Hyperpolarization and lysophosphatidylcholine induce inward currents and ethidium fluorescence in rabbit ventricular myocytes

Abstract

Strong electric pulses produce reversible or irreversible membrane breakdown (electroporation). We analysed the permeation properties of minute pores caused by hyperpolarization or lysophosphatidylcholine (LPC) by comparing the amount of charge carried by irregular inward currents (I_hi) with changes in ethidium bromide (EB) fluorescence in isolated rabbit ventricular myocytes. Forty-second negative pulses from a holding potential of −20 mV induced I_hi whose conductance increased with hyperpolarization; the mean conductance (G_hi) was 63.6 ± 9.9 pS pF⁻¹ (mean ± s.e.m., n = 9) at −160 mV. EB fluorescence increased during voltage pulses in parallel with the time integral of I_hi (Q_hi), with the magnitude of the increases in nuclear EB fluorescence being 5.3 times greater than in the cytoplasm at −160 mV. Similar hyperpolarization-induced parallel increases in I_hi and EB fluorescence were also obtained in Na⁺-free, N-methyl-d-glucamine (NMDG) solution. LPC (10 μm) induced large (101.2 ± 21.2 pS pF⁻¹, n = 16), rapid (rise times, 1-10 ms) I_hi with slow relaxation rates at −80 mV that reflected increases in G_hi to 94.3 ± 24.8 pS pF⁻¹ (n = 8) at 6 min. Plots of EB fluorescence vs. Q_hi were well fitted by a common Hill's equation with a Hill coefficient of 0.97. Taken together, our findings indicate that hyperpolarization and LPC produced pores having the same filter properties for the permeation of small ions, including ethidium⁺, and that I_hi (carried in part by Ca²⁺) generated by membrane breakdown are capable of supplying sufficient ions to evoke abnormal excitation and contraction in cardiac myocytes.

The plasma membrane lipid bilayer is a barrier to the diffusion of charged molecules. High electrical resistivity accompanies this barrier function and is the basis for normal cellular electrical activity. When exposed to a high-intensity electrical field, however, the lipid bilayer is reversibly or irreversibly permeabilized (electroporation, Tsong, 1991; Weaver, 1993). Such electroporation has been utilized to facilitate uptake of DNA and other macromolecules into cells. Moreover, a recent rabbit heart study revealed that electroporation is involved in the therapeutic efficacy of electric shock for ventricular fibrillation (Al-Kharda et al. 2000). On the other hand, electric shock may also exert a transient detrimental effect on cardiac haemodynamics in human subjects (Tokano et al. 1998). This complication may be due to transient electrical leaks that affect the membrane potential and to diffusion of Ca²⁺ ions through such leaks. For example, countershock-type electric field stimulation (100-200 V cm⁻¹) induces sustained depolarization and contracture in cultured chick embryo myocardial cells (Jones et al. 1978), presumably due to electroporation, while short DC field shocks > 50 V cm⁻¹ induce similar sustained depolarizations in isolated guinea pig papillary muscles (Kodama et al. 1994).

One way to demonstrate electroporation is to characterize the permeation of markers through the passive pores. For example, microlesions in chick embryo myocardial cells caused by field stimulation at 50-200 V cm⁻¹ enable incorporation of fluorescein isothiocyanate-labelled dextrans (FITC-dextrans) with molecular masses of 4-10 kDa (Jones et al. 1987). In addition, calcein (623 Da) and propidium iodide (668 Da) have been used with flow cytometry to optimize electroporation protocols in cultured prostate cancer cells (Cantella et al. 2001). The internal Ca²⁺ concentration ([Ca²⁺]_i) is also a useful indicator of electroporation (Teruel & Meyer, 1997), though Ca²⁺ released from internal stores predominates in HeLa cells at low stimulus intensities (Bobanovi'c et al. 1999), and Ca²⁺ influx through voltage-gated channels would be expected to contaminate influx through passive pores (Cheek et al. 2000).

Ethidium bromide (EB) is a small, fluorescent cyanide dye (ethidium⁺, 314 Da) that intercalates strongly with double helical DNAs and RNAs, which causes striking increases in fluorescence (Olmsted & Kearns, 1977). Often used for labelling dead cells (Lim et al. 2001), it is also quickly taken up by intact cultured cells (Burns, 1972; Tramier et al., 2000) and by Drosophila salivary gland cells (Favard et al. 1997). EB behaves as a lipophilic cation that accumulates in mitochondria, where it inhibits oxidative phosphorylation (Peña et al. 1977). The absence of EB fluorescence in living cells is largely explained by the strong inhibition of its intercalation with intact chromatinized DNA (Tramier et al. 2000). On the other hand, when used as an electro-permeabilization indicator, EB exhibited polarization-dependent distribution in the cytoplasm of cultured cells (Tekle et al. 1991). In addition, following electroporation of several mammalian cell lines in low salt medium, EB and Ca²⁺ were utilized to show smaller pores in the anode-facing membrane, while propidium iodide and ethidium homodimer were more permeable through larger pores in the cathode-facing membrane (Tekle et al. 1994). Mitochondrial DNA is the main target for EB in the cytoplasm (Hayashi et al. 1994), but RNA may also be involved (Burns, 1972). Whether EB intercalation with nuclear DNA develops within a short time after membrane permeabilization is not yet known.

We previously demonstrated that La³⁺-suppressible irregular inward currents (I_hi) are activated by hyperpolarizing voltage clamp pulses to potentials between −110 and −170 mV for 10-30 s in isolated rabbit ventricular myocytes (Akuzawa-Tateyama et al. 1998). In the present study, we used EB as a fluorescent marker that could pass through small pores produced by mild electroporation. We show that EB fluorescence increases in parallel with the time integral of I_hi at voltages ranging from −120 to −180 mV, and that similar changes in I_hi and EB fluorescence are evoked by 1-α-lysophosphatidylcholine (LPC), an amphiphilic phospholipid that acumulates in ischaemic myocardial tissues, where it affects various ion channels (Lopaschuk, 2001) and has potent cytolytic activity (Weltzien, 1979; Katz & Messineo, 1981).

Methods

This study was conducted in accordance with the guiding principles for the care and use of animals in the field of physiological sciences set out by the Physiological Society of Japan (1994). The protocol was approved by our University's Animal Experimentation Committee.

Single cell preparation

Ventricular myocytes were enzymatically isolated from the hearts of Japanese white rabbits (1.6 kg) during Langendorff perfusion at 37 °C. The rabbits were anaesthetized by injection of urethane (1.25 g (kg body weight)⁻¹) into the auricular vein, after which the ascending aorta was cannulated under artificial respiration, and the heart was excised and sequentially superfused with normal Tyrode solution for 5 min, Ca²⁺-free Tyrode for 5 min, collagenase solution for 36 min and Kraft-Brühe (KB) solution (composition given below) for 5 min. The digested ventricles were cut into small pieces with scissors and gently shaken in a warmed water bath for 1 min. The tissue fragments were then passed through a coarse metal filter into a beaker containing KB solution, after which the cells that precipitated to the bottom were aspirated in a Pasteur pipette and dispersed into a second beaker containing normal Tyrode solution at 37 °C. The cells were kept at 37 °C for an additional 15 min, then allowed to cool to 25 °C for experimentation.

Patch clamp instrumentation and procedures

Membrane currents were recorded under voltage-clamp configurations using the whole-cell patch clamp technique with an EPC-8 patch clamp amplifier (HEKA Electronic, Lambrecht, Germany) (Hamill et al. 1981). Using a Pasteur pipette, cells were gently dispersed onto a cover glass fixed at the bottom of an experimental chamber, where they were superfused continuously at about 1 ml min⁻¹ using a peristaltic pump. The depth of solution in the chamber was < 1 mm. A 3 m KCl-agar electrode placed at the exit of the chamber served as the indifferent electrode.

Patch pipettes were pulled from plain capillary tubes made of soda lime glass (Chase Instruments Corp., NY, USA), coated with Sylgard (Silpot 184, Dow Corning) and heat-polished under a microscope; the resultant electrode tip resistances ranged from 1 to 3 MΩ when filled with the pipette solution. Once a giga-seal was established, the patch membrane was disrupted by applying a short, negative pressure pulse. Membrane voltage and currents were filtered at 1 kHz using a Bessel-type filter and continuously recorded at a sampling frequency of 4 kHz using a PowerLab Chart system (Chart V. 4.0.1, AD Instruments, Castle Hill, Australia) running on a Macintosh computer (Power Macintosh 7500/100, Apple Computer). Analysis of patch clamp data was carried out using Igor Pro V. 4.02A (Wavemetrics Inc., OR, USA) running on a Macintosh computer (Power Macintosh G4, Apple Computer). Membrane capacitance was estimated by integrating capacitative currents elicited by 10 mV depolarization steps from 0 mV. We used small cells with capacitances of 55.0 ± 2.0 pF (mean ± s.e.m., n = 90). Pulses of 300 ms duration were applied from a holding potential (HP) near the resting potential in normal Tyrode solution to verify the normal electrical activity of the myocytes. K⁺-free solution containing 5 μm glibenclamide, a K⁺_ATP channel blocker (Findlay, 1992), was introduced before switching to the test solution containing 25 μm EB. In the test solution, 40 s negative-going voltage steps were applied from an HP of −20 mV with 2 min intervals in between.

Measurement of marker fluorescence

EB fluorescence was recorded using a Nikon TE 300 microscope equipped with a xenon lamp, an electronic shutter and a cooled, slow-scan, black-and-white CCD camera (1024 pixels × 1024 pixels; model C4749-25, Hamamatsu Photonics, Hamamatsu, Japan). The fluorescence images were collected using a ×40 objective lens and an appropriate filter set (XF103-2, Omega Optical, VT, USA; excitation: peak 525 nm with full width at half-maximum (FWHM) of 45 nm; emission: peak 645 nm with FWHM 75 nm). Exposure times were 222 ms. Images and signal intensities from selected regions of cells were analysed using Aquacosmos V. 1.3 (Hamamatsu Photonics, Hamamatsu, Japan).

Time-lapse recordings, with 20 s intervals between frames, of both transparent and fluorescent images were started at the time EB-containing test solutions entered the chamber. The images recorded after the initiation of contraction were not adopted for analysis. After subtracting background fluorescence recorded from a cell-free field from the cell images, EB fluorescence was measured in the nuclei and cytoplasm and presented as mean values from each region. EB fluorescence from myocytes was quantified by comparing it to a calibration curve previously constructed by plotting EB concentration vs. steady-state fluorescence intensity measured in solutions containing 12.5 to 1250 μm EB in a cell-free chamber. Fluorescence intensities are given in relative units: 1 denotes the background intensity in the presence of 25 μm EB, which was almost constant in each experiment.

Solutions and drugs

Normal Tyrode solution had the following composition (mm): NaCl 135; KCl 5.4; CaCl₂ 1.8; MgCl₂ 1; Hepes 5; glucose 10; pH was adjusted to 7.40 with NaOH. Collagenase solution contained 45 mg collagenase (Sigma type I) in 50 ml of low (20 μm) Ca²⁺-containing normal Tyrode solution. KB solution contained (mm): KOH 110; taurine 10; oxalic acid 10; glutamic acid 70; KCl 25; KH₂PO₄ 10; EGTA-Tris 0.5; Hepes 5; glucose 10; pH adjusted to 7.40 by KOH. K⁺-free solution was made by removing KCl from normal Tyrode solution and adding 5 mmNaCl and 5 μm glibenclamide (Sigma Chemical Co.). NMDG solution contained (mm): N-methyl-d-glucamine chloride 140; CaCl₂ 1.8; MgCl₂ 1; Hepes 5; glibenclamide 0.02; glucose 10; pH was adjusted to 7.40 with Tris. Pipette solution contained (mm): KCl 130; MgCl₂ 1; 1,2-bis(2-aminophenoxy) ethane-N,N,N‘,N‘-tetraacetic acid (BAPTA) 5; MgATP 3; Hepes 5; pH was adjusted to 7.40 with KOH. Low Cl⁻ pipette solution was made by replacing 119 mm KCl in the above pipette solution with isomolar potassium aspartate. Glibenclamide (5 mm) and LPC (10 mm) were dissolved in ethanol as stock solutions. A low concentration of ethanol (0.1 %) in test solutions did not appreciably affect electroporation. EB (Molecular Probes, Eugene, OR, USA) was dissolved in double-distilled water at 10 mg ml⁻¹ for the stock solution.

Statistics

Values are expressed as means ± standard errors of the mean (s.e.m.). Statistical comparisons were made using Student's t test for unpaired and paired samples. Values of P < 0.05 were considered statistically significant.

Results

Inward currents and ethidium fluorescence increase produced by hyperpolarizations

Membrane currents

Successive hyperpolarizing steps to potentials between −80 and −180 mV from an HP of −20 mV elicited I_hi that did not exhibit unitary currents on faster recordings and were partially inhibited by 0.5 mm La³⁺ (Fig. 1), as was previously reported for hyperpolarizations from an HP of −80 mV in the absence of EB (Akuzawa-Tateyama et al. 1998). I_hi fluctuated irregularly, but upon hyperpolarization there was clearly an instantaneous increase in amplitude, which then gradually increased further during the course of the pulse (Fig. 1B). The initial rapid increase is explained by the increase in the driving force for the passage of ions through the opened pores, while the subsequent slower, irregular increase is thought to reflect a gradual increase in I_hi conductance (G_hi). The fact that the I_hi-V relationship was nearly linear with a reversal potential of near 0 mV (Fig. 5) indicates that the progressive increases in amplitude that occurred with increasing hyperpolarizations reflected mainly the increase in G_hi. The appearance of spiky inward Na⁺ currents upon depolarization to −20 mV in K⁺-free solution following the pulses suggests that ion channel properties were essentially preserved despite the large, sustained hyperpolarizations. The same negative pulse sequence elicited similar I_hi in Na⁺-free, NMDG solution (Fig. 2) and also in K⁺ -free solution when low Cl⁻ (10 %) pipette solution was used. Figure 2 shows that I_hi developed gradually during hyperpolarizing steps and was deactivated almost completely by depolarization to −20 mV. The time integral of I_hi recorded using low Cl⁻ pipette solution was 114.6 ± 16.5 pQ pF⁻¹ (n = 4) at −140 mV and 271.7 ± 31.4 at −160 mV, values analogous to those obtained using normal pipette solution (Fig. 3).

Figure 1. Hyperpolarization-induced inward currents and increases in cellular EB fluorescence.

A, from an HP of −20 mV, 40 s negative pulses to between −80 mV and −180 mV were applied in 20 mV increments with an interval of 2 min. Images of intracellular EB fluorescence were recorded with 20 s intervals starting at the entry (time 0) of K⁺-free solution containing 25 μm EB and 5 μm glibenclamide; the pipette solution contained 5 mm BAPTA. B, membrane current (I), C, its time integral (Q) and D, mean EB fluorescence in the two nuclei (red) and the cytoplasm (blue) are given on the same time axis. E, typical EB images taken at points a-d in D. Inward currents (I_hi) appeared at the third pulse (-120 mV) and increased progressively with increasing hyperpolarization. All depolarizations to the HP elicited large, voltage-gated, inward currents. Q increased steeply during the pulse and then slowly between pulses because of a small sustained inward current at the HP. Mean nuclear and cytoplasmic EB fluorescence increased progressively as a mirror image of Q. Nuclear EB fluorescence rose steeply after the last pulse without a corresponding increase in Q (arrow). E a, nuclear EB fluorescence was initially undetectable; b, nuclear fluorescence appeared and cytoplasmic fluorescence increased; c, nuclear EB fluorescence increased together with cytoplasmic fluorescence that was brightest at the right caudal region (indicated by an arrow); d, nuclear fluorescence began to increase without a large increase of Q, while cytoplasmic EB fluorescence became more homogeneous and diminished in the caudal region.

Figure 5. Current-voltage relationship for I_hi.

A, hyperpolarization to −160 mV evoked a large increase in I_hi. A triangle ramp pulse from −160 mV to 10 mV with 1-s ascending and descending times was then applied. B, I_hi during the ramp voltage pulse is shown vs. voltage (V). The slow depolarization from −160 to 10 mV evoked a transient voltage-gated inward current that was absent during repolarization. I_K1 recoded as the terminal current elicited by 300 ms pulses near the resting membrane potential in normal Tyrode solution are also plotted. I_K1 peaked at 2.6 ± 0.4 pA pF⁻¹ (n = 10) at −70 mV.

Figure 2. Hyperpolarization-induced activation of irregular inward currents in Na⁺ -free solution.

40 s negative step pulses to −120, −140 and −160 mV were applied. At the HP (-20 mV) the conductance of I_hi was nearly zero, but inward current gradually developed during each hyperpolarization. The time integral of I (Q) rose with some delay and was markedly enhanced by increasing hyperpolarization.

Figure 3. Voltage dependence of I_hi.

A, dependence of Q_hi (charge) carried by I_hi (left axis) - obtained as the time integral of mean I_hi (I_hi,m right axis) during 40 s hyperpolarization steps - on E_m (membrane potential); •, K⁺-free solution (means ± s.e.m., n = 9); □, NMDG solution (n = 6, unpaired cells). Because of development of contracture, I_hi for −160 and −180 mV were not obtained in NMDG solution. For each solution, Q_hi was significantly larger at potentials more negative than −120 mV than at −80 mV. B, voltage dependence of G_hi (I_hi conductance). The conductance was calculated from I_hi.m assuming a linear I_hi-V relationship and a reversal potential of 0 mV.

Fluorescence signals

The mean intensity plots (Fig. 1D) and representative images (Fig. 1E) illustrate that EB fluorescence rose more in the nuclei than in the cytoplasm. The rise in fluorescence in the two nuclei followed similar time courses and reached similar maxima, and in all cases, the mean EB fluorescence increased cumulatively with successive application of negative pulses. It was notable, however, that cytoplasmic EB fluorescence was spatially variable, forming a striation pattern, longitudinal lines or dotted lines with irregular flecks. This variability is probably explained by the influx of EB at various membrane sites and subsequent diffusion and intercalation with nucleic acids and proteins in the cytoplasm. The decline in fluorescence in the caudal portion of the cell, shown in Fig. 1E (c, arrow), during the period between acquisition of the images in panels c and d may reflect the diffusion of EB from that region to other regions of the cytoplasm. With increases in intensity, cytoplasmic fluorescence became more spatially homogeneous (compare panels c and d in Fig. 1E).

If EB did permeate through pores produced by membrane breakdown, the influx should be proportional to the charge carried by I_hi (Q_hi). When Q_hi was estimated from the time integral of I_hi, it was found to rise more steeply during hyperpolarizations than at the HP (Fig. 1C), with EB fluorescence in both the nuclei and cytoplasm rising approximately in parallel with it (Fig. 1D). In addition, with large hyperpolarizations, the nuclear fluorescence continued to rise after depolarization to the HP - e.g. after the cessation of the large I_hi - induced by a pulse to −180 mV (Fig. 1D, arrow in the plot and Fig. 1D, d).

Dependence of I_hi and EB fluorescence on membrane potential

Figure 3 summarizes the voltage-dependent increases in the time integral of I_hi (Q_hi) and the mean amplitude of I_hi during the pulse (I_hi.m; A) and G_hi (B) during a series of 40 s step pulses to between −80 and −180 mV. G_hi was calculated assuming a linear current- voltage relationship for I_hi with a reversal potential at 0 mV (cf. Fig. 5). Q_hi and I_hi.m increased with hyperpolarization, and the slope of the relationship became steeper as the hyperpolarization increased. G_hi increased with hyperpolarization gradually at first, but at potentials more negative than −120 mV was significantly (P < 0.05) larger than at −80 mV. Similar increases were also obtained in NMDG solution at potentials between −80 and −140 mV.

The increases in EB fluorescence produced by 40 s pulses were estimated as the difference between the mean EB fluorescence in the first image, obtained at the onset of the pulse, and in the third image, obtained at the end of the pulse. Plotting nuclear (Fig. 4A) and cytoplasmic (Fig. 4B) fluorescence vs. membrane potential revealed that both signals increased with increasing hyperpolarization. Levels of nuclear fluorescence recorded at potentials more negative than −100 mV were significantly greater than at −80 mV, while cytoplasmic fluorescence was greater at potentials more negative than −140 mV. Nuclear fluorescence was significantly brighter than cytoplasmic fluorescence at all potentials, and the ratio of these increased with increasing hyperpolarization from 1.8 at −80 mV to 5.3 at −160 mV and 6.9 at −180 mV. In addition, the increases in EB fluorescence seen during hyperpolarizing pulses were always significantly larger than the changes occurring before or after the pulses (not shown). The nuclear EB fluorescence before pulses tended to increase gradually with increasing hyperpolarization, while corresponding cytoplasmic fluorescence decreased (Fig. 4). In NMDG solution, the magnitudes of the increases in nuclear and cytoplasmic EB fluorescence during the pulses also increased with hyperpolarization to potentials between −80 and −140 mV.

Figure 4. Voltage dependence of EB fluorescence.

A, nuclear fluorescence. Relative changes in EB fluorescence (means ± s.e.m. of relative units) induced by a 40 s test pulses in the cells used to construct Fig. 3 are plotted against test potential; •, EB fluorescence changed during 40 s negative pulses in K⁺-free solution; □, similar estimation in the presence of NMDG⁺ solution; ○ with a dotted line, fluorescence recorded 40 s before the pulses in K⁺-free solution. Note that at −160 and −180 mV, EB fluorescence was elevated before the pulse. This may have been due to continued EB uptake after the preceding pulse, and may have affected measurements of nuclear fluorescence during subsequent pulses. Nuclear EB fluorescence was significantly larger at potentials more negative than −100 mV than at −80 mV in K⁺-free solition, while it was more negative than −120 mV in NMDG solution. B, cytoplasmic fluorescence. •, relative changes in fluorescence during 40 s hyperpolarizing pulses in K⁺-free solution; □, data obtained in the presence of NMDG solution; ○ with a dotted line, changes recorded in K⁺-free solution during the 40 s before each pulse. The EB fluorescence before pulses decreased with increase of hyperpolarization, presumably due to translocation of accumulated EB. Cytoplasmic EB fluorescence was significantly larger at potentials more negative than −140 mV than at −80 mV in K⁺-free solution, and significanct increase was observed only between −80 and −120 mV in NMDG solution. The difference between nuclear and cytoplasmic EB fluorescence at each E_m was always significant.

Current-voltage relationship of I_hi

The I_hi-V relationship was determined by applying ramp steps from hyperpolarized potentials (Fig. 5A). In response to depolarization from −160 to 10 mV, a transient inward current was evoked at around −60 mV (Fig. 5B). This current was absent during the repolarization from 10 to −160 mV, although the relationship was virtually superimposable otherwise. The I_hi reversal potential was −0.7 ± 3.5 mV (n = 7), which is consistent with the nonspecific nature of G_hi. I_K1, recorded in normal Tyrode solution as a terminal current elicited by 300 ms pulses before switching to K⁺-free solution, are also plotted. The I_K1 conductance at potentials negative to the resting potential was larger than G_hi, i.e. the conductance between −90 and −100 mV was ≈1000 pS pF⁻¹. However, because outward-going I_K1 was 2.6 ± 0.4 pA pF⁻¹ (n = 10) at its peak (at −70 mV), G_hi activated at potentials more negative than −160 mV could induce a net inward current, thereby triggering an action potential.

Effects of LPC on I_hi and EB fluorescence

To assess the effects of LPC on I_hi and EB fluorescence, a train of 40 s negative pulses to −80 mV with 2 min intervals in between was initiated 2 min after switching to K⁺-free solution containing EB and 5 μm glibenclamide (Fig. 6). LPC (10 μm) was introduced about 1 min before the second pulse. Before application of LPC, the negative pulse induced a small I_hi and a small increase in EB fluorescence (Fig. 6B, Ea, and Fa). Upon introduction of LPC, the same hyperpolarization progressively increased both I_hi and Q. EB fluorescence increased almost as a mirror image of the increases in Q, though after the last pulse, nuclear fluorescence continued to increase steeply with no additional accumulation of charge. Cytoplasmic EB fluorescence always increased in parallel with Q, and there was never any secondary increase. Images of EB fluorescence showed that the marker was quite irregularly distributed in both the cytoplasm and nuclei (Fig. 6Fa-d).

Figure 6. Effects of 10 μm LPC on membrane currents and cellular EB fluorescence.

A, negative step pulses to −80 mV were applied with 2 min intervals in between. LPC was applied 1 min after recording the first pulse. B, in the presence of LPC, I_hi often increased abruptly (arrows 1-3). The first response (arrow 1) is expanded in panel D. E, I_hi and EB fluorescence (red circles, the two nuclei; blue circle, cytoplasm) increased roughly in parallel with Q (C). Changes in EB fluorescence occurred mainly during pulses, though nuclear fluorescence also increased steeply after the last pulse (from c to d). F, fluorescent and transparent (at a) images recorded at times a-d in E. a, at this gain, only slight nuclear fluorescence is detectable. b, record obtained after a large, rapid I_hi; note that nuclear fluorescence is clearly detectable, and cytoplasmic fluorescence is increased in all areas, but especially near the left and right edges. c, nuclear fluorescence is high; cytoplasmic fluorescence has increased more slowly and is more homogeneously distributed, though dot-like hot spots are detectable at the upper edge between the two nuclei. d, nuclear fluorescence intensity is saturated at this gain; cytoplasmic fluorescence appears homogeneous.

In the presence of LPC, large and rapid I_hi appeared abruptly during the negative pulses (Fig. 6B, arrows 1 and 3), and similar increases in G_hi also occurred at the HP (arrow 2). As shown in the expanded sweep (D), these I_hi rose to a maximum within 1-10 ms and then decayed very slowly, usually on a time scale of 10 s. The peak current density was 8.1 ± 1.7 pA pF⁻¹, (n = 16) and G_hi was 101.2 ± 21.2 pS pF⁻¹, which is analogous to that activated by a −180 mV pulse and sufficient to elicit action potentials (see Fig. 5). With continued exposure to LPC, the frequency of the rapid I_hi increased: after exposing cells to LPC for 5 min, the rapid I_hi occurred during > 50 % of pulses. Their slow decay made it possible for I_hi to increase progressively during a train of negative pulses. Consequently, the integral of I_hi (Q_hi) rose steeply, pulse-by-pulse, reaching 301.6 ± 53.8 pQ pF⁻¹ (n = 8; Fig. 7A), which corresponds to 94.3 ± 16.8 pS pF⁻¹, at the third pulse applied 5 min after introduction of LPC. The Q_hi increase produced by 10 μm LPC was significantly larger than the time-matched control from the first image after introduction of LPC. A lower concentration of LPC (3 μm) increased Q_hi to 91.5 ± 39.5 pQ pF⁻¹ (n = 5) in 6 min, but this increase was not significant compared with the time-matched control of 28.4 ± 5.4 pQ pF⁻¹ (n = 6). The I_hi reversal potential determined by ramp pulses applied in the presence of LPC was −4.4 ± 3.8 mV(n = 3). As with hyperpolarization-induced currents, 0.5 mm La³⁺ suppressed LPC-induced I_hi to 31 ± 10 % (n = 5) of control.

Figure 7. LPC-induced increases in I_hi and nuclear EB fluorescence.

40 s step pulses to −80 mV were applied from −20 mV with 2 min intervals in between. Increases in Q_hi (time integral of I_hi during the pulse) and EB fluorescence during the pulse are shown. Estimation was conducted in K⁺-free solution (control, ▪; n = 6), K⁺-free solution with 10 μm LPC (•; n = 8 or 6) and NMDG solution with 10 μm LPC (○; n = 5). LPC-containing solution was introduced 1 min before the second pulse. A, the increases in Q_hi during the pulses are plotted vs. time; B, nuclear EB fluorescence; C, cytoplasmic EB fluorescence. In K⁺-free solution, increases of Q_hi, nuclear and cytoplasmic EB fluorescence produced by 10 μm LPC were significantly larger than control from 1 min after its introduction. Symbols depict means ± s.e.m.

Both nuclear and cytoplasmic EB fluorescence increased in parallel with Q_hi after the introduction of 10 μm LPC (Fig. 7), which also produced similar increases of I_hi and EB fluorescence in NMDG solution. The presence of 3 μm LPC induced small and non-significant increase of EB fluorescence in K⁺-free solution: in the nuclei 0.132 ± 0.044 and in the cytoplasm 0.039± 0.007 relative units (n = 5).

The dependence of EB fluorescence on Q_hi was evaluated from the data presented in Figs 3, 4 and 7 (Fig. 8). Surprisingly, increases in Q_hi induced by hyperpolarization or LPC resulted in similar increases in EB fluorescence. Plots of nuclear EB fluorescence vs. Q_hi induced by large negative step pulses or LPC were well fitted by a common Hill's curve for which the Hill coefficient + 0.965, y_max (maximal EB fluorescence increase in 40 s) + 0.936 relative units, and k (dissociation constant, the charge needed to result in 50 % of y_max) + 826 (pQ pF⁻¹).

Figure 8. The dependence of nuclear EB fluorescence on the charge carried by I_hi.

The relations shown were determined from Figs 3, 4 and 7. Increases in EB fluorescence during hyperpolarizing pulses (relative units) are plotted vs. charge (Q_hi); •, negative pulses increasing in amplitude successively from −80 mV to −180 mV in K⁺-free solution; □, −80 mV pulses in the presence of 10 μm LPC from control before LPC introduction until the fifth pulse applied after 7 min in K⁺-free solution. The increase of fluorescence was well fitted by a common dotted line given by Hill's equation:

where k₀ (0.037) corresponds to the y intercept, k₁ (0.936) to y_max (maximum of y), k₂ (0.965) to the rate (Hill coefficient), and k₃ (826) to the magnitude of Q_hi necessary to induce a half y_max; χ² + 0.001954.

Discussion

We have demonstrated clear I_hi dependent increases in EB fluorescence in viable myocardial cells, with greater increases in nuclear than cytoplasmic fluorescence. Whether activated by hyperpolarization or by LPC, both cytoplasmic and nuclear EB fluorescence increased in parallel with increases in the integral of I_hi. Since ion channels are poorly permeable to ethidium cations (but see Cario-Toumanianz et al. 1998), this strongly suggests that Na⁺, Ca²⁺, Mg²⁺ and ethidium⁺ all flowed into cells through electrically or chemically formed pores in the plasma membrane. Indeed, the fact that substitution of Na⁺ with NMDG⁺ (195 Da) had little effect on I_hi or increases in EB fluorescence is consistent with permeation through pores having openings larger than those of typical protein channels.

When exposed to short electrical field stimulations, reversible electric breakdown of the plasma membrane usually occurs when the membrane voltage reaches 0.5-1.5 V (Weaver, 1993). Voltage clamp pulses demonstrated the membrane breakdown by the appearance of irregular membrane currents. In cell-attached patches on isolated frog ventricular myocytes, sudden current increases appear at threshold voltages of 0.6-1.1 V during 7 ms, 2 V ramp pulses (O'Neil & Tung, 1991) and at 0.4 V during 5 ms monophasic or 10 ms biphasic pulses (Tovar & Tung, 1992). In a whole-cell clamp study of human embryonic kidney HEK293 cells, 50-100 ms, 200-400 mV pulses are required to induce currents via membrane breakdown (Akinlaja & Sachs, 1998). Thus, the threshold membrane potential of the electroporation decreases with prolongation of the pulse duration. When the strength-duration relationship for the electroporation threshold was determined in a planar lipid bilayer over a range of 10 ms to 10 s, it was found that at 10 ms the threshold potential is 282 mV, but it is reduced to 167 mV at 10 s and is reduced further in a dose-dependent way by C₁₂E₈, a non-ionic surfactant (Troiano et al. 1998). Analogously, in cardiac myocytes, I_hi activated by 40 s negative steps were accelerated by LPC, an intrinsic surfactant.

The noise-like fluctuation of hyperpolarization-evoked I_hi (Fig. 1 and Fig. 2) (Fig. 6 in Akuzawa-Tateyama et al. 1998) did not exhibit the clear step-like unitary currents seen with 500 pS pores in unmodified planar lipid bilayers (Melikov et al. 2001), which suggests that I_hi were generated by small conductances through numerous pores. On the other hand, in the presence of LPC, macroscopic I_hi appeared to consist mainly of a large, rapid component with a peak conductance of 101 pS pF⁻¹ and a slow relaxation (Fig. 6). The amplitude of this inward-going current was 7.1 pA pF⁻¹ at −70 mV and would therefore be expected to overcome the outward-going I_K1 (Fig. 5, peak amplitude + 2.6 pA pF⁻¹) and elicit action potentials. The large, rapid openings may reflect extensive membrane breakdown that is probably facilitated by an LPC-induced increase of membrane fluidity (Weltzien, 1979). Despite the difference in the activation rate, pores produced in the presence of LPC appeared not to have larger openings, as the permeation efficacy of EB was not enhanced by LPC (Fig. 8).

The recording of I_hi in NMDG solution completely rules out a contribution to I_hi by Na⁺ channel currents and by I_f that are Na⁺ and K⁺ permeable and activated by large hyperpolarizations in ventricular myocytes with a half-activation voltage of −150 mV in dogs and −145 mV in guinea-pigs (Ranjan et al. 1998)_. Strong inward rectification through K⁺ channels also does not explain I_hi, as E_K would be more negative than −150 mV, and I_hi showed only slight rectification (Fig. 5). Involvement of I_K(ATP) was excluded, as activation of I_K(ATP) in the absence of glibenclamide produced prominent outward current at −20 mV (Y.-M. Song and R. Ochi, unpublished observations). Finally, the contribution of Cl⁻ channels to I_hi is negligible, as replacement of 90 % of the Cl⁻ in the pipette solution with aspartate had little effect on Q_hi. Thus, ion channel currents played little or no role in the generation of I_hi.

The longitudinal fluorescence pattern in the myocytes is thought to originate from mitochondrial EB fluorescence, while the striated pattern may be indicative of EB binding to proteins regularly arrayed within the sarcomere. Overlap of longitudinal and transverse fluorescence would produce the dotted longitudinal pattern. Thus, both cytoplasmic and nuclear EB fluorescence is affected by the availability of various target molecules. In the present study, increases in cytoplasmic and nuclear EB fluorescence occurring during the hyperpolarizing pulses paralleled the accumulation of charge, which suggests that EB flowed into the myocyte together with primary external cations and then quickly (within 20 s) completed its interaction with target molecules, thereby enhancing the fluorescent signal. The nuclear signal followed the increase in G_hi quite faithfully (Figs 3, 4 and 7) and increased along with the increases in Q_hi, which confirmed that influx of EB was crucial among various intermediary reactions for the rapid increase of fluorescence. Nevertheless, when large increases in I_hi were evoked, steep increases in nuclear EB fluorescence developed that persisted for 40-80 s after the pulse (Fig. 1 and Fig. 6, after the last pulse). Perhaps the large influx of external ions facilitated DNA-intercalation within the nuclei preceding the development of contraction.

The increase in intracellular Ca²⁺, Na⁺ and ethidium⁺ mediated by I_hi during a 40 s negative pulse to −160 mV was calculated from the Q_hi. Assuming free diffusion of major extracellular and intracellular ions through the pores, 50 % of the influx would be due to the movement of cations from the extracellular compartment, and the influx of a given cation would be proportional to its extracellular concentration. A ventricular cell was assumed to be a cylinder, 100 μm long and 20 μm in diameter, with a membrane capacity of 50 pF. When hyperpolarized to −160 mV, the increase in the current integral was 407 pQ pF⁻¹ (Fig. 4); a net increase of 20 nQ occurred within the cell, half of which was due to the cation influx. Approximating an equivalent charge of 10⁵ Q, a 10 nQ influx would increase the internal concentration (m) of Na⁺ by:

The expected increase in [Ca²⁺]_i (m) would be:

and that of [EB]_i (m) would be:

Similar increases would be also expected with −80 mV pulses after 5-7 min in the presence of 10 μm LPC. The increase in [Ca²⁺]_i mediated by I_hi could be as large as 4.1 × 10⁻⁵m. Reflecting this large increase in [Ca²⁺]_i, gradual shortening of the myocytes developed after large I_hi elicited by −160 or −180 mV pulses or by application of 10 μm LPC, even though the pipette solution contained 5 mm BAPTA (not shown). The increase in EB fluorescence elicited by a −160 mV pulse was 0.37 relative units in the nuclei and 0.07 units in the cytoplasm (Fig. 4), which corresponds to 9.3 × 10⁻⁶m and 1.8 × 10⁻⁶m, respectively. These increases in nuclear and cytoplasmic EB fluorescence are, respectively, 16.3 times and 3.2 times greater than those predicted by the calculation. This means that a large part of the observed increases in EB fluorescence was probably due to increases in the quantum yield of EB intercalated between hydrophobic substrates represented by nucleic acids, which is analogous to labelling of nuclei in dead cells (Olmsted & Kearns, 1977; Lim et al. 2001).

The LPC concentration in the coronary sinus of ischaemic patients may rise by as much as 100 μm in response to rapid atrial pacing (Sedis et al. 1990). LPC causes depolarization of cardiac myocytes, and numerous reports describe its modulation of various ion channels (Lopaschuk, 2001). The depolarization may originate from blockade of the inward rectifier K⁺ channel (I_K1) (Kiyosue & Arita, 1986) or a negative shift of the activation-voltage relationship of the Na⁺ channel (Undrovinas et al. 1992). However, LPC also directly potentiates Ca²⁺ accumulation during periods of increased membrane permeability (Sedis et al. 1983; Gailis et al. 2001), and we have shown that 10 μm LPC induces large I_hi that can trigger action potentials in ventricular cells. For those reasons, we think electroporation (or membrane breakdown) should be considered as the primary mechanism by which LPC induces depolarization and Ca²⁺ accumulation.

Strong electric field pulses elicit localized elevations in Ca²⁺ in cultured adherent cells due to electroporation (Teruel & Meyer, 1997). In addition, low intensity electroporative pulses induce reversible permeabilization of HeLa cell membranes, allowing limited Ca²⁺ signalling like that evoked by hormonal stimulation (Bobanovi'c et al. 1999). One might expect electroporation-induced Ca²⁺ signals to exert an effect on cellular function, and once the plasma membrane pores are formed, permeation of other ions, such as Na⁺, K⁺, Mg²⁺ and Cl⁻, could also invoke significant functional changes within cells. I_hi induced by large hyperpolarizations or by 10 μm LPC is sufficient to induce arrhythmic excitation and contracture in myocytes; even when the effect is small, it could affect the activity of cells secondarily by causing them to expend ATP to pump out Na⁺. In situ, I_hi is negligible in normal cells at a normal membrane potential. But during the pacing therapies or in the presence of metabolically compromised cells that release synthesized LPC, even if its concentration is only in the micromolar range, I_hi could be sufficient to induce a meaningful increase of cation uptake.