Relationship between transient outward K+ current and Ca2+ influx in rat cardiac myocytes of endo‐ and epicardial origin

1 The transient outward K+ current (Ito) is a major repolarizing ionic current in ventricular myocytes of several mammals. Recently it has been found that its magnitude depends on the origin of the myocyte and is regulated by a number of physiological and pathophysiological signals. 2 The relationship between the magnitude of Ito, action potential duration (APD) and Ca2+ influx (QCa) was studied in rat left ventricular myocytes of endo‐ and epicardial origin using whole‐cell recordings and the action potential voltage‐clamp method. 3 Under control conditions, in response to a depolarizing voltage step to +40 mV, Ito averaged 12.1 ± 2.6 pA pF−1 in endocardial (n= 11) and 24.0 ± 2.6 pA pF−1 in epicardial myocytes (n= 12; P < 0.01). APD90 (90 % repolarization) was twice as long in endocardial myocytes, whereas QCa inversely depended on the magnitude of Ito. L‐type Ca2+ current density was similar in myocytes from both regions. 4 To determine the effects of controlled reductions of Ito on QCa, recordings were repeated in the presence of increasing concentrations of the Ito inhibitor 4‐aminopyridine. 5 Inhibition of Ito by as little as 20 % more than doubled QCa in epicardial myocytes, whereas it had only a minor effect on QCa in myocytes of endocardial origin. Further inhibition of Ito led to a progressive increase in QCa in epicardial myocytes; at 90 % inhibition of Ito, QCa was four times larger than the control value. 6 We conclude that moderate changes in the magnitude of Ito strongly affect QCa primarily in epicardial regions. An alteration of Ito might therefore allow for a regional regulation of contractility during physiological and pathophysiological adaptations.

The transient outward K¤ current (Ito) is a major repolarizing ionic current in ventricular myocytes of several species (Josephson et al. 1984;Giles & Imaizumi, 1988;Litovsky & Antzelevitch, 1988;Beuckelmann et al. 1993;Varro et al. 1993;Wettwer et al. 1993). Accordingly, its inhibition leads to a significant prolongation of the AP (Josephson et al. 1984;Mitchell et al. 1984). Recently, several studies have demonstrated that the magnitude of Ito is considerably altered in response to a number of physiological and pathophysiological influences like á1_adrenergic activation (Apkon & Nerbonne, 1988), cardiac hypertrophy (Tomita et al. 1994), heart failure (N abauer et al. 1993), hypo-or hyperthyroidism (Shimoni et al. 1992(Shimoni et al. , 1995, diabetes (Jourdon & Feuvray, 1993), anoxia (Thierfelder et al. 1994) and development (Jeck & Boyden, 1992). This suggests that changes in the magnitude of Ito might constitute a pathway for altering APD and consequently Ca¥ influx and cardiac contractility. Moderate reductions in Ito have been observed in several pathophysiological conditions (Wickenden et al. 1998;N abauer & K a ab, 1998), but the effects of this reduction on Ca¥ influx in different regions of the left ventricle have not yet been investigated. Journal of Physiology (1999), 519.3, pp. 841-850 841 Relationship between transient outward K¤ current and Ca¥ influx in rat cardiac myocytes of endo-and epicardial origin Tilmann Volk* †, Thi Hong-Diep Nguyen*, Jobst-Hendrik Schultz* and Heimo Ehmke * † 1. The transient outward K¤ current (Ito) is a major repolarizing ionic current in ventricular myocytes of several mammals. Recently it has been found that its magnitude depends on the origin of the myocyte and is regulated by a number of physiological and pathophysiological signals. 2. The relationship between the magnitude of Ito, action potential duration (APD) and Ca¥ influx (QCa) was studied in rat left ventricular myocytes of endo-and epicardial origin using whole-cell recordings and the action potential voltage-clamp method. 3. Under control conditions, in response to a depolarizing voltage step to +40 mV, Ito averaged 12·1 ± 2·6 pA pF¢ in endocardial (n = 11) and 24·0 ± 2·6 pA pF¢ in epicardial myocytes (n = 12; P < 0·01). APD90 (90% repolarization) was twice as long in endocardial myocytes, whereas QCa inversely depended on the magnitude of Ito. L-type Ca¥ current density was similar in myocytes from both regions. 4. To determine the effects of controlled reductions of Ito on QCa, recordings were repeated in the presence of increasing concentrations of the Ito inhibitor 4-aminopyridine. 5. Inhibition of Ito by as little as 20% more than doubled QCa in epicardial myocytes, whereas it had only a minor effect on QCa in myocytes of endocardial origin. Further inhibition of Ito led to a progressive increase in QCa in epicardial myocytes; at 90% inhibition of Ito, QCa was four times larger than the control value. 6. We conclude that moderate changes in the magnitude of Ito strongly affect QCa primarily in epicardial regions. An alteration of Ito might therefore allow for a regional regulation of contractility during physiological and pathophysiological adaptations.
We therefore used whole-cell recordings and the AP voltage-clamp method to analyse the quantitative effects of a graded inhibition of Ito on the shape and size of Ca¥ influx. Since Ito is not uniformly expressed in the left ventricle (Litovsky & Antzelevitch, 1988;Fedida & Giles, 1991;Clark et al. 1993;Wettwer et al. 1994), we have separately investigated myocytes from endocardial and epicardial regions.

Isolation of myocytes
Female Sprague-Dawley rats, weighing 180-220 g, were anaesthetized with intraperitoneal injection of Inactin (thiobutabarbital-sodium, 100 mg (kg body mass)¢; Byk Gulden, Konstanz, Germany). The heart was quickly excised and placed into cold (4°C) cardioplegic solution, where it stopped beating immediately. Left ventricular myocytes were isolated according to the method described by Isenberg & Klockner (1982). Briefly, the aorta was cannulated and retrogradedly perfused with nominally Ca¥-free modified Tyrode solution at 37°C for 5 min. Perfusion pressure was 75 mmHg, and all solutions were bubbled with 100% oxygen. The perfusion was continued for 15 min with 20 ml of the same solution containing collagenase (Type CLS II, 200 U ml¢; Biochrom KG, Berlin, Germany) and protease (Type XIV, 0·7 U ml¢; Sigma), and the solution was recirculated. Finally, the heart was perfused with modified Tyrode solution containing 100 ìÒ Ca¥ for another 5 min. After the perfusion, the left ventricular free wall was separated from the rest of the heart, and epi-and endocardial tissue was carefully isolated from the left ventricular free wall with fine forceps and placed into separate cups.
To further disaggregate the tissue pieces, they were gently shaken at 37°C for some minutes, filtered through a cotton mesh and allowed to settle for half an hour. Cells were stored at room temperature in modified Tyrode solution containing 100 ìÒ Ca¥. Only single rodshaped cells with clear cross striations and no spontaneous contractions were used for experiments. All animal experiments were conducted in accordance with institutional guidelines and approved by local authorities (permit no. 37-9185·81Ï114Ï95).

Patch-clamp technique
The ruptured-patch whole-cell configuration was used as described previously (Hamill et al. 1981). Myocardial cells were transferred to an elongated chamber (2·5 mm ² 20 mm), mounted on the stage of an inverted microscope (Axiovert 25, Zeiss, Jena, Germany), and initially superfused with control solution. All experiments were performed at room temperature (22-26°C). During recordings, liquid flow was stopped to minimize electric noise. The flow rate was 15 ml min¢ and solution exchange of the bath was achieved within seconds. Patch pipettes were pulled from borosilicate glass (GC150-15, Clark Electromedical Instruments) using a P-87 Puller (Sutter Instruments). Pipette resistance (Rpip) averaged 3·6 ± 0·1 MÙ (n = 38) with potassium glutamate in the pipettes and control solution in the bath. Currents were recorded using an EPC_9 amplifier (HEKA Elektronik), controlled by a Power-Macintosh (Apple Computer) and the Pulse-Software (HEKA Elektronik). Membrane voltage (Vm) and APs were recorded in the zero current-clamp mode and ionic currents in the voltage-clamp mode. For AP voltage-clamp recordings, APs were recorded at the beginning of the experiments and used as a voltage template in the voltage-clamp mode of the amplifier (Doerr et al. 1989). Membrane capacitance (Cm) and series resistance (Rs) were calculated using the automated capacitance compensation procedure of the EPC_9 amplifier. Rs averaged 6·5 ± 0·3 MÙ (n = 38) and was compensated by 85% leading to an average effective Rs of •1·0 MÙ. Accordingly, at the largest recorded currents of about 6 nA observed at a pipette potential (Vpip) of +60 mV, the voltage error was less than 6 mV or 10%. Cm averaged 120·5 ± 5·8 pF (n = 35) and was similar in both endoand epicardial myocytes. The reference electrode of the amplifier headstage was bathed in pipette solution in a separate chamber and was connected to the bath solution via an agar-agar bridge filled with pipette solution. Vpip and Vm were corrected for liquid junction potentials occurring at the bridge-bath junction (13·3 mV for potassium glutamate pipette solution and control solution in the bath). Whole-cell currents were low-pass filtered at 1 kHz and sampled at 5 kHz. APs were sampled at 1 kHz. Whole-cell data were analysed using the Pulse-Fit (HEKA Elektronik) and IGOR (WaveMetrics, Lake Oswego, OR, USA) software. Data are given as means ± s.e.m., and n is the number of experiments. Statistical significance was calculated using the appropriate version of Student's t test or the Mann-Whitney U test after an initial twoway ANOVA using the software PRISM (Graph-Pad Inc., San Diego, CA, USA). Differences with P < 0·05 were considered statistically significant.

Identification and separation of currents
The Ca¥-independent depolarization-induced outward current observed in rat ventricle has been separated into two distinct K¤ currents: a rapidly inactivating component (Ito) and a sustained (steady-state) component (Iss) (Apkon & Nerbonne, 1991). Ito is sensitive to 4_AP, while Iss is blocked by tetraethylammonium (TEA). Nevertheless, quantification and separation of the two components has been a matter of debate (Shimoni et al. 1995), because both currents are rate dependent and inhibition of Ito by 4_AP shows a reverse use dependence resulting in a reduced inhibition at higher stimulation rates (Campbell et al. 1993). We therefore used a low stimulation rate of 0·33 Hz or less for all voltage-clamp, AP voltage-clamp or AP recordings. Ito was calculated as the transient component of the outward current, that is the difference between the peak and the steady-state current, which correlates quite well with the 4_AP-sensitive current (data not shown) and thus gives a good estimate for Ito. Using the AP voltage-clamp method, AP-induced Ca¥ currents and Ca¥ influx were estimated by subtracting recordings made in the presence of 300 ìÒ Cd¥ in the bath solution from those in the absence of Cd¥. This difference represents the Cd¥-sensitive current. Cd¥ at a concentration of 300 ìÒ is a relatively nonspecific blocker of Ca¥ channels (Fox et al. 1987) but has only minor effects on Ito, whereas other, more specific inhibitors of Ca¥ channels like dihydropyridines or D600 are potent inhibitors of Ito (Gotoh et al. 1991;Lefevre et al. 1991). Cd¥ also inhibits the Na¤-Ca¥ exchange current (Kimura et al. 1987) and the currents depending on Ca¥ influx. Therefore, the pipette solution contained 10 mÒ of the Ca¥ chelator EGTA, which should have been sufficient to prevent Ca¥-dependent currents from activating. In the absence of intracellular Ca¥, the Na¤-Ca¥ exchanger does not contribute to the whole-cell current (Allen & Baker, 1985). Since previous studies (Tytgat et al. 1990;Bouchard et al. 1995) could not identify T-type currents in rat left ventricle, their contribution to the Cd¥-sensitive component is very unlikely. Because we wanted to evaluate the influence of Ito on the AP-induced Ca¥ influx, we could not replace intracellular K¤ by impermeable cations. Ito is known to be influenced by divalent cations (Agus et al. 1991): at a concentration of 300 ìÒ, Cd¥ shifts the steady-state activation and inactivation curves by about 15-20 mV to more positive potentials. Maximal activation of Ito, however, is observed at about Vm = 20 mV. This value was always exceeded in the AP voltage-clamp experiments, since the overshoot of the rat AP reaches values of about Vm = 30 mV. Furthermore, we started the AP voltage-clamp experiments from the normal resting membrane potential of Vm = −90 mV, which is out of range of the steadystate inactivation. Thus, the influence of a 20 mV shift in the steady-state activation and inactivation on the present results should be minimal. Subtracting current recordings made in the absence of Ca¥ from those made with Ca¥ present in the bath solution is another way of estimating Ca¥ currents. However, Ito-like K¤ currents, which are activated by removal of external Ca¥, have been described (Inoue & Imanaga, 1993) and therefore this protocol can only be used in the absence of internal K¤ to suppress these currents. Finally, Ca¥ currents typically exhibit run-down during an experiment. Scamps et al. (1990) found a reduction in Ca¥ currents of less than 10% within 30 min. In our experiments, the estimation of Ca¥ currents was usually performed within the first 10 min, and therefore rundown is negligible. We conclude therefore that under our experimental conditions the Cd¥-sensitive currents appear to give the best estimate of the L-type Ca¥ current.

Transient outward K¤ current
At the beginning of each experiment, after the conventional ruptured patch whole-cell configuration was established, the resting membrane potential was recorded in the zero current-clamp mode of the amplifier. The average Vm was −87·0 ± 0·4 mV (n = 36, no. of rats = 11) and was not significantly different in endo-and epicardial myocytes.
Only cells with a Vm of at least −85 mV were used for experiments. Panels A and B in Fig. 1 show representative current recordings of Ito in, respectively, an epicardial and an endocardial myocyte, activated by depolarizing voltage pulses. Sodium currents were inactivated by an initial depolarization to Vpip = −50 mV for 20 ms, and Ca¥ currents were inhibited by the presence of 300 ìÒ Cd¥ in the bath solution. Ito activated at pipette potentials above Vpip = −20 mV and continued to increase up to Vpip = 60 mV. Ito was much larger in the epicardial myocyte, but had similar activation and inactivation properties in both cells. Figure 1C shows average current-voltage (I-V) relations of Ito recorded in 11 endocardial and 12 epicardial myocytes. Ito was quantified as the difference between the peak current and the current at the end of the voltage pulse (after 600 ms; see Methods). Currents were normalized to cell capacitance to correct for different cell sizes and are thus given in picoamps per picofarad (pA pF¢). The threshold of activation was around Vpip = −20 mV in myocytes of both regions, and Ito increased linearly up to the maximal applied voltage of Vpip = 60 mV. At Vpip = 40 mV, Ito averaged 12·1 ± 2·6 pA pF¢ in endocardial myocytes and 24·0 ± 2·6 pA pF¢ in epicardial myocytes (P < 0·01). The current decay was best fitted monoexponentially and the inactivation time constant was identical in both regions Impact of Ito on Ca¥ influx J. Physiol. 519.3 843 Figure 1. Magnitude of Ito in epi-and endocardial myocytes Representative outward currents recorded from an epicardial (A) and an endocardial (B) myocyte. Current was normalized to cell capacitance to compensate for different cell sizes. Currents were activated by voltage pulses of 600 ms duration from a holding potential of −90 mV to values ranging from −60 to +60 mV in steps of 20 mV. Prior to each depolarization, Vm was clamped on −50 mV for 20 ms to inactivate Na¤ currents. C, average I-V relation of Ito recorded from 12 epicardial (0) and 11 endocardial (1) myocytes. Ito was quantified by subtracting the current at the end of the voltage pulse (600 ms) from the peak current.
AP, Ito and Ca¥ influx in endo-and epicardial myocytes Figure 2 compares APs, AP-induced Cd¥-sensitive currents, Ito and Cd¥-sensitive currents activated by a rectangular voltage step to Vpip = 0 mV in representative endo-and epicardial myocytes; average results are summarized in Table 1. The epicardial AP showed a pronounced early repolarization, whereas the endocardial AP was considerably longer and the plateau potential was more positive (Fig. 2A). The strong early repolarization observed in epicardial myocytes is reflected in APD0mV (i.e. the time span from AP upstroke until Vm falls to below 0 mV), which was on average about three times shorter (27·9 ± 8·5 ms, n = 13) than in endocardial myocytes (85·2 ± 20·6 ms, n = 11, P < 0·05). The APD90 (i.e. the time span from AP upstroke until 90% repolarization) differed only by a factor of two (118 ± 20 vs. 209 ± 38 ms, P < 0·05). Figure 2B shows Cd¥-sensitive currents recorded with the AP voltage-clamp method. The duration of the Cd¥-sensitive current was much greater in the endocardial than in the epicardial myocyte. Total Ca¥ influx (QCa), calculated as the integral of the AP-induced Cd¥-sensitive current and normalized to the cell capacitance, was larger in the endocardial cell; average values were 526 ± 107 fC pF¢ (n = 11) in endocardial myocytes and 250 ± 71 fC pF¢ (n = 13) in epicardial myocytes (P < 0·05  Figure 2. AP voltage-clamp recordings from epi-and endocardial myocytes A, APs recorded in a representative epicardial (left) and endocardial (right) myocyte. Early repolarization was much stronger in the epicardial myocyte. B, Cd¥-sensitive current recorded with the AP voltage-clamp method using the APs depicted in A. Peak current was higher in the epicardial myocyte, but total QCa (integral of the current) was much higher in the endocardial myocyte. C, I-V relation, drawn using the AP_induced Cd¥-sensitive current (B) and the corresponding AP (A). The peak of the inward current was recorded at a more positive potential in the endocardial myocyte (20 vs. 5 mV). D, Cd¥-sensitive current, activated by a rectangular depolarizing voltage pulse to 0 mV. The Na¤ current was inactivated by a 20 ms step from a holding potential of −90 mV to −50 mV. Activated by the same command potential, Cd¥sensitive currents were similar in size and inactivation kinetics in the epi-and endocardial myocyte. E, outward current, activated by a rectangular depolarizing voltage pulse to +40 mV. The Na¤ current was inactivated by a 20 ms step from a holding potential of −90 mV to −50 mV. Ito was much more prominent in the epicardial myocyte, which explains the strong early repolarization and the smaller QCa. All recordings were made in the same epi-or endocardial myocyte.
currents were not significantly different between both regions (epicardial, 11·5 ± 0·8 pA pF¢, n = 13; endocardial, 10·2 ± 0·5 pA pF¢, n = 11; P = 0·14). The I-V relations shown in Fig. 2C were computed using the cell's AP and the corresponding AP-induced Cd¥sensitive current. It should be noted that they differ from the classical bell-shaped I-V relation of L-type Ca¥ currents in that their shape also depends on the timedependent activation and inactivation properties of the Ca¥ channels. Thus, the magnitude of the current recorded depends not only on the Vm at which it is recorded, but also on the time when Vm is reached during the AP. These I-V curves explain the differences in the peak of the APinduced Cd¥-sensitive currents: in the epicardial myocyte, the current peaked at Vpip = 5 mV, whereas it peaked at Vpip = 20 mV in the endocardial myocyte. The bell-shaped I-V relation of the L-type Ca¥ current has its maximum around 0 mV (McDonald et al. 1994), thus the AP-induced Ca¥ current would be maximal if the membrane potential was 0 mV immediately after the overshoot, when the open probability of the Ca¥ channels is maximal. The strong early repolarization to values of Vm 0 mV in the epicardial myocytes favours a high peak Ca¥ current. In contrast, the weaker early repolarization observed in endocardial myocytes leads to more depolarized plateau potentials and thus a lower electrochemical driving force for Ca¥ ions. The average values for Vm, at which the peak AP-induced Cd¥sensitive current was observed (Vm,peak), were 2·4 ± 2·5 mV in epicardial myocytes and 14·2 ± 3·4 mV in endocardial myocytes (P < 0·05). Although in endocardial myocytes Vm,peak was significantly higher than in epicardial myocytes, both values were close to the maximum of the bell-shaped L-type Ca¥ current I-V. This explains why the average peak values of the AP-induced Cd¥-sensitive currents were not significantly different.

Quantification of the effect of 4_AP on Ito
4_AP is an inhibitor of Ito with a well-characterized pharmacology (Castle & Slawsky, 1993). In the present study, we wanted to use 4_AP for a graded inhibition of Ito and therefore established a dose-response relation to estimate the inhibitory effect of defined concentrations of 4_AP under our experimental conditions. Figure 3A

. Influence of inhibition of Ito on APs and AP-induced Ca¥ currents in epi-and endocardial myocytes
A, APs and corresponding AP-induced Cd¥-sensitive currents recorded in a representative epicardial myocyte subjected to increasing inhibition of Ito with 4_AP. B, similar recordings in an endocardial myocyte. Cell capacitance was similar with 136 pF in the epicardial and 144 pF in the endocardial myocyte. The resting membrane potential was not affected by 4_AP. With increasing concentrations of 4_AP in the bath solution, APs became longer and the plateau potential increased. The increase in APD was accompanied by an increased duration, but reduced peak of the AP-induced Cd¥-sensitive current. These effects were more pronounced in the epicardial myocyte. With maximal inhibition of Ito, APs and AP-induced Cd¥-sensitive currents were similar in both myocytes. Artefacts in the current recordings prior to the activation of the Cd¥-sensitive current resulting from capacitive and Na¤ currents were cut away. The spike of the AP-induced Cd¥-sensitive current recorded from the endocardial myocyte at 4 mÒ 4_AP is an artefact secondary to the slow upstroke of the corresponding AP and was not observed in similar recordings.
Ito until it was almost completely inhibited at 4 mÒ ( Fig. 3A and B). Since we could not detect any differences in the inhibitory effect of 4_AP on Ito between endo-and epicardial myocytes, all experiments were pooled (Fig. 3C). The doseresponse relation was fitted assuming Michaelis-Menten kinetics for the effect of 4_AP on Ito.

Influence of inhibition of Ito on AP and Ca¥ influx in endo-and epicardial myocytes
To evaluate the effect of an inhibition of Ito on Ca¥ influx, the following experimental protocol was used: after establishing the whole-cell configuration, APs were recorded under control conditions and in the presence of 0·3, 1 and 4 mÒ 4_AP, which corresponds to an inhibition of Ito by 0, 20, 50 and 86% as demonstrated in Fig. 3C. Subsequently, by using the AP voltage-clamp method, the membrane potential was clamped on these four APs in the absence and presence of Cd¥ in the bath solution. Subtraction of the corresponding currents revealed the AP-induced Cd¥sensitive currents for each AP. Figure 4 shows APs and corresponding AP-induced Cd¥-sensitive currents under increasing inhibition of Ito in a representative epicardial (Fig. 4A) and endocardial (Fig. 4B) myocyte. 4_AP had no effect on the resting membrane potential. Under control conditions, the AP of the endocardial myocyte was longer than that of the epicardial myocyte (see also Fig. 2), and the plateau potential was more positive. However, with increasing inhibition of Ito, the increase in APD and plateau potential in the epicardial myocyte was greater than in the endocardial myocyte. At maximal inhibition of Ito by 4 mÒ 4_AP, both APs had approximately the same duration and plateau potential. The peak of the AP-induced Cd¥sensitive currents decreased with increasing inhibition of Ito, but the duration of the current increased. Both effects were more pronounced in the epicardial myocyte. The high peak current observed in the endocardial myocyte at 4 mÒ 4_AP is an artefact due to a sudden increase in access resistance during the AP recording. The resultant slow upstroke of this AP led to an artificial high peak current which was not recorded in other similar experiments. Figure 5A summarizes similar results obtained in 11 endocardial and 13 epicardial myocytes. Under control conditions, APD0mV was three times longer in endocardial myocytes than in epicardial myocytes. With increasing inhibition of Ito, APD0mV increased in both cell types until similar values were reached at maximal inhibition. In Fig. 5B, Ca¥ influx is plotted versus the concentration of Impact of Ito on Ca¥ influx J. Physiol. 519.3 847 Figure 5. Effect of inhibition of Ito on APD0mV, QCa, ICd,peak and Vm,peak in epi-and endocardial myocytes Curves were calculated from average data recorded in every experiment. Single recordings were made similar to those shown in Fig. 4. The effects of 0, 0·3, 1 and 4 mÒ 4_AP on APD0mV (A), QCa (B), the peak AP-induced Cd¥-sensitive current (ICd,peak) (C), and the membrane voltage at which ICd,peak was recorded (Vm,peak) (D) are illustrated. 0, data recorded in epicardial myocytes; 1, data recorded in endocardial myocytes. ICd,peak was estimated as the maximal inward current after the initial capacitive current. Note that differences between epicardial and endocardial myocytes disappear with increasing inhibition of Ito. Significance of difference from control values (0 mÒ 4_AP) was tested using Student's paired t test (*P < 0·05; ** P < 0·01; *** P < 0·001). 4_AP in the bath solution. Under control conditions, Ca¥ influx was significantly larger in endocardial myocytes. Inhibition of Ito by 20% (0·3 mÒ 4_AP) more than doubled Ca¥ influx from an average of 250 ± 71 to 636 ± 72 fC pF¢ in epicardial myocytes (P < 0·001). In myocytes of endocardial origin, the effect of inhibition of Ito was much less pronounced: at 20% inhibition, Ca¥ influx increased by a factor of 1·3 from 526 ± 107 to 678 ± 112 fC pF¢ (P < 0·01). With maximal inhibition of Ito, Ca¥ influx increased by a factor of 4·0 to 1011 ± 136 fC pF¢ in epicardial myocytes (P < 0·001), but only by a factor of 1·9 to 994 ± 168 fC pF¢ in endocardial myocytes (P < 0·01). Interestingly, at 20% inhibition of Ito, Ca¥ influx attained similar values in epi-and endocardial myocytes. Figure 5C illustrates the effect of inhibition of Ito on the peak of the AP-induced Cd¥-sensitive current. With increasing inhibition of Ito, the peak current decreased significantly in both endo-and epicardial myocytes. At maximal inhibition of Ito, the peak current was reduced by about 50%. In parallel with the decrease in the peak Cd¥sensitive current, the membrane potential at which the peak Cd¥-sensitive current was observed during the AP (Vm,peak) increased (Fig. 5D). Under control conditions, Vm,peak was higher in endocardial myocytes, while peak Cd¥-sensitive current was lower (see Fig. 2C). With maximal inhibition of Ito, Vm,peak increased by 29·9 ± 2·0 mV in epicardial and by 15·9 ± 3·0 mV in endocardial myocytes (P < 0·001). Thus, progressive inhibition of Ito led to a higher plateau potential which reduced the driving force for the AP-induced Ca¥ current. This effect was more pronounced in epicardial myocytes.

DISCUSSION
The present investigation demonstrates that a differential distribution of Ito is the most important cause for different AP waveforms in endo-and epicardial myocytes from adult rat left ventricle. Furthermore, the results indicate that the magnitude of Ito has a major impact on Ca¥ transients and the amount of Ca¥ influx during an AP.

Accuracy of estimated Ca¥ influx
We tried to quantify AP-induced Ca¥ currents and Ca¥ influx under as much physiological conditions as possible. The composition of internal and external solutions resembled mostly the normal distribution of ions. However, some compromises had to be made. Internal Ca¥ was buffered using 10 mÒ EGTA in the pipette solution to inhibit Ca¥-dependent currents (as discussed in Methods). Buffering of internal Ca¥ may slow L-type Ca¥ current inactivation in two ways: first, Ca¥ ions entering the cell via L-type channels are buffered and thus cannot bind to the proposed Ca¥-binding site of the channel (de-Leon et al. 1995); second, Ca¥ release into the fuzzy space close to the L-type channel is prevented by emptying of the sarcoplasmic reticulum (Sham, 1997). However, the normal Ca¥induced inactivation is well preserved, even in the presence of EGTA, probably because the proposed Ca¥-binding site is close to the channel mouth and cannot be reached quickly enough by EGTA (de-Leon et al. 1995). Furthermore, using the AP voltage-clamp method, Linz & Meyer (1998) demonstrated that Ca¥ currents recorded in the presence of 10 mÒ EGTA in the pipette solution differed only slightly in the late phase of the AP from those recorded in the absence of internal Ca¥ buffering. Finally, we performed experiments with minimal internal Ca¥ buffering (0·5 mÒ EGTA, free [Ca¥] = 100 nÒ) and found a similar increase in Ca¥ influx induced by inhibition of Ito in epicardial myocytes compared with that found in the experiments in which 10 mÒ EGTA was present in the pipette solution (n = 4, data not shown). Because we used 4_AP to inhibit Ito in order to determine its contribution to APD and Ca¥ influx, a key question concerns the specificity of 4_AP for Ito. 4_AP does not affect L-type Ca¥ currents (Clark et al. 1993). Apkon & Nerbonne (1991) have shown that the transient component of the depolarization-activated outward current is sensitive to 4_AP, whereas the steady-state current is insensitive to 4_AP, but can be blocked by TEA. Thus, by using 4_AP we should have inhibited only the transient component of the depolarization-activated outward current (Ito). However, Kv1.5, a channel gene which codes for a slowly inactivating current, and which is expressed in the rat ventricle at the mRNA (Dixon & McKinnon, 1994) and at the protein level (Gidh-Jain et al. 1996), is also inhibited by 4_AP (Po et al. 1993) and thus may contribute to the 4_AP-sensitive current. Consistent with an alteration of the delayed rectifier current, we observed in some experiments a reduction in the steady-state current at the end of the voltage pulse induced by 4_AP. On average, however, only a minor component of the total delayed rectifier current was inhibited by 4_AP. In the present study, stimulation rates of 0·33 Hz or less were used, whereas normal heart rates in rats are much higher. This higher stimulation rate could potentially alter the effects of Ito on AP waveform and Ca¥ influx. Interestingly, Shimoni et al. (1995) found Ito to be more rate sensitive in endocardial than in epicardial myocytes and identified differences in recovery kinetics as the cause: Ito in endocardial myocytes recovered with smaller time constants. As a consequence, at higher stimulation rates APD increased to a larger extent in endocardial than in epicardial myocytes. Thus at higher stimulation rates, the difference in Ca¥ influx between endo-and epicardial myocytes and its modulation by Ito may be even larger than observed in the present study.

Differences between endo-and epicardial myocytes
The gradient in APD and the values we obtained for Ito and L-type Ca¥ current density in endo-and epicardial myocytes are similar to those observed in previous studies under similar conditions (Clark et al. 1993;Shimoni et al. 1995). Clark et al. (1993) did not observe any differences in L-type Ca¥ currents or the steady-state current at the end of the voltage pulse between these two regions. In addition, they found the inwardly rectifier current to be similar in the two regions. Ca¥ influx has been quantified using a similar approach by Bouchard et al. (1995). The values they obtained for Ca¥ influx and the AP-induced peak Ca¥ current were between the values we have observed in endoand epicardial regions. Since the regional origin of the myocytes they used was not further specified, they might have used myocytes from both regions or from in-between, where Ca¥ influx has intermediate values (data not shown). Terracciano & MacLeod (1997) estimated smaller values for Ca¥ entry than were found in the present study. This discrepancy may be due to a different technical approach: they used conventional microelectrodes without intracellular buffering of Ca¥, and thus a Ca¥-dependent outward current, which was blocked by the addition of Cd¥, may have reduced the recorded inward currents. In addition, they did not specify whether the myocytes were of left or right ventricular or of endo-or epicardial origin.

Effect of inhibition of Ito
Because of its immediate activation by depolarization and its transient nature, Ito plays an important role in the early repolarization of the AP. It influences the height of the plateau potential and the APD, especially in the early phase. Changes in shape and duration of the AP alter the Ca¥ influx: the increase in plateau potential leads to a reduced electrochemical driving force for Ca¥, thereby reducing the AP-induced peak Ca¥ current. A longer AP increases the total time of the Ca¥ current, thus increasing Ca¥ influx. It should be noted that the effects of inhibition of Ito vary among different regions and species. In canine left ventricular epicardial myocytes, the plateau potential increased and the APD decreased, whereas in endocardial myocytes both plateau potential and APD increased after inhibition of Ito (Litovsky & Antzelevitch, 1988). In contrast, in rabbit epicardial myocytes inhibition of Ito increased both APD and plateau potential, while rabbit endocardial myocytes were affected only to a very small extent (Fedida & Giles, 1991). In rat ventricle, inhibition of Ito also led to an increase in the plateau potential and a prolongation of the AP in both epi-and endocardial myocytes, but to a greater extent in epicardial myocytes (Clark et al. 1993). It has been shown for á1-adrenoceptor activation that an increased Ca¥ influx, facilitated by a longer AP, is responsible for the increase in contractility (Fedida & Bouchard, 1992). In most studies of the effects of chronic alterations on Ito (for review, see Wickenden et al. 1998;N abauer & K a ab, 1998), its magnitude was found to be reduced in the range of 20-40%, and this reduction was held to be at least partly responsible for the observed increase in APD. According to the present study, a 30% reduction in Ito in epicardial myocytes would increase Ca¥ influx by a factor of about 2·8, whereas similar changes in Ito in endocardial myocytes would only have a minor effect. Interestingly, the mechanism of reducing Ito in order to increase Ca¥ influx and thus contractility appears to be limited to epicardial regions of the left ventricle. This may explain the greater increase in unloaded cell shortening in epicardial than endocardial myocytes observed by Clark et al. (1993). An alteration in Ito might therefore allow for a regional regulation of contractility during physiological and pathophysiological adaptations.