Fig 2: HBMSCs plated and cultured at high density differentiated toward functional ECs. To determine whether HBMSCs plated and cultured at high density differentiated toward EC over time, their morphology, EC functional properties, and expression of EC markers were determined. (A) Representative phase contrast images of (I) HBMSCs under standard culture conditions (MSC-St) for 48 h, (II) HBMSCs cultured at low density (HBMSC-L) for 14 days, or (III) HBMSCs cultured at high density (HBMSC-H) for 14 days. (IV) HUVECs under standard culture conditions were used as a positive cellular control. For culture conditions, see Materials and methods. Images were obtained using an Olympus (CK X41) microscope (4x objective). Inserts represent an enlarged region of each image. Scale bars = 200 µm. (B) (I–III) Immunofluorescence analysis showing DiI-Ac-LDL uptake in (I) HUVECs, (II) HBMSCs under standard culture conditions (HBMSC-St) for 48 h, and (III) HBMSCs cultured at high density (HBMSC-H) for 28 days. DiI-Ac-LDL uptake (red), DAPI (blue). Images obtained using an Olympus upright widefield fluorescence (BX51) microscope (20x objective). Scale bars = 20 µm. (IV–VI) Immunofluorescence analysis of VCAM-1 in cells stimulated with 10 µg/ml TNFa for 24 h. (IV) HUVECs under standard culture conditions for 48 h, (V) HBMSCs under standard culture conditions (HBMSC-St) for 48 h, and (VI) HBMSCs at high density (HBMSC-H) for 28 days. VCAM-1 (green), DAPI (blue). Images were obtained using a Nikon C1 upright confocal microscope (60x objective). Scale bars = 20 µm. (C,D) Immunofluorescence analysis of VE-cadherin, PECAM-1, vWF, and VEGFR1 expression in (I, IV) HUVECs under standard culture conditions for 48 h, (II, V) HBMSCs under standard culture conditions for 48 h, or (III, VI) HBMSCs cultured at high density (HBMSC-H) for (C) 28 days or (D) 14 days. VE-cadherin, PECAM-1, vWF, and VEGFR1 (green), phosphorylated VEGFR1 (red), DAPI (blue). Images were obtained using a Nikon C1 upright confocal microscope (60x objective) (C, I–VI; and D, IV–VI) or using an Olympus IX71 Deltavision microscope (40x objective) (D, I–III). Scale bars = 10 µm. White arrows highlight specific regions of localised immunoreactivity. Representative images from two independent experiments.

Fig 3: High density HBMSCs expressed increased endothelial cell markers but reduced smooth muscle cell markers, by 3 days. The levels of vascular cell markers were determined in HBMSCs plated and cultured at high density for 3 days. (A) ELISA assay of VEGF-A derived from medium taken from HBMSCs cultured at low density for 24 h, or at high cell density up to 3 days, normalised against corresponding ß-actin levels. (B–E) Immunoblot analysis of (B) vWF, (C) VEGFR1, (D) SM-MHC-1, and (E) smoothelin-B (SMTH) in HBMSCs cultured at low density (HBMSC-L) for 24 h or at high cell density up to 3 days. Pixel density was normalised to ß-actin and plotted as bar graphs. * P < 0.05 compared to HBMSCs cultured after plating at low cell density.

Fig 4: VEGF-A inhibited HBMSC differentiation to vSMCs but promoted EC differentiation. To determine whether secreted VEGF-A inhibited vSMC marker expression in HBMSCs cultured at high density, the effects of blocking VEGF-A signaling was determined. (A,B) HBMSCs cultured at high density for 24 h were (A) exposed to 1 µg/ml VEGF neutralisation antibody (VEGF-I) (lane 2) or (B) transfected with VEGF-A siRNA (lane 2), and then SM-MHC-1 expression was determined by immunoblot analysis. Untreated HBMSCs (Cont) or scrambled (Scr) siRNA-transfected HBMSCs (lane 1) were used as respective controls. Pixel density was normalised to ß-actin and plotted as a bar graph. * P < 0.05 compared to controls. Data are representative of two independent experiments for each analysis. (C,D) HBMSCs cultured at high density for 14 days were exposed to 1 µg/ml VEGF-I (lane 2) or 0.5 µM VEGFR tyrosine kinase inhibitor (VEGFR-I) (C, lane 3, and D, lane 5), and then (C) VEGFR1 and (D) VE-cadherin expression was determined by immunoblot analysis. Untreated HBMSCs were used as controls (Cont) (C, lane 1, and D, lanes 1 and 4). Pixel density was normalised to ß-actin and plotted as a bar graph. * P < 0.05 compared to untreated HBMSCs. A representative of two independent experiments is shown in each case. (E) HBMSCs cultured at high density for 14 days were exposed to 50 ng/ml recombinant VEGF-A, and then VEGFR transcript expression was determined by semiquantitative RT-PCR analysis. Lanes 1–2, GAPDH in untreated (control) HBMSCs. VEGFRs 1–3 in untreated (control) HBMSCs, lanes 3–5, or following exposure to recombinant VEGF-A, lanes 6–8, respectively. Two different primer pairs for VEGFRs 1–3 gave similar results. Red boxes denote unstimulated cells; blue boxes denote VEGF-A-stimulated cells. (F) Immunoblot analysis of VEGFR1 protein level in untreated HBMSCs (lane 1), or following exposure to recombinant VEGF-A for 14 days (lane 2). * P < 0.05 compared to untreated HBMSCs cultured after plating at high cell density (HBMSC-H). In each case, fresh medium supplemented with VEGF-A or inhibitors was added to cells every 2 days.

Fig 5: Notch activation stimulated HBMSCs at low density to express EC markers. To determine whether experimental activation of Notch could induce vWF and VEGFR1 protein expression in HBMSCs plated at low density and cultured for 24 h, HBMSCs were treated with Notch activators. (A,B) Immunoblot analysis of (A) VEGFR1 or (B) vWF protein expression from HBMSCs cultured after plating at low density (L) (lanes 1 and 2), or high cell density (H) (lanes 3 and 4), in the absence (-) (lanes 1 and 3) or presence (+) (lanes 2 and 4) of 5 mM EDTA. (C) To verify that HBMSCs cultured at low density could be induced to up-regulate vWF protein expression by activation of Notch signaling, Notch was activated using 5 µM Jagged-1. Immunoblot analysis of vWF protein expression in HBMSCs cultured after plating at low density (lanes 1 and 2) or high cell density (lanes 3 and 4) in the absence (-) (lanes 1 and 3) or presence (+) (lanes 2 and 4) of 5 µM Jagged-1. Membranes were reprobed with ß-actin as loading controls. Data are representative of two independent experiments.