Extracellular Vesicles Work as a Functional Inflammatory Mediator Between Vascular Endothelial Cells and Immune Cells

Extracellular vesicles (EV) mediated intercellular communication between monocytes and endothelial cells (EC) might play a major role in vascular inflammation and atherosclerotic plaque formation during cardiovascular diseases (CVD). While critical involvement of small (exosomes) and large EV (microvesicles) in CVD has recently been appreciated, the pro- and/or anti-inflammatory impact of a bulk EV (exosomes + microvesicles) on vascular cell function as well as their inflammatory capacity are poorly defined. This study aims to unravel the immunomodulatory content of EV bulk derived from control (uEV) and TNF-α induced inflamed endothelial cells (tEV) and to define their capacity to affect the inflammatory status of recipients monocytes (THP-1) and endothelial cells (HUVEC) in vitro. Here, we show that EV derived from inflamed vascular EC were readily taken up by THP-1 and HUVEC. Human inflammation antibody array together with ELISA revealed that tEV contain a pro-inflammatory profile with chemotactic mediators, including intercellular adhesion molecule (ICAM)-1, CCL-2, IL-6, IL-8, CXCL-10, CCL-5, and TNF-α as compared to uEV. In addition, EV may mediate a selective transfer of functional inflammatory mediators to their target cells and modulate them toward either pro-inflammatory (HUVEC) or anti/pro-inflammatory (THP-1) mode. Accordingly, the expression of pro-inflammatory markers (IL-6, IL-8, and ICAM-1) in tEV-treated HUVEC was increased. In the case of THP-1, EC-EV do induce a mixed of pro- and anti-inflammatory response as indicated by the elevated expression of ICAM-1, CCL-4, CCL-5, and CXCL-10 proteins. At the functional level, EC-EV mediated inflammation and promoted the adhesion and migration of THP-1. Taken together, our findings proved that the EV released from inflamed EC were enriched with a cocktail of inflammatory markers, chemokines, and cytokines which are able to establish a targeted cross-talk between EC and monocytes and reprogramming them toward a pro- or anti-inflammatory phenotypes.

inTrODUcTiOn Atherosclerosis is a chronic and progressive inflammatory vas cular disorder that largely contributes to the development of cardiovascular diseases (CVD) including coronary artery and peripheral vascular disease (1). Tightly regulated inflammatory interactions between two major cellular players, monocytes (MC) and endothelial cells (EC), play a pivotal role in atherosclerotic plaque formation in the arterial intima (2). EC have been known as the major functional coordinator in the cardiovascular homeo stasis and maintaining cardiac functions (3). Accumulating epidemiological and clinical evidence in CVD since 1970 suggest that traditional risk factors such as smoking, elevated blood sugar, hypertension, diabetes, infection, homocysteine, ischemia, and oxidant damage evoke endothelial dysfunction and reprogram them toward either pro and/or antiinflammatory actions (4). Accordingly, overexpression of adhesion molecules [e.g., vas cular cell adhesion molecule1, intercellular adhesion molecule (ICAM)1] on the EC surface together with the secretion of cytokines and chemokines lead to the recruitment of circulating MC into the intima (5,6). Functionally, transmigrated MC will initiate the formation of atherosclerotic plaques, termed fatty streaks, in the arterial walls that, in turn, will lead to CVD (7). So far, the communication of EC with their neighboring EC as well as circulating MC during development of CVD is largely unknown. In recent times, the findings of extracellular vesicles (EV) have opened new perspectives in the understanding of cell-cell communication during the development of several dis eases including CVD (8). EV, traditionally classified as exosomes (40-100 nm), microvesicles (100 nm-1 μm), and apoptotic bodies (>1 μm), have received extensive attention as a novel cell freesignaling conveyors of bioactive molecules in the body fluids and, which can have dramatic impact on the fitness of their reci pient cells (9,10). However, many studies have been focusing on the participation of a certain fraction of EV (e.g., exosome) in the progression of CVD at RNA level (11,12). In spite of that, the protein profile of EV and their mode of action at the site of inflamed vascular cells are still not well defined. In this study, we first aim to unravel the immunomodulatory content of EV bulk derived from inflammatorytriggered EC, thereafter, to under stand their pathological and functional impact on the cellular profiles and behavior of recipient cells.
In order to understand the underlying mechanism of the involvement of EV in the crosstalk between two CVD key graPhical aBsTracT | Immunomodulatory content of bulk extracellular vesicles (EV) (exosomes + microvesicles) derived from untreated (uEV) and TNF-α treated endothelial cells (tEV) and unraveling the functional inflammatory impact of uEV and tEV on the phenotype and behavior of monocytes (THP-1) and endothelial cells (HUVEC) in vitro.
players (EC and MC), transmission electron microscopy (TEM), nanosight tracking analysis (NTA), and western blot were used to confirm the presence of EV (exosomes + microvesicles) in the culture supernatant of a human vascular endothelial cell model (HUVEC), either untreated (uEV) or treated with TNFα to induce an inflammatory stress (tEV). Furthermore, human inflammation antibody arrays were used to discover the immunomodulatory content of both uEV and tEV. Thereafter, HUVEC and a circulating human MC model (THP1) were exposed to uEV or tEV. Relevant pro/antiinflammatory mark ers [IL1β, IL4, IL6, IL6R, IL8 (CXCL8), IL10, IL13, TNFα,  ICAM1, CCL2 (MCP1), CD40, HSP70, CXCL10 (IP10), CCL4 (MIP1), CCL5 (RANTES), TIMP2] were evaluated at the protein in both cell types. Furthermore, the functional inflammatory effect of uEV and tEV was assessed using in vitro monocyte adhesion and migration assays. We discovered that EV may selectively transfer functional inflammatory media tors to their target cells. Accordingly, they were dramatically altering the cellular profile of their recipients toward either pro inflammatory (HUVEC) or anti/proinflammatory (THP1) via the expression of several inflammatory markers. In addition, these biologically active EV induced the THP1 migration and the adhesion of THP1 into HUVEC. Altogether, our cur rent findings for the first time highlighted that the EV released from inflamed EC were enriched with a cocktail of inflammatory proteins, chemokines, and cytokines. These findings also dem onstrate that ECEV are able to establish a targeted crosstalk between EC and MC as well as reprogramming them toward a pro or antiinflammatory phenotypes, resulting in the adhesion and mobilization of MC.

cells and culture conditions
HUVEC (BD Bioscience, cat # 354151) at passages three to six were seeded at a density of 600,000 cells in EBM2 (Lonza) supplemented with EGM2 MV SingleQuot Kit (Lonza) and 5% vesiclesdepleted fetal bovine serum (System Bioscience). When HUVEC were grown up to 70-75% confluency, cells were washed twice with HEPES buffer saline (Lonza) and cells were then inflammatory triggered by adding 10 ng/ml TNFα in refreshed medium for overnight (13). Afterward, the supernatants were collected for the EV isolation. All collected supernatant samples containing EV were stored at −80°C until EV isolation procedures.

eV isolation
A modified differential centrifugation method was used to collect the bulk ECEV containing large EV (microversicle) and small EV (exosomes) from cell culture supernatant of unstimulated (uEV), TNFα stimulated (tEV), and cellfree medium (cEV). Briefly, collected supernatant from the same number of parent cells was first centrifuged at 300 g for 5 min at 4°C to eliminate cell debris. To remove remaining debris and apoptotic bodies, another centrifugation step was done on the supernatant passed through a 0.22µm filter (VWR, Belgium) for 20 min at 2,000 g at 4°C (14). Afterward, to pellet the ECEV, the supernatant was centrifuged at 110,000 g for 3 h at 4°C. All ultracentrifugation (UC) steps were performed using an L90 Beckman centrifuge (Beckman Instruments, Inc., Fullerton, CA, USA) equipped with a Ti70 rotor (Beckman Instruments) (15). Based on the downstream analysis, pellets were suspended in 1 ml of HEPES (Lonza), RIPA or extraction buffers (Abcam).

nanosight Tracking analysis
Extracellular vesicles size distribution and concentration were analyzed based on the tracking of light scattered by vesicles moving under Brownian motion using the NanoSight NS300 system (Sysmex Belgium N.V.) equipped with a 532nm laser. The data were captured and analyzed using NTA software 3.2 (NanoSight Ltd.). Samples were diluted with PBS over a range of concentrations to obtain between 20 and 50 particles per frame. Samples were injected into the sample chamber and measured three times for 60 s at 25°C with manual shutter and gain adjust ments for three individual samples.

Transmission electron Microscopy
Transmission electron microscopy samples were prepared and analyzed as previously described (16). The size and morphology of ECEV were evaluated using a Tecnai G2 transmission electron microscope (TEM; Tecnai G 2 spirit twin, FEI, Eindhoven, the Netherlands) at 120 kV. The microscope was provided with a bot tom mounted digital camera FEI Eagle (4k × 4k pixels) to acquire images of the evaluated samples. Digital processing of the images was performed with the FEI imaging software (TEM Imaging and Analysis version 3.2 SP4 build 419).

live imaging
Labeling of ECEV and cEV was performed by adding 50 µg/ml CellMask™ orange plasma membrane tracking label for 10 min at 37°C into the supernatant. Free dye was removed from labeled EV using Amicon ® Ultra centrifugal columns (10 kDa cutoff) after isolation procedures. Labeled EVs were added to approximately 1 × 10 6 of HUVECs cell per well in an eightwell culture plate (Ibidi GmbH, Martinsried, Germany). In the case of THP1, labeled EV were added into polydlysinecoated glass coverslips (Sigma) which were seeded overnight with 8 × 10 5 undifferenti ated THP1 in sixwell plates. Following 2-24 h of incubation, the live cell imagining of internalized of EV was performed using Zeiss LSM 510 META confocal laser scanning microscope (Jena, Germany) on an Axiovert 200 M motorized frame for TICS, STICS, and STICCS analyses. The microscope was coupled to a 30 mW aircooled argon ion laser emitting at 488 nm under the control of an acoustooptic modulator (~11 µW irradiance at the sample position) for onephoton excitation. To provide a suitable environment for sustaining cells during the imaging steps, the microscope was equipped with an airtight chamber (Tempcontrol 37-2 digital, PeCon, Erbach, Germany) with con trolled temperature at 37°C. Cellfree mediumderived EV served as a negative control. Nuclei were stained with Hoechst 33342.

Protein Quantification
Extracellular vesicles protein lysates in RIPA buffer for western blotting, EV protein lysates in extraction buffer (ab193970, Abcam Ltd., Cambridge, UK) for ELISA and inflammatory cytokine arrays and EV suspensions for migration and adhesion assays were quantified using the Pierce BCA Protein Assay Reagent Kit (Thermo Scientific Pierce, USA) following the manufacturer's protocol. Optical density of standards and samples were measured at OD595 nm using a Multiskan™ FC Microplate Absorbance Reader (Thermo Scientific, Belgium).

inflammatory cytokine arrays
To simultaneously detecting and semiquantifying of 40 inflam matory markers in EV and cell lysates, human cytokine anti body C1, C2, and C3 arrays were purchased from RayBiotech (Boechout, Belgium). Experiments were done according to the manufacturers' instructions. Briefly, 25 µg of EV lysate or cell lysate proteins in extraction buffer (ab193970, Abcam Ltd., Cambridge, UK) were added in to a preblocked membrane and incubated overnight at 4°C with gentle shaking. Thereafter, the membrane incubated with the primary biotinconjugated antibody for 2 h, followed by incubation with HRPconjugated streptavidin antibodies for 1 h. The signal intensity of each array was scanned by densitometry using the ImagerQuant™TL detection system. Intensity of each dots was then quantified using ImageJ open source software (National Institutes of Health, USA). Heat maps of inflammationrelated protein expression was analyzed using GENEE open source software. , and other known CVD marker [CD40 (ab99990) and HSP70 (ab187399)] were performed and normalized for 1 µg total protein of cell lysates and EV lysates using Human ELISA Kits (Abcam Ltd., Cambridge, UK), according to their manufacturer's instruc tions. Cellfree mediumderived EV (cEV) served as a negative control. Optical density of standards and samples were measured using a Multiskan™ FC Microplate Absorbance Reader (Thermo Scientific, Belgium).

Western Blotting
The equivalent of 5 µg of EV proteins in RIPA buffer containing protease inhibitor cocktail (SigmaAldrich) were first separated by SDSPAGE with 8 or 12% polyacrylamide gels under 200 V for 30-45 min. The proteins were then electrophoretically trans ferred to a polyvinylidene fluoride membrane for minimum 1 h at 350 mA. The membranes were blocked with PBS Marvel 5% for 2 h and incubated with 1:1,000 dilution of primary antibodies against CD9, CD63, ICAM1, GM130 (negative control), and βactin (reference protein) overnight at 4°C. Next, rabbit anti mouse HRPconjugated secondary antibody at 1:2,000 dilution (Agilent, USA) were added in to the membrane for 1 h at room temperature (RT). The blots were developed with Pierce™ ECL Western Blotting Substrate. The corresponding bands were detected by the ImagerQuant™TL detection system. Intensity of each bands (2×) was quantified using ImageJ open source software (National Institutes of Health, USA) (17).

immunofluorescence staining
HUVECs were first grown into fourwell culture slides (Sarstedt, Berchem, Belgium) up to 70-75% confluency. Cells were then stimulated with PBS, 10 ng/ml TNFα (ImmunoTools), uEVs or tEVs for 24 h. An equal amount of EV with total protein concentration (10 µg/ml) was added to the cell cultures with the use of the BCAassay results. After treatment, HUVEC were fix ated and permeabalized with 4% paraformaldehyde for 10 min at RT and then rinsed with PBS twice. Specimens were incubated with the corresponding primary antibody against ICAM1 (1:500 in PBS) for overnight at RT. After three times washing with PBS (Lonza), the secondary antibody donkeyantimouse Alexa 488 (1:1,000 in PBS, Thermo Fisher Scientific) was applied into each chamber for 1 h at RT in the dark. Nuclei were stained with DAPI. Images were taken with a Leica DM4000 B LED micro scope along with a digital microscope camera Leica DFC450 C (Leica, Diegem, Belgium). ImageJ open source software (National Institutes of Health, USA) was used to calculate the mean of fluorescence intensity (MFI) for each protein of interest under different treatments in HUVEC and THP1. The MFI was meas ured by subtracting the multiplication of the area of the selected cell and the mean fluorescence of the background readings from the integrated density of each cell.
Transmembrane Migration assay THP1 cells were harvested from RPMI1640 medium sup plemented with 10% FBS and washed twice with PBS, then, incubated in serum free medium for 2 h. EV samples in the experiments were diluted in RPMI1640 medium containing 0% FBS. The migration capacity of THP1 was determined using 8 µm pore polycarbonate filter transwell plates (ThinCert Cell Culture Inserts, Greiner bioone, Vilvoorde, Belgium). Briefly, 300 µl of the above prepared THP1 (10 6 cells/ml) were seeded on top of the transwell insert and the lower chambers were filled with 500 µl RPMI1640 medium containing 0% FBS with or without 10 µg/ml of uEV and tEV samples. RPMI1640 sup plemented with 10% FBS (Thermo Fisher Scientific) and 50 ng/ml recombinant human MCP1 (PeproTECH, Rocky Hill, CT, USA) were used as positive controls. After overnight incubation (~16 h) at 37°C, the number of cells that passed through the membrane were counted in the lower chambers using trypan blue 0.4% (Thermo Fisher Scientific). The percentage of migrated cells for each condition in three independent experiments with three technical replicates (n = 9) were calculated.
cell adhesion assay HUVEC were first grown into fourwell microscope slides (Sarstedt-Germany) up to 70-75% confluency. HUVEC were treated with PBS, 10 ng/ml TNFα (ImmunoTools) and a 10 µg/ ml uEVs and tEVs overnight (~16 h). Nuclei of HUVEC were stained with Hoechst33342 staining solution (Thermo Fisher Scientific) and cells washed twice with PBS to remove the non engulfed EV and dye residuals. THP1 cells were also grown in RPMI1640 medium supplemented with 10% FBS. THP1 were stained with 5 µM Calcien AM, for 15 min at 37°C, washed twice with PBS. Fluorescently labeled THP1 were coincubated with the pretreated HUVEC for 60 min at 37°C. Afterward, HUVEC were thoroughly washed with PBS (6×) to remove the non adherent THP1 cells. Images were taken with a Leica DM4000 B LED microscope supplemented with a digital microscope camera Leica DFC450 C (Leica, Belgium). ImageJ open source software (National Institutes of Health, USA) was used to calculate the percentage of adhered THP1 monocytes to HUVEC under dif ferent treatments (18).

statistical analysis
Data were presented as mean ± SD of three independent experiments in two technical replicates (n = 6) or three techni cal replicates (n = 9). Oneway analysis of variance (ANOVA) with a multiple comparisons test (Tukey's multiple comparison test) and Student's test using the statistical packages GraphPad Prism 7.04 software (GraphPad Software, Inc., USA) were applied to evaluate the statistical significance between differ ent treatments. Twotailed tests at value of *p < 0.05 and were considered as statistically significant. NS represented as not significant, p > 0.05. resUlTs cross internalization of ec-eV into Vascular ec (hUVec) and circulating immune cells (ThP-1) Extracellular vesicles bulk were pelleted from HUVEC cell cul ture supernatant using a modified differential UC. UCpurified EV contained a mixture of large (microvesicles) and small EV (exosomes) (TEM image: Figure 1A and NTA analysis: Figure S1 in Supplementary Material). In line with previous data, UCisolated EV from either untreated EC (uEV) or EC treated with TNFα (tEV) were enriched with both classical EV membranebound biomarkers including CD9, CD63, CD81, and ICAM1 (16). Comparative marker analysis of selected classical (CD9 and CD63) and inflammatory (ICAM1) asso ciated markers was performed on the bulk of uEV and tEV using western blot. CD9 (24 kDa), CD63 (30-70 kDa), and ICAM1 (90 kDa) were highly enriched in EV bulk derived from TNFα stimulated HUVEC (tEV) in comparison with EV derived from unstimulated cells (uEV) (Figure 1B). GM130 (a Golgirelated protein) was used as a negative marker protein for EV. The absence of the GM130 (130 kDa) in uEV and tEV confirmed the purity of samples. Within 3 h EV derived from EC (HUVEC) were taken up by HUVEC ( Figure 1D) and THP1 ( Figure 1F) from EVsupplemented culture medium and predominantly accumulated in the perinuclear region. No vesicles were detected in the control experiments (EV isolated from cellfree medium) (Figures 1C,E). Altogether these observations confirmed that inflammatorytriggered EC secreted a bulk of EV containing large and smallsized vesicles which were taken up by vascular EC (HUVEC) and circulating immune cells (THP1).

ec-eV immunomodulatory content and Their Mode of action
There is insufficient evidence concerning the mode of action of released EV during an inflammatory stress response. In order to obtain a complete overview on the inflammatory content of the ECEV in this study, antibodypairbased assays and ELISA were used to detect several EVassociated inflamma tory mediators simultaneously. Therefore, we first assayed the immunomodulatory content of uEV and tEV lysate using human inflammatory arrays C1 and C2 (Figure 2A). These arrays include several inflammatory markers such as cytokines, growth factors, cellular adhesion, and inflammationassociated markers. Among 40 pro and antiinflammatory proteins, GMCSF, IL6, IL8, ICAM1, CXCL10, CCL5, TNFα, and TNFR were significantly higher expressed in the tEV as com pared to uEV ( Figure 2B). We also observed that the detected intensity for CCL2 in the tEV was slightly higher than uEV ( Figure 2B).
To further confirm the array defined markers and quantify the EV pro and antiinflammatory protein content, ELISA based assays for GMCSF , IL1β, IL4, IL6, IL6R, IL8, IL10,  IL13, ICAM1 were statistically significantly increased in the tEV as compared to uEV (Figure 2C). These data already show that EV derived from inflammationtriggered EC are highly enriched with several key proinflammatory mediators, chemokines whereas antiinflammatory mediators (IL10 and IL13) were barely expressed in them. In order to find out the role of these inflam matory EV in the cytokine and chemokine networks during inflammatory mediated crosstalk between EC and MC as well as their functional effect on these two recipients, we further  investigated the physiological impact of EV derived from TNF α stimulated HUVEC (tEV) and nonstressed (unstimulated) cells (uEV) on two major CVD cell culture models, HUVECs (reference cell culture model for EC) and THP1 (reference cell culture model for MC) at both protein and RNA levels and functional behavior in vitro. In addition, negligible amounts of cytokines and chemokines were detected in EV derived from cellfree medium treated with 10 ng/ml TNFα as negative control ( Figure 2C).

ec-eV alter the inflammatory Profile of Mc (ThP-1) and ec (hUVec)
To assess whether ECEV shuttle the inflammatoryassociated proteins and induce their expression in HUVEC and THP1 at the protein level, we performed an semiquantitatively protein array and ELISA on the cell lysate of recipient cells treated with uEV, tEV, TNFα (positive control), and PBS (negative control) for 18 h. First, a membranebased inflammation array C3 was used to detect the differentially expressed cytokines, growth factors, cellular adhesion, and inflammationassociated mark ers, concurrently (Figures 2A,B in Supplementary Material). Expression of a wide range of inflammatory markers was evi dent in the TNFα and EV treated HUVEC and THP1. Heat map analysis of differentially expressed proteins revealed that among 40 human inflammatory markers, a series of chemotactic cytokines and adhesion promoters including ICAM1, IL6R, CXCL10, CCL2, CCL4, CCL5, TIMP2, and several ILs were the most highly expressed in both cell types (Figures 3A,B).
Collectively, these results suggest that EV content may selec tively transfer inflammatory markers to recipients and altered their cellular profiles differently. In particular, they promoted a pro inflammatory behavior in HUVEC, whereas they reprogrammed THP1 toward a mixed of pro and antiinflammatory phenotype as indicated by elevated expression of ICAM1, CCL4, CCL5, and CXCL10.

ec-eV increase the expression of adhesion Molecule (icaM-1) in Mc (ThP-1) and ec (hUVec)
Intercellular adhesion molecule1 expression is one the key candi date for inflammationassociated disorders. In our protein studies, we discovered the expression of this marker was significantly induced in HUVEC and THP1 treated with ECEV. Therefore, to understand if ECEV can actively induce inflammation in EC and MC, the induction of ICAM1 as a key candidate of inflam mation was immunofluorescently visualized and quantified ( Figure 5). In the line with ELISA results, expression of ICAM1 in HUVEC after TNFα and tEV exposure was significantly enhanced (p < 0.0001 and p = 0.0157, respectively) ( Figure 5). A low level of ICAM1 was expressed in PBS treatment HUVEC. Upon stimulation with tEV in THP1, ICAM1 expression was increased (p = 0.0037) whereas only a modest enhancement (p = 0.17) was detected in the uEVtreated THP1.

ec-eV Promote ThP-1 adhesion and Migration
The activation, adhesion, and transendothelial migration of MC into the intima occurs rapidly during development of athero sclerosis. As ECEV are enriched with a cocktail of chemotaxis and migration associated factors, we further investigate whether these EV are actively involved in MC adhesion and migration. The chemotactic effect of uEVs or tEVs on the migration of THP1 were compared with the condition without and with THP1 migration capacities (0% FBS and 10% FBS, respectively). Our data were demonstrated a chemotactic effect of ECEV on THP1 by promoting their transmembrane migration in the presence of ECEV using in an in vitro transwell migration assay. As shown in Figure 6A, when THP1 was incubated with uEV and tEV, THP1 migration enhanced by 32 ± 22.5 and 35 ± 16.7%, respectively (mean ± SD, n = 9) compared to 0% FBS ( Figure 6A). In the response to 10% FBS and MCP1, positive controls, THP1 migration were increased up to 80.5 ± 20 and 64 ± 10.1%, respectively.
Also, a functional adhesion assay was performed to dis cover the effect of ECEV at the crossing of inflammation and development of vascular disease by measuring the adhesion of THP1 monocytes to HUVEC monolayer under static condi tions. As shown in Figures 6B,C, preincubation of HUVEC with either tEV or TNFα effectively increased the adhesion of THP1 (p = 0.002 and p = 0.004, respectively) as compared to PBStreated HUVECs. Exposure of HUVEC to uEV has a slight but not significant effect on THP1 adhesion as compared to PBStreated cells (p = 0.35) but there was a statistically signifi cant difference between uEV and tEV (p = 0.015) treated cells. These observations confirmed the role of EV in the transferring biologically active inflammatory modulators across cells which lead to amplifying the inflammatory response by EC activation and MC adhesion and migration.

DiscUssiOn
The combined of small and large eV Together are associated With inflammation FigUre 4 | ELISA analysis of inflammatory markers including interleukins (IL1-β, IL-6, IL-6R, IL-8, IL-10, and IL-13), cellular adhesion and inflammation-associated marker [intercellular adhesion molecule (ICAM-1)] and chemokines (CCL-2, CCL-4, CCL-5, and CXCL-10) and TIMP-2 in cells extraction of HUVEC (a) and THP-1 (B) treated with PBS, TNF-α (10 ng/ml), uEV (10 µg/ml total proteins), and tEV (10 µg/ml total proteins). Values are given as mean ± SD of three independent biological individuals in two technical replicates (n = 6). p-Values were calculated by one-way analysis of variance with a multiple comparisons test (Tukey's multiple comparison test) and *p < 0.05 (PBS vs treatments) and # p < 0.05 (uEV vs tEV) were considered as statistically significant. observed that during the development of inflammationassociated disorders, a multitude of diseased and healthy cells are constitu tively releasing a heterogeneous mixture of both small and large EV into the body fluids (8). These combined EV are actively contributing to natural intercellular communication via transport ing several bioactive molecules such as peptides, proteins, lipids, mRNAs, and miRNAs (19). Yet, the content and functionality of the combined fraction of both small and large EV are still  missing. To our knowledge, this is the first study to investigate the immunomodulatory content of the combined small and large EV derived from inflamed vascular cells and to discover their effect on the cellular fitness and function of recipients. In order to isolate a combined fraction of both small and large EV, the collected supernatant was first centrifuged at 300 and 2,000 g to eliminate cell debris and apoptotic bodies, respectively (14).
Pelleting of large and small EV together were then happened at 110,000 g. Principally, in the differential centrifugation method, the most commonly used protocol for EV isolation, small and large EV are separated at different gforces and kfactors. As frac tioning of large EV (microvesicles) and small EV (exosome) from different cell types could be done at gforces of 10,000-20,000 and >100,000 g, respectively (14). Therefore, the copelleting of small and large EV was done by skipping the 10,000-20,000 g centrifugation step (Figure 1A; Figure S1 in Supplementary Material).

ec-eV contain several inflammatory Mediators
Several studies have demonstrated that the initiation and progres sion of inflammationassociated disorders such as atherosclerosis and CVD are governed by interactions between EC and MC via multiple inflammatory mediators, the best recognized of which are cell adhesion molecules (e.g., ICAM1), chemoattractants (e.g., CCL2, CCL4, and CCL5), growth factors (e.g., GMCSF), and cytokines (e.g., IL6, IL8) (2,20,21). Although, it is well known that chemokines and cytokines are effectively involved in a complex inflammatory interaction between EC and circulating immune cells, little is known about the ECEV immunomodula tory content and their role in the chemokine network between the two key drivers (EC and MC) after an inflammatory stress response.
In our previous study, we already demonstrated that an elevated level of ICAM1(+) small EV were released from inflammation triggered EC (16). To our knowledge, this study presents the first complete overview of the common immunomodulatory content of the combined fraction of both small and large EV released from inflammatorytriggered EC. Our data suggest that beyond the higher expression of adhesion markers (ICAM1) in EV derived from inflammationtriggered vascular EC, these EV contain several proinflammatory mediators including chemotactic mediators such as IL6, IL8, CXCL10, monocyte chemoattractant protein1 (CCL2), macrophage inflammatory protein (CCL4 and CCL5) together with key antiinflammatory mediators (IL10 and IL13). These EV enriched with a cocktail of inflammatory agents may contribute in the earliest phase of atherosclerosis and CVD which is initiating by endothelial dysfunction, recruiting monocytes/macrophages toward EC and then rolling and transendothelial migration of MC into the intima.

ec-eV Mediate inflammatory responses in ec and Mc
Previous studies have shown that RNA content of EVEC are mainly playing a central role in the educating recipient cells toward inflammatory gene activation or suppression responses (22). However, we show that ECEV harbor a wide range of inflammatory proteins, suggesting that EVassociated proteins could attribute to the functional activity of recipient cells. Several studies have already demonstrated that EV may trans fer inflammationassociated protein (e.g., ICAM1) into their target cells (23,24). Here, comparing whole protein profiles of cell lysate with the EV content (Figures 2 and 3) highlight that EV may be able to selectively transfer the specific inflammatory associated mediators to target cells (e.g., CCL5 and CXCL10 to THP1 and ICAM1, IL6 and IL8 to HUVEC) thereby modulate cells toward either proinflammatory (HUVEC) or pro/antiinflammatory (THP1) statues. Moreover, the elevated expression of ICAM1, IL6, and IL8 in tEVtreated HUVEC, suggest that EV may translocate these proinflammatory media tors and promote vascular endothelial inflammation. In fact, ICAM1 together with IL6, IL8 play an important role in the progression of atherosclerosis through triggering the transen dothelial migration of immune cells to the site of inflammation and the activation of proinflammatory cascades in target cells (5,7,21). We also provide evidence that chemokineenriched EV (tEV) can modulate the expression of antiinflammatory markers including CCL5 and CXCL10 in THP1. Overall, a broad range of proinflammatory proteins in HUVEC and pro/antiinflammatory proteins in THP1 were significantly induced by the bulk of both uEV and tEV compared to the control. It is likely that specific modulators contained in EV may play these extensive inflammatory effects and regulate the expression of a large number of inflammatoryassociated genes. The changes in the phenotype and behavior of recipient cells in this time frame of treatment (an overnight incubation) can be associated with either the transfer of the EV cargo into cells or de novo synthesis of inflammatory markers induced by the EV cargo or can be due to a mixture of both pathways. While the impact of ECEV on the two target cells was investigated in this study, the actual mechanistic pathway of EV involved in these effects as well as their uptake/transfer pathway into recipients are still unclear and needs to be further investigated. Yet, another key mediator for the inflammatory effect of EV would be their RNA cargo. Further investigation is therefore required to detect the RNAsassociated inflammation in the EV derived from inflammatorytriggered EC, profile changes at the transcriptional level and to discover their functional contribution in MC adhesion and mobilization. In this work, both tEV and uEV were first isolated from the same number of parent cells. The total protein concentration of tEV was higher than uEV from the same number of parent cells. In addition, as presented in the Figure S1 in Supplementary Material, higher concentrations of particle number/ml EV was detected in TNFα stimulated HUVEC (tEV) when compared to nonstressed (unstimulated) cells (uEV). In the next step, to understand the effect of EV fractions we normalized EV samples for functional assays where we considered both criteria (particle number and protein concentration). As recommended by Tkach et al. 2018, the combined quantification of total protein and particle number is the best way to quantify materials present in an EV preparation (25). Adjusting both uEV and tEV to 10 µg/ml the total protein was fairly balanced to the same number of EV (e.g., ~1E9). In our opinion, target cells were undergoing the same uEV and tEV treatments.

ec-eV are related to Migration and adhesion of Mc
The majority of studies has been focused on the functional prop erties of EV derived from MC at the crossing of inflammation and development of vascular disease (24,26). ECEV are most likely to be an important coordinator in the cardiovascular homeostasis, maintaining cardiac functions and development of CVD due to the position of their parental origin, EC, at the inter face of vascular cells and immune cells. Here, our study shows that ECEV are not only mediated the THP1 adhesion into HUVEC but are also capable of promoting their transmembrane migration in vitro. Actually, docking of EV proteins into HUVEC and THP1 induces the expression of key chemotactic mediators including IL6, IL8, CXCL10, monocyte chemoattractant pro tein1 (CCL2), and macrophage inflammatory protein (CCL4 and CCL5), leading to increased THP1 adhesion and mobiliza tion. At the functional level, our results also support the idea that ECEV are involved in the chemokine networks between EC and MC at sites of inflammation, in particular, promoting the MC adhesion into EC and recruiting them to inflamed sites. Taken together, our study revealed that ECEV are actively associated with the vascular endothelial inflammation, MCassociated inflammatory response and MC adhesion and migration. In addition, our results extended the previous findings regard ing EV mediated an inflammatory crosstalk between EC and their neighboring EC and circulating MC and we for the first time report that this intercellular communication seems most likely occurring via EVmediated transferring of inflammatory chemokines and cytokines to their local and distant recipients. It should be noted that our conclusion is driven from in vitro studies at the protein and functional levels. In vivo (animal) and ex vivo experiments are planned to explore further the involvement of EV in the communication networks at RNA, protein and functional levels.
aUThOr cOnTriBUTiOns BH, LM, NA, and DM designed the experiments. BH conducted most of the experimental work, data analysis, and interpretation, and drafted and edited the manuscript. SK conducted parts of the experimental work and revised the manuscript. LM, NA, and DM conducted several critical revisions.