 3T3-L1 adipocytes. Upon 8-day treat-
ment of SB1501, we observed a significant increase in the
maximal mitochondrial respiration compared to the DMSO-
treated cells, in accordance with elevated oxygen consumption
rate by upregulated level of UCP1 (Figure 4d).[51–53] In addition,
the copy number of mitochondrial DNA (mtDNA) was increased
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1109
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
with a half-life (t1/2) of 8.6 h in male ICR mice (Figure S8; PK
1
2
3
4
5
6
7
8
9
study). We then orally administered SB1501 (10 mg/kg) to db/db
mice daily for 5 weeks. For the control groups, rosiglitazone
(15 mg/kg) and vehicle were orally administrated in parallel.[54]
As shown in Figure 5a, bodyweight profiles showed significant
weight-loss effects in the SB1501-treated group. The SB1501-
treated group started to show significant weight-loss from day
27, resulting in an average bodyweight-loss of 5.2 g at the end
of 5-weeks of SB1501 administration without significant change
in the food intake (Figure 5b, c). In contrast, the vehicle- and
rosiglitazone-treated groups showed weight-increases of 2.5 g
and 11.6 g, respectively (Figure 5b). We observed significant
weight-loss in epicardial white adipose tissue (EWAT) and
marginal loss in brown adipose tissue (BAT) upon SB1501
treatment (Figure 5d).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Figure 4. Cellular effects of SB1501 on mitochondrial function and morphol-
ogy. a) Representative fluorescence images of 3T3-L1 adipocytes stained
with SF44 for LDs (green) and MitoTracker Red for mitochondria (red) upon
treatment with SB1501 (40 μM) for 8 days. Scale bars: 2 μm. Quantification of
b) average volume and c) number of mitochondrial networks per cell from
the images in (a). Each dot represents a single cell. d) Measurement of
oxygen consumption rate (OCR) by using Seahorse XFe24 Analyzer [Seahorse
Bioscience] in 3T3-L1 adipocytes after treatment of differentiating 3T3-L1
cells with SB1501 (40 μM) or DMSO for 8 days. N=7 for each group. e)
Quantification of mitochondrial DNA (MitoDNA) in 3T3-L1 adipocytes treated
with SB1501 (40 μM) for 8 days. All data are presented as mean � SD. n=3
biologically independent experiments. *p<0.05, ***p<0.001 followed by
one-way Dunnett’s test (b,c) or unpaired Student’s t-test (d,e,f). ns: not
statistically significant.
by 1.9-fold (p=0.0009), which further suggested that SB1501
induces mitochondrial biogenesis (Figure 4e). These observa-
tions suggest that SB1501 would increase the cellular energy
expenditure of adipocytes by elevating mitochondrial bio-
genesis and perturbing mitochondrial dynamics.
Figure 5. Improvement of metabolic profiles in a db/db mouse model after
SB1501 administration. a) Monitoring body weight in db/db mice from day 1.
N=3 for each group. b) Bodyweight change on day 34. c) Food intake per
day in db/db mice. Food intake of each group was measured 8 times
between day 20 and day 30, and averaged (n=3). Each dot represents an
average food intake per day per group. d) Measurement of adipose tissue
weight on day 34. IWAT: inguinal white adipose tissue; EWAT: epicardial
white adipose tissue; BAT: brown adipose tissue. e) Glucose tolerance tests
(GTTs) in db/db mice after administration of vehicle (Veh), rosiglitazone (Rosi,
15 mg/kg), or SB1501 (low: 10 mg/kg, high, 20 mg/kg) for 34 days. f)
Quantification of the area under the curve (AUC) for the GTTs. All data are
presented as mean � SEM; *p<0.05, **p<0.01, ***p<0.001 followed by
one-way Dunnett’s test. ns: not statistically significant.
Anti-obesity effects of SB1501 in a db/db mouse model
Next, we investigated whether SB1501 shows anti-obesity
effects in vivo. When we first performed a pharmacokinetic
study to decide the administration route, oral administration of
SB1501 (10 mg/kg) showed 42.9% oral bioavailability and an
average maximum plasma concentration (Cmax) of 0.89 (μg/mL)
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1110
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
Insulin resistance is a major metabolic characteristic of
SB1501 induces a browning process in IWAT and BAT
1
2
3
4
5
6
7
8
9
obesity and acts as a major factor in the pathogenesis of many
obesity-related diseases, including type 2 diabetes.[55] Therefore,
anti-obesity therapeutics could exhibit beneficial effects on
obesity-induced insulin resistance. To assess the effect of
SB1501 on insulin resistance, we performed glucose tolerance
tests (GTTs) on day 34 of treatment. Oral administration of
10 mg/kg (low) and 20 mg/kg (high) SB1501 exhibited a
significant reduction in blood glucose level compared to the
vehicle-treated group (Figure 5e). Accordingly, the net AUC was
significantly reduced in the SB1501-treated group (Figure 5f).
Moreover, the SB1501-treated group with a high dosage
(20 mg/kg) ameliorated glucose tolerance to a level similar to
that of rosiglitazone (Figure 5f). Together, SB1501 reduced
adipose tissue mass and bodyweight while improving glucose
tolerance in the db/db mouse model. These data suggest that
SB1501 has therapeutic potential as an anti-obesity agent for
the treatment of obesity and obesity-related insulin resistance.
Finally, we analyzed the phenotypic changes in adipose tissues
upon SB1501 treatment. Histological analysis of adipose tissues
showed reduced LD sizes in both inguinal white adipose tissue
(IWAT) and BAT in the SB1501-treated group compared to the
vehicle- and rosiglitazone-treated groups (Figure 6a). In con-
trast, the rosiglitazone-treated group showed an increase in LD
sizes in adipose tissues along with weight-increase (Figure 6a).
Immunofluorescence imaging showed the upregulated expres-
sion of PGC-1α and UCP1 in both adipose tissues along with
reduced LD sizes (Figure 6b, c). Of note, the induction of UCP1
in IWAT increases their thermogenic capacity by dissipating
fatty acids as heat, which is called the process of browning of
IWAT.[56] IWAT and BAT are metabolically active and mainly
utilize lipids driven by upregulation of UCP1.[57] In addition, the
formation of small multilocular LDs is a representative pheno-
type of the browning process.[58] These data suggest that
SB1501 induces thermogenesis of adipose tissues through
upregulation of the PGC-1α–UCP1 regulatory axis, thereby
reducing fat accumulation. We further analyzed the gene
expression alterations in both adipose tissues. As expected,
SB1501 significantly increased the mRNA levels of Pgc-1α and
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Figure 6. Changes in adipose tissue morphology and protein/mRNA expression in a db/db mouse model. a) Changes in adipose tissue morphology were
monitored by H&E staining. b), c) Representative immunofluorescence images of adipose tissues stained with PGC-1α and UCP1. Scale bars: 25 μm. d), e)
Quantitative analysis of mRNA expression in adipose tissues. Fold expressions of each gene were normalized to the vehicle-treated group. All data are
presented as mean � SEM; *p<0.05, **p<0.01, ***p<0.001 followed by one-way Dunnett’s test.
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1111
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
Ucp1 in IWAT (Figure 6d). Along with the upregulation of the
brown adipocyte marker (Cidea), the genes for mitochondrial
biogenesis and (Cox IV) and FAO (Acox1 and Cpt1b) were
increased in IWAT. Cox IV and Cidea are also known as the
thermoregulatory gene. In BAT, we observed significant
increases in the mRNA level of Ucp1 in the SB1501-treated
group; however, Pgc-1α was not significantly perturbed (Fig-
ure 6e). Nonetheless, increases in the expression levels of genes
regulating mitochondrial biogenesis and FAO in BAT were
evident. We also observed that SB1501 reduces fat accumu-
lation and upregulates the PGC-1α signaling pathway in liver
tissue (Figure S9). These data collectively show that SB1501
reduces fat accumulation in the multiple key metabolic tissues,
including liver and adipose tissues, via perturbing the PGC-1α–
UCP1 signaling pathway.
EC50 value in HeLa cells. We also demonstrated the LD-reducing
effect of SB1501 in differentiating 3T3-L1 adipocytes. Interest-
ingly, SB1501 reduced fat accumulation of 3T3-L1 adipocytes in
1
2
3
4
5
6
7
8
9
a
treatment time-dependent manner and showed better
efficacy under treatment at the late stage of differentiation. In
the process of revealing the SB1501’s mechanism of action, we
identified that the reduction of cellular LD accumulation by
SB1501 is mainly associated with the perturbation of mitochon-
drial biogenesis and FAO. Mitochondrial FAO uses fatty acids as
a cellular energy source and is regarded as a potential
therapeutic target for obesity and obesity-related diseases.
Long-term treatment of SB1501 in differentiating 3T3-L1
adipocytes upregulated the mRNA and protein expression of
PGC-1α and its downstream proteins, mainly involved in
mitochondria biogenesis (PPARα and TFAM) and FAO (UCP1,
CPT1, CD36, and LPL). Consistent with those data, SB1501
increased both mitochondrial mass and maximal respiration. In
addition, we observed an increase in mitochondrial fusion,
consistent with the previously reported correlation between
mitochondrial fusion and cellular fat reduction.[65] Together, our
data show that SB1501 reduces cellular fat accumulation by
elevating mitochondrial biogenesis and FAO via the PGC-1α–
UCP1 signaling pathway.
We then demonstrated the therapeutic potential of SB1501
as an anti-obesity agent in the db/db mouse model. Oral
administration of SB1501 significantly reduces body weight,
compared to the vehicle-treated group. In addition to its
weight-loss effect, SB1501 improved glucose tolerance, which is
related to obesity-induced insulin resistance. Histopathological
analysis of the adipose tissues also revealed that the treatment
of SB1501 notably decreased the size of LDs in both IWAT and
BAT, along with the formation of multiple small LDs in IWAT.
The size and fat accumulation of LDs in IWAT and BAT are
strongly linked to the induction of UCP1-mediated thermo-
genesis, which is primarily regulated by PGC-1α. Treatment of
SB1501 enhanced the mRNA and protein expression of PGC-1α
and UCP1 in adipose tissues, consistent with the in vitro studies
in 3T3-L1 adipocytes. These data show that SB1501 upregulates
UCP1, which dissipates energy as heat and plays a key role in
the thermogenesis of mitochondria, resulting in increased FAO
and reduced fat accumulation in adipose tissues. Furthermore,
SB1501 significantly increased the mRNA expression levels of
thermogenesis-related genes and brown adipocyte markers,
including UCP1, PGC-1α, Cidea, and Cox IV in IWAT, which
suggests that SB1501 upregulates UCP1-mediated thermogene-
sis and induces browning process in IWAT. Collectively, we
found that SB1501 showed an anti-obesity effect by activating
adipose tissues via the upregulation of PGC-1α and UCP1 as
well as beneficial effects on liver fat accumulation in a db/db
mouse model. At present, it is unclear whether SB1501
exhibited direct pharmacological effects in the liver or its LD-
reducing effect in adipose tissues results in a reduction of
ectopic fats in the liver. Nevertheless, our data show that
SB1501 has the potential to improve obesity-induced systemic
insulin resistance in vivo.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Conclusion
Obesity is defined as a state of increased body weight, and
chronic imbalance between energy intake and consumption is
the main driver of obesity. Current treatment strategies for
obesity predominantly focus on decreased energy intake and
preventing the absorption of food. Appetite suppressants, such
as phentermine, phendimetrazine, diethylpropion, and mazin-
dol, cause a decrease in appetite by increasing the level of
neurotransmitters that control hunger and fullness within the
central nervous system.[59] Besides, orlistat is a well-known
prescribed obesity medication that suppresses lipase activities
in the stomach and small intestine and blocks the absorption of
TGs into fatty acids, which are absorbed into the intestinal tract,
thereby showing a weight-loss effect.[60] However, current anti-
obesity agents (appetite suppressors and lipase inhibitors) have
limited effectiveness and undesirable side effects, including
insomnia, dizziness, headaches, vomiting, and oily bowel
movements.[61] Therefore, many anti-obesity therapeutics target-
ing different mechanisms are currently under investigation.[62,63]
Mitochondria are the main organelles for energy production,
and promoting mitochondrial activity is currently used in the
anti-obesity approach.[64] However, small molecules targeting
mitochondria remain challenging due to complex mechanisms,
and only a few have anti-obesity efficacy with a suitable
therapeutic window.
In this study, our research focused on enhancing cellular
energy expenditure as an alternative strategy for developing a
new anti-obesity agent. We aimed to discover a novel bioactive
compound that reduces cellular LD accumulation. For the
primary screening, we performed image-based high-throughput
screening to monitor cellular LDs with fluorogenic LD bioprobe
(SF44) in HeLa cells. Among 3,000 small molecules from an in-
house pDOS library, we discovered hit compound 8 containing
benzopyran-substituted 1,2,3-triazole core skeleton, which
reduces cellular LDs by 67% without cytotoxic effect. Through
the systematic SAR study, we discovered the lead compound
12, namely SB1501, containing an electron-donating methoxy
substituent at the ortho position of the phenyl group. SB1501
afforded the best LD-reducing efficacy with a twofold higher
In conclusion, our study demonstrated that SB1501 has the
potential as an anti-obesity agent. Our findings showed that
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1112
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
33342 [Thermo Scientific] were added to individual wells. After
SB1501 improves obesity and obesity-induced insulin resistance
in vivo. However, it still remains unclear what is the molecular
target of SB1501. To understand its exact mechanism of action,
we pursued the target identification of SB1501 with both a
photoaffinity-based approach using FITGE[66,67] and a label-free
approach using thermal stability shift, namely TS-FITGE,[68,69] but
we haven’t yet revealed the exclusive target proteins of SB1501,
probably due to its low efficacy, tight SAR, and complex
mechanism associated with multiple target proteins. The
extensive target identification and validation of SB1501 are
currently ongoing, and these results will be reported in due
course.
1
2
3
4
5
6
7
8
9
°
30 min incubation at 37 C, without washing, fluorescence images
of each well were acquired with InCell Analyzer 2000 [GE Health-
care]. Fluorescence images were captured using excitation/emission
filters at CFP/Cy3 channels for SF44 and DAPI channel for Hoechst
33342. Fluorescence intensities were analyzed using the InCell
Analyzer 1000 workstation 3.6 program [GE Healthcare]. Hit
compounds reducing cellular LD organelle count were identified by
calculating relative LD organelle count [%] for each well compared
to the DMSO control well. In parallel, cell viability screening was
performed to assess the cytotoxicity of each compound with WST
assay kit [DoGenBio].
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
High-resolution imaging of LDs in 3T3-L1 adipocytes. 3T3-L1
preadipocytes were seeded on glass-bottom confocal dishes
[Corning]. After reaching 100% confluence, cells were differentiated
as described above. During the differentiation, cells were treated
with SB1501 (40 μM) at different duration after the induction of
differentiation while maintaining the cells with fresh DMEM
containing 10% FBS and 5 μg/mL insulin. After 8 days, cells were
stained with SF44 (20 μM) in DMEM containing 10% FBS and 5 μg/
mL insulin for 30 min. After washing with phosphate-buffered saline
(PBS) twice, DMEM containing 10% FBS and 5 μg/mL insulin was
added for live-cell imaging. Fluorescence microscopy imaging was
performed using the DeltaVision imaging system equipped with
60× objective lenses of Olympus IX-71 inverted microscope under
Experimental Section
Cell lines. HeLa human cervical cancer cells and 3T3-L1 cells (mouse
embryonic fibroblasts) were obtained from American Type Culture
Collection. HeLa cells were maintained in RPMI medium containing
10% fetal bovine serum (FBS) and 1% antibiotic-antimycotic
°
solution [GIBCO] at 37 C in an atmosphere of 5% CO2. 3T3-L1 cells
were maintained in Dulbeco’s modified Eagle’s medium (DMEM)
containing 10% heat-inactivated bovine calf serum (BCS) [GIBCO]
°
and 1% antibiotic-antimycotic solution at 37 C in an atmosphere of
5% CO2.
°
5% CO2 at 37 C. Every single LD within a cell was identified from
the fluorescence images, then the mean area and fluorescence
intensity were analyzed for the quantitative analysis of images
using InCell Developer Toolbox v1.9.3. [GE Healthcare].
Animal administration. Ten-week-old C57BLKS/J-Leprdb/Leprdb (db/
db) male mice were obtained from Central Lab Animal Inc. (Seoul,
Korea) and were housed in 12-h light/12-h dark cycles. Once they
were 11 weeks of age, rosiglitazone (oral gavage, 15 mg/kg/day)
and SB1501 (oral gavage, 10 and 20 mg/kg/day) were administrated
for 5 weeks. All animal experiments were performed in accordance
with the research guidelines of the Seoul National University
Institutional Animal Care and Use Committee.
Quantitative reverse transcription PCR (qRT-PCR). 3T3-L1 cells
were differentiated as described above. During the differentiation,
cells were treated with 40 μM of SB1501 for 8 days while
maintaining the cells with fresh DMEM containing 10% FBS and 5
μg/mL insulin. After 8 days, total RNA was isolated using RNA
isolation kit [Qiagen]. Total RNA of adipose tissues were also
isolated as described above. cDNA was synthesized using a reverse
transcriptase kit and oligo(dT) primer according to the manufac-
turer’s instruction [Bioneer]. Quantitative real-time PCR was per-
formed with cDNA, primers, and SYBR Green [Kapa Biosystems]. The
Applied BiosystemsTM StepOnePlus real-time PCR system was used
for qRT-PCR. All the primers for qRT-PCR were obtained from
Bioneer.
3T3-L1 differentiation. 3T3-L1 preadipocyte cells were seeded on
°
culture plates and incubated for 2 days at 37 C in an atmosphere
of 5% CO2. After reaching 100% confluence, differentiation was
induced by treating the cells with 1 μM dexamethasone, 10 μM
rosiglitazone, and 5 μg/mL insulin in DMEM media containing 10%
FBS for 2 days. Cells were maintained for 8 days by exchanging
with fresh DMEM containing 10% FBS and 5 μg/mL insulin every
2 days.
Western blot analysis. 3T3-L1 cells were differentiated as described
above. During the differentiation, cells were treated with 40 μM of
SB1501 for 8 days while maintaining the cells with fresh DMEM
containing 10% FBS and 5 μg/mL insulin. After 8 days, cell lysates
were obtained by RIPA buffer [Cell Signaling Technology] contain-
ing 1× protease inhibitor cocktail [Roche]. The proteome was
analyzed by SDS-PAGE and transferred into the PVDF membrane.
After blocking with 2% BSA in Tris-buffered saline containing
Tween 20 (TBST) for 1 h, membranes were treated with the desired
Pharmacokinetics study. The following procedure obtained phar-
macokinetic data. We tested two different routes of administration:
intravenous (iv) injection at a concentration of 10 mg/kg (n=1) and
oral treatment at a dose of 10 and 20 mg/kg in ICR male mice (n=
2). SB1501 was prepared as a solution (DMSO, PEG400, and distilled
water at 5:47:50, v/v/v %). In oral treatment, we obtained blood
samples at 30 min, 1, 2, 4, 8, and 24 h after oral administration. In
the case of iv injection, blood samples were taken at 5 min, 30 min,
1 h, 2 h, 4 h, 8 h, and 24 h after injection. After the centrifugation of
plasma for purification, the concentration of SB1501 was analyzed
using Agilent 6460 LC/MS/MS system [Agilent] using electron spray
ionization. Pharmacokinetic parameters were obtained after the
analysis of the plasma concentrationÀ time plot with WinNonlin
software [Pharsight].
°
primary antibody (1:1000 in 1% BSA-TBST) overnight at 4 C. After
washing with TBST 3 times, membranes were exposed to HRP-
conjugated secondary antibody (1:2000 in 1% BSA-TBST) for 1 h at
room temperature and developed by ECL Prime [GE Healthcare].
Chemiluminescent signal was scanned by ChemiDocTM MP imaging
system [BioRad].
Immunofluorescence imaging of 3T3-L1 cells. 3T3-L1 preadipo-
cytes were seeded on glass-bottom dishes. After reaching 100%
confluence, cells were differentiated as described above. During the
differentiation, cells were treated with 40 μM of SB1501 for 8 days
while maintained in fresh DMEM containing 10% FBS and 5 μg/mL
insulin. After 8 days, cells were stained with Mitotracker Red
(200 nM) in DMEM containing 10% FBS for 30 min. After washing
Image-based high-throughput screening. HeLa cells were cultured
on black 96-well plates for 1 day. Small molecules (10 μM) were
added to the designated wells using a pin tool and incubated for
24 h. Oleic acid stock (100 mM) was prepared in isopropanol, and
freshly dissolved to 10 mM in DMSO before use. Oleic acid (5 μM)
and serum-free media were used as positive/negative controls.
After exchanging with fresh medium, SF44 (5 μM) and Hoechst
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1113
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
with phosphate-buffered saline (PBS), cells were incubated with
fixation buffer (3.7% paraformaldehyde in PBS) for 15 min and
permeabilization buffer (0.1% TritonX-100 in PBS) for 10 min. After
blocking the cells with 4% BSA in PBS for 1 h at room temperature,
the anti-UCP1 antibody (1:200 in 1% BSA-PBS) was incubated
Measurement of oxygen consumption rate. The oxygen consump-
tion rate (OCR) of 3T3-L1 adipocytes was measured using a
Seahorse XFe24 extracellular flux analyzer [Agilent] according to
the manufacturer’s protocols. Briefly, 3T3-L1 cells were seeded on
XF24 cell culture microplates and differentiated as described above
in the presence or absence of SB1501 (40 μM). To measure the
mitochondrial respiration, 1.5 μM of oligomycin, 0.5 μM of FCCP
1
2
3
4
5
6
7
8
9
°
overnight at 4 C. After washing with PBS three times, cells were
incubated with FITC-conjugated secondary antibody (1:200 in 1%
BSA-PBS) for 2 h at room temperature. After washing 5 times with
PBS, the nucleus was stained with Hoechst 33342 and washed out.
Fluorescence microscopy imaging was performed using the
DeltaVision imaging system equipped with 60× objective lenses of
Olympus IX-71 inverted microscope. The temperature was main-
(carbonyl
cyanide-p-trifluoromethoxy-phenylhydrazone),
and
0.5 μM of rotenone/antimycin A were added through the injection
ports. During Seahorse analysis, 5 μg/mL insulin was added to the
assay media. BCA protein assay was performed to measure the
protein concentration of cell lysates. Data were normalized with a
lysate protein concentration of each sample.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
°
tained at 23 C.
Lipolysis assay. 3T3-L1 preadipocytes were seeded on 12-well
plates [Nunc]. After reaching 100% confluence, cells were differ-
entiated as described above. At day 8, cells were washed with
lipolysis wash buffer and incubated with lipolysis assay buffer
Glucose tolerance tests. For the glucose tolerance tests (GTTs),
overnight (16 h) fasted db/db mice were orally injected with glucose
(1 g/kg, 20% glucose solution). Blood samples were drawn at 15,
30, 60, 90, and 120 min after glucose injection by taking 3 μL of
blood from the tip of the tail vein.
°
containing isoproterenol (ISO, 100 nM) for 2 h at 37 C in an
atmosphere of 5% CO2. After 2 h, lipolysis assay buffers were
harvested into 1.5 mL ep-tube and mixed the lipolysis reaction
mixture containing glycerol probe and glycerol enzyme mix. A
lipolysis assay was performed with Lipolysis assay kit [Abcam]
according to the manufacturer’s instruction. BCA protein assay
[ThermoFisher Scientific] was performed to measure the protein
concentration of the cell lysates. Absorbance values obtained from
lipolysis assay were normalized with a lysate protein concentration
of each sample.
Immunohistochemistry of adipose tissues. The adipose tissues
were isolated from mice, fixed in 4% paraformaldehyde, and
embedded in a paraffin block. The paraffin blocks were trimmed (5-
μm-thick) and mounted onto a glass slide. The tissue sections were
dewaxed in xylene and rehydrated in a graded series of ethanol
solutions (100, 90, 80, and 70%). For antigen retrieval, the slides
were incubated in boiling sodium citrate buffer (10 mM, pH 6.0) for
30 min. The antigen retrieval step was skipped for slides stained
with hematoxylin and eosin. After blocking the slices with 5% horse
serum, sections were stained with primary antibodies against PGC-
1α (sc-13067; Santa Cruz Biotechnology) and UCP1 (ab10983;
Abcam). Secondary staining with anti-rabbit Alexa Fluor 594 goat
antibody (A32740; Invitrogen) was followed. The sections were
counterstained with hematoxylin and eosin dye. Fluorescence
microscopy imaging was performed using the DeltaVision imaging
system equipped with 60× objective lenses of Olympus IX-71
Quantification of triglyceride concentration. 3T3-L1 cells were
differentiated as described above. During the differentiation, cells
were treated with 40 μM of SB1501 for 8 days while maintained in
fresh DMEM containing 10% FBS and 5 μg/mL insulin. After 8 days,
cells were homogenized with 5% NP-40 in ddH2O. Triglyceride
concentration within the homogenized cells was measured with
Triglyceride Quantification Assay Kit [Abcam] according to the
manufacturer’s instruction.
°
inverted microscope. The temperature was maintained at 23 C.
Quantification of mitochondrial DNA. 3T3-L1 cells were differ-
entiated as described above. During the differentiation, cells were
treated with 40 μM of SB1501 for 8 days while maintained in fresh
DMEM containing 10% FBS and 5 μg/mL insulin. After 8 days,
Statistical analysis. The results are shown as mean � SD for in vitro
studies and SEM for in vivo studies. The unpaired student’s t-test
and one-way Dunnett’s test were performed by GraphPad Prism8. p
values of <0.05 were considered significant.
genomic DNA was prepared with
a DNA prep kit [Qiagen]
according to the manufacturer’s instruction. 10 ng of gDNA was
mixed with a qPCR reaction mixture containing 1× KAPA SYBR
FAST qPCR Master Mix, 200 nM of forward/reverse primers for 16S
rRNA, or hexokinase 2 for a final volume of 20 μL. The PCR
Author Contributions
°
temperature cycling was used; initial denaturing at 95 C for 1 min,
A.J., M.K., and S.B.P. conceived the idea and designed the
research outline. M.K. designed and synthesized all compounds.
A.J., J.H., and H.N. executed the phenotypic screening, bioimag-
ing, immunohistochemistry protocol, and western blot. J.-I.K.,
I.J.H., and J.B.K. designed and performed all animal experiments.
A.J., M.K., and S.B.P. wrote the manuscript. All authors
contributed to the discussion and approved the final version of
this manuscript.
°
followed by 40 cycles of denaturing at 95 C for 15 s, and extension
°
at 60 C for 1 min. The copy number of 16S rRNA was normalized
by the copy number of hexokinase 2.
Quantitative analysis mitochondrial network in 3T3-L1 adipo-
cytes. 3T3-L1 preadipocytes were seeded on glass-bottom confocal
dishes. After reaching 100% confluence, cells were differentiated as
described above. During the differentiation, cells were treated with
40 μM of SB1501 for 8 days while maintaining the cells with fresh
DMEM containing 10% FBS and 5 μg/mL insulin. After 8 days, cells
were stained with MitoTracker Deep Red (200 nM), SF44 (20 μM),
and Hoechst 33342 (2 μg/mL) in DMEM containing 10% FBS and 5
μg/mL insulin for 30 min. Carbonyl cyanide m-chlorophenylhydra-
zone (CCCP) was treated for 2 h on day 8. After washing with PBS
twice, DMEM containing 10% FBS and 5 μg/mL insulin was added
for live-cell imaging. Fluorescence microscopy imaging was
performed using the DeltaVision imaging system equipped with
60× objective lenses of Olympus IX-71 inverted microscope under
Acknowledgements
This work was supported by National Creative Research Initiative
Grants (2014R1A3A2030423 to S.B.P.; 2020R1A3B2078617 to J.B.K.)
and the Bio
& Medical Technology Development Program
(2012M3A9C4048780 to S.B.P.) through the National Research
Foundation of Korea (NRF) funded by the Korean Government
(Ministry of Science & ICT). A.J., M.K., and H.N. are grateful to the
°
5% CO2 at 37 C. Quantification of the mitochondrial network was
performed with 3D Object Counter of Image J software.
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1114
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/cmdc.202100062
ChemMedChem
[38] P. Morigny, M. Houssier, E. Mouisel, D. Langin, Biochimie 2016, 125, 259–
266.
[39] M. Schweiger, M. Romauch, R. Schreiber, G. F. Grabner, S. Hutter, P.
Kotzbeck, P. Benedikt, T. O. Eichmann, S. Yamada, O. Knittelfelder, C.
Diwoky, C. Doler, N. Mayer, W. De Cecco, R. Breinbauer, R. Zimmermann,
R. Zechner, Nat. Commun. 2017, 8, 14859.
BK21 Fellowship Program. J.H. is grateful to the Fostering Core
Leaders of the Future Basic Science Program/Global Ph.D. Fellow-
ship Program (NRF-2016H1A2A19 07084).
1
2
3
4
[40] F. Zalatan, J. A. Krause, D. E. Blask, Endocrinology 2001, 142, 3783–3790.
[41] N. Mayer, M. Schweiger, M. Romauch, G. F. Grabner, T. O. Eichmann, E.
Fuchs, J. Ivkovic, C. Heier, I. Mrak, A. Lass, G. Hofler, C. Fledelius, R.
Zechner, R. Zimmermann, R. Breinbauer, Nat. Chem. Biol. 2013, 9, 785–
787.
[42] M. E. Patti, S. Corvera, Endocr. Rev. 2010, 31, 364–395.
[43] E. E. Blaak, D. P. van Aggel-Leijssen, A. J. Wagenmakers, W. H. Saris, M. A.
van Baak, Diabetes 2000, 49, 2102–2107.
[44] H. Zong, J. M. Ren, L. H. Young, M. Pypaert, J. Mu, M. J. Birnbaum, G. I.
Shulman, Proc. Natl. Acad. Sci. USA 2002, 99, 15983–15987.
[45] M. J. Betz, S. Enerbäck, Nat. Rev. Endocrinol. 2018, 14, 77–87.
[46] D. G. Nicholls, Biochim. Biophys. Acta. 2006, 1757, 459–466.
[47] K. Schneider, J. Valdez, J. Nguyen, M. Vawter, B. Galke, T. W. Kurtz, J. Y.
Chan, J. Biol. Chem. 2016, 291, 7754–7766.
[48] M. Liesa, O. S. Shirihai, Cell Metab. 2013, 17, 491–506.
[49] T. Kita, H. Nishida, H. Shibata, S. Niimi, T. Higuti, N. Arakaki, J. Biochem.
2009, 146, 787–796.
[50] P. H. Ducluzeau, M. Priou, M. Weitheimer, M. Flamment, L. Duluc, F.
Iacobazi, R. Soleti, G. Simard, A. Durand, J. Rieusset, R. Andriantsitohaina,
Y. Malthièry, J. Physiol. Biochem. 2011, 67, 285–296.
[51] Q. Zheng, J. Lin, J. Huang, H. Zhang, R. Zhang, X. Zhang, C. Cao, C.
Hambly, G. Qin, J. Yao, R. Song, Q. Jia, X. Wang, Y. Li, N. Zhang, Z. Piao,
R. Ye, J. R. Speakman, H. Wang, Q. Zhou, Y. Wang, W. Jin, J. Zhao, Proc.
Natl. Acad. Sci. USA 2017, 114, E9474–E9482.
[52] H. Wang, M. Willershauser, A. Karlas, D. Gorpas, J. Reber, V. Ntziachristos,
S. Maurer, T. Fromme, Y. Li, M. Klingenspor, Mol. Metab. 2019, 20, 14–27.
[53] F. Shamsi, R. Xue, T. L. Huang, M. Lundh, Y. Liu, L. O. Leiria, M. D. Lynes,
E. Kempf, C. H. Wang, S. Sugimoto, et al., Nat. Commun. 2020, 11, 1421.
[54] P. Seale, H. M. Conroe, J. Estall, S. Kajimura, A. Frontini, J. Ishibashi, P.
Cohen, S. Cinti, B. M. Spiegelman, J. Clin. Invest. 2011, 121, 96–105.
[55] M. P. Czech, Nat. Med. 2017, 23, 804–814.
[56] M. Harms, P. Seale, Nat. Med. 2013, 19, 1252–1263.
[57] K. Ikeda, P. Maretich, S. Kajimura, Trends Endocrinol. Metab. 2018, 29,
191–200.
5
6
7
8
Conflict of Interest
The authors declare no conflict of interest.
9
Keywords: lipid droplets · metabolic diseases · obesity · PGC-
1alpha-UCP1 · phenotype-based drug discovery
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
[1] S. E. Kahn, R. L. Hull, K. M. Utzschneider, Nature 2006, 444, 840–846.
[2] P. Kopelman, Obes. Rev. 2007, 8 Suppl 1, 13–17.
[3] J. E. Hall, J. M. do Carmo, A. A. da Silva, Z. Wang, M. E. Hall, Nat. Rev.
Nephrol. 2019, 15, 367–385.
[4] S. Verma, M. E. Hussain, Diabetes Metab. Syndr. Obes. 2017, 11, 73–79.
[5] G. B. D. O. Collaborators, A. Afshin, M. H. Forouzanfar, M. B. Reitsma, P.
Sur, K. Estep, A. Lee, L. Marczak, A. H. Mokdad, M. Moradi-Lakeh, M.
Naghavi, et al., N. Engl. J. Med. 2017, 377, 13–27.
[6] M. Glandt, I. Raz, J. Obes. 2011, 2011, 636181.
[7] A. Elangovan, R. Shah, Z. L. Smith, Obes. Surg. 2020.
[8] S. Henness, C. M. Perry, Drugs 2006, 66, 1625–1656.
[9] D. K. Patel, F. C. Stanford, Postgrad. Med. 2018, 130, 173–182.
[10] I. J. Onakpoya, C. J. Heneghan, J. K. Aronson, BMC Med. 2016, 14, 191.
[11] D. Serra, P. Mera, M. I. Malandrino, J. F. Mir, L. Herrero, Antioxid. Redox
Signaling 2013, 19, 269–284.
[12] E. S. Childress, S. J. Alexopoulos, K. L. Hoehn, W. L. Santos, J. Med. Chem.
2018, 61, 4641–4655.
[13] M. Goldgof, C. Xiao, T. Chanturiya, W. Jou, O. Gavrilova, M. L. Reitman, J.
Biol. Chem. 2014, 289, 19341–19350.
[14] U. S. Eggert, Nat. Chem. Biol. 2013, 9, 206–209.
[15] D. C. Swinney, Clin. Pharmacol. Ther. 2013, 93, 299–301.
[16] E. Kim, S. Lee, S. B. Park, Chem. Commun. 2012, 48, 2331–2333.
[17] M. Konige, H. Wang, C. Sztalryd, Biochim. Biophys. Acta. 2014, 1842, 393–
401.
[18] J. J. Fuster, N. Ouchi, N. Gokce, K. Walsh, Circ. Res. 2016, 118, 1786–1807.
[19] S. Lee, E. Kim, S. B. Park, Chem. Sci. 2013, 4, 3282–3287.
[20] M. Zhu, B. J. Lim, M. Koh, S. B. Park, ACS Comb. Sci. 2012, 14, 124–134.
[21] F. J. Ruiz-Ojeda, A. I. Ruperez, C. Gomez-Llorente, A. Gil, C. M. Aguilera,
Int. J. Mol. Sci. 2016, 17, 1040.
[22] S. R. Farmer, Cell Metab. 2006, 4, 263–273.
[23] E. Ji, M. Y. Jung, J. H. Park, S. Kim, C. R. Seo, K. W. Park, E. K. Lee, C. H.
Yeom, S. Lee, Int. J. Obes. (Lond) 2014, 38, 1035–1043.
[24] Y. K. Park, J. Lee, V. S. Hong, J. S. Choi, T. Y. Lee, B. C. Jang, PLoS One
2014, 9, e109344.
[25] N. A. Ducharme, P. E. Bickel, Endocrinology 2008, 149, 942–949.
[26] M. A. Aon, N. Bhatt, S. C. Cortassa, Front. Plant Physiol. 2014, 5, 282.
[27] N. Turner, C. R. Bruce, S. M. Beale, K. L. Hoehn, T. So, M. S. Rolph, G. J.
Cooney, Diabetes 2007, 56, 2085–2092.
[58] V. Peirce, S. Carobbio, A. Vidal-Puig, Nature 2014, 510, 76–83.
[59] K. Suzuki, C. N. Jayasena, S. R. Bloom, Exp. Diabetes Res. 2012, 2012,
824305.
[60] A. M. Heck, J. A. Yanovski, K. A. Calis, Pharmacotherapy 2000, 20, 270–
279.
[61] R. J. Rodgers, M. H. Tschop, J. P. Wilding, Dis. Model. Mech. 2012, 5, 621–
626.
[62] P. Flachs, M. Rossmeisl, O. Kuda, J. Kopecky, Biochim. Biophys. Acta.
2013, 1831, 986–1003.
[63] J. Lee, J. M. Ellis, M. J. Wolfgang, Cell Rep. 2015, 10, 266–279.
[64] J. A. Harper, K. Dickinson, M. D. Brand, Obes. Rev. 2001, 2, 255–265.
[65] R. Putti, R. Sica, V. Migliaccio, L. Lionetti, Front. Plant Physiol. 2015, 6,
109.
[66] J. Park, S. Oh, S. B. Park, Angew. Chem. Int. Ed. 2012, 51, 5447–5451;
Angew. Chem. 2012, 124, 5543–5547.
[67] S. Lee, Y. Nam, J. Y. Koo, D. Lim, J. Park, J. Ock, J. Kim, K. Suk, S. B. Park,
Nat. Chem. Biol. 2014, 10, 1055–1060.
[68] H. Park, J. Ha, J. Y. Koo, J. Park, S. B. Park, Chem. Sci. 2017, 8, 1127–1133.
[69] H. Park, S. B. Park, Chem. Sci. 2019, 10, 3449–3458.
[28] Y. Zhang, L. W. Castellani, C. J. Sinal, F. J. Gonzalez, P. A. Edwards, Genes
Dev. 2004, 18, 157–169.
[29] B. N. Finck, D. P. Kelly, J. Clin. Invest. 2006, 116, 615–622.
[30] P. J. Fernandez-Marcos, J. Auwerx, Am. J. Clin. Nutr. 2011, 93, 884S–890.
[31] C.-F. Cheng, H.-C. Ku, H. Lin, Int. J. Mol. Sci. 2018, 19, 3447.
[32] H. P. Harding, Y. Zhang, S. Khersonsky, S. Marciniak, D. Scheuner, R. J.
Kaufman, N. Javitt, Y. T. Chang, D. Ron, Cell Metab. 2005, 2, 361–371.
[33] P. Keller, J. T. Petrie, P. De Rose, I. Gerin, W. S. Wright, S. H. Chiang, A. R.
Nielsen, C. P. Fischer, B. K. Pedersen, O. A. MacDougald, J. Biol. Chem.
2008, 283, 14355–14365.
[34] K. Makki, P. Froguel, I. Wolowczuk, ISRN Inflamm. 2013, 2013, 139239.
[35] A. E. Achari, S. K. Jain, Int. J. Mol. Sci. 2017, 18, 1321.
[36] M. Schweiger, T. O. Eichmann, U. Taschler, R. Zimmermann, R. Zechner,
A. Lass, Methods Enzymol. 2014, 538, 171–193.
Manuscript received: January 25, 2021
Accepted manuscript online: February 3, 2021
Version of record online: March 10, 2021
Correction added on March 24, 2021, after first online publication: Assign-
ment of author first and last names were corrected. Figure 3b was corrected
as the 2nd & 6th entries of the x-axis were mislabeled. None of the minor
corrections affect the scientific content and conclusions of the article.
[37] A. Yang, E. P. Mottillo, Biochem. J. 2020, 477, 985–1008.
ChemMedChem 2021, 16, 1104–1115
www.chemmedchem.org
1115
© 2021 Wiley-VCH GmbH