Materials and Methods

Myocyte Preparation

Guinea pig ventricular myocytes were isolated by using a retrograde coronary perfusion method previously published except for minor modifications (Zaza et al., 1998). The investigation conforms to the Guide of the Care and Use of Laboratory Animals published by the National Institutes of Health (NIH publication 85-23, revised 1996); all experiments were carried out according to the guidelines issued by the Animal Care Committee of the University of Milan. In brief, female guinea pigs weighing 200 to 300 g were anesthetized with a xylazine (1.5 mg/kg b.wt.) and ketamine (20 mg/kg b.wt.) mix, killed through cervical dislocation, and exsanguinated. Hearts were quickly removed and perfused in sequence with 1) modified Tyrode's solution (37°C) containing 130 mM NaCl, 4.5 mM KCl, 0.75 mM CaCl₂, 5 mM MgCl₂, 23 mM HEPES-NaOH, 21 mM d-glucose, 1 mM NaH₂PO₄, 20 mM taurine, 5 mM creatine, and 5 mM pyruvate, adjusted to pH 7.3, and equilibrated with 100% O₂; 2) a nominally Ca²⁺-free Tyrode's solution containing 3.3 μM EGTA and adjusted to pH 7.0; 3) a 0.075 mM CaCl₂Tyrode's solution containing 140 U/ml collagenase (type 2; Worthington Biochemicals, Freehold, NJ), 0.17 U/ml protease (type XIV; Sigma-Aldrich, St. Louis, MO), and bovine serum albumin (1 mg/ml). The ventricles were then chopped at 37°C and triturated in nominal Ca²⁺-free solution adjusted to pH 7.3. The supernatant was collected every 5 min, filtered through a nylon mesh, and centrifuged at 500 rpm for 3 min. To separate myocyte and nonmyocyte cells, the pellets were resuspended (50%, v/v) in a continuous Percoll gradient containing 0.9% NaCl and centrifuged at 500 rpm for 15 min. Finally, isolated myocytes were resuspended in 1 mM CaCl₂ Tyrode's solution containing gentamycin (10 μg/ml) and stored at room temperature until use. Rod-shaped, Ca²⁺-tolerant myocytes, obtained with this procedure, were used within 12 h from dissociation. Measurements were performed only in quiescent myocytes with clear-cut striations.

Recording Techniques

Myocytes suspension was placed in a 30-mm Petri dish, with a plastic ring to reduce total volume to ≈1 ml. The dish was perfused at 2 ml/min with standard external Tyrode's solution, containing 154 mM NaCl, 4 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 5 mM HEPES-NaOH, and 5.5 mMd-glucose, adjusted to pH 7.35. The cell under study was held within 300 μm from the tip (1 mm) of a thermostated multiline pipette allowing for rapid switch between solutions. The solution temperature was monitored at the pipette tip with a fast-response digital thermometer (BAT-12; Physitemp, Clifton, NJ) and kept at 35 ± 0.5°C except when otherwise specified (seeResults).

Membrane potential and current were measured in the whole cell configuration (Axopatch 200-A; Axon Instruments, Inc., Foster City, CA) by borosilicate glass pipettes with a tip resistance between 2.5 and 4 MΩ. The pipette solution contained 110 mM K⁺-aspartate, 23 mM KCl, 0.4 mM CaCl₂ (calculated free-Ca²⁺= 10⁻⁷ M), 3 mM MgCl₂, 5 mM HEPES-KOH, 1 mM EGTA-KOH, 0.4 mM GTP-Na⁺ salt, 5 mM ATP-Na⁺ salt, and 5 mM creatine phosphate, pH 7.3. Pipette solutions were modified in specific cases, as detailed in the relevant sections of Results. Membrane capacitance and series resistance (<5 MΩ) were measured in every cell; the latter was compensated by approx. 70% of its initial value. An average junction potential of about 5 mV, measured upon moving the electrode tip from Tyrode to “intracellular” (K⁺-aspartate), was left uncompensated. Potential and current signals were filtered at 2 KHz, digitized through a 12-bit A/D converter (Digidata 1200, sampling rate 5 KHz; Axon Instruments, Inc.) and simultaneously recorded by an adapted VCR system. Trace acquisition and analysis was controlled by dedicated software (pClamp 8.0; Axon Instruments, Inc.).

Experimental Protocols

Different ionic currents have been studied, each requiring specific conditions detailed here.

Na⁺/K⁺ Pump Current (I_NaK) Recordings.

I_NaK was measured as the holding current recorded at −40 mV in the presence of Ni²⁺ (5 mM), nifedipine (5 μM), and Ba²⁺ (1 mM) to minimize contamination by changes in Na⁺/Ca²⁺ exchanger, Ca²⁺ and K⁺ currents, respectively. Tetraethylammonium-Cl (20 mM) and EGTA (5 mM) were added to the pipette solution and intracellular K⁺ was replaced by Cs⁺. To optimize the recording conditions, I_NaK was enhanced by increasing intracellular Na⁺ (to 10 mM) and extracellular K⁺ (to 5.4 mM). In each cell, the current was recorded at steady state during exposure to increasing concentrations of the drug under test and, finally, to a saturating concentration of ouabain (20 μM) (Fig. 2a). BecauseI_NaK inhibition was expressed as percentage of ouabain effect, the latter was used as the asymptote for the estimation of EC₅₀ values. The rate of onset of I_NaK inhibition was measured by linear interpolation of the early phase of the change in current.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Drug effects onI_NaK. a) membrane current recorded at −40 mV from an individual cell during exposure to increasing concentrations (micromolar) of PST2744 (PST) followed by 20 μM ouabain (ouab). The effect of a single digoxin (digo) concentration (micromolar) applied after ouabain washout is shown for comparison. B, dose dependence ofI_NaK inhibition by PST2744 (open circles,n = 9, 14, 17, and 6) and digoxin (filled circles,n = 7, 9, 13, and 7). c, myocyte twitch amplitude is plotted as a function of I_NaK inhibition at various PST2744 (0.5, 2.5, 10, and 20 μM) and digoxin (0.5, 0.8, 2.5, and 5 μM) concentrations. Data of twitch amplitude are fromMicheletti et al. (2002).

High-Threshold Ca²⁺ Current (I_CaL) Recordings.

I_CaL was recorded during depolarizing pulses to 0 mV from a holding potential of −40 mV as the difference between currents recorded in the absence and presence of 10 μM nifedipine, respectively (nifedipine-sensitive current,I_nife). To prevent contamination by K⁺ currents, intracellular K⁺ was replaced by equimolar Cs⁺, the latter was also added to the superfusing solution at a concentration of 10 mM. A low EGTA concentration (0.5 mM) was used to control basal Ca²⁺ levels without suppressing Ca²⁺ transients.

Transient Inward Current (I_TI) Recordings.

I_TI was elicited by repolarization during a Ca²⁺ transient. Because the aim was to test the effects of PST2744 and digoxin independently of the inhibition of the Na⁺/K⁺ pump, the latter was blocked during all measurements by exposure to a K⁺-free solution containing 20 μM ouabain. To minimize contamination by components independent of the Ca²⁺ transient,I_TI was obtained as the subtraction of traces recorded, respectively, under loading (Fig. 4, inset 1) and depletion (Fig. 4, inset 2) of the sarcoplasmic reticulum. Under such conditions, I_TI should be mostly carried by the Na⁺/Ca²⁺exchanger (NaCaX) (Fedida et al., 1987; Zygmunt et al., 2000), whose driving force is transiently increased upon repolarization.I_TI was measured at room temperature under ruptured-patch conditions with low intracellular EGTA concentration (1 mM). To prevent contamination by K⁺ currents, intracellular K⁺ was replaced by equimolar Cs⁺ and the extracellular solution contained Ba²⁺ (1 mM) and Cs⁺ (2 mM). Because the contour of I_TI was also affected by digoxin, effects were more accurately quantified by measuring “average I_TI” as the ratio of I_TI time integral over the integration interval (avg I_TI).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Drug effects on I_TI in the presence of Na⁺/K⁺ pump blockade. a and b,I_NaCa measured asI_TI elicited by repolarization during a Ca²⁺ transient in the presence of 20 μM ouabain. The voltage protocols (inset) included a conditioning train (1 Hz) followed by a test pulse, after which I_TI was measured (shown by the box). The duty cycle of the conditioning train was such as to favor cell Ca²⁺ loading in protocol 1 and Ca²⁺ depletion in protocol 2. Protocols 1 and 2 were applied in each cell, and I_TI was obtained by subtraction between the respective current traces. c and d, effects of PST2744 (PST 4 μM, n = 14) and digoxin (digo 1 μM, n = 12) on avgI_TI. ∗, P < 0.05.

Na⁺/Ca²⁺ Exchanger Current (I_NaCa) Recordings.

I_NaCa was measured as the current sensitive to 5 mM Ni²⁺ (Kimura et al., 1987) by the voltage protocol shown in the inset of Fig. 5. As forI_TI, measurements were performed in the presence of Na⁺/K⁺ pump blockade (see described above), and Na⁺concentration in the pipette solution was increased to 20 mM to improve measurement of outward I_NaCa. Contamination by I_CaL and cytosolic Ca²⁺ oscillations were minimized by adding 0.2 μM nisoldipine and 5 mM EGTA to the extra- and intracellular solutions, respectively. Current/voltage relations ofI_NaCa were obtained by applying repolarizing voltage ramps (dV/dt = −85 mV/s), which allowed measurement of I_NaCa reversal potential (E_rev) (Hobai et al., 1997). Drug effects on forward and backward NaCaX operation modes were separately quantified by measuringI_NaCa at potentials symmetrical toE_rev (±60 mV fromE_rev). This type of measurement provides an estimate of inward and outward NaCaX “conductances”.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5

Drug effects on I_NaCa in the presence of Na⁺/K⁺ pump blockade. a,I_NaCa measured as the current sensitive to 5 mM Ni²⁺ (see text for details). Current/voltage curves were obtained by applying slow hyperpolarizing V ramps (−0.085 V/s, inset); inward and outward I_NaCa were measured at ±60 mV from the current reversal potential, respectively. a and b, examples of the effects of PST2744 (PST 4 μM) and digoxin (digo 1 μM) on I_NaCa current/voltage curves. Drug effects on inward and outward I_NaCa are summarized in c and d, respectively (n = 10 for PST and n = 11 for digo). ∗, P < 0.05.

Delayed Rectifier Current (I_Kr andI_Ks) Recordings.

Delayed rectifier currents were measured as the amplitude of the “tail” elicited upon returning to the holding potential after an activating pulse (protocols at the top of Figs. 6 and 7). To isolateI_Ks,I_Kr was blocked by 5 μM E-4031, and tail contamination by drug-induced changes inI_CaL andI_NaCa were prevented by ouabain (20 μM), Ni²⁺ (5 mM), and nisoldipine (0.2 μM). Under such conditions, the time-dependent outward current was completely suppressed by 10 μM chromanol 293B, thus confirming its identity with I_Ks.I_Kr was isolated by subtracting currents recorded in the absence and presence of 5 μM E-4031 (Sanguinetti and Jurkiewicz, 1990).

• Download figure
• Open in new tab
• Download powerpoint

Figure 6

Drug effects on I_Ks. The voltage protocol used to measure I_Ks is shown at the top (see text for details). Examples of PST2744 (PST 4 μM) and digoxin (digo 1 μM) effects are shown in a and b, respectively. Drug-induced changes in the amplitude ofI_Ks tail current density are summarized in c (n = 5 for PST and n = 4 for digo). ∗, P < 0.05.

• Download figure
• Open in new tab
• Download powerpoint

Figure 7

Drug effects on I_Kr. The voltage protocol used to measure I_Kr(current sensitive to 5 μM E-4031, I_E-4031) is shown at the top (see text for details). Examples of PST2744 (PST 4 μM) and digoxin (digo 1 μM) effect are shown in a and b, respectively; because I_Kr flowing during the activating step was very small, only I_Kr tail currents are displayed at high magnification. Drug-induced changes in the amplitude of I_Kr tail current amplitude are summarized in c (n = 6 for PST andn = 8 for digo). ∗, P < 0.05.

Substances

Nifedipine (Sigma-Aldrich) and nisoldipine, E-4031, and chromanol 293B stock solutions were prepared by dissolving the substances in ethanol, water, and dimethyl sulfoxide, respectively. PST2744 and ouabain (Sigma-Aldrich) were dissolved in water, and digoxin (Sigma-Aldrich) was dissolved in dimethyl sulfoxide; E-4031, chromanol 293B, and nisoldipine were generous gifts from Sanofi Recherche (Montpellier, France), Roche Diagnostics, and Bayer Pharmaceuticals (Milano, Italy), respectively.

Statistical Analysis

Means were compared by Student's t test or analysis of variance for paired or unpaired observations as appropriate. A probability level P < 0.05 was used to define significance throughout the study (N.S., not significant). In the text and figures, values are presented as mean ± standard error.

Results

Effect on Membrane Potential.

Myocytes were studied in the ruptured patch mode, during stimulation by 2-ms current pulses delivered through the patch pipette at a cycle length of 0.5 s (Fig. 1, a–c).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Drug effects on the action potential. a, APD₉₀ measured in an individual cell during stimulation at 2 Hz is plotted as a function of time. Exposure to PST2744 (PST) and digoxin (digo) is marked by the bars. The insets show action potentials recorded at the times marked on the APD graph by the labels a–f. b and c, summary of E_diast and APD₉₀changes observed during exposure to the specified equi-inotropic drug concentration pairs (n = 5, 7, and 9 for both agents). ∗, P < 0.05 versus control; #,P < 0.05 PST versus digo.

Both PST2744 and digoxin depolarized diastolic membrane potential slightly and only at concentrations belonging to the rightmost portion of the inotropy dose-response curve (Micheletti et al., 2002). The effect of equi-inotropic concentrations of PST2744 and digoxin on diastolic membrane potential was similar (Fig. 1b).

In the majority of cells, action potential duration (APD) was modulated by PST2744 in a biphasic manner, i.e., a transient prolongation was followed by a larger shortening (Fig. 1a). Mirror-like effects were observed upon washout, i.e., return to control values was preceded by transient APD shortening beyond the level achieved during steady-state drug perfusion (Fig. 1a). A biphasic response was less consistently observed with digoxin. Transient APD prolongation at washin occurred in three of seven cells with 2.5 μM PST2744 (10.2 ± 3.5%) and in zero of seven cells with the equi-inotropic digoxin concentration of 0.8 μM. At higher concentrations, transient prolongation occurred in seven of nine cells with PST2744 (10 μM, 27.8 ± 3.7%) and in two of nine cells with digoxin (2.5 μM, 6.9%). As previously observed for inotropic effects (Micheletti et al., 2002), the onset and decay of APD changes were faster with PST2744 than with digoxin. The extent of APD shortening induced by equi-inotropic PST2744 and digoxin concentrations was similar (Fig. 1c). APD shortening occurred mainly through depression of the plateau phase (Fig. 1a, insets). Diastolic oscillations of membrane potential (DADs) were seldom observed and only at the highest concentration tested; their incidence was too small to allow a meaningful comparison between the two drugs.

Effect on I_NaK.

This set of experiments was designed to compare directly the concentration dependence of Na⁺/K⁺ pump inhibition by PST2744 and digoxin (Fig.2, a–c). The effect of each drug concentration on I_NaK was expressed as percentage of the ouabain-induced change (see Materials and Methods). As occurred for inotropic effects (Micheletti et al., 2002), digoxin (EC₅₀ = 2.66 μM) was more potent than PST2744 (EC₅₀ = 6.75 μM) (Fig. 2b) in inhibiting I_NaK. However, the extent of I_NaK inhibition associated to a given inotropic effect was slightly greater for PST2744 than for digoxin (Fig. 2c). Thus, digoxin may require less Na⁺/K⁺ pump inhibition than PST2744 to induce the same inotropic response. When measured at the same concentrations, the onset of I_NaKinhibition was faster for PST2744 than for digoxin (e.g., at 2.5 μM, 4.9 ± 0.9 versus 2.1 ± 0.9 pA/s; P < 0.05).

Effect on I_CaL.

The purpose of this set of experiments (Fig. 3, a–d) was to test whether inhibition of I_CaL might contribute to drug-induced changes in the action potential contour. Equi-inotropic drug concentrations corresponding to the mid-portion of the inotropy dose-response curve (Micheletti et al., 2002) were used. Neither PST2744 (4 μM; Fig. 3a) nor digoxin (1 μM; Fig. 3b) significantly affected the peak amplitude ofI_CaL (PST2744, −0.52 ± 5.5%, N.S.; digoxin, −6.23 ± 4.2%, N.S.). The current decay was best fitted by a biexponential function. The faster exponential component, reflecting Ca²⁺-dependentI_CaL inactivation, was significantly accelerated by both PST2744 (τ_fast, −17.05 ± 3.2%, P < 0.05) and digoxin (τ_fast, −13.53 ± 2.8%,P < 0.05) (Fig. 3d). The slow component was unchanged by PST2744 and slightly slowed by digoxin (τ_slow, 5.4 ± 1.5%, P < 0.05).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Drug effects on I_CaL. a and b, nifedipine-sensitive currents (10 μM,I_nife) in control and during drugs superfusion at equi-inotropic concentrations (PST2744 4 μM,n = 4 and digoxin 1 μM, n = 6). Current traces were recorded during steps to 0 mV from a holding potential of −40 mV every 2 s; to emphasize changes inI_CaL inactivation, each trace was normalized to the peak amplitude. c and d, average results on peakI_nife and τ_fast ofI_CaL inactivation. ∗,P < 0.05.

Effect on I_TI.

Representative examples of the effects of PST2744 and digoxin onI_TI are shown in Fig.4, a and b, respectively. PST2744 (4 μM) consistently reduced avg I_TI(−42.8 ± 7.9%, P < 0.05; Fig. 4c). Although digoxin (1 μM) also caused a small decrease inI_TI in part of the cells, its effect was inconsistent and failed to achieve significance (−8.8 ± 11.5%, N.S.; Fig. 4d).

Effect on I_NaCa.

Representative examples of the effects of PST2744 and digoxin onI_NaCa are shown in Fig.5, a and b, respectively. BaselineI_NaCa current and itsE_rev were similar between PST2744 and digoxin groups. PST2744 (4 μM; Fig. 5a) decreased inward (Fig. 5c) and outward (Fig. 5d) I_NaCaconductance similarly (inward, −14.9 ± 4.8%, P< 0.05; outward, −13.5 ± 4.1%, P < 0.05) and slightly shifted E_rev in the negative direction (−26.3 ± 3 versus −23.6 ± 3.2 mV,P < 0.05). Digoxin (1 μM; Fig. 5b) affected neitherI_NaCa norE_rev (−18.5 ± 2.1 versus −17.1 ± 1.9 mV; N.S.)

In summary, PST2744, but not digoxin, markedly inhibitedI_TI (i.e.,I_NaCa triggered by a Ca²⁺ transient), but only slightly reducedI_NaCa conductance measured under conditions minimizing cytosolic Ca²⁺fluctuations. Because such effects were measured in the presence of ouabain, they were not secondary to Na⁺/K⁺ pump inhibition.

Effect on Delayed Rectifier Currents.

Evaluation of PST2744 and digoxin effects on delayed rectifier currents was prompted by the arrhythmogenic potential of their inhibition and by the observation of a difference in transient APD changes upon drug washin and washout (see above). PST2744 (4 μM) and digoxin (1 μM) were applied during separate measurement of I_Ks (Fig.6, a–c) andI_Kr (Fig. 7, a–c) (Sanguinetti and Jurkiewicz, 1990) (see Materials and Methods). I_Kr was marginally reduced by PST2744 (−5.99 ± 1.5%, P < 0.05) and unaffected by digoxin (2.99 ± 5.3%; N.S.).I_Ks was slightly but significantly reduced by both PST2744 and digoxin to a similar extent (−21.5 ± 4.0 versus −14.6 ± 2.6%; N.S.). Thus, concentrations of PST2744 and digoxin exerting similar inotropic effect slightly reducedI_Ks. PST2744 also inhibitedI_Kr but to a very small extent.

Discussion

Effects on Membrane Potential.

The aim of this set of experiments was to look for peculiarities in the modulation of action potential contour that might suggest a mechanism for the different proarrhythmic effects of PST2744 and digoxin.

Slight depolarization of diastolic membrane potential was similarly induced by PST2744 and digoxin and was expected as a result of Na⁺/K⁺ pump inhibition. Transient APD prolongation followed by shortening is commonly observed with cardiac glycosides and is respectively interpreted as primary and secondary effects of Na⁺/K⁺pump inhibition (Levi et al., 1994). Indeed, although removal ofI_NaK would primarily prolong APD, the concomitant increase in intracellular Na⁺ would move the NaCaX electrochemical balance to shiftI_NaCa in the outward (i.e., repolarizing) direction, thus causing APD shortening (Levi, 1993).

Transient APD changes (wash-in prolongation and wash-out overshoot shortening) were more prominent for PST2744 than for digoxin. This can be hardly attributed to modulation of delayed rectifier currents; indeed, the two agents differed only slightly in their effects on I_K, which were small even in absolute terms. An alternative and more plausible explanation was provided by the faster rate of onset and dissipation of PST2744 effects, probably attributable to a difference between PST2744 and digoxin in the kinetics of Na⁺/K⁺ pump inhibition. This might reduce the overlap betweenI_NaK inhibition and intracellular Na⁺ accumulation, thereby emphasizing the transient changes in APD.

DADs, expected in view of the previously described after- contractions (Micheletti et al., 2002), did not occur in the present setting. This might reflect partial buffering of drug-induced Ca²⁺ loading by cell dialysis through the recording pipette.

Effects on Na⁺/K⁺ Pump Current.

According to the classical interpretation of glycosides action, inotropy should be directly related to Na⁺/K⁺ pump blockade. However, other mechanisms may contribute to their inotropic effect (Tian et al., 2001; Sagawa et al., 2002) and affect the relationship between Na⁺/K⁺ pump inhibition and inotropy (Wasserstrom et al., 1991; Wasserstrom, 2002). This set of experiments was designed to assess whether differences might exist in such a relationship between PST2744 and digoxin. Data from a previous study were used for the concentration dependence of the inotropic effect (Micheletti et al., 2002). Comparison of such data with the present results shows that the extent of Na⁺/K⁺ pump inhibition required to produce a given inotropic effect was slightly smaller for digoxin than for PST2744. The ratio between Na⁺/K⁺ pump inhibition and inotropy for various digitalis compounds may correlate with the ability to facilitate opening of Ca²⁺ release channels of the sarcoplasmic reticulum (Wasserstrom et al., 1991). Such an action has been demonstrated for digoxin (Sarkozi et al., 1996; Sagawa et al., 2002) and, if not shared by PST2744, might justify the difference observed between the two agents. The onset of Na⁺/K⁺ pump inhibition was faster for PST2744 than for digoxin. This is consistent with the previous observation that association and dissociation rate constants of digitalis compounds to the Na⁺/K⁺ pump are decreased by the glycosidic group (aglycones bind faster than the respective glycoside) (Yoda, 1974), which is not present in PST2744 structure.

It is fair to stress that precise quantitation of Na⁺/K⁺ pump inhibition byI_NaK measurements could be obtained only if subsarcolemmal Na⁺ activity remained strictly constant (Levi et al., 1994). This would require instantaneous equilibration of subsarcolemmal space with pipette solution, a condition only grossly approximated under real experimental conditions. Thus, the curves shown in Fig. 2c are meant to compare the two agents, rather than to establish the relation between Na⁺/K⁺ pump inhibition and inotropy in absolute terms.

Effects on I_TI and Na⁺/Ca²⁺ Exchanger Current.

This group of experiments was aimed at testing the primary effects (i.e., independent of Na⁺/K⁺ pump inhibition) of PST2744 and digoxin on I_TI and its charge carrier NaCaX.

At variance with digoxin, PST2744 significantly inhibitedI_TI. This current is directly responsible for digitalis-induced DADs; therefore inhibition ofI_TI can be viewed as an antiarrhythmic effect and might account for the different toxicity of the two agents. Two mechanisms may account for I_TIreduction by PST2744: 1) a direct inhibition of NaCaX; or 2) a reduction in the subsarcolemmal Ca²⁺concentration transiently available after repolarization to drive the exchanger (e.g., a decrease in the amplitude or duration of the Ca²⁺ transient induced by the depolarizing step). Although the former mechanism would take place at sarcolemmal level, the latter would imply a change in the function of sarcoplasmic reticulum components. Clues to discriminate between such mechanisms are provided by the second set of experiments, in whichI_NaCa was dissected by pharmacological means under conditions unlikely to induce significant fluctuations of subsarcolemmal Ca²⁺. In this case, NaCaX conductance was still consistently reduced by PST2744, but the effect was small compared with that on I_TI. Thus, although PST2744 may directly inhibit NaCaX to a small extent, other mechanisms (e.g., changes in the kinetics of intracellular Ca²⁺ transients) are likely to contribute to the observed inhibition of I_TI. PST2744 also slightly shifted E_rev in the negative direction. This effect cannot be explained by an increase in intracellular Na⁺. First, because the Na⁺/K⁺ pump was fully blocked, and second, because although an increase in intracellular Na⁺ should cause largerI_NaCa conductance (LabHeart version 4.8, model; D. M. Bers and J. L Puglisi), the latter was decreased by PST2744. Thus, PST2744-induced change in NaCaX conductance is likely to result from a direct action on the exchanger, rather than from a change in transarcolemmal ion distribution.

Effects on Delayed Rectifier and Ca²⁺ Currents.

Due to its proarrhythmic potential, inhibition of delayed rectifier currents has recently become a major concern in drug safety; this motivated an analysis of PST and digoxin effects onI_Kr andI_Ks. While both agents exerted some effects, their functional relevance is questionable for the following reasons. Although major inhibition of delayed rectifier currents is generally required to modulate repolarization (Bosch et al., 1998), the effect on I_Ks were relatively small and those on I_Kr almost negligible. Second, I_Ks inhibition is generally associated with a relatively low proarrhythmic risk, because of the lack of reverse use dependence of its effect on APD (Bosch et al., 1998). Finally, although induction of early after-depolarizations and arrhythmias related to I_K inhibition may be secondary to APD prolongation (Brachmann et al., 1983; January et al., 1988; Antzelevitch and Sicouri, 1994), the net steady-state effect of PST2744 and digoxin was a shortening of APD (see above). Nonetheless, it is fair to stress that some APD prolongation might theoretically occur in other cardiac tissues expressing drug-sensitive currents in a different proportion (e.g., Purkinje fibers and M cells).

PST2744 and digoxin did not affectI_CaL peak amplitude; as expected (Hancox and Levi, 1996), both drugs acceleratedI_CaL inactivation. On the other hand, high concentrations of cytosolic Cs⁺ may obstruct Ca²⁺ fluxes through the SR membrane (Kawai et al., 1998); thus, drug-induced increase in Ca²⁺-dependent inactivation ofI_CaL might have been underestimated. Thus, PST2744 and digoxin effects on the action potential, as well as the different toxicity of two agents, cannot be attributed toI_CaL modulation.

Relevance to Arrhythmogenic Effects.

The cell targets of drug action evaluated in this study encompass those most likely to be involved in arrhythmogenic effects based on electrophysiological alterations. Among the actions observed, the one not shared by digoxin and most likely to account for the lower arrhythmogenic effect of PST2744 is inhibition of I_TI. Indeed,I_TI is directly involved in the genesis of DADs, the main mechanism of digitalis-induced arrhythmias (Bers, 2002). The extent of I_TIreduction was large compared with direct inhibition ofI_NaCa, thus suggesting that PST2744 might modulate the amplitude and/or kinetics of intracellular Ca²⁺ transients, a hypothesis that requires further investigation. Although in guinea pig and human ventricular myocytes I_TI is mostly carried by NaCaX (Fedida et al., 1987; Koster et al., 1999), other Ca²⁺-sensitive conductances may contribute to it in other species (Zygmunt et al., 1998). On the other hand, Ca²⁺-sensitive conductances would still be affected by drug-induced changes in cytosolic Ca²⁺ dynamics. Although other differences have been disclosed between the two agents, they concern quantitatively small effects, which, as discussed above, are unlikely to contribute to the arrhythmogenic profile.

Digitalis-induced triggered activity may more frequently originate from Purkinje fibers or M cells (Antzelevitch and Sicouri, 1994), probably because a longer plateau favors Ca²⁺ overload. Nonetheless, the conductances involved in the genesis of DADs are the same in all cell types. Thus, except for a potential difference in modulation of APD (see above), the findings of the present study may be relevant to all ventricular cell types.

Androgen steroids have been shown to counteract the tendency of estrogens to inhibit repolarizing currents (Pham and Rosen, 2002). In the present studies, performed on myocytes freshly isolated from female hearts, PST2744 slightly inhibited delayed rectifier currents and only minimally affected I_CaL. Thus, it seems unlikely that the antiarrhythmic effect of PST2744 may be related to the “repolarizing” effect of androgen steroids. Moreover, in spite of its chemical structure, PST2744 has no affinity for steroid receptors (Micheletti et al., 2002).