Author + information
- Received January 25, 2013
- Revision received April 27, 2013
- Accepted April 30, 2013
- Published online July 23, 2013.
- Sevil Korkmaz, PhD∗,†,
- Edgar Zitron, MD†,‡,
- Anna Bangert†,‡,
- Claudia Seyler, PhD†,‡,
- Shiliang Li, MD∗,†,
- Peter Hegedüs, MD∗,†,
- Daniel Scherer, MD†,‡,
- Jin Li, MD†,‡,
- Thomas Fink†,‡,
- Patrick A. Schweizer, MD†,‡,
- Evangelos Giannitsis, MD†,‡,
- Matthias Karck, MD∗,†,
- Gábor Szabó, MD, PhD∗,†,
- Hugo A. Katus, MD†,‡ and
- Ziya Kaya, MD†,‡∗ ()
- ∗Department of Cardiac Surgery, University of Heidelberg, Heidelberg, Germany
- †DZHK (German Centre for Cardiovascular Research), partner site Heidelberg/Mannheim, University of Heidelberg, Heidelberg, Germany
- ‡Department of Internal Medicine III, Cardiology, University of Heidelberg, Heidelberg, Germany
- ↵∗Reprints requests and correspondence:
Dr. Ziya Kaya, Department of Cardiology, University of Heidelberg, Im Neuenheimer Feld 410, 69120 Heidelberg, Germany.
Objectives This study sought to test the hypothesis that inducing an autoimmune response against the cardiac sodium channel (NaV1.5) induces arrhythmias.
Background Sporadic evidence supports the concept that autoantibodies may cause cardiac arrhythmias but substantial experimental investigations using in vivo models have been lacking to date. The NaV1.5 is essential for cardiac impulse propagation and its dysfunction has been linked to conduction disease.
Methods Rats were immunized with a peptide sequence derived from the third extracellular loop of the first domain of NaV1.5. After 28 days, we evaluated in vivo both the electrical and mechanical parameters of cardiac function. Histopathology, myocardial gene and protein expression were assessed. Whole-cell patch-clamp was used to measure sodium current (INa) density in isolated cardiomyocytes.
Results NaV1.5-immunized rats had high titers of autoantibodies against NaV1.5. On ECG recording, NaV1.5-immunized animals showed significantly prolonged PR-intervals. During Holter ECG-monitoring we observed repeated prolonged episodes of third-degree atrioventricular and sinoatrial block in every NaV1.5-immunized animal, but not in controls. Immunization had no effect on cardiac function. In comparison to controls, myocardial NaV1.5 mRNA and protein levels were decreased in immunized rats. INa density was reduced in cardiomyocytes incubated with sera from NaV1.5-immunized rats and from patients with idiopathic atrioventricular block (AVB) in comparison to sera from respective controls. In patients with idiopathic AVB, we observed autoantibodies against NaV1.5 that were absent in sera from healthy controls.
Conclusions Provocation of an autoimmune response against NaV1.5 induces conductance defects probably caused by a reduced expression level and an inhibition of NaV1.5 by autoantibodies, resulting in decreased INa.
Cardiac arrhythmias contribute substantially to morbidity and mortality and have consequences that range from asymptomatic to life-threatening disorders. Among different pathophysiological mechanisms involved in arrhythmogenesis, a heterogeneous group of sporadic experimental and clinical studies have supported the hypothesis that autoantibodies may play a pathogenetic role in arrhythmogenesis (1). Various antiheart antibodies, including those targeting beta1-adrenergic receptors, muscarinic M2-acetylcholine receptors, Na+/K+-ATPase, cardiac troponin I, and cardiac myosin heavy chain, can be detected in the serum of patients with dilated cardiomyopathy and have been suggested to play a pivotal part in the pathomechanism of dilated cardiomyopathy (2). Among these, antibodies against beta1-adrenergic receptors and muscarinic M2-acetylcholine receptors have been associated with cardiac arrhythmias (3). Whereas the inflammatory process may itself be arrhythmogenic, changes in the electrophysiological properties of the myocardium also play a crucial role. In line with this notion, autoantibodies targeting the cardiac KCNH2 ion channel (human ether-a-go-go–related gene) have recently been associated with an exclusively electrophysiological phenotype with QT prolongation in a case report (4).
Most of the currently available data rely on clinical associations between the occurrence of distinct autoantibodies and arrhythmias (5,6). Unlike for cardiomyopathy phenotypes, a causative experimental investigation in an animal model has not been performed to date that evaluates whether autoimmunity against a specific cardiac target may suffice to monocausally induce an electrophysiological phenotype. However, such evidence may be highly relevant to motivate and guide future clinical investigations.
In cardiomyocytes, voltage-gated sodium channels are crucial for the initiation and conduction of action potentials. The alpha subunit of the cardiac sodium channel NaV1.5 is encoded by the SCN5A gene and is the predominant isoform in the heart. Heterozygous mutations in SCN5A leading to cardiac sodium channelopathies have been associated with a range of cardiac phenotypes including progressive cardiac conduction defect, sick sinus syndrome, long QT syndrome type 3, Brugada syndrome, atrial fibrillation, and even dilated cardiomyopathy (7–9). NaV1.5 is located in the sarcolemma of cardiomyocytes and consists of four homologous domains (DI-DIV), with each domain composed of six transmembrane segments (S1–S6) connected to each other by alternating extracellular and cytoplasmic loops.
Hence, we studied the functional effects of antibodies induced against the third extracellular loop of the first domain of the NaV1.5.
See the Online Appendix for further details.
Male Lewis rats (8 weeks old; Charles River, Sulzfeld, Germany) were housed at 22 ± 2°C under 12-h light/dark cycles and were fed a standard laboratory rat diet and water ad libitum. The rats were acclimatized for at least 1 week before experiments. All animals received humane care in compliance with the Principles of Laboratory Animal Care, formulated by the National Society for Medical Research; the Guide for the Care and Use of Laboratory Animals, prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (publication no. 86-23, revised 1996); and the German animal protection code. Approval was also granted by the local ethics review board (G-135/10).
The NaV1.5 peptide (GNLRHKCVRNFTELNGTNGSVEADGLVW) corresponding to the third extracellular loop of the NaV1.5 (residues 275–302), between the fifth and sixth transmembrane segment of the first domain was synthesized and purified through high-performance liquid chromatography with a purity of >90% (Peptide Specialty Laboratories, Heidelberg, Germany).
Experimental groups and immunization of animals
Rats were randomly assigned to one of two groups: the NaV1.5-peptide immunized group (n = 10) or the control group (n = 11). Immunization (on days 0, 7, and 14) was performed through a subcutaneous injection of 0.1 mL of emulsion containing 500 μg of NaV1.5 peptide dissolved in complete Freund's adjuvant supplemented with 5 mg/mL of Mycobacterium tuberculosis H37Ra (Sigma, St. Louis, Missouri). Control rats received buffer with supplemented complete Freund's adjuvant. The experiment was terminated on day 28.
Enzyme-linked immunosorbent assay for autoantibodies detection
Serum samples were obtained after centrifugation of blood collected from rats on days 0, 7, 14, 21, and 28. Antibody levels were determined using the enzyme-linked immunosorbent assay technique, as described elsewhere (10).
Standard 12-lead electrocardiography (ECG) was recorded for 1 h on days 21 and 28.
On day 25, rats were implanted with transmitter devices (ECG Device, Data Sciences International, St. Paul, Minnesota) for ECG recordings for a period of 24 h.
In vivo hemodynamic measurements
On day 28, left ventricular (LV) pressure–volume analysis of cardiac function was performed using a 2-Fr microtip pressure–volume catheter (SPR-838, Millar Instruments, Houston, Texas), as described elsewhere (11).
The heart was removed, and the right atrium and atrioventricular septum were cut away from the ventricles. The right atrium was opened by a longitudinal incision through the tricuspid valve and into the superior vena cava (n = 4–6). Additionally, 5 or 6 hearts were dissected longitudinally into 2 symmetric sections. Formalin-fixed, paraffin-embedded tissues were stained with hematoxylin and eosin (H&E) and Masson's trichrome to determine the extent of inflammation and fibrosis, respectively, as described elsewhere (12).
Quantitative polymerase chain reaction
Myocardial gene expression of NaV1.5, a tissue inhibitor of metalloproteinase (TIMP)-1 and matrix metalloproteinases (MMP)-2, -9, and -14 was performed by quantitative polymerase chain reaction. Sample quantifications were normalized to the ribosomal protein L32 expression.
Myocardial protein expression of NaV1.5 was performed by western blot using pooled serum from NaV1.5-immunized rats or primary antibodies specific to NaV1.5 (Abcam, Cambridge, United Kingdom).
Ventricular cardiomyocyte isolation and treatment
Adult ventricular cardiomyocytes were isolated using methods described elsewhere (13). After 1 h, the medium of isolated cardiomyocytes was changed to a medium containing 1:10 diluted pooled sera from control rats or from NaV1.5-immunized rats, or 1:2 diluted pooled sera from healthy human controls or from patients with idiopathic atrioventricular block (AVB), respectively, and then incubated again for at least 1 h. Cardiomyocytes were examined only after this incubation period. Instantaneous effects of direct application of sera were due to the limited availability of the rat serum not investigated.
Whole-cell patch clamp recordings
The effects of pooled sera from NaV1.5-immunized rats (n = 6) on the density of voltage-gated sodium current (INa), transient outward potassium current (Ito) and voltage-gated L-type calcium current, and the effects of pooled sera from patients with idiopathic AVB (n = 10) on the density of INa and Ito, were examined using whole-cell patch clamp on isolated rat ventricular cardiomyocytes. As negative controls, pooled sera from control rats (n = 5) or pooled sera from healthy human controls (n = 10) were used.
Screening for autoantibodies
A peptide microarray–based screening for autoantibodies against epitopes of pre-defined 168 different antigens associated with cardiovascular disease was performed with two pooled serum samples (PEPperPRINT GmbH, Heidelberg, Germany). Patients from our outpatient center diagnosed with idiopathic AVB in the absence of structural heart disease were included into the first group (n = 10; clinical characteristics shown in results section). Healthy individuals without a history of cardiovascular disease, with normal ECG findings, with normal LV function, and without abnormalities on cardiac magnetic resonance tomography investigation served as the control group (n = 10). All participants gave written informed consent.
All data are expressed as mean ± SEM. The Student t test was used to analyze the differences between the groups. A value of p < 0.05 was considered statistically significant.
Baseline characteristics of the immunized rats and humoral immune response induced by NaV1.5 immunization
In rats immunized with a NaV1.5-peptide, there were no significant differences in body weight, heart weight, ratio of heart to body weight, heart rate, systolic blood pressure, diastolic blood pressure, and mean arterial pressure compared with the control group (Table 1).
To study the humoral immune response following immunization, total IgG autoantibody titers against NaV1.5 in serum were measured on days 0, 7, 14, 21, and 28. Antibodies were detectable on day 14 and displayed a high IgG response from day 21 on. All rats immunized with NaV1.5 peptide had higher levels of autoantibodies against NaV1.5-peptide compared with those in the control group (Fig. 1A).
Western blot analysis performed using sera from rats immunized with NaV1.5 against ventricular myocardial homogenates revealed a band approximating 227 kDa, corresponding to the predicted molecular weight of NaV1.5 (Fig. 1B).
ECG alterations in animals immunized with NaV1.5
In 12-lead ECGs, rats immunized with NaV1.5-peptide showed a significant prolongation of PR intervals in comparison to those in control animals (Fig. 2). This effect was observed on days 21 and 28 at similar magnitude (51 vs. 45–46 ms; p < 0.05). In contrast, heart rate and QRS interval were not significantly different between the two groups (Table 2). No relevant ST-segment alterations or Brugada-like ECG patterns were observed. Mean QT/QTc intervals were longer in NaV1.5-immunized animals in comparison to those in controls on day 21 (91 vs. 76 ms; p < 0.05). However, this effect was attenuated on day 28 (85 vs. 79 ms), and the resulting difference was no longer significant.
Due to the alterations observed in the baseline 12-lead ECG recordings, we also obtained continuous telemetric 24-h Holter ECG recordings (Fig. 3). In control animals, we did not observe ECG signs of conduction defects. Those animals exhibited largely stable PP and RR intervals (Fig. 3A). Notably, we did not find episodes of second- or third-degree AVB or sinoatrial block in any of the control animals. Only rarely, we observed the loss of a single atrioventricular conduction of a P wave resulting in RR-interval prolongation of up to a maximum of 0.5 s (Fig. 3B). By contrast, all animals immunized with NaV1.5 peptide exhibited episodes of intermittent third-degree AVB and sinoatrial block with RR intervals of up to 1.0 s (Figs. 3C, 3D, 3E, and 3F). To quantify the number of longer conduction defects, we used a cutoff value of ≥0.6 s for RR intervals. We observed in rats immunized with NaV1.5 peptide a mean of three respective episodes with RR intervals ≥0.6 s during a 24-h recording and none in the control rats (Fig. 3G).
Ventricular arrhythmias such as monomorphic and polymorphic ventricular premature beats (Online Fig. 1) were only rarely observed, without significant differences between NaV1.5-peptide–immunized rats and control rats.
Effects of NaV1.5 immunization on cardiac function
Cardiac indexes derived from pressure–volume analysis are shown in Table 3. Immunization of rats with NaV1.5 peptide had no effect on LV end-diastolic pressure, end-systolic volume, end-diastolic volume, stroke volume, load-dependent (dP/dtmax) and load-independent (slope of dP/dtmax-end-diastolic volume and preload recruitable stroke work) contractility parameters, ejection-phase indexes (ejection fraction and cardiac output), indexes of the active phase of relaxation (dP/dtmin, time-constant of left-ventricular pressure decay), or end-diastolic stiffness (end-diastolic pressure–volume relationship). Only LV end-systolic pressure was increased in immunized rats.
Association between NaV1.5 immunization and histopathologic signs of inflammation or fibrosis
On histopathologic examination of longitudinal sections (Online Fig. 2) and sinoatrial and atrioventricular nodes (Online Fig. 3) of myocardial tissue, immunization with NaV1.5 peptide led to neither inflammatory nor fibrotic reaction, as evidenced by H&E and Masson's trichrome staining, respectively.
NaV1.5 immunization and down-regulation of NaV1.5 mRNA and protein expression
Quantitative polymerase chain reaction from myocardial RNA extracts revealed that mRNA expression for NaV1.5 was significantly decreased in the immunized group compared with that in the control rats (Fig. 4A), whereas TIMP-1, MMP-2, MMP-9, and MMP-14 mRNA levels remained unchanged (Figs. 4B, 4C, 4D, and 4E). Protein expression of NaV1.5 was also significantly decreased in the immunized group compared with the control rats (Fig. 1B and Online Fig. 4).
Effects of sera from NaV1.5 immunized rats on sodium current (INa) density in ventricular cardiomyocytes
Figure 5A shows representative INa and Ito recordings from ventricular cardiomyocytes. On whole-cell patch clamp analyses, INa in cardiomyocytes incubated in medium to which pooled sera from NaV1.5-immunized rats had been added was significantly reduced compared with that in cardiomyocytes incubated in medium to which pooled sera from control rats had been added. Respective INa densities were 4.20 ± 0.27 picoamperes/picofarad (pA/pF) in NaV1.5-immunized rats versus 5.95 ± 0.37 pA/pF in control rats (p < 0.05; n = 7 rats, 17–19 myocytes) (Fig. 5B). To evaluate the specificity of this effect for INa, we also determined Ito density in those cardiomyocytes simultaneously. There was no difference in Ito after incubation with either serum (n = 7 rats, 17–19 myocytes) (Fig. 5B). Accordingly, we did not observe a change in cardiac voltage-gated L-type calcium channel current density in cardiomyocytes incubated with sera from NaV1.5-immunized rats compared with sera from the control group (n = 4–5 rats, 19 myocytes) (Online Fig. 5).
Properties of pooled sera from patients with idiopathic AVB
Pooled serum samples from patients with idiopathic AVB in the absence of structural heart disease (Table 4) were compared to pooled serum samples from healthy controls (n = 10 each). In a peptide microarray–based screening of those sera, we observed autoantibodies against different consensus motifs of the NaV1.5 (Online Table 1) that were absent in sera from healthy controls. There were two peptide sequences corresponding to the third extracellular loop of the NaV1.5, between the fifth and the sixth transmembrane segment of the first domain. One of them overlapped with the peptide used in the present study.
Hence, we also tested functional effects in those sera. We found that sera from the patients with idiopathic AVB significantly reduced INa density in cardiomyocytes compared with that in sera from healthy controls (4.77 ± 0.36 vs. 5.93 ± 0.28 pA/pF; p < 0.05; n = 21–23 myocytes) but had no significant effect on the Ito density measured simultaneously (Fig. 5C).
To study the effect of autoimmunity against the alpha subunit of cardiac voltage-gated sodium channel NaV1.5 in vivo, an autoimmune response was induced in rats through immunization with a NaV1.5-peptide. Consequently, high levels of autoantibodies targeting the third extracellular loop of the first domain could be detected in blood, reflecting successful immunization. Immunized animals developed an exclusively electrical phenotype with conduction defects on the sinoatrial and the atrioventricular level without signs of structural heart disease or myocardial inflammation. On the cellular level, this phenotype is probably caused by a reduced density of INa in cardiomyocytes.
The NaV1.5 is composed of 4 structurally homologous domains (DI–DIV) each consisting of 6 transmembrane segments (S1–S6) (14). The residues between S5 and S6 form the channel pore (P loop) and control ion selectivity and permeation. We chose a peptide sequence between S5 and S6 within the third extracellular loop (residues 275–302) of the first domain for immunization because it is accessible from the extracellular side and many single amino acid mutations herein have been linked to sodium channelopathies with different phenotypes including Brugada syndrome, cardiac conduction defects, and sick sinus syndrome (15).
NaV1.5-immunized rats exhibited a prolonged PR interval and intermittent third-degree AVB and sinoatrial block with RR intervals of 0.6 to 1.0 s. Notably, intermittent conduction blocks with RR intervals ≥0.6 s were observed in every NaV1.5-immunized animal but in none of the control animals. Rarely, we observed in control rats the loss of a single atrioventricular conduction. Overall, these findings are in line with a conduction defect/sick sinus syndrome phenotype, which is clinically known from a subset of sodium channelopathies with loss-of-function mutations in the SCN5A gene (15). Immunized animals also exhibited a QT/QTc-interval prolongation on day 21 that was attenuated on day 28. In long QT syndrome type 3, which is caused by mutations in SCN5A, phenotype variability has been described in affected families exhibiting conduction disease, sick sinus syndrome, and long QT syndrome (16). Furthermore, it has been demonstrated clinically and in transgenic models that a single structural modification of SCN5A by mutations may be associated with variable phenotypes such as loss-of-function phenotypes characterized by conduction defects and gain-of-function phenotypes resulting in long QT syndrome type 3 (16,17). Interestingly, in our model of antibody-induced SCN5A interference, we also observed an interindividual variability of electrical phenotypes. Hence, it is plausible to speculate that the variable QT prolongation observed in the immunized animals may be interpreted as a consequence of autoantibody binding and modification of the tertiary structure of the NaV1.5.
In whole-cell patch clamp measurements, we observed a reduction of INa density in cardiomyocytes incubated with sera from NaV1.5-immunized rats without a modification of the biophysical properties of INa. Of note, Ito in the studied cardiomyocytes was not affected by incubation with these sera, which is in line with a selective effect on INa. In line with our findings, Xu et al. (18) previously generated immunoglobulins targeting the third extracellular loop of the first domain for in vitro purposes and observed an inhibitory effect of those antibodies on NaV1.5 currents in an expression system. In addition to the effects on ion currents of antibody binding to SCN5A, there was a significant reduction of myocardial NaV1.5 mRNA levels and protein levels in immunized animals. This finding indicates that the functional effects may result both from reduced expression and inhibition of NaV1.5 in the plasma membrane by autoantibodies, resulting in a decrease in INa. Both clinically and in animal models, INa inhibition has been shown to cause conduction defects and sick sinus syndrome (8,15) and hence provides a plausible pathophysiological link between autoimmunity and the observed phenotype.
In immunized animals, we observed PR prolongation and intermittent third-degree AVB without prolongation of QRS intervals. Impulse propagation through the open atrioventricular node depends on L-type calcium channels, and hence these effects may appear surprising at first sight. Interestingly, however, it has been described in detail that sodium channels are localized in the pathways leading into the open node, namely the inferior nodal extension and the transitional zone (19). Accordingly, it has been proposed that these pathways are the regions in which conduction defects are to be expected when sodium channel function is impaired. As the density of sodium channels in those regions is lower than in the working myocardium, it has been further proposed that these regions may be particularly vulnerable to sodium channel dysfunction caused by mutations or drugs (19). In line with this understanding, there have been clinical reports from patients with sodium-channel mutations that showed conduction defects with AVB in the absence of QRS-interval prolongation (20,21). Overall, our findings of PR prolongation and intermittent third-degree AVB without QRS lengthening would be compatible with these observations.
Although the down-regulation of NaV1.5-expression following immunization is a consistent finding on both the mRNA and protein levels, the underlying intracellular pathways are not clear. The regulation of NaV1.5 expression, membrane targeting, and degradation is complex and has only incompletely been elucidated to date (reviewed recently by Rook et al.  and previously by Herfst et al. ). Transcriptional regulation of the underlying SCN5A gene may be modulated by intracellular signaling, by alternative splicing, and also by alterations in cellular electrophysiology such as altered sodium-channel gating. Functional INa density is further modulated by anchoring proteins, signaling pathways, and accessory subunits. To date, relatively little is known about the mechanisms determining endocytosis and degradation of NaV1.5. Autoantibody binding often induces alterations of the tertiary protein structure as demonstrated in structural analyses (24,25). Thereby, the interaction with accessory subunits and associated proteins can be disturbed and intracellular signaling pathways may be activated (25). Furthermore, antibody binding has been shown to modify degradation of membrane proteins and respective mRNA levels (26,27). Hence, potential mechanisms could be related to an internalization of NaV1.5 following binding of autoantibodies, entry of an endolysosomal degradation pathway, and subsequent negative-feedback regulation on the level of DNA transcription. Alternatively, other pathways may be activated following alterations of the tertiary structure of the channel protein, modifications of channel endocytosis, and degradation, or as a consequence of altered cellular electrophysiology due to direct INa reduction after autoantibody binding (22–25). Additionally, in the sera from patients with idiopathic AVB, we observed autoantibodies binding additional peptide sequences of the NaV1.5, which could also be examined in future research.
We did not assess the potential role of the cellular immune response on immunization against NaV1.5 but confined the investigation solely to the humoral arm of the immune system.
We have demonstrated for the first time, in vivo, that autoimmunity against an extracellular sequence motif of the NaV1.5 induces cardiac conductance defects without signs of structural heart disease or myocardial inflammation in rats. The exclusively electrical phenotype results both from a reduced expression level and an inhibition of NaV1.5 in the plasma membrane by autoantibodies, causing a decrease of INa. The findings of this study suggest a potential role of autoimmunity against NaV1.5 in conduction defects observed in patients without structural heart disease.
The authors thank Renate Öttl, Vesna Vukovic, and Patricia Kraft for their excellent technical assistance. They also thank Elke Calamia and Dr. Felix Lasitschka for histological studies; Prof. Thomas Hilbel for technical support in 12-lead ECG recording; Nicole Herzog, Tobias Mayer, and Nadine Weiberg for their help with adult cardiomyocyte isolation; and Anna-Maria Müller for proofreading the manuscript.
For an expanded Methods section, and a supplemental table and figures, please see the online version of this article.
This work was supported in part by grants from Else Kröner-Fresenius-Stiftung (to Z.K.), DZHK (German Centre for Cardiovascular Research), BMBF (German Ministry of Education and Research), and the Medical Faculty of the University of Heidelberg, Germany (postdoctoral fellowships to S.K. and D.S.). The authors have reported that they have no other relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrioventricular block
- maximal slope of the systolic pressure increment
- maximal slope of the diastolic pressure decrement
- hematoxylin and eosin
- cardiac voltage-gated sodium-channel current
- cardiac transient outward potassium current
- left ventricular
- matrix metalloproteinase
- cardiac voltage-gated sodium channel α-subunit
- cardiac voltage-gated sodium channel α-subunit gene
- tissue inhibitor of metalloproteinase
- Received January 25, 2013.
- Revision received April 27, 2013.
- Accepted April 30, 2013.
- American College of Cardiology Foundation
- Kaya Z.,
- Leib C.,
- Katus H.A.
- Nakamura K.,
- Katayama Y.,
- Kusano K.F.,
- et al.
- Baba A.,
- Yoshikawa T.,
- Ogawa S.
- Bezzina C.R.,
- Rook M.B.,
- Groenewegen W.A.,
- et al.
- Makiyama T.,
- Akao M.,
- Shizuta S.,
- et al.
- Goser S.,
- Andrassy M.,
- Buss S.J.,
- et al.
- Radovits T.,
- Korkmaz S.,
- Loganathan S.,
- et al.
- Kaya Z.,
- Goser S.,
- Buss S.J.,
- et al.
- Bezzina C.,
- Veldkamp M.W.,
- van Den Berg M.P.,
- et al.
- Remme C.A.,
- Verkerk A.O.,
- Nuyens D.,
- et al.
- Yoo S.,
- Dobrzynski H.,
- Fedorov V.V.,
- et al.
- Wang D.W.,
- Viswanathan P.C.,
- Balser J.R.,
- George A.L. Jr..,
- Benson D.W.
- Andersen L.M.,
- Andreasen P.A.,
- Svendsen I.,
- Keemink J.,
- Ostergaard H.,
- Persson E.