Binding of ATP-sensitive potassium channel (KATP) openers to cardiac membranes: correlation of binding affinities with cardioprotective and smooth muscle relaxing potencies.

Full text of DOI:10.1021/jm970762d

J . Med. Chem. 1998, 41, 271-275
271
Bin d in g of ATP -Sen sitive P ota ssiu m
Sch em e 1
Ch a n n el (KATP ) Op en er s to Ca r d ia c
Mem br a n es: Cor r ela tion of Bin d in g
Affin ities w ith Ca r d iop r otective a n d
Sm ooth Mu scle Rela xin g P oten cies
1
1
Karnail S. Atwal,* Gary J . Grover, Nicholas J . Lodge,
Diane E. Normandin, Sarah C. Traeger,
Paul G. Sleph, Robert B. Cohen,
1
Catherine C. Bryson, and Keneth E. J . Dickinson
Bristol-Myers Squibb Pharmaceutical Research Institute,
P.O. Box 4000, Princeton, New J ersey 08543-4000
Sch em e 2
Received November 10, 1997
KATP openers (e.g., cromakalim (1), pinacidil (2)) are
an important class of compounds with potential clinical
indications for diseases such as hypertension, asthma,
1
hair growth, ischemia, and urinary incontinence. The
major drawback of the first generation agents (e.g., 1,
of 3 was 51-61 Ci/mmol. The KATP opener pinacidil (2)
and compounds 4, 9, and 10 employed in these studies
were all racemates.
2
) is their lack of tissue selectivity which limits clinical
2
utility. The search for tissue selective agents has been
hampered by the lack of high-volume screening meth-
ods, notably those associated with radioligand binding.
Although binding sites for the pinacidil related KATP
As recently reported, the key discovery for the iden-
tification of binding sites for 3 in skeletal muscle
membranes was the presence of nucleotides (e.g., MgATP)
3
opener [ H]P-1075 (3) have been reported in intact
5
3
4
in binding medium. Having established the protocol
tissues and cells, efforts to identify binding sites in
membrane preparation which are usually amenable to
high volume screening have not been successful. We
for the observation of consistent binding in skeletal
muscle membranes, we carried out experiments to
identify binding sites for 3 in canine cardiac membranes.
Our initial studies indicated the existence of specific
binding in canine and rat myocardial membranes with
similar pharmacological characteristics. For example,
have recently reported that 3 binds to skeletal muscle
membranes with high affinity.5 In this communication
we report for the first time the identification of binding
sites for 3 in canine myocardial membranes and the
correlation of binding affinities to cardioprotective and
vasorelaxation potencies for a series of KATP openers.
3
has a Ki value of 27 ( 5 and 30 ( 3 nM in canine and
rat heart myocardial membranes, respectively. The
KATP blocker glyburide also has similar Ki values in the
rat (290 ( 100 nM) and dog (130 ( 20 nM) cardiac
membranes. To evaluate a large number of analogues
for binding affinities, we employed canine cardiac
membranes due to higher supply of the tissue.8 Binding
assays included 240 mM sucrose, 3 mM sodium ortho-
vanadate, 2 mM MgCl2, 50 mM Tris-HCl, pH 7.4, in a
9
total volume of 1 mL. Displacement studies were
conducted by incubation of 3 (14 nM) and inhibitors with
membranes (500-700 µg) for 60 min. Nonspecific
binding was defined with 10 µM nontritiated 3. The
nucleotide requirements for observation of binding were
similar to those described in skeletal muscle mem-
5
branes. Ki values for various compounds were calcu-
lated from IC50 values using the equation of Cheng and
1
0
Most of the analogues (4-10) reported in this paper
have been described previously.6 The remaining com-
pounds can be prepared by treatment of 15 with
diphenyl cyanocarbonimidate (16) to give the intermedi-
ate 17 which is converted to the desired products (11-
Prusoff.
The binding to cardiac membranes was saturable,
reversible and following addition of nontritiated 3, the
radioligand dissociated from membranes with first-order
kinetics (t1/2 ) 21 min). The Kd value calculated as the
ratio of the dissociation and association rate constants
was 39 nM. At 14 nM 3, specific binding represented
87% of total binding. Saturation analysis indicated a
single homogeneous population of high-affinity binding
sites with a Kd value of 28 ( 5 nM and Bmax value of 73
( 8 fmol/mg of protein (n ) 4). The binding was
maximal at pH 8-9, and it was sensitive to proteolysis
1
4) by treatment with amines (18) (Scheme 1). Details
7
of this method are described. Physical properties of
new analogues are summarized in Table 1. The radio-
ligand (3) was prepared from the corresponding olefin
1
9 by catalytic tritiation (Scheme 2). Compound 19 was
generated from 3-aminopyridine and 2,2-dimethylallyl-
amine as described in Scheme 1. The specific activity
S0022-2623(97)00762-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/13/1998

2
72 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 3
Communications to the Editor
Ta ble 1. Physical Properties of Compounds 11-14 and 19
compd mol formula
appearence
mp (°C)a
microanalysis
1
1
1
1
1
1
2
3
4
9
C18H27N5O2S
C18H27N5O2S
C13H16N4I‚0.17H2O
C13H16N7I
colorless solid
colorless solid
colorless solid
off white foam
colorless solid
173-174 (A)
209-210 (A)
152-153 (B)
C, H, N, S
C, H, N, S
C, H, N
b
C12H15N5‚0.2H2O‚1.2ethyl ether
168-170 (C)
C, H, N
a
Solvent for crystallization: A, trituration with isopropyl ether; B, hexanes-CH2Cl2; C, trituration with ethyl ether. b Also reported
in ref 15.
these compounds are close to unity indicating competi-
tive displacement.
The classical KATP blocker glyburide which is known
to inhibit pharmacological effects of KATP openers in-
hibits binding of 3 to cardiac membranes with a Ki value
of 130 nM and a slope factor of 0.67 ( 0.15 nH. Analysis
of the slope factor which is significantly less than unity
by a two-site binding model indicates putative high- and
low-affinity binding components. The high-affinity
binding sites have an apparent Ki value of 70 nM which
represents 70%, of the specific binding. The low-affinity
sites have an apparent Ki value of 4.3 µM. The slope
factor of less than unity for glyburide indicates either
heterogeneous binding or a negative allosteric interac-
tion with the binding sites for compound 3. These data
suggest that glyburide antagonizes the pharmacological
effects of KATP openers in the heart by binding to a site
that is in some manner associated with those occupied
by KATP openers. Previous studies using intact tissues
have reported similar observations for displacement of
3
F igu r e 1. Competition of KATP openers for the [ H]P1075
binding sites on cardiac membranes. Membranes were incu-
bated with [ H]P1075 (14 nM) and the stated concentrations
of inhibitors. Binding was assessed as described and calculated
as percent control. Result represent mean ( SEM of three to
seven experiments.
3
3
3 by glyburide.
To investigate the relevance of binding affinities to
pharmacological activity in the myocardium, we com-
pared binding affinities to cardioprotective potencies as
measured by EC25 values for increase in time to con-
tracture in isolated globally perfused rat hearts.11 The
EC25 value is defined as the concentration necessary to
cause increase in time to contracture by 25% over
by trypsin treatment. The binding was insensitive to
calcium channel blockers, â-adrenergic antagonists, and
several other classes of pharmacological agents. These
characteristics are similar to those described for binding
in skeletal muscle membranes.5
11
baseline value. Time to contracture is defined as the
The structure-activity relationships for displacement
of radioligand were determined by competition experi-
ments using cromakalim (1), pinacidil (2) and its
analogues, and the KATP blocker glyburide. The data
on selected compounds is shown in Figure 1 and the Ki
values for all compounds are given in Table 2. As
expected, 3 is the most potent inhibitor and its calcu-
lated Ki value (27 nM) is similar to the Kd (39 nM)
determined from saturation analyses and kinetic de-
terminations. Pinacidil is 10-fold less potent than 3.
The 4-cyano analogue (4) of pinacidil is about 4-fold less
potent than 3. The size of the alkyl side chain is critical
to potency as the methyl (5) and isopropyl (6) analogues
are less potent than 4. Replacement of the alkyl side
chain with other bulky groups (7, 8) maintains some
inhibitory potency. The nature and the position of the
electron-withdrawing group on the aromatic ring ap-
pears to be crucial as the unsubstituted (9), the 3-cyano
time necessary during total global ischemia to increase
end diastolic pressure by 5 mmHg. In general, com-
pounds (2-4, 7, 13, 14) that are potent cardiopro-
tectants are also potent in displacing 3 from its binding
sites. The analogues (5, 6, 11, 12) with poor cardiopro-
tective potencies are also weak in inhibiting radioligand
binding to cardiac membranes. For a correlation plot,
we excluded analogues for which potency values could
not be quantitated. Thus, 10 analogues with a 30-fold
difference in potencies could be used for a quantitative
correlation between cardioprotective potencies and bind-
ing affinities.
When the EC25 values for cardioprotective potencies
were plotted against the Ki values for displacement of
the radioligand, we obtained a good correlation coef-
2
ficient (r ) 0.88) (Figure 2). If we take into account
the analogues with poor potencies which could not be
included in the correlation graph, the overall cor-
respondence between cardioprotective potencies and
binding affinities is excellent. These results suggest
that binding sites in cardiac membranes may be related
to those mediating cardioprotective activity of KATP
openers. The cardiac membranes used in the present
study were P3 pellets which contain sarcolemmal and
mitochondrial membranes (data not shown). It is pres-
ently uncertain which organelle provided the binding
(10), and the sulfonamido analogues (11, 12) are less
potent than 4. Replacement of the cyano group in 4
with iodine maintains potency either with (14) or
without an additional azido (13) substituent. The
binding is stereoselectively inhibited by the benzopyran
KATP opener cromakalim as its 3S,4R-enantiomer 20 (Ki
)
1.1 µM) is 70-fold more potent than the 3R,4S-
enantiomer 21 (Ki ) 72 µM). The slope factors for all

Communications to the Editor
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 3 273
3
Ta ble 2. Ki Values for [ H]P1075 Binding Sites in Canine Cardiac Membranes, Cardioprotective and Vasorelaxation Potencies of
KATP Openers
a
3
Ki values were calculated from IC50’s determined by displacement of [ H]P1075 from canine cardiac membranes and Hill coefficients
b
for competition curves were calculated by curve fitting of data to a single binding site model. EC25 value is the concentration necessary
to increase time to the onset of contracture (TTC) in globally ischemic rat hearts. c IC50 value is the concentration necessary for relaxation
of the methoxamine contracted rat aorta by 50%.
sites for 3. There is some evidence to suggest that the
precontracted rat aorta by 50%. As shown in Table 2,
there is 5 orders of magnitude difference in vasorelaxant
potencies with compound 13 being the most potent (IC50
) 1.4 nM) and compound 5 being the least potent (IC50
) 130 µM). When the IC50 values for vasorelaxation
i
mechanism of cardioprotection by KATP openers does not
1
2
involve sarcolemmal membrane binding sites. A mi-
tochondrial site of action is proposed for cardioprotection
by KATP openers based on some preliminary studies in
semipurified mitochondrial preparations.1 Although
preliminary studies to identify binding sites in purified
mitochondrial membranes were unsuccessful, further
work is necessary to identify the organelle that support
binding in cardiac membranes.
3
were plotted against the K values for binding affinities,
2
a weak correlation is obtained (r ) 0.71) (Figure 3).
These data indicate that the binding sites for KATP
openers in the cardiac membranes might be distinct
from those mediating vasorelaxation in rat aorta. These
results are supported by numerous studies indicating
the existence of distinct structure-activity relationships
for vasorelaxant and cardioprotective potencies of KATP
We also attempted to correlate the binding affinities
in myocardial membranes with vasorelaxant potencies.
The IC50 values for vasorelaxant potencies were deter-
mined by relaxation of the methoxamine constricted rat
aorta, as described previously.7 The IC50 is defined as
the concentration necessary to cause relaxation of the
7
openers. It is also possible that the benzopyran deriva-
tives and pinacidil analogues express their vasorelaxing
effects by binding to distinct sites in the smooth muscle

2
74 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 3
Communications to the Editor
attempt to correlate binding affinities in membrane
preparations with pharmacological activities of KATP
openers. Despite a good correlation, it is premature to
speculate that the binding sites in cardiac membranes
are relevant to functional potencies (e.g., cardioprotec-
tion) of KATP openers. However, these results are highly
significant as they provide an important step forward
in understanding the mechanism of action of KATP
openers at the molecular level. KATP is thought to be
composed of at least two distinct subunits: sulfonylurea
1
4
receptor and an inward rectifier protein. The binding
conditions reported herein would help unravel the site
of action of KATP openers using purified membrane
preparations containing both the sulfonylurea receptor
and the inward rectifier protein. Further, our work
should facilitate the identification of binding sites in
membrane preparations from other tissues and subcel-
lular compartments where KATP openers are known to
express pharmacological activity. The establishment of
high throughput binding assays using membranes
preparations should facilitate the identification of struc-
turally distinct compounds with tissue selectivity.
i
F igu r e 2. Correlation between binding affinities (K ’s) and
cardioprotective potencies as measured by EC25 values for the
increase in time to contracture for 2-4, 7-10, 13, 14, and 20.
2
A good correlation (r ) 0.88) was obtained for the specified
compounds.
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails, including NMR, MS, and elemental analysis (3 pages).
Ordering information is given on any current masthead page.
Refer en ces
(
1) Gopalakrishnan, M.; J anis, R. A.; Triggle, D. J . ATP-Sensitive
K+ Channels: Pharmacological Properties, Regulation, and
Therapeutic Potential. Drug Dev. Res. 1993, 28, 95-127.
2) Atwal, K. S. Advances in the Structure-Activity Relationships,
Mechanism of Action, and Therapeutic Utilities of ATP-Sensitive
Potassium Channel Openers. Drug Dev. Res. 1994, 33, 250-
(
2
62.3.
+
(
3) (a) Bray, K. M.; Quast, U. A specific binding site for K channel
openers in rat aorta. J . Biol. Chem. 1992, 267, 11689. (b) Manley,
P. W.; Quast, U.; Anders, H.; Bray, K. Synthesis of and
Radioligand binding Studies with a Tritiated Analog: Receptor
Interactions of Structurally Different Classes of Potassium
Channel Openers and Blockers. J . Med. Chem. 1993, 36, 2004.
(
c) Quast, U.; Bray, K. M.; Anders, H.; Manley, P. W.; Baumlin,
+
3
Y.; Dosogne, J . Binding of the K channel opener [ H]P1075 in
rat isolated aorta: relationship to functional effects of openers
and blockers. Mol. Pharmacol. 1993, 43, 474.
(
4) Dickinson, K. E. J .; Cohen, R. B.; Bryson, C. C.; Normandin, D.
E.; Conder, M. L.; Gonzales, S.; McCullough, J . R.; Atwal, K. S.
Characterization of KATP channels in smooth muscle cells and
3
cardiac myocytes with [ H]P1075. FASEB J . 1993, 7, A354.
i
F igu r e 3. Correlation between binding affinities (K ’s) and
vasorelaxation potencies as measured by IC50 values for
(
5) Dickinson, K. E. J .; Bryson, C. C.; Cohen, R. B.; Rogers, L.;
Green, D. W.; Atwal, K. S. Nucleotide Regulation and Charac-
terization of Potassium Channel Opener Binding to Skeletal
Muscle Membranes. Mol. Pharmacol. 1997, 52, 1473-1481.
relaxation of the precontracted rat aorta for 2-4, 6-10, 13,
2
0, and 21. A weak correlation (r2 ) 0.71) was obtained for
the specified compounds.
(6) Atwal, K. S.; Moreland, S.; McCullough, J . R.; O’Reilly, B. C.;
Ahmed, S. Z.; Normandin, D. E. Aryl Cyanoguanidine Potassium
Channel Openers. Bioorg. Med. Chem. Lett. 1992, 2, 83-86.
(7) Atwal, K. S.; Grover, G. J .; Ferrara, F. N.; Ahmed, S. Z.; Sleph,
P. G.; Dzwonczyk, S.; Normandin, D. E. Cardioselective anti-
ischemic ATP-sensitive potassium channel openers 2. Structure-
Activity Studies on Benzopyranyl Cyanoguanidines; Modification
of the Benzopyran. J . Med. Chem. 1995, 38, 1966-1973.
as a somewhat improved correlation coefficient (r2
)
0
.85) is obtained by excluding cromakalim enantiomers
2
0 and 21 from the correlation plot.
Using the pinacidil analogue 3 as a radioligand, we
(
8) Preparation of canine cardiac membranes: Dogs of either sex
were anesthetized with Na pentobarbital (35 mg/kg) and their
hearts rapidly removed and placed in ice-cold saline. The
ventricles were dissected free of atria, fat, and vessels and
weighed (80-100 g). The ventricles were minced and placed in
ice-cold 250 mM sucrose and 50 mM Tris-HCl, pH 7.4 (1:9 w/v).
The tissue was homogenized four times in a Waring blender at
high speed for 20 s each. This homogenate was then further
homogenized with a Brinkman polytron 3 × 10 s at setting 7.
The homogenate was centrifuged at 1100g for 10 min and the
pellet was designated P1. The supernatant was filtered through
two layers of cheesecloth and centrifuged at 12800g for 10 min
have demonstrated the existence of binding sites for
KATP openers in canine cardiac membranes. The key
to the identification of binding sites is the presence of
nucleotides (e.g., MgATP) in the binding medium.
Using a limited number of mostly pinacidil analogues,
we have shown that there is a good correlation between
binding affinities in cardiac membranes and cardiopro-
tective potencies. A reasonably good correlation is also
seen between binding affinities in canine cardiac mem-
branes and smooth muscle relaxing potencies for the
pinacidil class of compounds, although the correlation
is much weaker if we include benzopyranyl KATP open-
ers. As far as we know, this study represents the first
(
P2). The supernatant was centrifuged at 105000g for 35 min
and the pellet (P3) resuspended in 50 mL of 0.6 M KCl, 4 mM
and then
imidazole, pH 7.4, incubated on ice for 1-3
h
centrifuged at 100000g for 35 min. The pellet was resuspended
in sucrose/Tris buffer at a protein concentration of approximately
10 mg/mL and stored frozen at -80 °C until use.

Communications to the Editor
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 3 275
(
9) [3H]P1075 binding assay: Assays were conducted in 12 × 75
G. J .; D’Alonzo, A. J .; Hess, T. A.; Sleph, P. G.; Darbenzio, R. B.
Glyburide-Reversible Cardioprotective Effects of BMS-180448
is Independent of Action Potential Shortening. Cardiovasc. Res.
1995, 30, 731.
mm tubes in a total volume of 1 mL. The assay buffer consisted
of 240 mM sucrose, 3 mM sodium orthovanadate, 2 mM MgCl
0 mM Tris-HCl, pH 7.4, and 2 mM MgATP. The nucleotides
2
,
5
(
sodium or trilithium salts), 3 (14 nM, 51-61 Ci/mmol), and
(
13) (a) Garlid, K. D.; Paucek, P.; Yarov-Yarovoy, V.; Sun, X.;
Schindler, P. A. The Mitochondrial KATP Channel as a Receptor
for Potassium Channel Openers. J . Biol. Chem. 1996, 271, 8796-
inhibitors were incubated with membranes (500-700 µg) for 60
min. Nonspecific binding was defined with 10 µM nontritiated
3
. Bound and free radioligand were separated by filtration
8
799. (b) Paucek, P.; Mironova, G.; Mahdi, F.; Beavis, A. D.;
through a GF/C glass fiber filter on a Brandel cell harvester with
Woldegiorgis, G.; Garlid, K. D. Reconstitution and Partial
Purification of the Glibenclamide-sensitive ATP-dependent K+
Channel from Rat Liver and Beef Heart. J . Biol. Chem. 1992,
3
× 2 mL washes of 50 mM Tris, pH 7.4, at 4 °C. The filters
were transferred to 6 mL plastic scintillation vials, soaked
overnight in 5 mL of scintillation fluid (Packard’s Opti-fluor),
and counted in a Packard liquid scintillation counter.
3
6, 26062.
(
14) (a) Inagaki, N.; Gonoi, T.; Clement, J . P., IV; Namba, N.;
Inazawa, J .; Gonzalez, G.; Aguilar-Bryan, L.; Seino, S.; Bryan,
J . Reconstitution of IKATP: An Inward Rectifier Subunit Plus the
Sulfonylurea receptor Science 1995, 270, 1166. (b) Inagaki, N.;
Gonoi, T.; Clement, J . P., IV; Wang, C.; Aguilar-Bryan, L.; Bryan,
J .; Seino, S. A Family of Sulfonylurea Receptors Determines the
Pharmacological Properties of ATP-Sensitive K+ Channels.
Neuron 1996, 16, 1011-1017.
(
10) Cheng, Y. C.; Prusoff, W. H. Relationship between the inhibition
constant (K ) and the concentration of the inhibitor which causes
0 per cent inhibition (I50) of an enzymatic reaction. Biochem.
i
5
Pharmacol. 1973, 22, 3099-3109.
(
11) Grover, G. J .; Newburger, J .; Sleph, P. G.; Dzwonczyk, S.; Taylor,
S. C.; Ahmed, S. Z.; Atwal, K. S. Cardioprotective Effects of the
Potassium Channel Opener Cromakalim: Stereoselectivity and
Effects on Myocardial Adenine Nucleotides. J . Pharmacol. Exp.
Ther. 1991, 257, 156-162.
12) (a) Yao, Z.; Gross, G. J . Effects of the KATP Opener Bimakalim
on Coronary Blood Flow, Monophasic Action Potential Duration
and Infarct Size in Dogs. Circulation 1994, 89, 1769. (b) Grover,
(15) Gadwood, R. C.; Groppi, V. E. Azidophenylcyanoguanidines as
Photoaffinity Probes. US 5525742, 1996.
(
J M970762D