INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

CALCIUM (Ca²⁺)-sensitizing drugs increase the force of myocardial contraction by increasing the responsiveness of myofilaments to Ca²⁺ rather than by increasing the amount of Ca²⁺ released to activate myofilaments (15). Many Ca²⁺ sensitizers also possess some degree of phosphodiesterase (PDE) III-inhibiting activity. For the thiadiazinone derivative EMD-53998 {5-[1-(3,4-dimethoxybenzoyl)-1,2,3,4-tetrahydrochinolin-6-yl]-6-methyl-3,6-dihydro-2H-1,3,4-thiadiazin-2-one}, these two actions are enantiospecific, with the positive enantiomer, EMD-57033, acting predominately as a Ca²⁺ sensitizer and the negative enantiomer, EMD-57439, acting predominately as a PDE III inhibitor (5,16, 23). The isolation of very specific Ca²⁺-sensitizing action, as with EMD-57033, makes it important to understand the effects of these drugs on ventricular relaxation; although Ca²⁺-sensitizing compounds have generated considerable enthusiasm, there are concerns about their ultimate usefulness because a specific Ca²⁺-sensitizing effect might also prolong relaxation (3, 15, 22) or increase diastolic tone (5, 7, 19).

We have shown previously (21) that, in addition to slowing relaxation by increasing left ventricular (LV) wall stress, 2 × 10⁶ M EMD-57033 also has a stress-independent effect to slow relaxation in the isolated rabbit heart (buffer perfused at 30°C). Hgashiyama et al. (11) also showed that EMD-57033 significantly prolonged relaxation in the isolated rabbit heart (blood perfused at 37°C). Furthermore, they noted a trend toward a stronger inotropic effect in ejecting versus isovolumic beats. However, they did not evaluate whether ejection influenced the effect of EMD-57033 on relaxation. Accordingly, the first purpose of our study was to develop a broader analysis of stress-dependent relaxation, which included ejecting beats, to evaluate whether the effects of positive inotropic interventions on relaxation were different in ejecting versus isovolumic beats.

We then applied this analysis to compare the lusitropic effects of EMD-57033 and another recently introduced Ca²⁺ sensitizer, the novel 1,5-benzodiazocine derivative BA-41899 (5-methyl-6-phenyl-1,3,5,6-tetrahydro-3,6-methano-1,5-benzodiazocine-2,4-dione) (8, 9). Like EMD-57033, BA-41988 possesses enantiospecific effects: the Ca²⁺-sensitizing activity resides in the positive enantiomer CGP-48506 (8), which is devoid of PDE I-IV activity at concentrations 3 × 10⁴ M in both human (17) and guinea pig (27) myocardium, making it the most selective Ca²⁺ sensitizer introduced to date (8). Although in vitro studies have suggested that CGP-48506 slows relaxation (8, 9, 17, 18, 24, 27), some of these data also suggested that, unlike EMD-57033, CGP-48506 does not adversely affect resting cell length (24) or force (18). On the basis of these and other observations, it has been suggested both that the site of action of CGP-48506 is different from that of EMD-57033 and that CGP-48506 might have fewer adverse effects on diastolic function than EMD-57033 (18, 24). Accordingly, the second purpose of our study was to test the hypothesis that CGP-48506 would have less effect than EMD-57033 on the stress-independent effect on relaxation.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Isolated Heart Preparation

The heart was transferred to a perfusion support system, which was maintained at a constant temperature of 30°C. A thin latex balloon (constructed from an adult esophageal balloon) was secured to the flared ostium of the piston cylinder of the volume-servo system. A suture attached to the end of the balloon was passed through the mitral orifice and out through a puncture in the LV apex. A purse-string suture was then tightened around the flared ostium to secure the balloon within the LV chamber. A 5-Fr Millar Micro-Tip catheter pressure transducer was advanced through a side port in the piston cylinder into the center of the balloon. The perfusate was then changed to a modified Tyrode solution, which contained a lower K⁺ concentration, to allow contractions. This solution was composed of (in mM) 147 Na⁺, 7.44 K⁺, 138 Cl, 1.24 Ca²⁺, 1.08 Mg²⁺, 21 , 0.36 , and 11.1 glucose, with 2.5 U/l insulin. The heart beat isovolumically and was paced at 1 beat/s using punctate stimulation with the electrodes placed at the LV apex.

A single-beat, variably preloaded Frank-Starling (FS) protocol (2) was conducted to identify the LV volume (V_max) at which the peak developed pressure of the Frank-Starling curve was obtained. LV volume was then adjusted so that it was 80% of V_max, and this was the baseline volume (V_BL) used throughout the experiment.

Preparation of Solutions

Experimental Protocols

6

6

6

After V_BL was established, but before treatment with an inotropic drug, two protocols were conducted while in the control state: 1) a single-beat, isovolumic, variably preloaded FS protocol and 2) a single-beat, ejecting, variably afterloaded (AL) protocol. Records taken during these protocols yielded a family of functional indexes to serve as control values. After these two protocols were completed, perfusion was switched from control perfusate to a perfusate containing one of the three inotropic drugs, followed by a 15-min period for a stable baseline to be established, as indicated by stable peak pressure. The FS and AL protocols were then once again conducted, and records were taken during the influence of the inotropic drug.

FS Protocol. The single-beat FS protocol has been described in detail elsewhere (2, 20, 21). Briefly, with the heart beating isovolumically at V_BL, 80 ms before a selected beat the volume was commanded to be 1 of 10 volumes (50-140% of V_BL, in 10% increments) (Fig. 1A). This selected beat is called the volume-perturbed beat. The contraction and relaxation of the volume-perturbed beat took place isovolumically at the commanded volume. The heart period of the volume-perturbed beat was prolonged to 115% of the basal period to ensure complete relaxation during the diastolic period at the commanded volume. The ensuing beat also took place at the commanded volume, and, after this beat, volume and heart period were returned to baseline. Fifteen seconds were allowed for transients to die out, and then the procedure was repeated at one of the other commanded volumes. This sequence was repeated until records (digitally sampled at 250 Hz) had been taken for the set of 10 commanded volumes. Stability during the protocol was assessed by variation in peak isovolumic pressure of the beat before the volume-perturbed beat among the full set of records. This variation was always <3%.

View larger version (40K): [in this window] [in a new window]   View larger version (17K): [in this window] [in a new window]   Fig. 1.   A: 10 superimposed pressure and volume records for volume-perturbed isovolumic beats generated from a single-beat Frank-Starling (FS) protocol. Set of 10 beats provides information necessary for determining FS relationship, passive pressure-volume relationship, and isovolumic relaxation-stress relationship. B: pressure variables: Ppeak, maximal isovolumic pressure; Ppass, minimum pressure measured during a prolonged diastole; and DPpeak, peak developed pressure (Ppeak  Ppass). Timing variables: Tpeak, time from pacing pulse to occurrence of DPpeak; T75-25, characteristic time of relaxation, measured as time necessary for pressure to fall from 75 to 25% of DPpeak; and Ttotal, total duration of pressure development measured from time of pacing pulse to time during relaxation when pressure falls to 10% of DPpeak.

(1)

AL Protocol. The AL protocol is one we have used previously to construct end-systolic pressure-volume relationships and has been described in detail elsewhere (2). Briefly, with the heart beating isovolumically at V_BL, a beat was selected, called the pressure-clamped beat, in which pressure was not allowed to rise above one of eight commanded pressures (100-40% of P_peak, in 10% decrements) (Fig. 2A). This was achieved by allowing the beat to proceed isovolumically until pressure rose to the commanded value, at which point pressure was clamped. Pressure clamping, which was achieved by volume withdrawal, continued until maintenance of the commanded pressure required volume infusion. At that point, volume was held constant at the end-systolic volume (V_es) while pressure fell as the heart relaxed. The heart was refilled to V_BL before the ensuing beat and was allowed to beat isovolumically for 15 s, and then the next record was taken. This sequence was repeated for all eight commanded pressures.

View larger version (23K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Fig. 2.   A: 8 superimposed pressure and volume records for pressure-clamped ejecting beats generated from a single-beat variably afterloaded (AL) protocol. Set of 8 beats provides information necessary for determining ejecting relaxation-stress relationships. B: characteristic features of an ejecting beat measured for these analyses: Tej, time from pacing pulse to time of maximal pressure-volume ratio during ejection (end ejection); Pej, left ventricular (LV) pressure at Tej; Ppass, minimum pressure at end-systolic volume (Ves); and Pes, end-systolic pressure (Pej  Ppass).

(2)

(3)

Data Analysis

Frank-Starling relationship and passive pressure-volume relationship. The effects of each treatment on the fully relaxed, pressure-volume relationship and the Frank-Starling relationship were evaluated within each treatment group. The passive pressure-volume relationships were compared visually. The Frank-Starling relationships were compared using a multiple linear regression in which DP_peak was related to volume (V) according to

(4)

where D is a dummy variable to encode control (D = 0) and drug treatment (D = 1) conditions and the b terms are regression coefficients (b₀, DP_peak intercept; b_v, initial "slope"; and b_v2, quadratic term). The term b_dD allows for a drug-induced change in the DP_peak-axis intercept, which would be seen as a parallel shift in the Frank-Starling relationship. The term b_vdV · D allows for a drug-induced change in the initial slope, which would be seen as a nonparallel shift in the Frank-Starling relationship.

Wall-Stress Dependence of Relaxation. Similarly, the difference in the T_75-25-_peak relationships for isovolumic beats under control conditions and during perfusion with a drug was analyzed using multiple linear regression in which T_75-25 was related to _peak according to

(5)

where D is a dummy variable as defined for Eq. 4 and the c terms are regression coefficients that form the basis for interpreting the effects of a drug. If drug treatment did not change the relationship between T_75-25 and _peak, (i.e., the control and drug treatment data fell on the same regression line), then any effect of this agent to slow relaxation would be ascribable entirely to its effect in increasing _peak. However, if drug treatment changed the slope (i.e., c_d is significant in Eq. 5) or intercept (i.e., c_d is significant in Eq. 5) of the relationship between T_75-25 and _peak, this would indicate that the drug had an additional effect on the duration of LV relaxation that occurred independently of its effect in changing _peak.

(6)

View larger version (8K): [in this window] [in a new window]   Fig. 3.   T75-25- relationship determined in 1 left ventricle during control state. Data illustrate interpretation of Eqs. 8 and 9. For isovolumic beats, stroke circumference (Cs)  0 and regression fit follows isovolumic  data points as indicated by curve A. For variably ejecting beats, Cs  0 and regression fit follows + data points as indicated by curve B. Note that isovolumic and ejecting curves increasingly diverge with increasing Cs (i.e., lower ). Regression coefficients fc and fc in Eqs. 8 and 9 quantify this effect of ejection on T75-25- relationship.

(7)

(8)

(9)

cd

3

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Frank-Starling and Passive PressureVolume Relationships

As we have shown previously for EMD-57033 (21), there was little or no effect of these three inotropic drugs on the fully relaxed, passive pressure-volume relationship [data not shown; for example, see Fig. 3 in Tobias et al. (21)].

Timing of Contraction Events in Isovolumic Beats

6

6

6

6

View larger version (26K): [in this window] [in a new window]   Fig. 4.   Examples of volume-perturbed isovolumic beats from FS protocol and pressure-clamped ejecting beats from AL protocol. Isovolumic beats have been normalized by subtracting Ppass from pressure of perturbed beat [P(t)] and then dividing by DPpeak so that pressure varies from 0 to 1. Two superimposed isovolumic beats are shown in each panel. Pressure-clamped ejecting beats have been normalized by subtracting Ppass from P(t) and then dividing by clamp pressure percentage such that pressure varies from 0 at end diastole to clamp-pressure level as a percentage of Ppeak. Two superimposed ejecting beats are also shown in each panel. Left: a pair of isovolumic beats and a pair of ejecting beats for which pressure-clamp level was set to 40% of Ppeak. Right: same pair of isovolumic beats and a pair of ejecting beats for which pressure-clamp level was set to 80% of Ppeak. A: each pair of beats shows a control beat (dashed line) superimposed with a beat measured with 6 × 106 M CGP-48506 (solid line). B: each pair of beats shows a control beat (dashed line) superimposed with a beat measured with 2 × 106 M EMD-57033 (solid line). C: each pair of beats shows a control beat (dashed line) superimposed with a beat measured with 1.25 × 106 M dobutamine (solid line). Effects of these 3 drugs on contraction and relaxation timing are readily apparent. See text for average values of timing variables.

As expected, T_peak and T_total were significantly shortened by 1.25 × 10⁶ M dobutamine [T_peak was 314 ± 13 ms in the control state and 262 ± 14 ms with dobutamine (16.6% shortening; P < 0.001) and 693 ± 53 ms in the control state and 575 ± 40 ms with 1.25 × 10⁶ M dobutamine (17% shortening; P < 0.001)].

T_75-25-_peak Relationship in Isovolumic Beats

d

6

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Effect of 6 × 106 M CGP-48506 (A), 2 × 106 M EMD-57033 (B), and 1.25 × 106 M dobutamine (C) on T75-25-peak midwall stress (peak) relationship in isovolumic beats (T75-25 in s and peak in mmHg). All observations from all 8 hearts in each group are shown for control () and drug treatment (). Dashed line is regression fit to control observations using Eq. 5, ignoring dummy variables, and represents average control curve for all 8 hearts. Similarly, solid line is regression fit to drug-treatment observations using Eq. 5, ignoring dummy variables, and represents average drug-treatment curve for all 8 hearts. See text and Table 1 for details of regression fits to Eq. 5.

In contrast, as we have observed previously (21), 2 × 10⁶ M EMD-57033 shifted the T_75-25-_peak relationship upward such that relaxation at any given _peak was significantly slower with EMD-57033. Both c_d and c_d were significant (P < 0.001; Table 1). Thus the upward shift caused by EMD-57033 was not parallel (i.e., both slope and intercept increased with EMD-57033).

Dobutamine (1.25 × 10⁶ M) shifted the T_75-25-_peak relationship downward such that relaxation at any given _peak was significantly shorter with dobutamine. Both c_d and c_d were significant (P < 0.001; Table 1). Thus the downward shift caused by dobutamine was not parallel (i.e., both slope and intercept decreased with dobutamine).

Timing of Contraction Events in Ejecting Beats

View larger version (12K): [in this window] [in a new window]   Fig. 6.   Effect of 6 × 106 M CGP-48506 (A), 2 × 106 M EMD-57033 (B), and 1.25 × 106 M dobutamine (C) on Tej-end-systolic midwall stress (es) relationship in ejecting beats (Tej in s and es in mmHg). For clarity of presentation, individual observations from 8 hearts are not shown (between-heart variation is similar to that shown in Fig. 5). Dashed line is regression fit to control observations using Eq. 7, ignoring dummy variables, and represents average control curve for all 8 hearts. Similarly, solid line is regression fit to drug-treatment observations using Eq. 7, ignoring dummy variables, and represents average drug-treatment curve for all 8 hearts. Note biphasic effect of Ca2+ sensitizers. Both drugs shorten Tej for lowest values of es, but this effect crosses over to a prolongation of Tej as es increases. Dobutamine shortens Tej for all es, but this effect diminishes as es increases.

The effect of both Ca²⁺ sensitizers on this relationship is biphasic (Fig. 6). For the lowest levels of _es (i.e., the largest ejections), T_ej is reached more quickly when the heart is perfused by either CGP-48506 or EMD-57033 than in the control. However, at intermediate levels of _es, this effect crosses over to an effect in which at the highest levels of _es (i.e., smallest ejections) T_ej is reached at a later time in the presence of either CGP-48506 or EMD-57033. From the regression fits to Eq. 7 in each drug group, these crossover points are estimated to be 37 mmHg for 6 × 10⁶ M CGP-48506 and 31 mmHg for 2 × 10⁶ M EMD-57033.

In contrast, 1.25 × 10⁶ M dobutamine shortened T_ej at all values of _es, with the T_ej-_es relationship for baseline control and dobutamine conditions converging only at the highest _es (the point of convergence calculated from the fit to Eq. 7 is 106 mmHg).

T_75-25-_es Relationship in Ejecting Beats

View larger version (10K): [in this window] [in a new window]   Fig. 7.   Effect of 6 × 106 M CGP-48506 (A), 2 × 106 M EMD-57033 (B), and 1.25 × 106 M dobutamine (C) on T75-25-es relationship in ejecting beats (T75-25 in s and es in mmHg). For clarity of presentation, individual observations from 8 hearts are not shown (between-heart variation is similar to that shown in Fig. 5). Dashed line is regression fit to control observations using Eq. 6, ignoring dummy variables, and represents average control curve for all 8 hearts. Similarly, solid line is regression fit to drug-treatment observations using Eq. 6, ignoring dummy variables, and represents average drug-treatment curve for all 8 hearts. There are qualitative differences among effects of these drugs in ejecting beats compared with their effects in isovolumic beats. For example, CGP-48506 has no effect on T75-25-peak relationship in isovolumic beats (see Fig. 5) but has a biphasic effect in ejecting beats which slows relaxation slightly at low es and then speeds relaxation slightly at higher es. See text and Table 1 for details of regression fits to Eq. 6.

For the CGP group, the coefficients d_d and d_d were both significant (P < 0.001; Table 1). Thus there was a significant CGP-48506-induced increase in intercept and decrease in slope of the T_75-25-_es relationship, as can be seen by the average fit lines drawn in Fig. 7A. Hence, unlike the response to this drug in isovolumic contractions, relaxation was prolonged slightly by CGP-48506 at the lowest _es (i.e., largest ejections). This effect is opposite to the effect of this drug on T_ej, with the overall result that the T_total-_es relationship in ejecting beats was largely unaffected by 6 × 10⁶ M CGP-48506 (Fig. 8A).

View larger version (9K): [in this window] [in a new window]   Fig. 8.   Effect of 6 × 106 M CGP-48506 (A), 2 × 106 M EMD-57033 (B), and 1.25 × 106 M dobutamine (C) on Ttotal-es relationship in ejecting beats (Ttotal in s and es in mmHg). For clarity of presentation, individual observations from 8 hearts are not shown (between-heart variation is similar to that shown in Fig. 5). Dashed line is average regression fit to control observations. Similarly, solid line is average regression fit to drug-treatment observations. Summed effects of a drug on Tej and T75-25 to determine Ttotal in ejecting beats closely parallel effects of drug on T75-25-peak relationship in isovolumic beats.

The upward shift in the T_75-25-_es relationship that occurred with 2 × 10⁶ M EMD-57033 was larger than that for CGP-48506; the coefficients d_d and d_d were both significant (P < 0.006; Table 1), indicating a significant EMD-57033-induced increase in intercept and decrease in slope of the T_75-25-_es relationship, as can be seen by the average fit lines drawn in Fig. 7B. Hence, relaxation was prolonged by EMD-57033 at all _es, but the prolongation was slightly less at higher _es. The overall result, combining the dependence of both T_ej and T_75-25 on _es, was that the T_total-_es relationship in ejecting beats was shifted upward by 2 × 10⁶ M EMD-57033. This shift was such that the slope of the T_total-_es relationship was increased by EMD-57033. Thus T_total was relatively more prolonged with EMD-57033 at high _es compared with at low _es (Fig. 8B).

Dobutamine (1.25 × 10⁶ M) shifted the T_75-25-_es relationship downward such that the time of relaxation at any given _es was significantly shorter with dobutamine. The coefficients d_d and d_d were both significant (P < 0.001; Table 1). Thus both the slope and intercept decreased with dobutamine (Fig. 7C). The overall result, combining the dependence of both T_ej and T_75-25 on _es, was that the T_total-_es relationship in ejecting beats was shifted downward by 1.25 × 10⁶ M dobutamine. Dobutamine decreased the slope of the T_total-_es relationship, with the result that T_total was shortened a relatively larger amount with dobutamine at high _es compared with at low _es (Fig. 8C).

T_75-25- Relationship in Both Isovolumic and Ejecting Beats

View larger version (11K): [in this window] [in a new window]   Fig. 9.   Effect of 6 × 106 M CGP-48506 (A), 2 × 106 M EMD-57033 (B), and 1.25 × 106 M dobutamine (C) on T75-25-wall stress () relationship in data combined from both isovolumic and ejecting beats (T75-25 in s and  in mmHg). For clarity of presentation, individual observations from 8 hearts are not shown (between-heart variation is similar to that shown in Fig. 5). Dashed lines show regression fit to control observations using Eq. 8, ignoring dummy variables, and represent average control fit for all 8 hearts. Similarly, solid lines show regression fit to drug-treatment observations using Eq. 8, ignoring dummy variables, and represent average drug-treatment fit for all 8 hearts. See Fig. 3 for explanation of relationship between isovolumic and ejecting data; see text and Table 1 for details of regression fits to Eq. 8.

The effect of ejection on relaxation could enter either directly (i.e., the term f_cC_s in Eq. 8) or as a work-related term (i.e., f_cC_s · _es in Eq. 8). In a regression model this complex, it is difficult to determine unequivocally which of these two possibilities is involved. For example, in the control state, f_c was significant for the CGP and EMD groups (P < 0.001; Table 1), whereas f_c was significant for the CGP and dobutamine groups (P < 0.001; Table 1). In the absence of a more consistent finding in the control state, we will not attempt to interpret either effect separately but rather will treat these as a lumped effect of C_s.

Drug-induced changes in the stress-independent effects of ejection are represented by the coefficients f_cd and f_cd in Eq. 8. If either coefficient is significant, the effect of the drug treatment on relaxation (as judged by the T_75-25- relationship) is different in ejecting versus isovolumic beats. The coefficient f_cd in Eq. 8 was not significant in any of the drug treatment groups (all P > 0.18; Table 1). In the CGP group, f_cd was significant (P < 0.001; Table 1), indicating that CGP-48506 affects relaxation in ejecting beats differently from in isovolumic beats. Furthermore, the coefficient f_d was not significant (P = 0.13; Table 1), indicating that there isno CGP-48506-induced offset in the combined T_75-25- relationship. Thus the only effect of CGP-48506 is to influence relaxation in ejecting beats but not in isovolumic beats (these effects are readily apparent visually in Fig. 9 and can also be appreciated by comparing Figs. 5A and 7A). In contrast, in the EMD group, f_cd was not significant (P = 0.27; Table 1), indicating that EMD-57033 does not affect relaxation in ejecting beats differently from in isovolumic beats; that is, EMD-57033 slows relaxation similarly in both isovolumic and ejecting beats as indicated by the significance of the coefficient f_d (P < 0.001; Table 1), which signifies a simple EMD-57033-induced offset in the combined T_75-25- relationship (Fig. 9). Finally, in the dobutamine group both f_d and f_cd were significant (P < 0.009; Table 1). Thus dobutamine speeds relaxation in both ejecting and isovolumic beats (i.e., f_d is significant and negative), but it diminishes the effect of ejection to speed relaxation (i.e., f_cd is significant and positive) (Fig. 9).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

We applied a comprehensive analysis of LV relaxation that encompassed both isovolumic and ejecting beats to identify the lusitropic effects caused by different positive inotropic interventions. Our principal findings are that the two specific Ca²⁺ sensitizers, EMD-57033 and CGP-48506, have very different effects on LV relaxation and that differences in the effects of these two drugs depend on the mode of LV contraction. At the concentrations we used, neither EMD-57033 nor CGP-48506 has been shown to have significant confounding effects that might also influence relaxation. EMD-57033 does not significantly inhibit PDE III activity at concentrations <5 × 10⁶ M (19), and CGP-48506 does not significantly inhibit PDE III activity at concentrations <3 × 10⁴ M (17, 27). Further, at these concentrations neither drug affects Ca²⁺ binding to troponin C (19, 24). Finally, although CGP-48506 has been reported to increase L-type Ca²⁺-channel conductance, this effect was at a high concentration, 10⁴ M (10). Accordingly, we assume that the differences we observed reside in differential effects of these drugs "downstream" from troponin C.

Effect of Ca²⁺ Sensitizers on Relaxation in Isovolumic Beats

EMD-57033 shifted the T_75-25-_peak relationship in isovolumic beats upward. Thus, in addition to the effect of EMD-57033 to slow relaxation because its positive inotropic effect increased _peak, it also has a stress-independent negative lusitropic effect that further slowed relaxation by increasing T_75-25 at any given _peak (21). In contrast, CGP-48506, when administered to give a similar increase in inotropic state, has a negligible effect on relaxation; that is, it did slow relaxation, as would any positive inotropic intervention, because it increased _peak. However, this is the only effect of CGP-48506, because the T_75-25-_peak relationships with and without CGP-48506 were nearly superimposed.

For comparison, and as further validation of this framework for analyzing LV relaxation, we also studied the effect of 1.25 × 10⁶ M dobutamine on the T_75-25-_peak relationship. We chose a -agonist because its effect via the adenosine 3',5'-cyclic monophosphate-dependent protein kinase led us to predict that it would cause a downward shift in the T_75-25-_peak relationship [consistent with its effect as shown in Zile et al. (25)]. Indeed, our analysis shows that dobutamine has a strong positive lusitropic effect.

Effect of Ca²⁺ Sensitizers on Relaxation in Ejecting Beats

Furthermore, as shown in Eq. 8 and Fig. 9, by combining the measurements of T_75-25, _peak from isovolumic beats, and _es and C_s from ejecting beats, we can determine directly whether the lusitropic effect of these positive inotropic drugs, as judged by the T_75-25- relationship, is different in ejecting versus isovolumic beats. This analysis shows that EMD-57033 affects stress-dependent relaxation in a similar way in both ejecting and isovolumic beats, whereas both CGP-48506 and dobutamine affect stress-dependent relaxation in ejecting beats differently from how they affect it in isovolumic beats. For CGP-48506 and EMD-57033, the result of this combined analysis agrees with the interpretation made when comparing separate data for isovolumic and ejecting beats in T_75-25-_peak Relationship in Isovolumic Beats and T_75-25-_es Relationship in Ejecting Beats. However, for dobutamine the result of this combined analysis suggests that its effect was different for isovolumic versus ejecting beats, which is difficult to appreciate from a visual comparison of the separate data shown in Figs. 5C and 7C. In summary, EMD-57033 slows relaxation similarly in both ejecting and isovolumic beats. CGP-48506 affects relaxation differently in ejecting versus isovolumic beats: it does not affect relaxation in isovolumic beats but reduces the effect of ejection to speed relaxation. Unlike CGP-48506, dobutamine affects relaxation in both ejecting and isovolumic beats but, like CGP-48506, it reduces the effect of ejection to speed relaxation.

Evaluation of the lusitropic effect of such drugs in ejecting beats is more complex than in isovolumic beats because, in addition to an effect on duration of relaxation, as quantified by T_75-25, these drugs potentially affect the time of relaxation onset. Hence, we performed an additional analysis, relating T_ej to _es (Fig. 6) according to Eq. 7. The effect of the two Ca²⁺ sensitizers is qualitatively similar in this analysis. Both drugs reduce T_ej by a small amount at low _es and increase T_ej at larger _es. These two drugs differ quantitatively in that CGP-48506 reduces T_ej over a broader range of _es than does EMD-57033. In contrast, dobutamine reduces T_ej over most of the range of _es in the ejecting beats. The effect of these drugs on T_75-25 and T_ej, combined, produces an overall effect on the duration of beat. The T_total-_es relationship in ejecting beats is affected by these three inotropic drugs in virtually the same manner as the T_75-25-_peak relationship in isovolumic beats was affected.

The more complex nature of relaxation in ejecting beats has been highlighted previously, and the determinant of relaxation rate in ejecting beats has been suggested to be, variously, load (26) or ejection timing (12). Our approach is much like that of Zile et al. (25, 26), who related relaxation to ventricular loading and showed that the -agonist isoproterenol shifted the relaxation-load relationship (25). We treated load ( in our study) as the determinant of relaxation rate (T_75-25) in ejecting beats because it allowed us to analyze both isovolumic and ejecting beats in a similar manner. Moreover, doing so allowed us to combine data from both isovolumic and ejecting beats to evaluate whether an intervention had a different effect on isovolumic versus ejecting beats. We treated ejection timing as another important feature of relaxation in ejecting beats but did not treat ejection timing as a determinant of relaxation rate, as did Hori et al. (12). Because both T_ej (Fig. 6) and T_75-25 (Fig. 7) depend on _es, it is apparent that T_ej and T_75-25 could be analyzed in such a way as to treat T_ej as a "determinant" of T_75-25. However, it was not our intent to decouple T_ej from _es so as to identify which was really the determinant of T_75-25. Rather, we treated _es as the determinant of T_75-25, because doing so allowed a consistent framework for analyzing the effects of interventions on relaxation in both isovolumic and ejecting beats.

Mechanism(s) of Action of CGP-48506 and EMD-57033

6

5

In a detailed study comparing EMD-57033 and CGP-48506 in both live and skinned fiber preparations, Palmer et al. (18) concluded that CGP-48506 and EMD-57033 increased the Ca²⁺ sensitivity of myofilaments via different mechanisms. One conclusion from their findings was that CGP-48506 did not affect the tension cost (i.e., did not shift the ATPase-force relationship) and that, therefore, assuming the number of cross bridges and force per cross bridge were unaffected [following Brenner (1)], CGP-48506 did not affect the cross-bridge detachment rate (g_app). On the basis of this finding and others, they proposed that CGP-48506 affected the transition from the detached to the weakly bound state. In contrast, they proposed that the site of action of EMD-57033 was at the weak-to-strong transition.

Our results from the whole heart are consistent with the interpretation that CGP-48506 does not affect g_app. However, we do not think our results allow us to speculate further about specific differences in the site(s) of action(s) of these two drugs that would be implied by a differential effect in ejecting versus isovolumic beats. Nevertheless, our results show that the Ca²⁺-sensitizing effects of CGP-48506 and EMD-57033 must be mediated by different sites of action on the myofilaments.

Using T_75-25- Relationship to Evaluate Lusitropy

The implication of a positive T_75-25- relationship is that any positive inotropic drug will slow LV relaxation because it increases LV wall stress. When comparing one beat in the absence of a positive inotropic drug to another beat in the presence of drug, the net effect on relaxation will be the balance of this effect due to increased wall stress and any stress-independent lusitropic effect the drug may have. Accordingly, evaluation of a drug's effect by comparing only two beats will not allow discrimination between the negative lusitropic effect that would be present with any inotropic interven- tion of the same magnitude (i.e., slowed relaxation due to increased wall stress) and the stress-independent lusitropic effect of the drug.

The necessity for such an evaluation is depicted in Fig. 10, which shows a family of four hypothetical T_75-25- relationships. Three of these depict the results of this study, whereas the fourth depicts a hypothetical positive inotropic agent (Fig. 10, line C) with a smaller positive lusitropic effect than dobutamine (line D). Assume, for argument's sake, that all four drugs result in the same positive inotropic effect, and thus the same increased wall stress, as indicated by the vertical dashed line in Fig. 10. The arrows emanating from point 1 (i.e., the hypothetical baseline state) on the control curve (Fig. 10, line B) show the summed effect of the increased wall stress plus the stress-independent effect of the drug (i.e., shift of the curve up or down) on relaxation. Of most significance for the purposes of this illustration, consider the effect of the hypothetical inotrope (Fig. 10, line C), which has a net effect to shift the LV from point 1 to point 4. The net effect of this drug is, in fact, to slow relaxation when judged only by comparing a beat at point 1 with a beat at point 4. However