Factor VIIa releases phosphatidylserine-enriched extracellular vesicles from endothelial cells by activating acid sphingomyelinase

Abstract

Background:

Our recent studies showed that activated factor (F) VII (FVIIa) releases extracellular vesicles (EVs) from the endothelium. FVIIa-released EVs were found to be enriched with phosphatidylserine (PS) and contribute to the hemostatic effect of FVIIa in thrombocytopenia and hemophilia.

Objective:

To investigate mechanisms by which FVIIa induces EV biogenesis and enriches EVs with PS.

Methods:

FVIIa activation of acid sphingomyelinase (aSMase) was evaluated by its translocation to the cell surface. The role of aSMase in the biogenesis of FVIIa-induced EVs and their enrichment with PS was investigated using specific siRNAs and inhibitors of aSMase and its downstream metabolites. Wild-type and aSMase^−/− mice were injected with a control vehicle or FVIIa. EVs released into circulation were quantified by nanoparticle tracking analysis. EVs hemostatic potential was assessed in a murine thrombocytopenia model.

Results:

FVIIa activation of aSMase is responsible for both the externalization of PS and the release of EVs in endothelial cells. FVIIa-induced aSMase activation led to ceramide generation and de novo expression of transmembrane protein 16F. Inhibitors of ceramidases, sphingosine kinase, or sphingosine-1-phosphate receptor modulator blocked FVIIa-induced expression of transmembrane protein 16F and PS externalization without interfering with FVIIa release of EVs. In vivo, FVIIa release of EVs was markedly impaired in aSMase^−/− mice compared with wild-type mice. Administration of a low dose of FVIIa, sufficient to induce EVs release, corrected bleeding associated with thrombocytopenia in wild-type mice but not in aSMase^−/− mice.

Conclusion:

Our study identifies a novel mechanism by which FVIIa induces PS externalization and releases PS-enriched EVs.

1 |. INTRODUCTION

Extracellular vesicles (EVs) are membrane-enclosed, nano-sized vesicles shed by cells that are involved in intercellular communication via the transfer of cargoes such as nucleic acids, proteins, and lipids [1]. EVs are found to be present in various biological fluids [1]. Emerging evidence indicates that EVs play pivotal roles in multiple physiological and pathologic processes [2–6]. EVs might play an integral role in regulating the coagulation system with their procoagulant and anticoagulant properties and contribute to hemostasis or thrombosis [7–14]. The prothrombotic properties of the EVs are mainly attributed to the expression of procoagulant proteins, such as tissue factor (TF) [15], and the increased presence of negatively charged phospholipids, such as phosphatidylserine (PS), on the EVs’ surface [14,16].

Interestingly, proteases generated during blood coagulation were shown to induce the release of EVs from certain cell types, including endothelial cells. Thrombin was shown to induce the release of EVs from the endothelial cells (EEVs) by caspase-2-dependent activation of ROCK-II [17]. Activated protein C (APC) was shown to generate EEVs via endothelial cell protein C receptor (EPCR)-dependent activation of protease-activated receptor 1 (PAR1) [18]. Our earlier studies showed factor (F) VIIa (FVIIa) binds EPCR [19,20] and activates PAR1-mediated cell signaling in endothelial cells in an EPCR-dependent manner [21,22]. Our recent studies showed that FVIIa also promotes the release of EVs from endothelial cells via EPCR-PAR1-mediated signaling [23]. Our studies suggest that FVIIa-released EEVs confer hemostatic effects in therapeutic settings and reduce bleeding associated with hemophilia and severe thrombocytopenia [23]. The hemostatic potential of FVIIa-released EEVs is TF-independent and appears to stem from the increased presence of PS in the outer leaflet of FVIIa-released EEVs [23].

Recombinant FVIIa (rFVIIa) is used as a bypassing therapeutic agent to treat persons with hemophilia with inhibitors [24–26]. rFVIIa is also used as a clinically approved drug or off-label to treat bleeding associated with congenital and acquired bleeding disorders [27,28]. Evidence indicates that rFVIIa provides a hemostatic response in therapeutic settings in a platelet-dependent mechanism [28]. However, bleeding associated with congenital or acquired platelet defects was also treated successfully with rFVIIa [29]. Interestingly, rFVIIa was shown to be effective in treating bleeding in a patient with severe thrombocytopenia in whom platelet-enhancing treatment was unresponsive [30]. The above data indicate that platelet-independent mechanisms may also play a pivotal role in FVIIa conferring the hemostatic effect.

Activated platelets play a crucial role in supporting hemostasis by providing the PS-enriched membrane surface [16]. FVIIa-released EEVs with increased PS levels on the surface could function as mini platelets in supporting hemostasis. At present, mechanisms by which FVIIa releases EEVs and their enrichment with PS are unknown. Understanding these mechanisms may provide novel mechanistic insights into the biogenesis of EVs and their selective enrichment with PS. They could also provide insights into the pathophysiologic functions of FVIIa-released EVs and their therapeutic potential.

In the present study, for the first time, we provide a novel mechanism by which FVIIa releases EVs and enriches them with PS. Our studies show that FVIIa, via the EPCR-PAR1-dependent intracellular activation of the β-arrestin1-phosphoinositide 3-kinase (PI3K)-RAC (Rho family)-alpha serine/threonine-protein kinase(AKT) pathway, induces the translocation of acid sphingoe (aSMase) to the cell surface. The translocation of aSMase to the cell surface leads to ceramide generation and, consequently, sphingosine-1-phosphate (S1P)-mediated S1P receptor (S1PR) activation. The activation of S1PRs, in turn, induces the expression of transmembrane protein 16F (TMEM16F) that acts as a phospholipid scramblase and promotes PS externalization. FVIIa activation of aSMase, independent of downstream signaling events, also promotes the release of EEVs. The biogenesis of EEVs coupled with FVIIa-induced PS externalization to endothelial cell surface leads to the generation of PS-enriched EEVs from endothelial cells exposed to FVIIa. Our in vitro data are supported by the in vivo analyses in a murine model system that show aSMase deficiency not only impairs FVIIa release of EVs but also attenuates the incorporation of PS into FVIIa-released EVs and thus fails to protect against bleeding induced by thrombocytopenia.

2 |. MATERIALS AND METHODS

2.1 |. Mice

Wild-type (WT) C57BL/6J mice were obtained from Jackson Laboratories and bred in-house. Breeding pairs of aSMase^−/− mice of C57BL/6J background were provided by Dr Richard Kolesnick, Memorial Sloan Kettering Cancer Center, New York.

2.2 |. Endothelial cells

Primary human umbilical vein endothelial cells (HUVECs) were obtained from Lonza and cultured in endothelial basal medium (EBM-2; Lonza) supplemented with endothelial-specific growth factors (Lonza). Murine bEnd.3 endothelial cells were obtained from American Type Culture Collection. Murine brain endothelial cells from WT and transgenic mice were isolated and cultured as described in our recent publication [31]. Before treating cells with FVIIa (25 nM) or a control vehicle (CV), they were maintained for 1 hour in a serum-free medium.

2.3 |. Silencing of aSMase, EPCR, PAR1, β-arrestin1, β-arrestin2, S1PR1, TMEM16F, or TMEM16D

HUVECs were transfected with siRNA specific for aSMase, EPCR, PAR1, β-arrestin1, β-arrestin2, S1PR1, TMEM16F, or TMEM16D (200 nM). As controls, cells were transfected with a control RNA having a scrambled nucleotide sequence. The transfection was carried out using Lipofectamine RNAiMAX reagent (ThermoFisher Scientific) in a serum-free medium, and the cells were used in experiments 48 hours posttransfection.

2.4 |. Isolation of EVs

EVs from cell culture supernatant medium and murine blood were isolated and quantified by nanoparticle tracking analysis as described in our recent publications [23,32,33].

2.5 |. Factor Xa generation and prothrombinase assays

The rate of activation of factor X (FX) and prothrombin on the surface of endothelial cells and the EVs was measured in a chromogenic assay as described previously [23,34].

2.6 |. Animal studies

To investigate the role of aSMase in FVIIa-induced EVs generation in vivo, WT and aSMase^−/− mice were injected with saline or FVIIa (0.25mg/kg body weight in 100 μL) via the tail vein. Two hours following FVIIa administration, blood was drawn into citrate anticoagulant, and EVs were isolated from platelet-free plasma and quantified as described in our recent publication [23].

To investigate the role of aSMase in FVIIa-EEVs hemostatic effect in vivo, WT mice were depleted of platelets by administering platelet depletion antibody (CD42b monoclonal antibody, 1 mg/kg, Emfret, Germany). Five hours following the antibody administration, mice were injected via the tail vein with EVs (1 × 10⁹ EVs/mouse) isolated from control- or aSMase-silenced cultured endothelial cells that were treated with a CV or FVIIa (25nM). Five minutes later, mice were subjected to the saphenous vein incision. Blood coming from the incision site was absorbed into Kim wipes to measure blood loss, as described in our earlier publication [35]. A small aliquot of blood from the injury site was also collected into citrate anticoagulant, and thrombin-antithrombin (TAT) complex levels were measured as an index for thrombin generation. A small aliquot of blood collected from the submandibular vein was also subjected to TAT analysis as a measure of systemic thrombin generation.

In the second set of experiments, platelet-depleted WT and aSMase^−/− mice were administered with saline or FVIIa (0.25 mg/kg) via the tail vein injection. Two hours later, when the FVIIa-induced EVs population in the plasma reached the maximum [23], and exogenously administered FVIIa was cleared from the circulation [36], a saphenous vein incision was performed. Blood loss volume and TAT generation at the wound site, as well as in the systemic circulation, were measured as described above.

2.7 |. Statistical analysis

All in vitro experiments described here were repeated at least 3 times. Six mice were used for each experimental group. The data were presented as mean ± SEM. The Mann–Whitney U-test was used to determine statistically significant differences between the 2 groups. One-way analysis of variance, followed by Tukey’s or Dunnett’s post hoc multiple comparison tests, was used to determine the statistical significance among >2 groups.

2.8 | See Supplementary Methods for extended experimental details

3 |. RESULTS

3.1 |. FVIIa triggers aSMase translocation to the plasma membrane, which critically regulates PS externalization to the outer leaflet of the endothelial cell membrane

Our earlier studies indicate that FVIIa-released EEVs contain significantly higher levels of PS on their surface as compared with EEVs released under basal conditions [23,34]. Our earlier studies also suggest that FVIIa treatment might lead to the externalization of PS to the outer leaflet of the endothelial cell membrane [23]. However, the mechanism by which FVIIa induces PS externalization and the role of this process in the biogenesis of PS-enriched EEVs by FVIIa is unknown. Sphingomyelinases that alter the lipid composition of the plasma membrane could play a role in the externalization of PS [37] as well as in the biogenesis of EVs [38,39]. aSMase is generally localized in lysosomal compartments in unperturbedd cells but could be secreted or translocated to the cell surface upon activation by various stimuli [40]. Therefore, we investigated whether FVIIa treatment induces the translocation of aSMase to the outer leaflet and its role in FVIIa-induced PS externalization and the biogenesis of EEVs. HUVECs were treated with a CV or FVIIa, and cell surface aSMase and PS levels were analyzed by immunostaining intact and nonpermeabilized cells with aSMase antibody and the binding of fluorescein isothiocyanate (FITC)-annexin V, respectively. No detectable levels of aSMase or PS were found on the surface of endothelial cells under basal conditions, whereas they were readily detectable in FVIIa-treated cells (Figure 1A–C). FVIIa treatment increased both the cell surface aSMase and PS levels in a time-dependent manner (Supplementary Figure S1A–C). aSMase expression on the cell surface was detectable at 1 hour of FVIIa treatment and reached the maximum at 2 hours. PS externalization was noticeable at 2 hours after adding FVIIa and reached the peak at 4 hours (Supplementary Figure S1A–C). Total cellular aSMase levels, measured by western blotting, showed that FVIIa treatment did not increase the cellular expression of aSMase (Figure 1D, E). Next, we evaluated whether FVIIa-induced externalization of PS to the cell surface is dependent on aSMase by silencing aSMase gene expression by siRNA (Figure 1F–J). aSMase silencing markedly reduced the FVIIa-induced PS externalization on the endothelial cell surface. Consistent with the crucial role of aSMase in FVIIa-induced PS externalization, we found no externalization of PS in FVIIa-treated murine endothelial cells isolated from aSMase^−/− mice (Figure 1K–M). In additional studies, functional inhibition of aSMase by desipramine (20 μM) was also shown to attenuate FVIIa-induced PS externalization (Figure 1N–P).

FIGURE 1.

FVIIa promotes aSMase translocation and aSMase-dependent PS externalization in endothelial cells. (A–C) FVIIa induces the translocation of aSMase to the cell surface and PS externalization. HUVECs, grown on glass coverslips, were treated with a CV or FVIIa (25 nM) for 4 hours. At the end of treatment, cells were washed and incubated with FITC-conjugated annexin V or rabbit anti-aSMase antibody for 1 hour. After washing, cells were fixed with 4% p-formaldehyde and the fixed cells were incubated with AF546-conjugated secondary antibodies for 1 hour (to immunostain aSMase) and 4’,6-diamidino-2-phenylindole (DAPI) for 30 minutes (to stain nucleus). The immunostained cells were subjected to confocal microscopy (A) and the fluorescence intensity of aSMase immunostaining (B) and FITC-annexin V binding (C) on the cell surface was quantified. (D–E) FVIIa treatment does not increase the total cellular aSMase levels. HUVECs, grown on a 12-well culture plate, were treated with a CV or FVIIa for 4 hours. Cells were lysed and subjected to immunoblotting to probe for aSMase (D), and the band intensities in the immunoblot were quantified by densitometric analysis (E). (F–J) aSMase silencing attenuates FVIIa-induced PS externalization. HUVECs were transfected with scrambled RNA or aSMase-specific siRNA. After 48 hours, cells were lysed, and aSMase expression was analyzed by western blotting using a human aSMase-specific antibody (F), and band intensities were quantified by densitometric analysis (G). Scr RNA or aSMase siRNA-transfected cells were treated with a CV or FVIIa for 4 hours and then immunostained for aSMase or stained with FITC-annexin V to detect PS as described in panel A. The stained cells were subjected to confocal microscopy (H) and the fluorescence intensities of aSMase (I) and annexin V (J) were quantified. (K–M) FVIIa does not induce PS externalization in aSMase-deficient endothelial cells. Brain endothelial cells were isolated from C57BL6/J wild-type or aSMase knock-out (aSMase^−/−) mice and grown on glass coverslips. Murine endothelial cells were treated with a CV or FVIIa for 4 hours and stained for aSMase and PS. The stained cells were subjected to confocal microscopy (K) and the fluorescence intensities of aSMase (L) and PS (M) staining were quantified. (N–P) pharmacologic inhibition of aSMase blocks FVIIa-induced PS externalization. HUVECs were treated with a functional inhibitor of aSMase, desipramine (Desip.; 20 μM), for 18 hours. The supernatant medium was replaced with fresh serum-free medium, and then the cells were treated with a CV or FVIIa. Cell surface aSMase and PS levels were analyzed (N) and quantified as described above (O–P). ***, P < .001; ****, P < .0001; ns, no statistically significant difference. aSMase, acid sphingomyelinase; CV, control vehicle; FVIIa, activated factor VII; FITC, fluorescein isothiocyanate; HUVECs, human umblical vein endothelial cells; PS, phosphatidylserine.

3.2 |. aSMase deficiency impairs the FVIIa-induced release of EEVs and their PS-dependent coagulant activity

Time course studies showed that FVIIa releases EEVs in a time-dependent manner and EEVs generation was maximal at 4 hours (Supplementary Figure S2A). EEVs released from endothelial cells treated with FVIIa for 4 hours showed maximum effect in supporting FX activation (Supplementary Figure S2B). aSMase silencing significantly reduced the FVIIa-induced release of EEVs (Figure 2A). Furthermore, aSMase silencing also markedly decreased FVIIa-EEVs’ ability to support FX or prothrombin activation (Figure 2B and 2C, please note that equal numbers of EEVs were used in the assay). Similar results were obtained with murine brain endothelial cells isolated from aSMase^−/− mice. FVIIa release of EEVs was significantly lower in endothelial cells isolated from aSMase^−/− mice than that in endothelial cells isolated from WT mice (Figure 2D). FVIIa-released EEVs from aSMase^−/− cells exhibited significantly lower PS-dependent coagulant activity (Figure 2E–F). Functional inhibition of aSMase by desipramine also significantly attenuated FVIIa release of EEVs (Figure 2G) and impaired their coagulant activity (Figure 2H). Similar results, ie, aSMase-dependent release of PS-enriched EVs, were obtained using endothelial cells derived from a different vascular bed (human aortic endothelial cells, data not shown), indicating that the above phenomenon is likely to be universal for all endothelial cell types.

FIGURE 2.

aSMase silencing impairs the release of FVIIa-induced EVs from endothelial cells and down-regulates FVIIa-EEVs’ hemostatic activity. (A) aSMase silencing inhibits the release of FVIIa-induced EVs. HUVECs were transfected with scrambled RNA (Scr RNA) or aSMase siRNA. The transfected cells were treated with a CV or FVIIa. After 4 hours of the treatment, EVs were isolated from the cell supernatant and quantified by nanoparticle tracking analysis. (B, C) aSMase silencing blocks the enhanced hemostatic potential of FVIIa-released EEVs. EVs isolated from HUVECs transfected with Scr RNA or aSMase siRNA and then treated with a control vehicle or FVIIa for 4 hours were quantified by nanoparticle tracking analysis. Equal numbers of EVs were suspended in a calcium-containing buffer. EVs were then incubated with a buffer (Control), annexin V (Anx V, 400 nM), TF neutralizing antibody (10 μg/mL), or control immunoglobulin G (IgG, 10 μg/mL) for 1 hour. Then, FVIIa (10 nM) and FX (175 nM) were added to the EVs’ suspension and the rate of FX activation was measured in a chromogenic assay (B). (C) EVs were isolated and treated as described in panel B. The ability of EVs to support prothrombin activation was measured by adding FVa (10 nM) and FXa (1.0 nM), followed by the substrate prothrombin (1.4 μM). Thrombin generated in the reaction mixture was measured in a chromogenic assay. (D–F) aSMase deficiency attenuates the FVIIa release of EVs from murine endothelial cells and impairs the hemostatic effect of FVIIa-released EEVs. Brain endothelial cells were isolated from wild-type and aSMase^−/− mice and cultured ex vivo. Confluent monolayers of endothelial cells were treated with a CV or FVIIa for 4 hours. EVs were isolated and quantified by nanoparticle tracking analysis. Equal number of EVs isolated from CV- and FVIIa-treated cells were evaluated for their ability to activate FX (E) or prothrombin (F) in the presence of and absence of annexin V. (G, H) Functional inhibition of aSMase reduces FVIIa release of EEVs and the hemostatic potential of FVIIa-released EEVs. HUVECs were treated with desipramine (Desip.) as described in Figure 1N, and then treated with a CV or FVIIa for 4 hours. EVs were isolated and quantified (G) and equal numbers of EEVs were subjected to analysis for their ability to activate FX in presence and absence of annexin V (H). *P < .05; ****P < .0001; ns, no statistically significant difference. aSMase, acid sphingomyelinase; CV, control vehicle; EVs, extracellular vesicles; EEV, endothelial extracellular vesicles.

As reported earlier [23], we found no evidence for apoptosis in FVIIa-treated cells, either in the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Supplementary Figure S3A) or in the analysis of apoptotic markers, Bcl-2, Bax, caspase3, and the cleavage of poly (ADP-ribose) polymerase-1 (PARP1) (Supplementary Figure S3B–E). Therefore, FVIIa-induced PS externalization and EV release in endothelial cells were independent of apoptosis.

3.3 |. FVIIa-mediated translocation of aSMase to the cell membrane and subsequent PS externalization is dependent on the EPCR-PAR1 signaling axis

To determine whether FVIIa-induced translocation of aSMase to the cell surface and PS externalization is dependent on FVIIa signaling through EPCR-PAR1 axis, we knocked down EPCR (Figure 3A, B) or PAR1 (Figure 3C, D) in endothelial cells before treating with FVIIa. Either EPCR or PAR1 silencing completely attenuated FVIIa-induced aSMase translocation and PS externalization (Figure 3E–G). In additional studies, we isolated brain endothelial cells from WT, EPCR KO, EPCR OX, and PAR1 mutant (R41Q or R46Q) mice and challenged them with FVIIa. FVIIa promoted aSMase translocation and, subsequently, PS externalization in endothelial cells derived from WT, EPCR OX, and PAR1 R46Q mice but not EPCR KO and PAR R41Q mice (Figure 3H–J). These data confirm that FVIIa-mediated aSMase translocation and PS externalization depend on FVIIa signaling via the EPCR-PAR1 axis.

FIGURE 3.

FVIIa-induced aSMase translocation to the cell surface and the subsequent PS externalization to the outer leaflet depend on the EPCR-PAR1 signaling axis. (A–D) EPCR and PAR1 silencing. HUVECs were transfected with scrambled (Scr) RNA, EPCR siRNA, or PAR1 siRNA. After 48 hours, cells were lysed and EPCR or PAR1 expression was analyzed by western blotting using a human EPCR- or PAR1-specific antibody, respectively. Band intensities were quantified by densitometric analysis. (E–J) FVIIa-induced aSMase translocation and PS externalization are dependent on EPCR and PAR1. EPCR or PAR1 silenced HUVECs were treated with a CV or FVIIa for 4 hours and cell surface aSMase and phosphatidylserine (annexin V binding) levels were analyzed and quantified (E–G) as described in Figure 1H. (H–J) Brain endothelial cells isolated from wild-type, EPCR-deficient (EPCR KO), EPCR overexpressing (EPCR OX), PAR1 R41Q mutant (PAR1 R41), or PAR1 R46Q mutant (PAR1 R46) mice were treated with a CV or FVIIa. After 4 hours, cell surface aSMase and PS levels were analyzed and quantified. ***P < .001; ****P < .0001; ns, no statistically significant difference. aSMase, acid sphingomyelinase; EPCR, endothelial cell protein C receptor; CV, control vehicle; FCIIa, activated factor VII; HUVECs, human umblical vein endothelial cells; PAR1, protease-activated receptor-1; PS, phosphatidylserine.

3.4 |. FVIIa-induced aSMase translocation and concomitant PS externalization are dependent on intracellular activation of the β-arrestin1-PI3K-AKT pathway

Earlier studies indicated that FVIIa-EPCR-PAR1 signaling is transmitted through β-arrestin1-dependent activation of the PI3K-AKT pathway [22,31]. Therefore, we examined whether FVIIa-induced aSMase translocation and subsequent PS externalization depend on the activation of the β-arrestin1-PI3K-AKT pathway. The knockdown of β-arrestin1 (Figure 4A, B) but not β-arrestin2 (Figure 4C, D) suppressed FVIIa-induced aSMase translocation and PS externalization (Figure 4E–G). Additional studies showed that the PI3K-specific inhibitor, LY294002 (LY; 25 μM), completely abrogated FVIIa-induced aSMase translocation and PS externalization, whereas its inactive analog, LY303511 (NC LY; 25 μM) had no effect (Figure 4H–J). Similarly, inhibition of AKT by a specific inhibitor, AKT Inhibitor VIII (10 μM), completely attenuated FVIIa-induced aSMase translocation and externalization of PS (Figure 4K–M). Analysis of AKT activation showed that both PI3K and AKT-specific inhibitors markedly reduced FVIIa-induced AKT activation in endothelial cells (Figure 4N, O). Consistent with our earlier findings [23], inhibition of β-arrestin1 or the PI3K-AKT pathway also attenuated the FVIIa-mediated release of EVs from endothelial cells (Supplementary Figure S4). Overall, the above data indicate that FVIIa-induced aSMase translocation and the concomitant PS externalization in endothelial cells are transmitted through intracellular activation of the β-arrestin1-PI3K-AKT pathway (Figure 4P).

FIGURE 4.

FVIIa-induced aSMase translocation and PS externalization are dependent on intracellular activation of the β-arrestin1-PI3K-AKT signaling pathway. (A–D) Silencing of β-arrestins. HUVECs were transfected with scrambled (Scr) RNA, β-arrestin 1 siRNA, or β-arrestin 2 siRNA. After 48 hours, cells were lysed to analyze the expression of β-arrestin 1 (A) or β-arrestin 2 (C) by western blotting, and band intensities were quantified by densitometry (B, D). (E–G) FVIIa-induced aSMase translocation and PS externalization were dependent on β-arrestin 1. HUVECs transfected with Scr RNA, β-arrestin 1 siRNA, or β-arrestin 2 siRNA were challenged with a CV or FVIIa for 4 hours. aSMase and PS levels at the cell surface were analyzed by confocal microscopy following immunostaining of aSMase and FITC-annexin V binding (E) and the fluorescence intensities of the signals were quantified (F, G). (H–J) PI3K inhibition attenuates FVIIa-induced aSMase translocation and PS externalization. HUVECs were pretreated with LY294002 (LY, 25 μM) or negative control of LY294002 (LY303511/NC LY, 25 μM). After 1 hour, cells were challenged with a CV or FVIIa for 4 hours and the translocation of aSMase or PS externalization was analyzed, and the signals were quantified. (K–M) Inhibition of AKT blocks FVIIa-induced aSMase translocation and PS externalization. HUVECs were treated with a CV (Control) or AKT inhibitor (AKT Inhibitor VIII; 10 μM) for 1 hour. Then, cells were treated with a CV or FVIIa for 4 hours and aSMase translocation and PS externalization were analyzed. (N, O) FVIIa activation of AKT and its inhibition by LY compound and AKT inhibitor. HUVECs, grown in a 12-well plate, were pretreated with PI3K inhibitor (LY), a nonactive compound of LY (NC LY), or AKT inhibitor (AKT Inh. VIII) for 1 hour. Then the cells were treated with a CV or FVIIa for 1 hour and the activation of AKT was assessed by immunoblotting using antibodies specific to phosphorylated AKT (N) and the band intensities were quantified by densitometry (O) to determine the extent of AKT activation. (P) Schematic representation of FVIIa-EPCR-PAR1 signaling leading to PS externalization. ***P < .001; ****P < .0001; ns, no statistically significant difference. AKT, RAC (Rho family)-alpha serine/threonine-protein kinase; aSMase, acid sphingomyelinase; CV, control vehicle; FVIIa, activated factor VII; FVIIa-EPCR-PAR1, activated factor VII-endothelial cell protein C receptor-protease-activated receptor-1; HUVECs, human umblical vein endothelial cells; PS, phosphatidylserine.

3.5 |. FVIIa-induced aSMase translocation on endothelial cell surface leads to externalization of PS via the induction of phospholipid scramblase, TMEM16F

Phospholipid scramblases play a pivotal role in the externalization of negatively charged phospholipids, PS, from the inner to the outer leaflet of the plasma membrane [41]. Endothelial cells are known to express TMEM protein family members [42]. Among them, TMEM16D and TMEM16F are shown to be lipid scramblases and contribute to PS externalization in immune cell types [43]. To investigate whether the expression of TMEM16D or TMEM16F is increased in response to FVIIa treatment, HUVECs were treated with FVIIa for varying times, and the expressions of TMEM16D and TMEM16F were analyzed by western blotting. FVIIa treatment increased the expression of TMEM16F in a time-dependent manner, with a significant increase at 2 hours and reaching a maximum at 4 hours (Figure 5A, B). FVIIa treatment did not alter the expression of TMEM16D (Figure 5C, D). aSMase silencing impaired FVIIa-induced TMEM16F expression (Figure 5E, F). Next, to investigate the contribution of TMEM16D or -F in the externalization of PS, we knocked down the expression of TMEM16D or -F with specific siRNAs in HUVECs (Figure 5G–J) and analyzed FVIIa-driven PS externalization (Figure 5K–P). The data show that the knockdown of TMEM16F, but not TMEM16D, completely abolished FVIIa-induced externalization of PS without altering aSMase translocation to the cell surface. The activity of TMEM16F is dependent on Ca²⁺ [44]. To strengthen the observation that increased TMEM16F activity as the result of increased TMEM16F expression might be responsible for FVIIa-induced PS externalization, endothelial cells were treated with an intracellular Ca²⁺ chelator, BAPTA-AM, before treatment with FVIIa. Ca²⁺ chelation did not interfere with FVIIa-induced aSMase translocation but prevented the externalization of PS (Supplementary Figure S5A–C). Overall, these data suggest that FVIIa-mediated, aSMase activation-dependent, increased expression of TMEM16F is responsible for the externalization of PS. In additional studies, we investigated the role of TMEM16D or -F in FVIIa-induced EEVs generation and EEVs’ hemostatic activity. Knockdown of either TMEM did not interfere with FVIIa-induced EEVs generation (Supplementary Figure S6A, B). As expected, TMEM16F silencing, but not TMEM16D, impeded FVIIa-EEVs’ ability to support FX and prothrombin activation (Supplementary Figure S6C–F).

FIGURE 5.

FVIIa-mediated aSMase translocation to the plasma membrane induces PS externalization via the induction of TMEM16F. (A–D). FVIIa induces the expression of TMEM16F and not TMEM16D. HUVECs were treated with FVIIa for varying times and the expressions of TMEM16F (A, B) and TMEM16D (C, D) were analyzed by immunoblot analysis followed by densitometry. (E, F) FVIIa-induced TMEM16F expression was dependent on aSMase. HUVECs were transfected with scrambled (Scr) RNA or aSMase siRNA and the transfected cells were treated with a CV or FVIIa for 4 hours. The expression of TMEM16F was analyzed by immunoblotting (E) and the band intensities were quantified by densitometric analysis (F). (G–J) Silencing of TMEM16F, and TMEM16D. HUVECs were transfected with scrambled (Scr) RNA, TMEM16F siRNA, or TMEM16D siRNA. Forty-eight hours posttransfection, cells were lysed and the expressions of TMEM16F and TMEM16D were analyzed by immunoblotting (G, I) and the immunostained band intensity signals were quantified by densitometry (H, J). (K–P) Silencing of TMEM16F and not TMEM16D blocks FVIIa-induced PS externalization. TMEM16F (K–M) or TMEM16D (N–P) silenced cells were treated with a CV or FVIIa for 4 hours. aSMase translocation to the cell surface and PS externalization was analyzed by immunostaining of aSMase and FITC-annexin V binding to the cells followed by confocal microscopy (K, N) and the fluorescence intensity signals were quantified (L, M, O, and P). *P < .05; ***P < .001; ****P < .0001; ns, not statistically significant. aSMase, acid sphingomyelinase; CV, control vehicle; FVIIa, activated factor VII; HUVECs, human umblical vein endothelial cells; PS, phosphatidylserine; TMEM, transmembrane protein.

3.6 |. FVIIa-driven aSMase translocation to endothelial cell surface leads to the generation of ceramide and activation of S1PR pathway to induce TMEM16F expression and PS externalization

aSMase translocation to the plasma membrane leads to sphingomyelin (SM) breakdown, resulting in a significant drop in SM levels in the plasma membrane [45]. The breakdown of SM generates ceramide, which in turn is converted to sphingosine by the action of ceramidases [46]. Sphingosine is readily converted to S1P by sphingosine kinases (SphKs) [46] and S1P is well-known for activating S1PRs to induce various signaling cascades [47]. To determine whether FVIIa-induced aSMase activation leads to the activation of the S1PR pathway to induce TMEM16F expression and PS externalization, we first analyzed the production of ceramide in cells treated with FVIIa. FVIIa treatment induced ceramide production in endothelial cells (Supplementary Figure S7A, B). Knockdown of aSMase fully blocked FVIIa-induced ceramide generation (Figure 6A, B). Inhibition of ceramidases or the downstream SphKs with specific inhibitors, B13 (200 μM) and SK1-II (20 μM) respectively, completely abrogated FVIIa-induced PS externalization without affecting aSMase levels on the cell surface (Figure 6C–E). Furthermore, blocking S1P receptors by adding an S1P receptor modulator, FTY720 (10 μM), fully attenuated FVIIa-induced PS externalization (Figure 6F–H). Consistent with the above data, silencing of S1PR1 (Supplementary Figure S8A, B) significantly down-regulated FVIIa-induced externalization of PS in endothelial cells (Supplementary Figure S8C–E). Additional studies showed that inhibition of ceramidases, SphKs, or S1P receptor activation significantly down-regulated FVIIa-mediated expression of TMEM16F (Figure 6I–L). Similarly, knockdown of S1PR1 also attenuated FVIIa-induced expression of TMEM16F in endothelial cells, indicating the role of S1PR1 signaling in the FVIIa-induced PS externalization (Supplementary Figure S9A, B). In contrast to the above, inhibition of ceramidases, SphKs, or S1P receptor activation did not affect the FVIIa release of EEVs (Supplementary Figure S11A, B). However, they impaired PS-dependent FVIIa-EEVs’ coagulant activity (Supplementary Figure S11C, D).

FIGURE 6.

FVIIa-driven aSMase translocation leads to the generation of ceramide and activation of S1PR that induce TMEM16F expression and PS externalization. (A,B) FVIIa induces ceramide generation in endothelial cells in an aSMase-dependent manner. HUVECs, transfected with scrambled (Scr) RNA or aSMase siRNA were treated with a CV or FVIIa for 2 hours. Ceramide generation was analyzed by immunostaining with ceramide-specific antibody (A) and the fluorescence intensities of immunostaining signals were quantified (B). (C–E) Inhibition of ceramidases or sphingosine kinases blocks FVIIa-induced PS externalization without impairing aSMase translocation. HUVECs were treated with ceramidase inhibitor, B13 (200 μM), or sphingosine kinase inhibitor, SK1-II (20 μM), for 24 hours. Cells were then challenged with a CV or FVIIa for 4 hours, and surface aSMase and PS levels were analyzed by immunostaining aSMase and the binding of FITC-annexin V, respectively (C) and the fluorescence intensity of signals were quantified (D, E). (F–H) Inhibition of the S1P receptor blocks FVIIa-induced PS externalization. HUVECs were treated with an S1P receptor modulator, FTY720 (10 μM) for 1 hour, and then with a CV or FVIIa for 4 hours. Cell surface aSMase and PS levels were analyzed (F) and quantified (G, H). (I, J) Inhibition of ceramidases or sphingosine kinase blocks FVIIa-induced TMEM16F expression. HUVECs were pretreated with an inhibitor of ceramidase (B13) or sphingosine kinase (SK1-II) as described in panel C and treated with a CV or FVIIa for 4 hours. The expression of TMEM16F was analyzed by western blotting (I) and the immunostained band signals were quantified by densitometric analysis (J). (K–L) Inhibition of S1P receptor activation attenuates FVIIa-induced TMEM16F expression. HUVECs were pretreated S1P receptor inhibitor, FTY720 for 1 hour and then treated with a CV or FVIIa for 4 hours. The expression of TMEM16F was analyzed by western blotting (K) and the immunostained band signals were quantified by densitometry (L). (M) Schematic representation of a potential pathway by which FVIIa-EPCR-PAR1 signaling leads to PS externalization. ***P < .001; ****P < .0001; ns, not statistically significant. aSMase, acid sphingomyelinase; CV, control vehicle; FVIIa, activated factor VII; FVIIa-EPCR-PAR1, activated factor VII-endothelial cell protein C receptor-protease-activated receptor-1; HUVECs, human umblical vein endothelial cells; PS, phosphatidylserine; TMEM, transmembrane protein.

To support the conclusion that FVIIa-induced S1PR activation via aSMase-mediated S1P generation is responsible for PS externalization, endothelial cells were treated exogenously with S1P for varying times, and TMEM16F expression was analyzed. S1P treatment induced TMEM16F expression (Supplementary Figure S10A, B) in endothelial cells. Knockdown TMEM16F abolished the S1P-induced PS externalization (Supplementary Figure S10C, D). Furthermore, inhibition of S1P receptor activation by the receptor modulator, FTY720 impaired S1P-mediated TMEM16F expression (Supplementary Figure S10E, F) as well as PS externalization (Supplementary Figure S10G, I). S1P treatment, on the other hand, did not affect aSMase translocation to the cell surface (Supplementary Figure S10G, H).

Overall, the above data suggest that FVIIa-induced aSMase translocation to the plasma membrane leads to the generation of S1Ps and activation of S1PR autocrine signaling to induce TMEM16F expression, thereby promoting the externalization of PS. FVIIa-induced aSMase activation also leads to biogenesis of EVs independent of the downstream S1PR activation. Both processes together leads to generation of PS-enriched EVs in endothelial cells exposed to FVIIa (Figure 6M).

3.7 |. aSMase deficiency impairs FVIIa-induced EVs generation in vivo and reduces FVIIa’s hemostatic potential in thrombocytopenia

To investigate the role of aSMase in the FVIIa-induced release of EVs in vivo, WT and aSMase^−/− mice were administered with saline or FVIIa (0.25 mg/kg body weight). Measurement of EVs in circulating blood at 2 h following FVIIa administration, at which time FVIIa-induced EVs reach peak levels in the blood [23], showed a marked decrease in FVIIa-released EVs in aSMase^−/− mice (Figure 7A). Analysis of EVs by western blot analysis showed a marked increase in the endothelial cell-specific marker, VE-cadherin, in the EVs isolated from FVIIa-injected WT mice. VE-cadherin levels were significantly lower in the EVs isolated from aSMase^−/− mice treated with FVIIa. No significant differences were found in monocyte-, platelet- and RBC-specific markers between EVs isolated from a CV and FVIIa-treated mice (Figure 7B–F). In additional studies, we analyzed plasma EVs by flow cytometry for VE-cadherin and a second endothelial cell-specific marker, CD31. The data showed that approximately 7% of EVs isolated from the plasma of control mice were positive for VE-cadherin, whereas 22% of EVs isolated from the plasma of FVIIa-treated mice were positive for VE-cadherin. We obtained similar results with CD31 staining (see Supplementary Figure S12). In contrast to the above, we found no significant increase in VE-cadherin and CD31 positive EVs in aSMase^−/− mice treated with FVIIa over a CV (Supplementary Figure S12). Similar to that observed in vitro system, FVIIa-released EVs from WT mice, not aSMase^−/− mice, support PS-dependent and TF-independent FX and prothrombin activation (Figure 7G, H).

FIGURE 7.

aSMase deficiency impairs FVIIa-induced EVs generation from the endothelium in vivo and impedes EEVs’ hemostatic potential. (A) aSMase deficiency blunts FVIIa release of EVs into circulation. WT or aSMase^−/− mice were administered saline or FVIIa (0.25 mg/kg body weight) via the tail vein. After 2 hours, blood was collected, EVs were isolated from the plasma and quantified by nanoparticle tracking analysis. (B–F) Decrease in endothelium-originated EVs in aSMase^−/− mice. EVs, isolated from the blood of WT or aSMase^−/− mice treated with saline or FVIIa (0.25 mg/kg) for 2 hours were subjected to immunoblot analysis to probe for VE-cadherin (endothelial cell marker), CD14 (monocyte marker), CD41 (platelet marker), or TER119 (RBC marker) (B). (C–F) Band intensities of immunoblots were quantified by densitometric analysis. (G–H) Phosphatidylserine-dependent procoagulant activity of FVIIa-released EVs was lower in EVs from aSMase−/− mice. EVs, isolated from WT or aSMase^−/− mice treated with saline or FVIIa as described in panel A, were assayed for their ability to support the activation of FX (G) or prothrombin (H). (I–K) FVIIa-released EVs from control- and not aSMase-silenced endothelial cells correct bleeding-associated with thrombocytopenia. Wild-type mice were injected with CD42b antibodies (1 mg/kg) via the tail vein to deplete platelets. Five hours following the administration of platelet-depleting antibodies, mice were injected with equal numbers of EVs (1 × 10⁹/mouse) isolated from scrambled (Scr) or aSMase siRNA-transfected endothelial cells treated with a CV, FVIIa, or saline. Immediately following the administration of EVs, mice were subjected to the saphenous vein incision, and the blood coming from the injury site was collected on filter paper. Hemoglobin was extracted from the filter paper, and hemoglobin levels were extrapolated to the blood volume using a standard curve derived from known volumes of blood (I). Blood coming from the wound site was collected periodically (up to 15 minutes; if a clot was formed, it was dislodged) directly into citrate anticoagulant, and thrombin generation in the blood was measured as the amount of TAT complexes formed (J). TAT generation was also measured in the blood, collected from the submandibular vein (K). (L–O) aSMase-dependent release of endogenous EVs by FVIIa confers hemostasis in thrombocytopenic mice in the saphenous vein injury model. WT or aSMase^−/− mice were injected with CD42b antibodies (1 mg/kg) via the tail vein. Platelet counts in the blood were measured before the administration of the antibody and 5 hours following the antibody administration (L). Platelet-depleted WT or aSMase^−/− mice were injected with saline or FVIIa (0.25 mg/kg) via the tail vein. Two hours following FVIIa administration, mice were subjected to saphenous vein incision, and blood loss at the site of injury (M) and TAT generation at the injury site (N) and in the systemic circulation (O) were measured. *P < .05; **P < .01; ***P < .001; ****P < .0001; ns, no statistically significant difference. aSMase, acid sphingomyelinase; CV, control vehicle; FVIIa, activated FVIIa; EV, extracellular vesicle; EEV, endothelial extracellular vesicles; TAT, thrombin-antithrombin; WT, wild-type.

Next, to determine how aSMase affects the hemostatic potential of FVIIa-released EVs in vivo, an equal number of EVs (1 X10⁹ EVs per mouse) isolated from CV- or FVIIa-treated endothelial cells that were transfected with scrambled (Scr) RNA or aSMase siRNA were injected into platelet-depleted WT mice. Bleeding was induced by a saphenous vein incision. As shown in Figure 7I, administration of FVIIa-EEVs from Scr RNA transfected cells significantly reduced the blood loss compared with saline control, but FVIIa-EEVs from aSMase-silenced cells failed to prevent bleeding. The measurement of TAT levels in the blood collected at the injury site showed a significant increase in TAT levels in platelet-depleted mice injected with FVIIa-EEVs from Scr RNA transfected endothelial cells, but not from aSMase-silenced cells (Figure 7J). Control EEVs did not increase the TAT levels at the site of injury. FVIIa-released EEVs, either isolated from control or aSMase-deficient cells, did not increase systemic plasma TAT levels (Figure 7K).

To confirm the crucial role of aSMase in generating hemostatic effective EVs endogenously in the circulation in response to FVIIa treatment, platelet-depleted WT and aSMase^−/− mice were administered with saline or FVIIa (0.25 mg/kg body weight). After 2 h of FVIIa administration, by which time the levels of FVIIa-released EVs in the circulation would reach the peak and exogenously administered FVIIa would be cleared from the circulation [23], mice were subjected to the saphenous vein incision and the blood loss and TAT generation at the wound site were measured. The administration of platelet-depleting antibodies reduced the blood platelet level to >90% (Figure 7L). FVIIa administration to the platelet-depleted WT mice reduced blood loss and increased TAT generation at the wound site (Figure 7M, N). In contrast, FVIIa failed to correct bleeding in platelet-depleted aSMase^−/− mice. No significant systemic TAT generation was observed in either group of mice (Figure 7O). These observations indicate that aSMase plays a crucial role not only in FVIIa-mediated EVs generation but also in imparting hemostatic potential to FVIIa-released EVs.

4 |. DISCUSSION

Our recent studies revealed that FVIIa binding to EPCR on the endothelium leads to the generation of EVs via the activation of PAR1 both in vitro and in vivo [23]. FVIIa-released EEVs showed higher procoagulant activity, due to the presence of negatively charged phospholipid, PS, on the outer surface, than that shown by EEVs generated under basal conditions [23,34]. These studies also suggested that selective enrichment of FVIIa-released EEVs with PS might be due to the externalization of PS in endothelial cells in response to FVIIa [23]. However, the underlying mechanism involved in the FVIIa-induced PS externalization and its role in the biogenesis of PS-enriched EVs were unknown. In the present study, we provide compelling evidence that FVIIa activation of aSMase via the EPCR-PAR1 axis, and the subsequent generation of ceramide and activation of the S1PR-mediated signaling pathway induces the expression of TMEM16F that acts as a phospholipid scramblase to induce PS externalization. FVIIa-induced activation of aSMase also plays a crucial role in the FVIIa release of EVs from endothelial cells. The release of EVs from endothelial cells where PS was externalized in response to FVIIa is responsible for the release of PS-enriched EVs by FVIIa.

Earlier studies showed that exposure to bacterial SMases promotes PS externalization in erythrocytes, platelets, and endothelial cells [37,48,49]. Munzer et al. [49] showed that activation-dependent PS exposure was significantly blunted in platelets from aSMase^−/− mice, indicating that aSMase plays a role in PS externalization in platelets. Although diverse stress stimuli were shown to activate aSMase, and the activation of aSMase plays a key role during cellular differentiation, growth arrest, and apoptosis [40,50,51], there was no strong evidence for aSMase to regulate the externalization of PS. Under basal conditions, aSMase was primarily localized within acidic compartments in the cytoplasm, but upon activation with diverse stimuli, aSMase translocates from endolysosomes to the plasma membrane [45]. It is possible that activation of aSMase and its subsequent translocation to the plasma membrane would lead to the breakdown of SM in the plasma membrane, which results in the generation of ceramide [45,52]. Ceramide is readily converted to sphingosine by cellular ceramidases [46], which in turn is phosphorylated by SphKs to produce S1P [46]. S1P, through activating S1P receptors [47], could induce various biological responses [53,54]. One such response could be PS externalization. Consistent with the above possibility, our current study shows that FVIIa activation of aSMase leads to the activation of the S1P-S1PR pathway, and the activation of the S1P-S1PR pathway induces PS externalization. Our present data that show FVIIa treatment produces ceramide generation in endothelial cells and the inhibition of ceramidases, SphKs, or S1PR receptor activation significantly attenuates the FVIIa-induced PS externalization, provide compelling evidence to the above hypothesis. Typically, agonist-induced activation of aSMase and its translocation to the plasma membrane is fairly a rapid process and requires only a few minutes [55,56]. Our observation that FVIIa-EPCR-PAR1 signaling-induced aSMase translocation requires a much longer time (~1 h) could be due to the slower assembly of the signaling complex and delayed intracellular signaling. FVIIa-induced aSMase activation was dependent on AKT activation. It may be pertinent to note here that our earlier studies showed that although FVIIa-induced AKT activation was noticeable at 15 minutes, it required 1 hour to reach the maximum [31].

Previous studies have indicated that phospholipid scramblases play a pivotal role in the externalization of PS [41]. Endothelial cells are known to express many members of TMEM16 family proteins [42]. Some of the TMEM16 family proteins, such as TMEM16D and TMEM16F, were shown to function as phospholipid scramblases and externalize PS in certain cell types [43]. Our studies provide convincing evidence that FVIIa induces the expression of TMEM16F in endothelial cells, and the increased expression of TMEM16F is responsible for FVIIa-induced PS externalization in endothelial cells as the knockdown of TMEM16F abrogates the FVIIa-induced PS externalization. The observation that the inhibition of aSMase, ceramidases, SphKs, or S1PR significantly attenuate FVIIa-induced expression of TMEM16F and PS externalization strongly indicates that FVIIa activation of aSMase, and the consequent generation of ceramide and S1P, and S1P activation of S1PR signaling pathway are responsible for FVIIa-induced PS externalization. The observation that inhibition of ceramidases and SphKs completely block FVIIa-induced PS externalization while they inhibit TMEM16F expression significantly, but not completely, leaves the possibility open that other mechanisms may also play a role in FVIIa-induced externalization of PS in endothelial cells. Although our present data on the potential transactivation of the S1P-S1PR signaling pathway by FVIIa and its consequences are limited to the expression of TMEM16F and PS externalization, our findings may have wider consequences, as the S1PR-mediated signaling modulates diverse endothelial activities, including cell survival, regulation of proinflammatory responses, and endothelial barrier stability [57,58].

Previous studies have shown APC-PAR1 transactivates S1PR1, and this could play a role in APC-mediated protection against barrier disruption [59,60]. However, the mechanism by which APC-PAR1 transactivates S1PR1 and the role of S1PR1 transactivation in other APC-mediated cytoprotective effects are unknown. In a recent study, Molinar-Inglis et al. [61] demonstrated that APC-PAR1-stimulated transactivation of S1PR1 is mediated by β-arrestin2-dependent SphK1 activation in caveolae. FVIIa-EPCR-PAR1 transactivation of S1PR reported in the present study appears unique and differs from APC-PAR1 transactivation of S1PR1. FVIIa-EPCR-PAR1 response is mediated through β-arrestin 1, not β-arrestin 2, -dependent signaling pathways. More importantly, our data show that FVIIa-EPCR-PAR1 activation of aSMase via β-arrestin 1-dependent signaling and the subsequent generation of ceramide and S1P plays a crucial role in the transactivation of S1PR. Here, it may be pertinent to point out that we found no evidence that APC, similar to FVIIa, activates aSMase and induces PS externalization in endothelial cells (unpublished data from the laboratory, Kaushik Das, 2022). S1PRs consist of a family of 5 receptors, and endothelial cells express at least 3 of them, S1PR1, S1PR2, and S1PR3, with S1PR1 being the primary receptor [62]. Our current data indicate that S1PR1 activation is responsible for FVIIa-induced externalization of PS in endothelial cells.

FVIIa activation of aSMase, in addition to playing a role in PS externalization, also plays a crucial role in the biogenesis of EVs from endothelial cells, as silencing or functional inhibition of aSMase significantly reduced FVIIa-release EEVs. However, FVIIa release of EEVs is independent of aSMase-mediated transactivation of S1PR. A limited number of studies suggest the involvement of sphingolipids and their enzymes in EVs biogenesis and their release [38]. Bianco et al. [39] showed that ATP-induced EVs biogenesis in glial cells is mediated by P2X₇-receptor activation-dependent translocation of aSMase from lysosomes to the outer leaflet of the plasma membrane. Consistent with these data, our studies in macrophages showed that ATP-induced translocation of aSMase to the plasma membrane plays a role in releasing TF-positive EVs [63]. Changing local lipid composition in the plasma membrane could alter the membrane fluidity and curvature. Breakdown of SM in the membrane could lead to increased efflux of cholesterol from the membrane and thus increase membrane fluidity [64,65]. The generation of ceramide, which has a cone-shaped structure, could induce curvature to the membrane, favoring membrane blebbing and evagination of EVs [66]. FVIIa translocation of aSMase to the outer leaflet and the consequent hydrolysis of SM and ceramide generation is likely to be responsible for the FVIIa release of EVs in endothelial cells. This, coupled with SM catabolism-derived metabolite S1P, and S1P-induced S1PR activation-dependent PS externalization on the endothelial cell surface, contribute to the enrichment of PS in FVIIa-released EVs.

Consistent with our in vitro data, our in vivo findings provide strong evidence for the crucial role of aSMase in the FVIIa release of EVs and enrichment of these EVs with PS. aSMase deficiency not only significantly blunted the release of EVs into the circulation following FVIIa administration but FVIIa-released EVs also exhibited diminished PS-dependent coagulant activity in ex vivo studies. Although analysis of EVs isolated from plasma indicates that FVIIa-released EVs are derived from endothelial cells and this process is dependent on aSMase, we cannot completely rule out at present the possibility that a fraction of FVIIa-released EVs into the circulation could have come from other cell types, including blood cells. The potential significance of aSMase-dependent FVIIa-released EVs in therapeutic settings is quite evident in the ability of FVIIa to reduce bleeding in severe thrombocytopenia in WT mice, and not in aSMase^−/− mice, even after the exogenously administered FVIIa is cleared from the circulation. FVIIa is successfully used in the treatment of bleeding disorders in severe thrombocytopenia [30,67–70]. Our present findings indicate that FVIIa-induced hemostatic response in thrombocytopenia probably comes from FVIIa-released EVs that could act as mini platelets by providing PS-rich membrane surface for coagulant reactions. Our current study shows that this process is dependent on aSMase. It is pertinent to note here, as with the activated platelets, the procoagulant effect of FVIIa-released EVs is limited to the wound site, as we found no evidence for the systemic activation of coagulation following the release of EVs.

In summary, our studies identify a novel mechanism by which FVIIa releases PS-enriched EVs from endothelial cells and their potential role in correcting bleeding associated with thrombocytopenia. Furthermore, our studies indicate that sphingolipid metabolism could influence the therapeutic effect of FVIIa by modulating the release and composition of FVIIa-released EVs. Finally, our studies discover a unique mechanism by which FVIIa-EPCR-PAR1 signaling transactivates S1PR.

Essentials.

• Activated factor (F) VII (FVIIa) treatment releases extracellular vesicles (EVs) from endothelial cells, and these EVs exhibit hemostatic and anti-inflammatory properties.

• FVIIa-endothelial cell protein C receptor signaling activates acid sphingomyelinase (aSMase) in endothelial cells.

• FVIIa release of EVs is dependent on aSMase.

• FVIIa-induced aSMase activation leads to phosphatidylserine externalization via ceramide generation, S1PR activation, and expression of phospholipid scramblase transmembrane protein 16F.