PPARα Is Essential for Microparticle-Induced Differentiation of Mouse Bone Marrow-Derived Endothelial Progenitor Cells and Angiogenesis

Background Bone marrow-derived endothelial progenitor cells (EPCs) are critical for neovascularization. We hypothesized that microparticles (MPs), small fragments generated from the plasma membrane, can activate angiogenic programming of EPCs. Methodology/Principal Findings We studied the effects of MPs obtained from wild type (MPsPPARα+/+) and knock-out (MPsPPARα−/−) mice on EPC differentiation and angiogenesis. Bone marrow-derived cells were isolated from WT or KO mice and were cultured in the presence of MPsPPARα+/+ or MPsPPARα−/− obtained from blood of mice. Only MPsPPARα+/+ harboring PPARα significantly increased EPC, but not monocytic, differentiation. Bone marrow-derived cells treated with MPsPPARα+/+ displayed increased expression of pro-angiogenic genes and increased in vivo angiogenesis. MPsPPARα+/+ increased capillary-like tube formation of endothelial cells that was associated with enhanced expressions of endothelial cell-specific markers. Finally, the effects of MPsPPARα+/+ were mediated by NF-κB-dependent mechanisms. Conclusions/Significance Our results underscore the obligatory role of PPARα carried by MPs for EPC differentiation and angiogenesis. PPARα-NF-κB-Akt pathways may play a pivotal stimulatory role for neovascularization, which may, at least in part, be mediated by bone marrow-derived EPCs. Improvement of EPC differentiation may represent a useful strategy during reparative neovascularization.


Introduction
Bone marrow-derived endothelial progenitor cells (EPCs) possess the capacity to proliferate, migrate, and differentiate into endothelial lineage cells [1]. EPCs are mobilized from bone marrow into the circulation during vascular injury, and then home to the site of neovascularization and thereby contribute to postnatal neovascularization [2][3][4]. Moreover, decreased levels of EPCs correlate with subsequent increases in cardiovascular events suggesting that EPCs are important predictors for cardiovascular mortality and morbidity [5,6]. It has also been proposed that other populations of bone marrow-derived cells that are phenotypically different from EPCs can also be recruited to the injured endothelium and contribute to the regeneration of the endothelium [7]. Characterization of these cells indicates that they mainly express hematopoietic lineage markers such as CD11b, CD14 and CD45 [8], whereas EPCs express CD31, CD106, and Flk-1 [9].
One potential limitation for the use of autologous bone marrow EPCs for therapeutic vasculogenesis is the decline in the number and function of progenitor cells in patients with coronary artery disease [10], diabetes [11], metabolic syndrome [12] and severe heart failure [13]. Therefore, enhancing EPC differentiation and homing towards sites of injury will likely improve efficiency of cell therapy; this approach is a subject of intense investigation. Emerging evidence indicates that PPARa, a member of the nuclear hormone receptor superfamily activated by ligands and several fatty acids [14], improves nitric oxide (NO)-mediated vasodilatation in conductance and resistance vessels and also protects against myocardial ischemic injury [15].Recent reports suggest that stimulation of PPARa, but not PPARb or PPARc, enhances cardiac progenitor differentiation through the involvement of NADPH oxidase activity [16,17] and that activation of PPARa induces neovascularization through a VEGF-dependent mechanism [18,19]. Collectively these studies suggest that stimulation of PPARa may stimulate cell differentiation.
The role of cell-to-cell communication involving membrane fragments such as microparticles (MPs) [20,21] during neovascularization is unknown. MPs are small vesicles released from the plasma membrane of stimulated or apoptotic cells; their protein composition is not completely elucidated, but it differs in function depending on the cell origin and the stimuli for their generation [22]. Importantly, MPs can transfer receptors and organelles between cells and deliver mRNA and proteins into cells [20]. Thus, MPs can be considered as real vectors of biological messages such as induction of angiogenesis or differentiation since endothelial, platelet or lymphocytic MPs can modulate in vitro and in vivo angiogenic properties of EPCs and ECs [23][24][25][26]. However, the mechanism(s) by which circulating MPs modulate angiogenesis have not elucidated. Because PPARs can be harbored by MPs [27], we hypothesized that the effects of circulating MPs on EPC differentiation may be mediated, at least in part, by the PPARa pathway. Therefore, the aim of this study was to determine whether circulating MPs induce bone marrow-derived cell differentiation towards an endothelial lineage and increase angiogenic-related properties of these cells and ECs.

Results
The number and origin of circulating MPs does not differ between wild type (WT) and PPARa null mice Flow cytometry analysis of circulating MPs ( Figure 1A and B) showed that the total number of circulating MPs was not significantly different between PPARa null mice (20566 360 MPs/mL plasma) and their corresponding WT littermates (16786417 MPs/mL plasma), indicating that PPARa deletion did not affect the circulating levels of MPs ( Figure 1C). These MP concentrations were used for further experiments. There was also no significant differences in the cellular origins of MPs between the two groups of mice ( Figure 1C) with only MPs from PPARa WT (MPs PPARa+/+ ) expressing PPARa receptors ( Figure 1D). Circulating MPs PPARa+/+ , but not MPs PPARa2/2 , increase the differentiation of bone marrow-derived cells towards the EPC lineage Phenotypic characterization of bone marrow-derived cells using multiple markers indicates that, at day 7, the presence of populations of bone marrow-derived cells displaying double labeling of Sca-1 + /CD31 + , Sca-1 + /CD45 + , Sca-1 + /CD14 + and F4/80 + / CD31 + whose levels were enhanced compared to those at day 0 ( Figure S1). It should be noted that the percentages of bone marrowderived cells expressing Sca-1 + /Flk-1 + and CD133 + /Flk-1 + were not significantly different at day 0 and 7. This characterization indicates that, at day 7, bone marrow-derived cells are mainly cells expressing monocytic and endothelial markers; and moreover that the EPC subpopulation belongs to the so-called ''early EPCs''.
To investigate the effects of PPARa carried by MPs on in vitro bone marrow-derived cell differentiation, circulating MPs isolated from PPARa-deficient or WT mice were added to adherent bone marrowderived cells from WT mice. Thereafter, EPCs were identified by Dil-Acetyl-LDL-uptake with concomitant lectin binding, as previously described [28], and double expression of Sca-1 and Flk-1 determined. After 7 days, isolated bone marrow-derived cells cultured in the presence of MPs PPARa+/+ displayed enhanced adhesion but with no changes in cell proliferation (data not shown). At day 7, confocal microscopy characterization of adherent cells incubated with MPs PPARa+/+ , but not with MPs PPARa2/2 , also demonstrated increased Dil-ac-LDL/lectin uptake typical for endothelial lineage ( Figure 2A). As illustrated in Figure 2B, quantification indicates that only MPs PPARa+/+ significantly increased the number of doublepositive (Dil-Ac-LDL + /lectin + ) cells. Similar results were obtained by flow cytometry, revealing double-positive staining with Sca-1/Flk-1. Indeed, MPs PPARa+/+ increased the number of differentiated EPClike attaching cells by 36% ( Figure 2C). In contrast, the number of Sca-1 + Flk-1 + cells were significantly lower in EPCs treated with MPs PPARa2/2 when compared with non-treated EPCs ( Figure 2C). In addition, MPs PPARa+/+ had no effect on the number of cells expressing both F4/80 and CD31 markers whereas the number of F4/80 + CD31 + cells treated with MPs PPARa2/2 were significantly reduced ( Figure 2D).

Pharmacological PPARa activation induces EPC differentiation
To evaluate the presence of a functional PPARa pathway in bone marrow-derived cells and also whether direct activation of PPARa leads to EPC differentiation, we first analyzed PPARa expression in bone marrow-derived cells and secondly, we used a selective PPARa agonist, WY-14643 (5 mM) to stimulate EPC differentiation of cells isolated from WT mice. As shown by Western blot, PPARa was expressed on bone marrow-derived cells at day 7 of culture ( Figure S2A) and that PPARa levels were not modified in bone marrow-derived cells after incubation with MPs PPARa+/+ ( Figure S2A), suggesting that PPARa transfer between MPs and bone marrow-derived cells did not take place. Interestingly, the PPARa agonist significantly increased (,42%) EPC differentiation from WT but not from PPARa-null mice as assessed by Sca-1 + /Flk-1 + staining ( Figure S2B and S2C). In addition, when bone marrow-derived cells from PPARa-null mice were treated with MPs from WT mice, no changes in EPC differentiation were observed ( Figure S2D). Altogether, these results demonstrated the presence of functional PPARa receptors on bone marrow-derived cells from WT mice that induce differentiation and that it was unlikely that PPARa protein was transferred from MPs to bone marrow-derived cells.
MPs PPARa+/+ -induced EPC differentiation is associated with up-regulation of pro-angiogenic factor mRNA levels To confirm and extend the results obtained by immunofluorescence and flow cytometry, we further analyzed the expression of differentiation marker genes profiles by q-PCR. Mice endothelial markers and angiogenic gene expression were analyzed using a panel of 52 transcripts (Table S1). The data in Figure 3, which was obtained from cultured bone marrow-derived cells, shows that MPs PPARa+/+ enhances the expression of pro-angiogenic factors such as VEGF A, SDF-1, and angiopoietin-2, while the expression of thrombospondin-1 (TSP-1), which has a potent anti-angiogenic activity, was significantly reduced. In addition, mRNA levels of endothelial markers such as VE-cadherin, PECAM-1 and ICAM-1 were increased. Any significant changes of mRNA levels of proangiogenic factors and endothelial markers were observed when bone marrow-derived cells were treated with MPs PPARa2/2 .

MPs PPARa+/+ enhance EC-specific marker expression and Akt activity
We investigated the molecular mechanisms underlying the effects of MPs on EPC differentiation. To further confirm the effects of MPs harboring PPARa on bone marrow-derived cells differentiation at day 7 of growth, we examined the protein expression levels in cellular extracts of pro-angiogenic and EC specific markers and Akt activity in the bone marrow-derived cells ( Figure 4). Consistent with the results obtained by q-PCR, Western blot analysis revealed that MPs PPARa+/+ significantly enhanced VEGF A but not its receptor FLT-1. Besides, ICAM-1 but not eNOS expression was significantly enhanced upon MPs PPARa+/+ stimulation. In contrast, MPs isolated from PPAR a -deficient mice significantly decreased the expression of eNOS, VEGF and FLT-1. To demonstrate whether Akt signalling mediates EPC differentiation, bone marrow-derived cells from PPARa WT mice were incubated in the absence or presence of either MPs PPARa+/+ or MPs PPARa2/2 , and Akt phosphorylation on Ser 473 indicative for Akt activity was detected by Western blot analysis. MPs PPARa+/+ , but not MPs PPARa2/2 , stimulated Akt phosphorylation indicating an increased activity of Akt in differentiating bone marrow-derived cells ( Figure 4).

MPs PPARa+/+ favor in vitro capillary-like structure formation of ECs
To examine the effects of MPs on the capacity of ECs to form capillary-like structure, ECs were seeded on MatrigelH for 24 hours in the absence or presence of either MPs PPARa+/+ or MPs PPARa2/2 . In the absence of stimulation, ECs from both PPARa WT and KO mice form capillary-like structures suggesting that deletion of PPARa did not affect the angiogenic capacity of ECs ( Figure S3A and S3B). Interestingly, MPs PPARa+/+ induced significant increases in the number of capillary-like structures (,55%) relative to untreated ECs from PPARa WT mice ( Figure 5A and 5B). To determine whether PPARa accounts for the effects induced by MPs PPARa+/+ , we assessed the effects of PPARa deletion on MP-induced capillary formation. Capillary formation was absent abolished by MPs PPARa2/2 ( Figure 5A), suggesting that PPARa harbored by MPs is essential for MP-induced capillary-like structure formation of ECs.

NF-kB pathway is involved in MP-induced EPC differentiation and in vitro EC angiogenesis
NF-kB is a transcription factor that is complexed with I-kBa and b in the cytoplasm, which are selectively phosphorylated, ubiquinated, and degraded, leading to NF-kB translocation. Phosphorylation of I-kBa results in the activation of NF-kB and so allows its translocation to the nucleus.
Western blot analysis of bone marrow-derived cell extracts demonstrated that MPs PPARa+/+ , but not MPs PPARa2/2 , significantly enhanced I-kBa phosphorylation on Ser32/36 ( Figure 6A), highlighting the activation of NF-kB signaling. To elucidate whether activation of the NF-kB pathway is related to EPC differentiation induced by MPs PPARa+/+ , we used BAY11-7082 (5 mM), an inhibitor of I-kBa phosphorylation. As shown in Figure 6B, MPs PPARa+/+ significantly increase the number of double-positive EPCs for Sca-1 + /Flk-1 + staining. These effects were prevented by incubation with BAY11-7082, suggesting a potential role of the NF-kB pathway in EPC differentiation induced by MPs PPARa+/+ . In addition, the weak decrease in EPC differentiation induced by MPs PPARa2/2 was not modified by BAY11-7082 ( Figure 6B) suggesting that NF-kB pathway is not implicated in the effects of MPs PPARa2/2 .
In another set of experiments, aortic ECs isolated from WT mice were preincubated with BAY11-7082 for 30 minutes, and then later incubated for 24 hours in the absence or presence of MPs PPARa+/+ . As observed during EPC differentiation experiments, BAY11-7082 also abolished MPs PPARa+/+ -induced capillary network formation of mature ECs on MatrigelH ( Figure 6C), suggesting the involvement of the NF-kB pathway in in vitro effects of MPs PPARa+/+ on EC-mediated angiogenesis.
MPs PPARa+/+ promote in vivo bone marrow-derived cell-mediated angiogenesis Finally, to investigate the effects of MPs on the formation of the new vessels, in vivo angiogenesis was studied by using MatrigelH plug assay. At day 7, a blush of vessel proliferation was clearly observed in plugs from mice injected with bone marrowderived cells treated with MPs PPARa+/+ , but not with MPs PPARa2/2 ( Figure 7A). These effects were further quantified by measuring the hemoglobin content of the plugs. Vascularization of the plugs was significantly enhanced (,25%) in mice injected with bone marrow-derived cells treated with MPs PPARa+/+ when compared with those untreated or treated with MPs PPARa2/2 ( Figure 7B). No further effect was observed at day 15 (not shown). Conversely, when compared with control, hemoglobin content of new-formed blood vessels was significantly reduced (,20%) in plugs mixed with MPs PPARa2/2treated bone marrow-derived cells. Vascularization of the plugs induced by MPs PPARa+/+ was abolished when bone marrowderived cells were pre-treated with an inhibitor of NF-kB pathway, suggesting a key role of NF-kB activation in the effects evoked by MPs PPARa+/+ (Figure 7A and 7B).

Discussion
Our study demonstrates for the first time that PPARa harbored by MPs mediate the functionality and the differentiation of bone marrow-derived EPCs. The major findings of our study are that MPs PPARa+/+ : (i) increase EPC differentiation and EC marker expression, but not those of macrophage differentiation; (ii) enhances EPC differentiation that is mediated, at least in part, by PPARa-NF-kB and Akt pathways; (iii) modulate in vitro angiogenic-related properties of ECs; and (iv) stimulate in vivo bone marrow-derived cell-associated neovascularization. Altogether, these findings provide a novel mechanism underlying EPC differentiation and angiogenesis. Bone marrow-derived cells contribute to tissue regeneration. In particular, EPCs and macrophages play critical roles in angiogenesis and potentially help to repair injured endothelium [29,30]. In this study, we show that, after 7 days of culture, bone marrow-derived cells are mainly EPCs (Sca-1 + CD31 + ) and monocyte/macrophages (Sca-1 + CD14 + and F4/80 + CD31 + ), but after MPs PPARa+/+ treatment, EPC but not macrophage differentiation was observed.
Several studies confirm that MPs favor angiogenesis. Whereas platelet-derived MPs promote angiogenesis through ERK pathway activation [25], endothelial-derived MPs increase capillary-like structure formation via plasminogen generation [23]. We recently reported that MPs from apoptotic/activated lymphocytes evoke angiogenesis through the up-regulation of adhesion and proangiogenic factors [26]. However, no data are available concerning circulating MPs. Here, we show that circulating MPs from WT mice favor both in vitro and in vivo angiogenesis through the modulation of EPC differentiation and angiogenic function, while MPs PPARa2/2 decreased both EPC and macrophage differentiations. We cannot rule out the possibility that PPARa deletion can induce other cellular modifications, as previously shown in T cells from PPARa KO mice [31,32]. Although the mechanisms by which MPs PPARa+/+ induces both EPC differentiation and neovascularization are not completely elucidated, we demonstrated that effects evoked by MPs PPARa+/+ involve NF-kB signaling that was associated with Akt activation. Moreover, PPARa is not transferred from MPs into bone marrow-derived cells since PPARa expression is not enhanced in these cells after treatment with MPs PPARa+/+ . On the other hand, PPARa is expressed and functional in bone marrow-derived cells, suggesting that lipids harbored by MPs PPARa+/+ , but not those carried by MPs PPARa2/2 , may act as endogenous ligands. In agreement with previous studies [33,34], it is possible that mRNA or miRNA delivery from MPs PPARa+/+ to target cells (in this case bone marrow-derived cells or ECs) may be responsible for changes of their phenotype. In this context, it has been described that MPs from EPCs may activate angiogenic program in ECs [34]. Further experiments are necessary to determine the molecular mechanism by which MPs transfer their message. Activated PPARa is required during in vitro differentiation of murine stem cells towards cardiac cells [16,17] while PPARa activators induce keratinocyte differentiation [35] and phenotypic modulation of rat oval cells toward the hepatocyte lineage [36]. Biscetti and coworkers [19] recently reported that iloprost-induced angiogenesis and VEGF up-regulation are dependent on the presence and proper function of PPARa gene. Besides, PPARc agonists favorably modulated bone marrow-derived progenitor cells to promote endothelial lineage [37]. Our findings demonstrate that the PPARa gene is necessary for MPs to develop proangiogenic activity since circulating MPs form PPARa deficient mice did not induce either EPC differentiation, in vivo neovascularization or in vitro EC angiogenesis. In addition, the selective pharmacological PPARa agonist WY-14643 evoked EPC differentiation in bone marrow-derived cells from WT PPARa mice, but not in those derived from KO PPARa mice, suggesting that PPARa activation accounts for pro-angiogenic effects on EPCs.
To elucidate the potential molecular mechanisms underlying EPC differentiation and pro-angiogenic properties of MPs harboring PPARa, we evaluated the expression of pro-angiogenic factors at the mRNA level in ex vivo expanded bone marrowderived cells pretreated with MPs PPARa+/+ or MPs PPARa2/2 . A significant increase of mRNA levels of pro-angiogenic factors, such as VEGF A, Ang-2 and SDF-1 was observed in bone marrowderived cells treated with MPs PPARa+/+ , suggesting that MPs from PPARa mice may be involved in the enhanced homing ability of EPCs and in their pro-angiogenic abilities. Since SDF-1 pretreatment during EPC expansion stimulates EPC adhesion and increases the efficiency of cell therapy for ischemic vascular diseases, increases EPC number and recruitment of bone marrowderived EPCs into site of ischemia [38], one can advance the hypothesis that MPs PPARa+/+ can be used as an expansion agent of EPCs. In addition, expressions of EC markers such as VEcadherin, PECAM-1, and ICAM-1 were increased, indicating that MPs PPARa+/+ evoke differentiation of bone marrow-derived cells towards an endothelial lineage. In contrast, TSP-1 which possesses potent anti-angiogenic activity [39] was down-regulated in bone marrow-derived cells treated by MPs PPARa+/+ and conversely upregulated in pretreated bone marrow-derived cells with MPs PPARa2/2 . These results emphasize the key role of PPARa in modulating angiogenic potential of bone marrow-derived cells.
Supporting the results obtained with bone marrow-derived cells, MPs PPARa+/+ also significantly enhanced the expression of the proangiogenic factor VEGF and EC marker ICAM-1, whereas eNOS, VEGF and FLT-1 protein levels were significantly decreased in bone marrow-derived cells treated with MPs PPARa2/2 . These data are in accordance with those recently reported by Biscetti et al [18] showing that PPARa activation evokes angiogenesis through VEGF up-regulation. Collectively, these results indicate that MPs PPARa+/+ are able to promote the angiogenic program of both EPCs and ECs so as to enhance their proangiogenic properties. However, it remains to be determined whether MPs PPARa+/+ act directly on bone marrow-derived cells to induce EPC differentiation and pro-angiogenic factor release (i.e. VEGF) or whether bone marrow-derived cells play a paracrine angiogenic role. In this respect, it has been shown the essential paracrine role of recruited bone marrow-derived circulating cells occurs by secreting pro-angiogenic factors rather than serving as endothelial progenitors in an adult model of neovascularization [29]. These authors have shown that SDF-1 secreted by perivascular cells in response to VEGF is responsible for positioning of bone marrow-derived circulating cells near to blood vessels. In this study, we show that MPs PPARa+/+ enhances both VEGF and SDF-1 mRNA, suggesting that similar mechanisms may occur.
The anti-inflammatory actions of PPARs have been largely described to inhibition of the NF-kB pathway [40,41]. Surprisingly, in our study the effects of MPs PPARa+/+ on EPC differentiation and EC angiogenesis are sensitive to NF-kB inhibition. Furthermore, MPs PPARa+/+ effects were associated with increased phosphorylation of I-kBa (Ser32/36) in bone marrow-derived cells. These findings suggest a positive regulation of NF-kB pathway by PPARa, as described in other studies [42]. In this context, PPARs have been recently reported to be involved in neural stem cells (NSC) acquisition of a specific fate, and the control of NSC proliferation, migration and differentiation through NF-kB pathway [43]. A similar signaling through PPARa significantly may contribute to EPC differentiation via stimulation of NF-kB pathway.
The Akt pathway plays a central role in the functioning of mature ECs. Thus, activation of Akt promotes endothelial cell survival by inhibiting apoptosis and is involved in EPC differentiation [44]. Our study not only extends these findings by demonstrating that Akt activation is associated with EPC differentiation, but also suggests Akt as a target for MPs PPARa+/+ in modulating EPC differentiation and functions. On the basis of these findings, it is possible that one of the mechanisms that mediates EPC differentiation induced by MPs PPARa+/+ is the crosstalk between PPARa-NF-kB and Akt pathways.
Finally, to further validate our obtained data concerning the effects of MPs PPARa+/+ on in vitro EPC differentiation, we investigated the in vivo effects of untreated or MP-treated bone marrow-derived cells on neovascularization. MPs PPARa+/+ -treated bone marrow-derived cells enhanced the formation of new vessels as assessed by increased hemoglobin content in MatrigelH plugs; these effects were attenuated by inhibition of the NF-kB pathway. These findings demonstrate that isolated bone marrow-derived cells expanded by in vitro MP treatment are able to generate in vivo functional vessels and confirm our results obtained in vitro.
In conclusion, the results of this study confirm the proangiogenic ability of circulating MPs harboring PPARa, which are able to induce in vitro EPC differentiation and angiogenesis through the NF-kB pathway. Therefore, MP treatment may significantly contribute to EPC differentiation and stimulation of neovascularization following ischemic injury. The PPARa-NF-kB-Akt pathways may play a pivotal stimulatory role in neovascularization, which is, at least in part, mediated through bone marrowderived EPCs. Improvement of EPC differentiation may greatly aid reparative neovascularization.

Circulating MP isolation
All animal studies were carried out using approved institutional protocols and were conformed the Guide for the Care and Use of Laboratory Animals published by US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Mice (6 to 8 week old) were anesthetized by isofluorane for all procedures. Circulating MPs were isolated from venous citrated blood collected from PPARa null (KO) and corresponding wild-type (WT). PPARa KO mice were generated by the group of Gonzalez, as previously described [45]. Briefly, blood was centrifuged at 1,900 g for 3 minutes and platelet-rich plasma was separated from whole blood. Then, platelet-rich plasma was centrifuged at 5,000 g for 4 minutes to obtain platelet-free plasma (PFP). Sixty mL of PFP were frozen and stored at 280uC until subsequent use. In order to pellet MPs for in vitro and in vivo studies, circulating MPs were concentrated from PFP by three series of centrifugations at 21,000 g for 45 minutes and resuspended in 0.9% saline salt solution and stored at 4uC until subsequent use. The physiological circulating concentration of mice MPs was used in all experiments. Washing medium for the last supernatant was used as control.

Characterization of MP phenotype
Membrane MP subpopulations were discriminated in PFP according the expression of membrane-specific antigens. Phenotype of endothelial MPs was performed using anti-CD54 labeling, characterization of platelet, leukocyte and erythrocyte MPs was performed using anti-CD61, anti-CD45, anti-Ter-119/erythroid cell and anti-CD133 labeling, respectively. Irrelevant mouse IgG was used as an isotype-matched negative control for each sample.
For numeration studies, 8 mL of PFP were incubated with either 1 mL of specific antibody (BioLegend, San Diego, CA). After 45 minutes of incubation at room temperature, samples were diluted in 300 mL of 0.9% saline salt solution. Annexin V (BioVision Inc., Mountain View, CA) binding was used to numerate phosphatidylserine-expressing circulating MPs (2 mL of annexin V/5 mL PFP). Then, in order to enumerate MPs, an equal volume of sample and Flowcount beads were added and samples were analyzed in a flow cytometer 500 MPL system (Beckman Coulter, Roissy, France) as previously described [46,47].
In parallel, after 7 days of treatment, adherent cells were detached with scratching with accutase (Sigma-Aldrich), and then incubated for 45 minutes at room temperature simultaneously with anti-Ly-6A/E (Sca-1)-PE/Cy7 and anti-Flk-1-PE Abs or anti-F4/80-PE/Cy7 and anti-CD31 followed incubation with goat anti-rat IgG-FITC. Isotype-matched antibodies served as controls. After washing, cells were analyzed by flow cytometer 500 MPL system (Beckman Coulter). Ten thousand events were acquired by sample.

Aortic EC isolation and characterization
Aortic ECs were isolated from either PPARa+/+ or PPARa2/2 mice (6 to 8 week old) as previously described [48]. The endothelial phenotype of primary aortic ECs was verified by flow cytometry (data not shown).

In vitro capillary network formation on MatrigelH
ECs isolated from either PPARa+/+ or PPARa2/2 aorta were seeded (10 5 cells/cm 2 ) on culture dishes coated with extracellular matrix gel from Engelbreth-Holm-Swarm murine sarcoma (MatrigelH, Sigma-Aldrich). Briefly, 200 mL of MatrigelH substrate diluted with serum-free medium (1:1 dilution) was added into a four-well plate and allowed to solidify for 1 hour at 37uC. Then, cells were incubated with medium containing 5% of FBS and allowed to adhere for 1 hour. ECs were incubated for 24 hour in the absence or in the presence of MPs from either PPARa+/+ or PPARa2/2 mice. Capillary-like structures formation was examined using an inverted phase contrast microscope (MOTIC AE21) and quantified using ImageJ software.

In vivo MatrigelH plug assay
For in vivo studies of bone marrow-derived cell-induced angiogenesis, isolated bone marrow-derived cells from WT mice were incubated in the absence or in the presence of MPs isolated from PPARa+/+ or PPARa2/2 mice for 7 or 15 days. MatrigelH plugs were prepared on ice by mixing 500 mL of MatrigelH (Sigma-Aldrich) with recombinant bFGF (300 ng/mL, Peprotech, Rocky Hill, NJ) and bone marrow-derived cells treated or untreated with MPs. This mixture was injected subcutaneously on the back of WT mice. At day 7, MatrigelH plugs were removed and homogenized in lysis buffer and incubated for 24 hour at 4uC and then, disrupted with a Polytron (PRO250, Monroe, CT). Hemoglobin concentration was measured in the supernatants with Drabkin's reagent (Sigma-Aldrich) according to the manufacturer's instructions.

Quantitative-PCR (q-PCR) analysis
Bone marrow-derived cells isolated from PPARa+/+ mice were grown for 7 days in the absence or presence of MPs isolated from PPARa+/+ or PPARa2/2 mice. Cells were washed with PBS and lysed with Tri-reagent lysis buffer. Cell lysates were frozen in liquid N 2 and used to investigate the expression of mRNA for 52 transcripts related to differentiation and angiogenesis by quantitative real-time reverse transcription-polymerase chain reaction (q-PCR) ( Table S1). q-PCR analyses were carried out by Service Commun de Cytométrie et d'Analyses Nucléotidiques from Angers University, using a Chromo 4 TM (Bio-Rad, Hercules, CA) and SYBR Green detection. Primers were designed using Primer3 software (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www. cgi). Quantifications were realized according to the DCt method and the relative gene expression levels were normalized using the geometric mean of three housekeeping genes as previously described [49].

Data analysis
Data are represented as mean 6 SEM; n represents the number of mice. Statistical analyses were performed by Mann-Whitney Utests (non-parametric) using Prism software package 4.00 (Graph-PAD Software, San Diego, CA). P,0.05 was considered to be statistically significant. Figure S1 Flow cytometry characterization of isolated cells derived from bone marrow from PPARa+/+ mice at day 0 and after 7 days of culture. At day 7, cells displaying a double labeling for Sca-1/CD31, Sca-1/CD45, Sca-1/CD14 and F4/80/CD31 were significantly increased when compared to cells at day 0. Also, Sca-1/Flk-1 and CD133/Flk-1 cells were slightly increased when compared to cells at day 0. *P,0.05, **P,0.01, ***P,0.001 vs isolated cells at day 0. EPCs had no effect on PPARa expression compared with control indicating that of PPARa transfer did not take place. Protein expression was normalized by b-actin. (B) Flow cytometric analysis of EPC differentiation showed that PPARa activation (treatment of EPCs isolated from WT mice with WY-14643 (5 mM), a specific PPARa agonist) increased significantly Sca-1+/Flk-1+ cells indicating an increased differentiation. (C) Treatment of EPCs isolated from PPARa2/2 with PPARa agonist had no effect on EPC differentiation in vitro.