Insulin Resistance and Oxidative Stress Influence Colony-Forming Unit-Endothelial Cells Capacity in Obese Patients




The aim of this study was to investigate the relationship between a sub-population of endothelial progenitor cells (EPC), namely colony-forming unit-endothelial cells (CFU-EC), their colony-forming capacity and variable clinical parameters, including insulin resistance and oxidative stress, in obese individuals. Thirty-eight obese adults (aged 42.5 ± 12.7), with BMI 32.3 ± 4.0 and 13 normal-weight controls (aged 48.2 ± 12.9; BMI 23.2 ± 2.3) were studied. CFU-EC colony-forming capacity was impaired in the group of obese individuals compared to the normal-weight controls (P = 0.001). The inverse correlation between homeostasis model assessment-insulin resistance (HOMAIR) index and CFU-EC number (r = −0.558, P < 0.0001) as well as positive total antioxidant status of plasma (TAS)/CFU-EC relation were noticed during the study. Additionally, correlations between the concentration of triglycerides (TG), high-density lipoproteins (HDLs), and body composition parameters in the obese participants were established. Our results demonstrate that insulin resistance and oxidative stress have a significant impact on the CFU-EC colony formation in obesity. Moreover, in multivariate regression analysis, in both studied groups, the HOMAIR index and HDL concentration were independent predictors of the number of CFU-EC. Endothelium dysfunction, which can be present in obesity, may in part be caused by EPC function impairment in this condition.


Increasing obesity is a growing worldwide problem, particularly in the western population (1). Twenty-six-year follow-up of the Framingham Heart Study showed that obesity was an independent risk factor for cardiovascular disease (CVD) (2). BMI is highly correlated with most risk factors for coronary heart disease (3) and is associated with an increased risk of recurrent coronary events following acute myocardial infarction (4). Increased waist circumference was an independent predictor for all cardiovascular events except stroke in the population of obese patients with CVD (5). Obesity contributes to the development of insulin resistance and is associated with dyslipidemia. Moreover, it was demonstrated that insulin resistance was a significant prediction factor of CVD (6). The high triglycerides (TG) and low high-density lipoproteins (HDL) levels that are often accompanied by obesity are also related to insulin resistance (7). The observations of Meigs et al. (8) on obese subjects showed that insulin resistance is associated with systemic oxidative stress. It has been documented that insulin resistance may be reduced by altering the individual's diet or by treating the individual with antioxidants (9,10). Ceriello and Motz (10) described oxidative stress as a pathogenic mechanism in which insulin resistance is linked with dysfunction of pancreatic β cells and endothelium.

Until 1997, when Asahara et al. (11) performed their study, the processes of injured endothelium repair had seemed to be restricted to peripheral blood circulating mature endothelial cells. Authors described peripheral blood-derived cells forming colonies termed putative endothelial progenitor cells (EPC), which were able to incorporate injured endothelium in mouse and rabbit models of ischemic limb. Since that time, various methods, including colony-forming unit-endothelial cells (CFU-EC) assay, have been used to indicate that the number of EPC correlates with various pathological conditions such as diabetes mellitus (DM), CVD, or rheumatoid arthritis (12,13,14,15). A strong correlation between the number of CFU-EC and combined Framingham factor risk score has been previously established in healthy individuals (12). Moreover, there are only a few reports regarding this method of evaluating the CFU-EC sub-population of EPC with respect to obesity. Recently, MacEneaney et al. (16) showed that the number of EPC measured by the CFU-EC assay was decreased in the overweight and obese subjects as compared to the normal-weight controls.

The aim of this study was to determine whether obesity accompanied by insulin resistance and oxidation stress affected CFU-EC colony-forming capacity. Moreover, the putative relationship between the number of CFU-EC colonies and anthropometric, clinical, and biochemical parameters was evaluated in the studied groups.

Methods and Procedures


The study groups' characteristics are presented in Table 1.

Table 1.  Anthropometric, clinical, and biochemical characteristics of the participants
inline image

The study group consisted of 38 patients (19 men and 19 women, mean age 42.5 ± 12.7) presenting to the metabolic disorders and hypertension outpatient clinic with simple obesity and 13 healthy controls (6 men and 7 women, mean age 48.2 ± 12.9). All participants underwent routine physical examination. Anthropometric measurements of individuals wearing light clothing and no shoes were conducted. Weight was measured to the nearest 0.1 kg, and height was measured to the nearest 0.1 cm. BMI was calculated by dividing weight (kg) by height squared (m2). Obesity was defined as BMI ≥30 kg/m2. Waist circumference was measured to the nearest 0.5 cm. Dependent (%FAT) fat content was determined by impedance analysis using a Bodystat analyzer (1500 MDD; Bodystat, Isle of Man, UK). Resting seated blood pressure was measured three times and an average value was calculated according to The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) guidelines. All patients had normal renal and liver function. No patient had a history of coronary artery disease, stroke (including transient ischemic attack), congestive heart failure, or malignancy. Exclusion criteria were secondary obesity, DM, chronic diseases, and current use of any medication, including dietary supplements. Based on patient history, physical examination, and basic laboratory analysis, any patient exhibiting clinically evident acute and/or chronic inflammatory process (within respiratory, digestive and genitourinary tract, oral cavity, pharynx, and paranasal sinuses) was excluded. Patients from both the study and control group were asked to restrain from smoking for the time of the study. The percentage of smokers was similar in both groups.

According to the design of the study, the obese patients had significantly greater BMI values (P < 0.0001), fat percentage (P = 0.0006), and waist circumference (P < 0.0001) compared to the normal-weight control group. The obese participants also had a significantly higher concentration of plasma total cholesterol total cholesterol concentration (TCH) (P = 0.03) than the normal-weight controls. HDL cholesterol concentration was considerably decreased in the obese group (P = 0.012) compared to the control group. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were higher in obese individuals as compared to controls. Insulin concentration and the evaluated homeostasis model assessment-insulin resistance (HOMAIR) index was significantly elevated (P < 0.001 for both parameters) with respect to the obese participants compared to controls. The obese group was characterized by significantly decreased total antioxidant status of plasma (TAS) level indicating the presence of higher oxidative stress compared to the control group. The obese participants had a significantly higher C-reactive protein level as compared to normal-weight controls (P = 0.0001). No significant differences were found with respect to the other studied clinical and biochemical parameters (Table 1).

The study was approved by the Human Subjects Oversight Committee of the Poznan University of Medical Sciences, Poznan no. 260/06 and all participants provided informed consent.

EPC colony-forming assay

EPC colony-forming capacity was determined according to manufacturer protocol and as described by others (12,16). Briefly, mononuclear cells obtained after Ficoll centrifugation from blood samples were cultured 5 × 106 per well for 48 h on six-well fibronectin-coated dishes (BD Biosciences, San Jose, CA) in EndoCult liquid medium kit (StemCell Technologies, Vancouver, British Columbia, Canada) at 37 °C, 5% CO2 with ≥95% humidity. Experiment was performed in duplicate using two wells of a 6-well plate per sample. Then 1 × 106 nonadherent cells per well in duplicate were reseeded on 24-well fibronectin-coated dishes (BD Biosciences) for subsequent culture. After a 5-day assay, CFU-EC were counted by two independent investigators. Endothelial-like phenotype was confirmed by Ulex europeus agglutinin 1 binding and acetylated low-density lipoprotein cholesterol LDL (acLDL) uptake as described by others (12). Briefly, CFU-EC were incubated with 2.4 µg/ml 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine-labeled acLDL (Invitrogen, Carlsbad, CA) for 1 h then fixed in 2% paraformaldehyde in phosphate-buffered saline and counterstained with 10 µg/ml fluorescein isothiocyanate-labeled U. europeus agglutinin 1 (Sigma-Aldrich, St Louis, MO). Staining with DAPI Ultra Cruz mounting medium (Santa Cruz Biotechnology, Santa Cruz, CA) was performed to visualize nuclei. The fluorescent images were recorded under a fluorescence microscope Imager Z1 Zeiss with AxioVision software. To confirm immune phenotype of the CFU-EC after 5 days of culture, cells were harvested, incubated with CD14-FITC, CD45-APC, CD31-PE, and CD34-PerCP antibodies (BD Biosciences), and evaluated by flow cytometry assay in FACS Canto flow analyzer and FACS Diva software (BD Biosciences) (data not shown).

Laboratory measurements

Blood samples were taken after an overnight fast and after placing each participant in the supine position for 30 min. Plasma insulin was determined by radioimmunoassay (Biosource-Europe S.A., Nivelles, Belgium). Insulin resistance was estimated using the HOMAIR index calculated by the following equation: (fasting insulin (mU/l) × fasting glucose (mmol/l). HOMAIR index has emerged as a practical and simple method for estimating insulin resistance. This index was extensively validated in comparison with the gold-standard method for the evaluation of insulin resistance, the hyperinsulinemic euglycemic glucose clamp technique. TAS was measured using TAS Randox kit (Randox Laboratories, Crumlin, UK) and spectrophotometry (SPECORD M40; Carl Zeiss, Jena, Germany). TCH, LDL cholesterol, HDL cholesterol, TG, creatinine and glucose were measured using commercial kits. C-reactive protein was measured by a high sensitivity modified laser nephelometry technique (Berhing Diagnostics, Marburg, Germany).

Statistical analysis

Data are shown as mean ± s.d. All calculations and statistics were performed with Statistica for Windows (version 6). The differences between the groups were tested by one-way ANOVA. When the P value was <0.05, the groups were compared by the appropriate test (Student's t-test for unpaired samples). Simple associations between variables were calculated by the Pearson coefficient of correlation. Two explorative multiple-regression models, stepwise regression, and all possible subset regression were used to identify the combination of variables with the best predictive value. The following variables or transformations of these were tested both in simple-regression analyses and in the multiple-regression models: age, gender, BMI, body fat %, waist circumference, SBP, DBP, TCH, LDL cholesterol, HDL cholesterol, TG, fasting glucose, insulin, HOMAIR index, TAS, and CFU-EC. P value of <0.05 was considered significant. Logarithmic transformation was used to normalize non-normally distributed dependent variables.


Number of CFU-EC and basic statistical correlations

CFU-EC colonies appeared as clusters of round cells with radiating elongated spindle-like cells at the periphery (Figure 1a) and were able to bind U. europeus agglutinin 1 and uptake acLDL particles (Figure 1c,d). CFU-EC analysis revealed that almost 95% of cells formed CFU-EC shown surface expression of CD45 marker, about 30% of cells were positive for CD31 and 40% of cells were positive for CD14 marker. The fraction of CD34+ cells accompanied about 1% of cells derived from CFU-EC (data not shown). The number of peripheral blood-derived CFU-EC cultured in vitro from obese patients (15 ± 14) was significantly (nearly 50%) lower compared to the normal-weight control group (36 ± 28; Figure 2). No significant differences were found between the number of CFU-EC among males and females in the overall examined population. Strong inverse correlation was found between CFU-EC count and insulin concentration (r = −0.552; P < 0.001) and HOMAIR index (r = −0.558; P < 0.001) in obese participants. Similarly, converse relation of CFU-EC count was documented in the obese group with respect to the SBP (r = −0.728; P < 0.001) and TG concentration (r = −0.648; P < 0.001). By contrast, the number of CFU-EC number positively correlated with TAS (r = 0.376; P = 0.018) and HDL cholesterol level (r = 0.721; P < 0.001) in the obese group (Table 2).

Figure 1.

Colony-forming unit-endothelial cells (CFU-EC) derived from peripheral blood (PB) during 5-day colony assay. (a) Without staining (b) DAPI staining for visualization of nuclei, (c) counterstaining with FITC-labeled lectin from Ulex europeus agglutinin 1 (UEA-1), (d) counterstaining with DiI-labeled acetylated LDL (DiI-acLDL). The fluorescent images were recorded under fluorescent microscope (original magnification ×40). DAPI, 4′-6-diamidino-2-phenylindole; DiI, 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine; FITC, fluorescein isothiocyanate; LDL, low-density lipoprotein.

Figure 2.

Number of colony-forming unit-endothelial cells (CFU-EC) in lean and obese subjects. Values are mean ± s.d. **P < 0.001 Student's t-test for unpaired samples. CFU-EC, colony-forming unit-endothelial cell.

Table 2.  Correlation between CFU-EC number and clinical parameters in obese participants and overall studied population
inline image

Additional significant data correlations were established in the overall examined population. Strong inverse relation was documented between CFU-EC number and weight (r = −0.501; P < 0.001), BMI (r = −0.442; P = 0.001), %fat (r = −0.376; P = 0.006), insulin concentration (r = −0.668; P < 0.001), HOMAIR index (r = −0.672; P < 0.001) (Figure 3a), and TG (r = −0.666; P < 0.001). As was seen in the obese participants, the overall population that was studied also demonstrated positive correlation of CFU-EC number and TAS (r = −0.564; P < 0.001) (Figure 3b) and HDL cholesterol concentration (r = 0.740; P < 0.001) (Figure 3c and Table 2).

Figure 3.

Correlation between the number of colony-forming unit-endothelial cells (CFU-EC) and (a) HOMAIR index, (b) TAS, and (c) HDL in the overall studied population. Simple associations between variables were calculated using the Pearson coefficient of correlation. HDL, high-density lipoprotein; HOMAIR, homeostasis model assessment-insulin resistance; TAS, total antioxidant status.

Multivariant regression analysis

In multivariant regression analysis, in both studied groups together, HOMAIR index and HDL concentration were independent predictors of the number of CFU-EC (β = −0.270; P = 0.012 and β = 0.376; P < 0.001, respectively). The HOMAIR index and HDL concentration remained predictors of the number of CFU-EC after adjusting for age, gender, smoking status, BMI, SBP, DBP, TCH, LDL cholesterol, TG, fasting glucose, creatinine, TAS, and C-reactive protein.


A significant impact of insulin resistance and oxidative stress on the CFU-EC colony formation together with the HOMAIR index and HDL concentration as independent predictors of the number of CFU-EC in obesity are new findings demonstrated in our study. Endothelial injury and dysfunction develops as a consequence of existing CVD factors, including obesity, age, male gender, hypertension, DM, hyperlipidemia, smoking, and physical inactivity. Moreover, both insulin resistance and oxidative stress lead to endothelial injury and dysfunction in obese individuals (17,18). EPC believed to repair endothelium in adult life have revolutionized the view of postnatal vasculogenesis (11). The lack of a specific EPC marker was concluded by recent statement that EPC is a heterogeneous group comprised of distinct sub-populations as clonogenic endothelial colony-forming cells or CFU-EC investigated by means of different culture assays (19,20,21,22). Thus far, particular EPC sub-populations have been related to various pathological conditions, which may indicate a putative role for each of them in the repair of injured endothelium. Hill et al. (12) showed that the integrity of endothelium, measured by flow-mediated bronchial reactivity, was significantly associated with an elevated number of CFU-EC in healthy individuals. Cumulative cardiovascular risk formulated by Framingham risk score was also found to be significantly correlated with the number of CFU-EC in this study (12). Considering that obesity found to be an independent risk factor for CVD, the diminished number of CFU-EC and decreased colony-forming capacity that was observed in the obese participants of our study were to be expected. Endothelial cell colony-forming potential of bone marrow EPC in Zucker-fatty rats was significantly decreased, by ∼50%, compared to the bone marrow EPC derived from lean rats (23). Additionally and congruent with our results, MacEneaney et al. (16) recently showed that CFU-EC in overweight and obese individuals were significantly fewer in number compared to normal-weight individuals. On the other hand, the study presented by Hristov et al. (14) failed to demonstrate such a relationship of CFU-EC to obesity in patients with coronary artery disease. Notably, obesity in the Hristov et al. study was defined as BMI ≥25, whereas our study considered participants obese when they had BMI equal to or >30. Furthermore, the CFU-EC assay in the Hristov et al. study was performed on a limited subgroup of patients. With respect to other lipid parameters influencing the number of CFU-EC, the study conducted by Noor et al. (24) revealed that blood HDL level correlated to CFU-EC in that HDL concentration was significantly lower in individuals with low-colony counts. The vasculoprotective effect of HDL via the improvement of function of circulating EPC has previously been documented by Petoumenos et al. (25) during their study of patients after coronary angiography and in the animal model. Treatment of wild-type mice with recombinant HDL resulted in a significant increase in the number of spleen-derived CFU-EC (25). HDL cholesterol levels and TCH/HDL cholesterol ratio were also significantly related to the number of CFU-EC in patients with large range of lipidemia (26). The above reports are in line with our results in that a decrease in the number of CFU-EC correlate to a decrease in HDL cholesterol plasma levels in obese individuals. Also, a decreased number of CFU-EC in the obese patients was seen to correlate with increased levels of TG, in our study which is similar to the results obtained by Pellegatta et al. in patients with large range of lipidemia (26). Moreover, in our study HDL concentration demonstrated to be independent predictor of the number of CFU-EC in obese patients. Recently, it has been determined that both low HDL levels and increased TG levels are associated with insulin resistance (27). Our findings are also similar to those presented by MacEneaney et al. (16), who observed that HOMAIR values and insulin levels were significantly elevated in obese patients compared to overweight individuals and normal-weight controls. Mouquet et al. (28) reported a strong inverse correlation between CFU-EC number and HOMAIR index in patients with metabolic syndrome (a cluster of metabolic alterations including often insulin resistance, hypertension, and obesity). Similarly, the number of CFU-EC was decreased in patients with metabolic syndrome compared to controls (28). In the present study, we confirmed that HOMAIR index remained independent predictor of the number of CFU-EC in obese patients. Systemic oxidative stress plays a major role in many pathological conditions such as cancer, DM, and CVD (18) is associated with fat accumulation and related to BMI (29). Meigs et al. (8) observed the correlation between an increase in insulin resistance and an increase in systemic oxidative stress in a population without DM. NADPH oxidase serves as a major source of reactive oxygen species in endothelial cells and adipocytes (29,30). Increased vascular endothelial expression of the NADPH p47phox subunit in overweight and obese adults and evidence of endothelial oxidative stress have been recently reported by Silver et al. (31). Also, the pathogenesis of chronic obstructive pulmonary disease (32) and rheumatoid arthritis (33) is believed to be associated with high levels of oxidative stress. Interestingly, the number of CFU-EC was decreased in both chronic obstructive pulmonary disease (34) and rheumatoid arthritis (15) patients vs. control individuals. In our study, an increase in oxidative stress was inversely related to the number of CFU-EC in obese participants in terms of the role of diminished antioxidant capacity in development of endothelial injury in obesity. The fact that the studied groups were characterized by significantly elevated levels of SBP and DBP may be a limitation of the present study. We found that SBP and DBP values were significantly related to CFU-EC number in obese participants as compared to normal-weight controls. These results are similar to those of other studies conducted on metabolic syndrome patients or demonstrated with respect to the general population (12,28). Therefore, the patients enrolled into our study represent a group of individuals with obesity accompanied by features such as insulin resistance, serum oxidant/antioxidant imbalance, and increased blood pressure, which increase the susceptibility of the endothelium to additional injury. Moreover, in multivariant regression analysis we confirmed that HOMAIR index and HDL concentration remained predictors of the number of CFU-EC after adjusting for SBP, DBP in both studied groups.

Since they were first identified, EPC have given hope for therapeutic approaches to the treatment of vascular diseases. Recently, Rohde et al. (35) and others (21,22) determined that the CFU-EC sub-population of EPC actually exhibited an immune cell aggregation, and were not able to incorporate into injured vessel walls. The positive labeling with the lectin and uptake of acLDL particles is not unique for endothelial phenotype but it is also observed for myeloid cells. CFU-EC possesses myeloid progenitor cell activity (21). Rohde et al. (35) also found that CFU-EC colony formation is not depended on hematopoietic CD34+ cells. Similarly to results reported by others (21,35) our study showed that most cells were CD45+ (hematopoietic lineage-specific antigen) after 5 days of culture, with only a minor population of CFU-EC comprising cells exhibiting expression of the hematopoietic stem cell marker CD34. Additionally, in the context of data regarding CFU-EC as an aggregate of immune cells we confirmed that elevated C-reactive protein level in obese participants did not correlate with CFU-EC number. However, it would be worth estimating other serum inflammation markers among them tumor necrosis factor with respect to its relation to the number of CFU-EC, as this marker is also usually elevated in obese individuals. Despite inconsistent data on the exact composition of CFU-EC (21,35,36), it has been confirmed that CFU-EC express and secrete proangiogenic factors and metalloproteinases (37,38) and may promote proliferative vascular lesions in idiopathic pulmonary artery patients by activation of endothelium through matrix metalloproteinase release (38). Moreover, recent studies in vitro presented that CFU-EC that are cocultured with endothelial colony-forming cells, stimulated endothelial colony-forming cells to proliferate and repair injured endothelium (19).

In conclusion, our findings indicate that the number of CFU-EC colonies and forming capacity is decreased in obese individuals as compared to the controls, and is significantly related to an alteration of the serum oxidant/antioxidant balance and the presence of insulin resistance in the examined groups. Moreover, insulin resistance and HDL concentration were found to be an independent predictor of CFU-EC number in the studied population. Recent studies have confirmed that CFU-EC cooperate with endothelial colony-forming cells in the angiogenic response, though further studies are needed to fully elucidate the precise function of these cells in obesity and vascular repair.


This work was supported by Grant No. NN 402 253934 (2008–2010) from Polish Ministry of Science and Higher Education. The authors wish to acknowledge Dr Lianeri for her editorial assistance.


The authors declared no conflict of interest.