Arylcyanoguanidines as activators of Kir6.2/SUR1KATP channels and inhibitors of insulin release

Article abstract of DOI:10.1021/jm031018y

Phenylcyanoguanidines substituted with lipophilic electron-withdrawing functional groups, e.g. N-cyano-N′ -[3,5-bis-(trifluoromethyl)phenyl]-N″-(cyclopentyl)guanidine (10) and N-cyano-N′-(3,5-dichlorophenyl)-N′-(3-methylbutyl)guanidine (12) were synthesiz

Full text of DOI:10.1021/jm031018y

3202
J . Med. Chem. 2004, 47, 3202-3211
Ar ylcya n ogu a n id in es a s Activa tor s of Kir 6.2/SUR1KATP Ch a n n els a n d In h ibitor s
of In su lin Relea se
Tina M. Tagmose,† Søren C. Schou,† J ohn P. Mogensen,† Flemming E. Nielsen,† Per O. G. Arkhammar,†
Philip Wahl,† Birgit S. Hansen,† Anne Worsaae,† Harrie C. M. Boonen,† Marie-He´le`ne Antoine,‡
Philippe Lebrun,‡ and J ohn Bondo Hansen*,†
Discovery, Novo Nordisk A/ S, Novo Nordisk Park, DK 2760, Måløv, Denmark, and Laboratoire de Pharmacodynamie et de
The´rapeutique, Universite´ Libre de Bruxelles, 808, Route de Lennik, B-1070 Bruxelles, Belgium
Received September 3, 2003
Phenylcyanoguanidines substituted with lipophilic electron-withdrawing functional groups, e.g.
N-cyano-N′-[3,5-bis-(trifluoromethyl)phenyl]-N′′-(cyclopentyl)guanidine (10) and N-cyano-N′-
(3,5-dichlorophenyl)-N′′-(3-methylbutyl)guanidine (12) were synthesized and investigated for
their ability to inhibit insulin release from beta cells, to repolarize beta cell membrane potential,
and to relax precontracted rat aorta rings. Structural modifications gave compounds, which
selectively inhibit insulin release from âTC6 cells (e.g. compound 10: IC50 ) 5.45 ( 1.9 µM)
and which repolarize âTC3 beta cells (10: IC50 ) 4.7 ( 0.5 µM) without relaxation of
precontracted aorta rings (10: IC50 > 300 µM). Inhibition of insulin release from rat islets was
observed in the same concentration level as for âTC6 cells (10: IC50 ) 1.24 ( 0.1 µM, 12: IC50
) 3.8 ( 0.4 µM). Compound 10 (10 µM) inhibits calcium outflow and insulin release from
perifused rat pancreatic islets. The mechanisms of action of 10 and 12 were further investigated.
The compounds depolarize mitochondrial membrane from smooth muscle cells and beta cell
and stimulate glucose utilization and mitochondrial respiration in isolated liver cells.
Furthermore, 10 was studied in a patch clamp experiment and was found to activate Kir6.2/
SUR1 and inhibit Kir6.2/SUR2B type of KATP channels. These studies indicate that the observed
effects of the compounds on beta cells result from activation of KATP channels of the cell
membrane in combination with a depolarization of mitochondrial membranes. It also highlights
that small structural changes can dramatically shift the efficacy of the cyanoguanidine type of
selective activators of Kir6.2/SUR2 potassium channels.
In tr od u ction
in blood pressure. In pancreatic beta cells high glucose
concentrations increase the ATP/ADP ratio to close KATP
channels. This will depolarize the cellular membrane
and open the voltage-gated Ca2+ (L-type) channels to
stimulate insulin secretion. Pharmacological activation
of the KATP channels in the presence of high glucose
concentrations will prevent insulin secretion. The weak
and nonselective KATP channel opener (KATPCO) diaz-
oxide and the SUR1 selective KATPCOs NNC55-0118
and NN414 (Figure 1) inhibit insulin release.4 The
cyanoguanidine pinacidil is a potent and SUR2 selective
The ATP-sensitive potassium (KATP) channels are
present in various types of cells such as cardiac, smooth
muscle, and pancreatic beta cells. The open-state prob-
ability of the KATP channels, which is regulated by
intracellular nucleotides such as ATP and ADP, couples
the metabolic state to cellular membrane potential.
Pharmacological modulation of the KATP channels by
blockers (e.g. sulfonylurea and benzoic acid insulin
secretagogues) and potassium channel openers has
provided important drugs for treatment of, for example,
metabolic and cardiovascular diseases.1,2
The KATP channel exists as an octameric complex of
the sulfonylurea receptor (SUR) and the pore-forming
inwardly rectifying potassium channel (Kir) in a 4+4
stoichiometry.3 The genes for two closely related sulfo-
nylurea receptors SUR1 and SUR2 have been cloned.
Two different splice variants of SUR2, SUR2A and
SUR2B, have been reported. SUR1 combines with
Kir6.2 to form the KATP channels of pancreatic beta cells,
whereas the cardiac type consists of SUR2A and Kir6.2
and the smooth muscle type of SUR2B and Kir6.1 or
Kir6.2. Activation of KATP channels of vascular smooth
muscle causes relaxation and subsequently reduction
K
ATPCO, which inhibits insulin release only at high
concentrations.4,5 The potassium channel blockers and
potassium channel openers interact with the sulfonyl-
urea receptor in a compound specific and complex
manner, which involves both transmembrane regions
and the two nucleotide binding folds.1,6,7
The existence of mitochondrial ATP-sensitive potas-
sium (mitoKATP) channels has been proposed.8 Opening
of mitoKATP results in an increased K+ influx into the
mitochondria, which could result in uncoupling of the
respiratory chain due to futile cycling of K+ via the K+/
H+ antiporter.9 In consequence, this will increase energy
expenditure and decrease cellular ATP levels, which
subsequently will reduce insulin secretion. The molec-
ular structure of mitoKATP channels is probably multi-
meric and similar to the plasma membrane KATP
channels.10 Diazoxide has been shown to exhibit a
pronounced differential pharmacology between sar-
* To whom correspondence should be addressed. Telephone: +454-
4434857. Fax: +4544434547. E-mail: jbha@novonordisk.com.
† Novo Nordisk A/S.
‡ Universite´ Libre de Bruxelles.
10.1021/jm031018y CCC: $27.50 © 2004 American Chemical Society
Published on Web 05/04/2004

Arylcyanoguanidine KATP Channel Activators
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 12 3203
resulted in compounds which activate beta cell KATP
channels to inhibit glucose-induced insulin secretion at
concentrations which do not relax smooth muscle.21 In
this study, we wish to present an exploration of the
synthesis and mechanism of action of a series of 3,5-
disubstituted phenylcyanoguanidines (Scheme 1).
Ch em istr y
The most commonly used methods for the preparation
of cyanoguanidines are (1) treatment of an aromatic
isothiocyanate with an amine to give the corresponding
thiourea, conversion of the thiourea to its carbodiimide
and treatment of this with cyanamide to give the
cyanoguanidine;22,23 (2) treatment of an aromatic isothio-
cyanate with cyanamide and conversion of the obtained
N-cyanothiourea to the cyanoguanidine by treatment
with an amine in the presence of 1-(3-dimethylamino-
propyl)-3-ethylcarbodiimide hydrochloride;24 (3) treat-
ment of diphenyl N-cyanocarbonimidate25 or dimethyl
N-cyanoamido-S,S-dimethylthio-carbamate26 with an
aromatic amine and subsequently treating the obtained
N-cyano-isourea or N-cyano-isothiourea with alkyl-
amines.
F igu r e 1. Structures of modulators of ATP-sensitive potas-
sium channels.
colemmal KATP and mitoKATP of bovine heart, being
about 1000 times more potent at opening mitoKATP
channels.10 The mitochondrial KATP channels could play
a role in protecting cardiomyocytes during ischemia and
reperfusion, and it has been shown by Garlid et al.11
that diazoxide significantly protected hearts against
ischemic injury at concentrations lower than those
required for the opening of sarcolemmal KATP. The
protective effect of potassium channel openers only
requires modest activation of mitoKATP, which does not
produce any detectable mitochondrial uncoupling.12
It has furthermore been found that the cyanoguani-
dine derivative P1075 uncouples oxidative phosphoryl-
In the synthesis of the present compounds, we used
diphenyl N-cyanocarbonimidate as starting material as
described previously21 (Scheme 1). The phenylcyano-
guanidines were obtained in moderate to good yields.
Resu lts a n d Discu ssion
In Vitr o Scr een in g. Pinacidil relaxes rat aortic rings
precontracted with phenylephrine while it does not
inhibit glucose-induced insulin release from âTC6 beta
cells nor hyperpolarize âTC3 beta cells, as reported
previously.4 Diazoxide is, in contrast, a weak but
nonselective KATPCO as indicated by the effects on beta
cells and vascular tissue.4 The screening assay for
measuring changes in membrane potential of the âTC3
cells has been modified compared to the one that was
used previously.21,27 Using the new procedure, diazoxide
was found to be more potent on âTC3 cells compared to
what was reported previously.
The size and bulkiness of the N-alkyl side chain of
the examined 3,5-bistrifluoromethyl cyanoguanidines
2-10 is important for the activity of the compounds.
With two exceptions, the compounds were able to
hyperpolarize the âTC3 cell membrane and to inhibit
glucose-induced insulin secretion from âTC6 cells. The
tert-amylamino (8) and the isopropylamino (4) deriva-
tives were poorly active on beta cells indicating that two
methyl groups R to the nitrogen considerably reduce
beta cell activity. The 4-methylbutylamino derivative
13
ation through activation of mitoKATP and that high
concentrations of pinacidil can uncouple mitochondria
by virtue of an intrinsic protonophoretic action.14
K
ATPCOs such as diazoxide, NNC 55-0118, and NN414
are able to preserve beta cell function in vivo15-17 and
in vitro,18 an effect possibly mediated through an
inhibition of insulin release (beta cell rest). In high
concentrations, and probably through deactivation of
mitochondria, NNC 55-0118 furthermore protects beta
cells against streptozotocin, alloxan, and interleukin 1â-
induced toxicity.19,20 Compounds, which are able to
inhibit insulin release and depolarize mitochondrial
membranes, therefore could have beneficial effects in
preservation of beta cell function in patients suffering
from type 1 or type 2 diabetes.
We recently found that replacement of the pyrido
group in pinacidil with a 3,5-disubstituted phenyl group
Sch em e 1. Synthesis of Phenylcyanoguanidinesa
a
Reagents and conditions: (a) RT, 14 h; (b) R3NH2, 75 °C, 8 h.

3204 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 12
Tagmose et al.
Ta ble 1. Biological Activity of Compounds 2-19 and Reference KATPCOs on Smooth Muscle and Beta Cells
relaxation of
inhibition of insulin release
membrane potential
rat aorta rings
âTC6 cells
âTC3 cells
compound
R1, R2
R3
IC50, µMa
IC50, µMa
efficacya
%
IC50, µMb
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
CF3 , CF3
CF3 , CF3
CF3 , CF3
CF3 , CF3
CF3 , CF3
CF3 , CF3
CF3 , CF3
CF3 , CF3
CF3 , CF3
OCH3,CF3
Cl,Cl
Cl,Cl
Cl,Cl
Cl,Cl
Cl,Cl
Cl,Cl
F,F
CH2CH2CH3
CH2CH2CH(CH3)2
CH(CH3)2
CH(CH3)CH(CH3)2
S-CH(CH3)C(CH3)3
R-CH(CH3)C(CH3)3
C(CH3)2CH2CH3
1-(3-chlorophenyl)cyclobut-1-yl
cyclopentyl
cyclopentyl
CH2CH2CH(CH3)2
CH(CH3)2
S-CH(CH3)Ph
R-CH(CH3)Ph
12.4 ( 2.69
5.6 ( 2.52
61 ( 51.8
17.6 ( 3.38
11.9 ( 0.60
20.3 ( 3.19
0.9 ( 0.24
27.9 ( 6.21
>300
75 ( 30.1
18.7 ( 5.12
52 ( 27.3
30.0 ( 12.8
43.8 ( 30.1
0.3 ( 0.08
75 ( 16.0
85 ( 11.7
11 ( 2.40
12.8 ( 2.5c
0.8 ( 0.2
5.0 ( 2.9
2.0 ( 0.1
41.4 ( 11.2
5.2 ( 2.3
13.7 ( 4.6
4.4 ( 0.8
NA
2.8 ( 0.8
5.45 ( 1.9
NA
53 ( 5.2
77 ( 3.6
20 ( 9.8
69 ( 6.7
67 ( 7.8
68 ( 5.6
-
66 ( 2.5
65 ( 5
-
19 ( 4
4.7 ( 0.3
NS
5.5 ( 0.4
12 ( 5
18 ( 6
NS
20 ( 7.0
4.7 ( 0.5
NS
4.2 ( 2.0
NA
60 ( 7.9
-
0.8 ( 0.2
NS
39.8 ( 13.9
16.1 ( 5.6
48.5 ( 13.2
28.7 ( 19.4
NA
1.2 ( 0.8
22.4 ( 3.7c
>100
27 ( 4.9
22 ( 7.1
32 ( 5
18 ( 1.2
-
25 ( 10
NS
24 ( 0.6
12 ( 0.4
NS
NS
13.7 ( 0.25
NS
C(CH3)2CH2CH3
cyclopentyl
CH2CH2CH(CH3)2
18
19
23 ( 5.9
diazoxide
23 ( 3.8c
pinacidilc
a
b
Values are means of at least four experiments ( SEM. Values are means of at least three experiments ( SD. c The values are taken
from Nielsen et al.4 (NA ) not active, efficacy was found to be less than 10%; NS ) not significant).
Ta ble 2. Effects of 10, 12 on Insulin Release from Rat Islets,
(3) is the most potent compound, being approximately
Membrane Potential of Mitochondria from Muscle and Beta
Cells and Glucose Utilization in Hepatocytes (HEP-G2 cells)
10 times as potent as diazoxide with respect to both
inhibition of insulin release and hyperpolarization of
âTC3 cell membranes. There appears to be only small
differences between the potencies of the two 1,2,2-
trimethylpropylamino (pinacidil-like) enantiomers 6 and
7. It has previously been shown that modification of the
N-alkyl side-chain of the pyrido-cyanoguanidine, P1075,
reverses the efficacy and produces very potent KATP
channel blockers (e.g. PNU-99963) of vascular smooth
muscle.27 The 1-(3-chlorophenyl)cyclobutylamino side
chain taken from PNU 9996328 in combination with 3,5-
bis(trifluoromethyl)phenyl resulted in the potent inhibi-
tor of insulin secretion 9. In our hands, PNU-99963 did
not affect glucose-stimulated insulin release from âTC6
cells or membrane potential of âTC3 cells (data not
shown) and had only minimal effects on isolated rat
aorta
insulin release
rat islets
HEP-G2
EC50 (µM)
IC50 (µM)
A10 (mito) âTC3 (mito) [Emax: % of full
compd
[eff: %]a
ED50 (µM)b ED50 (µM)c stimulation]d
10
1.24 ( 0.1
[65 ( 8]
3.8 ( 0.4
[72 ( 7.4]
1.0 ( 0.4
2 ( 5
>30
>30
10.4 ( 0.2
[95 ( 10]
22 ( 6
12
[98 ( 16]
a
b
Values are an average of two experiments ( SEM. Values
are from one experiment with eight data points run in duplicate,
the EC50 and SD values extracted using WinCurveFit 1.0.2 with
a four-parameter nonlogistic curve fit. c Values are an average of
d
three experiments. Values are an average of three experiments
( SEM.
less on the substitutions on the phenyl group. The most
potent compounds, which are as potent as pinacidil,
have 3,5-dichloro or 3,5-bistrifluoromethyl phenyl groups
and the tert-amylamino (P1075-like) side chain, 8 and
16. Furthermore, these two compounds are relatively
selective for vascular smooth muscle. The cycloalkyl
derivatives 10, 11, and 17 had little or no activity on
the aortic rings, which make compound 10 beta cell
selective.
The activities of selected compounds on smooth
muscle and beta cells have been examined in further
detail. 10 and 12 have been tested with respect to
inhibition of glucose-stimulated insulin release from
isolated rat islets, on membrane potential of mitochon-
dria of A10 smooth muscle cells and âTC3 beta cells and
on stimulation of glucose utilization of rat hepatocytes
(Table 2).
(EC50 ) 113 ( 25 µM). Substituting one trifluoromethyl
group of 10 with a methoxy group to get 11 considerably
reduced beta cell activity.
Changing the trifluoromethyl groups to chloro groups
gave compounds 12-17 with similar or slightly less
potent effects on beta cells. In contrast, the 3,5-difluo-
rophenyl derivative 18 was not active on beta cells,
indicating that electron-withdrawing substituents on
the phenyl is necessary in order to obtain activity on
beta cells. Retaining the 4-methylbutylamino side-chain
but substituting the 3,5-bis-(trifluoromethyl)phenyl with
a 3-pyrido group (3 vs 19) resulted in a considerable loss
of efficacy with respect to inhibition of insulin release
from âTC6 cells and total loss of effects on âTC3
membrane potential.
The compounds reduced glucose-stimulated insulin
release from isolated islets. The potency on islets was
comparable or higher than on the âTC6 beta cells. In
The effects of the compounds on rat aortic rings are
dependent on the size of the N-alkyl substituent and

Arylcyanoguanidine KATP Channel Activators
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 12 3205
comparison, on rat islets the IC50 for diazoxide averages
20.28 ( 11.7 µM.4
The J C-1 fluorescence probe was used to measure
changes in membrane potential of mitochondria in the
A10 smooth muscle cell line and in the âTC3 cell line.
FCCP was used as a positive control. While concentra-
tions of more than 30 µM of either compound were
needed in order to significantly affect âTC3 mitochon-
drial membrane, 10 and 12 potently depolarized the
mitochondrial membranes of A10 cells as indicated by
the recorded changes in the J C-1 fluorescence (Table
2). The measured changes were in the concentration
ranges similar to the ones at which plasma membrane
KATP channel activation (Table 1) and inhibition of
insulin release is observed. The discrepancies between
the effects of the two compounds on mitochondria from
âTC3 versus A10 cells could in part be due to different
sensitivities of the experimental methods. In compari-
son, 10 was found to depolarize mitochondrial mem-
branes of isolated rat or mouse islets at concentrations
above 10 µM (data not shown, P. O. G. Arkhammar,
unpublished; S. Sandler, unpublished). Furthermore,
compounds 10 and 12 were tested for stimulation of
state-4 respiration in liver mitochondria. The com-
pounds increase O2 respiration in a concentration range
from 1 to 10 µM. The effect observed in the liver
mitochondria can be due to several effects; an activation
of mitochondria KATP channels, interference with mito-
chondria transport systems including exchange of a
proton, a protonophoric uncoupling, inhibition of the
respiratory chain or due to a combination of the men-
tioned effects.
F igu r e 2. Effect of 10 µM 10 on 86Rb outflow from rat
pancreatic islets perifused throughout in the absence (b) or
presence (O) of 10 µM glibenclamide. Basal media contained
extracellular Ca2+ (2.56 mM) and glucose (5.6 mM). Mean
values (( SE mean) refer to six individual experiments.
On the basis of the fact that compound 10 in screen-
ings assays was able to inhibit insulin release and
depolarize mitochondrial membranes without affecting
smooth muscle, we selected this compound for additional
examination.
Effect of 10 on P er ifu sed Ra t P a n cr ea tic Islets.
86Rb Ou tflow . Compound 10 (10 µM) provoked a rapid
and sustained increase in 86Rb outflow from pancreatic
islets perifused throughout in the presence of 5.6 mM
glucose (Figure 2). The withdrawal of the drug from the
perifusate was followed by an immediate reduction in
86Rb outflow rate, indicating that the cationic response
to the compound did not result from a damage of
pancreatic islet cells. When the perifusate was enriched
with the hypoglycaemic sulfonylurea glibenclamide (10
µM), the stimulatory effect of 10 was markedly reduced
(Figure 2).
45Ca Ou tflow a n d In su lin Relea se. In the presence
of 16.7 mM glucose and extracellular Ca2+ in the
perifusing medium, the addition of 10 (10 µM) elicited
an immediate and sustained inhibition of both 45Ca
outflow and insulin release (Figure 3, upper and lower
panel). During the period of exposure to 10 (60th-68th
min), the release of insulin represented 11.81 ( 0.72%
(P < 0.05) of that recorded before the administration of
the drug (40th-44th min). The removal of 10 from the
perifusate was accompanied by a slow increase in both
45Ca outflow and insulin output (Figure 3). When the
same experiment was repeated in the absence of extra-
cellular Ca2+, the basal rate of 45Ca outflow (40th-44th
min) was lower (P < 0.05). Under the latter experimen-
F igu r e 3. Effects of compound 10 on Ca2+ outflow and insulin
release from isolated pancreatic islets. Mean values (( SE
mean) refer to six individual experiments. Upper panel: Effect
of 10 µM 10 on 45Ca outflow from rat pancreatic islets perifused
throughout in the presence of 16.7 mM glucose. Basal media
contained extracellular Ca2+ (b, 2.56 mM) or were deprived
of Ca2+ and enriched with EGTA (O, 0.5 mM). Lower panel:
Effect of 10 µM 10 on insulin release from rat pancreatic islets
perifused throughout in the presence of 16.7 mM glucose. Basal
media contained extracellular Ca2+ (b, 2.56 mM).
tal condition, the addition of 10 (10 µM) provoked a
modest but sustained and reversible rise in the 45Ca
outflow rate (Figure 3, upper panel). In islets exposed
throughout to external Ca2+ but in the absence of
glucose, 10 (10 µM) again provoked a slight increase in
45Ca outflow (data not shown).

3206 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 12
Tagmose et al.
stimulation of isotopic exchange between influent 40Ca
and effluent 45Ca.31 Thus, the inhibitory effect of 10 on
45Ca outflow can be viewed as the result of a reduction
of Ca2+ entry into the islet cells. In agreement with such
a proposal, the decrease in 45Ca outflow mediated by
10 did not occur when the islets were exposed to
noninsulinotropic glucose concentrations or to Ca2+
deprived media.
The data from the radio isotopic experiments indi-
rectly suggest that the drug could be a potent activator
of KATP channels. Indeed, the enhancing effect of 10 on
86Rb outflow is reminiscent of that evoked by other KATP
channel openers27,29,30 and is sensitive to glibenclamide,
a hypoglycaemic sulfonylurea known to close KATP
channels.32,33
All this experimental evidence indicates that the
inhibitory effect of 10 on the insulin releasing process
results from KATP channel activation. The failure of 10
to counteract the increment in 45Ca outflow mediated
by a high K+ concentration (50 mM) further supports
the hypothesis that the primary effect of the compound
is to raise the K+ permeability of the pancreatic beta
cell. Indeed, this radioisotopic response to 50 mM K+,
which is mediated by the opening of L-type voltage-
dependent Ca2+ channels, is known to be sensitive to
F igu r e 4. Effect of a rise in the extracellular concentration
of K+ from 5 to 50 mM on 45Ca outflow from rat pancreatic
islets perifused throughout in the absence (O) or presence of
10 µM 10 (b). Basal media contained extracellular Ca2+ (2.56
mM) and glucose (2.8 mM). Mean values (( SE mean) refer
to five individual experiments.
Ca2+ entry blockers but resistant to K+
openers.27,29-31
channel
ATP
Last, it should be noted that 10, unlike other KATP
channel openers,4,27,29,30 elicited a glibenclamide-resis-
tant modality of 86Rb extrusion. Indeed, the hypogly-
caemic sulfonylurea markedly reduced but failed to
totally abolish the capacity of 10 to increase 86Rb outflow
from prelabeled and perifused rat pancreatic islets.
Moreover, the drug provoked an increase in 45Ca outflow
from islets perifused in the absence or the presence of
noninsulinotropic glucose concentrations. An identical
pattern was revealed in islet exposed to 16.7 mM glucose
and Ca2+-deprived media. These findings suggest that
10 could promote intracellular calcium redistribution
with subsequent changes in 45Ca outflow or activation
of a Ca2+-sensitive but glibenclamide-resistant modality
of 86Rb extrusion.34
In the last series of radio isotopic experiments, we
examined the effect of 10 (10 µM) on the KCl-induced
changes in 45Ca outflow. A rise in the extracellular
concentration of K+ from 5 to 50 mM provoked a rapid
and marked increase in 45Ca FOR from pancreatic islet
perifused in the presence of 2.8 mM glucose and
extracellular Ca2+ (Figure 4). When the same experi-
ments were repeated in the presence of 10 (10 µM)
throughout, the basal rate of 45Ca outflow (40th-44th
min) was higher. Thus, 45Ca outflow before the increase
in extracellular K+ averaged 0.75 ( 0.10%/min in the
absence and 1.08 ( 0.11%/min in the presence of 10 µM
10 (P < 0.1), respectively. The presence of 10 (10 µM)
in the basal medium failed, however, to counteract the
cationic response to high extracellular K+ concentration
(Figure 3).
The radio isotopic experiments clearly revealed that
the addition of micromolar concentrations of 10 pro-
voked a rapid, pronounced and sustained increase in
86Rb outflow from prelabeled and perifused rat pancre-
atic islets. Although the measurement of 86Rb fractional
outflow rate underestimates the real changes in K+
fluxes,29 such data suggest that the drug increases the
membrane K+ permeability.27,29,30
The K+ channel activation mediated by 10 may be
expected to hyperpolarize the beta cell plasma mem-
brane and, in turn, inhibit the L-type voltage-dependent
Ca2+ channels, reduce the Ca2+ inflow and, ultimately,
inhibit the secretory process.
This proposed mechanism of action is attested by the
finding that 10 induced a marked reduction in 45Ca
outflow and insulin release from rat pancreatic islets
perifused in the presence of 16.7 mM glucose and
extracellular Ca2+. In islets exposed throughout to Ca2+
and insulinotropic concentrations of glucose, the 45Ca
fractional outflow rate is known to reflect a sustained
Electr op h ysiology. To further explore its mecha-
nism of action, 10 was studied using the patch clamp
technique. We have previously described that certain
phenyl cyanoguanidines increase ion current flow through
recombinant Kir6.2/SUR1 channels.21 Compound 10
gave reproducible activation of whole-cell ion currents
through Kir6.2/SUR1 channels expressed in HEK293
cells and in patch clamp experiments on Kir6.2/SUR1
expressed in Xenopus oocytes using the inside out
configuration (Figure 5). In accordance with the lack of
effect on phenylephrine-contracted aorta rings, com-
pound 10 did not activate the Kir6.2/SUR2B channels
expressed in Xenopus oocytes. In contrast, it was found
(Figure 6) that the compound dose dependently inhibits
the currents through metabolically activated (3 mM Na-
azide) Kir6.1/SUR2B channels, indicating that the
structural modifications (3-pyridyl to 3,5-bis(trifluor-
methyl)phenyl and 2-(3,3-dimethyl)butyl to cyclobutyl)
have changed the efficacy of the cyanoguanidine deriva-
tive from an selective activator of Kir6.2/SUR2 KATP
channels to an activator of Kir6.2/SUR1 and a blocker
of Kir6.2/SUR2B channels.

Arylcyanoguanidine KATP Channel Activators
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 12 3207
pling mechanism or an activation of mitoKATP channels.
The ability of 10 to inhibit glucose-induced insulin
release is presumably mediated through a modulation
of intracellular Ca2+ concentrations caused by activation
of Kir6.2/SUR1 KATP channels or depolarization of
mitochondrial membranes, which will mobilize intra-
cellular Ca2+ stores. The combined effects of inducing
beta cell rest and deactivating mitochondria could
improve the ability of this compound to protect beta cells
against metabolic and toxic stress. A detailed presenta-
tion of the beta cell protective effects of 10 will be
published (Sandler et al.).
Compound 10 was furthermore found to inhibit ion
currents through the Kir6.2/SUR2B KATP channels of
smooth muscles. This supports previous findings that
structural modifications of selective KATP channel open-
ers such as pinacidil could change the KATP channel
efficacies and that it therefore is possible to identify
compounds which are activators of one KATP channel
subtype and blockers of another.28,35,36
Exp er im en ta l Section
Ch em istr y. Reagents, starting materials and solvents were
purchased from common commercial suppliers and were used
as received. All dry solvents were dried overnight over molec-
ular sieves (0.3 or 0.4 nm). Evaporation was carried out on a
rotary evaporator at bath temperatures <40 °C and under
appropriate vacuum. Flash chromatography was carried out
on a Biotage flash 40 using Biotage flash columns (KP-SIL,
60 Å particle size 32-63 µm). Melting points were determined
with a Bu¨chi B545 apparatus and are uncorrected. Proton
NMR spectra were recorded at ambient temperature using a
Brucker Avance DPX 200 (200 MHz), Brucker Avance DPX
300 (300 MHz), and Brucker Avance DPX 400 (400 MHz) with
tetramethylsilane as an internal standard. Chemical shifts (δ)
are given in ppm and splitting patterns are designated as
follows: s, singlet; d, doublet; dd, double doublet; dt, double
triplet t, triplet, tt, triplet of triplets; q, quartet; quint, quintet;
sext, sextet; m, multiplet, and br ) broad. The 70 eV EI solid
mass spectra were recorded on a Finnigan MAT-TSQ 70 mass
spectrometer. Liquid chromatography-mass spectrometry
(LC-MS) analysis was performed on HP1100 MSD equipped
with a Waters Xterra MS C-18 × 3 mm column. Reactions were
followed by thin-layer chromatography performed on silica gel
60 F254 (Merck) or ALUGRAMSIL G/UV254 (MACHEREY-
NAGEL) TLC aluminum sheets. Elemental analyses (C, H,
N) were performed by Novo Nordisk, Microanalytical Labora-
tory, Denmark.
F igu r e 5. Effects of 10 on Kir6.2/SUR1 channels. A: Whole
cell recording from an HEK293 cell expressing Kir6.2/SUR1.
The points indicate the current obtained in response to a 10
mV depolarizing pulse applied every 10 s from a holding
potential of -80 mV. Diazoxide (300 µM) and 10 (10 µM) were
applied as indicated by the horizontal bars. B: Effect of 10 on
macroscopic currents through Kir6.2/SUR1 channels in inside-
out patches. Mean macroscopic conductances (G) are expressed
as a fraction of the mean slope conductance in nucleotide and
drug free solution (Gcontrol). Data were obtained in the presence
of 100 µM MgATP, 100 µM MgATP plus 1 µM 10, or 100 µM
MgATP plus 10 µM 10.
N-[3,5-Bis(tr iflu or om eth yl)p h en yl]-N′-cya n o-O-p h en -
ylisou r ea (1a ). A solution of diphenylcyanocarbonimidate (2
mmol, 476 mg), 3,5-bis(trifluoromethyl)aniline (2 mmol, 458
mg), and triethylamine (2 mmol, 202 mg) in dichloromethane
(15 mL) was stirred under nitrogen for 12 h. After concentra-
tion, the residue was stirred with toluene (5 mL) for 2 h and
the solid was collected by filtration giving 550 mg of the title
1
compound (73.6%); mp 190.5-191.5 °C; EI SP/MS; H NMR.
F igu r e 6. Effects of 10 on Kir6.2/SUR2B channels. Repre-
sentative time course of 10 induced block of Kir6.2/SUR2B
current activated by Na-azide. Currents were evoked by 10
mV hyperpolarizing steps from a holding potential of -10 mV
every 30 s.
N-Cyan o-N′-[3,5-bis(tr iflu or om eth yl)ph en yl]-N′′-(n -pr o-
p yl)gu a n id in e (2). 1a (0.94 mmol, 350 mg) and n-propyl-
amine (1 mL) was stirred in a sealed flask for 19 h at 75 °C.
After concentration, the residue was dissolved in dichlo-
romethane, washed with 1 N aqueous HCl (2x), water, dried
(Na2SO4) and concentrated. The residue was purified by flash
chromatography (ethyl acetate/heptane 1:2) to give the title
compound (90 mg, 28%) as white crystals. Mp 142.5-143.5
Con clu sion
N-Aryl-cyanoguanidines have been optimized to get
compounds, e.g. 10, which, in vitro, inhibit glucose-
stimulated insulin release. Compound 10 activates
Kir6.2/SUR1 channels of beta cells and depolarize
mitochondrial membranes through an unknown uncou-
1
°C; H NMR. Anal. (C13H12F6N4) C, H, N.
N -Cya n o-N ′-[3,5-b is(t r iflu or om e t h yl)p h e n yl]-N ′′-(3-
m eth ylbu tyl)gu a n id in e (3). A solution of 1a (1 mmol, 373
mg), 3-methylbutylamine (1.15 mmol,100 mg) and triethyl-
amine (1.5 mmol, 150 mg) in acetonitrile (2 mL) was stirred

3208 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 12
Tagmose et al.
for 24 h at 60 °C. After concentration, the residue was purified
by column chromatography (heptane/ethyl acetate 1:1) to give
the title compound (100 mg, 27%). 1H NMR. Anal. (C13H12F6N4)
C, H, N.
N-Cya n o-N′-[3-m eth oxy-5-(tr iflu or om eth yl)p h en yl]-O-
p h en ylisou r ea (1b). To a solution of 3-methoxy-5-trifluoro-
methylaniline (11 mmol, 2.10 g) in dichloromethane (25 mL)
was added diphenyl cyanocarbonimidate (10 mmol, 2.38 g), and
triethylamine (11 mmol, 1.53 mL). The reaction mixture was
stirred under nitrogen for 16 h at room temperature. After
concentration, the residue was stirred with water and then
the water was decanted off. The residue was purified by flash
chromatography (ethyl acetate/heptane 1:2) to give 0.413 g of
N -Cya n o-N ′-[3,5-b is(t r iflu or om e t h yl)p h e n yl]-N ′′-(1-
isop r op yl)gu a n id in e (4). 1a (0.94 mmol, 350 mg) and
isopropylamine (1 mL) was stirred in a sealed flask for 19 h
at 75 °C. After concentration, the residue was dissolved in
dichloromethane, washed with 1 N aqueous HCl (2×), and
water, dried (Na2SO4), and concentrated. The residue was
crystallized in ethyl acetate and heptane to give the title
compound (114 mg, 45%) as white crystals. Mp 176-182 °C;
1H NMR. Anal. (C13H12F6N4), C, H, N.
1
the title compound (12%); Mp 168.5-169.5°C. EI SP/MS; H
NMR.
N-Cya n o-N′-cyclop en t yl-N′′-[3-m et h oxy-5-(t r iflu or o-
m et h yl)p h en yl]gu a n id in e (11). To a solution of 1b (0.52
mmol, 175 mg) in dry acetonitrile (1 mL) were added cyclo-
pentylamine (1.04 mmol, 0.113 mL) and triethylamine (0.080
mL). The mixture was stirred for 19 h at 75 °C under nitrogen.
After concentration, the residue was dissolved in dichloro-
methane, washed with 1 N aqueous HCl (2×), and water, dried
(Na2SO4), and concentrated. The residue was purified by flash
chromatography (ethyl acetate/heptane 1:2) to give a syrup.
Crystallization in ethyl acetate/heptane 1:4 gave gray crystals
(110 mg, 65%). Mp 101-102 °C; EI SP/MS; 1H NMR. Anal.
(C15H17F3N4O) C, H, N.
N-Cya n o-N′-[3,5-b is(t r iflu or om et h yl)p h en yl]-N′′-(1,2-
d im eth ylp r op yl)gu a n id in e (5). A solution of 1a (0.8 mmol,
300 mg), 3-methyl-2-butylamine (0.88 mmol, 0.101 mL), and
triethylamine (0.88 mmol, 0.123 mL) in acetonitrile (2 mL) was
stirred for 7 h at 75 °C. After concentration, the residue was
dissolved in ethyl acetate, washed with water, dried (Na2SO4),
and concentrated. The residue was purified by column chro-
matography (heptane/ethyl acetate 4:1) to give the title
compound (143 mg, 59%) as white crystals. Mp 134-136 °C;
1
EI SP/MS; H NMR; Anal. (C15H16F6N4) C, H, N.
N-Cyan o-N′-(3,5-dich lor oph en yl)-O-ph en ylisou r ea (1c).
To a solution of 3,5-dichloroaniline (11 mmol,1.79 g) in
dichloromethane (25 mL) were added diphenylcyanocarbon-
imidate (10 mmol, 2.38 g) and triethylamine (11 mmol, 1.53
mL). The reaction mixture was stirred under nitrogen for 65
h at room temperature. After concentration, the residue was
stirred with water and the water was decanted off. The residue
was stirred with toluene, and the solid was collected by
filtration, giving 2.04 g of the title compound (67%). EI SP/
(S)-N-Cya n o-N′-[3,5-bis(tr iflu or om eth yl)p h en yl]-N′′-(1-
m eth yl-2,2-d im eth ylp r op yl)gu a n id in e (6). To a solution of
1a (1.0 mmol, 373 mg) in dry acetonitrile (3.5 mL) were added
(S)-3,3-dimethyl-2-butylamine (1.1 mmol, 111 mg, 145 µL) and
triethylamine (1.1 mmol, 111 mg, 153 µL). The mixture was
stirred in a sealed flask for 16 h at 80 °C under nitrogen. After
concentration, the residue was purified by flash chromatog-
raphy (ethyl acetate/heptane 1:2) to give the title compound
(152 mg, 40%) as white crystals. Mp 167.9-168.7 °C. Anal.
(C16H18F6N4) C, H, N.
1
MS; H NMR.
N-Cya n o-N′-(3,5-d ich lor op h en yl)-N′′-(3-m eth ylbu tyl)-
gu a n id in e (12). To 1c (0.98 mmol, 300 mg) in dry acetonitrile
(2 mL) were added 3-methylbutylamine (2.16 mmol, 0.255 mL)
and triethylamine (1.08 mmol, 0.150 mL). The reaction
mixture was stirred for 16 h at 85 °C under nitrogen. The
precipitated material was filtered off and recrystallized from
ethyl acetate to give the title compound (160 mg, 54%) as white
crystals. Mp 146.5-151.5 °C; 1H NMR. Anal. (C13H16Cl2N4) C,
H, N.
N-Cyan o-N′-(3,5-dich lor oph en yl)-N′′-(1-isopr opyl)gu an i-
d in e (13). 1c (0.98 mmol, 300 mg) and isopropylamine (1 mL)
were stirred in a sealed flask for 19 h at 75 °C. After
concentration, the residue was dissolved in dichloromethane,
washed with 1 N aqueous HCl (2×) and water, dried (Na2-
SO4), and concentrated. The residue was crystallized in ethyl
(R)-N-Cya n o-N′-(3,5-bis(tr iflu or om eth yl)p h en yl)-N′′-(1-
m eth yl-2,2-d im eth yl-p r op yl)gu a n id in e (7). To a solution
of 1a (1.0 mmol, 373 mg) in dry acetonitrile (3.5 mL) were
added (R)-3,3-dimethyl-2-butylamine (1.1 mmol, 111 mg, 145
µL) and triethylamine (1.1 mmol, 111 mg, 153 µL). The
mixture was stirred in a sealed flask for 16 h at 80 °C under
nitrogen. After concentration, the residue was purified by flash
chromatography (ethyl acetate/heptane 1:3) to give the title
compound (236 mg, 62%) as white crystals. Mp 167.9-168.7
1
°C; H NMR. Anal. (C16H18F6N4) C, H, N.
N-Cya n o-N′-[3,5-b is(t r iflu or om et h yl)p h en yl]-N′′-(1,1-
d im eth ylp r op yl)gu a n id in e (8). A solution of 1a (0.78 mmol,
290 mg), tert-amylamine (0.85 mmol, 0.100 mL), and trieth-
ylamine (0.85 mmol, 0.119 mL) in acetonitrile (2 mL) was
stirred for 25 h at 75 °C followed by workup as described for
5 to give the title compound (80 mg, 28%) as white crystals.
Mp 149-150 °C; EI SP/MS. 1H NMR. Anal. (C15H16F6N4‚
0.15H2O) C, H, N.
acetate/heptane 1:3 to give the title compound (115 mg, 43%)
1
as white crystals. Mp 156-158.5 °C; H NMR. Anal. (C11H12
-
Cl2N4) C, H, N.
(S)-N-Cya n o-N′-(3,5-d ich lor op h en yl)-N′′-(1-p h en yleth -
yl)gu a n id in e (14). A solution of 1c (1.3 mmol, 400 mg), trieth-
ylamine (1.4 mmol, 0.200 mL), and (S)-1-phenylethylamine (1.4
mmol, 0.185 mL) in acetonitrile (3.5 mL) was stirred in a
sealed flask for 18 h at 80 °C. The cooled reaction mixture
was concentrated, and the residue was purified by flash
chromatography using ethyl acetate/heptane 1:3 as eluent to
give the title compound (269 mg, 63%) as a clear oil. LC/MS;
1H NMR; Anal. (C16H14Cl2N4) C, H, N.
N-Cya n o-N′-[3,5-bis(tr iflu or om eth yl)p h en yl]-N′′-[1-(3-
ch lor op h en yl)cyclobu t-1-yl]gu a n id in e (9). To a suspension
of 1a (1.5 mmol, 570 mg) and 1-(3-chlorophenyl)cyclobut-1-
ylamine (1.8 mmol, 390 mg) in dry acetonitrile (5.0 mL) was
added triethylamine (3.75 mmol, 379 mg, 525 µL). The mixture
was refluxed for 14 h under nitrogen. After concentration, the
residue was purified by flash chromatography using ethyl
acetate/heptane 1:3 to give the title compound (153 mg, 22%)
as white crystals. Mp 181.1-182.4 °C; 1H NMR. Anal. (C20H15
ClF6N4) C, H, N.
-
(R)-N-Cya n o-N′-(3,5-d ich lor op h en yl)-N′′-(1-p h en yleth -
yl)gu a n id in e (15). A solution of 1c (1.3 mmol, 400 mg), trieth-
ylamine (1.4 mmol, 0.200 mL), and (R)-1-phenylethylamine
(1.4 mmol, 0.183 mL) in acetonitrile (3.5 mL) was stirred in a
sealed flask for 18 h at 80 °C. The cooled reaction mixture
was concentrated, and the residue was purified by flash
chromatography using ethyl acetate/heptane 1:3 as eluent to
N-Cya n o-N′-[3,5-b is(t r iflu or om et h yl)p h en yl]-N′′-(cy-
clop en tyl)gu a n id in e (10). To a suspension of 1a (0.400 g,
1.1 mmol) in dry acetonitrile (2.5 mL) were added triethyl-
amine (0.164 mL, 1.2 mmol) and cyclopentylamine (0.116 mL,
1.2 mmol). The homogeneous solution was stirred at 85 °C
under N2 for 3.5 h. The solvent was evaporated, and the
residue was dissolved in ethyl acetate and washed with water
twice. The organic layer was dried (Na2SO4) and concentrated.
The crude product was purified by flash chromatography using
ethyl acetate/heptane 1:2 to give the title compound. Yield 82%
1
give the title compound (218 mg, 51%) as a foam. LC/MS; H
NMR. Anal. (C16H14Cl2N4) C, H, N.
N-Cya n o-N′-(3,5-d ich lor op h en yl)-N′′-(1,1-d im eth ylp r o-
p yl)gu a n id in e (16).22 To a solution of 1c (0.8 mmol, 250 mg)
in dry acetonitrile (2 mL) were added tert-amylamine (0.9
mmol, 0.105 mL) and triethylamine (0.125 mL). The mixture
was stirred for 20 h at 70°C under nitrogen. After concentra-
(0.320 g). mp 156-159 °C. EI SP/MS. 1H NMR. Anal. (C15H14
F6N4‚0.2H2O) C, H, N.
-

Arylcyanoguanidine KATP Channel Activators
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 12 3209
tion, the residue was dissolved in ethyl acetate, washed with
water (2×), dried (Na2SO4), and concentrated. The residue was
purified by flash chromatography (ethyl acetate/heptane 1:4)
to give the title compound (158 mg, 65%) as crystals. Mp
Plates were then incubated at 37 °C, 5.0% CO2, for 15 min.
The loading buffer was removed, and the cells were washed
once with 100 µL of KRW buffer preheated to 37 °C. Finally,
160 µL of KRW (supplemented with 2.5 mM glucose) preheated
to 37 °C was added to all wells, and the plate was placed in a
NovoStar fluorescence plate reader together with a polypro-
pylene plate (Greiner 650201) used to plate out 5× concen-
trated dilution series of test compounds in KRW containing
0.5% DMSO (starting at the highest concentration of the of
the test compound at a 1:3 dilution). Compounds were tested
as multiple dilution series with 8 points in each. After 15 min.,
an initial reading was taken from the plate. Compounds were
then added manually from the compound plate with a 12-
channel multipipet (40 µL/well) and the assay incubated for
30 s, 5 min, and 10 min. A second reading was taken using
the same measurement protocol without any adjustment of
the signal gains after 30 s, 5 min, and 10 min.
1
158.5-160°C (lit.22 151-153 °C); EI SP/MS; H NMR. Anal.
(C13H16Cl2N4) C, H, N.
N-Cya n o-N′-cyclop en tyl-N′′-(3,5-d ich lor op h en yl)gu a n i-
d in e (17). To a solution of 1c (0.8 mmol, 250 mg) in dry
acetonitrile (2 mL) were added cyclopentylamine (0.9 mmol,
0.089 mL) and triethylamine (0.125 mL). The mixture was
stirred for 2.5 h at 82 °C under nitrogen. After concentration,
the residue was purified by flash chromatography (ethyl
acetate/heptane 1:2) to give the title compound (160 mg, 66%)
as crystals. Mp 147.5-148.5 °C; EI SP/MS; 1H NMR. Anal.
(C13H14Cl2N4) C, H, N.
N-Cyan o-N′-(3,5-diflu or oph en yl)-O-ph en ylisou r ea (1d).
To a solution of 3,5-difluoroaniline (11 mmol, 1.42 g) in
dichloromethane (25 mL) were added diphenylcyanocarbon-
imidate (10 mmol, 2.38 g) and triethylamine (11 mmol, 1.53
mL). The reaction mixture was stirred under nitrogen for 65
h at room temperature. After concentration in vacuo at 45 °C,
the residue was stirred with toluene. The title compound was
isolated as a white powder by filtration of the precipitate. Yield
The assay on A10 smooth muscle cells were made with a
similar protocol, however loaded in a 50/50 mixture of KRW
and loading buffer and measured in a FLIPR plate reader
which is a more sensitive system than the NovoStar, and the
measurements were solely based on the green fluorescence
from J C-1.
1
Mea su r em en ts of 86Rb, 45Ca Ou tflow a n d In su lin
Relea se fr om P er ifu sed Ra t P a n cr ea tic Islets. Experi-
ments were performed with pancreatic islets isolated by the
collagenase method from fed Wistar rats. The methods used
to measure 86Rb (42K substitute) outflow, 45Ca outflow, and
insulin release from perifused islets have been described
previously.30,31 Groups of 100 islets were incubated for 60 min
in a bicarbonate-buffered medium (in mM: 115 NaCl, 5 KCl,
2.56 CaCl2, 1 MgCl2, 24 NaHCO3) containing 16.7 mM glucose
and either 86Rb ion (0.15-0.25 mM; 50 µCi mL-1) or 45Ca ion
(0.02-0.04 mM; 100 µCi mL-1). After incubation, the islets
were washed four times with a nonradioactive medium and
then placed in a perifusion chamber. The perifusate was
delivered at a constant rate (1.0 mL min-1). From the 31st to
the 90th min of perifusion, the effluent was continuously
collected over successive periods of 1 min each. An aliquot of
the effluent (0.5 mL) was used for scintillation counting while
the remainder was stored at -20 °C for insulin radioimmu-
noassay.37 At the end of the perifusion, the radioactive content
of the islets was also determined. The outflow of 86Rb or 45Ca
ion (cpm min-1) was expressed as a fractional outflow rate (%
of instantaneous islet content min-1, FOR).
1.63 g (60%); Mp 184-188 °C; EI SP/MS; H NMR.
N-Cya n o-N′-(3,5-d iflu or op h en yl)-N′′-(3-m et h ylb u t yl)-
gu a n id in e (18). A solution of 1d (1.5 mmol, 0.40 g) in dry
acetonitrile (2.5 mL) were added 3-methylbutylamine (1.6
mmol, 0.19 mL) and triethylamine (1.6 mmol, 0.225 mL). The
reaction mixture was stirred for 18 h at 85 °C under nitrogen.
After concentration, the residue was stirred with water.
Filtration and recrystallization of the precipitate from ethyl
acetate/heptane 1:2 gave the title compound. Yield 0.290 g
(75%). Mp 118°-122 °C; EI SP/MS; 1H NMR. Anal. (C13H16F2N4)
C, H, N.
N-Cya n o-N′-(3-m et h ylb u t yl)-N′′-(3-p yr id yl)gu a n id in e
(19). To a solution of 3-aminopyridine (2.3 mmol, 0.215 g) in
dichloromethane (7.5 mL) were added diphenylcyanocarbon-
imidate (2.5 mmol, 0.60 g) and triethylamine (2.5 mmol, 0.359
mL). The reaction mixture was stirred under nitrogen for 17
h at room temperature. After concentration, the residue was
stirred with water, the water was decanted off. The residue
was crystallized in toluene to give 0.459 g N-cyano-O-phenyl-
N′-(3-pyridyl)isourea (77%); Mp 123-125 °C; EI SP/MS; 238
1
(M+). H NMR (CDCl3): δ 7.13 (d, 2H), 7.33 (m, 2H), 7.40 (t,
2H), 7.75 (br d, 1H), 8.48 (d, 1H), 8.70 (br s, 1H), 9.0 ppm (br,
1H).
According to the experiments, the media were enriched with
glucose (Merck, Darmstadt, Germany), glibenclamide (ICN
Biomedicals, Inc., OH) or 10. Glibenclamide and 10 were
dissolved in dimethylsulfoxide which was added to both control
and test media. At the final concentrations used, dimethyl
sulfoxide fails to affect islet function.29,30 When high concentra-
tions of extracellular K+ were used, the concentration of
extracellular NaCl was lowered to keep osmolarity constant.
To a solution of N-cyano-O-phenyl-N′-(3-pyridyl)isourea (1.5
mmol, 0.35 g) in dry acetonitrile (3 mL) was added triethyl-
amine (1.6 mmol, 0.22 mL) and 3-methylbutylamine (1.6 mmol,
0.19 mL). The mixture was stirred for 35 h at 80 °C under
nitrogen. After concentration, the residue was purified by flash
chromatography (ethyl acetate/methanol 9:1) followed by re-
crystallization to give the title compound (0.13 g, 38%) as white
Results are expressed as the mean (( SEM). The statistical
significance of the differences between mean data was assessed
by use of Student’s t-test.
crystals. Mp 136.5-137 °C; EI SP/MS: 1H NMR. Anal. (C12
17N5) C, H, N.
Biology. Effects on membrane potential in âTC3 cells,
-
H
Mea su r em en ts of Glu cose Oxid a tion in Hu m a n Hep a -
tocyte Cell Lin e HEP -G2. The assay measures indirectly the
activity of the respiratory chain in HEP-G2 cells by using d-(5-
3H(N))-glucose. Upon oxidation, the 3H-proton will be incor-
porated into water. The water is thereafter separated from the
D-(5-3H(N))-glucose by evaporation. Finally, the radioactivity
in the water is determined using a Topcounter. Cells are
seeded at a concentration of 2 × 104 cells/100 µL/ well and
then incubate at 37 °C and 5% CO2 overnight.
The next day the compounds to be tested are diluted in
DMSO (Sigma, St. Louis, MO) to 100 times final concentration,
and then they are diluted to final concentration in assay
medium containing 10 µCi/mL D-(5-3H(N))-glucose (Perkin-
Elmer Life Sciences Inc., Boston, MA). The medium is removed
from the cells, and 200 µL of the compounds dilutions are
added in duplicates. The cells are then incubated for another
3 h at 37 °C and 5% CO2. Finally the cells are lysed by adding
50 µL of 10% TCA (trichloroacetate), and the radioactive water
inhibition of glucose-stimulated insulin release in vitro in âTC6
cells isolated or rat islets (static incubations) and relaxation
of phenylephrine-contracted rat aortic rings in vitro were
determined according to procedures described in Nielsen et
al.4
Mitoch on d r ia l Mem br a n e P oten tia l. The âTC3 cells
were seeded out in 96-well plates 2 days prior to experiments
at 50 000 cells per well. They were cultured in DMEM cell
culture medium supplemented with 1000 mg/L glucose, 10%
FCS, penicillin, and streptomycin. At the day of experiment,
loading buffer (KRW buffer (containing in mM: NaCl (140),
KCl (3.6), NaH2PO4 (0.5), MgSO4 (0.5), NaHCO3 (2.0), CaCl2
(1.5), HEPES (10), pH 7.4 (adjusted with NaOH, 1 M)
supplemented with 5 µM J C-1 (Molecular Probes, Leiden, The
Netherlands, T-3168) and containing 10% FCS) prewarmed
to 37 °C was added to all wells still containing culture medium
(100 µL in each).

3210 J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 12
Tagmose et al.
is thereafter separated from the D-(5-3H(N))-glucose by evapo-
ration. The results are expressed in mean ( SEM of three
separate experiments.
Su p p or tin g In for m a tion Ava ila ble: Spectroscopic data
and elementary analysis for compounds 1a -d , 2-19. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Mitoch on d r ia Resp ir a tion . Mitochondria were prepared
according to standard procedure36 from Sprague-Dawley male
rats (approximately 250 g) in a medium containing 250 µM
sucrose, 5 mM HEPES, and 1 mM EDTA (pH 7.2). Mitochon-
drial state IV respiration rate was measured using a Clark-
type oxygen electrode. Mitochondria were incubated at 37 °C
in a total volume of 10 mL containing 1.6 mg mitochondrial
protein/mL in medium containing 220 mM D-mannitol, 5 mM
MgCl2, 2 mM HEPES, and 5 mM K3PO4 (pH 7.4) in the
presence of 10 mM succinic acid, 2.5 µM rotenone, and 1 µg/
mL oligomycin.
Refer en ces
(1) Seino, S.; Miki, T. Physiological and Pathophysiological Roles
of ATP-Sensitive K+ Channels. Prog. Biophys Mol. Biol. 2003,
81, 133-176.
(2) Ashcroft, F. M.; Gribble, F. M. ATP-Sensitive K+ Channels and
Insulin Secretion: Their Role in Health and Disease. Diabeto-
logia 1999, 42, 903-919.
(3) Seino, S. ATP-Sensitive Potassium Channels:
a Model of
Heteromultimeric Potassium Channel/Receptor Assemblies. Annu.
Rev. Physiol. 1999, 61, 337-362.
Electr op h ysiology. Ma m m a lia n Cells. Whole cell cur-
rents were recorded from HEK293 cells stably expressing
human Kir6.2/SUR1 channels, using an EPC9 patch clamp
amplifier (HEKA Electronic GmbH, Lambrecht, Germany).
Cells were clamped at -70 mV and currents evoked by
repetitive 250 ms, 10 mV depolarizing voltage steps. Currents
were filtered at 2 kHz and sampled at 10 kHz. The internal
solution contained (in mM): 120 KCl, 1 MgCl2, 5 EGTA, 2
CaCl2, 20 HEPES, 0.3 MgADP, 5 MgATP (pH 7.3). The
external solution was (in mM): 140 NaCl, 3 KCl, 1 CaCl2, 1
MgCl2, 20 mannitol, 10 HEPES (pH 7.2 with NaOH).
Oocyt es. Oocyte collection and mRNA injection Female
Xenopus laevis were anaesthetized with MS222 (2 g/L added
to the water). From each animal was removed one ovary via a
minilaparotomy, and the incision was sutured and the animal
allowed to recover. Once the wound had completely healed,
the second ovary was removed in a similar operation, and the
animal was then killed by decapitation while under anaes-
thesia. Oocytes were coinjected with about 0.2 ng of Kir6.2
(Genbank D50582) and about 2 ng of mRNA encoding SUR2B
(Genbank AF061324). The final injection volume was 50 nL/
oocyte. Isolated oocytes were maintained in Barth’s solution
and manually defollicated after 30 min collagenase (Sigma
type V) incubation. Injected oocytes were studied 3 to 7 days
after injection.
Xen op u s Oocyte Electr op h ysiology: Tw o-Electr od e
Volta ge Cla m p (TEVC). Whole-cell currents were recorded
from oocytes expressing Kir6.2/SUR2B at 20 to 24 °C using a
two-electrode voltage-clamp amplifier (Warner OC725) and
analyzed using in-house software.24 Currents were filtered at
0.1 kHz and digitized at 0.24 kHz. TEVC electrodes were
pulled from thin-walled borosilicate glass and had resistances
between 0.4 and 1 MΩ when filled with 3 mol/L KCl. KATP
currents were activated by metabolic inhibition with 3 mmol/L
azide, and currents were recorded in extracellular solution
containing (mmol/L): KCl 90, MgCl2 2.5, HEPES 10 (pH 7.2
with KOH). The holding potential was set to the zero current
potential (-10 or -20 mV). Hyperpolarizations of 10 or 20 mV
amplitude of 3 s duration were applied every 30 s. The test
compounds were prepared as 200 mmol/L stock solutions in
DMSO. In control experiments, the maximal DMSO concen-
tration applied (0.15%) was without effect on the KATP current.
Xen op u s Oocyte Electr op h ysiology: Ma cr op a tch es.
Currents were recorded from giant inside-out patches excised
from oocytes expressing Kir6.2/SUR1 channels using an EPC7
patch-clamp amplifier. Currents were evoked by repetitive 3
s voltage ramps from -110 mV to +100 mV, filtered at 0.2
kHz, and digitized at 0.4 kHz. The external (pipet) solution
contained (in mM): 140 KCl, 1.2 MgCl2, 2.6 CaCl2, 10 HEPES
(pH 7.4 with KOH). The intracellular (bath) solution contained
(in mM): 110 KCl, 2 MgCl2, 1 CaCl2 10 EGTA, 10 HEPES (pH
7.2 with KOH); final [K+]: 140 mM). The slope conductance
(G) was measured by fitting a straight line to the current-
voltage relation between -20 mV and -100 mV. Drug effects
were calculated as the conductance in the presence of drug
(G) relative to the conductance in drug and nucleotide-free
solution (Gc).
(4) Nielsen, F. E.; Bodvarsdottir, T. B.; Worsaae, A.; MacKay, P.;
Stidsen, C. E.; Boonen, H. C. M.; Pridal, L.; Arkhammar, P. O.
G.; Wahl, P.; Ynddal, L.; J unager, 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 Beta-Cells. J . Med.
Chem. 2002, 45, 4171-4187.
(5) Lebrun, P.; Devreux, V.; Hermann, M.; Herchuelz, A. Pinacidil
Inhibits Insulin Release by Increasing K+ Outflow From Pan-
creatic B-Cells. Eur. J . Pharmacol. 1988, 156, 283-286.
(6) Dabrowski, M.; Ashcroft, F. M.; Ashfield, R.; Lebrun, P.; Pirotte,
B.; Egebjerg, J .; Hansen, J . B.; Wahl, P. The Novel Diazoxide
Analogue 3-Isopropylamino-7-Methoxy-4H-1,2,4-Benzothiadi-
azine 1,1-Dioxide Is a Selective Kir6.2/SUR1 Channel Opener.
Diabetes 2002, 51, 1896-1906.
(7) Hansen, A. M. K.; Christensen, I. T.; Hansen, J . B.; Carr, R. D.;
Ashcroft, F. M.; Wahl, P. Differential Interactions of Nateglinide
and Repaglinide on the Human Beta-Cell Sulphonylurea Recep-
tor 1. Diabetes 2002, 51, 2789-2795.
(8) Inoue, I.; Nagase, H.; Kishi, K.; Higuti, T. ATP-Sensitive K+
Channel in the Mitochondrial Inner Membrane. Nature 1991,
352, 244-247.
(9) Garlid, K. D. Mitochondrial Volume Control. In Integration of
Mitochondrial Function; Lemasters, J . J ., Hackenbrock, C. R.,
Thurman, R. G., and Westerhoff, H. V., Eds.; Plenum Publish-
ing: New York, 1988; pp 257-276.
(10) Grover, G. J .; Garlid, K. D. ATP-Sensitive Potassium Chan-
nels: a Review of Their Cardioprotective Pharmacology. J . Mol.
Cell. Cardiol. 2000, 32, 677-695.
(11) 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 Mitochondrial ATP-Sensitive K+ Channels.
Possible Mechanism of Cardioprotection. Circ. Res. 1997, 81,
1072-1082.
(12) Krenz, M.; Oldenburg, O.; Wimpee, H.; Cohen, M. V.; Garlid, K.
D.; Critz, S. D.; Downey, J . M.; Benoit, J . N. Opening of ATP-
Sensitive Potassium Channels Causes Generation of Free Radi-
cals in Vascular Smooth Muscle Cells. Basic Res. Cardiol. 2002,
97, 365-373.
(13) J ilkina, O.; Kuzio, B.; Grover, G. J .; Kupriyanov, V. V. Effects
of K(ATP) Channel Openers, P-1075, Pinacidil, and Diazoxide,
on Energetics and Contractile Function in Isolated Rat Hearts.
J . Mol. Cell. Cardiol. 2002, 34, 427-440.
(14) Kowaltowski, A. J .; Seetharaman, S.; Paucek, P.; Garlid, K. D.
Bioenergetic Consequences of Opening the ATP-Sensitive K(+)
Channel of Heart Mitochondria. Am. J . Physiol. Heart. Circ.
Physiol. 2001, 280, H649-H657.
(15) Alemzadeh, R.; Slonim, A. E.; Zdanowicz, M. M.; Maturo, J .
Modification of Insulin Resistance by Diazoxide in Obese Zucker
Rats. Endocrinology 1993, 133, 705-712.
(16) Rasmussen, S. B.; Sorensen, T. S.; Hansen, J . B.; Mandrup-
Poulsen, T.; Hornum, L.; Markhoist, H. Functional Rest Through
Intensive Treatment With Insulin and Potassium Channel
Openers Preserves Residual Beta-Cell Function and Mass in
Acutely Diabetic BB Rats. Hormone Met. 2000, 32, 294-300.
(17) Carr, R. D.; Brand, C. L.; Bo¨dvarsdottir, T. B.; Hansen, J . B.;
Sturis. J . NN414, a SUR1/Kir6.2-Selective Potassium Channel
Opener, Reduces Blood Glucose, Improves Glucose Tolerance in
the VDF Zucker Rat. Diabetes 2003, 52, 2513-2518.
(18) Maedler, K.; Sturis, J .; Lehmann, R.; Spinas, G. A.; Donath. M.
Y. The SUR1/Kir6.2 Selective Potassium Channel Opener
NN414 Prevents Glucose-Induced Apoptosis in Human Pancre-
atic â-Cells. Diabetes 2002, 51, A386.
(19) Kullin, M.; Li, Z. C.; Hansen, J . B.; Bjork, E.; Sandler, S.;
Karlsson, F. A. K-ATP Channel Openers Protect Rat Islets
Against the Toxic Effect of Streptozotocin. Diabetes 2000, 49,
1131-1136.
Ack n ow led gm en t. We thank Tinna Fremming,
Anette Kock, Ann-Charlott Nielsen, J ette Møller, Vibeke
Nielsen, and Berit Gerlach for technical assistance.

Arylcyanoguanidine KATP Channel Activators
J ournal of Medicinal Chemistry, 2004, Vol. 47, No. 12 3211
(20) Kullin, M.; Li, Z.; Hansen, J . B.; Welsh, N.; Karlsson, F. A.;
Sandler, S. Protection of Rat Pancreatic Islets by Potassium
Channel Openers Against Alloxan, Sodium Nitroprusside and
Interieukin-1 Beta Mediated Suppression - Possible Involve-
ment of the Mitochondrial Membrane Potential. Diabetologia
2003, 46, 80-88.
(28) Khan, S. A.; Higdon, N. R.; Hester, J . B.; Meisheri, K. D.
Pharmacological Characterization of Novel Cyanoguanidines As
Vascular KATP Channel Blockers. J Pharmacol. Exp. Ther.
1997, 283, 1207-1213.
(29) Lebrun, P.; Antoine, M. H.; Herchuelz, A. K+ Channel Openers
and Insulin Release. Life Sci. 1992, 51, 795-806.
(21) Tagmose, T. M.; Mogensen, J . P.; Agerholm, P. C.; Arkhammar,
P. O. G.; Wahl, P.; Worsaae, A.; Hansen, J . B. Phenylcyano-
guanidines As Inhibitors of Glucose-Induced Insulin Secretion
From Beta Cells. Bioorg. Med. Chem. Lett. 2001, 11, 1749-1752.
(22) Yoshiizumi, K.; Ikeda, S.; Goto, K.; Morita, T.; Nishimura, N.;
Sukamoto, T.; Yoshino, K. Synthesis and Structure-Activity
Relationships of Novel Phenylcyanoguanidine Derivatives As
Potassium Channel Openers. Chem. Pharm. Bull. (Tokyo) 1996,
44, 2042-2050.
(23) Manley, P. W.; Quast, U. Structure-Activity Studies of Potas-
sium Channel Opening in Pinacidil- Type Cyanoguanidines,
Nitroethenediamines, Thioureas, and Ureas. J . Med. Chem.
1992, 35, 2327-2340.
(24) Atwal, K. S.; McCullough, J . R.; Hedberg, A.; Conder, M. L.;
Ahmed, S. Z.; Cucinotta, G.; Normandin, D. E. The Discovery of
a Novel Calcium Channel Blocker Related to the Structure of
Potassium Channel Opener Cromakalim. Bioorg. Med. Chem.
Lett. 1992, 2, 1475-1478.
(25) Yoshiizumi, K.; Ikeda, S.; Nishimura, N.; Yoshino, K. Synthesis
and Structure-Activity Relationships of Thienylcyanoguanidine
Derivatives as Potassium Channel Openers. Chem. Pharm. Bull.
(Tokyo) 1997, 45, 2005-2010.
(30) Lebrun, P.; Antoine, M. H.; Ouedraogo, R.; Kane, C.; Dunne,
M.; Hermann, M.; Herchuelz, A.; Masereel, B.; Delarge, J .; de
Tullio, P.; Pirotte, B. Activation of ATP-Dependent K+ Channels
and Inhibition of Insulin Release: Effect of BPDZ 62. J .
Pharmacol. Exp. Ther. 1996, 277, 156-162.
(31) Lebrun, P.; Malaisse, W. J .; Herchuelz, A. Evidence for Two
Distinct Modalities of CA2+ Influx into Pancreatic B Cell. Am.
J . Physiol. 1982, 242, E59-E66.
(32) Ashcroft, F. M.; Rorsman, P. Electrophysiology of the Pancreatic
Beta-Cell. Prog. Biophys. Mol. Biol. 1989, 54, 87-143.
(33) Malaisse, W. J .; Lebrun, P. Mechanisms of Sulfonylurea-Induced
Insulin Release. Diabetes Care 1990, 13 Suppl 3, 9-17.
(34) Antoine, M. H.; Hermann, M.; Herchuelz, A.; Lebrun, P. Ionic
and Secretory Response of Pancreatic Islet Cells to Minoxidil
Sulfate. J . Pharmacol. Exp. Ther. 1991, 258, 286-291.
(35) Tagmose, T. M.; Mogensen, J . P.; Agerholm, P. C.; Arkhammar,
P. O. G.; Wahl, P.; Worsaae, A.; Hansen, J . B. Phenylcyano-
guanidines As Inhibitors of Glucose-Induced Insulin Secretion
From Beta Cells. Bioorg. Med. Chem. Lett. 2001, 11, 1749-1752.
(36) Dabrowski, M.; Larsen, T.; Ashcroft, F. M.; Hansen, J . B.; Wahl,
P. Potent and Selective Activation of the Pancreatic Beta-Cell
Type K-ATP Channel by Two Novel Diazoxide Analogues.
Diabetologia 2003, 46, 1375-1382.
(26) Schou, C.; Ottosen, E. R.; Petersen, H. J .; Bjorkling, F.; Latini,
S.; Hjarnaa, P. V.; Bramm, E.; Binderup, L. Novel Cyano-
guanidines With Potent Oral Antitumour Activity. Bioorg. Med.
Chem. Lett. 1997, 7, 3095-3100.
(27) Lebrun, P.; Arkhammar, P.; Antoine, M. H.; Nguyen, Q. A.;
Hansen, J . B.; Pirotte, B. A Potent Diazoxide Analogue Activat-
ing ATP-Sensitive K+ Channels and Inhibiting Insulin Release.
Diabetologia 2000, 43, 723-732.
(37) Leclercq-Meyer, V.; Marchand, J .; Woussen-Colle, M. C.; Giroix,
M. H.; Malaisse, W. J . Multiple Effects of Leucine on Glucagon,
Insulin, and Somatostatin Secretion From the Perfused Rat
Pancreas. Endocrinology 1985, 116, 1168-1174.
J M031018Y