Induction of Smooth Muscle Cell Migration During Arteriogenesis Is Mediated by Rap2
Objective—Collateral artery growth or arteriogenesis is the primary means of the circulatory system to maintain blood flow in the face of major arterial occlusions. Arteriogenesis depends on activation of fibroblast growth factor (FGF) receptors, but relatively little is known about downstream mediators of FGF signaling.
Methods and Results—We screened for signaling components that are activated in response to administration of FGF-2 to cultured vascular smooth muscle cells (VSMCs) and detected a significant increase of Rap2 but not of other Ras family members, which corresponded to a strong upregulation of Rap2 and C-Raf in growing collaterals from rabbits with femoral artery occlusion. Small interfering RNAs directed against Rap2 did not affect FGF-2 induced proliferation of VSMC but strongly inhibited their migration. Inhibition of FGF receptor-1 (FGFR1) signaling by infusion of a sulfonic acid polymer or infection with a dominant-negative FGFR1 adenovirus inhibited Rap2 upregulation and collateral vessel growth. Similarly, expression of dominant-negative Rap2 blocked arteriogenesis, whereas constitutive active Rap2 enhanced collateral vessel growth.
Conclusion—Rap2 is part of the arteriogenic program and acts downstream of the FGFR1 to stimulate VSMC migration. Specific modulation of Rap2 might be an attractive target to manipulate VSMC migration, which plays a role in numerous pathological processes.
- collateral circulation
- growth factors
- peripheral arterial disease
- peripheral vasculature
- vascular biology
Cardiovascular diseases are still the leading cause of death in Western societies, with coronary artery disease being responsible for approximately 50% of this burden. However, the heart of human beings is not defenseless against a slowly occurring closure of artery vessels but responds by collateral arterial growth. This process, which has been termed arteriogenesis, takes place in virtually all organs of the body. It is fundamentally different from angiogenesis in that it relies on the growth of preexisting collateral arterioles and not on the sprouting of capillaries. Arteriogenesis is initiated by shear stress leading to an inflammatory microenvironment and to the activation of growth factor cascades that spur collateral vessel growth and not by hypoxia, which mainly triggers angiogenesis.1 Furthermore, arteriogenesis is able to completely restore perfusion after occlusion of arteries, whereas angiogenesis improves the local blood supply only marginally, because far too many capillaries would be needed to replace a conducting artery.2 The ability of arteriogenesis to restore normal blood flow has raised the hope of stimulating this process to combat vascular ischemic diseases. However, the complexity of the regulatory mechanisms driving arteriogenesis, which includes the interplay of different cell types and many growth factors and cytokines, has slowed down therapeutic applications.
On the cellular level, arteriogenesis is characterized by dedifferentiation, proliferation, and migration of vascular smooth muscle cells (VSMCs) and by remodeling of the extracellular matrix. The formation of a thick neointima through phenotypically modulated and proliferating smooth muscle cells is a hallmark of collateral vessel growth. A crucial step in this process is the formation of dedifferentiated proliferating smooth muscle cells, which involves the loss of muscle specific structural proteins such as desmin.3 It has been demonstrated that one of the major driving forces of arteriogenesis is the fibroblast growth factor (FGF) system. Intraarterial infusion of a nontoxic sulfonic acid polymer, which blocks FGF signaling, reduces the size of growing collateral arteries markedly.4 Moreover, femoral artery occlusion leads to an increase of the expression and of the kinase activity of FGF receptor-1 (FGFR1). FGF-2, which is a ligand of FGFR1, acts as a potent arteriogenic factor in various animal models5 and induces proliferation and loss of desmin in cultured smooth muscle cells.3 We have previously demonstrated that activation of the mitogen-activated protein kinase (MEK)/extracellular signal-regulated kinase (Erk) pathway downstream of FGF-2 signaling is indispensable for the arteriogenic response of smooth muscle cells.3 The activity of the MEK/Erk pathway is tightly correlated with arterial growth in arterio-venous shunt models, which is the most efficient way to stimulate arteriogenesis, and loss of mitogen-activated protein kinase activity coincides with inhibition of shunt-induced arteriogenesis.1
In the current study, we explored specific signaling components downstream of the broadly active MEK/Erk pathway to obtain new tools for enhancement of collateral vessel growth. We found that Rap2 but not other small GTP-binding proteins, such as Rap1, H-Ras, RalA, and Ran, was upregulated in growing collateral vessels. Our study reveals that Rap2 is indispensable for FGF-2 induced cell migration but is not involved in the FGF-2-mediated regulation of VSMC proliferation.
Materials and Methods
Models of Femoral Artery Ligation, Laser Doppler Imaging, and Adenovirus Production
Rabbits were subjected to femoral artery occlusion and analyzed as previously described.5 Polyanetholesulfonic acid (PAS, 15 mg/kg per day, ICN) infusion was performed as described before.4 Ligation of the arteria profunda femoris and laser Doppler imager measurements were done as previously reported.6 Animal experiments were performed with permission of the local authorities. Adenoviral vectors expressing constitutive active (CA) Rap2, dominant-negative Rap2 and dominant-negative FGFR1 were generated in HEK293 cells using standard techniques, collected, and purified using Vivapure AdenoPACK 100 (Sartorius, Göttingen, Germany). The viral titer was determined with QuickTiter Adenovirus Titer Immunoassay Kit (Cell Biolabs, San Diego, CA). Animals were infected with a titer of 5×109 infectious units.
VSMC Isolation, Culture, Cell Cycle, and Scratch Assay
VSMCs were isolated and cultured in basic medium as described previously.3 Smooth muscle cells were stimulated with 10 ng/mL platelet-derived growth factor-AB (PDGF-AB) or 10 ng/mL of FGF-2. UO126 (5 μmol/L) was added 1 hour before stimulation. Migration of VSMCs was analyzed in 1-mm-wide scratches. In all experiments, VSMCs of passages 2 and 3 were used. DNA synthesis and cell number were determined as previously described.3 For flow cytometry, cells were fixed in 70% ethanol in PBS overnight at 4°C, centrifuged, and resuspended in PBS containing 50 μg/mL RNase A and 50 μg/mL propidium iodide. Different phases of the cell cycle were recorded with a FACSCalibur (Becton Dickinson, Heidelberg, Germany).
Preparation of Protein Extracts and Western Blot Analysis
Western blot analysis was performed by standard procedures3 using pan-actin antibody from Cell Signaling Technology (Danvers, MA) and Rap1, Rap2, H-Ras, RalA, Ran, and C-Raf antibodies from Becton Dickinson (Heidelberg, Germany).
Reverse Transcription–Polymerase Chain Reaction Analysis and Knockdown Experiments
RNA was extracted at the indicated times with Trizol reagent (Invitrogen, Karlsruhe, Germany). Synthesis of cDNA from DNA-free RNA was performed with SuperScript II RT (Invitrogen) followed by standard polymerase chain reaction using the following forward and reverse primer sequences: 5′-ACAAAGTGGTGGTGCTAGGC-3′ and 5′-TTTACGCGGATGATCTGGTC-3′ for Rap2; 5′-AAACGGCTACCACATCCAAG-3′ and 5′-TGCCCTCCAATGGATCCTCG-3′ for 18S rRNA. Nontargeting control small interfering RNA (siRNA) and siRNA targeting Rap2a and Rap2b were obtained from Thermo Scientific Dharmacon (Schwerte, Germany). The following target sequences were used: 5′-GCUUUAAGUUUACAAGAUA-3′, 5′-GGCUCAGCUUUAACUGUAU-3′, 5′-CUAGCAACCAGAUAAAUGA-3′, 5′-GCCUAAUCGUAGAUGUUAU-3′, 5′-CAAUAAUCUUUACAGAAUA-3′, 5′-GUAUCAAGGCCUAAGACUC-3′, 5′-GUUGGAUGUAAAGUGAUGU-3′, and 5′-GUGUUGUGCUGUGAGGCAA-3′. For knockdown experiments, confluent cultures of VSMCs derived from mouse aortas were exposed to 1 μmol/L siRNA in basic medium containing 1% serum for 3 days.
The following statistical tests were used to demonstrate significance: 1-way ANOVA and subsequent Bonferroni multiple comparisons by Kruskal-Wallis test and subsequent Dunn's multiple comparisons. Probability values <0.05 were considered statistically significant.
FGF-2 Specifically Induces Expression of the Small GTP-Binding Protein Rap2 in Cultured Smooth Muscle Cells
To identify putative signaling components that are specifically upregulated during arteriogenesis in smooth muscle cells, we treated isolated VSMCs with FGF-2, which activates FGFR1, a major driving force of arteriogenesis.5 PDGF-AB, which also elicits mitogenic effects on VSMCs but fails to induce efficient dedifferentiation of VSMCs and is absent in growing collaterals, was used as a negative control. Before treatment with FGF-2 and PDGF-AB, cells were kept for 3 days in serum-free basic medium to induce quiescence. This procedure resulted in maximal stimulation of VSMCs on addition of growth factors. We detected a 2.5-fold upregulation of Rap2 after 48 and 72 hours in FGF-2 stimulated cultures, which was absent in PDGF-AB treated cultures (Figure 1A and 1B). Likewise, neither transforming growth factor-β, insulin-like growth factor-1, nor oncostatin M, which all are known to have potent effects on VSMCs, affected the protein level of Rap2 (data not shown). Upregulation of Rap2 protein was preceded by a 1.8-fold increase of Rap2 mRNA 24 hours after stimulation of VSMCs by FGF-2, at a time point when no increase of the Rap2 protein was present (Supplemental Figure I[available online at http://atvb.ahajournals.org] versus Figure 1A). At 72 hours, the levels of Rap2 mRNA had returned to control values although Rap2 protein was still elevated. Apparently, the relatively long half-life of Rap2 protein allows increased Rap2 signaling even when Rap2 transcription has returned to baseline levels.
Next, we wanted to know whether other members of the Ras-related family of GTP-binding proteins, such as Rap1,7 were also upregulated on FGF-2 treatment. Surprisingly, we found no change of Rap1 expression by Western blot analysis after addition of FGF-2 or PDGF-AB (Figure 1C). Similar results were obtained for H-Ras, RalA, and Ran, suggesting a specific function of Rap2 within the family of Ras-related proteins for arteriogenesis.
To determine the precise onset of Rap2 expression in relation to the FGF-2 triggered cell cycle progression, we measured incorporation of tritiated thymidine (DNA synthesis), cell number, and protein levels of p27/Kip1 under serum-free conditions and after addition of different growth factors. p27/Kip1 served as a marker for G1 because its expression decreases significantly as cells reach S phase. Approximately 8% of cells in serum-free control cultures were in S phase and 4% in the G2/M phase (Supplemental Figure IIA and IIB), which seems to contradict the low levels of thymidine incorporation under the same conditions (Figure 2A). However, a certain percentage of cells escape quiescence even under serum-free conditions and enter the S and G2/M phases during the 3-day culture period, which might account for basal thymidine incorporation without stimulation. Analysis of the dose-response curve of FGF-2 and PDGF-AB revealed that 10 ng/mL was sufficient to evoke the maximal rate of DNA synthesis in VSMC cultures (Supplemental Figure IIC). Under our conditions, S phase was initiated 12 hours after addition of PDGF-AB or FGF-2, as indicated by an increase of thymidine incorporation (Figure 2A) and by degradation of the cell cycle inhibitor p27/Kip1 (Figure 2B and 2C) well before the increase of Rap2 protein at 48 hours. The rate of DNA synthesis peaked at 60 hours and decreased thereafter, whereas the number of cells continuously increased from 48 hours on (Figure 2D). Fluorescence-activated cell sorting analysis with the DNA intercalating agent propidium iodide did also reveal a significant increase in the number of cells in S and G2/M phase after stimulation with PDGF-AB or FGF-2 (Supplemental Figure IIA and IIB). Taken together, our data demonstrated that accumulation of Rap2 in FGF-2 stimulated cells occurred long after exit from G0 during late S phase and remained elevated during the transition to the M phase (Figure 2E).
Previously, we demonstrated that the mitogenic effect of FGF-2 on VSMCs depends on the MEK/Erk module.3 Here, we found that FGF-2 had a lasting effect on Erk1/2 phosphorylation, which was upregulated even >24 hours after initiation of the treatment, indicating that the pathway was not subject to substantial negative feedback regulation under our conditions (Supplemental Figure IIIA). As expected, administration of UO126 completely abolished Erk1/2 phosphorylation. It did not matter whether the inhibitor was added 1 hour before or 23 hours after stimulation with FGF-2 during the course of the 24-hour experiment (Supplemental Figure IIIA), indicating continuously active FGF signaling and instant inhibition by UO126. We next investigated whether Rap2 upregulation in response to FGF-2 was also mediated by MEK/Erk. Addition of UO126 1 hour before FGF-2 stimulation significantly reduced the increase of Rap2 expression in VSMCs (Figure 3B), demonstrating that MEK/Erk signaling was instrumental for Rap2 expression. Interestingly, inhibition of Erk signaling >11 hours after addition of FGF-2 resulted only in a minor decrease of Rap2 protein, suggesting that Erk-mediated accumulation of Rap2 occurred during the first 11 hours after stimulation (Supplemental Figure IIIB). In contrast, Rap2 RNA levels were only slightly inhibited by UO126, suggesting a posttranscriptional regulation of Rap2 by UO126-sensitive processes (Supplemental Figure IIIC).
Rap2 Is Essential for Migration of VSMCs But Not for Their Proliferation
FGF-2 signaling is indispensable for the arteriogenic response of smooth muscle cells, which goes along with induction of proliferation, loss of desmin, and increased cell migration.3,8 To analyze which of these processes is controlled by Rap2, we knocked down Rap2 in VSMCs using specific siRNAs. Western blot analysis revealed a 75% decrease of Rap2 in VSMCs by Rap2 siRNA but not of control siRNAs proving effectiveness and specificity of the approach (Figure 3D). Interestingly, the loss of Rap2 did not affect DNA synthesis and cell proliferation (Figure 3A) but prevented FGF-2-induced migration of VSMCs in a scratch assay. Addition of the Rap2 siRNA almost completely blocked migration of VSMCs in vitro in comparison to control siRNAs (Figure 3C). We did not observe a decrease of cell viability or an increase in the rate of apoptosis (data not shown). Treatment of VSMCs with UO126 yielded similar results (Figure 3C and data not shown), although inhibition of migration by Rap2 siRNA appeared to be more efficient.
Rap2 Expression Increases During Collateral Growth
To further investigate the role of the FGF-2/FGFR1/Ras/Raf/MEK/Erk signaling cascade in the activation of Rap2 expression during arteriogenesis, we studied the expression of Rap2 in collateral vessels of rabbits after femoral artery occlusion (Figure 4C). For this experiment, we used rabbits because it is technically not possible to cleanly dissect collateral vessels from mice or rats.5,9 Collaterals were excised from operated and nonligated control legs at 0.5, 1, 3, and 7 days and subjected to Western blot analysis. On induction of collateral vessel growth, we detected a strong increase of the expression of C-Raf, which serves as the main effecter of GTP-bound Ras signaling and defines the entry point to Erk signaling (Figure 4A). In addition, Rap2 was strongly activated in growing collateral vessels but not in vessels isolated from sham-operated legs. The level of Rap2 increased from day 3 onward and became statistically significant at day 7, coinciding with the peak of C-Raf expression (Figure 4A and 4B). A second Rap2 band appeared in some samples at approximately 24 kDa, most likely because of posttranslational modifications, such as geranylgeranylations or farnesylations10 (Figure 4B and 4D).
Inhibition of FGF-2/FGFR1 Axis In Vivo Inhibits Upregulation of Rap2 and Interferes With the Growth of Collateral Vessels
To prove that the upregulation of Rap2 was dependent on the FGF-2/FGFR1 axis during collateral growth, we treated rabbits for 1 week with PAS or PBS using osmotic minipumps as described previously.4 This nontoxic sulfonic acid polymer has been shown to block the action of FGF-2 via complex formation. Continuous infusion of PAS for 1 week resulted in a marked reduction of Rap2 levels in collaterals of the occluded leg that went along with a significant reduction of the size of growing collateral arteries4 (Figure 4D). As PAS might also bind and neutralize other heparin binding growth factors, we constructed an adenovirus that carries a dominant-negative (DN) form of the FGFR1 receptor11 and infected tissues harboring growing collateral vessels before ligation of the femoral artery. Efficiency of adenoviral mediated gene transfer into growing arteries was ascertained by analysis of green fluorescent protein expression mediated by an adeno-green fluorescent protein virus (Figure 5B). We found that DN-FGFR1 significantly reduced collateral vessel growth and hindlimb perfusion (Figures 5 and 6), demonstrating the importance of the FGF-2/FGFR1 axis during arteriogenesis in vivo.
CA Rap2 Increases Collateral Perfusion and Growth Whereas DN Rap2 Represses Collateral Perfusion and Growth In Vivo
To prove that Rap2 is involved in collateral vessel growth in vivo, we generated adenoviral constructs expressing either DN (Rap2N17) or CA Rap2 (Rap2V12) mutants.12 Expression of DN-Rap2 massively reduced blood flow increase in collateral vessels 3 days after artery ligation. The blood flow of animals that received DN-Rap2 remained low even 14 days after ligation, when control animals had reached almost normal levels (Figure 5A and 5C). In addition, we found that diameters and wall areas of collateral vessels were significantly decreased in DN-Rap2 animals, nearly matching effects caused by DN-FGFR1. In contrast, expression of CA-Rap2 accelerated blood flow recovery resulting in improved perfusion of injected legs 7 days after ligation compared with control animals (Figure 5A and 5C). Moreover, we noted increased diameter and wall area of collateral vessels in injected animals 14 days after femoral artery occlusion (Figure 6A and 6B), which corresponded well to the functional improvement. Taken together, the effects of CA- and DN-Rap2 viruses clearly demonstrated the importance of Rap2 for blood flow recovery after arterial occlusion in vivo.
The formation of a collateral blood flow as response to the occlusion of a large artery is generally not achieved by formation of one large collateral vessel but rather by growth of several already existing arterioles. During the course of collateral artery development, many of the smaller contributing vessels start to regress, which might contribute to an inefficient rearrangement of the collateral circulation. In fact, the maximal blood flow that is restored after femoral artery occlusion is not optimal in larger animals and often reaches less than 50% of the original conductance.1,2 Moreover, there are often strong individual variations within 1 species. The reasons for this phenomenon are not well understood but might be due to ineffective activation of signaling cascades required to promote arteriogenesis. An obvious conclusion from these findings is stimulation of signaling cascades involved in the arteriogenic process to achieve optimal formation of collateral vessels.13 In our current study, we demonstrated that the small GTP-binding protein Rap2 is a crucial part of such a cascade because it is instrumental for migration of VSMCs, an important feature of collateral vessel growth.
Rap2 acts downstream of the Ras/Raf/MEK/Erk pathway, which plays a central role in formation of collateral vessels. We found that the protein level of C-Raf increased significantly during arteriogenesis at a time of accelerated collateral vessel growth and increased conductance, which emphasizes the importance of mitogen-activated protein kinase signal amplification during arteriogenesis. In general, members of the Ras family are activated through binding of growth factors to a transmembrane receptor tyrosine kinase leading to activation of signaling cascades. Interestingly, only FGF-2, which is well known to stimulate smooth muscle cell proliferation, migration, and differentiation, induced Rap2 expression in VSMCs, whereas other growth factors, including insulin-like growth factor-1, transforming growth factor-β1, oncostatin M, and PDGF-AB, failed to exert similar effects. Accumulation of Rap2 protein in VSMCs significantly lagged behind FGF-2-induced expression of Rap2 mRNA, suggesting the presence of an additional, posttranscriptional mode of regulation, which might control the concentration of Rap2. In fact, miRNA-dependent regulatory circuits have been described to control protein accumulation in VSMCs14 although other mechanisms, such as protein destabilization by the ubiquitin ligase system, might also account for this phenomenon. Exogenously administrated FGF-2 has been demonstrated to increase the number of collaterals and improve blood flow in therapeutic as well as experimental studies.15–17 The activity of FGF-2 on arterial smooth muscle cell proliferation depends on FGFR1,8 whereas expression of FGFR1 and the activity of its kinase domain increase during collateral growth.5 Our findings that PAS- and DN-FGFR1-mediated inhibition of FGF signaling markedly reduced blood flow recovery, as well as collateral vessel remodeling, concomitant with repression of Rap2 expression after femoral artery occlusion, demonstrate the importance of the FGFR signaling axis for activation of Rap2 expression and arteriogenesis in vivo. We postulate that activation of Rap2 expression by FGF signaling is instrumental for efficient collateral vessel remodeling because DN-Rap2 reduced increases in blood flow, diameter, and collateral wall area. This hypothesis is also supported by the ability of constitutively active Rap2 to accelerate arteriogenesis resulting in improved collateral blood flow 7 days after ligation, which even slightly surpassed the performance of normal vasculature before arterial occlusion. We assume that this effect is due to increased migratory performance of VSMCs, although it needs to be pointed out that our adenovirus-mediated gene transfer protocol resulted in overexpression of DN-FGFR1, DN-Rap2, and CA-Rap2 in all cells of growing collateral vessels. Hence, it is principally possible that also cell types other than VSMCs contributed to the observed effects.
So far, relatively little is known about Rap2, which belongs to the family of small GTP binding proteins that possess a homology to Ras. The family consists of H-Ras, K-Ras, N-Ras, M-Ras, Rap1, Rap2, TC21, RalA, RalB, and Rheb, which are attached to biological membranes via lipid modifications. The closest relative to Rap2 is Rap1, which is ≈60% identical to Rap2. Rap1 has been claimed to share some effectors with Ras (reviewed in18), whereas its role in the regulation of integrin activation and integrin-dependent adhesion via association with RAPL and RIAM is much better characterized.19,20 Recently, it has also been demonstrated that Rap1 mediates phosphorylation of focal adhesion kinase and vascular endothelial growth factor–induced Akt1 activation and is required for ischemia-induced angiogenesis.7 Rap1 recruits C-Raf to membrane domains, which prevents full activation and thereby antagonizes Ras. Other reports suggest that Rap1, like Ras, activates the MAP kinase cascade (for review, see18). Conflicting reports exist about the function of Rap2, which might be due to context dependent effects, different assay systems, or a combination of both. siRNA-mediated knockdown of Rap2 reduced androgen-mediated DNA synthesis in human prostate cancer cells,21 whereas our own results in VSMCs using a similar approach demonstrated that Rap2 lacks stimulatory or inhibitory effects on the cell cycle. The absence of inhibitory effects on the cell cycle does also concur with the increase of Rap2 after 7 days of collateral vessel growth because this period is characterized by intense proliferation of VSMCs. Rap2, in contrast to Rap1, shows low affinity to GTPase-activating proteins, resulting in a prolonged half-life of the GTP-bound active form in adhering cells, which leads to weaker but extended downstream signaling.22 Our experimental data clearly demonstrate that Rap2 is required for migration of VSMCs without affecting cell proliferation. It is tempting to speculate that an extended activity of Rap2 due to its lower affinity to GTPase-activating proteins might help to keep VSMC in a migratory state even when the initial surge of cell proliferation caused by activation of Ras/Raf/MEK/Erk pathways has subsided.
In summary, our data provide evidence that Rap2 is part of the arteriogenic program and indispensable for VSMC migration. Rap2 seems to be a more attractive target to specifically modulate migration of VSMCs than the FGFR1/Ras/Raf/MEK/Erk pathway, which is broadly active in numerous biological processes (Figure 6C). Increased activity of Rap2 might promote arteriogenesis by enabling migration of VSMCs, whereas its inhibition might interfere with unwanted processes, such as atherosclerosis and restenosis after therapeutic angioplasty, which involves migration of VSMCs.
Sources of Funding
This work was supported by the Max-Planck-Society, the Excellence Cluster Cardiopulmonary System, the University of Giessen-Marburg Lung Center, and the Cell and Gene Therapy Center of the University of Frankfurt.
The authors thank Brigitte Matzke, Jutta Wetzel, and Kerstin Richter for excellent technical assistance and Gerhard Stämmler for statistical analyses. The help of Frederic Pipp, Inka Eitenmüller, and Andreas Kampmann in providing rabbit tissue samples is greatly acknowledged.
- Received January 9, 2010.
- Accepted June 23, 2011.
- © 2011 American Heart Association, Inc.
- Heil M,
- Schaper W
- Deindl E,
- Hoefer IE,
- Fernandez B,
- Barancik M,
- Heil M,
- Strniskova M,
- Schaper W
- Carmona G,
- Gottig S,
- Orlandi A,
- Scheele J,
- Bauerle T,
- Jugold M,
- Kiessling F,
- Henschler R,
- Zeiher AM,
- Dimmeler S,
- Chavakis E
- Ueno H,
- Li JJ,
- Masuda S,
- Qi Z,
- Yamamoto H,
- Takeshita A
- Ito WD,
- Arras M,
- Winkler B,
- Scholz D,
- Schaper J,
- Schaper W
- Ryu J,
- Futai K,
- Feliu M,
- Weinberg R,
- Sheng M
- Yang HT,
- Deschenes MR,
- Ogilvie RW,
- Terjung RL
- Ohba Y,
- Mochizuki N,
- Matsuo K,
- Yamashita S,
- Nakaya M,
- Hashimoto Y,
- Hamaguchi M,
- Kurata T,
- Nagashima K,
- Matsuda M