Human coronary artery endothelial cells (HCAEC) were obtained from PromoCell (Heidelberg, Germany) isolated from left and right main coronary arteries, circumflex and anterior coronary arteries and were CD31 and Dil-Ac-LDL uptake positive. The cells were cultured in EC-MV media containing 5% (v/v) FCS and growth supplements (PromoCell) and remained CD31 positive. Human coronary artery smooth muscle cells (HCASMC) were also isolated and provided by PromoCell as smooth muscle α-actin positive, CD31 negative cells. The cells were cultured in SMC-MV2 media containing 5% (v/v) FCS and growth supplements (PromoCell) and were tested for the expression of smoothelin-B which is a marker associated with contractile SMCs from muscular arteries (see discussion for details). Both primary cell types were guaranteed for >15 divisions and were used at around 9 divisions. ECV304 cell line was cultured in M199 medium containing 10% (v/v) FCS.

The sequence for full length filamin A was cloned into pcDNA 3 c-myc plasmid and stop codons were engineered into positions Gly2370, Gly2467 and Ala2595 in order to delete the repeat units 22–24, 23–24, or 24 alone. The successful mutations were verified by DNA sequencing (MWG-eurofins, Wolverhampton, UK). The pCMV6-Ac-TF-tGFP plasmid for the expression of full-length wild-type human TF (TF_Wt-tGFP) was obtained from OriGene (Rockville, USA) and mutations were prepared as previously described (46) to express proteins with Ser253Asp (TF_Asp253-tGFP), truncated at Ser241 (TF_ΔCT-tGFP), or tGFP alone.

Human coronary artery endothelial cells (HCAEC) were used throughout the investigation because these cells are the closest representatives to the aims of this investigation. In contrast to monocytes or tumor cells, these cells do not either constitutively express TF and/or release large amounts of microvesicles spontaneously. Furthermore, unlike tumor cell derived microvesicles, the resultant microvesicles do not contain other material that alter the behavior of vascular cells (e.g., pro-angiogenic mediators). HCAEC have also been shown to express PAR2 on the cell surface (47). HCAEC (5 × 10⁵) and also in some experiments ECV304 cells, were transfected with 0.5 µg of pCMV6-Ac-TF-tGFP plasmid variants using TransIT-2020 (Geneflow, Litchfield, UK) according to the manufacturer's instructions and permitted to express the proteins for 48 h prior to use. The expression of TF was previously confirmed by measuring TF mRNA and measuring the total and cell-surface antigen levels by western blot and flow cytometry (46, 47). All cell samples, including the non-transfected cells, were then transferred and adapted to serum-free medium for 1 h. The release of TF-containing microvesicles was induced by incubation with PAR2 activating peptide (PAR2-AP); SLIGKV (20 µM; Sigma Chemical Company Ltd, Poole, UK). Following 80 min incubation with PAR2-AP, the conditioned media were cleared of any cell debris by centrifuging for 10 min at 2,500 g on a microcentrifuge. The samples (1 ml aliquots) were then placed in 11 × 34 mm polycarbonate centrifuge tubes (Beckman Coulter, High Wycombe, UK), and the cell-derived microvesicles were sedimented at 100,000 g on a TL-100 ultracentrifuge at 20°C, using a TLA 100.2 rotor (Beckman) for 1 h, as described before (46, 48). Sedimented microvesicles were washed with PBS. The pellet was resuspended in pre-filtered (0.1 μm) PBS (200 µl), divided into batches, and frozen at −80°C or used immediately. The specificity of the agonist in the incorporation and release of TF was previously demonstrated and compared to the ability of factor Xa to promote the release of TF-containing microvesicles (46). Additionally we showed that freezing of the microvesicles did not alter the microvesicle size or number as determined by nanoparticle tracking analysis (48), the TF antigen levels (46, 47) or the TF associated thrombin generation and factor Xa-generation. Furthermore, pre-incubation of the microvesicles with antibodies against TF, or immune-purification of the microvesicles using TF antibodies did not alter the size distribution of the microvesicles (Supplementary Figure S1), or the total TF content (48).

Previously, the microvesicles released following PAR2-AP activation were compared to those released following activation with TNFα (10 ng/ml) or IL-1β (10 ng/ml) (46) and microvesicles released from THP-1 cells, isolated peripheral blood mononuclear cells (PBMC), or HCASMC, in parallel (46, 47). The properties of TF-containing microvesicles from these various sources prepared by different procedures were established previously; in particular the numbers and size distributions of the isolated microvesicles were determined by nanoparticle tracking analysis (NTA) using Nano-Sight LM10 and NTA software (NanoSight Ltd, Amesbury, UK). The Nanosight instrument was calibrated using FluoSpheres® carboxylate-modified microspheres with diameters of 0.1 μm and 1.0 μm (Invitrogen). Microvesicle samples were diluted 1:5 in 0.1 μm-filtered PBS and analyzed using the NanoSight LM10 by tracking particles over 60 s using a camera level of 12 and shutter speed of 21.26 frames/s. A typical NTA trace for microvesicles is included in Supplementary Figure S1. Additionally, size and density of the TF-containing microvesicles were established previously by density gradient centrifugation (48) and the phospholipid content assessed by thin layer chromatography (46). Importantly, the density of the released microvesicles was measured by the concentration of phosphatidylserine, which was determined using the Zymuphen MP-assay kit (Hyphen BioMed/ Quadratech, Epsom, UK) against the standards provided with the kit. This function is the most relevant to this study. The TF content of the microvesicles was also measured by flow cytometry (46–48), TF-specific ELISA (Quantikine, R&D Systems, Abingdon, UK) and western blot (49). Although on the whole the resultant microvesicles were comparable in TF content and microvesicle identity, repeated studies in our laboratory has indicated the release of more consistent TF-containing microvesicles following activation by PAR2-AP and therefore this procedure has been adopted by our laboratory as the most consistent and reproducible procedure for preparing TF-MV (45–49). Previous examination of the microvesicles for exosome markers Tsg101 and CD9 by western blot showed that exosomes were not present, which was likely due to the short incubation times (26, 46, 47). Similarly, exosome marker antibodies (anti-Tsg101 or CD9 antibodies) were not capable pulling down TF-microvesicles (48).

HCASMC were transfected with the pcDNA 3 c-myc constructs using TransIT 2020 reagent to express the wild type or mutant forms of filamin A, lacking repeats domains (22–24, 23–24 and 24 alone). The cells were permitted to express the proteins for 48 h and then employed in migration assays as below. The expression of the proteins was confirmed using a biotin-conjugated mouse anti-c-myc antibody (Miltenyi Biotech Ltd., Woking, UK), probed with HRP-conjugated streptavidin (Cytivia Life Science/ Fisher Scientific UK Ltd., Loughborough, UK) and developed using TMB stabilized substrate (Promega Corp., Southampton, UK). Parallel sets of samples were also examined using a rabbit anti-filamin A antibody (EP2405Y) (Epitomics Inc, Burlingame, CA, USA). The cells were then probed with goat anti-rabbit-alkaline phosphatase antibody (Santa Cruz Biotechnology) and developed using Western Blue substrate (Promega). In some experiments, HCASMC were transfected using TransIT-2020, to express TF-tGFP variants as described for the EC. In other experiments, HCASMC were transfected with a set of peptides corresponding to the cytoplasmic domain of TF with different serine-phosphorylation patterns (50), as stated in the results section. Briefly, the peptides (0.5 µg) were diluted in PBS (100 µl) and Chariot reagent (6 µl; Active Motif, La Hulpe, Belgium) was also diluted in PBS. The two reagents were then mixed and incubated for 30 min. HCASMC were adapted to serum-free medium (400 µl) for 1 h before and the Chariot-peptide mix (200 µl) was added to the cells. The cells were incubated for a further 60 min, the medium replaced and used in the experiments.

In some experiments, the isolated MV were pre-incubated for 1 h, with a mouse anti-human-TF antibody, 10H10 (20 µg/ml; BD Bioscience, Wokingham, UK) capable of blocking TF signaling, a mouse anti-human-TF antibody, HTF-1 (20 µg/ml; eBioscience/Thermo Scientific, Warrington, UK) to block TF-fVIIa protease/procoagulant activity, an inhibitory polyclonal rabbit anti-human fVIIa antibody (10 µg/ml; Abcam, Cambridge, UK) or the respective control isotype IgG antibodies (20 µg/ml; New England Biolabs, Hitchin, UK). The MV were then added to the lower chamber in migration assays as below. Alternatively, HCASMC were treated with a mouse anti-human PAR2 antibody, SAM11 (20 µg/ml; Santa Cruz Biotechnology, Heidelberg, Germany), capable of blocking PAR2 signaling, or PAR2-agonsit peptide (PAR2-AP) to induce PAR2 signaling. SAM11 antibody was added with the HCASMC in the upper chamber at the start of the experiment and PAR2-AP was added to the lower chamber. PDGF-BB and PDGF-AA (10 ng/ml) were also added to sets of cells, in the bottom chamber as positive and negative controls, respectively. Additionally, to examine the influence of TF alone, and the contributions of other factors, HCASMC were stimulated with combination of recombinant Innovin TF (0–260 pg/ml; Dade Behring, Deerfield, USA), fVIIa (3 nM; Enzyme Research Lab., Swansea, UK), fXa (6 nM; Enzyme Research Lab.) and tissue factor pathway inhibitor (TFPI; 1 nM; American Diagnostica/Axis-Shield Ltd, Dundee, UK). To examine the contribution of activation of FAK, HCASMC were pre-incubated with either FAK inhibitor-14 (0–100 μM; 1,2,4,5-benzenetetraamine tetrahydrochloride; Sigma Chemical Company Ltd, Poole, UK) to inhibit FAK autophosphorylation, or AG82 {2-[(3,4,5-trihydroxyphenyl)methylene]-propanedinitrile}; (0–10 μM; Calbiochem/Merck-Millipore, Gillingham, UK) that prevents receptor-kinase mediated p125FAK tyrosine phosphorylation. Other sets of cells were pre-incubated with calyculin-A (1 nM) in order to prevent TF de-phosphorylation.

We have previously shown that exogenous-TF can become incorporated into the recipient cell membrane (26, 28, 47, 51). The HCASMC migration assay relies on the ability of cells to traverse a membrane to the lower chamber using Boyden chambers (8 μm pore size) (VWR International Ltd., Leicestershire, UK). For each assay, identical number of cells (approximately 3 × 10⁴ in 250 μl of complete SMC-MV2 media) were placed in the upper chamber. If required, the cells were supplemented with antibodies directed towards the cellular receptors (e.g., SAM11), or other reagents. Complete media (250 μl) was placed in the lower chamber containing the test reagent as described above and indicated in the results section. The chambers were incubated at 37°C for 18 h. The cells were then fixed with glutaraldehyde (3% v/v), washed with PBS and then, the cells on the upper side of the chamber were scraped off. To quantify the number of migrating cells, the scraped chambers were stained with crystal violet solution (Active Motif) for 30 min, washed and 10 fields of view were photographed. The cell-associated crystal violet was then released from the cells by incubation with 1% (w/v) SDS (200 μl) for 20 min. The absorptions were then measured at 595 nm and migrated cell numbers were determined from a standard curve prepared using HCASMC. In some experiments, the chambers were washed and stained with haematoxylin for visualization and 10 fields of view were counted manually or subsequently stained with crystal violet as above. For clearer demonstration of HCASMC migration, alternate images have been presented in the results section displaying cells stained with crystal violet at ×4, or at ×10 magnification.

Total RNA was isolated using the Monarch total RNA extraction kit (New England Biolabs, Hitchin, UK) from 2 × 10⁵ cells and 100 ng of total RNA was used for each reaction. The relative amounts of FAK mRNA was determined using the primer (forward 5′-AATCCTGGAGGAAGAGAAGGC-3′; reverse 5′-TGTTGCTGTCGGATTAGACGC-3′) and semi-qualitative analysis carried out in comparison to β-actin (forward 5′-CCAGAGCAAGAGAGGCATCC-3′; reverse 5′-CTGTGGTGGTGAAGCTGTAG-3′). One-step RT-PCR reaction was carried out with 100 ng of isolated total RNA using Ready-To-Go RT-PCR Beads (Amersham Pharmacia Biotech, Inc., Giles, UK) for 30 cycles for FAK or 18 cycles for β-actin. RT-PCR products were visualized on a 2% (w/v) agarose gel, stained with SYBR Green I (FMC BioProducts, Rockland, USA).

Western blot analysis was carried out to determine the amount of Smoothelin-B protein in the HCASMC. Cells (10⁵) were lysed in loading buffer and electrophoresis was carried out on a denaturing 8% (w/v) polyacrylamide gel, transferred to a nitrocellulose membrane and blocked with Tris-buffered saline Tween 0.01% (v/v) (TBST) pH 8. The membranes were probed overnight with a rabbit anti-human Smoothlein antibody (diluted 1:3,000 v/v; Abbexa, Cambridge, UK), or a goat anti-human GAPDH (V18) antibody (diluted 1:3,000 v/v; Santa Cruz Biotechnology, Heidelberg, Germany). The membranes were then washed and developed with goat anti-rabbit, or donkey anti-goat alkaline phosphatase-conjugated antibodies (Santa Cruz), diluted 1:3,000 (v/v), and visualized using the Western Blue stabilized alkaline phosphatase-substrate (Promega).

Presented data include the calculated mean values ± the calculated standard error of the mean, and the number of experiments is stated in each column. Statistical analysis was carried out using the GraphPad Prism version 9.0 (GraphPad Software, Boston, Massachusetts USA). Significance was determined using one-way ANOVA (analysis of variance) and Tukey's honesty significance test.

HCASMC were incubated with MV isolated from HCAEC expressing TF_Wt-tGFP, TF_Asp253-tGFP, TF_ΔCT-tGFP, tGFP alone, or from non-transfected cells (control MV). The incorporation of TF-tGFP into the cells and interaction with filamin A (probed with EP2405Y antibody) was monitored qualitatively, by confocal microscopy (Supplementary Figure S2). In addition, the association of TF-tGFP with the c-terminal of filamin A was confirmed in situ by proximity ligation assay as before (28, 45), and using a rabbit monoclonal antibody against the c-terminal of filamin A (EP2405Y), in conjunction with first a mouse monoclonal antibody to TF (10H10) and also, with a mouse monoclonal antibody to tGFP protein (2HB, OriGene) (Supplementary Figure S3). The MV were isolated using an established ultracentrifugation procedure, confirmed and quantified as before (48). Selected experiments were repeated using MV derived from transfected ECV304 line. These cells do not express TF under normal conditions but the transfected cells express the recombinant protein efficiently, and the cells release large quantities of MV on PAR2 activation. Typically, the yield of MV ranged between 0.25–0.31 nM and the concentration of TF in the MV from cells that expressed TF-tGFP was between 1 and 3 ng/ml. For the HCASMC migration experiments, 0.5 pg/ml (or equivalent amount of control MV) was placed in the bottom chamber of the Boyden chamber. Additional sets of HCASMC were supplemented with PDGF-BB or PDGF-AA which was added to the bottom chamber, as positive and negative controls, as well as sets of non-treated sample. The reagents were placed in the bottom chamber and the rate of cell migration across the Boyden chamber measured. Incubation of HCASMC with MV containing TF_Wt-tGFP resulted in increased cell migration compared to cells expressing either tGFP alone, or the control MV. The increases were comparable to but lower than those obtained using the positive control (PDGF-BB) (Figures 1A,B and Supplementary Figure S4). However, the rate of migration was lower with TF_ΔCT-tGFP indicating that the cytoplasmic domain of TF is essential for full effectiveness of the signal. Interestingly, incubation of cells with MV containing TF_Asp253-tGFP resulted in reduced cell migration to levels comparable to the negative control (PDGF-AA).

In addition, HCASMC were incubated with combinations of recombinant TF, fVIIa, fXa and TFPI. Incubation of cells with the combination of TF and fVIIa resulted in increased cell migration but was not significantly altered by supplementation with fXa (Figures 2A,B). In contrast, neither concentration of recombinant TF was capable of promoting HCASMC migration in the absence of fVIIa. Furthermore, stimulation with fVIIa alone was less effective in inducing HCASMC migration while fXa alone was only marginally effective. In accordance with previous studies, HCASMC expressed endogenous TF but the levels were low in our study which is also in agreement with the levels reported in healthy vessels (52–55). However, neither the zymogen nor activated fVII/VIIa were detected in HCASMC when examined by western blot (not shown), using a previously verified polyclonal anti-fVIIa antibody (56). Moreover, inclusion of recombinant TFPI reduced the induction of migration. In an attempt to examine the state of HCASMC, the expression of the Smoothelin was examined by western blot analysis. Smoothelin-B was shown to be present in resting HCASMC but not following treatment with TF-fVIIa (Figure 3). However, other markers were not examined in this study. Similarly, the expression of Smoothelin-B was suppressed in HCASMC treated with MV containing TF_Asp253-tGFP, or with PDGF-BB, was partially reduced in cells treated with MV containing TF_ΔCT-tGFP but was not altered in cells treated with tGFP-MV (Figures 3C,D).

To further examine the underlying mechanisms of induction of HCASMC migration by TF, MV were isolated from HCAEC and ECV304 cell line that were transfected to express TF_Wt-tGFP, and then pre-incubated with an inhibitory polyclonal antibody against fVIIa (56, 57), a monoclonal anti-TF antibody capable inhibiting the procoagulant activity (HTF-1) (58), or a monoclonal antibody that blocks TF-mediated signaling (10H10) (59), before adding the MV to the lower chamber. These antibodies are directed to the extracellular domain of TF and fVIIa also associates with the extracellular domain. Alternatively, the HCASMC were pre-incubated with a monoclonal antibody capable of blocking PAR2 activation and signaling (SAM11). Appropriate isotype antibodies were analyzed alongside according to the species of the antibody used. The inclusion of these antibodies did not alter the cell migration by TF-MV, and the isotypes IgG antibodies showed no influence otherwise (Figure 4). The effectiveness of TF-tGFP-containing MV to induce HCASMC migration was largely abolished on blocking PAR2 signaling by pre-incubation of HCASMC with SAM11 (Figures 4A,B and Supplementary Figure S5). Similarly, neutralization of TF-fVIIa protease activity, by pre-incubation with an inhibitory anti-fVIIa antibody, or incubation with HTF-1 antibody significantly reduced the rate of HCASMC migration while, the blocking of TF signaling using 10H10 antibody was ineffective. Stimulation of HCASMC with the PAR2 activating peptide (PAR2-AP, SLIGKV) or PAR1 activating peptide (PAR1-AP, SFLLRN) did not alter the rate of HCASMC migration, while pre-incubation of HCASMC with a monoclonal antibody to block β1-integrin completely abolished any cellular migration (not shown).

In order to examine the role of TF phosphorylation, HCASMC were transfected with peptides corresponding to the 18 amino acids of the cytoplasmic domain of TF and the rate of migration examined as above. The transfection of a similar FITC-labelled peptide into the cells was confirmed by flow cytometry which indicated up to 85% transfection rate in both washed and unwashed cells and mean fluorescence intensity of up to 3 orders of magnitude (Supplementary Figure S6). Additionally, transfection of the peptides does not influence cell viability (46, 50). Transfection with either form of the peptide, incorporating a phosphorylated-Ser253 (pSer253 or pSer253/pSer258) resulted in significant reduction in the number of migrating cells in response to TF-fVIIa (Figures 5A,B). In contrast, the peptide incorporating phospho-Ser258 (pSer258) and the non-phosphorylated form were largely ineffective.

Phosphorylation of Ser253 is known to be essential for the interaction of TF with the repeat domain 22–24 within filamin A (43, 45). Additionally, TF is known to colocalize with filamin A at the leading edges of migrating SMCs (19). To further assess the contribution of filamin A to TF-mediated cell migration, HCASMC were transfected to express three truncated forms of filamin A without the repeat domains 22–24, 23–24 and 24 alone. The expression of the filamin A variants (filamin A_Wt, filamin A_Δ22−24, filamin A_Δ23−24 and filamin A_Δ24) was confirmed by western blot, probed using an antibody against the c-myc and anti-filamin A (Supplementary Figure S7). Moreover, the lack of any detrimental outcome on cell viability was established by determining the cell numbers in the transfected and control cells. HCASMC migration across Boyden chamber membranes was then stimulated with the HCAEC-derived MV containing TF_Wt-tGFP, or devoid of TF (control MV). Removal of either set of repeat domains (22–24, 23–24 or 24 alone) resulted in significantly reduced HCASMC migration (Figures 6A,B). Next, whether repeat-24 is required specifically for TF-induced migration, or also affects other/non-specific migration mechanisms/processes, was examined. Expression of either filamin A_Wt, or filamin A_Δ24 resulted in similar rates of migration in unstimulated HCASMC, cells stimulated with control MV, or in response to PDGF-BB (10 ng/ml) (Figures 6A,C). These were in contrast to the difference in migration observed on stimulation with TF-MV. Consequently, this study indicates that repeat-24 is likely to participate in TF-specific promotion of migration in HCASMC.

While the incorporation of TF within the focal adhesion points may establish the direction of cell migration (44), focal adhesion disassembly is vital for trailing edge retraction. In the above experiments the stimulation of HCASMC with MV carrying TF_Asp253-tGFP resulted in the retardation of cell migration. In addition, transfection of HCASMC with phospho-Ser253 peptide prevented the stimulation with MV carrying TF_Wt-tGFP. To examine the involvement of TF-de-phosphorylation in chemotaxis, first PP2A phosphatase activity was inhibited in HCASMC (46). Pre-incubation of HCASMC with calyculin-A (1 nM) prior to stimulation with MV carrying TF_Wt-tGFP severely suppressed TF-induced cell migration (Figures 7A,B). However, cell migration in response to PDGF-BB was also affected by inclusion of calyculin-A. Therefore, no definitive conclusions were derived from these experiments. To further explore the possibility that the TF-de-phosphorylation is required for migration, HCASMC were transfected to express TF_Wt-tGFP, TF_Asp253-tGFP, or tGFP alone and the rate of cell migration in response to fVIIa, examined. Since unlike TF_Wt-tGFP, TF_Asp253-tGFP retains the ability to interact with filamin A, the observed reduction in the rate of HCASMC migration suggests that TF-de-phosphorylation is likely to be essential as part of the chemoattractant or migratory mechanism (Figures 7C,D).

The above data suggest that HCASMC migration in response to TF may be distinct to that induced by growth factor engagement. To further decipher the mechanisms involved in TF-mediated cell migration and to differentiate between the TF-induced and growth factor/cytokine-promoted cell migration, HCASMC were pre-incubated with either FAK inhibitor-14 (0–100 μM; 1,2,4,5-benzenetetraamine tetrahydrochloride) to inhibit FAK autophosphorylation, or AG82 that prevents receptor-kinase mediated p125FAK tyrosine phosphorylation. Preincubation of cells with FAK inhibitor-14 inhibited the TF-MV mediated SMC migration at both concentrations used. However, pre-incubation of the cells with AG82 (up to 10 μM) did not interfere with TF-fVIIa induced cell migration (Figures 8A,B) but reduced response to PDGF-BB (not shown). Furthermore, examination of the expression of FAK mRNA, in HCASMC following incubation with TF_Wt-tGFP containing MV, indicated the upregulation of FAK expression at 18 h post-treatment (Figures 8C,D), but not on incubation with control MV.