Author + information
- Received January 19, 2005
- Revision received March 7, 2005
- Accepted May 9, 2005
- Published online September 6, 2005.
- Luciano Cominacini, MD⁎,⁎ (, )
- Maurizio Anselmi, MD†,
- Ulisse Garbin, MD⁎,
- Anna Fratta Pasini, MD⁎,
- Chiara Stranieri, BSc, PhD⁎,
- Massimiliano Fusaro, MD†,
- Cristina Nava, MD⁎,
- Pierfrancesco Agostoni, MD†,
- Dritan Keta, MD†,
- Pietro Zardini, MD†,
- Tatsuya Sawamura, MD, PhD‡ and
- Vincenzo Lo Cascio, MD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Luciano Cominacini, Dipartimento di Scienze Biomediche e Chirurgiche, Sezione di Medicina Interna D - Università di Verona, Policlinico G.B. Rossi - P.le L.A. Scuro 10, 37134 Verona, Italy
Objectives The purpose of this study was to investigate the effect of circulating levels of oxidized low-density lipoprotein (ox-LDL) on nuclear factor-kappa B (NF-kB) activation in peripheral blood mononuclear cells (PBMC) of patients with unstable angina (UA) or stable angina (SA) and control subjects.
Background Nuclear factor-kB might be involved in atherosclerosis, as is suggested by the presence of activated NF-kB in human atherosclerotic lesions.
Methods Levels of plasma ox-LDL and circulating NF-kB in PBMC (and in separated lymphocytes and monocytes) were measured in 27 control subjects and 29 SA and 27 UA patients. In in vitro studies, the effect of ox-LDL and of the sera derived from a subgroup of UA patients and control subjects on monocytic NF-kB activation was also evaluated.
Results The UA and SA patients had higher levels of circulating ox-LDL and NF-kB in PBMC than control subjects (p < 0.001). The increase in circulating NF-kB was mainly due to the activation of monocytes. In the in vitro studies, ox-LDL dose-dependently increased the activation of NF-kB in monocytes, but not in lymphocytes derived from healthy volunteers. This increase was related to the expression of lectin-like ox-LDL receptor-1 on monocytes. The incubation of monocytes with the sera derived from the UA patients induced a significant increase in NF-kB activation compared with the sera derived from the control subjects.
Conclusions The data suggest that the activation of NF-kB in monocytes of UA patients is, at least in part, induced by circulating molecules such as ox-LDL, which has been found to be particularly elevated in UA patients.
There is increasing evidence that inflammation plays an important role in atherogenesis and might determine plaque vulnerability (1). Many of the genes involved in the acute inflammatory response that are pivotal in the atherogenic process are activated by the nuclear factor-kappa B (NF-kB) (2). Nuclear factor-kB resides inactive and bound to the inhibitory protein-kappa B (I-kB) in the cytoplasm of many cell types, including T-lymphocytes, monocytes, macrophages, endothelial cells, and smooth muscle cells (3,4). Numerous stimulants—including cytokines and oxidants (5) such as oxidized low-density lipoprotein (ox-LDL) (6)—alter I-kB, causing nuclear translocation of NF-kB.
Nuclear factor-kB was demonstrated to be present in human coronary plaque and enhanced in patients with unstable angina (UA) (7,8). Recent work has shown an association between increasing circulating level of NF-kB in peripheral blood mononuclear cells (PBMC) and UA (8). So far, it is unclear whether NF-kB activation in the circulation of patients with UA represents cause or effect, with peripheral activation of the transcription factor occurring in response to a stimulus in the coronary circulation or the coronary event itself occurring as a reaction to circulating inflammation.
Recent data indicate that circulating levels of ox-LDL are high in acute coronary syndromes (ACS) and, in particular, in UA (9). The mechanisms leading to this increase are unclear. However, it is known that plaque instability correlates with the location of macrophages, T cells, and mast cells within the plaque (1). Moreover, macrophage-rich plaques have been recently shown to contain higher concentration of ox-LDL than macrophage-poor plaques (10,11) and to be associated with elevated levels of ox-LDL in plasma (10). Taken together, this evidence suggests that plasma and plaque levels of ox-LDL may be correlated with the vulnerability to rupture of atherosclerotic lesions.
There are many scavenger receptors that can bind ox-LDL (12). Currently, however, the majority of these receptors have been shown to be mostly non-signaling molecules and to play a key role in foam cell formation (13). In contrast, ox-LDL/lectin-like ox-LDL receptor-1 (LOX-1) interaction has already been shown to modulate cellular function (14) and to induce NF-kB activation in endothelial cells (15).
Therefore, the aim of this study was to assess the role of plasma ox-LDL and LOX-1 on circulating NF-kB activation in patients with UA.
This study was approved by our hospital ethical committees, and informed consent was obtained from all patients before their enrolment.
Major requirements for enrollment in all the groups were: absence of infectious or acute/chronic inflammatory diseases, known malignancy, absence of acute/chronic renal failure, or hepatic failure. Three groups of patients were studied.
Stable angina (SA) group
This group comprised patients with typical effort anginal pain associated with documentation of inducible ischemia during exercise stress test, defined as ST-segment depression >1 mm on the electrocardiogram during bicycle ergometry or ipo/akinesia of one or more normocontractile segments of the cardiac wall during stress-echo; no significant worsening of the symptoms in the previous two months; and no anginal episodes in the week before enrollment.
This group comprised patients with at least two episodes of rest anginal pain or one episode lasting more than 20 min in the previous 48 h (class IIIB unstable angina, according to Braunwald’s classification) (16), preferably, but not necessarily, associated with electrocardiographic modifications (T-wave inversion, ST-segment depression, transient ST-segment elevation) and a normal value of I-troponin on admission and during the first 24 h.
For both SA and UA patients, exclusion criteria were: previous coronary artery bypass grafting, recent (<6 months) myocardial infarction, recent (<6 months) percutaneous coronary intervention (PCI), congestive heart failure, and coronary tree free of significant coronary artery disease (CAD) (defined as at least one stenosis with minimal luminal diameter >70% of the arterial lumen by visual estimate) detected at coronary angiography.
This group comprised patients affected by valvular or congenital heart disease, but without a clinical history of CAD, who underwent coronary angiography for a pre-surgical evaluation. The only exclusion criterion was detection of significant CAD during coronary angiography.
The following data were obtained from all patients: age, gender, presence of CAD risk factors (hypertension, cigarette smoking, diabetes mellitus, hypercholesterolemia, family history of CAD), use of medications, previous myocardial infarction, and previous revascularization (PCI or coronary artery bypass grafting).
Venous blood samples were obtained from SA and control patients the morning after the day of admission. In UA patients, samples were drawn within 24 h from the index event (the last episode of anginal pain or admission if angina occurred previously). Blood was collected from each patient and drawn into pyrogen-free blood collection tubes without additives. Multiple aliquots of serum were placed into sterile 5-ml screw-capped polypropylene vials and stored at −80°C in plastic bags. Samples were kept frozen for no longer than 30 days, with an average of 21 days. For thawing, serum tubes were placed overnight in a refrigerator at 2°C to 8°C. The presence of flocculent material was removed by aseptic filtration through a sterile 0.45-μm filter. The samples were frozen and thawed only once.
Plasma levels of total cholesterol, high-density lipoprotein (HDL) cholesterol, LDL cholesterol, triglycerides, glucose, white blood cell count with differential count, and high-sensitivity C-reactive protein were measured with standard techniques used in the Central Laboratory of Verona University Hospital. Levels of plasma ox-LDL and circulating NF-kB in PBMC and in separated lymphocytes and monocytes and of LOX-1 messenger ribonucleic acid (mRNA) and protein expression were measured by investigators blinded to the clinical and angiographic characteristics of the patients.
PBMC and monocyte isolation
Peripheral blood mononuclear cells were separated, as previously described (17), in all subjects enrolled in the study. Isolation of monocytes and lymphocytes was performed in 10 subjects of each group. Monocytes and lymphocytes were isolated from PBMC by negative or positive selection with a cocktail of hapten-conjugated antibodies and magnetic microbeads coupled to an anti-hapten monoclonal antibody (No touch monocyte isolation kit; Miltenyi Biotec, Auburn, California) and depletion on a column in a magnetic field (VarioMACS, Miltenyi Biotec). Monocyte and lymphocyte purity was >97%, as assessed by flow cytometry (data not shown).
Endotoxin contamination of cell cultures, involving the use of ox-LDL and lipoprotein-depleted serum (LPDS), was routinely excluded with the chromogenic limulus amebocyte lysate assay (Sigma, St. Louis, Missouri). Furthermore, all cell cultures were set up in the presence of 10 μg/ml polymixin B to neutralize any potential lipopolysaccharide contamination.
Nuclear factor-kB activation was measured by two methods: 1) by a sensitive multi-well colorimetric assay for active NF-kB (TRANS-AM; Active Motif, Rixensart, Belgium), as previously described (18); and 2) by flow cytometry, to differentiate NF-kB activation in lymphocytes and monocytes. As a reference, recombinant p65 (Active Motif) was used.
Flow cytometric analysis was performed according to a previously published procedure (19). Whole blood cells were labeled with phycoerythrin-conjugated anti-CD14 monoclonal antibodies and peridinin chlorophyll protein-conjugated anti-CD3 monoclonal antibodies. The cells were then labeled with mouse anti-NF-kB (nuclear-localized signal) antibodies (IgG3; Boehringer Mannheim, Mannheim, Germany), recognizing an epitope overlapping the nuclear location signal of NF-k-p65, and therefore, the activated form of NF-kB. The cells were then labeled with fluorescein isothiocyanate-conjugated rat anti-mouse IgG3 monoclonal antibodies (Pharmingen, San Diego, California). Immunofluorescence staining was analyzed with a FACScan flow cytometer, equipped with CellQuest software (Becton-Dickinson Biosciences, San Jose, California).
Oxidized LDL assay
The ox-LDL were measured with the enzyme-linked immunosorbent assay Mercodia Oxidized LDL ELISA kit, in which the wells of the microtiter plates are coated with the capture monoclonal antibody (mAb)-4E6 (9) (Mercodia AB, Uppsala, Sweden), following the method described by Holvoet et al. (20). As a standard solution, Cu2+-modified LDL, ranging from 50 to 500 ng/ml, was used.
LDL isolation and oxidation
Whole blood from healthy volunteers, containing ethylenediamine tetraacetic acid (EDTA) (1 mg/ml), was processed for LDL separation as previously described (21). Cu2+-modified LDL was prepared as previously reported (15).
LOX-1 mRNA and protein expression in separated lymphocytes and monocytes
Separated monocytes and lymphocytes were from pools derived from control subjects and SA and UA patients. Total RNA was extracted from cells by the RNAse mini kit (Qiagen, Venlo, the Netherlands). Reverse-transcriptase polymerase chain reaction (RT-PCR) was performed with the TwoStep RT-PCR kit (Invitrogen, Carlsbad, California). For each reaction, 1 μg of total RNA served as a template. For amplification, a primer pair specific for human LOX-1 (sense primer, 5′-TTACTCTCCATGGTGGTGCC-3′; antisense primer, and 5′-AGCTTCTTCT DCTTGTTGCC-3′) was used; beta-ACTIN (5′-ATCTGGCACCACACCTTCTAC-3′ and 5′-GAGGCGTACAGGGATAGCAC-3′) was used as an internal standard in the PCR mixture. A 199-bp human LOX-1 complementary DNA (cDNA) fragment and a 182-bp human beta-ACTIN cDNA fragment were enzymatically amplified by 30 and 22 repeated cycles, respectively. An aliquot of each reaction mixture was then subjected to electrophoresis on 1.5% Tris-acetate EDTA agarose gel containing ethidium bromide. The intensity of the bands was measured with an image analysis scanning system (Alpha Imager 2000; Packard Instruments, Meriden, Connecticut).
The LOX-1 protein expression was analyzed in monocytes and lymphocytes by flow cytometry, as described previously, using a specific anti-LOX-1 mAb (22).
Oxidized LDL-dependent activation of NF-kB in monocytes
Purified monocytes (3 × 105/ml, 200 μl/well) from healthy donors were cultured in 96-well trays (Costar, Cambridge, Massachusetts) in RPMI 1640 with L-glutamine (GIBCO; Invitrogen) for 20 h at 37°C, with increasing amounts of ox-LDL (from 10- to 40-μg/ml medium as measured by the previously specified enzyme-linked immunosorbent assay).
In some experiments, blocking anti-LOX-1 mAb (20 μg/ml) or control mouse immunoglobulin G (IgG) (50 μg/ml) was also added to cell culture. Nuclear factor-kB was measured in cellular extract of monocytes as previously indicated.
Plasma-dependent activation of NF-kB in monocytes
Purified monocytes (3 × 105/ml, 200 μl/well) from healthy volunteers were cultured in 96-well trays (Costar) in RPMI 1640 with L-glutamine (GIBCO) for 20 h at 37°C with 40% serum from 10 UA patients with the highest ox-LDL levels or control subjects with the lowest ox-LDL values. For processing of serum, 80 μl of serum from each patient or control subject was added to the monocyte culture immediately after thawing, at the start of the culture period.
In some experiments, anti-LOX-1 mAb (20 μg/ml) or control human IgG (50 μg/ml) was also added to the cell culture.
As a further control, monocytes were also incubated with the corresponding LPDS in which all the lipoproteins were taken away by ultracentrifugation at a density >1.21 g/ml, following the previously indicated method (21).
Continuous data are expressed as mean ± SD values, if normally distributed. Median (interquartile range) was used for variables not normally distributed. Normal distribution of the data was determined with the Shapiro-Wilk test. Differences between continuous data were analyzed by the two-tailed unpaired Student ttest. Statistical comparison among three groups was performed by one- or two-way analysis of variance and post-hoc multiple comparison with Student-Newmann-Keuls’ test, if a parametric distribution was assessed. If the data were not parametric, analysis of variance on ranks and post-hoc Dunn’s tests were used. Relationship between variables was assessed by linear regression. A probability value <0.05 was considered statistically significant. All data were analyzed with SPSS 11.01 for Macintosh (SPSS, Chicago, Illinois).
Baseline characteristics of the patients
During a period of 14 months, 102 patients were enrolled in the study; of these patients, only 82 (27 control subjects, 29 with SA, and 27 with UA) fully satisfied the enrollment criteria. Baseline clinical characteristics of the patients are listed in Table 1.In the control group, there were more women and fewer patients with hypercholesterolemia than in the SA and UA groups (p < 0.05); in the SA group, there were more patients with a history of ACS or PCI than in the UA group (p < 0.01) and more patients taking aspirin than in the control group (p < 0.05).
Data on total, LDL, and HDL cholesterol, triglycerides, glucose plasma levels, white blood cell count, and C-reactive protein are shown in Table 2.There were no significant differences in total, LDL, or HDL cholesterol, triglycerides, or glucose levels; the total white blood cell counts and C-reactive protein levels were higher in UA patients than in SA or control patients (p < 0.001).
Circulating ox-LDL and NF-kB
Figures 1Aand 1B show the levels of ox-LDL (expressed in μg/ml) and activated NF-kB (expressed in ng/μg cell protein) in the three groups of patients. The ox-LDL level differed significantly among groups. In particular, post-hoc tests revealed a significant difference between UA and SA patients and control subjects and between SA patients and control subjects (p from <0.001 to <0.01). Also, NF-kB in PBMC (Fig. 1B) results were significantly different in the three groups of patients. The UA and SA patients had higher levels of circulating NF-kB than control patients (p < 0.001). Furthermore, a statistically significant difference was observed between SA and UA patients (p < 0.01).
In the whole population, a statistically significant correlation was detected between ox-LDL in plasma and NF-kB in PBMC (r = 0.68, p < 0.001) (Fig. 2).
Percentage of CD3+ T cells and CD14+ monocytes exhibiting NF-kB activation
By flow cytometry, there was only a slight increase in the percentage of CD3+ T cells exhibiting NF-kB activation in UA and SA patients and in control subjects, and there were no significant differences among the three groups (Table 3).In contrast, the percentage of CD14+ monocytes exhibiting NF-kB activation was high, and there were significant differences among the three groups (Table 3). In particular, UA and SA patients had a higher percentage of NF-kB activation in CD14+ monocytes than control patients (p < 0.001) (Table 3). A statistically significant difference was also observed between SA and UA patients (p < 0.01) (Table 3).
NF-kB activation in separated lymphocytes and monocytes
The results on NF-kB activation in CD14+ monocytes were confirmed by directly measuring NF-kB activation in separated monocytes and lymphocytes derived from a subgroup of UA, SA, and control subjects (Fig. 3).On average, activated NF-kB levels (expressed in ng/μg cell protein) were higher in monocytes than in lymphocytes in UA, SA, and control subjects (p < 0.001). There were no differences in NF-kB levels of lymphocytes among the three groups of subjects; conversely, monocytes of UA and SA patients had higher levels of activated NF-kB than control patients (p < 0.001). A statistically significant difference was also observed between SA and UA patients (p < 0.01).
LOX-1 mRNA and protein expression in monocytes and lymphocytes
To examine whether LOX-1 was expressed in lymphocytes and monocytes, we measured the level of LOX-1 mRNA and protein. Both mRNA and peptide were present in monocytes but not in lymphocytes (Figs. 4Aand 4B).
Oxidized LDL-dependent activation of NF-kB in monocytes and lymphocytes
Different amounts of Cu2+ox-LDL (from 10 to 40 μg/ml, as measured in the previously specified assay), incubated overnight with monocytes derived from healthy volunteers, dose-dependently increased NF-kB activation (Fig. 5)(p < 0.001). Oxidized LDL had no effect in lymphocytes (Fig. 5). In some experiments, anti-LOX-1 mAb (20 μg/ml) or control mouse IgG (50 μg/ml) was also added to the cell culture. The presence of the blocking antibody significantly reduced the ox-LDL-dependent activation of NF-kB in monocytes (Fig. 5) (p < 0.001).
Plasma-dependent activation of NF-kB in monocytes
Monocytes from healthy volunteers were evaluated for NF-kB activation after they were cultured for 20 h in a medium supplemented with either 40% serum from 10 UA patients with the highest ox-LDL plasma concentrations (range from 41 to 62 μg/ml) or 40% serum from 10 control subjects with the lowest ox-LDL levels (range from 4 to 7 μg/ml). The incubation of monocytes with the sera derived from the UA patients with the highest values of ox-LDL induced a significant increase in NF-kB activation compared with the sera derived from the control subjects with the lowest values (Fig. 6)(p < 0.001).
The NF-kB activation was partially blocked by anti-LOX-1 mAb (Fig. 6) (p < 0.001). The incubation of monocytes with the LPDS of UA patients still induced a significant increase of NF-kB compared with the LPDS of the control subjects (Fig. 6). The increase, however, was much lower than that determined by sera in toto (p < 0.001) and was not inhibited by anti-LOX-1 mAb (Fig. 6).
This study shows that circulating levels of ox-LDL and NF-kB were significantly higher in patients with UA than in SA or control patients. A difference in ox-LDL and activated NF-kB between UA and SA patients was also found.
Several studies have investigated the correlation between clinical manifestations of CAD and circulating levels of ox-LDL. The results of these studies are quite variable, given the heterogeneity of the populations studied and of antibodies used that detect different oxidation-specific epitopes of ox-LDL (9,11,23). In our study, circulating ox-LDL concentrations were measured using the 4E6 antibody that is directed against an epitope generated by the substitution of lysine residues of apoprotein B 100 with aldehydes (9). Holvoet et al. (9) found elevated ox-LDL levels in patients with ACS using the 4E6 antibody, but failed to separate UA and SA patients. These contradictory results might be related to patient selection. In fact, in the study by Holvoet et al. (9), troponin was not measured, and it was not clearly specified whether their UA patients were only in class III or also in classes II or I, according to Braunwald’s classification (16).
In this study, we also showed that circulating NF-kB was more elevated in patients with UA than in patients with SA or in control subjects without severe CAD. Our results also show that NF-kB was significantly higher in SA patients than in control subjects. Although for UA patients our results are in agreement with previous findings (8), this is the first demonstration that circulating activated NF-kB is also higher in SA patients than in control subjects, indicating a progressive increase of activated NF-kB from control subjects without severe CAD to UA patients. At variance with our results, Ritchie (8) did not find any difference in circulating NF-kB between SA patients and control subjects. A likely explanation of this discrepancy might be the fact that Ritchie (8) measured NF-kB by electrophoretic mobility shift assay with subsequent densitometric semiquantitative evaluation of the electrophoretic bands, whereas in this study, circulating NF-kB was evaluated with a sensitive multi-well colorimetric assay that was reported to be at least 10 times more sensitive than electrophoretic mobility shift assay (18).
We then evaluated which cells in PBMC specifically contributed to the increase in circulating NF-kB activation. It was found that the increase in circulating NF-kB activation was mainly due to the activation of monocytes. A number of circulating agents, including cytokines and oxidants (5), or a locally produced substance at the site of plaque rupture might be responsible. In this context, for in vitro experiments, we found that ox-LDL dose-dependently increased the activation of NF-kB in monocytes derived from healthy volunteers. In contrast, ox-LDL had no effect in lymphocytes. These results agree with previous data showing a deregulation of monocytic NF-kB by ox-LDL (6). The increase in monocytic NF-kB activation induced by ox-LDL was related to the expression of LOX-1 on the monocytes, but not on lymphocytes, and was dependent on the binding of ox-LDL to this receptor, because blocking anti-LOX-1 mAb significantly reduced NF-kB activation. The finding that ox-LDL did not activate NF-kB in lymphocytes might, therefore, depend on the fact that, as previously shown (24), LOX-1 mRNA and protein were not expressed in these cells.
Finally, it has to be mentioned that the amounts of ox-LDL that increased monocytic NF-kB activation in these in vitro experiments ranged between 20 and 40 μg/ml medium, as measured with the same enzyme-linked immunosorbent assay used for plasma samples. Thus, they were in the same order of concentration found in the plasma of UA patients.
From these results we then evaluated the effect of sera derived from UA patients and control subjects on NF-kB activation in monocytes derived from healthy volunteers. The results show that the sera of UA patients increased NF-kB activation at a much higher extent than those of control subjects. Because a number of circulating agents could potentially increase monocytic NF-kB activation, the LPDS of UA patients were then used to evaluate the magnitude of lipoprotein effect on NF-kB activation. In these conditions (i.e., in absence of lipoproteins and, therefore, of ox-LDL), the increase in monocytic NF-kB activation was about 70% lower than that obtained with the complete sera. Furthermore, the fact that the contemporary incubation of plasma from UA patients with blocking anti-LOX-1 mAb decreased the NF-kB activation on the same order of magnitude demonstrates that ox-LDL greatly contributes to the activation of NF-kB in monocytes of UA patients.
Taken together, these results show that the activation of NF-kB in monocytes of UA patients might be related to the binding of ox-LDL to LOX-1. These conclusions are in line with the findings that ox-LDL/LOX-1 interaction has already been shown to modulate cellular function (14) and to induce NF-kB activation in endothelial cells (15).
In conclusion, the data of this study suggest that monocytic activation of NF-kB in UA patients is not a primary event, but rather, induced, at least partially, by circulating molecules such as ox-LDL, which have been found to be particularly elevated in UA patients. The activation of NF-kB in these cells might participate in up-regulating the expression of some genes such as interleukin-8, interleukin-1, and tissue factor, proposed to be involved in atherogenesis (5).
This work was supported in part by grants from the Ministry of University and Scientific Research of Italy; the Ministry of Culture, Sports, Science, and Technology of Japan; the ministry of Health, Labour, and Welfare of Japan; the Organization for Pharmaceutical Safety and Research; Takeda Science Foundation; and DNO Medical Research Foundation.
- Abbreviations and Acronyms
- acute coronary syndrome
- coronary artery disease
- high-density lipoprotein
- immunoglobulin G
- lectin-like ox-LDL receptor-1
- lipoprotein-depleted serum
- monoclonal antibody
- messenger ribonucleic acid
- nuclear factor-kappa B
- oxidized low-density lipoprotein
- peripheral blood mononuclear cells
- percutaneous coronary intervention
- reverse transcriptase-polymerase chain reaction
- stable angina
- unstable angina
- Received January 19, 2005.
- Revision received March 7, 2005.
- Accepted May 9, 2005.
- American College of Cardiology Foundation
- Beg A.A.,
- Baldwin A.S.
- Brand K.,
- Eisele T.,
- Kreusel U.,
- et al.
- Ritchie M.E.
- Holvoet P.,
- Vanhaecke J.,
- Janssens S.,
- Van de Werf F.,
- Collen D.
- Nishi K.,
- Itabe H.,
- Uno M.,
- et al.
- Ehara S.,
- Ueda M.,
- Naruko T.,
- et al.
- Kunjathoor V.V.,
- Febbraio M.,
- Podrez E.A.,
- et al.
- Cominacini L.,
- Pasini A.F.,
- Garbin U.,
- et al.
- Braunwald E.
- Ichiyama T.,
- Nishikawa M.,
- Yoshitomi T.,
- et al.
- Holvoet P.,
- Stassen J.M.,
- Van Cleeput J.,
- Collen D.,
- Vanhaecke J.
- Tsimikas S.,
- Bergmark C.,
- Beyer R.W.,
- et al.