OUP user menu

Abnormalities in intracellular Ca2+ regulation contribute to the pathomechanism of Tako-Tsubo cardiomyopathy

Holger M. Nef, Helge Möllmann, Christian Troidl, Sawa Kostin, Sandra Voss, Pirmin Hilpert, Christopher B. Behrens, Andreas Rolf, Johannes Rixe, Michael Weber, Christian W. Hamm, Albrecht Elsässer
DOI: http://dx.doi.org/10.1093/eurheartj/ehp240 2155-2164 First published online: 13 June 2009


Aims The Tako-Tsubo cardiomyopathy (TTC) is characterized by a transient contractile dysfunction that has been assigned to excessive catecholamine levels after episodes of severe emotional or physical stress. Several studies have indicated that β-adrenoceptor stimulation is associated with alteration in gene expression of Ca2+-regulatory proteins. Thus, the present study investigated the gene expression of crucial proteins [sarcoplasmic Ca2+ ATPase (SERCA2a), sarcolipin (SLN), phospholamban (PLN), ryanodine receptor (RyR2), and sodium-calcium exchanger (NCX)] involved in the Ca2+-regulating system in TTC.

Methods and results In 10 consecutive patients, TTC was diagnosed by coronary angiography, ventriculography, and echocardiography. Endomyocardial biopsies were taken during the phase of severely impaired left ventricular (LV) function and after functional recovery. Non-diseased LV tissue from three donor hearts not used for transplantation served as healthy controls. Expression levels of Ca2+-regulatory proteins were analysed by means of real-time PCR, western blot, and immunohistochemistry. SLN, predominantly expressed in the atrial component, showed a remarkable ventricular expression in TTC patients. Gene expression of SERCA2a was significantly down-regulated. Conversely, PLN/SERCA2a ratio was increased. For PLN, dephosphorylation was documented using western blot and immunostaining of PLN-Ser16 and PLN-Thr17. No changes could be documented for NCX and RyR2.

Conclusion In TTC, ventricular expression of SLN and dephosphorylation of PLN potentially result in a reduced SERCA2a activity and its Ca2+ affinity. Thus, the TTC is associated with specific alteration of Ca2+-handling proteins, which might be crucial for contractile dysfunction.

  • Tako-Tsubo cardiomyopathy
  • Apical ballooning
  • Calcium-handling proteins
  • Contractile dysfunction


The Tako-Tsubo cardiomyopathy (TTC) is a form of reversible systolic dysfunction of the mid- and apical left ventricle (LV) with abnormal changes in electrocardiography in the absence of an obstructive coronary artery disease.1 Most commonly, post-menopausal woman are affected. The syndrome was initially recognized in the Japanese population.2,3 However, an increasing number also in the western population has been reported. The underlying mechanisms responsible for this syndrome have not yet been fully elucidated.

Current theoretic concepts include epicardial multivessel vasospasm4 or microvascular dysfunction.5 The most established hypothesis relates to the role of stress in patients with TTC. In the majority of cases, triggering conditions that preceded onset were said to involve exposure to endogenous (emotional) or exogenous stresses (trauma, surgical procedure, and exacerbation of a pre-existing condition). This suggests that increased sympathetic activity plays a major role in the origin of this syndrome. In a recent study, signs of sympathoadrenergic overstimulation contributing to myocardial dysfunction could be documented by specific structural analysis of myocardial biopsies.6

Several studies have indicated that β-adrenoceptor stimulation is associated with alteration in gene expression of Ca2+-regulatory proteins.79 Furthermore, it is known that an altered Ca2+ cycling plays a dominant role in the development of pathological hearts.10

Thus, in this study, we investigated in patients with TTC the most important Ca2+-handling proteins such as sarcoplasmic Ca2+ ATPase (SERCA2a), ryanodine receptor (RyR2), sodium-calcium exchanger (NCX), phospholamban (PLN) and its phosphorylation sites PLN-Thr17 and PLN-Ser16. Additionally, we examined the role of sarcolipin (SLN), which has been recently identified as having a similar function as PLN. This study aims to investigate whether an altered Ca2+-handling system might be associated with TTC.


Patients and Ethics

A period of 2 years was arbitrarily chosen for recruitment of TTC patients. As this is an observational study, no formal power or sample size calculations have been performed. From March 2005 until March 2007, TTC was diagnosed in 15 patients referred to our hospital due to the symptoms of an acute coronary syndrome. All patients showed typical features of TTC including transient ST-elevation and mild increase in cardiac enzymes. Each patient was carefully assessed with history and physical exam. Among these 15 patients, one patient showed an inverse TTC making successful biopsy of the affected region impossible, in 1 patient biopsy could not be taken due to haemodynamic instability, and three patients refused their participation in the study. Finally, 10 patients fulfilled eligibility criteria and gave their written informed consent to participate in the study.

The study was approved by the Ethics Committee of the University of Freiburg and the investigation conformed to the principles outlined in the Declaration of Helsinki.

Clinical procedures

As previously described, coronary angiography and ventriculography of the LV were performed upon admission.11 Standard projections were obtained. Coronary artery disease was defined as >50% reduction in the lumen diameter. Left ventricular ejection fraction (LVEF), end-diastolic volume, and end-systolic volume were calculated by means of the area-length method using a quantitative LV analysis (QLVA, CMS, Medis Medical Imaging Systems, Leiden, The Netherlands).

Echocardiography was performed using SONOS 5500 with a 2.5 MHz transducer (Philips Medical Systems, Erlangen, Germany). Left ventricular ejection fraction was calculated from the apical two- and four-chamber views using the modified biplane Simpson method. For evaluation of regional wall abnormalities, the LV was divided into 16 segments according to the American Society of Echocardiography.12

For cardiovascular magnetic resonance imaging (CMR), a 1.5 Tesla scanner (Siemens Sonata®, Erlangen, Germany) and a six-element phased-array surface coil was used as described previously.11

Endomyocardial biopsies

Endomyocardial biopsies were obtained from each patient on the day of admission in the phase of severely depressed contractile function (acute) and after functional recovery (rec). As previously described, six biopsies were taken from the anterolateral and/or apical region of the LV guided by real-time three-dimensional echocardiography (IE33, Siemens Medical Systems, Erlangen, Germany) and fluoroscopy.13 Using this technique, the biopsy could be taken exactly from the dysfunctional region. After functional recovery, the formerly affected region could be reproducibly identified. The tissue was immediately immersed in liquid nitrogen.

Control tissue

The control tissue (con) was obtained from three donated hearts for which suitable recipients were not found at the time of surgery. The age of the donors was 26, 31, and 39 years. Two donors were female and one male. The donors did not have chronic illness but had experienced severe trauma after car accidents. Histological analysis of these hearts showed normal myocardial morphology.

RNA isolation and cDNA synthesis

Deep-frozen biopsies were homogenized in a Mixer Mill (Retsch Technology, Haan, Germany). Total RNA was isolated using the RNeasy Micro Kit (Qiagen, Venlo, The Netherlands) followed by DNase treatment (Turbo DNAfree, Ambion, Foster City, CA, USA). Biotinylated cRNA probes were synthesized by in vitro transcription using the ENZO BioArray RNA transcript labelling kit (ENZO Biochem Inc., New York, NY, USA).

Real-time polymerase chain reaction

Gene-specific RT–PCR primers were selected using the data from the PrimerBank database.14 cDNA was synthesized for 50 min at 50°C in a 20 µL reaction containing 1× first-strand buffer, 300 ng total RNA, 200 ng of random hexamer oligonucleotides, 10 mM DTT, 0.5 mM dNTPs, 40 U RNase inhibitor, and 200 U Superscript II reverse transcriptase (Invitrogen GmbH, Karlsruhe, Germany). Real-time PCR was performed in a 25 µl reaction, 96-well format (1.0 µl cDNA; 200 nM each primer; 1X SYBR Green Super Mix; Bio-Rad, Munich, Germany) using an 7500 real-time PCR system (Aplied Biosystems, Foster City, CA, USA). Three samples were measured in each experimental group in triplicate, with a minimum of two independent experiments. The relative amount of target mRNA normalized to HPRT1 was calculated according to the method described by Pfaffl and coworkers.15

Western blot analysis

Electrophoresis and blotting of the proteins were performed by using the NuPAGE electrophoresis system (Invitrogen GmbH) according to the instructions of the manufacturer. Bis-Tris gels 4–12% were used with MES or MOPS running buffer. After blotting, nitrocellulose membranes were stained with Ponceau S red (Sigma-Aldrich, Munich, Germany) and photographed (ChemiDoc XRS, Bio-Rad).

Membranes were blocked overnight with Tris–HCl 10 mmol/L, NaCl 150 mmol/L, pH 7.4 buffer (TBS) containing 5% non-fat dry milk and 0.1% Tween 20, incubated for 1 h with the first antibody (SERCA2a, Abcam, Cambridge, UK; PLN-A1, Upstate, Billerica, MA, USA; PLN-Ser16, Upstate; PLN-Thr17, SantaCruz Inc., CA, USA; NCX, Swant, Bellinzona, Switzerland; RyR2 MA3-925, Affinity BioReagents, Golden, CO, USA; PP1, Upstate; and pan-actin #4968, Cell Signaling Technology Inc., Danvers, MA, USA) diluted in a specific concentration with TBS containing 5% non-fat dry milk, washed six times with TBS containing 0.1% Tween 20 (TTBS), and then incubated for 1 h with the secondary antibody labelled with peroxidase (1:5000, goat anti-rabbit IgG, Sigma-Aldrich). After washing with TTBS, membranes were developed with a chemoluminescent substrate (FEMTO Kit, Pierce, Rockford, IL, USA). Quantification of immunoblots was done by scanning on ChemiDoc XRS (Bio-Rad) using Quantity One software. In order to equalize data, normalization of all specific targets was performed using pan-actin.

Immunohistochemistry and confocal microscopy

The tissue samples were mounted with Tissue Tek (Sakura Finetek Inc., Torrance, CA, USA) and cryosections were incubated with the first antibody (PLN-Ser16, Upstate; PLN-Thr17, SantaCruz Inc.; SLN, 1:50, SantaCruz Inc.; phalloidin, 1:100, Sigma-Aldrich) followed by treatment with a biotinylated second antibody when non-directly labelled antibodies were used. The directly labelled antibodies were conjugated to Cy3. The last incubation was carried out with fluoroisothiocyanate-linked streptavidin-Cy2 (Rockland Immunochemicals Inc., Gilbertsville, PA, USA). Nuclei were counterstained with Draq5™ (Alexis Corporation, Lausen, Switzerland).

The sections were viewed in a Leica TCS SP laser scanning confocal laser microscope (TCS SP, Leica Camera AG, Solms, Germany) equipped with appropriate filter blocks using a Silicon Graphics Octane workstation (Silicon Graphics, Sunnyvale, CA, USA) and three-dimensional multichannel image processing software (Bitplane AG, Zurich, Switzerland).


Data are expressed as mean ± SD. A non-parametric test for paired samples (Wilcoxon's signed-rank test) was used to compare the variables EF in the acute phase and after functional recovery. For comparison of the gene expression, western blot, and immunohistochemistry, Kruskal–Wallis test was used followed by post hoc analysis with Dunn's multiple comparison test to control the increase in the type I error.

Due to the observational character of our study, no adjustments for multiple testing have been made.

The statistical analysis was performed with SPSS 12 (SPSS, Chicago, IL, USA). A two-sided P-value of <0.05 was considered statistically significant.


Patients characteristics

The mean age of the patients was 61.4 ± 8.9. Prior cardiovascular history was uneventful, with no CAD, chest pain, myocardial infarction, valvular heart disease, or heart failure. Each patient experienced a stressful incident on the day of admission. All patients reported sudden onset of chest pain mimicking an acute coronary syndrome (Table 1).

View this table:
Table 1

Patient characteristics

No.AgeSex (F/M)BP (mmHg)Troponin T (ng/mL)NT-proBNP (ng/L)ECGEjection fraction (%)Emotional stressor
162F120/700.1184572++3056Fierce argument
268M160/110<0.014263++3060Visiting of a classical concert
363F170/700.3593305++4054Death of mother
455F120/600.4813954060Fierce argument
668F140/70<0.01605+3453Car accident
777F130/750.87615494056Familial conflict
848F130/800.162636+4360Fierce argument
1057F120/800.2862014+3063Fierce argument
Mean ± SD61.4 ± 8.9F, 90%; M, 10%142.0 ± 20.9/80.5 ± 16.50.31 ± 0.283088 ± 3039ST-E, 60%T-inv, 70%33.5 ± 6.956.7 ± 3.9
  • BP, blood pressure; ECG, electrocardiogram; F, female; M, male; ST-E, ST-segment elevation; T-inv, T-wave inversion.

Clinical results

In all patients, obstructive coronary artery disease or spontaneous vasospasm could be excluded by means of coronary angiography (Figure 1A and B). The ventriculography showed the typical contractile pattern of TTC, i.e. hypokinesia or akinesia in the anterolateral, apical, and diaphragmatic segments associated with a hypercontractile base of the LV. The mean EF in acute phase was 33.5 ± 6.9% as calculated by LV analysis (Figure 1C and D).

Figure 1

Representative coronary angiography from a female patient presenting with Tako-Tsubo cardiomyopathy (A, left coronary artery; B, right coronary artery). Left ventriculography showed the typical contractile pattern of wall motion abnormalities including hypo- to akinesia of the apical segments (C, systole; D, diastole). The diagnosis of Tako-Tsubo cardiomyopathy was confirmed in the acute phase (E) and regeneration of contractile dysfunction (F) was documented in cardiac magnetic resonance imaging.

In the CMR, cine-sequences confirmed the diagnosis of TTC (Figure 1E and F). Moreover, in the acute phase, the late-enhancement sequences excluded pathological signal activity ruling out necrosis or inflammatory processes. Recovery of EF could be documented after 12 ± 4 days by echocardiography and CMR (56.7 ± 3.9%).

Expression of Ca2+-handling proteins

As recently published, a microarray analysis showed that important Ca2+-regulating proteins were significantly regulated in TTC.13 SERCA2a, as a key regulator of intracellular Ca2+, showed an unaltered transcription level when compared intra-individually. However, in comparison to the used control tissue, a significant decrease was detected in the acute phase (con: 1.00 ± 0.24, acute: 0.51 ± 0.06, rec: 0.78 ± 0.14; Figure 2A). PLN mRNA levels were not significantly different in the acute phase and after functional recovery (con: 1.00 ± 0.21, acute: 7.44 ± 1.94, rec: 6.79 ± 1.63; Figure 2B).

Figure 2

RT–PCR analysis of Ca2+-regulating proteins. Sarcoplasmic Ca2+-ATPase showed a significant decrease in mRNA levels in the acute phase (A), whereas phospholamban was significantly up-regulated (B). No significant changes were observed for sodium-calcium exchanger (C) and ryanodine receptor (D) expression (Con, control; SERCA2a, sarcoplasmic Ca2+-ATPase; NCX, sodium-calcium exchanger; RyR2, ryanodine receptor).

Neither NCX (con: 1.00 ± 0.13, acute: 2.55 ± 1.87, rec: 4.44 ± 1.05; Figure 2C) nor RyR2 (con: 1.00 ± 0.27, acute: 0.53 ± 0.18, rec: 0.48 ± 0.19; Figure 2D) showed changes regarding mRNA levels. Correspondingly, protein amount of these proteins was unaltered (data not shown).

By means of RT–PCR, a significantly increased transcription of SLN in the LV could be demonstrated when compared with the myocardium taken from the same patients after functional recovery (acute: 136.2 ± 109.5, rec: 103.2 ± 25.4; Figure 3A).

Figure 3

Sarcolipin mRNA expression by PCR (A). Semi-quantitative analysis after immunostaining of sarcolipin (B). Sarcolipin showed a significant up-regulation in the acute phase (C), when compared with tissue taken after functional recovery (D), in healthy left ventricle the specific labelling could not document sarcolipin expression (E), labelling of sarcolipin in healthy left atrium (F).

Sarcolipin protein amount

The relative expression level of SLN in the different phases of TTC was obtained by immunostaining from biopsies taken in the acute phase and after functional recovery. Correspondingly to RT–PCR, also protein amount of SLN was significantly increased in the acute phase in comparison to biopsies taken after functional recovery (con: 1.48 ± 0.19, acute: 48.84 ± 2.10, rec: 23.67 ± 7.35, P < 0.05; Figure 3BD). However, in tissue from potentially healthy myocardium, SLN expression was predominantly present in the atrium, whereas signal activity in the ventricle was absent (left atrium: 25.81 ± 5.21, P < 0.05; Figure 3B and F).

Phospholamban/sarcoplasmic Ca2+ ATPase ratio

It has been proposed that the PLN/Serca2a ratio determines cardiac contractility.16 Thus, we investigated the protein expression of SERCA2a (Figure 4A) and PLN (Figure 4B) and analysed the PLN/SERCA2a ratio. As showed in western blot analysis, protein amount of SERCA2a was significantly reduced in the acute phase when compared intra-individually to biopsies taken after functional recovery. Conversely, PLN protein amount in the acute phase and after functional recovery was unaltered. PLN/SERCA2a ratio was significantly increased in TTC (con: 1.13 ± 0.39, acute: 4.15 ± 0.40, rec: 1.82 ± 0.20; Figure 4C).

Figure 4

Western blot analysis of sarcoplasmic Ca2+ ATPase (A) and phospholamban (B). Phospholamban/sarcoplasmic Ca2+ ATPase ratio was significantly increased in the acute phase of Tako-Tsubo cardiomyopathy (C).

PLN-Ser16 and -Thr17 phosphorylation

To investigate whether the phosphorylation status of PLN is altered, Ser16- and Thr17-phosphorylation has been examined. Figure 5 compares the Ser16- and Thr17-phosphorylation of control tissue, biopsies in the acute phase, and biopsies after functional recovery.

Figure 5

Western blot analysis of PLN-Ser16 (A) and PLN-Thr17 (B). Representative confocal micrographs showing PLN-Ser16(C and D) and PLN-Thr17 immunolabelling (E and F) in the same patient with acute Tako-Tsubo cardiomyopathy (C and E) and after functional recovery (D and F). Nuclei are stained blue with Draq5 and f-actin is stained red with phalloidin conjugated with TRITC.

Both the PLN-Ser16-phosphorylation (con: 1.00 ± 0.20, acute: 0.42 ± 0.06, rec: 0.63 ± 0.12, Figure 5A) and the Thr17-phosphorylation of PLN (con: 1.00 ± 0.15, acute: 1.67 ± 0.15, rec: 2.87 ± 0.19, Figure 5B) were significantly decreased in TTC when compared intra-individually to biopsies taken after functional recovery.

Immunohistological analysis revealed that PLN-Ser16 signal was localized in the cardiomyocyte cytoplasm as punctate labelling (Figure 5C and D). It was evident that in comparison with biopsies after functional recovery (Figure 5C), the fluorescence intensity of PLN-Ser16 was apparently lower in biopsies from the acute phase (Figure 5D). Correspondingly, PLN-Thr17 immunofluorescent signal was higher in the biopsies taken after functional recovery in comparison to the acute phase (Figure 5E and F).

Activity of protein phosphatase PP1

Because phosphorylation of PLN is also regulated by protein phosphatase (PP1) as the major isotype of Ser/Thr protein phosphatase in cardiomyocytes, its expression was examined in TTC by means of western blot. In the acute phase, PP1 tended to higher levels when compared with the samples taken after functional recovery (con: 1.00 ± 0.04, acute: 1.60 ± 0.14, rec: 0.93 ± 0.20; Figure 6).

Figure 6

Protein phosphatase 1 showed a more pronounced expression in the acute phase compared with the biopsies after functional recovery.


Although much is known about clinical characteristics of the new cardiac phenomenon of TTC, the main pathomechanistic concepts remain unclear. Several of the so far available studies focused on the hypothesis of an emotional triggered catecholamine excess. Akashi et al.17 first described elevated serum levels of norepinephrine in patients presenting with TTC. Furthermore, Wittstein et al.18 reported supraphysiological levels of plasma catecholamines, which were even higher than in patients with acute myocardial infarction (Kilip III). Recently, we could demonstrate the typical histological signs of catecholamine toxicity in TTC, which are described as focal, mononuclear, inflammatory areas of fibrotic response and characteristic contraction bands.6,19

The stimulation of the β-adrenoceptors by catecholamines is widely known to result in disturbances of contractility and alteration in gene expression of Ca2+-regulatory proteins.7,8 In this regard, Linck et al.9 could demonstrate that an intense β1-adrenoceptor-Gs and β2-adrenoceptor-Gs signalling might be responsible for initiating alteration in mRNA levels of PLN and SERCA2a in the mammalian heart. It was concluded that altered levels of SERCA2a and PLN probably account for the documented changes in contractile parameters.

In this context, this study provides for the first time evidence of a disturbed Ca2+-regulating system in TTC, which might be mainly triggered by supraphysiological levels of catecholamines. Thereby, the two small homologous intrinsic membrane proteins SLN and PLN, which are confined to the sarcoplasmic reticulum, are potentially critical regulators of cardiac contractility. We detected a significantly increased ventricular SLN expression in the acute phase and after functional recovery. In contrast, analysis of SLN expression in the normal heart showed that SLN is expressed predominantly in the atrial component and its expression is low20 or even absent in the present study. Sarcolipin functions as a regulator for SERCA2a by lowering Ca2+ affinity.21 Several studies suggest that SLN can either mediate its inhibitory effect on SERCA2a directly or in addition with PLN inducing a super-inhibitory effect on the SERCA pump.22 When SLN is overexpressed in ventricular cardiomyocytes, it appears to be targeted to the same subcompartments as PLN where it has the ability to form a stable SLN/PLN/SERCA2a complex.23 Two independent reports describe how cardiac-specific overexpression of SLN in mouse reduced the affinity of SERCA2a for Ca2+.23,24 In these studies on SLN transgenes, a slowed cardiomyocytes relaxation and impaired cardiac function was reported. Thus, the novel finding of an unusually ventricular SLN expression in TTC implicates its contribution to contractile dysfunction in the acute phase. The question remains, whether ventricular expression of SLN is a detrimental effect in the acute phase since its expression, although to a lower level, persists also after functional recovery. However, the interaction of SLN with PLN has not been fully understood and needs to be clarified.

The expression of PLN, as a main regulator of SERCA2a activity, was significantly increased in TTC. Accordingly, the PLN/SERCA2a ratio showed a significant increase. Alterations in the stoichiometric ratio of PLN to SERCA2a have been implicated as important determinants of reduced LV function.25 Therefore, an increased PLN/SERCA2a ratio represents a crucial factor of depressed SR function and an altered Ca2+ cycling in failing human myocardium.26

The functional interaction between SERCA2a and PLN is regulated by the phosphorylation of PLN. In the unphosphorylated state, PLN inhibits SERCA2a by lowering its apparent Ca2+ affinity.27 The inhibitory function of PLN is reversed by its phosphorylation. Phospholamban can be phosphorylated at distinct sites by different protein kinases: Ser16 by cyclic adenosine monophosphate-dependent protein kinases,28 and Thr17 by Ca2+-calmodulin-dependent kinases.29 Phosphorylation of Ser16 in PLN can occur independently of Thr17 in vivo30 and may be a prerequisite for Thr17 phosphorylation during β-agonist stimulation.31 Our finding that Ser16-phosphorylated PLN was significantly reduced in the acute phase without any change in the level of Thr17-phosphorylated PLN indicates a reduction in the PKA-dependent PLN phosphorylation in TTC.32

The abnormalities in the PKA-dependent phosphorylation pathway could result from an increase in the amount of the protein phosphatase PP1. It is assumed that >90% of myocardial PP activity occurs by PP1 and PP2A.33 Several lines of evidence suggest that PP1 is accounted for the most phosphatase activity, which dephosphorylates both the PLN-Ser16 and PLN-Thr17.34 Protein phosphatase activity has been reported to be increased after β-adrenergic stimulation.35

Taken together, this study provides evidence that an altered gene expression of the Ca2+-regulating proteins contribute to the pathomechanism of TTC. Both, SLN overexpression and a dephosphorylated PLN are factors of a potential reduction of SERCA2a activity and its Ca2+ affinity. Furthermore, an increased PLN/SERCA2a ratio represents a major determinant of contractile dysfunction. Thus, the TTC is associated with specific alteration of Ca2+-handling proteins which might be crucial for contractile dysfunction.


This study has the inherent limitations of any small observational case series. Direct measurements of SERCA2a activity and also analysis of intracellular Ca2+ levels were not achievable given the small sample amount of human endomyocardial biopsy. Nonetheless, the documented alterations were observed in all patients.

The myocardial dysfunction seen in brain-dead organ donors might be affected by alteration in Ca2+-handling proteins or a change in β-AR-receptor density.36 Moreover, it is known that donor hearts, particularly in individuals who may have experienced brain injury or trauma, may indeed be the subject of catecholamine toxicity. Despite the fact that these LV controls showed normal immunohistological and electron microscopic findings, the data from each patient were additionally paired analysing changes between the acute phase and after functional recovery.


This study was supported by the Kerckhoff-Stiftung and by a grant from the Max-Planck Institute Bad Nauheim (P.S. PFOR 406).

Conflict of interest: none declared.


The authors thank Monika Rieschel and Brigitte Matzke for their excellent technical assistance.


  • These authors contributed equally to this work.


View Abstract