In rat ventricle, two Ca2+-insensitive components of K+ current have been distinguished kinetically and pharmacologically, the transient, 4-aminopyridine (4-AP)-sensitiveI to and the sustained, tetraethylammonium (TEA)-sensitiveI K. However, a much greater diversity of depolarization-activated K+ channels has been reported on the level of mRNA and protein. In the search for electrophysiological evidence of further current components, the whole cell voltage-clamp technique was used to analyze steady-state inactivation of outward currents by conditioning potentials in a wide voltage range. Peak (I peak) and late (I late) currents during the test pulse were analyzed by Boltzmann curve fitting, producing three fractions each. Fractionsa andb had different potentials of half-maximum inactivation (V 0.5); the third residual fraction, r, did not inactivate. Fractions a forI peak andI late had similar relative amplitudes andV 0.5 values, whereas size andV 0.5 of fractionsb differed significantly betweenI peak andI late. Onlyb ofI peak was transient, suggesting a relation withI to, whereasa, b, and r ofI late appeared to be three different sustained currents. Therefore, four individual outward current components were distinguished:I to(b ofI peak),I K(a), the steady-state currentI ss(r), and the novel currentI Kx(b ofI late). This was further supported by differential sensitivity to TEA, 4-AP, clofilium, quinidine, dendrotoxin, heteropodatoxin, and hanatoxin. With the exception ofI to, none of the currents exhibited a marked transmural gradient. Availability ofI K was low at resting potential; nevertheless,I K contributed to action potential shortening in hyperpolarized subendocardial myocytes. In conclusion, on the basis of electrophysiological and pharmacological evidence, at least four components contribute to outward current in rat ventricular myocytes.
- isolated myocytes
- rat ventricle
- transient current
- sustained current
- cloned channels
action potential waveforms differ between atrial and ventricular myocytes as well as between subepicardial and subendocardial myocytes, as shown for many species including dog, rat, and human (3, 14). This heterogeneity can be traced back to differences in outward current (2, 21, 25, 42). In particular, the transient outward current (I to) is more prominent in subepicardial than in subendocardial myocytes (rat ventricle; see Ref. 14). In rat ventricular myocytes two components of outward current are distinguished kinetically and pharmacologically (4,14, 40), the rapidly activating and inactivatingI to, which is sensitive to blockade by 4-aminopyridine (4-AP), and the rapidly activating but slowly inactivating delayed-rectifier-like current (I K), which can be blocked by tetraethylammonium (TEA). Furthermore,I to andI K differ with respect to the potential dependence of availability of the underlying channels (4). In some cardiac preparations, e.g., Purkinje fiber, dog ventricle, and human atrium,I to can be subdivided into a cytosolic Ca2+-insensitive, 4-AP-sensitiveI to1 and a cytosolic Ca2+-sensitive, 4-AP-insensitiveI to2 (6). In other preparations such as rat ventricle, only the cytosolic Ca2+-insensitive, 4-AP-sensitiveI to1 has been described (23). In the present paper, onlyI to1 is considered, and for reasons of simplicity it is referred to asI to.
With molecular biological approaches, a multitude of depolarization-activated K+-channel genes of theKv1,Kv2, andKv4 families and raterg andKvLQT1 have been identified in adult and embryonic rat ventricle, respectively (7, 16, 17). Among those genes, Kv4.2 andKv4.3 encode for K+-channel proteins withI to-like properties (19, 39, 45), whereas three other gene products (Kv1.2, Kv1.5, Kv2.1) are channel proteins withI K-like properties (9, 20, 22, 28, 36, 44). On the basis of these reports, more than two components of K+ outward current of rat ventricular myocytes are expected to be distinguishable, provided that the gene products differ in electrophysiological and/or pharmacological properties.
Here we report that it is in fact possible to differentiate at least four components of outward current by making use of steady-state inactivation kinetics. The sensitivity of these components to block by TEA, 4-AP, dendrotoxin (DTX), heteropodatoxin (HpTx3), and hanatoxin and other tools yielded pharmacological profiles that were compared with those reported for cloned channels. Because the amplitude of total outward current declines from subepicardial to subendocardial cells within the ventricular wall, we have also characterized the contribution of each current component to this differential current distribution. Preliminary results have been published in abstract form (22a and 22b).
All studies complied with the German home office regulations governing the care and use of laboratory animals. Male Wistar rats (body wt 200–250 g) were killed by cervical dislocation. As described previously (41), the hearts were perfused on a Langendorff apparatus at 37°C for 5–7 min with nominally Ca2+-free saline solution (composition in mM: 100 NaCl, 10 KCl, 5.0 MgSO4, 1.2 KH2PO4, 20 glucose, 50 taurine, and 5.0 MOPS, adjusted to pH 7.0 with NaOH). The rat hearts were then perfused for 15 min with collagenase-containing solution (collagenase type I, 0.5 g/l, Sigma C-0130, Munich, Germany) supplemented with CaCl2 (200 μM) and albumin (1 g/l). After enzyme perfusion, the hearts were chopped into small pieces and dissociation was continued by gentle stirring of the tissue pieces in fresh enzyme solution for 5–15 min. On some occasions tissue batches were dissected from the apex and the base of the heart, and their dissociation was continued separately to obtain subepicardial and subendocardial myocytes, respectively (14). Single ventricular myocytes were collected in a low-Ca2+solution, the Ca2+ concentration of which was slowly increased (0.2 mM steps in intervals of 10 min) until a final concentration of 0.6 mM was reached. The cells were stored at room temperature and used within 12 h.
Whole cell voltage-clamp technique.
Myocytes were transferred to a small Perspex chamber (volume 0.5 ml) placed on the stage of an inverted microscope (Olympus IMT-2 or Zeiss Axiovert-10). The chamber was continuously perfused at a constant rate (1.2 ml/min). Only rod-shaped myocytes with clear striations were used. For action potential and membrane current recordings, the single-electrode voltage-clamp technique was applied. Heat-polished pipettes made from borosilicate filament glass (OD 1.5 mm, Hilgenberg, Malsfeld, Germany) were used to form gigaohm seals with gentle suction; on average, the seal resistance was 2.8 GΩ (range 1–12 GΩ). The patched membrane was then disrupted by a pulse of suction to establish continuity of the interior of the electrode with the cytosol. Voltage or current clamp was achieved using a List L/M-EPC-7 or an Axopatch 200 amplifier. For stimulus protocol design and data acquisition, the Axolab TL-125 interface and pCLAMP 5.5 software (Axon Instruments, Foster City, CA) were used.
To account for variabilities in cell size, the membrane capacitance was measured before compensation by means of fast depolarizing ramp pulses (from −40 to −35 mV, duration 5 ms) at the beginning of each experiment. Because the membrane conductance is very low and constant in this range, a change in current level is caused by the capacitive properties of the cell membrane. The average membrane capacity of the rat myocytes investigated in this study was 203 ± 55 pF (mean ± SD; range 92–336 pF;n = 144 myocytes). Absolute current amplitudes (in pA) were then divided by the cell capacity and expressed as picoamperes per picofarad. Access resistance was kept below 5 MΩ. Series resistance was routinely compensated by 80%. With a current amplitude of 4 nA, the voltage error after compensation amounted to <4 mV. The input resistance (R in = ΔV/ΔI; where ΔV and ΔI are change in voltage and current, respectively) calculated in the range of −80 to −70 mV, i.e., at resting membrane potential (RMP), amounted to 70 ± 15 MΩ (mean ± SE;n = 86 myocytes) and is consistent with previously published data on ventricular myocytes from various species (2, 21, 29, 33, 43). Therefore, total outward current should be contaminated by <3% because of the contribution of leak currents.
All experiments were carried out at room temperature (20–24°C) to improve both stability of myocytes and voltage control when measuring large, rapidly activating currents. The stimulation rate was 0.1 Hz for all experiments unless otherwise indicated.
Measurement of action potentials.
Action potentials in rat myocytes were measured in the current-clamp mode after injection of current (duration 3–5 ms, amplitude 0.8–1.0 nA, stimulation rate 0.1 Hz, room temperature). For these experiments, pipettes were pulled with tip resistances of 3.0–5.0 MΩ when filled with a solution containing (in mM) 140 KCl, 4.0 MgCl2, 5.0 CaCl2, 10.0 EGTA, 10.0 HEPES, and 4.0 Na2ATP, adjusted to pH 7.3 with KOH. Thus the free Ca2+concentration was buffered to 50 nM (free Mg2+ concentration 300 μM) as calculated by the computer program EQCAL (Biosoft, Cambridge, UK). The bath solution was composed of (in mM) 150 NaCl, 5.4 KCl, 2.0 MgCl2, 10 glucose, 10 HEPES, and 1.2 CaCl2, adjusted to pH 7.4 with NaOH. Action potentials were analyzed for RMP and action potential amplitude. Action potential duration (APD) was measured at 20, 50, and 90% of repolarization (APD20, APD50, APD90).
Measurement of outward K+ current.
To measure outward current in ventricular myocytes of rat heart, the bath was perfused with a solution similar to that for action potential recording, except that 0.6 mM CaCl2 and 0.1 mM CdCl2 were used to block Ca2+ channels. Electrodes had tip resistances of 1.5–2.5 MΩ when filled with the same solution as used for action potential recordings. Current-voltage relations (range −40 to +60 mV) were measured with 300-ms clamp steps in 10-mV increments after Na+-current inactivation by a 40-ms clamp step to −40 mV from the holding potential of −80 mV. For steady-state inactivation, 2,000-ms conditioning clamp steps (range −140 to +20 mV; 10-mV increments) were followed by a test clamp step to +60 mV (duration 300 ms); a step of 5 ms at −40 mV was interposed between conditioning and test clamp steps to keep step amplitude and capacitive current constant.
Samples of HpTx3 and hanatoxin were kindly provided by NPS Pharmaceuticals (Salt Lake City, UT) and Dr. Kenton Swartz (National Institutes of Health, Bethesda, MD), respectively. Clofilium tosylate was a gift of Eli Lilly (Indianapolis, IN), and quinidine hemisulfate was from Merck (Darmstadt, Germany). All drugs were dissolved in H2O; aliquots of concentrated stock solutions were stored at −20 C until use. Enzymes used for cell isolation (collagenase type I) and BSA were obtained from Sigma Chemicals. All other chemicals were purchased from commercial suppliers and were of laboratory grade.
Steady-state inactivation curves were obtained by plotting normalized current (I/I max) at the test potential as a function of the conditioning potential (V m) and fitting a Boltzmann function to the data points whereV 0.5 andk are the potentials of half-maximal inactivation and the slope factor, respectively. However, a single Boltzmann function did not adequately describe the data, whereas in almost all cases the sum of two Boltzmann functions plus a residual component significantly improved the goodness of fit wherea andb are the fractional amplitudes of the two functions and r is the residual component, i.e., (1 − a −b). Fits of theoretical equations to the experimental data were performed using pCLAMP software (Clampfit) or Prism (Graphpad Software, San Diego, CA).
The results are expressed as means ± SE or SD ofn experiments. Statistical differences were analyzed by means of Welch’s approximatet-test, which does not assume equal variances, or an appropriate nonparametric test for paired or grouped data. Correlation between two parameters was tested with the nonparametric rank test according to Spearman.
Steady-state inactivation of total outward current.
A typical family of outward current traces after variousV m (Fig.1 A) revealed distinct differences in inactivation pattern for peak and late outward current (I peak andI late, respectively). With aV m of −140 mV (trace 1), the test current at +60 mV rapidly reached its peak value at the beginning of the clamp step (I peak) and declined to 70% ofI peak toward the end of the clamp step (I late), indicating that a large fraction of outward current did not inactivate during the test clamp step. In the voltage range negative to the resting potential the amplitude of the test pulse current became smaller, with less decrease inI peak than inI late causing an apparent increase in the transient component (comparetraces 1 and2 in Fig. 1). After two to three conditioning steps with little change in test current giving rise to a plateau, I peakwas strongly diminished byV m between −60 and −30 mV (trace 3). The remainder of the current was inactivated by V mup to −20 mV, and the residual current positive to this potential was resistant to any inactivation (trace 4).
Because the pattern of voltage dependence appeared to be more complex than the simple sum of the known currentsI to andI K,I peak andI late were evaluated separately. Normalization of data to the maximum outward current resulted in biphasic steady-state inactivation curves forI peak andI late, which were best fitted by the sum of two Boltzmann functions with fractionsa andb in addition to the residual fractionr (Fig.1 B; seemethods). BecauseI peak andI late were evaluated separately and each consists of these three fractions, a total of six fractions, i.e., a,b, andr ofI peak anda, b, and r ofI late, could be distinguished.
The parameters derived from the steady-state inactivation curves of 141 myocytes are summarized in Table 1. Very few of these myocytes possessed monophasic steady-state inactivation curves that could not be fitted reliably with the sum of two Boltzmann functions because of very small a orb ofI late. In 138 cells, a contributed 25% of total outward current toI peak and 27% toI late. TheV 0.5 were also similar, i.e., −95 mV forI peak and −93 mV forI late. Therefore, on the basis of electrophysiological properties,a ofI peak anda ofI late could represent a single current component and be well defined as the delayed rectifier current (I K). The residual current fraction r that did not inactivate even at very positiveV m amounted to 28% of the total outward current for bothI peak andI late and thus could be defined as the steady-state current (I ss). In marked contrast, b ofI peak andI late differed significantly in size, i.e., 47% of total outward current (I peak) versus only 9% (I late), as well as in V 0.5values (I peak: −38 mV;I late: −28 mV). In addition to these differences,b ofI peak had a marked transient time course with a time constant of 59.5 ± 3.1 ms (+60 mV, n = 22) for its exponential current decay, so that at the end of the 300-ms test clamp step >99% of current had been inactivated (compare Ref. 40). Hence,b ofI peak and ofI late could well represent two separate current components.
This hypothesis was tested by correlation analysis of fractional current amplitudes (Fig. 2). If identical,a, b, and r ofI peak are expected to correlate significantly with the respective fractions forI late. This was indeed the case for a (Fig.2 A) as well as forr (Fig.2 C). In the latter case, the regression line passed through the origin, providing convincing evidence for correlation of current fractions and hence for their identity as a single current component. In the case of the two fractions a, however,I peak was always smaller thanI late, which could be interpreted as current activating during the clamp step. In marked contrast to the significant correlation betweenI peak andI late ina andr, respectively, the amplitudes ofI peak andI late forb were not at all related to each other (Fig. 2 B). From this analysis it is concluded that b ofI peak and ofI late represent two independent current components, the former apparently representingI to and the latter a putatively novel sustained current designatedI Kx. Therefore, on the basis of steady-state inactivation, a total of four outward current components, i.e.,I K,I to,I Kx, andI ss, could be distinguished.
It must be pointed out, however, that the conditioning clamp step had to last long enough for complete inactivation of all current components. This was checked by varying the duration of conditioning clamp steps between 400 and 8,000 ms; the results are summarized in Table 2. The modified clamp protocol slightly affected the relative contribution of the individual components to the total outward current but never resulted in complete disappearance of any one current component. For instance,I ss decreased from 33 ± 2% of total current after 400-ms conditioning steps to 16 ± 2% after 8,000-ms steps, whereasI K increased from 20 ± 1% to 29 ± 2% under these conditions. The contribution of I to did not significantly depend on the duration of the conditioning clamp step, whereas I Kxbecame larger after long-lasting conditioning pulses.
Voltage dependence of activation of outward current components.
The steady-state inactivation data presented so far allow us to distinguish a total of four outward current components, i.e.,I K,I to,I Kx, andI ss. To obtain data on the activation kinetics and voltage dependence of the four current components, outward currents were activated by stepping to voltages in the range of −40 to +60 mV from the threeV m of −140, −70, and −20 mV (Fig. 3,A–C) followed by digital subtraction of current tracings. This procedure was aimed at further characterizing the individual current components (Fig. 3,D–E). Because channel availability should be at its maximum with aV m of −140 mV (compare Fig. 1), the amplitude of activated outward current is large (Fig. 3 A). BothI peak andI late are decreased after conditioning steps to −70 mV (Fig.3 B), where steady-state inactivation of I K should be complete (Fig. 1; Table 1). Therefore, digital subtraction of these two sets of current tracings should result in the isolation ofI K (Fig. 3,D andF). Currents activated from aV m of −20 mV should reflect activation ofI ss (Fig. 3,C andH). Hence, the difference betweenV m of −20 and −70 mV should represent the sum ofI to andI Kx (Fig. 3,E andG), which cannot be separated electrophysiologically because steady-state inactivation occurs in overlapping voltage ranges (Fig. 1; Table 1).
I K appears as a rapidly activating and slowly inactivating current with an activation threshold negative to −30 mV (Fig. 3,D andF). Assuming K+ as the major, but not only, charge carrier (compare Fig. 6) and a reversal potential (E rev) of −65 mV (Ref. 41; tail current analysis cannot be conducted with difference currents),I K is half-maximally activated at −34 ± 6 mV (slope 14 ± 2 mV, n = 12). Current activation accelerated at more depolarized potentials and could be approximated by a third-order power function yielding an activation time constant of 3.5 ± 0.8 ms at +40 mV. Current inactivation was voltage independent and followed a monoexponential time course with an inactivation time constant (τin) of 205 ± 32 ms at +40 mV. On the contrary,I to activates and inactivates rapidly with an activation threshold positive to −30 mV (Fig. 3, E andG). On average, activation ofI to is 2.6 times faster than that ofI K (activation time constant ofI to: 1.0 ± 0.1 ms at +40 mV). Current inactivation ofI to was voltage independent positive to 0 mV, ∼3.9 times faster thanI K, and followed a monoexponential time course with a τin of 48 ± 7 ms at +40 mV. Half-maximal activation ofI Kx(b ofI late) occurred at −8 ± 2 mV (slope 12 ± 2 mV) and forI to at +1 ± 2 mV (slope 13 ± 1 mV, n = 12; inactivating b ofI peak). Finally, I ss(Fig. 3, C andH) was characterized by an almost instantaneous activation and no inactivation within 300 ms, an activation threshold positive to −10 mV, and half-maximal activation at +7 ± 3 mV (slope 14 ± 1 mV). In conclusion, the current componentsI K andI ss isolated by means of a subtraction approach display distinct differences in terms of their activation and inactivation kinetics and voltage dependence. For componentsI to andI Kx, however, the kinetic differences are discrete, and therefore pharmacological tools are required for current separation.
Sensitivity of outward current components to pharmacological tools.
So far, electrophysiological evidence in support of four outward current components has been presented. In another approach to channel differentiation, we made use of several pharmacological tools that selectively block individual currents. For instance, 4-AP selectively blocks I to (4,12), TEA attenuatesI K (4, 35), and quinidine or clofilium reduces both current components (10, 26, 35). DTX is a potent blocker of a delayed-rectifier-like current flowing through the cloned Kv1.2 channel (13, 36). HpTx3 selectively blocks cloned and native Kv4.2 channels (32), whereas hanatoxin blocks both Kv4.2 and Kv2.1 channels, as shown in aXenopus expression system (36). Because the selectivity and efficacy of these agents are usually maintained after expression of cloned channels in mammalian cell lines, the sensitivity profile can be used for channel identification (6, 22).
To investigate the effects of HpTx3 and the other pharmacological tools on the electrophysiologically distinct current components, we examined outward currents at test steps to +60 mV after selectedV m, i.e., −140, −60, −30, and +20 mV (seetraces 1–4 in Fig.4 A) within the range of −140 to +20 mV. The control current traces in the absence of any blocker mark the transitions between the current components (i.e., fromI K toI ss; compare Fig.1). In comparison with control recordings, the Kv4.2 blocker HpTx3 (2 μM; Fig. 4 A) had little effect on current componentsI K andI ss (no reduction of current amplitude between traces 1and 2 and betweentraces 3 and4, respectively). However, HpTx3 markedly reduced the amplitude ofI to(trace 2) and in addition shifted the steady-state inactivation ofI to to less negative potentials, as evidenced by the persistence of a transient current in the presence of the toxin (trace 3 in Fig. 4 A). Between traces 2 and3, however, both peak and late current decreased to a similar extent in the presence of HpTx3, suggesting that a sustained current component, presumablyI Kx, undergoes steady-state inactivation there. This is supported by the quantitative analysis of steady-state inactivation curves displayed in Fig.4 B. Relative amplitudes,V 0.5 values, andk were not affected by HpTx3, except for I peak in the range of −60 mV to 0 mV (Fig.4 B). The transient current componentI to was reduced in its amplitude to 33 ± 4% of control and shifted on the voltage axis by some 30 mV to the right. However, the block ofI to was incomplete with 2 μM of HpTx3 and amounted to 72 ± 5 % at +20 mV but only 54 ± 2% at +60 mV (data not shown). Therefore, a portion of I to is expected to remain unblocked in the presence of HpTx3 (Fig. 4).
Using the Kv4.2 and Kv2.1 blocker hanatoxin (500 nM; data not shown), we observed a distinct decrease of the amplitude ofI to to 49 ± 9% of control and a significant amplitude reduction ofI K to 62 ± 11% of control (n = 6 experiments). The former blocking effect appears to confirm the results obtained with HpTx3; however, the interpretation of the latter effect as block of Kv2.1 must remain preliminary. We were unable to use higher concentrations to achieve a more complete block because hanatoxin is a very rare toxin.
Besides HpTx3 and hanatoxin, several other K+-channel blockers were tested, and their effects onI K,I to,I Kx, andI ss (i.e.,a, b, and r ofI peak andI late) are summarized in Fig. 5. 4-AP (100 μM and 1 mM) hardly affectedI K, slightly reduced I ss, and blocked both I toand I Kx in a concentration-dependent manner. With 1 mM 4-AP, block ofI Kx was significantly larger than block ofI to. TEA showed a complex blocking pattern in the concentration range of 1–10 mM: the compound reducedI K andI ss in a concentration-dependent manner with a larger maximum block ofI K than ofI ss, i.e., reduction to <20% vs. 70% of predrug control.I to andI Kx were affected differently by 10 mM TEA:I Kx was reduced to some 20% of predrug control, whereasI to was not significantly impaired at this concentration. Clofilium (3 μM) significantly reduced all current components, which confirms the nonselective nature of this blocker. With 30 μM clofilium,a andb ofI late were reduced to a larger extent than were the respective fractions ofI peak. This difference in sensitivity to block must be interpreted with caution because clofilium is known to cause time-dependent block (10, 26). Quinidine (5 μM) reducedI K but did not significantly impair the relative amplitudes of componentsI to,I Kx, orI ss. However, quinidine accelerated the apparent inactivation ofI to (not shown). Even at the maximum effective concentration of 100 nM, DTX neither blocked nor reduced any of the outward current components, suggesting the functional absence of Kv1.2 in rat ventricular myocytes. Therefore, of the investigated K+-channel blockers, only TEA (10 mM) selectively blockedI Kx without any effect on I to.
In conclusion, the pharmacological data presented so far are in line with our electrophysiological data and appear to support the hypothesis that four components contribute to outward current in rat ventricular myocytes. In particular, outward current fractionb ofI peak and component a resemble the well-characterized currentsI to andI K, respectively (4, 6, 14). The residual component ris similar to a steady-state current (I ss), which has been occasionally mentioned in the literature but has never received much attention (4, 35, 42). Finally, the HpTx3-insensitive fraction b ofI late appears to be a novel sustained current, which we have termedI Kx.
Ion selectivity of outward current components.
Ion selectivity is another criterion to distinguish between different ion channels. Other ions than K+could contribute as charge carriers in generating the four current components. In particular, a nonselective cation current carried by K+, Na+, and Ca2+ or an anion background current carried by Cl− could be involved. To test for ion selectivity, the intracellular ion concentration was varied by substituting K+ in the pipette solution with either Cs+ or TEA, both of which permeate poorly through K+channels, or by lowering extracellular Cl− from 163 to 13 mM by substituting sodium methanesulfonate for NaCl. The results of the respective experiments are summarized in Fig.6. Replacement of K+ by Cs+ in the pipette solution significantly depressedI K, abolishedI Kx(b ofI late), and reduced the amplitude ofI to(b ofI peak) andI ss to <50%. The latter current components were further decreased when TEA was present in the pipette solution. With 90% of extracellular Cl− replaced by the membrane-impermeant methanesulfonate, onlyI ss was reduced to 70%, whereas the other current components were not affected. These results suggest that the majority of outward current was carried by K+ and thatI ss consisted of two separate currents, one of which appeared to be carried by K+ and the other of which was most likely a Cl− current.
Transmural distribution of outward current components.
Within the ventricular wall,I to was found to be more prominent in subepicardial than in subendocardial myocytes of rat ventricle (14), and a similar distribution has been reported for K+ channels at the mRNA and protein levels (7, 16). Myocytes of different transmural location can be obtained from rat hearts by isolating myocytes separately from the apex and the base, yielding subepicardial and subendocardial cells, respectively (14). Using this approach, we consistently observed that subepicardial myocytes possessed a large rapidly inactivating transient outward current, whereas subendocardial myocytes isolated from the base of the heart were characterized by a smallI to component (data not shown).
When the size ofI to in absolute values (pA/pF) was plotted againstI to expressed as a fraction of total outward current (Fig.7 A), the data points from subendocardial myocytes clustered at the lower part and those from subepicardial cells at the upper part of the relation, as expected from the known transmural gradient ofI to. In addition, not all myocytes presently investigated have been isolated according to their origin within the ventricular wall. In fact >70 of the total of 141 cells were obtained from the whole free left ventricular wall. The majority of these myocytes are expected to stem from the midmyocardial region, but some of them could also be derived from either subendocardial or subepicardial regions, and this was confirmed indirectly by the widespread distribution of their amplitudes in the center part of this plot. Therefore,I to appears to possess a strong transmural gradient.
If the size of any of the other current components also depends on the site of origin within the ventricular wall, these fractions should correlate withI to, which is used as a marker for the transmural gradient. Of the currents tested,I K andI Kx did not correlate significantly withI to (Fig. 7,B andC), whereasI ss showed a small but significant negative correlation withI to (Fig.7 D). From these results it is concluded that the transmural gradient of outward current is caused by the differences inI to.
Outward current component IK and repolarization of action potential.
The outward current componentsI to,I Kx, andI ss should contribute to the shape of the action potential, as judged by the potential range of their availability. For componentI K, however,V 0.5 was −93 mV (see Table 1), and therefore this current component should be largely inactivated at normal RMP. Hence, its role for the action potential is less obvious than with the other current components. Here we tested whether increasing the availability ofI K by hyperpolarizing the membrane could influence the shape of the action potential. Action potentials recorded from subendocardial myocytes are much longer than those from subepicardial myocytes because of their profound difference inI to (Fig.8 A). Action potentials measured at hyperpolarized potentials were markedly shortened in duration, and this effect was significantly greater in subendocardial than in subepicardial cells (Fig.8 B). This observation was consistent with the theoretically expected increase inI K availability under hyperpolarizing conditions. BecauseI K was half-inactivated at −93 mV (Table 1), its availability at a normal resting potential of −70 mV amounted to 8%. Average hyperpolarization by 12 mV should have increased the availability to 24%, and this additional repolarizing current strongly reduced APD in subendocardial myocytes. The much lesser effect on APD in subepicardial cells was probably caused by the larger repolarizing force ofI to. Conversely, the blocking effect of TEA (10 mM) on outward currents did not produce any prolongation in APD in subepicardial myocytes (data not shown), whereas action potentials in subendocardial cells were markedly prolonged at both normal and hyperpolarized resting potentials (Fig.8 C). These data suggest thatI K may contribute to repolarization at least in subendocardial myocytes.
Outward current in rat ventricular myocytes consists of at least four different components that are distinguished on the basis of time course, potential range of steady-state inactivation, sensitivity to pharmacological blockers, and gradient of amplitude within the ventricular wall. In addition toI to, two delayed rectifier-like currents (I K,I Kx) and at least one noninactivating background component (I ss) were identified.
Dissection of outward current components.
In native rat ventricular myocytes, two major outward current components are regularly detected, i.e., the transient, 4-AP-sensitive K+ currentI to and the sustained TEA-sensitive K+ currentI K (4, 6, 14). For I to, half-maximum steady-state inactivation (V 0.5) is found at potentials between −29 and −46 mV (4, 43). This difference inV 0.5 values from the various studies may be caused in part by divalent cations (i.e., Cd2+ or Co2+) that are used to block Ca2+ current and are known to shift steady-state inactivation curves to the right (1). The steady-state inactivation ofI K has a more shallow potential dependence thanI to;V 0.5 values are reported between −77 and −114 mV (4, 11). Occasionally, a noninactivating residual outward current is observed that persists in the presence of 4-AP and TEA and contributes 10–30% to peak outward current (4, 19, 35, 40, 42).
In our experiments, trial protocols to estimate steady-state inactivation of outward current revealed that conditioning steps of −100 mV were not sufficiently negative for complete current availability. This was achieved only with strongly negative conditioning pulses to −140 mV. The pattern of voltage dependence observed under these conditions appeared to be more complex than the simple sum of the well-characterized currentsI to andI K. Therefore, we decided to evaluateI peak andI late separately and to temporarily use a special nomenclature for the various current components. Normalized steady-state inactivation curves exhibited three distinct current fractions (a,b, andr) for each of the separately analyzed I peakand I late. Because the amplitudes of a ofI peak andI late were found to correlate significantly, they were supposed to represent a single current component. By analogy, r ofI peak andI late were also considered as one current component. This reduced the number of distinguishable outward current components to four:a, r, and b ofI peak andb ofI late.
It should be pointed out that despite the significant correlation,I peak ofa was smaller thanI late according to the regression line. However, only in the case of an ideal noninactivating current (I peak =I late), should one expect a positive correlation, with the regression line characterized by a slope of 1 and an intercept at the origin. In the case of an inactivating current as shown here (I peak >I late; Fig.3 D), a significant positive correlation should also be observed, albeit with a different regression line (slope < 1 but > 0, intercept at origin). A Cole-Moore shift might also contribute to the fact that the regression line misses the origin. In the majority of the cells, this led to the impression of an increase in transient current with less negativeV m (between −140 and −80 mV). Differences in both activation and inactivation time constants ofI K andI to could confound the time course; using a subtraction approach, we found thatI K apparently activated and inactivated more slowly thanI to (Fig. 3; seeVoltage dependence of activation of outward current components). Therefore,I to determines peak current amplitude, whereas the slower-activatingI K is underestimated. Indeed, it has been reported that the delayed rectifierI K activates 10-fold more slowly than the transient outward currentI to (4).
I K was a delayed rectifier-like current with a shallow steady-state inactivation curve at rather negative potentials (V 0.5 −93 mV; Table 1, Figs. 1, 3); it was insensitive to 4-AP but was concentration dependently blocked by TEA in low millimolar concentrations (Fig. 5). Furthermore,I K was inhibited by quinidine and by clofilium (Fig. 5). Although the block by clofilium (30 μM) ofI late was significantly stronger than that ofI peak, this was not an argument against a single current component but could be explained on the basis of the time-dependent blocking mechanism of clofilium (10, 26). Therefore, the properties ofI K resemble those previously reported (4, 11). The transient current componentI to had a steep steady-state inactivation curve with a midpoint at −38 mV (Table1, Fig. 1), was blocked by millimolar concentrations of 4-AP and by HpTx3 (Fig. 4), but was insensitive to TEA (Fig. 5). In addition, this component was predominant in subepicardial myocytes. Such properties are identical with the published characteristics of the transient outward current (4, 6, 12).
In addition, we have presented evidence for another current component termed I Kx that was partially superimposed onI to but clearly distinct from it.I Kx andI to differed significantly with respect to the midpoints of steady-state inactivation curves, i.e., −28 vs. −38 mV (Table 1). Their relative amplitudes did not correlate (Fig. 2); 10 mM TEA blockedI Kx but did not affect I to (Fig.5), whereas HpTx3 blockedI to but did not influence I Kx(Fig. 4). These data suggest thatI Kx is a separate entity and cannot be considered as a noninactivating part ofI to. On the other hand, the differences in effects of 4-AP or clofilium onI Kx andI to amplitudes did not allow this conclusion. With 4-AP the difference of block was too small, and with clofilium the difference could be attributed to time-dependent channel block (10, 26).I Kx could represent the small, sustained outward current inhibited by nanomolar concentrations of isoproterenol (33). However,I Kx was not altered by the adenylyl cyclase activator forskolin (data not shown). Furthermore, I Kxdoes not resemble the sustained outward current Iso present in human atrial myocytes, because Iso was absent in ventricular cells and was TEA insensitive (2).
In every myocyte, >25% of total outward current persisted asI ss atV m positive to −20 mV (Table 1, Fig. 1). This current component was attenuated by the K+-channel blockers 4-AP, TEA (10 mM), and clofilium (Fig. 5), was markedly reduced by substituting Cs+ or TEA for K+ in the intracellular solution (Fig. 6), and was inhibited by lowering the extracellular Cl− concentration (Fig. 6). These findings suggest that K+ and Cl− contribute toI ss. At present, we can only speculate about its nature. For instance, a Ba2+-sensitive background K+ current has been described to be active at action potential plateau, albeit in guinea pig ventricular myocytes (5). However, in our cells the relative amplitude of the residual current was only slightly reduced on exposure to Ba2+ (1 mM, −15%;n = 4 experiments). Nonselective currents carried by monovalent cations have been reported in human atrium (2, 15) and in rat ventricle (27). In rat, this current is blocked in a voltage-dependent manner by extracellular Ca2+ and could therefore contribute to I ssunder our conditions.I ss was significantly reduced after substitution of extracellular Cl− with methanesulfonate (Fig. 5 B), indicating that a Cl− conductance contributes to background current (see also Ref. 24). However, the poor selectivity of Cl−-channel blockers precludes more detailed characterization of the Cl− conducting pathway (Ref.24; unpublished observations).
The data presented so far support the hypothesis that outward current in rat ventricular myocytes consists of more than two distinct components, i.e.,I K,I to,I Kx, andI ss. These components are distinguished on the basis of their time courses, potential dependence of availability, and pharmacological profile. The properties of I Kand I to are consistent with published data. However, the sustained K+ currentI Kx and the noninactivating steady-state currentI ss appear to be novel phenotypes that could nevertheless match those identified by K+-channel genes.
Relation to cloned voltage-dependent K+ channels.
In rat ventricle, a multitude of depolarization-activated K+ channels have been identified at the mRNA level, whereas only two current phenotypes,I to andI K, have been distinguished (7, 13, 16, 30). Heterologous expression of Kv channels allows their pharmacological profiling. Our data on fractionb ofI peak (transient time course, inactivation kinetics, pharmacological profile, transmural gradient) are consistent with the idea of its identity toI to and confirm the role of proteins of the Kv4 family in generating I toin rat ventricle (compare HpTx3 data). The kinetic properties of component a resemble those of the delayed rectifierI K. However, the present data and our indirect experimental approach do not allow a definite conclusion about the nature of the Kv channel responsible forI K. In particular, component a, i.e.,I K, is a sustained current without transmural gradient and is sensitive to block by TEA and hanatoxin (blocker of Kv2.1 and Kv4.2; Ref. 36) but is insensitive to 4-AP, dendrotoxin (blocker of Kv1.2; Ref. 13), and HpTx3 (blocker of Kv4.2; Ref. 32). This pattern could give rise to the hypothesis that Kv2.1 might underlieI K. Finally, although the current componentsI Kx andI ss were unaffected by either of the toxins used, this lack of effect cannot be interpreted in terms of absence of the respective Kv gene products (particularly Kv1.2). Moreover, the reason for this finding is unclear and requires further investigation. In any case, Kv channel gene products can only be related to native currents with great caution, because of the inherent differences in heteromultimeric composition and accessory subunits of K+ channels between expression systems and native myocytes (31).
In conclusion, the great diversity in expression of K+ channels in myocardial cells determines the regional variability of cardiac action potential waveform (6, 8). The underlying K+channels are subject to developmental change, to modulation by neurotransmitters, or to differential pathophysiological alteration (e.g., hypertrophy-associated action potential prolongation because of decreased I to and diminished expression of Kv4.2/3; Refs. 34, 37, 38). The possible consequences include increased susceptibility to arrhythmias and altered pump function of the heart. Under physiological conditions, the observed diversity of K+ currents and action potential waveforms has pronounced effects on patterns of myocyte shortening and the inotropic state (18, 33).
We have shown that outward current in rat ventricular myocytes consists of more than the two previously described currents. In addition toI to andI K, a small sustained K+ current (I Kx) and a noninactivating steady-state current (I ss) contribute to total outward current. Knockout of individual K+-channel genes by means of antisense oligonucleotides in cultured myocytes should provide further insight into rat ventricular outward current components and their (patho)physiological roles in cellular repolarization and modulation of contractility.
The skillful technical assistance of Doris Petermeyer is gratefully acknowledged. The authors thank NPS Pharmaceuticals (Salt Lake City, UT) and Dr. Kenton Swartz (National Institutes of Health, Bethesda, MD) for the gifts of heteropodatoxin and hanatoxin, respectively.
Address for reprint requests: H. M. Himmel, Institut für Pharmakologie und Toxikologie, Universitätsklinikum Carl Gustav Carus, TU Dresden, Karl-Marx-Str. 3, D-01109 Dresden, Germany (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 1999 the American Physiological Society