INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

ABNORMALITIES OF SYSTOLIC and diastolic function are key aspects in the pathophysiology of human heart failure. The impaired systolic function, evident as a decreased ejection fraction, has been directly related to contractile defects of individual myocytes. Myocyte contractile defects are thought to result primarily from changes in the calcium uptake and release mechanisms (11) and from the sensitivity of the myofibrils to calcium (29). Diastolic dysfunction includes a restrictive left ventricular filling pattern, increase in chamber stiffness, incomplete relaxation, and interstitial fibrosis (28). At the myocyte level, contraction is initiated by calcium influx with subsequent calcium release from the sarcoplasmic reticulum (SR) (22), and relaxation is a result of calcium removal from the cytoplasm by sarco(endo)plasmic reticulum Ca²⁺-ATPase (SERCA) and the sarcolemmal sodium/calcium exchanger. Other mechanisms of calcium removal (sarcolemmal Ca²⁺-ATPase and mitochondrial uptake) are probably too slow to participate in the decay of the calcium transient (2, 3, 30).

In human heart failure, the expression of calcium regulatory proteins has been shown to be abnormal, with decreased levels of SERCA and increased levels of sodium/calcium exchanger protein (1, 12, 13, 18, 19, 23, 26). It has been suggested that a critical defect of myocyte function is the decreased capacity of the SR to accumulate calcium as a result of decreased SERCA activity (12). Recently, it has been (17) shown that in human heart failure the SR calcium load is decreased, resulting in lower steady-state and caffeine-induced calcium transients. Given that SR function is depressed and sodium/calcium exchange function is increased in heart failure, it is possible that calcium influx via reverse-mode sodium/calcium exchange and calcium efflux via forward-mode sodium/calcium exchange are more important processes in the contraction and relaxation of failed myocytes. Increased calcium influx via reverse-mode exchange activity during the action potential could be a source of intracellular calcium to help "replace" the SR as a mechanism for the activation of the myofibrils.

Prolongation of the action potential is a consistent finding in human heart failure (5, 6). The mechanism of action potential prolongation likely involves changes in potassium ion channel expression (6). It has been suggested that the increase in the duration of the action potential results in a greater influx of calcium through a prolongation of the plateau phase (5, 11). Contractions in the failing heart are longer and slower, suggesting that the prolonged action potential duration may be related to slower contraction and relaxation. If electrogenic sodium/calcium exchange makes a significant contribution to contraction and relaxation in failed human myocytes, then this might be the process that links the prolonged action potentials and contractions.

In this study, we examined the relationships between cellular contraction and relaxation to sodium/calcium exchange activity and the repolarization phase of the action potential. Our hypothesis was that the sodium/calcium exchanger made a significant contribution to contraction and relaxation in the failed human heart. Our objectives were to separate contraction and relaxation due to sodium/calcium exchange activity from other mechanisms of contraction (SERCA, L-type calcium current) and relaxation (SERCA, sarcolemmal Ca-ATPase, etc.) and to define the relationship between the myocyte action potential duration and relaxation. Furthermore, we sought to determine whether calcium influx via sodium/calcium exchange plays an important role in contraction of failed human myocytes.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell isolation. We isolated left ventricular myocytes from 29 failed human explants. Etiologies included idiopathic dilated cardiomyopathy (11 hearts) and ischemic dilated cardiomyopathy (18 hearts). A blood vessel in a piece of left ventricular free wall (10 × 8 cm) was cannulated and perfused with calcium-free Krebs solution containing (in mM) 130 sodium chloride, 25 NaHCO₃, 12.5 dextrose, 5.4 potassium chloride, 1 lactic acid, 1.2 MgSO₄, 1.2 NaH₂PO₄, and 2 pyruvic acid for 30 min, followed by collagenase-containing Krebs solution (180 U/ml) for ~40 min at 37°C. The resulting rod-shaped myocytes had clear striations and did not demonstrate spontaneous contractions.

Electrophysiology. All experiments were performed on single myocytes with no obvious contact with neighboring cells. Suction-type patch pipettes were prepared from borosilicate glass (1B150F, World Precision Instruments, Sarasota, FL) using a two-stage pipette puller (model P-87, Sutter Instruments, Novato, CA). The pipette tip was lightly heat polished before use and had a tip resistance of 2 to 4 M. The pipette was attached to a patch-clamp amplifier (Axoclamp II, Axon Instruments, Foster City, CA). A drop of cell suspension was placed in a Lucite perfusion chamber (~20 × 10 × 10 mm) mounted on the heated stage of an inverted microscope (Nikon Diaphot-TMD). The experimental chamber was perfused with various bath solutions (see above) at a rate of 1 to 2 ml/min. Before beginning a voltage-clamp experiment, we superfused cells with the bath solution for 3-5 min.

Drugs. The sodium/calcium exchange inhibitor Kanebo 7943 (KB-7943) was a kind gift from Kanebo (Osaka, Japan) (15) and was prepared as a 10 mM stock solution in DMSO and stored at 0°C until use. Thapsigargin (Sigma), a specific inhibitor of SERCA (16), was added to the bath solution, yielding a final concentration of 1 µM.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Action potential experiments. Contractions induced by action potentials in failed human myocytes were often biphasic with an initial rapid component followed by a second, slower component (Fig. 1A). Evidence of a biphasic contraction was apparent in 12 of 28 cells studied at 37°C in response to an action potential at 0.5 Hz. The rapid component began just after the upstroke of the action potential and had a fast rate of shortening and relaxation [dL/dt, 12.79 ± 7.48 µm/s; +dL/dt, 12.89 ± 7.79 µm/s, respectively; n = 12, means ± SD], suggesting that it was mediated by SR calcium release and uptake. The second component had a significantly slower rate of shortening and relaxation [dL/dt, 2.20 ± 1.10 µm/s; and +dL/dt, 2.13 ± 1.09 µm/s, respectively; n = 12, means ± SD], suggesting that it was caused by a different process. At stimulation frequencies of 0.2 to 1 Hz, the relaxation of the slower component followed the repolarizing downstroke of the action potential, whereas relaxation of the rapid component occurred during the plateau of the action potential and did not change with rate-dependent shortening of the action potential. At stimulation frequencies >0.5 Hz, the action potential duration shortened and only monophasic contractions were observed. Under these conditions relaxation followed the repolarizing downstroke of the action potential. These results suggested that the myocyte relaxation mechanism has voltage-dependent and -independent components and suggests that removal of intracellular calcium by forward-mode sodium/calcium exchange together with calcium uptake by the SR are the processes involved in relaxation.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Action potentials (APs) and accompanying contractions were recorded using perforated-patch technique in Tyrode solution with 1 mM Ca2+ at 0.2, 0.5, and 1.0 Hz at 37°C (A) and at 25°C (B). Relaxation of contraction was closely linked to AP downstroke at 37°C but not at 25°C, showing effect of temperature-dependent sodium/calcium exchanger in relaxation. C: relationship of contraction duration to AP duration for 18 myocytes (10 hearts) recorded at 37°C compared with 6 myocytes (4 hearts) recorded at 25°C. Index of time to 90% relaxation (CONT90) to time to 90% repolarization (APD90) was significantly longer at 25°C (P < 0.05).

Effect of temperature. Previous studies have shown that the sodium/calcium exchanger has a higher temperature sensitivity than SERCA in guinea pig myocytes (27). Therefore, contractions at 25°C should rely less on the sodium/calcium exchanger. The duration of the contraction was longer at 25°C than at 37°C (Table 1), and the relaxation phase occurred well after the repolarization of the action potential (Fig. 1B). Under these conditions, relaxation had no obvious voltage dependence, supporting the idea that electrogenic sodium/calcium exchange played less of a role in relaxation with the SR functioning as the principal mechanism of relaxation at 25°C. Action potential and contraction data from 18 myocytes (10 hearts) recorded at 37°C (0.2, 0.5, and 1 Hz) were compared with data from six myocytes (4 hearts) recorded at 25°C. The index derived by dividing the duration of the contraction by the duration of the action potential was significantly greater at 25 vs. 37°C, indicating that relaxation was influenced to a lesser extent by the repolarization of the action potential. (Fig. 1C).

Voltage-clamp experiments. To further explore the basis of the two components of contraction, failed myocytes were studied using voltage-clamp techniques. From a holding potential of 50 mV, after six conditioning pulses to +10 mV (to load the SR), a 1,500-ms step was made to +10 mV to maximally activate L-type calcium current and trigger contraction. We repeated this protocol and shortened the duration of the test pulse (after conditioning pulses) with each subsequent step by 150 ms. In Tyrode solution with physiological filling solution at 37°C, the long initial contraction (1,500-ms pulse) was usually biphasic with a rapid and a slow component. The relaxation rate of the fast component did not change with repolarization until pulse duration was less than 500 ms. Relaxation of the slower component always followed the repolarization at the end of the test pulse (Fig. 2A). These results showed that relaxation of the second component was voltage dependent and was consistent with a sodium/calcium exchange mechanism.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Voltage-clamp experiments were carried out on isolated myocytes at 37°C in Tyrode solution with 1 mM Ca2+ from a holding potential of 50 mV to a test potential of +10 mV and stepped back to 80 mV (A) or 50 mV (B). The initial step was 1,500 ms with subsequent steps shortened by 150 ms. Myocytes were conditioned with 6 pulses to +10 mV before each test step. Myocytes displayed either a biphasic (A) or monophasic (B) contraction with relaxation that followed repolarization of test pulse. Whole cell currents included a prominent early outward, reverse-mode, sodium/calcium exchange current (arrowhead). Similar results were seen in 24 myocytes (11 hearts). Early repolarization during inactivation of calcium current prevented maximal activation and contraction. In sodium- and potassium-free solution (C), there was no evidence of reverse-mode exchange current, and contraction was not linked to repolarization. Similar results were seen in 4 myocytes (2 hearts). Likewise, experiments at 25°C in Tyrode solution with 1 mM Ca2+ demonstrated no sign of a reverse-mode exchange current, and relaxation did not follow repolarization (4 myocytes, 2 hearts) (D).

Effect of thapsigargin. To further explore the basis of the two components of contraction, action potentials and contractions were recorded at 37 °C before (Fig. 3A) and after exposure to 1 µM thapsigargin (Fig. 3B). Thapsigargin specifically inhibits SERCA (16), leaving the L-type calcium current and reverse-mode sodium/calcium exchange as the principal sources of calcium influx to activate the myofilaments and forward-mode sodium/calcium exchange as the primary route of calcium removal. Action potential characteristics after exposure to thapsigargin did not change significantly (Fig. 3B). Thapsigargin eliminated the first component of contraction. The remaining thapsigargin-insensitive contraction was very slow to reach peak, consistent with direct activation of the myofilaments from calcium influx during the action potential (Fig. 3B). The relaxation of this slower contraction depended on the repolarization of the action potential as before. In voltage-clamp experiments, thapsigargin abolished the rapid but not the slower component of contraction, consistent with elimination of SR calcium release. Different methods were used to "condition" the cell and to establish steady-state contractions in the absence of SR function. Action potentials at 0.5 Hz, voltage steps to +10 mV (maximal L-type calcium current, Fig. 3C) and to +50 mV (physiological overshoot, L-type calcium current and exchanger current, Fig. 3D), and to +135 mV (exchanger current only, data not shown) were used to initiate contractions and condition the cell. Slow contractions were initiated, and in all cases the relaxation was closely linked to the repolarization at the end of the voltage step (Fig. 3, C and D, bottom). These results provide further proof that the relaxation of the first component of contraction was SR mediated, and the relaxation of the second component was voltage dependent and therefore exchanger mediated under these conditions. Contractions initiated by voltage steps to +10 mV (predominantly L-type calcium current), to +50 mV (L-type calcium current and exchanger current), and to +135 mV (exchanger current only) resembled the slow component of the action potential-triggered contraction and suggested an important role for reverse-mode calcium entry in the initiation of contraction. These results showed that in the absence of a functional SR, the sodium/calcium exchanger played a significant role both in contraction and relaxation of the myocyte. Similar results were seen in eight myocytes (four hearts).

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Myocytes were exposed to 1 µM thapsigargin. AP characteristics were unaffected at 0.5 and 1 Hz (A, control; B, thapsigargin). Accompanying contractions were slower at both frequencies and were closely linked to the AP downstroke during relaxation (B). Voltage-clamp experiments used test steps to +10 mV (C) or +50 mV (D) to condition and to initiate contractions. Under these conditions, contractions were linked to repolarization. Contractions initiated by step to +50 mV were much larger than those triggered by L-type calcium current (ICaL). Similar results were seen in 8 myocytes (4 hearts).

Effect of KB-7943. Myocytes at 37°C were exposed to 10 µM KB-7943, a specific blocker of sodium/calcium exchange (15). The duration of the action potential was significantly decreased after exposure to 10 µM KB-7943 (Fig. 4A, mean action potential duration at 90% repolarization at 0.5 Hz: 37°C, 1,100 ± 360 ms control vs. 533 ± 144 ms KB-7943, P < 0.05, n = 3), suggesting that the sodium/calcium exchange current was an important contributor to the action potential plateau. The accompanying contractions were monophasic with relaxation not clearly linked to the repolarization of the action potential. These results show that when contraction and relaxation are strictly SR-mediated (exchanger blocked), the rate of relaxation is independent of voltage. In voltage-clamp experiments, the current records of myocytes exposed to 10 µM KB-7943 had no obvious exchanger current (Fig. 4B), and contractions were monophasic (no exchanger component) and not linked to the repolarization after the voltage step (Fig. 4C). These results showed that in the absence of the exchanger, relaxation does not depend on voltage. Similar results were seen in four myocytes (three hearts).

View larger version (19K): [in this window] [in a new window]   Fig. 4.   APs and accompanying contractions from failed myocytes before (A, left) and after exposure to 10 µM Kanebo 7943 (KB-7943) (a sodium/calcium exchange blocker) for 5 min (A, right). B: voltage-clamp experiments were carried out on isolated myocytes at 37°C in Tyrode solution with 1 mM Ca2+, exposed to 10 µM KB-7943. Myocytes exposed to KB-7943 had inward ICaL without reverse-mode exchange current (B). Contractions were small and had delayed relaxation that did not follow repolarization at the end of the test pulse (C). Similar results were seen in 4 myocytes (3 hearts).

Reverse-mode exchange. To understand the contribution of reverse-mode exchange to contraction, we used a protocol that stepped (without conditioning pulses) from 50 mV to +50 mV, the approximate plateau potential, where L-type calcium current is small and reverse-mode exchange current should be large. These experiments were carried out after a 60-s rest period. The net current was outward during the entire pulse, suggesting reverse-mode exchange occurred continuously during repeat loading steps (Fig. 5A, top). Repeat steps (without conditioning pulses) from 50 mV to +50 mV initiated contractions with increasingly larger first components (Fig. 5A, bottom) with each successive pulse, suggesting that reverse-mode exchange loaded the SR. Similar to experiments using conditioning pulses, complete relaxation relied on the repolarization at the end of the test pulse. When conditioning pulses were included in the protocol (Fig. 5B), the rapid component of contraction (SR mediated) was followed by a slower exchanger-mediated contraction that did not relax until repolarization at the end of the test pulse. When the test steps were preceded by conditioning pulses (Fig. 5B), the net current during the pulse was initially outward followed by net inward current for the remainder of the pulse, suggesting that the loading protocol maintained higher levels of intracellular calcium.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Voltage-clamp experiments were carried out on isolated ventricular myocytes at 37°C in Tyrode solution with 1 mM Ca2+. Multiple test steps from a holding potential (Vhold) of 50 mV to a test potential (Vtest) of +50 mV without conditioning pulses were used to demonstrate loading of sarcoplasmic reticulum by reverse-mode sodium/calcium exchange (A). Whole cell current (top) and contractions are recorded (bottom). Arrowhead indicates 1st, 2nd, and 10th steps and progressively larger contractions with each step. B: similar experiment with conditioning pulses and decreasing pulse duration. The slow component of contraction followed the decreasing pulse duration and showed the voltage dependence of relaxation. Without conditioning pulses net exchange current was outward; with conditioning pulses net current was inward, suggesting a higher level of intracellular calcium. Both current traces showed a rapid early reverse-mode exchange current. C: a single myocyte subjected to repetitive test pulses from a holding potential of 50 mV. Steps to +10 mV activated ICaL and small contractions (left). Steps to +50 mV activated reverse-mode exchange current, small ICaL, and progressively larger contractions (middle). Steps to +80 mV activated large reverse-mode exchange current and large contractions followed by "tail" calcium currents and smaller contractions (right). Similar results were seen in 6 myocytes (4 hearts).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

These results demonstrate for the first time the important role of the sodium/calcium exchanger in contraction and relaxation of failed human ventricular myocytes. Contractions had two distinct components in response to an action potential or a voltage-clamp pulse and suggested that contraction of human myocytes is mediated through two mechanisms: a rapid SR-mediated phase and a slower exchanger-mediated phase. Depending on the apparent degree of SR dysfunction, contractions varied from rapid biphasic ones (Figs. 1, 2A, and 3A) to slower monophasic ones (Fig. 2B). The kinetics of the rapid component were independent of voltage. Once the contraction was initiated a rapid shortening and relengthening of the fast component occurred. We concluded this to be a SR-mediated event because this component was blocked by the SERCA inhibitor thapsigargin (Fig. 3).

As the pulse duration became very short in voltage-clamp experiments (i.e., 150-450 ms), the repolarization decreased the magnitude of the contraction, suggesting that the exchanger was able to rapidly remove intracellular calcium, effectively competing with the SR for calcium and abbreviating the contraction. This reduced magnitude of contraction due to shortening of the test pulse may partially explain the negative force-frequency relationship observed in failed human myocytes attributable to smaller calcium transients (9, 20). Sipido et al. (25) observed smaller calcium transients at higher stimulation frequencies in failed human myocytes and suggested that this was the basis for the negative force-frequency relationship. A similar phenomenon was observed in rat myocytes when rapid repolarization interrupted the rising phase and abbreviated the intracellular calcium transient (7). The sodium/calcium exchanger has been shown to compete with the SR for removal of intracellular calcium during very brief depolarizations (4). The rate of shortening of the second component was slower and clearly demonstrated a process of initiation of contraction separate from the SR. Our results suggest that reverse-mode sodium/calcium exchange was the source of activator calcium for direct activation of the myofilaments producing the second slow component.

The appearance of biphasic contractions in response to action potentials varied from cell to cell. We believe that the relative amount of SR dysfunction dictated this result. Approximately one-half of the myocytes we studied had single-phase contractions, the rate of shortening and relengthening of which was less. It appeared that these myocytes had severely limited SR function and depended to a greater extent on sodium/calcium exchange as a source of activator calcium. It was always possible to elicit some element of a biphasic response using a long voltage-clamp step (1,500 ms), suggesting that the SR plays at least some role in contraction even in severely diseased cells.

Complete relaxation of failed myocytes was closely linked to the repolarization at the end of the action potential or voltage-clamp test pulse. The kinetics of relaxation of the rapid (SR) component of contraction were independent of voltage and matched the rate of shortening of the rapid component. The kinetics of the slow component were strongly voltage dependent and always depended on the repolarization at the end of the action potential or voltage-clamp test pulse. Because the action potential duration is prolonged in human heart failure, the linking of relaxation to the action potential may contribute to the slower relaxation seen in failed human myocytes.

The identity of the two components of contractions has been addressed by Isenberg et al. (14). In guinea pig and bovine isolated myocytes, the rapid component was independent of voltage, and they concluded that the rapid component depended on intracellular calcium stores (SR).

In our experiments, further evidence about the identity of the two components of contraction was provided by repeating our experiments at 25°C. Because the sodium/calcium exchanger is highly temperature sensitive (21, 27), the effect of the exchanger on relaxation could be reduced by recording action potentials and contractions at 25°C. Although the duration of the action potentials and contractions were longer at 25°C, the relaxation phase was no longer linked to the repolarization of the action potential. Likewise, when voltage-clamp experiments were carried out at 25°C, relaxation was no longer dependent on the repolarization and changing pulse duration. We believe by using this method of reducing exchanger activity we were able to visualize the contribution of the exchanger to relaxation.

We further explored the role of the exchanger in contraction and relaxation of failed myocytes using KB-7943, a specific blocker of sodium/calcium exchange (15). KB-7943 caused significant shortening of the action potential, suggesting that the sodium/calcium exchange current is a significant contributor to the action potential plateau as suggested by Schouten et al. (24). Contractions resulting from shortened action potentials and from voltage-clamp steps in the presence of KB-7943 were small and independent of repolarization. In the voltage-clamp experiments, the characteristic exchange current was blocked by KB-7943, leaving only L-type calcium current as a source of activator calcium. The meager contractions in the absence of sodium/calcium exchange are strong evidence of the importance of the exchanger in the initiation of contraction and SR loading in failed myocytes. The slow rate of relaxation in the presence of KB-7943 provides further evidence of the prominent role of the exchanger in relaxation.

Thapsigargin has been successfully used as a selective blocker of SERCA in cardiac myocytes (16). By exposing myocytes to thapsigargin, we were able to virtually eliminate SR function. Thapsigargin had no effect on the action potential waveshape but dramatically slowed the rate of contraction similar to the result reported by Davia et al. (8). In voltage-clamp experiments when L-type calcium current was the primary source of activator calcium, the contractions were very small and likely resulted from direct activation of the myofibrils. When the voltage-clamp step activated reverse-mode sodium/calcium exchange, the resulting contraction was much larger but still slow, consistent with direct activation of the myofibrils. Without a functional SR, reverse-mode exchange contributes to the direct activation of the contractile machinery at physiological overshoot voltages (+50 mV, Fig. 4B) more effectively than L-type calcium current (+10 mV, Fig. 4C). With the SR incapacitated, the sodium/calcium exchanger was left as the principal mechanism of relaxation and was closely linked to repolarization. Voltage-dependent relaxation was apparent regardless of the stimulus for the contraction. These data strongly support the hypothesis that the exchanger is a primary mechanism of relaxation in failed human myocytes.

Further evidence supporting the key role of sodium/calcium exchange in contraction and relaxation comes from experiments showing the capability of the exchange-mediated calcium influx to progressively increase the size of contraction with repetitive voltage-clamp steps (Fig. 5, B and C). This result shows indirectly that reverse-mode exchange can load the SR and initiate contraction. Consistent with other experiments, relaxation of these contractions was voltage dependent, indicating reliance on the exchanger for relaxation.

The results of the present investigation suggest that the upregulation of the exchanger in human heart failure (1, 18, 26) provides an extra source of activator calcium via reverse-mode exchange for activation of the myofibrils in the face of severe SR dysfunction and also functions as a major process of relaxation via forward-mode exchange. Because the SR is reported to be downregulated in heart failure (1, 12, 13, 18, 19, 26), we believe the upregulation of the exchanger may be an adaptation for SR loss.

The limitations of our study surround the difficulty in obtaining normal human tissue for comparison. Similar studies on normal human myocytes are needed before firm conclusions about human heart failure can be made. There was a large degree of heterogeneity among the different etiologies of heart failure and even among individual myocytes from a single preparation. The heterogeneity may reflect the chronic and complex nature of end-stage human heart failure. No difference between failed hearts of different etiologies could be discerned.

In summary, we conclude that the sodium/calcium exchanger plays a key role in both contraction and relaxation of failed human myocytes. Our data suggest that in human heart failure there are varying degrees of contractile dysfunction that reflect the decreased level of SERCA. The increased exchanger activity may be an adaptation for this dysfunction, providing an alternative and/or complementary mechanism of contraction and relaxation.