Design and synthesis of rosiglitazone-ferulic acid-nitric oxide donor trihybrids for improving glucose tolerance

Article abstract of DOI:10.1016/j.ejmech.2018.10.006

Glucose intolerance is associated with metabolic syndrome and type 2 diabetes mellitus (T2DM) while some new therapeutic drugs, such as rosiglitazone (Rosi), for T2DM can cause severe cardiovascular side effects. Herein we report the synthesis of Rosi-ferulic acid (FA)-nitric oxide (NO) donor trihybrids to improve glucose tolerance and minimize the side effects. In comparison with Rosi, the most active compound 21 exhibited better effects on improving glucose tolerance, which was associated with its NO production, antioxidant and anti-inflammatory activities. Furthermore, 21 displayed relatively high stability in the simulated gastrointestinal environments and human liver microsomes, and released Rosi in plasma. More importantly, 21, unlike Rosi, had little stimulatory effect on the membrane translocation of aquaporin-2 (AQP2) in kidney collecting duct epithelial cells. These, together with a better safety profile, suggest that the trihybrids, like 21, may be promising candidates for intervention of glucose intolerance-related metabolic syndrome and T2DM.

Full text of DOI:10.1016/j.ejmech.2018.10.006

Accepted Manuscript
Design and synthesis of rosiglitazone-ferulic acid-nitric oxide donor trihybrids for
improving glucose tolerance
Jingchao Liu, Zhangjian Huang, Wenhuan Ma, Sixun Peng, Yunman Li, Katrina M.
Miranda, Jide Tian, Yihua Zhang
PII:
S0223-5234(18)30868-7
10.1016/j.ejmech.2018.10.006
EJMECH 10792
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 1 May 2018
Revised Date: 30 September 2018
Accepted Date: 1 October 2018
Please cite this article as: J. Liu, Z. Huang, W. Ma, S. Peng, Y. Li, K.M. Miranda, J. Tian, Y.
Zhang, Design and synthesis of rosiglitazone-ferulic acid-nitric oxide donor trihybrids for improving
glucose tolerance, European Journal of Medicinal Chemistry (2018), doi: https://doi.org/10.1016/
j.ejmech.2018.10.006.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Design and Synthesis of Rosiglitazone-Ferulic
Acid-Nitric Oxide Donor Trihybrids for
Improving Glucose Tolerance
†
,‡,#
†,‡ ,#
§,#
†,‡
Jingchao Liu,
Zhangjian Huang,
Wenhuan Ma, Sixun Peng, Yunman
,
§
,∥
⊥
,†,‡
Li,* Katrina M. Miranda,* Jide Tian, Yihua Zhang*
†
State Key Laboratory of Natural Medicines, China Pharmaceutical University,
Nanjing 210009, China
‡
Jiangsu Key Laboratory of Drug Discovery for Metabolic Diseases, China
Pharmaceutical University, Nanjing 210009, China
§
State Key Laboratory of Natural Medicines, Department of Physiology, China
Pharmaceutical University, Nanjing 210009, China
∥
Department of Chemistry and Biochemistry, University of Arizona, Tucson,
Arizona 85721, United States
⊥
Department of Molecular and Medical Pharmacology, University of California, Los
Angeles, California 90095, United States
#
These authors contributed equally to this work.

ACCEPTED MANUSCRIPT
Graphical Abstract

ACCEPTED MANUSCRIPT
Design and synthesis of rosiglitazone-ferulic acid-nitric oxide donor trihybrids
for improving glucose tolerance
†
,‡,#
†,‡ ,#
§,#
†,‡
,§
Jingchao Liu,
Zhangjian Huang,
Wenhuan Ma, Sixun Peng, Yunman Li,* Katrina
,
∥
⊥
,†,‡
M. Miranda,* Jide Tian, Yihua Zhang*
†
State Key Laboratory of Natural Medicines, China Pharmaceutical University, Nanjing 210009,
China
‡
Jiangsu Key Laboratory of Drug Discovery for Metabolic Diseases, China Pharmaceutical
University, Nanjing 210009, China
§
State Key Laboratory of Natural Medicines, Department of Physiology, China Pharmaceutical
University, Nanjing 210009, China
∥
Department of Chemistry and Biochemistry, University of Arizona, Tucson, Arizona 85721,
United States
⊥
Department of Molecular and Medical Pharmacology, University of California, Los Angeles,
California 90095, United States
#
These authors contributed equally to this work.
Abstract:
Glucose intolerance is associated with metabolic syndrome and type 2 diabetes mellitus (T2DM)
while some new therapeutic drugs, such as rosiglitazone (Rosi), for T2DM can cause severe
cardiovascular side effects. Herein we report the synthesis of Rosi-ferulic acid (FA)-nitric oxide
(NO) donor trihybrids to improve glucose tolerance and minimize the side effects. In comparison
with Rosi, the most active compound 21 exhibited better effects on improving glucose tolerance,
which was associated with its NO production, antioxidant and anti-inflammatory activities.
Furthermore, 21 displayed relatively high stability in the simulated gastrointestinal environments
and human liver microsomes, and released Rosi in plasma. More importantly, 21, unlike Rosi,
1

ACCEPTED MANUSCRIPT
had little stimulatory effect on the membrane translocation of aquaporin-2 (AQP2) in kidney
collecting duct epithelial cells. These, together with a better safety profile, suggest that the
trihybrids, like 21, may be promising candidates for intervention of glucose intolerance-related
metabolic syndrome and T2DM.
Keywords: Glucose intolerance, rosiglitazone, nitric oxide, ferulic acid, type 2 diabetes
mellitus.
1
. Introduction
Type 2 diabetes mellitus (T2DM) is characterized as glucose intolerance, insulin resistance
and insufficient insulin production, leading to hyperglycemia [1]. Although several classes of
new drugs, such as dipeptidyl peptidase 4 (DPP-4) inhibitors, sodium/glucose cotransporter 2
(SGLT-2) inhibitors and the peroxisome proliferator-activated receptor γ (PPARγ) agonists have
been developed for the treatment of T2DM [2, 3], their cardiovascular safety remains a major
concern in clinical applications [4]. Rosiglitazone (Rosi), a thiazolidinedione-based PPARγ
agonist, can increase insulin receptor sensitivity and has been used for treatment of T2DM in the
clinic for several years. However, due to its potential risk of heart failure, it was restricted only
for prescription by the FDA in 2013 [5]. Currently, the exact mechanisms underlying the heart
failure of Rosi are still not clear. Previous studies have suggested that the side effect of Rosi may
be attributed to the body weight gain and fluid retention at least in some patients [6-17]. Further
studies have indicated that Rosi stimulates the expression of aquaporin 2 (AQP2), which is
responsible for water reabsorption in the renal collecting duct, implicating fluid retention [18-
2
4]. Hence, development of new compounds to inhibit AQP2 expression may reduce the Rosi-
related side effect to improve its safety.
Chronic hyperglycemia is associated with oxidative stress, reactive oxygen species (ROS)
production, which can exhaust endogenous thiol and tetrahydrobiopterin (BH ) [25],
4
deteriorating the redox imbalance and endothelium dysfunction by reducing nitric oxide (NO)
generation [25-28]. Furthermore, a reduction in NO bioavailability can result in platelet
aggregation, atherosclerosis and hypertension, leading to heart failure in T2DM patients [28]. In
addition, NO can increase energy expenditure and enhance glucose and lipid metabolism by
2

ACCEPTED MANUSCRIPT
improving insulin resistance [29-32]. Actually, conjugates by coupling an NO donor, such as
diazeniumdiolate (NONOate), organic nitrate or furoxan with nateglinide (2a and 2b),
glibenclamide (3a and 3b), or tolbutamide (4), as shown respectively in Figure 1, can display
comparable or higher anti-diabetic activity with reduced adverse effects by slow releasing NO
[33-35]. In addition, ferulic acid (FA) is an antioxidant with anti-hypertensive [36] and anti-
platelet aggregation [37] activities. Importantly, FA can also improve glucose tolerance [36].
Accordingly, to improve the efficacy and safety of Rosi, we synthesized trihybrids of Rosi-FA-
nitrate or Rosi-FA-diethylamine diazeniumdiolate (DEA/NO) and biologically evaluated for
their effects on glucose tolerance.
Figure 1. Chemical structures of rosiglitazone and representative NO donating anti-diabetic
agents.
2
. Results and discussion
2
.1 Chemistry
Synthesis of the target compounds 9a-d, 15a-d, 21, 26a-d and their corresponding
intermediates 8a-d, 14a-d, 20 and 25 is outlined in Scheme 1. For 8a-d, the starting material FA
was reacted with dibromoalkanes to give monobromo intermediates 6a-d, which were treated
with AgNO in acetonitrile to generate corresponding nitrates 7a-d [38]. The nitrates were then
3
condensed with chloromethyl chloroformate to form 8a-d (Scheme 1A). For 14a-d, FA was
esterified to give methyl ester 10, which underwent similar reactions as for 6a-d and 7a-d,
3

ACCEPTED MANUSCRIPT
successively, to offer nitrates 12a-d [39]. The nitrates were demethylated under basic conditions
to give acids 13a-d [39], which were treated with chloromethyl chlorosulfate in the presence of
catalytic n-Bu NHSO in H O/CH Cl to yield 14a-d (Scheme 1B). For 20, tert-butyl ester of FA
4
4
2
2
2
1
6 was firstly synthesized according to the reported method [36]. Intermediate 17 was generated
by the same procedure as for the synthesis of 8a-d, which was then reacted with the potassium
salt of Rosi in the presecnce of catalytic NaI to generate 18. After removal of tert-butyl in 18 by
trifluoroacetic acid, 19 was obtained, which was treated with chloromethyl chlorosulfate to yield
chloromethyl ester 20 (Scheme 1C). Intermediate 25 was synthesized using similar synthetic
approach above to build carboxylic chloromethyl ester 23 at first, followed by coupling with the
Rosi potassium salt and de-tert-butylation, successively (Scheme 1D). Finally, the target
compounds 9a-d and 15a-d were synthesized by coupling potassium salt of Rosi with
corresponding chloromethyl-containing intermediates 8a-d and 14a-d, respectively, in the
presence of catalytic amount of NaI in DMF (Schemes 1A, 1B and 1E). The target compound 21
was produced referring to the published procedures with minor modification (Scheme 1C) [40].
Briefly, 20 was coupled with DEA/NO in the presence of 15-crown-5 in anhydrous THF under
argon protection at 0 °C. The target compounds 26a-d were generated by the condensation of 25
with 7a-d in the presence of DCC/DMAP, respectively (Scheme 1D).
a
Scheme 1. Synthetic routes of compounds 9a−d, 15a−d, 21 and 26a−d
4

ACCEPTED MANUSCRIPT
O
O
O
O
a
O
b
O
OH
O
nBr
O
nONO2
A
HO
HO
HO
5
6a-d
7a-d
O
O
N
O
O
O
n ONO2
O
O
c
O
O
nONO2
d
O
O
N
N
S
O
O
Cl
O
O
8a-d
9a-d
O
O
O
B
O
O
nO
e
O
a
O
O
b
O
f
5
O
HO
Br
O2NO
nO
10
11a-d
12a-d
O
O
O
O
O
O
O
OH
O
Cl
N
O
g
d
N
N
S
O
O2NO
nO
O2NO
nO
n ONO2
O
13a-d
14a-d
15a-d
O
O
O
O
O
O
O
O
O
O
O
c
O
O
d
g
N
O
C
O
N
N
S
O
O
HO
Cl
O
16
17
18
O
O
O
N
O
O
O
O
O
O
OH
O
O
Cl
h
N
O
N
N
S
O
O
N
N
S
O
O
O
19
O
O
O
O
N
O N
N
i
O
O
N
O
N
N
S
O
O
O
21
O
O
O
O
O
D
d
O
g
O
N
HO
Cl
O
N
N
S
O
O
O
O
O
O
22
23
O
2
4
O
O
O
nONO2
O
O
O
O
OH
h
j
N
O
N
N
N
S
O
N
N
S
O
O
O
O
O
O
25
26a-d
6-9a, 11-15a, 26a n = 3; 6-9b, 11-15b, 26b n = 4; 6-9c, 11-15c, 26c n = 5; 6-9d, 11-15d, 26d n = 6
a
Reagents and conditions: (a) dibromoalkanes, TEA, DMF, 50 °C, overnight; (b) AgNO3,
acetonitrile, protected from light, reflux, overnight; (c) chloromethyl chloroformate, pyridine,
CH Cl
, 0−25 °C; (d) rosiglitazone potassium salt, DMF, catalytic NaI, argon protected, 25 °C;
e) MeOH, catalytic H SO , reflux, 8 h; (f) LiOH, MeOH/THF/H O, 25 °C, overnight; (g)
2
2
(
2
4
2
Na CO , H O, 25 °C, 20 min, and then chloromethyl chlorosulfate, n-Bu NHSO , CH Cl , 0−25
2
3
2
4
4
2
2
°
C, overnight; (h) trifluoroacetic acid, CH Cl , 25 °C, overnight; (i) DEA/NO, anhydrous THF,
2 2
1
5-crown-5, 0 °C, argon protected, 3 h; (j) DCC/DMAP, CH Cl , 25 °C, overnight.
2 2
5

ACCEPTED MANUSCRIPT
1
13
The structures of all target compounds were characterized by H NMR, C NMR and high
resolution MS. The purity of compounds was determined by HPLC and all compounds with >
9
2
2
5% purity were used in biological evaluations.
.2 Biological evaluations
.2.1 Suppressive effect on cellular glucose output.
Trihybrids 9a-d, 15a-d, 21 and 26a-d (10 ꢀM) were first tested for their suppressive effect on
cellular glucose output by determining glucose uptake rate in hepatic cells [41]. Human non-
tumor hepatic L-02 cells were treated with dexamethasone (1 ꢀM) to induce insulin resistance
and cultured in serum-free medium in the presence of vehicle (0.1% DMSO) or each compound
for 48 h. During the last 24-h culture, the cells were stimulated with 1 nM insulin and the glucose
contents in the supernatants were measured for calculating glucose uptake rates. As shown in
Table 1, treatment with 9b, 15a, 15b or 21, like Rosi, significantly suppressed the hepatic
glucose output in these cells as compared with the vehicle-treated model group and the
suppressive effect of 21 was significantly greater than that of Rosi. Preliminary analysis of
structure-activity relationship revealed that the suppressive activity of glucose output (presented
as glucose uptake percent rate) of trihybrid 21 (47.8 ± 2.8) bearing the diazeniumdiolate as an
NO donor moiety was several fold higher than trihybrids with the nitrate as an NO donor moiety
9
a-d (7.1 ± 0.8 to 12.2 ± 1.2), 15a-d (7.1 ± 1.4 to 11.5 ± 2.6) and 26a-d (6.8 ± 2.8 to 10.1 ± 2.8).
The better effect of 21 was probably attributed to higher levels of NO produced by
diazeniumdiolate(s) than nitrate(s) [38, 42-47]. In addition, the effects of 9a-d were similar to
1
5a-d and slightly better than 26a-d, which might be attributed to the various linker(s) used in
these trihybrids, leading to different oil-water partition coefficients and abilities to penetrate the
cells.
Table 1. Trihybrids’ suppressive glucose output in insulin resistant L-02 cells.
Comp.
Glucose uptake (%)
Comp.
Glucose uptake (%)
6

ACCEPTED MANUSCRIPT
*
*
*
**
*
control
model
Rosi
13.8 ± 1.9
4.5 ± 0.7
11.8 ± 2.4
9.6 ± 2.2
12.2 ± 1.2
7.1 ± 0.8
8.0 ± 2.6
11.5 ± 2.6
15b
15c
15d
21
11.1 ± 2.3
7.1 ± 1.4
8.5 ± 2.7
47.8 ± 2.8
10.0 ± 2.4
7.8 ± 1.7
10.1 ± 2.8
6.8 ± 2.8
*
*
**#
9
a
*
9b
26a
26b
26c
26d
9
c
9d
*
*
1
5a
Data are expressed as mean ± SD of each compound from five separate experiments. *P <
#
0
.05, **P < 0.01, ***P < 0.001 vs. the model group. P < 0.01 vs. the Rosi, determined by
Student T-test.
Subsequently, the effects of the active compounds 9b, 15a, 15b and 21 on 2-(N-(7-nitrobenz-
2
-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose (2-NBDG) uptake were tested in insulin resistant
murine 3T3-L1 adipocytes [48]. The 2-NBDG is a fluorescent glucose analogue and sensitively
detected by fluorospectrometer [49]. Murine 3T3-L1 preadipocytes were cultured sequentially in
different media to induce their differentiation. The mature adipocytes were treated with vehicle
(
control) or TNF-α (4 ng/mL) daily for 4 days to induce their insulin resistance [48].
Subsequently, the insulin resistant adipocytes were pre-treated with vehicle (model) or 10 ꢀM
b, 15a, 15b, 21 or Rosi and challenged with 100 ꢀM 2-NBDG at 37 °C for 30 min in the dark in
9
the presence or absence of exogenous insulin. After being washed, the contents of intracellular
fluorescent signals were detected. In comparison with that in the control, the levels of 2-NBDG
uptake in the model group were reduced by 49% in the presence of insulin, demonstrating insulin
resistance (Figure 2). Pre-treatment with Rosi, 9b, or 21, but not 15a or 15b, significantly
increased the levels of insulin-stimulated 2-NBDG uptake in insulin-resistant adipocytes (Figure
2
A). The most active compound 21 significantly improved 2-NBDG uptake with an EC value
50
7

ACCEPTED MANUSCRIPT
of 1.08 ± 0.13 ꢀM, which was better than Rosi (EC50 = 1.72 ± 0.23 ꢀM). Hence, 21 was selected
for further evaluation.
Figure 2. Compound 21 significantly improves 2-NBDG uptake in insulin-resistant 3T3-L1
adipocytes. The induced mature adipocytes were treated with TNF-α (4 ng/ml) to induce insulin
resistance. Subsequently, the insulin-resistant adipocytes were treated in triplicate with vehicle
(model) or the indicated compound (10 µM) for 48 h and challenged with 100 ꢀM 2-NBDG in
the absence (basal uptake) or presence of insulin (insulin stimulated uptake) for 30 min. The
relative levels of 2-NBDG uptake in individual groups of cells were determined. Data are
expressed as the mean ± SD of each group from three separate experiments and the levels of 2-
NBDG uptake in the control cells (without insulin resistance) were designated as 100%. (A) The
levels of 2-NBDG uptake in individual groups of cells. (B) Effect of each moieties in 21 on 2-
#
##
NBDG uptake.
P < 0.001 vs. the control group. *P < 0.05, ***P < 0.001 vs. the model group.
The difference among the groups was analyzed by one-way ANOVA and post hoc Tukey’s test.
Given that 21 is composed of Rosi, FA and DEA/NO moieties, we next evaluated the effect of
each moiety on 2-NBDG uptake in insulin-resistant 3T3-L1 adipocytes. Pre-treatment with 21,
Rosi, 10 (ferulic acid methyl ester), or 27 (obtained by removal of DEA/NO from 21, Supporting
Information, Scheme S1) at 10 ꢀM significantly increased the insulin-stimulated 2-NBDG uptake
in insulin-resistant 3T3-L1 adipocytes and the effect of 21 was higher than that of Rosi, 10, or 27
8

ACCEPTED MANUSCRIPT
(
Figure 2B). Hence, individual moieties of 21 can promote the insulin-stimulated 2-NBDG
uptake in insulin-resistant 3T3-L1 adipocytes.
2
.2.2 Compound 21 improves glucose tolerance in mice.
Diabetic db/db mice spontaneously develop diabetes with severe glucose tolerance
impairment. To determine the effect of 21 on the glucose tolerance, adult db/db mice were
randomized and treated with vehicle (0.5% CMC-Na, model), Rosi at 27 ꢀmol/kg (∼10 mg/kg)
and 21 at 13.5 ꢀmol/kg (∼10 mg/kg) or 27 ꢀmol/kg (∼20 mg/kg) daily by gavage for 18 days (n
=
8 per group). A group of control C57BL/6 mice received the same dose of vehicle. While the
control mice maintained euglycemic throughout the observation period the model group of mice
displayed high levels of fasting blood glucose and their fasting blood glucose levels increased
with time (Figure 3A). In contrast, treatment with Rosi or 21 at 9 days post treatment stopped the
up-trend of fasting glucose and resulted in gradually reduced levels of fasting blood glucose in
mice. More importantly, the levels of fasting blood glucose in the mice that had been treated with
2
1 were significantly lower than that of those with Rosi. Further analysis indicated that treatment
with Rosi or 21 significantly reduced the AUC values of glucose tolerance and the effect of 21
was significantly stronger than that of Rosi in mice at 12 days post treatment (Figure 3B and
Figure S1 of the Supporting Information). Given that Rosi treatment can result in body weight
(BW) gain [14], we tested the BW values in individual mice. While treatment with Rosi
significantly increased the BW, treatment with 21 at 13.5 or 27 ꢀmol/kg for 18 days, the mean
BW gain (the average BW on day 18 minus the average of initial BW on day 1) of db/db mice
was lower than that of Rosi at 27 ꢀmol/kg by 39% and 52%, respectively (Figure S2 and Table
S1 of the Supporting Information). Thus, the significantly reduced levels of fasting blood
glucose and AUC values of glucose tolerance by 21 indicated that treatment with 21 significantly
improved glucose tolerance in diabetic mice.
Additionally, the serum levels of triglyceride (TG), low-density lipoprotein (LDL) and high-
density lipoprotein (HDL) were tested. The db/db mice treated with compound 21 (27 ꢀmol/kg)
exhibited lower levels of serum triglyceride and higher levels of HDL in comparison with those
by treatment of CMC-Na, respectively, suggesting that 21 might affect lipid regulations to some
extent (Figure S3 of the Supporting Information).
9

ACCEPTED MANUSCRIPT
Figure 3. Treatment with 21 improves glucose tolerance in mice. Diabetic db/db mice were
randomized and treated with vehicle 0.5% CMC-Na (model), Rosi or 21 by gavage daily for 18
consecutive days. A group of C57BL/6 mice served as the control. Their fasting blood glucose
levels were determined every three days and oral glucose tolerance OGTT was examined at 12
days post treatment, followed by calculating the AUC values. Data are expressed as mean ±
SEM of each group (n=8). (A) The dynamic changes in the levels of fasting blood glucose. (B)
#
##
The AUC values of OGTT. P < 0.001 vs. the control; *P < 0.05, **P < 0.01, ***P < 0.001 vs.
§
§§
the model; P < 0.001 vs. the Rosi-treated group, calculated by One-way ANOVA and post hoc
Tukey’s test.
2
.2.3 Compound 21 slowly produces NO, contributing to the enhanced 2-NBDG uptake in
insulin-resistance adipocytes.
2
Compound 21 contained the NO donor of O -esterified DEA/NO, which did not undergo
spontaneous decay [50, 51]. Next, we analyzed the contribution of NO to the enhanced 2-NBDG
uptake in insulin-resistance adipocytes using a fluorescent NO detector of 4-amino-5-
methylamino-2,7-difluorofluorescein diacetate (DAF-FM-2DA) [52]. As shown in Figure 4,
DEA/NO released NO rapidly [53] while 21-produced NO gradually increased with time, which
were completely abrogated by treatment with carboxy-PTIO (C-PTIO), an NO scavenger [54].
10

ACCEPTED MANUSCRIPT
However, there was no detectable NO in the cells treated with Rosi. To determine the role of NO
in the 21-enhanced 2-NBDG uptake, the insulin-resistant adipocytes were treated with vehicle
(model), Rosi (10 ꢀM), 21 (3 and 10 ꢀM) or/and C-PTIO (20 ꢀM) for 48 h, respectively and
challenged with 2-NBDG in the presence or absence of insulin for 30 min. The levels of 2-
NBDG uptake in individual groups of cells were determined (Figures 4E and F). Treatment with
C-PTIO alone did not change the levels of 2-NBDG uptake in insulin-resistant 3T3-L1
adipocytes. Treatment with 21 at both concentrations significantly enhanced the insulin-
stimulated 2-NBDG uptake, which was mitigated by the addition of C-PTIO. Therefore, the NO
2
produced from O -esterified DEA/NO contributed to the 21-enhanced insulin-stimulated 2-
NBDG uptake in insulin-resistant 3T3-L1 adipocytes.
2
.2.4 Compound 21 induced adipogenesis.
Rosi is a crucial stimulator of adipogenesis [55]. To determine the effect of 21 on
adipogenesis, pre-adipocyte 3T3-L1 cells were induced for differentiation in the presence or
absence of 21 or Rosi and the degrees of adipogenesis were measured by AdipoRed staining and
fluorescent microscopy. We found that treatment with 21, like Rosi, significantly increased the
contents of intracellular triglyceride in 3T3-L1 cells (Figure 3), suggesting that 21 retained the
capacity of Rosi to stimulate adipogenesis [56] The detailed results are shown in Figure S4 of the
Supporting Information.
11

ACCEPTED MANUSCRIPT
Figure 4. The NO produced by 21 contributes to the 21-enhanced insulin-stimulated 2-NBDG
uptake in insulin-resistant 3T3-L1 adipocytes. (A-D) The levels of NO production. Adipocytes
were treated the indicated compounds for varying periods and exposed to DAF-FM-2DA (10
ꢀ
M) in pH 7.4 PBS at 37 °C for 30 min. The fluorescent signals in individual groups of cells
12

ACCEPTED MANUSCRIPT
were measured. (E-F) The levels of 2-NBDG uptake. Adipocytes were treated with indicated
compounds for 48 h and challenged with 2-NBDG in the presence or absence of insulin for 30
min. The levels of 2-NBDG uptake in individual groups of cells were measured. Data are
expressed as mean ± SD of each group of cells from three separate experiments. **P < 0.01,
*
**P < 0.001 vs. the model group, calculated by One-way ANOVA and post hoc Tukey’s test.
#
#
###
&
&&
&&&
P < 0.01, P < 0.001, P < 0.05, P < 0.01,
analysis.
P < 0.001, calculated by Student’s t-test
2
.2.5 Compound 21 has potent antioxidant activity.
Oxidative stress and high levels of ROS production are associated with insulin resistance [26,
8, 57-61]. Thus, reduction in the ROS levels will be a useful strategy for improving glucose
4
tolerance [48]. Accordingly, the antioxidant activity of individual compounds was determined in
the insulin-resistant adipocytes. While high levels of ROS were detected in the insulin-resistant
adipocytes, treatment with Rosi (10 ꢀM), 21 (3 or 10 ꢀM) for 48 h significantly decreased the
levels of ROS production, which were mitigated by the presence of C-PTIO (20 ꢀM) in the
insulin resistant 3T3-L1 adipocytes (Figure 5A). Further analysis indicated that treatment with
1
0 ꢀM 21 significantly increased the levels of intracellular glutathione (GSH) and superoxide
dismutase (SOD) but decreased the levels of malondialdehyde (MDA, a common lipid
peroxidation product) in the insulin-resistant adipocytes (Figures 5B-D). More importantly, these
effects of 21 were significantly stronger than that of Rosi. Collectively, 21 exhibited potent
antioxidant activity in mature 3T3-L1 adipocytes.
13

ACCEPTED MANUSCRIPT
Figure 5. Compound 21 exhibits potent antioxidant activity in the insulin-resistant 3T3-L1
adipocytes. Following induction of insulin resistance by treatment with TNF-α, adipocytes were
treated with vehicle (model) or individual compounds at the indicated dose in the presence or
absence of C-PTIO (20 µM) for 48 h. The control cells did not received TNF-α treatment and
treated with vehicle alone. The levels of intracellular ROS were determined after staining with
DCFH-DA and the levels of intracellular GSH, SOD and MDA were determined using specific
#
##
kits. Data are expressed as mean ± SD of each group from three individual experiments. P <
§
§§
0
.001 vs. the control. *P < 0.05, **P < 0.01, ***P < 0.001 vs. the model group. P < 0.05, P <
§
§§
0
.01, P < 0.001 vs. the Rosi-treated group, calculated by One-way ANOVA and post hoc
&&&
Tukey’s test.
P < 0.001 calculated by Student’s t-test.
2
.2.6 Stability of 21 in different media.
14

ACCEPTED MANUSCRIPT
To examine in vitro stability, 21 was incubated at 37 °C in the simulated gastric fluid (pH 1.5),
intestinal fluid (pH 6.8), human liver microsomes and Sprague-Dawley (SD) rat plasma,
respectively. At 4 and 24 h post incubation, 21 remained reasonably intact in the simulated
gastric fluid (∼97 and ∼80%) and intestinal fluid (∼97 and ∼75%), as well as at 0.5 and 2 h in
human liver microsomes (∼94% and ∼85%) (see Supporting Information, Tables S2 and S3).
However, incubation of 21 in rat plasma reduced its contents by ∼23% and ∼53%, but increased
levels of Rosi from ∼20% to ∼39% at 0.5 and 24 h post incubation (Figure 6). These results
suggest that 21 may function as a prodrug of Rosi and mainly release it in plasma.
Figure 6. Metabolic hydrolysis of 21 in SD rat plasma. Compound 21 (100 ꢀM) was incubated
in SD rat plasma at 37 °C for 0.5, 2, 4, 7, 16 or 24 h. After the addition of triple volumes of
acetonitrile, the contents of 21 and Rosi in the organic phase were analyzed by HPLC. Data are
expressed as mean ± SD from three separate experiments.
2
.2.7 Compound 21 has a better safety profile.
2
1 does not reduce hematocrit (Hct) value in db/db mice. Treatment with Rosi can lead to
water retention, which reduces the Hct values and be associated with increased risk of heart
failure, particularly in T2DM patients [10, 62]. To determine the safety profile of 21, we first
examined the Hct values in individual groups of mice. Treatment of Rosi at 27 ꢀmol/kg, but not
15

ACCEPTED MANUSCRIPT
2
1 at 13.5 or 27 ꢀmol/kg, significantly reduced the Hct value in db/db mice, compared with the
model group (Figure 7A), a hall mark of increased water retention in mice [10, 63].
1 protects the cardiovascular system. TNF-α is an important factor for insulin resistance
2
and the pathogenesis of heart failure [64-66], and blood cardiac troponin I (cTnI) level is a
biomarker of myocardial damage [67-69]. Further analysis indicated that treatment with 21 at 27
ꢀ
mol/kg, but not others, significantly reduced the levels of serum TNF-α in db/db mice (Figure
B). This suggest that 21 may have anti-inflammatory activity, which is important for improving
7
glucose tolerance and protecting from hyperglycemia-related cardiac complication [70].
Furthermore, treatment with 21, like Rosi, significantly decreased the levels of serum cTnI in
db/db mice and the effect of 21 was significantly stronger than that of Rosi and tended to be
dose-dependent (Figure 7C). The cardiac protective effect of 21 may be attributed to NO and FA,
which have been demonstrated to protect cardiac tissues from ischemic injuries [71, 72].
Figure 7. Serum hematocrit values (A), and concentrations of TNF-α (B) and cTnI (C). db/db
Mice were treated with indicated compounds for 18 days. The tail vein blood was collected and
analyzed to determine the hematocrit values, and the concentrations of TNF-α and cTnI by using
the corresponding commercial ELISA kits. Results are expressed as mean ± SD for six
#
##
#
##
###
individual experiments. P < 0.001 vs. model group (A), P < 0.1, P < 0.01 and P < 0.001
vs. control group (B and C), *P < 0.01, **P < 0.01, ***P < 0.001 by One-way ANOVA.
In addition, treatment with 21 significantly inhibited the ADP- and AA-induced platelet
aggregation in platelet-rich plasma (PRP) with IC values of 2.14 ± 0.28 and 0.69 ± 0.11 mM,
50
which were much lower than that of Rosi (IC50 = 5.69 ± 0.62 and 6.34 ± 0.74 mM) (see
16

ACCEPTED MANUSCRIPT
Supporting Information, Table S4). Given that inhibition on ether-a-go-go-related gene (hERG)
channel can induce long QT syndrome [73-76], a severe cardiac adverse effect of drugs, the
potential inhibitory effects of 21 and Rosi on hERG channel were evaluated by patch-clamp in
Chinese hamster ovary cells that had been stably expressed for the hERG. Compound 21
displayed significantly less inhibition (IC > 40.0 ꢀM ) on hERG activation than Rosi (IC =
5
0
50
2
9.5 ꢀM) (see Supporting Information, Table S5).
2
1 has lower in vivo acute toxicity. To further investigate the safety profile of 21, the acute
toxicity of both Rosi and 21 was tested on ICR mice at indicated doses as shown in Table 2. The
animals were observed for abnormal behavior and mortality up to two weeks post treatment.
Mice treated with Rosi at the lowest dose (500 mg/kg) all survived. However, treatment with
Rosi at the highest dose (5000 mg/kg) killed all the mice. Thus, the median lethal dose (LD )
5
0
value of Rosi was calculated to be 2584.7 mg/kg. Comparatively, for 21, all mice survived at the
dose of 2000 mg/kg, and three of ten mice died at the highest dose (5000 mg/kg) on day 14,
suggesting the LD value of 21 exceeds 5000 mg/kg.
5
0
a
Table 2. Acute toxicity of Rosi and 21 in ICR mice.
a
Dose (mg/kg) Number of mice Total mortality LD50 (mg/kg)
5
4
3
2
1
000
000
000
000
000
10
10
10
10
10
10
10
7
5
3
2584.7
Rosi
1
5
00
0
a
9
5% confidence limits: 1952.8−3325.7 mg/kg.
5
000
4000
000
10
10
10
3
1
0
21
> 5000
2
17

ACCEPTED MANUSCRIPT
a
Mice were treated once with Rosi or 21 at indicated doses via oral gavage on day 1, and the
health status of each mouse was observed daily for 14 days. The total mortality number on day
1
4 was listed.
2
.2.8 Compound 21 inhibits the membrane translocation of aquaporin-2.
Although the mechanism underlying the Rosi-induced water retention has not been fully
understood [18], previous studies have indicated that Rosi can stimulate the apical membrane
translocation of intracellular vesicle AQP2, promoting water reabsorption in the kidney
collecting duct epithelial cells [19, 21, 23]. To understand the molecular mechanisms underlying
the action of 21, rat kidney inner medullary collecting duct (IMCD) cells were transfected for the
expression of human AQP2 and the cells were treated with vehicle (control), 10 or 50 ꢀM 21,
Rosi, 100 ꢀM forskolin (FK, a positive control) for 30 min. The membrane-associated proteins
were biotinylated and the levels of membrane-associated biotinylated and total AQP2 were
determined by Western blot. As shown in Figure 8, there is no obvious difference in the levels of
both glycol- and nature AQP2 expression among the different groups of cells. In comparison
with that of the control, treatment with FK or Rosi significantly increased the relative levels of
membrane-associated biotinylated AQP2, particularly for the glycol-AQP2, in the cells,
consistent with a previous report [21]. In contrast, the relative levels of membrane-associated
biotinylated AQP2 in the 21-treated cells were significantly lower than that of Rosi-treated cells
and treatment with 10 µM 21 almost had no stimulatory effect on the membrane translocation of
AQP2.
Finally, immunofluorescent confocal analysis indicated that while little apical membrane-
associated AQP2 was detected in the control cells, clearly bright anti-AQP2 staining was
observed in the membranes (upper panels) and apical membrane line (bottom panels) of the FK-
or Rosi-treated cells. In contrast, treatment with 50 µM 21 induced low levels of membrane
translocation of AQP2 and treatment with 10 µM 21 resulted in no membrane translocation of
AQP2 in the cells. The mild stimulatory effect of 21 should not significantly increase water
reabsorption in the kidney collecting duct and may explain why treatment with 21 had less
increase the BW than Rosi in mice.
18

ACCEPTED MANUSCRIPT
Figure 8. Compound 21 has little effect on the membrane translocation of AQP2 in IMCD cells.
IMCD cells were transfected with the plasmid for expression of human AQP2. The cells were
treated with vehicle (control) or FK (100 ꢀM), Rosi (10 or 50 ꢀM), or 21 (10 or 50 ꢀM) for 30
min. After the membrane-associated proteins were biotinylated and the membrane-associated
biotinylated and total AQP2 in the different groups of cells were determined by Western blot. In
addition, the different groups of cells were stained with FITC-anti-AQP2 and the distribution of
AQP2 in the different groups of cells were photoimaged under fluorescent confocal microscope.
Data are representative images or expressed as the mean ± SD of each group from three separate
experiments. (A) Western blot analysis of biotinylated and total AQP2. (B) Quantitative analysis
of ratios of membrane-associated biotinylated to total AQP2. (C) Immunofluorescence analysis
of AQP2 distribution in the cells. The upper panels: connected cells. The bottom panels: The
#
##
apical membrane.
P < 0.001 vs. the control group calculated by One-way ANOVA and post
*
***
hoc Tukey’s test. P < 0.05 and P < 0.001 calculated by t-test.
3
. Conclusions
19

ACCEPTED MANUSCRIPT
Given that treatment with Rosi is associated with cardiovascular side effects, particularly in
T2DM patients [5-13, 15], we designed and synthesized new trihybrids of Rosi, FA and an NO
donor. Our data indicated that treatment with 21 not only suppressed the insulin-stimulated
glucose output in non-tumor hepatocytes and promoted 2-NBDG uptake in 3T3-L1 adipocytes,
respectively, but also ameliorated glucose intolerance in diabetic mice without significantly
increasing body weights. The anti-diabetic effects of 21 were better than Rosi and associated
with its ability to stimulate NO production, reduce oxidative stress and decrease serum TNF-α
and cTnI levels in mice. Furthermore, 21 displayed relative stability in the simulated
gastrointestinal fluids and human liver microsomes, and released Rosi mainly in plasma. More
importantly, while treatment with Rosi stimulated the membrane translocation of AQP2, 21 had
little such stimulatory role in IMCD cells. These, together with a better safety profile, suggest
that the trihybrids, like 21, may be promising candidates for intervention of glucose intolerance-
related metabolic syndrome and T2DM.
4
. Experimental section
4
.1 Chemistry.
1
Nuclear magnetic resonance (NMR) spectra were obtained from a Bruker Avance 300 ( H,
1
3
1
13
3
3
00 MHz; C, 75 MHz) or a Bruker Avance 500 ( H, 500 MHz; C, 125 MHz) spectrometer at
00 K using TMS as an internal standard. Mass spectrometry (MS) spectra were recorded on a
Mariner mass spectrometer. Melting points (mp) were measured by a Mel-TEMP II apparatus
and uncorrected. TLC was performed on silica gel GF/UV 254, and column chromatography was
conducted by silica gel (200−300 mesh). The purities of target compounds were characterized by
HPLC analysis (Shimadzu DGU-20A3R) and HRMS (Agilent Technologies LC/MSD TOF).
The target compounds 9a−d, 15a−d, 21 and 26a−d with a purity of >95% were used for
subsequent experiments. Intermediates 6a−d, 7a−d, 11a−d, 12a−d, 13a−d and 16 were
synthesized as previously described [36, 38, 39].
General Procedure A for the Preparation of Intermediates 8a
chloroformate (0.22 mL, 2.5 mmol) was added dropwise to the stirring solution of 7a
mmol) in CH Cl (10 mL) at 0 °C, respectively, followed by the addition of pyridine (0.20 mL,
.5 mmol). The reaction mixture was gradually warmed up to room temperature and was
−
d and 17. Chloromethyl
−
d or 16 (1
2
2
2
20

ACCEPTED MANUSCRIPT
monitored by TLC. After 1 h, the reaction was quenched by pouring into water (50 mL) and
extracted by CH Cl (3 × 20 mL). The combined organic layer was washed successively by
2
2
water (3 × 20 mL) and brine (3 × 20 mL). The organic layer was dried by anhydrous sodium
sulfate, filtered, and evaporated in vacuum to give intermediates 8a−d or 17 which was used
immediately without further purification.
General Procedure B for the Preparation of Intermediates 14a
424 mg, 4 mmol) was added to the stirring suspension of intermediates 13a
mmol) in water (8 mL) at room temperature. The reaction mixture turned transparent gradually
in 20 min, and then dichloromethane (CH Cl , 16 mL) and n-Bu NHSO (68 mg, 0.2 mmol)
2 3
−d, 20 and 23. Na CO
(
−d, 19 or 22 (1
2
2
4
4
were added successively. After the reaction was cooled to 0 °C, chloromethyl chlorosulfate (0.12
mL, 1.2 mmol) was added dropwise. Then, the reaction mixture was gradually warmed up to
room temperature and stirred overnight. The reaction solution was diluted in CH Cl (30 mL)
2
2
and washed successively by water (3 × 20 mL) and brine (3 × 20 mL). The organic layer was
dried by anhydrous sodium sulfate, filtered, and evaporated in vacuum to give intermediates
1
4a
General Procedure C for the Preparation of Intermediates 19 and 25. Trifluoroacetic acid
7.8 mL, 105 mmol) was added to the stirring solution of 18 (2.32 g, 3.5 mmol) or 24 (1.9 g, 3.5
mmol) in CH Cl (30 mL) at room temperature. The reaction mixture was stirred overnight and
−d, 20 or 23 which was used immediately without further purification.
(
2
2
monitored by TLC. After completion of the reaction, the mixture was evaporated in vacuum. The
crude product was purified by column chromatography (MeOH/CH Cl = 1:10) to yield the title
2
2
compound, respectively.
General Procedure D for the Preparation of Compounds 9a
intermediates 8a d, 14a−d, 17 or 23 (1 mmol) was dissolved in DMF (10 mL) and added
−d, 15a−d, 18 and 24. The
−
dropwise to the stirring mixture of rosiglitazone potassium salt (316 mg, 0.8 mmol) in DMF (5
mL) with catalytic amount of NaI at room temperature under argon protection. The reaction
mixture was stirred overnight and then poured into water (80 mL). The mixture was then
extracted by EtOAc (3 × 20 mL) and washed successively by water (3 × 20 mL) and brine (3 ×
2
0 mL). The organic layer was dried by anhydrous sodium sulfate, filtered, and evaporated in
vacuum. The crude product was purified by column chromatography (MeOH/CH
2 2
Cl = 1:200) to
yield the title compound, respectively.
21

ACCEPTED MANUSCRIPT
General Procedure E for the Preparation of Target Compounds 26a−d. Intermediate 25
(
200 mg, 0.41 mmol) was dissolved in CH Cl (10 mL) and 7a
−
d (0.41 mmol), DCC (102 mg,
.49 mmol) as well as DMAP (3 mg, cat.) were added successively. The reaction mixture was
Cl (30 mL) and water (30 mL). The
2
2
0
stirred overnight at room temperature and extracted by CH
2
2
organic phase was successively washed by water (3 × 20 mL) and brine (3 × 20 mL). The
organic layer was dried by anhydrous sodium sulfate, filtered, and evaporated in vacuum. The
crude product was purified by column chromatography (MeOH/CH Cl = 1:200) to yield the title
2
2
compound, respectively.
E)-3-(Nitrooxy)propyl 3-[3-methoxy-4-((((5-(4-(2-(methyl(pyridin-2-yl)amino)ethoxy)benzyl)-
(
2
,4-dioxo-3-thiazolidinyl)methoxy)carbonyl)oxy)phenyl]acrylate (9a). This title compound was
1
obtained in 11% yield as a light yellow solid: mp 56–58 °C. H NMR (CDCl
3
, 500 MHz, δ ppm)
8
1
.13 (ddd, 1H, J = 4.5 Hz, 2.4 Hz, 0.9 Hz, ArH), 7.64 (d, 1H, J = 15.9 Hz, CH═), 7.44−7.41 (m,
H, ArH), 7.16−7.15 (m, 1H, ArH), 7.12−7.09 (m, 4H, ArH), 6.84 (d, 2H, J = 9.2 Hz, ArH), 6.54
(ddd, 1H, J = 6.7 Hz, 5.4 Hz, 0.8 Hz, ArH), 6.49 (d, 1H, J = 8.7 Hz, ArH), 6.38 (d, 1H, J = 16.0
Hz, CH═), 5.67−5.62 (m, 2H, CH ), 4.59 (t, 2H, J = 5.8 Hz, CH ONO ), 4.49 (dd, 1H, J = 9.4
2
2
2
2 2
Hz, 4.0 Hz, CHS), 4.31 (t, 2H, J = 6.6 Hz, CH O), 4.15 (t, 2H, J = 5.7 Hz, CH O), 3.95 (t, 2H, J
=
5.7 Hz, CH N), 3.87 (s, 3H, CH O), 3.47 (dd, 1H, J = 14.2 Hz, 4.0 Hz, CH), 3.11 (s, 3H, CH ),
2 3 3
1
3
3
1
1
6
.17−3.06 (m, 1H, CH), 2.14 (p, 2H, J = 6.2 Hz, CH ); C NMR (CDCl , 125 MHz, δ ppm)
2 3
72.41, 169.64, 166.39, 158.45, 158.30, 151.30, 151.24, 147.86, 144.40, 141.29, 137.30, 133.72,
30.32, 127.32, 122.67, 121.15, 118.15, 114.85, 111.77, 111.54, 105.74, 69.99, 66.28, 66.13,
+
0.61, 56.04, 51.99, 49.43, 37.79, 29.71, 26.50; ESI-MS: 711.2 [M + H] ; HRMS calculated for
+
C H N O S [M + H] 711.1972, found 711.1986.
33
35
4
12
(E)-4-(Nitrooxy)butyl 3-[3-methoxy-4-((((5-(4-(2-(methyl(pyridin-2-yl)amino)ethoxy)benzyl)-
2
,4-dioxo-3-thiazolidinyl)methoxy)carbonyl)oxy)phenyl]acrylate (9b). This title compound was
1
obtained in 14% yield as a light yellow solid: mp 52–53 °C. H NMR (CDCl , 500 MHz, δ ppm)
3
8
1
.13 (ddd, 1H, J = 4.5 Hz, 2.4 Hz, 0.9 Hz, ArH), 7.62 (d, 1H, J = 15.9 Hz, CH═), 7.45−7.41 (m,
H, ArH), 7.16−7.14 (m, 1H, ArH), 7.11−7.08 (m, 4H, ArH), 6.85−6.82 (m, 2H, ArH), 6.54
(
ddd, 1H, J = 6.7 Hz, 5.4 Hz, 0.8 Hz, ArH), 6.49 (d, 1H, J = 8.7 Hz, ArH), 6.38 (d, 1H, J = 16.0
Hz, CH═), 5.67−5.62 (m, 2H, CH ), 4.51−4.49 (m, 3H, CH and CHS), 4.24 (t, 2H, J = 6.6 Hz,
CH O), 4.15 (t, 2H, J = 5.7 Hz, CH O), 3.95 (t, 2H, J = 5.8 Hz, CH N), 3.87 (s, 3H, CH O), 3.47
2
2
2
2
2
3
3
(dd, 1H, J = 14.2 Hz, 4.0 Hz, CH), 3.11 (s, 3H, CH ), 3.12−3.06 (m, 1H, CH), 1.88−1.81 (m, 4H,
22

ACCEPTED MANUSCRIPT
1
3
2
1
1
2
7
× CH
2
); C NMR (CDCl
3
, 125 MHz, δ ppm) 172.40, 169.62, 166.57, 158.45, 158.30, 151.29,
51.24, 147.87, 144.06, 141.24, 137.29, 133.80, 130.31, 127.32, 122.65, 121.10, 118.47, 114.85,
11.76, 111.51, 105.73, 72.68, 66.28, 66.13, 63.68, 56.04, 51.98, 49.43, 37.78, 29.70, 25.09,
+
+
3.72; ESI-MS: 725.2 [M + H] ; HRMS calculated for C34
H
37
N
4
O
12S [M + H] 725.2129, found
25.2117.
(E)-5-(Nitrooxy)pentyl 3-[3-methoxy-4-((((5-(4-(2-(methyl(pyridin-2-yl)amino)ethoxy)benzyl)-
2
,4-dioxo-3-thiazolidinyl)methoxy)carbonyl)oxy)phenyl]acrylate (9c). This title compound was
1
obtained in 10% yield as a light yellow solid: mp 47–49 °C. H NMR (CDCl , 500 MHz, δ ppm)
3
8
.13−8.12 (m, 1H, ArH), 7.61 (d, 1H, J = 15.9 Hz, CH═), 7.42−7.440 (m, 1H, ArH), 7.15−7.13
m, 1H, ArH), 7.09−7.07 (m, 4H, ArH), 6.84−6.81 (m, 2H, ArH), 6.52 (ddd, 1H, J = 6.7 Hz, 5.4
Hz, 0.8 Hz, ArH), 6.48 (d, 1H, J = 8.7 Hz, ArH), 6.38 (d, 1H, J = 16.0 Hz, CH═), 5.66−5.61 (m,
(
2
2
3
1
1
1
6
H, CH ), 4.49 (dd, 1H, J = 9.4 Hz, 4.0 Hz, CHS), 4.44 (t, 2H, J = 5.8 Hz, CH ONO ), 4.20 (t,
2 2 2
H, J = 6.0 Hz, CH O), 4.13 (t, 2H, J = 5.5 Hz, CH O), 3.94 (t, 2H, J = 5.8 Hz, CH N), 3.85 (s,
2
2
2
H, CH
3
O), 3.45 (dd, 1H, J = 14.2 Hz, 4.0 Hz, CH), 3.10 (s, 3H, CH
3
), 3.11−3.05 (m, 1H, CH),
13
.79−1.71 (m, 4H, 2 × CH ), 1.54−1.50 (m, 2H, CH ); C NMR (CDCl , 125 MHz, δ ppm)
2
2
3
72.41, 169.64, 166.66, 158.43, 158.31, 151.27, 151.25, 147.87, 143.84, 141.18, 137.27, 133.86,
30.32, 127.35, 122.62, 121.08, 118.68, 114.83, 111.75, 111.50, 105.71, 73.08, 66.26, 66.13,
+
4.14, 56.03, 51.95, 49.40, 37.74, 28.25, 26.47, 22.34; ESI-MS: 739.2 [M + H] ; HRMS
+
calculated for C35
H
39
N
4
O
12S [M + H] 739.2285, found 739.2293.
(E)-6-(Nitrooxy)hexyl 3-[3-methoxy-4-((((5-(4-(2-(methyl(pyridin-2-yl)amino)ethoxy)benzyl)-
2
,4-dioxo-3-thiazolidinyl)methoxy)carbonyl)oxy)phenyl]acrylate (9d). This title compound was
1
obtained in 12% yield as a light yellow solid: mp 42–44 °C. H NMR (CDCl , 500 MHz, δ ppm)
3
8
1
.14 (ddd, 1H, J = 4.5 Hz, 2.4 Hz, 0.9 Hz, ArH), 7.63 (d, 1H, J = 15.9 Hz, CH═), 7.46−7.43 (m,
H, ArH), 7.17−7.14 (m, 1H, ArH), 7.13−7.09 (m, 4H, ArH), 6.86−6.84 (m, 2H, ArH), 6.55
(
ddd, 1H, J = 6.7 Hz, 5.4 Hz, 0.8 Hz, ArH), 6.51 (d, 1H, J = 8.7 Hz, ArH), 6.39 (d, 1H, J = 16.0
Hz, CH═), 5.68−5.64 (m, 2H, CH ), 4.51 (dd, 1H, J = 9.4 Hz, 4.0 Hz, CHS), 4.46 (t, 2H, J = 5.8
Hz, CH ONO ), 4.22 (t, 2H, J = 6.0 Hz, CH O), 4.16 (t, 2H, J = 5.5 Hz, CH O), 3.97 (t, 2H, J =
2
2
2
2
2
5
3
.8 Hz, CH N), 3.89 (s, 3H, CH O), 3.50 (dd, 1H, J = 14.2 Hz, 4.0 Hz, CH), 3.13 (s, 3H, CH ),
2 3 3
1
3
2 2
.13−3.08 (m, 1H, CH), 1.78−1.72 (m, 4H, 2 × CH ), 1.49−1.46 (m, 4H, 2 × CH ); C NMR
(
CDCl , 125 MHz, δ ppm) 172.39, 169.60, 166.76, 158.47, 151.27, 147.87, 143.77, 141.16,
3
1
37.30, 133.92, 130.30, 127.28, 122.63, 121.08, 118.78, 114.86, 111.76, 111.44, 105.73, 73.19,
23

ACCEPTED MANUSCRIPT
6
7
6.31, 66.11, 64.43, 56.04, 52.01, 49.46, 37.85, 37.82, 28.52, 26.71, 25.60, 25.40; ESI-MS:
+
+
53.2 [M + H] ; HRMS calculated for C H N O S [M + H] 753.2442, found 753.2431.
3
6
41
4
12
(E)-[5-(4-(2-(Methyl(pyridin-2-yl)amino)ethoxy)benzyl)-2,4-dioxo-3-thiazolidinyl]methyl 3-
[
3-methoxy-4-(3-(nitrooxy)propoxy)phenyl]acrylate (15a). This title compound was obtained in
1
2
3% yield as a light yellow solid: mp 55–57 °C. H NMR (CDCl , 500 MHz, δ ppm) 8.14–8.13
3
(
m, 1H, ArH), 7.63 (d, 1H, J = 15.9 Hz, CH═), 7.43 (ddd, 1H, J = 9.1 Hz, 7.1 Hz, 2.0 Hz, ArH),
.10–7.02 (m, 4H, ArH), 6.86–6.82 (m, 3H, ArH), 6.54 (dd, 1H, J = 7.0 Hz, 5.0 Hz, ArH), 6.49
d, 1H, J = 8.6 Hz, ArH), 6.25 (d, 1H, J = 15.9 Hz, CH═), 5.67–5.62 (m, 2H, CH ), 4.68 (t, 2H, J
7
(
2
=
2 2 2
6.1 Hz, CH ONO ), 4.49 (dd, 1H, J = 9.3 Hz, 3.9 Hz, CHS), 4.14–4.12 (m, 4H, 2 × OCH ),
3
3
.93 (t, 2H, J = 5.6 Hz, NCH ), 3.87 (s, 3H, OCH ), 3.47 (dd, 1H, J = 14.1 Hz, 3.8 Hz, CH),
2
3
13
3 2 3
.13–3.06 (m, 4H, CH N and CH), 2.25 (p, 2H, J = 6.0 Hz, CH ); C NMR (CDCl , 125 MHz, δ
ppm) 172.66, 169.96, 165.41, 158.40, 158.29, 150.39, 149.73, 147.86, 146.36, 137.29, 130.34,
1
6
27.70, 127.42, 122.77, 114.79, 114.38, 113.02, 111.77, 110.38, 105.72, 69.86, 66.27, 64.82,
+
3.13, 55.91, 51.94, 49.41, 37.78, 26.89; ESI-MS: 667.2 [M + H] ; HRMS calculated for
+
C H N O S [M + H] 667.2074, found 667.2085.
32
35
4
10
(E)-[5-(4-(2-(Methyl(pyridin-2-yl)amino)ethoxy)benzyl)-2,4-dioxo-3-thiazolidinyl]methyl 3-
[
3-methoxy-4-(4-(nitrooxy)butoxy)phenyl]acrylate (15b). This title compound was obtained in
1
3
1
7
2% yield as light yellow solid: mp 53–55 °C. H NMR (CDCl , 500 MHz, δ ppm) 8.11 (ddd,
3
H, J = 4.9 Hz, 1.9 Hz, 0.7 Hz, ArH), 7.61 (d, 1H, J = 15.9 Hz, CH═), 7.42–7.39 (m, 1H, ArH),
.08–7.00 (m, 4H, ArH), 6.81–6.79 (m, 3H, ArH), 6.53–6.50 (m, 1H, ArH), 6.46 (d, 1H, J = 8.7
Hz, ArH), 6.24 (d, 1H, J = 15.9 Hz, CH═), 5.65–5.59 (m, 2H, CH
2
), 4.51 (t, 2H, J = 6.2 Hz,
CH ONO ), 4.47 (dd, 1H, J = 9.3 Hz, 4.0 Hz, CHS), 4.09 (t, 2H, J = 6.6 Hz, CH O), 4.02 (t, 2H,
2
2
2
J = 5.7 Hz, CH O), 3.90 (t, 2H, J = 5.6 Hz, CH N), 3.84 (s, 3H, CH O), 3.43 (dd, 1H, J = 14.2
2
2
3
1
3
3 2
Hz, 3.9 Hz, CH), 3.07 (s, 3H, CH N), 3.07–3.03 (m, 1H, CH), 1.92–1.91 (m, 4H, 2 × CH ); C
NMR (CDCl , 125 MHz, δ ppm) 172.67, 170.00, 165.44, 158.35, 158.28, 150.72, 149.58,
3
1
1
6
47.86, 146.46, 137.29, 130.37, 127.45, 127.22, 122.90, 114.75, 114.09, 112.44, 111.76, 110.17,
05.72, 72.99, 68.11, 66.21, 63.08, 55.86, 51.89, 49.36, 37.74, 37.68, 25.32, 23.88; ESI-MS:
+
+
81.2 [M + H] ; HRMS calculated for C H N O S [M + H] 681.2230, found 681.2221.
3
3
37
4
10
(E)-[5-(4-(2-(Methyl(pyridin-2-yl)amino)ethoxy)benzyl)-2,4-dioxo-3-thiazolidinyl]methyl 3-
[
3-methoxy-4-(5-(nitrooxy)pentoxy)phenyl]acrylate (15c). This title compound was obtained in
1
1
6% yield as a light yellow solid: mp 45–46 °C. H NMR (CDCl
3
, 300 MHz, δ ppm) 8.15−8.13
24

ACCEPTED MANUSCRIPT
(m, 1H, ArH), 7.64 (d, 1H, J = 15.9 Hz, CH═), 7.47−7.41 (m, 1H, ArH), 7.11−7.03 (m, 4H,
ArH), 6.85−6.81 (m, 3H, ArH), 6.56−6.48 (m, 2H, ArH), 6.25 (d, 1H, J = 15.9 Hz, CH═),
5
.69−5.61 (m, 2H, CH ), 4.51−4.44 (m, 3H, CH ONO and CHS), 4.13 (t, 2H, J = 5.6 Hz,
2
2
2
CH
2
2 2 3
O), 4.04 (t, 2H, J = 6.4 Hz, CH O), 3.93 (t, 2H, J = 5.6 Hz, CH N), 3.88 (s, 3H, CH O), 3.47
(
dd, 1H, J = 14.1 Hz, 3.8 Hz, CH), 3.13−3.05 (m, 4H, CH and CH), 1.94−1.76 (m, 4H, 2 ×
3
1
3
CH
2
2 3
), 1.65−1.55 (m, 2H, CH ); C NMR (CDCl , 75 MHz, δ ppm) 172.67, 170.00, 165.50,
1
1
2
58.38, 150.87, 149.52, 147.87, 146.54, 137.30, 130.34, 127.39, 127.04, 122.94, 114.76, 113.96,
12.31, 111.75, 110.10, 105.70, 73.10, 68.43, 66.25, 63.10, 55.93, 51.95, 49.39, 37.79, 28.54,
+
+
39 4
6.53, 22.41; ESI-MS: 695.2 [M + H] ; HRMS calculated for C34H N O10S [M + H] 695.2387,
found 695.2396.
E)-[5-(4-(2-(Methyl(pyridin-2-yl)amino)ethoxy)benzyl)-2,4-dioxo-3-thiazolidinyl]methyl 3-
(
[
3-methoxy-4-(6-(nitrooxy)hexyloxy)phenyl]acrylate (15d). This title compound was obtained in
1
3
1
7
6% yield as a light yellow solid: mp 39–41 °C. H NMR (CDCl , 500 MHz, δ ppm) 8.13 (ddd,
3
H, J = 4.9 Hz, 1.9 Hz, 0.8 Hz, ArH), 7.64 (d, 1H, J = 15.9 Hz, CH═), 7.43 (ddd, 1H, J = 8.9 Hz,
.1 Hz, 2.0 Hz, ArH), 7.10–7.03 (m, 4H, ArH), 6.85–6.81 (m, 3H, ArH), 6.54 (ddd, 1H, J = 7.0
Hz, 5.0 Hz, 0.6 Hz, ArH), 6.49 (d, 1H, J = 8.6 Hz, ArH), 6.24 (d, 1H, J = 15.9 Hz, CH═), 5.67–
5
4
3
1
.62 (m, 2H, CH ), 4.49 (dd, 1H, J = 9.3 Hz, 4.0 Hz, CHS), 4.44 (t, 1H, J = 4.0 Hz, CH ONO ),
2 2 2
.13 (t, 2H, J = 5.8 Hz, CH O), 4.04 (t, 2H, J = 5.6 Hz, CH O), 3.92 (t, 2H, J = 5.6 Hz, CH N),
2
2
2
.88 (s, 3H, CH
3
O), 3.47 (dd, 1H, J = 14.2 Hz, 3.9 Hz, CH), 3.13–3.07 (m, 4H, CH
3
N and CH),
13
.87–1.84 (m, 2H, CH ), 1.78–1.73 (m, 2H, CH ), 1.54–1.46 (m, 4H, 2 × CH ); C NMR
2
2
2
3
(CDCl , 125 MHz, δ ppm) 172.67, 169.97, 165.51, 158.39, 158.28, 151.02, 149.55, 147.82,
1
7
7
46.57, 137.31, 130.34, 127.42, 126.97, 122.96, 114.79, 113.92, 112.34, 111.77, 110.19, 105.74,
3.23, 68.66, 66.28, 63.11, 55.97, 51.93, 49.43, 37.79, 28.81, 26.68, 25.60, 25.45; ESI-MS:
+
+
09.3 [M + H] ; HRMS calculated for C35
41 4
H N O10S [M + H] 709.2543, found 709.2549.
(E)-tert-Butyl 3-[3-methoxy-4-(((5-(4-(2-(methyl(pyridin-2-yl)amino)ethoxy)benzyl)-2,4-dioxo-
3
7
1
7
-thiazolidinyl)methoxy)carbonyloxy)phenyl]acrylate (18). This title compound was obtained in
1
3% yield as an off-white solid: mp 69–71 °C. H NMR (CDCl , 500 MHz, δ ppm) 8.14 (ddd,
3
H, J = 5.0 Hz, 2.0 Hz, 0.9Hz, ArH), 7.53 (d, 1H, J = 15.9 Hz, CH═), 7.44 (ddd, 1H, J = 8.9 Hz,
.1 Hz, 2.0 Hz, ArH), 7.13−7.07 (m, 5H, ArH), 6.87−6.84 (m, 2H, ArH), 6.55 (ddd, 1H, J = 7.1
Hz, 5.0 Hz, 0.7Hz, ArH), 6.51 (d, 1H, J = 8.6 Hz, ArH), 6.32 (d, 1H, J = 15.9 Hz, CH═),
.68−5.64 (m, 2H, CH ), 4.51 (dd, 1H, J = 9.5 Hz, 4.0 Hz, CHS), 4.16 (t, 2H, J = 5.7 Hz, CH O),
5
2
2
25

ACCEPTED MANUSCRIPT
3
.97 (t, 2H, J = 5.7 Hz, CH
2 3
N), 3.88 (s, 3H, CH O), 3.50 (dd, 1H, J = 14.2 Hz, 4.0 Hz, CH), 3.13
1
3
(
s, 3H, CH N), 3.13−3.08 (m, 1H, CH), 1.54 (s, 9H, 3 × CH ); C NMR (CDCl , 125 MHz, δ
3
3
3
ppm) 172.39, 169.59, 166.01, 158.47, 151.29, 151.20, 147.90, 142.53, 140.92, 137.28, 134.21,
1
5
30.30, 127.29, 122.53, 120.98, 120.91, 114.86, 111.76, 111.27, 105.71, 80.74, 66.31, 66.09,
+
6.01, 52.00, 49.45, 37.85, 37.80, 30.94, 28.20; ESI-MS: 664.2 [M + H] ; HRMS calculated for
+
C
34
H
38
N
3
O
9
S [M + H] 664.2329, found 664.2336.
(E)-3-[3-methoxy-4-(((5-(4-(2-(methyl(pyridin-2-yl)amino)ethoxy)benzyl)-2,4-dioxo-3-
thiazolidin)methoxy)carbonyloxy)phenyl]acrylic acid (19). This title compound was obtained in
1
9
5
8% yield as a white solid: mp 125–126 °C. H NMR (CDCl
3
, 500 MHz, δ ppm) 8.27 (d, 1H, J =
.9 Hz, ArH), 7.80 (t, 1H, J = 8.1 Hz, ArH), 7.69 (d, 1H, J = 15.9 Hz, CH═), 7.17–7.10 (m, 5H,
), 4.49
dd, 1H, J = 9.1 Hz, 4.1 Hz, SCH), 4.26 (t, 2H, J = 4.8 Hz, CH ), 4.11 (t, 2H, J = 4.8 Hz, CH ),
ArH), 6.90–6.80 (m, 4H, ArH), 6.40 (d, 1H, J = 15.9 Hz, CH═), 5.67–5.62 (m, 2H, CH
2
(
2
2
3
.89 (s, 3H, OCH ), 3.47–3.43 (m, 1H, CH), 3.37 (s, 3H, NCH ), 3.11 (dd, 1H, J = 14.3 Hz, 9.1
3
3
13
Hz, CH); C NMR (CDCl
3
, 75 MHz, δ ppm) 172.28, 169.79, 169.42, 157.59, 152.88, 151.32,
1
1
51.08, 145.30, 142.70, 141.43, 139.35, 133.69, 130.52, 128.06, 122.60, 121.28, 118.15, 114.73,
12.02, 111.63, 110.66, 66.15, 65.66, 56.06, 51.68, 51.50, 39.51, 37.71; ESI-MS: 608.2 [M +
+
+
H] ; HRMS calculated for C H N O S [M + H] 608.1703, found 608.1697.
3
0
30
3
9
2
O -[(E)-3-(3-methoxy-4-((5-(4-(2-(methyl(pyridin-2-yl)amino)ethoxy)benzyl)-2,4-dioxo-3-
thiazolidinyl)methoxy)carbonyloxy)phenyl)acryloyl]methyl 1-(N,N-diethylamino)diazen-1-ium-
1
,2-diolate (21). Chloromethyl ester 20 (150 mg, 0.2 mmol) was diluted in 5 mL anhydrous
tetrahydrofuran (THF) and added to the stirring slurry mixture of 74 mg (0.48 mmol) DEA/NO
27c) in 5 mL anhydrous THF in the presence of catalytic amount of 15-crown-5 under argon
(
protection at 0 °C. The reaction was stirred for 3 h at 0 °C and poured into water (50 mL)
followed by extraction with EtOAc (3 × 20 mL) and successive wash with water (3 × 20 mL)
and brine (3 × 20 mL). The organic layer was dried by anhydrous sodium sulfate, filtered, and
evaporated in vacuum to give the crude product which was purified by column chromatography
(
MeOH/CH Cl = 1:200) to yield the title compound in 53% yield as a light yellow solid: mp
2 2
1
6
7
7–69 °C. H NMR (CDCl , 500 MHz, δ ppm) 8.14 (ddd, 1H, J = 4.5 Hz, 2.4 Hz, 0.9 Hz, ArH),
3
.71 (d, 1H, J = 16.0 Hz, CH═), 7.46−7.42 (m, 1H, ArH), 7.17−7.09 (m, 5H, ArH), 6.86−6.84
(m, 2H, ArH), 6.55 (ddd, 1H, J = 6.7 Hz, 5.4 Hz, 0.8 Hz, ArH), 6.51 (d, 1H, J = 8.6 Hz, ArH),
6
.38 (d, 1H, J = 16.0 Hz, CH═), 5.96 (s, 2H, CH ), 5.68−5.63 (m, 2H, CH ), 4.51 (dd, 1H, J =
2
2
26

ACCEPTED MANUSCRIPT
9
3
3
.4 Hz, 4.0 Hz, CHS), 4.16 (t, 2H, J = 5.7 Hz, CH
2 2
O), 3.97 (t, 2H, J = 5.8 Hz, CH N), 3.89 (s,
H, CH O), 3.49 (dd, 1H, J = 14.2 Hz, 4.0 Hz, CH), 3.22 (q, 4H, J = 7.1 Hz, 2 × NCH ), 3.13 (s,
3
2
13
H, CH N), 3.14−3.08 (m, 1H, CH) , 1.12 (t, 6H, J = 7.1 Hz, 2 × CH ); C NMR (CDCl , 125
3
3
3
MHz, δ ppm) 172.37, 169.57, 164.80, 158.46, 158.29, 151.32, 151.18, 147.88, 145.59, 141.47,
1
6
37.28, 133.52, 130.29, 127.26, 122.71, 121.30, 117.50, 114.86, 111.76, 111.60, 105.71, 87.59,
+
6.32, 66.13, 56.06, 52.01, 49.45, 48.13, 37.85, 37.81, 11.48; ESI-MS: 753.3 [M + H] ; HRMS
+
calculated for C H N O S [M + H] 753.2554, found 753.2563.
3
5
41
6
11
tert-Butyl
[(5-(4-(2-(methyl(pyridin-2-yl)amino)ethoxy)benzyl)-2,4-dioxo-3-thiazolidinyl)
methyl]succinate (24). This title compound was obtained in 52% yield as a light yellow solid: mp
1
5
7
3–55 °C. H NMR (CDCl , 500 MHz, δ ppm) 8.14 (ddd, 1H, J = 4.9 Hz, 2.0 Hz, 0.9 Hz, ArH),
3
.44 (t, 1H, J = 8.3 Hz, ArH), 7.11–7.08 (m, 2H, ArH), 6.85–6.82 (m, 2H, ArH), 6.54 (ddd, 1H,
J = 7.1 Hz, 5.0 Hz, 0.8 Hz, ArH), 6.51 (dt, 1H, J = 8.6 Hz, 0.8 Hz, ArH), 5.55–5.50 (m, 2H,
CH ), 4.46 (dd, 1H, J = 9.4 Hz, 3.9Hz, CHS), 4.16 (t, 2H, J = 5.7 Hz, CH O), 3.97 (t, 2H, J = 5.7
2
2
Hz, CH
2
N), 3.46 (dd, 1H, J = 14.2Hz, 3.9 Hz, CH), 3.13 (s, 3H, CH
3
N), 3.08 (dd, 1H, J = 14.2
1
3
Hz, 9.4 Hz, CH), 2.58–2.50 (m, 4H, 2 × CH CO), 1.44 (s, 9H, 3 × CH ); C NMR (CDCl , 125
2
3
3
MHz, δ ppm) 172.51, 171.05, 171.03, 169.77, 162.52, 158.40, 158.29, 147.85, 137.29, 130.32,
1
27.39, 114.76, 111.75, 105.72, 80.93, 66.30, 63.20, 51.88, 49.45, 37.80, 37.75, 36.46, 31.42,
+
+
3
0.01, 28.89, 28.03; ESI-MS: 544.2 [M + H] , 566.2 [M + Na] ; HRMS calculated for
+
+
C
27 34 3 7 33 3 7
H N O S [M + H] 544.2177, found 544.2127, C27H N O NaS [M + H] 566.1937, found
5
66.1931.
-[(5-(4-(2-(Methyl(pyridin-2-yl)amino)ethoxy)benzyl)-2,4-dioxo-3-thiazolidinyl)methoxy]-4-
4
oxobutanoic acid (25). This title compound was obtained in 97% yield as a white solid: mp 118–
1
1
7
2
19 °C. H NMR ((CD ) CO, 500 MHz, δ ppm) 8.23 (ddd, 1H, J = 6.1 Hz, 1.8 Hz, 0.7 Hz, ArH),
3
2
.94 (ddd, 1H, J = 9.0 Hz, 7.0 Hz, 1.8 Hz, ArH), 7.26 (d, 1H, J = 9.2 Hz, ArH), 7.19–7.17 (m,
H, ArH), 6.92–6.87 (m, 3H, ArH), 5.53–5.45 (m, 2H, CH ), 4.85 (dd, 1H, J = 8.1 Hz, 4.4 Hz ,
2
2 2
CHS), 4.31 (t, 2H, J = 5.3 Hz, CH O), 4.16 (t, 2H, J = 5.3 Hz, CH N), 3.42–3.38 (m, 1H, CH),
13
3
.40 (s, 3H, CH N), 3.24–3.20 (m, 1H, CH), 2.58–2.53 (m, 4H, 2 × CH CO); C NMR
3
2
(
(CD ) CO, 125 MHz, δ ppm) 173.34, 171.45, 170.46, 158.57, 154.70, 142.66, 140.07, 131.42,
3 2
1
4
28.96, 115.04, 112.53, 111.62, 66.10, 63.64, 51.83, 51.32, 38.70, 37.25, 29.09, 28.75; ESI-MS:
+
+
88.1 [M + H] ; HRMS calculated for C H N O S [M + H] 488.1491, found 488.1499.
2
3
26
3
7
27