Summary. Background: Fucoidan, an antithrombotic polysaccharide, can induce endothelial colony-forming cells (ECFC) to adopt an angiogenic phenotype in vitro. Objectives: We evaluated the effect of fucoidan on vasculogenesis induced by ECFC in vivo. Methods: We used a murine hindlimb ischemia model to probe the synergic role of fucoidan-treatment and ECFC infusion during tissue repair. Results: We found that exposure of ECFC to fucoidan prior to their intravenous injection improved residual muscle blood flow and increased collateral vessel formation. Necrosis of ischemic tissue was significantly reduced on day 14, to 12.1% of the gastronecmius cross-sectional surface area compared with 40.1% in animals injected with untreated-ECFC. ECFC stimulation with fucoidan caused a rapid increase in cell adhesion to activated endothelium in flow conditions, and enhanced transendothelial extravasation. Fucoidan-stimulated ECFC were resistant to shear stresses of up to 21 dyn cm−2. Direct binding assays showed strong interaction of fucoidan with displaceable binding sites on the ECFC membrane. Bolus intramuscular administration of fucoidan 1 day after surgery reduces rhabdomyolysis. Mice injected with fucoidan (15 mg kg−1) had significantly lower mean serum creatine phosphokinase (CPK) activity than control animals. This CPK reduction was correlated with muscle preservation against necrosis (P < 0.001). Conclusions: Fucoidan greatly increases ECFC-mediated angiogenesis in vivo. Its angiogenic effect would be due in part to its transportation to the ischemic site and its release after displacement by proteoglycans present in the extracellular matrix. The use of ECFC and fucoidan together, will be an efficient angiogenesis strategy to provide therapeutic neovascularization.
Cardiovascular disease, due essentially to atherosclerosis, causes about half of all deaths in western countries. Peripheral vascular disease of the lower limbs is characterized by chronic obstruction of the arteries supplying the leg, gradually leading to critical limb ischemia. In end-stage disease, severe hypoperfusion results in ischemic ulceration and gangrene, with a high risk of amputation (10%–40%) in the first years . Limb loss is best prevented by revascularization with endovascular or surgical approaches but this is impossible in 30% of patients with critical limb ischemia. No medical treatments have been shown to reduce the amputation rate at 6 months . In this context, activation of local angiogenesis is a promising approach.
Infusion of in vitro-expanded bone marrow progenitor cells can enhance neovascularization in animal models of hindlimb ischemia [3,4]. Among these cells, endothelial progenitor cells are involved in regeneration of injured endothelium and neoangiogenesis after tissue ischemia, and are therefore a candidate cell therapy product [3,5,6]. Different cell populations have been isolated, which play a role in angiogenesis, but only one population, called ‘endothelial colony-forming cells’ (ECFC) has been shown to possess all the characteristics of true endothelial progenitor and to form neovessel in vivo [3,4,7,8]. However, proangiogenic approaches to critical limb ischemia have so far given disappointing results, owing partly to insufficient homing and survival of transplanted progenitor cells at the ischemic site [2,9–13]. The ability of these cells to mobilize and migrate to ischemic sites appears to be an important factor in the success of such therapy [3,9,14,15]. Proangiogenic potential can be enhanced by activating ECFC prior to their injection and thereby amplifying the biochemical signaling cascades that contribute to the production of cytokines involved in the neovascularization process [3,4,6,16–18].
We have previously reported that fucoidan enhances the proangiogenic properties of ECFC in vitro . Fucoidan is a marine sulphated polysaccharide with antithrombotic properties, which contrary to heparin does not significantly increase the bleeding risk, as it has a lower capacity to activate antithrombin . Like proteoglycans, fucoidan interacts with a wide range of proteins and thus exhibits other biological properties including anti-inflammatory activity. It may act as a regulator of tissue remodelling  and has been shown to promote neovascularization when infused intramuscularly, together with the proangiogenic growth factor FGF-2, in a rat model of hindlimb ischemia [20,22]. Here we studied the interaction of fucoidan with infused ECFC during functional neovascularization in an athymic nude mouse model of hindlimb ischemia. We also investigated the mechanisms by which fucoidan enhances angiogenesis.
Cell culture, characterization and pretreatment
Endothelial cells from human umbilical cords (HUVEC) and ECFC from human umbilical cord blood collected from consenting mothers (n = 20), were isolated, expanded and characterized as previously described . The study was approved by the local ethics committee of ‘Hôpital des Instructions et des Armées de Begin’ (201008043234797) and protocol conformed to ethical guidelines of the Declaration of Helsinki. Flow cytometry was used to assess cell surface antigen expression of the endothelial lineage . One day before all experiments, cells were growth-arrested for 24 h in EBM-2, 2% FCS and released from growth arrest by adding EBM-2, 5% FCS, with or without 10 μg mL−1 of fucoidan concentrations for 24 or 72 h at 37 °C, then washed, detached with versene/0.01% collagenase (1/1) and washed twice with buffered Hank’s, 0.5% BSA, before use in angiogenesis and adhesion assays as previously described . Supernatants and cells were separately analyzed after fucoidan stimulation. All the following experiments were performed in triplicate during the first 30 days of culture.
Mouse model of hindlimb ischemia
Animal care conformed to French guidelines (Services Vétérinaires de la Santé et de la Production Animale). The experiments were approved by the IDF Paris Institutional Research Ethics Committee (CREEA, P2.CBV.031.07). Unilateral hindlimb ischemia was induced by femoral artery ligation in 7-week-old male athymic nude mice (Harlan) as previously described . Five hours later the animals received an intravenous injection of normal saline (n = 10), ECFC (105 cells per animal; n = 10) or ECFC incubated in starvation medium overnight then stimulated with fucoidan 10 μg mL−1 for 24 h, and washed (105 cells per animal; n = 10). To evaluate the effect of fucoidan on critical ischemia, three groups of animals underwent the surgical procedure and received two intramuscular injections 1 and 2 days after surgery of normal saline, fucoidan solution (15 mg kg−1) or low-molecular-weight heparin (Lovenox, 2.5 mg kg−1; Sanofi Aventis, Paris, France). Laser Doppler perfusion imaging (Perimed France, Craponne, France) was used to appreciate revascularization on day 14 . Hindlimb macroscopic necrosis was quantified before sacrifice as previously described . A score was calculated for each animal, as follows: 0, no necrosis; 1, necrosis of one toe; 2, necrosis of two or more toes; 3, foot necrosis;, 4, leg necrosis; 5, autoamputation of the entire leg. Vessel density was evaluated by high-definition microangiography Erase (Trophy system) and was expressed as the ratio of the ischemic to non-ischemic leg percentage of pixels per image in the quantification area occupied by vessels . Peripheral venous blood samples were collected 14 days after surgery. The quantitative determination of total creatine phosphokinase (CPK) and lactate dehydrogenase (LDH) concentrations in plasma were determined by enzymatic assays according to procedures provided by Beckman Coulter (Villepinte, France).
Computed-assisted histomorphometric analysis
The gastronecmius muscles from both hindlimbs were fixed in 4% formol and embedded in paraffin. Histological studies were based on computed-assisted analysis of different area (ICS framework viewer; Tribvn, Chatillon, France) as previously described . For each muscle, four sections, 4 μm thick and 300 μm apart, were stained with hematoxylin and eosin (H&E) then digitized for further analysis. The mean of surface area of each tissue type (necrotic, ischemic or healthy) was quantified in each muscle and reported as a percentage of the entire surface area of the section (n = 4). The ratio between the mean entire cross-sectional areas of the ischemic and non-ischemic gastronecmius muscles in each animal was used as an index of muscular trophicity.
Shear-flow adhesion assays
Tethering, rolling and flow adhesion experiments were conducted with a parallel-plate flow chamber in physiological shear stress conditions as previously described . To distinguish between the adhesion of pretreated ECFC and that of detached endothelial cells, ECFC were stained with calcein. HUVEC monolayers (7.5 105) were grown on coverslips, maintained at 37 °C for 6 days, then placed in the flow chamber and stimulated by exposure to a shear rate of 0.4 dyn cm−2 for 30 min. Calcein-labeled ECFC in adhesion buffer (cation-free HBSS, 10 mm HEPES, 1 mm CaCl2, 1 mm MgCl2, 2 mg mL−1 BSA, pH 7.4) were then perfused for 15 min at 37 °C at a shear rate of 0.4 dyn cm−2 and the coverslips were washed with adhesion buffer for 10 min. Adherent cells were visualized by phase-contrast microscopy. All experiments were performed in real time and videotaped for offline analysis. Images from 60 random microscope fields (1 cm2) were collected. Data are expressed as the number of adherent cells per cm2. Detachment of adherent ECFC was studied by increasing the flow rate from 0.4 to 22 dyn cm−2. In some experiments, fucoidan-stimulated or untreated ECFC were pretreated for 2 h at 37 °C with 0.5 U mL−1 heparinase I, 0.1 U mL−1 heparinase III and 0.2 U mL−1 chondroitinases ABC (Sigma-Aldrich, Saint-Quentin Fallavier, France).
Rat aortic ring assay
Angiogenesis was also studied with aortic explants cultured in three-dimensional matrix gels as described by Zhu et al. . Experiments conformed to the guidelines of the University Paris-Descartes Institutional Animal Care and Use Committee (C75.06.02). The thoracic aorta was isolated from 8-week-old male Wistar rats (Charles River Breeders; Charles River, L' Arbresle, France), immediately transferred to a culture dish containing warm serum-free minimum essential medium, and sliced into 1–2-mm rings after carefully removing fibroadipose tissue. The aortic rings were rinsed and embedded in rat tail type I collagen gel. The gels were overlaid with DMEM-10% FCS supplemented with 100 U mL−1 penicillin and 100 μg mL−1 streptomycin. Fucoidan (10 μg mL−1), FGF-2 (50 ng mL−1) or vascular endothelial growth factor (VEGF) (40 ng mL−1) was added to the medium in the appropriate wells. The medium was changed on days 3, 6 and 9 of culture. Angiogenesis was scored by counting the number of branching endothelial sprouts.
Data are expressed as means ± SEM of at least three independent experiments. Significant differences were identified by anova and Fisher’s test. The Statview software package (SAS France, Brie Comte Robert, France ) was used for all analyses. Differences with probability values below 0.05 were considered significant.
Having previously shown that fucoidan can induce ECFC to adopt a proangiogenic phenotype in vitro , we first assessed the ability of fucoidan-stimulated ECFC to promote neovascularization in a mouse model of hindlimb ischemia. We focused on ECFC because this cell type is currently proposed as the cell type at the origin of newly formed vessels . Cells were isolated from human umbilical cord blood on the basis of CD34 expression. The presence of Weibel-Palade bodies and combined expression of endothelial markers (CD31,Tie-2,KDR,Flt-1,CD144) unequivocally confirmed the endothelial phenotype of the ECFC thus obtained. Furthermore, ECFC do not express leuko-monocytic markers such as CD14 and CD45 .
Mice treated with fucoidan-stimulated ECFC are protected from ischemia-induced necrosis
Respectively, 30% and 25% of mice that received normal saline or untreated ECFC exhibited macroscopic necrosis (Fig. 1A). No mice treated with fucoidan-stimulated ECFC had macroscopic necrosis (P < 0.05 vs. CTRL, n = 10).
Hindlimb perfusion was, respectively, 34% and 99% higher with fucoidan-stimulated ECFC than with untreated ECFC (P < 0.001) and normal saline (P < 0.001) (Fig. 1B). The ischemic/non-ischemic limb angiography score was 11% higher with fucoidan-stimulated ECFC than with untreated ECFC, but the difference was not statistically significant (Fig. 1C).
Fucoidan-stimulated ECFC protect ischemic tissue against necrosis
The results of histomorphometric analysis of distal gastronecmius muscle sections on day 14 in mice treated with normal saline, control ECFC and fucoidan-stimulated ECFC are shown in Fig. 2. Digitized muscle sections were divided into three areas: (i) preserved areas with a normal histology (Fig. 2C), (ii) ischemic areas containing cellular alterations but no necrosis (Fig. 2D) and (iii) necrotic tissue with muscle cell destruction (Fig. 2E). Animals treated with fucoidan-stimulated ECFC showed less necrosis (P < 0.01; Fig. 2E) and better muscle preservation than animals treated with control ECFC (P < 0.05; Fig. 2C) or with saline solution alone (P < 0.01; Fig. 2C).
Cellular and molecular mechanisms involved in fucoidan enhancement of proangiogenic activity
We next examined the potential mechanisms by which fucoidan-stimulation improved proangiogenic properties of ECFC. We investigated whether ECFC stimulation with fucoidan enhances their homing to ischemic tissue via ECFC adhesion to activated endothelium, their extravasation and differentiation into a vascular network.
Fucoidan-stimulated ECFC show enhanced adhesion to activated HUVEC in flow conditions
We used a flow-based adhesion assay with HUVEC to investigate the effect of fucoidan stimulation on treated ECFC adhesion to activated endothelium in conditions simulating the shear forces encountered by ECFC when they adhere to vascular endothelial cells. In our experimental conditions, fucoidan-stimulated ECFC adhered much more rapidly than control ECFC. As shown in Fig. 3(A), short ECFC exposure to fucoidan caused a rapid increase in adhesion to HUVEC (1000 treated cells cm−2 vs. 545 control cells cm−2; P < 0.001). Fucoidan-stimulated ECFC adhered firmly to the endothelium and were more resistant than control cells to washing at shear rates up to 21 dyn cm−2 (Fig. 3B). This property is not correlated to fucoidan stimulation on cell surface integrins αLβ2 (LFA-1), αMβ2 (MAC-1) or α4β1 (VLA4) expression levels and gene expression (data not shown).
Fucoidan-stimulated ECFC show enhanced transmigration through activated HUVEC and enhanced differentiation into vascular tubes
Fluorescence microscopy of labelled fucoidan-stimulated ECFC that migrated through the activated endothelium showed that fucoidan enhanced ECFC motility by a factor of 2.5 compared with 40 ng mL−1 VEGF (P < 0.01) (Fig. 3C). Similar results were obtained when VEGF was present in the bottom well of the chamber. Fucoidan enhanced ECFC transmigration through activated HUVEC monolayers towards VEGF (P < 0.01). Fucoidan also enhanced the proangiogenic activity of heparin growth factors present in EGM-2 culture medium (P < 0.05) (Fig. 3D). Real-time RT-PCR showed no change in mRNA levels of the proangiogenic cytokines VEGF-A, FGF-2, angiopoietin and SDF-1 (data not shown). In the same experimental conditions, no significant increase in proangiogenic cytokine release was observed in media conditioned by fucoidan-stimulated ECFC, as measured by ELISA. However, gelatine zymography of the same conditioned medium revealed that tube formation coincided with increased secretion of pro-MMP-9 (2-fold, P < 0.05; Fig. 4A,B).
Fucoidan binds to the ECFC surface
Fluorescent DTAF-fucoidan that retains its full proangiogenic activity bound rapidly to the ECFC outer membrane in a time-dependent manner (Fig. S1A) and was internalized within 15 min at 37 °C, probably by endocytosis. After 24 h of incubation, DTAF-fucoidan was found in endocytotic vesicles that localized in the perinuclear region (Fig. S1B). DTAF-fucoidan interacted with displaceable binding sites on ECFC, as heparin and cold fucoidan both induced concentration-dependent displacement of labelled fucoidan when added to the culture medium (P < 0.001; Fig. S1C).
Bound fucoidan acts as a glycosaminoglycan
Because fucoidan is a sulphated polysaccharide and as heparan-sulphate proteoglycans (HSPG) play key roles in the angiogenic process, presenting chemokines to circulating cells, we investigated the role of fucoidan bound to the ECFC surface in ECFC adhesion and differentiation processes. ECFC were treated for 120 min with enzymes that selectively degrade either heparan sulphate or chondroitin sulphate, and were then incubated with 10 μg mL−1 fucoidan at 37 °C overnight. Enzyme treatment abolished ECFC adhesion to the endothelium in flow conditions (−65%, P < 0.001; Fig. 3A) and also reduced tube formation in Matrigel (−50%, P < 0.01). Preincubation of enzyme-treated cells with fucoidan enhanced ECFC adhesion by 82% (342 cells cm−2 with fucoidan-stimulated ECFC without HSPG [HSPGd] vs. 187 cells cm−2 with HSPGd-ECFC; P < 0.001) (Fig. 3A). These fucoidan-pretreated HSPGd-ECFC showed poor resistance to detachment when tested at shear stresses up to 20 dyn cm−2 (Fig. 3B). This fucoidan pretreatment had no effect on HSPGd-ECFC extravasation or tube formation (data not shown).
Fucoidan promotes angiogenesis in the aortic ring model
Rings of rat aorta embedded in collagen gel immediately after excision generate branching microvessels, and this process was enhanced by FGF-2 and VEGF in a concentration-dependent manner, as previously reported . Unstimulated control rings exhibited only rare and short endothelial sprouts. As shown in Fig. 5, aorta rings treated with FGF-2 (Fig. 5A) and VEGF (Fig. 5B) showed approximate 2.5- and 2-fold increases in microvessel outgrowth, respectively, after 9 days of culture. Fucoidan significantly increased the number of neovessel sprouts induced by FGF-2 (Fig. 5A; P < 0.01) but not by VEGF (Fig. 5B). The distance between the ring and the edge of the cell front around the ring, which reflects both cell proliferation and migration, was 1.4-fold longer in the presence of both fucoidan and FGF-2 than in the presence of FGF-2 alone (Fig. 5C).
Fucoidan protects ischemic tissue against necrosis better than heparin
Histomorphometric analysis of distal gastronecmius muscle sections on day 14 in mice treated with intramuscular injection of normal saline, low-molecular-weight heparin (2.5 mg kg−1) and fucoidan (15 mg kg−1) is shown in Fig. 6. Animals treated with fucoidan showed less necrosis and better muscle preservation than animals treated with saline solution alone and heparin-treated animals (P < 0.001 and P < 0.01, respectively; Fig. 6A–C). This correlates with a significant decrease of blood levels of CPK 14 days after surgery (Fig. 6D). Serum CPK and LDH are enzymes whose levels can increase significantly when muscle damage or inflammation has occurred. In the control animals, a serum CPK activity of 2200 ± 600 per IU was observed. Fucoidan pretreatment markedly attenuated the CPK activity (≈70%P = 0.002, ≈60%P = 0.0320) compared with the control and heparin groups. No significant changes were observed regarding the serum lactate dehydrogenase activity.
Fucoidan, an antithrombotic polysaccharide, can induce ECFC to adopt an angiogenic phenotype in vitro . Human ECFC treatment with fucoidan induced a proangiogenic phenotype, involving both an early angiogenic event (proliferation and migration) and a late event (differentiation into vascular cords). This study shows that the beneficial effect of ECFC infusion in a mouse model of hindlimb ischemia is amplified significantly by fucoidan pretreatment, preventing necrosis. This tissue protection is associated with enhanced neoangiogenesis and reduction of rhabdomyolysis.
The angiogenic process involves ECFC rolling and adhesion to activated endothelium and their extravasation to target sites. Our results provide strong evidence that fucoidan prestimulation enhanced each of these steps, by inducing pro-MMP-9 secretion involved in cell homing . These actions were associated with direct fucoidan interaction with the ECFC surface. Fucoidan bound to the cell membrane and was internalized. We have previously shown that fucoidan stimulation leads to ECFC actin cytoskeleton changes, which may be involved in extravasation . Experiments underway in our laboratory indicate that fucoidan interaction is partially inhibited by blocking antibodies against P-selectin, CD11b (αM) and CD18 (β2) (30–40%, P < 0.001). This is in keeping with the reported high affinity of selectin and integrin αMβ2 for a high-molecular-weight commercial fucoidan preparation [26,27]. Because of its spatial structure, fucoidan can mimic the clustering of sulphated, sialylated and fucosylated oligosaccharides on the cell surface and can provide the appropriate structural backbone for selectin binding. Fucoidan interaction with P-selectin does not affect the ECFC arrest on activated endothelium in flow conditions, as VLA4 (α4β1) can mediate both tethering and firm adhesion . The possible interaction of αMβ2 or other integrins with fucoidan remains to be addressed. We found that fucoidan interacted with displaceable binding sites on ECFC, allowing it to be released in the ischemic tissue environment after displacement by heparan sulphate present in extracellular matrix. There, it can interact with heparin-like growth factors and enhance neoangiogenesis through the transduction of intracellular signals required to induce the angiogenic phenotype. Its effect on signaling pathways is under investigation.
Tissue ischemia is also accompanied by inflammation and fucoidan has been shown to possess anti-inflammatory properties . Parenchymal tissue protection against inflammatory processes and myofibroblastic remodelling have been observed after daily subcutaneous injection of fucoidan in a rat cardiac allograft model . During ischemia, hindlimb muscle destruction causes permeabilization of muscle fibers and the release of their contents into the bloodstream, including muscle enzyme such as CPK. Serum CPK activity will thus reflect the magnitude of acute muscle damage due to ischemia. Bolus administration of fucoidan injected directly in the quadriceps muscle of ischemic mice was shown to induce 70% and 60% reduction in elevated CPK activities compared with, respectively, control and heparin mice. This could account for the anti-necrotic effect observed here, although further investigations are needed. This effect seems to be specific to the polyfucose; intramuscular injection of low-molecular-weight heparin in the same experimental conditions does not protect the muscle from necrosis. This is in keeping with previous studies showing the important role of the fucosyl backbone in fucoidan properties . Finally, in the aortic ring model, the angiogenic activity of fucoidan in the presence of FGF-2 stems from its ability to potentiate FGF-2 proangiogenic activity. These results corroborate the ones observed by Luyt et al.  showing that fucoidan can induce angiogenesis in vivo by modulating the proangiogenic properties of heparin-binding growth factor FGF-2. However, regeneration of ischemic tissue was only observed when fucoidan was associated with FGF-2  (C. Boisson-Vidal, unpublished data). Therefore, the angiogenic potency of the fucoidan-stimulated ECFC is likely to be attributable to both the beneficial effects of fucoidan on the proangiogenic potential of transplanted ECFC and the influence of fucoidan and ECFC on the ischemic hindlimb.
So far, efforts to stimulate angiogenesis have been limited to the introduction of exogenous growth factors such as FGF-2 or VEGF by means of protein delivery or gene transfer. Our results indicate that a similar response could be obtained with a sulphated polysaccharide. Of particular interest, fucoidan partially restored the adhesion of ECFC to activated endothelium after their treatment with enzymes that remove HSPG located at the cell surface or in the extracellular matrix and that play crucial roles in physiological processes and tissue remodelling . Fucoidan shares several characteristics with HSPG, including sulphate groups and short polysaccharide chains. Owing to its ionic structure, fucoidan also mimics some properties of heparan sulphate and can thus bind proangiogenic heparin-binding growth factors, thereby increasing local concentrations at the cell surface or protecting these factors against enzymatic degradation [20,31]. We cannot rule out the possibility that free and cell-associated fucoidan may bind proangiogenic growth factors and confer a stable, receptor-compatible conformational change upon growth factors, which might increase their affinity for their receptors and improve the angiogenesis process. Fucoidan can modulate the biological activity of growth factors and disrupt heparan sulphate-growth factor/cytokine complexes, thus releasing chemokines sequestered by heparan sulphate [19,20,32]. Intramuscular fucoidan injection increases plasma levels of SDF-1 and is thus able to promote ECFC mobilization [20,32]. Fucoidan can also act on surrounding cells such as smooth muscle cells, neutrophils and/or macrophages attracted by cytokines released by the ischemic site.
Figure 7 summarizes the multiple roles of fucoidan in therapeutic angiogenesis and in the proangiogenic activity of transplanted ECFC. Fucoidan immobilization on transplanted ECFC allows these cells to home rapidly to ischemic sites, by enhancing ECFC adhesion to activated endothelium and extravasation. These effects involve upregulating the secretion of pro-MMP-9 involved in stem cell homing. Locally released fucoidan binds heparin-like growth factors and regulates their mitogenic activity, leading to increased angiogenesis. Fucoidan can also specifically displace sequestered SDF-1 from its HSPG anchors and thus contribute to its release into the circulation, promoting ECFC mobilization.
In conclusion, our results show that the therapeutic effects of fucoidan-stimulated ECFC are due to both individual and synergistic effects of the two agents. The fucoidan-ECFC tandem warrants further studies as a means of promoting revascularization of ischemic tissues and as a potential candidate for cell therapy in peripheral ischemia.
G. Sarlon, F. Zemani and C. Boisson-Vidal designed and performed the research, analyzed the data and wrote the manuscript. L. David: rat aortic ring experiments, data collection and image analysis. J. P. Duong Van Huyen: computed-assisted histomorphometric analysis, data collection and interpretation of the results. B. Dizier: in vivo experiments, data collection and Doppler analysis. F. Grelac: zymography, data collection and image analysis. S. Colliec-Jouault: obtained funding. I. Galy-Fauroux: cell culture supervisor. P. Bruneval: histomorphometric study supervisor. J. Emmerich, A. M. Fischer: critical revisions of the manuscript for important intellectual content.
We thank A. Lokajczyk for technical assistance, J. Vilar and J. S. Silvestre for angiographic assays, M. Tible for completing computed-assisted histomorphometric analyses, C. D’Audigier, D. Smadja and the Biochemistry Laboratory of HEGP for evaluating levels of serum muscular enzymes, and C. Martin and all the technicians from the animal facilities of IMTCE. We are indebted to the nursing services of Hôpital des Diaconnesses (Paris, France) and Hôpital des Instructions et des Armées de Begin (Saint Mandé, France) for providing umbilical cord blood samples. This work was supported by funds from the French Research Institute for Exploitation of the Sea (IFREMER, GDR-08/5210007) and from Fondation de France (RAF05030KKP). F. Zemani was the recipient of grants from Groupe d’Etude et de Recherches sur l’Hémostase and the French Hematology Society. G. Sarlon was supported by the Société de Chirurgie Vasculaire de Langue Française. CNRS paid the salary of C. Boisson-Vidal. Therapol paid the salary of L. David.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.