Relaxation of Rabbit Ventricular Muscle by Na-Ca Exchange and Sarcoplasmic Reticulum Calcium Pump. Ryanodine and Voltage Sensitivity

Transcription

1 334 Relaxation of Rabbit Ventricular Muscle by Na-Ca Exchange and Sarcoplasmic Reticulum Calcium Pump Ryanodine and Voltage Sensitivity Donald M. Bers and John H.B. Bridge We studied relaxation during rapid rewanning of rabbit ventricular muscles that had been activated by rapid cooling. Rewanning from 1 to 30 C (in <0.5 second) activates mechanisms that contribute to the reduction of intracellular calcium concentration and thus relaxation (e.g., sarcoplasmic reticulum [SR] calcium pump and sarcolemmal Na-Ca exchange and calcium pump). Rapid rewanning in normal Tyrode's solution induces relaxation with a half-time (t^) of 217±14 msec (mean±sem). During cold exposure, changing the superfusate to a sodiumfree, calcium-free medium with 2 mm CoCl 2 (to eliminate Na-Ca exchange) slightly slows relaxation upon rewanning in the same medium (t w =279±44 msec). Addition of 10 mm caffeine (which prevents SR calcium sequestration) to normal Tyrode's solution during cold superfusion slows relaxation somewhat more (t 1/2 =37±31 msec) than sodium-free, calciumfree solution. However, if both interventions are combined (sodium-free+caffeine) during the cold exposure and rewarming, the relaxation is greatly slowed (t 1/1 =2,580±810 msec). These results suggest that either the SR calcium pump or, to a lesser extent, sarcolemmal Na-Ca exchange can produce rapid relaxation, but if both systems are blocked, relaxation is very slow. If muscles are equilibrated with 500 nm ryanodine before cooling, relaxation upon rewarming is not greatly slowed (t 1/2 =2±37 msec) even if sodium-free, calcium-free solution is applied during the cold and rewarming phases (t l/2 =305± msec). This result suggests that ryanodine does not prevent the SR from accumulating calcium to induce relaxation. Relaxation in the presence of 10 mm caffeine appears to depend on a simple 3 : 1 Na-Ca exchange since relaxation is slowed by extracellular sodium reduction but stays constant with simultaneous reduction of extracellular sodium concentration and extracellular calcium concentration (where [Na] 3 /[Ca] is held constant). Furthermore, relaxation in the presence of caffeine is slowed by membrane depolarization in a manner expected of a voltage-sensitive Na-Ca exchange. (Circulation Research 1989;5: ) During the cardiac action potential, calcium enters the cell via calcium channels (and perhaps also via Na-Ca exchange). This calcium may directly activate the myofilaments or may induce release of additional calcium from the sarcoplasmic reticulum (SR). 1-2 For relaxation to Division of Biomedical Sciences (D.M.B.), University of CaHBmia, Rirorside, California, and the Nora Eccles Harrison Cardiovascular Research and Training Institute and Department of Medicine (J.H.B.B.), University of Utah, Salt Lake City, Utah. Supported by grants from the National Institutes of Health (HL-W77 and HL-13348). J.H.B.B. is also supported by the Nora Becles Treadwefl Foundation and the Richard A. and Nora Eccles Harrison Fund for Cardiovascular Research. D.M.B. is the recipient of a National Institutes of Health Research Career Development Award (HL-Q152 ). Address for reprints: DT. Donald M. Bers, Division of Biomedicai Sciences, University of California, Riverside, CA Received October 12, 1988; accepted January 13, occur, calcium must be removed from the myofilaments. At least three mechanisms are thought to be involved in reducing myoplasmic calcium: 1) the ATP-dependent SR calcium pump, 2) the sarcolemmal Na-Ca exchange, and 3) the sarcolemmal Ca- ATPase pump. It is also possible that mitochondria contribute to relaxation by sequestering calcium 3 and other intracellular constituents that have very slow calcium-binding kinetics (relative to troponin C) could also be involved in relaxation. It is generally assumed that the SR calcium pump is the prtaeipaj aecfeaeisfla responsible for relaxation. However, it is becoming increasingly dear that Na-Ca exchange can transport sufficient calcium rapidly enough to be involved in contraction and relaxation. 47 Here we investigate the relative contributions of the SR calcium pump and Na-Ca exchange in cardiac muscle relaxation.

2 Bers and Bridge Na-Ca Exchange and SR Calcium Pump in Cardiac Relaxation 335 Rapid cooling of mammalian cardiac muscle (to 0-1 C in < 1 second) induces the SR to release calcium to the cytoplasm and consequently to activate a contracture (see Figure I.) 8-11 These rapid cooling contractures (RCCs) have been used as a relative index of SR calcium content that is available for release Cooling to 1 C apparently inhibits the mechanisms responsible for lowering cytoplasmic calcium (e.g., SR calcium pump and Na-Ca exchange) as well as other ion transport systems (e.g., membrane ion channels and Na,K- ATPase pump). Rewarming after an RCC leads to a transient increase of force that can be directly attributed to an increase in myofilament calcium sensitivity at higher temperatures while intracellular calcium concentration ([Ca],) is still elevated. Rewarming also reactivates the mechanisms responsible for relaxation, such as the SR calcium pump and Na-Ca exchange, which had been inhibited by the cold. Thus, the relaxation upon rewarming after the "rewarming spike" is likely to reflect the removal of calcium from the cytoplasm. The inhibition of ion transport in the cold allows us to modify the extracellular medium, for example, by removing extracellular sodium (Na,,) or adding caffeine, in the cold during an RCC and to study the way in which the rewarming relaxation is affected. Thus, cooling allows us to investigate the influence of ions, caffeine, and ryanodine on mechanical relaxation; hence, the contribution of Na-Ca exchange and the SR calcium pump on relaxation can be assessed. This approach may also allow us to investigate the action of ryanodine. Ryanodine is a neutral plant alkaloid that appears to bind specifically to the SR calcium-release channel and spanning (or foot) processes in skeletal and cardiac muscle Although the biochemical characterization of the ryanodine receptor is progressing rapidly, the way in which cellular calcium movements are modified in vivo by ryanodine are more difficult to evaluate.io,i2.i8-2o Rousseau et al 21 incorporated the SR calcium-release channel into lipid bilayers and found that ryanodine "locked" the channel in an open state, but with a lower than normal conductance. Another complicating factor is that whereas physiologically effective concentrations of ryanodine (1 nm-10 jim) accelerate calcium loss in SR vesicles,- very high ryanodine concentrations (>0.1 mm) may block SR calcium release.- Indirect approaches have been used to assess the calcium releasable from cardiac SR (e.g., extracellular calcium microelectrodes, caffeine, and RCCs) and have resulted in conclusions about ryanodine action that fit reasonably well with those from experiments with isolated vesicles. These studies have suggested that ryanodine increases SR calcium leak into the cytoplasm (particularly at rest). However, it also seems clear that during repetitive stimulation or with a steady level of calcium entry, the SR is able to accumulate a substantial amount of calcium in the presence of ryanodine (compared with control) This suggests that in vivo the SR calcium pump may be able to compete with the ryanodine-induced SR calcium leak, particularly when intracellular calcium is elevated. One aim of the present study is to evaluate whether, in the presence of ryanodine, the SR can still accumulate calcium rapidly enough to be responsible for the rapid phase of relaxation. Materials and Methods Thin papillary muscles or ventricular trabeculae ( mm in diameter) were obtained from the hearts of New Zealand White rabbits after intravenous administration of pentobarbital sodium ( 75 mg/kg). The ends of the muscle were tied with fine suture. One end of the muscle was attached to a fixed post and the other to a piezoresistive transducer (AE 875, SensoNor, Horten, Norway) in a 0.15-ml supervision chamber. The muscle was stimulated at 0.5 Hz by platinum plates in the lateral chamber walls during equilibration (~1 hour) and between protocols. The superfusate was a normal Tyrode's solution (NT) containing (mm) NaCl 140, KC1, MgCl 2 1, CaCl 2 2, glucose 10, and HEPES 5 at ph All solutions were equilibrated with 100% O 2, and the bath temperature was maintained at 30 C (except during cooling contractures). The flow rate in the chamber was 35 ml/min. In sodiumfree solutions, NaCl was replaced isoosmotically with UC1. In sodium-free, calcium-free solutions, CaCl 2 was replaced by 2 mm CoCl 2 to increase the rate of displacement (and washout) of interstitial calcium and also to inhibit calcium entry that might occur due to incomplete washout of interstitial calcium. Caffeine (10 mm) was added as a powder, and ryanodine ( nm) was added from an aqueous stock (1 mm). Ryanodine was purchased from Penick(lot #704-RWP-2, Lyndhurst, New Jersey). When relaxation was examined in test solutions that were sodium free, calcium free and/or containing caffeine, the muscles were also cooled in the same test solution. Switching to the test solution after initial cooling gave similar results, but our protocol maximized the time for replacement of the interstitial solution with the test solution before rewarming. Solenoid pinch valves were situated at the bath inlet, and the perfusion lines leading to these valves were jacketed with either water (at 30 C) or propylene glycopwater (1 = 3 at -2.0 C). At this flow rate, switching to the cold solution cooled the muscle surface to below 3 C in 1 second or less. Rewarming was similarly rapid. The muscle surface temperature was increased from 0 to 24 C in less than 300 msec and to 27 C in less than 1 second. The time courses of cooling and rewarming are somewhat faster than those previously described The RCCs and rewarming relaxations induced in this manner are highly reproducible.

3 33 Circulation Research Vol 5, No 2, August 1989 K. NT 3 S RCC B. NT 3 min RCC C. ONa OCa 0. Caffeine E. ONa OCa Caffeine 20 sec FIGURE 1. Tracings of a series of rapid cooling contractures (RCCs) in one rabbit ventricular strip equilibrated in normal Tyrode's solution (NT) and stimulated at 0.5 Hz between successive RCCs. In the first RCC, the muscle was at0 Cfor 3 seconds (A), and the subsequent RCCs were 3 minutes in duration. There is a break in the record for the 3-minute RCCs. The muscle was cooled 0.5 seconds after the last stimulated twitch in NT in each panel The cold and rewarmingsolution in each trace were: A and B, NT; C, sodium-free (ONa), calcium-free (OCa) with 2 mm cobalt; D, NT+10 mm caffeine; and E, sodium-free, calcium-free+2mm cobalt+10 mm caffeine. Results Mechanical Relaxation From Cooling Contractures When rabbit ventricular muscle is cooled to 0 C, contractures are induced that relax when the muscle is rewarmed to 30 C (Figure 1). We induced RCCs 0.5 second after the last stimulated twitch and maintained the muscle at 0 C for 3 seconds. This is sufficient time to reach the peak of the RCC. On rewarraing the muscle to 30 C, the usual rewarming spike and rewanning relaxation are observed. If the muscle is maintained in the cold solution for 3 minutes, the contracture slowly declines (Figure IB). However, the rewarming relaxation rate is not affected by the slow decline in contracture tension that takes place in the cold solution. When the rates of relaxation of the 3-secoad and 3-miaute RCC are normalized to the peaks of the rewarmiog spikes, they are nearly identical (Figure 2, Table 1), despite the fact that the rewarding spice of the former is almost twice that a the latter. Thav^he relaxation rate does not depend on the tension from which relaxation began. Moreover, the half-time for RCC relaxation ( 200 msec) isoirfy about ttvice as leug as the half-time for twitch relaxation (~ W0 msec). We attribute this difference to tbe fnjte time required to rewarm the entire muscle to 30 C. We assume tbat rewarming induces Time (sec) FIGURE 2. Graph of the rewarming relaxation phase of the two rapid cooling contractures (RCCs) in Figure 1A and IB after normalization for the peak value of tension at the start of relaxation. That is, tension at t=0 seconds corresponds to the peak of the rewarming spike and is defined as 100%. relaxation because some processes remove calcium from the vicinity of the myofilaments. The Effect of Ionic Composition on Mechanical Relaxation To investigate the contribution of Na-Ca exchange to the rewarming relaxation, muscles are cooled and rewarmed in solutions lacking sodium and calcium (Figure 1C and Figure 3). More than 3 seconds are required to exchange the extracellular space. However, we find that 3 minutes is sufficient to induce most of the effects attributable to the change in solution, while minimizing the relaxation of the RCC. We assume that at cold temperatures transsarcolemmal transport in sufficiently slow that the extracellular space can be largely exchanged without significantly altering intracellular ionic contents. If both the cold and rewarming solutions are sodium-free and calcium-free, rewarming relaxations are only slightly slowed (Figure 1C and TABLE 1. Influence of Na» Caffeine, and Ryanodlne on the Rewarming Relaxation of Rapid Cooling Conbracturej NT (peak) NT (3 min) ONa OCa Caffeine ONa caff NT Ryanodine ONa ryanodine n RCC height [% of control NT (3-min)] 108± ±5 104±12 90±9 89 ± Spike height [% of control NT (3-min)] ± ±11 37±4 Half-time of relaxation (msec) 202 ±30 217±14 279±44 37±31 2,580±810 2Q2±20 2±37 305± Values are mean±sem. RCC, rapid cooling contracture; NT, normal Tyrode's solution; ONa, sodium-free; OCa, calcium-free. The values for RCC amplitude and rewarraing spike amplitude are normalized to the values during the control 3-minute NT RCC.

4 Bers and Bridge Na-Ca Exchange and SR Calcium Pump in Cardiac Relaxation Time (sec) FIGURE 3. Graph of the rewarming relaxation phase of the four rapid cooling contractures in Figure IB, 1C, ID, and IE (respectively, from left to right). Tension is normalized as in Figure 2. ONa, sodium-free; Caff, caffeine; NT, normal Tyrode 's solution. Figure 3, Table 1). This result indicates that the relaxation rate is largely independent of external sodium and calcium. Since Na-Ca exchange cannot extrude calcium in the absence of external sodium, we conclude that some other process, possibly the SR calcium pump, is responsible for relaxation under these circumstances. Extracellular calcium was also replaced by cobalt in this protocol to prevent reversal of the Na-Ca exchange by the sodium-free solution upon rewarming, that is, intracellular sodium (Na^-dependent calcium influx. In the absence of cobalt (or EGTA), calcium entry may take place because more time is required to wash out interstitial calcium than sodium. 2 Cobalt accelerates the washout of interstitial calcium and also inhibits calcium entry. Qualitatively similar results were obtained with 5 mm EGTA (rather than CoCl 2 ), but EGTA was somewhat less effective. Recovery from this cobalt exposure seemed to be complete, based on steady-state twitch tension (see Figure 1) or repeated control RCCs (Table 1). The RCC is not changed noticeably by a drastic alteration in external ion composition (sodium free, calcium free, and with CoCl 2 ; Figures IB and 1C, Table 1). This supports earlier findings from which we concluded that the RCC amplitude is principally determined by the SR calcium release and is not dependent on transsarcolemmal calcium movements in the cold solution. We have also confirmed this finding with RCCs in isolated myocytes where diffusional limitations, which could complicate the interpretation of multicellular experiments, are minimized (D.M. Bers and L.V. Hryshko, unpublished observations). The Effect of Caffeine and Ionic Composition on Mechanical Relaxation Muscles were cooled (and rewarmed) in solutions containing 10 mm caffeine (Figure ID). Caffeine was used to inhibit the reaccumulation of SR calcium so that relaxation would depend mainly on other processes. The amplitude of the rewarming spike is greatly increased by caffeine, presumably due to the action of caffeine to increase the calcium sensitivity of the myofilaments. 27 ' 28 The rate of relaxation, however, is only slightly slowed in the presence of caffeine (Figure 3, Table 1). This finding suggests that processes other than the SR calcium pump can induce relaxation. If the cold and rewarming solutions are sodium free and calcium free with 10 mm caffeine, relaxation is greatly slowed (Figure IE). This is clear from the normalized data in Figure 3 (ONa+Caff) and for pooled data in Table 1. Under these circumstances, complete relaxation was not always observed until NT was returned. Full relaxation (or to the plateau before NT) required an average of 14±2 seconds. Thus, in the presence of caffeine, relaxation becomes strikingly dependent on external sodium. Because rapid relaxation in the presence of caffeine requires Na o, we infer that Na-Ca exchange produces relaxation. We conclude that either the SR calcium pump or, to a slightly lesser extent, the Na-Ca exchange system is able to remove calcium from the myoplasm rapidly enough to effect nearly normal relaxation. However, if both of these systems are unavailable (Figure IE), relaxation can proceed only very slowly. It is not possible to deduce the mechanism responsible for the slow relaxation in Figure IE, but candidates are the sarcolemmal Ca-ATPase pump, mitochondria, or incompletely inhibited SR calcium pump (or Na-Ca exchange). We also conclude that the SR calcium pump and Na-Ca exchange are the main mechanisms responsible for removing calcium from the myoplasm during relaxation. Furthermore, these parallel processes must compete for myoplasmic free calcium during relaxation. Ryanodine and Relaxation Previous results indicate that although ryanodine induces a calcium leak in the SR, the SR can still accumulate calcium w e investigated the way that relaxation is modified by ryanodine. Relaxation was first measured before ryanodine treatment (Figures 4A and 4B). The muscle was then treated with 500 nm ryanodine for 30 minutes during continued 0.5-Hz stimulation. We find this to be sufficient for the maximal ryanodine effect and postrest contractions, and RCCs (for rests^l second) are strongly depressed, as previously described To observe RCCs in the presence of ryanodine, the muscle must be cooled very soon after the twitch (<1.5 seconds). 10 Rewarming the muscle in the presence of ryanodine revealed that relaxation was not appreciably slowed (see Figure 5 and Table 1). Differences often were most apparent in the terminal phase of relaxation, which was slowed slightly. The relaxation rate was also not much altered in ryanodine-treated muscle in sodiumfree, calcium-free medium (Figure 4D and Figure 5;

5 338 Circulation Research Vol 5, No 2, August 1989.ao'c o 1UU A. NT B. ONa OCa c a c <u. to ^^ eak a. k» O 50 t \ \ \ \ -, *\,«Na C. Ryan 0. ONa OCa Ryan JU l/ar" 20 sec FIGURE 4. Tracings of a series of rapid cooling contractures (RCCs) in one rabbit ventricular strip equilibrated in normal Tyrode's solution (NT) (supplemented with 500 nm ryanodine [Ryan] in panels C and D) and stimulated at 0.5 Hz between successive RCCs. Between panels B and C, the muscle was equilibrated with 500 nm ryanodine for 40 minutes. The duration of each RCC was 3 minutes. After the peak tension was attained in each RCC, the recording speed was temporarily reduced (for a variable interval). This is why there is an apparently steeply declining phase in each trace and also why the rewarming time in each trace is not exactly aligned. The cold and rewarming solution in each trace were: A, NT; B, sodium-free (ONa), calcium-free (OCa) with 2 mm cobalt; C, NT+500 nm ryanodine; and D, sodium-free, calcium-free+2 mm cobalt+500 nm ryanodine. ONa+Ryan). Ryanodine alone increases the halftime of relaxation only slightly (202 ±20-2 ±37 msec) and only slightly more in sodium-free medium (to 305 ± msec). This result with ryanodine is in striking contrast with the result with caffeine. That is, in the presence of caffeine, relaxation is highly := o M Time (sec) FIGURE 5. Graph of normalized rewarming relaxation phase of the rapid cooling contractures in Figure 4A, 4C, and 4D. Tension is normalized as in Figure 2. NT, normal Tyrode's solution; Ryan, ryanodine; ONa, sodium-free. n 40 Na/.04 Ca 140 Ha \ Tine (sec) FIGURE. Graph of normalized rewarming relaxation phases of a series of rapid cooling contractures in one muscle. The cold and rewarming solution were normal Tyrode's solution (NT)+10 mm caffeine (140 Na, solid curve), 40 mm sodium Tyrode+10 mm caffeine (40 Na, broken curve) and 40 mm sodium Tyrode with calcium reduced to 4 um+10 mm caffeine (40 Na/. 04 Ca, dotted curve). The muscle was reequilibrated in NT with stimulation at 0.5 Hz between the rapid cooling contractures. dependent, whereas in the presence of ryanodine it is not. We infer that in the presence of ryanodine, the SR can still accumulate calcium and cause relaxation. Membrane Potential Dependence of Na o -Dependent Relaxation If relaxation in the presence of caffeine is produced by Na-Ca exchange and if the Na-Ca exchange is voltage sensitive, then Nao-dependent relaxation should also be voltage sensitive. Na-Ca exchange in cardiac sarcolemma has a stoichiometry of 3 Na : 1 Ca and is both electrogenic and potential dependent in isolated vesicles We varied membrane potential (E m ) by changing extracellular potassium concentration ([K] o ) in the cold and rewarming solutions. However, to raise [K] o by 100 mm (to approach E m =0), we must reduce Na o concentration ([Na] o ) to 40 mm to maintain ionic strength and osmolarity. We measured the rate of relaxation in the presence of 10 mm caffeine and 40 mm sodium (using 100 mm lithium as substitute for sodium, Figure ). As expected, relaxation is slowed. If relaxation in the presence of caffeine is mainly due to a 3 Na : 1 Ca Na-Ca exchanger, it should be possible to simuhaneousry change [Na] o and extracellular calcium concentration ([Ca] o ) in a predictable manner to maintain a constant driving force on the Na-Ca exchanger. It is a simple matter to show that keeping the ionic activity (a) ratio (ana^/aca,, constant fulfills this goal. The equation for the Na-Ca exchange reversal potential (E,) as described by Mullins 32 for «=3 is as follows: E r =3E Na -2Ec. (1)

6 Bers and Bridge Na-Ca Exchange and SR Calcium Pump in Cardiac Rdaxation 339 TABLE 2. Influence of Simultaneous Reduction of Sodium and Calduin Concentration and of Depolarization on Relaxation In the Presence of 10 mm Caffeine Cold and rewarming solution Normal Tyrode's solution 40 mm sodium Caffeine Caffeine with 40 mm sodium Caffeine with 40 mm sodium, 4/tM calcium mm potassium 1 mm potassium 3 mm potassium 10 mm potassium mm potassium Values are mean±sem from six experiments. Then and _ 3RT, ana. RT, aca n o r,-, in m F anaj F Half time of relaxation (msec) 233±20 20±18 37±34 2,193±9 388±38 43±4 710±112 1,980 ± ±45 i(anajj_ E,- ]n (anaj (3) aca,, aca< where R is the universal gas constant, T is the absolute temperature, and F is the Faraday constant. Thus, if" anaj and Ca, do not change, then E r and the driving force on the exchanger (F^-Em) for a given E m will be constant as long as the (anaj 3 / aca,, ratio is constant. When both [Na] o and [Ca] o are simultaneously reduced to 40 mm and 4 fim, respectively, maintaining a constant (anaj'/aca,, ratio*, the relaxation is almost the same as with caffeine alone (Figure, Table 2). This is consistent with the contention that in the presence of caffeine, relaxation is mainly controlled by a simple 3:1 Na-Ca exchanger. Since relaxation is restored to normal by simultaneous Nao and Ca,, reduction, this condition can serve as the control for potassium substitution experiments, which are illustrated in Figure 7. The curve labeled K in Figure 7 is the same as the curve labeled 40 Na/.O4 Ca in Figure. The other relaxation time courses in Figure 7 were obtained by replacing 10, 30, or 100 mm of the LiCl with KC1 in the cold and rewarming solutions (with 40 mm Na, 4 fim Ca,* and 10 mm caffeine). Increasing Contaminant calcium must be considered in micromolar calcium solutions. The solutions for this series of experiments were made up with 34 /im added calcium, thus allowing a generous margin for contaminant calcium (12 jim, to give a total of 4 fim calcium). This is likely to be higher than the actual contaminant calcium (which we typically measure to be 3- ^im) but was chosen for the following two reasons: 1) During the 3-minute RCC, it is unlikely that interstitial [Ca] gets down as low as superfusate [Ca] (see above), but probably gets closer to (2) Time (sec) FIGURE 7. Graph of normalized rewarming relaxation phases of a series of rapid cooling contractures in one muscle. Between and before each rapid cooling contracture, the muscle was stimulated at 0.5 Hz in normal Tyrode's solution. The cold and rewarming solutions for each curve contained 40 mm sodium Tyrode with calcium reduced to 4 fim and 10 mm caffeine with variable amounts of potassium (replacing lithium) as indicated (in mm) by the labels on the curves. [K] o (and consequent depolarization) slows relaxation substantially. Pooled results from six experiments like this are summarized in Table 2, and Figure 8 shows the relaxation half-times for the different [K] and estimated E ra values. The apparent voltage dependence is steepest from -35 to 7 mv (the most positive potential used). This is expected since calculation of E r from Equation 1 (for ana,=7.5 mm and [Ca]i=200 nm) yields -34 mv. Thus, relaxation is slowed as we depolarize toward E, for Na-Ca exchange and is more dramatically slowed above the expected E r. The fact that relaxation is not completely prevented by depolarization above E r suggests that some other means of relaxation is still functional. When the muscle is depolarized by high [K] o, it is possible that non-inactivating voltage-dependent calcium channels (i.e., calcium window current) are opened such that a background calcium entry is slowing relaxation. To investigate this possibility, we repeated the experiments in Figure 7 with or superfusate [Na]. 2 * The half-time for twitch-tension decline upon abrupt reduction of [Ca] in muscles of the size used here is typically about seconds, and the interstitial [Ca] change may be extrapolated to be about 99% complete in 3 minutes. Thus, in some muscles the interstitial [Ca] may still be slightly higher than 4 /im at the time of rewarming. 2) During long RCCs, it is likely that intracellular [Na] slowly rises due to sodium-pump inhibition. Indeed, after long RCCs (3-10 minutes), a transient inotropic effect is observed that resembles cardiac gjycoside-induced inotropy. We can also estimate from Equation 3 that a 1 mm rise in ana, (from 7.5 to 8.5 mm) could be compensated by reducing [Ca] 0 an additional 11 jtm, and Er would still be unchanged. While these quantitative limitations must be kept in mind, we feel that the simple interpretation of the results stated in the text is correct (i.e., that in the presence of 10 mm caffeine, relaxation is largely determined by a 3:1 Na-Ca exchange).

7 340 Circulation Research Vol 5, No 2, August 1989 IT) a: ID X IT3 1 10K. 3K / /y'''caffeine 1K K _.. r Na-fpee B Calculated E (mv) FIGURE 8. Graph of relaxation half-times for rapid cooling contractures in the presence of 10 mm caffeine (circles) and sodium-free, calcium-free medium (squares), as a function of membrane potential (E m ), where E m is calculated assuming [K]j=140 mm and that E m is the same as the potassium reversal potential, E K. The rewarming relaxations were also repeated in the presence of 10 (im nifedipine (open symbols and dotted curves). This set of experiments (nifedipine) is from a different group of six muscles than that used in Figure and Figure 7 and Table 2, but a similar protocol is used. without 10 ym nifedipine in the cold and rewarming solution. Inclusion of 10 jum nifedipine in the cold and rewarming solution was sufficient to reduce the post-rcc twitches by more than 95%. Nifedipine accelerated relaxation somewhat at high [K] but had no effect at lower [K] (Figure 8). Thus there may be a small calcium current (with [Ca] 0 =4 yum) in the higher [K] relaxations in Figure 7 and Figure 8, but the conclusion is not changed. Relaxation in the absence of caffeine but in sodiumfree, calcium-free medium was insensitive to changes in E m in either the absence or presence of 10 /xm nifedipine (Figure 8, Na-free). Thus, the relaxation in the absence of Na-Ca exchange (presumably due to SR calcium uptake) is not E m sensitive. This contrasts with the E m dependence of relaxation in the presence of caffeine (presumably due to Na-Ca exchange). Discussion Rapid cooling of mammalian cardiac muscle to 0-1 C causes a release of SR calcium that then activates a contraction (RCC) initially described by Kurihara and Sakai 9 and Bridge. 8 The amplitude of these RCCs is a useful indicator of releasable SR calcium in intact cardiac muscle, which is a difficult quantity to assess. The transient increase in force upon rewarming (or rewarming spike) can be attributed to the warming-induced increase of myofilament calcium sensitivity observed in chemically skinned fibers. 13 In the present study, we have focused our attention on the relaxation observed after the peak of the rewarming spike. Relaxation under these circumstances should be due to the rapid reactivation of calcium transport systems that were inhibited at 0 C. There is undoubtedly temporal overlap between the rewarminginduced myofilament sensitization and calcium removal from the sarcoplasm. However, the point where tension begins to decline marks the time where calcium extrusion is the dominant process (and the myofilament sensitization may be nearly complete). Thus the time course of relaxation may provide information about the calcium movements responsible for relaxation, but at times near the peak of tension, the kinetics will be complicated by the myofilament sensitization and temperature inhomogeneity. Rewarming to 30 C is fast, but not instantaneous, and the final approach from C is the slowest phase. Thus, if there is a large difference in temperature sensitivity of the SR calcium pump and Na-Ca exchange, relaxation will appear more dependent on the process that is closer to its 30 C condition. We take advantage of the cold time to modify the ambient solution when relaxation is activated. These changes in solution have little effect on the RCC itself but can markedly change the time course of relaxation. In effect, we are using the RCC to hold the intracellular ionic conditions relatively constant with high [Ca]j while we modify the interstitial medium. It may be noted that we are measuring the relaxation of force rather than calcium removal from the cytoplasm or myofilaments per se. We use the terms almost interchangeably to facilitate discussion, with the implicit assumption that these processes are specifically and closely related (although the relation can be modified by caffeine, see below). The time course of relaxation does not appear to vary much as a function of force or [Ca] f at the time of rewarming (e.g., see Figure 2, Table 1). While this may suggest some functional characteristics of the calcium transport systems, for the present study we take advantage of this empirical finding simply so that we can compare relaxation time courses from different absolute force levels. Sarcoplasmic Reticulum Calcium Pump and Na-Ca Exchange in Rapid Cooling Contracture Relaxation A major conclusion of this study is illustrated in Figure 3. That is, either the SR calcium accumulation (which is caffeine-sensitive) or, to a slightly lesser extent, the Na-Ca exchange (which is dependent on Na o ) can induce relaxation at nearly normal rates, but when both systems are inhibited, relaxation is dramatically slowed. This result suggests that the SR calcium pump and Na-Ca exchange are two main mechanisms responsible for relaxation and they may be expected to compete with each other. The extent of the competition between the SR calcium pump and Na-Ca exchange is difficult to quantitate from the present experiments. One reason for this difficulty is the complicating effect of caffeine increasing myofilament calcium sensitivity Thus, if Na-Ca exchange is responsible

8 Bers and Bridge Na-Ca Exchange and SR Calcium Pump in Cardiac Relaxation 341 for relaxation in the presence of caffeine, it must pump [Ca]j down to a lower level to achieve a given level of (or full) relaxation. This effect of caffeine would then bias the results (e.g., Figure 1 and Figure 3, Table 1) toward making Na-Ca exchange (and other relaxation mechanisms) seem less able to produce relaxation than they might be in the absence of caffeine. Despite this effect, Na-Ca exchange still appears capable of inducing relaxation at nearly normal rates in the presence of caffeine. Comparison of relaxation in NT to that in sodiumfree, calcium-free medium may allow a simple estimate of the contribution of Na-Ca exchange without the complications of caffeine. The relaxations are not simply exponential, so that simple mathematical modeling is not straightforward. However, the relaxation half-time is increased by 22% when sodium-free, calcium-free medium is used to block Na-Ca exchange. This suggests that Na-Ca exchange might contribute significantly to relaxation under normal conditions. This conclusion is directly supported by experiments where muscles were recooled 1 second after the rewarming relaxation was triggered. 35 The amplitude of this second RCC was 0-80% of that of the first RCC. This finding suggests that the SR reaccumulated most of the calcium released at the first RCC. Furthermore, the second RCC in this protocol was about the same as the first RCC when the cold and rewarming solutions were sodium-free and calcium-free (preventing Na-dependent calcium extrusion). 35 These results, along with the present study, suggest that under normal conditions the SR is responsible for most of relaxation (~75%), and Na-Ca exchange may be responsible for a smaller but substantial component (~25%). Of course, these estimates are approximate and may be expected to vary under different conditions. Extension of Conclusions to Twitch Relaxation Net calcium efflux during individual cardiac contractions has been reported using extracellular calciumselective microelectrodes and calcium-sensitive dyes 4-3 ' 37 and attributed to Na-Ca exchange. Barcenas-Ruiz et al 7 and Bielfeld et al 38 have also reported an inward current attributed to Na-Ca exchange and associated with relaxation in isolated cardiac myocytes after depolarizing voltage-clamp pulses. Indeed there is extensive evidence for Na- Ca exchange electrogenicity and a transport stoichiometry of 3 Na^l Ca in cardiac sarcolemma (e.g., References 29 and 39). The present results with simultaneous Na o Ca o reduction (Figure ) support the conclusion that, in the presence of caffeine, relaxation depends on a simple 3:1 Na-Ca exchange, and the voltage dependence of relaxation (Figure 7 and Figure 8) is also consistent with this interpretation. It may be noted that for Na: Ca stoichiometries of 2 or 4, [Ca] 0 would have to have been reduced to 13 or 13 /JM, respectively, to keep E r constant rather than ~40 /um as in Figure. Also, stoichiometry of 2 Na : i Ca need not be inherently voltage-sensitive (though it could be). Based on these results, depolarization may be expected to decrease the relative contribution of Na-Ca exchange to relaxation and favor SR calcium accumulation. 5 This may explain why rabbit ventricular muscle (with its broad action potential plateau near 0 mv) accumulates SR calcium when stimulation is resumed after a rest 19 whereas rat ventricular muscle (with its very short, plateauless action potential) loses calcium when stimulation is resumed. 3 Indeed, a prominent calcium extrusion is observed in steady-state contractions in rat ventricle. 3 Thus, the action potential plateau in most mammalian ventricular muscle may serve to prevent calcium extrusion and enhance SR recycling. It may be anticipated that these two major mechanisms for removing calcium from the myofilaments will compete with one another. The fractional contributions of the SR calcium pump and Na-Ca exchange will also be expected to vary depending upon conditions and in different species. Ryanodine and Sarcoplasmic Retriculum Calcium Accumulation The results with ryanodine (Figure 4 and Figure 5, Table 1) indicate that this agent does not prevent the SR from taking up calcium. That is, the major part of relaxation is not much slowed by ryanodine even in the absence of extracellular sodium. This would seem at odds with the conclusions from numerous studies that suggest that ryanodine induces an SR calcium leak Qn the contrary, we feel these results are readily reconciled. If ryanodine induces a calcium leak by locking SR calcium release channels in an open subconductance state, 21 this would indeed shift the resting pump-leak balance favoring loss of SR calcium. This explains the acceleration of the rest decay of twitches and RCCs induced by ryanodine However, if [Ca]j is high, as it may be during an RCC (or twitch), the pump rate will be increased and may exceed the leak rate allowing SR calcium uptake. As cytoplasmic [Ca] falls during relaxation, the pump rate would decline and the leak may again exceed the pump rate. At the same time, the [Ca] gradient becomes more in favor of calcium moving from SR to cytoplasm. Thus, when the leak exceeds the pump rate, the SR will be drained of releasable calcium. This may explain the slow final tail of relaxation induced by ryanodine (Figure 5). This also fits very well with the fact that RCCs of up to normal magnitudes can be induced in the presence of ryanodine but only immediately after a contraction. As little as 1-3 seconds after a twitch in ryanodine, RCCs are almost completely eliminated. This also corresponds exactly with results with Ca o microelectrodes where cellular uptake of calcium with stimulation attributable to SR calcium loading is still observed with ryanodine, but the cellular loss of calcium immediately after

9 342 CircoUtton Research Vol 5, No 2, August 1989 contractions is greatly accelerated by ryanodine. 10 This calcium efflux may represent SR calcium leaked to the cytoplasm and subsequently extruded from the ce]].io.i9.2o Thus, in the presence of ryanodine, the SR seems to work as a transient calcium buffer. That is, the SR can accumulate calcium and contribute to rapid relaxation, but then calcium leaks more slowly out of the SR (t^ 1 sec) such that other calcium extrusion mechanisms can nearly keep up with the leak to remove the calcium from the cytoplasm. References 1. Fabiato A: Calcium-induced release of calcium from the cardiac sarcoplasmic reticulum. Am J Physiol 1983; 245:C1-C14 2. Bers DM: Ca influx and SR Ca release in cardiac muscle activation during postrest recovery. Am J Physiol 1985; 248:H3-H Harrison SM, Miller DJ: Mitochondrial contribution to relaxation demonstrated in skinned cardiac muscle of the rat (abstract). / Physiol (Lond) 1984;353:55P 4. Bers DM: Mechanisms contributing to the cardiac inotropic effect of Na-pump inhibition and reduction of extracellular Na. / Gen Physiol 1987;90: Bridge JHB, Spitzer KW, Ershler PR: Relaxation of isolated ventricular cardiomyocytes by a voltage-sensitive process. Science 1988;241: Bers DM, Christensen DM, Nguyen TX: Can Ca entry via Na-Ca exchange directly activate cardiac muscle contraction? JMol Cell Cardiol 1988;20: Barcenas-Ruiz L, Beukelmann DJ, Wier WG. Sodiumcalcium exchange in heart: Membrane currents and changes in [Ca 2+ ],. Science 1987;238: Bridge JHB: Relationships between the sarcoplasmic reticulum and transarcolemmal Ca transport revealed by rapidly cooling rabbit ventricular muscle. J Gen Physiol 198; 88: Kurihara S, Sakai T: Effects of rapid cooling on mechanical and electrical responses in ventricular muscle of guinea pig. J Physiol (Lond) 1985;31: Bers DM, Bridge JHB, MacLeod KT: The mechanism of ryanodine action in cardiac muscle assessed with Ca selective microelectrodes and rapid cooling contractures. Can J Physiol Pharmacol 1987;5: Bcrs DM: SR Ca loading in cardiac muscle preparations based on rapid cooling contractures. Am J Physiol 1989; 25:C109-C Bers DM: Ryanodine and the calcium content of cardiac SR assessed by caffeine and rapid cooling contractures. Am J Physiol 1987;253:C408-C415 Harrison SM, Bers DM: The influence of temperature on the calcium sensitivity of the myofilaments of skinned ventricular muscle from the rabbit. / Gen Physiol 1989;93: Inui M, Saito A, Fleischer S: Isolation of the ryanodine receptor from cardiac sarcoplasmic reticulum and identity with the feet structures. J Biol Chem. 1987;22: Inui M, Saito A, Fleischer S: Purification of the ryanodine receptor and identity with feet structures of junctional terminal cisternae of sarcoplasmic reticulum from fast skeletal muscle. J Biol Chem 1987;22: Lai FA, Enckson H, Block BA, Meissner G: Evidence for a junctional feet-ryanodine receptor complex from sarcoplasmic reticulum. Biochem Biophys Res Common 1987;143:7O4-7O9 17. Lai FA, Anderson K, Rousseau E, Liu QY, Meissner G: Evidence for a Ca 2+ channel within the ryanodine receptor complex from cardiac sarcoplasmic reticulum. Biochem Biophys Res Commun 1988;151: Sutko JL, Ito K, Kenyon, JL: Ryanodine: A modifier of sarcoplasmic reticulum calcium release. Biochemical and functional consequences of its action is striated muscle. Fed Proc 1985;44: Bers DM, MacLeod KT: Cumulative extracellular Ca depletions in rabbit ventricular muscle monitored with Ca selective microelectrodes. Ore Res 198;58: MacLeod KT, Bers DM: The effects of rest duration and ryanodine on extracellular calcium concentration in cardiac muscle from rabbits. Am J Physiol 1987;253:C398-C Rosseau E, Smith JS, Meissner G: Ryanodine modifies conductance and gating behavior of single Ca J+ release channel. Am J Physiol 1987;253:C34-C Fleischer S, Ogunbunmi EM, Dixon MC, Fleer EAM: Localization of Ca 2+ release channels with ryanodine in junctional terminal cisternae of sarcoplasmic reticulum of fast skeletal muscle. Proc NatlAcad Sci USA 1985;82: Meissner G: Ryanodine activation and inhibition of the Ca 2+ release channel of sarcoplasmic reticulum. J Biol Chem 198;21: Lattanzio FA Jr, Schlatterer RG, Nicar M, Campbell KP, Sutko JL: The effects of ryanodine on passive calcium fluxes across sarcoplasmic reticulum membranes. / Biol Chem 1987;22: Jones LR, Besch HR, Sutko JL, Willerson JT: Ryanodineinduced stimulation of net Ca ++ uptake by cardiac sarcoplasmicreticulumvesicles. JPharmacolExp Ther 1979;2O9: Chapman RA: Control of cardiac contractility at the cellular level. Am J Physiol 1983;245:H535-H Fabiato A, Fabiato F: Techniques of skinned cardiac cells and of isolated cardiac fibers with disrupted sarcolemmas with reference to the effects of catecholamines and of caffeine. Recent Adv Stud Card Struct Metab 197;9: Wendt IR, Stephenson DG: Effects of caffeine on Caactivated force production in skinned cardiac and skeletal muscle fibres of the rat. Pftugers Arch 1983;398: Reeves JP, Hale CC: The stoichiometry of the cardiac sodium-calcium exchange system. J Biol Chem 1984; 259: Bers DM, Philipson KD, Nishimoto AY: Sodium-calcium exchange and sidedness of isolated cardiac sarcolemmal vesicles. Biochim Biophys Ada l Q 80;01: Reeves JP, Sutko JL: Sodium-calcium exchange activity generates a current in cardiac membrane vesicles. Science 1980;208: Mullins LJ: The generation of electric currents in cardiac fibers by Na/Ca exchange. Am J Physiol 1979;23:C103-C Bers DM, Bridge JHB: Effects of acetylstrophanthidin on twitches, microscopic tension fluctuations and cooling contractures in rabbit ventricle. JPhysiol (Lond) 1988;4O4: Shattock MJ, Bers DM: The inotropic response to hypothermia and the temperature-dependence of ryanodine action in isolated rabbit and rat ventricular muscle: Implications for E-C coupling. Ore Res 1987;1: Bers DM, Christensen DM: SR Ca-pump and Na/Ca exchange in relaxation of rabbit ventricular muscle (abstract). J Mol Cell Cardiol 1988;20(suppl IV):S41 3. Shattock MJ, Bers DM: Rat vs rabbit ventricle: Ca flux and intracellular Na assessed by ion-selective microelectrodes. Am JPhysiol 1989;25:C813-C Hilgemann DW: Extracellular calcium transients at single excitations in rabbit atrium measured with tetramethylmurexide. / Gen Physiol 198;87: Bielfeld DR, Hadley RW, Vassilev PM, Hume JR: Membrane electrical properties of vesicular Na-Ca exchange inhibitors in single atrial myocytes. Ore Res 198^9: Philipson KD: Sodium-calcium exchange in plasma membrane vesicles. Anna Rev Physiol 1985;47: Hilgemann DW, Delay MJ, Langer GA: Activationdependent cumulative depletions of extracellular free calcium in guinea pig atrium measured with antipyiylazo III and tetramethyimurexirje. Ore Res 1983^3: Sutko JL, Willerson JT: Ryanodine alteration of the contractile state of rat ventricular myocardium. Comparison with dog, cat and rabbit ventricular tissues. Ore Res 1980;4: KEY WORDS relaxation sarcoplasmic reticulum Na-Ca exchange ryanodine

10 Relaxation of rabbit ventricular muscle by Na-Ca exchange and sarcoplasmic reticulum calcium pump. Ryanodine and voltage sensitivity. D M Bers and J H Bridge Circ Res. 1989;5: doi: /01.RES Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 1989 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Circulation Research is online at: