OUP user menu

Key role of myosin light chain (MLC) kinase-mediated MLC2a phosphorylation in the α1-adrenergic positive inotropic effect in human atrium

Michael Grimm, Pascal Haas, Birthe Willipinski-Stapelfeldt, Wolfram-Hubertus Zimmermann, Thomas Rau, Klaus Pantel, Michael Weyand, Thomas Eschenhagen
DOI: http://dx.doi.org/10.1016/j.cardiores.2004.09.019 211-220 First published online: 1 January 2005

Abstract

Objective: Mechanisms of the positive inotropic response to α1-adrenergic stimulation in the heart remain poorly understood, but recent evidence in rat papillary muscle suggests an important role of regulatory myosin light chain (MLC2) phosphorylation. This study investigated α1-adrenergic contractile effects and the role of MLC kinase (MLCK)-dependent phosphorylation of MLC2 in human atrial muscle strips.

Methods: Force measurement was performed on electrically stimulated atrial muscle strips (n=140; 20 hearts) in the presence of the beta-blocker nadolol. MLC2a phosphorylation was determined by 2D-polyacrylamide gel electrophoresis and Western blotting of muscle strips that were immediately freeze-clamped following force measurements.

Results: The α1-agonist phenylephrine (PE; 0.3–100 μM) exerted a concentration-dependent, monophasic, sustained positive inotropic effect (86% above basal) that was accompanied by an 80% increase in MLC2a phosphorylation. Desinhibition of MLC phosphatase by the Rho kinase inhibitor Y-27632 (10 μM) reduced the effect of PE by 16%. The MLCK inhibitor wortmannin (10 μM) completely abolished both the PE-induced increase in force and MLC2a phosphorylation. The structurally unrelated MLCK inhibitor ML-7 (10 μM) had similar effects. Neither Y-27632 nor wortmannin or ML-7 affected β-adrenergic force stimulation. In contrast to our findings in atrial muscle strips, we observed no increase in MLC2v phosphorylation after PE in human ventricular muscle strips and wortmannin failed to inhibit PE-induced force of contraction.

Conclusion: α1-Adrenergic receptors mediate a prominent increase in contractile force in human atria that depends on MLCK activity and is accompanied by an increase in MLC2 phosphorylation.

Keywords
  • Myosin light chain kinase
  • Wortmannin
  • Calcium sensitivity
  • Cardiac contractility
  • Human heart

1. Introduction

Intact cardiovascular function depends on adaptive changes of cardiac contraction and relaxation to meet the changing needs of the body. The most important regulatory inputs to the heart are brought along by the sympathetic nervous system and are mediated by β-adrenergic and α1-adrenergic receptors. In contrast to β-adrenergic effects, the α1-adrenergic responses are cAMP-independent. The α1-adrenergic signal transduction involves activation of the heterotrimeric Gq protein and phospholipase Cβ which in turn generates the second messengers IP3 and DAG, a signaling pathway that is shared by endothelin and angiotensin II. Recent evidence from a large clinical trial [1] and experiments with α1A/C double knockout mice [2] suggests that functioning α1-adrenoceptor signaling may be more critical for the maintenance of normal heart function than previously thought. Both the role and the exact mechanism of inotropic effects induced by α1-adrenoceptor activation are still controversial [3]. It has recently been proposed that the inotropic effect of α1-adrenergic stimulation in rat papillary muscles and in human ventricular myocardium depends on MLCK-mediated myosin light chain phosphorylation [4]. This view is supported by evidence that stimulation with the Gq-coupled receptor agonists phenylephrine [5], endothelin [6], or angiotensin II [5] leads to increases in cardiac MLC2 phosphorylation. Others had reported either no change in ventricular MLC2v phosphorylation following α1-adrenergic stimulation [7] or only moderate changes in MLC2v phosphorylation induced by treadmill exercise [8]. Thus, the importance of this pathway for acute inotropic effects remains equivocal. Examination of the physiological role of MLC2 phosphorylation in the heart by a transgenic approach in mice showed that phosphorylation of MLC2 may be critical for cardiac function [9]. Interestingly, the effects on atrial compartments were much more profound than the effects on the ventricles, suggesting a more important role of this regulatory pathway in atria than in ventricles. In line with this reasoning is the fact that the α1-adrenoceptor mediated positive inotropic effect in various species, including man, is much more pronounced in atria than in ventricle [10]. Thus, we hypothesized that the question whether or not MLCK and MLC2 phosphorylation are critical for the α1-adrenergic positive inotropic effect can be best studied in human atrium. Human ventricular muscles were studied for comparison. By using new phospho-specific MLC2 antibodies, we show that the acute positive inotropic effect of α1-adrenoceptor stimulation in human atria is accompanied by a robust increase in MLC2 phosphorylation that is MLCK-dependent and modulated by MLC phosphatase activity.

2. Material and methods

The protocol of the study was approved by the Ethics Committee of the University of Erlangen (30.11.00, no. 2308), and the investigation conforms with the principles outlined in the Declaration of Helsinki.

2.1. Reagents

All standard reagents, if not otherwise stated, were obtained from Sigma. Wortmannin and ML-7 were supplied by Alexis Biochemicals. Antisera against phosphorylated MLC2 (anti-muRLCP, anti-huRLCP) and against MLC2a were kindly provided by N. Epstein (Bethesda, MD) and P. Doevendans (Utrecht, The Netherlands), respectively. The monoclonal antibody against MLC2v (anti-MLC20k) was purchased from Sigma. Yoshitomi Pharmaceutical Industries (Osaka, Japan) provided the Rho kinase inhibitor Y-27632.

2.2. Muscle strip preparation

Right atrial appendages were obtained from a total of 20 patients undergoing open-heart surgery for coronary artery disease (CABG). Ventricular strips were obtained from left or right ventricular myocardium of 14 patients (age 2–72 years, 10 male). Some presented with terminal heart failure (heart transplantation because of dilated cardiomyopathy, n=4; ischemic cardiomyopathy, n=2; hypertrophic nonobstructive cardiomyopathy, n=1). Other tissues were obtained during valve replacement (aortic stenosis, n=2; pulmonary regurgitation, n=2;) or surgical correction of congenital heart defects (tetralogy of Fallot, n=2; atrial septal defect, n=1). All patients or their representatives gave written informed consent before surgery. The excised myocardium was kept at room temperature in continuously oxygenated (95% O2+5% CO2) modified Tyrode's solution (composition in mM: NaCl 119.8, KCl 5.4, CaCl2 1.8, MgCl2 1.05, NaH2PO4 0.42, NaHCO3 22.6, Na2EDTA 0.05, ascorbic acid 0.28, glucose 5.0) that was made cardioplegic by the addition of 2,3-butanedione monoxime (BDM, 30 mM) and immediately transferred to the laboratory (10 min). All preparation steps were carried out in specially designed dissection chambers. Thin trabeculae were excised from the subendocardial myocardium with microscissors and were individually mounted vertically in a cylindrical glass tissue bath (25 ml bath volume) and attached to an isometric force transducer (Ingenieurbüro Jäckel, Hanau, Germany) with a stainless steel wire. Contraction experiments were started immediately after BDM-washout with modified Tyrode's solution.

2.3. Force measurement

Experiments were performed on electrically stimulated (rectangular pulses, 1 Hz, 5 ms, 80–100 mA, 37 °C, pH 7.4) muscle strips in modified Tyrode's solution. Muscles were gradually stretched in 0.2-mm steps. Preload was adjusted to 50% of maximal twitch force. Experimental conditions (37 °C, pH 7.4, 1.8 mM Ca2+) were stable throughout the series of experiments. After stabilization of contractile force and at least three changes of Tyrode's solution, nadolol was added at 0.3 μM. Inhibitors were added 10–15 min later. After an additional incubation time of 15–20 min, inotropic and lusitropic responses to phenylephrine (0.3–100 μM) or isoprenaline (10 μM) were measured in both the absence and presence of pharmacological inhibitors (Fig. 1A). Steady state force of contraction in the presence of nadolol before addition of inhibitors or vehicle was termed “basal1”. “Basal2” was defined as the steady state force of contraction in the presence of a pharmacological inhibitor or the vehicle before addition of phenylephrine (Fig. 1A). Contraction amplitude and relaxation time (time to 90% relaxation) were processed with BMON2 software (Ingenieurbüro Jäckel). Each muscle strip was photographed with a digital camera (Coolpix 950, Nikon) during the isometric contraction experiment, and muscle strip dimensions were determined using a special software (LSM 5 Image Browser, Zeiss). The steel wire was used for calibration. Mean length of the muscle strips was 4.6 mm (2.7–9.2 mm); the mean diameter was 1.03 mm (0.7–1.39 mm). Following the contraction experiments, muscle strips were freeze-clamped with forceps precooled in liquid nitrogen and stored at −80 °C until Western blot analyses.

Fig. 1

(A) Representative simultaneous recording of contractile parameters in a human atrial preparation. Top: force of contraction; Bottom: time to peak tension (TPT) and time from peak force to 90% relaxation (RT90). Arrows: cumulatively increasing concentrations of 0.3–100 μM phenylephrine (PE) and addition of 10 μM isoprenaline (ISO). Dashed line: basal2. (B) Time course curves showing the inotropic response of human atrial trabeculae exposed to 10 μM isoprenaline (ISO) or 30 μM phenylephrine (PE). Stimulation with isoprenaline and phenylephrine was recorded simultaneously in two muscle strips from the same atrium. The experiment was performed three times using muscle strips from three different atria.

2.4. Two-dimensional gel electrophoresis

Human ventricular and atrial tissue samples were homogenized in acetone containing 10% trichloroacetic acid (TCA) and 10 mM 1,4-dithio-dl-threitol (DTT) to fix the phosphorylation status of MLC2. After 1 h at −20 °C, samples were centrifuged (14,000 × g, 10 min) and washed three times with acetone. Precipitated samples were resuspended in rehydration buffer [8 M urea, 2% CHAPS (w/v), IPG buffer (Amersham Biosciences, Upsala, Sweden), and bromphenol blue] and loaded on immobiline dry strips (pH gradient 4.5–5.5, Amersham Biosciences). Second dimension separation was done by SDS-PAGE (12%). Gels were silver-stained or blotted onto nitrocellulose, as described below.

2.5. Analysis of MLC2 Phosphorylation

Whole frozen muscle strips were homogenized (glass/teflon) 3 × 15 s in ice-cold homogenization buffer containing 20 mM Tris, 5 mM EDTA, 50 mM NaF, 2 μg/ml aprotinin, and 1% Triton X-100. Following a 10 min incubation period on ice, samples were centrifuged at 14,000 × g for 10 min at 4 °C. The pellet containing the myofilament fraction was resuspended in homogenization buffer and spun down again. Subsequently, pellets were dissolved in standard Laemmli sample buffer and subjected to SDS-PAGE and immunodetection. The phospho-specific MLC2 antibody was made and kindly provided by N. Epstein [11]. Detection of total and phosphorylated MLC2 with the anti-MLC2a or anti-MLC2v (1:10,000) and the anti-RLCP antibodies (1:7500), respectively, was accomplished using standard immunoblot procedures. Immunoreactive bands were detected using the ECL Plus Detection System (Amersham Biosciences) with subsequent exposure of Hyperfilms. Bands of nonsaturated images were quantified by densitometry, using Total Lab software (Nonlinear Dynamics, Newcastle upon Tyne, UK). Total MLC2 signals served as loading control and did not significantly differ among different treatment groups or among separate experiments. Phosphorylation status was determined by quantifying phosphorylated MLC2 signals and expressed as percentage of time control means ± S.E.M.

2.6. Immunohistochemistry and confocal laser scanning microscopy

Muscle strips were fixed overnight at 4 °C in 4% formaldehyde in phosphate-buffered saline, pH 7.4. Immunohistochemistry was performed on cryostat sections (7 μm) using the primary antibody against phosphorylated MLC2 (muRLCP, 1:500, N. Epstein) and anti-α-actinin (1:800, Sigma). Confocal imaging was performed with a Zeiss LSM 5 Pascal system using a Zeiss Axiovert microscope.

2.7. Statistical analysis

Data were calculated as means ± S.E.M. and analyzed using Student's t test (Microsoft Excel 2002 for Windows). Repeated measures one-way ANOVA with Dunnett's post hoc test was performed using StatView version 5.0 for Windows, SAS Institute, Cary, NC, USA. Nonparametric (Spearman) correlation and nonlinear regression was performed using GraphPad Prism version 4.00 for Windows, GraphPad Software, San Diego, CA, USA. A P value less than 0.05 was considered significant.

3. Results

3.1. Effect of MLCK inhibition on force development in atrial muscles

In human atrial muscle strips, baseline twitch tension after equilibration with nadolol was 4.6 ± 0.5 mN/mm2 (n=68). Nadolol had a mean negative inotropic effect of −26% (representative experiment shown in Fig. 1A), most likely due to antagonism of endogenously released norepinephrine. Addition of the α1-adrenoceptor antagonist prazosin at 0.1 μM had no additional effect (not shown). Phenylephrine exerted a concentration-dependent sustained positive inotropic effect that developed significantly slower than that of isoprenaline (Fig. 1B) and was accompanied by a small increase in relaxation time in most (representative example in Fig. 1A), but not all, muscles. We next examined whether MLCK activity is necessary for the positive inotropic effect of α1-adrenoceptor stimulation. To test for involvement of MLCK, we evaluated the positive inotropic effect of phenylephrine in the presence of two different pharmacological inhibitors, wortmannin and ML-7. Wortmannin is also a potent inhibitor of phosphatidylinositol 3-kinase (PI3K, Ki 5 nM). To test the involvement of PI3K on the same parameters, we used wortmannin at a concentration of 0.1 μM. While trabeculae were almost stable during the 20 min incubation period under control conditions (−7%), wortmannin and ML-7 reduced twitch tension by 17–36% (Fig. 2A). Cumulative concentration–response curves for the inotropic effect of α1-adrenoceptor stimulation are shown in Fig. 2B. Inhibition of MLCK by wortmannin shifted the concentration–response curve for phenylephrine by about 2 log units to the right and reduced the maximal effect of phenylephrine by 80%. Inhibition of MLCK by ML-7 had similar but smaller effects. In contrast, inhibition of PI3K did not affect the positive inotropic effect of phenylephrine (Fig. 2B). Neither wortmannin nor ML-7 affected maximal twitch tension under combined stimulation with phenylephrine (100 μM) and isoprenaline (10 μM; Fig. 2C), indicating that inhibitors of MLCK did not inhibit PKA-dependent signaling pathways, and that general toxic effects of the substances can be excluded. Indeed, the concentration–response to isoprenaline in the absence of phenylephrine after preincubation with MLCK-inhibitors was not significantly different from the response to isoprenaline in control trabeculae (Fig. 2D).

Fig. 2

Effects of MLCK inhibition on twitch tension in human atrial trabeculae. (A) Summary of data on basal twitch tension. (B) Effect of wortmannin (Wort) or ML-7 on the concentration–response curves of phenylephrine in atrial trabeculae. Baseline twitch tensions averaged 3.7 ± 0.7 (Ctr), 3.6 ± 0.9 (Wort, 0.1 μM), 3.8 ± 0.8 (Wort, 10 μM), and 4.8 ± 0.8 mN/mm2 (ML-7). (C) Positive inotropic effects of combined 100 μM phenylephrine and 10 μM isoprenaline in the absence or presence of the indicated inhibitors. (D) Effect of 10 μM wortmannin (Wort) or ML-7 on the concentration–response curves of isoprenaline in atrial trabeculae. Values are mean ± S.E.M. Numbers in parentheses: numbers of trabeculae examined. *P<0.05 vs. Ctr.

3.2. Phenylephrine-induced MLC2a phosphorylation

In contrast to the monophosphorylated MLC2v isoforms (Fig. 5C) that have been shown to be phosphorylatable by phenylephrine in vivo [5], the MLC2a exists in mono- or diphosphorylated forms (Fig. 3A–D), and phenylephrine-induced phosphorylation has not been described previously. In order to study the effect of phenylephrine on the phosphorylation of MLC2a, we performed 2D gel electrophoresis and Western blots, using affinity-purified polyclonal antibodies specific for MLC2a (Fig. 3C) and for phosphorylated MLC2a (Fig. 3D). The specificity of the phospho-antibody had been tested by Epstein and coworkers, showing detectable responses to both expressed MLCK and a serine/threonine phosphatase, PP1 [11]. As expected, the more acidic spots represented the mono- and diphosphorylated MLC2a, and the spot corresponding to the unphosphorylated form was not detected by the phospho-specific antibody (Fig. 3D). Expression of the phosphorylated MLC2a protein was further analyzed by examining cryo sections of an atrial muscle. Immunofluorescent labeling with the antiserum against phosphorylated MLC2a showed the expected A-band pattern when compared to the Z-disc pattern of sarcomeric α-actinin (Fig. 3E). We next determined MLC2a protein and phosphorylation levels of muscle strips, immediately freeze-clamped after completion of the force measurements. Fig. 3F shows a representative blot from a 1D-gel and the statistical analysis of 5–11 experiments. When electrically paced atrial muscles were stimulated with 30 μM phenylephrine for 10 min, the level of MLC2a phosphorylation was increased by a mean of 80% compared to time controls. Inhibition of MLCK activity decreased the basal phosphorylation level to about 50% of time controls and completely blocked the stimulatory effect of phenylephrine. There was a significant positive correlation of MLC2a phosphorylation and force of contraction in atrial muscle strips (Fig. 3G).

Fig. 5

Effects of MLCK inhibition on force of contraction in human ventricular trabeculae from failing hearts. (A) Comparison of the positive inotropic effects of phenylephrine and combined 100 μM phenylephrine/10 μM isoprenaline (PE/ISO) in human atrial and ventricular trabeculae. Numbers in parentheses: numbers of trabeculae examined. *P < 0.05 vs. atrium. (B) Positive inotropic effect of phenylephrine in the absence of inhibitors (Ctr) or in the presence of 10 μM wortmannin (Wort) or ML-7. Baseline twitch tensions averaged 4.5 ± 1.3 (Ctr), 4.0 ± 0.8 (Wort), and 4.1 ± 1.2 mN/mm2 (ML-7). Addition of 10 μM isoprenaline increased force to 6.2 ± 1.3 (Ctr), 6.4 ± 1.8 (Wort), and 5.8 ± 2.1 mN/mm2 (ML-7) above baseline. Numbers in parentheses: numbers of trabeculae examined. *P<0.05 vs. Ctr. (C) Top: silver stain of a 2D gel from human ventricular tissue. Bottom: magnified section of the top panel showing two nonphosphorylated (2v and 2v*) and phosphorylated (2vP and 2vP*) ventricular myosin light chain 2 isoforms. Note the specificity of the antibody against phosphorylated MLC2v. (D) Representative Western blot (for 6 independent experiments) of corresponding total (2v) and phosphorylated (2vP) myosin light chain 2 in human ventricular muscle strips. Muscle strips were freeze-clamped at the time of maximal force development after addition of phenylephrine or time-matched conditions, as indicated. (E) Relation of MLC2v phosphorylation and force of contraction in human ventricular muscle strips. Preparations were stimulated with or without phenylephrine or wortmannin alone or in combination.

Fig. 3

(A) Silver-stained 2D gel of TCA-extracted proteins in a human atrium. (B) Magnified section of the same 2D gel as in panel (A). (C and D) Western blot of the same section of a 2D gel, using the same protein extract as in panel (A). Detection of MLC2a (arrows, C) and phosphorylated MLC2a (D) showing the specificity of the antibodies. (E) Merged image of the red (left) and green channels (right) of a micrograph, showing immunofluorescence localization of phosphorylated MLC2a (left) and sarcomeric α-actinin in a human atrial myocyte. Shown is the typical A-band pattern of phosphorylated MLC2a (left) and the Z-disc pattern of α-actinin (right) in the same myocyte. Note the M-line (arrows). (F) Representative Western blot (for 5–7 independent experiments) of corresponding total (MLC2a) and phosphorylated (MLC2a-P) myosin light chain 2 in human atrial trabeculae (top panel). Muscles were freeze-clamped at the time of maximal force development after addition of phenylephrine or time-matched conditions, as indicated. Bottom panel summarizes normalized densitometric data from several Western blots. Numbers in parentheses: numbers of trabeculae examined. *P<0.05 vs. Ctr. (G) Correlation of MLC2a phosphorylation with force of contraction in human atrial muscle strips. Preparations were stimulated with or without phenylephrine or wortmannin alone or in combination [same strips as in panel (F)].

3.3. Involvement of Rho kinase in phenylephrine effects

The α1-AR-Gq-RhoA pathway has been shown to be involved in changes of Ca2+ sensitivity in failing hearts [12]. Rho kinase, a downstream target of RhoA, phosphorylates and thereby inhibits the myosin-binding subunit of myosin phosphatase, which dephosphorylates MLC2 [13]. Therefore, Rho kinase and MLCK could act in concert to increase the phosphorylation state of MLC2a, which then would result in an increased force of contraction. In accordance with these assumptions, the Rho kinase inhibitor Y-27632 (10 μM) by itself induced a 30% decrease in basal twitch tension (Fig. 4A) and caused a significant right-shift in the phenylephrine concentration–response curve by about 1 log unit (Fig. 4B). However, the maximal inotropic effect of phenylephrine in the presence of Y-27632 was slightly reduced but did not differ significantly from that in the absence of Y-27632. Inhibition of Rho kinase had a pronounced inhibitory effect on basal MLC2a phosphorylation (−64%) and on the absolute phosphorylation level in the presence of phenylephrine at 30 μM (Fig. 4C; last compared to second lane) but did not reduce the relative stimulatory effect of phenylephrine on MLC2a phosphorylation (Fig. 4C,D).

Fig. 4

Effect of Rho kinase inhibition with Y-27632 (10 μM, 20 min) on twitch tension and MLC2a phosphorylation in human atrial trabeculae. (A) Effect of Y-27632 on basal1 twitch tension compared to control. (B) Cumulative concentration–response curves of phenylephrine in the same preparations as in panel (A). Addition of isoprenaline (ISO) increased force to 6.7 ± 0.8 mN/mm2 (Ctr) and 5.8 ± 0.5 mN/mm2 (Y-27632) above baseline. *P<0.05 vs. Ctr. (C) Representative Western blot of corresponding total and phosphorylated MLC2a in human atrial muscle strips. Muscle strips were freeze-clamped at the time of maximal force development after addition of phenylephrine or time-matched conditions, as indicated. (D) Normalized densitometric data from several Western blots. Mean force of contraction at time of freeze-clamping was 4.8 ± 0.9 (Ctr), 8.2 ± 1.1 (PE), 2.6 ± 0.8 (Ctr+Y-27632), and 3.3 ± 0.6 mN/mm2 (PE+Y-27632). Numbers in parentheses: numbers of trabeculae examined. *P<0.05 vs. Ctr.

3.4. Effect of MLCK inhibition on force development in ventricular muscle preparations

In ventricular preparations, basal twitch tension in the presence of nadolol averaged 5.1 ± 1.5 mN/mm2 (n=38). Nadolol had a small negative inotropic effect of −10%. Wortmannin (10 μM) and ML-7 (10 μM) reduced basal twitch tension in ventricular muscles by 27% and 10%, respectively, whereas time controls remained almost stable during the 20 min incubation period. Phenylephrine concentration dependently increased force of contraction to 37% above basal (Fig. 5A). Addition of a high concentration of isoprenaline (10 μM) increased twitch tension to 122% above baseline. Thus, the effect of phenylephrine in ventricular muscles amounted to 32 ± 4% of the maximum effect elicited by the combined effect of phenylephrine and isoprenaline. This compares with 55 ± 3% for atrial muscles (Fig. 5A). Wortmannin did not affect the positive inotropic effect of phenylephrine, whereas ML-7 slightly but significantly inhibited it (Fig. 5B). When freeze-clamped ventricular muscle strips were evaluated by 2D gel electrophoresis and silver stain, 4 bands were visible at about 20 kDa and at a pH range of 4.5–5.5 (Fig. 5C). They represent MLC-2v, 2v*, and the monophosphorylated forms of both [14]. In contrast to atria, we could not observe a consistent phenylephrine-induced increase in MLC2v phosphorylation level (Fig. 5D). Neither phenylephrine (99 ± 21% of Ctr, n=7) nor wortmannin (105 ± 25% of Ctr, n=6) alone or in combination (93 ± 17% of Ctr, n=6) exerted a detectable effect on MLC2v phosphorylation compared to control (100 ± 12%, n=8; Fig. 5D). Accordingly, no correlation between force of contraction and the degree of MLC2 phosphorylation could be detected under these various conditions (Fig. 5E).

4. Discussion

In the present study, we examined the role of MLCK and MLC2 phosphorylation in the positive inotropic effect of α1-adrenergic stimulation in atrial and for comparison ventricular muscle strips from human hearts. The main new findings in human atria are as follows. (1) The relatively large positive inotropic effect of phenylephrine in human atria (to ∼180% of baseline) amounted to 55% of maximal isoprenaline response. (2) The α1-adrenergic, but not the β-adrenergic, positive inotropic effect was almost completely blocked by two chemically different pharmacological inhibitors of MLCK. (3) Phenylephrine induced a robust 80% increase in MLC2a steady state phosphorylation. (4) The MLCK inhibitors reduced basal MLC2a phosphorylation and blocked the phenylephrine-induced increase. (5) Pharmacological inhibition of Rho kinase partly inhibited the phenylephrine effect on force of contraction and MLC2a phosphorylation. (6) The MLC2 phosphorylation level, determined under various experimental conditions in freeze-clamped specimens, significantly correlated with contractile force in atrial muscle strips. Taken together, these experiments provide solid evidence for a critical role of MLCK-mediated MLC2a phosphorylation in the positive inotropic effect of α1-adrenergic stimulation in human atria.

The conclusion that MLCK-mediated hyperphosphorylation plays a critical role for the positive inotropic effect of α1-adrenergic stimulation in human atria is in accordance with previous studies performed on ventricular preparations. The studies showed that phenylephrine increased MLC2 phosphorylation by 15% in rat ventricular myocytes [4,5] and by approximately 45% in rat ventricle in vivo [5], and that inhibition of MLCK reduced the positive inotropic effect of phenylephrine by 50% in ventricular muscle strips from failing human hearts. Unexpectedly, our own experiments on human ventricular preparations failed to detect a consistent phenylephrine-induced increase in MLC2v phosphorylation, a significant blockade of phenylephrine-effects by the MLCK inhibitor wortmannin, or an effect of MLCK inhibitors on basal MLC2v phosphorylation. This apparent discrepancy could either reflect true differences in α1-adrenergic mechanisms between human atria and ventricles, or between species (rat vs. human), or could be specific for the investigated samples (subendocardial trabeculae, see below), or may be due to the small size of the α1-adrenergic effect in human ventricle. At this point, it is not possible to differentiate between these possibilities. Clearly, some technical aspects need consideration. First, we had access to less ventricular than atrial samples, making statistical evaluation of an effect half the size that in atria more difficult. Second, the variability of patient age, medication, and disease diagnosis was larger for the ventricular than for atrial samples. Third, atrial muscles have been obtained on a regular basis, i.e., 2–3 times a week, whereas ventricular muscles were mostly excised on the rare occasion of heart transplantations, making the experimental conditions less stable. Fourth, the magnitude of the α1-adrenergic effect varies substantially in human ventricle, and it is known for long that some muscles do not react at all [15]. In addition, a significant intersample variability exists in terms of MLC2a and MLC2v phosphorylation under basal unstimulated conditions. Inasmuch as the phosphorylation assay evaluated individual muscles in an unpaired fashion, a 15% increase in phosphorylation level is very easily overlooked, in fact impossible to substantiate. Finally, ML-7 had a small inhibitory effect on the positive inotropic effect of phenylephrine, very similar to that observed by Andersen et al. [4] with ML-9. Thus, our experiments cannot exclude that phenylephrine increases MLC2v phosphorylation in an MLCK-dependent manner also in human ventricular muscle. But it is apparent that this effect is small compared to human atrium and may be affected by the disease state [3]. These quantitative aspects are likely reasons why several earlier studies came to the conclusion that the MLC2 phosphorylation state is not affected by acute inotropic interventions [7,16–18]. In this context, the recent observation that an MLCK- and MLC2-phosphorylation gradient exists from (high to low) apex to midmyocardium and subepi- to subendocardial myocardium in the ventricle could offer another explanation for apparent discrepancies [11]. According to these data, levels of MLCK are expected to be low in our subendocardial ventricular trabeculae, and MLCK-mediated MLC2 phosphorylation is less relevant than in subepicardial muscle.

In contrast, the effect of phenylephrine in human atrial muscles on force of contraction and MLC2a phosphorylation as well as its blockade by MLCK inhibitors was large and robust. It should be noted that, to our knowledge, MLCK inhibition is the only intervention that almost completely antagonized the effect of phenylephrine. We and others have evaluated several pharmacological inhibitors of PKC, NCX, NHE, and MAP kinases to block phenylephrine effects in human atrial and rat papillary muscles, respectively, and found them to be either without or with only moderate effect [4,19]. Accordance only exists that α1-adrenergic stimulation in heart muscle induces an increase in intracellular Ca2+ transients that is significantly smaller than the accompanying increase in force [20], indicating Ca2+ sensitization (β-adrenergic stimulation, in contrast, induces a relatively larger increase in Ca2+ transients, indicating desensitization). Thus, at least two different mechanisms have to be considered, the one(s) underlying the small increase in Ca2+ transients and the other(s) underlying Ca2+ sensitization. The phenylephrine-induced prolongation of relaxation time (Fig. 1) is compatible with Ca2+ sensitization. At the molecular level, Ca2+ sensitivity increases with the number of attached myosin cross-bridges, i.e., cross-bridges in the force-generating state. MLC2 phosphorylation introduces negative charges in the light chain that causes the cross-bridge to move closer to the actin filament and/or accelerates phosphate release in the cross-bridge cycle, thus increasing the probability of attachment and force generation. In consequence, MLC2 phosphorylation increases force at a given Ca2+ concentration; that is, it increases the Ca2+ sensitivity of the myofilaments (see Ref. [14,21] for review). MLCK and PKC have both been shown to phosphorylate MLC2, and the present data do not rule out the importance of PKC, which could also act as a cofactor to increase MLCK-mediated MLC2 phosphorylation [22,23]. To better define its role, specific PKC isotypes and their downstream targets have to be identified using more specific approaches than simple pharmacological inhibition. What remains unclear at present is the mechanism of the small increase in Ca2+ transients, necessary to activate the Ca2+/calmodulin-dependent MLCK [14]. Several possibilities have been discussed. One is an IP3-induced SR release of Ca2+ [24–26]. A second one is an increase in L-type calcium currents [27], which has recently been supported by experiments using perforated-patch techniques. Under these conditions, phenylephrine increased ICa,L by 60–70% in ventricular myocytes from guinea pig [28] and from rat hearts [29], possibly via direct Ca2+ channel phosphorylation by PKC. Finally, activation of NHE1 could favor an influx of Na+ that is exchanged against Ca2+ via the NCX [30,31]. Whatever the source of Ca2+ and the exact mechanism, it is generally accepted that the increase in Ca2+ transient is not sufficient to account for the α1-adrenergic effect [32]. Our experiments in human atria suggest a model in which the initial increase in Ca2+ transient activates MLCK, which then phosphorylates MLC2a. The latter accounts for the major increase and stabilization of contractile force. In parallel, the Gq-Rho-Rho kinase-mediated inhibition of MLC2 phosphatase could amplify and stabilize the effect.

Limitations of the study. In important aspects, the study bases on pharmacological inhibitors of MLCK and Rho kinase, whose specificity cannot be proven. Indeed, radioligand binding experiments showed that ML-7 and another chemically related MLCK inhibitor, ML-9, but not wortmannin, displaced 3H-prazosin from rat heart membrane binding in a concentration-dependent manner (data not shown). We were aware of this limitation and therefore used two different inhibitors for MLCK from different chemical classes, and they gave very similar results. In addition, the MLCK- and Rho kinase inhibitors did not affect the inotropic response to the β-adrenergic agonist isoprenaline nor did they affect contractile effects of phenylephrine in ventricular muscles. This rules out a general toxicity. Finally, it is presently not possible to study α1-adrenergic positive inotropic effects in human preparations with other tools. Nevertheless, the proposed model for the positive inotropic effect of α1-adrenergic stimulation has to be proven in future experiments with more specific molecular tools in simpler experimental systems.

Acknowledgments

This study was supported by the German Ministry for Education and Research to T.E. (BMBF 01EC9803). We are very grateful to Dr. N. Epstein (NIH, Bethesda, MD) for the phospho-specific MLC2 antibodies and to Dr. P. Doevendans (Utrecht, The Netherlands) for the MLC2a antibody.

Footnotes

  • 1 Both authors contributed equally to this work.

  • Time for primary review 27 days

References

  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
View Abstract