INTRODUCTION

Top Abstract Introduction Materials Results Discussion References

PREVIOUS STUDIES have documented a presence of impaired myocardial relaxation in the hypertrophied left ventricle (18, 25). This abnormality of left ventricular (LV) diastolic function has recently been suggested to be related to the enhanced activity of the renin-angiotensin system, in particular at the local cardiac level (10, 23, 31). Angiotensin II is known to promote influx of calcium and increase its intracellular level via type 1 (AT₁) receptors in cardiomyocytes (1, 8, 19). This effect of angiotensin II is, however, potentially detrimental for the calcium homeostasis in the hypertrophied myocardium, where intrinsic abnormalities of sarcoplasmic calcium reuptake already exist (5, 14), and may aggravate an abnormality of inactivation process during myocardial relaxation. Indeed, it has been reported that angiotensin II impairs myocardial relaxation and elevates diastolic pressure (31) and that these effects are prevented by an angiotensin-converting enzyme (ACE) inhibitor (10) and also by an AT₁-receptor antagonist (22) in experimental LV hypertrophy (LVH). A recent clinical study also reported an improvement of LV diastolic function by enalaprilat infusion in patients with aortic stenosis (11).

The presence of the impaired myocardial inactivation can increase sensitivity of LV relaxation to loading conditions in the diseased hearts (4). It has been suggested that imbalance between the reduced LV load and the impaired myocardial inactivation leads to further exacerbation of ventricular relaxation abnormality in pressure-overload LVH (25). In this setting, it can be hypothesized that angiotensin II may affect this load-sensitive LV relaxation by modifying the intracellular calcium handling and the myocardial inactivation process.

To test this hypothesis, we studied the effects of acute ACE inhibition and AT₁-receptor blockade on LV relaxation in a canine model of pressure-overload LVH due to perinephritic hypertension (7, 16).

	METHODS AND MATERIALS

Top Abstract Introduction Materials Results Discussion References

Study group. Sixteen dogs underwent surgery for having one kidney wrapped, and they subsequently developed perinephritic hypertension and LVH. Eleven dogs without kidney wrapping served as normal controls. There was no difference in body weight between the control and LVH groups (21 ± 1 vs. 21 ± 2 kg).

Six control dogs (CTL-E group) and eight LVH dogs (LVH-E group) received 0.25 mg/kg iv of enalaprilat (Renitec, Merck Sharp & Dohme, Brussels, Belgium). The other five control dogs (CTL-L group) and eight LVH dogs (LVH-L) received 0.30 mg/kg iv of a nonpeptide AT₁-receptor-selective antagonist, L-158,809 {5,7-dimethyl-2-ethyl-3-[[2'-(1H-tetrazoil-5-yl)[1,1'-biphenyl]-4-yl]methyl]-3H-imidazo[4,5-]pyridine, Merck Research Laboratories, Blue Bell, PA}. Our previous study (16) has shown that these doses of enalaprilat and L-158,809 achieved the maximal effects on systemic blood pressure and also significantly improved LV diastolic stiffness in the LVH dogs.

The study was in accordance with the "Guiding Principles in the Care and Use of Animals" approved by the Council of the American Physiological Society and also with the guidelines by the Committee for Animal Research of the Belgian Fonds National de la Recherche Scientifique Médicale.

Experimental procedures. Surgical procedure for the preparation of perinephretic hypertension has been described elsewhere (7, 16). In brief, while the dog was under general anesthesia with intravenous thiopental sodium (5 mg/kg) and enflurane gas (2-3 vol%), we exposed the left kidney and wrapped silk around its surface under sterile conditions. Care was taken to avoid inadvertent stenosis of the renal artery. Then the kidney was returned to its initial position and the skin incision was closed. The dogs were allowed to recover and develop hypertension during the following 6 wk.

At the time of the experiment, all dogs had recovered from the surgery. Symptoms of heart failure and infection were absent, and plasma creatinine and urea nitrogen levels were within normal ranges in all dogs. The dog was intubated, and ventilation was maintained under general anesthesia with intravenous pentobarbital sodium (20 mg/kg). The chest was opened at the left fifth intercostal space, the pericardial sac was incised, and the heart was exposed in the pericardial cradle. A high-fidelity micromanometer (Serel JSI 0400, Alleur, Belgium) was introduced into the LV cavity through the apex. A polyethylene band (5 mm in width) was placed around the inferior vena cava just above the diaphragm for vena caval occlusion. One pair of 5-MHz ultrasonic crystals (Triton Technology, San Diego, CA) was implanted to measure anteroposterior external (epicardial) diameter at midventricular level. Another pair of crystals was implanted to measure wall thickness in the anterior wall of the midventricle, adjacent to the anterior crystal for the diameter (16). An 8-Fr USCI catheter, filled with heparinized saline, was advanced to the descending thoracic aorta via the right femoral artery and connected to a Gould P23 ID transducer with zero reference at the level of the left ventricle. Another Tygon catheter was inserted into the right femoral vein for volume infusion.

After the baseline hemodynamic data were obtained, 100 ml of saline warmed at 37°C were slowly injected into the femoral vein (16). When the increases in LV diastolic pressure and dimensions reached the plateau levels, the inferior vena caval occlusion was performed until the stable minimal dimensions were observed, which was accomplished within 10-20 s in all dogs. LV peak systolic pressure was decreased from 128 ± 7 (SE) to 89 ± 6 mmHg in the control group and from 172 ± 7 to 104 ± 7 mmHg in the LVH group. Then the occlusion was released. During the vena caval occlusion, respiration was suspended at end expiration to avoid its influence on LV pressure and dimensions.

When the hemodynamic conditions were judged to have returned to the preocclusion basal levels, enalaprilat or L-158,809 was slowly (over 5 min) injected into the inferior vena cava. After the administration of the drug, the second set of steady-state hemodynamic data was obtained, and the volume infusion and the subsequent vena caval occlusion were repeated in the same manner.

Hemodynamic data analysis. Analog hemodynamic signals were digitized and processed by a Hewlett-Packard A900 computer (16, 17). A first derivative of LV pressure (dP/dt) was obtained by digital differentiation of the pressure data. The end diastole was defined at the peak of R wave on the electrocardiogram, and the end systole was defined at the time of minimum dP/dt (dP/dt_min). LV anteroposterior internal diameter was derived by the external diameter minus (2 × wall thickness). Fractional shortening of the internal diameter was obtained by end-diastolic dimension minus end-systolic dimension divided by end-diastolic dimension.

LV circumferential wall stress was calculated assuming a cylindrical geometry (13):  = PD/2h, where  is wall stress, P is pressure, D is internal diameter, and h is wall thickness. End-systolic elastance of the left ventricle was assessed by linear regression analysis of wall stress () and internal diameter (D'_es) at the time of the maximal elastance during vena caval occlusion (27):  = E_es (D'_es  D_d), where the slope E_es is the end-systolic elastance, and D_d is the diameter axis intercept. To avoid the influence of reflex change, the data for and D'_es over the first 6-8 s from the beginning of pressure decrease were used for the analysis (30). The correlation coefficients for this relationship were 0.951 ± 0.019 in the control group and 0.963 ± 0.010 in the LVH group.

Assessment of LV relaxation rate. LV relaxation rate was assessed by a time constant during the first 80 ms after dP/dt_min. The relaxation time constant (T) was derived from the regression of dP/dt versus LV pressure (28): dP/dt = (1/T) · (P  P_B), where P is LV pressure, and P_B is the variable pressure asymptote. The correlation coefficients for T calculation were 0.993 ± 0.001 in the control group and 0.984 ± 0.004 in the LVH group.

When ACE inhibitors or angiotensin II antagonists are given intravenously, it is difficult to separate their direct cardiac effects from their systemic vascular effects on loading conditions, which also influence LV relaxation rate (4). Therefore, we compared T for cardiac beats at the matched levels of loading conditions before and after the infusion of the drug. The changes of LV end-systolic wall stress during vena caval occlusion were compared before and after the drug infusion, and the overlapped range of end-systolic wall stress was determined for each dog. LV hemodynamics were studied for the beats at the highest and lowest matched levels of end-systolic wall stress before and after the drug infusion. For these beats at either level of end-systolic wall stress, LV end-diastolic wall stress was not significantly different before and after the drug infusion (see Tables 3 and 4). Therefore, LV relaxation was considered to be studied over the comparable range of loading conditions before and after the drug administration. The beats analyzed were in normal sinus rhythm without bundle branch block or postextrasystolic potentiation.

Statistical analysis. Data are presented as means ± SE. Student's unpaired t-tests were used for the comparison of the baseline hemodynamics between the pooled control and LVH groups (Table 1). Two-way analysis of variance and Bonferroni-Dunnett tests for the multiple comparisons were used for the asessment of the hemodynamic data during vena caval occlusion. Differences were considered significant if the probability value was <0.05.

	RESULTS

Top Abstract Introduction Materials Results Discussion References

Effects of hypertension on heart weight. The LV-to-body weight ratio and LV relative wall thickness (a ratio of end-diastolic wall thickness to internal diameter) were significantly greater in the LVH group than those in the control group (Table 1). The right ventricular-to-body weight ratio was not different between the two groups.

LV hemodynamic data. Table 1 also summarizes the baseline LV hemodynamics in the pooled control (n = 11) and LVH (n = 16) groups. Compared with the control group, LV systolic and end-diastolic pressures were significantly higher, and LV dP/dt_max and the slope for the end-systolic stress-diameter relationship (E_es, end-systolic elastance) were also greater in the LVH group. No significant difference was observed for the fractional shortening between the two groups. LV relaxation time constant T was significantly prolonged and P_B was smaller in the LVH group, suggesting an impairment of LV diastolic relaxation.

Enalaprilat and L-158,809 caused similar decreases in LV systolic pressure in the control groups as well as in the LVH groups (Table 2). E_es or dP/dt_max was not affected by enalaprilat or L-158,809 in either the control or LVH group (Fig. 1). No significant change was observed for LV diameter or fractional shortening.

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Examples for left ventricular (LV) wall stress-internal diameter loops during inferior vena caval occlusion at baseline (A) and after enalaprilat infusion (B) in a dog with hypertensive LV hypertrophy. Note that no substantial change is observed for slope of end-systolic stress-diameter relationship (dotted line superimposed on the loop) before and after enalaprilat infusion.

LV relaxation and its load sensitivity. In the control group, T did not significantly change during load reduction by vena caval occlusion. In the LVH groups, however, T became significantly prolonged during vena caval occlusion, indicating that LV relaxation was deteriorated in a load-sensitive manner. Figure 2 shows examples for beat-to-beat phase-plane plots of LV pressure and dP/dt during vena caval occlusion. In the control dog (Fig. 2A), the slope of dP/dt-pressure relationship during the isovolumic relaxation phase was substantially unchanged. In the LVH dog (Fig. 2B), however, the relaxation phase dP/dt-pressure relationship became less steep when LV load was manipulated from high (upward arrows) to low (downward arrowheads) levels, reflecting a presence of load-sensitive LV relaxation.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Examples for beat-to-beat LV pressure-first derivative of LV pressure (dP/dt) phase-plane plots in a control dog (A) and a dog with LV hypertrophy (LVH) (B). In control dog, slope of LV pressure-dP/dt relationship during relaxation phase was substantially unchanged when load was changed from high (upward arrows) to low (downward arrowheads) levels by vena caval occlusion. However, in LVH dog, relaxation phase pressure-dP/dt relationship became less steep when LV load was changed from high (upward arrows) to low (downward arrowheads) levels, indicating presence of load-sensitive deterioration of LV relaxation.

After enalaprilat or L-158,809 was administered, T remained unaltered during vena caval occlusion in the control group (Tables 3 and 4), and there was no significant difference in T before and after the drug infusion at either level of LV load. In the LVH group, there was no significant difference in T at the matched high load before and after infusion of enalaprilat or L-158,809. However, the prolongation of T during vena caval occlusion was attenuated, and, as a consequence, T at the matched low load was significantly improved from 52 ± 5 to 44 ± 3 ms in the LVH-E group and from 52 ± 5 to 43 ± 4 ms in the LVH-L group (both P < 0.05 before and after the drugs). Figure 3 shows a representative example for phase-plane plots of LV pressure and dP/dt for the beats at the matched high and low LV load in a dog from the LVH-E group. The pressure-dP/dt relationship during the relaxation phase became steeper and T was improved by enalaprilat infusion for the beat at the low LV load.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Effect of enalaprilat infusion on LV pressure-dP/dt phase-plane plots in a dog with hypertensive LVH. Representative two beats at high and low LV load were selected from data during vena caval occlusion before and after enalaprilat infusion. At baseline (A), pressure-dP/dt relationship during isovolumic relaxation period became less steep by load reduction and relaxation time constant (T) was markedly prolonged from 45 to 73 ms. After enalaprilat infusion (B), pressure-dP/dt relationship for the beat at low load became steeper, which was accompanied by an improvement in T and magnitude of dP/dtmin without changes in LV peak and end-systolic pressures.

	DISCUSSION

Top Abstract Introduction Materials Results Discussion References

The present study showed that LV relaxation was impaired in dogs with hypertensive LVH even though systolic function was preserved (Table 1, Fig. 1). This result is consistent with a previous study (32) reporting that LV inotropic state was enhanced in the LVH dogs, whereas LV relaxation rate was impaired. Although the enhanced inotropic state is usually accompanied by the accelerated relaxation in the normal left ventricle (4), the dissociation between inotropic state and relaxation is often observed in the hypertrophied left ventricle. For instance, postextrasystolic LV contraction potentiates the inotropic state by increasing calcium influx into the myocardium but rather impairs ventricular relaxation in patients with pressure-overload LVH (26). Similar dissociation between baseline LV inotropic state and relaxation in our LVH dogs (Table 1) may suggest the presence of imbalance between calcium influx and sarcoplasmic reuptake, which affects the inactivation process of myocardium.

Another major determinant of LV relaxation rate is the loading condition (4). LV relaxation is little affected (33) or accelerated (12) by load reduction in the normal hearts. Interestingly, previous studies have suggested that myocardial relaxation in the heart with pressure-overload LVH is more sensitive to loading conditions (21) and may be significantly deteriorated by extensive load reduction. It has been reported that LV time constant was further prolonged after extensive load reduction by aortic valvuloplasty and subsequent nitroprusside infusion in patients with aortic stenosis (25). Another study (6) also observed a similar change in LV relaxation, albeit to a lesser extent, during moderate unloading by nitroprusside infusion alone. In addition, we previously reported that LV regional myocardial relaxation rate in hypertrophic nonobstructive cardiomyopathy was more severely impaired at the ventricular wall with pronounced hypertrophy, where regional wall stress is expected to be excessively reduced below the normal range (18). The results of the present study confirm this paradoxical load-sensitive LV relaxation in pressure-overload LVH (Tables 3 and 4, Figs. 2 and 3). When LV load was extensively reduced during vena caval occlusion, the relaxation time constant was remarkably prolonged in the LVH groups. In contrast, the time constant was insensitive to the changes in loading conditions in the control group over the comparable range of afterload and preload (Tables 3 and 4).

The underlying mechanism for such load-sensitive relaxation in LVH seems to be at least partly explained by the imbalance between the external load and the intrinsic myocardial inactivation process (4, 25). Excessive reduction of LV load will result in an earlier onset of isometric relaxation, whereas the sarcoplasmic calcium reuptaking process remains impaired in the hypertrophied myocardium. Under such conditions, LV relaxation rate is more dependent on the kinetics of the sarcoplasmic calcium sequestering system rather than on the mechanical detachment of myofilament, which leads to further deterioration of LV relaxation. This concept is also supported by another previous study (15) which shows that an increase in afterload improved LV relaxation and early diastolic filling in hypertrophic nonobstructive cardiomyopathy.

Interestingly, administration of enalaprilat or L-158,809 significantly improved the load-dependent LV relaxation in the LVH group (Tables 3 and 4, Fig. 3) without any substantial changes in LV inotropic state. This finding tempted us to consider that angiotensin II directly impairs sarcoplasmic calcium removal and myocardial inactivation at cellular levels. As previously suggested, the cardiac renin-angiotensin system (9, 23) is highly activated in pressure-overload LVH even when the circulating renin-angiotensin activity returns to a normal level (31). It has been shown that cardiac ACE is diffusely distributed within the myocardium, and its density is significantly higher in the left ventricle of rats with pressure-overload LVH (31). We have also observed that enalaprilat and L-158,809 were still effective to improve LV diastolic stiffness even in LVH dogs with normal plasma angiotensin II level (16), suggesting a potential role of local cardiac angiotensin II in the pathogenesis of LV diastolic dysfunction. The locally produced angiotensin II elevates intracellular calcium levels via AT₁ receptor and phosphoinositide second-messenger signaling pathway (1, 8) and also increases myofilament calcium sensitivity by stimulating Na⁺/H⁺ antiport (19). In the normal myocardium, these effects of angiotensin II mediate positive inotropic and lusitropic effects (1). In the hypertrophied myocardium, however, these effects can be rather detrimental for intracellular calcium reuptake and myocardial relaxation because of the coexistence of the impaired sarcoplasmic reticulum calcium sequestration (5, 14).

Another possible mechanism for the improvement of the load-sensitive LV relaxation may be related to the modification of regional nonuniformity by enalaprilat and L-158,809 (4). Nonuniformity of LV regional wall dynamics is commonly observed in disease conditions such as myocardial ischemia (17) and hypertrophy (18) and has been suggested to impair LV relaxation rate. Some previous studies observed a nonuniform distribution of cardiac ACE and/or angiotensin II receptors at different regions of the left ventricle (31, 34), and the nonuniform effects of angiotensin II on regional myocardium may induce asynergy and/or asynchrony of regional wall dynamics.

Enalaprilat and L-158,809 exerted similar systemic vasodilatory effects in the control and LVH groups (Table 2). This reduction in afterload could contribute to improvement in LV relaxation (12). In the present study, however, LV relaxation was compared at the matched levels of loading conditions (Tables 3 and 4), and therefore the improvement of LV relaxation cannot be entirely explained by the reduction in the systemic load. Yet, it is still possible that vascular effects of enalaprilat and L-158,809 may have changed the input impedance of resistance vessels and its reflection flow wave (24), thereby altering LV relaxation by affecting the loading sequence on the left ventricle (20).

In the hypertrophied heart, hemodynamic stress is known to induce myocardial demand ischemia because of impaired coronary vasodilatory reserve due to structural and functional alterations of coronary circulation (2). Administration of ACE inhibitor and AT₁-receptor antagonist may improve the coronary blood flow, thereby improving LV diastolic function. However, the baseline myocardial blood flow was shown to be preserved to a similar level in the control and LVH dogs by our previous study (16) as well as by other investigators (13), and an effect of short vena caval occlusion on myocardial perfusion is considered to be negligible (29).

In summary, the present data support the concept that angiotensin II mainly affects diastolic function in this model of pressure-overload LVH and that these deleterious effects on diastolic LV relaxation are ameliorated by ACE inhibition and AT₁-receptor blockade. By contrast, in the control group, LV relaxation rate remained unchanged after enalaprilat or L-158,809 infusion, which may suggest the absence of substantial effects of angiotensin II on diastolic function in the normal canine heart. Thus this study provides new insights for understanding the mechanisms of diastolic dysfunction in the hypertrophied heart.