Impact of the nature of the substituent at the 3-position of 4 H -1,2,4-benzothiadiazine 1,1-dioxides on their opening activity toward ATP-sensitive potassium channels

Article abstract of DOI:10.1021/jm200100c

Figure Presented. The synthesis of diversely substituted 3-isopropoxy-, 3-isopropylsulfanyl-, 3-isopropylsulfinyl-, and 3-isobutyl-4H-1,2,4- benzothiadiazine 1,1-dioxides is described. Their activity on pancreatic β-cells (inhibitory effect on the insulin releasing process) and on vascular and uterine smooth muscle tissues (myorelaxant effects) was compared to that of previously reported K_ATP channel openers belonging to 3-isopropylamino-4H-1,2,4-benzothiadiazine 1,1-dioxides. The present study aimed at evaluating the impact on biological activity of the isosteric replacement of the NH group of 3-alkylamino-4H-1,2,4-benzothiadiazine 1,1-dioxides by a O, S, S( - O), or CH₂ group. By comparing compounds bearing identical substituents, the following rank order of potency on pancreatic β-cells was observed: 3-isopropylamino > 3-isobutyl > 3-isopropoxy > 3-isopropylsulfanyl > 3-isopropylsulfinyl-substituted 4H-1,2,4- benzothiadiazine 1,1-dioxides (NH > CH₂ > O > S > S( - O)). A molecular modeling study revealed that 3-isopropoxy-, 3-isopropylsulfanyl-, and 3-isopropylamino-substituted compounds adopted a similar low-energy conformation (preferred orientation of the isopropyl chain). Moreover, no direct relationship was detected between the conformational freedom of the different classes of benzothiadiazines (from the most to the lowest conformationally constrained compounds: NH > O > S > CH₂) and their biological activity on insulin-secreting cells. Therefore, the present study confirmed the critical role of the NH group at the 3-position for the establishment of a strong hydrogen bond responsible for optimal activity expressed by 3-alkylamino-4H-1,2,4-benzothiadiazine 1,1-dioxides on insulin-secreting cells. Radioisotopic and fluorimetric experiments conducted with 7-chloro-3-isopropoxy-4H-1,2,4-benzothiadiazine 1,1-dioxide 10c demonstrated that such a compound, bearing a short branched O-alkyl group instead of the NH-alkyl group at the 3-position, also behaved as a specific K_ATP channel opener. Lastly, the present work further identified 3-(alkyl/aralkyl)sulfanyl-substituted 7-chloro-4H-1,2,4-benzothiadiazine 1,1-dioxides as a class of promising myorelaxant drugs acting on uterine smooth muscles, at least in part, through the activation of K_ATP channels.

Full text of DOI:10.1021/jm200100c

ARTICLE
pubs.acs.org/jmc
Impact of the Nature of the Substituent at the 3-Position of
4H-1,2,4-Benzothiadiazine 1,1-Dioxides on Their Opening
Activity toward ATP-Sensitive Potassium Channels
Bernard Pirotte,^,†,* Pascal de Tullio,^,† Stꢀephane Boverie,† Catherine Michaux,‡ and Philippe Lebrun§
†Laboratoire de Chimie Pharmaceutique, Centre Interfacultaire de Recherche du Mꢀedicament (Drug Research Center),
Universitꢀe de Liꢁege, C.H.U., 1 Avenue de l’H^opital, B-4000 Liꢁege, Belgium
‡Unitꢀe de Chimie Physique Structurale et Thꢀeorique, Facultꢀes Universitaires Notre Dame de la Paix, 61, rue de Bruxelles,
B-5000 Namur, Belgium
§Laboratoire de Pharmacodynamie et de Thꢀerapeutique, Universitꢀe Libre de Bruxelles, Facultꢀe de Mꢀedecine, 808,
Route de Lennik, B-1070 Bruxelles, Belgium
S Supporting Information
b
ABSTRACT: The synthesis of diversely substituted 3-isopro-
poxy-, 3-isopropylsulfanyl-, 3-isopropylsulfinyl-, and 3-isobutyl-
4H-1,2,4-benzothiadiazine 1,1-dioxides is described. Their ac-
tivity on pancreatic β-cells (inhibitory effect on the insulin
releasing process) and on vascular and uterine smooth muscle tissues (myorelaxant effects) was compared to that of previously
reported KATP channel openers belonging to 3-isopropylamino-4H-1,2,4-benzothiadiazine 1,1-dioxides. The present study aimed at
evaluating the impact on biological activity of the isosteric replacement of the NH group of 3-alkylamino-4H-1,2,4-benzothiadiazine
1,1-dioxides by a O, S, S(dO), or CH2 group. By comparing compounds bearing identical substituents, the following rank order of
potency on pancreatic β-cells was observed: 3-isopropylamino > 3-isobutyl > 3-isopropoxy > 3-isopropylsulfanyl > 3-isopropyl-
sulfinyl-substituted 4H-1,2,4-benzothiadiazine 1,1-dioxides (NH > CH2 > O > S > S(dO)). A molecular modeling study revealed
that 3-isopropoxy-, 3-isopropylsulfanyl-, and 3-isopropylamino-substituted compounds adopted a similar low-energy conformation
(preferred orientation of the isopropyl chain). Moreover, no direct relationship was detected between the conformational freedom
of the different classes of benzothiadiazines (from the most to the lowest conformationally constrained compounds: NH > O > S >
CH2) and their biological activity on insulin-secreting cells. Therefore, the present study confirmed the critical role of the NH group
at the 3-position for the establishment of a strong hydrogen bond responsible for optimal activity expressed by 3-alkylamino-4H-
1,2,4-benzothiadiazine 1,1-dioxides on insulin-secreting cells. Radioisotopic and fluorimetric experiments conducted with 7-chloro-
3-isopropoxy-4H-1,2,4-benzothiadiazine 1,1-dioxide 10c demonstrated that such a compound, bearing a short branched O-alkyl
group instead of the NH-alkyl group at the 3-position, also behaved as a specific KATP channel opener. Lastly, the present work
further identified 3-(alkyl/aralkyl)sulfanyl-substituted 7-chloro-4H-1,2,4-benzothiadiazine 1,1-dioxides as a class of promising
myorelaxant drugs acting on uterine smooth muscles, at least in part, through the activation of KATP channels.
’ INTRODUCTION
muscle cells.3 A putative mitochondrial KATP channel (mitoKATP
channel) has also been reported in the literature, but the exact identity
of the pore-forming subunits still remains controversial.4ꢀ6 The latter
channel is expected to be involved in myocardial preconditioning and
cytoprotection in different tissues.4ꢀ6
‘Potassium channel openers’ (PCOs) represent a pharmaco-
logical class of drugs able to activate the KATP channels. (ꢀ)-
Cromakalim (1), (()-pinacidil (2), and diazoxide (3) (Figure1) are
typical examples of such compounds.7,8 Among these reference
PCOs, diazoxide remains an important pharmacological tool and is
currently used by some research groups as a ‘selective’ opener of the
cardiac mitoKATP channels.9 This compound is known to activate the
SUR1- and SUR2B-type KATP channels but appears to be only
weakly active on the SUR2A/Kir6.2 channels.3,10 Thus, at least in
Among the wide variety of potassium channels, the ATP-
sensitive potassium channels (KATP channels) represent a particular
type for which opening and closing processes are mainly linked to
changes in intracellular levels of adenine nucleotides (ADP, ATP).1,2
KATP channels are also known to be complex octameric structures
combining two kinds of transmembrane proteins, the ‘sulfonylurea
receptor’ (SURx) subunit and the ‘inwardly rectifying potassium
channel’ (Kir6.x) subunit. The assembly of the Kir6.x (Kir6.1 and
Kir6.2) and the SURx (SUR1, SUR2A, and SUR2B) subunits in
multiple combinations led to the existence of different isoforms of
KATP channels diversely distributed throughout tissues.2 For example,
four SUR1 subunits combine with four Kir6.2 subunits to form the
SUR1/Kir6.2 KATP channel subtype as found in the endocrine
pancreas and the brain, whereas a SUR2A/Kir6.2 channel subtype
is expressed in the cardiac and the skeletal muscle cells, and a SUR2B/
Kir6.1 or a SUR2B/Kir6.2 channel subtype is found in smooth
Received: October 21, 2010
Published: March 24, 2011
r
2011 American Chemical Society
3188
dx.doi.org/10.1021/jm200100c J. Med. Chem. 2011, 54, 3188–3199
|

Journal of Medicinal Chemistry
ARTICLE
The previously reported 6- and/or 7-substituted 3-oxo-3,4-
dihydro-2H-1,2,4-benzothiadiazine 1,1-dioxides 8 were used as start-
ing materials.12ꢀ14 Usually, direct alkylation of 3-oxo-3,4-dihydro-
2H-1,2,4-benzothiadiazine 1,1-dioxides preferentially occurred on
the nitrogen atom at the 2-position, and poor yields of the product
from O-alkylation were obtained. We have observed that O-alkyla-
tion can be favored if the steric hindrance of the alkylating agent was
increased. When the bulky isopropyl iodide was used, a reasonably
good yield of the product of O-alkylation was obtained. To separate
the latter from the product of N-alkylation, the mixture of com-
pounds 9 and 10 was suspended in an aqueous solution of sodium
hydroxide. Compounds of general formula 10 were solubilized by
forming a sodium salt (by deprotonation at the 4-position), while
compounds of general formula 9remained insoluble. After separation
of the insoluble material by filtration, the filtrate regenerated the
desired compound as a precipitate after acidification.
For the synthesis of the 3-isopropylsulfanyl-4H-1,2,4-ben-
zothiadiazine 1,1-dioxides 12, the ‘3-oxo’ derivatives 8 were con-
verted into the corresponding 3-thioxo-3,4-dihydro-2H-1,2,4-benzo-
thiadiazine 1,1-dioxides 11 by means of phosphorus pentasulfide.12,13
Alkylation always occurred on the sulfur atom, generating the desired
compounds 12 in good yields.
The synthesis of 3-isobutyl-4H-1,2,4-benzothiadiazine 1,1-
dioxides 15 was performed according to the process described for
the 3-cyclopentyl-substituted analogue of diazoxide (Scheme 2).16
The selective acylation of the primary amine function was accom-
plished by using 1 equiv of the acyl chloride in the presence of
pyridine at low temperature. Ring closure occurred by heating the
intermediates 14 in an aqueous alkaline medium. After acidification,
the expected compounds 15 precipitated and were collected by
filtration.
Figure 1. Typical examples of reference PCOs (1ꢀ3) and recently
described benzothiadiazine-type analogues of diazoxide (4ꢀ7).
cardiomyocytes into which SUR2A/Kir6.2 KATP channels are ex-
pressed at the plasma membrane, the cell protective activity of
diazoxide could be linked to the opening of the mitoKATP channels.11
In our efforts to develop SUR1-specific PCOs, we previously
prepared several series of original diazoxide analogues character-
ized by the introduction of an alkylamino side chain at the 3-position
of the benzothiadiazine ring.12ꢀ15 Compared to diazoxide, such a
chemical modification induced an improvement of the KATP channel
opening activity.12ꢀ14 Moreover, according to the nature of the alkyl
chain at the 3-position and the nature of the substituents on the
aromatic ring, these original series of drugs provided either potent
and selective activators of the endocrine pancreatic KATP channels
(i.e., 4, 5, 6; Figure 1) or potent myorelaxant drugs devoid of activity
on pancreatic β-cells (i.e., 7; Figure 1).12ꢀ14 These previous works
on benzothiadiazine dioxides also highlighted the fact that the
presence of a very short branched alkylamino chain at the 3-position
(preferably an isopropylamino chain) was responsible for the
selectivity toward SUR1-type channels.12ꢀ15
The aim of the present study was to assess the biological
impact of the isosteric replacement, by an isopropoxy, an isopropyl-
sulfanyl, an isopropylsulfinyl, or an isobutyl chain, of the isopropy-
lamino side chain at the 3-position (concretely the replacement of the
NH group by a O, S, S(dO), or CH2 group) on the putative
inhibitory effect of these diazoxide analogues on the insulin releasing
process and the smooth muscle contractile activity.
The mechanism of action of active representatives from
different series was also determined. Moreover, a molecular modeling
approach was used to highlight the differences between these drugs,
especially regarding the conformational space and the most prefer-
able conformations adopted by typical examples of compounds.
Finally, the pKa values of representative drugs were determined to
predict their ionization state at physiological pH.
The sulfoxide derivative 16 was obtained by treating an
alkaline solution of compound 12c with bromine, which, under
these conditions, was converted into sodium hypobromite by
dismutation (Scheme 3).
’ RESULTS AND DISCUSSION
Activity on Pancreatic β-Cells. The ability of the newly synthe-
sized compounds to inhibit the glucose-induced insulin secretion from
isolated rat pancreatic islets is reported in Table 1. The in vitro data are
expressed as the percentage of residual insulin release recorded at
different drug concentrations and are compared to the results obtained
with diazoxide and the previously described 6- or 7-substituted
3-isopropylamino-4H-1,2,4-benzothiadiazine 1,1- dioxides.12ꢀ15
The 3-isopropylamino-substituted benzothiadiazine 1,1-diox-
ides 17 clearly behaved as the most potent compounds inhibiting the
insulin secretory process. Among this series, the 6-chloro-substituted
compound 17g was the most active. The 3-isobutyl-substituted
benzothiadiazine 1,1-dioxides 15 expressed a moderate activity,
roughly similar to that of diazoxide, another example of 3-alkyl-4H-
1,2,4-benzothiadiazine 1,1-dioxide, except for 15c, which was some-
what more potent than the reference compound at inhibiting insulin
release (diazoxide: IC50 = 18.4 μM;17 15c: estimated IC50 = 5.8 μM;
see Table 2). 3-Isopropoxy-4H-1,2,4-benzothiadiazine 1,1-dioxides
10 were found to be equipotent or even slightly less potent than the
3-alkyl-substituted benzothiadiazine 1,1-dioxides 15. Finally, exam-
ination of the results obtained at 10 μM and 50 μM (when available)
showed that 3-isopropylsulfanyl-4H-1,2,4-benzothiadiazine 1,1-diox-
ides 12 expressed an activity on pancreatic β-cells somewhat less
pronounced than their ‘isopropoxy-substituted’ counterparts 10. By
’ CHEMISTRY
Access to 6- and/or 7-substituted 3-isopropoxy-4H-1,2,4-ben-
zothiadiazine 1,1-dioxides 10 and 3-isopropylsulfanyl-4H-1,2,4-ben-
zothiadiazine 1,1-dioxides 12 is described in Scheme 1.
3189
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
Scheme 1a
a Reagents: (i) NaOH, ICH(CH3)2, DMF, Δ; (ii) P2S5, pyridine, Δ; (iii) NaOH, ICH(CH3)2, CH3CH2NO2, DMF, Δ; (iv) NaOH, R-Br,
CH3CH2NO2, DMF, Δ.
Scheme 2a
dioxides 15, that the best choice of substituent at the 7-position was
the chlorine atom (Cl > Br > F > H), although, in the 3-alkylamino
series 17, there was no major differences between the activity of the
7-chloro and the 7-bromo compounds (compare 4 and 17d). In the
3-isopropoxy and the 3-isopropylsulfanyl series 10 and 12, the
7-bromo compounds became the most effective at inhibiting the
insulin releasing process (Br > Cl > F), although differences were
small and sometimes not significant. Interestingly, in the 3-isopro-
pylsulfanyl series 12, the 6-bromo and 6-chloro compounds appeared
to exert a more pronounced activity than their corresponding
7-bromo and 7-chloro counterparts (compare 12h with 12d and
12g with 12c). The 6,7-dichloro substitution provided the most
potent compound in the 3-isopropylsulfanyl series (see compound
12i), but such a disubstitution has a lower impact in the 3-isopropoxy
(10i) and 3-isopropylamino (17h) series. Compared to the 7-halo-
substituted compounds, the presence of a methyl or a methoxy group
at the 7-position did not improve the activity on pancreatic β-cells.
Determination of the Acidic Character. In order to explain
the differences in activity as a result of the replacement of the
amino group (NH) at the 3-position by a methylene group
(CH2), an oxygen atom (O), or a sulfur atom (S), we expected
that the acidic character of the molecules, linked to the presence
of a labile proton at the 4-position, varied in accordance with the
nature of the ‘bridge’ at this 3-position.
a Reagents: (i) (CH3)2CHCH2COCl, pyridine, dioxane; (ii) NaOH 1%
in H2O, Δ.
Scheme 3a
Thus, we have determined the pKa value of a representative of
each series of compounds, namely the 7-chloro-substituted
compounds 4, 10c, 12c, and 15c (Table 2) and observed that
the acidic character decreased as follows (from the most to the
less acidic compound): S (12c: 7.01) > O (10c: 8.00) > CH2
(15c: 8.52) > NH (4: 9.51).12 By comparison, the pKa value of
diazoxide has been reported to be 8.62,12 close to the pKa value of
its 3-isobutyl-substituted analogue 15c.
In accordance with such pKa values, we may predict that the
3-isopropylsulfanyl-substituted compounds 12 are essentially
ionized at the physiological pH of 7.4. Considering that the
nonionic form of the molecule could be the active form, it is
a Reagent: (i) Na2CO3, NaOH, Br2.
oxidizing the sulfur atom of 12c to give the corresponding sulfoxide
16, no gain of activity was observed.
As a first conclusion, and when comparing the activity of the
7-chloro-substituted compound in each series of benzothiadia-
zine dioxides (compounds 4, 10c, 12c, 15c, and 16), we can
deduce the following rank order of potency: 3-NHCH(CH3)2
(4) > 3-CH2CH(CH3)2 (15c) > 3-OCH(CH3)2 (10c) >
3-SCH(CH3)2 (12c) g 3-S(dO)CH(CH3)2 (16).
Concerning the nature of the substituent introduced on the ben-
zene ring, we observe, for the 3-isobutyl-substituted benzothiadiazine
3190
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
Table 1. Effects of Diversely Substituted 4H-1,2,4-Benzothiadiazine 1,1-Dioxides on Insulin Secretion from Rat Pancreatic
Islets
no.
10a
10b
10c
10d
10e
10f
10i
12a
12b
12c
12d
12e
12g
12h
12i
15a
15b
15c
X6
X7
X
RISa (50 μM)
RISa (10 μM)
RISa (1 μM)
RISa (0.1 μM)
H
H
H
H
H
H
Cl
H
H
H
H
H
Cl
Br
Cl
H
H
H
H
H
H
H
H
H
H
H
Cl
Cl
H
F
O
O
O
O
O
O
O
ꢀ
90.4 ( 3.1 (23)
87.5 ( 4.3 (23)
77.3 ( 4.6 (14)
62.5 ( 3.2 (16)
79.5 ( 2.8 (15)
80.8 ( 4.4 (15)
73.7 ( 4.3 (15)
93.9 ( 4.2 (15)
88.7 ( 5.4 (21)
104.0 ( 5.8 (23)
90.3 ( 3.8 (24)
77.0 ( 4.7 (16)
80.2 ( 4.4 (23)
87.0 ( 5.1 (23)
70.0 ( 3.7 (16)
85.2 ( 4.1 (16)
75.3 ( 5.6 (15)
43.2 ( 2.2 (15)
61.8 ( 3.7 (15)
92.3 ( 3.6 (15)
34.9 ( 2.0 (16)
3.7 ( 0.6 (13)
4.8 ( 0.4 (32)
8.1 ( 0.8 (12)
8.5 ( 0.7 (14)
8.5 ( 0.9 (24)
7.5 ( 0.8 (15)
6.3 ( 0.7 (12)
73.9 ( 4.4 (16)
ꢀ
ꢀ
44.1 ( 3.3 (15)
17.5 ( 2.0 (11)
ꢀ
ꢀ
ꢀ
Cl
Br
CH3
OCH3
Cl
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
H
F
S
S
S
S
S
S
S
S
ꢀ
ꢀ
ꢀ
55.1 ( 1.8 (15)
51.3 ( 3.9 (19)
47.3 ( 2.6 (14)
ꢀ
ꢀ
ꢀ
Cl
Br
CH3
H
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
44.0 ( 3.5 (16)
33.8 ( 2.5 (14)
12.9 ( 1.1 (13)
ꢀ
ꢀ
ꢀ
H
ꢀ
ꢀ
Cl
ꢀ
ꢀ
H
F
CH2
CH2
CH2
CH2
SdO
NH
NH
NH
NH
NH
NH
NH
NH
ꢀ
ꢀ
ꢀ
ꢀ
ꢀ
Cl
Br
Cl
21.0 ( 1.1 (22)
ꢀ
85.8 ( 4.6 (16)
ꢀ
ꢀ
15d
16
ꢀ
89.8 ( 4.2 (21)
ꢀ
ꢀ
ꢀ
17ab
17bb
4b
H
F
73.7 ( 5.2 (15)
47.3 ( 3.7 (23)
36.2 ( 2.4 (31)
34.9 ( 2.8 (12)
71.3 ( 3.5 (15)
67.6 ( 4.3 (20)
13.2 ( 1.2 (16)
13.2 ( 1.0 (26)
87.5 ( 5.0 (15)
ꢀ
3.3 ( 0.7 (13)
5.7 ( 0.5 (35)
6.7 ( 1.7 (12)
3.5 ( 0.3 (13)
4.4 ( 0.7 (12)
6.4 ( 0.7 (16)
5.0 ( 0.4 (12)
26.7 ( 1.6 (16)
96.9 ( 4.9 (16)
90.4 ( 3.5 (23)
91.0 ( 5.3 (13)
ꢀ
Cl
Br
CH3
OCH3
H
17db
17eb
17fb
17gb
17hb
diazoxideb
ꢀ
76.7 ( 4.3 (15)
84.9 ( 4.5 (21)
ꢀ
Cl
a RIS: percentage of residual insulin release from rat pancreatic islets incubated in the presence of 16.7 mM glucose (mean ( SEM (n)). b Published
results (refs 12ꢀ15).
tempting to speculate that the marked ionization of the 3-alkyl-
sulfanyl-substituted derivatives could justify their poor activity on
the pancreatic tissue. Moreover, considering that the binding
sites of KATP channel modulators (activators and blockers) have
been reported to be located at the cytoplasmic face of the cell
membranes,18,19 the ionization state of the molecules could have
a critical impact on the capacity of the drugs to permeate the
biological membranes and to reach the binding sites.
Previous works with two other series of potent SUR1-selective
KATP channel activators, i.e., 3-alkylamino-4H-pyrido[4,3-e]-
1,2,4-thiadiazine 1,1-dioxides (pKa values between 8.1 and
8.216) and 3-alkylamino-4H-thieno[3,2-e]-1,2,4-thiadiazine 1,1-
dioxides (pKa values between 8.2 and 8.520), reinforce the
assumption that a pKa value higher than 8 warrants a sufficient
fraction of nonionized molecules at physiological pH that can
cross the cell membranes and exert their biological activity.
Molecular Modeling. The modification of the nature of the
‘bridge’ introduced at the 3-position might also have an impact
on the conformational space of the molecules and on the
freedom of rotation of the bond linking the substituent at the
3-position, thus influencing the spatial orientation of the alkyl
(isopropyl) chain.
potent series of compounds, namely the 3-alkylamino-4H-
1,2,4-benzothiadiazine 1,1-dioxides, we hypothesized that the
isopropyl chain had to be firmly blocked in the spatial orienta-
tion, as found with compound 4. For such a compound as well as
for the other 3-alkylamino-4H-1,2,4-benzothiadiazine 1,1-diox-
ides, crystal structures indicated that the ‘guanidine’ moiety was
always found to be in the plane (assuming optimal electron
delocalization between the three nitrogen atoms) with the two
NH groups systematically oriented in parallel.21,22
In addition, the crystallographic data obtained with 4,22 10c,23
and 12c24 indicated that the three compounds adopted the 4H-
tautomeric form in the solid state. Indeed, the bond length
between C(3) and N(4) was always greater than the bond length
between N(2) and C(3) [C(3)ꢀN(4) for 4, 10c, and 12c =
1.352 Å, 1.339 Å, and 1.360 Å, respectively; N(2)ꢀC(3) for 4,
10c, and 12c = 1.330 Å, 1.289 Å, and 1.310 Å, respectively],22ꢀ24
assuming that the double bond in the thiadiazine ring is located at
N(2)ꢀC(3) rather than at C(3)ꢀN(4). Therefore, the hydro-
gen atom must be found at the N(4) position (4H-tautomerism).
We have used these crystallographic data as the starting low
energy conformation (for 15c, no X-ray data were available
because no convenient monocrystal was obtained), and we
further explored the conformational space resulting from the
rotation of the bond linking the substituent at the 3-position
Considering that the ‘active’ conformation of 3-substituted
1,2,4-benzothiadiazine 1,1-dioxides is provided by the most
3191
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
Table 2. Effects of Selected 4H-1,2,4-Benzothiadiazine 1,1-Dioxides, Diazoxide, and Pinacidil on Insulin Release, on
Contractile Activity of Kþ-Depolarized Rat Aorta Rings, and on Oxytocin-Induced Contraction of Rat Uterus (bolus
of oxytocin)
rat uterus % residual contractionc
no.
pancreatic islets IC50 (μM)a
rat aorta rings EC50 (μM)b
100 μM
50 μM
10 μM
pKa value
4
0.55 ( 0.10 (3)d
19.1e
36.3 ( 2.2 (6)d
22.7 ( 1.9 (7)
23.0 ( 3.6 (6)
5.8 ( 1.5 (4)
75.6 ( 5.1 (4)
68.7 ( 10.6 (4)
2.3 ( 1.9 (4)
55.0 ( 4.3 (4)
67.7 ( 4.0 (4)
38.1 ( 2.2 (6)
95.0 ( 1.7 (4)
85.8 ( 6.2 (4)
54.7 ( 14.8 (4)
64.7 ( 2.8 (4)
76.3 ( 4.9 (4)
42.7 ( 2.4 (6)
94.4 ( 4.9 (4)
88.0 ( 6.0 (4)
98.8 ( 11.4 (4)
73.9 ( 6.3 (4)
93.8 ( 2.2 (4)
44.9 ( 5.1 (6)
9.51d
8.00
7.01
8.52
8.62d
ꢀ
10c
12c
∼50e
15c
5.8e
diazoxide
pinacidil
18.4 ( 2.2 (5)d
202.3 ( 28.3 (3)d
22.4 ( 2.1 (11)d
0.62 ( 0.17 (15)d
a IC50: drug concentration giving 50% inhibition of insulin release (mean ( SEM (n)). b EC50: drug concentration giving 50% relaxation of the 30 mM
KCl-induced contraction of rat aortic rings (mean ( SEM (n)). c % residual contraction: percentage of residual contraction of rat uterus induced by
20 mU oxytocin injected as a bolus in the superfusion system (mean ( SEM (n)). d Published results (refs 12 and 17). e Estimated IC50 value.
modeling study and by previously published X-ray data for
diversely substituted 3-alkylamino-4H-1,2,4-benzothiadiazine
1,1-dioxides.21,22
The oxygen atom and the sulfur atom also possess electronic
lone pairs and the possibility for delocalization of the lone pairs in
the ‘amidinic’ (NdCꢀNH) moiety of the thiadiazine ring could
also favor the planar conformation. In fact, the lowest energy
conformer in each series has been found to be the one with the
torsion angle T1 near 0ꢀ (see Figure 5). Moreover, in both series,
and probably for the same steric reasons, the preferred orienta-
tion of the isopropyl chain was similar to that found with the
3-isopropylamino-substituted compound.
Our molecular modeling approach further revealed that the
3-alkoxy- and 3-alkylsulfanyl-substituted 1,2,4-benzothiadiazine
1,1-dioxides preferentially adopt a conformation similar to that
Figure 2. Chemical structure of 3-isopropylamino- (4), 3-isopropoxy-
observed with the most potent (biologically efficient on the
insulin-releasing process) 3-alkylamino-substituted compounds.
Such a feature, however, was not sufficient to ensure a marked
biological activity. These data rather confirm the critical role of
the NH group at the 3-position of 3-alkylamino-4H-1,2,4-ben-
zothiadiazine 1,1-dioxides for the establishment of a strong
hydrogen bond with the receptor binding site responsible for
optimal activity on pancreatic β-cells.
Pharmacological Evaluation on Smooth Muscles. Further
pharmacological investigations were conducted with the 7-chloro-
substituted benzothiadiazine dioxides 10c, 12c, and 15c on two
types of smooth muscle cells. The three compounds were compared
to the previously studied compound 4, to diazoxide, and to pinacidil
for their myorelaxant activity on rat aortic rings precontracted with
30 mM KCl (Table 2). The following EC50 values (drug concen-
tration giving 50% relaxation of the 30 mM KCl-induced contrac-
tion of rat aortic rings) were obtained: 4 (NH): EC50 = 36.3 μM;12
10c (O): EC50 = 22.7 μM; 12c (S): EC50 = 23.0 μM; 15c (CH2):
EC50 = 5.8 μM.
(10c), 3-isopropylsulfanyl- (12c), and 3-isobutyl- (15c) substituted
7-chloro-4H-1,2,4-benzothiadiazine 1,1-dioxide with indication of the
torsion angles T1 (atoms 1ꢀ2ꢀ3ꢀ4) and T2 (atoms 2ꢀ3ꢀ4ꢀ5).
(torsion angle T1 defined by atoms 1ꢀ2ꢀ3ꢀ4 in Figure 2) and
also resulting from the rotation of the bond linking the ‘isopropyl’
moiety to the ‘bridge’ at the 3-position (torsion angle T2 defined
by atoms 2ꢀ3ꢀ4ꢀ5 in Figure 2). Particular attention was paid to
the preferred orientation adopted by the isopropyl chain, in
each case.
As shown in Figure 3, the greatest freedom of rotation of the
bond linking the substituent at the 3-position (defined by the T1
angle) was observed with the compound possessing a ‘methy-
lene’ (CH2) bridge. The following rank order of conformational
freedom is given as follows: CH2 > S > O > NH.
Regarding the 3-alkylamino-substituted derivatives, free rota-
tion was excluded because the optimal delocalization of the lone
pairs of the nitrogen atoms in the guanidine moiety was only
achieved when the system was coplanar. Two situations may
warrant such optimal delocalization (see conformations A and B,
Figure 4), but for steric reasons, the first conformation A was
preferred. Such a feature was confirmed by the present molecular
As a result, the compound exhibiting the most potent vascular
myorelaxant activity was the 3-alkyl-substituted benzothiadiazine
dioxide 15c, being also poorly tissue selective. Indeed, its EC50
value on the vascular smooth muscle cells (EC50 = 5.8 μM) was
3192
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
Figure 3. 2D conformational scan (T1, T2) of compounds 4 (left upper panel), 10c (right upper panel), 12c (left lower panel), and 15c (right lower
panel). The pictured conformations are within 10 kcal/mol of energy.
although none of the tested compounds exhibited a strong
inhibitory effect (Table 2). The rank order of potency on the
uterine smooth muscle was 3-alkylsulfanyl (S) > 3-alkyl (CH2) >
3-alkoxy (O) > 3-alkylamino (NH). On the rat uterine tissue, the
reference compound diazoxide was found to be less potent than
the other 3-alkyl-substituted benzothiadiazine dioxide 15c.
Among the compounds tested on the different tissues (aortic,
uterine, pancreatic), the 3-alkylamino-substituted compound 4
Figure 4. Two possible conformations adopted by 3-isopropylamino-
remained the most selective toward the endocrine pancreas
4H-1,2,4-benzothiadiazine 1,1-dioxides, assuming optimal delocaliza-
versus smooth muscle tissues.
tion of the electron lone pairs of the nitrogen atoms in the guanidine
Because the reference SUR2B-type PCO pinacidil was unable
to completely suppress the uterine contractions induced by a
bolus of oxytocin (about 40% residual contraction in the
presence of 100 μM pinacidil), we decided to explore other
experimental conditions in order to have access to EC50 values.
When oxytocin was continuously superfused (50 mU/L) on the
muscle preparation, a regular set of reproducible small contrac-
tions was observed. Under such experimental conditions, pina-
cidil completely suppressed the oxytocin-induced contractions,
and concentrationꢀresponse curves led us to calculate EC50
values.
The EC50 value of pinacidil (EC50 = 1.3 μM) was found to be
shifted to higher values in the continuous presence of 1 μM
(EC50 = 8.1 μM) or 10 μM (EC50 = 28.3 μM) of the KATP
channel blocker glibenclamide26 in the physiological medium
(Table 3). This finding confirms the involvement of KATP
channels in the myorelaxant effect of pinacidil on rat uterus.
According to the initial results on rat uterus reported in
Table 2, 3-alkylsulfanyl-substituted compounds can be expected
to represent a promising series of myorelaxant drugs. As a result,
we decided to examine, in the new experimental test conditions,
compound 12c and three other of its analogues bearing a
different hydrocarbon chain linked to the sulfur atom at the
3-position (see compounds 12jꢀl; Scheme 2). Table 3 reports
the results obtained with these compounds as well as with
diazoxide and pinacidil on rat uterus and rat aorta rings.
moiety.
found to be similar to the estimated IC50 value on pancreatic β-
cells (IC50 = 5.8 μM) (Table 2). Such a feature was previously
noticed with the reference compound diazoxide, another 3-alkyl-
substituted 1,2,4-benzothiadiazine 1,1-dioxide (EC50 = 22.4 μM;
IC50 = 18.4 μM; Table 2) .16 Thus, for this class of compounds, it
was observed that the bulky isobutyl chain induced a more marked
effect on vascular smooth muscles and on pancreatic β-cells than the
methyl chain (e.g.: diazoxide). However, and according to recent
published data obtained with a series of 7-chloro-3-cycloalkyl-4H-
1,2,4-benzothiadiazine 1,1-dioxides,25 it was expected that a sub-
stituent more bulky than an isobutyl chain (i.e., a cyclohexyl chain)
at the 3-position should induce a loss of activity on SUR1-type KATP
channels (pancreatic β-cell type channels), while keeping a high
level of activity on the SUR2B-type KATP channels (smooth muscle
cell type channels).25
The less vasorelaxant compound among the selected ben-
zothiadiazines (see Table 2) was the 3-alkylamino-substituted
benzothiadiazine dioxide 4, being also the most selective toward
the pancreatic tissue.
When the data obtained for rat uterus were examined, another
smooth muscle tissue expressing KATP channels, it was found that
the 3-alkylsulfanyl-substituted compound 12c was the most
powerful myorelaxant (inhibition of the contraction induced
by 20 mU oxytocin injected as a bolus in the superfusion system)
3193
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
Figure 5. The most stable conformation of 4 (green), 10c (blue), 12c (pink), and 15c (orange).
Table 3. Effects of Selected 3-Alkylsulfanyl-4H-1,2,4-benzothiadiazine 1,1-Dioxides, Diazoxide, and Pinacidil on the Contractile
Activity of Kþ-Depolarized Rat Aorta Rings and on Oxytocin-Induced Contraction of Rat Uterus (continuous superfusion of
oxytocin)
rat uterus EC50 (μM)b
no.
12c
R
rat aorta rings EC50 (μM)a
no glibenclamide
þ 1 μM glibenclamide
þ 10 μM glibenclamide
CH(CH3)2
CH2CH(CH3)2
CH2C6H5
CH(CH3)C6H5
ꢀ
23.0 ( 3.6 (6)
6.2 ( 0.9 (4)
>30
ndd
nd
12j
13.5 ( 1.8 (4)
11.7 ( 0.5 (4)
4.4 ( 0.3 (12)
>30
nd
nd
12k
5.4 ( 0.1 (4)
nd
nd
12l
5.4 ( 0.1 (4)
6.8 ( 0.9 (4)
nd
24.2 ( 2.4 (12)
nd
diazoxide
22.4 ( 2.1 (11)c
0.62 ( 0.17 (15)c
pinacidil
ꢀ
1.3 ( 0.3 (18)
8.1 ( 0.8 (16)
28.3 ( 3.3 (16)
a EC50: drug concentration giving 50% relaxation of the 30 mM KCl-induced contraction of rat aortic rings (mean ( SEM (n)). b EC50: drug
concentration giving 50% relaxation of the oxytocin-induced contraction of rat uterus continuously superfused with 50 mU/L oxytocin in the absence or
presence of glibenclamide (mean ( SEM (n)). c Published results (refs 12 and 17). d Not determined.
Compound 12c exhibited an EC50 above 30 μM. The three
analogues 12j, 12k, and 12l were found to be more potent than 12c
with EC50 values of 13.5 μM, 11.7 μM, and 4.4 μM, respectively.
The marked myorelaxant effects of 12j, 12k, and 12l was confirmed
interest of 3-alkylsulfanyl-substituted 4H-1,2,4-benzothiadiazine
1,1-dioxides as a new class of tocolytic agents.
Radioisotopic and Fluorimetric Experiments. Additional in
vitro experiments have been performed to specify the mechanism
of action of the new benzothiadiazine dioxides. In the first series
of experiments, we characterized the effects of selected com-
pounds on 86Rb (42K substitute), 45Ca outflow, and insulin release
from perifused rat pancreatic islets.
The addition of the chloro-substituted 3-isopropoxy-4H-1,2,4-
benzothiadiazine 1,1 dioxide 10c to prelabeled pancreatic islets
exposed throughout to 5.6 mM glucose and extracellular Ca2þ
provoked a concentration-dependent increase in the rate of 86Rb
outflow (Figure 6). Thus, the magnitude of the increase in 86Rb
outflow observed during exposure to 10c averaged 0.23 (
0.05%/min after the addition of 10 μM and 1.32 ( 0.30%/
min after the addition of 50 μM 10c, respectively (P < 0.05). The
withdrawal of the drug from the perifusate was followed by a
decrease in 86Rb outflow (Figure 6).
on vascular smooth muscle cells (EC50 = 6.2 μM for 12j; EC50
=
5.4 μM for 12k and 12l). Moreover, the EC50 value of the most
potent compound 12l on rat uterus was shifted to higher values
when the experiment was repeated in the continuous presence of
glibenclamide in the physiological medium (Table 3), indicating
that the myorelaxant effect of the drug was, at least in part, mediated
by the activation of KATP channels. However, due to the assumption
that access and binding to KATP channels should be mediated by
nonionized drugs and according to the ionization state of 3-alkyl-
sulfanyl-substituted benzothiadiazine dioxides at physiological pH,
the unexpected marked activity of these compounds on rat uterus
probably involve additional myorelaxant mechanisms. Further
chemical and pharmacological developments are required to eluci-
date the whole mechanism of action and to confirm the possible
3194
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
Figure 7. Effects of 10c (50 μM) on 45Ca outflow (upper panel) and
insulin output (lower panel) from prelabeled rat pancreatic islets
perifused throughout in the presence of 16.7 mM glucose. Basal media
contained extracellular Ca2þ (b) or were deprived of Ca2þ (O). Mean
values refer to four to six individual experiments.
Figure 6. Effects of 10c (10 μM, 2; 50 μM, b) and 10c (50 μM) in the
continuous presence of glibenclamide (10 μM, O) on 86Rb outflow from
prelabeled and perifused rat pancreatic islets. Basal media contained
5.6 mM glucose and extracellular Ca2þ. Mean values ((SEM) refer to
six individual experiments.
insulinotropic glucose concentration (data not shown). Third, the
simultaneous measurement of insulin release from pancreatic islets
perifused throughout in the presence of 16.7 mM glucose and
extracellular Ca2þ further revealed an inhibitory effect, displaying a
time course identical to that of the 45Ca outflow responses, of
compounds 10c (Figure 7, lower panel) and 17f (data not shown)
on the insulin secretory rate.
In islets exposed throughout to 2.8 mM glucose and extra-
cellular Ca2þ, the 45Ca response to a sudden rise in the extra-
cellular concentration of Kþ (5 to 50 mM) was unaffected by the
presence of 10c (50 μM) in the basal medium (data not shown).
Thus, the magnitude of the increase in 45Ca outflow evoked by
50 mM Kþ averaged 1.24 ( 0.05%/min in the absence and
1.12 ( 0.07%/min in the presence of 50 μM 10c (P > 0.05). Such
a finding further indicates that 10c fails to interact directly at the
level of the voltage-sensitive Ca2þ channels.26
Altogether, these experimental data indicate that the 3-alkoxy-
substituted benzothiadiazine 10c, as well as the 3-alkylamino-
substituted benzothiadiazine 17f, activates the plasma membrane
KATP channels and ultimately inhibits the insulin releasing
process through a reduction in Ca2þ entry. Incidentally, the
ionic and secretory responses to compounds 10c (see Figures 6
and 7) and 17f (data not shown) were always rapidly reversible,
implying a lack of damaging effect of such compounds to the
insulin-secreting cells.
In the last series of radioisotopic experiments, we determined
the effects of the 3-alkyl-substituted benzothiadiazine 15c, an
original compound exhibiting a marked myorelaxant activity, on
the rate of 86Rb outflow from prelabeled and perifused rat aortic
rings. Figure 8 clearly shows that 15c (100 μM) induced a fast,
sustained, and reversible increase in 86Rb outflow. The presence
of the KATP channel blocker glibenclamide26 in the perifusing
medium strongly reduced the stimulatory effect of 15c. Such
observations further suggest that the vasorelaxant effect of 15c
When the physiological medium was enriched with the
hypoglycemic sulfonylurea glibenclamide, a KATP channel blocker,26
the enhancing effect of 10 μM (data not shown) or 50 μM 10c was
completely suppressed (Figure 6). Experiments performed with
compound 17f indicated that, under the same experimental condi-
tions, the drug also elicited a reversible and glibenclamide-sensitive
increase in 86Rb outflow from prelabeled and perifused rat pancreatic
islets (data not shown). These findings indirectly indicate that
3-alkoxy-substituted benzothiadiazine 1,1-dioxides (such as 10c)
and 3-alkylamino-substituted compounds (such as 17f) provoke an
increase in membrane Kþ permeability and further suggest that the
Kþ permeability changes are mediated by the activation of ATP-
sensitive Kþ channels.14,26
An increase in KATP channel activity might be expected to shift
the membrane potential below the threshold required for the
opening of voltage-sensitive Ca2þ channels, thereby reducing the
Ca2þ entry, the cytosolic Ca2þ concentration, and the secretory
process. Such a cascade of events is corroborated by additional
radioisotopic and fluorimetric experiments.
First, compound 10c (Figure 7, upper panel), as well as com-
pound 17f (data not shown), reduced 45Ca outflow from prelabeled
rat pancreatic islets exposed throughout to 16.7 mM glucose and
extracellular Ca2þ. Under such experimental conditions, namely in
the presence of an insulinotropic glucose concentration and extra-
cellular Ca2þ in the physiological medium, a decrease in the 45Ca
outflow rate reflects a reduction in 40Ca2þ entry into the islets
cells.14,26 The lack of effect of compound 10c (Figure 7, upper panel)
and 17f (data not shown) on 45Ca outflow from pancreatic islets
exposed to Ca2þ-free media corroborates this interpretation. Second,
calcium fluorimetry experiments conducted on isolated pancreatic
islet cells clearly revealed the capacity of compound 10c to counter-
act the rise in cytosolic Ca2þ concentration mediated by an
3195
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
biological activity, at least in part, through the activation of KATP
channels.
Further radioisotopic and fluorimetric experiments conducted
with 7-chloro-3-isopropoxy-4H-1,2,4-benzothiadiazine 1,1-diox-
ide 10c demonstrated that such a compound bearing a short
branched O-alkyl group instead of an NH-alkyl or an alkyl group
at the 3-position also behaved as a specific opener of the ATP-
sensitive potassium channels. This study is the first report on the
identification of a 3-alkoxy-4H-1,2,4-benzothiadiazine 1,1-diox-
ide as a KATP channel opener.
’ EXPERIMENTAL SECTION
Chemistry. Melting points were determined on a Stuart SMP3
capillary apparatus and are uncorrected. IR spectra were recorded as KBr
pellets on a Perkin-Elmer 1000 FTIR spectrophotometer. The 1H NMR
spectra were recorded on a Bruker Avance (500 MHz) instrument using
DMSO-d6 as the solvent with TMS as an internal standard; chemical
shifts are reported in δ values (ppm) relative to that of internal TMS.
The abbreviations s = singlet, d = doublet, t = triplet, q = quadruplet, m =
multiplet, and b = broad are used throughout. Elemental analyses (C, H,
N, S) were realized on a Thermo Scientific FlashEA 1112-elemental
analyzer and were within (0.4% of the theoretical values. This analytical
method certified a purity g95% for each tested compound. All reactions
were routinely checked by TLC on silica gel Merck 60 F254. The
synthesis of compounds 10a,b, 10dꢀf, 10i, 12a,b, 12d,e, 12gꢀk, 14a,
b, 14d, 15a,b, and 15d is detailed in the Supporting Information. The
synthesis of 10a and 12a has been reported previously.28,29
Figure 8. Effect of 15c (100 μM) on 86Rb outflow from prelabeled rat
aortic rings perifused throughout in the absence (b) or presence (O) of
glibenclamide (10 μM). Mean values ((SEM) refer to four to five
individual experiments.
7-Chloro-3-isopropoxy-4H-1,2,4-benzothiadiazine 1,1-dioxide Mono-
hydrate (10c). 7-Chloro-3-oxo-3,4-dihydro-2H-1,2,4-benzothiadiazine
1,1-dioxide (8c)12 (4.3 mmol) was solubilized in methanol (10 mL) by
adding, under stirring, 1 equiv of NaOH (4.3 mmol). The resulting
solution was evaporated to dryness under reduced pressure, and the
residue was dissolved in DMF (20 mL). Isopropyl iodide (6.45 mmol)
was added to the mixture, and the solution was heated at 70ꢀ80 ꢀC
during 4 to 6 h. The reaction gave rise to the formation of two major
compounds: 7-chloro-2-isopropyl-3-oxo-3,4-dihydro-2H-1,2,4-ben-
zothiadiazine 1,1-dioxide (9c) and 7-chloro-3-isopropoxy-4H-1,2,4-
benzothiadiazine 1,1-dioxide (10c). The solvent was eliminated by dis-
tillation under reduced pressure, and the residue was suspended in water
(30 mL). The mixture was alkalinized under stirring by adding a 5% m/v
aqueous solution of NaOH until pH 14. The insoluble material was
eliminated by filtration, and the filtrate was treated with charcoal. After
filtration, the filtrate was acidified to pH 10 by means of 12 N HCl. The
precipitate was eliminated by filtration, and the filtrate was supplemented
with 12 N HCl until pH 1. The resulting precipitate of the title compound
was collected by filtration, washed with water, and dried (yields: 40%); mp:
227ꢀ229 ꢀC; IR (KBr): 3582, 3522, 2986, 1613, 1580, 1523, 1482, 1329,
1312, 1288, 1255, 1172 cmꢀ1; 1H NMR (DMSO-d6): δ 1.35 (d, 6H, 2 ꢁ
CH3), 5.15 (m, 1H, CH), 7.29 (d, 1H, 5-H), 7.70 (d, 1H, 6-H), 7.79 (s, 1H,
8-H), 12.22 (bs, 1H, N-H). Anal. (C10H11ClN2O3S.H2O) C, H, N, S.
General Synthetic Pathway to 7-Substituted 3-Isopropyl-
sulfanyl-4H-1,2,4-benzothiadiazine 1,1-Dioxides (12). The
appropriate 6/7-substituted 3-thioxo-3,4-dihydro-2H-1,2,4-benzothia-
diazine 1,1-dioxide (11)12,13,15 (2.0 mmol) was dissolved in methanol
(10 mL) supplemented with NaOH (2.0 mmol). The solvent was
evaporated under reduced pressure, and the residue was solubilized in
nitroethaneꢀDMF 5:1. Isopropyl iodide (2.2 mmol) was added to the
solution, and the mixture was heated for 3 h at 80 ꢀC. Then, the solvents
were removed by distillation under reduced pressure, and the residue
was dissolved in a 10% w/v aqueous solution of NaOH. The resulting
solution was treated with charcoal and filtered, and the filtrate was
treated with 12 N HCl until pH 1. The resulting precipitate was collected
by filtration, washed with water, and dried; yields: 40ꢀ60%.
results from the activation of KATP channels, as already reported
for the reference compound diazoxide, another 3-alkyl-substi-
tuted benzothiadiazine dioxide.27
’ CONCLUSION
Diversely substituted 3-isopropoxy-, 3-isopropylsulfanyl-,
3-isopropylsulfinyl-, and 3-isobutyl-4H-1,2,4-benzothiadiazine 1,1-
dioxides were synthesized and their activity on pancreatic β-cells
(inhibition of insulin release) and vascular and uterine smooth
muscle tissues (myorelaxant effects) was compared to that of
previously reported KATP channel openers belonging to the 3-iso-
propylamino-4H-1,2,4-benzothiadiazine 1,1-dioxide series. The im-
pact on biological activity of the isosteric replacement of the NH
group of 3-alkylamino-substituted benzothiadiazines by a O, S,
S(dO), or CH2 group was examined. The following rank order of
potency on insulin-secreting cells was observed: 3-isopropylamino-
(NH) > 3-isobutyl- (CH2) > 3-isopropoxy- (O) > 3-isopropylsulfa-
nyl- (S) > 3-isopropylsulfinyl- (S(dO)) substituted 4H-1,2,4-ben-
zothiadiazine 1,1-dioxides.
The present study also confirmed the critical role of the NH
group at the 3-position for the establishment of a strong
hydrogen bond responsible for optimal activity expressed by
3-alkylamino-4H-1,2,4-benzothiadiazine 1,1-dioxides on the in-
sulin-secreting cells. Compared to the three other series of drugs,
the latter series of compounds also expressed the highest
selectivity for the pancreatic tissue versus smooth muscle tissues
(aorta and uterus).
Interestingly, 3-(alkyl/aralkyl)sulfanyl-substituted 7-chloro-
4H-1,2,4-benzothiadiazine 1,1-dioxides were identified as potent
myorelaxant drugs acting on uterine smooth muscles. The most
potent compound, R/S-7-chloro-3-(1-phenylethyl)sulfanyl-4H-
1,2,4-benzothiadiazine 1,1-dioxide (12l), was found to exert its
3196
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
7-Chloro-3-isopropylsulfanyl-4H-1,2,4-benzothiadiazine 1,1-Diox-
ide (12c). The title compound was obtained according to the general
synthetic pathway starting from 7-chloro-3-thioxo-3,4-dihydro-2H-
1,2,4-benzothiadiazine 1,1-dioxide (11c); mp: 220ꢀ226 ꢀC; IR
(KBr): 3228, 3179, 3077, 1604, 1553, 1506, 1475, 1305, 1194, 1157,
1135, 1109 cmꢀ1; 1H NMR (DMSO-d6): δ 1.41 (d, 6H, 2 ꢁ CH3), 3.85
(m, 1H, CH), 7.30 (d, 1H, 5-H), 7.73 (d, 1H, 6-H), 7.84 (s, 1H, 8-H),
12.58 (bs, 1H, N-H). Anal. (C10H11ClN2O2S2) C, H, N, S.
Measurements of Insulin Release from Incubated Rat
Pancreatic Islets. The method used to measure insulin release from
incubated rat pancreatic islets was previously described.14,17,26,30
Measurement of the Contractile Activity in Rat Aorta. The
method used to measure the myorelaxant effect of the drugs on 30 mM
KCl-precontracted rat aortic rings was previously described.14,17,26,30
Measurement of the Myorelaxant Activity on Rat Uterus.
First Model: Contractions Induced by Bolus of Oxytocin Injected in the
Superfusion System. Fed Wistar rats (150ꢀ200 g) were treated the day
before killing with diethylstilboestrol diproprionate [i.m. injection of
0.1 mL/100 g of a 1 mg/mL oily solution of diethylstilboestrol
diproprionate (Sigma)]. The rats were anaesthetized and then sacrificed.
The two uterine horns were removed, cleared of adhering fat and
connective tissue, and separated. Each horn was superfused with a
Tyrode solution (in mM: NaCl 137, KCl 2.7, CaCl2 1.8, MgCl2 1.1,
NaH2PO4 0.4, NaHCO3 11.9, glucose 5.6) bubbled continuously with a
mixture of O2 (95%) and CO2 (5%). The superfusate was maintained at
37 ꢀC. After a stabilization period of 30 min, injection of 20 mU oxytocin
(200 μL of a 0.1 U/mL solution of the hormone in 9 % NaCl) in the
superfusion channel was repeated at 10 min intervals until the recorded
contractions (AUC) were constant. The mean of the three last injections
gave the 100% of the contractile response to oxytocin. For each drug
concentration added in the medium (10, 50, and 100 μM), injection of
20 mU oxytocin was repeated at least three times. The contractile
responses recorded in the presence of different drug concentrations
added in the superfusate medium (mean of the three AUC) were
expressed as a percentage of the reference value (contractile response
to oxytocin in the absence of drug).
R/S-7-Chloro-3-(1-phenylethyl)sulfanyl-4H-1,2,4-benzothiadiazine
1,1-Dioxide (12l). The title compound was obtained according to the
general synthetic pathway starting from 7-chloro-3-thioxo-3,4-dihydro-
2H-1,2,4-benzothiadiazine 1,1-dioxide (11c) and 1-phenylethyl bro-
mide instead of isopropyl iodide; mp: 216ꢀ218 ꢀC; IR (KBr): 3250,
1
1599, 1547, 1508, 1479, 1298, 1159 cmꢀ1; H NMR (DMSO-d6): δ
1.75 (d, 3H, CH3), 5.00 (q, 1H, CH), 7.25ꢀ7.50 (m, 6H, C6H5 þ 5-H),
7.75 (dd, 1H, 6-H), 7.85 (d, 1H, 8-H), 12.55 (s, 1H, N-H). Anal.
(C15H13ClN2O2S2) C, H, N, S.
General Synthetic Pathway to 5-Substituted 2-(3-Methyl-
butyrylamino)benzenesulfonamides (14). The appropriate
5-substituted aminobenzenesulfonamide (13) (3.2 mmol) was dissolved
in dioxane (12 mL) and supplemented with pyridine (3.2 mmol) and
2-methylbutyryl chloride (3.8 mmol). The mixture was stirred at room
temperature for 1 h. The solvent was removed by distillation under
reduced pressure, and the residue was solubilized with a 5% w/v aqueous
solution of NaOH. The resulting solution was adjusted to pH 6ꢀ7 by
means of 1 N HCl, and the resulting precipitate was collected by
filtration, washed with water, and dried; yields = 70ꢀ80%.
5-Chloro-2-(3-methylbutyrylamino)benzenesulfonamide (14c). Start-
ing from 2-amino-5-chlorobenzenesulfonamide (13c); mp: 175ꢀ177 ꢀC.
Anal. (C11H15ClN2O3S) C, H, N, S.
Second Model: Contractions Induced by a Continuous Superfusion
of Oxytocin. After a stabilization period of 30 min, each horn was
superfused with Tyrode solution containing oxytocin at a low concen-
tration (50 mU/L). After a period of several minutes, the uterine
contractions were recorded during 15 min. The mean of three successive
contractions (AUC: area under the curve) provided 100% of the
contractile response to oxytocin. This sequence of events was repeated
with a superfusate solution containing oxytocin (50 mU/L) and the
tested drug at increasing concentrations. The contractile responses
recorded in the presence of different drug concentrations (mean of
three AUC) were expressed as a percentage of the reference value
(contractile response to oxytocin in the absence of drug). For several
drugs, tested at increasing concentrations, the experiment was repeated
in the continuous presence of 1 or 10 μM glibenclamide. Results were
expressed as the percentage of residual contraction, and an EC50 value
was calculated corresponding to the drug concentration giving 50%
residual contraction induced by oxytocin.
7-Chloro-3-isobutyl-4H-1,2,4-benzothiadiazine 1,1-Dioxide (15c).
The solution of 5-chloro-2-(3-methylbutyrylamino)benzenesulfon-
amide (14c) (2.2 mmol) in a 1% w/v aqueous solution of NaOH
(32 mL) was refluxed for 30 min. After cooling, the solution was adjusted
to pH 6ꢀ7 by means of 1 N HCl. The resulting precipitate was collected
by filtration, washed with water, and dried (yields: 85%); mp: 235ꢀ
238 ꢀC; IR (KBr): 3283, 3191, 3118, 2959, 1622, 1610, 1580, 1525,
1483, 1289, 1157, 1143, 1109 cmꢀ1; 1H NMR (DMSO-d6): δ 0.96 (d,
6H, 2 ꢁ CH3), 2.13 (m, 1H, CH), 2.41 (m, 2H, CH2), 7.37 (d, 1H, 5-H),
7.73 (d, 1H, 6-H), 7.84 (s, 1H, 8-H), 12.12 (bs, 1H, N-H). Anal.
(C11H13ClN2O2S) C, H, N, S.
R/S-7-Chloro-3-isopropylsulfinyl-4H-1,2,4-benzothiadiazine 1,1-Di-
oxide (16). The suspension of 7-chloro-3-isopropylsulfanyl-4H-1,2,4-
benzothiadiazine 1,1-dioxide (12c) (0.5 g, 1.72 mmol) in an aqueous
solution of sodium carbonate (0.22 g/25 mL) was supplemented under
stirring with 2 N NaOH until complete dissolution. The alkaline
solution was then supplemented, under stirring at room temperature,
with bromine (0.1 mL). After 10 min, the mixture was adjusted to pH 2
by means of 12 N HCl, and the resulting precipitate was collected by
filtration, washed with water, and suspended in methanol (15 mL) under
stirring during 1 h. The insoluble material was collected by filtration,
washed with methanol, and dried (0.40 g, 76%); mp: 257ꢀ260 ꢀC; IR
(KBr): 3146, 2977, 1606, 1594, 1570, 1507, 1476, 1326, 1172, 1060,
1027 cmꢀ1; 1H NMR (DMSO-d6): δ 1.19 (d, 3H, CH3-A), 1.38 (d, 3H,
CH3-B) 3.40 (m, 1H, CH), 7.83ꢀ7.95 (m, 3H, 5-H, 6-H, 8-H), 12.70
(bs, 1H, N-H). Anal. (C10H11ClN2O3S) C, H, N, S.
Measurements of 86Rb Outflow from Rat Perifused Pan-
creatic Islets and Rat Aortic Rings. The methods used for
measuring 86Rb (42K substitute) outflow from prelabeled and perifused
rat pancreatic islets or from prelabeled and perifused rat aortic rings were
previously described.17,26,30
Measurements of 45Ca Outflow and Insulin Release from
Perifused Rat Pancreatic Islets. The methods used for simulta-
neously measuring 45Ca outflow and insulin release from prelabeled and
perifused rat pancreatic islets were previously described.26,30
Measurements of Cytosolic Ca2þ Concentration from
Isolated Rat Pancreatic Islets Cells. The method used for measur-
ing the cytosolic Ca2þ concentration ([Ca2þ]i) from single islet cells was
previously described.26,30
Conformational Studies. Quantum mechanical calculations at
the HF/6-31G* level have been used to characterize the intrinsic conforma-
tional preferences of 10c, 12c, 4, and 15c in the gas phase. All calculations
have been performed with the Gaussian03 program. Starting from the
crystal structure of 10c,23 12c,24 and 4,22 a conformational scan was
performed by varying both so-called dihedral angles T1 and T2 from 0 to
360ꢀ. In the case of 15c, a QM-minimized structure was first calculated as a
starting point, because no crystal structure was available.
Ionization Constants. The pKa values of the compounds were
determined by means of UV spectrophotometry using a Perkin-Elmer
UV/vis 554 spectrophotometer at 25 ꢀC. UV spectra of compounds
were taken in different aqueous buffers of pH, ranking from 5 to 12. The
pKa values were calculated by the DebyeꢀH€uckel equation at the
wavelength giving the maximum absorbance of the ionized form.31
3197
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
’ ASSOCIATED CONTENT
Mitochondrial ATP-sensitive Kþ Channels. Possible Mechanism of
Cardioprotection. Circ. Res. 1997, 1072–1082.
(10) Moreau, C.; Prost, A. L.; Dꢀerand, R.; Vivaudou, M. SUR, ABC
Proteins Targeted by KATP Channel Openers. J. Mol. Cell. Cardiol. 2005,
38, 951–963.
(11) Abdallah, Y.; Wolf, C.; Meuter, K.; Piper, H. M.; Reusch, H. P.;
Ladilov, Y. Preconditioning with Diazoxide Prevents Reoxygenation-
induced Rigor-type Hypercontracture. J. Mol. Cell Cardiol. 2010,
48, 270–276.
S
Supporting Information. General synthetic pathways to
b
7-substituted 3-isopropoxy-4H-1,2,4-benzothiadiazine 1,1-diox-
ides (10) and 7-substituted 3-isobutyl-4H-1,2,4-benzothiadia-
zine 1,1-dioxides (15); synthesis of compounds 10a,b, 10dꢀf,
10i, 12a,b, 12d,e, 12gꢀk, 14a,b, 14d, 15a,b, and 15d; elemental
analysis results for the new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
(12) de Tullio, P.; Becker, B.; Boverie, S.; Dabrowski, M.; Wahl, P.;
Antoine, M. H.; Somers, F.; Sebille, S.; Ouedraogo, R.; Hansen, J. B.;
Lebrun, P.; Pirotte, B. Toward Tissue-selective Pancreatic B-cells KATP
Channel Openers Belonging to 3-Alkylamino-7-halo-4H-1,2,4-ben-
zothiadiazine 1,1-Dioxides. J. Med. Chem. 2003, 46, 3342–3353.
(13) Boverie, S.; Antoine, M. H.; Somers, F.; Becker, B.; Sebille, S.;
Ouedraogo, R.; Counerotte, S.; Pirotte, B.; Lebrun, P.; de Tullio, P.
Effect on KATP Channel Activation Properties and Tissue Selectivity of
the Nature of the Substituent in the 7- and the 3-Position of 4H-1,2,4-
Benzothiadiazine 1,1-Dioxides. J. Med. Chem. 2005, 48, 3492–3503.
(14) de Tullio, P.; Boverie, S.; Becker, B.; Antoine, M. H.; Nguyen,
Q. A.; Francotte, P.; Counerotte, S.; Sebille, S.; Pirotte, B.; Lebrun, P. 3-
Alkylamino-4H-1,2,4-benzothiadiazine 1,1-Dioxides as ATP-Sensitive
Potassium Channel Openers: Effect of 6,7-Disubstitution on Potency
and Tissue Selectivity. J. Med. Chem. 2005, 48, 4990–5000.
(15) Pirotte, B.; de Tullio, P.; Nguyen, Q. A.; Somers, F.; Fraikin, P.;
Florence, X.; Wahl, P.; Hansen, J. B.; Lebrun, P. Chloro-Substituted
3-Alkylamino-4H-1,2,4-benzothiadiazine 1,1-Dioxides as ATP-Sensitive
Potassium Channel Activators: Impact of the Position of the Chlorine
Atom on the Aromatic Ring on Activity and Tissue Selectivity. J. Med.
Chem. 2010, 53, 147–154.
’ AUTHOR INFORMATION
Corresponding Author
*Phone: 32-4-3664365. Fax: 32-4-3664362. E-mail: B.Pirotte@
ulg.ac.be.
Author Contributions
^These authors contributed equally to the work.
’ ACKNOWLEDGMENT
This study was supported by grants from the National Fund
for Scientific Research (F.N.R.S., Belgium) from which P. de
Tullio is a Research Associate and P. Lebrun is a Research
Director. The authors gratefully acknowledge the technical
assistance of S. Counerotte, Y. Abrassart, F. Leleux, A.-M.
Vanbellinghen, and A. Van Praet.
(16) de Tullio, P.; Pirotte, B.; Lebrun, P.; Fontaine, J.; Dupont, L.;
Antoine, M. H.; Ouedraogo, R.; Khelili, S.; Maggetto, C.; Masereel, B.;
Diouf, O.; Podona, T.; Delarge, J. 3-and 4-Substituted 4H-Pyrido[4,3-e]-
1,2,4-thiadiazine 1,1-Dioxides as Potassium Channel Openers: Synth-
esis, Pharmacological Evaluation, and StructureꢀActivity Relationships.
J. Med. Chem. 1996, 39, 937–948.
’ ABBREVIATIONS
KATP channel, ATP-sensitive potassium channel; Kir, inwardly
rectifying potassium channel; PCO, potassium channel opener; SUR,
sulfonylurea receptor
(17) Lebrun, P.; Becker, B.; Morel, N.; Ghisdal, P.; Antoine, M. H.;
de Tullio, P.; Pirotte, B. KATP Channel Openers: Tissue Selectivity of
Original 3-Alkylaminopyrido- and 3-Alkylaminobenzothiadiazine 1,1-
Dioxides. Biochem. Pharmacol. 2008, 75, 468–475.
(18) Schwanstecher, M.; Schwanstecher, C.; Dickel, C.; Chudziak,
F.; Moshiri, A.; Panten, U. Location of the Sulphonylurea Receptor at
the Cytoplasmic Face of the Beta-cell Membrane. Br. J. Pharmacol. 1994,
113, 903–911.
(19) Stephan, D.; Salamon, E.; Weber, H.; Russ, U.; Lemoine, H.;
Quast, U. KATP Channel Openers of the Benzopyran Type Reach their
Binding Site via the Cytosol. Br. J. Pharmacol. 2006, 149, 199–205.
(20) Nielsen, F. E.; Bodvarsdottir, T. B.; Worsaae, A.; MacKay, P.;
Stidsen, C. E.; Boonen, H. C.; Pridal, L.; Arkhammar, P. O.; Wahl, P.;
Ynddal, L.; Junager, F.; Dragsted, N.; Tagmose, T. M.; Mogensen, J. P.;
Koch, A.; Treppendahl, S. P.; Hansen, J. B. 6-Chloro-3-alkylamino-4H-
thieno[3,2-e]-1,2,4-thiadiazine 1,1-Dioxide Derivatives Potently and
Selectively Activate ATP Sensitive Potassium Channels of Pancreatic
β-Cells. J. Med. Chem. 2002, 45, 4171–4187.
’ REFERENCES
(1) Burke, M. A.; Mutharasan, R. K.; Ardehali, H. The Sulfonylurea
Receptor, An Atypical ATP-binding Cassette Protein, and Its Regulation
of the KATP Channel. Circ. Res. 2008, 102, 164–176.
(2) Seino, S. ATP-sensitive Potassium Channels: a Model of Hetero-
multimeric Potassium Channel/Receptor Assemblies. Annu. Rev. Phy-
siol. 1999, 61, 337–362.
(3) Shi, N. Q.; Ye, B.; Makielski, J. C. Function and Distribution of
the SUR Isoforms and Splice Variants. J. Mol. Cell. Cardiol. 2005,
39, 51–60.
(4) Ardehali, H.; O’Rourke, B. Mitochondrial KATP Channels In Cell
Survival and Death. J. Mol. Cell. Cardiol. 2005, 39, 7–16.
(5) Szewczyk, A.; Skalska, J.; Gzab, M.; Kulawiak, B.; Maliꢀnska, D.;
Koszela-Piotrowska, I.; Kunz, W. S. Mitochondrial Potassium Channels:
From Pharmacology to Function. Biochim. Biophys. Acta 2006,
1757, 715–720.
(21) Dupont, L.; Pirotte, B.; de Tullio, P. 7-Iodo-3-isopropylamino-
4H-1,2,4-benzothiadiazine 1,1-Dioxide. Acta Crystallogr., Sect. C: Cryst.
Struct. Commun. 1999, C55, 1152–1154.
(6) Ye, B.; Kroboth, S. L.; Pu, J. L.; Sims, J. J.; Aggarwal, N. T.;
McNally, E. M.; Makielski, J. C.; Shi, N. Q. Molecular Identification and
Functional Characterization of a Mitochondrial Sulfonylurea Receptor 2
Splice Variant Generated by Intraexonic Splicing. Circ. Res. 2009, 105,
1083–1093.
(7) Sebille, S.; de Tullio, P.; Boverie, S.; Antoine, M. H.; Lebrun, P.;
Pirotte, B. Recent Developments in the Chemistry of Potassium
Channel Activators: the Cromakalim Analogs. Curr. Med. Chem. 2004,
11, 1213–1222.
(8) Mannhold, R. KATP Channel Openers: Structure-activity Rela-
tionships and Therapeutic Potential. Med. Res. Rev. 2004, 24, 213–266.
(9) Garlid, K. D.; Paucek, P.; Yarov-Yarovoy, V.; Murray, H. N.;
Darbenzio, R. B.; D’Alonzo, A. J.; Lodge, N. J.; Smith, M. A.; Grover,
G. J. Cardioprotective Effect of Diazoxide and Its Interaction With
(22) Dupont, L; Pirotte, B .; de Tullio, P. Crystal Structure of
7-Chloro-3-isopropylamino-4H-1,2,4-benzothiadiazine
1,1-Dioxide,
C10H12ClN3O2S. Z. Kristallogr. NCS 220 2005, 353–354.
(23) Dupont, L .; Boverie, S.; Pirotte, B.; de Tullio, P. Crystal
Structure of 7-Chloro-3-isopropoxy-4H-1,2,4-benzothiadiazine 1,1-Di-
oxide Monohydrate, C10H13ClN2O4S H2O. Z. Kristallogr., New Cryst.
3
Struct. 2005, 220, 565–566.
(24) Dupont, L.; Boverie, S.; de Tullio, P.; Pirotte, B. Crystal
Structure of 7-Chloro-3-isopropylsulfanyl-4H-1,2,4-benzothiadiazine
1,1-Dioxide, C10H11ClN2O2S2. Z. Kristallogr., New Cryst. Struct. 2005,
220, 467–468.
3198
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199

Journal of Medicinal Chemistry
ARTICLE
(25) Lachenicht, S.; Fischer, A.; Schmidt, C.; Winkler, M.; Rood, A.;
Lemoine, H.; Braun, M. Synthesis of Modified 4H-1,2,4-Benzothiadia-
zine-1,1-dioxides and Determination of their Affinity and Selectivity for
Different Types of KATP Channels. Chem. Med. Chem. 2009, 4, 1850–
1858.
(26) Lebrun, P; Arkhammar, P.; Antoine, M.-H.; Nguyen, Q.-A.;
Hansen, J. B.; Pirotte, B. A Potent Diazoxide Analogue Activating ATP-
sensitive Kþ Channels and Inhibiting Insulin Release. Diabetologia 2000,
43, 723–732.
(27) Antoine, M. H.; Berkenboom, G.; Fang, Z. Y.; Fontaine, J.;
Herchuelz, A.; Lebrun, P. Mechanical and Ionic Response of Rat Aorta
to Diazoxide. Eur. J. Pharmacol. 1992, 216, 299–306.
(28) Pecorari, P.; Albasini, A.; Raffa, L. 1,2,4-Benzothiadiazines.
XLVIII. Heat-induced Transformations of 3-Alkoxy Derivatives of
1,2,4-Benzothiadiazine-1,1-dioxides. Il Farmaco, Ed. Sci. 1973, 28,
203–213.
(29) Raffa, L.; Di Bella, M.; Melegari, M.; Vampa, G. 1,2,4-Ben-
zothiadiazine Derivatives. XX. Synthesis of 3-Thio-1,2,4-benzothiadia-
zine 1,1-Dioxide and its Derivatives. Il Farmaco, Ed. Sci. 1962,
17, 320–330.
(30) Becker, B.; Antoine, M. H.; Nguyen, Q. A.; Rigo, B.; Cosgrove,
K. E.; Barnes, P. D.; Dunne, M. J.; Pirotte, B.; Lebrun, P. Synthesis and
Characterization of a Quinolinonic Compound Activating ATP-sensitive
Kþ Channels in Endocrine and Smooth Muscle Tissues. Br. J. Pharmacol.
2001, 134, 375–385.
(31) Albert, A.; Serjeant, E. P. The Determination of Ionization
Constants; Chapman & Hall: London, 1971; pp 44ꢀ64.
3199
dx.doi.org/10.1021/jm200100c |J. Med. Chem. 2011, 54, 3188–3199