School of Biomedical Sciences, University of Leeds, Leeds
LS2 9JT, United Kingdom
The inotropic effects of ACh and
adenosine on ferret ventricular cells were investigated with the action
potential-clamp technique. Under current clamp, both agonists resulted
in action potential shortening and a decrease in contraction. Under
action potential clamp, both agonists failed to decrease contraction
substantially. In the absence of agonist, application of the short
action potential waveform (recorded previously in the presence of
agonist) also resulted in a decrease in contraction. Under action
potential clamp, application of ACh resulted in a
Ba2+-sensitive outward current with the characteristics of
muscarinic K+ current (IK,ACh); the
presence of the muscarinic K+ channel was confirmed by PCR
and immunocytochemistry. In the absence of agonist, on application of
the short ACh action potential waveform, the decrease in contraction
was accompanied by loss of the inward Na+/Ca2+
exchange current (INaCa). ACh also inhibited the
background inward K+ current (IK,1).
It is concluded that ACh activates IK,ACh,
inhibits IK,1, and indirectly inhibits
INaCa; this results in action potential shortening, decrease in contraction, and, as a result of the inhibition of IK,1, minimum decrease in excitability.
 |
INTRODUCTION |
PARASYMPATHETIC
NERVES extensively innervate the ventricles, and parasympathetic
stimulation has a negative inotropic effect on the ventricular muscle
(29, 36). Previously, the parasympathetic transmitter ACh
was thought only to have a negative inotropic effect on ventricular
muscle after the muscle had first been potentiated by a 
-agonist
(28). However, it is now well established that ACh does
have a negative inotropic effect in the absence of -stimulation in
ferret, rat, and dog ventricular cells (5, 30, 31, 39). In
the dog, ACh has a substantial effect on left ventricular subepicardial cells, a smaller effect on midwall cells, and little or no effect on
subendocardial cells (39). The negative inotropic effect of ACh has been suggested to be the result of the activation of the
muscarinic K+ current (IK,ACh)
despite the fact that the muscarinic K+ channel (a
heteromultimer of Kir3.1 or GIRK1 and Kir3.4 or GIRK4), although
detectable in the atrium, has not been detected in the ventricles
(26). It is postulated that the activation of
IK,ACh results in a shortening of the action
potential and thus a decrease in the contraction. An ACh-induced
decrease in the basal L-type Ca2+ current
(ICa,L) may also be involved in the ACh-induced
changes in the action potential and contraction (27, 39).
The action potential-clamp technique was devised by Starzak and Needle
(37) and first used with cardiac tissue by Doerr et al.
(13). Whereas the conventional voltage-clamp technique uses square-wave command pulses, the action potential-clamp technique involves the use of an action potential waveform as the voltage-clamp command. The action potential-clamp technique has been used to study
the time course and amplitude of ionic currents during the action
potential: current recorded during an action potential voltage-clamp
command waveform after the application of a blocker has been subtracted
from that under control conditions; ICa, the Na+/Ca2+ exchange current
(INaCa), the delayed rectifier K+
current (IK), and the background inward
rectifier K+ current (IK,1) have
been studied in this way (1, 17, 19, 20). The principal
aim of this study was to investigate the ACh-induced changes in current
during the ventricular action potential. Adenosine was studied in the
same way. This is the first study to investigate agonist-induced
changes in membrane current during the action potential. Another aim of
the study was to investigate the mechanisms underlying the negative
inotropic effect of ACh during the action potential. In part, this
involved the detection of Kir3.1 mRNA by screening ventricular tissue
cDNA using the PCR technique and Kir3.1 and Kir3.4 proteins in
ventricular cells by immunocytochemistry.
| |
MATERIALS AND METHODS |
Experiments on single cells.
Experiments were performed on isolated ferret ventricular cells.
Ferrets were anesthetized by intraperitoneal injection of 90-150
mg pentobarbitone sodium. The heart was rapidly excised and placed in
"isolation solution" (see below) containing 750 µM
CaCl2. The aorta was cannulated and the heart retrogradely perfused at ~20-23 ml/min with isolation solution containing 750 µM Ca2+ for sufficient time to clear all the blood
(~2-3 min). The perfusate was then switched to
Ca2+-free isolation solution (containing 0.1 mM EGTA) for 4 min. Finally, the heart was perfused for 10 min with "enzyme
solution": isolation solution containing 1 mg/ml collagenase
(Worthington type II, Lorne laboratories), 0.1 mg/ml protease (Sigma;
Poole, UK), and 50 µM CaCl2; this solution was
recirculated through the heart. The ventricles were cut away from the
atria, finely chopped, placed in a conical flask, and digested with the
enzyme solution supplemented with 10% (wt/vol) bovine serum albumin
for a further 5 min. The tissue was shaken gently during this period.
This process was repeated four times, and the cells from each 5-min
period were harvested by filtration and pelleted by centrifugation at
400 rpm for 40 s. Cells were washed by resuspending them in
isolation solution containing 750 µM Ca2+ and then
recentrifuging them. The cells were stored in isolation solution
containing 750 µM Ca2+ at 4°C until required. The
isolation procedure was performed at 37°C. Isolation solution
contained (in mM) 130 NaCl, 5.4 KCl, 1.4 MgCl2, 0.4 NaH2PO4, 10 creatine, 20 taurine, 10 HEPES, and 10 glucose; pH 7.3 at room temperature. This solution was equilibrated with O2.
Cells were pipetted into a small tissue bath (volume, 0.2 ml) attached
to the stage of an inverted microscope (Nikon Diaphot). The cells were
allowed to settle for several minutes onto the glass bottom of the
chamber before being superfused at a rate of ~1.6 ml/min with Tyrode
solution of the following composition (in mM): 136.9 NaCl, 5.4 KCl, 2 CaCl2, 0.57 MgCl2, 0.37 NaH2PO4, 5 HEPES, and 5.6 glucose; pH 7.4 at
37°C. In some experiments 5 µM
1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA)-AM was added to the Tyrode solution to buffer intracellular Ca2+. A 1 mM BAPTA-AM stock solution was made by dissolving
the drug in DMSO. Cells were perfused with Tyrode solution containing 5 µM BAPTA-AM (0.5% DMSO) for 5-10 min. During this period the
size of the contraction dwindled as a result of the buffering of
intracellular Ca2+. When the contraction was almost
completely abolished, the BAPTA-AM was washed off. On washoff of
BAPTA-AM, the contraction did not recover (BAPTA remained trapped
within the cells as a result of the cleavage of the AM group by
intracellular esterases). As a control, cells were exposed to Tyrode
solution containing 0.5% DMSO only for 10 min; this had no effect
(n = 4). Measurements were only made after more than 2 min of washoff of BAPTA-AM. We used the muscarinic and purinergic
agonists acetylcholine chloride and adenosine (Sigma). A
10
2 M stock solution of each drug was prepared in
double-distilled water; this was added to the Tyrode solution in
appropriate quantities. The pH of Tyrode solution containing a drug was
checked and readjusted if necessary.
Miniature solenoid valves (Lee Products; Gerrards Cross,
Buckinghamshire, UK) controlled which of the four solutions flowed to
the chamber. The fluid level in the chamber was controlled using the system described by Cannell and Lederer (8). The temperature of the solution was maintained at 37 ± 0.5°C by a heating coil wrapped around the glass inflow tube immediately before
the chamber. Solution temperature was monitored by a thermistor mounted
in the side of the chamber and controlled by a feedback circuit, which
regulated the flow of current to the heating coil.
Ventricular cells were impaled with conventional microelectrodes
(15-30 M
) filled with 1 M KCl. The switch clamp technique was
used to voltage clamp cells (Dagan 8800 amplifier, Dagan; switching
frequency, ~4.5 kHz) to minimize intracellular dialysis and allow
stable recordings of up to and over 1 h. The capacitance of the
microelectrode was compensated just before impalement. The electrode
potential upon switching between current injection and voltage
recording was monitored on an oscilloscope throughout an experiment to
ensure that the electrode potential settled between current injection
pulses. Membrane current was typically filtered at 1 kHz (low-pass
filter). Membrane potential and current were also displayed on an
oscilloscope (Tektronix 5000 series) to assess the quality of the
voltage clamp. Timing units (Hi-Med; Reading, UK) were used to control
voltage-clamp protocols and synchronize other recording equipment. A
pulse generator (3) was used to drive the action
potential-clamp experiments. Briefly, action potentials at a
stimulation rate of 1 Hz were recorded from a cell, and an average of
five were transferred to the pulse generator as a text file. The pulse
generator was then used to apply the action potential voltage-clamp
command waveform (via the Dagan 8800 amplifier) repetitively at a rate
of 1 Hz.
Cell length was recorded using an optical system based on a photodiode
array (7); twitch shortening (the shortening of a cell
during a twitch contraction) was measured electronically (6). Action potential duration was measured electronically at 
65 mV (22). Membrane current and potential, cell
length, twitch shortening, and action potential duration were displayed on a six-channel chart recorder (Gould 2600S or RS 3600) and
simultaneously recorded on videotape using a pulse code modulator
(Neuro-Corder DR-890, Neuro Data Instruments) and a video recorder.
Membrane current and potential and cell length were also digitized
using an analog-to-digital converter (1401 plus, Cambridge Electronic Design; Cambridge, UK) and stored on an IBM-compatible computer running
the Cambridge Electronic Design voltage-clamp software.
Cells were stimulated to contract by either a 5-ms current pulse
(current clamp mode) or application of the action potential clamp, at 1 Hz.
Data are presented as means ± SE (number of cells).
Computer modeling.
Computer simulations were carried out using Version 7 of the Oxsoft
HEART program (33). The "simplified guinea pig
ventricular cell" model was modified to simulate the ferret
ventricular action potential. The modifications made were the
following: extracellular K+ concentration set to 5.4 mM
(same as experiments); conductance for tetrodotoxin-sensitive
Na+ current (INa), GNA = 1.05;
Ca2+ permeability for L-type Ca2+ current
(ICa,L), PCA = 0.07; conductance for
transient outward K+ current (Ito),
GTO = 0.03; speed of inactivation variable for Ito, SPEED[18] = 6; maximum
outward K+ current for IK, IKM = 0.8; conductance for background Na+ current
(Ib,Na) GBNA = 0.0012; voltage shift of the
voltage dependence of the background IK,1,
SHIFTK1 = 10; steepness of the voltage dependence of
IK,1, and STEEPK1 = 2; and conductance for
IK,1, GK1 = 1.
cDNA cloning and sequencing.
Standard molecular biology techniques were carried out as described by
Berger and Kimmel (2) and Sambrook et al.
(35). Oligonucleotide primers 1 (bases 1240 to 1263) and 2 (bases 1501 to 1533) were designed by comparison with known Kir3.1
sequences (26): sense primer 1, 5'-AAACTCCTGAGGATGAGTTCT-3'; antisense primer 2, 5'-TGAAATAATGCCTCGAGGGGTGTTTTGCTATGT-3'.
Total RNA was purified from the ferret ventricle and atrium
(10). Total RNA was also purified from the dog ventricle
(obtained at the end of an experiment carried out by a colleague). Poly (A+) RNA, purified from total RNA by
oligo(dT)-cellulose affinity chromatography (35), was
used to generate cDNA by the rapid amplification of cDNA ends [RACE
technique (14)] using a Marathon cDNA amplification kit
according to the manufacturer's instructions (Clontech Laboratories;
Palo Alto, CA).
PCR amplification of ~500 pg cDNA (55°C annealing temperature, 50 cycles, Taq polymerase), using oligonucleotides 1 and 2 as
primers, generated a single reproducible product when visualized under
ultraviolet light after electrophoresis on a 1.5% (wt/vol) agarose
gel, containing 50 µg/ml ethidium bromide, for 1 h at 5 V/cm in
40 mM Tris-acetate and 1 mM EDTA buffer (see Fig. 12). No products were
generated when PCR was performed in the absence of template cDNA (data
not shown). The size of the PCR product was estimated by comparison to
fragments of DNA of known size (
-HindIII markers;
GIBCO-BRL; Paisley, UK). The PCR product obtained from the ferret
ventricle was cloned into pCR2.1 T/A cloning vector (Invitrogen;
Groningen, The Netherlands) and then sequenced on an Applied Biosystems
373 DNA Sequencer (Perkin Elmer) using fluorescent cycle sequencing.
Immunocytochemistry.
Immunocytochemistry on ferret ventricular and atrial cells was carried
out as previously described (12). After isolation, heart
cells were plated on Bunsen burner flame-treated, polysine-coated slides (BDH; Poole, UK) and were single labeled with either anti-Kir3.1 (Alomone Labs; Jerusalem, Israel) or anti-Kir3.4 (gift from Dr. G. B. Krapivinsky, Harvard Medical School, Boston, MA) primary antibodies
followed by an anti-rabbit secondary antibody conjugated to FITC and
examined by a confocal laser scanning microscope (Leica; Heidelberg,
Germany) as previously described (12). The images recorded
were single optical sections of 10 atrial cells and 6 ventricular cells
for each antibody used. The anti-Kir3.1 antibody was previously
characterized by immunocytochemistry on Chinese hamster ovary and
single rat heart cells (12), whereas the anti-Kir3.4 antibody was previously characterized by Western blotting on Sf9 cells (25). Various control experiments were
carried out with these antibodies and are fully described
(12). In this study, no labeling of heart cells was
detected when the secondary antibody was applied only (i.e., primary
antibodies not applied; not shown).
| |
RESULTS |
Effect of ACh on the action potential and contraction under current
clamp conditions.
Typical effects of 1 µM ACh on a ferret ventricular cell under
current-clamp conditions are shown in Fig.
1A. The cell was stimulated at
1 Hz. Slow time-base records of contractions and action potential
duration are shown on the left. On application of ACh during
the period shown by the bar, there was an abrupt shortening of the
action potential by 66% accompanied by a decrease in the contraction
of 55%. Fast time-base records of action potentials and contractions
under control conditions and at the time of the peak effect of ACh are
shown on the right of Fig. 1A. In five cells,
contraction was maximally decreased by 75 ± 4% by 1 µM ACh
under current-clamp conditions (see Fig. 2, trace i). After a peak was reached, the effects of ACh faded with time during the
remainder of the 2-min exposure to ACh; the duration of the action
potential and the amplitude of the contraction both increased (Fig.
1A). This is attributed to desensitization to ACh. On
washoff of ACh, action potential duration and contraction rapidly
increased; action potential duration returned to its control value,
whereas there was a small rebound increase in the contraction beyond
its control value (Fig. 1A). The contraction declined to its
control value over ~2 min (not shown).
View larger version (62K):
[in this window]
[in a new window]
|
Fig. 1.
Effect of 1 µM acetylcholine (ACh) on the action potential (AP)
duration (APD) and contraction under current-clamp (A) and
AP-clamp (B) conditions. Left, slow time base
records of cell length or twitch shortening (top) and APD
(bottom). ACh (1 µM) was applied for 2 min as indicated by
bar. Right, fast time base records of the AP or AP command
waveform (top) and accompanying contractions
(bottom) before (trace i) and during
(trace ii) the application of ACh. Fast time base
records correspond to the times shown in left. Data in
A and B were obtained from the same cell. The
cell was under current-clamp conditions in A. In
B, the cell was under voltage-clamp control and a control AP
recorded earlier from the same cell under current-clamp conditions was
used as the voltage-clamp command waveform.
|
|
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 2.
Inotropic effects of 1 µM ACh and 100 µM adenosine
under different conditions. Filled and hatched bars represent ACh and
adenosine data, respectively. Bars i and iv show
the minimum amplitude of contraction during an application of 1 µM
ACh or 100 µM adenosine under current-clamp conditions. Bars
ii and v show the amplitude of contraction when a short
action potential recorded in the presence of the agonist was used as
the voltage-clamp command waveform. Contraction amplitude was measured
after application of the AP command waveform at the same time the
minimum contraction occurred in the presence of ACh under current
clamp. Bars iii and vi show the minimum amplitude
of contraction during an application of 1 µM ACh or 100 µM
adenosine under AP clamp (APClamp) conditions. Values are means + SE of the amplitude of contraction (as a percentage of the control)
shown. Parentheses represent number of cells.
|
|
Effect of ACh under action potential-clamp conditions.
Typical effects of 1 µM ACh on the contraction of a ferret
ventricular cell under action potential-clamp conditions are shown in
Fig. 1B (all data in Fig. 1 were obtained from the same
cell). Before the experiment shown, the cell was stimulated at 1 Hz
under current-clamp conditions, and the steady-state action potential was recorded. Recording conditions were then switched from current clamp to voltage clamp. The control action potential was used as the
voltage-clamp command waveform. This was repetitively applied to the
cell at a rate of 1 Hz. Application of ACh under these conditions, of
course, did not result in a shortening of the action potential, and the
decrease in contraction (16%) was much reduced compared with that
observed under current-clamp conditions (cf. Fig. 1, A and
B). Figure 2 shows that in
five cells, 1 µM ACh maximally decreased contraction by 7 ± 3%
under action potential-clamp conditions (Fig. 2, bar iii),
whereas under current-clamp conditions it decreased contraction by
75 ± 4% (Fig. 2, bar i; P < 0.001, paired t-test). This much reduced the inotropic effect of
ACh under action potential-clamp conditions suggests that the inotropic effect under current-clamp conditions (Fig. 1A) is primarily
the result of the action potential shortening caused by ACh. Notice that the small effect of ACh on contraction under action
potential-clamp conditions faded with time as a result of
desensitization (Fig. 1B).
The changes in the membrane current during the experiment in Fig.
1B are shown in Fig. 3. The
voltage-clamp command waveform (the control action potential) is again
shown in Fig. 3A. As expected, under control conditions,
little or no current was recorded during the action potential command
waveform (Fig. 3B). On application of 1 µM ACh, an outward
current was activated during the action potential. Four seconds after
the start of the application of ACh, the current was ~2 nA in
amplitude in this example, but 60 s after the start, the current
had declined to ~1 nA as a result of desensitization (Fig.
3B). The ACh-activated outward current explains the
ACh-induced shortening of the action potential under current-clamp
conditions. The current-voltage relationship of the ACh-activated
outward current was obtained by plotting the current against the
membrane potential during the action potential command waveform. Figure
3C shows current-voltage relationships 4 and 60 s after
the start of the ACh application. The current-voltage relationships are
those expected of an inward-rectifying K+ current such as
IK,ACh (see DISCUSSION). Similar
data were obtained from 14 cells.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 3.
Effect of 1 µM ACh on membrane current during the control AP
(voltage-clamp command waveform). A: voltage-clamp command
waveform. AP used as the voltage-clamp command waveform was recorded
from the same cell under current-clamp conditions. B:
membrane currents recorded during the control action potential command
waveform under control conditions (Con) and after 4 and 60 s
exposure to 1 µM ACh. C: current-voltage relationship of
the ACh-activated outward current after 4 and 60 s exposure to
ACh. Currents in B were plotted against the membrane
potential during the AP command waveform. Arrow indicates the time
sequence of the data.
|
|
Simulation of the effect of ACh.
If the inotropic effect of ACh is primarily the result of the action
potential shortening, then it should be possible to simulate the
inotropic effect of ACh by using the shortened action potential recorded in the presence of ACh as the voltage-clamp command waveform. Figures 1 and 4 were obtained from the
same cell. Before the experiment shown in Fig. 4, the experiment in
Fig. 1A was carried out: the cell was stimulated at 1 Hz
under current-clamp conditions, 1 µM ACh was applied, and the
shortest action potential in the presence of ACh was recorded.
ACh was washed off, and the preparation was allowed to return to steady
state. The experiment in Fig. 4 was then started. Recording conditions
were switched from current clamp to voltage clamp, and the short action
potential recorded previously in the presence of ACh was used as the
voltage-clamp command waveform and repetitively applied to the cell at
1 Hz. Despite the fact that ACh was absent, there was a decrease in contraction comparable to that observed on application of ACh under
current-clamp conditions. In five cells, application of the short ACh
action potential command waveform resulted in a 74 ± 3% decrease
in contraction (Fig. 2, bar ii), which is not significantly
different (P = 0.526, paired t-test) from
the decrease in contraction (75 ± 4%; Fig. 2, bar i)
when ACh was applied under current-clamp conditions. This confirms that
the inotropic effect of ACh is primarily the result of the action
potential shortening rather than another action such as a decrease of
ICa,L. However, the inotropic effect of the ACh
action potential command waveform was not identical to that of
ACh; during an exposure to ACh under current-clamp conditions, the
inotropic effect faded as a result of desensitization (Fig.
1A), whereas during the application of the ACh action
potential command waveform, the inotropic effect did not fade (Fig. 4).
View larger version (40K):
[in this window]
[in a new window]
|
Fig. 4.
Simulation of the inotropic effect of ACh. A:
slow time base records of membrane potential (top), cell
length (middle), and APD (bottom). Cell was
switched from current clamp to AP clamp for 2 min as indicated by the
bar. The shortest action potential recorded earlier under current-clamp
conditions in the presence of 1 µM ACh (Fig. 1A) was used
as the voltage-clamp command waveform. B: fast time base
records of the control AP (trace i) and the ACh
AP command waveform (trace ii) (top)
and accompanying contractions (bottom).
|
|
Changes in membrane current during the experiment in Fig. 4 are shown
in Fig. 5. Figure 5A shows the
control action potential (current-clamp conditions) and the ACh action
potential command waveform and accompanying contractions. When the ACh
action potential command waveform was first applied, there was a slight
decrease in the amplitude of the contraction as a result of a decrease in the time to peak of the contraction (Fig. 5A). With time,
the contraction then declined to a steady state (Fig. 5A;
see also Fig. 4). Figure 5B shows the first and steady-state
membrane currents recorded during the ACh action potential command
waveform. When the action potential clamp was first applied, a
prolonged inward current was recorded. Once a steady state was reached
during the application of the action potential clamp, the inward
current was much abbreviated. The difference between the two currents is shown in Fig. 5C: there is a small transient outward
current followed by a large inward current. A similar result was
obtained from 11 cells. A substantial component of the inward current
during the ACh action potential command waveform is likely to be inward INaCa triggered by the Ca2+
transient (see DISCUSSION). The decrease in the inward
current during the repetitive application of the ACh action potential command waveform in Fig. 5B (responsible for the inward
difference current in Fig. 5C) might be the result of a
decline in inward INaCa caused by a decline in
the underlying Ca2+ transient (presumably responsible for
the decrease in the contraction).
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 5.
Membrane current during the short ACh action potential
(voltage-clamp command waveform). A: fast time base records
of the control AP under current-clamp conditions (Con) and the ACh AP
command waveform under APClamp conditions (top) and
accompanying contractions (bottom). Contractions accompanied
the control action potential (Con), the first ACh AP command waveform
(1st), and the ACh AP command waveform once contraction had reached a
steady state (ss). B: currents during the first ACh AP
command waveform (1st) and the ACh AP command waveform once contraction
had reached a steady state (ss). C: difference ( ) in
current between the 1st and ss ACh AP command waveforms. Difference
current was calculated by subtracting the current at ss from the
current during the 1st ACh AP command waveform.
|
|
Figure 6 shows how we tested this
hypothesis. We have shown that buffering intracellular Ca2+
by BAPTA-AM results in the loss of contraction (and, presumably, the
underlying Ca2+ transient) as well as the loss of a
Ito during the ferret ventricular action
potential; the current displays the characteristics expected of inward
INaCa triggered by the Ca2+
transient (20). If the hypothesis above is correct, then
buffering intracellular Ca2+ by BAPTA-AM by eliminating the
Ca2+ transient should abolish the inward
INaCa during the ACh action potential command
waveform as well as the difference in the first and steady-state
currents. The result in Fig. 6A, obtained under control
conditions, is similar to the result in Fig. 5; it shows the membrane
current during the ACh action potential command waveform. Once again,
the current that flowed during the first ACh action potential command
waveform was more inward than the current that flowed under
steady-state conditions. This is clear from the difference current
shown in Fig. 6A, bottom. After this result was
obtained, the cell was exposed to 5 µM BAPTA-AM for 7 min to buffer
intracellular Ca2+. After the application and washoff of
BAPTA-AM (see MATERIALS AND METHODS), the action potential
under current-clamp conditions was shortened (Fig. 6B). This
effect has been described previously and has been attributed to the
loss of inward INaCa (20). After the application and washoff of BAPTA-AM, the ACh action potential command waveform was again applied (Fig. 6B). As predicted,
buffering intracellular Ca2+ reduced the inward
current flowing during the ACh action potential command waveform (Fig.
6B, middle). The change in current on repetitive application of the ACh action potential command waveform was also abolished (Fig. 6B, middle), and, as a result,
the inward difference current was abolished (Fig. 6B,
bottom). Similar results were obtained from four cells. The
ACh-activated outward current (Fig. 3B) must overcome the
inward current in Figs. 5B and 6A to be able to
shorten the action potential.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 6.
Effect of buffering intracellular Ca2+ on the inward
current during the short ACh AP (voltage-clamp command waveform).
Top, fast time base records of an AP during current clamp
under Con conditions (A) or after the application of
1,2-bis(2-aminophenoxy) ethane-N,N,N',N'-tetraacetic
acid-AM (BAPTA) (B) and the ACh AP command waveform during
voltage clamp (APClamp, APClp). Middle, currents
during the 1st ACh AP command waveform and the ACh AP command waveform
once contraction had reached a steady state or at the equivalent time
after the application of BAPTA-AM (ss). Bottom, difference
in current between the 1st and ss ACh AP command waveforms. Difference
current was calculated by subtracting the current at ss from the
current during the first ACh AP command waveform. Records were obtained
before (A) and after (B) the buffering of
intracellular Ca2+ by the application and washoff of 5 µM
BAPTA-AM.
|
|
In the experiments shown in Figs. 5 and 6A, repetitive
application of the ACh action potential command waveform resulted in the loss of a Ito; this is best seen in the
difference currents in Fig. 5C and the bottom panel of Fig.
6A. Figure 6B, bottom, shows that this
Ito was reduced after the application of
BAPTA-AM (see DISCUSSION for interpretation). Similar
results were obtained from two cells.
Activation of IK,1 by ACh.
The shortening of the action potential in the presence of ACh is
attributable to the ACh-activated outward current shown in Fig.
3B. The current-voltage relationship of this current
suggests that it is an IK,1. To test this
possibility, we investigated the effects of Ba2+, a potent
blocker of inward-rectifying K+ channels. First, the effect
of Ba2+ alone was investigated. Figure
7A shows the effect of 2 mM
Ba2+ on a cell under current-clamp conditions. Under
control conditions, a stimulus resulted in a normal action potential.
After the application of Ba2+, there was a large
depolarization. The cell was no longer excitable, and instead there
were small spontaneous oscillations of the membrane potential
(accompanied by contractions; not shown). Figure 7B shows a
control action potential command waveform that was repetitively applied
to the same cell at a rate of 1 Hz under voltage-clamp control.
Application of Ba2+ under these conditions, of course, did
not result in a depolarization of the resting membrane. Instead it
resulted in the loss of an outward current. The
Ba2+-sensitive outward current was obtained by subtracting
the current after the application of Ba2+ from the current
before the application of Ba2+ and is shown in Fig.
7B, bottom. The current-voltage relationship of
this current was obtained by plotting the current against the membrane
potential during the control action potential command waveform and is
shown in Fig. 7C. The current displays inward rectification
more pronounced than that of the ACh-activated current (Fig.
3C). The current is a combination of all
Ba2+-sensitive currents but is likely to be primarily the
background IK,1. Similar results were obtained
from five cells.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 7.
Effect of 2 mM Ba2+ on the AP and membrane
current. A: AP under control conditions and membrane
potential after application of 2 mM Ba2+. B:
control AP recorded under current-clamp conditions used as the APClamp
waveform (top) and accompanying current during application
of 2 mM Ba2+ (bottom; current during the AP
clamp in the presence of Ba2+ was subtracted from the
current during the AP clamp before the application of
Ba2+). C: current-voltage relationship of the
Ba2+-sensitive current. Ba2+-sensitive current
was plotted against the membrane potential during the APClamp
waveform.
|
|
Figure 8 shows the effect of
Ba2+ on the ACh-activated outward current. Under action
potential-clamp conditions, application of ACh resulted in the usual
development of an outward current during the control action potential
command waveform (Fig. 8A, top). The
current-voltage relationship of this current is shown in Fig.
8B; the current shows the same degree of inward
rectification as the current in Fig. 3C. To investigate
whether this current could be blocked by Ba2+, we first had
to record the membrane current in the presence of Ba2+
alone (see above) and then in the presence of ACh and Ba2+.
The trace in Fig. 8A, bottom, shows the membrane
current activated by ACh in the presence of Ba2+. This was
calculated by subtracting the current in the presence of
Ba2+ alone from the current in the presence of ACh and
Ba2+. Figure 8B shows the current-voltage
relationship of the ACh-activated current in the presence of
Ba2+. Ba2+ caused a large reduction in the
ACh-activated current (Fig. 8). Similar results were obtained from five
cells. The results support the hypothesis that ACh activates an
IK,1.
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 8.
Effect of 2 mM Ba2+ on the ACh-activated
outward current during the control AP (voltage-clamp command waveform).
A: control AP recorded under current-clamp conditions used
as the voltage-clamp command waveform (top), ACh-activated
current (middle; current in the presence of 10 µM ACh was
subtracted from the current before the application of ACh), and
ACh-activated current in the presence of 2 mM Ba2+
(bottom; current in the presence of Ba2+ only
was subtracted from that in the presence of 10 µM ACh and
Ba2+). B, current-voltage relationship of the
ACh-activated current under normal conditions (ACh) and in the presence
of 2 mM Ba2+ (ACh + Ba2+). Currents in
A were plotted against the membrane potential during the
APClamp waveform.
|
|
Inhibition of IK,1 by ACh.
To confirm that the ACh-activated outward current is a
K+ current, we attempted to measure its reversal potential.
This was not possible using the action potential-clamp technique,
because the membrane potential never exceeds the K+
equilibrium potential (EK) during the action
potential. Instead, we used a ramp-clamp protocol that ramped the
membrane potential from +33 to 120 mV (below
EK) over 200 ms (roughly similar to the duration
of the action potential) at a rate of 1 Hz (holding potential, 80
mV). Current-voltage relationships were obtained by plotting membrane
current against the membrane potential during the ramp clamp. Figure
9A shows the current-voltage
relationship of total membrane current before and early and late during
an application of ACh. ACh had the expected effect on membrane current at potentials more positive than EK; it caused
an outward shift in current that faded with time as a result of
desensitization as in the previous series of action potential-clamp
experiments. ACh also had little effect on current near
EK. However, at potentials more negative than
EK, early during the application of ACh there was an inward shift in current, but later during the application an
outward shift in current occurred. The two different effects of ACh can
be more clearly seen in Fig. 9B, which shows current-voltage relationships of the ACh-dependent current. The ACh-dependent current
was calculated by subtracting the current under control conditions from
the current in the presence of ACh. Early during the application of
ACh, the current-voltage relationship of the ACh-dependent current is
characteristic of an inward-rectifying K+ current such as
IK,1, and the current reversed close to the
suspected value of EK (see
DISCUSSION). However, later during the application of ACh
the current-voltage relationship was greatly altered, and the
ACh-dependent current was outward at potentials both more negative and
more positive than EK.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 9.
Effect of 10 µM ACh on membrane current during ramp clamps.
A: current-voltage relationships of total membrane current
obtained from ramp clamps before (Con) and early (2 s) and late (>15
s) during an application of 10 µM ACh. B: current-voltage
relationships of the ACh-dependent current early and late during the
application of ACh. ACh-dependent current was obtained by subtracting
current under control conditions from current after the application of
ACh. C and D: same as A and
B, but from a different cell. In addition, D
shows the current-voltage relationship of ACh-dependent current in the
presence of 2 mM Ba2+. Current in the presence of
Ba2+ only was subtracted from current in the presence of
ACh and Ba2+.
|
|
In some cells, ACh caused an outward shift of current at potentials
more negative than EK both early and late during
the ACh application. An example is shown in Fig. 9C. At
potentials above EK, ACh caused the expected
outward shift in current that faded with time. At potentials below
EK, ACh also caused an outward shift in current;
the outward shift in current increased with time. Current-voltage
relationships for the ACh-dependent current are shown in Fig.
9D; the current-voltage relationship had a U-shape both
early and late during the ACh application. Figure 9D also shows a current-voltage relationship for ACh-dependent current in the
presence of Ba2+. The cell was exposed to 2 mM
Ba2+ and subsequently to ACh and Ba2+. The
current shown was obtained by subtracting current in the presence of
Ba2+ only from current after the application of ACh and
Ba2+. Ba2+ abolished the ACh-dependent outward
current at potentials both negative and positive to
EK. The U-shape of the current-voltage relationship of ACh-dependent current was observed in four cells. Ba2+ block of the current was observed in four cells.
A possible explanation of the U-shape of the current-voltage
relationship of the ACh-dependent current is that ACh, as well as
activating IK,ACh, inhibits
IK,1. This would explain why ACh has little
effect on the membrane current close to EK.
According to this hypothesis, the outward shift of current at
potentials positive to EK is the activation of
IK,ACh, whereas the outward shift at potentials
negative to EK is the inhibition of
IK,1. Why the activation of
IK,ACh should dominate at potentials positive to
EK and the inhibition of
IK,1 should dominate at potentials negative to
EK is considered below.
Computer model of the activation of IK,ACh and
inhibition of IK,1 by ACh.
The consequences of the simultaneous activation of
IK,ACh and inhibition of
IK,1 by ACh were investigated in a modified
Oxsoft HEART model of the ventricular action potential (see
MATERIALS AND METHODS for details). Figure
10A shows calculated
current-voltage relationships for IK,ACh, and
Fig. 10B shows calculated current-voltage relationships for
IK,1; on application of ACh, it was assumed that
IK,ACh was activated (maximum current at
positive potentials, 1 nA) and IK,1 was
inhibited by 75%. Figure 10C shows calculated current-voltage relationships for IK,ACh plus
IK,1. These are assumed to be roughly equivalent
to the current-voltage relationships in Fig. 9, A and
C. As observed experimentally, on application of ACh, the
current was shifted in the outwards direction at potentials both
positive and negative to EK. The current-voltage
relationship of the ACh-dependent current had the characteristic
U-shape (cf. Figs. 10D and 9, B and
D). The reason why the activation of
IK,ACh dominates at potentials positive to
EK and the inhibition of
IK,1 dominates at potentials negative to
EK is further considered in the
DISCUSSION.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 10.
Computer simulation of the action of ACh on the ferret ventricular
AP incorporating activation of muscarinic K+ current
(IK,ACh) and inhibition of background inward
rectifier K+ current (IK,1).
A and B, current-voltage relationships for
IK,ACh (A) and
IK,1 (B) under control conditions and
after the application of ACh. Data were calculated using equations
taken from the Oxsoft HEART model of the ventricular AP. ACh was
assumed to activate IK,ACh (maximum current, 1 nA) and decrease IK,1 by 75%. C:
current-voltage relationships for "total" membrane current
(IK,1 plus IK,ACh) under
Con conditions and after the application of ACh. Total current was
calculated as the sum of the individual currents in A and
B. D: current-voltage relationship of the
ACh-dependent current. Current under Con conditions was subtracted from
current after the application of ACh. E: membrane potential
recorded from a ferret ventricular cell before (trace i) and
after (traces ii and iii) the application of 1 µM ACh. Stimulus current is shown below. AP are shown in traces
i and iii, but the stimulus-evoked depolarization only
is shown in trace ii. F: membrane potential from
a modified version of the Oxsoft HEART model of the ventricular AP
before (trace i) and after (traces ii
and iii) the application of ACh. Stimulus current is shown
below. Action potentials are shown in traces i and
iii, but the stimulus-evoked depolarization only is shown in
trace ii. In trace ii
IK,ACh was activated only (maximum current, 1 nA), whereas in trace iii IK,1 was
inhibited by 75% in addition to the activation of
IK,ACh.
|
|
A clue concerning the possible physiological importance of the
inhibition of IK,1 is provided by the
observation in Fig. 10E. Figure 10E shows action
potentials recorded from a ferret ventricular cell. Although under
control conditions a 1.5 nA stimulus was sufficient to trigger an
action potential (Fig. 10E, trace i), after
application of 1 µM ACh the stimulus was no longer sufficient and a
stimulus-evoked depolarization (Fig. 10E, trace
ii) only was seen. When the stimulus was increased to 1.8 nA, an
action potential (shortened compared with the control) was again
triggered (Fig. 10E, trace iii). This observation, which was
made in other cells, demonstrates that in the presence of ACh there was
a decrease in excitability. Figure 10F shows simulated
action potentials from the modified Oxsoft HEART model. In Fig.
10F, a 1.55 nA stimulus was used to trigger a control
action potential (Fig. 10F, trace i). After activation of
IK,ACh (maximum current, 1 nA), the stimulus was
no longer suprathreshold and a stimulus-evoked depolarization (Fig.
10F, trace ii) only was seen. This demonstrates that, in the
model at least, it is the activation of IK,ACh
that depresses excitability (in the model, the stimulus threshold was
increased from 1.15 to 1.80 nA after activation of
IK,ACh). After the activation of
IK,ACh, if the stimulus was increased to 1.80
nA, an action potential was again triggered, but it was short (not shown). Figure 10F, trace iii, shows that when
IK,1 was decreased by 75%
(IK,ACh was still activated), the 1.55 nA
stimulus was again able to trigger an action potential. Thus the
simultaneous inhibition of IK,1 was able to
restore excitability (the stimulus threshold fell to 1.55 nA). In
conclusion, it is proposed that the activation of
IK,ACh depresses excitability of ventricular cells. This is presumably undesirable. It is suggested that the effect
of ACh on excitability is blunted as a result of the twin actions of
ACh on IK,ACh and IK,1.
By the way, in the model the activation of
IK,ACh resulted in the shortening of the action potential (Fig. 10F), and the simultaneous inhibition of
IK,1 had little antagonistic effect on the
shortening. This proposal is considered further in the
DISCUSSION.
Adenosine.
In addition to activation by ACh, IK,ACh is
known to be activated by other agonists, including adenosine
(4). Adenosine has a negative inotropic effect on the
ferret ventricle (16). To test whether the conclusions of
the present study are restricted to ACh, we investigated the inotropic
effect of 100 µM adenosine using the action potential-clamp
technique. Figure 11 shows three experiments carried out on the same ferret ventricular cell. Figure 11A shows the typical effect of 100 µM adenosine under
current-clamp conditions. On application of adenosine there was a
shortening of the action potential accompanied by a decrease in
contraction, which can be seen in both the slow and fast time base
records. In eight cells, 100 µM adenosine caused a maximal reduction
of 42 ± 4% in contraction under current-clamp conditions (Fig.
2, bar iv). Like that of ACh, the effect of adenosine faded
with time (Fig. 11A). Figure 11B shows an
experiment in which a control action potential, recorded under
current-clamp conditions, was used as the voltage-clamp command
waveform. In this case, application of adenosine had no observable
effect on contraction. In four cells under action potential-clamp
conditions, adenosine caused a 5 ± 4% reduction in contraction
(Fig. 2, bar vi). The current-voltage relationship of the adenosine-activated current was the same shape as
the ACh-activated current (Fig. 3C) and as expected of
IK,ACh (not shown). The negative inotropic
effect of adenosine was simulated in the same way as before for ACh
(Fig. 4). Figure 11C shows that switching from current-clamp
to voltage-clamp mode, using the shortest action potential under
current-clamp conditions in the presence of adenosine (Fig.
11A) as the voltage-clamp command waveform, caused a
decrease in contraction comparable to the decrease in Fig.
11A (once again there was no fade in the effect). In five cells, application of the shortened action potential command waveform caused a 45 ± 5% reduction in contraction (Fig. 2,
bar v).
View larger version (55K):
[in this window]
[in a new window]
|
Fig. 11.
Effect of 100 µM adenosine. Left: slow time base
records of cell length (top) and APD (bottom).
Right: fast time base records of AP s or action potential
command waveforms (top) and accompanying contractions
(bottom) at the times indicated in left. In
A, the cell was under current clamp and 100 µM adenosine
was applied for 2 min as indicated by the bar. In B, the
cell was under APClamp control throughout. A control action potential
recorded under current-clamp conditions was used as the APClamp command
waveform. Adenosine (100 µM) was applied for 2 min as indicated by
bar. In C, control was switched from current clamp to
voltage clamp for 2 min as indicated by the bar. Shortest AP in the
presence of adenosine (under current-clamp conditions, labelled
trace ii in A) was used as the APClamp command
waveform. All data are from the same cell.
|
|
Detection of Kir3.1 mRNA in ventricle.
The electrophysiological experiments above suggest that the inotropic
effect of ACh (and adenosine) on ferret ventricular cells is
principally the result of the activation of
IK,ACh. The muscarinic K+ channel is
known to be a heteromultimer of Kir 3.1 and Kir 3.4. Northern blot
analysis of Kir3.1 mRNA expression in the rat and guinea pig heart has
previously shown expression of Kir3.1 in the atria in the two species,
but not in the ventricles (26). Because of this conflict,
we have reexamined the issue of whether there is expression of Kir3.1
in the ferret ventricle. Total RNA from the ferret atrium, ferret
ventricle, and dog ventricle was purified to yield mRNA from which cDNA
was generated by the RACE technique. With use of Kir3.1-specific
oligonucleotide primers (see MATERIALS AND METHODS), the
Kir3.1 sequence was amplified by PCR. Figure
12E shows PCR products
subjected to agarose gel electrophoresis and shows that Kir3.1 was
detected in the ferret ventricle as well as the atrium. Kir3.1 was also
detected in the dog ventricle (Fig. 12E). We have previously
obtained electrophysiological evidence of the presence of
IK,ACh in the dog ventricle (39). In the case of the ferret ventricle, the PCR product was cloned into
pCR2.1 vector, and the sequence was determined to confirm that it was
Kir3.1.
View larger version (50K):
[in this window]
[in a new window]
|
Fig. 12.
Detection of Kir3.1 and Kir3.4 protein by immunocytochemistry and
confocal microscopy and detection of Kir3.1 PCR products by agarose gel
electrophoresis. A-D: Kir3.1 (A,
C) and Kir3.4 (B, D) protein labeling
in ventricular (A, B) and atrial (C,
D) cells. A-D are at same scale
(D, 56 µm wide). E: Kir3.1 sequences amplified
by PCR from cDNA obtained from various cardiac tissues: lanes
1 and 2, ferret atrium; lanes 3 and
4, ferret ventricle; lanes 5 and 6,
dog ventricle.
|
|
Detection of Kir3.1 and Kir3.4 proteins in ventricular and atrial
cells.
The muscarinic K+ channel protein Kir3.1 was not detected
by immunocytochemistry in tissue sections through the ferret ventricle (Kir3.4 not tested). Immunocytochemistry is more sensitive when applied
to isolated cells compared with tissue sections, because in isolated
cells there is better access of the antibodies to the antigenic sites.
Figure 12, A and B, shows ventricular cells labeled for Kir3.1 or Kir3.4. Both Kir3.1 (Fig. 12A) and
Kir3.4 (Fig. 12B) labeling was observed (although it was
weak) and was absent from the outer cell membrane and was present in
the t tubules. Similar results were obtained from six Kir3.1 and seven
Kir3.4 ventricular cells. Kir3.1 was detected in tissue sections
through the ferret atrium. Figure 12, C and D,
shows atrial cells labeled for Kir3.1 or Kir3.4. Both Kir3.1 (Fig.
12C) and Kir3.4 (Fig. 12D) labeling were strong
and present in the outer cell membrane (including the intercalated
disks). Similar results were obtained from 10 Kir3.1 and 10 Kir3.4
atrial cells. In summary, the immunocytochemical data show low
expression of both Kir3.1 and Kir3.4 proteins in the ventricle and high
expression of both Kir3.1 and Kir3.4 proteins in the atrium.
| |
DISCUSSION |
This study has proven that the negative inotropic effects of ACh
and adenosine are the result of the action potential shortening in the
presence of the agonists. It has also suggested that ACh activates
IK,ACh and inhibits INaCa
and IK,1. The first two effects cause the action
potential shortening, whereas the third may preserve cell excitability.
Presence of Kir3.1 in the ventricle.
This study and previous studies (18, 24, 31, 39) have
provided electrophysiological evidence of IK,ACh
in ferret, rat, human, cat, guinea pig, and dog ventricular cells.
Previously, a Northern blot by Kubo et al. (26) failed to
show the presence of Kir3.1 mRNA in the rat ventricle. However, the PCR
technique is a more sensitive test because it amplifies cDNA derived
from low-abundance mRNAs. Using the PCR technique, we were able to show
the presence of Kir3.1 mRNA in the ferret and dog ventricle (Fig.
12E). However, the presence of Kir3.1 mRNA in the ferret ventricle does not prove that the Kir3.1 protein is present in ferret
ventricular cells and for this reason we examined Kir3.1 and Kir3.4
protein expression in ferret ventricular cells using immunocytochemistry. Labeling of Kir3.1 and Kir3.4 was detected in the
t tubules of ferret ventricular cells (Fig. 12). It is interesting that
other channels (
1C, Task-1, Kv4.2, Kir2.1) are also
found in the t tubules of ventricular cells (11, 21, 32,
38).
Agonist-induced changes in membrane current.
ACh activated an outward current during a control action potential
command waveform (Fig. 3B). This current may be
IK,ACh for several reasons. First, it was
activated by ACh. Second, the ACh-activated current displayed a typical
current-voltage relationship of IK,ACh- (Figs.
3C and 8B). IK,ACh shows
inward rectification, but the degree of rectification of
IK,ACh is less than that of, for example,
IK,1 (34, 41); this is illustrated
by the calculated data in Fig. 10, A and B. The
degree of rectification of the ACh-activated current (Figs.
3C and 8B) is typical of that of
IK,ACh. Third, the ACh-activated current was
blocked by 2 mM Ba2+ (Fig. 8). Ba2+ at 2 mM is
known to block IK,ACh fully (40).
Fourth, during an application of ACh, the ACh-activated outward current
faded with time (Fig. 3B). Fade during an exposure to ACh is
a characteristic feature of IK,ACh and is the
result of desensitization to ACh (e.g., Ref. 41). Finally,
the reversal potential of the ACh-activated current was close to
EK. Using the ramp clamp, in one cell at least,
we were able to observe reversal of the ACh-activated current close to
the suspected value of EK early during an
application of ACh (Fig. 9B).
In all cells studied, ACh eventually inhibited inward current at
potentials negative to EK (Fig. 9). ACh has a
similar action in dog left ventricular subepicardial cells
(39). This may be the result of inhibition of
IK,1. In support of this, the ACh-induced change
in current was abolished in the presence of 2 mM Ba2+ (Fig.
9D); IK,1 is known to be abolished
under these conditions (9). The voltage dependence of the
ACh-induced changes in membrane current (Fig. 9, B and
D) can be explained by a dual effect of ACh on
IK,ACh and IK,1. At
approximately 80 mV, there was little change in membrane current;
this is expected, because this potential is close to
EK and at this potential both
IK,ACh and IK,1 are small
and, furthermore, it is possible that at potentials near EK any effects of ACh on
IK,ACh and IK,1 are equal
and opposite. At potentials positive to IK, the
results indicate that activation of IK,ACh was
more important than inhibition of IK,1. The
likely reason for this is shown by Fig. 10,
A-D. IK,1 shows
greater inward rectification than IK,ACh over
this range of potentials. For example, at plateau potentials at ~0
mV, IK,1 may be negligible, whereas IK,ACh, although it rectifies at more positive
potentials, does not decline below a maximum value (34,
40). Therefore, at the more positive potentials a decrease in
IK,1 will have little or no effect on total
current, whereas the activation of IK,ACh will.
At potentials negative to EK, the results
indicate that the activation of IK,ACh was less
important than the inhibition of IK,1, and the
likely reason for this result is again shown by Fig. 10,
A-D. At these potentials,
IK,1 steeply increases with hyperpolarization,
and Fig. 10, A-D, suggests that it is
greater in amplitude than IK,ACh. Therefore, the
change in total current may be dominated by the decrease in
IK,1 at these potentials. Figure 10,
E and F, suggests a possible reason for the
simultaneous inhibition of IK,1 by ACh: it may
blunt the depression of excitability caused by the activation of
IK,ACh. Excitability will be determined by the
slope conductance at the resting potential (close to
EK). Figure 10A shows that the
activation of IK,ACh alone will result in a
large increase in the slope conductance and thus decrease in
excitability. When the slope of the current-voltage relationship of the
control IK,1 at IK (this
will largely determine the slope conductance under control conditions)
is compared with that of IK,ACh (after
activation of IK,ACh), the two are comparable,
which shows that activation of IK,ACh alone will
result in a large fractional increase in the slope conductance at the
resting potential. The central panel of Fig. 10F confirms
that in the model, at least, activation of
IK,ACh does result in a decrease in
excitability. On the other hand, Fig. 10C shows that the
simultaneous activation of IK,ACh and inhibition
of IK,1 results in little change in the slope
conductance near EK. In this way, the
simultaneous inhibition of IK,1 blunts the
depression of excitability; in the model, this is confirmed by Fig.
10F.
When a short ACh action potential command waveform was applied, an
inward current was recorded (Fig. 5). During an application of ACh, the
ACh-activated outward current in Fig. 3B must be able to
overcome this current to be able to shorten the action potential (the
two currents are expected to be roughly equal and opposite). Strict
comparison of the currents in Figs. 3C and 5C is
not possible, because they were recorded from different cells and
during different action potential command waveforms, but the two
currents are roughly equivalent in amplitude. It is well known that
after a short square wave voltage-clamp pulse, an inward tail current
is recorded (e.g., Ref. 15). The inward tail current is
analogous to the current in Fig. 5B. It is also well known
that a substantial fraction of the inward tail current after a square
wave voltage-clamp pulse is inward INaCa
triggered by the Ca2+ transient. Figure 6 shows that this
is also likely to be true for the inward current recorded during the
short ACh action potential command waveform: after the suppression of
the Ca2+ transient by BAPTA-AM, the inward tail current was
greatly diminished. Figures 5B and 6A show that
during repetitive application of the ACh action potential command
waveform, the inward current diminished and Fig. 6B suggests
that it was the result of the expected reduction in the underlying
Ca2+ transient.
Figures 5C and 6A show that during repetitive
application of the ACh action potential command waveform, there was a
reduction in a transient outward current. Figure 6B suggests
that this too was a consequence of the expected reduction in the
Ca2+ transient. Janvier et al. (20) showed
that the suppression of the Ca2+ transient in ferret
ventricular cells by BAPTA-AM resulted in the loss of a transient
outward current as well as a transient inward current. Whereas the
transient inward current exhibited the properties of inward
INaCa triggered by the Ca2+
transient, the transient outward current exhibited the properties of
Ca2+-activated Cl current triggered by the
Ca2+ transient. The time course and amplitude of the
transient outward current in the study of Janvier et al.
(20) is very similar to those of the transient outward
current in the present study (Figs. 5C and 6A).
It is suggested that in the present study during the application of the
ACh action potential command waveform, a reduction in the
Ca2+ transient resulted in a reduction in the
Ca2+-activated Cl current.
Figure 7 shows a Ba2+-sensitive current during the action
potential. Inward rectifying K+ currents are most sensitive
to Ba2+ and, therefore, the Ba2+-sensitive
current is likely to be IK,1. Furthermore, the
current-voltage relationship of the Ba2+-sensitive current
shows the marked inward rectification characteristic of
IK,1 (Fig. 7C) (34).
Agonist-induced changes in the action potential.
The shortening of the ferret ventricular action potential in the
presence of ACh and adenosine is likely to be primarily the result of
the activation of IK,ACh. The current-voltage
relationship of IK,ACh (Figs. 3C,
8B, and 9) indicates that over the range of potentials
during the action potential plateau, a significant outward current
flows that will tend to shorten the action potential. The fade of the
action potential shortening in the presence of both agonists (Figs.
1A and 11A) can be explained by
IK,ACh, because IK,ACh
faded with time during application of both agonists (Fig. 3 and
unpublished observation). It is likely that the density of the
muscarinic K+ channel is less in the ventricle than the
atrium, and it is interesting that despite this, the effects on the
ventricular action potential are still substantial. A reduction in
ICa,L may also explain some of the action
potential shortening, although in dog ventricular cells 10 µM ACh
only caused an 8% reduction in ICa,L at a
holding potential of 80 mV (39). The results in
Figs. 5 and 6 suggest that on application of ACh the reduction of the
Ca2+ transient underlying the reduction of the contraction
leads to an indirect reduction in inward INaCa,
and this is also expected to contribute to the shortening of the action
potential. It is already known that suppression of inward
INaCa by buffering intracellular Ca2+ by BAPTA-AM leads to a shortened action potential
(20) (see also the control action potential and the action
potential after the application of BAPTA-AM in Fig. 6). The inhibition
of IK,1 by ACh may be expected to antagonize the
shortening of the action potential caused by ACh. However, in the
model, the simultaneous inhibition of IK,1 had
little effect on the shortening of the action potential caused by
IK,ACh (Fig. 10F). This is because at plateau potentials there is little or no IK,1
(Fig. 10B).
Agonist-induced changes in contraction.
In ferret ventricular cells, the negative inotropic effect of ACh and
adenosine is likely to be primarily the result of the shortening of the
action potential. With the use a control action potential waveform as
the voltage-clamp command, the negative inotropic effect of ACh and
adenosine was almost abolished (Fig. 2). Furthermore, use of the
agonist-shortened action potential as the voltage-clamp command
waveform caused a reduction in contraction comparable to that observed
on application of the agonist under current-clamp conditions (Fig. 2).
With both agonists there could be a small reduction in contraction when
the cell was under action potential-clamp control (Fig. 2). This may be
the result of a small reduction in ICa,L by the
agonists. The agonist-induced shortening of the action
potential is expected to lead to a reduction in the Ca2+
transient and thus the contraction by reducing the Ca2+
content of the sarcoplasmic reticulum by: 1) promoting
Ca2+ efflux on the Na/Ca exchanger and 2)
reducing Ca2+ influx via ICa,L
(because of the shortened action potential, ICa,L flows for a shorter time). Direct evidence
of the stimulation of Ca2+ efflux via the Na/Ca exchanger
is the inward current measured on application of the ACh action
potential command waveform (Figs. 5B and 6A). The
evidence that this is partly composed of inward INaCa (corresponding to Ca2+ efflux)
has already been discussed.
During an application of ACh or adenosine, the slow increase in
contraction under current-clamp conditions (Figs. 1A and
11A) can be explained by fade of the action potential
shortening (primarily the consequence of the fade of
IK,ACh as a result of desensitization) or a fade
of a decrease in ICa,L
(39). The slow increase in contraction during an
application of ACh has also been suggested to be the result of a slow
increase in intracellular Na+ (23). During
action potential clamp the small inotropic effect faded with time (Fig.
1B), and this at least could not have been the result of a
change in action potential duration.
In summary, it is concluded that the negative inotropic effects of ACh
and adenosine are primarily the result of the activation of
IK,ACh and the consequent shortening of the
action potential. The reduction in the intracellular Ca2+
transient (underlying the decrease in the contraction) causes a
reduction in inward INaCa, and this also
contributes to the action potential shortening. A small reduction in
ICa,L may also play a role in the negative
inotropic effect. ACh inhibits IK,1 and
this may blunt a reduction in cell excitability following activation of
IK,ACh. Although the present study was carried
out in the ferret, the results are unlikely to be specific to the ferret, because, in ventricular cells from a range of species, IK,ACh has been recorded and ACh shown to have a
negative inotropic effect in the absence of -stimulation (5,
18, 24, 30, 31, 39).
This work was supported by grants from the British Heart Foundation and
the Wellcome Trust. N. C. Janvier was a Wellcome Prize student.
Address for reprint requests and other correspondence:
M. R. Boyett, School of Biomedical Sciences, Univ. of Leeds,
Leeds LS2 9JT, UK (E-mail:
m.r.boyett{at}leeds.ac.uk).
The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement"
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
TOP
ABSTRACT
INTRODUCTION
