Pioglitazone Improves In Vitro Viability and Function of Endothelial Progenitor Cells from Individuals with Impaired Glucose Tolerance

Background Evidence suggests that the PPARγ-agonist insulin sensitizer pioglitazone, may provide potential beneficial cardiovascular (CV) effects beyond its anti-hyperglycaemic function. A reduced endothelial progenitor cell (EPC) number is associated with impaired glucose tolerance (IGT) or diabetes, conditions characterised by increased CV risk. Aim To evaluate whether pioglitazone can provide benefit in vitro in EPCs obtained from IGT subjects. Materials and Methods Early and late-outgrowth EPCs were obtained from peripheral blood mononuclear cells of 14 IGT subjects. The in vitro effect of pioglitazone (10 µM) with/without PPARγ-antagonist GW9662 (1 µM) was assessed on EPC viability, apoptosis, ability to form tubular-like structures and pro-inflammatory molecule expression. Results Pioglitazone increased early and late-outgrowth EPC viability, with negligible effects on apoptosis. The capacity of EPCs to form tubular-like structures was improved by pioglitazone in early (mean increase 28%; p = 0.005) and late-outgrowth (mean increase 30%; p = 0.037) EPCs. Pioglitazone reduced ICAM-1 and VCAM-1 adhesion molecule expression in both early (p = 0.001 and p = 0.012 respectively) and late-outgrowth (p = 0.047 and p = 0.048, respectively) EPCs. Similarly, pioglitazone reduced TNFα gene and protein expression in both early (p = 0.034;p = 0.022) and late-outgrowth (p = 0.026;p = 0.017) EPCs compared to control. These effects were prevented by incubation with the PPARγ-antagonist GW9662. Conclusion Pioglitazone exerts beneficial effects in vitro on EPCs isolated from IGT subjects, supporting the potential implication of pioglitazone as a CV protective agents.


Introduction
The peroxisome proliferator-activated receptor-gamma (PPARc) agonist pioglitazone is an insulin-sensitizing agent that is currently used in the treatment of type 2 diabetes mellitus. Pioglitazone has also been shown to exert favourable cardiovascular (CV) effects in slowing atherosclerosis progression [1] and may reduce the risk of myocardial infarction, stroke and premature death in high risk diabetic patients [2]. Impaired glucose tolerance (IGT) is a prediabetic condition characterized by insulin-resistance, predisposing to increased cardiovascular disease (CVD) risk [3]. Recent studies have demonstrated that pioglitazone reduces atherosclerotic plaque inflammation [4] and development [5,6], and improves endothelial function [7] in subjects with IGT, pointing towards potential beneficial CV properties in pre-diabetic conditions. Experimental studies support the hypothesis that putative pioglitazone CV protective effects extend beyond its metabolic action [8,9], although the precise molecular mechanisms remain to be elucidated. Robust evidence has demonstrated that CV function and angiogenesis are significantly modulated by endothelial progenitor cells (EPCs), a subset of bone marrow-derived stem cells [10] that play a critical role in the maintenance of endothelial homeostasis contributing to vessel repair following endothelial damage [11]. Reduced EPC number and function are associated with the presence of traditional CV risk factors and with the development of atherosclerosis [12,13], suggesting that endothelial injury, in the absence of sufficient circulating EPCs, promotes progression of vascular disease. In humans, the number of circulating EPCs is reduced in diabetes [14], metabolic syndrome [15] and insulin resistance [16,17]. EPCs can be isolated and differentiated ex vivo from the circulating mononuclear cell fraction and show specific endothelial markers and properties. Currently, two cell subpopulations, both capable of mediating angiogenesis, have been described and isolated: early (or circulating angiogenic cells CAC) and late-outgrowth EPCs [18,19].
Clinical studies have demonstrated that pioglitazone improves the number and migratory capacity of EPCs isolated from diabetic subjects [20] and from patients with coronary artery disease (CAD) and normal glucose tolerance [21], a phenomenon paralleled by a reduction in circulating inflammatory markers. Pioglitazone antiinflammatory effects have been demonstrated in human studies [20,21] and in vitro in activated endothelial cells [22]. However, the capacity of EPCs to form tubular-like structures and the expression of inflammatory molecules in EPCs in the presence of pioglitazone remains to be investigated. To date, no studies are available that have examined the direct effects of pioglitazone on EPCs isolated from IGT subjects. Therefore, the aim of our study was to evaluate whether the addition of pioglitazone in vitro can prove benefit in EPC obtained from IGT subjects in terms of apoptosis, viability and tube formation capacity. We also sought to investigate in vitro potential changes in EPC pro-inflammatory molecule expression in the presence of pioglitazone. This preclinical study shows that pioglitazone improves in vitro angiogenic capacity of EPCs isolated from IGT subjects and reduces inflammation.

Ethics Statement
The study protocol was in accordance with the Declaration of Helsinki and was approved by the Ethical Committee of Parma University. A written informed consent was obtained from all subjects.

Study Population
IGT subjects were consecutively recruited from those attending the CV prevention outpatient clinic of the Department of Internal Medicine, University of Parma. In subjects with impaired fasting plasma glucose an oral glucose tolerance test (OGTT) (75 g) was performed and IGT was defined according to criteria of the American Diabetes Association [23]. Inclusion criteria were IGT males or females aged 18-65 years. Exclusion criteria were stringent to exclude possible confounding factors and included the following: previous history of CVD, diabetes, medical history of neoplastic disease, acute or chronic illnesses, alcohol intake .80 g/day and chronic/intermittent medications. All participants had their medical history and anthropometric parameters recorded. Venous plasma samples (60 ml) were drawn for the determination of the main biochemical and metabolic parameters and for early EPC isolation. Five subjects (35.7%) refused to perform a second blood test for late-outgrowth EPC isolation.

Isolation and Culture of EPCs
EPC isolation was performed according to validated protocols [24,25]. Peripheral blood mononuclear cells (PBMCs) were isolated by Lymphoprep (SENTINEL, Milan, Italy) density gradient centrifugation. PBMCs were cultured into six-well tissue fibronectin-coated (10 mM) culture plates at a density of 5610 6 cells/well, and grown in endothelial cell growth medium (EGM-2) (Lonza, Milano, Italy) supplemented with hEGF (human epidermal growth factor), hydrocortisone, gentamicin and amphotericin-B, 2% FBS (fetal bovine serum), VEGF (vascular endothelial growth factor), hFGF-B (human fibroblast growth factor), R3-IGF-1 (R3-insulin-like growth factor-1), ascorbic acid, heparin in 5% CO 2 at 37uC. Culture medium was changed on day 5 and subsequently every 2 days. On day 7, adherent cells displaying an elongated spindle-shaped morphology were characterized to confirm early EPC/CAC phenotype [24]. For the isolation of late-outgrowth EPCs culture experiments were prolonged up to 3 weeks [24]. Late-outgrowth EPCs were isolated in a subgroup of 9/14subjects (64.3%). At day 21 adherent cells showed an endothelial-like morphology and were characterized to confirm the endothelial phenotype. The effect of pioglitazone was tested by culturing cells in three different conditions: 10 mM pioglitazone [21,26] (kindly provided by Takeda Chemical, Osaka, Japan), in the presence or absence of 1 mM of the PPARc-antagonist GW9662 (Sigma Aldrich, Milan, Italy), or vehicle (0.2% DMSO) as a control. Culture stimuli were added at the time of isolation (day 0) and then every time the medium was changed, up to the time cells were processed. All experiments were performed at day 7 and at day 21 for early and late-outgrowth EPCs, respectively.

Characterization of Cultured EPCs by Flow Cytometry
In order to evaluate the phenotype of cultured EPCs, 5610 5 adherent cells were detached with trypsin-EDTA and analyzed by flow cytometry (FACS Calibur, BD Biosciences, Franklin Lakes, NJ) for the expression of CD45, CD14, CD31, CD146, CD34, CD3, CD19, CD64 (BD Biosciences) and KDR (kinase insert domain receptor) (R&D Systems, Minneapolis, MN, USA) surface markers and at least 2610 4 events were acquired for each analysis.

Apoptosis Assay
Early stages of apoptosis in EPCs were investigated with Annexin V-FITC Apoptosis Detection Kit Plus (BioVision, Mountain View, CA). Early and late-outgrowth EPCs were harvested by trypsinization and 5610 5 cells were stained with Annexin-V-FITC and propidium iodide in binding buffer for 10 min at room temperature following manufacturer's instructions. Stained cells were analysed by flow cytometry (FACS Calibur, BD) using CellQuest software (Becton Dickinson) by a single blinded operator.

Viability Assay
Cell viability was evaluated by VisionBlue fluorescence cell viability assay kit (Biovision) following manufacturer's instructions. At day 0, PBMCs were seeded on fibronectin-coated 96-well tissue culture plates at a density of 5610 5 cells/well in a volume of 200 ml of EGM-2 and cultured to obtain early and late-outgrowth EPCs in the three culture conditions as previously described. After washing, adherent cells were resuspended in 100 ml EGM-2 plus 10 ml VisionBlue reagent followed by incubation for 2 hours at 37uC. The fluorescent product was measured in a fluorescent plate reader (excitation: 540 nm, emission: 586 nm) Cary Eclipse fluorescence spectrophotometer (Varian/Agilent, Santa Clara, CA, USA). To test the effect of pioglitazone on EPC viability in a oxidative stress condition, H 2 O 2 (500 mM, 24 h) was added to lateoutgrowth EPC culture medium. Data were normalized for vehicle control values.

EPC Function (Tube Formation Assay)
The capacity of early and late-outgrowth EPCs to cooperate to tubular-like structure formation when co-cultured with mature endothelial cells, was examined in Matrigel culture (BD Biosciences) [27]. Briefly, 50 ml of Matrigel (BD Biosciences) was added to prechilled 96-well plates. Matrigel was allowed to polymerize for 30 min at 37uC. Human umbilical vein endothelial cells (HUVECs) (2610 4 ), were cultured in triplicate in EGM-2 medium on Matrigel with or without the addition of EPCs (3610 3 ) obtained in the different culture conditions. After 24 hours, the number of closed circles and sprouts formed by tubular-like structures, total tube length, mean tube length and number of tubes were counted in each well. The capacity of EPCs to participate to tube formation was expressed as the ratio of tube formation assay measurements in the culture of HUVECs plus EPCs to the mean values of tube formation assay measurements in the culture of HUVECs alone. Each conditions was tested in triplicate and intra-individual coefficient of variation of total tube length was 4.4662.80% and 3.8261.56% for early and lateoutgrowth EPCs respectively.

EPC Pro-inflammatory Molecule Assessment
EPC inflammatory response to pioglitazone was investigated in early and late-outgrowth EPCs isolated from healthy donor buffy coats. PBMCs were isolated and EPCs were cultured in the three different conditions as described in the dedicated paragraph. Measurements were performed in 5 independent experiments.
Expression of adhesion molecules in EPCs. Adhesion molecule expression was assessed in 5610 5 early and lateoutgrowth EPCs detached with trypsin-EDTA and analyzed by . Values from ELISA assays were normalized to total cell number. NF-kB activation assay. NF-kB activation was assessed with a sensitive multi-well colorimetric assay for nuclear NF-kB (p50) (TRANS-AM; Active Motif, Rixensart, Belgium) [29] following manufacturer's instructions.

Statistical Analysis
Continuous normally distributed variables are expressed as mean6SD; skewed variables are reported as median (25 th -75 th interquartiles). One-way repeated measures ANOVA followed by Bonferroni post-hoc test or Student's unpaired t test were used to compare different experimental conditions. Early EPC function response to pioglitazone was expressed as [total tube length (pio)total tube length (vehicle) ]/total tube length (vehicle) . Statistical significance was accepted at p,0.05. All analyses were performed using a Windows-based SPSS statistical package (Version 19.0, Chicago, IL, USA).

Characteristics of the Study Population
The study population included 7 male and 7 female IGT subjects (mean age was 5866 years). Clinical, metabolic and biochemical variables are reported in Table 1.

Early and Late-outgrowth EPC Characterization
Early EPCs exhibited a spindle-like shape ( Figure 1A) and displayed a range of endothelial cell-surface markers, including KDR (92%) CD31 (75%), CD146 (18%) ( Figure 1B) and vWF ( Figure 1C). Consistent with previous reports [30], these cells also expressed leukocyte cell-surface markers such as CD45 (99%), and  CD14 (98%) ( Figure 1B). Macrophage marker CD68 expression resulted of 16% by immunofluorescence and that of the stem cell marker CD34 very low (2%). Early EPCs were double positive for Ac-LDL uptake and lectin binding properties ( Figure 1C). According to literature [25], late-outgrowth EPCs were able to form rare colonies of cells with a cobblestone-like morphology (Figure 2A). Lateoutgrowth EPCs expressed endothelial markers VE-cad and vWF together with CD34 ( Figure 2C), as shown by immunofluorescence. Compared to early EPCs, late EPCs showed a higher expression of the endothelial markers CD31 (98%), KDR (96%) and CD146 (92%) ( Figure 2B), whereas that of the monocyte antigen CD14, CD45 ( Figure 2B) and CD68 was negative. Late-outgrowth EPCs were also double positive for Ac-LDL uptake and lectin binding capacities ( Figure 2D). Lymphocyte contamination was excluded by the absence of CD3 and CD19 expression in early and late-outgrowth EPCs. eNOS gene expression was observed in both EPC populations ( Figure 1C and 2E).

Effect of Pioglitazone on EPC Apoptosis and Viability
The addition of pioglitazone to culture medium was not associated with changes in rate of apoptosis in either early or lateoutgrowth EPCs (data not shown).

Effect of Pioglitazone on EPC Function
The capacity of EPCs to form tubular structures is one of the most important functional properties of EPCs [27]. The addition of pioglitazone improved the capacity of EPCs to form tubular-like structures ( Figure 4C) expressed as: number of closed circles formed by early (mean increase of 28%, p = 0.005) and lateoutgrowth (mean increase of 30%; p = 0.037) EPCs ( Figure 4B); total tube length in early (mean increase of 11%; p = 0.026) and late-outgrowth (mean increase of 12%, p = 0.031) EPCs ( Figure 4A); tube number in early (mean increase of 12%; p = 0.001) and late-outgrowth (mean increase of 23%; p = 0.007) EPCs; number of sprouts in early (mean increase of 20%; p = 0.001) and late-outgrowth (mean increase of 31%; p = 0.011) EPCs compared to control. The observed effects were PPARc mediated, since the addition of GW9662 together with pioglitazone did not significantly improve EPC function.

Effects of Pioglitazone on PPARc Gene Expression
PPARc mRNA was expressed in both early and late-outgrowth EPC. The addition of pioglitazone significantly (p,0.05) increased PPARc expression in both cell populations and was prevented by the addition of GW9662 ( Figure 6A). These results were also shown to be PPARc-mediated, since the addition of GW9662 prevented these effects ( Figure 6B).

EPC Pro-inflammatory Profile
Effect of pioglitazone on inflammatory cytokine/ chemokine expression. The presence of pioglitazone significantly decreased TNFa gene expression, measured by RT-PCR in both early (p = 0.034) and late-outgrowth (p = 0.026) EPCs ( Figure 6C). TNFa levels measured by ELISA paralleleled the mRNA expression findings as in both early (p = 0.022) and in lateoutgrowth (p = 0.017) EPCs ( Figure 6D). The levels of IL-6, IL-8 and MCP-1 gene and protein expression remained unchanged in cells following treatment with pioglitazone (data not shown).
Effect of pioglitazone on NF-kB gene expression and activation. Gene expression assay showed no difference in NF-kB gene expression in early (p = 0.31) or late-outgrowth (p = 0.19) EPCs in the presence of pioglitazone. Similar findings were observed in NF-kB activation assay showing no differences between vehicle and pioglitazone-cultured early (p = 0.81) or late-outgrowth (p = 0.78) EPCs (data not shown).

Discussion
The main finding of this pre-clinical study is the demonstration of a direct in vitro favourable vascular effect of pioglitazone. The addition of pioglitazone in vitro improved viability and capacity to form tubular-like structures, in early and late-outgrowth EPCs obtained from IGT subjects. Furthermore, pioglitazone showed a potential anti-inflammatory action in reducing the levels of EPC adhesion molecules and TNFa expression. These observed effects were PPARc-specific, since they were prevented by the addition of the PPARc-antagonist.
Recent studies have demonstrated that pioglitazone treatment improves plaque stability and endothelial function [5][6][7] in IGT patients, suggesting that its beneficial effects may also occur in prediabetic conditions. Experiments conducted in cultured EPCs supported a positive action of pioglitazone on key regulators of atherosclerosis independently of its metabolic action [31,32]. Our study highlights an additional putative mechanism whereby pioglitazone therapy may reduce CV risk in IGT individuals, independent of the insulin-sensitising action.
In the present study the effects of pioglitazone were examined in two distinct EPC populations isolated ex vivo retaining a complementary function in vascular repairing mechanisms [19]. It is recognised that early EPCs show a leukocyte origin [30,33], and have also been referred to as circulating angiogenic cells (CACs) since they secrete regulators of angiogenesis [30] and appear to promote neovascularization in animal models of hindlimb ischemia [34] and myocardial infarction [35]. Lateoutgrowth EPCs retain the ability to proliferate and showed a high vascular regenerative potential [36,37]. Given the different features and origins of these two EPC populations, the evaluation of pioglitazone effects on both phenotypes serves to provide a more complete insight into EPC biology.
Pioglitazone treatment has been shown to increase the number of circulating EPCs and prevent apoptosis in wild-type mice in a phosphatidylinositol 3-kinase-dependent manner [38]. Furthermore, pioglitazone treatment increased EPC levels in diabetic patients [20] and in subjects with CAD and normal glucose tolerance [21].
The main finding of our study is that pioglitazone improved early and late-outgrowth EPC function, as measured by EPC capacity to participate to tubular-like structure formation in vitro. These results extend previous observations where pioglitazone treatment ameliorated EPC functional properties in terms of migratory response in both animal and human studies. Pioglitazone increased migratory response and adhesion capacity in type 2 diabetic subjects [20]. Similarly, EPCs cultured from mice treated with pioglitazone have shown a higher migratory capacity compared to placebo control [38]. Analogous effects in EPC migration were observed in subjects with CAD and normal glucose tolerance treated with pioglitazone for 30 days [21]. Studies have attempted to investigate possible mechanisms to explain the observed effects of pioglitazone on EPC function [39]. The synthesis of nitric oxide via endothelial cells has been shown to be a key regulator of EPC release [40]; however, this hypothesis has not been confirmed by others [38]. Other evidence has shown that pioglitazone treatment reduces basal and stimulated NADPH oxidase activity with consequent reduction in oxidative stress and EPC apoptosis [21]. In agreement with this observation and other studies [20,38], we observed that pioglitazone prevented H 2 O 2induced cell death in late-outgrowth EPCs, suggesting a protective role of this drug against oxidative stress-induced cell death.
Pioglitazone anti-inflammatory effects, in terms of a reduction in pro-inflammatory cytokine and adhesion molecule expression in mature stimulated endothelial cells [22,41,42], were also demonstrated for the first time in our study in unstimulated early and late-outgrowth EPCs. However, pioglitazone did not modify other inflammatory cytokines nor NF-kB gene expression and activity. This is in contrast with previous reports indicating that antiinflammatory effects of pioglitazone are sustained by a transrepression mechanism by which the activated PPARc interferes with the activity of pro-inflammatory transcription factors, such as NF-kB [43,44]. However, the hypothesis that the receptor can directly interact with other signaling proteins such as ERK [45], MAPK [46] and PKC1a [47], has also been described.
Future studies are required to unravel the precise mechanisms by which these anti-inflammatory effects in EPCs are exerted.
Unexpectedly, in our study EPC function response to pioglitazone was inversely correlated with some clinical parameters in IGT subjects. This finding is partially supported by the work by Werner et al. [21] showing that pioglitazone exerted beneficial effects on EPC number and migratory capacity in patients with normal glucose tolerance, supporting the idea that this agent may also be beneficial for normoglycemic individuals when presenting a higher CV risk.
The present study presents some limitations. The lack of a control group did not allow us to observe any possible alteration in the function of EPC isolated from IGT subjects, as already demonstrated in the presence of diabetes [20]. The sample size was also lower than desired, which precluded the possibility of subgroup analysis and assessment of associations between multiple parameters. Although we demonstrated a role of pioglitazone in reducing EPC pro-inflammatory molecule expression, the precise molecular mechanisms mediating pioglitazone effects on EPC biology remain unclear and still need to be addressed. The inflammatory response was evaluated in unstimulated conditions in contrast with other studies in which inflammation was assessed following diverse stimuli (i.e. LPS) in which pioglitazone showed a significant reduction of TNFa expression levels in mononuclear cells [48] and in adypocytes [49] cultured ex vivo. This approach would have been useful to confirm and strengthen our findings.
In conclusion, EPCs have clearly emerged as a new dimension in vascular biology. In patients with metabolic alterations such as compensatory hyperinsulinemia, impaired fasting glucose and IGT, EPC number and function are impaired. This preclinical study shows the in vitro effect of pioglitazone in ameliorating EPC angiogenic capacity and inflammation. Improvement in EPC function may represent a clinically relevant effect of pioglitazone that could potentially benefit patients with vascular disease before the progression to overt diabetes mellitus. In this study we have shown for the first time that pioglitazone exerts beneficial in vitro effects on EPCs isolated from IGT subjects, supporting a novel clinical therapeutic implication of pioglitazone in reducing/ preventing CV risk in pre-diabetic states, although potential clinical effects need to be confirmed in human randomized studies.