Author + information
- Received October 11, 2007
- Revision received December 31, 2007
- Accepted January 21, 2008
- Published online June 3, 2008.
- Alexander Jabs, MD,
- Sebastian Göbel, MS,
- Philip Wenzel, MD,
- Andrei L. Kleschyov, PhD,
- Marcus Hortmann, MS,
- Matthias Oelze, PhD,
- Andreas Daiber, PhD and
- Thomas Münzel, MD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Thomas Münzel, II Medizinische Klinik und Poliklinik, Johannes Gutenberg Universität Mainz, Langenbeckstrasse 1, 55131 Mainz, Germany.
Objectives This study sought to analyze mechanisms that mediate vascular dysfunction induced by sirolimus.
Background Despite excellent antirestenotic capacity, sirolimus-eluting stents have been found to trigger coronary endothelial dysfunction and impaired re-endothelialization.
Methods To mimic the continuous sirolimus exposure of a stented vessel, Wistar rats underwent drug infusion with an osmotic pump for 7 days.
Results Sirolimus treatment caused a marked degree of endothelial dysfunction as well as a desensitization of the vasculature to the endothelium-independent vasodilator nitroglycerin. Also, sirolimus stimulated intense transmural superoxide formation as detected by dihydroethidine fluorescence in aortae. Increased superoxide production was mediated in part by the vascular nicotinamide adenosine dinucleotide phosphate (NADPH) oxidase as indicated by a marked stimulation of p67phox/rac1 NADPH oxidase subunit expression and by increased rac1 membrane association. In addition, superoxide production in rat heart mitochondria was up-regulated by sirolimus, as measured by L012-enhanced chemiluminescence. As a consequence, electron spin resonance measurements showed a 40% reduction in vascular nitric oxide bioavailability, which was further supported by decreased serum nitrite levels.
Conclusions Sirolimus causes marked vascular dysfunction and nitrate resistance after continuous treatment for 7 days. This impaired vasorelaxation may, in part, be induced by up-regulated mitochondrial superoxide release as well as by an up-regulation of NADPH oxidase-driven superoxide production. Both processes could contribute to endothelial dysfunction observed after coronary vascular interventions with sirolimus-coated stents.
Endothelial dysfunction and oxidative stress play an essential role in the progression of atherosclerosis and predict the risk of cardiovascular events in patients with coronary heart disease (CHD) (1). Coronary stent implantation has become a standard treatment modality for patients with symptomatic CHD, and drug-eluting stents (DES) have further improved angiographic and clinical outcomes (2,3). Despite excellent antirestenotic capacity, sirolimus-eluting stents can trigger coronary endothelial dysfunction, impairment of coronary collateral function, late stent thrombosis, and delayed re-endothelialization in some patients (4–9).
Although local drug-related effects are very likely to modulate coronary vascular function, little is known about potential influences of sirolimus on the endothelium. Sirolimus (rapamycin) induces smooth muscle and endothelial cell cycle arrest in the late G1 phase by forming a complex with FK506-binding protein 12 (FKBP12) that inhibits the protein Ser-Thr kinase mammalian target of rapamycin (mTOR), with mTOR being a central element in signaling pathways involved in the control of cell growth and proliferation (10–12). Recently, Long et al. (13,14) showed that acute in vitro sirolimus treatment as well as genetic deletion of the sirolimus-receptor isoform FKBP12.6 increased protein kinase C (PKC)–mediated endothelial nitric oxide synthase (eNOS) threonine 495 phosphorylation, leading to decreased vascular nitric oxide (NO) production and endothelial dysfunction. However, the effects of a continuous in vivo sirolimus exposure as in a stented vessel (15) on vascular function have not been characterized in detail. Also, the role and sources of oxidative stress in sirolimus-induced vascular dysfunction are not completely understood.
Therefore, in the present study, we established an animal model of vascular dysfunction induced by chronic sirolimus treatment. We found increased reactive oxygen species (ROS) production, and hypothesized that this could contribute to sirolimus-induced vascular dysfunction. We further assessed the transmural distribution of vascular ROS formation, as well as the cellular pathways that could lead to increased superoxide production. Both nicotinamide adenosine dinucleotide phosphate (NADPH) oxidase and mitochondria were found to contribute significantly to the vascular superoxide load after sirolimus treatment.
Male Wistar rats (n = 34 in total, 300 to 330 g; Charles River, Sulzfeld, Germany) were anesthetized by isoflurane, and a subcutaneous osmotic minipump (model 1003D, Durect Corp., Cupertino, California) containing either sirolimus (sirolimus group) or the vehicle alone (dimethyl sulfoxide, control group) was implanted. A continuous sirolimus infusion rate of 5 mg/kg/day was used because antiproliferative and immunomodulating effects of sirolimus have been described in the rat neointima model at a dosage between 1.5 and 6 mg/kg/day (16). Systemic levels of sirolimus were measured by high-performance liquid chromatography in whole-blood ethylene diamine tetra-acetic acid samples by the local university clinic core laboratory. Sirolimus concentration in rat blood was 14.2 ± 5.7 μg/l at the time that the animals were euthanized. After 7 days, animals were euthanized under isoflurane anesthesia. Aortae and hearts were carefully removed and processed as described later in the text. All animal investigations conformed to the Guide for the Care and Use of Laboratory Animals as adopted by the U.S. National Institutes of Health. Sirolimus was kindly provided by Wyeth Pharma (Münster, Germany). All other chemicals were purchased from Sigma-Aldrich (St. Louis, Missouri), Merck (Darmstadt, Germany), or Fluka (Buchs, Switzerland).
Vascular Reactivity Studies
Vasodilator responses to the endothelium-dependent vasodilator acetylcholine (ACh) and the endothelium-independent vasodilator nitroglycerin (NTG) were determined in organ chambers by isometric tension studies, as described (17).
Analysis of NO Formation by Spin Trapping as NO-Fe(DETC)2 Complex
Vascular NO formation was detected in rat aortic segments (3 mm) using electron paramagnetic resonance as described (18). Briefly, after an equilibration period, medium was supplemented with 1 μmol/l ACh plus 200 μmol/l colloid Fe(II)-diethyldithiocarbamate (Fe[DETC]2) and further incubated for 1 h. Vascular rings were then frozen in liquid nitrogen and spectra of electron paramagnetic resonance were recorded at 77 K using an X-band radiospectrometer MS200 (Magnettech, Berlin, Germany). In control experiments, addition of the inhibitor of NO synthase, NG-nitro-L-arginine methyl ester (L-NAME; 1 mmol/l), completely prevented the formation of NO-Fe(DETC)2 in aortae.
Detection of NO Synthesis as Serum Nitrite
Nitrite, the oxidation product of NO, was analyzed in rat serum by ozone chemiluminescence after chemical reduction to NO as a measure of NO synthesis because nitrite levels correlate with NO biosynthesis (19).
Analysis of Vascular Superoxide Production
Oxidative fluorescent microtopography
The fluorescent dye dihydroethidine (DHE) was used to detect superoxide in situ (20). To assess the influence of eNOS-derived superoxide, vessels were incubated with L-NAME (1 mM) as described (18,20). Also, in some experiments, aortic cross-sections were pre-incubated with the NADPH oxidase inhibitor apocynin (1 mM) (21).
Phorbol myristate (PMA)–induced ROS production
Phorbol myristate–stimulated vascular ROS production was measured using the chemiluminescence indicator reagent Diogenes, a luminol-peroxidase based assay (50% of total reaction volume; National Diagnostics, Atlanta, Georgia) (22). Briefly, aortae were isolated in chilled buffer, cut into 3-mm segments, and incubated for 30 min at 37°C in Hanks buffered salt solution (PAA Laboratories, Pasching, Austria). Then, a mixture of Diogenes containing dimethylsulfoxide (0.1%, vehicle for PMA) and PMA (final concentration 100 nM; Calbiochem, San Diego, California) was added. Chemiluminescence was quantified using a Mithras Microplate Luminometer (Berthold, Bad Wildbad, Germany). Ten-second readings were obtained for each ring over 60 min, and photon counts were normalized for the dry weight of aortic tissue.
Real-Time Quantitative Reverse-Transcriptase Polymerase Chain Reaction (qRT-PCR)
The messenger ribonucleic acid (mRNA) expression was analyzed with qRT-PCR as previously described (19). Briefly, total ribonucleic acid from rat aorta was isolated (RNeasy Fibrous Tissue Mini Kit, Qiagen, Hilden, Germany), and 0.5 μg of total ribonucleic acid was used for RT-PCR analysis with the QuantiTect Probe RT-PCR kit (Qiagen). TATA box binding protein (TBP), eNOS, and Nox1 were obtained from MWG-Biotech (Ebersberg, Germany). A TaqMan Gene Expression assay Nox2/gp91phox was purchased as probe-and-primer set (Applied Biosystems, Foster City, California). Sequences of the primers and TaqMan probes were (forward, reverse, and probe) CTTCGTGCCAGAAATGCTGAA, TGTTCGTGGCTCTCTTATTCTCATG, and AATCCCAAGCGGTTTGCTGCAGTCA for TBP; GAGCAGCACAAGAGTTACAAAATCC, TCCACCGCTCGAGCAAAG, and CCACTGGTATCCTCTTGGCGGCG for eNOS; and ACCCCCTGAGTCTTGGAAGTG, GGGTGCATGACAACCTTGGT, and AGGATCCTTCGCTTTTATCGCTCCCG for Nox1. The comparative Ct method was used for relative mRNA quantification; gene expression was normalized to the endogenous control, TBP mRNA; and the amount of target gene mRNA in each sample was expressed relative to that of control.
Detection of Superoxide Formation in Isolated Rat Heart Mitochondria
Mitochondria were prepared from sirolimus-treated and control rats as described (23). Mitochondrial suspensions were diluted to a final protein concentration of 0.1 mg/ml in 0.5 ml of phosphate-buffered saline containing the chemiluminescence dye 8-amino-5-chloro-7-phenylpyridol[3,4-d]pyridozine-1,2-(2H,3H) dione (L012; 100μM), and ROS production was measured after stimulation with succinate (4 mM final concentration) or malate/glutamate (2.5 mM), respectively, as described (23). In a separate series of experiments, mitochondria were treated with cyclosporine A (0.2 μM), glibenclamide (10 μM), diazoxide (100 μM), rotenone (5 μM), antimycin A (20 μg/ml) or myxothiazole (20 μM) before the superoxide measurements (23,24).
Expression and Membrane Association of Protein by Western Blotting
Sample preparation and Western blotting was performed as described (19,25). Incubations were performed with a mouse monoclonal antibody against eNOS (dilution 1:1,000), p67phox (dilution 1:500), rac1 (dilution 1:1000), Nox2 (dilution 1:500) (Transduction Laboratories, Rockford, Illinois), goat polyclonal antibody Nox1 (dilution 1:100; Santa Cruz Biotechnologies, Santa Cruz, California), and rabbit polyclonal p47phox (dilution 1:500, Upstate/Millipore, Billerica, Massachusetts). To determine membrane association of p67phox, p47phox, and rac1, protein homogenates were divided into cytosolic and membrane fractions by ultracentrifugation (1 h, 100,000g, 4°C).
Results are expressed as mean ± SEM. The half maximal effective concentration values were obtained by logit transformation. Statistically significant differences were determined using an unpaired 2-sample Student t test (SPSS 9.0.1 for Windows, SPSS Inc., Chicago, Illinois). Values of p < 0.05 were considered significant.
Effects of Continuous Sirolimus Treatment on Vascular Function and NO Bioavailability
Sirolimus caused endothelial dysfunction as indicated by a markedly reduced vasodilator response of intact rat aortae to the endothelium-dependent vasodilator ACh (Fig. 1A,Table 1). Also, vascular relaxation to the endothelium-independent vasodilator nitroglycerin was impaired (Fig. 1B, Table 1). A significant reduction of NO formation as measured by electron paramagnetic resonance was observed in aortic segments from sirolimus-treated rats (Fig. 1C). Accordingly, serum nitrite was significantly lower in the sirolimus group (Fig. 1D).
Effects of Sirolimus on Vascular ROS Generation
Dihydroethidine staining revealed a strong increase in superoxide throughout all vascular wall layers in aortae from sirolimus-treated animals (Figs. 2A and 2B). Pre-incubation with the eNOS inhibitor L-NAME specifically increased DHE signaling in the endothelial monolayer of control animals (Fig. 2C), whereas neither transmural DHE fluorescence pattern nor intensity was changed in aorta of sirolimus-treated rats (Fig. 2D). In contrast, incubation of aortic segments with the NADPH oxidase inhibitor apocynin reduced vascular DHE fluorescence in the sirolimus group, albeit not to control levels (Figs. 2E and 2F). Quantitatively, vascular ROS formation in response to activation of protein kinase C with PMA was substantially increased in aortic segments of sirolimus-treated animals (Fig. 3).
Effects of Sirolimus on the Membrane Association of NADPH Oxidase Subunits
Sirolimus up-regulated the expression of the NADPH oxidase subunits p67phox and rac1 in aortic segments (Table 2). Specifically rac1 was found increased in the membrane fraction, that is, in its active membrane-associated form, whereas p67phox and p47phox showed only nonsignificant increases in membrane association (Fig. 4,Table 2). No significant differences were found in the protein expression of the NADPH oxidase isoforms NOX1 and NOX2, although NOX1 mRNA seemed to be up-regulated in sirolimus-treated animals (Table 3). For eNOS, neither mRNA nor protein expression showed any changes induced by sirolimus (Table 3).
Effects of Sirolimus on Mitochondrial ROS Formation
Mitochondrial ROS formation was measured using L012-derived chemiluminescence. A significant increase in mitochondrial ROS was observed in the sirolimus group (Fig. 5A). To analyze the underlying intracellular mechanisms contributing to increased mitochondrial ROS formation, we studied the effects of several modulators of mitochondrial function on ROS generation in separate series of experiments: pre-incubation with either cyclosporine A, glibenclamide, or rotenone reduced mitochondrial ROS formation after sirolimus treatment (Fig. 5B). In contrast, incubation with diazoxide did not change ROS generation in mitochondria from sirolimus-treated animals, but triggered a strong increase in the control group to levels similar to those observed in the sirolimus group (Fig. 5B). Also, the respiratory chain inhibitors Antimycin A and myxothiazol induced a 15- to 20-fold increase in mitochondrial ROS formation in both control and sirolimus groups (not shown). These findings were consistent for stimulation with the complex II substrate succinate (5 mM) (Fig. 5B) and the complex I substrates malate/glutamate (2.5 mM, not shown).
The results of the present study show that continuous sirolimus treatment causes impaired endothelium-dependent and -independent vascular relaxation, reduced vascular NO bioavailability, and increased vascular ROS formation. This increase is triggered in part by the vascular NADPH oxidase and by increased mitochondrial ROS release. Our findings may provide new insights into the mechanisms that link sirolimus pharmacotherapy to vascular dysfunction, specifically in view of recent clinical data showing impaired coronary endothelial function after sirolimus-eluting stent implantation (4–6).
Sirolimus Induces Vascular Dysfunction
Endothelial-dependent and -independent vasodilator responses were significantly impaired after 7 days of sirolimus treatment (Figs. 1A and 1B, Table 1). Reduced endothelium-dependent vasorelaxation has also been described for short-time in vitro sirolimus incubation of murine aortic rings (13,14); however, reduced vascular smooth muscle relaxation to endothelium-independent vasodilators was not observed in these previous studies. Our present findings indicate that sirolimus exposure leads to both endothelial and smooth muscle cell dysfunction. Because steady state arterial tissue concentrations are approached only after several days of drug exposure (26), a continuous sirolimus treatment as used in the current study seems essential to assess drug effects in the media. Also, a significant −40% reduction in vascular NO bioavailability was observed in the sirolimus group (Fig. 1C). Reduced NO formation also has been shown after acute in vitro sirolimus exposure of vascular homogenates (13), whereas a slight increase of NO content was reported in human microvascular endothelial cells (27). However, the marked decrease in serum nitrite (Fig. 1D), the oxidation product of NO whose concentration correlates with NO biosynthesis (19,28), further supports our finding of reduced NO availability after sirolimus treatment. Likewise, sirolimus has been shown to induce oxidative stress in cultured human microvascular endothelial cells (27), whereas in another in vitro study, no effect of sirolimus on superoxide production was observed in human umbilical vein endothelial cells (29).
Sirolimus Causes Increased Vascular ROS Formation
Because vascular NO is metabolized by the superoxide anion, we hypothesized that continuous sirolimus exposure may stimulate vascular superoxide production. Indeed, we detected a strong increase in superoxide levels that was not restricted to the endothelium cell layer but was observed throughout the vascular wall (Figs. 2A and 2B). To check for an involvement of the vascular NADPH oxidase, which is activated by protein kinase C (PKC), vessels were exposed to the direct PKC activator PMA. Indeed, phorbol ester-induced vascular ROS formation was found significantly up-regulated in the sirolimus group (Fig. 3).
Sirolimus Activates NADPH Oxidase-Dependent ROS Production
To further assess the involvement of the NADPH oxidase, aortic cross-sections were pre-incubated with the NADPH oxidase inhibitor apocynin (Figs. 2E and 2F), which was indeed able to drastically reduce vascular superoxide production. Cytosolic NADPH oxidase subunits p67phox and rac1 were significantly up-regulated in sirolimus-treated aortae; a trend toward increased expression was also seen for p47phox (Table 2). The rac1 protein was found specifically increased in the membrane fraction, that is, in its active form (Fig. 4). Because the small GTP-binding protein rac1 is known to be activated during endothelial NADPH oxidase activation (30), our present findings are compatible with increased NADPH oxidase activity in response to sirolimus treatment (Fig. 4). Up-regulation of rac1 among other NADPH oxidase subunits has also been described for the FKBP12 binding immunosuppressant FK506 in a rat renal transplant model (31). No significant changes were found for NOX1 and NOX2 protein expression, indicating that increased NADPH oxidase mediated ROS formation after sirolimus treatment is caused by stimulation of cytosolic NADPH oxidase subunit expression and by rac1 membrane translocation (Fig. 6) (32). Previously, we have shown that increased NADPH oxidase-mediated superoxide production may act as a kindling radical leading to eNOS uncoupling, for example, in diabetes (20) and in angiotensin II hypertension (33).
Sirolimus Does Not Induce eNOS Uncoupling
The DHE-stained vessels were incubated with the eNOS inhibitor L-NAME. NG-nitro-L-arginine methyl ester did not induce any changes in DHE signaling in the sirolimus group, whereas endothelial superoxide increased in control subjects (Figs. 2C and 2D). Hence, eNOS-derived NO quenches basal levels of superoxide in control animals, whereas uncoupled eNOS seems not to be a significant source of superoxide after sirolimus treatment. In addition, eNOS mRNA and protein expression was not changed by sirolimus (Table 3), as also observed for acute sirolimus treatment (13). Importantly, DHE fluorescence was not reduced to control levels by the NADPH oxidase inhibitor apocynin (Fig. 2F), pointing to other ROS sources being activated in response to sirolimus treatment.
Mitochondrial ROS Contribute to Sirolimus-Induced Oxidative Stress
The mitochondrial respiratory chain is the major source of ROS in most mammalian cells, and excess production of mitochondrial ROS has been implicated in aging and cardiovascular diseases such as atherosclerosis, hypertension, and diabetes (34,35). A pronounced increase in mitochondrial ROS was induced by continuous sirolimus treatment (Fig. 5A), whereas short-term in vitro incubation of isolated mitochondria with sirolimus had no effect on superoxide production (not shown). This effect was partially reversed by mitochondrial pre-incubation with the mitochondrial permeability transition pore (mPTP) blocker cyclosporine A, the KATP channel inhibitor glibenclamide, and the complex I inhibitor rotenone (Figs. 5B and 6). Treatment with the KATP channel opener diazoxide did not change sirolimus-induced mitochondrial ROS formation, but triggered mitochondrial ROS in the control group (Fig. 5B). These findings indicate that sirolimus treatment could lead to KATP channel opening by increasing intracellular superoxide, whereas direct effects of sirolimus on mitochondrial ion channels are not supported by our present data. The mitochondrial respiratory chain complexes I to III have been identified as major sites of ROS production in human umbilical vein endothelial cells and human coronary arteriolar endothelial cells (36,37). Opening of mPTP is known to induce the efflux of superoxide from the mitochondrial matrix into the cytoplasm, an effect that can be triggered by the opening of KATP channels, which itself can be induced by cytosolic ROS (38). Therefore, it is tempting to speculate that sirolimus-induced increases in cytosolic ROS from NADPH oxidases may contribute to the opening of KATP channels and subsequent ROS release from mitochondria via the mPTP (Fig. 6).
The current study can only provide limited insight into the pathophysiological mechanisms leading to impaired coronary endothelial function in a clinical setting after stent implantation (4–6), given limitations such as the comparability of rodent and human vasoreactivity, divergent vasodilator responses in different vascular beds, absence of pre-existing atherosclerotic plaques, and lack of the specific drug-release kinetics provided by sirolimus-eluting stents (15). Also, it seems conceivable that oral medication in CHD patients could interfere with sirolimus-induced vascular ROS production (19,39,40) whereas coronary inflammatory cell chemotaxis may further increase local oxidative stress after angioplasty (41,42). Given the species-specific elimination and degradation half-life of sirolimus (43), comparisons of drug concentrations among different species must be interpreted cautiously. Also, none of the previous clinical studies could conclusively explain the phenomenon of endothelial dysfunction at 6 months (4,5), whereas drug release is almost completed within 30 days. Herein, either the high tissue binding capacity and prolonged elution of sirolimus (26) and/or the persistent presence of the stent polymer (44) may contribute to endothelial dysfunction and impaired re-endothelialization. However, recent clinical data identify endothelial dysfunction as early as 2 weeks after sirolimus-eluting stent implantation (6). In addition, possible vascular side effects or adaptive mechanisms of a continuous systemic sirolimus treatment, as in organ transplant patients, cannot be extrapolated from the current data. However, our findings clearly show sirolimus-induced vascular dysfunction and identify intracellular signaling pathways that contribute to reduced vascular NO availability and increased ROS formation, and thereby clearly point to mechanisms leading to vascular side effects of sirolimus.
Summary and Clinical Implications
Continuous sirolimus exposure causes impaired endothelium-dependent and -independent vascular relaxation, reduced vascular NO formation, and increased transmural ROS production. This increase is triggered by NADPH oxidase expression and membrane association as well as stimulation of mitochondrial ROS release. These drug effects could very well contribute to clinically observed vascular dysfunction after sirolimus-eluting stent implantation, and their specific antagonization could help to optimize efficacy and safety of future pharmacological antirestenotic approaches.
The authors thank Jörg A. Schreiner for expert technical assistance. This article contains results that are part of the thesis work of Sebastian Göbel.
Supported by a vascular biology grant from Cordis GmbH, Langenfeld, Germany (Drs. Jabs and Münzel), a research grant from the Johannes Gutenberg University (Drs. Daiber, Hortmann, and Wenzel), the Deutsche Stiftung für Herzforschung, Frankfurt/M., Germany (Drs. Wenzel and Münzel), and the Deutsche Forschungsgemeinschaft, Bonn, Germany (SFB 553-C17, Drs. Daiber and Münzel).
- Abbreviations and Acronyms
- coronary heart disease
- drug-eluting stent(s)
- endothelial nitric oxide synthase
- FK506 binding protein 12
- NG-nitro-L-arginine methyl ester
- mitochondrial permeability transition pore
- mammalian target of rapamycin
- nicotinamide adenosine dinucleotide phosphate
- nitric oxide
- nitric oxide synthase
- protein kinase C
- phorbol myristate
- real-time quantitative reverse-transcriptase polymerase chain reaction
- reactive oxygen species
- Received October 11, 2007.
- Revision received December 31, 2007.
- Accepted January 21, 2008.
- American College of Cardiology Foundation
- Heitzer T.,
- Schlinzig T.,
- Krohn K.,
- et al.
- Williams D.O.,
- Abbott J.D.,
- Kip K.E.
- Togni M.,
- Windecker S.,
- Cocchia R.,
- et al.
- Hofma S.H.,
- van der Giessen W.J.,
- van Dalen B.M.,
- et al.
- Obata J.,
- Kitta Y.,
- Takano H.,
- et al.
- Meier P.,
- Zbinden R.,
- Togni M.,
- et al.
- Jensen L.O.,
- Maeng M.,
- Kaltoft A.,
- et al.
- Luscher T.F.,
- Steffel J.,
- Eberli F.R.,
- et al.
- Harris T.E.,
- Lawrence J.C. Jr..
- Long C.,
- Cook L.G.,
- Wu G.Y.,
- Mitchell B.M.
- Oelze M.,
- Warnholtz A.,
- Faulhaber J.,
- et al.
- Lavigne M.C.,
- Malech H.L.,
- Holland S.M.,
- Leto T.L.
- Wenzel P.,
- Mollnau H.,
- Oelze M.,
- et al.
- Mollnau H.,
- Oelze M.,
- August M.,
- et al.
- Levin A.D.,
- Vukmirovic N.,
- Hwang C.W.,
- Edelman E.R.
- Lauer T.,
- Preik M.,
- Rassaf T.,
- et al.
- Krotz F.,
- Keller M.,
- Derflinger S.,
- et al.
- Krotz F.,
- Engelbrecht B.,
- Buerkle M.A.,
- et al.
- Khanna A.K.,
- Pieper G.M.
- Zhang D.X.,
- Gutterman D.D.
- Warnholtz A.,
- Wendt M.,
- Munzel T.
- Therade-Matharan S.,
- Laemmel E.,
- Carpentier S.,
- et al.
- Liu Y.,
- Zhao H.,
- Li H.,
- et al.
- Wassmann S.,
- Laufs U.,
- Muller K.,
- et al.
- Warnholtz A.,
- Nickenig G.,
- Schulz E.,
- et al.
- Bauriedel G.,
- Jabs A.,
- Skowasch D.,
- et al.
- Ferron G.M.,
- Jusko W.J.