Am J Physiol Cell Physiol 286: C398–C405, 2004.
First published October 1, 2003; 10.1152/ajpcell.00016.2003.
Dietary cholesterol alters Na⫹/K⫹ selectivity at intracellular Na⫹/K⫹ pump
sites in cardiac myocytes
Kerrie A. Buhagiar,1,2 Peter S. Hansen,1,2 Benjamin Y. Kong,3
Ronald J. Clarke,3 Clyne Fernandes,1,2 and Helge H. Rasmussen1,2
1
University of Sydney and 2Department of Cardiology, Royal North Shore Hospital;
and 3School of Chemistry, University of Sydney, Sydney 2006, Australia
Submitted 12 January 2003; accepted in final form 23 September 2003
Buhagiar, Kerrie A., Peter S. Hansen, Benjamin Y. Kong, The sensitivity of the pump to Na⫹ i reflects binding to three
Ronald J. Clarke, Clyne Fernandes, and Helge H. Rasmussen. cytosolic sites. Binding of Na⫹ at one site occurs inside the
Dietary cholesterol alters Na⫹/K⫹ selectivity at intracellular Na⫹/K⫹ membrane dielectric in a voltage-dependent, highly selective
pump sites in cardiac myocytes. Am J Physiol Cell Physiol 286: manner. Binding at the two other sites is nonselective and
C398–C405, 2004. First published October 1, 2003; 10.1152/ occurs in competition with intracellular K⫹ (K⫹ i ) (1). The
ajpcell.00016.2003.—A modest diet-induced increase in serum cho-
lesterol in rabbits increases the sensitivity of the sarcolemmal
K⫹ ⫹
i /Nai antagonism exhibited by the Na /K
⫹ ⫹
pump varies
Na⫹/K⫹ pump to intracellular Na⫹, whereas a large increase in between different tissues, is particularly pronounced in the
cholesterol levels decreases the sensitivity to Na⫹. To examine the heart (32), and is subject to regulation by intracellular signal
mechanisms, we isolated cardiac myocytes from controls and from pathways (4, 5). The latter can be demonstrated experimentally
rabbits with diet-induced increases in serum cholesterol. The myo- as a dependence of regulation of the pump on the K⫹ concen-
cytes were voltage clamped with the use of patch pipettes that tration in patch pipette filling solutions ([K⫹]pip) used to
contained osmotically balanced solutions with Na⫹ in a concentration measure Ip of voltage-clamped isolated ventricular myocytes
of 10 mM and K⫹ in concentrations ([K⫹]pip) ranging from 0 to 140 (5). Regulation of voltage-dependent binding of Na⫹ i can be
mM. There was no effect of dietary cholesterol on electrogenic demonstrated as a dependence on the test potential (Vm) used to
Na⫹/K⫹ current (Ip) when pipette solutions were K⫹ free. A modest voltage clamp myocytes (3, 12).
increase in serum cholesterol caused a [K⫹]pip-dependent increase in Interest in the sarcolemmal Na⫹/K⫹ pump has traditionally
Ip, whereas a large increase caused a [K⫹]pip-dependent decrease in Ip.
been focused on its role in excitation-contraction coupling and
Modeling suggested that pump stimulation with a modest increase in
serum cholesterol can be explained by a decrease in the microscopic
its putative role as a “receptor” for cardiac glycosides. How-
association constant KK describing the backward reaction E1 ⫹ ever, strong evidence is emerging indicating that the pump has
2K⫹ 3 E2(K⫹)2, whereas pump inhibition with a large increase in serum a much broader role in cell function by interacting with and
cholesterol can be explained by an increase in KK. Because hyper- modulating multiple intracellular signal pathways (35). The
cholesterolemia upregulates angiotensin II receptors and because mechanism for the effect of dietary cholesterol supplementa-
angiotensin II regulates the Na⫹/K⫹ pump in cardiac myocytes in a tion on the sensitivity of the Na⫹/K⫹ pump to Na⫹ i may
[K⫹]pip-dependent manner, we blocked angiotensin synthesis or an- therefore have implications for the pathogenesis and treatment
giotensin II receptors in vivo in cholesterol-fed rabbits. This abolished of cardiovascular manifestations of hypercholesterolemia.
cholesterol-induced pump inhibition. Because the ⑀-isoform of protein Voltage and [K⫹]pip dependence of Ip are regulated by
kinase C (⑀PKC) mediates effects of angiotensin II on the pump, we distinctly different intracellular signal pathways (3–5, 12) and
included specific ⑀PKC-blocking peptide in patch pipette filling so- are subject to control by hormones that can be manipulated
lutions. The peptide reversed cholesterol-induced pump inhibition.
therapeutically (12, 14, 15). The primary objective of this study
partial reactions; protein kinase C; angiotensin converting enzyme was to examine the dependence of cholesterol-induced changes
inhibitors; arteriosclerosis; insulin resistance in Ip on [K⫹]pip and Vm. Because we found that cholesterol-
induced changes in Ip were dependent upon [K⫹]pip and inde-
pendent of Vm, we also examined how the changes are influ-
A MODEST DIET-INDUCED INCREASE in serum cholesterol in rabbits enced by an in vitro experimental manipulation of a messenger
can induce an increase in electrogenic Na⫹/K⫹ pump current pathway known to regulate K⫹ ⫹
i /Nai antagonism of the pump
(Ip) in cardiac myocytes. In contrast, a large increase in serum and an in vivo pharmacological intervention known to alter this
cholesterol is associated with pump inhibition to levels below antagonism.
those of myocytes from rabbits fed a standard diet. The
changes in Ip can only be demonstrated when the intracellular MATERIALS AND METHODS
Na⫹ concentration is near physiological levels. There is no
Treatment protocols. We fed male New Zealand White rabbits
effect of hypercholesterolemia on Ip measured when intracel- standard chow for 4 wk or the same chow supplemented with
lular Na⫹ (Na⫹ i ) is at high levels expected to nearly saturate cholesterol as described previously (10). Most rabbits were given
Na⫹/K⫹ pump sites (10). These findings indicate that dietary chow containing 1% cholesterol for 4 wk. A small number were given
cholesterol is a determinant of the sensitivity of the sarcolem- 0.3% cholesterol for 1 wk. Some rabbits given cholesterol-containing
mal Na⫹/K⫹ pump to intracellular Na⫹. chow were treated with the angiotensin-converting enzyme (ACE)
Address for reprint requests and other correspondence: H. H. Rasmussen, The costs of publication of this article were defrayed in part by the payment
Dept. of Cardiology, Royal North Shore Hospital, Pacific Highway, St. of page charges. The article must therefore be hereby marked “advertisement”
Leonards, Sydney, NSW 2065, Australia (E-mail: helger@med.usyd.edu.au). in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
C398 0363-6143/04 $5.00 Copyright © 2004 the American Physiological Society http://www.ajpcell.org
Downloaded from journals.physiology.org/journal/ajpcell (054.146.058.037) on November 29, 2021.
, CHOLESTEROL AND NA⫹/K⫹ PUMP ION SELECTIVITY C399
inhibitor captopril in a dose of 8 mg䡠kg⫺1 䡠24 h⫺1 (15) or the were compared using both linear regression and a Mann-Whitney rank
angiotensin (AT1) receptor antagonist losartan in a dose of 25 sum test. P ⬍ 0.05 was regarded as significant in all comparisons.
mg䡠kg⫺1 䡠24 h⫺1 (5). The drugs were given the last 8 days before Differential rate equations describing the kinetics of the Na⫹/K⫹
death. pump’s partial reactions were integrated numerically using backward
Measurement of Ip. Single ventricular myocytes were isolated as differentiation formulae within a subroutine of the Numerical Algo-
described previously (15) and stored at room temperature in Krebs- rithms Group (NAG) Fortran Library. Integration yielded the enzyme
Henseleit buffer solution until used for patch-clamp studies. In some steady-state turnover number, which can be compared with the steady-
experiments, myocytes were incubated with 10 nM angiotensin II state pump current, Ip. The best fit of simulations to experimental data
(ANG II) before patch-clamp studies (14). For measurement of Ip at was determined using nonlinear regression by applying Newton’s
a fixed Vm of ⫺40 mV, wide-tipped (4–5 m) patch pipettes were method, also within a subroutine of the NAG Fortran Library. The
filled with solutions containing (in mM) 9 Na-glutamate, 1 NaH2PO4, nonlinear regression procedure involved a two-step iterative process.
5 N-2-hydroxyethylpiperazine-N⬘-2-ethenesulphonic acid (HEPES), 2 First, the set of simultaneous differential rate equations describing the
MgATP, 5 ethylene glycol-bis(-aminoethyl ether)-N,N,N⬘,N⬘-tet- enzyme’s partial reactions considered were solved numerically for
raacetic acid (EGTA), and 0 KCl to 140 mM. Osmotic balance of each [K⫹]pip to obtain values of the pump turnover number at each
pipette solutions was maintained with 150 to 10 mM tetramethylam- value of [K⫹]pip. The normalized reductions in enzyme turnover at
monium chloride (TMA-Cl). The solutions were titrated to a pH of increasing values of [K⫹]pip were then compared with the experimen-
7.05 ⫾0.01 at 35°C with 1 M TMA-OH. In experiments designed to tal values of the normalized current inhibition, Ii. Ii was defined as the
determine the relationship between Ip and Vm, we used patch pipettes difference between Ip recorded using K⫹-free and K⫹-containing
containing (in mM) 10 Na-glutamate, 1 KH2PO4, 5 HEPES, 5 EGTA, patch pipette solutions. The sum of the squares of the residuals
2 MgATP, 60 TMA-Cl, 20 tetraethylammonium chloride, 70 CsOH, between the experimental and calculated values were minimized by
and 50 aspartic acid. Solutions were titrated to a pH of 7.05 ⫾ 0.01 varying the value of KK alone, the microscopic association constant
at 35°C with 1 M HCl. Filled patch pipettes had resistances of for binding of K⫹ ions to the two nonspecific binding sites on the E1
0.8–1.1 M⍀. form of the pump. Each iterative variation of the value of KK thus
Myocytes were suspended in a tissue bath mounted on an inverted required the differential rate equations to be again solved numerically
microscope for determination of Ip. The bath was perfused with so that the residuals could be calculated.
modified Tyrode’s solution which contained (in mM) 140 NaCl, 5.6
KCl, 2.16 CaCl2, 1 MgCl2, 0.44 NaH2PO4, 10 glucose, and 10 RESULTS
HEPES. The solution was titrated to a pH of 7.40 ⫾ 0.01 at 35°C with
NaOH. When the whole cell configuration had been established, we Effect of an increase in serum cholesterol on the dependence
switched to a superfusate that was identical except that it was of Ip on [K⫹]pip. We gave rabbits chow containing 1% choles-
nominally Ca2⫹ free and contained 0.2 mM CdCl2 and 2 mM BaCl2. terol for 4 wk to examine whether a decrease in Ip with a large
Ip was identified as the shift in holding current induced by superfusion increase in serum cholesterol is dependent on [K⫹]pip. The
of 100 M ouabain ⬃12 min after the whole cell configuration had
been established. Ip is reported normalized for membrane capacitance
dietary cholesterol supplementation resulted in a serum cho-
unless otherwise indicated. Membrane currents were recorded using lesterol of 20.6 ⫾ 2.9 mM, a level similar to that reported
the single electrode voltage-clamp mode of an Axoclamp-2B ampli- previously for an identical feeding protocol and higher than
fier and Axotape or pCLAMP software (Axon Instruments, Foster levels in control rabbits on cholesterol-free chow (⬃1 mM)
City, CA). Voltage-clamp protocols were generated with pCLAMP. (10). We measured Ip in myocytes from 12 rabbits fed standard
Reagents and chemicals. TMA-Cl was “purum” grade and pur- chow and in myocytes from 10 rabbits fed a cholesterol-
chased from Fluka (Switzerland). All other chemicals were “analyti- containing chow using a [K⫹]pip of 0, 35, 70, or 140 mM. We
cal” grade and purchased from BDH (Australia). Cholesterol, ANG II, used a different [K⫹]pip in experiments on myocytes from the
and ouabain were purchased from Sigma Chemical. Captopril was same rabbit to reduce effects of inter-rabbit variability. Figure
purchased from Bristol-Myers Squibb Pharmaceuticals (Australia) 1 shows representative traces of membrane currents during
and losartan was donated by Merck, Sharpe, and Dohme (Australia).
The ⑀PKC-blocking peptide was provided by Professor Mochly-
measurement of Ip of myocytes from control rabbits and rabbits
Rosen (Stanford University School of Medicine, CA). fed cholesterol. Experiments were included when no drift in
Statistical analysis. Results are expressed as means ⫾ SE unless holding current could be identified on the digital display of the
indicated otherwise. Student’s t-test for unpaired data was used for voltage-clamp amplifier before and after superfusion of
statistical comparisons. We used Dunnett’s test when the same control ouabain. Holding currents were sampled five times with an
group was used for more than one comparison. Ip/Vm relationships electronic cursor every 5 s before and after superfusion of
Fig. 1. Ouabain-induced shifts in holding current (Ih) of myo-
cytes isolated from control rabbits (con) and rabbits given chow
supplemented with 1% cholesterol for 4 wk (chol). The K⫹
concentrations refer to patch pipette solutions ([K⫹]pip) and Cm
indicates the cell capacitance. The shift in Ih, and hence elec-
trogenic pump current (Ip), was similar in myocytes isolated
from control rabbits and rabbits fed cholesterol when [K⫹]pip
was 0 mM, whereas a difference in Ip between myocytes from
control rabbits and cholesterol-fed rabbits was apparent when
[K⫹]pip was 70 mM.
AJP-Cell Physiol • VOL 286 • FEBRUARY 2004 • www.ajpcell.org
Downloaded from journals.physiology.org/journal/ajpcell (054.146.058.037) on November 29, 2021.
