Ethylenedisalicylic acid derivatives as dual inhibitors of PTP1B and IKKβ and their antiobesity and antidiabetic effects in mice

Article abstract of DOI:10.1002/bkcs.10786

Ethylenedisalicylic acid (EDSA) was developed as a novel scaffold for dual inhibitors of protein tyrosine phosphatase (PTP) 1B and IkB kinase β (IKKβ). EDSA is a modified version of methylenedisalicylic acid (MDSA) and contains two salicylate moieties connected by an ethylene moiety. In this study, derivatives of EDSA were synthesized and their inhibitory potencies against PTP1B and IKKβ were investigated. Many of these derivatives exhibited half-maximal inhibitory concentrations (IC₅₀s) in the low micromolar range. The metabolic effects of ESA4, 7, and 8 were further examined in a mouse model system. ESA4 and ESA8 significantly suppressed diet-induced weight gain, whereas ESA7 had a marginal effect. ESA4, 7, and 8 also lowered fasting glucose levels and accelerated glucose clearance rates after glucose injection. These observations indicate that EDSA scaffold-based compounds are promising candidates for the treatment of obesity and diabetes.

Full text of DOI:10.1002/bkcs.10786

Article
DOI: 10.1002/bkcs.10786
BULLETIN OF THE
N. G. Aher et al.
KOREAN CHEMICAL SOCIETY
Ethylenedisalicylic Acid Derivatives as Dual Inhibitors of PTP1B
and IKKβ and their Antiobesity and Antidiabetic Effects in Mice
Nilkanth G. Aher,† Ji-Won Park,‡ Byung Ho Park,† Chan Kyung Kim,†
Inn-Oc Han,§ and Hyeongjin Cho†,
*
†Department of Chemistry, College of Natural Sciences, Inha University, Incheon 22212, Korea.
*E-mail: hcho@inha.ac.kr
‡Department of Life Science, College of Natural Sciences, Inha University, Incheon 22212, Korea
§Department of Physiology and Biophysics, College of Medicine, Inha University, Incheon 22212, Korea
Received February 11, 2016, Accepted March 18, 2016, Published online May 23, 2016
Ethylenedisalicylic acid (EDSA) was developed as a novel scaffold for dual inhibitors of protein tyrosine
phosphatase (PTP) 1B and IkB kinase β (IKKβ). EDSA is a modified version of methylenedisalicylic acid
(MDSA) and contains two salicylate moieties connected by an ethylene moiety. In this study, derivatives
of EDSA were synthesized and their inhibitory potencies against PTP1B and IKKβ were investigated.
Many of these derivatives exhibited half-maximal inhibitory concentrations (IC50s) in the low micromolar
range. The metabolic effects of ESA4, 7, and 8 were further examined in a mouse model system. ESA4
and ESA8 significantly suppressed diet-induced weight gain, whereas ESA7 had a marginal effect. ESA4,
7, and 8 also lowered fasting glucose levels and accelerated glucose clearance rates after glucose injection.
These observations indicate that EDSA scaffold-based compounds are promising candidates for the treat-
ment of obesity and diabetes.
Keywords: PTP1B, IKKβ, Dual inhibitor, Antiobesity, Antidiabetic
Introduction
Several were examined in a mouse model system and found
to effectively suppress weight gain or ameliorate insulin
resistance induced by HFD. As an extension of these stud-
ies, we investigated the possible pharmacologic use of ethy-
lenedisalicylic acid (EDSA) a congener of MDSA. The
benzylic position could be labile to metabolic oxidation
thus accelerating the whole-body clearance of the drug. We
envisioned that the introduction of a methyl group might
influence favorably on the half-life of the drug by sterically
hindering the benzylic oxidation.
Obesity and diabetes have reached epidemic proportions
worldwide. The global prevalence of obesity (BMI ≥
30 kg/m2) in adults doubled from 6.5% in 1980 to 13% in
2014, and its prevalence is highest in North America (27%)
and lowest in Southeast Asia (5%).1 On the other hand, the
global prevalence of diabetes in 2014 was estimated to be
9%.1 These two conditions are closely related; e.g., 60% of
diabetic cases are directly linked to overweightedness,2 and
also increase the risks of related diseases like hypertension
and coronary heart disease.1 Protein tyrosine phosphatase
(PTP) 1B is ubiquitously present and is known to partici-
pate in insulin signaling and cancer-related signaling
events.3 Furthermore, transgenic mice lacking PTP1B ame-
liorate diet-induced weight gain and the development of
insulin resistance.4,5 On the other hand, IkB kinase β
(IKKβ) is a major player in inflammation and other immune
responses,6 and recently, was found to be involved in insu-
lin signaling. IKKβ-deficient transgenic mice were robust
to the induction of insulin resistance by high-fat diet (HFD)
or leptin deficiency.7,8 Unlike PTP1B deletion, IKKβ defi-
ciency did not suppress weight gain in mice, but lowered
serum free fatty acid (FFA) levels.8 Obesity and insulin
resistance are associated with chronic low-grade inflamma-
tion and it has been suggested IKKβ might integrate inflam-
mation and metabolic disorders.9
Experimental
Synthesis the ESA Derivatives. 2,4,6-Trimethyl-1,3,5-
trioxane (A). Acetaldehyde trimer was prepared as previ-
ously reported.11
Dimethyl 3,30-ethylidenedisalicylate (B). A solution of
methyl salicylate (3.04 g, 20.0 mmol) in MeOH/H2O (4:1,
50 mL) was cooled to −5ꢀC. Concentrated H2SO4 (60 mL)
was then added over 40 min with stirring and cooling in an
ice-bath. The acetate trimer (A) (2.0 mL, 15 mmol) was
then added over 1 min and the reaction mixture stirred at
0ꢀC for 30 min. After stirring at room temperature for 2 h,
the reaction mixture was poured into ice water (200 g) and
extracted with EtOAc (200 mL × 3,100 mL × 1). Com-
bined extracts were sequentially washed with water
(100 mL) and brine (100 mL), dried (Na2SO4), filtered, and
concentrated. The dark brown residue obtained was purified
by column chromatography (hexane/EtOAc 9:1) to afford
We previously described methylenedisalicylic acid
(MDSA) derivatives that inhibit both PTP1B and IKKβ.10
Bull. Korean Chem. Soc. 2016, Vol. 37, 855–863
© 2016 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Wiley Online Library
855

Article
BULLETIN OF THE
ISSN (Print) 0253-2964 | (Online) 1229-5949
KOREAN CHEMICAL SOCIETY
B as a liquid (1.22 g, 37%), which solidified on standing.
(s, 3H), 4.04 (s, 3H), 4.07 (s, 3H), 7.11 (d, J = 9.2 Hz,
1H), 7.15 (d, J = 8.8 Hz, 1H), 8.00 (dd, J = 8.8, 2.2 Hz,
1H), 8.16 (dd, J = 8.8, 2.2 Hz, 1H), 8.39 (d, J = 2.2 Hz,
1H), 8.56 (d, J = 2.2 Hz, 1H), 11.31 (s, 1H); 13C NMR
(CDCl3, 100 MHz) δ 26.2, 52.6, 112.1, 117.8, 118.0,
128.9, 131.4, 132.8, 135.4, 137.0, 164.8, 165.3,
170.1, 192.6.
Unreacted starting material was also recovered (0.76 g,
1
25%). Compound B: mp 62–63ꢀC; H NMR (400 MHz,
CDCl3) δ 1.60 (d, J = 7.2 Hz, 3H), 3.95 (s, 6H), 4.05 (q,
J = 7.2 Hz, 1H), 6.92 (d, J = 8.4 Hz, 2H), 7.27 (dd,
J = 2.4 and 8.4 Hz, 2H), 7.67 (d, J = 2.4 Hz, 2H), 10.64
(s, 2H); 13C NMR (CDCl3, 100 MHz) δ 22.2, 43.1, 52.5,
112.2, 118.0, 128.3, 135.4, 137.1, 160.3, 170.7.
Compound F. Compound D (1.714 g, 2.87 mmol) and
Pd(PPh3)2Cl2 (241 mg, 0.34 mmol) were added to 1,4-
dioxane (43 mL) in an oven-dried round bottom flask. After
flushing with N2, 2-(tributylstannyl)furan (2.71 mL,
8.60 mmol) was added using a syringe and the reaction
mixture was allowed to stir for 3 h at 85ꢀC. Additional 2-
(tributylstannyl)furan (0.5 mL) was then added to avoid
production of the unwanted monofurannylation product,
which appeared slightly above compound F on TLC plates.
After stirring at 85ꢀC for additional 1.5 h, the reaction mix-
ture was cooled and poured into 10% aq. NH4OH solution
(70 mL). The resulting solution was extracted with EtOAc
(200 mL × 3). The organic extracts were combined,
sequentially washed with water (200 mL × 2) and brine
(60 mL), dried (Na2SO4), and concentrated. The crude
product was a solid–liquid mixture, which was not recrys-
tallized successfully. The mixture was purified by column
chromatography (hexane/EtOAc 2:1) to obtain compound
F (1.392 g, 85%) as a solid, which was recrystallized in
EtOAc for further purification. Compound F: mp
167–169ꢀC; 1H NMR (400 MHz, CDCl3) δ 1.75 (d,
J = 7.2 Hz, 3H), 2.42 (s, 6H), 3.88 (s, 3H), 4.30 (q,
J = 7.2 Hz, 1H), 6.49 (dd, J = 3.2, 1.6 Hz, 2H), 6.74 (d,
J = 3.2 Hz, 2H), 7.50 (d, J = 1.6 Hz, 2H), 7.80 (d,
J = 2.0 Hz, 2H), 7.89 (d, J = 2.0 Hz, 2H); 13C NMR
(100 MHz, CDCl3) δ 21.5, 21.9, 52.5, 110.2, 112.094,
124.3, 125.9, 129.7, 130.4, 142.7, 143.3, 145.0, 148.9,
165.1, 169.4.
Dimethyl 5,50-dibromo-3,30-ethylidenedisalicylate (C).
Bromination of compound 5 was carried out as previously
described.12 Compound B (2.18 g, 6.61 mmol) was dis-
solved in MeOH (30 mL), and a solution of Br2 (0.68 mL,
13.2 mmol) in MeOH (20 mL) was added dropwise. The
reaction mixture was then stirred overnight at room temper-
ature. The precipitate produced was filtered, washed with
cold MeOH (4 mL × 3), and recrystallized in EtOAc to
obtain C as a white solid (2.35 g, 73%). Compound C: mp
174–176ꢀC; 1H NMR (400 MHz, CDCl3) δ 1.59 (d,
J = 7.2 Hz, 3H), 4.01 (s, 6H), 4.06 (q, J = 7.4 Hz, 1H),
7.53 (d, J = 2.4 Hz, 2H), 7.63 (dd, J = 2.4, 0.4 Hz, 2H),
11.34 (s, 2H); 13C NMR (CDCl3, 100 MHz) δ 22.0, 42.8,
53.0, 111.7, 113.5, 127.8, 137.5, 138.3, 157.0, 170.2.
Dimethyl O,O0-diacetyl-5,50-dibromo-3,30-ethylidenedi-
salicylate (D). A mixture of compound C (1.40 g,
2.72 mmol), acetic anhydride (20 mL), and H3PO4
(1.0 mL) was stirred at 80ꢀC for 45 min. The reaction mix-
ture was then cooled to room temperature and poured into
a mixture of water (200 mL) and EtOAc (200 mL). The
organic layer was saved and the aqueous layer was
extracted with EtOAc (200 mL × 2). The organic layers
were then combined, washed with H2O (200 mL) and brine
(100 mL), dried (Na2SO4), concentrated to ~5 mL, and
mixed with hexane (10 mL). A minimum amount of EtOAc
(usually less than a few mL) was added to the resulting two
phase mixture until the solution became homogeneous.
Overnight storage at 4ꢀC induced the crystallization of
compound D (white crystalline solid). The mother liquor
was purified by column chromatography (hexane/EtOAc
9:1 to 3:1) to afford additional compound D. Combined
General Procedure for the Suzuki Coupling Reaction.
A
round-bottom flask was charged with compound
C (830 mg, 1.71 mmol), 3,5-di(trifluoromethyl)benzene-
boronic acid (1.32 g, 5.13 mmol), toluene (26 mL), and
2 M aq. Na2CO3 solution (2.6 mL, 3.8 mmol), and flushed
with high purity N2.13 (Ph3P)4Pd (150 mg, 10 mol%) and
95% EtOH (0.2 mL) were then added and heated at 95ꢀC
overnight under a N2 atmosphere. Additional (Ph3P)4Pd
(150 mg, 10 mol%) was added and heated continuously for
6 h. When the reaction was complete as determined by
TLC, the mixture was cooled, acidified with 0.25 M HCl
(40 mL) and extracted with EtOAc (50 mL × 2). Com-
bined organic layers were washed with H2O (15 mL × 2)
and brine (15 mL), dried (Na2SO4), filtered, and solvent
was removed under reduced pressure to give the crude
product. Purification by chromatography (hexane/EtOAc
9:1 to 3:1) afforded compound I as a white solid (658 mg,
51%). Interestingly, a methyl-to-ethyl ester exchange reac-
tion product was obtained as a byproduct (172 mg, 13%),
which eluted slightly earlier than I. It was also subjected
for the hydrolysis reaction. Compounds G, H, and J were
1
yield (1.454 g, 89%). Compound D: mp 147–148ꢀC; H
NMR (400 MHz, CDCl3) δ 1.66 (d, J = 7.2 Hz, 3H), 2.40
(s, 3H), 3.88 (s, 3H), 4.17 (q, J = 7.2 Hz, 1H), 7.60 (s,
2H), 7.80 (s, 2H); 13C NMR (100 MHz, CDCl3) δ 20.9,
21.8, 43.5, 52.8, 119.2, 125.3, 130.1, 136.4, 144.0, 147.1,
164.2, 168.6.
Compound E. Dimethyl 3,3-ethylidenediisalicylate (B)
(0.050 g, 0.15 mmol) in acetone (2 mL) was cooled to
0ꢀC, and 1.0 mL of Jones’ reagent was added. The mixture
was then stirred for 10 min at 0ꢀC, additional Jones’ rea-
gent (1.0 mL) was added, and stirring continued for
20 min. When the reaction was complete (monitored by
TLC), methanol (2 mL) was added to destroy excess Jones’
reagent. Solvent was removed by evaporation and the crude
product obtained was purified by column chromatography
(hexane/EtOAc 9:1) to afford compound E (0.012 g, 25 %
1
yield). Compound E: H NMR (400 MHz, CDCl3) δ 2.65
Bull. Korean Chem. Soc. 2016, Vol. 37, 855–863
© 2016 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
856

BULLETIN OF THE
Article
Ethylenedisalicylic Acid Derivatives and their metabolic effects
KOREAN CHEMICAL SOCIETY
1
synthesized using an identical procedure using appropriate
benzeneboronic acid derivatives.
ESA6. mp > 250ꢀC (decompose); H NMR (400 MHz,
DMSO-d6) δ 1.62 (d, J = 7.2 Hz, 3H), 4.28 (q, J = 7.2 Hz,
1H), 7.59 (d, J = 2.4 Hz, 2H), 7.77 (s, 8H), 7.78 (d,
J = 2.4 Hz, 2H); 13C NMR (100 MHz, DMSO-d6) δ 22.0,
42.6, 113.7, 125.0 (q, J = 267.1 Hz), 125.6, 128.4, 128.4
(q, J = 31.9 Hz), 129.3, 130.7, 136.2, 137.5, 141.8,
157.6, 173.0.
Compound H. mp 157–159ꢀC; 1H NMR (400 MHz,
CDCl3) δ 1.69 (d, J = 7.2 Hz, 3H), 3.97 (s, 6H), 4.16 (q,
J = 7.2 Hz, 1H), 7.37 (d, J = 2.0 Hz, 2H), 7.68 (s, 8H),
7.79 (d, J = 2.4 Hz, 2H), 11.24 (s, 2H); 13C NMR
(100 MHz, CDCl3) δ 22.2, 43.3, 52.8, 112.9, 124.4 (q,
J = 270.6 Hz), 125.3 (q, J = 3.8 Hz), 128.5, 129.4, 129.7
(q, J = 32.3 Hz), 129.9, 136.1, 136.9, 141.1, 157.7, 171.0.
Compound I. mp 244–246ꢀC; 1H NMR (400 MHz,
CDCl3) δ 1.72 (d, J = 7.2 Hz, 3H), 4.00 (s, 6H), 4.20 (q,
J = 7.2 Hz, 1H), 7.40 (d, J = 2.4 Hz, 2H), 7.83 (d,
J = 2.4 Hz, 2H), 7.86 (s, 2H), 8.02 (s, 4H); 13C NMR
(100 MHz, CDCl3) δ 22.4, 43.3, 52.9, 113.2, 121.4, 123.6
(q, J = 271.4 Hz), 127.7, 129.1, 129.8, 131.7 (q,
J = 32.9 Hz), 135.8, 136.9, 139.3, 157.545, 170.8.
1
ESA7. mp 128–138ꢀC; H NMR (400 MHz, DMSO-d6)
δ 1.65 (d, J = 7.2 Hz, 3H), 4.32 (1H, q, J = 7.2 Hz, 1H),
7.76 (d, J = 2.4 Hz, 2H), 7.83 (2H, d, J = 2.4 Hz, 2H),
8.06 (s, 2H), 8.21 (s, 4H); 13C NMR (100 MHz, DMSO-
d6) δ 22.0, 42.7, 113.9, 121.4, 124.0 (q, J = 271.1 Hz),
126.5, 130.0, 130.4, 130.8 (q, J = 32.6 Hz), 136.3, 137.6,
139.9, 157.5, 172.9.
1
ESA8. mp >250ꢀC (decompose); H NMR (400 MHz,
DMSO-d6) δ 1.63 (d, J = 7.2 Hz, 3H), 4.30 (q, J = 7.2 Hz,
1H), 7.64 (s, 2H), 7.67 (d, J = 7.2 Hz, 2H), 7.70 (t,
J = 7.2 Hz, 2H), 7.78 (s, 2H), 7.85 (d, J = 7.2 Hz, 2H),
7.88 (2H, s); 13C NMR (100 MHz, DMSO-d6) δ 22.1,
42.7, 113.7, 124.6, 124.9 (q, J = 270.7 Hz), 126.3, 128.3,
129.2, 129.6 (q, J = 31.1 Hz), 129.9, 133.9, 136.2, 137.6,
138.5, 157.6, 173.0.
PTP1B Assay. IC50 values of PTP1B were determined by
measuring the rate of p-nitrophenyl phosphate (pNPP)
hydrolysis at a series of inhibitor concentrations.14 Before
the experiment, PTP1B was diluted to 400 nM in enzyme
dilution buffer (25 mM HEPES, 5.0 mM EDTA, 1.0 mM
DTT, 1.0 mg/mL bovine serum albumin, pH 7.3) and the
inhibitors were dissolved in DMSO. To start the assay,
PTP1B was added to a reaction buffer containing pNPP
and inhibitor. The final assay solution contains: 40 nM
PTP1B, 2 mM pNPP, assorted concentrations of inhibitor,
5 mM EDTA, 50 mM HEPES (pH 7.0), and 10 % of the
enzyme dilution buffer. After 3 min incubation at 37ꢀC,
0.5 M NaOH was added to stop the reaction, and then
absorbance was measured at 405 nm.
IKKβ Assay. Kinase activity was measured using an
IKKα and β kinase assay/inhibitor screening kit (CY-1178,
CycLex, Nagano, Japan) according to the manufacturer’s
instructions but with minor modifications. The kit utilized
96-well plates pre-coated with a substrate (recombinant
IkBα) containing two serine residues that are phosphory-
lated by IKKα and IKKβ. For 100 μL reaction mixture, the
well was filled with 80 μL of kinase reaction buffer con-
taining ATP (62.5 μM) and 10 μL of inhibitor in 10%
DMSO. Then, 10 μL of IKKβ (10 mU, manufacturer’s defi-
nition) was added to initiate the enzyme reaction. After
30 min at 30ꢀC, the reaction was stopped by adding 1 M
H2SO4. The assay solution was then removed and the well
was washed with a buffer containing 2% Tween-20. Phos-
phorylated substrates on the well surface were treated
sequentially with anti-phospho-IkBα S32 antibody and
HRP-conjugated anti-mouse IgG, and the phosphorylated
substrates were quantitated using tetramethylbenzidine
(a chromogenic HRP substrate) and spectrophotometry by
measuring absorbance at 450 nm.
Compound J. mp 68–72ꢀC; 1H NMR (400 MHz,
CDCl3) δ 1.70 (d, J = 7.2 Hz, 3H), 3.99 (s, 3H), 4.18 (q,
J = 7.2 Hz, 1H), 7.40 (d, J = 2.4 Hz, 2H), 7.55 (t,
J = 7.8 Hz, 2H), 7.62 (d, J = 7.8 Hz, 2H), 7.76 (d,
J = 7.8 Hz, 2H), 7.79 (d, J = 2.4 Hz, 2H),7.83 (s, 2H),
11.25 (s, 2H); 13C NMR (CDCl3) δ 22.3, 43.4, 52.8, 112.9,
124.4 (q, J = 3.8 Hz), 124.4 (q, J = 270.7 Hz), 126.4 (q,
J = 3.8 Hz), 128.4, 128.8, 129.3, 130.5 (q, J = 31.8 Hz),
133.0, 136.1, 136.9, 138.2, 157.6, 171.0.
General Procedure for Production of ESA1–4.
Compound F (1.08 g, 1.88 mmol) was added into a mix-
ture of 2.0 M aq. NaOH (15.5 mL) and 1,4-dioxane
(20 mL). After stirring at 95ꢀC for 50 min, 1 M HCl
(ca. 30 mL) was added into the reaction mixture to make
pH ~ 1. The precipitate was saved and the aqueous layer
was extracted with EtOAc (90 mL × 3). The saved precipi-
tate was dissolved in the combined EtOAc extract and the
resulting solution was washed with H2O (50 mL × 2) and
brine (50 mL), dried (Na2SO4), and concentrated under
vacuum to obtain ESA4 as a beige solid in a quantitative
yield (870 mg). Similar results were obtained for ESA1,
ESA2, and ESA3 under the identical reaction conditions.
ESA4: mp > 250ꢀC (decompose); 1H NMR (400 MHz,
DMSO-d6) δ 1.59 (d, J = 7.2 Hz, 3H), 4.31 (q, J = 7.2 Hz,
1H), 6.60–6.62 (m, 2H), 7.03–7.05 (m, 2H), 7.662 (s, 2H),
7.744 (s, 2H), 7.834 (s, 2H); 13C NMR (100 MHz, DMSO-
d6) δ 22.2, 42.5, 111.3, 112.8, 113.8, 119.6, 127.8, 130.2,
137.3, 142.9, 149.1, 156.4, 173.1.
General Procedure for Production of ESA5–8.
Powdered compound I (545 mg, 0.71 mmol) was added to
a mixture of 1.0 M aq. NaOH (5.8 mL, 5.8 mmol) and
95% EtOH (16 mL). After stirring at 95ꢀC for 30 min, 1 M
HCl (ca. 7 mL) was added to reduce the pH of the reaction
mixture to make ~ 1. EtOH was evaporated off under vac-
uum and the aqueous so obtained was extracted with Et2O
(35 mL × 3). Combined extracts were washed with H2O
(20 mL × 2) and brine (20 mL), dried (Na2SO4), and con-
centrated under vacuum to obtain ESA7 as a thin film or
crispy foam (494 mg, 96%).
Bull. Korean Chem. Soc. 2016, Vol. 37, 855–863
© 2016 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
857

BULLETIN OF THE
Article
Ethylenedisalicylic Acid Derivatives and their metabolic effects
KOREAN CHEMICAL SOCIETY
Animal Experiments. Male 4-week old C57BL/6 wild-
type mice were purchased from Taconic Farms (Rockville,
MD, USA). Mice were fed a low-fat diet (LFD, 3.90 kcal/
g, D10012G) or a HFD (4.73 kcal/g, D12451), containing
16% or 45% fat, respectively (both diets were from
Research Diets, New Brunswick, NJ, USA). Upon arrival,
animals were individually housed in separate cages under a
12:12 h light:dark cycle at 23 ꢁ 2ꢀC. Animals had access
to pelleted LFD and tap water ad libitum for 1 week, and
were then randomly assigned to five experimental groups
(n = 7–8 per group). Two groups served as lean or obese
controls and were fed LFD or HFD throughout the 12-week
study period. Animals in the other three groups received
HFD for 8 weeks to induce obesity and diabetes, and then
HFD plus drugs for 4 weeks. To administer drugs, pow-
dered HFD (200 g) was mixed with a drug solution (20 mL
for ESA4, 15 mL for ESA7 and ESA8) and kneaded to a
dough. The drug solution was prepared by adding the drug
(333 mg for ESA4 or 500 mg for ESA7 or ESA8) dis-
solved in 0.8 mL DMSO in small aliquots to the solution
of polyglycerol fatty acid ester (0.2 g in 19.2 mL H2O)
with homogenization. HFD dough for the obese control
group was prepared in the same way without adding a drug.
Changes in mouse weights were monitored during 4-week
drug feeding period, after which glucose and insulin toler-
ance tests (GTT and ITT) were done. For GTTs, mice were
fasted for 6 h and then injected intraperitoneally (i.p.) with
glucose (1 g/kg of body weight). Blood glucose levels were
determined by sampling from the tail vein at 0, 20, 40, 60,
90, and 120 min after glucose administration using a gluc-
ometer (AccuCheck active, Roche Diagnostics, Ireland).
ITT was performed 3 days after GTT. Briefly, mice fasted
for 6 h were injected with insulin (1 unit/kg of body
weight) and blood glucose levels were determined as in
GTT. Mice were sacrificed 5 days after ITT by cardiac
puncture for blood samples. Serum levels of triglyceride,
total cholesterol, and nonesterified fatty acids (NEFAs)
were determined as described previously.10 Epididymal and
retroperitoneal fat pads and liver were removed and
weighed.
Figure 1. Reagent and conditions: (a) (CH3)3SiCl, 0ꢀC ! rt;
(b) conc. H2SO4–CH3OH–H2O, 6:4:1, 0ꢀC ! rt, 37%; (c) Br2,
MeOH, rt, overnight, 73%; (d) Ac2O, cat. H3PO4, 80ꢀC, 45 min,
89%; (e) 2 M NaOH, dioxane, 95ꢀC, 50 min, 99%; (f ) Jones’ rea-
gent, acetone, 0ꢀC, 30 min, 25%; (g) 2-(tributylstannyl)furan,
Pd(PPh3)2Cl2, 1,4-dioxane, 100ꢀC, 4.5 h, 85%; (h) ArB(OH)2,
(Ph3P)4Pd, 90–95ꢀC, 16 h, 64%; (i) 1 M NaOH, EtOH, 95ꢀC,
30 min, >90%.
Results and Discussion
substituted with various functional groups. The substituents
were chosen based on the predicted logD values
(ChemAxon, Budapest, Hungary) as well as the availability
of the corresponding reagents for Suzuki or Stille coupling
reactions. The calculated logD values of ESA5–8 were 0.7,
2.4, 4.2, and 2.4, respectively, all within the favorable
range for logP (−0.4 to +5.6) recommended in Lipinski’s
rule. The logD values of ESA3 and 4 were −1.1 and −1.2,
respectively, slightly deviating Lipinski’ criteria.
To derivatize the EDSA scaffold, compound B was bro-
minated ortho to the hydroxyl group, and the bromo com-
pound obtained (compound C) was used to produce the
other ESA compounds. ESA3 was the hydrolysis product
of compound C. To obtain the furanyl derivative (ESA4),
the phenolic hydroxyl groups of compound C were
EDSA was synthesized using a strategy similar to that pre-
viously used to synthesize MDSA in our laboratory
(Figure 1)15 by substituting the acetaldehyde trimer, 2,4,6-
trimethyl-1,3,5-trioxane, for trioxane (a cyclic trimer of
formaldehyde). Under strongly acidic conditions, acetalde-
hyde and 2 equiv of methyl salicylate condensed to form
compound B, the dimethyl ester of EDSA, which was sepa-
rated in a pure form without the complications experienced
during the preparation of MDSA dimethyl ester. ESA1 was
the hydrolysis product of B. ESA2 was obtained by the
oxidation of B to the quinone methide before hydrolytic
cleavage of the ester. For derivatization, the hydrogen
atoms at 5 and 50 of the salicylate moiety of ESA1 were
Bull. Korean Chem. Soc. 2016, Vol. 37, 855–863
© 2016 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
858

BULLETIN OF THE
Article
Ethylenedisalicylic Acid Derivatives and their metabolic effects
KOREAN CHEMICAL SOCIETY
protected as acetates (D) before Stille coupling, which was
performed using 2-(tributylstannyl)furan. To synthesize
ESA5–8, appropriate arylboronic acids were used in Suzuki
reactions without protecting the hydroxyl group of com-
pound C. In all cases, the byproducts produced moved
slightly faster on TLC plates than desired products. NMR
analysis showed that the byproducts were the methyl
ethyl mixed esters, formed by exchange of one of the meth-
oxy group with ethanol present in the reaction system.
The mixed ester was formed in 10–20% of the
dimethyl ester, and when hydrolyzed to a free acid, pro-
duced ESA7. The final step of the synthesis of ESA deriva-
tives was ester hydrolysis using 1,4-doxane or ethanol as
solvents.
structure of ESA4 is similar to SA51, previously reported
in our laboratory, and their in vivo effects can be compared.
Wild-type mice (5 weeks old) were used in this study, and
the results were compared to our previously published
results for obese-prone C57BL/6J mice.18,19 When four
groups of mice (n = 7–8) were fed HFD for 8 weeks, they
developed 45% heavier body weights than the LFD-fed
lean controls (n = 7) (39.8 ꢁ 3.1 vs. 27.4 ꢁ 1.3 g). The
obese control group and the three test groups were fed
HFD or HFD mixed with the three ESA compounds for
4 weeks, and changes in weight were monitored. Weight
gains during the 4-week drug treatment period are shown in
Table 2. Suppression of weight gain was significant for
ESA4 and ESA8, and marginal for ESA7 (p = 0.11). No
significant reduction in dietary intake was observed in the
drug-treatment groups, indicating that observed decreases
in weight gain in drug-fed mice was not due to reduced
food intakes. Feed efficiency (defined as the weight gain
divided by the calories consumed) was used to determine
the ability of mice to turn food into weight gain (Table 2).
Feed efficiencies in the drug-fed groups were much lower
than those in the obese control group, and compared to or
lower than those observed in the lean control group. These
results indicate that reduced weight gain was largely due to
greater metabolic dissipation of energy. However, intra-
abdominal fat (epididymal and perirenal pads) weights
showed no sign of drug-induced weight gain decrease at
the end of the 4-week treatment period (Table 2).
The GTT and ITT were performed after treatment. For
these tests, glucose or insulin was injected i.p. into 6-h
fasted mice, and then the changes in blood glucose levels
were measured. For both tests, fasting levels of glucose,
right before the glucose injection, were significantly lower
in drug-fed groups than in the HFD control group (Table 3
and Figure 2). For GTTs, clearance of glucose load was
measured after glucose injection (1 mg/g of body weight)
by measuring serum glucose levels at fixed times
(Figure 2). In all drug-fed groups, glucose challenge did
not result in a rise in glucose levels, though this was
observed in the HFD control group, and at all time points,
glucose levels were significantly lower in the drug-fed
groups than those in the HFD control group. Furthermore,
glucose levels in the drug-fed groups returned to basal
values in rates comparable to those observed in LFD con-
trol mice. These results indicate that the treatment of mice
with ESA4, 7, or 8 lowered fasting glucose levels and
accelerated blood glucose clearance.
PTP1B assays were performed using PTP1B overex-
pressed in Escherichia coli and pNPP as a substrate as pre-
viously described.14 For IKKβ assays, we used
a
commercial ELISA kit (CY-1178, CycLex, Nagano, Japan)
utilizing recombinant IkBα (substrate) immobilized on a
96-well plate, human full length IKKβ containing GST and
His tags overexpressed in a baculovirus system, an anti-
phosphoserine(S32)-IkBα antibody, and
a horseradish
peroxidase-based detection system. In previous studies, we
used a now discontinued kit, which utilized a soluble pep-
tide substrate and a different detection system. IKKβ assay
results obtained using the two different kits were compared.
The IC50 values of SA51 and SA37 obtained with the new
and previous kits were 6.3 and 12 μM, and 3.2 and
0.62 μM. These discrepancies were possibly caused by the
different substrates used.
The inhibitory potencies of ESA compounds were deter-
mined against PTP1B and IKKβ; half-maximal inhibitory
concentrations (IC50) are listed in Table 1. ESA1 weakly
inhibited PTP1B and IKKβ, but was 12-fold more potent
against PTP1B than MDSA. ESA3–5 moderately inhibited
PTP1B, which revealed no evocative differences compared
to the MDSA derivatives. ESA6–8 exhibited low micromo-
lar IC50 values against PTP1B, which were noticeably
lower than those of the MDSA derivatives previously
reported in our laboratory, including SA18, 37, and 51.10
On the other hand, ESA6–8 were less potent against IKKβ
than SA37 and SA51, the strongest group of IKKβ inhibi-
tors previously reported in our laboratory. The potencies of
the ESA compounds were compared with that of a stauros-
porine analog K252a (a known inhibitor of serine/threonine
kinases including IKKβ). The reported IC50 value of
K252a against IKKβ is 0.09–0.18 μM, which is 3–4-fold
lower than that reported for staurosporine.16 For K252a,
we obtained an IC50 value of 6.7 μM, which compares with
a value of 1.3 μM quoted by the manufacturer of the assay
kit used in this study. These data indicate that ESA4–8
have inhibitory potencies against IKKβ comparable to that
of staurosporin.
To assess systematic insulin sensitivity, ITT was per-
formed 3 days after GTT. Insulin was injected i.p. (0.75
unit/kg body weight), and blood glucose levels were meas-
ured at 0 (fasting glucose level), 20, 40, 60, 90, and
120 min after injection (Figure 2). Glucose levels declined
post-injection and reached minima at 20 or 40 min in all
three experimental groups. At 40 min post-injection, hypo-
glycemic responses in the HFD and LFD control groups
were significantly different (p < 0.05). On the other hand,
ESA4, 7, and 8 were selected to determine their effects
in C57BL/6 mice.17 They were chosen based on the inhibi-
tory potency against IKKβ. In addition, the chemical
Bull. Korean Chem. Soc. 2016, Vol. 37, 855–863
© 2016 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
859

BULLETIN OF THE
Article
Ethylenedisalicylic Acid Derivatives and their metabolic effects
KOREAN CHEMICAL SOCIETY
Table 1. Inhibitory effects of ESA compounds against PTP1B and IKKβ.
IC50 (μM)a
IC50 (μM)a
R
H
Compound
PTP1B
IKKβ
Compound
PTP1B
3600 ꢁ 290b
IKKβ
ESA1
ESA2
ESA3
ESA4
ESA5
ESA6
ESA7
ESA8
290 ꢁ 0
1400 ꢁ 760
190 ꢁ 100
32 ꢁ 10
>40
>>40
40
17 ꢁ 3
25
MDSA
>10b
Br
2-Furanyl
Ph
4-CF3-Ph
3,5-(CF3)2-Ph
3-CF3-Ph
SA50
SA51
SA18
78 ꢁ 9
39 ꢁ 7b
20 ꢁ 1b
6.3 ꢁ 0.7b, (3.2 ꢁ 0.9)c
4.7 ꢁ 1
78 ꢁ 8
3.4 ꢁ 0.1
2.5 ꢁ 0.4
7.6 ꢁ 1.1
27
18 ꢁ 4
22 ꢁ 6
SA37
67 ꢁ 5b
12 ꢁ 3b, (0.62 ꢁ 0.20)c
K252a
6.7 ꢁ 1.2
a
Results are means ꢁ standard deviations (SD) of ≥ 2 experiments except those without SDs. Assays were performed as described in the
Experimental section using pNPP as a substrate.
b
c
The assay was performed using a kit (CY-1178) manufactured by CycLex (Nagano, Japan).
Data reproduced from our previous publications10 obtained using an assay kit provided by Cell Signaling Technology (Danvers, MA, USA) no
longer supplied. Previously observed IC50 values were not consistent with those obtained in this study.
Table 2. Effect of ESA4, 7, and 8 on body weight gain, feed efficiency, and fat.a
Dose of drug (mg drug/day/
Feed efficiency
(wt gain/kcal × 100)
Epididymal and retroperitoneal
fat combined (g)
Mice group
kg body weight)b
Body weight gain (g)
HFD
HFD + ESA4
HFD + ESA7
HFD + ESA8
LFD
3.40 ꢁ 0.68
0.89
0.017
0.38
−0.049
0.39
3.20 ꢁ 0.13
3.50 ꢁ 0.20
3.55 ꢁ 0.11
3.52 ꢁ 0.24
1.20 ꢁ 0.14***
108
167
165
0.06 ꢁ 0.23**
1.37 ꢁ 0.63 (p = 0.11)
−0.17 ꢁ 0.67***
1.30 ꢁ 0.53 (p = 0.10)
a
HFD and LFD control groups were fed HFD or LFD for 12 weeks. Test groups were fed HFD for 8 weeks and, then, HFD + drugs for 4 weeks.
Values are mean ꢁ SD (n = 8 in HFD control group, n = 7 in others). Statistical significance of difference between HFD control and the test
groups was calculated by one-way ANOVA.
b
Calculated by (weight of drug in diet) × (average daily food intake per mouse)/(average mouse weight).
**p < 0.005, ***p < 0.001. Statistical significance was not determined for feed efficiency.
hypoglycemic responses in the ESA4 and ESA8-fed
groups were greater than in the HFD control group and
smaller than in the LFD control group, though these differ-
ences were not significant. Thus, no sign of enhanced
response to insulin was observed in any of the drug-treated
groups.
After sacrifice, serum levels of triglyceride, total choles-
terol, and NEFAs were measured and compared with
Bull. Korean Chem. Soc. 2016, Vol. 37, 855–863
© 2016 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
860

BULLETIN OF THE
Article
Ethylenedisalicylic Acid Derivatives and their metabolic effects
KOREAN CHEMICAL SOCIETY
Table 3. Fasting glucose levels after 4-week treatment of ESA compounds.a
HFD
ESA4
ESA7
ESA8
LFD
Before GTT
Before ITT
a
212 ꢁ 15
219 ꢁ 8
157 ꢁ 6*
178 ꢁ 11
188 ꢁ 5*
158 ꢁ 9*
163 ꢁ 9*
178 ꢁ 9**
153 ꢁ 6***
177 ꢁ 7**
Fasting glucose levels were measured just before GTT and ITT and represented as mean ꢁ SEM. Statistical significance of difference between
HFD control and the other groups was calculated by one-way ANOVA.
*p < 0.05, ** p < 0.005, *** p < 0.001.
Figure 2. Glucose tolerance test (A) and insulin tolerance test (B). Drug treatment groups were fed a HFD for 8 weeks followed by
HFD + ESA4, 7, or 8 for 4 weeks. Control groups continued on a HFD or LFD for all 12-week period. Mice fasted for 6 h were injected
i.p. with glucose (1.0 g glucose/kg body weight) or insulin (0.75 unit/kg body weight). Blood glucose levels (mean ꢁ SEM) in test groups
were compared with HFD controls by one-way analysis of variance (ANOVA) (not specified on the graph). In all test groups, fasting glu-
cose levels (0 min) were significantly different from HFD controls (details in Table 3). After 0 min, the p-values were < 0.001 for all of
the data points in (a). In (b), p < 0.001 except for ESA4 (90, 120 min), ESA8 (20, 40 min), LFD control (20 min) where p = 0.002.
those in the HFD control group (Table 4). In ESA7-fed
mice, total cholesterol levels were significantly lower
than in HFD controls. However, no significant reduction
in obesity-related blood parameters was observed in
other drug treatment groups, except for a marginal reduc-
tion in triglyceride level in ESA8-fed mice (p = 0.07).
No significant differences in NEFA levels and only
small differences in triglyceride levels were observed
even between HFD and LFD control mice, which was
possibly due to the use of wild-type mice. Epididymal and
retroperitoneal fat pad were not significantly lower in the
drug treatment groups than in the HFD control group
(Table 2).
To assess toxicity, we measured liver-to-body weight
ratios (Table 5). No significant change in the liver-to-body
weight ratios ratio was observed in either the drug-fed
groups or the HFD group, thus the ESA compounds induce
no overt toxicity.20
Bull. Korean Chem. Soc. 2016, Vol. 37, 855–863
© 2016 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
861

BULLETIN OF THE
Article
Ethylenedisalicylic Acid Derivatives and their metabolic effects
KOREAN CHEMICAL SOCIETY
Table 4. Effect of ESA4, 7, and 8 on blood parameters.a
Mice group
Total cholesterol (mg/dL)
Triglyceride (mg/dL)
NEFA (mEq/L)
HFD
317 ꢁ 17
299 ꢁ 9
32.8 ꢁ 1.7
0.19 ꢁ 0.01
0.17 ꢁ 0.01
0.19 ꢁ 0.01
0.15 ꢁ 0.01
0.17 ꢁ 0.01
HFD + ESA4
HFD + ESA7
HFD + ESA8
LFD
33.1 ꢁ 2.3
239 ꢁ 16*
305 ꢁ 21
187 ꢁ 15***
31.2 ꢁ 6.2
27.9 ꢁ 1.8 (p = 0.07)
27.2 ꢁ 1.3*
a
Values are mean ꢁ SD (n = 8 in HFD control, n = 7 in others). For triglyceride data, n = 7 in HFD control and n = 5 in ESA7 group after deletion
of 1 and 2 abnormal data respectively based on statistical criteria. Statistical significance of difference between HFD control and the other groups was
calculated by one-way ANOVA.
*p < 0.05, ***p < 0.001.
Acknowledgments. This study was supported by Inha
University and a grant from the Basic Science Research
Program through the National Research Foundation of
Korea (NRF) funded by the Ministry of Education, Science
and Technology (2010-0022853).
Table 5. Liver weights and liver-to-body weight ratio after
4-week treatment with ESA compounds.a
Liver-to-body weight
Mice group
HFD
Liver (g)
ratio (mg/g)
2.16 ꢁ 0.10***
5.01 ꢁ 0.19
3.82 ꢁ 0.17
4.63 ꢁ 0.37
3.99 ꢁ 0.28
4.65 ꢁ 0.18
HFD + ESA4 1.54 ꢁ 0.10
HFD + ESA7 1.92 ꢁ 0.15*
HFD + ESA8 1.61 ꢁ 0.12
References
LFD
a
1.33 ꢁ 0.07
1. World Health Organization, Global Status Report on Non-
communicable Diseases 2015, World Health Organization,
Geneva, 2014.
2. D. Yach, D. Stuckler, K. D. Brownell, Nat. Med. 2006,
12, 62.
Values are mean ꢁ SEM (n = 8 in HFD, n = 6 in LFD, n = 7 in
others). Statistical significance of difference between LFD control
and the other groups was calculated by one-way ANOVA. No
significant differences were observed in liver/body weight ratio.
* p < 0.05, ***p < 0.001.
3. S.-C. Yip, S. Saha, J. Chernoff, Trends Biochem. Sci. 2010,
35, 442.
4. M. Elchebly, P. Payette, E. Michaliszyn, W. Cromlish,
S. Collins, A. L. Loy, D. Normandin, A. Cheng, J. Himms-
Hagen, C.-C. Chan, C. Ramachandran, M. J. Gresser,
M. L. Tremblay, B. P. Kennedy, Science 1999, 283, 1544.
5. L. D. Klaman, O. Boss, O. D. Peroni, J. K. Kim,
J. L. Martino, J. M. Zabolotny, N. Moghal, M. Lubkin, Y.-
B. Kim, A. H. Sharpe, A. Stricker-Krongrad, G. I. Shulman,
B. G. Neel, B. B. Kahn, Mol. Cell. Biol. 2000, 20, 5479.
6. B. Hoesel, J. A. Schmid, Mol. Cancer 2013, 12, 86.
7. J. K. Kim, Y.-J. Kim, J. J. Fillmore, Y. Chen, I. Moore,
J. Lee, M. Yuan, Z. W. Li, M. Karin, P. Perret,
S. E. Shoelson, G. I. Shulman, J. Clin. Invest. 2001,
108, 437.
8. M. Yuan, N. Konstantopoulos, J. Lee, L. Hansen, Z.-W. Li,
M. Karin, S. E. Shoelson, Science 2001, 293, 1673.
9. K. E. Wellen, G. S. Hotamisligil, J. Clin. Invest. 2005,
115, 1111.
10. B. Kafle, D. Khadka, B. H. Park, C. K. Kim, H. Cho, Bull.
Korean Chem. Soc. 2016, in press.
11. J. Augé, R. Richard Gil, Tetrahedron Lett. 2002, 43, 7919.
12. O. El-Kabbani, P. J. Scammells, J. Gosling, U. Dhagat,
S. Endo, T. Matsunaga, M. Soda, A. Hara, J. Med. Chem.
2009, 52, 3259.
13. S. P. Khanapure, D. S. Garvey, D. V. Young, M. Ezawa,
R. A. Earl, R. D. Gaston, X. Fang, M. Murty, A. Martino,
M. Shumway, M. Trocha, P. Marek, S. W. Tam,
D. R. Janero, L. G. Letts, J. Med. Chem. 2003, 46, 5484.
14. B. Kafle, H. Cho, Bull. Korean Chem. Soc. 2013, 34, 1275.
15. S. Shrestha, K.-H. Lee, H. Cho, Bull. Korean Chem. Soc.
2004, 25, 1303.
In this study, we developed novel compounds based on
EDSA that inhibit both PTP1B and IKKβ. Of the com-
pounds produced, ESA4, 7, and 8 were administered in
diets to determine their effects in wild-type mice. ESA
compounds were not more potent inhibitors of IKKβ than
SA compounds in our previous studies. However, ESA
compounds were designed to be more favorable in lipophi-
licity and possibly in metabolic stability. ESA4 and ESA8
were found to suppress HFD-induced weight gain effec-
tively, and ESA7 afforded marginal effects. Furthermore,
these antiobesity effects were not due to reduced food
intake, and thus probably caused by elevated metabolic dis-
sipation of energy. ESA4, 7, and 8 also exhibited antidia-
betic effects as evidenced by GTT-determined glucose
clearance rates and fasting glucose levels, despite less obvi-
ous ITT findings. A wild-type strain of mice was used in
this study instead of an obesity-prone strain in our previous
studies using SA compounds, and thus it is difficult to
directly compare the in vivo effects of ESA and SA com-
pounds. However, ESA compounds were effective in both
weight and blood glucose control at the doses similar to or
lower than SA compounds. In our previous studies, SA
compounds were effective either only in weight control or
in blood glucose control. Collectively, the ESA derivatives
ESA4, 7, and 8 were found to be partly improved version
for the treatment of obesity and diabetes.
Bull. Korean Chem. Soc. 2016, Vol. 37, 855–863
© 2016 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
862

BULLETIN OF THE
Article
Ethylenedisalicylic Acid Derivatives and their metabolic effects
KOREAN CHEMICAL SOCIETY
16. T. M. Sadler, M. Achilleos, S. Ragunathan, A. Pitkin,
J. LaRocque, J. Morin, R. Annable, L. M. Greenberger,
P. Frost, Y. Zhang, Anal. Biochem. 2004, 326, 106.
17. K. Mekada, K. Abe, A. Murakami, S. Nakamura, H. Nakata,
K. Moriwaki, Y. Obata, A. Yoshiki, Exp. Anim. 2009,
58, 141.
18. S. Collins, T. L. Martin, R. S. Surwit, J. Robidoux, Physiol.
Behav. 2004, 81, 243.
19. M. S. Winzell, B. Ahrén, Diabetes 2004, 53, S215.
20. S. A. Bailey, R. H. Zidell, R. W. Perry, Toxicol. Pathol.
2004, 32, 448.
Bull. Korean Chem. Soc. 2016, Vol. 37, 855–863
© 2016 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.bkcs.wiley-vch.de
863