Ionic Mechanisms Underlying Abnormal QT Prolongation and the Associated Arrhythmias in ...

14 Pages·2007·1.16 MB·English

Original Paper
CCCCCeeeeelllllllllluuuuulllllaaaaarrrrr PPPPPhhhhhyyyyysssssiiiiiooooolllllooooogggggyyyyy
Cell Physiol Biochem 2007;19:225-238 Accepted: November 23, 2006
aaaaannnnnddddd BBBBBiiiiioooooccccchhhhheeeeemmmmmiiiiissssstttttrrrrryyyyy
Ionic Mechanisms Underlying Abnormal QT
Prolongation and the Associated Arrhythmias in
Diabetic Rabbits: A Role of Rapid Delayed Rectifier
K+ Current
Yiqiang Zhang1,2*, Jiening Xiao1,4*, Huixian Lin1,4, Xiaobin Luo1,4,
Huizhen Wang1,3, Yunlong Bai3,4,Jingxiong Wang1,2*, Haiqing Zhang1,

Baofeng Yang3,4 and Zhiguo Wang1,2,4
1Research Center, Montreal Heart Institute, Montreal, 2Department of Medicine, University of Montreal,
3Department of Pharmacology (State-Province Key Laboratory of China), Harbin Medical University, Harbin,
4Institute of Cardiovascular Research, Harbin Medical University, Harbin, *Authors with equal contributions
Key Words reduced by ~40%. The time-dependent kinetics of
Diabetes • QT prolongation • Arrhythmias • Ion these currents remained unaltered. The peak
currents amplitude of L-type Ca2+ current (I ) was reduced
CaL
by ~22% and the inactivation kinetics was slowed;
the integration of these two effects yielded ~15%
Abstract reduction of I . The inward rectifier K+ current (I )
CaL K1
Abnormal QT prolongation with the associated and fast sodium current (I ) were unaffected.
Na
arrhythmias is considered the major cardiac electrical Simulation with LabHEART, a computer model of
disorder and a significant predictor of mortality in rabbit ventricular action potentials, revealed that
diabetic patients. The precise ionic mechanisms for inhibition of I or I alone fails to alter APD whereas
to Ks
diabetic QT prolongation remained unclear. We inhibition of I alone results in 30% APD prolongation
Kr
performed whole-cell patch-clamp studies in a rabbit and inhibition of I alone causes 10% APD
CaL
model of alloxan-induced insulin-dependent diabetes shortening. Integration of changes of all these
mellitus. We demonstrated that heart rate-corrected currents leads to ~20% APD lengthening. Protein
QT interval and action potential duration (APD) were levels of the pore-forming subunits for these ion
prolonged by ~20% with frequent occurrence of channels were decreased to varying extents, as
ventricular tachyarrhythmias. Several K+ currents revealed by immunoblotting analysis. Our study
were found decreased in diabetic rabbits including represents the first documentation of I
Kr
transient outward K+ current (I ) that was reduced by channelopathy as the major ionic mechanism for
to
~60%, rapid delayed rectifier K+ current (I ) reduced diabetic QT prolongation.
Kr
by ~70% and slow delayed rectifier K+ current (I )
Ks Copyright © 2007 S. Karger AG, Basel
© 2007 S. Karger AG, Basel Zhiguo Wang, PhD or Baofeng Yang, MD, PhD; Res Center, Heart Institute225
1015-8987/07/0196-0225$23.50/0 5000 Belanger East, Montreal, PQ H1T 1C8 (Canada) or Dept Pharmacol
Fax [phone] Harbin Med University, Heilongjiang 150086, (P. R. China)
E-Mail [email protected] Accessible online at: Tel. [phone], Fax [phone] or Tel. [phone]
www.karger.com www.karger.com/cpb E-Mail [email protected] or [email protected]
Introduction The roles of several of above mentioned ion currents
in diabetic APD/QT prolongation have been studied with
The most prominent cardiac electrical disturbance experimental animal models. I and I (steady-state
to ss
in diabetic mellitus (DM) is the abnormality of QT interval. outward K+ current) were found reduced in IDDM
When QT interval is corrected for heart rate (QTc), values animals in most studies [13-18] and unaltered in others
longer than 440 ms are considered abnormal, and this [19]. I was found unchanged in IDDM [14, 20-21],
K1
limiting value is commonly exceeded in diabetic patients. excluding its contribution to APD prolongation in these
A prolongation of QT interval has been associated with animal models of diabetes. I was found smaller in
CaL
an increased risk of sudden cardiac death in patients with IDMM in some studies [17] but not different in other
diabetes due to occurrence of lethal ventricular studies [13, 22]. I was also found decreased [23]. These
Na
arrhythmias known as Torsade de Pointes or long QT studies contribute significantly to our current
syndrome (LQTS) [1-2]. Particularly noticeable is that in understanding of the ionic mechanisms for diabetic QT
recent years there have been increasing incidences of prolongation, but several important issues remained
the appearance of ECG disorders with no association unresolved. First, previous studies were conducted nearly
between QT alterations and neuropathy, micro- or macro- exclusively in rats and mice, the species in their adulthood
vascular diseases, cardiovascular drugs, age or gender, not expressing phenotypic and physiologically significant
being these situations additional risk factors [2-3]. LQTS I and I that are otherwise the major repolarizing
Kr Ks
occurs in both insulin-dependent type I (IDDM) and current determining the plateau phase and total APD in
insulin-independent type II DM (NIDDM) patients and humans. Second, inhibition of I , though expected to
to
the prevalence is as high as some 25%. Indeed, QT lengthen APD, has been shown to paradoxically shorten
prolongation has been suggested as a predictor of mortality APD in many species such as humans and rabbits [24-
in both IDDM and NIDDM [2-3]. 26]. Third, I or the equivalent has not been thus far
ss
QT interval reflects the total duration of ventricular identified in ventricular cells from species other than rats
depolarization and repolarization or integrated action and mice, such as humans, canines, rabbits, guinea pigs
potential duration (APD) of heart cells. Cardiac APD is and other I /I -bearing species. Therefore, its
Kr Ks
determined by the rate of membrane repolarization which physiological function and pathological role in diabetic QT
is governed by a delicate balance between inward and prolongation remain questionable. Finally, reduction of I
CaL
outward ion currents. Increase in inward currents and/or and I in diabetic hearts should, if anything, shorten, but
Na
decrease in outward currents tend to prolong APD thereby not lengthen APD/QT interval. Obviously, our
QT interval. L-type Ca2+ current (I ) is a key inward understanding of the ionic mechanisms underlying diabetic
CaL
current responsible for maintaining membrane QT prolongation is still incomplete and there is at need
depolarization or duration of plateau phase. The fast Na+ for further investigation into the issue in an animal model
current (I ,) is mainly responsible for the initial upstroke that possesses a similar set of ion current components as
Na
of an AP, but also contributes to the length of plateau humans. This study was designed (1) to have an overall
phase. The outward currents in a ventricular myocyte picture of alterations of ion currents/channels in diabetic
include a number of K+ channels that are of critical hearts under identical or similar experimental conditions
importance to determining cardiac repolarization thereby in rabbit model of diabetes, a species that possesses similar
QT interval; they are transient outward K+ current (I ) profiles of ion currents as humans and (2) to dissect the
to
[4], rapid delayed rectifier K+ current (I ) [5], slow relative contribution of these ion currents/channels in
Kr
delayed rectifier K+ current (I ) [5] and inward rectifier diabetic QT/APD prolongation and to shed light on the
Ks
K+ current (I ) [6]. I is mainly responsible for the initial potential underlying mechanisms.
K1 to
rapid membrane repolarization. I is a key repolarizing
Kr
current responsible for phase 3 repolarization in cardiac
cells of many species such as human (5) and rabbit [7- Materials and Methods
11], which is carried by the K+ channel subunit encoded
by the human ether-a-go-go related gene (HERG). The Preparation of Rabbit Model of Type I Insulin-Dependent
Diabetes Mellitus (DM)
slow component of delayed rectifier K+ current (I )
Ks Male New Zealand white rabbits weighing 1.6~2.0 kg
regulates only the late phase repolarization owing to its
(Charles River Canada Inc) were housed individually in stainless
slow activation kinetics [5]. Inward rectifier K+ current
steel wire-bottomed cages in a room with a 12:12-hrs light-dark
(I ) contributes to the final phase of repolarization [12]. cycle and standard laboratory rabbit chow and drinking water
K1
226 Cell Physiol Biochem 2007;19:225-238 Zhang/Xiao/Lin/Luo/Wang/Bai/Wang/Zhang/Yang/Wang
ad libitum. The animals were randomly assigned to control heart rate corrected QT intervals using Carlsson’s formula
(Ctl) and IDDM groups. The control animals were untreated (QTc=QT - 0.175 (RR-300)) [27].
age-matched rabbits studied in parallel to IDDM animals. To
establish diabetes, a single injection of 140 mg/kg (body weight) Isolation of Rabbit Ventricular Myocytes
of pre-warmed (37°C) alloxan monohydrate (Sigma-Aldrich), Myocytes were isolated from rabbit left ventricular
freshly dissolved in saline at a concentration of 100 mg/ml, was endocardium via enzymatic digestion with the procedures
administrated via marginal ear vein under local anesthesia. To similar to previously described [28]. Rabbits were anesthetized
prevent fatal hypoglycemia from massive insulin release, 10% by sodium pentobarbital (60 mg/kg, i.v.). Hearts were rapidly
glucose solution (100 mg/kg, s.c.) was administered 4 and 6 hrs excised and mounted on a Langendorff apparatus and perfused
after alloxan treatment. The blood was collected via marginal retrogradely with the following four solutions in a sequential
ear vein after local anesthesia for determining the plasma level order: 1 mM Ca2+ Tyrode (2 min), Ca2+-free Tyrode (3-5 min),
of glucose using a glucometer (TheraSense, USA) and the Ca2+-free Tyrode containing collagenase (Worthington type II)
blood glucose level was monitored weekly thereafter until 10 for 25-35 min. The left ventricular wall was shaved to obtain
weeks when the animals were sacrificed for further experiments. endocardial layer (=1.5 mm thick) and the samples were minced
Only those animals with serum glucose concentrations =15 in the storage solution and filtered. For I recordings, cells
Kr
mM were considered diabetic and were used for further studies. from the apical region and for I recordings cells from the basal
Ks
The plasma insulin was measured by an ELISA kit (Yanaihara area were used to enable better recordings as it has been reported
Inc, Insulin ELISA Kit; Cat# YK060). All procedures were in I density is larger in the apex and I is larger in the base [29].
Kr Ks
accordance with the guidelines set by the Animal Ethics The freshly isolated myocytes were gently centrifuged and
Committee of the Montreal Heart Institute. resuspended in the storage solution for patch-clamp studies.
The solution for cell storage contained (mM): 20 KCl, 10
Implantation of Telemeters and ECG Recording in KHPO, 25 glucose, 70 K-glutamate, 5 β-hydroxybutyric acid,
2 4
Conscious Rabbits 20 taurine, EGTA, 40 mannitol, and 0.1% albumin (pH 7.4).
Rabbits were anesthetized with ketamine (Vetalar®,
BioNiche Animal Health Canada, Belleville, ON, Canada) and Whole-Cell Patch-Clamp Recording
xylazine (Rompun®, Bayer Inc, Toronto, ON, Canada) mixture Patch-clamp techniques have been described in detail
(7:1) at a dosage of 1.2 ml/3 kg (i.m.). Abdominal hair was shaved elsewhere [5, 28, 30, 31]. Currents were recorded in the whole-
and skin was cleaned and sterilized with antiseptic. A small cell voltage-clamp mode and action potentials (APs) were
incision was then made on the skin for subcutaneous recorded in the current-clamp mode, with an Axopatch-200B
implantation of ECG telemeter (EMKA Technologie, Paris, amplifier (Axon Instruments). Borosilicate glass electrodes had
France) and the probes of the telemeter were fixed to the right tip resistances of 1-3 MΩ when filled with the internal pipette
and left underarm positions. Antibiotic cream (Polytopic®, solution. The pipette solution for K+ current recordings
Sabex Inc, Boucherville, QC, Canada) was applied to the closed contained (mM): 130 KCl, 1 MgCl, 5 Mg-ATP, 10 EGTA, and 10
2
skin wounds followed by adherent surgical dressing. Bandages HEPES (pH adjusted to 7.25 with KOH). The pipette solution
were then used to protect the wounds. Antibiotic solution (0.5 used for recording I contained (mM) 135 CsF, 5 NaCl, 10 EGTA
Na
ml, Longisil®, Vétoquinol N.-A. Inc) containing penicillin G and 5 Mg-ATP, 5 HEPES, and for I contained (mM): 20 CsCl,
CaL
benzathine (150,000 IU/ml) and penicillin G procaine (150,000 110 Cesium aspartate, 1 MgCl, 5 Mg-ATP, 10 EGTA, and 10
2
IU/ml) was applied i.m. daily for 5 days after the surgery. Seven HEPES (pH 7.25 with CsOH). The internal pipette solution for
days after implantation, the transducer was activated to record AP recording contained the same components as for K+ currents
the real-time ECG in conscious rabbits before induction of recording, except that EGTA was reduced to 0.05 mM. The
diabetes as the basal measurement. The ECG signal was extracellular (normal Tyrode’s) solution for I and I currents
to K1
acquired and analyzed by the EMKA Technologies’s IOX and AP recording contained (mM): 136 NaCl, 5.4 KCl, 1 CaCl, 1
2
acquisition software and ECG-Auto, respectively. ECG was MgCl, 5 glucose, and 10 HEPES (pH 7.4 with NaOH). The
2
monitored continuously for 24 hrs immediately after treatment extracellular solutions used for recording I contained (mM)
Na
with alloxan and from day 2 after alloxan, ECG was recorded for 132.5 CsCl, 5.0 NaCl, 1.0 MgCl, 1.0 CaCl, 11 glucose, and 10
2 2
20 min at an interval of every 3 hrs. ECG recorded in this way is HEPES. For I recordings, the solution contained (mM): 136
CaL
equivalent to the standard lead II ECG. TEA-Cl, 5.4 CsCl, 1 CaCl, 1 MgCl, 5 glucose, and 10 HEPES
2 2
(pH adjusted to 7.25 with CsOH). For I and I recordings, the
Kr Ks
Surface ECG Recording in Anesthetized Rabbits superfusate was changed to an NMG solution composed of
Standard lead II ECG was recorded before implantation of the following (in mM): 149 N-methyl-D-glucamine, 5 MgCl, 0.9
2
telemeters and right before being sacrificed for experiments (10 CaCl, and 5 HEPES (pH adjusted to 7.4 with HCl). In addition
2
weeks after establishment of diabetes). Sedation and induction to the use of different solutions to optimize recordings of the
of anesthesia were obtained with ketamine (65°mg/kg) and currents of interest and to avoid unwanted currents, ion channel
xylazine (13°mg/kg, i.m.). Three-lead surface ECG was recorded blockers were also used to prevent the contaminating currents.
with silver electrodes placed under the skin at optimized For studies on K+ currents, I was inactivated by holding the
Na
positions to obtain maximal amplitude recordings, enabling membrane at -50 mV and I was blocked by CdCl (200 µM) in
CaL 2
accurate measurements of QT intervals. The QT measurements the bathing solution. 4-Aminopyridine (1 mM) was used to
and simultaneously recorded RR intervals were used to derive inhibit I for recording other currents and glyburide (10 µM)
to
I and Diabetic QT Prolongation Cell Physiol Biochem 2007;19:225-238 227
Kr
plus Mg-ATP (5 mM) in the pipette solution to prevent ATP- incubated for 2 hrs with the HRP-conjugated donkey anti-goat
sensitive K+ current. Dofetilide (1 µM) and HMR1556 (1 µM) IgG (H+L) (1:600) in the blocking buffer. Both primary and
(Avanti Polar Lipid Inc., Alabaster, AL, USA) were used to secondary antibodies were purchased from Santa Cruz
block I and I , respectively, for I recordings. Experiments Biotechnology (Santa Cruz, CA). Bound antibodies were
Kr Ks to
were conducted at 36 ± 1°C except that I was recorded at 18 ± detected using the chemiluminescent substrate (Western Blot
Na
1°C. Junction potentials were zeroed before formation of the Chemiluminescence Reagent Plus, NEN Life Science Products,
membrane-pipette seal. Series resistance and capacitance were Boston, USA). GAPDH was used as an internal control for
compensated and leak currents were subtracted. equal input of protein samples, using anti-GAPDH antibody
The voltage protocols for current recordings are shown purchased from RDI (Flanders, NJ). Coomassie staining was
in the insets of the respective figures. The currents were all also performed to verify the sample quantity. Western blot bands
recorded immediately after membrane rupture and series were quantified using QuantityOne software by measuring the
resistance compensation in order to minimize the possible time- band intensity (Area x OD) for each group and normalizing to
dependent rundown, run-up, or negative shift of currents. GAPDH. The final results are expressed as fold changes by
Individual currents were normalized to the membrane capacity normalizing the data to the control values.
to control for differences in cell size, being expressed as current
density pA/pF. No significant difference in cell capacitance Data Analysis
was found between the two groups: 189±17.9 pF for Ctl and Group data are expressed as mean ± S.E. Statistical
209.9±8 pF (p=0.08) for IDDM. The amplitude of I was measured comparisons (ANOVA followed by Dunnett’s method) were
to
as the difference between the initial peak of I and the current carried out using Microsoft Excel. A two-tailed p< 0.05 was
to
level remaining at the end of the pulse. I was measured as the taken to indicate a statistically significant difference. Nonlinear
K1
magnitude of the current at the end of the pulse relative to zero least square curve fitting was performed with CLAMPFIT in
reference. I was expressed as dofetilide-sensitive currents by pCLAMP 8.0 or GraphPad Prism.
Kr
subtracting the currents recorded 10 min after dofetilide (1 µM)
from the baseline currents before dofetilide. I was taken as
Ks
HMR1556-sensitive currents in the presence of dofetilide. The
Results
amplitude of I and I was measured from both step currents
Kr Ks
at various test potentials (the difference between the current
Diabetic QT Prolongation and the Associated
level at the end of the pulse and zero level) and tail currents
(the difference between the peak tail current and zero level) at Arrhythmias
a repolarizing potential of -40 mV for I or at -20 mV for I . The A total of 55 rabbits were used in this study, among
Kr Ks
amplitude of I and I was measured as the difference between which 22 were in the control group, and 33 in the IDDM
CaL Na
the peak inward currents and the currents remaining at the end
group. Twenty-nine out of 33 rabbits (29/33, 88%)
of the pulse.
developed typical characteristics of type I diabetes 12-
24 hrs after single injection of alloxan, as indicated by the
Western Blot
The membrane protein samples were extracted from rabbit elevated non-fasting blood glucose level (22.6 ± 1.1 mM,
ventricles for immunoblotting analysis of ion channel proteins, p<0.05 vs control) as compared with the normal value
with the procedures essentially the same as described in detail (5.4 ± 0.7 mM) in control animals and by the reduced
elsewhere [28]. The protein content was determined with Bio-
plasma insulin level (from 42 ± 14 for Ctl to 0.08±0.05
Rad Protein Assay Kit (Bio-Rad, Mississauga, ON, Canada)
for IDDM IU/ml, p<0.05). The induced IDDM stabilized
using bovine serum albumin as the standard.
from day 3 after the onset until 10 weeks when the animals
Membrane protein sample (~150 µg) was fractionated by
SDS-PAGE (7.5%-10% polyacrylamide gels) and transferred to were used for cellular and molecular biology studies. Four
PVDF membrane (Millipore, Bedford, MA). The sample was of 33 rabbits of IDDM group (4/33, 12%) were resistant
incubated overnight at 4oC with the primary antibodies in to the alloxan treatment, either failed to develop diabetes
1:50~1:200. Affinity purified polyclonal primary antibodies (blood glucose <12 mM) or recovered from diabetes
against C-termini of human Na1.5 (the pore-forming α-subunit
v shortly after the onset. In addition, among the 29 IDDM
of I ), Ca1.2 (for the α subunit of I ), Kv4.3 (the pore-forming
Na v 1c CaL rabbits, 10 died before complete data collection.
α-subunit of I ), Kir2.1 (for I ), HERG (the pore-forming α-
to K1 Remarkably, heart rate-corrected QT interval (QTc
subunit of I ), and KCNQ1 (KvLQT1, the pore-forming α-
Kr
subunit of I ), and against N-terminus of KCNE1 (minK, the â- interval) were consistently prolonged in rabbits after
Ks
subunit of I ), were all raised in goat. HERG is used for rbERG treatment with alloxan (183 ± 5 ms) for 10 weeks
Ks
throughout the manuscript for simplicity as the rabbit ERG compared with the baseline values obtained before
channel sequence (Genbank accession U87513) is 93% and treatment (155 ± 2 ms, p<0.05, n=19), but not in control
96% homologous to human ERG at the nucleotide and amino
animals (159 ± 3 ms for baseline vs. 158 ± 4 ms for time-
acid levels, respectively. Inhibitory peptide for each antibody
matched comparison, n=22). These data indicate a 21%
was used to test the antibody specificity. Next day, the
membrane was washed in TTBS three times (10 min/each) and prolongation of QTc interval in the IDDM rabbits over
228 Cell Physiol Biochem 2007;19:225-238 Zhang/Xiao/Lin/Luo/Wang/Bai/Wang/Zhang/Yang/Wang
Fig. 1. Electrical disor-
ders in rabbits with
insulin-dependent dia-
betes mellitus (IDDM).
(A) Abnormal prolon-
gation of QT interval in
IDDM rabbits. Shown are
representative ECG recor-
dings from a control (Ctl)
and an IDDM (10 weeks)
rabbits, and mean data
(n=22 rabbits for Ctl and
n=19 for IDDM) of heart
rate-corrected QT interval
(QTc interval). The data
were obtained from
anesthetized rabbits. The
dash lines define the
measurements of QT
interval. *p<0.05 IDDM
vs. Ctl. (B) Ventricular
arrhythmias in IDDM
rabbits. The examples of
ventricular tachycardia
(VT) and ventricular
fibrillation (VF) were
recorded by an ECG
telemeter from an IDDM
rabbit 2 min before
sudden death. (C) and (D)
Blood glucose concentra-
tion and QTc interval,
respectively, as a function
of time after alloxan
administration in weeks
(wk). *p<0.05 vs. 0 wk;
+p<0.05 vs. Ctl at the same time point (n=22 rabbits for Ctl and n=19 for IDDM). *p<0.05 vs. Ctl. (E) Prolongation of action potential
duration (APD) in IDDM rabbits. The action potentials were elicited by a train of 10-consecutive stimuli of 2-ms duration and
twice threshold strengths at a frequency of 2 Hz. The recordings obtained from 10th pulses were taken as steady-state value and
used for data analysis. Mean data are from n=12 cells from 4 (Ctl) and 5 (IDDM) rabbits. *p<0.05 vs. Ctl.
the control animals. Also noticeable is that in the rabbits (VT) defined as a run of rapid ventricular deflections
that failed to develop IDDM, their QTc interval was (QRS complex) lasting longer than 30 s, were observed
normal (167 ± 3 ms) and comparable to that of control in the IDDM rabbits, which was otherwise absent in
rabbits (p>0.05). The QTc data are shown in Figure 1A. control animals. Of the 19 IDDM rabbits used for data
In the present study, the heart rate of diabetic rabbits analyses, 12 developed VT (12/19, 63%). The VT often
(205 ± 6 bpm) was only slightly slower than that of the predisposed to ventricular fibrillation (VF) leading to
control ones (211 ± 8 bpm, p>0.05). sudden death. We were able to record ECG with telemetry
Excessive QTc prolongation creates the substrates from 4 out of 10 rabbits right before they died from five
for arrhythmogenesis. This was indeed demonstrated in days after development of IDDM. The ECG clearly
our experiments with the IDDM rabbits. As depicted in showed runs of polymorphic VT and VF.
Figure 1B, arrhythmias, mainly of ventricular tachycardia To delineate the cellular mechanism underlying the
I and Diabetic QT Prolongation Cell Physiol Biochem 2007;19:225-238 229
Kr
Fig. 2. Comparison of
various K+ currents in
ventricular myocytes
between healthy and
IDDM rabbits. Currents
were recorded with the
voltage protocols shown
in the insets. Left panels:
representative recording
of I (A), I (B), I (C),
to K1 Kr
and I (D) in healthy (Ctl)
Ks
and IDDM rabbit hearts,
and right panels:
averaged I (current
density)-V relationships
(right panels). The inset
in the I I-V curves
K1
shows the outward
portion of I . I is
K1 Kr
defined as the dofetilide
(1 µM)-sensitive currents
and I the dofetilide-
Ks
resistant currents. The I-
V curves for I and I
Kr Ks
were constructed from
step currents and similar
results were obtained
with tail currents. The
values in the brackets
indicate the number of
cells used for data
analysis (the same
below). *p<0.05 IDDM
vs. Ctl.
QTc prolongation in our IDDM model, single cell action Functional Alterations of Ion Currents in
potentials (APs) were recorded in enzymetically dispersed Diabetic Hearts
myocytes from left ventricular endocardium. As illustrated To unravel the changes of ion currents that may
in Figure 1E, APD at 50% repolarization (APD ) and account for the QTc/APD prolongation and the associated
50
90% repolarization (APD ) was approximately 35% and arrhythmias in our IDDM animals, we performed whole-
90
24%, respectively, longer in IDDM than in healthy subjects, cell patch-clamp studies of the ion currents operating
which are somewhat longer than the 21% lengthening of under physiological conditions in ventricular myocytes,
QTc interval. including I , I , I , I , I and I . Our results revealed
to K1 Kr Ks CaL Na
230 Cell Physiol Biochem 2007;19:225-238 Zhang/Xiao/Lin/Luo/Wang/Bai/Wang/Zhang/Yang/Wang
Fig. 3. Comparison of L-
type Ca2+ current (I )
CaL
and fast Na+ current (I )
Na
in ventricular myocytes
between healthy and
IDDM rabbits. Currents
were recorded with the
voltage protocols shown
in the insets. (A) The
inactivation time co-
nstants (ô) of I were
CaL
obtained by the single
exponential fit to the
decaying phase of the
currents at potentials
from -10 to +30 mV. To
integrate the changes of
amplitude and inactiva-
tion kinetics, the areas
under the I traces were
CaL
calculated for more
rational comparison of
I sizes between
CaL
healthy and IDDM
rabbits. *p<0.05 IDDM
vs. Ctl.
reduction of multiple ion currents (I , I , I , and I ) in the kinetics of I did not seem to differ between control
to Kr Ks CaL Kr
cells isolated from IDDM rabbits, compared with healthy and IDDM either (data not shown).
rabbits. I was also reduced in IDDM rabbits relative to
Ks
I current density was approximately 60% smaller healthy control animals (Fig. 2D), but to a much less extent
to
in diabetic myocytes than in control ones and similar compared to I . Similar to other K+ currents, the gating
Kr
percentage of reduction was seen at all test potentials properties of I were unaltered by pathological conditions
Ks
ranging from -40 mV to +60 mV (Fig. 2A). The activation of IDDM (data not shown).
and inactivation kinetics remained unaltered (data not I was only slightly reduced in IDDM hearts. The
CaL
shown). reduction was voltage-independent with around 20%
I current density was found smaller in IDDM decreases at potentials ranging from -40 mV to +50 mV.
K1
myocytes only at non-physiological potentials negative to Noticeably, the inactivation process of I was moderately
CaL
-90 mV and no difference was seen at potential between but significantly slowed in IDDM, relative to control,
-90 mV and +10 mV, where I channels conduct outward myocytes (Fig. 3A). For example, at +10 mV the
K1
currents (the inset of Fig. 2B). inactivation time constant, obtained by the mono-
I was substantially depressed in IDDM cells, as exponential fit to the decaying phase of I , was 29.0 ±
Kr CaL
illustrated in Figure 2C. Intriguingly, the depression of I 1.2 ms (n=5 cells) for control myocytes and 24.0 ± 1.3
Kr
appears to be inversely voltage-dependent with greater ms (n=7 cells) for IDDM myocytes (p<0.05). Since
reduction at more negative potentials. For instance, at - slowing of I inactivation tends to maintain Ca2+ entry
CaL
40 mV I was some 80% smaller in IDDM than in control through the channels so as to counteract the reduction of
Kr
animals, whereas at +10 mV, I was only 60% smaller in I amplitude, we calculated the charge amount by
Kr CaL
IDDM than in control animals. The steady-state voltage- integrating the area under I traces at 0 and +10 mV in
CaL
dependent activation property of I was not changed and order to better quantify the changes of I (Fig. 3A). In
Kr CaL
I and Diabetic QT Prolongation Cell Physiol Biochem 2007;19:225-238 231
Kr
Fig. 4. Relative contributions of various ion currents to APD prolongation in ventricular myocytes from IDDM rabbits.
(A and B) Comparison of various ion currents at selected test potentials (+10 mV and -40 mV or -10 mV for I ). *p<0.05 IDDM vs.
Ks
Ctl. (C) Simulation of action potentials based on the changes of ion currents in IDDM hearts, using LabHEART software. The
labels in each panel indicate the simulation with the inhibition of a given current and the percent reduction of currents is indicated
in the brackets. Bottom left panel (IDDM) represents AP simulated with combinational changes of all ion currents examined.
this way, we found that I was reduced by ~15% in potentials: +10 mV and -40 mV. The voltage of +10 mV
CaL
IDDM hearts. was chosen for it roughly represents the phase 2 plateau
I density, by comparison, was not significantly level, and -40 mV was chosen because this potential falls
Na
changed by the pathological process of IDDM (Fig. 3B), roughly into the middle of phase 3 repolarization, of a
neither was the inactivation kinetics (Data not shown). rabbit ventricular AP. As illustrated in Figure 4A and 4B,
at +10 mV the current densities of I , I and I are
to Kr Ks
Relative Contributions of Various Ion Currents within the same range, around 1.5 pA/pF, and the degrees
to APD Lengthening in IDDM of depression of I and I in IDDM cells are also quite
to Kr
To estimate the relative contributions of various ion comparable (around 60%) and I was decreased by
Ks
currents to diabetic QTc/APD prolongation, we performed approximately 37%. At the same potential, I density
CaL
the following analyses. First, we compared the relative (4.2 pA/pF) was roughly the sum of the three K+ currents
changes of the ion currents that demonstrated differences mentioned above and the reduction of I (charge) was
CaL
between IDDM and control cells at the selected test around 15% when the slowing of inactivation was taken
232 Cell Physiol Biochem 2007;19:225-238 Zhang/Xiao/Lin/Luo/Wang/Bai/Wang/Zhang/Yang/Wang
into consideration by using the area under I traces (Fig.
CaL
3A). Evidently, the total decreases in K+ currents were
much greater than the decrease in I at +10 mV, which
CaL
is expected to result in net lengthening of the plateau phase
or APD . At -40 mV, the overall density of all currents
50
studied was considerably smaller than at +10 mV; for
I , -10 mV was used for analysis because I was
Ks Ks
minimal at -40 mV. Among the various ion currents, I
Kr
density was the greatest and more importantly, I
Kr
depression in IDDM cells was also to the greatest extent,
~80%. By comparison, the reduction of I at -10 mV
Ks
was smaller than at +10 mV (~27% at -10 mV vs. 37%
at +10 mV). Decreases in I and I at -40 mV were to
to CaL
a similar degree as at +10 mV, ~60% and 23%,
respectively. Obviously, the total reduction of outward
K+ currents tremendously exceeded that of the inward
current and slowing of phase 3 repolarization is expected;
hence, a decrease in I seems to be a major contributor.
Kr
To get further insight into the above analysis, we
investigated the changes of APD by simulation of AP
with LabHEART, an interactive computer model designed
for ion channels rabbit ventricular myocyte [32]. The model
allows us the opportunity for predicting changes of APD
based on the changes of ion currents, by inputting the
percent changes of each individual current or of any
combinations of different currents. In this way, we were
able to obtain the predicted APs as shown in Figure 4C.
Reduction of I (60%) alone hardly changed APD. 40%
to
decrease in I failed to alter APD either. I suppression
Ks Kr
by 70% (roughly the average between 60% and 80% at Fig. 5. Alterations of protein levels of K+ ion channel subunits
+10 mV and -40 mV, respectively), however, produced revealed by Western blot analysis. The relative quantification
remarkable lengthening (~30%) of APD and APD . of protein levels was attained by normalizing the band densities
50 90 to GAPDH, followed by further normalization to the values
In contrast, depression of I by only 22% resulted in
CaL from control hearts. The data were averaged from experiments
abbreviation of APD (changes of I amplitude, instead
CaL in triplicate with 7 hearts of healthy and IDDM rabbits,
of the integrated area, was used for maximum effect),
respectively, and are expressed as fold changes over control.
approximately 10% shortening of APD and APD . *p<0.05 IDDM vs. Ctl.
50 90
APD prolongation caused by I inhibition and APD
Kr
shortening caused by I inhibition should partially cancel
CaL
out each other. This was indeed verified when changes
of all currents were incorporated into the simulation; APD and the results are depicted in Figure 5. Kv4.3 (75 kDa),
was less prolonged than when only I was taken into the major pore-forming α-subunit of I in humans and
Kr to
account: APD was lengthened by 19% and APD by rabbits [28], was reduced by ~30% in IDDM rabbits.
50 90
23%. Kir2.1 (110 kDa), the major component of the inward
rectifier K+ channels I [33], remained unchanged in
K1
Altered Protein Levels of Various Ion Channel IDDM hearts. HERG, the pore-forming?-subunit of I
Kr
Subunits [34], demonstrated two discrete bands with molecular
To investigate whether the reduced densities of the mass of 155 kDa and 135 kDa, respectively. The former
ion currents tested in our experiments were due to represents the N-glycosylated form and the latter is non-
decreases in expression levels of the responsible channel N-glycosylated form of the channels. The N-glycosylated
proteins, we went on to carry out Western blot analysis HERG was reduced by around 45% and the non-N-
I and Diabetic QT Prolongation Cell Physiol Biochem 2007;19:225-238 233
Kr
Fig. 5. Alterations of protein
levels of K+ ion channel
subunits revealed by Western
blot analysis. The relative
quantification of protein levels
was attained by normalizing
the band densities to GAPDH,
followed by further normali-
zation to the values from
control hearts. The data were
averaged from experiments in
triplicate with 7 hearts of
healthy and IDDM rabbits,
respectively, and are expres-
sed as fold changes over
control. *p<0.05 IDDM vs. Ctl.
glycosylated HERG was reduced by some 55%, in arrhythmias. Second, our study is also the first to evaluate
IDDM relative to those in healthy rabbits. Unexpectedly, the relative contributions of various ion currents to the
KCNQ1 (80 kDa), the pore-forming -subunit of I [35], diabetic APD lengthening under identical or similar
Ks
decreased by approximately 25% in IDDM hearts. experimental conditions. Finally, I /HERG K+ channel
Kr
Moreover, minK (14 kDa), the auxiliary β-subunit of dysfunction in diabetic heart is caused by functional
KCNQ1, decreased by as high as 85%. However, despite impairment and expression down-regulation. Our study
the fact that minK protein was more pronouncedly suggests that diabetic QT prolongation results from
decreased than KCNQ1 protein, we did not observe dysfunction of multiple ion currents/channels and
marked differences in the activation or deactivation depression of I /HERG is the major ionic contributor.
Kr
kinetics of I (data not shown). The pore-forming α -
Ks 1c
subunit of I Ca 1.2 (210 kDa) was reduced by some Dysfunction of I /HERG as a Major Ionic
CaL v Kr
15%. No difference was found for Na 1.5 (200 kDa), Mechanism for Diabetic QT Prolongation and
v
the α-subunit of cardiac I [36] between IDDM and the Associated Arrhythmias
Na
healthy hearts (Fig. 6). Here we demonstrated remarkable depression of
I and I in IDDM animals, which has not been
Kr Ks
previously reported in the literature, in addition to
Discussion depression of I and I as already known. In agreement
to CaL
with the functional depression, the protein levels of HERG
Here we report a study on the electrical disturbances and KCNQ1/minK were also significantly reduced. These
in a rabbit model of type I diabetes (IDDM) and the findings open up an opportunity for evaluating the relative
related ionic and molecular alterations as possible contributions of the alterations of all these ion currents to
mechanisms. The diabetic animals had abnormal QT diabetic QT/APD prolongation. We used the LabHEART
prolongation and high incidence of ventricular computer model written for rabbit ventricular myocytes
tachyarrhythmias, resembling the clinical observations in [32] to address the issue. With the simulation, we showed
diabetic patients. Our study revealed alterations of multiple that 60% reduction of I alone does not produce any
to
ion currents/channels in IDDM hearts. While some of appreciable alterations of APD. This is not surprising for
the results reproduced the observations from previous direct demonstration of I inhibition causing LQTS has
to
studies by other laboratories, several novel findings are been missing to date, particularly in the clinical settings.
documented in the present study. First, this is the first While inhibition of I indeed can result in APD
to
published study thus far to take I and I into account prolongation in species devoid of I such as rats and
Kr Ks Kr
for diabetic QT prolongation and the associated mice, it paradoxically shortens APD in the species
234 Cell Physiol Biochem 2007;19:225-238 Zhang/Xiao/Lin/Luo/Wang/Bai/Wang/Zhang/Yang/Wang