Diastolic dysfunction in stunned myocardium: a state of abnormal excitation–contraction coupling that is limited by Na⁺–H⁺ exchange inhibition

Abstract

Objective: The systolic and diastolic effects of myocardial stunning were studied to evaluate the contributions of the endocardial and epicardial segments of the ventricular myocardial band, and determine if preconditioning by Na⁺–H⁺ exchange (NHE) inhibition effected post-stunning dysfunction. Methods: Thirteen Yorkshire–Duroc pigs (27.3–38.2 kg) underwent 15 min of mid-LAD clamping. Seven had no protective measures and six were pretreated with IV Cariporide 5 mg/kg 15 min before ischemia. Sonomicrometer crystals evaluated systolic dysfunction (impaired regional shortening) and diastolic dysfunction (contraction extending into early diastole). Results: Before ischemia, contraction started first on the endocardial side followed 82 ± 23 ms later by the subepicardium. Endocardial shortening stopped first, coinciding with negative dP/dt onset, while epicardial shortening phase persisted for 92 ± 33 ms more during occurrence of rapid LVP descent and development of peak negative dP/dt. Ischemia produced paradoxical bulging of both segments. Sixty minutes after ischemia systolic segment shortening recovered 36 ± 24% of baseline values without pretreatment, compared to 75.8 ± 15% with Cariporide (p ≪ 0.05). Global ejection force (maximum dP/dt) fell 32 ± 20% in the unprotected group, but was maintained by Cariporide pretreatment. Diastolic dysfunction always showed continued endocardial contraction into early diastole (occupying 38 ± 16% of diastole in untreated hearts), whereas Cariporide treatment reduced this dysfunction to 5 ± 10% (p ≪ 0.05). Persistent diastolic dysfunction raised left ventricle end diastolic pressure (LVEDP) 4 mmHg in untreated hearts, whereas Cariporide returned LVEDP to normal. Less elevation of creatine kinase MB (CK-MB) and conjugated dienes followed Cariporide pretreatment. Conclusions: Temporary LAD ischemia alters the normal sequential pattern of contraction responsible for ejection and suction by (a) reducing systolic contractile force, and (b) prolonging endocardial contraction into early diastole to disrupt the normal endocardial–epicardial sequence responsible for ventricular suction. NHE inhibition before ischemia limits postischemic systolic and diastolic dysfunction by re-establishing the expected shortening sequences within the ventricular myocardial band model.

1 Introduction

Prolonged reversible postischemic contractile dysfunction, or stunned myocardium, has been traditionally associated to systolic dysfunction with delayed and impaired contraction [1]. Diastolic abnormalities (i.e., dysfunction) have also been described to limit rapid filling [2], but the mechanisms responsible for alterations in the diastolic phase remain controversial. Some authors relate the abnormal rapid filling phase to adverse elastic myofiber recoil from systole [3], in relationship with the protein titin [4,5]. An alternate suggested mechanism is limited recoil of potential energy from prior systole during the period of isovolumetric relaxation [6]. However, new evidence [7] suggests that orderly sequential contraction that relates to the spatial orientation of myocardial fibers is responsible for a ‘systolic shortening’ component that produces active suction of blood from the auricles into the ventricles.

We suspect that disruption of the sequential ‘systolic contraction’ may be responsible for the excitation–contraction mismatch of myocardial stunning, thereby involving a reversible muscular phenomenon that impairs suction and then leading to the secondary requirement of a higher filling pressure. Diastolic dysfunction occurs in ∼30% of patients in congestive heart failure with normal systolic function [8], yet active treatment of a muscular cause is not currently part of the treatment profile [9]. A broad new avenue of treatment may become available if this active contractile concept cause of diastolic dysfunction is correct.

Examination of multiple factors in the pathogenesis of stunning suggests two principal theories that include generation of oxygen-derived free radicals and disruption of calcium homeostasis [10]. Three main calcium disturbances have been described [11]. First, an increase in intracellular calcium concentration during ischemia, but mainly during reperfusion due to extracellular influx, and a decrease in reuptake into the sarcoplasmic reticulum. Second, marked reduction in myofilament sensitivity to calcium during reperfusion, probably in response to this ion saturation. This is an uneven response because myocardial bands with rigor contracture have also been described during reperfusion and attributed to calcium overload. Third, the contractile proteins may be damaged due to proteolysis via calcium overload and activation of calpain-I, a calcium-activated neutral protease found in cardiac muscle.

During ischemia, a metabolic mismatch between anaerobic glycolysis and glucose oxidation leads to proton accumulation. In response to intracellular acidosis, the sarcolemmal Na⁺–H⁺ exchanger (NHE), quiescent under basal conditions, becomes activated. The resulting net Na⁺ influx by NHE causes intracellular Na⁺ overload, in part due to the ischemia-induced inhibition of the primary Na⁺ extrusion pathway, the Na⁺/K⁺ ATPase [12]. In an attempt to decrease Na_I⁺ overload, the reversible Na⁺/Ca⁺⁺ exchanger changes its normal Ca⁺⁺ efflux, enhancing Ca⁺⁺ entrance [13,14]. The so-called Ca⁺⁺ overload is responsible for rigor contracture, and by activation of calcium-dependent proteolytic and lipolytic enzymes it leads to irreversible cell injury [15,16]. Na⁺–H⁺ exchange inhibition by Cariporide has shown cardioprotective effects in many experimental models by attenuating both Na⁺ and Ca⁺ influx into cardiomyocytes [17].

This study will (a) focus upon the calcium cause that is tested in a practical regional coronary occlusion model that is comparable to that used during off pump surgery, (b) identify how the endocardial and epicardial structural elements of the ventricular myocardial band can interact to produce systolic and diastolic dysfunction, and (c) demonstrate how this dyssynergy can be favorably altered pharmacologically by interfering with the Na⁺–H⁺ exchange inhibition by Cariporide to reverse a muscular cause of diastolic dysfunction.

2 Material and methods

All animals received humane care in compliance with the 1996 NRC Guide for the Care and Use of Laboratory Animals, available at: http://www.nap.edu/readingroom/books/labrats/contents.html.

Eighteen Yorkshire–Duroc pigs (31.8 ± 3.9 kg) of either sex were pre-medicated (ketamine 15 mg/kg, diazepam 0.5 mg/kg intramuscularly) and anesthetized with pentobarbital, 30 mg/kg, intravenously and subsequent bolus injections of sodium pentobarbital. Support with a volume-controlled ventilator (Servo 900C, Siemens-Elema, Sweden) was started after tracheostomy and endotracheal intubation. The femoral artery and vein were cannulated and arterial blood gases measured to keep oxygen tension, carbon dioxide tension, and pH values within the normal range. A balloon-tipped catheter (Model 132F5, Baxter Healthcare Corp., Irvine, CA, USA) was advanced into the pulmonary artery through a jugular vein to measure pulmonary artery and wedge pressures, and cardiac output by thermodilution technique.

The pericardium was incised after median sternotomy and a solid-stated pressure transducer-tipped catheter (Model MPC-500, Millar Instruments Inc., Houston, TX, USA) was inserted through the apex to monitor left ventricular pressure (LVP). A segment of the left anterior descending coronary artery (LAD), distal of the first major diagonal branch, was circumferentially dissected. A 3/0 prolene suture in between a Teflon pledget was placed surrounding the LAD. A retrograde cardioplegia cannula was introduced transatrially into the coronary sinus for coronary blood sampling.

3 Experimental protocol

Pigs were divided in three different groups.

3.1 Unjeopardized

Control group: Five normal hearts did not have LAD ischemia, and were followed the same protocol as control procedure group.

3.2 Ischemia

Unprotected ischemia: Seven pigs underwent 15 min of LAD occlusion followed by 60 min of reperfusion.

HOE pretreatment: In six pigs with similar 15 min ischemic episode, 5 mg/kg Cariporide (HOE-642) was intravenously administered 15 min before LAD clamping.

4 Hemodynamic measurements

Left ventricular pressure was electronically differentiated to obtain dP/dt, the first LVP derivative. The cardiac cycle was defined automatically using dP/dt. Systole began 40 ms before peak positive dP/dt, and ended 20 ms before peak negative dP/dt, as defined previously by other groups [18,19].

Two pairs of 2 mm ultrasonic micro-transducer crystals (Sonometrics, London, Ont., Canada) were placed in the endocardial and epicardial sides of the anterior free wall of the left ventricle dependent of LAD perfusion. Adequacy of LAD crystal placement was checked within 20 s of temporal LAD ischemia, and sonomicrometer crystals orientation was chosen to get the best angle of contraction, as described previously [7]. Another pair of crystals was placed in the left circumflex coronary artery (LCx) muscular region as control. Data were recorded digitally via acquisition hardware and software (Sonometrics). Systolic shortening was defined as the change in segment length between start and end systole. Percent systolic shortening (SS%) was calculated as the ratio in percentage between systolic shortening and end diastolic length. Contraction during the first part of diastole was analyzed by the percentage of time of diastole in which contraction is in effect (DCT%).

Cardiac output was determined by mean of four central venous injections of 3 mL of 4 °C saline solution through a Swan–Ganz catheter.

4.1 Coronary sinus blood analysis

Myocardial injury was determined from analysis of coronary sinus blood samples taken 5 min before, and 30 and 60 min after the onset of myocardial ischemia.

4.2 Conjugated dienes (CD)

As a marker of oxidant-mediated lipid peroxidation, conjugated diene levels were determined spectrophotometrically in coronary sinus plasma after chloroform–methanol 2:1 (v/v) extraction. Conjugated diene concentration was expressed as absorbance (A) at a wavelength of 240 nm per 0.5 mL plasma.

4.3 Creatine kinase MB (CK-MB)

Myocardial damage was determined by measuring creatine kinase fraction MB (U/L) in coronary sinus plasma by a UV-spectrophotometric method (Sigma Chemical Co., St. Louis, MO, USA) as recommended by the German Society for Clinical Chemistry.

4.4 Myocardial biopsy

At the end of the experiment, the pigs were killed by bolus injections of pentobarbital 5 mg followed moments later by 15 mL cold hyperkalemic blood (KCL, 30 mEq/L). With the LAD and the ascending aorta cross clamped, 60 mL methylene blue was injected through the aortic root into the coronary arteries to stain the non-ischemic myocardium. The ischemic myocardium was calculated as percentage of dry weight between ischemic and non-ischemic myocardium.

4.5 Statistical analysis

Statistical analysis of data within and between groups was performed using multiple analyses of variance followed by application of the Student's t-test with Tukey–Kramer correction for multiplicity. Changes within and between groups were considered statistically significant when the p-value was less than 0.05. All data are expressed as mean ± SD.

5 Results

Before ischemia, regional shortening in the anterior LV wall was first detected in the endocardial side, corresponding to the Q wave of the EKG and initial slow LVP rise. Contraction at the epicardial side began 82 ± 23 ms after the initial muscle contraction, corresponding to peak positive dP/dt wave, S wave of QRS, and steep LVP rise for systolic ejection. Shortening was then present in both segments during systolic ejection, so that ‘co-contraction’ of both endocardial and epicardial fibers occurred during this cardiac phase. The first region to stop shortening was the endocardial side, coinciding with the onset of the negative dP/dt. The epicardial component continued to shorten for 92 ± 33 ms, coinciding with rapid LVP descent and peak negative dP/dt. Therefore, a hiatus existed, whereby subepicardial contraction continued beyond recognized end of systole defined by the dichrotic notch on the superimposed aortic and left ventricular pressure curves (Fig. 1A) [18,19].

With unprotected ischemia, bulging of the anterior wall persisted throughout the 15 min LAD coronary occlusion (Fig. 1B). All hearts presented at least one episode of ventricular fibrillation, treated by instant 10 Jules defibrillation. A 5 mg Lidocaine bolus was delivered after the first episode to prevent further fibrillation, but five of the six pigs again developed a reversible ventricular fibrillation episode. Hemodynamics are shown in Table 1. Mean arterial and left atrial pressures remained constant during the procedure, but cardiac output decreased by 16.6 ± 20% at 30 min and by 34.4 ± 27% at 60 min. Ejection force (maximum dP/dt) fell 32 ± 20% in the unprotected group when measured 60 min after reperfusion. This hemodynamic change was associated with a rise of left ventricle end diastolic pressure (LVEDP) from 8 ± 2.9 mmHg at baseline to 12.2 ± 7.2 mmHg during ischemia that persisted after 1 h of reperfusion.

Studies on regional contraction during ischemia showed distension between pairs of crystals during systole in both endocardium and epicardium, marking bulging or dyskinesia, as shown in Figs. 1B and 2². Recovery of the initial pre-systolic distance between crystals took the first half of diastole (51.1 ± 16.5%, Figs. 1C and 3³).

Reperfusion produced gradual disappearance of bulging, with regional segment shortening recovery of 19.7 ± 32% of baseline SS% at 15 min, and 30.8 ± 33.5% and 36.4 ± 24.4% after 30 and 60 min of reperfusion, respectively (Figs. 1C and 2). While epicardial contraction timings recovered almost immediately, the reperfusion phase was characterized by a delay in the expected termination of endocardial contraction. This prolongation of endocardial contraction (Figs. 1 and 4⁴) resulted in an endocardial timing sequence that resembled the ascending segment tracing, thereby abolishing the expected normal duration hiatus between cessation of endocardial shortening and ongoing epicardial shortening. This adverse sequential contraction change occurred during the period normally associated with ventricular suction for rapid diastolic filling.

Diastolic dysfunction was evident after unprotected ischemia, as endocardial contraction extended into diastole (Figs. 1c and 3) throughout the 1 h observation period following reperfusion, measuring 37.7 ± 12.1% of the time of diastole at 15 min and 27.9 ± 11.7% and 38.0 ± 16.4% at 30 and 60 min, respectively.

Cariporide pretreatment did not prevent bulging during ischemia or ventricular fibrillation during LAD occlusion. Cardiac output after reperfusion decreased to 22 ± 17% at 30 min and 28 ± 10% at 60 min with no difference with unprotected hearts at similar left atrial pressures (Table 1). Conversely, other variables improved as ejection force (maximum dP/dt) was maintained within baseline values (−10 ± 15% vs −32 ± 20%, although not statistically significant), and mean LVEDP was maintained in baseline levels at 8.6 ± 2.4 mmHg and 7.9 ± 1.7 mmHg at 30 and 60 min, respectively, after ischemia, as each value was p ≪ 0.05 versus unprotected ischemia.

Cariporide improved time related recovery of systole, as SS%, which was 51.6 ± 25% of baseline at 15 min after ischemia, increased to 69.4 ± 14% at 30 min and to 75.8 ± 15% at 60 min, both values p ≪ 0.05 versus unprotected ischemia.

Diastolic dysfunction was nearly reversed, as the percentage of time of early endocardial contraction during the first part of diastole (DCT%, Fig. 3) fell from 30.5 ± 19% at 15 min, 12.9 ± 18% at 30 min, and 4.6 ± 10% at 60 min postischemia, all p ≪ 0.05 versus unprotected ischemia. These diastolic shortening changes paralleled the avoidance of LVEDP changes described under hemodynamic recordings.

Creatine kinase MB in coronary sinus blood raised slightly after ischemia, and at 60 min was 59.3 ± 33% above baseline after unprotected ischemia but only 28.2 ± 6% above baseline after Cariporide.

Conjugated dienes at 30 min after ischemia were 40.2 ± 12% higher than baseline in unprotected ischemia compared with 27.3 ± 4.5% following Cariporide (p ≪ 0.05). CD values were similar at 60 min, at 33.8 ± 16% and 22.2 ± 3.6%, respectively.

Postmortem injection of methylene blue showed that mid LAD occlusion produced an ischemic zone of 26.5 ± 4.2% and 25.5 ± 2.3% of left and right ventricles, respectively, with no statistical differences among the groups.

6 Discussion

The systolic and diastolic effects of ventricular stunning after brief ischemic/reperfusion intervals relate to functional aspects of the geometric configuration of the ventricular myocardial band, and thus show a muscular component of diastolic dysfunction. Evidence of impaired systolic shortening, with prolongation of the endocardial (descending segment component of the ventricular myocardial band) was associated with insufficient relaxation, increased LV end diastolic pressure, and presumably limitation of the rapid filling aspect of ventricular suction. These adverse effects on diastolic dysfunction were reversed by Cariporide, an NHE inhibitor agent. The implications of how a muscular component of diastolic dysfunction can be effectively treated by interfering with the efficiency of calcium management in the stunned myocardium shall be discussed.

Muscular shortening data in the anterior free wall of the unjeopardized left ventricle corroborate the sequential contraction following the path of the ventricular myocardial band, first described by Torrent-Guasp (Fig. 5) [20]. The underlying considerations relate to ventricular structural configuration, whereby a myocardial fiber band divides the basal loop (that wraps around the free wall of the right ventricle, upper septum, and posterior basal portion of the LV) from the apical loop which contains a descending segment (endocardial) and ascending segment that includes epicardial anterior wall, right ventricle outflow track, and right ventricular side septal fibers.

Previous sonomicrometer studies demonstrate the normal contraction sequence that follows the band model, whereby the initial shortening of the basal loop occurs simultaneously with the inner fibers of the descending segment, and is subsequently followed by shortening of the ascending segment [7]. Ejection is caused by full contraction of the entire ventricular mass (‘co-contraction’) causing transmural thickening and strain, while ventricular suction is also a contractile event due to ongoing contraction of the ascending segment, rather than existing during an interval previously known as isovolumetric relaxation [21].

Geometry of the suction mechanism is closely linked to functional aspects of the ventricular band [22,21]. Responsible factors include ongoing shortening in the epicardial segment, which persists despite simultaneous cessation of shortening in the RV, posterior LV, and endocardial segment records. Rapid filling occurs during this timing phase, coincident with the rapid descent of LV pressure and the onset of the negative phase of dP/dt recording, whose negative magnitude is similar to the positive systolic phase. Consequently, the prior concept of isovolumetric relaxation becomes replaced with what is been recently termed late isovolumetric contraction[23] or systolic ventricular filling [22]. Accordingly, rapid filling becomes a ‘systolic’, rather than a ‘diastolic’ event. As a result, prior definitions of systole ending 20 ms before peak negative dP/dt, must now have become changed to systole existing 60 ms after peak negative dP/dt[18,19].

Identification of the possibility of dysfunction during late isovolumetric contraction introduces treatment options related to the sequential activity of the band which can be changed by altering calcium dynamics. This report shows that such functional disruption happens during stunning following transient ischemia and reperfusion, and codifies its pharmacological management.

Sonomicrometer data show that ‘myocardial stunning’ affects the systolic and diastolic phases of the cardiac cycle, only when diastole has been defined as occurring during the dichrotic notch of superimposed aortic and LV pressure tracings. Conversely, crystal tracings show that a late systolic event is responsible for the phase previously called diastolic dysfunction. Systolic dysfunction during ejection is a recognized [1,10] contractile deficit that recovers normal function after up to 24 h of the ischemic insult [1,10]. Although the diastolic component exists [2], explanation of such dysfunction during the phase of ‘isovolumetric relaxation’ is usually ascribed to loss of potential energy [6,24] or elastic recoil factors sometimes related to titin [5]. However, a repetitive high energy state exists during this time frame, since MRI studies define rapid ventricular twisting during this phase of rapid filling in the normal heart, and characterizes the capacity of inotropic agents to accelerate this diastolic phase. Such observations imply requirements for an oxygen-consuming muscular action, rather than ascribing these actions to the passive actions of collagen or titin, which cannot turn over oxygen in rapid order. This study identifies this late time interval as a systolic muscular event, relating to prolonged endocardial or descending segment contraction.

The resultant postischemic lower contractile force that is quantified by sonomicrometer tracings is associated with an excitation–contraction mismatch, characterized by a delay in the start and end of contraction, as shown in Figs. 1 and 4[25]. Dysfunction is most evident in the endocardium that occupies the descending segment of the ventricular myocardial band, a region that is more vulnerable to ischemic and reperfusion damage [26] than the ascending or epicardial component [26].

This dyssynchrony characterized by subsequent extension of endocardial contractile force into diastole may be a principal factor responsible for diastolic dysfunction. The active element may be the alteration of the rapid systolic phase of ventricular filling. The resultant impaired relaxation incurred by dyssyncrony between endocardial and epicardial shortening will also lead to higher ventricular stiffness during early diastole. The consequent compromise of rapid filling will simultaneously derail the mechanisms for suction so that the increased ventricular pressure rather than ventricular untwisting by clockwise muscle motion now becomes the principal filling determinant.

The hemodynamic studies of increased LVEDP during stunning provide direct confirmation of this adverse hemodynamic event showed. This contractile dysfunction may underlie the origin of diastolic dysfunction, that is reversed by NHE blockade, thereby opening the door for consideration of a new spectrum of pharmacological agents that can either limit calcium entry like the Na⁺–H⁺ exchanger inhibitors used in this study, or alter the efficiency of ionic calcium exchange with agents such as levosimendan [27].

This study, demonstrating rapid recovery of diastolic dysfunction, introduces the influence of Na⁺–H⁺ exchanger in myocardial stunning. A direct application of NHE blockage may occur in patients that undergo off-pump coronary bypass, where a 10–15 min period of ischemia is routine. Pretreatment may provide similar results as in our experiments if collateral blood supply is not sufficient, or limited inflow exists despite using an intra coronary shunt in patients with poor inflow through the native stenosis, as well as in on pump procedures where cardioplegic protection is not adequate. Diastolic dysfunction in congestive heart failure with retained systolic function may also be related to brief ischemia/reperfusion events in patients with underlying coronary artery disease [8,9]. However, the present studies of acute stunning differ from clinical studies of chronic ischemia [28].

Aside from evidence for this prolongation of late systole in hearts with diastolic dysfunction following stunning after ischemia [29], this adverse late systolic event that produces ‘diastolic dysfunction’ to curtail rapid ventricular filling has also been reported after hypertrophy during aortic stenosis [30], posttransplant dilatation [31], or tachycardia-induced cardiomyopathy [32]. Each condition may also alter the ventricular geometry as well as calcium dynamics, so that thickening of the muscle mass can exist with hypertrophy. Furthermore, dilation may change the oblique spatial orientation towards a sphere configuration, thereby introducing a more transverse pattern to apical loop fiber architecture that can alter contractile properties [33].

Surgical treatment options may become directed toward alleviating diastolic dysfunction by restoring more normal structural architecture (a) by returning normal ventricular mass by excluding the left ventricular obstructive element to normalize the interaction between descending and ascending segments, or (b) by directly altering ventricular form towards a conical shape to rebuild the helical heart configuration. Each management option addresses re-establishing a more normal anatomic structure, and thereby allowing coincident functional improvement of the relationship between the descending and ascending segments of the apical loop of the helical heart.

7 Conclusions

Contraction in the anterior free wall of the left ventricle is sequential, starting in the endocardium followed by the epicardium. While co-contraction of the whole wall thickness causes ejection, ongoing contraction in the epicardial side (following cessation of endocardial contraction) may be responsible for ventricular suction, thus supporting the new concept of sequential contraction through the myocardial band. Myocardial stunning affects systole by lowering the power of contraction, and also causes diastolic dysfunction by altering the normal endocardial–epicardial coordination to prolong endocardium shortening and interfere with the interaction between descending and ascending segments of the apical loop responsible for suction. Pretreatment by NHE inhibition limits systolic and avoids diastolic dysfunction, probably introducing a novel pharmacologic means of modifying the ionic aspects of calcium flux to treat diastolic dysfunction.

References