Chloromethylbenzamido-DiI–labeled endothelial progenitor cell (EPC) (red; 4,6-diamino-2-phenylindole–positive cell nuclei appear blue in A and B), isolated from the peripheral blood of lean or obese individuals, were analyzed for their capacity to adhere to fibronectin (A), migrate in the modified Boyden chamber assay (B), or incorporate into network-like structures provided by human umbilical vein endothelial cells (translucent cells) in the Matrigel angiogenesis assay (C; arrows point to EPCs). In some obese individuals, the studies were repeated after weight loss. (D) Cumulative results. *p < 0.05 and ***p < 0.001 versus lean; #p < 0.05, ##p < 0.01, and ###p < 0.001 versus obese.

After induction of hind limb ischemia and injection of chloromethylbenzamido-DiI (CM-DiI)–labeled endothelial progenitor cells (EPCs) from either lean or obese subjects, 5-µm-thick serial cryosections were analyzed for the presence of EPCs (red signal; arrows in A, top); costaining of cell nuclei with 4,6-diamino-2-phenylindole (DAPI) (blue signal) demonstrates that no artificial signals were counted (A, bottom). (B) To assess angiogenesis after ischemia, the density of CD31-positive capillaries per square millimeter (green signal; 4,6-diamino-2-phenylindole–positive cell nuclei appear blue) was determined in the ischemic (right) and the contralateral nonischemic (left) lower hind limb. *p < 0.05 and #p < 0.05 versus lean; **p < 0.01 versus nonischemic leg.

(A) Conditioned medium was harvested, and its effects on the angiogenic properties of EPCs were tested using the Matrigel assay. (B) The effect of conditioned medium alone on the sprouting of human umbilical vein endothelial cells was analyzed using the spheroid angiogenesis assay. (C) Cytokines in conditioned medium from lean versus obese individuals, before and after weight loss: representative findings of cytokine antibody arrays (3 independent experiments, each using conditioned medium pooled from 4 persons per study group; interleukin (IL)-6 and tumor necrosis factor (TNF)- were not detected by the assay) together with the corresponding quantitative analysis using specific enzyme-linked immunosorbent assays. *p < 0.05 and **p < 0.01 versus lean. CdM = conditioned medium; HUVEC = human umbilical vein endothelial cell; MCP = monocyte chemoattractant protein; WL = weight loss; other abbreviation as in Figure 2.

The angiogenic response of EPCs to stimulation with vascular endothelial growth factor (VEGF) (10 ng/ml for 24 h) was analyzed using the Matrigel (A) and the spheroid (B) angiogenesis assays. Results are expressed as percentage of unstimulated cells. (C), Expression of the indicated surface receptors was analyzed using flow cytometry: representative histograms (green line, lean EPCs; red line, obese EPCs; gray line, IgG control). *p < 0.05 and **p < 0.01 versus unstimulated cells; #p < 0.05 and ##p < 0.01 versus stimulated cells from lean individuals; p < 0.01 versus stimulated cells from obese individuals. CCR2 = chemokine (C-C motif) receptor 2; CXCR2 = chemokine (C-C motif) receptor 2; PE = phycoerythrin; VEGFR = vascular endothelial growth factor receptor; other abbreviations as in Figures 1 and 3.

Total protein lysates were prepared from unstimulated EPCs, and the basal phosphorylation of p38 MAPK was analyzed by Western blot. (A) Representative findings. (B) Quantitative analysis of 9 individuals per group. Results were normalized to total p38 protein and expressed as times-fold increase compared with levels found in lean subjects. The effect of p38 MAPK inhibition (using 20 µM of SB203580) or dimethylsulfoxide (DMSO) (negative control) on the interleukin-8 and monocyte chemoattractant protein 1 (MCP1) levels in conditioned medium from lean and obese EPC (C) and their angiogenic properties (Matrigel assay; D) was analyzed. *p < 0.05 and **p < 0.01 versus lean; #p < 0.05 versus control-treated obese EPC; ###p < 0.001 versus obese. MAPK = mitogen-activated protein kinase; other abbreviations as in Figures 2 and 3.