Extracellular vesicles activate a CD36 dependent signaling pathway to inhibit microvascular endothelial cell migration and tube formation

Abstract

Objective

Literature on the effect of cell-derived extracellular vesicles (EV), ≤1μm vesicles shed from various cell types during activation or apoptosis, on microvascular endothelial cell (MVEC) signaling is conflicting. Thrombospondin-1 and related proteins induce anti-angiogenic signals in MVEC via CD36. CD36 binds EV via phosphatidylserine (PS) exposed on their surface but the effects of this interaction on MVEC functions are not known. We hypothesized that EV would inhibit angiogenic MVEC functions via CD36.

Approach and Results

EV generated in vitro from various cell types or isolated from plasma inhibited MVEC tube formation in in vitro matrigel assays and EC migration in Boyden chamber assays. Exosomes derived from the same cells did not have inhibitory activity. Inhibition of migration required EC expression of CD36. In mouse in vivo matrigel plug assays EV inhibited cell migration into matrigel plugs in wild type but not in cd36 null animals. Annexin V, an anionic phospholipid binding protein, when incubated with EV partially reversed inhibition of migration, suggesting a PS-dependent effect. EV exposure induced reactive oxygen species (ROS) generation in MVEC in a NADPH oxidase and Src family kinase dependent manner and their inhibition by apocynin and PP2, respectively, partially reversed the EV-mediated inhibition of migration. Annexin V partially reversed EV-induced ROS generation in murine CD36 cDNA transfected HVUEC but not in CD36 negative HUVEC.

Conclusions

These studies establish a general inhibitory effect of EV on EC pro-angiogenic responses and identify a CD36-mediated mechanistic pathway through which EV inhibit MVEC migration and tube formation.

Introduction

CD36 is a 471 amino acid, 88 kDa, heavily glycosylated, multifunctional membrane protein present on many cell types including microvascular endothelial cells (MVEC)^1–5. As a class B scavenger receptor on macrophages, it recognizes patterns presented on pathogens or pathogen infected cells to facilitate clearance by the phagocytic cells^4,6,7. CD36 also binds endogenous danger signal ligands including modified lipids expressed on lipoproteins⁸ and apoptotic cells^2,9, advanced glycated proteins¹⁰, fibrillar β-amyloid protein¹¹ and S100A family pro-inflammatory peptides¹². Recognition of a specific protein domain known as thrombospondin type I repeats (TSR) present in anti-angiogenic proteins like TSP-1, TSP-2 and brain angiogenesis inhibitor 1 (BAI-1) by CD36 on MVEC mediates anti-angiogenic signaling. Interaction of TSR with CD36 induces pro-apoptotic signaling^13,14 and also down regulates vascular endothelial growth factor receptor-2 (VEGFR2)-mediated pro-angiogenic signaling to inhibit angiogenesis via both direct and indirect MVEC pathways¹⁵. Many in vivo and human studies have shown the importance of the anti-angiogenic role of CD36 in pathological processes like tumor growth and wound healing^13,16.

CD36 participates in clearance of apoptotic cells via oxidized anionic phospholipids phosphatidylserine (PS)^17,18 expressed on the cell surface⁹. Cell-derived extracellular vesicles (EV), ≤1μm vesicles shed from vascular and blood cells during activation or apoptosis, circulate in plasma and are biomarkers of cardiovascular disease. EV have implications in various patho-physiological conditions including thrombosis, arthritis and cancer. For example, we showed that recognition of EV by CD36 on platelets promotes platelet activation and thrombosis in mouse models¹⁹. The available literature on the role of EV on angiogenesis is controversial²⁰. One study of EC-derived EV showed that EV promote angiogenesis ²¹ whereas other groups have shown that they inhibit angiogenesis²². Similar contradictory results have been reported on lymphocyte-derived EV^23,24. These differences may to be due to different cargo carried by EV derived from different cell types or generated in vitro by different mechanisms. A common feature during EV formation regardless of cellular source and cargo is disruption of membrane asymmetry resulting in the exposure of anionic phospholipid PS on the outer leaflet of their membrane^25,26. Therefore, we hypothesized that EV interaction with CD36 via PS inhibits angiogenic activities.

In this manuscript we report that EV derived from multiple cellular sources or isolated from human subjects inhibit MVEC migration and tube formation in vitro and in vivo. Mechanistically, this was shown to be due to triggering of CD36-mediated antiangiogenic signals in MVEC involving Src-family kinase and NADPH-mediated ROS generation.

Materials & Methods

Materials and methods are available in the online-only data supplement

Results

EV from multiple cellular sources inhibit MVEC tube-like structure formation

Tube-like structure formation by MVEC cultured in matrigel was inhibited in a dose-dependent manner by EV generated in vitro from THP-1 cells (Figure 1A, top panel). At a ratio of one cell to 10EV, tube number was inhibited by 65% (p<0.05) and at a 1:20 ratio, inhibition was >85% (p<0.01). At both ratios, fewer than 50% of the number of segments and branch points were formed (p<0.05) (Figure 1A, lower panel). Since EV can originate from many cell types we tested if their cell of origin has a differential influence on the inhibition of MVEC tube-formation. EV prepared from different vascular cells (HUVEC, MVEC or RBC) in vitro were all found to significantly inhibit MVEC tube formation in matrigel (>70%; p<0.01) (Figure 1B) regardless of cell of origin.

Figure 1. Extracellular vesicles inhibit MVEC tube-like structure formation and migration.

(A.) MVECs, in EC media with growth factors, were added to matrigel coated tissue culture plates and incubated for 16 hours at 37°C in the absence (left panel) or presence of THP1-derived EV at ratios of 1 cell to 10 EV (middle panel) or 1:20 (right panel). Cells were washed, fixed with 4% paraformaldehyde, and stained with alexa-fluor 568 labeled phalloidin with 0.3% triton X-100 for 30 min. Images of the entire field were captured with fluorescence microscope. Total number of tubes formed was counted manually (lower left panel), and the number of segments and branch points formed were calculated using MetaMorph angiogenesis software (lower middle and right panels). Images are representative of n=3. (B.) MVEC were incubated with EV derived from HUVEC (top middle panel), MVEC (top right panel) or RBCs (lower left panel) at a 1:10 ratio and the number of tubes formed was quantified (lower right panel). Images are representative of n=3. (C.) MVEC treated with or without THP1-derived EV at a 1:10 ratio were allowed to migrate towards EC growth factors in a transwell insert having a membrane with 8.0μm pore for 15 hours (n=3). Results expressed as number of migrated cells.

EV inhibit MVEC migration via CD36

A critical step in angiogenesis and EC tube formation is cell migration. Using an in vitro transwell migration assay we found that in the presence of THP1-derived EV (1:10 ratio) more than 85% of MVEC migration was inhibited (p<0.001; Figure 1C) consistent with our hypothesis that EV inhibit tube formation by inhibiting EC migration. Since EV from all cell types tested inhibited migration of MVEC and a common feature of EV formation is disruption of membrane asymmetry ^25,27 with exposure of phosphatidylserine (PS) on the outer membrane leaflet, and since PS is a well-documented ligand for the anti-angiogenic receptor CD36, we hypothesized that EV-mediated inhibition of MVEC migration was CD36 and PS dependent. We thus compared THP1-derived EV effects on early passage MVEC that have high CD36 surface expression as shown by flow cytometry (Figure 2A, left) to HUVEC that do not express CD36 (Figure 2B, left). THP1-derived EV inhibited MVEC migration by more than 80%, whereas inhibition was significantly lower in HUVEC (<45%; p<0.001) suggesting a CD36-dependent inhibitory effect. To confirm CD36 dependence we generated CD36 expressing HUVEC using lentiviral mediated heterologous transfection of murine CD36 cDNA (Figure 2B middle panel). Surface expression of murine CD36 on HUVEC (mCD36 HUVEC) brought the inhibition of migration by THP1-derived EV to a level comparable to that seen in MVEC (Figure 2B, right). Levels of CD36 expression vary considerably among different batches of cultured MVEC (Figure 2A, left) and there was progressive loss of CD36 expression with serial passage of MVECs with no expression seen after passage 7 (Figure 2A middle panel). This allowed us to test the effect of EV on MVEC migration in cultures with varying expression levels of CD36. The bar graph in Figure 2A shows a significant difference in the level of THP1-derived EV mediated inhibition of migration (90% vs 73%) between 2 different batches of early passage MVEC with differing CD36 expression levels (p<0.001). Strikingly, passage 8 MVEC that lost CD36 expression showed a level of inhibition (33%) similar to that seen in CD36 negative HUVEC.

Figure 2. Extracellular vesicles inhibit EC migration via CD36 and PS interaction.

(A.) Human MVEC from different batches (left panel) and passages P5 to P7 (middle panel) were incubated with anti-CD36 or isotype control IgG followed by Alexa-fluor 488 conjugated secondary Ab and analyzed by flow cytometry for CD36 surface expression. Data were plotted with FlowJo software. (B.) HUVEC (left panel) or HUVEC transfected with murine CD36 (mCD36) (middle panel) were incubated with species specific anti-CD36 Ab or controls and analyzed as in (A.). EC were treated with THP1-derived EV (1:10 ratio) and allowed to migrate towards EC growth factors in a transwell insert having a membrane with 8.0μm pore for 15 hrs. Cells not treated with THP1 EV served as control (n=5). The number of cells migrated were counted and results expressed as a percent of cells inhibited from migrating by the EV compared to control (A and B, right panels). (C.) MVEC were incubated with EV isolated from plasma from healthy human subjects or patients with systemic vasculitis at a 1:10 ratio and then allowed to migrate in transwell assay as in (A.). MVEC incubated with EV-poor plasma (EVPP) served as control. The results are expressed as percent of cells inhibited from migrating by EV compared to EVPP (n=4). (D.) THP1-derived EV incubated with or without PE-conjugated Annexin V (75ng/ml) were analyzed by flow cytometry for phosphatidylserine exposure and data plotted with FlowJo software (left panel) (n=4). MVEC were subjected to migration assay, as in Panel A, in the presence of THP1-derived EV pretreated with Annexin V (5ng/ml) or vehicle control (right panel; (n=6).

To test if EV isolated from human plasma have a similar anti-angiogenic effect as that of EV generated in vitro we used differential centrifugation to isolate EV from plasma of healthy human donors (hEV) and from patients with systemic (vasculitis hEV) and tested their effect on migration of MVEC in a transwell assay. As shown in Figure 2C, when used at the same concentration, EV from both populations inhibited migration by more than 60%.

As noted above, a common feature of EV regardless of their cell of origin is surface expression of PS. By flow cytometry, we characterized THP1-derived EV by the anionic phospholipid binding protein, Annexin V (Figure 2D, left panel). EV pretreated with Annexin V had a significantly lower inhibitory effect on migration of MVEC (<31%) compared to control EV (74%; p<0.001) Figure 2D, right), similar to the level seen with CD36 negative late passage MVEC. Annexin V itself did not alter migration of MVEC. In sum, these data show that EV inhibit migration via a CD36 – PS interaction.

CD36-dependent inhibition of MVEC migration by EV is mediated by a signaling pathway involving NADPH oxidase, ROS, and src-family kinases

CD36 signaling in macrophages, smooth muscle cells^28,29 and platelets has been shown to involve generation of intracellular oxidant radicals. Since reactive oxygen species (ROS) can inhibit angiogenesis ^30,31 we determined the role of ROS in EV-induced anti-angiogenic MVEC signaling. Treatment of MVEC with THP1-derived EV induced oxidant radical formation in a time and concentration dependent manner (Figure 3A) as detected with the fluorescence-based indicator H₂DCFDA. Increase in DCF fluorescence over baseline was seen as early as 30 min (p<0.01) and was >2 fold at 4 hrs (p<0.001) with the higher ratio of EV:cells (20:1). Since NADPH is required for CD36 mediated ROS generation in macrophages ²⁸, we tested whether NADPH is involved in EV-mediated oxidant radical production. Pretreatment of MVEC with the NADPH oxidase inhibitor, apocynin at 0.01mmol/L, blocked ROS induced by THP1-derived EV at all timed points, with 94% inhibition seen at 4hrs (Figure 3B, upper left panel). Since superoxide is the predominant ROS species generated by NADPH, we pretreated MVEC with a pharmacologic superoxide dismutase (SOD) mimetic, MnTMPyP (Pentachloride (manganese (III) tetrakis (1-methyl-4-pyridyl)porphyrin.5Cl−) (0.01mmol/L), and found that it completely abolished THP1-derived EV-induced ROS to a level even lower than the basal ROS production (p<0.001) (Figure 3B, upper right panel). Finally, pretreating mCD36 HUVEC with dipheyleneiodinium (DPI) also abolished THP1-derived EV-induced ROS confirming the involvement of NADPH oxidase (Figure 3B, lower panel). We next tested if inhibition of NADPH-mediated ROS generation by apocynin influenced CD36-dependent inhibition of MVEC migration. Pretreating MVEC with apocynin partially reversed (23%) the inhibition of migration by THP1-derived EV (p<0.001). Apocynin itself did not affect MVEC migration (Figure 3C). To show CD36-dependence of ROS generation we pretreated EV with Annexin V prior to incubation with EC and found that ROS generation in mCD36-HUVEC was inhibited (40% at 4hrs; p<0.05) whereas ROS generation in HUVEC was not (Fig 3E), indicating that EV-induced ROS generation via a CD36-mediated pathway.

Figure 3. Extracellular vesicles inhibit MVEC migration by inducing reactive oxygen species generation involving CD36 and NADPH oxidase.

(A & B) MVEC or mCD36 HUVEC in tissue culture plates were loaded with 25μM ROS indicator dye CM-H₂DCFDA, treated with 10μM Apocynin or 10μM of MnTMPyP or 10uM of DPI or vehicle control, and then exposed to THP1-derived EV at a 1:10 or 1:20 ratio after basal fluorescence of the cells was recorded. Fluorescence was measured at 30min, 1, 2 and 4 hr timed intervals after EV exposure. Results are expressed as percent of basal fluorescence (n=6, #p<0.05, *p<0.01, **p<0.001). (C.) MVEC treated 10μM Apocynin or vehicle control were subjected to migration assay as in Figure 2 in the presence or absence of THP1-derived EV (n=6). (D.) Reactive oxygen species generation was measured as in (A) with HUVEC (lower panel) or HUVEC transfected with murine CD36 (mCD36) (top panel) in response to THP1-derived EV pretreated with or without Annexin V (5ng/ml) at a 1:10 ratio (n=a minimum of 9 from 3 different experiments), *p<0.05).

The Src-family tyrosine kinase Fyn is a critical downstream effector of thrombospondin-1/CD36 anti-angiogenic signaling in MVEC¹³. We found by immunoblot (Figure 4A) that THP1-derived EV induced phosphorylation of Fyn in MVEC in a time dependent manner, with peak phosphorylation at 15–30 min (p<0.05). A specific Src kinase inhibitor, PP2 (0.01mmol/L) blocked 85% of ROS generation by EV in MVEC at 4 hrs (Figure 4B) and partially rescued CD36-dependent inhibition of MVEC migration by THP1-derived EV (16%; p<0.001; Figure 4C). A control peptide PP3 had no effect on MVEC ROS generation or migration. These results suggest that CD36-dependent inhibition of MVEC migration by EV involves activation of src-family kinases and subsequent induction of ROS.

Figure 4. Extracellular vesicle dependent inhibition of migration involves activation of Fyn kinase to induce ROS.

(A.) Lysates from MVEC exposed to THP1-derived EV at 1:10 ratio for different time periods were analyzed by western blot with an antibody specific for Fyn phosphorylated at Tyr560 and an antibody to total Fyn (left panel). Fyn phosphorylation was expressed as a ratio of phosphorylated to total (right panel) (*p<0.05). (B.) ROS induced by THP1-derived EV (1:20 ratio) in MVEC pre-treated with 10μM PP2 was quantified as in Figure 3 using CM-H₂DCFDA (n=6, *p<0.01, **p< 0.001) (left panel). MVEC treated with 10μM PP2 were subjected to migration assays as in Figure 2 in the presence or absence of THP1-derived EV (n=6). PP3 served as control for PP2 (n=3) (right panel). (C.) Lysates of MVEC treated with THP1-derived EV (1:10 ratio) were probed for cleaved caspase 7 and 9 (n=2). Vehicle and staurosporin treated cells were used as controls (left panel). mCD36 HUVEC cells treated with vehicle, staurosporin, THP1-derived and platelet-derived EV were stained with Annexin V and number of apoptotic cells per field were counted (n=3 each with 3 fields; n.s = not significant). (D.) mCD36 HUVEC treated with THP1-derived EV (depleted of exosomes) or exosomes at a 1:10 ratio were allowed to migrate towards EC growth factors in a transwell insert having a membrane with 8.0μm pore for 15 hours (n=9). The number of cells migrated were counted and results expressed as a percent of cells inhibited from migrating compared to control.

We tested if EV induced apoptosis in MVEC and found that apoptosis markers cleaved caspase 7 and 9, were absent in western blots of THP1-derived EV-treated MVEC lysates (lanes 3 & 4) unlike staurosporin treated positive controls (lanes 5 & 6) (Figure 4C, left panel). In another set of experiments we confirmed that EV in our experiments did not induce apoptosis by using Annexin V staining in THP1 and platelet-derived EV treated mCD36 HUVEC. Staurosporin induced apoptosis in about 80 cells per field but THP1 and platelet-derived EV did not induce apoptosis similar to control (Figure 4C, right panel).

Since our protocol for isolating EV from cell culture supernatants also isolates exosomes that might be present we used differential centrifugation to separate exosomes from EV and found that while exosome depleted EV inhibited mCD36 HUVEC migration by 75%, the resuspended exosomes inhibited by only 12% (Figure 4D), indicating that the inhibitory effect observed in our studies was mainly due to EV.

Deletion of Cd36 impairs EV-induced inhibition of VEGF-mediated cell migration in mice

To understand the role of CD36 in the anti-MVEC migration activity of EV in vivo, we performed matrigel plug assays in cd36 null and wild type C57Bl6 (WT) mice in the presence and absence of THP1-derived EV. In both WT and cd36 null mice, VEGF present in the matrigel plug enhanced chemotaxis and migration (>350%; p<0.001) of cells into the matrigel plugs compared to basal migration in the absence of VEGF. THP1-derived EV when incorporated into the matrigel significantly inhibited VEGF-induced migration of cells into the plugs in WT mice (24%; p<0.001), but not in cd36 null animals (Figure 5). These results show a CD36 dependent anti-migratory activity of EV in vivo and corroborate our findings observed in vitro.

Figure 5. cd36 deletion partially rescues EV-induced inhibition of VEGF-mediated cell migration in vivo.

Matrigel alone or mixed with VEGF (50ng/ml) or VEGF plus THP1-derived EV (2500 EV/μl) was injected subcutaneously into the flanks of wild type or cd36 null mice. Plugs were harvested after 10 days, embedded in paraffin, and sections were stained with Masson’s Trichrome. The number of cells migrated into the matrigel was counted and results plotted as average number of cells migrated in each group (n=3 mice per group; 2 matrigel plugs per mouse).

Discussion

An anti-angiogenic role of CD36 was discovered more than 20 years ago when it was identified as the endothelial receptor responsible for mediating anti-angiogenic activities of Thrombospondin-1 (TSP-1)^13,32. Later oxLDL, another ligand for CD36, was shown to have anti-angiogenic effects³³. We now demonstrate that EV from multiple cellular sources also mediate MVEC signaling via CD36, inhibiting migration and tube formation. We have recently shown that exosomes, unlike EV, did not bind CD36 EV³⁴. With the literature having contradictory reports on the effect of EV on angiogenesis, we now provide evidence that in EC expressing CD36, EV inhibit migration and tube-formation. We have also identified a mechanism by which EV inhibit these activities via CD36 (Figure 6).

Figure 6. CD36-mediated signaling mechanisms triggered in MVEC by extracellular vesicles.

Extracellular vesicles induce ROS in EC in a CD36 and phosphatidylserine-dependent manner via Fyn kinase and NADPH oxidase leading to inhibition of endothelial cell migration and tube formation.

EV derived from leucocytes²⁴ and endothelial cells^21,22,35 were reported to have both pro- and anti-angiogenic activities³⁶. These studies showed that EV carrying Sonic Hedgehog (Shh) induced angiogenesis and those that did not have Shh were anti-angiogenic. Some studies have attributed the differential effect of EV on angiogenesis to the concentration of EV; lower concentrations could be pro-angiogenic and higher concentrations anti-angiogenic^20–22. The agent used to induce EV generation has also been implicated in the dual effect of EV on angiogenesis³⁷. In our work, we used EV generated in vitro from varied sources – EC derived EV (MVEC and HUVEC), circulating blood cell derived EV (RBC) and a tumor cell type derived (acute monocytic leukemia – THP1 cells) and generated by different agents including calcium ionophore, cycloheximide and tumor necrosis factor-α, and found that all inhibited MVEC tube-formation. To test EV that were not induced in vitro, we isolated endogenous circulating EV from healthy humans and subjects with vasculitis^38–40 and found them to inhibit MVEC migration as well. This result is also relevant in terms of the effect of physiological EV concentrations because circulating EV are a pool from various cell types including platelets, endothelial cells, RBCs and leucocytes with the predominant contribution from platelets. Therefore, our results show that irrespective of the agent used, the cell type of origin, or endogenous or induced in vitro, EV have an inhibitory activity on MVEC migration and tube-like structure formation although the degree of inhibition vary between different EV. We also found that the inhibitory effect of EV was not related to the presence of exosomes in the EV preparations, and that the inhibition was not due to induction of apoptosis by EV.

EV express different surface markers and carry a wide variety of proteins, RNA, micro-RNA, and other intracellular molecules depending upon their cell of origin and the stimulus that induced their formation. Although we did not extensively characterize the EV used in our studies, the only known common feature of EV from different sources is the exposure of PS during their formation on the outer leaflet of their membrane. CD36, a known anti-angiogenic receptor, has been shown to bind PS on apoptotic cells ^6,41–44 to facilitate their clearance by macrophages. Recent observations have shown that PS-CD36 interaction-mediated signaling induces a prothrombotic phenotype in platelets¹⁹. This led us to hypothesize that PS exposed on EV can interact with CD36 expressed on MVEC to induce anti-angiogenic signaling. We have shown by in vitro studies by blocking PS with anionic binding protein Annexin V, that the inhibitory effect of EV on MVEC migration is PS dependent consistent with a CD36-dependent process. Given the complexity of EV with respect to various cargo and surface proteins, it is not surprising that reports describing both pro- and anti-angiogenic effects of EV on angiogenesis have been published, but the importance of our work is in identifying a mechanism by which EV inhibit pro-angiogenic activities in endothelial cells expressing CD36.

We demonstrated CD36-dependence of the inhibitory effect of EV by two different strategies. We found that the inhibitory effect of EV on EC migration was significantly less in HUVEC that do not express CD36 compared to MVEC with rich expression of CD36 and that transfection of murine CD36 onto HUVEC restored the inhibition by EV. We also exploited the differential expression of CD36 on different batches and different passages of MVEC and found that MVEC with higher CD36 were inhibited more than MVEC with lower CD36 expression. Levels of inhibition in late passage MVEC that have lost CD36 expression were similar to HUVEC. Thus our results show a CD36 dependent inhibitory effect of EV on EC migration.

We also defined EV signaling events downstream of CD36 showing that EV activated the Src family kinase Fyn which is consistent with the established anti-angiogenic signaling pathway of CD36 in MVEC¹³. Our lab has previously shown that migration of macrophages is inhibited by CD36-mediated ROS generation via NADPH oxidase²⁸. We now showed that in MVEC, EV induced ROS, which was dependent on NADPH as shown by inhibitor studies with apocynin, DPI and SOD mimetic. These results are consistent with other reports suggesting EC-derived EV ^22,45 and RBC-derived EV⁴⁶ induced ROS via NADPH. We have also shown that inhibiting Src family kinases and NADPH oxidase can partially rescue EV-induced inhibition of MVEC migration which is consistent with a partial inhibition of ROS generation by Annexin V in mCD36-HUVEC underlining the functional relevance of these CD36 downstream signaling molecules in the inhibitory role of EV.

Our in vitro findings on the inhibition of cellular migration by EV is supported by our in vivo results that THP1-derived EV inhibited VEGF-induced migration of cells into matrigel plugs implanted in WT, but not cd36 null mice. Since angiogenesis is a process of the microvasculature, MVEC are a better in vitro model system and are more representative of angiogenic EC than HUVEC. Interestingly we observed inhibition of HUVEC migration by EV, though to a much lower extent than MVEC, suggesting that there is/are CD36 independent inhibitory pathway(s) triggered by EV that can be probed in the future. Our findings on the partial rescue effect on the inhibition of NADPH and Src family kinases suggest that there could be other CD36 dependent pathways that still need to be explored. Circulating EV levels are increased in many diseases including cardiovascular diseases and malignancies that are angiogenesis dependent. CD36 in the microvasculature and PS expressed on EV could be playing an important role in these conditions and our study can promote further research to better understand and treat these conditions.

In summary, we showed that EV from different sources in vitro and endogenous EV inhibit EC pro-angiogenic activities in a CD36-PS dependent manner involving NADPH and Fyn kinase mediated ROS generation. These studies identify a novel property of a CD36 ligand that may have relevance to angiogenesis and give a new direction that could be exploited therapeutically.

Significance.

Extracellular vesicles (EV), ≤1μm vesicles shed from various cell types, are present in our circulation and their levels are altered in disease states. Increasingly recognized for their varied biological activities, they have been shown to influence angiogenesis but with contradicting reports. Our studies have shown that MP, irrespective of their cell of origin, inhibit endothelial cell migration via the anti-angiogenic receptor, CD36, and have elucidated the mechanisms involved that can be exploited to design better therapies to treat angiogenesis dependent disorders.

Non-standard Abbreviations and Acronyms

MVEC

Microvascular endothelial cells

HUVEC

Human umbilical vein endothelial cells

EV

Extracellular vesicles

PS

Phosphatidylserine

TSP-1

Thrombospondin–1

TSR

Thrombospondin type I structure homology repeats

BAI-1

Brain angiogenesis inhibitor 1

hEV

Extracellular vesicles from human plasma

EVPP

Extracellular vesicle poor plasma