Endothelial Microparticle-Derived Reactive Oxygen Species: Role in Endothelial Signaling and Vascular Function

Endothelial microparticles are effectors of endothelial damage; however mechanisms involved are unclear. We examined the effects of eMPs on cultured endothelial cells (ECs) and isolated vessels and investigated the role of eMP-derived reactive oxygen species (ROS) and redox signaling in these processes. eMPs were isolated from EC media and their ability to directly produce ROS was assessed by lucigenin and liquid chromatography. Nicotinamide adenine dinucleotide phosphate oxidase (Nox) subunits were probed by Western blot. ECs were treated with eMPs and effects on kinase signaling, superoxide anion (O2 ∙−) generation, and nitric oxide (NO) production were examined. Acetylcholine-mediated vasorelaxation was assessed by myography in eMP-treated mesenteric arteries. eMPs contained Nox1, Nox2, Nox4, p47phox, p67phox, and p22phox and they produced ROS which was inhibited by the Nox inhibitor, apocynin. eMPs increased phosphorylation of ERK1/2 and Src, increased O2 ∙− production, and decreased A23187-induced NO production in ECs. Pretreatment of eMPs with apocynin diminished eMP-mediated effects on ROS and NO production but had no effect on eMP-mediated kinase activation or impairment in vasorelaxation. Our findings identify a novel mechanism whereby eMP-derived ROS contributes to MP bioactivity. These interactions may be important in conditions associated with vascular injury and increased eMP formation.


Introduction
The endothelium plays a critical role in the regulation of blood flow, cellular trafficking, coagulation, and inflammation [1,2]. Normal endothelial function requires a dynamic, controlled communication system between endothelial cells (ECs) and vascular smooth muscle cells, fibroblasts, and immune cells that employ electrical signals, cell-cell/cellmatrix contacts, cytokines/proteins, small molecules, and gases such as nitric oxide [2][3][4]. In addition, intercellular communication may involve the release of extracellular vesicles, which can function as intercellular carriers of ligands, enzymes, RNA, and miRNA [5][6][7]. Vascular cells have been shown to release vesicles of differing origins that contribute to many (patho)physiological processes including angiogenesis, coagulation, inflammation, and fibrosis [8].
Microparticles (MPs, sometimes referred to as microvesicles or ectosomes) are 100-1000 nm anuclear vesicles formed following cytoskeletal and membrane reorganization and release from cells into the extracellular milieu [7,8]. In biological samples, alterations in MP levels may be indicative of underlying pathology, correlate with measures of vascular dysfunction, and predict risk of adverse cardiovascular events [8][9][10]. MPs are also potent autocrine/paracrine signals for various cellular responses [11]. For example, endothelial MPs promote cell senescence, oxidative stress, coagulation, inflammation, and apoptosis [8,[12][13][14][15]. Amongst the many responses initiated by endothelial MPs in ECs is the induction of reactive oxygen species (ROS) production. Initial studies by Brodsky et al. showed that endothelial MPs induce O 2 2 Oxidative Medicine and Cellular Longevity the antiangiogenic effects of endothelial MPs [17]. Similarly ROS production has been shown in human umbilical vein ECs exposed to endothelial MPs [18][19][20]. We have observed that exogenous endothelial MPs induce ROS production in ECs through both mitochondrial and nicotinamide adenine dinucleotide phosphate (NADPH) oxidases [12,13]. More controversially, it has been suggested that endothelial MPs may directly produce ROS. Brodsky et al. examined de novo ROS production by endothelial MPs by DHE staining and observed p22 phox within endothelial MPs [16]. NADPH oxidase activity has also been reported in aortic endothelial MPs and may be influenced by glucose exposure [21]. However, to date, there has been no direct evidence that MPs possess the complete machinery necessary to produce ROS, nor has the role of de novo ROS production on MP-mediated EC responses been evaluated. The purpose of this study was to evaluate whether endothelial MPs produce ROS and to evaluate the potential regulatory role of MP-derived ROS on EC signaling and vascular function. 2.3. Cell Culture. ECs were isolated from aortas of C57BL6/J mice as we previously described [13,22]. Cells were cultured on attachment factor-coated polystyrene dishes in DMEM containing 10% FCS, 50 mg/L of EC growth supplement, 10 U/mL heparin, 100 U/mL penicillin/streptomycin, and 1x minimal essential amino acids and placed in a humidified incubator at 37 ∘ C and 5% CO 2 .

Endothelial Microparticle Isolation and Quantification.
Endothelial MPs were isolated from media collected from EC cultures using a modification of our previously described methods [13,23]. Samples were centrifuged at 2500 g for 10 minutes at 4 ∘ C to obtain cell-free media. MPs were then pelleted from cell-free media by centrifugation at 20000 ×g for 20 minutes at 4 ∘ C. Endothelial MPs were confirmed as such and quantified by flow cytometry using an Alexa-647-labeled Annexin V (0.5 g/mL final concentration in Annexin V binding buffer) to identify events as MPs. As a negative control, a subpopulation of MPs were resuspended in Annexin V binding buffer lacking calcium, which is necessary for Annexin V binding to phosphatidylserine. MPs were centrifuged and resuspended in assay-specific buffers (indicated below) or 1x phosphate buffered saline for cell treatment studies. A final concentration of 10 5 MPs/mL, previously identified to exert effects on cultured ECs, was used for treatments [12,13].
2.5. Measurement of ROS Production. ROS production was measured by lucigenin chemiluminescence and by DHE/HPLC in MPs and ECs. For chemiluminescence, MPs were isolated from 100 mL of media and lysed in a buffer containing (20 mM of KH 2 PO 4 , 1 mM of EGTA, 1 g/mL of aprotinin, 1 g/mL of leupeptin, 1 g/mL of pepstatin, and 1 mM of PMSF). Following lysis, 50 L of sample was added to 175 L of assay buffer (50 mM of KH 2 PO 4 , 1 mM of EGTA, and 150 mM of sucrose) and 5 M lucigenin. NADPH (1 mM) was added to the suspension (300 L) containing lucigenin. Luminescence was measured on a Berthold Orion II luminometer (Berthold Detection Systems, Pforzheim, Germany) in 1 s readings over a total of 30 cycles before and after stimulation with NADPH. A buffer blank was subtracted from each reading. As a negative control, the high-speed supernatant obtained during MP isolation was also assessed. The results are expressed as counts per mg of protein.
For HPLC studies, MPs were isolated from 100 mL of media and incubated in DHE (50 M final concentration in a buffer containing 1.3 mM CaCl 2 , 0.8 mM MgSO 4 , 5.4 mM KCl, 0.4 mM KH 2 PO 4 , 4.3 mM NaHCO 3 , 137 mM NaCl, 0.3 mM Na 2 HPO 4 , 5.6 mM glucose, and 100 M DTPA, pH 7.4) for 30 minutes in the presence and absence of the Nox inhibitor, apocynin (10 M). DHE oxidation was then assessed by HPLC as described previously [13]. As a negative control, the high-speed supernatant obtained during MP isolation was also assessed.

Nitric Oxide Production.
Measurement of EC nitric oxide production was performed using a fluorescence-based assay with live cell imaging [24][25][26]. Cells were loaded with the fluorescent NO-sensitive dye DAF (0.5 M final concentration) and incubated at 37 ∘ C for 30 minutes. Coverslips containing DAF-loaded cells were placed in a temperature regulated (37 ∘ C) chamber mounted on the stage of an inverted microscope in Krebs-Ringer phosphate (KRP) buffer (120 mM NaCl, 4.8 mM KCl, 0.54 mM CaCl 2 , 1.2 mM MgSO 4 , 11 mM glucose, and 15.9 mM sodium phosphate, pH 7.2). Changes in cellular DAF fluorescence intensities were simultaneously acquired for multiple cells (defined regions of interest) using a Stallion Digital Hi-Speed Multi-Channel Imaging System (Zeiss, Germany) with 495 nm excitation and 515 nm emission. Changes in NO induced by exposure to the calcium ionophore A23187 (1 M final concentration) were measured in relation to baseline fluorescence as described elsewhere [24][25][26]. Unstimulated and A23187 stimulated NO production was observed as the difference between the NO content measured in the presence (total) and absence (basal) of treatment and expressed as percentage of the basal considered as 100%.

Myography.
Second-order branches of mouse mesenteric artery were isolated and cleaned of adhering fat and connective tissue. Vessels segments (1.5-2 mm) were mounted on a wire myograph and equilibrated in a physiological saline solution (in mM; 130 NaCl, 14.9 NaHCO 3 , 4.7 KCl, 1.18 KH 2 PO 4 , 1.17 MgSO 4 , 1.6 CaCl 2 , 0.026 EDTA, and 5.5 glucose, pH 7.4, 37 ∘ C) which was bubbled continuously with 95% O 2 and 5% CO 2 . Endothelium-dependent relaxation was assessed in the presence and absence of endothelial MPs (10 5 MPs/mL) as previously described [27]. Vessels were preincubated with MPs for 30 minutes. Endotheliumdependent relaxation was then examined by measuring dilatory responses to acetylcholine (1 nM-10 M) in vessels precontracted with phenylephrine at a concentration which was found to cause approximately 80% maximal contraction.

Statistical Analysis.
Results are expressed as mean ± SEM and were analyzed using a one-way ANOVA with Bonferroni's posttest using Prism 5.0 (GraphPad Software, Inc., La Jolla, CA, USA). < 0.05 was considered statistically significant.

MPs Possess NADPH Oxidases and Generate ROS.
To determine whether MPs contain ROS-generating enzymes, we examined the presence of NADPH oxidase subunits in endothelial MPs. Both cultured ECs and endothelial MPs were found to contain Nox1, Nox2, Nox4, p22 phox , p47 phox , and p67 phox (Figure 1(a)). Compared with ECs, endothelial MPs were enriched in Nox4 and p22 phox , while Nox1 and p67 phox were found at lower levels. No NADPH oxidase subunits were detectable in the high-speed supernatant obtained during MP isolation.
As shown in Figure 1(b) lucigenin chemiluminescence revealed that endothelial MPs produced O 2 •− , a process that was attenuated by treatment with the NADPH oxidase inhibitor apocynin (5 M). Similarly, O 2 •− production was detectable by DHE/HPLC and was inhibited by apocynin (Figure 1(c)).

Effects of Endothelial MPs on Intracellular Signaling in Cultured ECs.
In order to assess EC signaling following endothelial MP exposure, ECs were treated with endothelial MPs for 0-24 hours and phosphorylation of ERK1/2, Src, p38, and Akt kinases was measured by Western blot analysis. As shown in Figure 2. MPs increased ERK1/2 phosphorylation at 30 minutes and 1 hour while Src was increased at 4 hours. Conversely, Akt and p38 phosphorylation were unchanged over 24 hours (not shown).

Effects of Endothelial MPs on eNOS Activation and NO Production in Cultured ECs.
Stimulation of ECs with endothelial MPs was associated with an impairment in eNOS activation and NO production. Following exposure to endothelial MPs, phosphorylation of eNOS (serine 1177, associated with enzyme activation) was decreased suggesting impaired NO production (Figure 3(a)), while phosphorylation at threonine 495 (associated with enzyme inhibition) was unchanged (Figure 3(b)). Consistent with a decrease in eNOS activity, we observed a decrease in calcium-dependent NO production assessed by DAF-FM. Exposure to A23187 (10 M) was associated with a 5-6-fold increase in fluorescence, indicative of NO production, and consistent with previous reports [28] (Figure 4). Preincubation of ECs with endothelial MPs significantly impaired A23187-induced NO production in ECs.
Oxidative Medicine and Cellular Longevity   Figure 4: Effects of endothelial MPs on NO production (a, b) and ROS production (c) in cultured mouse aortic ECs. ECs were exposed to endothelial MPs (10 5 /mL, 4 hours) or endothelial MPs pretreated with the NADPH oxidase inhibitor apocynin (Apo MPs, 10 5 /mL, 4 hours).

Effects of MP-Derived ROS on Endothelial Cell Signaling.
To determine if de novo production of ROS plays a role in the biological actions of endothelial MPs endothelial MPs were pretreated with apocynin and administered to cultured ECs. Apocynin pretreatment partially attenuated endothelial MP-mediated inhibition of NO production (Figures 4(a) and 4(b)) and MP-mediated induction of ROS production (Figure 4(c)) but had no effect on endothelial MP-mediated ERK1/2 or Src phosphorylation ( Figure 5).

Effects of Endothelial MPs on Endothelium-Dependent Vasorelaxation in Isolated Mesenteric
Arteries. Finally, to determine whether endothelial MPs influence endotheliumdependent vasorelaxation we assessed endotheliumdependent relaxation induced by acetylcholine (ACh) in phenylephrine precontracted arteries ( Figure 6). As shown in Figure 6, endothelium-dependent relaxation induced by ACh was impaired in mesenteric arteries pretreated with endothelial MPs. Preincubation of endothelial MPs with  Oxidative Medicine and Cellular Longevity apocynin did not alter MP-mediated effects on mesenteric artery relaxation.

Discussion
Endothelial MPs influence EC function through redoxsensitive processes; however the molecular mechanisms underlying this are unclear. In the present study we show that endothelial MPs contain NADPH oxidase subunits, produce ROS in an NADPH oxidase-dependent fashion, and induce downstream effects on EC signaling. At the functional level, endothelial MPs influence vascular tone by inhibiting endothelium-dependent vasorelaxation. EC regulation by MPs occurs in a paracrine manner, where MP-derived ROS downregulates eNOS activity and stimulates ROS production, without influencing MAP kinases or Src. Our findings highlight novel signaling mechanisms through de novo ROS production by MPs and indicate that MP-derived ROS differentially regulate redox and kinase signaling in ECs. Such processes may influence endothelial function in conditions associated with increased endothelial MP production, such as in atherosclerosis, hypertension, or diabetes. Microparticles typically possess membrane and cytoplasmic proteins/content of the parent cell from which they are derived. Here we show, similar to what has been demonstrated in ECs, that MPs express Nox isoforms and NADPH oxidase subunits, which are functionally active since ROS is produced in an apocynin-inhibitable manner. H 2 O 2 , a direct or indirect product of NADPH oxidase activation, is cell membrane-permeable and, as such, MP-derived ROS may influence ECs without direct physical contact. This observation may be of particular relevance in the in vivo setting where physical interaction between circulating MPs and the endothelium may be limited due to the presence of a glycocalyx, flow, or possibly atherosclerotic lesions. MPs also influence ECs through direct communication. We and others have reported that surface interactions involving ligandmediated receptor activation contribute to endothelial MP bioactivity [13,15,29] while uptake of endothelial MPs has been implicated in processes associated with coagulation and inflammation [18,30].
Although there is increasing interest in the role of endothelial MPs as redox biovectors of endothelial dysfunction and vascular inflammation, there is a paucity of information as to whether MPs themselves can generate ROS. Brodsky et al. reported that ROS are produced in isolated MPs and identified the presence of p22 phox [16]. However that study failed to identify the Nox machinery responsible for ROS production and in fact did not demonstrate Nox-dependence of the observed ROS production. Similarly, Jansen and colleagues reported ROS production by endothelial MPs and suggested that this may be dependent upon NADPH oxidase but the enzymatic machinery required for ROS production was again unclear [21]. Others have reported the presence of certain Nox machinery (p22 phox , Nox2) in platelet-derived exosomes [31]. Our data show that endothelial MPs possess Nox1, Nox2, Nox4, p22 phox , p47 phox , and p67 phox . Accordingly, ROS production in MPs may be mediated through multiple Nox isoforms. Of note is the observation that Nox1 and p67 phox were present at lower levels than seen in ECs while Nox4 and p22 phox were enriched in MPs.
We measured ROS production by two methods, lucigenin chemiluminescence and DHE/HPLC. Lucigenin assesses ROS production in lysed samples supplemented with NADPH such that endogenous sources of electrons are not required. On the other hand, DHE/HPLC methodology requires an intact system with an endogenous source of electrons for enzymatic activity. Thus it appears that endothelial MPs contain at least some source of NADPH that fuels ROS production. This observation is consistent with that reported by others where ROS was produced in MPs in the absence of NADPH supplementation [16,21]. The source of electrons in endothelial MPs is presently unclear since MPs have not been shown to contain NADPH-generating enzymes. One possibility may be that NADPH in MPs is a residual component that derives from the cytosolic material of parent ECs released during MP formation.
Using apocynin, a pharmacological NADPH oxidase inhibitor, we showed that endothelial MPs require de novo ROS production for inhibition of eNOS-derived NO production and for endothelial MP-mediated induction of ROS production in ECs. While endothelial MPs reduced eNOS phosphorylation and NO production, we did not see any alterations in Akt signaling and mechanisms responsible for impaired NO production are unclear at this time. It is possible that endothelial MPs interfere with intracellular calcium handling or reduce eNOS cofactors. Alternatively, endothelial MPs may influence levels of protein phosphatase 2a (PP2A) which dephosphorylates eNOS at serine 1177 [32]. A recent proteomic analysis found PP2A present in human endothelial cell-derived MPs and it is conceivable that eMPs may transfer bioactive PP2A to target cells but this has not been shown experimentally.
Endothelial MP-mediated effects on intracellular ERK1/2 and Src as well as effects on vasorelaxation were not altered in apocynin-treated MPs suggesting that these processes are not influenced by MP-derived ROS. Such findings highlight the pleiotropic nature of MPs with both ROS-dependent and ROS-independent pathways contributing to endothelial MP-mediated signaling. Accordingly it is our belief that the observed ROS effects are not generalized phenomena but highly regulated processes. Exact mechanisms whereby MPs mediate their ROS-independent effects are unclear. One possibility is the activation of cell surface receptors by endothelial MPs. For example, we have previously reported that inhibition of epidermal growth factor receptor (an upstream activator of ERK1/2 and Src) blocks endothelial MP-mediated responses in cultured mouse aortic endothelial cells [13]. Endothelial MPs may conceivably activate other cell surface receptors; however this remains an area in need of investigation.
In summary, we demonstrate that endothelial MPs produce ROS through Nox-dependent processes, which contribute to endothelial O 2 − generation and downregulation of NO generation without effect on ERK1/2 and Src activation or on MP-mediated impairment of vasorelaxation. Our findings indicate that MPs differentially influence EC function through redox-sensitive and insensitive processes. Modulation of vascular function by MPs appears to be NADPH oxidase/ROS-independent and may involve ERK1/2 and Src or other mechanisms. Cross talk between MPs and ECs may be important in conditions associated with vascular injury and increased endothelial MP formation.