Ca21 Content of Rabbit Ventricular Muscle

Transcription

1 813 Effects of Thapsigargin and Cyclopiazonic Acid on Twitch Force and Sarcoplasmic Reticulum Ca21 Content of Rabbit Ventricular Muscle Stephane Baudet, Rayan Shaoulian, Donald M. Bers Downloaded from by on November 3, 218 Thapsigargin (TG) and cyclopiazonic acid (CPA) are reported to be specific high-affinity inhibitors of the sarcoplasmic reticulum (SR) Ca2' pump in isolated membranes and cells, with TG causing complete pump inhibition at nanomolar concentrations. To evaluate the effectiveness of TG and CPA in small multicellular cardiac preparations, we used rapid cooling contractures (RCCs) to assess the SR Ca2' load. In contrast to observations in single myocytes, TG caused remarkably slow and incomplete SR Ca2' depletion in multicellular preparations. A 45-minute exposure to 5 JAM TG at 3 C and.5-hz stimulation only decreased RCCs by 76±5% (and 1,uM CPA reduced RCCs by 59±1% [mean± SEMI). In contrast, 1 minutes with 2 mm caffeine completely abolished RCCs. This confirms that there was still a caffeine-sensitive pool of Ca21 in the TG-treated muscle. The time constant of rest decay of RCCs was accelerated by both TG (from 83± 18 to 26±6 seconds) and CPA (from 68± 11 to 1±5 seconds). This might be expected since Ca21 leaking from the SR during rest cannot be taken back up as efficiently, favoring Ca2+extrusion by the sarcolemmal Na+-Ca2' exchanger. TG and CPA decreased twitch force (by 44±7% and 4±11%, respectively) and increased twitch duration, presumably because of the SR effects. We conclude that complete blockade of SR Ca2' uptake by TG or CPA in multicellular preparations cannot be assumed, even at high [TG] or [CPA], unless evaluated (eg, by RCC). (Circ Res. 1993;73: ) KEY WoRDs * sarcoplasmic reticulum * Ca2+-ATPase * excitation-contraction coupling rapid cooling contractures V entricular muscle contraction is activated by Ca2' influx through the sarcolemma, which subsequently releases Ca2+ from the sarcoplasmic reticulum (SR) via the Ca'+-induced Ca2' release mechanism."2 Relaxation is brought about by Ca2' dissociation from the contractile proteins, extrusion of that Ca2' by the sarcolemmal Na+-Ca2' exchanger, and reuptake by the SR Ca2+-ATPase.2-4 The SR Ca2 load is an important factor in control of cardiac muscle contraction and is itself determined by a balance between Ca2' efflux through the SR Ca2' release channel and Ca2 uptake via the SR Ca2+-ATPase.2,5,6 Until recently, caffeine and ryanodine have been the main agents used to inhibit SR function in intact cardiac cells. Although this approach has been helpful in showing that contraction could be supported to various extents by sarcolemmal Ca2 influx in different animal species,7 these agents activate SR Ca2' release rather than inhibiting the SR Ca ` pump. This could be a particular problem when using ryanodine, since the SR can still accumulate Ca2, albeit transiently.8 An additional impact that SR Ca2 uptake inhibition would have on contractile force is to alter the Ca2+-buffering Received December 14, 1992; accepted June 7, From the Department of Physiology, Loyola University Medical School, Maywood, Ill. Correspondence to Department of Physiology, Loyola University Medical School, 216 South First Ave, Maywood, IL 6153 (Dr Bers). role of the SR. In fact, the SR Ca'+ pump can limit the ability of Ca2' influx to reach the myofilaments.9 An agent that inhibits the SR Ca2' pump should block this SR action. Recently, two chemically unrelated compounds, thapsigargin (TG) and cyclopiazonic acid (CPA), have been used as specific inhibitors of SR/endoplasmic reticulum (ER) Ca2'-ATPase. TG or CPA can induce Ca2' loss from the SR/ER. Indeed, in cell types that don't have powerful Ca'+ extrusion mechanisms, these agents cause [Ca'+]i to rise.1,1" In cardiac and skeletal muscle SR vesicles, TG and CPA inhibit SR Ca2+ uptake at nanomolar concentrations. In isolated cardiac myocytes, TG decreases cell shortening by progressively depleting the SR of its Ca'+ but does not appear to affect SR Ca2 release (when the SR contains Ca21) or the L-type Ca2 current at the concentrations used.12-'4 Therefore, these two compounds could be valuable tools in the study of the impact of SR Ca 2+ pump inhibition on contractile force and relaxation in multicellular cardiac preparations. However, no studies have tested the ability of TG or CPA to prevent SR Ca21 uptake in such multicellular cardiac preparations. It is difficult to directly assess the SR Ca'+ content in intact functioning cardiac muscle. Rapid cooling contractures (RCCs) have been found to provide a valuable index of SR Ca21 content in mammalian cardiac muscle.25,6,8"15-'9 Rapidly cooling the cardiac muscle to C appears to cause release of all of the SR Ca21 while simultaneously inhibiting the Ca2+ transport systems

2 814 Circulation Research Vol 73, No 5 November 1993 Downloaded from by on November 3, 218 (eg, SR Ca2' pump and Na+-Ca2' exchange). The result is a slowly developing contracture (at C), where the amplitude is indicative of the SR Ca2' content at the moment of cooling. Although caffeine can also be used to release SR Ca2, it is much less useful in multicellular preparations than in isolated cells because caffeine diffuses slowly into the muscle (and doesn't slow Na+- Ca2' exchange). Thus, the caffeine-induced contractures in the many cells in a muscle are asynchronous, with cells at the outside contracting and relaxing (because of Na+-Ca2+ exchange) before cells in the core are activated. The aims of the present study were to determine, in multicellular preparations, the effects of TG and CPA (1) on SR Ca2' accumulation using RCCs as an indicator of SR Ca2' load, (2) on steady-state twitch parameters, and (3) on the balance of Ca2+ fluxes between the Na+-Ca2' exchanger and the SR Ca2+ pump during rest. Materials and Methods Muscle Preparations The protocol for muscle dissection, mounting, and RCCs has been previously described.5 In brief, thin papillary muscles or ventricular trabeculae (.1 to.7 mm in diameter) were obtained from the hearts of New Zealand White rabbits previously anesthetized with pentobarbital sodium (4 mg/kg IV). Each muscle was tied at both ends with a fine nylon thread; one end was fixed, and one end was attached to a force transducer (model AE 875, SensoNor, Horten, Norway; or Kulite Semiconductor, Leonia, NJ). Muscle shape was assumed to be cylindrical, and diameter was measured by means of a micrometer in the dissecting microscope. Force was expressed in millinewtons per square millimeter. The solutions fed the chambers by gravity (flow rate, -25 ml/min) and were maintained at the desired temperature by water jacketing at 3 C or by jacketing with a mixture of polypropylene glycol and water (1:3) at -4 C. Solenoid pinch valves, at the bath inlet, were gated by a stimulator, allowing synchronized solution changes. At this flow rate, switching to the cold solution cools the surface of the muscle to below 5 C in less than 2 milliseconds, and rewarming to 3 C is similarly rapid. Transducer signals were digitized at 5 Hz using Axotape software (Axon Instruments, Foster City, Calif). Solutions Muscles were superfused with a modified Tyrode's (NT) solution containing (mm) NaCI, 14; KCI, 6; MgC12, 1; CaCl2, 2; glucose, 1; and HEPES, 5 (ph 7.4 at 3 C). Solutions were gassed with 1% 2. Stock solutions of TG (5 mm) and CPA (5 or 3 mm) were prepared in dimethyl sulfoxide and kept in a freezer until use. Our experience has shown that some precautions must be taken with TG. First, according to Calbiochem Corp, La Jolla, Calif, this molecule is light sensitive, so the experiments were performed in the dark. Second, TG may be sensitive to oxidation, so that bubbling the TG-containing solutions with 2 was avoided. Lack of 2 bubbling by itself (without TG) did not alter RCC amplitude in control experiments. Finally, TG may adhere to the tubing.2 Therefore, after every experiment, all the perfusion lines were washed problems do not seem to exist with CPA, but the same precautions for washing the lines were used. Experimental Protocol After equilibration (~'9 minutes), at a stimulation frequency of.5 Hz, steady-state twitch tension was recorded, and an RCC was elicited in place of a twitch, to estimate relative SR Ca2' content under control conditions. To conserve TG and CPA, and also to limit tubing exposure, TG or CPA at the final concentrations in NT solutions was added directly in the bath via a Pasteur pipette. Then the bath was flushed with excess solution with the same drug concentration every 5 minutes along with the bath flow and drain stopped. During these no-flow conditions, stimulation was stopped. Thus, we have been less able to follow the time course of action of the drugs. Drug exposure was for 45 minutes, and then the flow of drug-free Tyrode's solution and recirculation were restored. This was possible because, at the high drug concentrations used (5 tm TG and 1,uM CPA), their effects on twitch force and RCC amplitude were irreversible. That is, even though TG did not abolish SR Ca2` uptake, the block was stable, with no recovery for more than 3 hours after TG washout. Pilot experiments (only NT applied with flow stopped for 45 minutes) and vehicle control experiments (up to 1% dimethyl sulfoxide) have shown that 2 minutes was usually necessary for the muscle to recover. Often the twitch amplitude was slightly depressed (by less than 1%), but more importantly, the SR Ca21 content, assessed by the RCC amplitude, was unchanged, allowing us to test the effects of TG and CPA. Finally, we tested the ability of caffeine (2 mm) and ryanodine (1,uM) to inhibit the RCC in TG- or CPA-treated muscles. The TG- or CPA-treated preparations were exposed to caffeine or ryanodine until twitch force reached a new steady state (, 1 to 2 minutes). Because the effects of ryanodine were irreversible, these effects were tested only after treating the muscle with either of the compounds. Chemicals and Statistics All salts, caffeine, and CPA were obtained from Sigma Chemical Co, St Louis, Mo. Caffeine was dissolved directly in NT. TG was obtained from Calbiochem. Ryanodine (stock solution of 1 mm in distilled water) was obtained from Agri Systems Corp, Wind Gap, Pa, and from Calbiochem. CPA and TG stock solutions were kept in a freezer. Results are expressed as mean±sem. Paired and unpaired Student's t tests were used when appropriate. A value P<.5 was considered statistically significant. Results The value of TG and CPA in physiological experiments is that they can completely abolish SR Ca2' uptake. Our first experiments were carried out to examine the conditions required to achieve complete SR Ca2' pump inhibition. If the SR Ca2' pump is completely blocked, then in the steady state the SR should be depleted of Ca2`. We have therefore used RCCs to assess the SR Ca2' content. Although we did not perform systematic dose-response curves for both compounds, we found that RCC ampliwith alcohol and several liters of deionized water. Such tude was not strongly affected by concentrations up to 2

3 Baudet et al Thapsigargin and Cyclopiazonic Acid in Cardiac Muscle 815 Downloaded from by on November 3, 218,uM for CPA and TG. Indeed, at 1 and 5,uM TG, RCC amplitude was only decreased by 29+7% (n=6) and 57±8% (n=6), respectively. These observations were rather surprising, particularly for TG, because of the high affinity of the drug in isolated membrane systems and its capacity to inhibit the intracellular Ca21 transient and contraction in isolated cardiac cells at concentrations less than 5,gM In isolated rabbit cells, the caffeineinduced contracture (a test for SR Ca21 content) is consistently abolished by 5,M TG.12 In an effort to abolish SR Ca21 uptake completely, TG and CPA concentrations have been increased to 5 and 1,uM, respectively. Even at 5,M TG, the steady-state RCC was only partially inhibited (by 76 ±6%; range, 48% to 1%). This surprising result could have been related to the size of the muscle (eg, if there were diffusion limitations, oxidation, or partitioning of TG). However, when the percent depression of the steady-state RCC was plotted against muscle diameter ( mm for the TG group and.38±.6 mm for the CPA group) or cross-sectional area, no correlation was found (P>.5, not shown). Another possibility was that the total amount of drug present in the bath was insufficient to saturate all the SR Ca2+ pump binding sites. A rough calculation shows that it is not likely to be the case. In fact, if the number of SR Ca21 pumps is approximately 6 gumol/kg wet wt21 and if the muscle has a mass of 2 mg in the worst case [ie, r- r' 1 length * density= v (.35 cm)2. (.5 cm). (1.6 g/cm3)], the total number of binding sites is 12 pmol. In.1 ml of a 5,uM drug-containing solution, there are 5 pmol TG in the bath, which was refreshed every 5 minutes. Thus, TG outnumbers the potential binding sites by more than three orders of magnitude. Characteristics of the Steady-State RCC Fig 1 illustrates the effects of TG on the steady-state RCC in two different muscles (similar qualitative results were obtained with CPA). In the control situation (Fig 1, A and D), cold NT was applied until RCC reached a steady level. After the muscle had been treated with 5,uM TG (Fig 1, B and E), the steady RCC amplitude was depressed by 64% and 41%, respectively. The sharp increase in force upon rewarming is attributed to the rapid increase in myofilament Ca21 sensitivity with increasing temperature.16 In some muscles, the rewarming spike persisted, despite the abolition of the RCC in CPA- or TG-treated muscles (not shown). This indicates that there was still Ca21 left in the SR but not enough to activate the contractile proteins during the cold (since cold decreases myofilament Ca 2+ sensitivity16). Overall, RCCs were depressed by 76+6% (range, 48% to 1%; n=9) by 5,uM TG and 59+1% (range, 36% to 1%; n=7; P>.5) by 1,uM CPA. This variability was not due to variability in muscle diameter (above) or an incapacity of the RCC to be completely abolished, because depletion of SR Ca21 using either caffeine (2 mm) or ryanodine (1 I M) completely abolished both the RCC and the rewarming spike (Fig 1, C and F, respectively). This confirms that the RCC in the presence of TG or CPA represents residual SR Ca2. When 2 mm caffeine was first applied (between Fig 1, B and C) there was usually a transient increase in diastolic force. This was probably A )/ B C Control TG-treated TG + caffeine t D IfLJ E F TG-treated TG + ryanodine 5 mn mm 5 mn.mm 1 s 1 s FIG 1. Tracings show the effect of 5,uM thapsigargin (TG) and subsequent addition of caffeine or ryanodine on the steady-state rapid cooling contracture (RCC, elicited 2 seconds after a twitch). In all tracings, the downward arrow indicates switching from 3 to C modified Tyrode's solution, and the upward arrow indicates the reverse procedure. A, The control steady-state RCC was 16 mn. mm-2, and the rewarming time to half-relaxation (t112) was 58 milliseconds. B, RCC elicited in a muscle exposed to 5,uM TG for 45 minutes decreased the RCC by 64%, and t1/2 was 42 milliseconds. C, Abolition of the RCC in the same TG-treated muscle after a 1-minute exposure to 2 mm caffeine is shown. D and E, In this muscle, control RCC amplitude was 17.6 mn * mm-2, and t112 was 416 millisconds. TG treatment depressed the RCC by 41% and increased t1/2 to 5 milliseconds. F, Ryanodine (1 um) applied for 1 minutes abolished the steady-state RCC. The slight increase in tension during rewarming is a mechanical artifact. For panels A through C, muscle 9271, the diameter was.6 mm; for panels D through F, muscle 92727, the diameter was.22 mm. due to release (and extrusion) of residual SR Ca2' in the presence of TG upon the application of caffeine. Rewarming Spike Relaxation In previous studies, the relaxation phase of the rewarming spike has been considered to reflect the rate of decline of [Ca2+]i due to SR Ca2' uptake and Ca2' efflux through the Na+-Ca2' exchanger.6"17 Since the SR Ca2' pump is dominant (by a factor of threefold to fourfold), we expect TG and CPA to slow relaxation of the rewarming spike. Since TG and CPA reduce RCC amplitude, it is important for this comparison that changes in the time to half-relaxation (tl/2) of the rewarming spike are independent of the amplitude of the RCC. This appears to be the case (not illustrated; see also Bers and Bridge6). In CPA-treated muscles, the t1/2 of relaxation was consistently slowed, from 355+±31 to 521±48 milliseconds (P<.2). In TG-treated muscles, the tl,2 of relax-

4 Downloaded from by on November 3, Circulation Research Vol 73, No 5 November It z 1. A /\ Control 1 \ C 22mN mm -2ri 5 ms Time (ms) B / \ Control 1/ \ CPA D / 'Ii Time ((s) FIG 2. Representative tracings of the effects of 5,uM thapsigargin (TG) and 1,uM cyclopiazonic acid (CPA) on steady-state twitch force and time course. Panels C and D are the twitches in panels A and B normalized to their respective peak tension. A, A 45-minute exposure to TG depressed twitch force by 48% (from 11.6 mn mm-2) and the steady-state rapid cooling contracture (RCC) by 62% (not shown). C, The normalization procedure emphasizes the prolongation of the twitch time course. Time to peak force (TTP) increased from 296 to 36 milliseconds, but the time to half-relaxation (t112) was not significantly affected: 238 milliseconds for control and 222 milliseconds after TG. B and D, CPA decreased the steady-state twitch by 44% (from 14.4 mn mm-2) and RCC by 7%. CPA also prolonged the twitch time course. TTP increased from 286 to 466 milliseconds, and t1/2 increased from 222 to 318 milliseconds. For panels A and C, muscle 9271, the diameter was.22 mm; for panels B and D, muscle 92729, the diameter was.44 mm. ation was not significantly changed (control: tl12, 456±4 milliseconds; TG: tl12, 38+±35 milliseconds; P>.5). Therefore, although the RCC amplitude was depressed by both TG and CPA, CPA appears to have the more expected effect on the relaxation phase. In some muscles, TG completely abolished the RCC. To examine whether the SR Ca2' pump was completely inhibited, three such preparations were exposed to 8 mm Ca2'-containing medium. Twitch force increased, and RCC amplitude increased to :e--5% of the control value. These results indicate that, even when the SR Ca 2+ pump appeared to be completely blocked by TG, Ca2' could still be accumulated by the SR when more Ca>2 was supplied. Effects of TG and CPA on the Steady-State Twitch Fig 2, A and C, are from a representative experiment illustrating the effects of 5,uM TG on the twitch., When the peaks were normalized (Fig 2, C), it is apparent that the time to peak force (TTP) was increased but that t112 was not appreciably changed. Overall, steady-state twitch force was decreased by 44±7%, from 19.5±2.2 mn - mm-2 (n=1). TTP was significantly increased by TG (from 284±11 to 362±17 milliseconds, P<.1), but tl2 was unaffected (237±13 milliseconds in the control condition and 238± 12 milliseconds after TG). Fig 2, B and D, illustrate representative effects of 1,uM CPA on the twitch. CPA clearly prolongs the twitch time course by increasing both TTP (by 63%) and tl/2 (by 43%). On the average, TTP increased significantly (from 295±12 to 492±21 milliseconds, P<.1) as did t1/2) (from 22±9 to 354±15 milliseconds, P<.1). Twitch force was also decreased by 4±11%. Effects of TG on Rest Decay of Twitches and RCCs In rabbit ventricular muscle, the amplitude of the first twitch or RCC decreases after increasing periods of rest (a process known as rest decay).7,22 During rest, Ca2 leaks from the SR and is either taken up by the SR or extruded through the Na+-Ca2' exchanger. Ca2+ reaccumulated by the SR does not contribute to rest decay. The ability of the Na+-Ca2' exchange to compete with SR Ca2+ reuptake determines the rate of rest decay of RCCs in rabbit ventricle.19 Therefore, if TG and CPA were effectively inhibiting Ca2+ uptake, rest decay of RCCs should be accelerated, because the balance of Ca2+ fluxes will be even more in favor of the efflux. Although rest decay of twitches is also observed, it is complicated by the contribution that transsarcolemmal Ca2+ fluxes have during the postrest twitch (eg, via the L-type Ca2+ current). Fig 3, A, illustrates the effect of TG on the amplitude of the first postrest contraction (ie, the time course of rest decay) normalized to the control twitch. In control conditions, the rest decay was typical for rabbit ventricular muscle. After TG treatment, the steady-state twitch was depressed by 47%, and rest decay was accelerated. The time constant for this rest decay (r) decreased from 72±11 to 13±1 seconds after TG. To confirm that this was actually due to a faster decline in SR Ca 2+ content, RCCs were elicited at steady state and after different rest periods (Fig 3, B). TG treatment decreased the steady-state RCC by 76% and also accelerated rest decay of RCCs (r decreased from 83±18 to 26±4 seconds). Results obtained with CPA were not qualitatively different from those obtained with TG, and a typical example is shown in Fig 4. CPA accelerated rest decay of both twitches and RCCs. On the average (n=7), CPA decreased r from 3 to 13 seconds for twitches and from 68±11 to 1±5 seconds for RCCs. In Fig 4, ryanodine was also added after equilibration and testing with CPA alone. Ryanodine accelerated the rest decay even more, as may have been expected (r=5 seconds). However, the twitch after a 5-minute rest was not further reduced by ryanodine. Thus, this "rested state" twitch (with CPA alone) may reflect activation by Ca2' influx alone. Discussion Our results show that TG and CPA both decreased twitch force by decreasing the amount of Ca 2+ inside the SR (assessed by RCCs), presumably due to partial

5 Baudet et al Thapsigargin and Cyclopiazonic Acid in Cardiac Muscle 817 A ` c) a Lt C D 2 o~~~~ Control (n= 1) *~~~~~ TG-treated (n=7) Q\ c) q 1 cd a) o 6 O Control * CPA-treated V CPA + ryanodine Downloaded from by on November 3, 218 U c) -., ' FIG 3. Graphs showing the effects of 5,uM thapsigargin (TG) on rest decay of twitches (A) and rapid cooling contracture (RCC) (B). In TG-treated preparations, the amplitudes of both twitches and RCCs have been normalized to their respective steady-state control values. Mean data points were fit using a single monoexponential decay [A. exp(-tlr)+b, where r is the time constant of rest decay]. A, After TG application, r decreased from 71 to 1 1 seconds; A, from 67% to 41%; and B, from 33% to 19%. B, The rest decay of RCCs was also accelerated (r decreased from 6 to 26 seconds; A, from 69% to 3%; and B, from 35% to %). It is possible that the remaining 19% of the rested-state twitch (B) is attributable to Ca2+ influx. inhibition of SR Ca2, uptake. However, despite high concentrations of TG or CPA, the inhibition of the SR Ca21 pump was incomplete. CPA appears to have the expected effects on twitch time course (especially a slowing of relaxation). Whereas TG slowed the overall contraction, it had surprisingly little effect on the t1/2 of relaxation. Finally, acceleration of rest decay of twitches and RCCs shows that, when SR Ca2' uptake is inhibited, Ca21 fluxes during rest are shifted toward Ca2+ extrusion by the sarcolemmal Na+-Ca2' exchanger. Incomplete Inhibition of SR Ca 2+ Uptake: Possible Mechanisms Both TG and CPA inhibit the SR Ca2+-ATPase, although by different mechanisms.23'24 The most surprising observation of this study has been the need to apply high concentrations of TG and CPA to observe a significant, but still incomplete, decrease of SR Ca2' content (assessed by RCCs). In fact, in single rabbit ventricular myocytes in the same lab, SR Ca2+ uptake is completely blocked, and the SR can be depleted of its C 1 Control B W9 o * CPA-treated V CPA + ryanodine 8 o r ao i, 1 O FIG 4. Graphs showing representative effects of 1,uM cyclopiazonic acid (CPA) and subsequent ryanodine (1,M) application on rest decay of twitches (A) and rapid cooling contractures (RCCs) (B). Data were normalized and fit as in Fig 3. A, CPA accelerated twitch force rest decay: r decreased from 94 to 5 seconds; A slightly increased from 67% to 75%; and B decreased from 31% to 15%. Subsequent ryanodine application drastically accelerated rest decay as follows: r, 5 seconds; A, 45%; and B, 13%. B, Rest decay of RCCs mirrored twitch rest decay as follows in control conditions:, 155 seconds; A, 99%; and B, %. After CPA application, RCC decay was accelerated as follows:, 36 seconds; A, 72%; and B, %. Ryanodine completely abolished RCC at steady-state and after rest in this CPA-treated preparation. Ca2` by a 9-second exposure to 5,M TG.12,25 We have also found that RCCs in single rabbit ventricular myocytes (and accompanying intracellular Ca2' transients) can be completely abolished by a 2-minute exposure to 5,M TG (J.W.M. Bassani and D.M. Bers, unpublished observations). This concentration of TG had almost no effect on twitch force and RCC amplitude in our preliminary experiments with multicellular preparations even after 1 hour of equilibration. This emphasizes the difference between the ability of TG to inhibit the SR Ca 2+ pump in isolated myocytes vs multicellular preparations. It also shows that the lack of complete SR Ca2+ pump inhibition in multicellular preparations is due to the preparation and not the method used to assess SR Ca2+ (ie, RCCs). As shown by our simple calculation, the number of potential binding sites is outnumbered by the amount of TG and CPA present in the bath by at least three orders of magnitude. Thus, there is plenty of ligand available. The possibility of high nonspecific binding is unlikely,

6 818 Circulation Research Vol 73, No 5 November 1993 Downloaded from by on November 3, 218 because titration experiments with permeabilized myocytes matched the expected high number of pump sites.25 Although there are other nonmyocyte binding sites (endothelial, neuronal, and smooth muscle cells), they would not be expected to exceed the concentration in myocytes and would be included in the number of sites expected in homogenates.21 Another possibility, linked to the observation of incomplete inhibition of SR Ca' uptake, is that the actual concentration of TG or CPA reaching its binding site is much lower than in the bath. For this to occur, TG or CPA could be degraded (or oxidized), creating a concentration gradient, which would be worsened by elevated muscle diameter. However, we did not find any relation between RCC inhibition and muscle diameter. Therefore, we have no satisfactory explanation for the high drug concentrations needed to inhibit the SR Ca>2 pump. Whatever the reason for the incomplete SR Ca' uptake depression, it is important to assess SR Ca>2 content directly in multicellular preparations before assuming that SR Ca2 uptake is completely abolished. Another important point in this characterization was to know whether the residual RCC observed with TG and CPA could be prevented by compounds known to fully inhibit the RCC in control conditions. The SR Ca>2 can also be depleted by caffeine and ryanodine, which keep the SR Ca 2> release channels opened, thereby preventing SR Ca>2 accumulation.8 Caffeine and ryanodine both abolished the RCC in the absence or presence of TG or CPA, confirming that the SR Ca>2 released by RCC after TG or CPA treatment was still caffeine and ryanodine sensitive. Effects of TG and CPA on the Steady-State Twitch Amplitude It is likely that the decrease in twitch force (by -45% with both compounds) was primarily due to a decrease in the SR Ca> content (ie, RCCs were reduced by 76% and 59% for TG and CPA, respectively). However, ryanodine, which completely abolishes RCCs, only decreases twitches in rabbit ventricle by 2% to 3% (with Ca> influx supporting the contraction).2 Since the SR can still transiently accumulate Ca2+ entering during a twitch with ryanodine,8 it is somewhat surprising that TG and CPA were more effective at inhibiting twitches (even without fully inhibiting the SR Ca21 pump). Thus, at these high concentrations, TG and/or CPA could have additional effects that decrease twitch force independent of their effects on the SR Ca>2 pump. One possibility is that myofibrillar Ca>2 sensitivity is decreased. Although Ca 2+ responsiveness was unaffected by ~1,uM TG,13 nothing is known of the effect at higher concentrations. CPA has been shown to increase myofibrillar Ca21 sensitivity of cardiac muscle fibers.26 Another possibility that could explain the discrepancies is if the L-type Ca> current were depressed. At low TG concentrations, the amplitude of the L-type Ca2+ current is unchanged.'3 Nothing is known of the effect at higher TG concentrations, but 1,uM CPA decreases Ca2+ current in rat ventricular cells (W. Balke, personal communication). Thus, nonspecific effects of high TG and CPA concentrations could complicate their use in multicellular preparations. On the other hand, the comparison with ryanodine effects may not be entirely duration and this could help to minimize the negative inotropic effect of this agent. Effects of TG and CPA on the Time Course of the Isometric Twitch CPA increased the time to peak twitch tension and t1l2 of relaxation of twitches and RCCs, consistent with partial inhibition of the SR Ca 2> pump.27 TG also increased TTP but did not slow relaxation. The absence of relaxation slowing with TG may indicate that other mechanisms compensate for the partial block of the SR Ca>2 pump at these high TG concentrations. Possibilities could include TG stimulation of Ca2' efflux by the Na+-Ca2' exchanger or an increased rate of Ca 2 dissociation from the myofilaments (eg, a decreased myofilament Ca'2 sensitivity). However, the fact that the contractile characteristics were stable for 3 hours after TG washout would make complicating lower-affinity effects at high TG concentrations seem less probable (ie, low-affinity effects would be expected to be more rapidly reversible). Effects of TG and CPA on Ca 2± Fluxes During Rest In rabbit ventricular muscle during rest, Ca>2 leaks from the SR, probably through the SR Ca2' channel.28 Part of this Ca21 is extruded by the Na+-Ca2+ exchanger (which leads to rest decay by decreasing SR Ca>2 content), and most of the rest is pumped back inside the SR (which, by contrast, slows rest decay). Therefore, the amount of Ca 2> inside the SR reflects the balance between these two transporters (sarcolemmal Na+-Ca 2 exchanger and SR Ca21 pump).6,19 According to this scheme, any drug that inhibits SR Ca2± reuptake or increases SR Ca>2 release during rest should accelerate rest decay, because the probability of the Ca>2 leaking from the SR being extruded by the Nat-Ca> exchanger is higher. Ryanodine, which does not inhibit the SR Ca21 pump, is known to greatly accelerate rest decay, simply because the Ca>2 taken up almost immediately leaks from the SR to be extruded. In this case, much more Ca21 leaks from the SR per unit time and is available for extrusion by Na+-Ca2' exchange. Our results show that TG and CPA have the expected accelerating effect on rest decay of twitch force and RCC, as indicated by the smaller time constant. Conclusions Our results have shown that both TG and CPA decreased the amplitude of twitches and RCCs (and hence, SR Ca2+ content). However, SR Ca2+ uptake depression by TG and CPA is incomplete, even at very high concentrations and long exposure times, which contrasts with more isolated preparations like single cells or SR vesicles. The SR Ca 2+ pump inhibition can probably explain the observed decrease of twitch force, but these agents might have other complicating effects at the high concentrations required. Both agents appear to shift the balance of Ca2+ fluxes in favor of Ca2+ extrusion by the sarcolemmal Na+-Ca2+ exchanger, thereby accelerating the rest decay of twitches and RCCs. Finally, CPA appears to have the more expected effect in that it significantly slows relaxation. In any event, the utility of TG and CPA in multicellular fair, because ryanodine can prolong action potential cardiac preparations may be seriously limited by incom-

7 Baudet et al Thapsigargin and Cyclopiazonic Acid in Cardiac Muscle 819 Downloaded from by on November 3, 218 plete effectiveness, particularly if no test for SR Ca2+ load is performed. Acknowledgments This study was supported by National Institutes of Health Grant HL-377. Dr Baudet and Mr Shaoulian were recipients of a Postdoctoral Fellowship and a Student Research Fellowship, respectively, from the American Heart Association, California Affiliate, Inc. References 1. Fabiato A. Calcium-induced release of calcium from the cardiac sarcoplasmic reticulum. Am J Physiol. 1983;245:C1-C Bers DM. Excitation-Contraction Coupling and Cardiac Contractile Force. Amsterdam, The Netherlands: Kluwer Academic Press; 1991: Wier WG. Cytoplasmic calcium in mammalian ventricle: dynamic control by cellular processes. Annu Rev Physiol. 199;52: Bassani RA, Bassani JWM, Bers DM. Mitochondrial and sarcolemmal Ca transport can reduce [Ca]i during caffeine contractures in rabbit cardiac myocytes. J Physiol (Lond). 1992;453: Bers DM. SR Ca loading in cardiac muscle preparations based on rapid cooling contractures. Am J Physiol. 1989;256:C19-C Bers DM, Bridge JHB. Relaxation of rabbit ventricular muscle by Na-Ca exchange and sarcoplasmic reticulum Ca-pump: ryanodine and voltage sensitivity. Circ Res. 1989;65: Bers DM. Ca influx and SR Ca release in cardiac muscle activation during postrest recovery. Am J Physiol. 1985;248:H366-H 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;65: Lipp P, Pott L, Callewaert G, Carmeliet E. Calcium transients caused by calcium entry are influenced by the sarcoplasmic reticulum in guinea-pig atrial myocytes. JPhysiol (Lond). 1992;454: Mason MJ, Garcia-Rodriguez C, Grinstein S. Coupling between intracellular Ca2+ stores and the Ca2+ permeability of the plasma membrane: comparison of the effects of thapsigargin, 2,5-di(tertbutyl)-1,4-hydroxyquinone, and cyclopiazonic acid in rat thymic lymphocytes. J Biol Chem. 1991;266: Thastrup, Cullen PJ, Drobak BK, Hanley MR, Dawson AP. Thapsigargin, a tumor promoter, discharges intracellular Ca stores by specific inhibition of the endoplasmic reticulum Ca-ATPase. Proc Natl Acad Sci U S A. 199;87: Bassani JWM, Bassani RA, Bers DM. Twitch-dependent SR Ca accumulation and release in rabbit ventricular myocytes. Am J Physiol. In press. 13. Kirby MS, Sagara Y, Gaa S, Inesi G, Lederer WJ, Rogers TB. Thapsigargin inhibits contraction and Ca transient in cardiac cells by specific inhibition of the sarcoplasmic reticulum Ca pump. J Biol Chem. 1992;267: Wrzosek A, Schneider H, Grueninger S, Chiesi M. Effect of thapsigargin on cardiac muscle cells. Cell Calcium. 1992;13: Bridge JHB. Relationships between the sarcoplasmic reticulum and transarcolemmal Ca transport revealed by rapidly cooling rabbit ventricular muscle. I Gen Physiol. 1986;88: Harrison SM, Bers DM. The influence of temperature on the calcium sensitivity of the myofilaments of skinned ventricular muscle from the rabbit. J Gen Physiol. 1989;93: Bers DM, Bridge JHB, Spitzer KW. Intracellular Ca transients during rapid cooling contractures in guinea-pig ventricular myocytes. JPhysiol (Lond). 1989;417: Bouchard RA, Bose D. Contribution of sarcolemmal sodiumcalcium exchange and intracellular calcium release to force development in isolated canine ventricular muscle. J Gen Physiol. 1992; 99: Bers DM, Christensen DM. Functional interconversion of rest decay and ryanodine effects in rabbit or rat ventricle depends on Na/Ca exchange. J Mol Cell Cardiol. 199;22: Sturek M, Stehno-Bittel L, Obye PK. Methods for simultaneous voltage-clamp, microfluorometry, and video of cells, II: physiology. In: Sperelakis N, Kuriyama H, eds. Electrophysiology and Ion Channels of Vascular Smooth Muscle Cells and Endothelial Cells. New York, NY: Elsevier Science Publishing Co, Inc; 1991: Levitsky DO, Benevolensky DS, Levchenko TS, Smirnov VN, Chazov El. Calcium-binding rate and capacity of cardiac sarcoplasmic reticulum. J Mol Cell Cardiol. 1981;13: Sutko JL, Bers DM, Reeves JP. Postrest inotropy in rabbit ventricle: Na+-Ca 2+ exchange determines sarcoplasmic reticulum Ca2 content. Am J Physiol. 1986;25:H654-H Sagara Y, Inesi G. Inhibition of the sarcoplasmic Ca transport ATPase by thapsigargin at subnanomolar concentrations. J Biol Chem. 1991;266: Seidler NW, Jona I, Vegh M, Martonosi A. Cyclopiazonic acid is a specific inhibitor of the Ca2-ATPase of sarcoplasmic reticulum. JBiol Chem. 1989;264: Hove-Madsen L, Bers DM. SR Ca uptake and thapsigargin sensitivity in permeabilized rabbit and rat ventricular myocytes. Biophys J. 1993;64: A12. Abstract. 26. Lahouratate P, Quiniou M-J, Leoty C. Cyclopiazonic acid. In: Frank GB et al, eds. Excitation-Contraction Coupling in Cardiac, Skeletal and Smooth Muscle. New York, NY: Plenum Publishing Corp; 1992: Balke CW, Wier WG. Selective and reversible inhibition of the Ca 2`-ATPase pump of the cardiac sarcoplasmic reticulum. Biophys J. 1992;61:A434. Abstract.