Preserved Frank–Starling mechanism in human end stage heart failure

Abstract

Objective: The goal of the present study was to examine the ability of failing myocardium to respond to enhanced preload with an increase in force development. Methods: The effect of various preload conditions (2.5–15 mN) on force development was studied in right ventricular trabeculae carneae from explanted human failing hearts with ischemic cardiomyopathy (ICM, n=5, 42 preparations) or idiopathic dilated cardiomyopathy (DCM, n=9, 77 preparations). To determine the severity of cardiac impairment we measured the positive inotropic effect of β-adrenoceptor stimulation and calcium (ISO/Ca²⁺ ratio) and the expression of atrial natriuretic peptide (ANP) mRNA in all hearts. Results: (1) Force of contraction increased with stepwise augmentation of preload (length at 2.5 mN preload to length of maximal force development) from 3.7±0.5 (ICM) and 2.7±0.4 (DCM) to 8.3±0.9 and 6.5±0.8 mN/mm², respectively (p<0.05). (2) The ISO/Ca²⁺ ratio was 0.40±0.04 (ICM) and 0.35±0.03 (DCM), respectively. (3) ANP mRNA was expressed in all preparations, albeit at greatly varying levels (ICM 22.5±6.1 and DCM 18.7±4.7 normalized arbitrary units). (4) Contraction experiments performed in left ventricular tissue (n=3, 32 preparations) essentially confirmed the results. Conclusion: The Frank–Starling mechanism is preserved in terminally failing human hearts irrespective of the underlying etiology. We found no relation between the severity of cardiac impairment as assessed by either ANP expression or the ISO/Ca²⁺ ratio and the ability of failing human myocardium to respond to enhanced preload with an increase in force development.

Human ventricle, Heart failure, Contractility, Atrial natriuretic factor

Time for primary review 23 days.

1 Introduction

The Frank–Starling mechanism represents the intrinsic capability of the heart to respond to enhanced preload with an increase in force development [1, 2]. This fundamental property of the heart reflects the length–active tension relation of striated muscles, i.e. that force development depends on initial rest length of sarcomeres [3]. Besides the Frank–Starling mechanism, cardiac output can be regulated by the autonomic nervous system and humoral factors [4, 5]and intracardiac autoregulatory mechanisms [6, 7].

In congestive heart failure, cardiac output and its regulation are compromised. Neurohumoral mechanisms to increase cardiac contractility are attenuated [8, 9], diastolic relaxation and calcium transients are prolonged [10], and the force–frequency relation is inverse [11]. Furthermore, alterations of the normal cardiac ultrastructure [12]have raised the hypothesis that the ability of the heart to utilize the Frank–Starling mechanism is attenuated as well [13].

Previous in vitro studies on human heart preparations have yielded conflicting results. Whereas one group reported the Frank–Starling mechanism in left ventricular preparations of failing human hearts to be abolished [14], others showed that the length–tension relationship is preserved, both in isolated, blood-perfused whole left ventricles and in muscle preparations from left ventricular myocardium [15]. Since the loss of preload-dependent increase in force generation would have important therapeutic consequences in the treatment of heart failure, the present study was designed to contribute to the question if the Frank–Starling mechanism is absent or not in ventricular muscle preparations of failing human hearts.

2 Methods

2.1 Myocardial tissue

Cardiac tissue was obtained at the time of transplantation from patients with end stage heart failure (New York Heart Association Class IV). The clinical and hemodynamic data are summarized in Table 1. In five cases the diagnosis was ischemic cardiomyopathy (ICM) and in nine cases idiopathic dilated cardiomyopathy (DCM). In addition, left ventricular tissue was obtained from one ICM and two DCM hearts. All patients gave written informed consent prior to surgery. The investigation conforms with the principles outlined in the Declaration of Helsinki and was approved by the local Ethics Committee. Patients receiving β-adrenoceptor blockers, calcium antagonists or catecholamines shortly before surgery were excluded from the study.

Table 1

Clinical and hemodynamical characteristics before surgery

Characteristic	ICM (n=5)	DCM (n=9)
Age (yr)	59.4±0.4*	48.3±4.1
Sex (m/f)	4/1	7/2
EF (%)	19.4±2.5	25.0±3.9
CI (ml/min/m2)	2.3±0.3	2.0±0.2
RVEDP (mmHg)	10.6±1.6	10.3±+2.4
Medication (number of pts)
Digitalis	4/5	7/9
ACE inhibitor	5/5	7/9
Diuretics	5/5	9/9
Nitrates	4/5	8/9

m/f=male/female; EF=ejection fraction; CI=cardiac index; RVEDP=right ventricular enddiastolic pressure and resting conditions; pts=patients. *p<0.05 vs. DCM.

Gross examination of the hearts revealed hypertrophy and dilatation of both ventricles. Preparations were placed into aerated, ice-cold bathing solution (modified Tyrode's solution containing (mmol/l): NaCl 119.8, KCl 5.4, CaCl2 1.8, MgCl₂ 1.05, Na₂HPO₄ 0.42, NaHCO₃ 22.6, Na₂EDTA 0.05, ascorbic acid 0.28, glucose 5.0, continuously gassed with 95% O₂+5% CO₂, pH 7.4) and transferred within 0.2–1.5 h to the laboratory. Contraction experiments were started immediately thereafter. Samples of right ventricular myocardium were taken shortly after explantation, frozen in liquid nitrogen and stored at −80°C until further processing.

2.2 Contraction experiments

Contraction experiments were performed on isolated, electrically driven right ventricular trabeculae carneae (ICM, n=5 hearts, 42 muscle preparations, DCM, n=9 hearts, 77 muscle preparations) and left ventricular trabeculae carneae (ICM, n=1 heart, 10 muscle preparations, DCM, n=2 hearts, 22 muscle preparations) as described previously [16]. In brief, isolated trabeculae (8–10 trabeculae from each heart) were suspended individually in 10 ml glass tissue chambers in Tyrode's solution (composition see above), continuously gassed with 95% O₂ and 5% CO₂ at 35°C. Muscle preparations were electrically paced at 0.5 Hz with rectangular pulses of 5 ms duration (field stimulation, Grass stimulator SD9, Grass, Quincy, MA, USA); the voltage was approximately 10–20% above the threshold. All preparations were allowed to equilibrate in drug-free bathing solution until complete mechanical stabilization (at least 90 min), the bathing solution was changed after 45 min. The isometric force of contraction was measured with an inductive force transducer (Scientific Instruments, Heidelberg, FRG) attached to a WeKagraph recorder (Föhr Medical, Egelsbach, FRG). The force of contraction is given in millinewton (mN) or millinewton per square millimeter (mN/mm²). After steady-state conditions were obtained, the muscle preparations were carefully and stepwise (2.5 mN) stretched from slack length (L₀) to the length at which maximal force was developed (L_max). To determine the length–tension relationship the resting tension (preload) and developed force were measured 5 min after the last stretch. After the experiments cross-sectional area (mm²) was calculated by determining the muscle diameter. To test the β-adrenoceptor-dependent contractile reserve and the mechanical performance of the muscle preparations the positive inotropic response of a maximal effective concentration of isoprenaline (Sigma, St. Louis, USA) and ionic calcium (14.4 mmol/l) was measured at the end of the experiments. Muscles in which the positive inotropic effect of calcium was less than 100% of baseline values were excluded from the evaluation. The ratio of the maximal positive inotropic effects (change in force of contraction compared to baseline) of isoprenaline to that of calcium (ISO/Ca²⁺ ratio) was used as an indicator of myocardial performance, i.e. the severity of heart failure.

2.3 Preparation of RNA and Northern blot analysis

Total RNA was extracted with the commercially available kit RNAzol™ (Biotecx Laboratories, Inc., Houston, USA) according to the manufacturer's instructions. In short: 100–150 mg of frozen right ventricular myocardium were thawed in 600 μl of RNAzol™ solution and homogenized by a Polytron® (Kinematica AG LITTAU, Switzerland), followed by phenol–chloroform extraction, isopropanol precipitation, and ethanol (75%) washing of precipitated RNA. RNA was solubilized in sterile and pyrogene free water. The concentration was determined photometrically at 260 nm. RNA was stored at −80°C. RNA blotting, cDNA labelling, hybridization and quantification were performed essentially as described previously [17]. 15 μg of total RNA from right ventricular myocardium from each heart were separated by electrophoresis on 1% agarose–formaldehyde gels and transferred to nylon membranes. Plasmid with the cDNA insert of human atrial natriuretic peptide (ANP) was a kind gift from Dr. M. Böhm, Cologne, Germany. A PstI fragment (677 bp) was used for Northern blot hybridization. To correct measurements for minor loading differences all membranes were rehybridized with -labelled Gsα (α-subunit of the stimulatory G-protein) which has been shown to be unchanged in human heart failure [18]. The blot was washed at a final stringency of 0.2×SSC/0.1% SDS at 65°C, exposed on imaging plates (BAS-IP MP 2040 P, Fuji, Japan) for 24 h or X-ray films for three days, and scanned by a phosphoimager (BAS 2000, Fuji, Japan). Hybridization signals were quantified using TINA 2.0 (Raytest, FRG) and Zerodescan (CSP Inc., USA).

2.4 Statistics

All values presented are arithmetic means±SEM. Statistical significance was estimated using the Student's t-test for paired or unpaired observations. A p value of less than 0.05 was considered to be significant.

3 Results

3.1 Clinical and hemodynamic characteristics

Table 1 illustrates the clinical and hemodynamic characteristics of the patients studied. Patients with ICM were older than patients with DCM (59 vs. 48 years, p<0.05). In both groups the right ventricular enddiastolic pressure was elevated under resting conditions (normal range: 0–8 mmHg) suggesting a compromised function of the right ventricle (Table 1). Other parameters including medication prior to surgery did not differ significantly between the two groups.

3.2 Contraction experiments

A typical recording of an isolated right ventricular muscle preparation from a patient with ICM is shown in Fig. 1. Stepwise increasing the preload resulted in a parallel increase in force of contraction. Fig. 2A (ICM) and 2C (DCM) summarize the increase in force of contraction of muscle preparation from individual hearts in % of 2.5 mN preload (L_2.5). Fig. 2B (ICM) and 2D (DCM) illustrate the mean developed force of all hearts at increasing preload. Compared to L_2.5 twitch amplitude increased by 124% (ICM) and 141% (DCM) at L_max. Previous studies on force of contraction in human ventricular myocardium using trabeculae carneae and similar experimental conditions [19]showed comparable contraction parameters at L_max (6.6±1.1 vs. 7.4±0.9 mN/mm² in this study). To further characterize the severity of heart failure the ratio of the maximal positive inotropic effect of isoprenaline (ISO) and calcium (Ca²⁺) was determined (ISO/Ca²⁺ ratio, Table 2). According to previous studies on muscle preparations or cells from human hearts this value is approximately 1 in non-failing hearts as compared to <0.5 in failing myocardium [20, 21]. In this study the resulting ISO/Ca²⁺ ratio was 0.40 (ICM) and 0.35 (DCM), respectively, and did not correlate with the maximal difference between the developed force of contraction at 2.5 mN and L_max (Fig. 5A). To characterize the quality of the muscle preparations we calculated the active-to-passive tension ratio at L_2.5 and L_max (Table 2).

Fig. 5

Relationship between the ISO/Ca²⁺ ratio (A) and ANP mRNA expression (B) in right ventricular myocardium and the maximal difference between the developed force of contraction at 2.5 mN and L_max (ΔmN/mm²) in failing hearts with ICM and DCM. FOC=force of contraction.

Fig. 2

Developed force (twitch amplitude) under various preload conditions. A (ICM) and C (DCM) show the increase in force of contraction of right ventricular muscle preparations from individual hearts in percent of force at 2.5 mN preload (every curve represents the pooled data of all muscle preparations of one heart). B (ICM) and D (DCM) illustrate the mean force of contraction at various preload conditions of all hearts in mN/mm². FOC=force of contraction at maximal preload (L_max) and 2.5 mN preload (L_2.5).

Fig. 1

Force measurement in an isolated right ventricular muscle preparation of a failing human heart (ICM). The muscle preparation was stretched from 2.5 mN to L_max. Increasing the preload stepwise resulted in a parallel increase in force of contraction. Arrows indicate changes in preload by approximately 2.5 mN.

Table 2

Contraction parameters at 2.5 mN (L_2.5) and L_max and inotropic response to a maximal effective concentration of calcium (14.4 mmol/l) and isoprenaline (0.1 mmol/l) in right ventricular muscle preparations. Values are increases in twitch amplitude over basal values at L_max and 0.5 Hz in ΔmN/mm²

	ICM (n=5)		DCM (n=9)
	L2.5	Lmax	L2.5	Lmax
Cross sectional area (mm2)	–	1.2±0.2	–	1.3±0.1
FOC (mN/mm2)	3.7±0.5	8.3±0.9	2.7±0.4	6.5±0.8
Max. PIE Ca2+ (ΔmN/mm2)	–	12.6±1.6	–	8.8±0.6
Max. PIE ISO (ΔmN/mm2)	–	6.5±0.9	–	2.8±0.5
ISO/Ca2+	–	0.40±0.04	–	0.35±0.03
Active-to-passive tension	1.8±0.3	0.7±0.1	1.4±0.1	0.9±0.1

FOC=force of contraction, PIE=positive inotropic effect, ISO=isoprenaline.

Since the left ventricle is expected to be the most affected in chronic heart failure we performed additional experiments in left ventricular tissue using the same protocol as described above. Compared to L_2.5 twitch amplitude increased by 206% at L_max (Fig. 3A and 3B). The contraction parameters, as well as the active-to-passive tension ratio are summarized in Table 3.

Fig. 3

Developed force (twitch amplitude) under various preload conditions. A shows the increase in force of contraction of left ventricular muscle preparations from individual hearts in percent of force at 2.5 mN preload (every curve represents the pooled data of all muscle preparations of one heart). B illustrates the mean force of contraction at various preload conditions of all hearts (n=32 muscle preparations) in mN/mm². FOC = force of contraction at maximal preload (L_max) and 2.5 mN preload (L_2.5).

Table 3

Contraction parameters at 2.5 mN (L_2.5) and L_max and inotropic response to a maximal effective concentration of calcium (14.4 mmol/l) and isoprenaline (0.1 mmol/l) in left ventricular muscle preparations. Values are increases in twitch amplitude over basal values at L_max and 0.5 Hz in ΔmN/mm²

	(DCM n=2; ICM n=1)
	L2.5	Lmax
Cross sectional area (mm2)	–	1.1±0.1
FOC (mN/mm2)	1.4±0.9	4.4±0.8
Max. PIE Ca2+(ΔmN/mm2)	–	12.2±2.2
Max. PIE ISO (ΔmN/mm2)	–	7.3±1.6
ISO/Ca2+	–	0.61±0.1
Active-to-passive tension	0.9±0.3	0.2±0.05

FOC=force of contraction, PIE=positive inotropic effect, ISO=isoprenaline.

3.3 Northern blot analysis

Since the expression of atrial natriuretic peptide (ANP) in ventricular myocardium is almost exclusively restricted to failing hearts and ANP mRNA levels were shown to correlate with the clinical severity of heart failure [22], we determined ANP mRNA levels in right ventricular myocardium of all hearts studied (due to technical reasons the ANP level of one preparation (DCM) was not determined). ANP mRNA was present in all hearts of the ICM and DCM group (Fig. 4A). However, ANP mRNA levels varied up to a factor of 30 in individual hearts. These high interindividual differences have been noted earlier [23]. There was no statistical difference in ANP mRNA expression between myocardium from ICM (22.5±6.1 arbitrary units) or DCM (18.2±4.7 arbitrary units, Fig. 4B). As shown in Fig. 5B ANP mRNA levels did not correlate with the developed force of contraction under maximal preload conditions.

Fig. 4

Northern blot analysis of ANP (atrial natriuretic peptide) mRNA and Gsα (α-subunit of the stimulatory G protein) mRNA (internal standard) expression in 15 μg total RNA from right ventricular myocardium of ICM and DCM hearts. (A) Autoradiograph of a representative Northern blot. (B) ANP mRNA levels normalized to Gsα mRNA in ICM and DCM hearts. The numbers of hearts are shown in the columns.

4 Discussion

This study was designed to evaluate the intrinsic capability of the terminally failing human heart to respond to enhanced preload with an increase in force development. Our data demonstrate that under in vitro conditions the Frank–Starling mechanism is preserved in right and left ventricular muscle preparations of failing human hearts, irrespective of the underlying cause of heart failure (ICM/DCM). It is important to note that the Frank–Starling mechanism did not correlate with the severity of myocardial dysfunction, as assessed by the decreased ISO/Ca²⁺ ratio and the expression of ANP mRNA.

The substantial role of preload in maintaining the cardiac output has been demonstrated in isolated myocardial preparations and in anesthetized animals [24–26]. During the development of congestive heart failure cardiac dilatation occurs, frequently accompanied by changes of the normal myocardial ultrastructure [12, 27]. In vivo studies suggested that the failing heart, when subjected to acute volume load, is incapable to further augment stroke volume [26]. This indicates that the Frank–Starling mechanism in vivo is exhausted, i.e. that the failing heart operates at or close to the maximum of the Frank–Starling curve [28]. Recent data implied that the failing human myocardium itself is unable to use the Frank–Starling mechanism [14]. In contrast to non-failing myocardium, left papillary muscle strips from failing human hearts did not respond to enhanced preload with an increase in force development. This finding was interpreted as a failure of the myofibrils to increase the Ca²⁺ sensitivity with an increase of the sarcomere length, one of the mechanisms that has been proposed as an explanation of the Frank–Starling mechanism [29]. However, these data are in contrast to others showing the full effectiveness of the Frank–Starling mechanism in whole heart preparations and left ventricular muscle strips from failing human hearts [15]. Possible explanations for this discrepancy could have been differences in the quality of isolated muscle preparations. The calculated active-to-passive tension ratio for muscle preparations from failing hearts was significantly lower in the study of Schwinger and coworkers compared to that of the other study (0.3 vs. 1.2) and may explain the observed difference.

We decided to use trabeculae carneae since they integrate possible paracrine effects of the endocardium [30]and may reflect the physiological conditions in situ better than muscle strips. With the preparation of muscle strips the endocardium and fibrotic areas are dissected away. Accumulation of extracellular matrix material and endomyocardial fibrosis contribute to the increased myocardial stiffness typically seen in diseased hearts [31]. This may account for the somewhat lower active-to-passive tension ratio seen in our study (0.8 vs. 1.2) compared to that of Holubarsch et al. [15].

The high ANP mRNA expression and the strongly reduced ISO/Ca²⁺ ratio show that the right ventricular muscle preparations used in this study were indeed terminally diseased. Recent data suggest that elevated plasma concentrations in ANP and ANP mRNA expression in ventricular myocardium positively correlate with the severity of cardiac impairment in DCM as well as in acute and chronic myocardial infarction [22, 32]. We found no relationship between the increase in force development with increasing preload (L_max–L_2.5, Fig. 5) and either the ANP levels or the ISO/Ca²⁺ ratio. Thus, at least under this in vitro condition, the effectiveness of the Frank–Starling mechanism seems to be independent of the severity of cardiac impairment. As shown in Figs. 2 and 3 there was no obvious difference in the preload dependent force generation between muscle preparations from the right or left ventricle. With this study we cannot answer the question whether the Frank–Starling mechanism is preserved yet reduced in the failing myocardium. However, the calculated L_max/L_2.5 ratio in terminally failing hearts seen in our study corresponds to values accomplished by others investigating the preload dependent force generation in non-failing hearts [14]. This supports the assumption that the Frank–Starling mechanism in vitro is not reduced in tissue from terminally failing hearts.

In addition to the classical sliding filament theory [3]a variety of mechanisms have been proposed to contribute to the Frank–Starling mechanism [33]: alterations in intracellular calcium release, activation of protein kinase C, changes in action potential, and stretch-activated ion channels. Whereas these experimental findings need further investigation, it is now generally believed that a length-dependent myofilament activation [34]plays an important role for the Frank–Starling mechanism. There is convincing evidence that the Ca²⁺ sensitivity of the contractile apparatus is length-dependent. Thus, at longer sarcomere length the affinity of troponin C for Ca²⁺ increases [35]. This effect may at least in part be due to the reduced distance between myosin and actin filaments at increased sarcomere length [29]. No specific information is available on interfilament spacing in heart failure. However, in failing hearts, fixed at the elevated filling pressure, sarcomere length was similar to that in normal myocardium [36]. Furthermore, experimental data on skinned fibres in human heart failure showed an unchanged Ca²⁺ sensitivity of the contractile protein system [37, 38], and the positive inotropic effect of calcium was found to be unaffected by the severity of heart failure [39].

In summary, our data demonstrate a preserved Frank–Starling mechanism in failing human hearts. In addition, the results show that the Frank–Starling mechanism does not correlate with the severity of heart failure, and thus favour the view that the failing heart critically depends on end-diastolic pressures in order to maintain a sufficient stroke volume.