Materials

Rivaroxaban was purchased from ChemScene LLC (Monmouth Junction, NJ, USA), and was dissolved in dimethyl sulfoxide (DMSO) before use. Warfarin sodium was purchased from Wako Pure Chemical Industries (Osaka, Japan).

Isolation and preparation of human CD34-positive cells

For EPC colony-forming assays, human mononuclear cells were isolated by density gradient centrifugation using Histopaque1077 (Sigma-Aldrich, St. Louis, MO, USA) from heparinized peripheral blood of healthy volunteers and patients with chronic coronary artery disease who underwent coronary stent implantation. The mononuclear cells were also isolated from umbilical cord blood (Riken BRC Cell Bank, Tsukuba, Japan) for other experiments. CD34-positive cells were purified using anti-CD34 monoclonal antibody-conjugated microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) and a magnetically activated cell sorter (auto MACS; Miltenyi Biotec) following the manufacturer’s protocol. The isolated cells contained approximately 95% pure CD34-positive cells.

EPC colony-forming assays

Human CD34-positive cells were prepared from the peripheral blood of healthy volunteers as well as coronary artery disease patients. Approximately 2 × 103cells were cultured in methylcellulose (H4236; StemCell Technologies, Vancouver, BC, Canada) with 100 ng/mL stem cell factor (SCF), 50 ng/mL vascular endothelial growth factor (VEGF), 20 ng/mL interleukin (IL)-3, 50 ng/mL basic fibroblast growth factor (bFGF), 50 ng/mL epidermal growth factor (EGF), 50 ng/mL insulin-like growth factor-1 (IGF-1; all from PeproTech, Rocky Hill, NJ, USA), and 2 U/mL heparin (Wako Pure Chemical Industries, Osaka, Japan) in 3.5-cm dishes. After 21 days, colony formation was observed. We counted the number of total colonies and colonies for small- or large-type EPCs in dishes using a phase-contrast microscope (CK30; Olympus, Tokyo, Japan). Small- and large-type EPCs were identified by visual inspection with an inverted microscope under 40 × magnification. Small-type EPCs were composed of round adhesive cells, and large-type EPCs were composed of spindle-shaped cells [17, 18].

Ex-vivo suspension culture of EPCs

CD34-positive cells isolated from umbilical cord blood were expanded as previously reported [17]. Briefly, 4 × 105 cells were cultured at 37 °C in an atmosphere containing 5% CO2 in Stem Span medium (StemCell Technologies) with 50 ng/mL VEGF, 20 ng/mL IL-6, 100 ng/mL SCF, 20 ng/mL thrombopoietin (PeproTech), and 100 ng/mL fms-related tyrosine kinase 3 ligand (PeproTech) in 24-well plates for 7 days. Ex-vivo expanded CD34-positive cells (1.5 × 106 cells) were seeded in 10-cm culture dishes coated with human fibronectin (Corning, Bedford, MA, USA) and then cultured in M199 medium (Gibco; Thermo Fisher Scientific, Grand Island, NY, USA) supplemented with 5% fetal bovine serum (Sigma-Aldrich), 10% dextran (molecular weight: 100,000–200,000; cat. no. D4876; Sigma-Aldrich), and endothelial growth medium containing VEGF, bFGF, EGF, IGF-1, and ascorbic acid (EGM-2; Lonza, Basel, Switzerland) at 37 °C in an atmosphere containing 5% CO2. After 7 days of culture, adhesive cells were used as EPCs.

Adhesion assay

EPCs (2 × 104) in suspension with or without rivaroxaban for 7 days were seeded onto human fibronectin-coated 96-well plates in M199 medium with 10% fetal bovine serum (FBS) and 10% dextran. After incubation with 5% CO2 for 6 h, nonadherent cells were removed by gently washing three times with phosphate-buffered saline (PBS). Adhesive cells were examined under a phase-contrast microscope equipped with a digital camera. The number of adherent cells per image was measured.

Tube formation assay

For analysis of tube formation, 2 × 103 EPCs and 1.5 × 104 human umbilical vein endothelial cells (HUVECs) were cultured with or without rivaroxaban in EBM-2 medium (Lonza) with 2% FBS and added to an equivalent amount of Matrigel (BD Falcon; Becton Dickinson, San Jose, CA, USA) in 96-well plates. After incubation at 37 °C in an atmosphere containing 5% CO2, gels were observed using phase-contrast microscopy. The number of circles per tube structure was counted in each image.

Flow cytometry

EPCs were washed with cold PBS and resuspended in PBS with FcR blocking reagent (Miltenyi Biotec), 2% FBS, and 2 mM ethylenediaminetetraacetic acid (EDTA) at 4 °C for 15 min. Cells were then stained with phycoerythrin (PE)-conjugated monoclonal antibodies specific for the following surface antigens: CD31 (Becton Dickinson), VEGF receptor (VEGFR)-1 (Becton Dickinson), VEGFR-2 (Becton Dickinson), protein receptor tyrosine kinase, epithelial-specific (Tie)-2 (R&D Systems, Minneapolis, MN, USA), and E-selectin (CD62E; Becton Dickinson). After incubation at 4 °C for 30 min, cells were washed twice with PBS and analyzed using a FACS Calibur flow cytometer (Becton Dickinson).

Western blot analysis

Cell lysates were prepared using ice-cold lysis buffer (50 mM Tris–HCl, 150 mM NaCl, 1 mM EDTA, 1 mM ethylene glycol tetraacetic acid, 1% NP-40, 10% glycerol, 20 mM Na4P2O7·10H2O, 200 mM NaF, and 1 mM Na3VO4 plus phosphatase and protease inhibitor). Samples (20–35 μg total protein for Akt/endothelial nitric oxide synthase [eNOS] experiments and 200 μg total protein for PAR-2 knockdown experiments) were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis using 6% or 8% gels and then transferred to Amersham Hybond P PVDF 0.45 membranes (GE Healthcare UK, Buckinghamshire, UK). The blots were blocked with Blocking One (Nacalai Tesque, Kyoto, Japan) for 1 h at room temperature, incubated overnight at 4 °C with primary antibodies, and washed three times with Tris-buffered saline/0.1% Tween 20 (TBS-T). Membranes were then incubated with horseradish peroxidase-conjugated anti-rabbit (anti-mouse for glyceraldehyde-3-phosphate dehydrogenase [GAPDH]) IgG antibodies, according to standard methods. Immunoreactive signals were enhanced with Chemi-Lumi One Super (Nacalai Tesque) or Chemi-Lumi One Ultra (Nacalai Tesque) and visualized using a WSE-6100H LuminoGraph I (ATTO, Tokyo, Japan). The intensity of each band was measured using a CS Analyzer 4 (ATTO). Quantitative analysis was achieved by normalization of the signal of each protein to that of GAPDH or β-actin. Anti-eNOS, anti-phospho-eNOS (Ser1177), anti-Akt, anti-phospho-Akt (Ser473), anti-β-actin, and anti-PAR-2 antibodies were obtained from Cell Signaling Technology (Danvers, MA, USA). Anti-GAPDH antibodies were obtained from Novus Biologicals (Centennial, CO, USA).

Transfection with synthetic small interfering RNA (siRNA)

PAR-2 siRNA and control siRNA (scrambled siRNA) were purchased from Invitrogen (Carlsbad, CA, USA). PAR-2 siRNA and control siRNA were transfected into EPCs to a final concentration of 5 nM with 0.3% Lipofectamine RNAiMax (Invitrogen). Transfected cells were incubated at 37 °C in an atmosphere containing 5% CO2 for 7 days before use in experiments.

Clinical experiments

In the clinical investigation, patients with chronic coronary artery disease (i.e., stable effort angina or old myocardial infarction) and atrial fibrillation, who underwent PCI using drug-eluting stents, were enrolled. These patients were receiving an anticoagulant warfarin in addition to single antiplatelet therapy with aspirin (100 mg/day) until undergoing PCI. Blood coagulation activity was adequately suppressed by warfarin treatment as the 2.0–3.0 international normalized ratio of the prothrombin time. Immediately after PCI, the patient received dual antiplatelet therapy by adding a P2Y12 inhibitor prasugrel (loading dose: 20 mg, maintenance dose: 3.75 mg/day) to aspirin. Simultaneously the patients switched anticoagulant therapy from warfarin to rivaroxaban. For rivaroxaban prescription, the dosage was selected based on recommendations in Japan (10 mg/day for patients with a creatinine clearance of 15–49 mL/min or 15 mg/day for patients with a creatinine clearance of ≥ 50 mL/min). In these patients, we observed changes in circulating EPCs during the acute phase after PCI and assessed stented-site vascular healing during the chronic phase.

We collected peripheral blood to observe the circulation EPCs before PCI under the warfarin prescription and 7 days after PCI under the rivaroxaban prescription. First, we measured the number of cells positive for both CD34 and kinase insert domain receptor (KDR) (CD34+/KDR+cells) as cell fractions including abundant EPCs using flow cytometry as previously described [19]. We then performed qualitative assessment using EPC colony-formation assays.

At follow-up coronary angiography 12 months after PCI, we assessed stented-site vascular healing by observing neointimal coverage over the stent strut using optical coherence tomography (OCT), as previously reported [19, 20]. Briefly, cross-sectional OCT images were analyzed at 0.6-mm intervals. In cross-sectional images, neointimal coverage was assessed for all struts, including uncovered struts and malapposed struts. The definition of malapposed struts was described by Sakuma et al. [19] The percentage of uncovered struts to total struts and that of malapposed struts to total struts was then calculated for all OCT cross-sections.

Statistical analysis

Values were presented as means ± standard deviations. The effects of serial rivaroxaban concentrations were analyzed using Fisher’s least significant difference procedure as follows: for Fig. 1, Friedman tests (paired nonparametric); for Figs. 2 and 3, analysis of variance followed by Student’s t tests as a post-hoc analysis (parametric); for Figs. 4, 5, 6 and 7, two-group comparisons were performed using Student’s t tests (parametric). Effects on patients were unable to be assessed due to the limited sample size. Results with P values less than 0.05 were considered significant.

Fig. 1

EPC colony-forming assays. A Colonies of primitive, dense, small, round EPCs showed higher proliferation potential and were composed of immature EPCs. Scale bar: 100 μm. B Colonies of definitive, large, spindle-type EPCs showed more vasculogenic properties and were composed of differentiating EPCs. Scale bar: 100 μm. C Effects of rivaroxaban on the numbers of total EPC colonies, small EPC colonies, and large EPC colonies. EPC: endothelial progenitor cell; DMSO, dimethyl sulfoxide. Bars are means ± standard deviations of nine samples

Fig. 2

Effects of rivaroxaban on adhesion and tube formation. A EPCs exposed to 0.1% DMSO (control) or rivaroxaban for 7 days were cultured for 6 h. The number of adhesive cells per low power field was counted. Bars are means ± standard deviations of five samples. B The adhesive cells were observed by phase-contrast microscopy. Scale bar: 100 μm. C EPCs exposed to 0.1% DMSO (control) or rivaroxaban for 7 days were cultured in Matrigel with human umbilical vein endothelial cells. The number of tubes per low power field was measured after 5 h. Bars are means ± standard deviations of five samples. D The tube formation observed by phase-contrast microscopy. Scale bar: 100 μm

Fig. 3

Flow cytometric analysis of the expression of endothelial cell markers on the surface of umbilical cord blood-derived CD34-positive cells. The effects of DMSO alone (control) and rivaroxaban on the percentages of cells positive for VEGFR-1, VEGFR-2, Tie-2, E-selectin, and CD31 were assessed. VEGFR: vascular endothelial growth factor receptor. Bars are means ± standard deviations of six or seven samples

Fig. 4

Western blot analysis of Akt/eNOS phosphorylation. A The effects of DMSO (control) and rivaroxaban on Akt phosphorylation levels were evaluated. p-Akt: phosphorylated Akt; GAPDH, glyceraldehyde-3-phosphate dehydrogenase. Bars are means ± standard deviations of five samples. B The effects of DMSO (control) and rivaroxaban on eNOS phosphorylation levels were evaluated. p-eNOS: phosphorylated endothelial nitric oxide synthase. Bars are means ± standard deviations of 5 samples. The original images without crop and correction are shown in the Additional file 2: Figure S2

Fig. 5

Transfection with synthetic PAR-2 siRNA into EPCs. A: The effect of transfection with synthetic PAR-2 siRNA into EPCs were evaluated by westernblotting. Values are means ± standard deviations of five samples. B: Expression of PAR-2 in EPCs transfected with PAR-2 siRNA compared with that in cells transfected with control siRNA. PAR, protease-activated receptor; siRNA, small interfering RNA. The full-length gels of PAR-2 protein are shown in the Additional file 3: Figure S3 and Additional file 4: Figure S4

Fig. 6

The bioactivities of EPCs transfected with PAR-2 siRNA. A EPCs transfected with control or PAR-2 siRNA for 7 days were cultured for 6 h. The number of adhesive cells per low power field was counted. Bars are means ± standard deviations of five samples. B The adhesive cells were observed by phase-contrast microscopy. Scale bar: 100 μm. C EPCs transfected with control or PAR-2 siRNA for 7 days were cultured in Matrigel with human umbilical vein endothelial cells. The number of tubes per low power field was measured after 5 h. Bars are means ± standard deviations of five samples. D The tube formation observed by phase-contrast microscopy. Scale bar: 100 μm

Fig. 7

Surface protein expression of endothelial cell markers in PAR-2-deficient. After PAR-2 siRNA transfection, effects of PAR-2 knockdown on surface protein expression of VEGFR-1, VEGFR-2, Tie-2, E-selectin, and CD31 were evaluated. Bars are means ± standard deviations of five samples (seven samples for VEGFR-2)