Decreased Sarcoplasmic Reticulum Calcium Content Is Responsible for Defective Excitation-Contraction Coupling in Canine Heart Failure

Transcription

1 Decreased Sarcoplasmic Reticulum Calcium Content Is Responsible for Defective Excitation-Contraction Coupling in Canine Heart Failure Ion A. Hobai, MD, PhD; Brian O Rourke, PhD Background Altered excitation-contraction (E-C) coupling in canine pacing-induced heart failure involves decreased sarcoplasmic reticulum (SR) Ca uptake and enhanced Na/Ca exchange, which could be expected to decrease SR Ca content (Ca SR ) and may explain the reduced intracellular Ca (Ca i ) transient. Studies in other failure models have suggested that the intrinsic coupling between L-type Ca current (I Ca,L ) and SR Ca release is reduced without a change in SR Ca load. The present study investigates whether Ca SR and/or coupling is altered in midmyocardial myocytes from failing canine hearts (F). Methods and Results Myocytes were indo-1 loaded via patch pipette (37 C), and Ca i transients were elicited with voltage-clamp steps applied at various frequencies. I Ca,L density was not significantly decreased in F, but steady-state Ca i transients were reduced to 20% to 40% of normal myocytes (N). Ca SR, measured by integrating Na/Ca exchange currents during caffeine-induced release, was profoundly decreased in F, to 15% to 25% of N. When Ca SR was normalized in F by preloading in 5 mmol/l external Ca before a test pulse at 2 mmol/l Ca, a normal-amplitude Ca i transient was elicited. E-C coupling gain was dependent on Ca SR but was affected similarly in both groups, indicating that intrinsic coupling is unaltered in F. Conclusions A decrease in Ca SR is sufficient to explain the diminished Ca i transients in F, without a change in the effectiveness of coupling. Therefore, therapeutic approaches that increase Ca SR may be able to fully correct the Ca handling deficit in heart failure. (Circulation. 2001;103: ) Key Words: sarcoplasmic reticulum calcium heart failure It is widely believed that heart muscle contraction is triggered by membrane depolarization via the process of Ca-induced Ca release (CICR 1,2 ): Ca entry through the Ca channels during the action potential triggers a large release of Ca from the sarcoplasmic reticulum (SR). The intracellular Ca (Ca i ) transient is thought to be the summation of local Ca release events occurring at junctions between the transversetubule sarcolemmal membrane and the SR ( diads ), where L-type Ca channels and SR Ca release channels are closely apposed. Such local control can explain the main features of CICR, including (1) a high gain (a large Ca release for a small Ca trigger), (2) a graded response (corresponding to the size of Ca current, I Ca,L ), and (3) the absence of uncontrolled regenerative Ca release (under normal conditions). 2,3 Local control is epitomized by the observation of Ca sparks in cardiomyocytes, 4,5 and the relationship between I Ca,L and spark frequency has been used as a measure of the intrinsic E-C coupling efficiency, or local gain. 6,7 In a larger sense, however, the total gain of CICR will depend on both the local effectiveness of coupling between I Ca,L and SR release ( ) and on the SR Ca load (Ca SR ). Limited information is available about which of these factors may change in human heart failure. Beuckelmann et al 8 showed that Ca i transients of ventricular myocytes isolated from failing human hearts have a reduced amplitude without a change in I Ca,L and also suggested a decrease in Ca SR (ryanodine had a small effect on the Ca i transient in failing cells). A decreased Ca SR has also been reported by Lindner et al 9 and Pieske et al, 10 but no information is available about whether is changed in the failing human heart, as reported in some rat models of heart disease. 6,7 The canine pacing-induced heart failure model shows a pattern of E-C coupling changes similar to that of humans: I Ca,L of normal amplitude triggers Ca i transients and contractions much smaller than normal, whereas SR Ca uptake is markedly impaired. 11 The present study investigates whether the defect in E-C coupling is due to a decrease in Ca SR,in, or in both. Methods Induction of heart failure, 12 isolation of midmyocardial cardiomyocytes, 11,13 single-cell electrophysiology studies, 11 and Ca measure- Received September 6, 2000; revision received October 12, 2000; accepted October 13, From the Institute of Molecular Cardiobiology, Division of Cardiology, Department of Medicine, Johns Hopkins University, Baltimore, Md. Correspondence to Brian O Rourke, PhD, Johns Hopkins University, Department of Medicine, Division of Cardiology, 844 Ross Building, 720 Rutland Ave, Baltimore, MD bor@jhmi.edu 2001 American Heart Association, Inc. Circulation is available at

2 1578 Circulation March 20, 2001 Figure 1. I Ca,L measured in presence of BAPTA at 37 C. Typical I Ca,L elicited with depolarizations from 40 to between 30 and 0 mv (A and C, for N and F, respectively) and between 10 and 40 mv (B and D). E, Current-voltage relationship for I Ca,L in N and F. F, Voltage-dependence of I Ca,L activation in N and F. See text for details. G, Ca entry via I Ca,L, calculated as current integral, over first 50 ms (as shown in inset and left 2 bars) and for whole 500-ms depolarization (right 2 bars). No significant difference between groups was observed. ments 11 were carried out as previously described. After 3 to 4 weeks of tachycardic pacing (240 bpm), mongrel dogs showed clinical signs and hemodynamic manifestations of heart failure, including elevated end-diastolic pressures and decreased contractility. 12 Isolated myocytes were whole-cell patch-clamped at 37 C. The external solution was K-free (to eliminate inward K currents) and contained (in mmol/l) NaCl 140, CaCl 2 2, MgCl 2 1, HEPES 5, glucose 10, and niflumic acid 0.1 (Sigma, Cl current blocker), ph adjusted to 7.4 with NaOH. Superfusing solutions were rapidly changed by use of a solenoid-controlled heated switching device. For selective measurement of I Ca,L, the pipette solution contained (in mmol/l) CsCl 110, MgATP 5, HEPES 10, MgCl 2 0.4, glucose 5, tetraethylammonium (TEA) 20, and BAPTA 5, ph 7.2. Cs and TEA inhibited outward K currents, and BAPTA was used to buffer Ca i. The pipette-to-bath liquid junction potential was minimal ( 2.7 mv) and was not corrected. For E-C coupling experiments, the external solution also contained 30 mol/l tetrodotoxin (TTX, Sigma; to block sodium current, I Na ). The pipette solution contained (in mmol/l) potassium glutamate 125, KCl 19, MgCl 2 0.5, MgATP 5, NaCl 10, and HEPES 10, ph 7.25, and 50 mol/l indo-1 (pentasodium salt, Calbiochem). The pipette-to-bath liquid junction potential was 20 mv and was corrected. Indo-1 fluorescence was excited at 365 nm and recorded at wavelengths of 405 and 495 nm. 11 Cellular autofluorescence was recorded before rupturing of the cell-attached patch and was subsequently subtracted. Ca i was calculated according to the equation Ca i K d [(R R min )/(R max R)], 14 with a K d of 844 nmol/l, 15 R min 1, R max 4, 2, and R 405/495 fluorescence ratio. Results We previously showed that basal I Ca,L density is unaltered in canine tachypacing-induced heart failure at room temperature 12 and with minimal Ca i buffering. 11 In the latter study, however, no provision was made to block potential contaminating currents (eg, background K currents, Ca-activated Cl current, and Na/Ca exchange [NCX] currents), and the kinetics of I Ca,L were not examined. Therefore, we first examined in detail the time course of I Ca,L inactivation and Ca influx via I Ca,L under controlled conditions in the presence of the fast Ca buffer BAPTA. Analysis of I Ca,L Depolarizations of 500 ms from 80 mv to various membrane potentials (V M ) were applied every 4 seconds and were preceded by a 100-ms prepulse to 40 mv to inactivate I Na. I Ca,L amplitude (measured as the peak inward minus end of pulse current) was similar in normal myocytes (N) and myocytes from failing canine hearts (F) over the potential range from 30 to 40 mv (Figure 1A through 1E). The voltage dependence of whole-cell Ca conductance [G I Ca,L /(V M E rev ), where E rev is the apparent reversal potential for I Ca,L ] normalized to maximum conductance (G max ) is shown in Figure 1F. Data were fitted with a Boltzmann equation 16 (G/G max 1/{1 exp[(v 0.5 V M )/k]}), and the activation parameters were unchanged in F (Table 1). From I Ca,L recordings obtained at 10 mv, Ca entry via L-type Ca channels was measured as the integral of I Ca,L for either the first 50 ms after depolarization ( I Ca,L ; the period most relevant to triggering CICR) or for the whole pulse. No statistically significant differences were found between N and F (Figure 1G and Table 1). To investigate whether the kinetics of I Ca,L inactivation are changed in F, the time course of inactivation of I Ca,L (at 10 mv) was fitted with a double exponential 16 : I Ca,L (t) A 1 exp( t/ 1 ) A 2 exp( t/ 2 ) C, where 1 and 2 are the fast and slow time constants, respectively, and A 1 and A 2 are the corresponding amplitudes of the exponential function. The 2 time constants, as well as the proportion of total I Ca,L inactivated with each of them, were also unchanged in F (Table 1). E-C Coupling Experiments I Ca,L,Ca i transients, and Ca SR were measured in indo-1 loaded myocytes after conditioning trains of voltage-clamp depolar-

3 Hobai and O Rourke E-C Coupling in Heart Failure 1579 TABLE 1. I Ca,L Measurements in the Presence of BAPTA N F Average SEM n Average SEM n P V 0.5,mV / / k / / Amplitude, pa/pf / / Integral 50 ms, fc/pf / / Integral 500 ms, fc/pf / / ,ms / / ,ms / / A 1 /(A 1 A 2 ) / / Apart from activation parameters, all other measurements were performed at 10 mv. A 1 /(A 1 A 2 ) measures the proportion of total Ca,L I that was inactivated with 1. For this and Table 2, statistical significance was tested with an unpaired t test, variance not assumed equal; P values are given. n is given as No. of cells/no. of hearts. F/N izations at different stimulation frequencies. Figure 2A illustrates typical membrane currents and Ca i transients elicited with depolarizations from 80 to 10 mv at 1-Hz stimulation (pulse duration 0.3 seconds). After steady state was attained, the train was interrupted and 10 mmol/l caffeine was applied rapidly, inducing Ca release from the SR. During caffeine application, Ca i rose rapidly and then decayed exponentially, extruded (mainly) by NCX, which generated an inward current. Ten seconds after caffeine was washed off, it was reapplied to ensure that all SR Ca was released with the first application. The second caffeine application produced no change in Ca i but induced a small ( 100-pA), slowly activating inward current (traces marked with asterisk in Figure 2B). Figure 2, C and D, illustrates a similar experiment in F. Compared with N, the fast component of the depolarizationevoked Ca i transient was markedly reduced. Caffeine application triggered a reduced rise in Ca i and only a small inward Figure 2. I Ca,L,Ca i transients, and Ca SR. A, Membrane current and Ca i signals elicited by a depolarization from 80 to 10 mv. Steady-state records obtained at 1-Hz stimulation. Removing external K (see Methods) induced an outward shift in membrane current, visible as a timeindependent outward current at 10 mv, which did not influence I Ca,L measurement (data not shown). B, Membrane current and Ca i signals elicited by caffeine (10 mmol/l) application (shown with bar). A second caffeine application (*) elicited no SR release but still elicited a slowly activating inward current. NCX current generated during Ca extrusion was measured as difference between 2 membrane currents. C and D, A similar experiment in F elicited much smaller Ca i transients and caffeine releases. Note 5 times expanded scale in D, bottom. E through G, Average Ca i / t (E), Ca i (F), and Ca SR (G) in N and F, as obtained at 0.25, 0.5, and 1 Hz stimulation frequencies. *Statistical significance.

4 1580 Circulation March 20, 2001 TABLE 2. E-C Coupling Mechanisms N F Freq, Hz Average SEM n Average SEM n P F/N Ca,L I, pa/pf / / / / / / Ca,L I, fc/pf / / / / / / Ca i, nmol/l / / / / / / Ca i / t, nmol L 1 ms / / / / / / SR Ca, mol/l / / / / / / Caffeine Ca i, nmol/l / / / / / / Freq indicates frequency of stimulation. n is given as No. of cells/no. of hearts. NCX current. Just as in N, a second caffeine application induced no Ca i rise and a slowly activating inward current. Similar results were obtained at frequencies of 0.5 and 0.25 Hz (pulse duration 0.5 seconds; not shown). I Ca,L amplitude and I Ca,L showed a trend toward lower values in F, but this difference was not statistically significant (Table 2; see Discussion). Compared with I Ca,L, a disproportionately large and unequivocal decrease was observed in the Ca i transients recorded in F. Both the rate of rise of Ca i ( Ca i / t; Figure 2E) and the amplitude of the early rapidly rising phase of the Ca i transient ( Ca i ; Figure 2F) were markedly reduced at all stimulation frequencies, the former being only 6% to 16% of N (Table 2). NCX current generated by caffeine Ca release was measured as the difference between currents recorded during the first and second caffeine applications. Ca SR (as total [Ca], Ca T ) was determined by integrating NCX current by use of the equation Ca T ( mol/l) 76 NCX current integral (pc)/ cell capacitance (pf), as previously described 17 (see also Negretti et al 18 ). Ca SR in N was between 80 and 120 mol/l, values comparable to reports in rat cells. 18 In contrast, F had markedly reduced Ca SR at all the stimulation frequencies used (14% to 25% of N, Figure 2G and Table 2). The amplitude of the caffeine-induced Ca i transients (another indication of Ca SR ) was similarly reduced in F (Table 2). A more exact Ca SR estimation should take into account the components of cytosolic Ca removal via non-ncx mechanisms (sarcolemmal Ca pump and mitochondrial Ca buffering), representing 28% and 12% of total non SR-mediated Ca transport in N and F, respectively. 17 Therefore, all Ca SR values reported here underestimate the difference between groups and could be corrected by multiplying by 1.28 and 1.12 for N and F, respectively. How Much of the Ca i Transient in N Is Due to Ca Entry? Depolarization-evoked Ca i transients were recorded in N immediately after a caffeine release, and thus with SR Ca load depleted (Figure 3A). Ca i / t in the first Ca i transient after caffeine was greatly reduced, to 17% of the precaffeine steady-state value (Figure 4B). I Ca,L was also slightly increased, indicating a partial relief from Ca-dependent inactivation (Figure 4A). Ca i / t in N cells with SR depleted was close to the steady-state Ca i / t in F, as discussed above. This result suggested that most of the Ca i / t in F was likely to be due to transsarcolemmal Ca entry via I Ca,L and reverse NCX, with little contribution from SR release. Although the caffeine release experiments indicated that there was still a residual Ca SR in F at steady state (14% to 25% of N), it appears that this pool is not readily releasable, perhaps because of a low at low Ca SR. 19 Increasing SR Load in F Because these experiments suggested that the marked decrease in Ca i transient amplitude in failing cells could be explained by the decrease in SR Ca load and did not necessitate the postulation of a decrease in, we tested whether adjustment of Ca SR to similar levels in N and F resulted in Ca i transients with similar characteristics. Myocytes from failing hearts were superfused with an external solution containing 5 mmol/l Ca while a train of depolarizations was applied at 0.5 Hz (see Figure 3B for the voltage protocol). After reaching a steady state in 5 mmol/l Ca (Figure 3C), the solution was rapidly changed (for 1 pulse) to 2 mmol/l Ca. In this way, Ca SR could be maintained at a higher level while I Ca,L was instantaneously restored to that

5 Hobai and O Rourke E-C Coupling in Heart Failure 1581 Figure 3. Effects of modifying Ca SR in N and F. A, Component of Ca i transient due to Ca entry is illustrated by comparison of Ca i transient in N at steady state (0.5 Hz; *) and in first pulse after caffeine-induced Ca release (unmarked traces). B through E, Increasing Ca SR in F, by protocol shown in B, results in a large steady-state Ca i transient in 5 mmol/l Ca (C) and a normalized Ca i transient during a short switch to 2 mmol/l Ca (D). Steady-state Ca SR in 5 mmol/l Ca was measured with caffeine (E). See text for details. typically observed at 2 mmol/l Ca. The Ca i transient recorded with the intercalated pulse in 2 mmol/l Ca (Figure 3D) was thus triggered by a control I Ca,L but had Ca SR equal to the steady-state load in 5 mmol/l Ca. The latter was measured by caffeine application (Figure 3E) immediately after steady state was attained once more in 5 mmol/l Ca. Mean I Ca,L was close to the steady-state I Ca,L in 2 mmol/l Ca shown in the previous experiments (Figure 4A), and Ca SR was close to N levels (Figure 4C). Ca i / t was also similar to N (Figure 4B), consistent with the hypothesis that the defect in E-C coupling is primarily due to impaired SR Ca loading. This experiment was performed in 16 cells from 3 F hearts (16/3) compared with 16/4 N. Figure 4D illustrates the results of this experiment in individual cells. The relationship between CICR gain [calculated as ( Ca i / t)/ I Ca,L 20 ] and Ca SR (as NCX current integral, in fc/pf, not transformed to [Ca T ]) was similar in N cells and F cells with increased SR load. Within individual cells, CICR gain was usually well correlated with Ca SR. When all cells were plotted together, more scatter in the data was evident, but it was evident that N and F data were interspersed. In each cell, we calculated as ( Ca i / t)/( I Ca,L Ca SR ) and found no difference between N and F (Figure 4E). Discussion Consistent with a previous study from this laboratory, 11 we report here that CICR gain (as defined to include both and Ca SR ) is decreased in heart failure: membrane depolarization triggered markedly reduced Ca i transients, although no significant difference could be seen in I Ca,L. The decrease in CICR gain could be explained by a similarly marked reduction in Ca SR, whereas no change in was observed. I Ca,L Measurements We measured I Ca,L both with Ca i buffered and in the presence of physiological Ca i transients. Both measurements showed a tendency toward a decreased I Ca,L density in F, but the difference was not statistically significant. This trend suggests that a decrease in I Ca,L could potentially contribute to the defective E-C coupling in this model but does not account for the majority of the difference between groups. CICR Gain Versus the Effectiveness of Coupling The concept of CICR gain usually means the magnitude of SR Ca release triggered by a given I Ca,L. 2,21 Because in these previous studies 2,21 Ca SR was kept constant, any variations in CICR gain were equivalent to variations in. In the present study, we needed to differentiate between Ca SR and, because a decrease in either could induce the decrease in CICR gain in heart failure. We measured SR release flux as Ca i / t, and we measured both Ca entry and Ca SR as integrals of I Ca,L and NCX current, respectively. CICR gain was dramatically reduced in F because of the decreased Ca SR. This deficit in F could be reversed when Ca SR was increased to normal levels, indicating that there is no decrease in. Pacing-Induced Canine Heart Failure Model After 3 to 4 weeks of tachycardic pacing, canine hearts exhibit a markedly decreased contractility in vivo (with

6 1582 Circulation March 20, 2001 Figure 4. A through C, Summary data for experiments shown in Figure 3. Left 2 bars show, for comparison, I Ca,L (A), Ca i / t (B), and Ca SR (C) in N and F at 0.5-Hz stimulation. Middle 2 bars show slightly increased I Ca,L (A) and profoundly decreased Ca i / t (B) evident in N for first pulse after caffeine release (compare with experiment shown in Figure 3A). Right 2 bars show that with unchanged I Ca,L (A) and normalized Ca SR (C), F cells showed Ca i transients with Ca i / t within normal range (B; compare with Figure 3, B through E). D, CICR gain [calculated as ( Ca i / t)/ I Ca,L plotted against Ca SR ] for N and F. For individual cells, CICR gain was well correlated with Ca SR, but this correlation was less evident when regression lines were fit to all of the data (r 0.32 and 0.48, P 0.21 and 0.06 for N and F, respectively). E, (see text) was not significantly different in F vs N. increased end-diastolic pressure and decreased systolic rate of rise of left ventricular pressure 12 ). Consistent with that, isolated cells show Ca i transients that are slowed and of decreased amplitude. 11 Previous studies from this laboratory indicated a main defect in Ca i removal mechanisms, 11 with both a decrease in SR Ca uptake and an increase in NCX Ca extrusion, which, in itself, might be expected to decrease Ca SR. This is also the prediction of a refined computer model of the failing heart cell. 22 The experiments reported here confirm the marked decrease in Ca SR in this model, which is largely responsible for the decreased Ca i transients. Although we purposely controlled the experimental conditions for this study, extrapolation of this conclusion to action potential evoked Ca transients in vivo requires consideration of the changes in action potential evident in heart failure, as well as differences in modulatory factors (external and internal), which may be altered by the disease. The present findings are inconsistent with the hypothesis that the E-C coupling abnormalities in heart failure are the result of defects at the level of the SR release channel, such as decreases in either the number of channels 23,24 or rate of release, 23 or the effectiveness of coupling ( ). Further investigation will be necessary to determine whether our conclusions are model-specific or can be generalized to humans.

7 Hobai and O Rourke E-C Coupling in Heart Failure 1583 Comparison With Other Heart Failure Models Although a decreased Ca SR is thought to be a major part of the defective E-C coupling in human heart failure, 9 there have been few estimates of Ca SR in animal models of heart failure and even fewer quantitative measurements. In a rat hypertrophy model without overt heart failure, cell contractility has been shown to be increased in parallel with an increase in Ca SR. 25 In a model of overt heart failure induced by combined aortic insufficiency and stenosis in rabbits, 26 a 26% decrease in (externally stimulated) myocyte twitches was reported, with a trend toward a decrease in both the Ca i transient and Ca SR (estimated as the amplitude of caffeine-induced Ca i transients; see note on method below). To the best of our knowledge, only 1 other animal model of heart failure, a postinfarction rat model, has shown a significant decrease in the Ca i transient associated with a reduction in Ca SR. 27 In contrast, changes in, with unchanged Ca SR, have been reported in 2 rat models, in 1 of which diminished Ca i transients were explained by a decrease in. 6 In another study, increases in Ca i transients in cells from spontaneously hypertensive rats could be accounted for by an increase in (again, with no change in Ca SR7 ). The change in could be due a change in the number of Ca sparks (of otherwise normal kinetics and amplitude) triggered by a given I Ca,L6 or to changes in the Ca spark amplitude ( big sparks 7 ). Whether a changed will induce a change in the steady-state Ca i transients in these (rat) models is controversial (because of compensatory changes in Ca 28 SR ), but the present experiments demonstrate that this is not the case in the canine pacing induced model. Relevance to Human Disease Midmyocardial cells isolated from failing human hearts showed unchanged I Ca,L and decreased Ca i transients. 8 The latter could be due to a decrease in SR Ca content, which has been qualitatively estimated from the amplitude of caffeineinduced Ca i transients by Lindner et al. 9 It is worthwhile to note that interpretation of the results of such experiments is not straightforward. For example, it is known that the peak of the Ca i rise can be blunted by Ca extrusion through NCX, because significant inward current can be recorded by the time Ca i reaches its peak. 29 Because NCX may be increased in F, it may blunt the peak of the caffeine transient more in F than in N, biasing the results. In the present study, we looked at both the caffeine-evoked Ca i transient and the NCX current integral under voltage-clamp conditions, and reassuringly, both parameters led to the same conclusion. A defective SR loading in heart failure was also indicated by Pieske et al. 10 Using rapid cooling contractures in intact cardiac muscle strips, they showed that in failing muscles, Ca SR decreased both during rest and with increasing frequency of stimulations, unlike the response of nonfailing muscle. 10 It is important to note that the E-C coupling changes reported here and previously 11 in the pacing canine model are remarkably similar to human disease. Just as in human disease, F cells showed decreased Ca i transients and a decreased Ca SR, with little change in I Ca,L. In the present study, we could also show that is not changed, because F cells with a normalized Ca SR showed normal Ca i transients. No information is yet available on whether the coupling effectiveness is changed in human heart failure. If this conclusion can be extrapolated to the human disease, it could have an important clinical significance, indicating that therapeutic strategies that could restore Ca SR (by stimulation or overexpression of sarcoplasmic/endoplasmic reticulum Ca2 -ATPase [SERCA], for example 30 ) could be expected to fully restore E-C coupling and cell contractility. If a decrease in were involved as well, this would not necessarily be the case. Acknowledgments This study was supported by NIH grants RO1-HL (to Dr O Rourke) and the Specialized Center of Research (SCOR) on Sudden Cardiac Death and Heart Failure (NIH P50-HL-52307). We are grateful to SCOR investigators Eduardo Marbán, David Kass, and Gordon Tomaselli for helpful discussions and guidance. We also thank Richard Tunin for help with dog preparation and surgery. References 1. Fabiato A, Fabiato F. Calcium-induced release of calcium from the sarcoplasmic reticulum of skinned cells from adult human, dog, cat, rabbit, rat, and frog hearts and from fetal and new-born rat ventricles. Ann N Y Acad Sci. 1978;307: Wier WG, Egan TM, Lopez-Lopez JR, et al. Local control of excitationcontraction coupling in rat heart cells. J Physiol. 1994;474: Stern MD. Theory of excitation-contraction coupling in cardiac muscle. Biophys J. 1992;63: Cheng H, Lederer WJ, Cannell MB. Calcium sparks: elementary events underlying excitation-contraction coupling in heart muscle. Science. 1993;262: Lopez-Lopez JR, Shacklock PS, Balke CW, et al. Local calcium transients triggered by single L-type calcium channel currents in cardiac cells. Science. 1995;268: Gomez AM, Valdivia HH, Cheng H, et al. Defective excitationcontraction coupling in experimental cardiac hypertrophy and heart failure. Science. 1997;276: Shorofsky SR, Aggarwal R, Corretti M, et al. Cellular mechanisms of altered contractility in the hypertrophied heart: big hearts, big sparks. Circ Res. 1999;84: Beuckelmann DJ, Nabauer M, Erdmann E. Intracellular calcium handling in isolated ventricular myocytes from patients with terminal heart failure. Circulation. 1992;85: Lindner M, Erdmann E, Beuckelmann DJ. Calcium content of the sarcoplasmic reticulum in isolated ventricular myocytes from patients with terminal heart failure. J Mol Cell Cardiol. 1998;30: Pieske B, Maier LS, Bers DM, et al. Ca 2 handling and sarcoplasmic reticulum Ca 2 content in isolated failing and nonfailing human myocardium. Circ Res. 1999;85: O Rourke B, Kass DA, Tomaselli GF, et al. Mechanisms of altered excitation-contraction coupling in canine tachycardia-induced heart failure, I: experimental studies. Circ Res. 1999;84: Kaab S, Nuss HB, Chiamvimonvat N, et al. Ionic mechanism of action potential prolongation in ventricular myocytes from dogs with pacinginduced heart failure. Circ Res. 1996;78: Hobai IA, Howarth FC, Pabbathi VK, et al. Voltage-activated Ca release in rabbit, rat and guinea-pig cardiac myocytes, and modulation by internal camp. Pflugers Arch. 1997;435: Grynkiewicz G, Poenie M, Tsien RY. A new generation of Ca 2 indicators with greatly improved fluorescence properties. J Biol Chem. 1985; 260: Bassani JW, Bassani RA, Bers DM. Calibration of indo-1 and resting intracellular [Ca] i in intact rabbit cardiac myocytes. Biophys J. 1995;68: Isenberg G, Klockner U. Calcium currents of isolated bovine ventricular myocytes are fast and of large amplitude. Pflugers Arch. 1982;395:

8 1584 Circulation March 20, Hobai IA, O Rourke B. Enhanced Ca-activated Na/Ca exchange activity in canine pacing-induced heart failure. Circ Res. 2000;87: Negretti N, Varro A, Eisner DA. Estimate of net calcium fluxes and sarcoplasmic reticulum calcium content during systole in rat ventricular myocytes. J Physiol (Lond). 1995;486: Shannon TR, Ginsburg KS, Bers DM. Potentiation of fractional sarcoplasmic reticulum calcium release by total and free intra-sarcoplasmic reticulum calcium concentration. Biophys J. 2000;78: Sipido KR, Wier WG. Flux of Ca 2 across the sarcoplasmic reticulum of guinea-pig cardiac cells during excitation-contraction coupling. J Physiol (Lond). 1991;435: Hussain M, Orchard CH. Sarcoplasmic reticulum Ca 2 content, L-type Ca 2 current and the Ca 2 transient in rat myocytes during -adrenergic stimulation. J Physiol. 1997;505: Winslow RL, Rice J, Jafri S, et al. Mechanisms of altered excitationcontraction coupling in canine tachycardia-induced heart failure, II: model studies. Circ Res. 1999;84: Yamamoto T, Yano M, Kohno M, et al. Abnormal Ca 2 release from cardiac sarcoplasmic reticulum in tachycardia-induced heart failure. Cardiovasc Res. 1999;44: Vatner DE, Sato N, Kiuchi K, et al. Decrease in myocardial ryanodine receptors and altered excitation-contraction coupling early in the development of heart failure. Circulation. 1994;90: Delbridge LM, Satoh H, Yuan W, et al. Cardiac myocyte volume, Ca 2 fluxes, and sarcoplasmic reticulum loading in pressure-overload hypertrophy. Am J Physiol. 1997;272:H2425 H Pogwizd SM, Qi M, Yuan W, et al. Upregulation of Na /Ca 2 exchanger expression and function in an arrhythmogenic rabbit model of heart failure. Circ Res. 1999;85: Zhang XQ, Tillotson DL, Moore RL, et al. Na /Ca 2 exchange currents and SR Ca 2 contents in postinfarction myocytes. Am J Physiol. 1996; 271:C1800 C Eisner DA, Trafford AW, Diaz ME, et al. The control of Ca release from the cardiac sarcoplasmic reticulum: regulation versus autoregulation. Cardiovasc Res. 1998;38: Bassani RA, Bassani JW, Bers DM. Mitochondrial and sarcolemmal Ca 2 transport reduce [Ca 2 ] i during caffeine contractures in rabbit cardiac myocytes. J Physiol. 1992;453: del Monte F, Harding SE, Schmidt U, et al. Restoration of contractile function in isolated cardiomyocytes from failing human hearts by gene transfer of SERCA2a. Circulation. 1999;100: