Fig 2: SDF-1a induces E-selectin expression in EPC.(A) Levels of mSDF-1a in sera 4 h after wound bed injection of ?mSDF-1a vs. PBS (n = 8 mice/group). (B) Measurement of levels of E-selectin in BM EPC by flow cytometry. BMC were incubated with Abs against CD34, KDR, and E-selectin or isotype-matched control Abs. EPC (CD34+/KDR+) were gated (P2), counted, and analyzed for levels of E-selectin (Q2-1). Bar graphs show % of EPC in BM mononuclear cells and relative levels of E-selectin in BM-EPC. Levels of E-selectin in EPC from mice injected with PBS were established as “1” and relative levels of E-selectin in EPC from mice injected with ?mSDF-1a were normalized accordingly (n = 8 mice/group). (C) Measurement of % of EPC in peripheral blood MNC and relative levels of E-selectin in circulating EPC as described in (B) (n = 8 mice/group). (D) Immunoblotting analysis of E-selectin expression upon ?hSDF-1a (100 ng/ml) stimulation in human EPC at various time points. ß-actin served as a loading control. (E) Human EPC were stimulated with ?hSDF-1a or BSA for 4 h, and total RNA was extracted. Expression of extracellular matrix and adhesion molecules were analyzed using RT2-PCRArray. Expression of E-selectin was upregulated upon ?mSDF-1a stimulation. Levels of mRNA in BSA-treated EPC were established as “1” and compared to those in ?mSDF-1a-treated EPC. Experiments were repeated three times in (D) and (E). Data are analyzed by 2-tailed Student’s t-test and presented as mean ± SEM.

Fig 3: SDF-1a induces expression of active form of E-selectin ligands in EC.(A) Increased expression of E-selectin ligand, CD44, in luminal EC of ischemic hindlimb wounds injected with ?mSDF-1a compared to PBS injected non-wounded hindlimb in NOD mice (n = 8 mice/group). (A) right: Co-expression (yellow) of CD44 (green) and CD31 (red) in vessels was detected by immunostaining. Representative images are shown (images of isotype-matched non-specific control Ab are not shown). Left: Quantification of CD44 expression in vessels. Data are presented as mean ± SD of ratio of CD44:CD31 signals from 5 random selected sections of high power field (LPF, X 20) of each wound sample. CD31 signal was established as “1” in each section and relative amount of CD44 signal was normalized accordingly. (B) Immunoblotting analysis of SDF-1a-induced expression of three E-selectin ligands in HMVEC at various time points (?hSDF-1a: 100 ng/ml). ß-actin is used as loading control. Experiments were repeated three times and similar results were obtained. (C) Binding of FITC-conjugated HECA452 to HMVEC stimulated with ?hSDF-1a or BSA. Fluorescent signals were quantified (top). Representative fluorescent images were exhibited (bottom). (D) Binding of FITC-conjugated HECA452 to human EPC stimulated with ?hSDF-1a or BSA (100 ng/ml). Fluorescent signals were quantified (top). Representative fluorescent images were exhibited (bottom). Data are analyzed by 2-tailed Student’s t-test and presented as mean ± SEM of fluorescent signals based on triplicate wells in each condition and totally three independent experiments in (C,D).

Fig 4: Illustration of a “double-lock” mechanism mediating EPC homing to ischemic wounds tissues for therapeutic neovascularization and wound repair.SDF-1α acts both locally and remotely to induce ischemic tissue endothelium and BMD-EPC to express E-selectin/ligands. Dual E-selectin/ligand pairs reciprocally expressed on activated endothelium and BMD-EPC mediate enhanced EPC-EC interactions and selective EPC homing.

Fig 5: E-selectin/ligands are required for EPC-dependent blood vessel formation.(A) Wound blood vessel perfusion with Dil dye. Upper: representative images of Dil-stained wound blood vessels detected by confocal laser scanning photography at day 7 are shown for each type of recipient mice (I–IV as illustrated in Fig. 4A). Lower: Quantification of vessel density in the entire area of residual wounds on day 7, as percent fluorescence. (B) Upper: representative images of non-invasive LDI measurements showing spontaneous restoration of blood flow into ischemic hindlimbs after femoral artery ligation/excision in different recipient mice (I–IV as illustrated in Fig. 4A). Lower: Quantitative data of LDI measurements. Ratio of ischemic versus normal hindlimb between two groups of mice at various time points. For both (A,B), data are analyzed by two-way ANOVA test and presented as mean ± SEM from each group (n = 6/group).