Vol. 279, Issue 3, H1421-H1433, September 2000
SPECIAL COMMUNICATION
Ratiometry of transmembrane voltage-sensitive fluorescent dye
emission in hearts
Stephen B.
Knisley,
Robert K.
Justice,
Wei
Kong, and
Philip L.
Johnson
Department of Biomedical Engineering of the School of
Engineering, The University of Alabama at Birmingham, Alabama
35294-0019
 |
ABSTRACT |
Transmembrane voltage-sensitive
fluorescence measurements are limited by baseline drift that can
obscure changes in resting membrane potential and by motion artifacts
that can obscure repolarization. Voltage-dependent shift of emission
wavelengths may allow reduction of drift and motion artifacts by
emission ratiometry. We have tested this for action potentials and
potassium-induced changes in resting membrane potential in rabbit
hearts stained with di-4-ANEPPS [Pyridinium,
4-(2-(6-(dibutylamino)-2-naphthalenyl) ethenyl)-1-(3-sulfopropyl)-, hydroxide, inner salt] using laser excitation (488 nm) and a
two-photomultiplier tube system or spectrofluorometer (resolution of
500-1,000 Hz and <1 mm). Green and red emissions produced upright
and inverted action potentials, respectively. Ratios of green emission
to red emission followed action potential contours and exhibited larger fractional changes than either emission alone (P < 0.001). The largest changes and signal-to-noise ratio (signal/noise)
were obtained with numerator wavelengths of 525-550 nm and
denominator wavelengths of 650-700 nm. Ratiometry lessened drift
56-66% (P < 0.015) and indicated decreases in
resting membrane potential. Ratiometry lessened motion
artifacts and increased magnitudes of deflections representing
phase-zero depolarizations relative to total deflections by
123-188% in intact hearts (P < 0.02). Durations
of action potentials at different pacing rates, temperatures, and
potassium concentrations were independent of whether they were measured
ratiometrically or with microelectrodes (P 
0.65). The
ratiometric calibration slope was 0.017/100 mV and decreased with time.
Thus emission ratiometry lessens the effects of motion and drift and
indicates resting membrane potential changes and repolarization.
potentiometric dye; Pyridinium,
4-(2-(6-(dibutylamino)-2-naphthalenyl) ethenyl)-1-(3-sulfopropyl)-,
hydroxide, inner salt; potassium; action potential
| |
INTRODUCTION |
TRANSMEMBRANE
voltage-sensitive fluorescent dyes have been used to monitor fast
events such as action potentials and membrane polarization produced by
electric stimulation. However, the fluorescence emission and fractional
fluorescence change depend on variables, including dye concentration,
dye internalization to the inner leaflet of the membrane, dye
photobleaching, and possible changes in excitation light intensity,
which are effects that can occur slowly and are not related to
transmembrane potential. Thus fluorescent dyes may not be suitable for
measuring slow changes of the resting membrane potential such as those
produced by hyperkalemia or ischemia.
However, optical mapping studies in hearts have not used all of the
information contained in the emitted light of a dye. The emission
spectrum of the dye di-4-ANEPPS [Pyridinium
4-(2-(6-(dibutylamino)-2-naphthalenyl) ethenyl)-1-(3-sulfopropyl)-,
hydroxide, inner salt] undergoes a reversible shift toward
shorter wavelengths in response to membrane depolarization, i.e.,
electrochromism (10). Cardiac optical mapping studies have
frequently measured the decrease in emitted light for wavelengths
longer than ~600-650 nm (red portion of the emission spectrum).
Our recent measurements of di-4-ANEPPS emission spectra in hearts
indicated that with an appropriate excitation wavelength, both red and
green portions of the emission spectrum can be measured. The emitted
light for shorter wavelengths (green) increases simultaneously with the
decrease in red emission during depolarization (17).
The availability of both red and green emission signals may allow
improvements in recording of changes in the membrane potential by using
the ratio of the green and red signals in hearts, as was developed by
Bullen and Saggau (5) for measurements in hippocampal
neurons during voltage-clamp pulses. Changes in fluorescence that are
common to both signals should theoretically cancel in a ratio.
Therefore, effects of dye photobleaching and variations in dye
concentration, which should affect both red and green signals, may be lessened.
Also motion artifacts may be lessened, which was described for
transmembrane voltage changes in arterioles and heart cell cultures and
calcium signals from hearts (2, 4, 23). Motion is a
limitation of optical mapping of transmembrane potentials in the heart,
requiring mechanical stabilization, pharmacological electromechanical
uncoupling, or use of temporal derivatives of the optical signal to
detect repolarization of the action potential (6, 8, 9, 12,
13).
This study considers ratiometry of epicardial di-4-ANEPPS emission in
isolated rabbit hearts with excitation by a laser beam. Ratiometry was
performed for action potentials and resting membrane potential changes
produced by potassium infusion. We measured emission in two optical
bands and, in separate experiments, 16 bands using a spectrophotometer
to determine the impact of the selection of emission wavelengths used
for the numerator and denominator.
| |
METHODS |
Heart preparation.
Hearts (n = 11) were removed from
pentobarbital-anesthetized New Zealand White rabbits in accordance with
Institutional Animal Care and Use Committee guidelines. Hearts were
initially perfused with a solution containing (in mmol/l) 129 NaCl, 5.4 KCl, 1.8 CaCl2, 1.1 MgCl2, 26 NaHCO3, 1 Na2HPO4, 11 glucose, and
0.04 g/l bovine serum albumin, at an aortic pressure of 60-80
cmH2O and a temperature measured with a miniature probe
(Physitemp) in a ventricular chamber of 35-36°C. Later studies
used injection of a 5- to 10-ml bolus of solution with potassium
concentration of 40 or 100 mM into the perfusing tubing or 1 min of
perfusion with a solution in which the potassium concentration was
increased to 20 mM by the addition of KCl. All solutions were
equilibrated to 95% O2-5% CO2 and pH of
7.3-7.4 with gas exchangers fabricated from silicone tubing.
Optical recordings were obtained from all hearts. In the first five
hearts, of which simultaneous microelectrode recordings were obtained
in two hearts, diacetyl monoxime (15 mM) was present in the solution
during all recordings to reduce cardiac movement (3, 21,
25). Two other hearts did not receive diacetyl monoxime and were
endocardially prefrozen with a 10-mm left ventricular catheter
containing liquid nitrogen for 6 min. This produced an epicardial layer
(~1 mm) of viable cardiac fibers that exhibited action potentials and
contractility (1, 20). Four other hearts were not
endocardially prefrozen and were studied both with diacetyl monoxime
(20 mM) during simultaneous optical and microelectrode recordings and
with no diacetyl monoxime during optical recordings to allow analysis
of effects of heart motion. These four hearts had atrioventricular
block produced by thermal ablation at the beginning of the experiment
and were paced for at least 10 min at approximately twice the threshold strength and 3-ms pulse duration from a bipolar electrode on the ventricular epicardium at intervals from 150 to 1,250 ms. One of the
hearts fibrillated during pacing at 150 ms. Orders of perfusion with
and without diacetyl monoxime and of different pacing rates were
varied. Three of the hearts were further studied at a temperature of
22-23°C by turning off the perfusate heater.
Potassium change.
Because the resting membrane potential in cardiac tissue is depolarized
by the addition of extracellular potassium, we used potassium bolus
injections or steady perfusion of hyperkalemic solution to alter the
resting membrane potential (18). In preliminary tests, the
delay from potassium bolus injection to the solution change at the
aortic cannula was determined with blue dye and with measurement of
transmittance of laser light through the cannula. Also elevated
potassium concentration in effluent from hearts after bolus injection
was verified with an electrolyte analyzer. In some experiments, a
custom-fabricated aortic cannula containing a potassium ion-selective
electrode (TIPK, World Precision Instruments, Sarasota, FL), reference,
and ground electrodes was used to indicate timing of the potassium
change just above the left ventricle. The potassium signal was recorded
simultaneously with optical signals.
Ratiometric fluorescence emission measurements.
Hearts were stained with di-4-ANEPPS dissolved in ethanol that was
gradually injected into the perfusion tubing (total 0.2-0.3 ml in
15 min). The heart gently contacted a transparent plate through which
excitation and emission light passed. In some experiments the plate
contained a 1-cm diameter hole to allow positioning of an intracellular
microelectrode near the optical recording site. Two optical recording
systems were used with different hearts (Fig.
1). In both systems, a 488-nm argon laser
excited the di-4-ANEPPS. This wavelength is near the excitation peak or
crossover wavelength (480 nm in bilayer), which minimizes transmembrane
voltage dependence of absorption (5, 10). Thus the
fractional change of red fluorescence is expected to be less than that
found when a longer excitation wavelength is used (5).
With a two-photomultiplier tube system, the laser scanned 63 spots in a
1-cm2 region of epicardium. Red fluorescence was collected
with a photomultiplier tube that had its photocathode covered with a
long-pass glass filter (>610 nm), whereas green fluorescence was
collected with another photomultiplier tube covered with an
interference filter (half-height bandwidth of 534-546 nm). The
photomultiplier tubes were positioned side by side and sufficiently far
from the heart so that the optical path from the heart was essentially
the same for both tubes (maximum angular difference <3°) and
incident fluorescence was approximately perpendicular to the filter
surface.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 1.
Optical recording systems for ratiometric measurements in rabbit
hearts stained with transmembrane voltage-sensitive fluorescent dye
di-4-ANEPPS. A: for two-photomultiplier tube (PMT) system,
blue laser beam (488 nm) steered by acousto-optic deflectors (AODs)
excited epicardial fluorescent dye. Green (540 ± 6 nm) and red
(>610 nm) fluorescence emission was collected with two PMTs that had
filters covering their photocathodes. B: for
spectrofluorometer system, stationary laser beam excited epicardial
fluorescent dye. Emission was collected with spectrometer containing
photodiodes that recorded 16 bands from just above the laser wavelength
to a maximum of 838 nm.
|
|
Before recordings, high-voltages (i.e., gains) for each photomultiplier
tube were adjusted during laser excitation to produce identical DC
fluorescence signals. The relative flux density of fluorescence in the
two wavelength bands was determined by moving both optical filters to
the opposite photomultiplier tube without changing the high voltages.
Before moving, flux densities were inversely proportional to
photomultiplier gains. After moving, the photomultiplier with larger
gain received the fluorescence wavelength band with larger flux
density, and hence the ratio of signals represented the square of the
ratio of flux densities. This method ignores differences in
photocathode sensitivity at different wavelengths and maximum filter
transmittances. The ratio was compared with theoretical ratios
calculated from areas under a gaussian curve with a peak at 620 nm and
standard deviation of 62 nm that was fit to a di-4-ANEPPS emission
spectrum with previously described methods (17).
In preliminary studies, the ratio of the two photomultiplier tube
signals during action potentials was determined online with an analog
circuit (MPY634BM, Burr-Brown, Tucson, AZ). For experiments included in
results, the system hardware was altered to digitize signals from both
of the photomultiplier tubes alternately at a rate of 500 Hz per tube.
Changes in the ratio were then calculated from digitized data.
In some hearts, a spectrofluorometer system was used that contained a
spectrograph (77400, Oriel, Stratford, CT) and a 16-element photodiode
array (PDB-C216, Photonic Detectors, Simi Valley, CA) as described
(17). Fluorescence emission in 15 or 16 adjacent bands of
wavelengths from just above the laser wavelength of 488 nm to 817 or
838 nm was digitized at a sampling rate of 1,000 Hz.
An intracellular glass microelectrode filled with 3 M KCl was
positioned near the optical recording sites in six hearts to measure
transmembrane potentials. The microelectrode signal was recorded
simultaneously with optical signals on a digitizing oscilloscope (Norland 3001A, Atkinson, WI) or a separate channel of each of the
optical recording systems. All recordings in this study were DC coupled.
Data analysis.
Fluorescence signals and ratios were calculated and graphed with a
computer (Sparcstation 20, Sun Microsystems, Mountain View, CA) using
PV Wave or Matlab. So that differences after versus before action
potential phase-zero depolarization or potassium infusion represent all
available data from the two-photomultiplier tube system, measurements
were performed on averaged recordings for all laser spots. Recordings
were not temporally smoothed. Changes in fluorescence and ratios
produced by phase-zero depolarizations or potassium-induced
depolarizations were determined from the mean values in 6-ms segments
before and after depolarization. The RMS noise (i.e., positive square
root of the average of the squares of the deviations of the
flourescence samples about their mean) was determined for 6-ms segments
before depolarization. Comparisons of action potential duration
measured with the microelectrode versus that measured ratiometrically
were performed with a single laser spot near the microelectrode.
Summaries are given as means ± SD. Statistical significance was
determined with the two-tailed t-test or analysis of
variance. A P value <0.05 was considered significant.
| |
RESULTS |
Dual-emission wavelength measurements.
Figure 2 shows results from the
two-photomultiplier tube system and a simultaneous signal from an
intracellular microelectrode. The green signal (wavelengths of 540 ± 6 nm), the red signal (wavelengths >610 nm), and their ratio are
shown. The red signal contains an inverted action potential,
i.e., downward deflection corresponding to the phase-zero
depolarization and upward deflection corresponding to the gradual
repolarization of the action potential. The green signal contains an
upright action potential. These recordings indicate a shift toward
shorter emission wavelengths when the membrane depolarizes.
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 2.
Green and red fluorescence signals recorded with
two-photomultiplier tube system (A), their ratio
(B), and simultaneous intracellular microelectrode signal
(C). The ratio of green to red fluorescence had larger
fractional change during action potential phase-zero depolarization
compared with either red or green signal and followed contour of
transmembrane voltage measured with microelectrode. The fluorescence
intensities of green and red signals (A) were proportional
to the values on the ordinate (i.e., 0 would correspond to 0 fluorescence); however, the constant of proportionality was not
necessarily the same for red and green because the two photomultiplier
tube gains were adjusted to produce a ratio near unity. Perfusing
solution contained diacetyl monoxime. Heart was not endocardially
prefrozen.
|
|
In all 16 recordings from ventricular epicardium of 5 rabbit hearts
using the two-photomultiplier tube system with solution potassium
concentration of 5.4 mM, the fractional fluorescence change (
F/F)
during the action potential phase-zero depolarization was 2.1 ± 0.8% for green fluorescence and 
3.8 ± 0.5% for red fluorescence. The fractional change of the ratio of the green fluorescence to the red fluorescence (Ratio/Ratio) was 6.1 ± 1.1%. Magnitudes of fractional changes were significantly larger for
the ratio versus either of the individual fluorescence signals (P < 0.001, paired 2-tailed t-test).
Figure 3 shows an example of the ratio of
the green and red signals during potassium infusion and the
simultaneous recording from a potassium ion-sensitive electrode located
in the aortic cannula ~1 cm above the heart. Data were recorded
without pacing the heart. Baseline drift in the ratio, which was not
negligible, has been canceled with linear subtraction. Resting membrane
potential decreased and the heart became quiescent when potassium
concentration was raised. The action potential amplitude and duration
also decreased during the first 20 s of hyperkalemia. Recovery
occurred after returning to the initial potassium concentration. In a
different heart that was paced continuously to prevent the ventricles
from becoming quiescent, again the emission ratio after subtraction of
drift followed depolarization and recovery of the resting membrane potential.
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 3.
Effect of hyperkalemia on the ratio of green to red fluorescence
recorded with two-photomultiplier tube system (A) and
simultaneous signal from potassium-selective electrode in perfusion
cannula (B). Baseline drift of ratio was determined from
diastolic values before and after hyperkalemia and subtracted. Ratio
indicated potassium-induced depolarization of the resting membrane
potential. The fractional change of the resting membrane potential
(Ratio/Ratio) at 70 s was +2.6% for the ratio. Corresponding
values for the green and red signals (not shown) were +0.5% and
1.8%. Perfusing solution contained diacetyl monoxime. Spikes in
B (shown by *) were produced by an electrically activated
valve that changed the solution. Perfusing solution contained diacetyl
monoxime. Heart was not endocardially prefrozen.
|
|
The estimated emission flux density ratio in the green collection band
to that in the red collection band for the two-photomultiplier tube
system was 0.083 ± 0.036 (P = 0.001, value vs. 1, two-tailed t-test, n = 3). This is
comparable to theoretical ratios calculated from areas under a curve
fit to the emission spectrum of the dye (METHODS) of 0.1 or
0.05 for the green bands of 540 ± 10 nm or 540 ± 5 nm,
respectively, to the red band of >610 nm.
Spectrofluorometric measurements.
To gain insight into the role of the emission wavelengths chosen for
ratiometric measurements, the emitted fluorescence was measured with a
spectrograph (77400, Oriel, Stratford, CT). Discrete wavelength bands
having a width of 21.9 nm were measured for wavelengths just above the
laser light wavelength of 488 to 817 or 838 nm (17). Of
six hearts with no diacetyl monoxime that were used to test the ability
to lessen motion artifacts with ratiometry, two were endocardially
prefrozen and four were not prefrozen.
Figures 4 and
5 show examples of the motion artifacts
and their reduction by emission ratiometry in endocardially prefrozen hearts. In Fig. 4, simultaneous recordings from two diodes that collected green and red emission wavelengths and their ratio are shown.
The green signal (Fig. 4A) and red signal (Fig.
4B) exhibited a rapid increase and decrease, respectively,
that corresponded to phase-zero depolarization. The rapid changes were
followed by motion artifacts that obscured repolarization in signals
from both diodes. The ratio of the green signal to the red signal (Fig. 4C) canceled the artifact and revealed the action potential
repolarization. Figure 5 shows an example in which the numerator and
denominator signals represented wider bands of the green and red
emission spectrum, such as those that may be sampled with bandpass or
longpass filters. The bands were produced by summing signals from the
spectrophotometer's diodes. The green signal (Fig. 5A)
represents emission wavelengths in the band of 510-576 nm. The red
signal (Fig. 5B) represents emission wavelengths of
598-838 nm. Both green and red signals contain motion artifacts.
The ratio of the green signal to the red signal (Fig. 5C)
eliminated motion artifacts and revealed action potentials.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 4.
Reduction of motion artifact by emission ratiometry with
spectrofluorometer system. Epicardial movement produced deflections in
individual photodiode recordings. Movement-induced deflections
were absent in the ratio of signals from photodiodes. A:
green signal; B: red signal; C: ratio of green to
red signal. Perfusing solution did not contain diacetyl monoxime. The
heart was endocardially prefrozen.
|
|
View larger version (29K):
[in this window]
[in a new window]
|
Fig. 5.
Reduction of photobleaching-induced drift and motion artifact by
emission ratiometry with spectrofluorometer system. The ordinates have
been normalized to the value at the beginning of the recording. Drift
in the green signal (A) was 17%, drift in the red signal
(B) was 12%, and drift in the ratio of green to red
signals (C) was 5%. Motion artifact was also reduced by
ratiometry. Perfusing solution did not contain diacetyl monoxime. The
heart was endocardially prefrozen.
|
|
Ratiometry also lessened effects of motion when the heart was not
endocardially prefrozen. In the example shown in Fig.
6, the phase-zero depolarization
expressed as a fraction of the peak-to-peak deflection during a cardiac
cycle was much larger for the ratio signal compared with either the
green (Fig. 6A) or red (Fig. 6B) signal.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 6.
Reduction of motion artifact by emission ratiometry with
spectrofluorometer system in intact heart. For the green and red
signals (A and B), magnitudes of deflections
during phase-zero depolarization were a small fraction of the
peak-to-peak magnitude of deflection during the cardiac cycle. For the
ratio (C), magnitude of phase-zero depolarization was
essentially the same as the peak-to-peak magnitude. Perfusing solution
did not contain diacetyl monoxime. The heart was not endocardially
prefrozen.
|
|
The ability of ratiometry to lessen motion artifacts was quantitatively
evaluated by measuring the ratio of the magnitude of change in the
signal during phase-zero depolarization to the total peak to peak
change during the cardiac cycle (Ph0/p p). This analysis
was performed on the first recording obtained in each heart without
diacetyl monoxime. The Ph0/p p will have a maximum
possible range of 0-1, will have a small value when motion causes
the signal to exceed the upper or lower values that occur at the
beginning and end of the phase-zero depolarization, and will have a
larger value when the effects of motion are reduced. Wavelength bands
were 510-576 nm and 598-751 nm for numerator and denominator.
In the two hearts that were endocardially prefrozen and received no
diacetyl monoxime, the Ph0/pp was 0.386 ± 0.104 for
the numerator, 0.700 ± 0.311 for the denominator, and 0.941 ± 0.083 for the ratio (P < 0.02 for ratio vs.
numerator, paired 2-tailed t-test). In the four hearts that
were not endocardially prefrozen and received no diacetyl monoxime, the
Ph0/pp was 0.273 ± 0.099 for the numerator,
0.351 ± 0.164 for the denominator, and 0.785 ± 0.144 for
the ratio (P < 0.02 for ratio vs. denominator or
numerator, paired 2-tailed t-test).
Effects of photobleaching, i.e., a slow decrease in emission, were
evident in long-duration recordings obtained with the
two-photomultiplier tube system and with the spectrofluorometer system
as seen in Fig. 5. Results from both systems indicated that the
ratiometry decreased drift but did not fully eliminate it. This may
occur if the fractional decrease in emission due to drift is different for different wavelengths. In all seven trials with the
spectrofluorometer in two endocardially prefrozen hearts that did not
receive diacetyl monoxime, the emission at wavelengths of 532-554
nm (green) decreased 53 ± 18% in 120 s, whereas the
emission at wavelengths of 663-685 (red) decreased 41 ± 16%
(P = 0.011, for red vs. green, paired 2-tailed
t-test). The ratio of the two emission signals (green/red) decreased only 18 ± 13% (P < 0.015 for value
vs. decrease in red or green emission, paired 2-tailed
t-test).
The signals obtained with ratiometry may depend on the choice of
emission wavelengths that are collected. To understand this, we
examined the emission spectrum and its change during an action potential or potassium bolus. Figure 7
shows the results for an action potential. The diastolic emission
spectrum is shown (Fig. 7A). Also, the changes in emission
(emission after the phase-zero depolarization minus emission before it)
for each diode are shown as a fraction of both the diastolic emission
(Fig. 7B) and RMS noise [i.e., signal-to-noise ratio
(signal/noise), Fig. 7C]. The emission of green
light increased while emission of red light decreased, which is
consistent with action potentials in previous figures.
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 7.
Fluorescence emission spectrum and spectral response of di-4-ANEPPS
during an action potential in rabbit heart recorded with
spectrofluorometer system. Relative intensities of emitted fluorescence
at various wavelengths during diastole (F, A), fractional
change in intensities during action potential phase-zero depolarization
(F/F, B) and signal-to-noise ratio (signal/noise)
(F/RMS noise, C) are shown as functions of emission
wavelength. Each measurement is from a single photodiode that collected
light in an optical bandwidth of 21.9 nm. The wavelength of maximum
emission was near 640 nm. Emission of green light (e.g., 520-550
nm) increased during phase-zero depolarization, whereas emission of red
light (e.g., 630-700 nm) decreased. Signal/noise was small at
wavelengths near 600 nm. Perfusing solution did not contain diacetyl
monoxime. Heart was endocardially prefrozen.
|
|
Figures 8 and
9 show the effects of altering numerator
and denominator wavelengths on the fractional change and signal/noise obtained with ratiometry of an action potential. The optimal fractional changes and signal/noise occurred when wavelengths near 525-550 nm
and 650-700 nm were used for the numerator and denominator, respectively. The highest fractional changes and signal/noise were
larger for the ratiometric measurements than for individual diodes
shown in Fig. 7 (e.g., fractional change of 8% for ratio vs. 6% for
diodes, and signal/noise of 60 for ratio vs. 35 for diodes).
View larger version (28K):
[in this window]
[in a new window]
|
Fig. 8.
Fractional change in ratiometric measurement (Ratio/Ratio)
during the same action potential as in Fig. 7. Results for ratios of
all pairs of diodes are shown. Each diode collected light in an optical
bandwidth of 21.9 nm. Large fractional changes occurred with ratios of
signals having wavelengths of ~525 nm and 600-750 nm. Fractional
changes were larger than those found with nonratiometric measurements
(Fig. 7B). Perfusing solution did not contain diacetyl
monoxime. The heart was endocardially prefrozen.
|
|
View larger version (21K):
[in this window]
[in a new window]
|
Fig. 9.
Signal/noise in ratiometric measurement (Ratio/RMS noise) during
the same action potential as in Fig. 8. Results for ratios of pairs of
diodes are shown. (Signal/noise of ratiometric measurements on
principal diagonal is undefined.) Large signal/noise occurred with
ratios of signals having wavelengths of ~550 nm and 650-700 nm.
Signal/noise was larger than that found with nonratiometric
measurements (see Fig. 7C). Perfusing solution did not
contain diacetyl monoxime. The heart was endocardially prefrozen.
|
|
Changes in emission for individual diodes and ratios during the
potassium-induced change in resting membrane potential (i.e., emission
after injection of a potassium bolus minus emission before injection)
were also examined. To correct for emission drift by photobleaching,
each signal from the diode after injection was multiplied by a factor
that represented the drift for that diode measured at a different laser
spot and in a different trial with no potassium bolus. The wavelengths
of emission used in the numerator and denominator that gave the largest
fractional changes and signal/noise during potassium infusion were
similar to those found during the action potential shown in Figs.
7-9. Also, the highest fractional changes and signal/noise during
potassium infusion were larger for the ratiometric measurements than
for individual diodes.
As a further test of the ability of ratiometry to reduce fluorescence
changes that are unrelated to the transmembrane potential, the
coefficients of variation (i.e., SD/mean) (22) were
determined for all seven trials in the two hearts that were
endocardially prefrozen and did not receive diacetyl monoxime.
Measurements were performed with the spectrofluorometer system for the
10-s segment of each recording beginning ~1 s after the onset of
laser illumination. This segment was selected because it contained
pronounced drift in green and red signals due to onset of laser
illumination, and it was after any effects of vibrations in the
apparatus caused by movement of the laser shutter. The hearts were in
sinus rhythm so that recordings contained diastolic intervals and
action potentials. The coefficient of variation represents all of the
variation in fluorescence, including effects of drift, motion and
transmembrane potential changes. However, because we found that
fractional changes due to transmembrane potential were larger for the
ratio, a decrease in the coefficient of variation for the ratio would
indicate reduction of fluorescence changes that are unrelated to the
transmembrane potential. Emissions at wavelengths just above the laser
wavelength of 488-575 nm (numerator) and 598-838 nm
(denominator) and their ratio were determined. These ranges contain the
optimal wavelengths indicated in the preceding analysis. The
coefficient of variation was 0.06 ± 0.03 (range of 0.01-0.1)
for the numerator, 0.05 ± 0.03 (range of 0.01-0.1) for the
denominator, and 0.02 ± 0.01 (range of 0.004-0.035) for
their ratio (P = 0.019 for ratio vs. numerator or
denominator, paired two-tailed t-test). Thus variability was
reduced for the ratio even though the ratio had increased fractional
fluorescence changes for action potentials and potassium-induced depolarization.
Ratiometric measurement of action potential repolarization.
The ability of the ratio to indicate action potential repolarization
was tested in six hearts with simultaneous microelectrode and optical
recordings during perfusion with warm solution containing 15 or 20 mM
diacetyl monoxime. Because multiple trials were performed at some laser
spots, the durations of recordings were shortened to 2-4 s to
lessen photobleaching. Figure 10 shows
recordings of the ratio and transmembrane voltage during pacing at
intervals of 150-1,250 ms in one heart. Ratio and microelectrode
signals indicate similar action potential contours and durations for a given pacing interval. Even though activation time at the laser recording spot differed from that at the microelectrode tip by 3.8 ± 1 ms as expected due to their spatial separation, the plots of ratio
versus transmembrane voltage indicate a linear relationship between the
two signals during most of the action potential duration.
View larger version (38K):
[in this window]
[in a new window]
|
Fig. 10.
Ratiometric and microelectrode recordings plotted vs.
time and vs. each other from one heart. Each row shows simultaneous
recordings during pacing at intervals of 150 ms (A), 250 ms
(B), 500 ms (C), and 1,250 ms (D).
Recordings were obtained in the sequence: B, A,
C, and D, and were completed within ~45 min.
Action potential durations increased with increases in pacing interval
in both ratiometric and microelectrode recordings. Graph at right in
each row illustrates linear relationship between ratiometric and
microelectrode recordings except during phase-zero depolarization
(i.e., part of the trace that falls below the line) as expected due to
spatial separation between laser spot and microelectrode tip. The
change in the ratio per 100 mV determined from slopes in right graphs
omitting phase-zero depolarization (0.015, 0.018, 0.01, and 0.005 from
top to bottom) decreased in the same sequence in which recordings were
obtained. Ratio was measured with emission wavelength bands of
510-576 nm and 598-751 nm for numerator and denominator,
respectively. The heart was perfused with solution at 36° containing
diacetyl monoxime and was not endocardially prefrozen.
|
|
The combined results shown in Fig. 11
indicate agreement of action potential duration at 75% repolarization
(APD75) measured ratiometrically with that measured from
microelectrode signals. Two-way analysis of variance of
APD75 in the six hearts during warm perfusion indicated
P = 0.003 for an effect of the pacing interval and
P = 0.847 for an effect of the recording method (i.e., microelectrode vs. ratiometry).
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 11.
Scatter plot of action potential duration at 75%
repolarization (APD75) measured from ratiometric recordings
and simultaneous microelectrode recordings in 6 hearts. Temperature was
36-37°, and pacing interval was 150-1,250 ms (open
circles). Additional measurements were obtained in 3 of the hearts at
room temperature during pacing at 1,250 ms (filled circles). Ratio was
measured with emission wavelength bands of 510-576 nm and
598-751 nm for the numerator and denominator, respectively.
Perfusing solution contained diacetyl monoxime. The hearts were not
endocardially prefrozen.
|
|
In the three hearts that were paced at rates of 150, 250, 500, and
1,250 ms and did not fibrillate, APD75 measured
ratiometrically (wavelength bands of 510-576 and 598-751 nm
for the numerator and denominator) and with microelectrodes,
respectively, was 99 ± 6 and 98 ± 9 ms during pacing at 150 ms, 116 ± 9 and 124 ± 4 ms during pacing at 250 ms,
144 ± 15 and 145 ± 14 ms during pacing at 500 ms, and
153 ± 33 and 155 ± 33 ms during pacing at 1,250 ms
(P > 0.1 for ratiometric measurement vs.
microelectrode measurement at same pacing interval, P 
0.05 for ratiometric or microelectrode measurements at 500 ms vs.
those at 150 ms, paired 2-tail t-tests).
The effect of a potassium bolus on the action potential duration was
recorded with the two-photomultiplier tube system in two hearts
perfused with warm solution containing 15 mM diacetyl monoxime. The
APD75 measured ratiometrically at a single laser spot or
with microelectrodes changed from 159 ± 25 ms to 102 ± 8 ms
when potassium was introduced (P = 0.005, paired
2-tailed t-test, n = 4 hearts × recording methods). Two-way analysis of variance of the
APD75 indicated P = 0.022 for an effect of
the potassium and P = 0.65 for an effect of the
recording method (i.e., microelectrode vs. ratiometry).
Calibration of ratiometric measurements.
The calibration factor between the ratio and the transmembrane voltage
during perfusion with solution containing normal potassium was
determined in the six hearts in which microelectrode recordings were
obtained (of which 2 were studied with the 2-photomultiplier tube
system and 4 were studied with the spectrofluorometer). For spectrofluorometric measurements, emission wavelength bands of 510-576 nm and 598-751 nm were used for the numerator and
denominator, respectively, which were wavelengths that produced large
fractional changes and low noise in the ratio described. When all
recordings were included, the calibration slope depended on the order
of the recording in each heart (P = 0.03, analysis of
variance, n = 30). A linear regression of the
calibration slope versus the order of the recording indicated that the
calibration slope decreased at an average rate of 0.0058 per recording.
This decrease may be due to dye internalization or photobleaching
(14). To estimate the calibration slope when these effects
are minimized, only a single ratiometric recording from each heart was
included. The criterion for selection was that the recording had the
largest Ratio/Ratio found in the heart, which consistently occurred
for recordings obtained within 20 min after staining the heart with d-4-ANEPPS.
In all six hearts, transmembrane voltages before and after phase-zero
depolarization were 85 ± 6 mV and 18 ± 7 mV
(P > 0.6 for hearts studied with photomultiplier tubes
versus hearts studied with spectrofluorometer, nonpaired 2-tail
t-test). In the two hearts studied with photomultiplier
tubes for which gains were initially set to produce a ratio of unity,
ratios before and after phase-zero depolarization were 0.98 ± 0.03 and 1.05 ± 0.02, and the calibration slope was (0.071 ± 0.017)/100 mV. In the four hearts studied with the
spectrofluorometer, ratios before and after phase-zero depolarization
were 0.21 ± 0.03 and 0.23 ± 0.04, and the calibration slope
was (0.017 ± 0.002)/100 mV.
Reduction of effects of laser noise.
The ability of emission ratiometry to reduce common mode noise due to
fluctuations of the excitation light intensity was studied by disabling
a regulator in the laser power supply that controlled laser light
intensity. Under this condition, the peak-to-peak magnitude of periodic
laser intensity fluctuations at a rate of 60 Hz was 1.7% of the
average laser intensity. For green and red emission signals, 60-Hz
fluctuations were approximately half as large as the deflections
produced during the phase-zero depolarization, as shown in Fig.
12, A and B.
However, for the ratio of these signals (Fig. 12C), any
60-Hz fluctuations were much smaller than the phase-zero depolarization.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 12.
Cancellation of the effect of fluctuations in laser intensity.
Laser light regulation, which normally stabilized laser
intensity, was eliminated to produce laser intensity
fluctuations at the AC power line frequency of 60 Hz. Both emission
fluctuations at 60 Hz and effects of motion at the heart's rate of 4 Hz in numerator (A) and denominator (B) were
lessened in ratio (C). Recording was obtained immediately
after the recording shown in Fig. 6. Perfusing solution did not contain
diacetyl monoxime. Heart was not endocardially prefrozen.
|
|
| |
DISCUSSION |
Emission ratiometry for transmembrane voltage measurements in
hearts.
The results indicate that emission ratiometry reduces drift and effects
of epicardial motion, which is indicated in individual recordings
(Figs. 4-6 and 12), in the measurements of drift over a 2-min
period and in the decreased coefficient of variation. The ratiometry
allows optical recordings of cardiac repolarization without requiring
electromechanical uncoupling drugs. This is a potentially important
advantage because the drugs can have undesired effects in myocardium
(3, 16, 25, 26). Also ratiometry reduces drift, which
affects long-duration recordings. Whereas slow changes in transmembrane
potential that occur over a longer time period than that studied here
may still be obscured by drift, our results show that ratiometry can
reveal changes in transmembrane potential that occur during a potassium bolus.
Photobleaching-induced drift was only partly canceled by ratiometry as
evident from drift in the ratio of green light to red light (Fig. 5).
The spectrofluorometric results indicated that there was a greater
decrease in recorded green emission compared with red. This is not
expected if photobleaching only destroys fluorophore molecules, in
which case the fractional decrease in emission should be the same for
all fluorescence wavelengths. A greater decrease in green emission may
result from a different emission spectrum of the remaining fluorophore,
possibly due to the environment of the fluorophore, or from a
phosphorescence emission at the longer wavelengths. (11,
24) Also, dye internalization may be enhanced with intense laser
illumination. If the emission spectrum and photobleaching decay rates
differ in different membranes within the cells, this would produce
drift of the ratio during photobleaching when dye has internalized, as
proposed by Kao, WY, Davis C, Kim YI, and Beach JM (unpublished
obervations). Also a change in the membrane environment, if it
occurs intracellularly, may alter spectral properties of internalized
dye because of changes in the membrane dipole potential
(14).
The emission ratiometry has potential advantages for studies in which
changes in absorption due to chemical effects unrelated to
transmembrane voltage may interfere with transmembrane voltage measurements. The ratio is expected to reject effects of changes in
absorption because they are common to both green and red emissions. The
finding that ratiometry detected similar transmembrane
voltage-dependent shifts of emission wavelengths for the cardiac action
potential and hyperkalemia indicates that any absorption changes due to potassium were rejected.
Regarding the possibility of improved calibration with ratiometry,
Bullen and Saggau (5) found that slopes of ratio versus transmembrane voltage were approximately constant among cells; however,
variable offset existed in the ratio that required subtraction. They
also found that when offset was subtracted, slopes indicated a
conversion factor between transmembrane voltage and ratio of 0.015/100
mV, which is similar to our mean slope of 0.017/100 mV found with the
spectrofluorometer. The calibration slope may depend on detector
sensitivities and optical bandwidths. Whereas gains and bandwidths were
constant among all diodes in our spectrofluorometer, they were
different in our photomultiplier tubes, which produced a mean slope of
0.071/100 mV when the gains were adjusted to produce an initial ratio
of unity. Had our spectrofluorometer diode gains been adjusted to
produce an initial ratio of unity (which would require increasing the
gains of diodes sensing green light by a factor of ~4.5), then a
slope of 0.077/100 mV can be calculated, which is close to the value
obtained with our two photomultiplier tubes. Despite similarities in
our mean slopes obtained with the spectrofluorometer compared with that
reported in neurons by Bullen and Saggau, the variation of slopes among
hearts was larger than the variation among neurons. This may be due to
nonspecific fluorescence in the heart in which various types of tissue
are present, as discussed by Bullen and Saggau.
The signal/noise obtained ratiometrically (Fig. 9) is optimal at
numerator wavelengths slightly longer than those that give the largest
fractional ratio change. This may be due to larger fluorescence
emission at the longer numerator wavelengths (Fig. 7), which increases
DC level and lessens effects of shot noise. In systems that use two
detectors for which the optical bandwidths of numerator and denominator
signals can be selected a priori, the shot noise should decrease when a
wider optical bandwidth is used. This is expected to increase
signal/noise if the additional wavelengths add to the change in ratio.
For the denominator, wavelengths will add to the signal in a wide band
of approximately >600 nm (Fig. 8). For the numerator, the widest
possible band is bounded by the excitation wavelength and ~580 nm,
above which wavelengths add no fractional change in the ratio. However,
a narrower numerator band gives a larger fractional change and
signal/ noise by excluding wavelengths that have a small change
(Figs. 8 and 9). Also, neither band needs to be wide if there is
sufficient light intensity to overcome shot noise, because the largest
signal/noise was obtained with narrow numerator and denominator bands
near 550 and 680 nm. Another issue impacting on the choice of bands is
that nonspecific fluorescence from surrounding cellular structures or
dye bound in nonspecific orientation may be rejected when optical
bandwidth is small (5).
Motion artifacts in optical action potential recordings may have
multiple causes that include possible changes in light transport properties (reflectance, scattering, or absorption of light) at the
excitation and emission wavelengths due to stretching of the tissue and
translational movement of the tissue when the light transport
properties or dye distribution among voltage-sensing and nonspecific
environments are heterogeneous. Ratiometry may eliminate distortion of
the action potential due to changes in light transport properties
provided that identical changes occur for the two emission wavelength
bands used for the numerator and denominator. Even when ratiometry
eliminates visible distortion in the optically recorded action
potential, translational movement would cause different phases of an
action potential to be recorded from different groups of cells. This
limitation is potentially important in cases of heterogeneous
repolarization and resting potential. In these cases, ratiometry may be
advantageous with endocardially prefrozen or mechanically stabilized
hearts that already have motion lessened by contact with adjacent
underlying necrotic tissue or external stabilizers.
Laser scanner method for ratiometry.
Recording systems in which spatial resolution of the recording is
determined by a narrow excitation laser beam have potential advantages
for emission ratiometry in hearts as well as microscopy (5,
7). The fact that localization of a recording is controlled by
the excitation beam ensures that both red and green emission signals
are collocal. For systems in which broad-field excitation is used and
spatial resolution is determined by locations of detectors in image
planes, collocal ratiometric recordings depend on exact alignment of
detectors (23). Effects of laser intensity fluctuations, which have been a limitation of laser scanner systems and required additional methods for cancellation (15, 19), are reduced by emission ratiometry because the effect is common to numerator and
denominator signals (Fig. 12). Also, the use of a narrow laser excitation bandwidth (~10
5 nm) allows short emission
wavelengths (e.g., 500 nm, Fig. 8) to be included in the numerator if necessary.
The emission spectrum in hearts.
The emission spectrum and its change during depolarization in hearts
are qualitatively similar to previous measurements of di-4-ANEPPS
emission in other preparations. For Retzius cells from the leech,
Fromherz and Lambacher (11) found an emission peak at 604 nm and emission crossover at 575 nm for excitation at 482 nm with a
positive change of ~6% in emission at 533 nm and a negative change
of 5% at 636 nm for 100 mV of depolarization. For hemispherical
bilayers, Fluhler et al. (10) found an emission peak at
640 nm and emission crossover at 620 nm with a positive change of 7.6%
at 540 nm and a negative change of ~2.2% at 680 nm for 100 mV of
depolarization. Our emission peak (Fig. 7) is similar to that of
Fluhler et al. (10), whereas our emission crossover for
the action potential is similar to that of Fromherz and Lambacher
(11). Both of these reports agree with our findings in
hearts in which positive changes in green emission and negative changes
in red emission occur during membrane depolarization.
| |
ACKNOWLEDGEMENTS |
This study was supported by National Institutes of Health Grant
HL-52003 and American Heart Association Grant 9740173N. S. B. Knisley is an Established Investigator Awardee of the American Heart Association.
| |
FOOTNOTES |
Address for reprint requests and other correspondence:
S. B. Knisley, The Univ. of Alabama at Birmingham, B122 Volker
Hall, 1670 University Blvd., Birmingham, AL 35294-0019 (E-mail:
sbk{at}crml.uab.edu).
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.
Received 8 November 1999; accepted in final form 8 March 2000.
| |
REFERENCES |
1.
Allessie, MA,
Schalij MJ,
Kirchhof CJHJ,
Boersma L,
Huybers M,
and
Hollen J.
Experimental electrophysiology and arrhythmogenicity: anisotropy and ventricular tachycardia.
Eur Heart J
10, SupplE:
2-8,
1989[Abstract/Free Full Text].
2.
Beach, JM,
McGahren ED,
Xia J,
and
Duling BR.
Ratiometric measurements of endothelial depolarization in arterioles with a potential-sensitive dye.
Am J Physiol Heart Circ Physiol
270:
H2216-H2227,
1996[Abstract/Free Full Text].
3.
Biermann, M,
Rubart M,
Moreno A,
Wu J,
Josiah-Durant A,
and
Zipes DP.
Differential effects of cytochalasin D and 2,3 butanedione monoxime on isometric twitch force and transmembrane action potential in isolated ventricular muscle: implications for optical measurements of cardiac repolarization.
J Cardiovasc Electrophysiol
9:
1348-1357,
1998[Web of Science][Medline].
4.
Brandes, R,
Figueredo VM,
Camacho SA,
Massie BM,
and
Weiner MW.
Suppression of motion artifacts in fluorescence spectroscopy of perfused hearts.
Am J Physiol Heart Circ Physiol
263:
H972-H980,
1992[Abstract/Free Full Text].
5.
Bullen, A,
and
Saggau P.
High-speed random-access fluorescence microscopy. II. Fast quantitative measurements with voltage-sensitive dyes.
Biophys J
76:
2272-2287,
1999[Web of Science][Medline].
6.
Dillon, SM.
Synchronized repolarization after defibrillation shocks: A possible component of the defibrillation process demonstrated by optical recordings in rabbit heart.
Circulation
85:
1865-1878,
1992[Abstract/Free Full Text].
7.
Dillon, S,
and
Morad M.
A new laser scanning system for measuring action potential propagation in the heart.
Science
214:
453-456,
1981[Abstract/Free Full Text].
8.
Efimov, IR,
Ermentrout B,
Huang DT,
and
Salama G.
Activation and repolarization patterns are governed by different structural characteristics of ventricular myocardium: Experimental study with voltage-sensitive dyes and numerical simulations.
J Cardiovasc Electrophysiol
7:
512-530,
1996[Web of Science][Medline].
9.
Efimov, IR,
Huang DT,
Rendt JM,
and
Salama G.
Optical mapping of repolarization and refractoriness from intact hearts.
Circulation
90:
1469-1480,
1994[Abstract/Free Full Text].
10.
Fluhler, E,
Burnham VG,
and
Loew LM.
Spectra, membrane binding, and potentiometric responses of new charge shift probes.
Biochemistry
24:
5749-5755,
1985[Medline].
11.
Fromherz, P,
and
Lambacher A.
Spectra of voltage-sensitive fluorescence of styryl-dye in neuron membrane.
Biochim Biophys Acta
1068:
149-156,
1991[Medline].
12.
Girouard, SD,
Pastore JM,
Laurita KR,
Gregory KW,
and
Rosenbaum DS.
Optical mapping in a new guinea pig model of ventricular tachycardia reveals mechanisms for multiple wavelengths in a single reentrant circuit.
Circulation
93:
603-613,
1996[Abstract/Free Full Text].
13.
Gray, RA,
Jalife J,
Panfilov AV,
Baxter WT,
Cabo C,
Davidenko JM,
and
Pertsov AM.
Mechanisms of cardiac fibrillation.
Science
270:
1222-1223,
1995[Abstract/Free Full Text].
14.
Gross, E,
Bedlack RS, Jr,
and
Loew LM.
Dual-wavelength ratiometric fluorescence measurements of the membrane dipole potential.
Biophys J
67:
208-216,
1994[Web of Science][Medline].
15.
Hill, BC,
and
Courtney KR.
Design of a multi-point laser scanned optical monitor of cardiac action potential propagation: application to microreentry in guinea pig atrium.
Ann Biomed Eng
15:
567-577,
1987[Web of Science][Medline].
16.
Jalife, J,
Morley GE,
Tallini NY,
and
Vaidya D.
A fungal metabolite that eliminates motion artifacts.
J Cardiovasc Electrophysiol
9:
1358-1362,
1998[Web of Science][Medline].
17.
Johnson, PL,
Smith W,
Baynham TC,
and
Knisley SB.
Errors caused by combination of Di-4-ANEPPS and Fluo3/4 for simultaneous measurements of transmembrane potentials and intracellular calcium.
Ann Biomed Eng
27:
563-571,
1999[Web of Science][Medline].
18.
Kishida, H,
Surawicz B,
and
Fu LT.
Effects of K+ and K+-induced polarization on (dV/dt)max threshold potential, and membrane input resistance in guinea pig and cat ventricular myocardium.
Circ Res
44:
800-814,
1979[Abstract/Free Full Text].
19.
Knisley, SB,
Blitchington TF,
Hill BC,
Grant AO,
Smith WM,
Pilkington TC,
and
Ideker RE.
Optical measurements of transmembrane potential changes during electric field stimulation of ventricular cells.
Circ Res
72:
255-270,
1993[Abstract/Free Full Text].
20.
Knisley, SB,
and
Hill BC.
Optical recordings of the effect of electrical stimulation on action potential repolarization and the induction of reentry in two-dimensional perfused rabbit epicardium.
Circulation
88:
2402-2414,
1993[Abstract/Free Full Text].
21.
Li, T,
Sperelakis N,
Teneick RE,
and
Solaro RJ.
Effects of diacetyl monoxime on cardiac excitation-contraction coupling.
J Pharmacol Exp Ther
232:
688-695,
1985[Abstract/Free Full Text].
22.
Mendenhall, W,
Scheaffer RL,
and
Wackerly DD.
Mathematical Statistics with Applications, edited by Payne M.. Boston, MA: Duxbury, 1986, p. 293.
23.
Rohr, S,
and
Kucera JP.
Optical recording system based on a fiber optic image conduit: assessment of microscopic activation patterns in cardiac tissue.
Biophys J
75:
1062-1075,
1998[Web of Science][Medline].
24.
Song, L,
Varma CAGO,
Verhoeven JW,
and
Tanke HJ.
Influence of the triplet excited state on the photobleaching kinetics of fluorescein in microscopy.
Biophys J
70:
2959-2968,
1996[Web of Science][Medline].
25.
Verrecchia, F,
and
Hervé JC.
Reversible blockade of gap junctional communication by 2,3-butanedione monoxime in rat cardiac myocytes.
Am J Physiol Cell Physiol
272:
C875-C885,
1997[Abstract/Free Full Text].
26.
Wu, J,
Biermann M,
Rubart M,
and
Zipes DP.
Cytochalasin D as excitation-contraction uncoupler for optical mapping action potentials in wedges of ventricular myocardium.
J Cardiovasc Electrophysiol
9:
1336-1347,
1998[Web of Science][Medline].
Am J Physiol Heart Circ Physiol 279(3):H1421-H1433
0363-6135/00 $5.00
Copyright © 2000 the American Physiological Society
TOP
ABSTRACT
INTRODUCTION
